Relevant Worlds : Current Perspectives on Language, Translation and Relevance Theory [1 ed.] 9781443815086, 9781847187796

Citation preview

Relevant Worlds

Relevant Worlds: Current Perspectives on Language, Translation and Relevance Theory

Edited by

Ewa Waáaszewska, Marta Kisielewska-Krysiuk, Aniela Korzeniowska and Maágorzata Grzegorzewska

Cambridge Scholars Publishing

Relevant Worlds: Current Perspectives on Language, Translation and Relevance Theory, Edited by Ewa Waáaszewska, Marta Kisielewska-Krysiuk, Aniela Korzeniowska and Maágorzata Grzegorzewska This book first published 2008 Cambridge Scholars Publishing 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2008 by Ewa Waáaszewska, Marta Kisielewska-Krysiuk, Aniela Korzeniowska and Maágorzata Grzegorzewska and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-84718-779-X, ISBN (13): 9781847187796

TABLE OF CONTENTS

Introduction ................................................................................................. 1 Current Perspectives on Language, Translation and Relevance Theory Ewa Waáaszewska, Marta Kisielewska-Krysiuk, Aniela Korzeniowska, Maágorzata Grzegorzewska Part I: Relevance Theory and Language Chapter One............................................................................................... 22 The Online Processing of Garden-Path Utterances: Accessibility-Driven or Relevance-Driven Agnieszka Solska Chapter Two .............................................................................................. 37 Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication Agnieszka Piskorska Chapter Three ............................................................................................ 48 Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance Marta Dynel Chapter Four.............................................................................................. 67 What’s in the Punchline? Maria Jodáowiec Chapter Five .............................................................................................. 87 Conversational Humour as a Power Game Katarzyna KosiĔska Chapter Six .............................................................................................. 105 Banter – an Attempt at a Relevance-Theoretic Account Ewa Karolina Nowik

vi

Table of Contents

Chapter Seven.......................................................................................... 119 On Echoic Imperatives: A Relevance-Theoretic Approach Keiko Abe Chapter Eight........................................................................................... 131 The Procedure Meaning of the Chinese Conjunction danshi/dan Zhou Lin, Liu Hongguang and Zou Lizhi Part II: Relevance Theory and Translation Chapter Nine............................................................................................ 156 The Performance of Relevance Theory in Translation Studies Pál Heltai Chapter Ten ............................................................................................. 171 ‘Prime Eagle’ and ‘Poor Tit’ in a Quagmire of Translation – on Processing Cost Savers Wojciech Kasprzak Chapter Eleven ........................................................................................ 188 Explicitation in Simultaneous Interpreting – the Quest for Optimal Relevance Ewa Gumul Chapter Twelve ....................................................................................... 206 The Translation of EU Texts and Relevance Mara Tsoumari Chapter Thirteen...................................................................................... 221 Theory of Relevance in the Translation of Official Documents Edyta ħraáka Chapter Fourteen ..................................................................................... 238 Translating Appellatives in Polish Soap Operas into English Agnieszka Szarkowska Chapter Fifteen ........................................................................................ 252 Translation and Relevance in the Early Modern Period Kirsti Sellevold

Relevant Worlds

vii

Chapter Sixteen ....................................................................................... 265 Paradise Lost: Caught in-between. John Milton’s Poem as a ‘Translated’ Version of Heroic Epic and a Source Text for Modern Translations Maágorzata Grzegorzewska Chapter Seventeen ................................................................................... 283 How the Relevance Principle Helps to Figure out when ‘the Fifth’ Should Be Translated as ‘the Fourth’. On Shifts of Coherence in Translation Izabela SzymaĔska Chapter Eighteen ..................................................................................... 296 James Kelman’s You Have to Be Careful in the Land of the Free: Relevance Theory and National Identity in Translation Aniela Korzeniowska Bibliographical Note ............................................................................... 309 Subject Index ........................................................................................... 310 Name Index ............................................................................................. 315

INTRODUCTION CURRENT PERSPECTIVES ON LANGUAGE, TRANSLATION AND RELEVANCE THEORY EWA WAàASZEWSKA AND MARTA KISIELEWSKA-KRYSIUK (PART I), ANIELA KORZENIOWSKA AND MAàGORZATA GRZEGORZEWSKA (PART II)

This volume is a collection of essays which were first presented at the third conference Interpreting for Relevance. Discourse and Translation that took place in Kazimierz Dolny, Poland in June 2006. The contributions fall into two groups: articles on Relevance Theory and Language, and Relevance Theory and Translation.

Part I: Relevance Theory and Language Relevance Theory is a comprehensive account of human communication and cognition developed by Dan Sperber and Deirdre Wilson and first fully presented in their seminal book ([1986] 1995), which has provoked a lot of response since then, both supporting and critical as well as offering revisions, supplementation, extensions and applications to various linguistic and pragmatic phenomena, and also other domains. The theory rests on the foundation provided by a theoretically recast notion of ‘relevance’, associated with two general principles. Relevance, described as the property of inputs (such as an item of information, a perceived phenomenon, an utterance) which automatically determines the direction of cognitive and communicative processes, is defined in terms of positive cognitive effects (true contextual assumptions, strengthenings or revisions of existing assumptions) and the processing effort required to achieve these effects. The relevance of an input will increase as long as the positive effects achieved increase or the processing effort expended

2

Introduction

decreases, and vice versa. According to the first of the two general principles, the Cognitive Principle of Relevance, human cognition automatically tends to maximise relevance. Apart from the tendency to maximise relevance, humans are endowed with the ability to metarepresent other people’s thoughts and intentions, which allows us to make fairly accurate predictions about particular interpretations likely to be relevant to others and use these predictions for various purposes, both benevolent and deceptive. The second principle, the Communicative Principle of Relevance, states that each act of communication conveys the presumption that the speaker has chosen the most relevant utterance in consideration of her abilities and preferences and that the hearer automatically expects the utterance to be sufficiently relevant to be worth putting in some effort to process it. The hearer starts processing an utterance from decoding its linguistically encoded meaning and then, following a path of least effort, he enriches the recovered meaning at the explicit level and complements it with derived implicit meanings (implicatures); he stops when he reaches an interpretation that satisfies his expectations of relevance. In fact, “[t]he mutual adjustment of explicit content and implicatures, constrained by expectations of relevance is the central feature of relevance-theoretic pragmatics” (Wilson 2005, 720). Within this approach to utterance understanding, the linguistically encoded meaning serves only as a pointer to the speaker’s meaning, a piece of evidence to be evaluated on a par with any other accessible information to arrive at the intended interpretation. Semantic representations are necessarily incomplete and need to be pragmatically adjusted and supplemented on the basis of considerations of relevance.1 Relevance Theory has been applied to the study of various linguistic and pragmatic phenomena, such as specific aspects of grammar: modality (e.g. Papafragou 2000, 2006), conditionals (e.g. Noh 2000), connectives (e.g. Blakemore 2000, 2002, 2007), adverbs (e.g. Ifantidou-Trouki 1993), definiteness (e.g. Rouchota 1994), and lexical meaning (especially under the influence of the dynamically developing field of lexical pragmatics, e.g. Wilson and Carston 2007), irony (e.g. Wilson and Sperber 1992; Wilson 2006), metaphor (e.g. Goatly 1997; Wilson and Carston 2006), humour (e.g. Curcó 1998; Yus 2003, 2004), politeness (e.g. EscandellVidal 1998, 2004; Jary 1998), literary discourse (e.g. Pilkington 2000; MacKenzie 2002), advertising (e.g. Tanaka 1994), translation and interpretation (e.g. Gutt 1991; Setton 1999), language teaching and 1

For detailed exposition of Relevance Theory, see e.g. Sperber and Wilson [1986] 1995, Carston 2002a; Wilson and Sperber 2004.

Current Perspectives on Language, Translation and Relevance Theory

3

learning (e.g. Maia de Paiva and Foster-Cohen 2004; Žegarac 2004). The theory is claimed to be potentially applicable to wider domains such as the psychology of reasoning (e.g. Sperber, Cara and Girotto 1995), and even the analysis of autism (Happé 1993). It might be legitimately claimed that all “these applications are a proof of [Relevance Theory’s] importance for contemporary research on language and communication” (Yus 1998, 327).2 The papers in Part I of this volume are a voice in a discussion of phenomena which have been of central interest to relevance theorists over recent years. The first two authors, Agnieszka Solska and Agnieszka Piskorska, evaluate crucial assumptions of Relevance Theory, comparing it in one way or another with the views of François Récanati. The other authors focus on the well-known notions of metarepresentation and procedural meaning, which emerge in the process of solving such empirical problems as the comprehension of jokes (Jodáowiec), echoic imperatives (Abe) and Chinese conjunctions (Lin, Hongguang, Lizhi). They also apply, re-examine and revise the basic tenets of Relevance Theory by addressing issues of humour (Jodáowiec, KosiĔska), banter (Nowik) and wittiness as a strategy of advertising (Dynel). Agnieszka Solska’s contribution “The Online Processing of GardenPath Utterances: Accessibility-Driven or Relevance-Driven” compares Relevance Theory (Carston 2002a, 2002b, forthcoming; Wilson and Sperber 2002) with the model of utterance comprehension offered by Récanati (1995, 2002, 2004) to see how each of these approaches explains the processing failure induced by garden-path utterances, which force the hearer to backtrack and revise the wrongly decoded syntactic structure – the phenomenon that is apparently problematic for both theories. The two models differ with respect to the nature of mechanisms they propose for the derivation of explicit and implicit contents of an utterance. In Relevance Theory, the two types of content are subject to uniform pragmatic processing, which is considered to be inferential and nonreflective. Récanati, on the other hand, distinguishes primary pragmatic processes, which yield the explicit content (the Gricean ‘what is said’), from secondary ones, which are used to derive implicatures. In his account, the former are associative in nature and governed by ‘accessibility’, understood as the most prominent activation of certain elements, whereas the latter are inferential and consciously available to the hearer. Solska points out some inconsistencies in the accessibility-driven model proposed by Récanati and convincingly argues that the relevance2

For an extensive bibliography on Relevance Theory, see e.g. Yus (1998, 2006).

4

Introduction

theoretic model provides a better account of the processing of garden-path utterances. In her article “Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication”, Agnieszka Piskorska also discusses two theoretical accounts, though at a different level, namely two models of communication, the code model and the model fundamentally based on inferential processes, the former traditional and the latter derived from Grice (1989), and subscribed to, among others, by the founders of Relevance Theory (Sperber and Wilson [1986] 1995) and Récanati (1991, 1998) alike. However, Piskorska’s analysis is not meant to show the superiority of either of the models because she sensibly assumes the position postulated by Relevance Theory, according to which human communication is necessarily inferential, even though it may be facilitated by the use of a code such as a linguistic system. In other words, inferential computations take place at both the explicit and the implicit levels, which implies that the explicit side of communication does not boil down to a code. The questions she asks go beyond a theoretical evaluation of the models and concern the conception of the nature of communication entertained by language users. In particular, the author asks to what extent speakers realise, if at all, the inferential nature of the explicit side of communication. The attempt to answer this question involves again a comparison of two approaches: Reddy’s (1979) ‘conduit metaphor’ vs. Récanati’s (1991, 1998) Availability Principle. While the ‘conduit metaphor’, which is present in most expressions we use to talk about communication, implicitly endorses the code model of communication, the Availability Principle can be given a strong interpretation endowing speakers of a language with the ability to realise the use of inferential elements so as to make theoretically valid judgements about what is explicit in an utterance. Basing on her analysis of sample conversational exchanges, Piskorska argues that language users are unaware of ubiquitous inferential processes, or are unable to recognise their role retrospectively. Thus, the strong interpretation of the Availability Principle is not confirmed by the data analysis. A weaker interpretation, according to which speakers’ judgements may be wrong, is found to be fully compatible with the ‘conduit metaphor’ and as such to contribute nothing new. It is the ‘conduit metaphor’ that best reflects speakers’ intuitions. The author follows this conclusion with an attempt to explain the popularity of the code model. The explanation essentially invokes the appeal of simplified ‘folk theories’, but in addition, the author, on the basis of an observation coming from Translation Studies, reaches the conclusion that the implicit-explicit distinction, the subject of a heated debate among

Current Perspectives on Language, Translation and Relevance Theory

5

relevance theorists and others, is of little importance to the average language user, unlike the distinction between what is coded and what is not coded, regardless of whether it is explicit or not. In “Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance”, Marta Dynel focuses on a very special type of ostensive stimulus, a pictorial advertisement, which she discusses in detail on the basis of press advertising messages. She attempts to account for the wittiness of advertising strategies and persuasive techniques from the relevance-theoretic perspective, with special emphasis on the notion of processing effort and cognitive effect. A witty advertisement is defined as the one which “hinges on some form of incongruity either inherent in the picture itself or originating from the juxtaposition of the image with the slogan.” It is perceived as a special type of ostensive stimulus since, contrary to the assumptions verbalised in the presumption of optimal relevance, the most relevant advertisements are those whose comprehension requires a great amount of effort. In order to account for the phenomenon, the author relies on the tenets of Relevance Theory enriched with cognitive postulates derived from psychological studies such as the postulate of unavailability (Cialdini 1993), according to which unavailability of goods (here meaning inferred with additional cognitive investment and contextual resources) corresponds to desirability. As Dynel claims, the postulate explains why the challenging and costly interpretation of advertising messages contributes to relevance and makes an advertisement successful, i.e. persuasive. She argues that although the comprehension of a witty picture requires longer inferential processing and thus more effort on the part of the recipient, both factors (against common expectations but in agreement with Greenwald and Leavitt’s (1984) conviction that the more effort is involved in processing the stimulus, the more retrievable it is from memory) contribute to the relevance of a visual stimulus as they facilitate remembering the brand and considering a purchase of an advertised product, thus enhancing the advertising effect. All in all, Dynel shows that Relevance Theory has great potential as it can be applied to explore, among many other phenomena, the rhetoric of advertising, and she puts forward a claim that for some types of discourse, here the discourse of advertising, the theory must be expanded and some of its basic premises modified. Another area of growing interest to relevance theorists (cf. Attardo 1994; Yus 2003, 2004) is the notion of humour, which is investigated by Maria Jodáowiec and Katarzyna KosiĔska. Although they are both concerned with the interpretation of jokes, each approaches the problem from a different perspective. Jodáowiec, in “What’s in the Punchline?”,

6

Introduction

focuses on verbal jokes and the punchline effect, which is explained with reference to the notion of weak communication (Sperber and Wilson [1986] 1995, 174-176) as opposed to strong communication. KosiĔska, on the other hand, in “Conversational Humour as a Power Game”, examines the use of conversational humour understood as “any spontaneous production of a humorous utterance within the natural context of a conversation” and treats it as an efficient tool in the social power game. Having presented a data-based overview of social and psychological functions of conversational humour, she focuses on the application of humour to claim or assert power in a speech event equated with a game understood as a set of “ulterior transactions, repetitive in nature, with a well-defined psychological payoff” (Berne 1977, 130). The author’s main assumption is that each act of conversational humour has a goal related to the speaker’s social position and depending on whether power is an attribute of both parties in a conversation and whether it is used for the sake of humour or just for the sake of itself, we can differentiate among friendly, hostile and ‘the guru effect’ (Sperber 2005) applications of humour. The distinction is shown to be also rooted in the relevancetheoretic distinction into types of cognitive effects. KosiĔska argues that the type of humour is determined by the way in which it achieves optimal relevance: friendly humour by strengthening existing assumptions, hostile humour by erasing existing assumptions or creating new ones, and ‘the guru effect’ humour by combining the aforementioned strategies. As was already mentioned, Jodáowiec addresses the issue of humour from a completely different direction and, in order to successfully account for joke production and comprehension, she finds it indispensible to employ the notion of metarepresentation and the relevance-theoretic apparatus with its important distinction between strong and weak communication. As she maintains, the treatment of utterance interpretation as a mind-reading activity (Wilson and Sperber 2004) and the theoretical notion of weak communication introduced by Sperber and Wilson ([1986] 1995, 59-60) are of crucial importance to explicating joke processing and explaining what makes a punchline a punchline, an important “obligatory ending of the joke ... which pulls the joke together and makes it funny” (Linde 1981, 92). Jodáowiec describes the punchline effect in relevancetheoretic terms as a cognitive overload of weakly manifest assumptions which are ‘experienced’ at the subrepresentational level (the weaker the assumptions, the less confidence the hearer can have that they closely reflect the speaker’s thoughts, cf. Sperber and Wilson 1987, 706) and claims that “as long as there is no access to a wide range of contextual assumptions to be made manifest” by a joke, the interpreter does not find a

Current Perspectives on Language, Translation and Relevance Theory

7

joke funny. She also points to the important role of encyclopaedic information in making humour idiosyncratic. The joke setting is intended to facilitate access to contextual assumptions in the encyclopaedic entry, but the different scope, quantity and content of assumptions that are stored in the entry and that are activated by the punchline account for the differences in humorous reactions. Towards the end of the paper, Jodáowiec points to numerous advantages of the relevance-theoretic analysis of humour and rightly notices that the treatment of verbal jokes in terms of weak communication has the significant advantage of placing processing humorous texts on a par with processing everyday discourse and does not require the employment of any special strategies of comprehension. Ewa Nowik’s paper “Banter – an Attempt at a Relevance-Theoretic Account” aims at defining and providing a relevance-theoretic analysis of an elusive and intriguing phenomenon of banter, which can be understood somewhat paradoxically as a way of being impolite in order to be polite. The author first tries to place banter within a particular approach to politeness, namely the social model of politeness proposed by Watts (2003) on the basis of Bourdieu’s (1990) concept of ‘habitus’. Nowik defines banter in terms of the speaker’s impolite but obviously non-serious behaviour aimed at demonstrating solidarity towards the hearer and emphasising the relation of intimacy between them, and enumerates three necessary conditions for the proper use of banter: small social distance, the balance of power and the mutual liking between the interlocutors. Then, Nowik goes on to incorporate her definition, formulated within a social perspective, into Relevance Theory, which has sometimes been criticised for ignoring the social aspects of language (e.g. Mey 1993). Here, the author follows Escandell-Vidal (2004), who considers the psychological and social approaches to pragmatics as being not so much mutually exclusive but complementary and consequently suggests that the theory should be enriched with a social module capable of accommodating norms, or representations based on generalisations extracted from observed instances of behaviour. In this modified relevance-theoretic model, the correct interpretation of an utterance depends not only on the workings of the inference module, postulated by the founders of Relevance Theory (Sperber and Wilson 2002), but also on the operation of the social module. Nowik thinks that the social module could be identified with the habitus and argues that banter should be analysed as a script made available via the habitus. Such a treatment of banter has important implications not only for research into politeness phenomena but also for the distinction between explicit and implicit meaning as endorsed by

8

Introduction

Relevance Theory (e.g. Sperber and Wilson [1986] 1995, Wilson and Sperber 1993; Carston 2002a; Sperber and Wilson 2002). Quite a different issue is explored in the paper contributed by Keiko Abe. As the title indicates, her paper “On Echoic Imperatives: A Relevance-Theoretic Approach” is a detailed consideration of a relevance-theoretic treatment of echoic imperatives, which are nonstandard imperatives in the sense that they express the speaker’s dissociation from the communicated proposition. Since the relevance of echoic utterances lies in “conveying the speaker’s attitude to an attributed utterance or thought” (Wilson and Sperber 1992, 59), what is crucial to an adequate account of this special use of imperatives is the notion of metarepresentation, which “involves a higher-order representation with a lower-order representation embedded inside it” (Wilson 2000, 130). In Abe’s account, a lower-order representation is an echoic source and represents the content of the imperative, whereas a higher-order representation additionally represents the speaker’s dissociative propositional attitude to the echoed proposition. In her analysis of the comprehension process of echoic imperatives, Abe provides a data-based overview of echoic sources, with a special focus on the sources expressed in an utterance, implicated by an utterance and anticipated in view of the hearer’s action or attitude. The analysis leads her to conclude that irrespective of the type of echoic source, the interpretation of echoic imperatives exploits the same cognitive mechanism and thus requires a unitary explanation, presented by Abe in the final section of the paper. According to her, the comprehension of echoic imperatives involves activating the echoic source as a contextual assumption with the propositional form of an imperative and metarepresenting it with the propositional attitude of dissociation. Finally, Zhou Lin, Liu Hongguang and Zou Lizhi, in “The Procedure Meaning of the Chinese Conjunction danshi/dan”, offer an intricate investigation of ‘procedural meaning’, a notion introduced by Diane Blakemore (1987, 1992) as an alternative to ‘conceptual meaning’ and defined as a ‘constraint on relevance’, i.e. encoded information about the inferential computations in which the encoded conceptual representations are involved. Lin, Hongguang and Lizhi approach the issue through a discussion of the Chinese conjunction danshi/dan, used to indicate contradictory relations between two clauses or, untypically, between an utterance and contextual information, and attempt to give a psychological explanation of the use of discourse connectives to establish coherence within a text. Having presented a relevance-theoretic perspective on discourse connectives as devices encoding procedural constraints on the interpretation

Current Perspectives on Language, Translation and Relevance Theory

9

of utterances (cf. Blakemore 1987, 1992, 2000, 2007), they advocate an experimental approach in proving the procedural character of the meaning of dan and its positive role in language interpretation. To achieve their goal, they bring together research from cognitive pragmatics, psychology and neural science, and test the assumption that connectives minimise the hearer’s processing effort experimentally, by timing and assessing the accuracy of responses to the sentences with and without dan, and by using ERPs (Event-related Brain Potentials) technology. The ERPs data of the experiments suggest that sentences without dan have larger negative-going brain activities and cost more processing effort than the corresponding sentences with dan, thus giving supporting evidence to behavioural results. The work done by Lin, Hongguang and Lizhi deserves special attention as an experimental methodology, which they employ in their research, “is especially necessary in pragmatics, where the exclusive reliance on intuition is particularly problematic” (Burton-Roberts 2007, 6).

Part II: Relevance Theory and Translation It is generally perceived that Translation Studies today cannot avoid reflection on culture, which is a polyphonic phenomenon made up of different voices, registers, ideological concepts and historical conditions of text production. This means that engagement with politics/ideology and, what follows, ethical choices that stem from such an engagement, are part and parcel of the translator’s everyday effort. Hence, the tendency to privilege foreignising policy and give due attention to, among others, “the substandard, marginalised, unorthodox, volatile and sedimented registers, everything, in short, that makes language teeming and heterogeneous” (Hermans 2007, 91). The richness of material justifies the plethora of attitudes. Instead of searching for one Master Narrative in translation theory, most scholars today are quite happy to allow for some kind of methodological ‘anarchy’, which is not understood as an invitation to abandon the discipline of thought, but provides an incentive to address the “massive need world-wide for professional translation as cultural mediation” (Snell-Hornby 2006, 169). The translator takes upon him/herself the difficult task of bridging the differences between cultures without trading the individual traces of the source culture for his/her own comfort and that of his/her target audience. This also involves putting the very discipline of Translation Studies and the translator him/herself in the limelight of theoretical self-reflection. Apart from elucidating the meanings of the source text (which in itself, for all sorts of reasons, may turn out to be a problematic task), this self-

10

Introduction

reflection opens the path for further discussion concerning the translator’s responsibilities, commitments and decisions. The Greek theoria means ‘speculation’, that is looking at something with intent, contemplating on the essence of, seeing and understanding, that often requires time. Also novelty is not always an advantage. This is clearly indicated in A companion to Translation Studies (2007), one of the most recent surveys presenting the state and field of interest today: ... the questions that are now asked by theory no longer have to do with a posteriori ideological and cultural factors that affect, not just translation, but also the translator. Thus, rather than expect new theories of translation, we should perhaps expect a prolonged period of eclectism (cf. Bassnett 2005). Alternatively, and in alignment with current trends in literary studies, we might well be entering a period of gestation in which the discipline seeks a new understanding of itself by turning to history: be this its history as a discipline, the history of theories of translation, the role that translation has played in book and publishing history, or a social-cultural history of the translator. (Kuhiwczak and Littau 2007, 5)

In this context, it may be slightly surprising that Relevance Theory enters the field of translation with claims that apparently minimise the importance of source-culture-oriented approaches. There seems to be little concern with identity or overt commitment to ethics. Instead, it clearly joins the ranks of those linguists today whose main line of emphasis is the resurrected interest in equivalence and language universals (cf. SnellHornby 2006). The policy of domestication boomerangs with a vengeance, which raises the following questions: How does Relevance Theory apply when dialogue and translation are perceived as intercultural communication? How can we communicate cultures if, in the translation act, we only transfer what we personally think as being of optimal relevance for our target reader? These issues must be addressed with reference to the most seminal work in the field, i.e. Ernst-August Gutt’s Translation and relevance. Cognition and context ([1991] 2000). The texts presented in Part II of the current volume take up the challenge. The collection of articles on translation opens with Pál Heltai’s comprehensive overview of “The Performance of Relevance Theory in Translation Studies”. The author’s main task is to illustrate the relationship between Relevance Theory and Translation Studies. This is not, by any means, as smooth a co-existence as the very title of Gutt’s study would suggest. Notably, both ‘theory’ and ‘studies’ are erased from the headline. Behind this seemingly innocent choice lay a provocative claim, recalled in the introduction to Heltai’s analysis, that it should be

Current Perspectives on Language, Translation and Relevance Theory

11

possible to provide “a theoretical account of translation without a translation theory” (Gutt 2000). In fact, however, Gutt is not willing to dispense with a theoretical framework, for Relevance Theory remains in his account a theory in its own right and it is taken for granted that its terms will satisfy all the needs of different kinds of translation analysis. One could expect such a statement to evoke a wave of critical responses and cause a revolution in the ways we produce and interpret translations. Surprisingly, however, the earthquake did not take place. As Heltai reports, Gutt’s immodest proposal “seems to have little impact on Translation Studies,” and there has been no paradigm shift in the field. If one recalls, in this context, the proper sense of ‘paradigm shift’, which was a concept first introduced by a famous philosopher and historian of science, Thomas S. Kuhn, one will have to ask, paraphrasing Kuhn’s own definition, whether the new approach was not “sufficiently unprecedented to [draw] an enduring number of adherents” away from the current ways of thinking about translation or whether it simply failed to present itself as “sufficiently open-ended to leave all sorts of problems for the redefined groups of problems to resolve” (Kuhn 1962, 21). Heltai does not go as far as to decide for us which of the two factors have prevailed and why even the cognition-based models of translation (which indeed make best candidates for citing Gutt) rarely engage in productive dialogue with Relevance Theory. However, he says there is a brighter future for Relevance Theory in translation heralded by a number of doctoral dissertations based on Gutt’s work. Yet, it should not pass without notice that as many as three articles included in the section on literary translation discuss the problems which apparently go beyond the explanatory power of Relevance Theory (Grzegorzewska, Korzeniowska, SzymaĔska). The cases addressed are all culture-based and refer to, respectively: the political relevance of accents and regional varieties of language (Korzeniowska), cultural specificity of religious concepts (SzymaĔska) and the difference between our times and the culture of the past (Grzegorzewska). It does not mean, though, that Heltai’s optimism is misplaced and the sober tone of his analysis makes one concede that “translation is not only translation between cultures, but translation between languages and cultures.” (Should we forget that language is also part of culture?) If we agree to play down the importance of the question in parentheses, we will be more inclined to agree that Translation Studies should not abandon the linguistic/cognitive aspects of translation which may profit from Relevance Theory. Inviting contributions for the conference we asked our participants to test the applicability of Relevance Theory, depending on the subject, the

12

Introduction

genre and the aim of the given translation. This resulted in a wide range of different, sometimes even conflicting responses, which we present here, hoping that the reader will draw from them the inspiration to follow similar lines of analysis. Most of the articles are focused on specific areas of translation practice, for example the translation of popular science, legal texts, film and fiction. Following Pál Heltai’s survey, we present Wojciech Kasprzak’s “‘Prime Eagle’ and ‘Poor Tit’ in a Quagmire of Translation – on Processing Cost Savers”, which discusses the translation of naturerelated terminology within the relevance-theoretic framework, postulating that writers and translators of nature texts employ ‘relevance defocus’, the idea that when encouraging interlanguage development, one cannot apply the principle of communicative relevance (Gozdawa-GoáĊbiowski 2004, 293), rather than “pander to the extremely modest expectations of the projected readers.” Acknowledging that covert translation (e.g. tourist guides originally written in Polish but which are to function as originals in their own right and can be said to achieve functional equivalence) and overt translation (belles-lettres which cannot reach this goal of functional equivalence),3 both being acts of communication, are inevitably subject to the Principle of Relevance, he draws our attention to how easy it is, however, to fail in our search for optimal relevance, especially when the translator is addressing a wide and varied audience. Keeping Gutt and his general theory of translation in mind all the time, the author differentiates between when processing cost saving is possible, showing, for example, how the target reader can reduce the reader’s processing effort, but also warns that “clarity ought to prevail over obscurity,” which does not mean we should remove all the challenges facing the translator, a postulate that does not follow from Relevance Theory. As the author emphasises, “given a free hand in deciding what is optimally relevant to the readers, the translator/writer is at the same time relieved of the responsibility for the highest scientific standard.” It is this highest scientific standard in nature writing that Wojciech Kasprzak believes is of primary importance, whereas the search for relevance will unavoidably entail the lowering of those standards. In following Gutt’s recommendation, the translator adopts easy forms of domestication, which initially may save the audience’s processing costs, but in the long run often results in misinformation and misunderstanding, which goes against the reliable transference of information. In his concern for 3

The terms were adopted by Gutt (2000, 47-50) from Juliane House’s (1981) model for quality assessment.

Current Perspectives on Language, Translation and Relevance Theory

13

reliability, the author encourages translators to deliberately inactivate the powerful Principle of Relevance, indicating that the reduction of someone’s effort will inevitably cause the intensification of someone else’s. From the above conclusions on the requirements concerning the satisfactory translation of nature texts, we turn to Ewa Gumul’s article on “Explicitation in Simultaneous Interpreting – the Quest for Optimal Relevance”. Here the author, on the basis of a study, employing both product-oriented and process-oriented (retrospective remarks) methods, wishes to contribute to the discussion on the nature of explicit and implicit information in translation and its relation to Relevance Theory. The aim of the study, the subject of which were twenty-eight advanced interpreting students, was to determine to what extent striving for optimal relevance of the translated discourse is interpreting students’ priority motivation to explicitate implicit information. In explaining the phenomenon of explicitation and presenting research findings done in the field in reference to simultaneous interpreting, the author clearly shows the relationship between explicit and implicit information and relevance in translation, as well as the concept of redundancy in language (e.g. repetitions). Also taking into consideration the constraints imposed on simultaneous interpreting and the fact that subconscious explicitation accounts for 94.15% of all cases of explicitating shifts detected, it appears that the latter is induced not only by the intention to achieve maximal contextual effects for minimal processing effort, but is primarily necessitated by the former. Hence the findings that strategic explicitation in simultaneous interpreting appears to be attributable to both interpreting constraints and the quest to achieve optimal relevance. Ewa Gumul’s conclusions very clearly show us that in this form of translation, no matter the factors that impede the interpreter’s actions, “the choices that result from the interpreting process are filtered through the meta-constraint of relevance, i.e. are ultimately and invariably geared at achieving maximal contextual effects for minimal processing effort.” She does stress, however, that the results of her study apply solely to interpreting students, suggesting that the same experiment on professional interpreters might bring different results. The above study was to examine what actually took place in the process of simultaneous interpreting with the findings clearly pointing to the adoption of Relevance Theory. Mara Tsoumari’s article, “The Translation of EU Texts and Relevance”, on the other hand, is an attempt at approaching the translation of EU texts in the light of Relevance Theory. Interest here lies in examining the characteristics of such texts and

14

Introduction

then connecting them with Relevance Theory as the latter is applied in the translation process. Attention here is paid to the features of EU originals and their ‘translations’, which the author calls ‘language versions’, quoting after the Treaty on the European Union. Discussing the very concept of ‘equally authentic texts’ and the fact that Eurospeak today often involves wordiness and obscurity, with many so-called original EU texts already being hybrid translations, there is a very clear attempt on the part of translators to “defog the foggy EU language.” Seeing a strong affinity between covert translations and EU translations, in which both are treated as originals in their own right and are in accordance with the notion of functional equivalence, the high degree of descriptive accuracy present in many of these texts also highlight interpretive resemblance, a quality perceived primarily in what is acknowledged to be ‘translation proper’ (or, following the above terminology, overt translation). Hence, the conclusion reached by the author, whose analyses and observations point to the fact that “covert translation is distinct from translation proper only in the way it is assumed to achieve its relevance regardless of any degree of interpretive resemblance to an original.” The open question Tsoumari leaves us with is how the European Union actually would like its documents to be accepted, a crucial issue she believes is still awaiting investigation. The translation of a different type of document is analysed by Edyta ħraáka in her article on the “Theory of Relevance in the Translation of Official Documents”. Her aim here is to show how the translation of birth and marriage certificates, for example, should be done in order to reach maximum communication effect with minimum effort according to the Theory of Relevance, regardless of the differences in legal establishments. Concerned with the Communicative Principle of Relevance and the ‘presumption of optimal relevance’, she first compares the data included in Polish, British and American certificates, to go on to the successful application of the Theory of Relevance in the translation of such documents. The emphasised goal, which should be achieved in the process, is maximum communication with minimum effort, at the same time differentiating between what is explicit and what implicit in the given texts. Following the principles of informative and communicative intentions, and finding satisfactory translation solutions with the help of Relevance Theory, the author firmly believes that a comparable level of communication between the source and target texts may be achieved. The article proves to be a strong recommendation of the Theory of Relevance when there is a need to create a new, fully communicative version of an official document that would also be legally valid in the source culture.

Current Perspectives on Language, Translation and Relevance Theory

15

Moving from the complexities of different legal establishments and the translation of their various types of documents with the help of Relevance Theory, we enter into a totally different medium, that of soap operas. Agnieszka Szarkowska’s “Translating Appellatives in Polish Soap Operas into English” investigates the translation techniques employed by translators with regard to one particular type of address, namely appellatives, here understood as calls and addresses. In discussing the difference between Polish and English, she concentrates on Polish source language film dialogues from three very popular soap operas and their target language subtitles. The impressive variety of idiolects and the problems therein for the translator are brought to the foreground. Taking into consideration what is understood as audience design and the nature of dialogues in audiovisual texts when looking at the needs of audiovisual translation (AVT) as such, and what happens to appellatives in subtitling in particular, the author points to the congruence between JiĜí Levý’s Minimax Strategy, which “promises a maximum effect with a minimum of effort” (1967), taken up later by Ernst-August Gutt in his understanding of Relevance Theory in reference to translation ([1991] 2000), and the Polish scholar Teresa Tomaszkiewicz, who uses the term ‘discursive economy’ when speaking about the “optimum transfer of information using as little energy of the recipient as possible” (2006, 85; translated by A. S.). Following this, we can say, after Agnieszka Szarkowska, the less effort put into the reading of a subtitled dialogue, the more relevant and rewarding it is for the audience. The remaining contributions were concerned with the translation of literary texts, which, of course, explains their engagement with theories of interpretation, re-reading and which, out of necessity, brings us back to the problem of cultural backgrounds and contexts. Space/geography and time/history play a very important role in all of these analyses. We begin with a journey back in time presenting Kirsti Sellevold’s thought-provoking account of the relevance-theoretic approach to Renaissance translations: “Translation and Relevance in the Early Modern Period”. Her text focuses on the advantages of applying the concepts of ‘interpretive resemblance’ and ‘resemblance in relevant respects’, developed by Gutt within the general framework of Relevance Theory, in the contemporary study of the early modern practices of translation, emulation and paraphrase. The three texts which she analyses are: Jacques Amyot’s preface to his translation of the Greek philosopher Plutarch’s Moralia (1572), Gabriel Chappuy’s preface to his translation of Doni’s Il mondi celesti (1578) and John Florio’s translation of Montaigne’s Essais (1603). Sellevold argues that, apart from the strictly

16

Introduction

linguistic insight into the patterns of “interlingual interpretive use in a secondary communication situation” (Heltai, this volume), a relevancebased analysis can facilitate our understanding of some important stages in the history of translation. Sellevold’s positive response to Relevance Theory in Translation Studies can be juxtaposed with Maágorzata Grzegorzewska’s article on the Polish translation of John Milton’s seminal work: “Paradise Lost: Caught in-between. John Milton’s Poem as a ‘Translated’ Version of Heroic Epic and a Source Text for Modern Translations”. This contribution focuses mainly on the problems of ideological manipulation involved in any rendering (or appropriation) of a text produced in a different cognitive environment. The issue at stake, the author argues, is the problematic notion of ‘intended interpretation’, which plays a vital role in all relevance-based analyses. Last but not least, Izabella SzymaĔska’s “How the Relevance Principle Helps to Figure out when ‘the Fifth’ Should Be Translated as ‘the Fourth’. On Shifts of Coherence in Translation” and Aniela Korzeniowska’s “James Kelman’s You Have to Be Careful in the Land of the Free: Relevance Theory and National Identity in Translation” raise the perennial question of untranslatability, which stems from linguistic, political, religious or other cultural factors. Paradoxically, however, these two case studies show the difficulties of a completely different nature. On the one hand, SzymaĔska warns against ‘false friends’, while on the other, Korzeniowska shows that the apparent ‘enemies’ (stubborn obstacles) best stimulate the translator’s creativity. Whereas SzymaĔska takes up the apparently straightforward and unproblematic example of the Ten Commandments, which, as it turns out, can be translated only with the help of a historical commentary, Korzeniowska provides samples of successful Polish renderings of apparently untranslatable accents which constitute the most relevant aspect of James Kelman’s politically involved novel. Thus, we conclude with the article which willy nilly brings us back to the important question of ethics. The translator’s choices and decisions may be fortunate or not, felicitous or disappointing, but every aesthetic evaluation should leave space for asserting whether these choices and decisions are right or wrong, honest or dishonest. This was perhaps the most valid and endurable lesson we drew from the conference participants’ translation practice reviewed from the perspective of their encounters with Relevance Theory.

Current Perspectives on Language, Translation and Relevance Theory

17

References Attardo, Salvatore. 1994. Linguistic theories of humor. Berlin: Mouton de Gruyter. Bassnett, Susan. 2005. Literature teaching in the twenty-first century: A hopeless endeavour or the start of something new? Cambridge Quarterly 34, 203-212. Berne, Eric. 1977. Beyond games and scripts. Selections from his major writings. Edited by C. M. Steiner, and C. Kerr. New York: Grove Press. Blakemore, Diane. 1987. Semantic constraints on relevance. Oxford: Blackwell. —. 1992. Understanding utterances. Oxford: Blackwell. —. 2000. Indicators and procedures: Nevertheless and but. Journal of Linguistics 36, 463-486. —. 2002. Relevance and linguistic meaning. The semantics and pragmatics of discourse markers. Cambridge: Cambridge University Press. —. 2007. Constraints, concepts and procedural encoding. In Pragmatics, edited by N. Burton-Roberts. London: Palgrave Macmillan. Bourdieu, Pierre. 1990. The logic of practice. Cambridge: Polity Press. Burton-Roberts, Noel. 2007. Introduction. In Pragmatics, edited by N. Burton-Roberts. London: Palgrave Macmillan. Carston, Robyn. 2002a. Thoughts and utterances: The pragmatics of explicit communication. Oxford: Blackwell. —. 2002b. Linguistic meaning, communicated meaning and cognitive pragmatics. Mind and Language 17, nos. 1-2: 127-148. —. Forthcoming. How many pragmatic systems are there. In Saying, meaning, referring: Essays on the philosophy of François Récanati, edited by M.-J. Frapolli. London: Palgrave Macmillan. Cialdini, Robert. 1993. Influence, science and practice. New York: Harper Collins College Publishers, Inc. Curcó, Carmen. 1998. Indirect echoes and verbal humour. In Current issues in Relevance Theory, edited by V. Rouchota, and A. Jucker. Amsterdam: John Benjamins. Escandell-Vidal, Victoria. 1998. Politeness: A relevant issue for Relevance Theory. Revista Alicantina de Estudios Ingleses 11: 45-57. —. 2004. Norms and principles. Putting social and cognitive pragmatics together. In Current trends in the pragmatics of Spanish, edited by R. Márquez-Reiter, and M. E. Placencia. Amsterdam: John Benjamins. Goatly, Andrew. 1997. The language of metaphors. London: Routledge.

18

Introduction

Gozdawa-GoáĊbiowski, Romuald. 2004. Relevance defocus in teaching the L2 system. In Relevance studies in Poland. Vol. I, edited by E. Mioduszewska. Warsaw: Institute of English Studies, University of Warsaw. Greenwald, Anthony, and Clark Leavitt. 1984. Audience involvement in advertising: Four levels. Journal of Consumer Research 11: 581-592. Grice, H. Paul. 1989. Studies in the way of words. Harvard: Harvard University Press. Gutt, Ernst-August. [1991] 2000. Translation and relevance. Cognition and context. 2nd ed. Manchester, Boston: St Jerome Publishing. Happé, Francesca. 1993. Communicative competence and theory of mind in autism: A test of Relevance Theory. Cognition 48: 101-119. Hermans, Theo. 2007. Literary translation. In A companion to Translation Studies, edited by P. Kuhiwczak, and K. Littau. Clevedon, Buffalo, Toronto: Multilingual Matters. Ifantidou-Trouki, Elly. 1993. Sentential adverbs and relevance. Lingua 90: 69-90. Jary, Mark. 1998. Relevance Theory and the communication of politeness. Journal of Pragmatics 30: 1-19. Kuhiwczak, Piotr, and Karin Littau, eds. 2007. A companion to Translation Studies. Clevedon, Buffalo, Toronto: Multilingual Matters. Kuhn, Thomas S. 1962. The structure of scientific revolutions. Chicago: Chicago University Press. Linde, Charlotte. 1981. The organisation of discourse. In Style and variables in English, edited by T. Shopen, and J. Williams. Cambridge, Mass.: Cambridge University Press. MacKenzie, Ian. 2002. Paradigms of reading: Relevance Theory and deconstruction. London: Palgrave Macmillan. Maia de Paiva, Beatriz-Mariz, and Susan Foster-Cohen. 2004. Exploring the relationships between theories of second language acquisition and Relevance Theory. Second Language Research 20: 281-288. Mey, Jacob. 1993. Pragmatics: An introduction. Oxford: Blackwell. Noh, Eun-Ju. 2000. Metarepresentation: A Relevance-Theory approach. Amsterdam, Philadelphia: John Benjamins. Papafragou, Anna. 2000. Modality: Issues in the semantics-pragmatics interface. Amsterdam: Elsevier. —. 2006. Epistemic modality and truth conditions. Lingua 116: 16881702. Pilkington, Adrian. 2000. Poetic effects: A Relevance Theory perspective. Amsterdam: John Benjamins.

Current Perspectives on Language, Translation and Relevance Theory

19

Récanati, François. 1991. The pragmatics of what is said. In Pragmatics. A reader, edited by S. Davis. Oxford: Oxford University Press. —. 1995. The alleged priority of literal interpretation. Cognitive Science 19: 207-232. —. 1998. Truth-conditional pragmatics. In Pragmatics: Critical concepts, edited by A. Kasher. London: Routledge. —. 2002. Does linguistic communication rest on inference? Mind and Language 17, nos. 1-2: 105-126. —. 2004. Literal meaning. Cambridge: Cambridge University Press. Reddy, Michael. 1979. The conduit metaphor – a case of frame conflict in our language about language. In Metaphor and thought, edited by A. Ortony. Cambridge: Cambridge University Press. Rouchota, Villy. 1994. On indefinite descriptions. Journal of Linguistics 30: 441-475. Setton, Robin. 1999. Simultaneous interpretation: A cognitive-pragmatic analysis. Amsterdam, Philadelphia: John Benjamins. Snell-Hornby, Mary. 2006. The turns of Translation Studies, Amsterdam, Philadelphia: John Benjamins. Sperber, Dan. 2005. The guru effect. Manuscript. (8 March 2006) Sperber, Dan, Francesco Cara, and Vittorio Girotto. 1995. Relevance Theory explains the selection task. Cognition 57: 31-95. Sperber, Dan, and Deirdre Wilson. [1986] 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. —. 1987. Précis of Relevance: Communication and cognition. Behavioral and Brain Sciences 10: 697-754. —. 2002. Pragmatics, modularity and mind-reading. Mind and Language 17, nos. 1-2: 3-23. Tanaka, Keiko. 1994. Advertising language: A pragmatic approach to advertisements in Britain and Japan. London: Routledge. Tomaszkiewicz, Teresa. 2006. Przekáad audiowizualny. Warszawa: Wydawnictwo Naukowe PWN. Watts, Richard J. 2003. Politeness. Cambridge: Cambridge University Press. Wilson, Deirdre. 1998. Relevance and Relevance Theory. In: MIT encyclopedia of cognitive science (MITECS), edited by R. Wilson, and F. Keil. Cambridge, Mass.: MIT Press. —. 2000. Metarepresentation in linguistic communication. In Metarepresentations: A multidisciplinary perspective, edited by D. Sperber. Oxford, New York: Oxford University Press.

20

Introduction

—. 2006. The pragmatics of verbal irony: Echo or pretence? Lingua 116: 1722-1743. Wilson, Deirdre, and Robyn Carston. 2006. Metaphor, relevance and the ‘emergent property’ issue. Mind and Language 21: 404-433. —. 2007. A unitary approach to lexical pragmatics: Relevance, inference and ad hoc concepts. In Pragmatics, edited by N. Burton-Roberts. London: Palgrave Macmillan. Wilson, Deirdre, and Dan Sperber. 1992. On verbal irony. Lingua 87: 5376. —. 1993. Linguistic form and relevance. Lingua 90: 1-25. —. 2002. Relevance Theory. UCL Working Papers in Linguistics 14: 249287. —. 2004. Relevance Theory. In The handbook of pragmatics, edited by L. Horn, and G. Ward. Oxford: Blackwell. Yus, Francisco. 1998. A decade of Relevance Theory. Journal of Pragmatics 30: 305-345. —. 2003. Humor and the search for relevance. Journal of Pragmatics 35, no. 9: 1295-1331. —. 2004. Pragmatics of humour strategies in El club de la comedie. In Current trends in the pragmatics of Spanish, edited by R. MárquezReiter, and M. E. Placencia. Amsterdam, Philadelphia: John Benjamins. —. 2006. Relevance Theory. In Encyclopedia of language and linguistics, edited by K. Brown. 2nd ed. Amsterdam: Elsevier. Žegarac, Vladimir. 2004. Relevance Theory and the in second language acquisition. Second Language Research 20, no. 3: 193-211.

PART I: RELEVANCE THEORY AND LANGUAGE

CHAPTER ONE THE ONLINE PROCESSING OF GARDEN-PATH UTTERANCES: ACCESSIBILITY-DRIVEN

OR RELEVANCE-DRIVEN? AGNIESZKA SOLSKA

1. Introduction In the pragmatic literature on utterance comprehension the derivation of utterance meaning is seen as involving a variety of distinct cognitive tasks including the recovery of a linguistically encoded sentence meaning, reference assignment, disambiguation and implicature derivation, to name just a few. The nature of these processes is the subject of a continuing debate. The position taken by relevance theorists (Carston 2002a, 2002b, forthcoming; Wilson and Sperber 2002) is that all aspects of utterance interpretation fall within the domain of a single pragmatic system, which is fundamentally inferential and whose operations do not require conscious reflection. This view is at variance with the model of utterance comprehension proposed by Récanati (1995, 2002, 2004), who argues that the processes involved in deriving the explicit content of an utterance, or ‘what is said,’ and the processes involved in the derivation of implicatures are essentially different and belong to distinct cognitive systems. Of these two types of processes only the latter are supposed to be inferential and consciously available to the hearer. Primary processes are said to be associative in nature, their application being governed by accessibility understood as the highest degree of activation in a conceptual network.

I am grateful to Deirdre Wilson for her encouragement and her comments on the issues discussed here.

The Online Processing of Garden-Path Utterances

23

The debates on the processes at work in utterance interpretation focus mostly on what is involved in the derivation of implied meanings. In this paper, however, I would like to compare and evaluate the two approaches with respect to how they account for the processing failure caused by linguistic rather than pragmatic factors. The comprehension of the socalled garden-path sentences, such as (1)–(7), is a case in point. (1) While Anna dressed the baby spat up on the bed. (i.e. ‘The baby spat up on the bed, while Anna dressed.’) (2) The horse raced past the barn fell. (i.e. ‘The horse which was raced [by someone] past the barn fell [down].’) (3) I convinced her children are noisy. (i.e. ‘I convinced her that children are noisy.’) (4) The old man the boat. (i.e. ‘The boat is manned by the old people.’) (5) Fat people eat accumulates. (i.e. ‘[The] fat [that] people eat accumulates [in their bodies]’) (6) The man who whistles tunes pianos. (i.e. ‘The man who whistles [all the time] tunes pianos [for a living].’) (7) While John hunted the deer ran into the woods. The deer ran into the woods, while John hunted. The sentences in question mislead the comprehender, or lead her ‘down the garden-path,’ in that they first lure her into selecting an interpretation which is not satisfactory, and then force her to abandon it and search for another. For example, sentence (1) invites the reader to treat the word dressed as a transitive verb followed by the object noun phrase the baby. Such an interpretation turns out to be incorrect when she reaches the apparently subjectless verb spat, which functions like an error signal belying the anticipatory assumption she has formed. At this point the reader backtracks and tries to revise her initial erroneous hypothesis. It is by no means certain that she will manage to correctly reanalyse the baby as the subject of the main clause preceded by a subordinate clause in which the verb dressed is used intransitively. It may happen that she will reject the sentence as ungrammatical or, as demonstrated by Christianson, et al. (2001), she may even continue to hold the mistaken belief that Anna did in fact dress the baby. Apart from the rare cases when the addressee does not fall into the verbal trap and correctly interprets the sentence at first pass, the stages in the on-line recovery of the linguistic meaning of a

Chapter One

24

typical garden-path sentence may be schematically represented in the way shown in diagram (8): (8)

Stages in the on-line recovery of the linguistic meaning of a garden-path utterance: rejection

backtracking/ À anticipatory error Æ ‘forward assumption Æ signal tracking’ Â

partial repair reanalysis

À Â complete repair

In this paper, I would like to outline the nature of the interpretive difficulties hearers encounter when confronted with garden-path utterances and to add a voice to the discussion of the pragmatic system (or systems) involved in the recovery of the utterance meaning.

2. A cognitive perspective on garden-path utterances Both relevance-theorists and Récanati mention garden-path utterances as examples of situations when the smooth running of communication is disrupted. However, of the two rival views, only the relevance-theoretic model locates utterance comprehension within a broader cognitive framework, which gives the theory the explanatory power to predict not only the high degree of communicative failure in understanding gardenpath utterances but also the fact that they are relatively rare in discourse. On the relevance-theoretic view, utterance interpretation is primarily a cognitive phenomenon. In keeping with the First or Cognitive Principle of Relevance, given in (9), cognition is seen as search for relevance understood as deriving maximum cognitive effects at the lowest processing effort, as spelled out in (10): (9) Cognitive Principle of Relevance (Wilson and Sperber 2002, 254): Human cognition tends to be geared to the maximization of relevance. (10) Relevance of an input to an individual (Wilson and Sperber 2002, 252):

The Online Processing of Garden-Path Utterances

(a)

(b)

25

Other things being equal, the greater the positive cognitive effects achieved by processing an input, the greater the relevance of the input to the individual at that time. Other things being equal, the greater the processing effort expended, the lower the relevance of the input to the individual at that time.

The Cognitive Principle of Relevance together with the definition of relevance instantly predict that structures which make it necessary for the comprehender to reprocess information in order to correct previously taken assumptions must increase the cost of achieving cognitive effects. As a result, such structures are less relevant to the individual who is processing them, which in turn increases the risk of communication breakdown. The main difference between utterances used in verbal communication and other cognitive stimuli lies in the fact that the overt demand the former make on the addressee’s attention guarantees to the addressee that what is being said is going to be optimally rather than maximally relevant to her. This idea is encapsulated in the so-called Communicative Principle of Relevance, presented in (11), with the key concept of the presumption of optimal relevance given in (12): (11) Communicative Principle of Relevance (Wilson and Sperber 2002, 256): Every act of ostensive communication conveys a presumption of its own optimal relevance. (12) Presumption of optimal relevance (Sperber and Wilson [1986] 1995, 270): (a) The ostensive stimulus is relevant enough for it to be worth the addressee’s effort to process it. (b) The ostensive stimulus is the most relevant one compatible with the abilities and preferences. Like other ostensive stimuli, garden-path constructions automatically trigger the comprehension process and the language user assumes that they will be optimally relevant to her. Expecting to have a valid interpretation, she will continue to work on them despite initial problems. Yet unlike most other utterances, garden-path structures thwart those initial expectations of relevance. The information load of a garden-path sentence is no different from that of its unambiguous paraphrase and the

26

Chapter One

comprehender will not be awarded for the extra cost she has expended by any communicative gains, except maybe the sense of amusement or satisfaction at having solved a little linguistic puzzle. The Communicative Principle of Relevance is a criterion which not only determines which interpretations of an utterance hearers will accept or reject but which also regulates what sort of utterances are produced: the encoded element normally makes them easier to process and this explains why garden-path utterances are relatively rare in discourse. The situations in which they appear by design rather than accident have little to do with natural language use and include experiments conducted by psychologists or psycholinguists attempting to gain insight into natural language processing or by scientists interested in designing computer programs that can emulate the syntactic and semantic capabilities of human language users. As a result, in the discussions of the garden-path phenomenon, the terms ‘comprehender’ or ‘language user’ are used more often than the more typical ‘hearer’ or ‘addressee’.

3. The two rival views on the derivation of utterance meaning The processing of garden-path utterances is disrupted at the subpragmatic level of linguistic decoding, which falls within the domain of a grammar competence system and which, according to both of the rival models, provides input to pragmatic processes. The relevance-theoretic view is that, while recovering the meaning of the linguistically encoded elements, hearers gain access to information about the concepts encoded by individual words and to procedural information about how to manipulate these concepts (Blakemore 2002). Concerned mainly with the pragmatic phase of utterance comprehension, Récanati is much less vocal on what is involved in the recovery of the linguistically encoded meaning. He mentions the activation of concepts associated with linguistic expressions and the cognitive schemata (i.e. scripts, frames or scenarios) which are activated by, or accessed through, linguistic expressions (Récanati 1995) and which play a role in the interpretation process. The main difference between the two models lies in their vision of the pragmatic phase of utterance interpretation. In the Relevance Theory view, during this phase the hearer draws a number of inferences in order to construct hypotheses about the explicit content of the utterance (i.e. its ‘explicatures’), about the intended contextual assumptions (‘implicated premises’), and the intended contextual implications (‘implicated

The Online Processing of Garden-Path Utterances

27

conclusions’). In forming them she follows the single general comprehension procedure given in (13): (13) Relevance-theoretic comprehension procedure (Wilson 2000, 420-421): Follow a path of least effort in computing cognitive effects (a) Consider interpretations in order of accessibility. (b) Stop when your expectation of relevance is satisfied. The pragmatic tasks the hearer has to perform while deriving explicitly communicated meanings include reference resolution, disambiguation, such mechanisms as loosening, broadening and metaphorical transfer of word meaning, as well as other pragmatic enrichment processes. In order to derive implicatures the hearer has to supply contextual assumptions, which allow her to form an appropriate hypothesis about the implicated premises and finally to combine the previous two in order to derive the intended contextual implications or implicated conclusions. The schematic outline of the relevance-theoretic model of the utterance interpretation process and the specific tasks involved in the derivation of the speaker’s meaning is presented in (14). The curvy lines indicate that the processes operating at the interface between linguistically encoded meaning and explicitly communicated meaning and those operating at the interface between explicitly and implicitly communicated meaning occur in parallel and are subject to a mechanism of mutual adjustment. This parallel adjustment process is visible in the reprocessing phase of gardenpath utterances, when new hypotheses are formed, tested and revised against the disambiguating input provided by the material occurring before and after the point of breakdown. In the binary view taken by Récanati (1995, 2002, 2004), the inferential phase of comprehension, governed by Gricean type norms, is restricted to the recovery of implicatures only. The derivation of the explicit utterance content, or ‘what is said’, is carried out by primary pragmatic processes, which do not engage the reflective capabilities of the hearer. Among them we find such mechanisms as reference assignment, sense selection (i.e. disambiguation of individual words) and sense construction, understood as constructing a possible interpretation of a complex phrase. Other tasks that may have to be performed include specifization, saturation (i.e. giving semantic value to a variable or a slot in a grammatical structure of a sentence) and (free) concept enrichment. The exact number and types of these processes is a matter that awaits further discussion.

Chapter One

28

(14)

Schematic outline of the utterance interpretation process – the RT model:

SUBPRAGMATIC PHASE: DECODING Recovery of the linguistically encoded meaning x recovery of conceptual representations x recovery of procedural information

PRAGMATIC PHASE: DRAWING INFERENCES Derivation of the explicitly communicated meanings, i.e. explicature(s) x reference assignment x disambiguation x loosening/broadening/ transfer of word meaning x free enrichment

Derivation of the implicitly communicated meanings, i.e. implicature(s) x constructing appropriate contextual assumptions (providing an implicated premise) x deriving a range of implicated conclusions

Récanati seems to be somewhat ambivalent on the issue of whether the primary processes are prior, both logically and temporarily to the secondary ones – the view he expressed in Récanati (2004, 74) – or whether they are subject to ‘a mutual adjustment mechanism’, which bridges the gap between them – the position he took in Récanati (2004, 47). In the schematic view of the utterance interpretation process in Récanati’s model, given in (15), I also use curvy lines indicating the possible interaction between the various processes but, as pointed out by Carston (forthcoming), it is unclear how in Récanati’s framework these two distinct kinds of processes are supposed to interact with each other and to effect the mutual adjustment of each other’s content. Another element whose place in the model seems not entirely obvious is the activation of abstract concept schemata. Since cognitive schemata, which can be described as abstract representations of a situation type, may be evoked by either the words used or by some salient features of the speech situations, it is arguable whether they belong to the linguistic or the pragmatic domain.

The Online Processing of Garden-Path Utterances

(15)

29

Schematic outline of the utterance interpretation process – Récanati’s model:

SUBPRAGMATIC PHASE DECODING Recovery of the linguistically encoded meaning x activation of individual concepts x activation of abstract concept schemata

PRAGMATIC PHASE PRIMARY ASSOCIATIVE PROCESSES Derivation of the explicitly communicated meanings, i.e. ‘what is said’ x x x x x x

reference assignment sense selection sense construction specifization saturation enrichment

SECONDARY INFERENTIAL PROCESSES Derivation of the implicitly communicated meanings, i.e. implicature(s) x deriving implicated meanings

4. Problems for Récanati’s model: reflection and inferences The mental backtracking performed by the hearer in order to work out the explicit content of garden-path utterances may seem to pose a problem to both models since it takes place before the autonomous system of unconscious syntactically driven mental computations has provided any input for the properly pragmatic phase, be it the inferential phase of Relevance Theory or the associative level of deriving ‘what is said.’ Nevertheless, neither relevance-oriented linguists nor Récanati deny the possibility of reflective, consciously carried out interpretation in situations when the ‘normal communicative flow’ is disrupted. Carston (2002b, 145) points out that “an effortful conscious search for an interpretation” might be required when dealing with certain instances of garden-path utterances or some cases of complex figurative use. On his part, Récanati (2004, 38) speaks in very general terms of situations “when the unreflective, normal process involved in comprehension yields weird results.” He claims, however, that the “genuine inference process” which takes place in such cases involves using “evidence concerning the speaker’s beliefs and intentions.” Considering garden-path utterances, it is not instantly obvious what the speaker’s beliefs and intentions might be unless we assume that the reasoning to be conducted by the hearer includes constructing such

Chapter One

30

propositions as ‘The writer mentions the old people, not the old man’. At any rate, the idea that in fleshing out meaning at the most basic level the hearer resorts to inference makes the model internally inconsistent. Such a problem does not arise in the relevance-theoretic model, in which “the distinction between normal communicative mechanisms and more reflective utterance processing crosscuts Récanati’s distinction between primary processes and implicatures” (Carston, online statement).

5. Accounting for the garden-path effect 5.1 Processing failure at the subpragmatic level of utterance interpretation Researchers point out that a classic garden-path sentence contains in its structure the so-called local ambiguity, i.e. a place of indeterminacy “in which some locally tenable choices will ultimately lead to global ungrammaticality and resultant processing failure” (Pritchett 1992, 6). Thus in sentences (1) and (7) the fronted subordinate clause creates a place of indeterminacy in which a noun phrase can be given either an object or a subject reading. Sentence (2) is locally ambiguous between the main clause reading and the relative clause complex noun phrase reading, while sentence (3) is locally ambiguous between the verbal complement reading of a pronoun and the complement clause reading of a noun phrase. In (4) and (5) the garden-path effect arises due to the lexical ambiguity of two elements and in (6) the ambiguity concerns the possible verbal or nominal interpretation of the key word. While it is clear that in undertaking the repair the reader has to disambiguate homophonous items and revise her decisions regarding the subcategorisation frame of the verbs used, it must be noted (Solska forthcoming) that local ambiguities alone cannot be what triggers the garden-path effect. For instance, sentences (16)-(22), modelled on the examples under analysis, do not seem to entrap the hearer in the processing failure even though they have exactly the same syntactic structures and the same, or at least similar, ambiguous items as their garden-path counterparts. (16) (17) (18) (19) (20)

While Anna dressed the bomb went off in the basement. The horse raced at the Belmont died. I convinced her viruses are dangerous. The old rock the boat. Fat food contains accumulates.

The Online Processing of Garden-Path Utterances

31

(21) The man who whistles shoes horses. (22) While the man hunted a meteorite crashed into the woods. Since the processing failure takes places at the subpragmatic level of interpretation, at which the comprehender accesses or activates simple or complex concepts encoded by words or phrases, it seems plausible to assume that this is where we should look for what leads to the garden-path effect. As was mentioned in section 3, in the relevance-theoretic view, words and phrases may encode concepts which become activated in the comprehender’s mind providing access to three types of information: (I)

lexical information, i.e. information about the natural language counterpart of the concept, (II) logical information consisting of a set of deductive rules which apply to logical forms of which that concept is a constituent, (III) encyclopaedic information, i.e. information about the extension and/or denotation of the concept; in other words, information about the objects, events and/or properties that instantiate the concept. In the process of utterance comprehension we draw from the encyclopaedic and possibly from the logical entries of the concepts which serve as raw material for the logical form we are trying to construct. In doing so, as always, we apply the relevance-theoretic comprehension strategy of following a path of least effort in the search for the most accessible interpretation which yields satisfactory contextual effects. According to Récanati’s model, linguistic expressions encode representations which become active in the hearer’s mind when the expression is uttered. The semantic value of these representations can undergo semantic composition with other representations. The comprehender typically chooses the most accessible semantic value, i.e. the one which is the most prominently activated in a conceptual network. Such a potential semantic value often has to be provided by the knowledge of the world or the concept schemata which are activated by linguistic expressions and which make certain elements in the situation they evoke more prominent in the comprehender’s mind. The factors which lower the “activation threshold of a representation” (Récanati 1995, 212), thus making it more accessible, include:

32

Chapter One

(I) the recency of processing, (II) frequency of processing and (III) close associative links to accessible representations, e.g. through schemas or world knowledge.

5.2 The two accounts on what triggers or blocks the gardenpath effect In order to determine how the two models account for the garden-path effect let us take a close look at what they can tell us about the recovery of the linguistically encoded meaning of the sentences under analysis. What can we learn from the accessibility-based model about the concepts and schemata which are activated by such words as baby, raced or bomb? What kind of predictions can the relevance-theoretic model make about the inferences that can be drawn from the encyclopaedic entries of such concepts as DEER as opposed to METEORITE, BARN as opposed to BELMONT? How do the two models explain why the comprehender of sentence (4) is more likely to access the structured string of concepts THE OLD MAN than MAN THE BOAT? What kind of information is made accessible by the logical address of such words as hunted or dressed? In the relevance-based account the conceptual address for the verb dress in sentences (1) and (16) contains an inference rule yielding the information ‘put clothes on someone’, which can invite the language user to search for a possible recipient of the action. Only the encyclopaedic entry for baby, not for bomb, contains information about being a potential recipient of such an action. This may account for comprehenders making plausible yet wrong assumptions about the baby being dressed, which in turn leads to the garden-path effect in (1) but not in (16). Alternatively, in the accessibility-driven model of verbal understanding, the garden-path effect arises because only the lexical item baby would activate the schema of caring for a baby, which would include such activities as dressing the infant. No such schema would be activated by the word bomb. As for what causes the garden-path effect in sentence (2), but not in (17), the explanation following from the Relevance Theory is that, unlike the phrase at the barn, the phrase at the Belmont can give access to information connected with a horse-racing event at which horses do not merely race but are raced against each other. This is why the locally ambiguous sentence (17) is unproblematic while sentence (2) triggers the garden-path effect. Similarly, the explanation following from Récanati’s account is that the phrase at the Belmont, unlike the phrase at the barn,

The Online Processing of Garden-Path Utterances

33

activates the script connected with the type of competition in which horses are made to race against each other. In both accounts the phrase her children in sentence (3) makes a more likely candidate for being the patient of the predicator convinced than the phrase her viruses in sentence (18). Besides, women having children is part of our knowledge of the world, hence the string her children is more likely to form a complex concept (or, to use Récanati’s terms, to be activated in the process of sense construction) than the string her viruses. This may account for the garden-path effect in (3) and the absence thereof in (18). According to both accounts, the strings the old man in sentence (4) and the old rock in sentence (19) both form tenable concepts. So do the strings rock the boat and man the boat. However, the string rock the boat can also function as an idiomatic phrase and is thus more likely to form a concept than the possible collocation man the boat. Moreover, it is more frequent in discourse, hence more accessible and more likely to be interpreted as a verb phrase. These two factors preclude the garden-path effect in (19) and trigger it in (4). Due to its sense and the frequency of occurrence, the phrase fat people in sentence (5) is highly likely to form a complex concept, which makes it more likely to cause the garden-path effect than the possible yet somewhat unusual phrase fat meals in sentence (20). In the Relevance Theory account, the encyclopaedic entry for WHISTLE in sentence (6) must include information about the kind of things that can be whistled, for example tunes but definitely not shoes, which may explain why the garden-path effect arises in (6) but not in (21). Alternatively, the accessibility account would invoke a script in which the activity of whistling is associatively linked to tunes but not to shoes. Finally, in the relevance-based model, the conceptual address for the verb hunt in sentences (7) and (22) gives access to the logical entry specifying the inference: ‘When a person hunts he or she engages in the pursuit, capture or killing of a prey’. Of the two concepts: DEER and METEORITE, only the encyclopaedic entry for the former contains information about being a potential quarry in a hunt. This may lead some comprehenders to make a justified yet mistaken assumption that it is the deer which is being hunted. Consequently, the garden-path effect arises in (7), but not in (22). On the other hand, the hunting script invoked by the accessibility model would be associated with an animal being hunted, the only likely candidate being the deer.

34

Chapter One

5.3 The inevitability of the garden-path effect As can be seen, in each model the mechanisms that trigger the garden-path effect are the same as those which allow effortless comprehension of ordinary sentences. In the accessibility model differently phrased sentences associatively activate different schemata in the comprehender’s mind and make the irrelevant reading of a key word more accessible than other readings. In the relevance-based model the comprehender draws inferences based on the contextual assumptions made accessible by the logical and the encyclopaedic entries of the linguistically encoded concept. As always, in drawing these inferences she is guided by her expectations of relevance. This time, however, the inferential processes which normally help resolve interpreting difficulties make her go astray and hamper the correct understanding of the sentence at first pass.

6. Inconsistencies of Récanati’s approach Since both accounts work well in predicting the garden-path effect, it comes as a surprise that Récanati does not welcome what seems to be the logical outcome of his framework. He emphatically denies the possibility that accessibility may lead to anything but success in deriving interpretations. He objects to Sperber’s observation that there are cases when accessibility leads the comprehender astray and argues that the garden-path effect is best explained by positing the notion of an accessibility shift and the distinction between successive stages of interpretation: Sometimes the first interpretation that comes to mind (the most accessible one) turns out not to be satisfactory and forces the hearer to backtrack. According to Sperber, the possibility of such garden-path effects shows that success, for a candidate semantic value, cannot be equated with sheer acceptability. The objection is misguided, I think. The most acceptable interpretation at some stage s in the interpretation process may well turn out to be unsatisfactory at some later stage s’, thereby resulting in a garden-path effect and the need to backtrack. This does not show that interpretational success cannot be cashed out in term of accessibility. At any given stage, the most accessible interpretation will be the winning one (at that stage). In garden-path utterances we have two successive stages to consider. Some interpretation is the most accessible one, hence winds, at s, but that interpretation fails to fit some schema, hence loses, at a later stage s’. (Récanati 1995, 227; 2004, 32)

The Online Processing of Garden-Path Utterances

35

Although there can be no doubt that understanding garden-path sentences is a two-step process, the fact remains that if the addressee has to reject her initial analysis and search for another it is precisely because choosing the most accessible candidate for semantic value was what caused the processing failure at step one. Another inconsistency that can be found in Récanati’s approach concerns the nature of the schemata driven interpretation process. Récanati (1995, 226) admits that the “role played by schemata explains why the process of utterance interpretation is to such a large extent top down and driven by world knowledge.” However, it seems doubtful that the salient information provided by a concept schema is always merely associatively activated in the hearer’s mind. Especially sentence (17) seems to require the comprehender to draw the inference (or, to use relevance-theoretic terminology, to form a contextual assumption) that the prepositional phrase at the Belmont has something to do with horse-racing. For many comprehenders, including the author of the present article, this is something they will in fact learn or have learnt from the sentence, not something they would know beforehand. Obviously, allowing inferences into the picture would belie the key notion of the accessibility model in which the primary processes of deriving ‘what is said’ are free from inferences.

7. Concluding remarks The inconsistencies in the accessibility-driven model which this paper has revealed do not necessarily invalidate the whole framework but they indicate that it might be necessary to revise some of its key elements. The examination of the online processes of interpreting garden-path sentences shows that there is some reason to doubt whether the conscious availability or unavailability of various cognitive processes runs along the lines proposed by Récanati. The analysis also seems to undermine the claim that the derivation of the explicit content of the utterance is never inferential. Compared with the accessibility-based account, the unified relevancetheoretic model is internally consistent. Moreover, grounded in a broader cognitive framework it is better equipped to account for the interpretive difficulties, be they pragmatic or subpragmatic in nature.

36

Chapter One

References Blakemore, Diane. 2002. Meaning and relevance: The semantics and pragmatics of discourse connectives. Cambridge: Cambridge University Press. Carston, Robyn. 2002a. Thoughts and utterances: The pragmatics of explicit communication. Oxford: Blackwell. —. 2002b. Linguistic meaning, communicated meaning and cognitive pragmatics. Mind and Language 17, nos. 102: 127-148. —. (forthcoming). How many pragmatic systems are there. In Saying, meaning, referring: Essays on the philosophy of François Récanati, edited by M.-J. Frapolli. London: Palgrave Macmillan. —. 2003. Conversational Implicatures and pragmatic mechanisms. European Society for Philosophy and Psychology. (30 May 2006) Christianson, Kiel, Andrew Hollingworth, John F. Halliwell, and Fernanda Ferreira. 2001. Thematic roles assigned along the garden-path linger. Cognitive Psychology 42: 368-407. Pritchett, Bradley L. 1992. Grammatical competence and parsing performance. Chicago: University of Chicago Press. Récanati, François. 1995. The alleged priority of literal interpretation. Cognitive Science 19: 207-232. —. 2002. Does linguistic communication rest on inference? Mind and Language 17: 105-126. —. 2004. Literal meaning. Cambridge: Cambridge University Press. Sperber, Dan, and Deirdre Wilson. 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. Solska, Agnieszka. (forthcoming). A relevance-theoretic perspective on garden-path sentences. In Proceedings of the Fifteenth Annual PASE Conference. Wilson, Deirdre. 2000. Metarepresentation in linguistic communication. In Metarepresentations: A multidisciplinary perspective, edited by D. Sperber. Oxford: Oxford University Press. Wilson, Deirdre, and Dan Sperber. 2002. Relevance Theory. UCL Working Papers in Linguistics 14: 249-287.

CHAPTER TWO WHY THE AVERAGE SPEAKER DOES NOT APPRECIATE THE ROLE OF INFERENTIAL PROCESSES IN COMMUNICATION AGNIESZKA PISKORSKA

1. Introduction Sperber and Wilson ([1986] 1995, ch. 1) write about the popularity of the code model of communication, notably among semioticians, who have been trying to find underlying codes in various domains of human activities. To Sperber and Wilson, however, semiotics with its dominance of the code model appears to be an intellectual failure, as it fails to recognise the fact that although codes do play a role in some forms of communication, they are only auxiliary to inferential processes, which constitute the core of human communication. In the relevance-theoretic view the existence of a code is not even a necessary condition for successful communication. What we really need a code for is to make communication more efficient and precise. In the case of verbal communication, for instance, the speaker produces a message in a natural language – a code. On receiving the message, the hearer first decodes it by means of his specialised mental module, and then the output of the decoding process is input to the inferential system.1 It is the inferential system that takes the final responsibility for arriving at the speaker-intended message and the coded element of the message limits the potential range of interpretations. As Sperber and Wilson demonstrate, the contribution of inferential processes to utterance interpretation is enormous. It can be observed at 1

This is not to say that the specialised language module operates prior to the inferential system on the whole utterance; the two systems must switch over constantly during the processing of a single utterance.

38

Chapter Two

every stage of processing, i.e. explicature formation (disambiguation, reference assignment, lexical pragmatic processes2) and implicature formation. The aim of this paper is to examine the status of both the code model and inferential model in the general understanding of language users. In other words, the question is whether the relevance-theoretic division of labour between coded and inferential communication is reflected in the popular conception of the nature of communication. Do we, as language users, realise the indeterminacy of the coded element of communication and appreciate the role of inferential processes? Is the inferential model of communication considered to be a valid description of reality? Linguists working both in the relevance-theoretic framework and outside it have made some observations on the subject. These observations and predictions they give rise to will be discussed in the following section. Then, language data intended to verify the predictions will be demonstrated. In the last section an attempt will be made to explain the picture emerging from the analysis presented.

2. Selected views on the perception of coded and inferential communication Some predictions related to the issue examined follow from the work of Michael Reddy (1979), discussed extensively in Lakoff and Johnson’s Metaphors we live by (1980). The main claim expressed in this book is that a great number, perhaps even the majority of concepts are metaphoric in nature. Abstract concepts, such as TIME, LOVE, ANGER are structured in terms of simpler concepts denoting physical objects. The conceptual metaphors, e.g. TIME IS A VALUABLE COMMODITY are reflected in the way we speak about time, therefore one can have time, waste time, spend time, etc. The contribution of Reddy (1979) to this model is the so called ‘conduit metaphor’, according to which: Ideas/meanings are objects Linguistic expressions are containers Communication is sending. (Lakoff and Johnson 1980, 9)

The speaker’s ideas are objects, words are containers into which she3 puts her ideas and sends them along a conduit to the hearer who takes the 2

Of more recent developments cf. e.g. Wilson (2003) on lexical pragmatic processes.

Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication

39

ideas out of words. There are numerous examples of linguistic expressions in English which exemplify the above mentioned conceptual metaphors, such as the following: It’s hard to get that idea across to him. I gave you that idea. It’s difficult to put my ideas into words. When you have a good idea, try to capture it immediately in words. Try to pack more thought into fewer words. The meaning is right there in the words. The sentence is without meaning.

Reddy (1979) estimates that the above mentioned conceptual metaphors account for at least 70 percent of the expressions we use for talking about language. These expressions are part of conventional language about language, and as Lakoff and Johnson (1980) observe, it is difficult for an average language user to see that they involve any kind of metaphor at all. The fact that the meaning is not ‘right there in the words’, but results from the interaction of the linguistic knowledge and contextdependent processes seems to escape our attention. Thus, the influence of the conduit metaphor on the way we understand the world distorts our perception of communication so that we equate it with its coded component. Let me now consider the potential implications of the Availability Principle put forward by François Récanati (1991, 1998). According to the principle, speakers are aware and capable of spelling out what they are communicating by an utterance. Moreover, they are capable of distinguishing which elements of an utterance are communicated explicitly (‘what is said’ in Récanati’s terminology, following the original Gricean distinction between what is said and what is implicated; cf. Grice 1989), and which are communicated implicitly. On the other hand, the distinction between the output of linguistic decoding and pragmatically supplied elements of the explicature is not available to the conscious inspection of the “unsophisticated speaker-hearer,” as Récanati (1991, 106) puts it. Although the Availability Principle was formulated to support the view that the explicature of an utterance goes beyond the minimal truthevaluable proposition – a claim not directly related to the topic of this paper, it may have some consequences for the present considerations (and vice versa). To see what these are it is necessary to narrow down the 3

Following Sperber and Wilson ([1986] 1995), for clarity, the speaker is pronominalised as she and the hearer as he.

Chapter Two

40

interpretation of the Availability Principle to one of the following possibilities, one of which should be true on the assumption that the speaker does not have conscious access to the linguistically encoded part of the utterance, and does have conscious access to the explicature: 1.

2.

The speaker is aware of the fact that the explicature is made up from the coded element and inferential element and that additionally there are implicatures; what the speaker is unaware of is which elements of the explicature are decoded and which are inferred; The speaker is unaware of the fact that the explicature is made up from the coded element and the inferred element, she takes the whole of the explicature as the basic, unanalysable component of the utterance meaning.

Neither of the above is directly endorsed by Récanati. Point 1 above offers a stronger interpretation of the Availability Principle, and it seems to be consistent with the spirit of Récanati’s paper, as it grants the speaker a greater degree of cognitive abilities, which is what Récanati argues for; after all, by postulating the Availability Principle the speaker’s intuitions on the content of the explicature are elevated to the status of theoretically valid judgements. On the other hand, Récanati (1991, 106) writes: To be sure, users of the language claim to have intuitions concerning what the sentences in their language mean; but these intuitions are not directly about their purported objects – linguistic meanings. They do not bear on the linguistic meanings of sentences, which are very abstract and inaccessible to consciousness, but on what would be said or communicated by the sentence were it uttered in a standard or easily accessible context.

This remark suggests that the weaker version of the Availability Principle was intended, on which speakers do not realise that what they communicate explicitly is not equivalent to what is linguistically encoded, or, speakers actually believe that what is communicated explicitly is the same as what is linguistically encoded. The strong interpretation may seem too far-fetched and impossible to verify, so perhaps the weaker one should be accepted as more realistic and reasonable. Unfortunately, only the strong one makes a non-trivial claim and would be therefore interesting, if it were correct. The weaker version is reasonable precisely because it is trivial, as it says that speakers’ intuitions can be wrong. In fact, since the weaker version accommodates the claim that

Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication

41

speakers take the explicature to be the linguistically coded element of communication, it is tantamount to the conduit metaphor. Thus, the picture resulting from the discussion so far is the following: if we take the strong interpretation of the Availability Principle, there are two opposing views. The ‘conduit metaphor view’ says that it is the code model of communication that prevails in popular beliefs, and the Availability Principle says that speakers are aware of the contribution of inferential processes to communication both at the explicit and implicit level, though they may not be aware of which specific elements of the explicature are provided inferentially, and which are decoded. The fact that these are different views does not mean that one of them is right and the other wrong; it is well known that humans are capable of holding different views, which are not necessarily fully consistent with one another. On the weaker reading of the Availability Principle, its predictions are fully compatible with those of the conduit metaphor.

3. Data analysis In order to find out how the two modes of communication are perceived, I would like to present some examples of verbal exchanges I have myself witnessed, or participated in. What they have in common is that one of the participants rejects what the other has said on the grounds that the inferential part of the explicature is not obvious and should be reformulated in a more precise, perhaps linguistically coded manner. Sometimes it is a source of moderate humour, sometimes a means of verbal aggression. The examples, either given in two languages, or translated from Polish into English, are as follows: (1)

A: Znasz jakiĞ dobry syrop na kaszel? literally: ‘Do you know a good syrup for a cough?’ ‘Do you know of any good cough syrup?’ B: To znaczy chcesz mieü kaszel, czy chcesz syrop przeciw kaszlowi? literally: ‘Does that mean you want a cough, or a syrup against a cough?’ ‘Does that mean that you want a cough or a cough syrup?’

The expression syrop na kaszel is ambiguous in Polish (either curing a cough or inducing it), yet it can be legitimately used by analogy to the fully sanctioned term ‘a medicine for something’. Here, the disambiguation leading to the speaker intended interpretation is rejected in

Chapter Two

42

favour of the other, nonsensical interpretation that A wants to have a cough. Although it is easy to arrive inferentially at the right explicature, B wants A to eliminate the ambiguity at the level of linguistic meaning, as if suggesting that it is better to guarantee the right interpretation in the coding phase rather than rely on the inferential phase. (2)

A appears not to be able to grasp B’s intended meaning B: I’m speaking Polish to you, I’m not speaking Chinese!

This is a metalinguistic remark, stating that if there is anything wrong with communication, it must be associated with one’s lack of command of the language, and not the inability to make the right inference. (3) A: Mum, what’s the English word for czajnik (ambiguous between ‘kettle’ and ‘teapot’)? B: Do you mean the one in which we boil water, or make tea? A: I mean the word czajnik only. This example comes from the area of second language acquisition. Children (and often adults) display a naïve belief that every lexical item (code unit) has an exact corresponding counterpart in a foreign language, disregarding the lexical pragmatic processes. (4)

A: I have a temperature. B: Everything has some temperature.

Here B rejects the obvious narrowing of the concept TEMPERATURE which gives the non-trivial meaning to the explicature of the utterance. B does so to block the implicature that A has an excuse for not doing what B wants her to, and B does it by questioning the inferential part of the explicature. (5)

A: When are you coming home? B: I’ve got to stay in the office for some time. A: I know that you have to stay there for some time!

Here, by emphatic repetition of the word some, A rejects the standard explicature of B’s utterance, on which B has to stay in the office for a considerable amount of time, and takes up the minimal linguistically encoded meaning of some.

Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication

(6)

43

A: Don’t complain. Everybody’s overworked nowadays. B: Except those unemployed.

B’s response relies on questioning the obvious inferential narrowing down of ‘everybody’ to some relevant referent, and, again, on exploiting the contextually unconstrained, linguistic sense of ‘everybody’. Note that examples (4), (5), and (6) are similar to the ones employed by Carston (1988) and Récanati (1991, 1998) in their respective attempts at defining the explicature of an utterance. If the exchanges like these above are continued, the person whose explicature has been rejected usually accepts the criticism, makes an excuse (unless he or she is a relevance-theorist), and goes on trying to be more precise, i.e. relying on the coded element of the message. The above examples are different from the ‘conduit metaphor’ expressions, as they do not directly express the claim that the ‘meaning is right there in the words’, except perhaps the metalinguistic (2) and (3). They involve a tacit assumption that the right way to say something is offered by the language itself and the inferential layer of interpretation is but an addition to it, sometimes facilitating processing, and sometimes leading to communication failure. Thus these examples provide further evidence supporting the ‘conduit metaphor’, i.e. the widespread belief that the key to communication is coding and decoding. The inferential layer of communication, which has been proved to be more essential than the coded layer, is not generally recognised as such. As far as the Availability Principle is concerned, the data from this section back up its weaker version, which is not surprising, since the predictions of the weaker reading of the principle were said to be equivalent to the conduit metaphor. If so, then, the Availability Principle turns out not to contribute any new, or important information on the validity of speakers’ intuitions.

4. Searching for an explanation Inferential processes are indispensable for communication and operate in every kind of discourse, both formal, and informal. Even in such registers in which ambiguity and imprecision are apparently to be avoided at all costs (e.g. legal discourse) pragmatic processes do apply and the message is context-dependent. Yet, speakers seem to consider that inferential communication does not come up to the standards of communication

Chapter Two

44

proper. What makes the code model of communication so attractive to the minds of human beings? There exist more scientific theories whose findings have not been fully internalised in the minds of average human beings. Ask an average Lotto player if the likelihood of drawing six consecutive numbers (e.g. 1, 2, 3, 4, 5, 6) is equal to the likelihood of drawing any other combination of six numbers. She will surely tell you that drawing consecutive numbers is less probable. Take the relativity theory, or quantum physics: although they are scientifically sound, they cannot be said to enjoy the status of ‘folk theories’, i.e. widespread beliefs about what the reality is like.4 However, there is a crucial difference between the relativity theory and a communication theory in that the former is based on a complicated mathematical apparatus and concerns phenomena which are not part of basic human experience, whereas the latter does not involve any mathematical formulae and pertains to a fundamental human activity. Still, in both cases the reason why Newton physics and the coded model of communication are generally considered to offer a true description of reality may be the same, namely, the need to understand the world and ourselves in terms of simple models. Besides, a possible explanation of the status of the code model can be sought in Translation Studies. The discipline has been using the terms ‘explicitation’ and ‘implicitation’ for almost half a century now (cf. Vinay and Darbelnet 1958). As one may guess, the term ‘explicitation’ does not correspond to the relevance-theoretic definition of ‘explicature’, it denotes the shifts the translator makes between what is inferable from the original into what is encoded in the translation, i.e. using words in the target text to encode the inferentially supplied content of the original. The following is an example of explicitation: (7)

Oil is running out. ‘ZáoĪa ropy naftowej na Ğwiecie wyczerpują siĊ’ Back translation: ‘Oil is running out in the world’

Here, the phrase in the world is added in the translation. The information that oil is running out in the world, as opposed to, say, a petrol station, is inferable from the original as a part of its explicature defined in relevance-theoretic terms. 4

‘Folk theories’ may be seen as analogous to ‘folk categories’ in the sense of Taylor (1989).

Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication

45

In an informal study I have myself conducted, I have found that the great majority of explicitations involve putting the inferable part of the original’s explicature into words of the target text, rather than reformulating truly implicit content in terms of explicature. It therefore seems that the distinction between the coded content and the inferable content of an utterance is of greater practical importance for the translator than the distinction between the explicature and implicatures of an utterance. Although the explicit/implicit distinction is theoretically sound, it does not directly bear on the translator’s decision-making process.5 The assumption that translation is a form of communication lies at the core of Translation Studies (e.g. Hatim and Mason 1990, Korzeniowska and Kuhiwczak 1994, Gutt 2000). Thus, it can be presumed that what is observable in translation may also carry over to communication in general and that the coded/inferred distinction is of greater utility for the average language user, as it is for the translator. This is, I believe, one of the reasons why speakers tend to think that saying something equals coding, as opposed to suggesting something, which involves inference.

5. Concluding remarks It may be thought that due to such ubiquity of inferential processes we, language users, should be aware of them, perhaps not in the sense that we apply them consciously in communication but that we are able to recognise their role retrospectively. The predictions offered by the conduit metaphor and the analysis of data offered in this paper indicate otherwise. One of the reasons why the code model enjoys the central status in the folk theory of communication is that it is simpler than the inferential model. Simplified models play a role in cognition as it is better to have at least a simplified account of some phenomenon than none at all. Another explanation of the popularity of the code model may be associated with the fact that the coded/inferred distinction plays a more prominent role in communication than the explicit/implicit distinction. What has been shown to be true about translation may also be true about monolingual communication. For example, when asked to clarify one’s message, what one does is shifting the inferable part of the message to the 5 Some translation scholars overtly acknowledge that they use the term ‘explicitness’ differently from the relevance-theoretic definition, e.g. Steiner (2005, 10) says: “With respect to Relevance Theory, then, the characteristic of our approach is that we measure explicitness as a property of the encoding, not as a property of the communicative act as such.”

46

Chapter Two

coded part rather than shifting the implicit to the still inferable, but explicit part of the message. If the claim that the coded/inferred distinction is of greater practical utility for language users and is therefore better internalised than the explicit/implicit distinction is true, then appeals to speaker’s intuitions in defining the explicature of the utterance are not justified.

References Carston, Robyn. 1988. Implicature, explicature, and truth-theoretic semantics. In Mental representations: The interface between language and reality, edited by R. M. Kempson. Cambridge: Cambridge University Press. Grice, H. Paul. 1989. Logic and conversation. In Studies in the way of words. Harvard: Harvard University Press. Gutt, Ernst-August. 2000. Translation and relevance. 2nd ed. Manchester: St Jerome Publishing. Hatim, Basil, and Ian Mason. 1990. Discourse and the translator. London, New York: Longman. Korzeniowska, Aniela, and Piotr Kuhiwczak. 1994. Successful PolishEnglish translation. Tricks of the trade. Warszawa: PWN. Lakoff, George, and Mark Johnson. 1980. Metaphors we live by. Chicago: The University of Chicago Press. Récanati, François. 1991. The pragmatics of what is said. In Pragmatics. A reader, edited by S. Davis. Oxford: Oxford University Press. —. 1998. Truth-conditional pragmatics. In Pragmatics: Critical concepts, edited by A. Kasher. London: Routlege. Reddy, Michael. 1979. The conduit metaphor – a case of frame conflict in our language about language. In Metaphor and thought, edited by A. Ortony. Cambridge: Cambridge University Press. Sperber, Dan, and Deirdre Wilson. 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. Steiner, Erich. 2005. Explicitation, its lexicogrammatical realization, and its determining (independent) variables – towards an empirical and corpus-based methodology. Reports of the project Languages in Contrast (Språk i kontrast) No. 36, December 2005. (10 September 2006) Taylor, John R. 1989. Linguistic categorization: Prototypes in linguistic theory. Oxford: Oxford University Press.

Why the Average Speaker Does Not Appreciate the Role of Inferential Processes in Communication

47

Vinay, Jean-Paul, and Jean Darbelnet. [1958] 1995. Comparative sylistics of French and English: A methodology for translation. Translated by J. C. Sager, and M.-J. Hamel. Amsterdam, Philadelphia: John Benjamins. Wilson, Deirdre. 2003. Relevance Theory and lexical pragmatics. Italian Journal of Linguistics/Rivista di Linguistica, 15, no. 2: 273-291. Reprinted in 2004. UCL Working Papers in Linguistics 16.

CHAPTER THREE WITTINESS IN THE VISUAL RHETORIC OF ADVERTISING AND THE QUEST FOR RELEVANCE MARTA DYNEL

1. Introduction From the viewpoint of marketing, the primary aim of advertising is to persuade, i.e. to provoke the target to take action, which prototypically means purchasing a given product or using a service.1 However, advertising has earned the status of an art form, which is subject to multifarious analyses from various scholarly perspectives. Not only are advertisements evaluated from the vantage point of pure marketing, but they are also discussed as aesthetic, rhetorical, psychological and linguistic phenomena. Notably, positive assessment within one of these dimensions need not be correlated with a marketing success, regarded as the most fundamental motivation for advertising.2 In this paper, we present advertisements in the perspective of rhetoric, simultaneously elaborating on the cognitive mechanisms which account for the persuasive techniques employed. Irrespective of discrepancies between heuristic approaches, both researchers and practitioners concede that a number of features of a wellformed advertisement can be distinguished. Due to the flood of products and advertisements, the aim of copywriters is to make a given product and, notably, its advertising message conspicuous (Sutherland and Sylvester 1993; Albin 2000). Apart from this primary feature, Jones (1997) 1

There are also social advertisements focused primarily on sensitising the target to a given problem, e.g. drink driving or child abuse. 2 Also, even an advertisement enjoying high likeability may fail to boost sales despite the enthusiastic approval received from the general public.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

49

identifies three characteristics of a successful advertisement. Firstly, it ought to communicate relevant information about the brand advertised. What is more important from our perspective, it ought to be likeable, amusing and entertaining, and, in consequence, offer the recipient a reward. It is often maintained that claims about the product are presented rationally while rewards are offered at the emotional level via music or imagery. We would contribute, however, that rewards may also stem from logical inferencing and thus that the cognitive benefit of having arrived at understanding with the effort invested, which squares with one tenet of Sperber and Wilson’s ([1986] 1995, 2004) framework. It is noteworthy that the overwhelming majority of advertising campaigns in colour magazines resort to pictorial representation, which is the focus of our analysis. We agree with Kochan (2003), who postulates that the analysis of advertising discourse entails predominantly the analysis of the image, optionally intertwined with the analysis of the verbal aspect, which happens to be marginal. The term ‘discourse’ is thus deemed as a vehicle for communication with the target, not necessarily by means of verbalisation. Whereas Sperber and Wilson ([1986] 1995, 2004) originally propose their framework for input, i.e. stimuli in general, and seem to focus primarily on linguistic utterances, we relate their principles to communicative acts embracing visual representation, which necessitates changes in the terminology used. Hence, we avoid using the terms: ‘speaker’, ‘hearer’ or ‘proposition’, giving preference to terminology congruent with advertising: ‘communicator’, ‘recipient’/‘target’ and ‘message’, respectively. The paper will elucidate our stipulation that the relevance-theoretic framework (Sperber and Wilson [1986] 1995, 2004; Wilson and Sperber 2000), if expanded, enjoys a great capacity for explaining the phenomenon of wittiness as a successful advertising strategy thanks to a few fundamental premises of the model. The most crucial of them is the necessity of balance of cognitive effects over processing effort, which will be modified to account for stimuli which cannot be processed with the least possible effort. We propose that the process of interpretation of witty advertising messages may be cognitively challenging and costly, but eventually rewarding. Therefore, arriving at understanding serves as a testament to the recipients’ intellect, perceptiveness, understanding and knowledge, which, in all likelihood, contributes to his/her positive attitude to the advertising message. Our second paramount aim will be to conduct a qualitative analysis of a few advertisements found in Polish quality weekly magazines (Newsweek, Polityka, Wprost and Przekrój). Basing on

50

Chapter Three

these examples, we will elaborate on particular strategies aimed at attracting recipients and engaging them in mental processing.

2. Functions of pictures On the whole, we can distinguish a number of functions an iconic representation serves in advertising. It is generally acknowledged that a picture focuses recipients’ attention and incites them to get acquainted with the advertising message (e.g. Albin 2000, Kochan 2003). Admittedly, this will be the outcome of its salience thanks to its size and colourfulness.3 The message is often conveyed in the form of verbalisation, ranging from a short slogan to a body of text. On the other hand, an advertising picture itself evokes emotional aura and symbolism value, producing a number of connotations and hence implicitly defines the product and the target group. Accordingly, an iconisation often bases on background knowledge, whether in the form of factual knowledge or social scripts, which appear indispensable in the process of comprehension. Also, a picture illustrates and visualises the textual content or motivates it, forming one coherent entity with the accompanying text. Of cardinal importance to us is also the fact that a picture, if witty, not only attracts but also engages the recipient in longer inferential processing. In dictionary entries, wittiness (wit) is commonly defined as the power of giving sudden intellectual pleasure by unexpected combining or contrasting of previously unconnected ideas or expressions (COD), the ability to say things which are both clever and amusing at the same time (DCE). We would postulate that witty advertising messages will be the manifestation of copywriters’ wittiness, but more significantly, they will also entail the addressee’s mental involvement in deciphering the meaning communicated implicitly and unconventionally. We thus label as witty any advertisement which hinges on some form of incongruity either inherent to the picture itself or originating from the juxtaposition of the image with the slogan. This incongruity leads to momentary surprise and/or incomprehensibility in the target’s mind and gives the latter the incentive 3

It cannot be ruled out, however, that in some cases it is the text rather than the picture that appears to be the most conspicuous, given its size or colour or the lack of any accompanying pictorial representation. On the whole, it can be hypothesised that the element attracting attention is the biggest and the most colourful one, which is thus the most conspicuous against the whole background. Therefore, the focus of interest may even be a minute black-and-white picture in the corner of an otherwise blank page.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

51

to devote more cognitive effort and follow the cognitive path delineated by copywriters. The target must strive to find motivation for the incongruity, by means of which he/she solves the riddle created by copywriters and finds congruities between the seemingly incongruous elements,4 the process of which may turn out to be amusing. As a result, the interpreter understands the advertising message, making inferences about the product, arguably hardly questioning its merits since his/her attention is focused on deriving the meaning as such. Notably, activating the recipient is maintained to be one of the most significant and fruit-bearing ploys of persuasion (DoliĔski 2003).

3. Wit as a rhetorical tool increasing the attractiveness and memorability of an advertising message One of the rudimentary characteristics of advertising discourse is that is should attract the audience’s attention (Sutherland and Sylvester 1993; DoliĔski 2003; Kochan 2003). Notably, the potential recipient is always liable to ignore a stimulus or to devote the minimum processing effort to it, should he/she not find it worth inferential processing. This is why the form of the message, viz. of an advertisement, must be motivating (Greenwald and Leavitt 1984; McQuarrie and Mick 1996), i.e. attractive. Since the exceptional form is intriguing, it distinguishes the message from the background or other similar stimuli and contributes to its memorability (Kochan 2003)., this perspective Nota bene on attractiveness is different but still congruent with the traditional rhetorical rule of attractiveness. The latter holds that the enticing value of the form also increases the attractiveness of the message conveyed (Burke 1969). Psychologically speaking, it follows that when the target concentrates on the formal aspect of the message, he/she will not (be able to) devote much attention to the verification of the basic information conveyed. Given the subject of this paper, we would conclude that a witty advertisement, demanding more effort to be understood, hinders the interpreter’s evaluation of the product as such. Therefore, not paying heed to its necessity or merits, the target will assess the product favourably, as he/she will have done with its witty advertisement. This is on a par with the postulate of the peripheral route to persuasion standing vis-à-vis the central route to persuasion (Petty and Cacioppo 1981). 4

We shall refrain from claiming that the incongruity is eventually resolved. Notably, whether incongruity is subject to resolution is the bone of contention in contemporary debates on the mechanisms of humour.

52

Chapter Three

The two determinants of wittiness are novelty and surprise, which render a message conspicuous. These two parameters are opposed to habituation, which relates to deafening and marginalising familiar, and thus insignificant, stimuli, while sensitisation relies on increasing the level of attention devoted to a new stimulus and more effort in order to process it (Sternberg 1996). Hence, the novelty of a witty form accounts for the message being conspicuous and attractive, which helps it focus the attention of the addressee. It has been long since acknowledged that the qualitative exceptionality of a stimulus which distinguishes it from multitudes of other stimuli facilitates its memorisation (Szewczuk 1966). Moreover, such attention-focusing forms will consequently demand considerable processing effort on the part of the recipient. This triggers further results, given that more painstaking effort during the processing of information renders it more retrievable from memory in comparison to effortlessly processed information (Greenwald and Leavitt 1984). If the interpreter arrives at the final resolution, a rich network of associations is set in motion, whereby multiple encodings arise, strengthening the existing conceptual memory linkages (McQuarrie and Mick 1996). Therefore, the additional cognitive activity expended in the process of interpretation increases the number of associative pathways stored in memory (Mitchell 1983). To recapitulate, the prolonged inferential process, as any form of engagement of the recipient, correlates positively with higher memorability of the text subjected to interpretation. It is even claimed that either form of activation, i.e. a physical action or mental interpretation, is the sine qua non for memorisation of a stimulus (Szewczuk 1966). The memory processes are also associated with the addressee’s willingness to remember the line, being elevated and flattered (cf. captatio benevolentiae and ingratiation). Decoding the attractive, albeit by no means transparent, message is immensely engaging and involves additional expenditure of mental effort (cf. Gibbs 1994; Noveck et al. 2001), but is only successful as an advertising strategy if the riddle is solved (Dorfles 1973). Solving a cognitive riddle i.e. perceiving and comprehending the witty proposition flatters the audience’s intellectual capabilities (van Mulken et. al 2005) and thus establishes rapport between the communicator, in this case the copywriter, and the audience (Norrick 2003), which may increase a positive attitude towards the product endorsed.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

53

4. Relevance Theory and the wittiness of an advertisement In this succinct overview and finetuning of Relevance Theory (Sperber and Wilson [1986] 1995, 2004; Wilson and Sperber 2000) we will concentrate only on just a few aspects of the original framework which we find the most essential for the analysis in question. We will thus neither support nor undermine those premises of the theory which are relatively insignificant for the approach assumed in this paper. The few tenets which we consider of most crucial importance will be modified and enriched with other cognitive postulates, mostly derived from psychological studies. Earlier we emphasised that the primary quality of a successful advertisement is its salience given the infinite number of advertisements. This ties in with the stipulation of ostensive-inferential communication, which entails “the use of an ostensive stimulus designed to attract the audience’s attention and focus it on the communicator’s means” (Sperber and Wilson 2004, 611). In other words, Relevance Theory presumes that what the communicator conveys must appear relevant to the interpreter so that the latter is motivated to embark on the analysis of the ostensive stimulus, and not other inputs, perceiving his/her5 cognitive effort to be overbalanced by cognitive effects as a result. Sperber and Wilson ([1986] 1995, 2004) motivate this with the Cognitive Principle of Relevance, which holds that human cognition is geared towards the maximisation of relevance. The audience will thus choose from all the inputs available only to focus on the stimulus which they regard as the most relevant, i.e. granting them the greatest cognitive effect at a given time. It is worth noting that the theory seems to be corroborated by the observation, derived from cognitive psychology, according to which humans are flooded with stimuli, which they cannot process, and thus must choose only some of them, in which they allocate their processing resources so as to maximise cognitive benefits (e.g. Zimbardo and Gerrig 2002). Relevance is hence a property of an input which renders it worth processing, whereas an input which does not guarantee benefits cannot be labelled as relevant. What is essential, the enhancement of knowledge must be of higher value than the costs incurred. The most fundamental premise of the framework is that the relevance of the input increases with the rising positive effects and decreasing processing effort expanded (Sperber and Wilson [1986] 1995, 2004; Wilson and Sperber 2000). The 5

Contrary to the common practice of relating to a speaker as if he were only a male, Sperber and Wilson use the form ‘she’. Nonetheless, we prefer to use the form he/she for the sake of political correctness.

54

Chapter Three

question arises, however, how interpreters choose the relevant stimuli from all the inputs available at a given time and how they weigh the potential effects and effort, which we would attribute to the motivation to endeavour to understand a stimulus, here understood as a pictorial advertisement. Sperber and Wilson (2004), supposedly, responding to various critiques of their postulates, maintain that the effort and effect are non-representational dimensions of mental processes and, therefore, they cannot be represented in numerals only to function as intuitive comparable judgements. Admittedly, this allows for much leeway in the interpretation of the balance of costs invested and effects retrieved. As stated earlier, relevance prototypically arises from the maximal deduction from new information and existing knowledge with minimal processing costs and is defined as an equation of contextual effects over processing effort. Initially, Sperber and Wilson ([1986] 1995) proposed that the increase in relevance is commensurate with the decline in the processing effort required. Further, however, they verified this postulate, adding that the speaker may assess their own effort sacrificed for the sake of creating an utterance and expect comparable effort from the recipient, rather than minimise his/her effort (Sperber and Wilson [1986] 1995). It is doubtful whether this will be the case of witty advertisements, as copywriters will not aim to overburden the target to more effort just because they have invested more effort in the creative process. According to the second clause relating to the presumption of optimal relevance, the communicator, guided by his/her abilities and preferences, will aim to make the ostensive stimulus as easily available to the audience as possible (Sperber and Wilson 2004). From our perspective, the abilities and preferences may lead the communicator to produce a stimulus which will be available, but not necessarily the most easily available so that it will be worth processing at all. In other words, the communicator may produce an ostensive stimulus, but simultaneously encourage the addressee to make some effort in order to process it. Pictorial advertisements, even if at first glance not entirely incomprehensible, must be regarded as worth processing for the sake of ultimate understanding the advertising message. Those who are willing to follow this cognitive path and succeed in inferencing, viz. invest effort and reap the credit in the form of understanding and self-satisfaction, are deemed the target. It is difficult, if not impossible, to judge whether the recipient decides to embark on the analysis only to understand the stimulus as such, perceiving it as an intellectual challenge, or to discover the advertisers’ aim, appreciating the fact that the latter must have used an implicit means of communication to attract their attention and engage them in mentally

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

55

processing the message. Notably, an overt message of persuasive function would put paid to the contemporary method of advertising which prefers subtler means of communication. Nowadays, marketers try to avoid transparently signalling their main aim, i.e. promoting the product (Sutherland and Sylvester 1998). However, it is undeniable that conscious readers will instantly perceive the stimuli in question as advertisements, regardless of how formally unconventional they are. In conclusion, upon encountering a page with a colourful picture and little text, the addressee will indubitably share the assumption that the stimulus is to serve as an advertisement. Hence, he/she may deem it as relevant in this context. Moreover, it could be hypothesised that contrary to the postulate of negative correlation between relevance and effort, it is those effortprovoking, i.e. witty/incongruous, messages that will emerge as the most relevant, being conspicuous against other stimuli whose meaning is more easily retrievable and thus may not even be considered as relevant and thus worth analysing. This phenomenon could be explicated by the psychological postulate of unavailability, according to which unavailability of goods corresponds to desirability (e.g. Cialdini 1993). In the case of witty advertisements, the unavailability pertains to meaning, which can be inferred only with additional cognitive investment and contextual resources. In essence, we would concur with humour researchers (e.g. Tanaka 1994; Yus 2003), expanding the original postulate of relevance. In the case of stimuli which escape instant understanding, supplementary mental effort will be required for a more relevant interpretation. Notably, stimuli which take longer to process sustain the attention of the addressee longer, yield more contextual effects producing a pleasant feeling consequent upon intellectual satisfaction (Tanaka 1994). Accordingly, we would even subvert the original postulate claiming that more beneficial effects are correlated with more effort-demanding, and thus costly, paths. As a result of all this, the audience is first intrigued, then involved in mental processing, which if successful renders them more congenial towards the advertisement and ultimately the product. Context is mostly subject to the hearer’s control, with the utterance having been regarded as relevant. Contrary to the common interpretation of the term, in Sperber and Wilson’s terms, context is defined as a psychological construct constituting the hearer’s assumptions used in inferencing. It is thus not restricted to the preceding discourse and its circumstances but embraces also encyclopaedic knowledge of various spheres. Such knowledge will be largely idiosyncratic and liable to alteration over time. This conceptualisation of context concurs with the

56

Chapter Three

psychological contribution of schema theory first proposed by Bartlett (1932). Contextual effects are obtained as the result of juxtaposition of and interaction between the newly presented stimulus with the already existing information to yield a contextual implication. The latter is here understood as the comprehension of an advertising message. Processing effort relies on the complexity of the stimulus and hence the accessibility of the context. Notably, in producing a message, the communicator should preconceive whether the hearer possesses all the indispensable knowledge to process the communicative act. The model is anchored in the hearer’s access to background beliefs and assumptions forming a context in which the incoming information is processed (Yus 2003). In Sperber and Wilson’s own words, “the speaker must make some assumptions about the hearer’s cognitive abilities and contextual resources” ([1986] 1995, 138). The communicator must adequately assess the target’s cognitive environment and accessibility of the advertising proposition. We should observe that this entails assessing the latter’s intellectual capacity and knowledge, not only linguistic but also extralinguistic (cultural, theoretical, etc.), which grants accessibility of the stimulus. In the context of this paper, those who are able to infer the meanings conveyed are thus defined as the target of the advertisement and the product.

5. Sample analysis As earlier assumed, the common denominator between the selected examples is wittiness understood as cleverness manifesting itself in any form of incongruity. Within our approach, a visual persuasive message initially produces an effect of surprise due to the discrepancy between the stimulus and the receiver’s cognitive environment, the effect of which is dubbed ‘cognitive dissonance’ (Festinger 1957). We would stipulate that the function of attracting the recipient, ascribable to any pictorial representation in advertising, is enhanced if the visualisation hinges on incongruity inherent to the image itself or stemming from the juxtaposition between a picture and a slogan. As a result, such an advertisement escapes instant understanding and demands more cognitive effort, which is all the greater if the advertising message is reliant on rich context, arguably not available to everybody. In the analysis below we aim to discuss six sample advertisements, conforming to the consecutive processes of perception and intellectual processing which take place in the recipient’s mind. The qualitative analysis of rhetorical techniques is structured in a way enabling readers to arrive at their independent interpretation of each

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

57

advertisement. Therefore, in each of the six sections, the picture is first presented and then analysed. Since the examples come from the Polish press, the slogans or accompanying captions have been translated into English so that those unfamiliar with the Polish language are not underprivileged. Admittedly, in two of the examples (2, 3), cultural constraints can preclude understanding to those unacquainted with the relevant aspects of Polish culture. However, this could also happen interculturally, since not all Poles would be capable at inferring the meaning.6 In essence, the lack of understanding may be the consequence of any scripts, whether culture-dependent or cross-cultural, e.g. dependent on the addressee’s: gender and age (5) or social and financial status (4). Comprehensibility facilitated by the availability of context is thus the determinant specifying the target of a given advertisement, but not necessarily a product, which may be advertised otherwise to cater for a different target. Each of the six advertisements and their potential target will now be discussed in detail. (1)

“Mirror, mirror on the wall…” 6

All the examples have been analysed as assignments with approximately a hundred university students during conversation classes. Although of Polish origin, sometimes not even a half of the students managed to infer the advertisers’ meaning, which seems to yield the conclusion that the target of an advertisement can be very narrow indeed.

Chapter Three

58

The advertisement of a car presented above is rather explicit, given the salient presentation of the product advertised. Also, the process of reaching implicatures demands relatively little inferencing, due to the alleged high availability of the context. However, the process is multistaged and quite complex. It must be appreciated that the picture is visually attractive thanks to the incongruity in the image of the car, which strikes recipients as absurd. Having encountered the incongruous stimulus, addressees will turn their eyes towards the slogan Mirror, mirror on the wall..., which alludes to the fairy tale about Snow White, in which the words were uttered by the evil queen, who wished to ascertain her beauty. Drawing a parallel between the fairy tale and the message, composed of the picture and the slogan, the target derived the implicature that the car can do nothing but marvel at itself in the mirror. Ultimately, the pictorial representation intertwined with the slogan is supposed to yield the cognitive effect, i.e. the inference concerning the beauty of the car. Arguably, another implicature that the target may derive pertains to the features of vanity or beauty which can be excelled, which can by no means be the advertisers’ aim. (2)

While the previous message was relatively easy to comprehend given the information provided, this advertisement promoting a brand of beer is a most interesting case of strong context dependence which gives rise to incomprehensibility should the interpreter not be equipped with background knowledge. The advertising ploy of unavailability hinges on pictorial representation, not accompanied by any text save the intelligible signs resembling Asian writing, which indeed creates a perceptive shock.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

59

Transparently, given the scarcity of information and its unavailability, the advertisement is directed at a very well defined target, i.e. those already acquainted with the product and its logo. Such a strategy will indubitably appeal to the target group who do recognise the brand and may also be aware of the exclusiveness of the advertisement, and hence be flattered by finding themselves among the chosen few. It is undeniable that a person unfamiliar with the product can only infer the type of product advertised, i.e. beer. The more perceptive may also infer the Polish origin of the beer, which also entails additional knowledge of the Polish flag. The provenance of the product appears to be implied by means of the colours used, white and red, alluding to the Polish flag composed of two horizontal stripes. Moreover, another implicature derived by the target contributing to the global cognitive effect may be that the Polish national product is of such a high quality that it is exported to, even as far as, Asian countries. On the whole, it must be appreciated that this advertisement proves how many effects, hence rewards, can be recovered solely from a picture as long as the target has sufficient context in the form of background knowledge. (3)

“She has just heard a hysterically funny joke” “A restrained English taste”

Another advertisement of beer is also couched in a rich context, which this time does not entail the knowledge of the brand but relates to broader cultural knowledge. The advertisers employ this strategy with a view to forming the image of the product. The aim is also achieved via incongruities on the text-picture axis, further enhanced by the

Chapter Three

60

incompatibility with the product. What strikes the recipient first is the solemn and even unhappy woman, whose facial expression is at odds with the text in the top left-hand corner of the picture. Failing to understand the implicature emerging from the interdependence of the picture and the slogan, the addressee will doubtless read the line at the bottom saying a restrained English taste, which is applicable to the brand of beer presented on the right. It is when the recipient acknowledges the bottle with the label that the advertising function is fulfilled. Admittedly, the message might serve its primary purpose, even if not all the implicatures are entertained. For the sake of garnering all cognitive effects, the recipient should arrive at the interpretation based on the stereotype presenting the English as a calm and restrained nation, who will remain poised regardless of the circumstances. Notably, the Englishness of the woman is not restricted to her emotional inhibition, but also the features of her appearance: reddish hair, pale complexion and plain clothes. The line at the bottom embracing the qualifier English contributes to the context and ushers in help for those who have failed to follow this inferential path, basing solely on background knowledge. The overall implicature is that the factor of Englishness ascribable to the visualisation applies also to the beer, which bears an English name. In essence, this advertising strategy is meant to convey the meaning that the brand of beer is indigenously English. This information, however, is false as the beer is produced in Poland. (4)

Like the second advertisement presented earlier, this one also presupposes the knowledge of the brand and hence narrows down the target to those in the know. However, this time the technique does not

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

61

pertain to the main product advertised but the company that endorses it, and arguably also promotes itself. Encountering the advertisement, the addressee will undoubtedly notice the surprising use of a carpet as shoe lining. Apart from that, the advertising picture is indeed aesthetically attractive and sublime showing a stylish high-heeled shoe, whose lining merges with the carpet made of the same fabric on which it is lying. Initially, it may turn out to distinguish the object in the focus of the advertisement. Only a few preoccupied with the latest fashion and posh brands will appreciate the fact that the stiletto is a product designed by Manolo Blahnik, not Brintons as its label on the lining of the slipper may suggest. For those who know the celebrity brand, but would not recognise it, there is an inconspicuous line in white letters in the top left-hand corner, which elucidates who the producer of the footgear is. However, the label with the company name relates to the carpet, which appears to be the product advertised. This juxtaposition of the supreme shoe brand with the carpet bears the implicature that the latter is also a luxury product of very fine quality. (5)

“The best protection guaranteed”

This advertisement is also anchored in a rich context and familiarity with another product but not any particular brand. The phenomenon of wittiness resides in an unexpected, yet most adequate, juxtaposition appreciated by those who possess the indispensable knowledge to observe the point of similarity. Without the necessary information, the whole meaning of the witty advertisement may go unnoticed, even if it should be

Chapter Three

62

acknowledged what product (a probiotic) is advertised. Arguably, the target will be primarily emancipated young and middle-aged women, who are the most likely to appreciate that the most salient image presents a packet of contraceptive pills. There is thus no incompatibility between this product and the slogan which certifies the infallibility of the product. The initial inference must be that the protection guaranteed is against pregnancy. Only later does the reader’s attention move to the bottom of the page which bears a small picture of the product advertised. The small image of a bottle is accompanied by information, which arguably only a few will read, as the probiotic and its merits are relatively well known.7 More meticulous addressees will also notice that the little slots do not contain pills but miniature bottles of the product, which is another means of drawing the parallel between the two products. Essentially, the cognitive effect conveyed via the comparison is that the probiotic affords almost perfect protection against viruses and germs only when used regularly. (6)

As was the case of the first example, this advertisement boasts the explicature of the promotion of a car. However, the advertiser has chosen to base the campaign in an intercultural context by alluding to a famous film, via three elements, which if appreciated, yield greater cognitive 7

In Poland, the product has been widely advertised in the press and on TV. Nevertheless, it is noteworthy that earlier advertising campaigns seem to have been targeted primarily at mothers and spouses, preoccupied with the health of the whole family.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

63

effects. Firstly, the advertising picture presents a car against the background of yellow flowers.8 Secondly, upon closer investigation, one can notice a vignette of a Japanese woman’s face among the flowers. Thirdly, the picture is accompanied by a very peculiar slogan Japan Beauty, which, apart from pointing to the provenance of the product, alludes to the film American Beauty. Notably, if the intention were only to stress the origin of the car the adjective Japanese would be used, though then the word ending would not coincide with that of the adjective in the title of the film.9 The picture itself mimics one of the scenes of the film, arguably the most memorable one, which presented an alluring young female lying naked among petals of red roses. What the target ought to infer is that while the girl was the object of the man’s desire, but was only a figment of his imagination, the car is desirable but attainable. The implicature may thus be that men will buy the car, finding it remarkably attractive and desirable. Surprisingly enough, according to the information in brochures, the model is recommended as a family car and the target is primarily females. Notably, even if oblivious to the allusion to the film, women are more likely to approve of the flowery poetics of the advertisement triggering stereotypical connotations of femininity and subtleness. In conclusion, it could be inferred that the target of the advertisement is either women, who either perceive the film allusion or not, or men who wish to buy their partner a car.

6. Conclusions The above seems to support the legitimacy of the relevance-theoretic model, albeit a slightly modified, as a heuristic tool explaining the mechanisms of wit in the visual rhetoric of press advertising. We maintained that a few cognitive linguistic postulates offered by Relevance Theory, whose scope is supported or broadened by a number of psychological findings, explain why witty advertising messages carry persuasive potential. We claimed that, in terms of advertising, the ostensive, i.e. the relevant, stimulus may be the one that is more effort provoking than other stimuli, which often correlates with its conspicuousness in the flood of advertising messages. Moreover, recipients willing to face the cognitive challenge nurse expectations, 8

The colour seems to be insignificant since there were also advertisements of the car with petals of different colours. 9 It must be remembered that the advertisement was targeted at the Polish audience, who may not have noticed the mistake but will have recognised the allusion to the film.

64

Chapter Three

obviously only subconsciously, that the invested effort will generate commensurate effects. We also postulated that witty advertisements are perceptually attractive and intellectually challenging, which eventually renders them memorable and acceptable to the recipient. More specifically, witty images, usually accompanied by short texts, trigger inferential processing and hence demand cognitive effort, often also with the use of context in the form of knowledge of various types, as a result of which addressees are granted cognitive rewards. Those coincide with the target’s understanding of the message, which ultimately produces their high self-esteem. To reformulate, having been attracted by the form of an advertisement and resolved, the riddle hinged on the incongruity between elements of the picture or the picture and the accompanying text, recipients feel flattered by proving their own cognitive capabilities and finding themselves among the target of a given advertisement. In consequence, they are bound to lean towards the advertisement and may also accept the persuasive message without much consideration of the product as such10 by developing a positive attitude towards it. What is noteworthy, it is also the longer inferential process that facilitates remembering the brand and hence considering potential purchase. In the second part of the paper, we attempted to verify these postulates via the analysis of six different advertisements, taking the inferential path which addressees are meant to follow.

References Albin, Krzysztof. 2000. Reklama. Przekaz, odbiór, interpretacja. Warszawa: PWN. Bartlett, Frederic. 1932. Remembering: A study in experimental and social psychology. Cambridge: Cambridge University Press. Burke, Kenneth. [1950] 1969. A Rhetoric of motives. Berkeley: University of California Press. Cialdini, Robert. 1993. Influence, science and practice. Harper Collins College Publishers, Inc. Contemporary Oxford Dictionary. Oxford. Dictionary of Contemporary English. Longman. DoliĔski, Dariusz. 2003. Psychologiczne mechanizmy reklamy. GdaĔsk: GdaĔskie Wydawnictwo Psychologiczne. 10 However, the rationality of purchase will also depend on the type of product, the level of involvement and hence the consequences of purchase.

Wittiness in the Visual Rhetoric of Advertising and the Quest for Relevance

65

Festinger, Leon. 1957. A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Gibbs, Raymond. 1994. The poetics of mind. Cambridge: Cambridge University Press. Greenwald, Anthony, and Clark Leavitt. 1984. Audience involvement in advertising. Four Levels Journal of Consumer Research 11: 581-592. Jones, John. 1997. Is advertising still salesmanship? In How advertising works, edited by J. Jones. London: Sage. —. 1999. The creative characteristics of successful television advertising. In The advertising business, edited by J. Jones. London: Sage. Kochan, Marek. 2003. Tekst i obraz w reklamie prasowej. In JĊzyk perswazji publicznej, edited by K. Mosioáek-KáosiĔska, and T. Zgóáka. PoznaĔ: Wydawnictwo PoznaĔskie. McQuarrie, Edward, and David Mick. 1996. Figures of rhetoric in advertising language. Journal of Consumer Research 22, no. 4: 424438. Mitchell, Andrew. 1983. Cognitive processes initiated by exposure to advertising. In Information processing research in advertising, edited by R. J. Harris. Hillsdale, NJ: Lawrence Erlbaum. Norrick, Neil. 2003. Issues in conversational joking. Journal of Pragmatics 35: 1333-1359. Petty, Richard, and John Cacioppo. 1981. Attitudes and persuasion: Classic and contemporary approaches. Dubuque: IA Brown. Sperber, Dan, and Deirdre Wilson. [1986] 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. —. 2004. Relevance Theory. In The handbook of pragmatics, edited by L. Horn, and G. Ward. Oxford: Blackwell. Sutherland, Max, and Alice Sylvester. 1993. Advertising, and the mind of consumer: What works, what doesn’t and why. St. Leonards: Allen & Unwin. Szewczuk, Wáodzimierz. 1966. Psychologia zapamiĊtywania. Warszawa: PWN. Tanaka, Keiko. 1994. Advertising language: A pragmatic approach to advertisements in Britain and Japan. London: Routledge. Van Mulken, Margot, Renske van Enschot-van Dijk, and Hans Hoeken. 2005. Puns, relevance and appreciation in advertisements. Journal of Pragmatics 37: 707-721. Wilson, Deirdre, and Dan Sperber. 2000. Truthfulness and relevance. UCL Working Papers in Linguistics 12: 215-254 Yus, Francisco. 2003. Humor and the search for relevance. Journal of Pragmatics 35: 1295-1331.

66

Chapter Three

Zimbardo, Philip, and Richard J. Gerrig. 2002. Psychology and life. Boston, Mass.: Allyn and Bacon.

CHAPTER FOUR WHAT’S IN THE PUNCHLINE? MARIA JODàOWIEC

1. Introduction It is one of the underlying assumptions of Relevance Theory that utterances are public interpretations of thoughts entertained by the speaker, which trigger inferential comprehension mechanisms in the hearer, resulting in the latter constructing mental representations similar to the original thoughts in the communicator’s mind (Sperber and Wilson [1986] 1995, 230). Utterance interpretation is then approached as a mind-reading activity (cf. Wilson and Sperber 2004), the outcomes of which depend only partially on the linguistic forms used, that is why the process goes well beyond mere decoding. Verbal comprehension is believed to be to a significant degree affected by how accurate the speaker’s judgement of both the hearer’s cognitive recourses available at a given moment and his cognitive abilities is (cf. Jucker, Smith, and Ludge 2003) as well as by the actual expectations of relevance that the hearer has at a given stage of the communicative exchange. As it will be argued below, this kind of approach, and, in particular, the treatment of communication as a scalar category, in which what gets communicated depends partly on ‘how far’ the interpreter is encouraged and prepared to go in the comprehension process, can be fruitfully employed in explicating what joke interpretation involves. Joke comprehension is undeniably complex (cf. Colston and Gibbs 2002, 75), but in real life it happens instantaneously (unless the recipient cannot get access to relevant conceptual information that is required, failing to appreciate the joke) and processing humorous texts seems to proceed similarly to everyday discourse comprehension: sometimes only when the punchline is reached does the hearer realise that a joke has been told. This suggests that no special strategies are at play. A pragmatic analysis of verbal jokes as attempted in this paper employs no other principles or rules

68

Chapter Four

but those that are used within Relevance Theory to account for communication per se. The present paper consists of four main parts. In section 2, the notions of strong and weak communication, the latter of which is central to the argument to be developed, are introduced. In section 3, verbal jokes are defined and the role of the setting in joke comprehension is discussed. In section 4, the punchline effect is analysed along relevance-theoretic lines. In the final section, some apparent advantages of adopting the relevancetheoretic perspective in exploring the joke comprehension process are briefly summarised.

2. Strong and weak communication Communication, as Sperber and Wilson ([1986] 1995, 1) contend, is an interaction involving two information-processing devices, one of which modifies the cognitive environment of the other, which results in the addressee constructing mental representations similar to those of the communicator. In an act of overt intentional communication, called ostensive-inferential, the communicator – or with reference to oral exchanges the speaker – produces an utterance, which enables the hearer to infer the communicator’s meaning (cf. Sperber and Wilson 2002, 12). On this model, utterance comprehension is postulated to involve the formulation and evaluation of a hypothesis about the intended speaker’s meaning on the basis of the evidence in the form of an utterance provided by the speaker for precisely this purpose (Sperber and Wilson 2002, 7). This means that, according to Relevance Theory, an utterance, similarly to other ostensive stimuli, does two important things: firstly, it attracts the audience’s attention, and secondly, it focuses their attention on the communicator’s intention and helps to reveal it. Given, as Wilson and Sperber convincingly argue, that “human cognition tends to be geared to the maximization of relevance” (2004, 610), since ostensive stimuli preempt attention and require mobilisation of cognitive resources, they automatically create an expectation of relevance. This is the essence of the Communicative Principle of Relevance, a mental law that describes the behaviour of communicators and addressees in intentional communication, and states that “every ostensive stimulus conveys a presumption of its own optimal relevance” (Wilson and Sperber 2004, 612). Thus the comprehension process, it is suggested, is constrained by an expectation that a given utterance will be worth the processing effort which the addressee must necessarily expend, and that it should be treated by the interpreter as the most relevant utterance that the communicator is willing

What’s in the Punchline?

69

and wants to produce in a certain communicative context (Wilson and Sperber 2004, 612-13). In this approach, it is postulated that an utterance makes manifest1 a number of assumptions, which the hearer uses in recovering what is being communicated. These assumptions come with varying degrees of strength. In other words, among the assumptions that an utterance will make manifest, some may be strongly, while others only weakly, made manifest in a certain communicative act (cf. Sperber and Wilson [1986] 1995, 5960; Carston 2002, 378; Wilson and Sperber 2004, 620; Wilson and Wharton 2006, 1569). So communication can be viewed as strong or weak communication. This is how Sperber and Wilson (1987, 706) describe this: An utterance that forces the hearer to supply a very specific premise or conclusion to arrive at an interpretation consistent with the principle of relevance has a very strong implicature. An utterance that can be given an interpretation consistent with the principle of relevance on the basis of different – though of course related – sets of premises and conclusions has a wide range of weak implicatures. Clearly, the weaker the implicatures, the less confidence the hearer can have that the particular premises or conclusions he supplies closely reflect the speaker’s thoughts, and this is where the indeterminacy lies.

This suggests that in weaker forms of communication, all that the communicator can hope for is, as Sperber and Wilson ([1986] 1995, 60) aptly put it, just “to steer the thoughts of the audience in a certain direction.”2 Let me illustrate this with an example. Let us analyse a brief verbal exchange like this: (1)

1

(a) (b) (c)

Mary: Are you coming to the party on Saturday? Peter: Yes, I am. Peter states/confirms that Peter is coming to the partyi on Saturdayx.

Manifestness is treated in this model as a technical term, and should be understood as “the degree to which an individual is capable of mentally representing an assumption and holding it as true or probably true at a given moment” (Carston 2002, 378). By claiming the audience’s attention, the communicator makes it manifest that she wants to inform the audience of something, in this way making manifest her informative intention, and additionally manifests the intention to inform the audience of her informative intention (cf. Wilson and Sperber 2004, 611). 2 Cf. Crook (2004) for an interesting suggestion that covert communication is a limiting case for weak communication.

Chapter Four

70

Peter’s utterance (1b) in a rather straightforward fashion communicates that he is coming to the party on Saturday. In the relevance-theoretic framework it is postulated that the explicature of the form (1c) ‘Peter states/confirms that Peter is coming to the partyi on Saturdayx’ will be recovered by the addressee guided in the interpretation process by the criterion of consistency with the Principle of Relevance. Notice that in order to arrive at this explicature, defined in Relevance Theory as “an ostensively communicated assumption which is inferentially developed from the incomplete conceptual representations … encoded by the utterance” (Carston 2002, 377), a pragmatic resolution of linguistic indeterminacies on the basis of contextual information must be carried out.3 For instance, in the very short utterance under discussion, reference assignment, ellipsis resolution, disambiguation and illocutionary force assignment must be performed in the process of arriving at explicitly communicated meaning. This means that the recovery of the communicated explicature involves a whole range of pragmatic enrichments (cf. Carston 2002, 121-124), the detailed analysis of which will not be provided, as it exceeds the scope of the present paper.4 The important point to make here is that the direct answer that Peter supplies to Mary’s question belongs to the strongest types of communication. However, if instead of the short utterance in (1b), in response to Mary’s question Peter produces (2b), the process of interpretation becomes slightly more complex, and what is communicated is partly left indeterminate. (2)

3

(a) Mary: (b) Peter:

Are you coming to the party on Saturday? Martha is coming to the party.

Hamblin and Gibbs (2003, 61, 74-75) define explicatures in terms of “what speakers pragmatically say.” 4 Technically, the explicature as presented in (1c) is oversimplified, as the subscripts only signal that the word party must be disambiguated and on Saturday should be given an adequate reading, which involves assigning appropriate time reference to the expression (e.g. ‘on the Saturday that is coming’). Yet, there are possibly other pragmatic processes which have not been indicated, for instance, those relating to lexical broadening or narrowing (in the example scrutinised, these will be relevant to explicating the meaning of the verb come, cf. Wilson 2004). A detailed account of explicatures to be formed would require an elaborate discussion of lexical-pragmatics issues and the explicature/implicature contrast, which goes beyond the problems focused on here and which might obscure the major claims that the present paper attempts to advance.

What’s in the Punchline?

71

The explicature recovered in the process of interpreting (2b), which can be spelled out as ‘Peter states that Marthay is coming to the partyi on Saturdayx’, is not enough in itself for Mary to recover Peter’s intended meaning. But it provides strong evidence that Peter wants Mary to access a contextual assumption that, processed jointly with his utterance, will yield an optimally relevant interpretation as an answer to Mary’s query about Peter’s coming or not coming to the forthcoming event. Let us assume that in the context of Peter’s utterance the assumption that Mary is led to recover is (2c): (2)

(c)

If Martha is coming to the party, Peter is surely coming too.

Observe that in her search for the interpretation that will satisfy the presumption of relevance and that will be consistent with the Principle of Relevance, Mary needs to come up with an implicated premise like that, because there is no other way for Peter’s utterance to achieve relevance. In the context of the implicated premise (2c), Peter’s utterance in (2b) is communicating that Peter is indeed coming to the party on Saturday. However, the extra processing effort that Mary needs to engage in recovering this interpretation, which she might have been spared if Peter had responded with a simple Yes, I am as in (1b), must be compensated for somehow. A relevance-theoretic approach predicts that it will be offset by extra contextual effects, which the simple answer Yes, I am does not communicate. So, a relevance-theoretic analysis of the interpretation process indicates that there are some other implicatures not so strongly communicated by the utterance in (2b), though easily accessible to Mary in the process of recovering Peter’s intended meaning. Let us assume that (2b) makes highly accessible to Mary assumptions (2d, e): (2)

(d) (e)

Peter likes Martha very much. Peter has designs on Martha.

By accessing these, Mary will form implicatures about Peter’s hopes to spend most of the time at the party with Martha and about his reasonable chances to use this opportunity to initiate a romance with Martha. Moreover, since utterance (2b) gives her immediate access to encyclopaedic information about Martha, Mary may access some other assumptions stored under the conceptual address MARTHA, possibly not unlike those in (2d-g), and draw further weak implicatures (cf. Sperber and Wilson [1986] 1995, 87-90).

Chapter Four

72

(2)

(d) (e) (f) (g)

Martha is much more attractive than Mary. Martha is an excellent dancer. Martha treats Mary with annoying superiority. Lads tend to ignore other girls when Martha is around.

So Mary may come up with some other implicatures, possibly that the party is likely to be little fun for her, as she feels uneasy in Martha’s presence. It seems crucial to emphasise at this juncture that while (2c) is very strongly implicated by Peter’s answer in (2b), because without accessing an implicated premise of this kind (though not necessarily this particular one), the utterance will fail to be relevant, assumptions (2d-g) or some others that may become manifest to Mary, are only weakly implicated. As they are only weakly implicated, Mary takes responsibility for the implicatures calculated on the basis of accessing these weakly manifest assumptions. I would like to argue that the notion of weak communication, which has been briefly characterised through this example, can be fruitfully applied to analysing the processes underlying the interpretation of the punchline in verbal jokes.

3. The role of the setting in joke-processing Before I proceed to discuss the punchline effect, let me characterise what I mean by a joke here and briefly describe the function of the first part of the joke. For the purposes of this paper, a ‘verbal joke’ will be defined as an ordered sequence of utterances, planned as a unit, with a humorous climax, or punchline, which is intentionally produced by the joke-teller to amuse the audience.5 The approach to jokes as mini-texts is quite common in humour studies (cf., among others, Sacks 1974; Togeby 1983; Raskin 1985; Schweinsberg 1987; Navon 1988; Giora 1991). Being often a part of some larger discourse or text, for instance, a conversation, novel, play, TV show etc., the text of a joke always stands out as a salient unit from whatever precedes or follows it. Frequently, jokes are introduced into the discourse, by metalinguistic formulaic expressions (cf. Togeby 1983, 358; Linde 1981, 92-95) of the type: Did you hear the one about... or Stop me if 5

Even those researchers who point out that a clear distinction between jokes, anecdotes, wordplay, irony and other humorous texts is virtually impossible to draw indicate that the presence of a punchline can be seen as a defining characteristic of verbal jokes (cf. inter alia Norrick 2003)

What’s in the Punchline?

73

you’ve heard this one..., which mark the onset of a joke. However, even if this ‘preface sequence’ (Sacks 1974, 340; cf. also Rutter 2000, 481) is absent, it is usually pretty obvious at which point the joke telling begins and when it ends (cf. Linde 1981, 92). It seems worth noting that a joke constitutes not only a linguistic unit, as a sequence of linearly ordered utterances that can be singled out as a complete text, but a joke is also a sociolinguistic whole, in that the joketeller holds the floor in the verbal exchange until the joke coda has been delivered unless, as part of the joke, the hearer is supposed to respond in some way (cf. Linde 1981, 93-94). In the approach that is being followed here, a verbal joke is treated as a structured text/discourse unit in which there are two parts, namely the setting and the punchline.6 Each of these parts, as it will be argued below, plays an important role in the joke interpretation process. The role of the setting in a joke follows from what Relevance Theory postulates about text processing in general. Since utterances are believed to automatically create expectations of relevance, an opening sentence of a story, novel, play etc. is postulated to achieve relevance by making accessible to the audience information to be used in the interpretation of what follows. In text processing then, each successive sentence contributes to the overall relevance of the text and provides the context for the interpretation of subsequent sentences or utterances (cf. Van der Henst, Politzer, and Sperber 2002, 5). In this way, in processing linearly ordered sequences of utterances that form a unit, which – as is being argued – is the case with interpreting verbal jokes, the hearer automatically uses each of the preceding portions of the text as supplying a context in which to interpret utterances that follow (cf. Sperber and Wilson [1986] 1995, 137141). In fact, the role of context and co-text in creating essential background for the interpretation of utterances is emphasised also by researchers working in psycholinguistic paradigms different from Relevance Theory (cf., for example, Colston and Gibbs 2002; Ivanko and Pexman 2003). In very general terms, the joke setting is intended to facilitate access to a set of immediately available assumptions essential for the comprehension of the punchline. Technically, when processing the setting of a joke, the hearer admits into what the originators of Relevance Theory 6

I will not be concerned in this paper with some important social and structural properties that distinguish jokes from stories, the issue which is dealt with in detail, among others, by Linde (1981).

74

Chapter Four

call the deductive device,7 a set of assumptions, each of which is a structured set of concepts. These concepts provide access to lexical, logical and encyclopaedic entries associated with them. In this way the setting contributes to creating the context, and directing “the hearer’s inferential processes to the intended interpretation” (Carston 2002, 47). Identifying the function of the setting in a verbal joke as providing the context for the interpretation of the punchline might indicate that the initial context in this situation is given to the interpreter, which would – in a sense – undermine one of the basic claims of Sperber and Wilson’s ([1986] 1995, 142), namely that context is selected by the addressee rather than given to him or her. A potential solution to this problem might be to follow those researchers who claim that joke production and comprehension should be approached as a special type of communication, which is governed by distinctive rules and constraints, different from other verbal exchanges (cf. among others, Raskin 1985; Graesser, Long, and Mio 1989; Curcó 1995). If this line of argument is pursued, it might be admitted that in the ‘joke-telling/reception mode of communication’ context is fully determined by the joke deliverer. However, one of the important claims to be made here is that a relevance-theoretic analysis of joke processing makes it possible to approach this communicative situation as involving no special mechanisms of comprehension. In fact, it is one of the significant advantages of adopting a relevance-theoretic analysis as applied here that joke comprehension can be explicated without recourse to any specialised strategies of production and interpretation that would be different from those implemented in dealing with non-jokes. This seems intuitively appealing since – as indicated in the introductory section above – it may happen that the audience will discover that they have just heard a joke only after the punchline occurs. Coming back to the role of the setting in jokes, it must be said that even though the setting provides context for the interpretation of the punchline, it is not the case that context is fully determined in advance. The point is that although the setting in an important way influences the assumptions that are activated in the hearer’s mind, the joke recipient actively participates in building up the background assumptions used during the interpretation process, the way it happens, as Sperber and Wilson ([1986] 1995, 141-142) maintain, in all other kinds of ostensive communicative situations, where the addressee needs to select a relevant subset of assumptions in which to process the utterance under 7

The term ‘deductive device’ should be understood here as “the inferential system which blindly performs its computations on the input that is given” (Carston 2002, 7).

What’s in the Punchline?

75

interpretation. As it is suggested here, in the course of processing the setting, a number of concepts become activated in the joke recipient’s mind, thus making the background knowledge that the individual has stored under them highly accessible. All this means that direct access to these concepts is facilitated or even provided by the words used in the text of the joke setting, which shows that the speaker (or writer) presenting a given joke exerts influence on the context accessed by the interpreter (no wonder some researchers refer to this process as ‘manufacturing’, cf. Dascal 1985). However, the control over context is shared between the joke-producer and joke-recipient, as, firstly, as hinted at above, the recipient chooses a relevant subset among those that are triggered, and secondly, what is stored under the concepts activated is pretty much idiosyncratic. So while it is mainly the joke-teller who is responsible for making sure that the audience will retrieve the right kind or kinds of context in which to process the punchline, his or her control over context is by no means absolute. To recapitulate the remarks on the function of the joke setting in preparing ground for the punchline, let me repeat that in the course of processing the setting, a number of concepts become activated in the joke recipient’s mind. Making the assumptions associated with the concepts triggered by the setting highly accessible results in their ‘being on a standby’ in the joke-recipient’s mind, and this, as will be argued shortly, in an important way contributes to creating the punchline effect. Before an analysis of what happens when the punchline is processed is offered here, an illustration of what the setting does for joke comprehension seems in place: (3)

(a)

John Paul II dies and goes to Heaven. God Himself greets him at the Pearly Gates. “Be thou hungry, Dear John Paul?” asks God. “I could eat,” replies the Pope. After all, reaching Heaven takes time and effort. So God opens a can of soup and reaches for a chunk of rye bread and they begin to share it. While eating this humble meal, John Paul looks down into Hell and sees the inhabitants devouring huge steaks, lobsters, pheasants, and pastries. Curious, but deeply trusting, he remains quiet. Next day at breakfast time canned soup and rye bread are served again. Once again, the Pope can see the denizens of Hell enjoying bacon and eggs, fried mushroom, baked beans, different types of jams and marmalades. Still he says

Chapter Four

76

(3) (b)

nothing. The menu remains the same at lunch. After dinner consisting of canned soup and rye bread, when God and John Paul comfortably settle for a game of chess, the Pope cannot contain himself any longer. Meekly, he asks, “God, I am grateful to be in Heaven with You as a reward for the pious, obedient life I led. But here in Heaven all I get to eat is canned soup and a piece of rye bread and in the Other Place they eat like emperors and kings! I just don’t understand it.” God sighs. “Dearest John Paul, let’s be honest. To employ a cook for just the two of us – wouldn’t that be wasteful?”

The initial context in this joke can be presented as a set of assumptions corresponding to the explicatures encoded by the sentences in the setting. This set can be roughly presented as being not very dissimilar to the following: John Paul II dies and goes to Heaven. God greets JP at the Pearly Gates. God asks JP in a very polite and formal manner if JP is hungry. JP replies formally and politely that he is hungry. This is understandable as reaching Heaven takes time and effort. God opens a can of soup and God reaches for a chunk of rye bread. God and JP begin to share food. While eating this humble meal, John Paul looks down into Hell and sees the inhabitants devouring huge steaks, lobsters, pheasants, and pastries. JP is curious, but JP is deeply trusting too. JP remains quiet . Next day at breakfast time canned soup and rye bread are served again. Once again, the Pope can see the denizens of Hell enjoying bacon and eggs, fried mushroom, baked beans, different types of jams and marmalades. Still JP says nothing. The menu remains the same (canned soup and rye bread) at breakfast and lunch the following day. After dinner consisting of canned soup and rye bread, when God and John Paul comfortably settle for a game of chess, the Pope cannot contain himself any longer. JP speaks to God in a very gentle and polite way. JP declares that JP is grateful to be in Heaven with God. JP treats his being in Heaven as a reward for the pious,

What’s in the Punchline?

77

obedient life he led. Then JP complains that in Heaven all he gets to eat is canned soup and a piece of rye bread and in Hell (which he calls the Other Place) they eat like emperors and kings. JP admits that JP just does not understand it. To be sure, this is a grossly oversimplified representation of the explicatures formulated by the hearer or reader who is processing the joke under scrutiny. However, going into a more elaborate, though certainly much more technically adequate account of mental representations of the assumptions made accessible by the setting of the joke would, in the first place, be extremely complex, and, secondly, might overshadow the point of focus in this paper, which has to do with the role of the punchline in jokes. In accordance with the description provided above, the key role of the setting in this joke (as well as in other discourse/text units belonging to the genre under investigation) is to afford access to encyclopaedic information in the interpreter’s conceptual store, accumulated under entries associated with the words used in the text. When the punchline is reached, these background assumptions become an essential part of the humorous reaction that is generated. In the joke under analysis, the text grants access to a whole range of concepts, such as JOHN PAUL II, DYING, HEAVEN, PEARLY GATES, GOD, HUNGRY and so on. These will be cross-referenced with a number of other concepts (cf. Carston 2002, 7), in this context possibly RELIGION, AFTERLIFE, ETERNAL REWARD, SAINT, ANGEL, FOOD and many others. Accessibility of a great number of concepts in this way will be crucial for the pivot of the joke to lead to the required interpretation effects.

4. Weak communication and punchline effect As hinted at above, the joke setting creates the context of assumptions available to the hearer or reader, which in an important way prepares the joke-recipient for the interpretation of the punchline. The punchline is an important “obligatory ending of the joke … which pulls the joke together and makes it funny” (Linde 1981, 92). I would like to argue that when the punchline is reached, a mental state is created in which a whole range of weakly communicated assumptions are made manifest to the hearer or reader of the joke. This means that the crucial function that the punchline performs can be defined in terms of it

78

Chapter Four

communicating in a very weak way a vast number of implicatures, some of which the joke-recipient may represent to himself or herself, but most of which may not surface as fully-fledged assumptions at all. In the joke under discussion, it is easy to observe that what the punchline does is in the first place to communicate strongly that John Paul II and God are the only two inhabitants of heaven. However, its major effect is that apart from strongly implicating that there are just two of them there, the punchline leads the audience to suddenly access a whole array of weakly communicated assumptions, which are formed on the basis of background information that the joke-recipient has immediate access to via concepts that have been made available to him or her by processing the setting. It is not unlikely that some interpreters will have accessed assumptions similar to those listed in (4): (4) Stinginess is so widespread that it reaches even afterlife. People’s hopes of going to Paradise after they die may be presumptuous. The eternal reward is really reserved for the very few chosen ones. All the saints that are believed to be inhabiting Heaven may not be there. God created man in His own image and He is ‘human’ in many ways. Mundane needs may be enslaving us for ever, making even the afterlife tough. I have tried to spell out the assumptions that the punchline makes highly accessible to me, but I certainly do not want to suggest that in order to appreciate the joke, each reader or hearer will actually represent to himself or herself most, some, or possibly any of these particular assumptions. In the first place, what has come to an individual’s mind may be quite different from what I have recovered, since, as Relevance Theory predicts, under the relevant conceptual addresses in our minds there may be different pieces of background information stored, and what is more, their order of accessibility will differ from individual to individual and from occasion to occasion. This indicates that the set actually available will vary across interpreters (cf. Sperber and Wilson [1986] 1995, 87-8, 190). Besides, cultural background, individual experience and personal beliefs will affect the assumptions stored under specific concepts, contributing to their inherently idiosyncratic character, so the assumptions made obtainable when the punchline is being processed will vary.

What’s in the Punchline?

79

But more importantly, my point is that in order for the joke punchline to produce the intended effect, all that must happen is for the recipient to become aware of, in very general terms, and probably in quite a weak way, a wide range of weakly manifest assumptions, and this is precisely what the joke ending does. As has already been mentioned, there are grounds to believe that only a few of these will surface in the hearer’s or reader’s mind. However, it is just as feasible to predict that he or she will not have explicitly formed any specific assumptions. It is postulated then that the punchline effect consists in creating in the audience’s minds a certain cognitive effect in which a vast range of assumptions become, all of a sudden, potentially retrievable to the recipient. It must be emphasised that for the punchline to be effective, its processing must lead to the phenomenon which I will tentatively refer to as ‘cognitive overload’. Since there will be a ‘constellation’ of assumptions made manifest by the punchline, very few, if any will actually be formed and most will not attain the status of fully-fledged assumptions that the reader or hearer formulates, even though access to them will abruptly open. It seems reasonable therefore to assume that much of what happens when the punchline is being processed takes place at the subrepresentational level: it is a fact that if we are asked why a given joke is funny, we find it notoriously difficult to explain what makes it humorous, precisely because most (if not all) of the assumptions that are made manifest are not represented in the recipient’s mind. They simply rapidly become manifest or more manifest to the interpreter as a result of processing the punchline, as if flashing through the back of their mind. This kind of analysis can easily explain why some jokes seem to be funnier than others in the opinion of certain individuals. If a given person has access to a rich conceptual field that the setting activates, the punchline is likely to create a considerable ‘conceptual overload’ and that is the reason for the favourable evaluation of the joke’s humorous qualities. The approach adopted can also throw some light on why in the joke about the pope’s going to heaven the punchline will become slightly less humorous if it is changed into something like: God sighs and says: “You may have failed to observe but there are only two of us here, and food should not be our priority.” Although the strongly communicated assumption that the original pivot in (3b) implies is preserved, a number of crucial assumptions related to God’s potential weaknesses, and mundane concerns that reach beyond our life on earth, are simply gone.

Chapter Four

80

The obvious problem, however, that appears for this model of joke comprehension is that if the nature of the punchline could be accounted for merely in terms of the weakly communicated import, any assumption communicated weakly should potentially be funny, and as we know, this is not the case. The exchange (2a-b) between Mary and Peter analysed above in section 2, even though an instance of weak communication, is not recognised as humorous at all. The interpreter does not find (2b) funny, as I will try to argue presently, as long as there is no access to a wide range of contextual assumptions to be made manifest or more manifest by Peter’s utterance in (2b), because ‘conceptual overload’ does not take place. There are good chances, though that things will change, if the interpretation of the utterance in question will create this kind of effect. So here is a slightly modified brief exchange between Mary and Peter: (5)

(a) (b)

Mary: Peter:

Are you coming to the party on Saturday? Martha is coming. At the last party she not only performed a striptease and table dancing, but she also splashed all the water out of the swimming-pool.

Peter’s answer (5b) is potentially – at least mildly – amusing, and changes the unfunny exchange in (2) into a joke. What the punchline in (5) does is to make accessible to the audience a vast spectrum of assumptions about wild parties. With reference to these, it may be interesting to note that some may have to do with certain images that will become available to the joke recipient. As Sperber and Wilson ([1986] 1995, 286 n. 17) indicate, “there is no reason … why [encyclopaedic entries] should not contain – or give access to – ‘images’ or whatever type of object can be used as sources of information in conceptual thinking,” so the images evoked might also contribute to creating the humorous effect. There may be some other, as my native-speaker informant was eager to point out, probably better punchlines to think of in the context of Mary and Peter’s exchange. The sexual undertones in (6b), at least for those who have easy access to a vast array of assumptions that Peter’s answer makes manifest in this respect, will contribute to creating the punchline effect. (6)

(a) (b)

Mary: Peter:

Are you coming to the party on Saturday? Martha has said she is coming. She seems ready to do it on the dining-room table in front of all the other guests.

What’s in the Punchline?

81

Jokes of this kind (so-called ‘semantic jokes’ in Giora’s 1991 framework), which in my earlier analyses of humour I referred to as ambiguity jokes (Jodáowiec 1991), are treated by some researchers as creating a cognitive dissonance, and then leading the hearer to resolving the incongruity (cf. the multiple-graded-interpretations and singlecovered-interpretation model of jokes that Yus puts forward in his 2003 paper; cf. also Yus 2004 as well as Giora’s 2003 graded salience hypothesis, and frame shifting in Ritchie 2005). However interesting the incongruity-resolution approaches to verbal jokes can be, they account for only one category of jokes, namely those in which two competing, clashing, incongruous interpretations are revealed when the punchline is reached. Further doubt with reference to the models just mentioned is raised by the following problem. It is difficult to see – to refer directly to Curcó’s explanations of humorous effects (1998, in Yus 2003) – how the inconsistencies in the organism’s representational system are minimised when the punchline is reached. Curcó suggests that the joke recipient is led to “find an interpretation that either eliminates or prevents the inconsistencies that the incongruous propositional contents might give rise to” (1998, 311-312, in Yus 2003). Even though I do not want to deny that processing jokes whose humour is based on ambiguity yields intellectual pleasure and contributes to the semantic surprise effect that the hearer experiences (cf. Giora 1991), finding a resolution to incongruities cannot account for all mechanisms that humorous texts employ. In fact, the joke about the pope analysed in the present paper, to my mind, introduces incongruities rather than resolves them, as the punchline brings forth a number of assumptions that undermine most people’s ideas of what heaven is supposed to be like. To recapitulate, it seems important to underline that the analysis of humorous effects of punchlines along relevance-theoretic lines as presented above seems attractive for a number of reasons. A relevancetheoretic interpretation of the punchline effect, based on the notion of weak communication, allows for a more homogenous and universal approach to jocular texts. Drawing on Sperber and Wilson’s argument that the processing of certain groups of words together primes particular bits of encyclopaedic knowledge, so that they get pushed into the foreground and acquire a marked status (I have argued that the setting plays just this role). A number of cognitive addresses become accessible and assumptions stored under these addresses are potentially open in the joke recipient’s mind. The punchline yields a huge number of contextual effects in terms of weakly communicated implicatures, which, being enabled

82

Chapter Four

through the priming the setting is responsible for, suddenly become accessible to the hearer or reader as a result of processing the punchline. Joke-processing, as analysed here, involves accessing a vast range of weak background assumptions, and this is, as postulated, a necessary condition for mini-texts to generate a humorous response. This special cognitive effect of assumption overload, which occurs as the punchline is being interpreted, results in an affective response which manifests itself as laughter. But it seems that gaining access to a very wide and underspecified array of weakly communicated assumptions is not a sufficient, though necessary, condition for humour to be generated. I think I am not prepared to formulate any conditions that would be sufficient for a text to make people laugh. Maybe the idea that the recovery of a vast range of weak assumptions must happen at the subrepresentational level could be profitably pursued. This effect is absent from the processing that Mary engages in when interpreting Peter’s utterance (2): there is no vast array of weakly communicated underspecified assumptions that will become manifest and not represented. That is, unless retrieving encyclopaedic information about Martha will yield this kind of effect. This means that if Martha is notorious for making a spectacle of herself at parties, the very utterance of Martha is coming to the party will become a punchline, because it will result in a vast range of weakly communicated assumptions becoming available in the addressee’s mind. Also, it cannot be denied that what is laughed at is determined by social and cultural aspects, so apart from the effect created through language, there are important socio-cultural considerations that will come into play.8 From the point of view of language processing though, the notion of weak communication provides a useful notion to explain what the punchline effect involves.

5. Advantages of the relevance-theoretic analysis My major goal in this paper has been to argue that weak communication, an important theoretical notion introduced by Sperber and Wilson in Relevance Theory, may be applied in an interesting way to explicating what makes a punchline a punchline. As Wilson and Wharton (2006, 1572) accurately observe, what is communicated is sometimes “nebulous, contextually shaded and hard to 8

If Martha is Mary’s sister, she will not be amused by the punchline which makes reference to Martha.

What’s in the Punchline?

83

pin down in conceptual terms,” but that does not mean that communicative goals are not achieved. In a similar vein, two decades earlier Stubbs (1986, 1, cited in Jucker, Smith, and Ludge 2003, 1738) remarked that “[w]hen we speak or write, we are rarely very clear, precise, or explicit about what we mean … This often appears superficially to be an inadequacy of human language: but only for those who hold a rather crude view of what is maximally efficient in communication.” Nowadays a number of researchers argue that the interpretation of utterances which does not lead to a very precise understanding does not in any sense render the comprehension process deficient (cf. Colston and Gibbs 2002, 75; Jucker, Smith, and Ludge 2003, 1737f). On the contrary, as Jucker, Smith, and Ludge (2003, 1737) argue, “vague expressions may be more effective than precise ones in conveying the intended meaning of an utterance.” So weak communication should not be thought of as a poorer form of communication; frequently it can be preferable to the strong kind. Analysing verbal jokes in terms of weak communication seems beneficial, as it provides a descriptively attractive and explanatory framework for analysing verbal jokes. The advantages of the relevancetheoretic approach as presented above can be summarised as follows: 1.

2.

3.

4.

Joke processing as analysed within Sperber and Wilson’s model of communication does not require postulating any special maxims or principles of interpretation associated with the special joke-processing mode (in contrast with other approaches; for instance, cf. Raskin’s 1985 model). The punchline effect studied in terms of a ‘cognitive overload’ brings into focus the complexity and intricacies of joke comprehension, emphasised by many researchers (cf. Yus 2003; Ritchie 2005). A relevance-theoretic analysis of joke interpretation throws some light on what accounts for subjectivity of judgments with reference to what is funny or what is not, and different degrees of amusement that a joke can generate in different people. The different scope, quantity, but also content of assumptions that the punchline makes manifest to different individuals on different occasions will result in the different intensity of humorous reactions that are generated. Applying the notion of weak communication to the analysis of punchlines helps to see why certain themes are better candidates for raising a laugh than others: entries that are associated with rich conceptual information are likely to be funnier, precisely

Chapter Four

84

5.

6.

because the audience will get access to a wide range of background knowledge. (That is why the British are likely to be amused by jokes about the French rather than that by those about the Sierra Leonians.) Jokes must be told in the right way: the more weakly manifest assumptions the setting will activate in the recipients mind, the funnier the punchline will appear to him or her. Extending a set of assumptions and accessing a lot of background assumptions at the stage of processing the setting of a joke, may lead to intensifying the humorous effect for reasons directly related to the nature of weak communication as such. The analysis advanced here can provide a justification why (some) jokes can be amusing for (some) people when they hear them for the second and third time: jokes lead to amusement even if we have heard them before as long as the process of interpretation involving accessing a wide range of weakly communicated assumptions can be replicated.

References Carston, Robyn. 2002. Thoughts and utterances. The pragmatics of explicit communication. Oxford: Blackwell. Colston, Herbert, and Raymond Gibbs. 2002. Are irony and metaphor understood differently? Metaphor and Symbol 17, no. 1: 57-80. Crook, John. 2004. On covert communication in advertising. Journal of Pragmatics 36, no. 4: 715-738. Curcó, Carmen. 1995. Some observations on the pragmatics of humorous interpretations: A relevance-theoretic approach. UCL Working Papers in Linguistics 7: 27-47. Dascal, Marcelo. 1985. Language use in jokes and dreams: Sociopragmatics vs. psychopragmatics. Language and Communication 5, no. 2: 95-106. Giora, Rachel. 1991. On the cognitive aspects of the joke. Journal of Pragmatics 16, no. 5: 465-485. —. 2003. On our mind: Salience, context and figurative language. Oxford: Oxford University Press. Graesser, Arthur, Debra Long, and Jeffrey Mio. 1989. What are the cognitive and conceptual components of humorous texts? Poetics 18: 143-163. Hamblin, Jennifer, and Raymond Gibbs. 2003. Processing the meanings of what speakers say and implicate. Discourse Processes 35, no. 1: 59-80.

What’s in the Punchline?

85

Jucker, Andreas, Sara Smith, and Tanja Ludge. 2003. Interactive aspects of vagueness in conversation. Journal of Pragmatics 35, no. 12: 17371769. Ivanko, Stacey, and Penny Pexman. 2003. Context incongruity and irony processing. Discourse Processes 35, no. 3: 241-279. Jodáowiec, Maria 1991. What makes jokes tick? UCL Working Papers in Linguistics 3: 241-253. Linde, Charlotte. 1981. The organisation of discourse. In Style and variables in English, edited by T. Shopen, and J. Williams. Cambridge, Mass.: Cambridge University Press. Navon, David. 1988. The seemingly appropriate but virtually inappropriate: Notes on characteristics of jokes. Poetics 17, no. 3: 207219. Norrick, Neil. 2003. Issues in conversational joking. Journal of Pragmatics 35, no. 9: 1333-1359. Raskin, Victor. 1985. Semantic mechanisms of humor. Dortrecht: D. Reidel Publishing. Ritchie, David. 2005. Frame-shifting in humor and irony. Metaphor and Symbol 20, no. 4: 275-294. Rutter, Jason. 2000. The stand-up introduction sequence: Comparing comedy comperes. Journal of Pragmatics 32, no. 4: 463-483. Sacks, Harvey. 1974. An analysis of the course of joke’s telling in conversation. In Explorations in the ethnography of speaking, edited by R. Bauman, and J. Sherzer. Cambridge: Cambridge University Press. Schweinsberg, Ilse. 1987. Telling jokes as a linguistic game. In On narratives. Proceedings from the 10th colloquium on speech communication, edited by H. Geissner. Frankfurt/M: Scriptor Verlag. Sperber, Dan, and Deirdre Wilson. [1986] 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. —. 1987. Précis of relevance: Communication and cognition. Behavioral and Brain Sciences 10: 697-754. —. 2002. Pragmatics, modularity and mind-reading. Mind and Language 17, nos. 1-2: 3-23. Togeby, Ole. 1983. The structure of the joke. In Papers from the seventh Scandinavian conference on linguistics, edited by F. Karlsson. Helsinki: University of Helsinki Press. Van der Henst, Jean-Baptiste, Guy Politzer, and Dan Sperber. 2002. When is a conclusion worth deriving? A relevance-based analysis of indeterminate relational problems. Thinking and Reasoning. 8, no. 1: 1-20.

86

Chapter Four

Wilson, Deirdre. 2004. Relevance and lexical pragmatics. UCL Working Papers in Linguistics 16: 343-360. Wilson, Deirdre, and Dan Sperber. 2004. Relevance Theory. In The handbook of pragmatics, edited by L. Horn, and G. Ward. Oxford: Blackwell. Wilson, Deirdre, and Tim Wharton. 2006. Relevance and prosody. Journal of Pragmatics 38, no. 10: 1559-1579. Yus, Francisco. 2003. Humor and the search for relevance. Journal of Pragmatics. 35, no. 9: 1295-1331. —. 2004. Pragmatics of humour strategies in El club de la comedie. In Current trends in the pragmatics of Spanish, edited by R. MárquezReiter, and M. E. Placencia. Amsterdam, Philadelphia: John Benjamins.

CHAPTER FIVE CONVERSATIONAL HUMOUR AS A POWER GAME KATARZYNA KOSIēSKA

1. Introduction The aim of this article is to examine the use of conversational humour (CH) as a tool in the power game engaged in daily interaction. An attempt will be made to explain some of the attendant phenomena using notions derived from the theory of relevance (Sperber and Wilson [1986] 1995, Wilson and Sperber 2002). Claims will be illustrated with real-life examples taken from the Santa Barbara Corpus of Spoken American English (SBC).

2. Defining CH Conversational humour can be defined as any spontaneous production of a humorous utterance within the natural context of a conversation (definition mine). It stands in opposition to canned jokes, whose reception remains relatively stable regardless of the situation in which they are encountered. Conversational humour is heavily context dependent (Boxer and CortésConde 1997), relying for its success on both the participants and the spatio-temporal surroundings. Interestingly, its nature makes it both collaborative and competitive (Holmes and Marra 2002) – it cannot exist without some form of reaction from the audience, and at the same time one of its core, subconscious functions is to display advantage over the interlocutors. This aspect already points to the game-like nature of the phenomenon.

Chapter Five

88

3. Recognised functions of CH Conversational humour has been recognised as performing a wide range of social and psychological functions (e.g. Hay 2000). Let us have a brief look at those that do not openly threaten the social position of the interlocutors.

3.1 Setting a joyful mood On the level of managing conversation flow, the core function of CH might be described as setting a joyful mood and relaxing the atmosphere. This aspect of humour is so powerful and conspicuous that laymen tend to label it as the crucial one. This use will be visible in all the examples quoted below.

3.2 Stressing mutual cognitive environments Another function of CH is stressing the mutual cognitive environments (Boxer and Cortés-Conde 1997). In the example below, a married couple, Jamie and Pete, together with their close friend Harold, are having a conversation when suddenly the overhead light goes on by itself. (1) 1. JAMIE: 2. HAROLD: 3. JAMIE: 4. HAROLD: 5. JAMIE: 6. HAROLD: 7. PETE: 8. HAROLD: 9. PETE: 10. JAMIE: 11. PETE:

Something’s @ wrong @@1 [I know, I thought we] got our electricity fixed. that’s kind of scary]. @@ I thought I turned this off, [though, [It’s been doing] -didn’t I ]? Oh, that’s right, [But] -[you] did.[That’s really scary] then. (SBC)

The mere fact of a malfunctioning electrical appliance would not be sufficient to generate laughter. Instead, what is important is that the interlocutors all know that this should not happen since they had the light repaired. By referring to this common element of their knowledge, they 1

The "@" symbol in transcripts indicates laughter.

Conversational Humour as a Power Game

89

stress the inclusive nature of the funniness of the event, and thus create a bonding instance of CH.

3.3 Delineating group boundaries The third function of conversational humour is also connected with bonding, but this time it delineates group boundaries instead of merely stressing group membership. In the following example, a group of friends are talking about the flexibility of children’s bones: (2) 1. MILES:

[So that’s why they always have] ... more flexibility. Cause as they’re growing, 2. PETE: [Hm]. 3. MILES: you’re making cartilage. So it al- -4. PETE: [Mhm]. 5. MILES: They] always have that, until they reach adulthood... in which case, 6. HAROLD: They get old and cranky [like the rest of] [us]. 7. MILES: [it stops]. 8. JAMIE: [@@]@[@@]@ 9. PETE: [Unhunh], 10. PETE: [and then they break]. 11. JAMIE: Old and rickety and brittle. (SBC) In line 6, by saying They get old and cranky like the rest of us, Harold uses the pronoun us in its inclusive version, thus indicating that he feels a member of the group. What is more, he takes a risk and makes a denigrating remark about that group, namely that its members are old and cranky. In a non-joke situation, this could possibly be a face-threatening act, turning Harold into an outcast. However, when presented as conversational humour, this utterance triggers mutual merriment and divides the social world into ‘us’, i.e. the conversationalists, and ‘them’ – the rest of the world.

3.4 Stigmatising an out-group Conversational humour can not only define an out-group, but it also has the power to stigmatise those belonging to it. What follows is a transcript of a conversation fragment where the same group of friends are talking about the attitude of women to safe sex.

Chapter Five

90

(3) 1. MILES:

2. 3. 4. 5. 6.

HAROLD: PETE: JAMIE: HAROLD: MILES:

7. PETE: 8. MILES: 9. PETE: 10. MILES: 11. HAROLD: 12. MILES: 13. HAROLD: 14. MILES: 15. PETE: 16. MILES: 17. JAMIE: 18. HAROLD: 19. MILES: 20. PETE:

in their twenties, I mean these guys, say these women don’t care if they use rubbers or not. This is what guys have been telling me, you know, and I’m thinking, I- I can’t believe that, cause in this city, with all the information available, It seems like, Right yeah Stupid. people, people have the information they need. It is just amazing, that people in their twenties apparently aren’t. Right. That’s really scary. So I guess it’s time for me to become a priest. @ It’s too dangerous out there. Oh, I don’t know. I couldn’t believe it. This ... woman I knew I was gonna ask out? God. You have to be Catholic first. Her sister, who’s an HIV counselor, @@ @@@ was telling me how her sister just won’t use rubbers at all. Hm. Well where would she put them. and I’m thinking, @@@@@ (SBC)

In line 1, by claiming that these women don’t care if they use rubbers or not, Miles not only creates an out-group consisting of ‘these women’, but also stigmatises them as careless and possibly lacking in intelligence. This remark in itself fails to generate laughter, so Miles elaborates on it, until in line 8 (So I guess it’s time for me to become a priest), he finally hits the right spot and makes the audience laugh. Having succeeded, he continues revelling in his own joke for a few more conversational turns, until he is sure that the reckless women have been sufficiently disparaged and satisfied that his witty remark has been appreciated.

Conversational Humour as a Power Game

91

3.5 Coping with problems One final, commonly recognised application of CH, is coping with problems. The problems can be either within or outside the speech situation. In the example below, two female friends, Doris and Lynne, are having a conversation on an entirely different topic, when Doris interrupts with a remark about their common friend. (4) 1. 2. 3. 4. 5. 6.

DORIS: LYNNE: DORIS: LYNNE: DORIS: LYNNE:

Rana Lee’s gonna have a baby by the way. Who is? Rana Lee. Brian’s wife? Brian’s wife. I didn’t even know she was pregnant, three months ago. Of course I guess they can get big in three months 7. DORIS: Well, I told you. 8. LYNNE: too. 9. LENORE: What’s her name? 10. LYNNE: Bran- what’s her name? 11. DORIS: Rana Lee. 12. LYNNE: Rana Lee? 13. DORIS: Unhunh. 14. LYNNE: Rana Lee. 15. LENORE: Hm. 16. LYNNE: @@@@@@@ I don’t know why I told you that @.SING Rana Lee SING. She had a brother that died from asthma too. (SBC). The apparent problem here is Lynne’s failure to behave appropriately. When faced with a remark about someone’s pregnancy, one is socially expected to express their joy resulting from the fact. In our example, not only does Lynne fail to do so, but she also ostentatiously shows that she does not remember the woman. When she finally realises who is being talked about (line 5), she tries to make a jocular remark about the pace of pregnancy development (line 6). The remark is not considered funny by Doris (line 7), so Lynne tries to cover the embarrassing silence by illcoordinated references to the subject of the exchange (lines 10, 12, 14). Finally, she resorts to nervous laughter and singing, and then abruptly changes the topic (line 16). On the whole, what she has managed to

92

Chapter Five

produce is an unsuccessful application of conversational humour to solve an ongoing problem within the speech exchange.

4. Power In addition to the recognised applications of conversational humour to set a joyful mood, stress mutual cognitive environments, delineate group boundaries, stigmatise outsiders and cope with problems, it might be tempting to see if – and how – CH can be used to claim or assert power in a speech exchange. Out of the many possible definitions of ‘power’, let us understand the term as “that opportunity existing within a social [relationship] which permits one to carry out one's own will even against resistance and regardless of the basis on which this opportunity rests” (Weber 1947, 152). In the context of conversation, power can most easily be exerted by controlling the topic or the turn taking sequence. The verbal and nonverbal reaction of the interlocutors can also be subject to manipulation. There are several taxonomies of types of power. For the purpose of this paper, let us choose the one suggested by Galbraith (1985), who divides power into ‘condign’ (based on force), ‘compensatory’ (achieved through the use of various resources) and ‘conditioned’ (being the result of persuasion). As will be shown, all types of power can be fought for in conversational humour. Among the sources of power, we can distinguish ‘personality’ (individuals), ‘property’ (the material resources of these individuals) and ‘organisational power’ (Galbraith 1985). Significantly, only ‘personality’ is a legitimate source of power in CH.

5. Games In order to analyse instances of conversational humour in terms of power games, let us choose to understand ‘games’ in their sociological aspect, as “sets of ulterior transactions, repetitive in nature, with a well-defined psychological payoff” (Berne 1977, 130). Thus defined, a game consists of a set of players, a set of moves (or strategies), and a specification of payoffs for each combination of strategies. Let us enrich this humanistic approach to games by two notions derived from the mathematical game theory (Dowding 1996). First of all, let us introduce the notion of ‘zero-sum games’, in which the total benefit to all players, for every combination of strategies, always adds to zero. Secondly, we are going to need the idea of a ‘Nash equilibrium’, which is

Conversational Humour as a Power Game

93

a situation when, after each player has chosen a strategy, no player can benefit by changing his or her strategy while the other players keep theirs unchanged.

6. CH as a power game The main aim of this paper is to demonstrate that each act of conversational humour carries with it an overt or covert goal connected with the producer’s social position. This position can be usurped, affirmed or augmented. The rationale for this phenomenon is that a witty contribution manages to promise (and usually achieve) greater relevance than an unmarked one, simply by the virtue of its unusual and stimulating form.

6.1 Friendly CH In all instances of CH discussed in section 3, there seems to be a ‘friendly’ employment of humour in the power game. It can be labelled as nonhostile, since there is the possibility of everyone winning in the power game. The game matrix below illustrates the different combinations of strategies, together with their pay-offs. Table 1. Power game matrix for friendly CH2

CH No CH

CH recognised 5,5 0,-5

No CH recognised -5,0 0,0

In the case of friendly CH, both parties – the initiator and the recipient(s) – gain the most in terms of power if humour is both produced and acknowledged (5,5). The speaker manages to assert her authority over the course of the conversation, while the hearer presents himself as an 2 The row headings indicate the actions of the initiator of CH (she can either produce a humorous utterance or avoid one and stick to unmarked discourse), while the column headings show the alternatives available for the audience (who can recognise the utterance as humorous and show it in their contribution, or fail to do so). In the grid, the pairs of numbers indicate the relative gain or loss of power in that particular speech exchange for the initiator and recipient respectively. Number 5 is taken as the highest possible gain of power, 0 indicates no change in status, while -5 is the highest loss of power. The ratings reflect the subjective judgment of the author and do not claim to be the ultimate ones.

Chapter Five

94

intelligent and sociable individual, accepting the joke as a desirable component of the conversation. If, in turn, the hearer fails to respond to the humorous cue, then the initiator suffers a serious face loss and the conversation slips out of her control. The audience have managed to assert their independence and desire not to be imposed on, but the positive influence on status is not as great as in the first case. The lower row, presented just for the sake of completion, shows the case of regular discourse, when no humour is produced. Standard conversation triggers no shifts in power status for either party, except for a mistaken recognition of CH in case of absence thereof, when the recipient shows he is not able to grasp the ongoing events.3

6.2 The guru effect On the borderline between friendly and hostile uses of CH in the power game, we could situate the employment of the ‘guru effect’, defined by Sperber (2005) as instances of employing obscurity in order to hint at a deeper meaning which, allegedly, would be impossible to convey in simpler words. In order to exploit the effect, the speaker should already have a position of power that she would be using (or abusing) in this way. In application to conversational humour, we could claim the guru effect to consist in exploiting one’s power position in order to assert the standards of funniness. We have failed to find instances of actual humorous employment of the guru effect in the Santa Barbara Corpus of Spoken American English, but apparently we could point at an instance of a misfire, or a failed employment of the effect. In the example below, a group of male friends are talking about a male dancer whom Miles met in a club. (5) 1. MILES: 2. JAMIE: 3. MILES:

He danced with a friend of @mine. Yeah, what happened. [But she didn’t wanna] dance with him anymore, I said, boy, he just wanted to get close to me. 4. HAROLD: Oo. 5. JAMIE: Oo. 6. PETE: [Mm=]. 3

Unless stated otherwise, the pronoun she is used when referring to the speaker/initiator, and he for the hearer/recipient.

Conversational Humour as a Power Game

95

7. MILES: 8. HAROLD: 9. MILES: 10. PETE: 11. JAMIE: 12. PETE: 13. MILES:

[I said, well, can’t blame him]. [Oba] ... Oba. @[@@@] [yeah]. Gross Mm. But I’m sure those guys get a lot of attention from women. They travel all over the world. I’m sure a lot of women throw themselves at em, so that’s what they expect from women. 14. HAROLD: A lot of groupies. 15. MILES: Yeah. ... Fringe benefits. 16. PETE: Hm. 17. MILES: @@@ 18. JAMIE: The only fringe benefit of being a dancer, probably. 19. HAROLD: Well in San Francisco, 20. MILES: But then again, 21. HAROLD: @you never know if it’s a benefit or not, either. 22. JAMIE: That’s true. 23. PETE: @ 24. MILES: @ (SBC)

In this example, in line 3 Miles introduces a remark that he considers conducive to laughter – about a male dancer who did not want to dance with Miles’ female friend, but instead went for Miles himself. Miles probably believed that it was funny that the dancer should prefer him to a woman. Since there is lack of recognition of the intended humour, Miles reintroduces the topic in line 7, implying that he is so handsome and attractive that even male dancers prefer him to women. Still failing to evoke a positive reaction, he nervously chuckles (line 9) and then makes a matter-of-fact remark about the dancers – that they get a lot of attention from women (line 13). In line 14, Harold politely reformulates a part of Miles’ contribution, which leads to a shift of Miles’ efforts – now he starts trying to make the aspect of groupies funny (line 15), and again embarrassingly laughs at the joke himself (line 17). It seems that the interlocutors do not even begin to perceive the topic as amusing until line 18, when Jamie jokingly refers to the groupies as the only fringe benefit of being a dancer. In line 21, Harold joins in, by questioning the positive nature of such a benefit. Only here does he start laughing, while Pete shows signs of amusement only in line 23.

96

Chapter Five

What we can see in this example is a desperate attempt of Miles’ to make the other conversationalists laugh at his initial remark. Although not taken up, he does not get discouraged, but keeps reformulating the stimulus, elaborating on it, or simply laughing himself in order to produce a humour-triggering cue. To an outside observer, his determination could be explained in two different ways. First of all, he could be seen as anxiously seeking confirmation of belonging to the group, looking for a positive token of recognizing his positive face by laughing at his joke. This way, we could see Miles as a shy individual striving for recognition. To a racially-biased sociologist, the fact that Miles is the only Afro-American in the group might confirm this explanation. The second option, however, is much more intriguing. Miles appears to be using – most likely subconsciously – a variation on the guru effect. He persists in presenting a topic as funny, even though the others keep rejecting its humorous nature. By repeatedly bringing it back, he appears to be affirming that he does indeed consider it funny, and that his position in the group allows him to get credit for the fact. On the nonverbal, social level, his communicated message might be: ‘Hey, guys, I’m your friend and I’m telling you this is funny, I’m sure it’s funny, so why don’t you believe me?’ Therefore, what we would thus encounter here is not a typical case of the guru effect, involving obscurity, but a variation on it – using one’s social position to warrant a certain reception of one’s words. What is more, the action seems unsuccessful, probably because of Miles’ lack of position in the group that would be actually strong enough to make him such a guru. Nevertheless, he does try to show his power among the friends by introducing a humorous remark, and the act is done in a way that seems entirely benign on the surface. Underneath, however, abusing the effect – or trying to subvert its legitimacy – can trigger serious changes in the power structure within the group. This places employing the guru effect on the borderline between friendly and hostile applications of humour in the power game.

6.3 Hostile CH In the hostile employment of conversational humour in the interactional power game, the primary purpose of the speakers seems to be not humour, but power itself. Quite often, they even refrain from pretending to be innocently joking, openly biting the conversationalists. This can be

Conversational Humour as a Power Game

97

achieved by manifesting a superior position, defining acceptable topics of the exchange, or setting formal standards for the conversation. 6.3.1 Manifesting a superior position First of all, linguistic humour can be used to justify an extralinguistic situation which would otherwise seem unacceptable. Apparent injustice may be presented in a funny framework, and real-life superiority can be strengthened by joking dominance. In the example below, a married couple, Marylin and Pete, are preparing dinner together. Suddenly, Pete drops a pepper down his wife’s blouse. (6) 1.

MARILYN:

2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

PETE: MARILYN: ROY: MARILYN: ROY: MARILYN: ROY: MARILYN: ROY: MARILYN:

@ God damn it, what’d you do? You @ son of a bitch @. @@@@@@ @@@@@ I threw a green pepper down your blouse. You threw a green pepper down my shirt. I thought it was funny. Hilarious. I’m laughing. @ @@@ I got quite a giggle out of it. @ (SBC)

In this situation, Marylin is a victim of what we might (perhaps jokingly) label as mild household violence. Her husband exerts power on her in two ways: first of all, the purely situational abuse of throwing a vegetable down her shirt asserts his physical control over the scene. Secondly, he disparages Marylin’s attempts at resistance by ridiculing them. Although the whole situation might seem to a casual observer to be reciprocal and consensual playfulness, in the recording we can hear a trace of irritation in Marylin’s protests. Although she accepts the imposed play frame by joining in the laughter, it seems obvious that her husband’s power is exerted on her by his defining what is acceptable as a topic of humour and what is not. Improper behaviour is made proper by being presented as funny.

Chapter Five

98

6.3.2 Defining topics Another use of conversational humour is also connected with the topic of the speech exchange, but this time the situation is as if reversed: the topic of a serious conversation is defined in a jocular way. This strategy is used as a way of defining the subject of the talk, which seems more socially acceptable than doing it blatantly on the record. In the following dialogue, another married couple, Pamela and Darryl, are having a conversation in bed before going to sleep. Pamela is in a pensive mood and starts reminiscing about a hard period in her life, when her marriage with Darryl helped her survive emotionally. (7) 1. PAMELA: what was hell in that ... that marriage became, became a way out for me. It was the flip side. It’s like sometimes you go through things, and you come out the other side of them, you come out so much better. And if I hadn’t had that, if I hadn’t had -2. DARRYL: It’s not the way with food. 3. PAMELA: What do you mean. 4. DARRYL: What goes in one way, 5. PAMELA: @@@@@ 6. DARRYL: @ doesn’t come out @@@@@@@@@@@. 7. PAMELA: @ Kay @ (SBC) What could have developed into a very deep and serious conversation about the nature of the relationship between the two people, and what could have touched upon issues vital to both of them, is interrupted by a down-to-earth, or even obscene, remark made by Darryl. By comparing Pamela’s psychological trials and tribulations with the workings of the digestive system, Darryl shows an utmost lack of respect for his wife’s feelings, and complete lack of interest in listening to her confessions. More importantly, since he does it in a witty, humorous way, he can always reject the implicature of denigrating his wife. Should Pamela have reacted in a different way, e.g. by demanding respect and attention, she would have been ridiculed as someone lacking a sense of humour. Therefore, by making an allegedly funny remark, Darryl not only enforces a change of topic, but also cunningly presents the shift as if it was approved by both parties involved.

Conversational Humour as a Power Game

99

6.3.3 Setting formal standards While the two previous hostile uses of conversational humour concerned the content of the speech exchanges, humour can also be employed to set formal standards on the interlocutors’ contributions. In the situation presented below, Miles is making a remark about some tribal masks hanging on the wall in Harold and Jamie’s house. (8) 1. 2. 3. 4. 5. 6. 7. 8. 9.

MILES: HAROLD: JAMIE: HAROLD: MILES: PETE: JAMIE: MILES: JAMIE:

Those two top masks, there. Yep. @@@ We need a verb. I guess the ones below are women, [@@@@@] [@@@] Is that why they look so different? [@@ ] (SBC)

In line 1, Miles introduces the topic of the masks, drawing the others’ attention to them. However, the initial cooperative token of interest (Yep in line 2) is quickly withdrawn and replaced by Harold’s factual remark about the faulty grammar of Miles’ utterance (line 4). The joke is very quickly grasped by Pete and Jamie, who laugh at it (lines 6 and 7), while the butt of the joke, Miles, refuses to acknowledge it and continues his matter-of-fact comment on the masks (lines 5 and 8). As a result, what we are faced with is an example of a possible fight for power in which both sides are active. Harold introduces humour to ridicule his interlocutor’s inadequate command of English and to set a requirement on full sentences in the conversation, while Miles openly resists being imposed on and carries on the serious discourse. Interestingly enough, the situation here stands at the opposite end of the spectrum than example 5, where it was Miles who was trying to enforce the non-bona fide mode of communication, while all the others were resisting. When considered from a broader sociological perspective, it might be interesting to investigate how the employment of CH changes throughout a longer conversation, and to seek a connection to the internal structure of the social group and the roles individuals take on at different points in time.

Chapter Five

100

6.3.4 Hostile CH: the matrix Table 2 presents the power game matrix for hostile conversational humour. As we can see, the recipient is in a no-win situation. If he recognises the humour as hostile, then he implicitly admits having lost in the power game: the initiator manages to manifest her superior position, change the topic or set formal requirements, while the hearer meekly accepts the manipulation. If the recipient mistakes the instance of hostile humour for a friendly one, the better for the initiator, since she still can maintain the false pretences of kindness and amiability, while at the same time exerting her power. The situation also does little social damage to the recipient, who can behave as if nothing inappropriate has happened and therefore apparently maintain his power status. The only situation when the speaker’s devious plan can go amiss is when the hostile humour falls flat, i.e. when it is either not recognised at all or when it is recognised as open hostility. In the latter case, the interlocutors are bound to be on their guard, becoming more likely to defect future instances of attempted hostile CH on the part of the speaker. Therefore, while hostile joking can be a powerful tool of weakening the social position of the others, it is also extremely risky for the person employing it. Table 2. Power game matrix for hostile CH

Hostile CH No CH

CH recognised as hostile 5,-5

CH recognised as friendly 5,0

No CH recognised -5,0

-

-

0,0

7. Concluding remarks Having analysed the practical aspects of a few example conversations, let us now deal with conversational humour in terms of the game theory. It would be interesting to find out what (if any) mathematical logic is inherent in the playful use of language, what types of power are related to what types of CH, and what might be the underlying cognitive mechanisms determining the differentiation into the three types of humour.

Conversational Humour as a Power Game

101

7.1 Non-zero-sum game First of all, it seems that all instances of CH (both friendly and hostile) are non-zero-sum games, which means that the total benefit to all players does not add up to zero. The only situations when the gains and losses cancel each other out are the borderline instances of regular conversation (no CH produced, no CH mistakenly recognised), and hostile CH recognised as hostile (when the initiator wins clearly at the expense of her audience). Apparently, in hostile CH more can be lost than gained. The most socially desirable pattern seems to be acknowledged friendly CH, in which everyone wins. The speaker manages to introduce the topic as funny and to affirm her social position, while the hearers assert their authority and independence in evaluating a humorous remark and present themselves as sociable individuals. The great potential for win-win situations can possibly be one of the factors making CH such a common occurrence in daily interaction.

7.2 Lack of Nash equlibria The aforementioned potential for win-win situations is augmented by considerable independence of the interlocutors. In friendly CH, once the initiator has chosen to produce a funny remark, the recipient can adjust his strategy to either win or lose in the game. Similarly, in hostile CH the audience have the choice of suffering losses or avoiding being hurt. As it seems, conversational humour is a game with no Nash equilibria: once each player has chosen a strategy, the others can benefit by changing theirs. Paradoxically perhaps, this makes CH an activity that is more individual than social, since one’s gain or loss is not inextricably tied with the moves of the other players, but is rather a matter of individual choice of game strategy.

7.3 Types of power It can be seen that different types of CH make use of different types of power. Friendly humour employs compensatory power: the initiator uses resources, such as her own wittiness or the funny potential of real-life situations, to carry out her laughter-triggering plan. On the other hand, humour involving the guru effect seems to feed on conditioned power, which is the result of persuasion. The instigator attempts to persuade the audience that her position in the group warrants

102

Chapter Five

appreciation of the jokes she cracks. Her authority in asserting a mirthevoking quality of an utterance is to be considered ultimate. Hostile CH, in turn, being based on force, thrives on condign power. The speaker aggressively imposes her judgment and decisions on the audience. The audience either resist being imposed on, or become intimidated and lose in the power struggle. Whatever the strategy, a conspicuous battle for domination is visible.

7.4 Source of differences The ultimate question of this paper considers the source of differences between the three types of conversational humour. The answer may be sought in the theory of relevance, or – more specifically – in the ways of manipulating the assumptions entertained by the audience. As a result, the three types of humour achieve optimal relevance in three different ways. Friendly humour consists in getting power in a cooperative way, which is performed by the strengthening of assumptions. The audience already believe that they share cognitive environments with the speaker; they are just reminded of the fact. They already know that they form one social group, they know who belongs and who does not belong. They are already convinced that outsiders are inferior. Collaboratively reaffirming these facts in a humorous way lets everyone win in the humour-based power game. On the other hand, hostile CH is oriented towards getting power competitively. It is accomplished by attempting to delete the assumptions entertained by the audience, or by creating new assumptions. The audience’s opinions about their social position and being esteemed by the initiator are heavily undermined. They are coerced into accepting a different topic of the conversation, a different way of talking, or even a different microcosmic representation of social structure and acceptability of certain actions. Their independence in making assumptions is threatened. Humour involving the guru effect, in turn, seems to be a tricky combination of the previous two. On the surface, the producer is using the existing beliefs of the hearers – beliefs concerning her great intelligence and authority. Underneath, however, we have seen how this false pretence is used to actually challenge the hearers’ system of beliefs and to create a new opinion about the initiator. Mixing the two strategies, this type of humour is particularly risky, but might yield great benefits to the speaker if mastered.

Conversational Humour as a Power Game

103

7.5 Conclusion On the whole, it seems that conversational humour is quite an efficient tool in the power game inherent in daily interaction between humans. Power goals can be achieved by friendly, hostile or persuasive methods. The cognitive rationale behind the mechanisms of these methods seems to be connected to the operations performed on the assumptions of the hearers, and thus on the different ways of achieving optimal relevance. The findings presented in this paper have no claims to being final or decisive. The categorisation is largely based on the author’s social and linguistic intuitions. The numerical judgments in the matrix tables are highly subjective. A more objective method of assessing the effects of the power game has yet to be found, if feasible at all. What is more, the assumptions of this paper have not yet been tested against a wider corpus of data. There probably exist more functions of CH, as there might be more mechanisms connected with their application. These are just some of the issues open to further research and investigation within the fascinating realm of conversational humour and the human struggle for power.

References Berne, Eric. 1977. Beyond games and scripts. Selections from his major writings. Edited by C. M. Steiner, and C. Kerr. New York: Grove Press. Boxer, Diana, and Florencia Cortés-Conde. 1997. From bonding to biting: Conversational joking and identity display. Journal of Pragmatics 27: 275-294. Dowding, Keith. 1996. Power. Minneapolis: University of Minnesota Press. Galbraith, John K. 1985. An anatomy of power. Boston: Houghton Mifflin. Hay, Jennifer. 2000. Functions of humor in the conversations of men and women. Journal of Pragmatics 32: 709-742. Holmes, Janet, and Meredith Marra. 2002. Having a laugh at work: How humor contributes to workplace culture. Journal of Pragmatics 34: 1683-1710. Sperber, Dan 2005. The guru effect. Manuscript. (8 March 2006) Sperber, Dan, and Deirdre Wilson. 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell.

104

Chapter Five

Weber, Max. 1947. The theory of social and economic organisation. Greenwood Pub Group. Wilson, Deirdre, and Dan Sperber. 2002. Relevance Theory. In The handbook of pragmatics, edited by L. Horn, and G. Ward. Oxford: Blackwell.

CHAPTER SIX BANTER – AN ATTEMPT AT A RELEVANCETHEORETIC ACCOUNT EWA KAROLINA NOWIK

1. Introduction The purpose of this paper is, firstly, to define the linguistic phenomenon of banter and, secondly, to provide its theoretical account in terms of the solutions proposed by Relevance Theory (Sperber and Wilson [1986] 1995). The first section of the paper is devoted to a brief characteristic of the phenomenon of banter as a mechanism which uses impoliteness in order to achieve the effect of politeness. I will attempt to define banter in relation to the concepts such as politeness and impoliteness. Having defined banter within a particular perspective on politeness (Watts 2003) and, following Victoria Escandell-Vidal (2004), I will proceed to consider how the social perspective, which is crucial for the definition of banter, can be integrated with the relevance-theoretic cognitive perspective on communication. Finally, I will attempt to propose certain theoretical solutions within the framework provided by Relevance Theory.

2. (Im)politeness Banter is a phenomenon situated at the intersection of the notions of politeness and impoliteness. The main difficulty connected with its classification seems to lie in the fact that it encompasses both sides of the polite-impolite divide. I understand by banter any utterance which uses impoliteness in order to achieve the effect of politeness. Thus greetings such as Come in, you silly bugger,1 when uttered to a good friend coming late 1

An example constructed on the basis of the example cited by Culpeper (1996, 352). In the original example Culpeper uses only the phrase you silly bugger.

106

Chapter Six

to a party at our house, would be a typical example of banter. Though being impolite on the surface, such a remark would probably be interpreted as a friendly invitation. The main purpose of banter is to mark familiarity or even intimacy between the interlocutors and thus to emphasise solidarity and friendly relations between the speakers. The preliminary definition of banter as offered above employs the concepts of ‘politeness’ and ‘impoliteness’. These notions have been widely discussed within the literature on the subject and have raised a significant amount of controversy. The definitions proposed within politeness theorising include the conceptualisation of (im)politeness in terms of facework, i.e. showing awareness of the hearer’s face (Brown and Levinson 1987), universal principles (Leech 1983) or terms of conversational contract (Fraser 1990). There have also been attempts at cultural definitions of politeness (e.g. Blum-Kulka 1992). In spite of an enormous amount of theoretical and empirical research on politeness, there is still no agreement on what actually politeness is and how it should be conceptualised. The theories devised so far have been criticised for a number of faults, for instance, for their excessive focus on one of the interlocutors (the speaker or the hearer) only. Another major objection that has been recently voiced (Eelen 2001) concerns the common neglect of one side of the (im)politeness continuum: the models account either for politeness or impoliteness only. And it is usually impoliteness that is either totally neglected or perceived as derivative of and subordinate to politeness. Moreover, the authors too often fail to notice that politeness and impoliteness form a continuum and that neither pole of this continuum can be properly described without reference to the other. Another cardinal sin that is commonly committed is the claim of universality: the norms/principles, etc. proposed in the models are believed to be universal for all kinds of polite behaviour in all societies and cultures. That claim is straightforwardly contradicted by the abundant evidence of cultural diversity of (im)polite behaviour. What is also commonly missing is the social aspect of the phenomenon, especially the way in which the society shapes the norms and conventions determining what counts as polite or impolite. Finally, the theories tend to overlook the dynamic interactivity of (im)politeness phenomena which should be viewed as a process-in-action and not as a static entity resulting from the application of rules or strategies. As a result, (im)politeness fails to be perceived as something created in between the speaker and the hearer and open to evaluation. As it has been argued above, the notion of (im)politeness is highly problematic in many respects. There is, however, one conceptualisation of

Banter – an Attempt at a Relevance-Theoretic Account

107

(im)politeness that seems to avoid the problems described above. It is the social model of politeness devised by Richard Watts (2003). Watts builds his model of (im)politeness upon Bourdieu’s (1990) theory of social practice. The key notion adapted from Bourdieu is the idea of the ‘habitus’. The habitus is understood as a set of dispositions to act in specific ways, which is shaped by the social structures internalised by an individual during the process of socialisation. Our habitus, which includes institutionalised forms of behaviour, rights and obligations of the individuals and the power structures, generates our cognitive practices and ensures their ‘correctness’: The habitus, a product of history, produces individual and collective practices – more history – in accordance with the schemes generated by history. It ensures active presence of past experiences, which deposited in each organism in the form of schemes of perception, thought and action, tend to guarantee the ‘correctness’ of practices and their constancy over time. (Bourdieu 1990, 54)

Social practice depends on the habitus combined with our social, cultural and material capital and the social field where the practice takes place. Watts plants this theory onto the linguistic field and proposes a subtype of the habitus: the linguistic habitus which comprises the ways of speaking that we believe to be appropriate in a given sociocultural context. As a result, our linguistic practice depends on our linguistic habitus (the knowledge which linguistic structures are appropriate in a specific type of interaction), our linguistic capital (i.e. our command of language) and the linguistic field (the type of interaction we are involved in). Politeness2 is understood as behaviour that is consistent with our habitus in a given situational context. Behaviour which contradicts the dispositions of the habitus is open to the interpretation of impoliteness. The value of Watts’s theory lies in the fact that it allows for the account of the universal phenomenon of (im)politeness in local terms. There are no global universal norms specifying what is polite and what is not: whether an utterance is judged as polite or impolite depends on the habitus of a given person. Moreover, utterances are not polite or impolite by themselves but 2

In Watts’s terminology this is in fact ‘politic’ and not ‘politeness’ behaviour. Watts (1992) is the father of the distinction between politeness and politic behaviour. What other theories name ‘politeness’, Watts calls ‘politic behaviour’. Thus behaviour that does not contradict the dispositions of the habitus is ‘politic behaviour’. Behaviour that exceeds these dispositions (being too politic for the occasion) is called ‘politeness behaviour’. Due to the space limits this distinction will not be analysed here in detail.

108

Chapter Six

open to the evaluation of (im)politeness. Such a conceptualisation of the politeness/impoliteness continuum accommodates and accounts for both its extremes.

3. Defining banter Having adopted a particular perspective on (im)politeness, it is now possible to offer a more detailed definition of the sociolinguistic phenomenon of banter. Employing the definition of (im)politeness proposed by Watts, I define banter as behaviour which consists in saying something that, firstly, is obviously not serious and, secondly, that would be obviously impolite in a non-intimate context.3 By impolite I understand, after Watts (2003), behaviour that is judged by the hearer, and/or the speaker as insufficient in relation to the dispositions of the habitus as determined by the situational features of the context. To the conditions sine qua non of the proper use of banter, I also include the situational context of small social distance, the balance of power and the mutual liking between the interlocutors. The aim of banter, I believe, is to demonstrate solidarity and to emphasise intimacy and mutual liking between the interlocutors. Thus, within the adopted view of (im)politeness, the purpose of banter is not to achieve politeness (i.e. not to abide by the dispositions of the habitus), but to achieve a particular effect of confirming and strengthening the bond between the interactants.

4. The social and the cognitive approach The notion of the habitus brings to the discussion a social perspective on language. Relevance Theory has often been accused of being asocial. It has been claimed that it deals with communication as a cognitive process and is not interested in its social context – since Relevance Theory is a cognitive approach. Yet, the notion of social conventions and norms as influencing our use and understanding of a language is very powerful and intuitively plausible. This is especially visible in the discussion of issues such as politeness phenomena. But are the social and the cognitive perspectives truly mutually exclusive? 3

The starting point for this definition was Leech’s (1983, 144) definition of banter. There are numerous reasons for the inadequacy of Leech’s definition but, due to the space limit, I will refrain from their discussion in this article. For a detailed discussion and criticism, see Nowik (2005).

Banter – an Attempt at a Relevance-Theoretic Account

109

It has been traditionally assumed that there are two separate and distinct ways of doing pragmatics: the social-cultural way and the cognitive way (Escandell-Vidal 2004, 2-5). The focus of the socio-cultural approaches was on external factors determining language use, such as cultural or politeness norms. Cognitive approaches, on the other hand, concentrated on the identification of the internal factors – the cognitive bases of communication, most often expressed in terms of principles. Norms were supposed to be culture-specific, whereas principles – universal. An interesting suggestion concerning this state of affairs has been put forward by Victoria Escandell-Vidal (2004), who reconsiders the abovementioned division and suggests that the two approaches should be viewed not only as not mutually exclusive but complementary. Her claim is that the social and the cognitive approaches are not two conflicting views but rather two different components of a larger system. Neither of them, it is claimed, can be reduced to the other.

5. An integrated theory of pragmatics The first move toward an integrated theory of pragmatics is the application of the commonly acknowledged distinction between representation and computation to the distinction between norms and principles. EscandellVidal (2004, 2) claims that norms are simply representations produced on the basis of generalisations of the observed behaviour. Principles, on the other hand, are computations – causal mechanical explanations, general laws explaining the workings of the system. Principles are very abstract and thus difficult to discover and verbalise. Adopting the relevance-theoretic perspective, the author sees two such principles: the Cognitive and the Communicative Principles of Relevance: 1) 2)

Human cognition aims at the maximisation of relevance. Every act of ostensive communication communicates a presumption of its own optimal relevance. (Sperber and Wilson [1986] 1995, 260)

These two principles constrain the way our mind works and, specifically, the workings of the processes of inference and decoding, which are crucial for human communication (Sperber and Wilson [1986] 1995). The question that arises is if and how Relevance Theory can accommodate the notion of norms.

Chapter Six

110

5.1 The inference module Relevance Theory assumes that the mind is modular, and that the process of inferencing takes place within the inference module, which is a processing device, a universal computational mechanism that works on representations. The representations that enter the module come from decoding and from memory. The task of the inference module is to combine information coming from decoding with accessible contextual assumptions in order to yield contextual effects. The operation of the module as well as its input and output conditions are constrained by the relevance-theoretic comprehension procedure (the so-called criterion of consistency with the Principle of Relevance): a) b)

Follow the path of least effort in computing cognitive effects. In particular, test interpretive hypotheses (disambiguations, reference resolutions, implicature, etc.) in order of accessibility. Stop when your expectations of relevance are satisfied. (Sperber and Wilson 2002, 18)

The input condition for the module is accessibility – apart from the assumptions coming from decoding, only the assumptions that are salient, i.e. activated by the context and retrieved from memory, are allowed into the module. Thus, clause a) of the relevance-theoretic comprehension procedure limits the workings of the module. Clause b), on the other hand, sets the output condition – representations produced by the module are the ones that satisfy the expectations of relevance. According to Escandell-Vidal (2004, 8-9), the workings of the module can be summarised as follows – there are two types of assumptions that enter the inference module: assumptions coming from decoding and assumptions retrieved from memory, activated by the situational context. Both kinds of assumptions are stored in the working memory, and enter the inference module on the condition that they are sufficiently salient and thus accessible. In the module, the assumptions undergo inference processes, and when the expected level of relevance is reached, the system produces an interpretation. If the level of relevance is not satisfactory, the processing starts anew – additional assumptions are added and the representation undergoes consecutive inference processes until the expectations of relevance are satisfied and an interpretation is delivered.

Banter – an Attempt at a Relevance-Theoretic Account

111

5.2 The social module More interestingly, however, Escandell-Vidal (2004, 10-12) proposes the existence of another module – the so-called ‘social module’.4 The social module, contrary to what has been assumed before, involves both: computation and representations. Computation is present in the social categorisation mechanism – our ability to extract generalisations on the basis of the regularly observed samples of behaviour. These generalisations have the form of norms,5 which can be stored in the longterm memory and brought to the working memory when activated by the context of the interaction. This is how the system fulfils the long-term function of the social module – it builds and updates the stock of representations (norms) concerning socially adequate behaviour. The mechanism itself is universal and common to all people, but the norms are culture-dependent and derived from the experience of a particular individual. The social categorisation system, it is postulated, is sensitive to features such as power, distance, age, sex, degree of imposition, etc. But the module has not only the long-term task but also a short-term task to fulfil. The latter consists in an on-line analysis of the current situation, including the incoming pieces of behaviour, which in turn, causes an appropriate readjustment of the communicative activity. When set in motion, the module is believed to be working in the following manner – the observed behaviour enters the social categorisation system, which, in turn, performs operations of two types: firstly, it extracts 4

Escandell-Vidal (1998, 47) strongly argues in favour of the claim that the faculty of social cognition should be a separate module because of its specific character – it invokes elements and establishes relationships with no direct perceptual motivation. Following Jackendoff, (1992), she claims that the main task of such a module would be to establish a coherent picture of self in the society. 5 Norms would, no doubt, constitute a part of specific knowledge, which enables us to interpret and participate in the events we have been through many times – it is acquired and based on previous experience. Specific knowledge may be organised into frames (or scripts, scenarios, schemas, etc. – Escandell-Vidal is aware of the differences between the definitions supplied by various authors; however, for the sake of simplicity, she decides to ignore the differences and decides to use the term ‘frame’). Frame is defined as “a structured set of organised knowledge” (Escandell-Vidal 1996, 634). It is worth noting that, even though it has not been precisely defined, this notion is not foreign to Relevance Theory – it has been employed under the label of ‘scenario’ or ‘cognitive script’ to describe bigger chunks of organised information, stored in the encyclopaedic entries of the concepts encoded or activated by an utterance (Sperber and Wilson [1986] 1995, 87-8; Wilson 1994; Carston 2002, 226-227).

112

Chapter Six

social regularities and stores them in our long-term memory in the form of norms. Secondly, it analyses and categorises the incoming pieces of behaviour in accordance with the norms already stored in the long-term memory. The behaviour is analysed from the angle of features such as power, distance, sex, age, etc. The resulting evaluation influences the forthcoming linguistic activity or the interpretation produced by the inference module.

5.3 An integration The two modules, Escandell-Vidal (2004, 12-13) argues, have to be integrated in the common task of producing the correct interpretation of a given utterance. The integration works two-ways: the representations established due to the workings of the social module have a bearing on the interpretation delivered by the inference module. And the interpretations obtained thanks to the operations of the inference module undergo social categorisation and evaluation within the social module. The integration of the modules works as follows: a given communicative situation is labelled by the social module in accordance with situational factors such as age, power, position etc. As a result, certain representations (norms) are activated and retrieved from the longterm memory. They enter the working memory and, consequently, raise certain expectations. The incoming pieces of communicative behaviour provide a number of assumptions coming from decoding, as well as the assumptions highlighted by the situational context and retrieved from memory. All these assumptions, if sufficiently salient, are processed in the inference module, and an interpretation of an utterance is delivered. If the interpretation produced contravenes the expectations raised by ‘social’ representations, the utterance is judged strange or impolite. At the same time, the representations (interpretations) obtained due to the workings of the inference module are analysed as new samples of behaviour and undergo social categorisation, perhaps serving to establish new norms or simply to confirm and strengthen the existing ones.

6. Banter as a script The concept of the habitus6 as a product of past experiences and, at the same time, the generator of the practices seems to fit perfectly into the 6 Escandell-Vidal (2004, 11) cites Bourdieu’s notion of the habitus in support of her claim that internalised knowledge of practices is acquired from the practices

Banter – an Attempt at a Relevance-Theoretic Account

113

mental architecture as described above. It would be a perfect candidate for the social module. And what is the role of the habitus in generating the meaning connected with the use of banter? I would argue that the interpretation of banter depends on the presence of what Relevance Theory has called a ‘script’ or ‘scenario’ (Sperber and Wilson [1986] 1995, 88; Wilson 1993) and what Escandell-Vidal (1996, 234) calls a ‘frame’, contained (or not) in our habitus. The script should be stored in the habitus and not in an encyclopaedic entry, because banter cannot be tied to any particular word or linguistic expression and its corresponding concept. An argument in favour of the script-based treatment of banter is also the complexity of the phenomenon in question. Let me start with a brief consideration of banter in terms of the speaker’s meaning by returning to the example cited at the beginning of this paper: imagine that you have come to a party at your friend’s house and you are greeted with the words: Come in, you silly bugger. In ordinary circumstances the hearer would treat this remark as a kind of joke, verbal play and s/he would not take offence in spite of its clearly offensive character. In this example we can distinguish at least three layers of meaning: the explicit meaning, which probably boils down to being invited inside (Come in), and assigning the reference of the silly bugger to ourselves. Then, we have the apparent impoliteness of the remark – after all, in a neutral context a silly bugger is not a positive expression. The impoliteness is the result of our evaluation of the representation obtained from linguistic decoding and inference when set against the norms stored in our habitus – these norms tell us that one should not evaluate others negatively in public. Having rejected the surface impoliteness, we arrive at the final interpretation of the utterance as an expression of solidarity and sympathy. The inferential route as presented above would definitely be too costly to go through each time the hearer encounters a banter remark. The processing cost, however, would be significantly lessened if we assumed that the pattern of inferencing is stored as a script. I do not want to claim here that, due to the presence of the script, all these layers of meaning disappear; they are distinguishable upon reflection. What I want to say is that the processing becomes faster, automatic and more effective. Therefore, it yields the desired cognitive effect at a lesser cost. Obviously, in order to have such a script in his/her inventory, the hearer would need to have had a history of previous experiences with banter. As a result of the repeated exposure to that type of communicative behaviour, the social themselves as a result of generalisation on observed behaviour, but she does not find it useful as an integral part of her framework.

114

Chapter Six

faculty would extract the general pattern of interpretation (impoliteness Æ solidarity, context: small distance, the balance of power and mutual affect) and preserve it as a norm. The derivation of the solidarity interpretation would be triggered by the apparent impoliteness in the context of the utterance. Such an approach to banter has significant implications. A number of authors have observed (e.g. Fraser 1990; Watts 1992; Escandell-Vidal 1996; Jary 1998) that behaviour that is compatible with a norm is anticipated and, therefore, goes unnoticed. It is deviation from the norm that captures our attention. That is why if both the speaker and the hearer happen to have the script for banter, banter remarks will largely go unnoticed. But if the hearer lacks the script in question, s/he will be unable to reach the interpretation of solidarity and intimacy and will stop the processing at the level of impoliteness. Obviously, this may lead to an unintended misunderstanding and certain unwanted implications of rudeness or even verbal aggression. Another consequence of treating banter as a script is its increased culture-dependence. Being based on the notion of (im)politeness, banter is culture-specific because, as has been stated above, (im)politeness itself is culture-specific. But banter as a script becomes even more linked to the notion of culture. For instance, the sufficient degree of familiarity or the required degree of impoliteness may vary, depending on a given culture.7 That is why, in order not to mistake banter for straightforward impoliteness, one needs to be well-rooted in a given community.

7. Explicit/implicit The final question that I would like to tackle in this paper concerns the position of banter in terms of the types of meaning it involves, in particular with regard to the explicit/implicit distinction as developed within Relevance Theory (Sperber and Wilson [1986] 1995; Carston 2002). Let us return to our rather worn-out example: Come in, you silly bugger. What we can definitely say is that at the level of the proposition expressed it is an invitation for the hearer to enter the house, said in the imperative mood. In order to reach this interpretation we need to decode and enrich the utterance. The possible processes that are in action here, apart from the straightforward linguistic decoding, involve the reference 7

Basing on my personal experience, I would claim that Polish banter is less aggressive and requires greater familiarity between the interactants than is the case with its British equivalent.

Banter – an Attempt at a Relevance-Theoretic Account

115

assignment of the noun phrase silly bugger, the ad-hoc concept construction performed over the expression bugger and the enrichment of the verb phrase come in. Thus, the hearer identifies him/herself as the referent of the expressions you and silly bugger. The concept BUGGER, which is an offensive term denoting a stupid or annoying person, is broadened to mean a person whom you like but who has done something slightly annoying or perhaps unwise. Finally, the verb phrase come in has to be enriched, so that we come in somewhere. As a result, we arrive at a representation such as the following: ‘Come in [to the house] you [the hearer] silly bugger*’.8 The meaning just presented is a straightforward development of the logical form, a base-level explicature – “an assumption communicated by an utterance U is explicit if and only if it is a development of a logical form encoded by U” (Sperber and Wilson [1986] 1995, 182). It is now recognised within Relevance Theory that an utterance typically has several explicatures; apart from the base-level explicature, it may have a number of higher-level explicatures (Wilson and Sperber 1993). A higher-level explicature is still a development of the logical form but it functions as a frame for a base-level explicature. It is constructed by embedding base-level explicatures under a higher-level description (a speech-act or propositional-attitude description). Linguistic phenomena believed to contribute to higher-level explicatures include e.g. mood or certain types of sentence and attitudinal adverbials (Wilson and Sperber 1993, 14; Blakemore 2002, 74-77; Carston 2002, 119-125, 128-132). In the case of our example, it is the imperative mood that will definitely be accommodated as a higher-level explicature. Moreover, since within the relevance-theoretic framework imperatives are interpreted as conveying the assumption that some states of affairs are both achievable and desirable to some extent to the speaker or the hearer, the possible higherlevel explicatures will include the following: (1) (2)

The speaker is telling the hearer to come into the house. It is desirable for the speaker or/and the hearer that the hearer enters the house.

Higher-level explicatures would also be the level to which the speech act of invitation should probably contribute: 8

The asterisk (*) is used to indicate that the concept contained in a given representation has been loosened (broadened) or strengthened (narrowed) and, as a result, is an ad-hoc concept. The notion of ad-hoc concept construction is discussed in detail in Carston (2002).

Chapter Six

116

(3)

The speaker invites the hearer to enter his/her house.

Since attitude description is also handled at the level of explicatures, we can assume that the (im)politeness of contributes to this very level of meaning. The speaker’s impolite, at least on the surface. The question that arises is (apparent) impoliteness gives rise to another separate explicature: (4)

higher-level the remark invitation is whether this higher-level

What the speaker said was impolite,

or whether it serves as a frame for the already higher-level explicature connected with the mood: (5)

The speaker told the hearer impolitely to come in,

and/or the speech act: (6)

The speaker impolitely [invited] the hearer to come in.

In the last two cases, the result would be a third-order representation: a base-level explicature embedded under a higher-level description which again would be embedded under another higher-level description. Finally, being guided by the search for relevance, and considering the contextual factors (small social distance, mutual liking and no obvious reasons for our interlocutor to offend us), we reject the surface impoliteness of the utterance and arrive at the final interpretation of the utterance as an expression of solidarity and sympathy: (7) By inviting the hearer to come in to the house in an apparently impolite manner the speaker wants to show solidarity and emphasise intimacy between him/her and the hearer. In this case, the higher-level explicature connected with the surface impoliteness of a banter remark is embedded into another higher-level explicature and definitely gives rise to at least third-level, and possibly even fourth-level, metarepresentation. Obviously, our cognitive processing is not linear. The final interpretation of familiarity and the resulting reinforcement of the bond between the interlocutors are reached by the mutual adjustment of the explicit and implicit content. We do not proceed step by step from the explicature, through the consecutive higher-level explicatures to the final

Banter – an Attempt at a Relevance-Theoretic Account

117

interpretation of solidarity; our interpretation is rather a matter of the online inference process. This process may be sped up due to the presence of the script for banter – the processing of banter by the hearers equipped with it becomes almost automatic and less costly than it might otherwise seem. .

8. Conclusions An account of banter is a demanding task. The combination of linguistic and social meaning it is intended to convey, its dependence on the widely disputed notion of (im)politeness, and, finally, its multiple layers of meaning do not make this task easier. That is why finding an adequate explanation of its mechanics is challenging but, at the same time, very rewarding. In this paper, I have tried to present a number of possible solutions connected with the phenomenon but, I must admit, they are only the tip of the iceberg. Apart from the combination of the social and the linguistic side of the phenomenon, explanation of the notions of politeness and impoliteness and an attempt at an account in terms of the explicit/implicit meaning, there is yet much work to be done. The issues of interest include, among others, the procedural/conceptual distinction and the attribution of intentions, but due to the space limits they have not been dealt with here.

References Blakemore, Diane. 2002. Relevance and linguistic meaning. The semantics and pragmatics of discourse markers. Cambridge: Cambridge University Press. Blum-Kulka, Shoshana. 1992. The metapragmatics of politeness in Israeli society. In Politeness in language: Studies in its history, theory and practice, edited by R. J. Watts, S. Ide, and K. Ehlich. Berlin: Mouton de Gruyter. Bourdieu, Pierre. 1990. The logic of practice. Cambridge: Polity Press. Brown, Penelope, and Stephen Levinson. 1987. Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Carston, Robyn. 2002. Thoughts and utterances. Oxford: Blackwell. Culpeper, Jonathan. 1996. Towards anatomy of impoliteness. Journal of Pragmatics 25: 349-367. Eelen, Gino. 2001. A critique of politeness theories. Manchester: St Jerome Publishing.

118

Chapter Six

Escandell-Vidal, Victoria. 1996. Towards a cognitive approach to politeness. Language Sciences 18: 629-650. —. 1998. Politeness: A relevant issue for Relevance Theory. Revista Alicanta de Estudios Ingleses 11: 45-57. —. 2004. Norms and principles. Putting social and cognitive pragmatics together. In Current trends in the pragmatics of Spanish, edited by R. Márquez-Reiter, and M. E. Placencia. Amsterdam: John Benjamins, a pre-print version.

(5 September 2006) Fraser, Bruce. 1990. Perspectives on politeness. Journal of Pragmatics 14, no. 2: 219-236. Jackendoff, Ray. 1992. Languages of the mind. Cambridge, MA: MIT Press. Jary, Mark. 1998. Relevance Theory and the communication of politeness. Journal of Pragmatics 30: 1-19. Leech, Geoffrey. 1983. Principles of pragmatics. Harmondsworth: Penguin. Nowik, Ewa K. 2005. Politeness of the impolite: Relevance Theory, politeness and banter. In Relevance Studies in Poland. Vol. II, edited by A. Korzeniowska, and M. Grzegorzewska. Warsaw: Institute of English Studies, University of Warsaw. Sperber, Dan. 1996. Explaining culture. A naturalistic approach. Oxford: Blackwell. Sperber, Dan, and Deirdre Wilson. [1986] 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. —. 2002. Pragmatics, modularity and mind-reading. Mind and Language 17, nos. 1-2: 3-23. Watts, Richard J. 1992. Linguistic politeness and politic verbal behaviour: Reconsidering claims for universality. In Politeness in language: Studies in its history, theory and practice, edited by R. J. Watts, S. Ide, and K. Ehlich. Berlin: Mouton de Gruyter. —. 2003. Politeness. Cambridge: Cambridge University Press. Wilson, Deirdre. 1994. Relevance and understanding. In Language and understanding, edited by G. Brown, K. Malmkjær, A. Pollit, and J. Williams. Oxford: Oxford University Press. Wilson, Deirdre, and Dan Sperber. 1993. Linguistic form and relevance. Lingua 90: 1-25.

CHAPTER SEVEN ON ECHOIC IMPERATIVES: A RELEVANCE-THEORETIC APPROACH1 KEIKO ABE

1. Introduction In Relevance Theory, when imperatives are used to communicate the speaker’s messages literally and seriously, their semantic structure is assumed as follows: [i] X regards Y as potential [ii] X regards Y as desirable to Z [X=Speaker, Y=Proposition, Z=Speaker or Hearer] (Wilson and Sperber 1988) The illocutionary acts performed by means of imperatives are categorised into a request-type and an advice-type depending upon whether Z is resolved as Speaker or Hearer. For example, Finish your breakfast as soon as possible can be understood either as an order or a piece of advice. When the hearer recognises that the speaker regards the proposition in the utterance desirable to the speaker, this is accepted as an order, and when the speaker regards the proposition desirable to the hearer, it is accepted as a piece of advice. 1

This paper is a revised version of my presentations at The 9th International Pragmatics Conference in Italy, 2005, and The 3rd Interpreting for Relevance Conference, in Poland, 2006. I’d like to express my gratitude to Professor Seiji Uchida for his constant encouragement and invaluable comments. I also thank Professor Deirdre Wilson for her insightful comments.

120

Chapter Seven

There is another use of imperatives which has not been studied sufficiently as such, although it is sometimes referred to in the course of discussion. Let us look at the following examples: (1) A: Pass me the salt. B: (critically) Pass you the salt! (Noh 2000, 93) (2) PETER: Can I open the window? MARY: Go ahead and let in some nice Arctic air. (Wilson and Sperber 1988, 90) (3) Context: Mary, seeing Peter about to throw a snowball, says threateningly. Go on. Throw it. Just you dare. (Wilson and Sperber 1988, 80) (4) That’s right. Misinterpret everything I say. (Blakemore 1992, 111) In this paper, these utterances are termed ‘echoic imperatives’.2 Roughly speaking, in (1), B echoes A’s expression, in (2) Mary echoes what is implied in Peter’s question, in (3) and (4) the speakers echo their anticipation of the hearer’s action or tendency. By echoing, the utterances express the speaker’s dissociation from what has been echoed. Such examples serve as evidence against the claim that imperatives have, in their standard use, an illocutionary force of ‘directive’, which is defined as “an attempt to get the hearer to perform the action described by the proposition expressed” (Wilson and Sperber 1988, 18). The purpose of this paper is to clarify the comprehension processes of such use of imperatives in the framework of Relevance Theory. It is expected that the theory will provide us with a unitary explanation of how the hearer understands the speaker’s utterance.

2. Echoic use in Relevance Theory According to Wilson, echoic utterances are one kind of indirect quotation (Wilson and Sperber 1992, 59) and their relevance lies not in giving information of what is echoed, but in “conveying the speaker’s attitude to an attributed utterance or thought” (Wilson 2000, 432). These attitudes range “from outright acceptance and endorsement to outright rejection and dissociation” (Wilson 2000, 432). It should be noted that “what exists is a continuum, with different blends of attitude and emotion giving rise to a whole range of borderline cases which do not fit neatly into any existing 2

This terminology is suggested by Uchida.

On Echoic Imperatives: A Relevance-Theoretic Approach

121

scheme” (Sperber and Wilson [1986] 1995, 240). This suggests that speakers’ attitudes are placed on a scale with the positive attitude at one end and the negative attitude at the other end and cannot be categorised into discrete units. As far as comprehension is concerned, it is assumed that “an echoic interpretation is acceptable only if it contributes to the relevance of the utterance for the hearer (or may seem to the speaker to do so)” (Sperber and Wilson 1998, 284). That is to say, an utterance is understood as echoic only if the interpretation is consistent with the Principle of Relevance. Wilson illustrates this with Mary’s utterance at the table “Could you pass the salt?” after Peter made the same request using the same expression. Here, Mary’s utterance is properly understood as her own request which has no relevance to what Peter has said. Interpreting Mary’s utterance as echoic requires extra processing efforts and is not consistent with the Principle of Relevance in this context. A similar illustration with imperatives is given below: (5) ‘Go!’ cried the people along the shore. ‘Go!’ cried the owner of the Wasp. ‘Go!’ yelled the policeman. (White 1945, 36) The utterances in (5) cannot be regarded as echoic, although the same linguistic form is used, because each utterance does not express the speaker’s attitude toward the others, and, in that sense, each should be regarded as independent from the others. There are two points that we should keep in mind: the identification of echoic sources and the mechanism of metarepresentation. Echoic sources vary, as we have seen above. They may be expressed in the preceding utterance, or may not be expressed explicitly. In order to understand the speaker’s intended message, the hearer should identify these sources according to the Principle of Relevance. Comprehension processes of echoic utterances are explained with the mechanism of metarepresentation. That is, an extra layer of representation is added “to the communicated content, since not only the attribution but also the speaker’s attitude must be represented” (Wilson 2000, 432). The mechanism of metarepresentation is defined as follows: A metarepresentation is a representation of a representation: a higher-order representation with a lower-order representation embedded within it. (Wilson 2000, 411)

122

Chapter Seven

It is through this higher order representation that the hearer understands the speaker’s attitude to the attributed thought. Therefore, in order to understand an echoic utterance, the hearer should recognise the source of the echo (attributed thought) and the speaker’s attitude to the hearer.

3. Analysis In the following, we will examine various kinds of echoic imperatives and trace the hearer’s comprehension processes. Here, we assume three kinds of echoic sources: (1) the source is expressed in the preceding utterance(s), (2) the source is implicated in the preceding utterance(s) and (3) the source is in the speaker’s anticipation of the hearer’s action or attitude. As Wilson (2002) suggests, the hearer takes these processes not sequentially but in parallel.

3.1 An echoic source which is expressed in the preceding utterance First of all, let us consider the cases where a source of echoing can be found in a linguistic form. Noh (2000, 75-76) calls this ‘metalinguistic’ use. According to the following example, B regards an imperative form in A’s utterance as improper in this context. (6) A: Pass me the salt. B: (critically) Pass you the salt! Concerning this utterance, Noh says the following: B may echo A’s utterance in order to express his disapproval of its abruptness or lack of politeness, suggesting that he would prefer something like “would you pass me the salt?” or “Can you pass me the salt?” (Noh 2000, 75-76)

That is, B criticises A’s use of an imperative form to ask for the salt and suggests that A should choose a more polite expression. The comprehension processes are: (a) A has said to B “Pass me the salt.” (b) B has said to A “Pass you the salt.” = B has echoed A’s utterance. (c) B has said to A [A regards [B pass A the salt] as potential and desirable to A]. < accepted as an explicature>

On Echoic Imperatives: A Relevance-Theoretic Approach

(d) (e)

123

B’s utterance will be optimally relevant to A. B regards A’s utterance [A regards [B pass A the salt] as potential and desirable to A] as ridiculous.3

By echoing A’s utterance, B tries to remind A of what A has said and makes A realise that B thinks A should use a more polite or less abrupt expression. In this example, the target of irony is A, the hearer of an echoic utterance and initiator of an echoic source. B’s dissociation is understood by adding an extra layer of representation. In the next example, the target of irony is different from the hearer of the echoic utterance. (7) Context: Applegate Skinner is being teased by one boy after another because of his name. “Quiet!” Bellowed the tent leader. “I want quiet in this tent. Leave Applegate alone!” “Leave Rotten Applegate alone!” whispered another boy. And some of the other boys had to pull their pillows over their heads so their snickering couldn’t be heard. (White 1970, 104) The comprehension processes are as follows: (a) (b) (c) (d) (e)

3

The tent leader said to the boys, “Leave Applegate alone.”

Another boy has whispered “Leave rotten Applegate alone.” = Another boy has echoed the tent leader’s utterance with the addition of rotten. Another boy has whispered [the tent leader regards [the boys leave Rotten Applegate alone ] as potential and desirable to the tent leader.] The boy’s utterance will be optimally relevant (to the other boys).

The boy regards the tent leader’s utterance [the tent leader regards [the boys leave Rotten Applegate alone] as potential and desirable to the tent leader] as ridiculous.

I owe this idea to Wilson’s comments on an earlier version of this paper.

Chapter Seven

124

Thus, the boy’s utterance will achieve relevance by expressing his mockery toward the tent leader (and Applegate). Here, the target of irony is the tent leader. However, the other boys are the hearers. By whispering, the boy tries to remind them of what the tent leader has said and to express his dissociation from it without letting the tent leader notice. In (e) the boy’s dissociation from the tent leader is understood through an extra layer of representation. With the addition of rotten to the original utterance, the boy makes fools of both the tent leader and Applegate. In the following example, the target of irony is not present when an echoic utterance is given. (8) Context: A is complaining to B about Bob. A: Well, Bob said “Fix my dinner.” B: Fix his dinner! Who does he think he is? (modified from Noh 2000, 39) Noh considers this “a case of free indirect speech to represent another original imperative utterance.” The target, Bob, is the person his refers to, and the echoic source is what Bob said. B’s utterance is understood as follows: (a) (b) (c) (d) (e)

A reported to B that Bob said “Fix my dinner” B has said to A “Fix his dinner.” = B has echoed A’s report. B has said to A [Bob regards [A fix Bob’s dinner] as potential and desirable to Bob] B’s utterance is optimally relevant to A. B regards A’s report [Bob regards [A fix Bob’s dinner] as potential and desirable to Bob] as ridiculous.

B’s utterance will achieve relevance by expressing B’s dissociative attitude to what A reported. Through echoing, B tries to remind A of what A reported and make A realise that B is criticising Bob’s request. B’s attitude is recognised through an extra layer of representation. Let us next, look at the following example: (9) Context: Templeton, being awoken from a sound sleep by Wilbur, retorts sullenly.

On Echoic Imperatives: A Relevance-Theoretic Approach

125

‘Listen to me!’ cried Wilbur. ‘… Please, please, please, Templeton, climb up and get the egg sac.’ The rat yawned. He straightened his whiskers. Then he looked up at the egg sac. ‘So!’ he said, in disgust, “So it’s old Templeton to the rescue again, is it? Templeton do this, Templeton do that, Templeton please run down to the dump and get me a magazine clipping, Templeton please lend me a piece of string so I can spin a web.” 4 (White 1952, 167)

The echoic expressions “Templeton do this, Templeton do that, Templeton please run down to the dump and get me a magazine clipping, Templeton please lend me a piece of string so I can spin a web” show Templeton’s disgust toward being made use of as a messenger boy. The targets of his irony are Wilbur and Charlotte. These utterances echo Templeton’s interpretation of what they want him to do. It should be noted that they might not always echo exactly and literally what has been uttered or regarded as intended. That is, the relationship between what is echoed (=source) and the echoic mention here is not ‘identity’ but ‘resemblance’. The comprehension processes are shown as: (a) Templeton was asked to run errands for Charlotte by Wilbur.

(b) Templeton has said to Wilbur “Templeton do this,” etc. = Templeton has echoed what he was told. (c) Templeton has said [Wilbur regards [Templeton do this, etc.] as potential and desirable to Wilbur] (d) Templeton’s utterance will be optimally relevant to Wilbur. (e) Templeton regards Wilbur’s utterance [Wilbur regards [Templeton do this, etc.] as potential and desirable to Wilbur] as ridiculous. Templeton’s disgust is understood through an extra layer of representation.

3.2 An echoic source which is implicated in the preceding utterance What follows is a more complicated case than the one discussed above. 4

Charlotte is a spider that saved Wilbur’s life through her wits. She is dying. Wilbur wants to take Charlotte’s egg sac home in order for her offspring to survive. Templeton is a rat.

Chapter Seven

126

(10) Context: Peter, after running a marathon, feels hot, while Mary has a cold and rather wants to keep warm. PETER: Can I open the window? I feel hot. MARY: Go ahead and let in some nice Arctic air. (modified from Wilson and Sperber 1988, 90) In this conversation, Mary echoes what Mary regards as being implicated in Peter’s utterance. Mary’s utterance is understood as follows: (a) (b) (c)

(d) (e) (f)

(g)

Peter said, “Can I open the window? I feel hot.” Mary has said to Peter “Go ahead and let in some nice Arctic air.” = Mary has echoed what she regards as being implicated in Peter’s question. Mary has said to Peter [Peter regards [Peter open the window and as the result let in some nice Arctic air] as potential and desirable to Peter] Mary’s utterance will be optimally relevant to Peter. Arctic air is very cold, and not nice for Mary. Mary regards Peter’s utterance [Peter regards [Peter open the window and as the result let in some nice Arctic air] as potential and desirable to Peter (and Mary)] as ridiculous.

Mary does not permit Peter to open the window.

In this example, the target of irony is Peter. Through echoing, Mary tries to remind Peter of what he implied in his question and make him realise her dissociation from that. Again, Mary’s attitude is understood through an extra layer of representation. Thus, Mary’s utterance has gained relevance as a negative response to what Mary regards as implicated in Peter’s question.

3.3 An echoic source which is expressed in anticipation of the hearer’s action or attitude Not only the hearer’s utterances but also her thoughts, attitudes or anticipated actions can be echoic sources. Look at the following examples:

On Echoic Imperatives: A Relevance-Theoretic Approach

127

(11) Context: Mary, seeing Peter about to throw a snowball, says threateningly. Go on. Throw it. Just you dare. (Wilson and Sperber 1988, 89) Here Mary, through echoing what she anticipates Peter is about to do, expresses her negative attitude toward it. Mary’s utterance is understood as follows: (a) (b) (c) (d) (e)

Peter is about to throw a snowball. [contextual assumption] Mary has said to Peter “Throw it. Just you dare.” = Mary has echoed what Mary anticipates Peter is about to do. Mary has said to Peter [Peter regards [Peter throw a snowball] as potential and desirable to Peter] Mary’s utterance will be optimally relevant to Peter. Mary regards Peter’s attitude [Peter regards [Peter throw a snowball] as potential and desirable to Peter] as ridiculous.

The target of irony is Peter. Through echoing, Mary tries to remind Peter of what he is about to do and make Peter realise that Mary does not like it. Peter understands Mary’s attitude through an extra layer of representation. We see a similar echoic source in (12). (12) Context: Mary thinks John has a tendency to misunderstand her. Mary says to John. Mary: That’s right. Misinterpret everything I say. (Blakemore 1992, 111) This utterance is understood as follows: (a) (b) (c) (d)

John has a tendency to misunderstand Mary. Mary has said to John “Misinterpret everything I say.” = Mary has echoed what she regards as John’s tendency. Mary has said to John [John regards [John misinterpret everything Mary says] as potential and desirable to John]

Mary’s utterance will be optimally relevant to John.

Chapter Seven

128

(e)

Mary regards John’s tendency [John regards [John misinterpret everything Mary says] as potential and desirable to John] as ridiculous.

The target of irony is John. Through echoing, Mary tries to remind John of his tendency and make him realise that Mary does not like it.

4. Conclusion Using a framework of Relevance Theory, we have arrived at a unitary explanation of how the hearer understands what the speaker communicates by means of echoic imperatives, whatever the echoic sources may be. The comprehension processes of echoic imperatives can be summarised as follows. (A)

The hearer identifies the echoic source and activates it as a contextual assumption. (B) The echoic source is represented in the form of imperative which has the following semantic structure. [i] X regards Y as potential [ii] X regards Y as desirable to Z (X = target of irony, Y = proposition, Z = Hearer or Speaker of an echoic utterance) (C) The hearer understands the speaker’s dissociation through a higher order of representation which metarepresents the attributed utterance, thought, etc. This mechanism is shown as: X0 regards [X regards Y as potential and desirable to Z] as ridiculous. (X0 = speaker of echoic utterance)

Echoic imperatives are different from the imperatives in standard use in identifying variables. X in (B) is a target of irony. The target is mostly the speaker of an echoic source but as in (7) and (8), it can also be, or include, someone else. X0 is the speaker of an echoic utterance, the one who is communicating the negative attitude to the attributed thought through echoing. In this paper we have discussed that what Wilson and Sperber propose as a semantic structure of imperatives in standard use includes echoic imperatives without any ad-hoc modification. In addition, the speaker’s propositional attitude is properly placed in a higher layer of representation,

On Echoic Imperatives: A Relevance-Theoretic Approach

129

which embeds the attribution. In conclusion, the comprehension processes of echoic imperatives are successfully shown as a double layer structure with a higher order representation, which represents the speaker’s propositional attitude and a lower order of representation which represents the content of the imperative.

Data White, E. B. 1945. Stuart Little. New York: Harper & Row. —. 1952. Charlotte’s web. New York: Harper & Row. —. 1970. Trumpet of the swan. New York: Harper & Row.

References Blakemore, Diane. 1992. Understanding utterances. Oxford: Blackwell. —. 2002. Relevance and linguistic meaning. The semantics and pragmatics of discourse markers. Cambridge: Cambridge University Press. Noh, Eun-Ju. 2000. Metarepresentation: A Relevance-Theory approach. Amsterdam: John Benjamins. Sperber, Dan, and Deirdre Wilson. [1986] 1995. Relevance: Communication and cognition. 2nd ed. Oxford: Blackwell. —. 1998. Irony and relevance: A reply to Seto, Hamamoto and Yamanashi. In Relevance Theory: Applications and implications, edited by R. Carston, and S. Uchida. Amsterdam: John Benjamins. Wilson, Deirdre. 2000. Metarepresentation in linguistic communication. In Metarepresentations: A multidisciplinary perspective, edited by D. Sperber. Oxford: Oxford University Press. —. 2002. Relevance Theory: From the basics to the cutting edge. ICU open lectures on cognitive pragmatics. Wilson, Deirdre, and Dan Sperber. 1988. Mood and the analysis of nondeclarative sentences. In Human agency: Language, duty and value, edited by J. Dancy, J. Moravczik, and C. Taylor. Stanford: Stanford University Press. —. 2004. Relevance Theory. In The handbook of pragmatics, edited by L. R. Horn, and G. Ward. Oxford: Blackwell. Yus, Francisco. 2000. On reaching the intended ironic interpretation. International Journal of Communication 10, nos. 1-2: 27-78.

CHAPTER EIGHT THE PROCEDURE MEANING OF THE CHINESE CONJUNCTION DANSHI/DAN ZHOU LIN, LIU HONGGUANG AND ZOU LIZHI

1. Previous studies on the Chinese adversative danshi /dan 1.1 Zhuanzhe (transitional) sentence It is generally accepted that the Chinese adversative dan/danshi (‘but’) is a function word, and is used in ‘zhuanzhe’ (hereafter ‘transitional’) sentences to indicate contradictory relations between A and B. The pattern of A, danshi/dan B is considered to be the most typical transitional sentence. There are two opinions on the transitional sentence of A, danshi/dan B. One is that there is a contradiction between A and B, and A and B are coordinate and contradictory. The other is that there is a contradiction between B and what is expected from A, and the main clause is B. Anyhow, both of the two opinions reach an agreement in that the conjunction danshi/dan links A and B, and indicates transition relations between them. Let us consider (1): (1) Ҫ‫ݙ‬ᖗᕜᰃᚆઔ, ԚdanҪᖡԣ≵᳝䅽ⴐ⊾⌕ߎᴹDŽ (1’) He is sad deeply in heart, dan (‘but’) he doesn’t let the tears burst out. In (1), dan indicates that there is a contradiction or transition between A (=he is sad deeply in heart) and B (=he doesn’t let the tears burst out).

We are very grateful to Prof. Seiji Uchida for his criticisms and constructive suggestions and to Dr. Akiko Yoshimura for her insightful comments.

132

Chapter Eight

According to traditional grammar, in the pattern of A, dan/danshi B, B does not follow from the meaning of A, but expresses a contradictory meaning to A or to the meaning of an expectation from A. Thus, there is a transition from A to B.

1.2 Semantic relation In the transitional sentence of A, danshi/dan B, the most basic semantic relation between A and B is the contradictory relation. Sometimes there is also a contrastive relation, too. In this paper, we mainly discuss contradictory relations between A and B. Adversative conjunctions can be used when the following conditions are met: i) Event A has occurred. ii) It is expected that A will make event ¬B (‘not B’) occur. iii) In fact, event B appears unexpectedly (Lv 1982; Wang 1982; Xing 1992). iv) Both A and B have truth-conditions (Hu 1981; Guo 1999). Wang (1982) gives the logical form of A, danshi B as follows: AĺM¬B (M stands for ‘possibility’, and ¬B stands for ‘not B’). That is to say that the transitional relation between A and B expressed in A, danshi/dan (‘but’) B is based on the presupposition that if A, so the most likely possibility is not B. Along the same line, Hu (1981) claims that A and B exist independently and simultaneously, and they do not influence each other. The semantic relation of the transitional sentence is in the form of A, but B, because of C, and its logical form is AġCĺB. For example, see (2): (2) ໻ᆊ㚃ᅮ᳝ѯϡᚙᜓ, Ԛdan䇕г≵ᬶ䇈ҔМ. (2’) All of us are a little unwilling, dan (‘but’) no one dare to say anything. In (2), the result of fact A (=all of us are a little unwilling) should not be B (=no one dare to say anything), but it turns out that it is B, thus a transition or contradiction is formed. Concerning a contrastive relation between A and B in the pattern of A, danshi/dan B, Chen (1995) argues that A, danshi/dan B with a contrastive

The Procedure Meaning of the Chinese Conjunction danshi/dan

133

relation can be expressed in the formula: S1 (R (a)), danshi S2 (¬R (b)).1 Let us consider (3): (3) ៥ (a)ᜓᛣএ(R), Ԛdanཌྷ (b) ϡ㚃 (¬R)DŽ (3’) I (a) am willing to go (R), dan (‘but’) she (b) is unwilling (¬R). In (3), there is a transitional relation between R (a) and ¬R (b); that is, a (=I) has the feature of R (=willing to go) while b (=she) a feature of ¬R (=unwilling).

1.3 Pragmatic use The Chinese conjunction dan (‘but’) indicates a transition relation from A to B, which are not logically or semantically contradictory without dan. Cheng and Cui (2004) argue that conjunctions add a kind of semantic meaning to a compound sentence. For example, consider (4): ᇣ㢇᯼ᰮϔⳈ೼‫خ‬԰Ϯ,Ԛᰃdanshi ᇣᔎⳟњϔᭈϾᰮϞⱘ⬉㾚DŽ (4’) Yesterday evening, Xiao Fang did her homework, danshi (‘but’) Xiao Qian watched TV the whole evening.

(4)

In (4), without dan, there is no contradiction between A (=Xiao Fang did her homework) and B (=Xiao Qian watched TV the whole evening). The speaker uses dan to imply his complaint that Xiao Qian did not do his homework, but watched TV. Most researchers accept that in the pattern of A, danshi/dan, B, the core meaning is in clause A. Huang and Liao (2002) share the viewpoint and point out that the latter clause conveys the speaker’s intentions. Cheng (2002) claims that the speaker can change the order of the preceding and subsequent segments to express her subjective attitude. Consider (5): (5) (a) ᴢᬭᄺᬜᵰད, Ԛdan⾥ⷨϡさߎDŽ(ϡ䗖ᑨ催᷵Ꮉ԰DŽ) (b) ᴢ⾥ⷨϡさߎ, ԚdanᬭᄺᬜᵰདDŽ(ϡ䗖ᑨ催᷵Ꮉ԰DŽ) (5’) (a) Xiao Li has good results in teaching, dan (‘but’) his research performance is not good. (He isn’t suitable to be a teacher at universities.) 1

S1 and S2 represent two clauses that have a transitional relation; a and b represent compared objects, and R and ¬R (‘not R’) represent features of a and b. R (a) and ¬R (b) represent that a has a feature of R, and b has a feature of ¬R.

134

Chapter Eight

(b) His research performance is not good, dan (‘but’) he has good results in teaching. (He isn’t suitable to be a teacher at universities.) In sentences (5a) and (5b), there is a transitional relation between teaching performance and research performance in the pattern of A, danshi/dan B. If a person wants to get a teaching post at a university, he should be capable of both teaching and good research performance. In (5a), the speaker intends to emphasise that Xiao Li’s research performance is not good as evidence that he is not suitable to be a teacher at universities. In (5b), the speaker intends to remind people that Xiao Li has good results in teaching, though he is not suitable to be a teacher at universities.

1.4 Problems In the area of transitional conjunctions and transitional sentences, researchers have not reached a consensus on many issues. On the whole, the semantic and pragmatic relations within clauses or sentences are typically realised by various connectives in the Chinese language. There are numerous explanations for that, but no one seems to have explained why the speaker tries to establish the intended coherence relation by using them. Moreover, there are some linguistic phenomena that have no explanation. Consider the following example and imagine that a teacher is pointing at a mistake or bad handwriting of a student and utters (6): (6) 䖯ℹᕜᖿ, Ԛᰃdanshi (ཌྷᣛⴔᴀᄤ) (6’) Yes! You have made great progress, danshi (‘but’) (She points to the notebook)… In (6), danshi does not link words, phrases or clauses, but an utterance and contextual information. For traditional Chinese grammar, it seems difficult to explain cases where connectives link an utterance and contextual information.

The Procedure Meaning of the Chinese Conjunction danshi/dan

135

2. Reanalysis of danshi/dan within the framework of Relevance Theory 2.1 Procedural meaning of but From the perspective of cognitive pragmatics, Sperber and Wilson ([1986] 1995), and Wilson and Sperber (1993) suggest that discourse connectives are not linking devices of lexical items but devices used to restrict inferential processes. Along the same line, Blakemore (1987, 1992, 2000) argues that connectives encoding procedural information are used as devices to provide necessary processing constraints on the interpretation of utterances. Blakemore (2000) develops the analysis of but in English and claims that the proposition introduced by but derives the effect of contradiction and elimination: ... the segment introduced by but communicates (explicitly or implicitly) a proposition that contradicts and leads to the elimination of a proposition which the speaker believes is manifestly inferable from a mutually manifest phenomenon, which may be coded communicative behavior, or simply something in the physical environment. (Blakemore 2000, 474)

Blakemore (2000) gives a detailed analysis of but, using the examples in (7) and (8) below. (7) Her husband is in hospital but she’s seeing other men. (8) (a) Her husband is in hospital. (b) She is not seeing other men. (c) She’s seeing other men. When the hearer processes the first segment of (7), repeated in (8a), he/she may draw an inference such as the one given in (8b). But (8c), which is introduced by but, contradicts (8b), which leads to the elimination of (8b). Sperber and Wilson ([1986] 1995), Wilson and Sperber (1993) and Blakemore (1987, 1992, 2000) suggest a convincing explanation for why communicators use connectives in verbal communication, and how connectives impose constraints on human inferential processes.

2.2 Reanalysis of danshi /dan The cognitive perspective of the relevance-theoretic approach helps us get a better understanding of the role which adversative connectives play in

Chapter Eight

136

utterance interpretation. We assume that dan/danshi is not a linking device but one used to constrain an inferential process, and that danshi/dan may be relevant in communicating a proposition that contradicts the inference drawn from the preceding segment and leads to the elimination of it, e.g. (9):2 (9) ϟ䲼њ(=It is raining), Ԛdanഄ≵⑓(=the ground is not wet). A,

dan (‘but’)

B

A’: ഄ䴶⑓DŽ(The ground is wet.) In (9), dan indicates that B is relevant in communicating that the proposition expressed by the ground is not wet contradicts A’ (=the ground is wet), which is inferable from A, and leads to the elimination of A’. Consider the following examples: (10) ៥ϡњ㾷䖭ḋⱘ䇁㿔DŽԚᰃ(danshi) ៥೼右ໄ䞠਀ߎњϔ⾡ᅝ䯆ⱘᖿФDŽ (10’) I don’t understand birds’ language, danshi (‘but’) I heard a calm happiness in their singing. (11) A: ៥ϡњ㾷䖭ḋⱘ䇁㿔DŽ A’: ៥೼右ໄ䞠਀ϡߎҔМDŽ B: ៥೼右ໄ䞠਀ߎњϔ⾡ᅝ䯆ⱘᖿФDŽ (11’) A: I don’t understand birds’ language. A’: I can’t tell anything from birds’ singing. B: I heard calm happiness in birds’ singing. In the case of utterance (10), the hearer will first hear A (=I don’t understand birds’ language), then B (=I heard calm happiness in birds’ singing), introduced by danshi. Since B is contradictory to A’ (=I can’t tell anything from birds’ singing), A’ will be eliminated. 2 In the utterance of A, danshi/dan B in the following examples, A stands for the preceding element such as an utterance, a gesture or a context, A’ for an inference or contextual implication drawn from A, and B for the subsequent segment.

The Procedure Meaning of the Chinese Conjunction danshi/dan

137

(12) 䖭ḋⱘџᴀᰃϡᑨᔧথ⫳ⱘˈԚᰃdanshi㒜コথ⫳њDŽ (12’) It shouldn’t have happened, danshi (‘but’) it happened at last. (13) A: 䖭ḋⱘџᴀᰃϡᑨᔧথ⫳ⱘDŽ A’: ϡӮথ⫳DŽ B: 㒜コথ⫳њDŽ (13’) A: It shouldn’t have happened. A’: It will not happen. B: It happened at last. In (12) and (13), the utterance of B (=it happened at last) introduced by dan completely contradicts A’ and leads to its elimination. Now, consider (14) and (15): (14) Ҫᬊܹϡ催, Ԛdan䖛ⱘᕜᖿФDŽ (14’) His salary is not high, dan (‘but’) he lives a very happy life. (15) (a) Ҫϡᰃᬊܹϡ催? (b) ᰃ, ԚdanҪ⹂ᅲ䖛ⱘᕜᖿФDŽ (15’) (a) His salary is not high, is it? (b) No, dan (‘but’) he really lives a very happy life. In (15), the reply No, dan (‘but’) he really lives a very happy life indicates that A (=his salary is not high) implies not B (=he may live an unhappy life). Actually, B (=he lives a really happy life) is a fact. In traditional Chinese grammar, adversative conjunctions are thought to link phrases, clauses, sentences and even paragraphs with conflicting relations. Let us consider (16): (16) 咥㡆ⱘ⠵Ѝ, Դ㾕䖛৫˛Ԛ(dan) ⹂ᅲ᳝DŽ (16’) The black peony, have you seen it? Dan (‘But’) it really exists. Dan in (16) links the interrogative sentence: Have you ever seen a black peony? and the sentence It really exists. It seems difficult for traditional Chinese grammar to explain example (16). The relevance-theoretic approach will explain it in the following way: (17) (a) 咥㡆ⱘ⠵ЍˈԴ㾕䖛৫˛ (b) ≵㾕䖛DŽ (c) ৃ㛑ϡᄬ೼DŽ (d) ⹂ᅲ᳝DŽ

138

Chapter Eight

(17’) (a) The black peony, have you seen it? (b) Never. (c) Maybe, there are no black peonies. (d) It really exists. Hearing utterance (16), the hearer first processes (17a), to which he will give an answer such as (17b); then from (17b), he draws inference (17c) – if you have never seen the black peony, it does not exist. The proposition expressed by (17d) introduced by dan contradicts (17c), so (17c) is eliminated. Examples (15)-(17) show that Relevance Theory can easily account for the cases in which one of the segments in the pattern of A, danshi/dan B is an interrogative.

3. The role of dan: evidence from ERP3 3.1 The aim of the present ERP experiments Theoretically, connectives are used to constrain the hearer’s inferential processes and to minimise the hearer’s processing effort in utterance interpretation by indicating how the speaker expects the relevance of his utterance to be achieved without the hearer putting in too much effort. We have carried out Experiment 1 and Experiment 2 using ERP in the Language Acquisition Lab of Capital Normal University in China in order to find evidence of the important role of dan in language interpretation. Each of the experiments includes 160 testing sentences of eight types: Type 1, Type 2, Type 3 and Type 4 (hereafter T1, T2, T3 and T4) in Experiment 1, and Type 5, Type 6, Type 7 and Type 8 (hereafter T5, T6, T7 and T8) in Experiment 2. T1, T3, T5, T7 contain the connective dan, 3

ERPs (Event-related Brain Potentials) provide extremely higher time resolution in the range of one millisecond, and are characterised by simultaneous online measures of negative or positive potentials, amplitude, latency, and scalp distribution. Amplitude reflects the level or extent of the brain activity to stimulation. Amplitude can be measured in relation to the features of the waveform, and the measurement of features can be accomplished by determining its amplitude in μv. Latency can be used as a device of evaluation about the speed and time of the neural activity and its processing, and can be measured in terms of the temporal relationship between the features of the waveform and stimuli. The measurement of the features can be accomplished by determining its latency in milliseconds. The ERP experiments were carried out with two computers, one SynAmps 32 of Neuroscan, Quike-cap, STIM2 and SCAN 4.

The Procedure Meaning of the Chinese Conjunction danshi/dan

139

and T2, T4, T6, T8 do not contain dan; that is to say, half of the sentences have the Chinese connective dan and half of them do not.

3.2 Participants The participants were undergraduate students and graduate students from two universities (7 females and 13 males). The age of the participants was between 20 and 27 years – on average, they were 22 years old. All of them were predominantly right-handed, with no diagnosed learning disabilities or neurological disorders. They had never joined similar experiments before.

3.3 Materials and designs T1 to T4 were designed to make the participants fill in the blanks. In Experiment 1, we aimed to compare mainly T1 with dan and T2 without dan. The comparison of T3 and T4 was designed to obtain evidence of the role of dan in language interpretation in a reverse way. T5 and T6 contained no testing words or phrases, designed to find out about the differences of the participants’ understanding of sentences with dan and without dan. T7 and T8 were designed to compare the differences of the participants’ understanding of the testing phrases, which were implicitly conveyed by the sentences with dan and without dan. The following are examples of T1-T8 from the testing materials.4 4

Each experiment had 160 testing trials. Each trial was made up of an incomplete sentence with a blank in Experiment 1 or a sentence in Experiment 2, an attention marker and a testing word. One trial was divided into three parts and presented on the screen of the stimulus displaying computer. The first number refers to the trials’ type and the second number to the screen order. The materials were presented at random, one at a time. The following figure shows the stimuli sequences:

 ᑆ  ϟ䲼њԚഄ䴶B B screen1

ˇ screen2

screen3:

next trial trial

Chapter Eight

140 T1 T2 T3 T4 T5 T6 T7

11 It has rained, dan (but) the ground is _____. 21 It has rained, the ground is _____. 31 It has rained, but the ground is _____. 41 It has rained, the ground is _____. 51 This cloth is thin, dan (but) it makes you warm. 61 This cloth is thin. It makes you warm. 71 This cloth is thin, dan (but) it makes you warm.

T8 81 This cloth is thin. It makes you warm.

12 + 13 dry 22 + 23 dry 32 + 33 wet 42 + 43 wet 52 + 53 _____ 62 + 63 _____ 72 + 73 Only thick cloth makes you warm. 82 + 83 Only thick cloth makes you warm.

The characteristics of T1 to T4 are as follows: 13 in T1 is definitely the right answer because of dan; 23 in T2 is a wrong answer on most occasions.5 We compared the accuracy rate, reaction time and ERP features of T1 and T2 to clarify the role of dan in language interpretation. According to Relevance Theory, connectives with procedural meanings restrict the number of assumptions made by hearers in searching for relevance in utterance interpretation. Let us consider (18), which is a testing trial from Experiment 1, to illustrate how dan influences the participants’ understanding in a positive way. (18) T1 11ϟ䖛䲼њ, Ԛdanഄ䴶___DŽ 12 + 13 ᑆ It has rained, dan (but) the ground is _____. dry When the participants read it has rained in the first screen of (18), with the help of encyclopaedic knowledge stored in the brain, they might have such assumptions as rain must bring a certain amount of water and water may make the ground dump and wet or if rain is very heavy, there may be a flood and the like. After all, it is a common phenomenon that when it rains, the ground becomes wet. So, from it has rained, they will arrive at the assumption the ground is wet. However, there is a contradiction between the ground is not wet and the ground is wet and the inference that the ground is wet would be eliminated. In this way, dan might constrain the participants’ understanding and lead them to make a right judgment to 13 in T1. Now, consider (19), one of T2 trials from Experiment 1:5 5

We set ‘wrong’ to T2 as a right answer.

The Procedure Meaning of the Chinese Conjunction danshi/dan

(19) T2 21ᑈ䖛ܿकˈ䑿ԧ___DŽ He is over eighty. He is ____.

141

22 + 23 ‫ع‬ᒋ healthy

When a man is getting old, he will become physically weak and will often be ill. However, there may be people who are over 80, but still in good health. We hypothesised that the accuracy rate and reaction time of T1 would be higher and faster than that of T2 because the participants might think over the two possibilities when they respond to T2 without dan, as mentioned above. The comparison of T3 and T4 is designed to shed light on the constraint role of dan in a reverse direction. We predicted that T2 and T3 will involve an N400 component, since their terminal words were not anticipated.6 There is a difference between T5 and T6, which is that 51 has the word dan. Consider some examples from the testing trials from Experiment 2. (20) T5 51䖭㸷᭭㭘, ԚdanֱᱪᗻདDŽ 52 + 53 _____ This cloth is thin, dan (but) it makes you warm. (21) T6 61䖭㸷᭭㭘, ֱᱪᗻདDŽ 62 + 63 _____ This cloth is thin. It makes you warm. The testing trials of both 53 and 63 are ‘——’, which means no testing words or phrase. The participants were asked to pass judgments on whether 51 and 61 are right. We predicted that it will be a little more difficult for the participants to judge T6 right or wrong than to judge T5 because the relations between two segments of the utterances are loose and the amount of relevance is low in T6 without dan. Concerning T7 and T8, the testing phrases are the phrases implicitly expressed by 71 and 81. Consider (22)-(25): 6

Kutas et al. (1980) found that when participants read sentences ended with semantically inappropriate words such as I take my coffee with cream and dog, the potential activity of N400 becomes larger. Furthermore, Kutas and Hillyard (1984) have measured the magnitude of N400 by cloze probabilities. For example, in the test sentence of I send a letter with ____, only the word stamp is appropriate. Here the word stamp has a very high cloze probability. It means that the context constraint of this sentence is very high. N400 will be elicited by a wrong filling word under a high context constraint. There is a connection between N400 and the predictability of the target word under certain contexts: the less the word is anticipated, the larger the wave of N400 is.

142

Chapter Eight

(22) T7 71䖭㸷᭭㭘, ԚdanֱᱪᗻདDŽ This cloth is thin, dan (but) it makes you warm. (23) T8 81䖭㸷᭭㭘, ֱᱪᗻདDŽ This cloth is thin. It makes you warm.

72 + 73 㸷᭭८ᠡֱᱪ Only thick cloth makes you warm. 72 + 73 㸷᭭८ᠡֱᱪ Only thick cloth makes you warm.

The testing phrase of (22) and (23) is that only thick cloth makes you warm. In (23), the sentence may just describe two facts: This cloth is thin and it makes you warm while in (22), the speaker uses dan to lead the hearer to eliminate the contextual assumption thin cloth does not keep you warm, and it would be easier for the participants to judge that only thick cloth makes you warm is not an appropriate answer. So, we predicted that it will be easier for the participants to pass judgment on (22) with dan than on (23) without dan. (24) T7 71ϧϮᰃ⠽⧚, Ԛdan䇁㿔ᄺ㊒䗮DŽ 72 + 73 मᄺ His specialty is physics, great learning dan (but) he knows a lot about Linguistics. (25) T8 81ϧϮᰃ⠽⧚, 䇁㿔ᄺ㊒䗮DŽ 82 + 83 मᄺ His specialty is physics. great learning He knows a lot about linguistics. (26) Contextual assumption: Physicists will not have good knowledge of linguistics. Fact 1: His specialty is physics. Contextual implication: He doesn’t know much about linguistics. Connective: dan Fact 2: he knows a lot about linguistics. Implicated meaning: great learning The testing phrases of (24) and (25) are the same. (25) without dan describes two facts, from which one might infer that he is a man of great learning. However, we supposed that it is easier for the participants to pass judgment on (24) with dan than on (25) without dan. When the hearer processes Fact 1, as (26) shows, he might infer that he doesn’t know much about linguistics, and Fact 2 introduced by dan in (26)

The Procedure Meaning of the Chinese Conjunction danshi/dan

143

leads to its elimination. Thus, he might infer that he is a man of great learning. T5, T6, T7 and T8 were designed to confirm the results of Experiment 1 – that the proposition introduced by dan stands in contradiction to the contextual assumption derived from the previous segment and the contextual assumption is eliminated. Carrying out Experiment 1 and Experiment 2, we aimed to confirm our expectation that sentences with dan will be understood more quickly and correctly than sentences without dan in different testing forms. We predicted that the rates of response accuracy and the speed of the response time will be higher in the types with dan than in the types without the word dan.

3.4 Results and discussion 3.4.1 Behavioural data In our two experiments, we designed four pairs of tests. Each pair includes two sentence types, one with dan and the other without dan. Figure 1 shows the accuracy rates from T1 to T8. 1 0.8 0.6 accurate ratio % 0.4 0.2 0 stim type

1

2

3

0.879 0.5755 0.793

4

5

6

7

0.68 0.8845 0.587 0.898

8 0.327

stim type

Figure 1. Accuracy rates of T1 to T8

It can be seen that each of the response accuracy rates of T1, T3, T5 and T7 with dan is apparently higher than those of T2, T4, T6 and T8 without dan. Moreover, different tests show similar results in that the accuracy rates of T1, a filling blank test, and T5, a right or wrong judgment test with

Chapter Eight

144

dan, are about 87.9% and 88.5%, respectively; while the accuracy rates of T2 and T6 without dan are about 57.6% and 58.7%, respectively. Figure 2 shows that the response times of T1, T5 and T7 with dan are faster than those of T2, T6 and T8 without dan.

1.5

1 reaction time (second) 0.5

0

1

2

3

4

5

6

7

8

reaction time (second) 1.012 1.187 1.168 1.073 0.952 1.025 1.319 1.43 stim type

Figure 2. Response times of T1 to T8

The behavioural results of Experiment 1 and Experiment 2 suggest that the response accuracy rates of the sentences with dan are apparently higher than those without dan under different conditions, and that the response times of 75% of sentences with dan are faster than those of sentences without dan. On the whole, the ERP data of our experiments have confirmed the behavioural data on the cognitive role of dan. Sentences with dan cost less effort than sentences without dan. Statistical analyses also reveal that there are evident differences between the type with dan and the type without dan. The accuracy rates of T1, T3, T5 and T7 with dan are 33% (P