Language and Social Cognition: Expression of the Social Mind 9783110216080, 9783110205862

Table of contents :
Frontmatter
Contents
Chapter 1. Linguistic structures as cues for social cognitive functions: Introduction
Chapter 2. The origin of the social approach in language and cognitive research exemplified by studies into the origin of language
Chapter 3. Fused bodies: Sense-making as a phenomenon of interacting, knowledgeable, social bodies
Chapter 4. Supracultural models, universalism and relativism: The language of personhood in Chinese and American cultures
Chapter 5. The development of Turkish and Finnish words related to privacy
Chapter 6. On collective cognition and language
Chapter 7. Conversational pragmatics and social cognition
Chapter 8. The creative construction of social orientation: Situated positioning with English as a lingua franca
Chapter 9. Constructing knowledge schemas in the workplace: A microanalysis
Chapter 10. Corporate self-presentation and self-centredness: A case for cognitive Critical Discourse Analysis
Chapter 11. Distributed cognition and play in the quest for the double helix
Chapter 12. Social aspects of verbal irony use
Chapter 13. Attribution theories wired into linguistic categories
Chapter 14. Tuned to hidden messages: Exploring recurrent word combinations in English
Chapter 15. Emotion talk and emotional talk: Cognitive and discursive perspectives
Chapter 16. From motion to emotion to interpersonal function: The case of fear predicates
Chapter 17. Metaphor in mental representations of space, time and society: The cognitive linguistic approach
Backmatter

Citation preview

Language and Social Cognition

≥

Trends in Linguistics Studies and Monographs 206

Editors

Walter Bisang Hans Henrich Hock Werner Winter

Mouton de Gruyter Berlin · New York

Language and Social Cognition Expression of the Social Mind

edited by

Hanna Pishwa

Mouton de Gruyter Berlin · New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

앝 Printed on acid-free paper which falls within the guidelines 앪 of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data Language and social cognition : expression of the social mind / edited by Hanna Pishwa. p. cm. ⫺ (Trends in linguistics. Studies and monographs ; 206) Includes bibliographical references and index. ISBN 978-3-11-020586-2 (hardcover : alk. paper) 1. Language and culture. 2. Sociolinguistics. 3. Social perception. I. Pishwa, Hanna. P35.L266 2009 306.44⫺dc22 2009020636

ISBN 978-3-11-020586-2 ISSN 1861-4302 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. ” Copyright 2009 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Cover design: Christopher Schneider, Laufen. Typesetting: OLD-Media OHG, Neckarsteinach. Printed in Germany.

Contents Chapter 1 Linguistic structures as cues for social cognitive functions: Introduction . Hanna Pishwa

1

Section I: Social cognition and language Chapter 2 The origin of the social approach in language and cognitive research exemplified by studies into the origin of language . . . . . . . . . . . . . . . . . Nathalie Gontier Chapter 3 Fused bodies: Sense-making as a phenomenon of interacting, knowledgeable, social bodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anders R. Hougaard and Gitte R. Hougaard Chapter 4 Supracultural models, universalism and relativism: The language of personhood in Chinese and American cultures . . . . . . . . . . . . . . . . . . . Jason D. Patent

25

47

79

Chapter 5 The development of Turkish and Finnish words related to privacy . . . . . 127 Mai Kuha and M. Ali Bolgün Chapter 6 On collective cognition and language . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Fazad Sharifian Section II: Social cognition in discourse Chapter 7 Conversational pragmatics and social cognition . . . . . . . . . . . . . . . . . . . 183 Thomas Holtgraves and Blake M. Anderson Chapter 8 The creative construction of social orientation: Situated positioning with English as a lingua franca . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Andreas Langlotz

vi   Contents Chapter 9 Constructing knowledge schemas in the workplace: A microanalysis . . 237 Susan Bridges and Brendan Bartlett Chapter 10 Corporate self-presentation and self-centredness: A case for cognitive Critical Discourse Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 Veronika Koller Chapter 11 Distributed cognition and play in the quest for the double helix . . . . . . . 289 L. David Ritchie Chapter 12 Social aspect of verbal irony use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 Roger J. Kreuz and Gina M. Caucci Section III: Social cognitive functions of single structures Chapter 13 Attribution theories wired into linguistic categories . . . . . . . . . . . . . . . . 349 Klaus Fiedler and Peter Freytag Chapter 14 Tuned to hidden messages: Exploring recurrent word combinations in English . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371 Rainer Schulze Chapter 15 Emotion talk and emotional talk: Cognitive and discursive perspectives 395 Monika Bednarek Chapter 16 From motion to emotion to interpersonal function: The case of fear predicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433 Eliza Kitis Chapter 17 Metaphor in mental representations of space, time and society: The cognitive linguistic approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455 Paul Chilton Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473

Chapter 1 Linguistic structures as cues for social cognitive functions: Introduction* Hanna Pishwa

1. Background Cognitive linguists have made revolutionary discoveries concerning the explanation of the nature of language and its functions in recent decades. The findings are, however, concerned with knowledge representations and the processes required for its management with a heavy emphasis on the effect of human perception. Social aspects have not been considered as explanatory factors for the properties of language and its use despite the usage-based foundation; Chilton (this volume) argues, however, that cognitive linguistics “assumes that social meanings exist in people’s minds”. But this has not been tackled yet. Even other linguistic fields have failed to include social aspects of language despite the assumption that it is meant to serve neither as a communication tool for individuals in isolation nor as a mere knowledge-retrieving device. Sociolinguistics, being based on sociology, provides a macro-view on language with the main goal of gaining information on the linguistic behavior of social groups rather than on social categories and the processing of social information in interaction. Critical discourse analysis (Koller, this volume) is the only exception: social issues such as attitudes, stereotypes, and power belong to its central topics. Research conducted by van Dijk has included these issues since the eighties with the objective of revealing biased language use (see, e.g., 1988). Social and interactional issues, in particular those of intersubjectivity, are largely dealt with in terms of pragmatics, however, without demarcating them from other contextual aspects, that is, no distinction is made between social and other kind of information. The cooperative principle has gained ground even in social psychology and cognition as a framework viable for explaining com* I would like to express my thanks to Rainer Schulze for suggestions.

2   Hanna Pishwa munication (see below). Social psychologists, who have recognized the mutual benefit of the two disciplines, have shown in their research, for instance, that certain linguistic structures reveal implicitly how people account for their own and others’ behavior (Fiedler, Semin, and Koppetsch 1991; Forgas 1985; Malle 2002; Bless, Fiedler, and Strack 2004: 128–132; Holtgraves 2002, 2006; Kreuz and Ashley 2006; Robinson 2003). A frequently studied area is also non-literal language, in particular, figurative language among social psychologists (e.g., Fussell and Moss 1997; Kreuz and Ashley 2006). Other linguistic topics have been relativity theory, politeness theory, and conversation analysis. This work has shown that there is a mutual benefit of viewing social cognitive aspects in terms of language: “the study of language can contribute greatly to our understanding of social behavior” (Holtgraves 2002: 190); on the back cover of his book we read that “understanding language use requires an understanding of its social dimension”. The above description of the present state indicates that there is a gap to be filled in linguistics. This is the goal of the volume, which is to my knowledge the first comprehensive attempt to relate language structure to social contents in a systematic way from a linguistic perspective. For such an enterprise, an investigation of the symbolization of social cognitive aspects by means of language is required by asking how language reflects the social nature of humans as well as how people express their ideas about themselves and others. The reason for the urgency of this task is not only scholarly curiosity, but also its usefulness for other disciplines such as communication studies. 2. Topics of social cognition At present, there is broad agreement on the social nature of cognition. Smith and Semin (2004: 54) argue for ‘situated cognition’ (see also Langlotz, this volume), adopting the view that cognition is “an adaptive process that emerges from the interaction between an agent and the world, both physical and social”. A central assumption of this approach is that “cognition is for action”, which implies that motivation and emotion are elements of this process. Gibbs holds that even “intentions are, in many cases, emergent products of interaction between individuals, and between individuals and the environment” (2001: 106). It is probably no accident that both of these approaches include “emergent” in their definition of cognition. In the complex event of processing social information, “cognitive, affective, and motivational aspects of the social world are inextricably linked in the way we perceive and respond to other people” (Forgas 2001: 66). The papers in this volume support this partly constructivist view by

Linguistic structures as cues for social cognitive functions   3

providing evidence from language use. There is, however, a lack of consensus concerning some areas of cognition such as intentions, which are considered to reside within the individual by many researchers, and to be socially conditioned by others (see above). Social cognition (Fiske and Taylor 1991; Kunda 1999; Moskowitz 2002) is concerned with the representation, processing and application of social information. Consequently, scholars in this field “investigate how individuals mentally construct social reality” (Bless, Fiedler, and Strack 2004: 6). They can also be said to be occupied with finding answers to the question of how humans understand themselves and others and their interaction. These two descriptions suggest that a large number of diverse issues have to be considered when relating social cognition to language, which adds complexity to research tasks. Therefore, the topics discussed in the following will be restricted mainly to those touched upon in the chapters. Before this presentation, however, a short overview of the processing of social information is provided. Objects of investigation i.e., knowledge representations, are shared with social psychology, while most processes are adopted from cognitive psychology.1 Thus, cognitive psychology and social cognition differ mainly concerning the stimuli, which are far more complex in social contexts than in cognitive psychology. The difference between the two kinds of objects becomes evident when we compare the categorization of cups and persons: While cups are categorized according to their form and function, persons are classified in terms of their gender, age, role, traits, moods, background knowledge etc. The similarity of the perceiver and the stimulus (Bless, Fiedler, and Strack 2004: 11–12) increases contextual dependence and variability, so that the same stimuli may be perceived differently due to the communicators’ moods, goals and even social status, to a greater degree than suggested by Gestalt theory for things. Due to the similarity of the perceiver and the stimulus we find a mutual influence: people’s image of their selves also depends on their construction of social reality, and they also tend to construe social reality the way they think about themselves. What is said above suggests that social cognitive processing involves several elaborate stages with constraints at each step. The sequence of the steps is basically the same as in cognitive psychology, starting with perception, followed by a preliminary categorization and then by a more or less thorough comparison with information in memory and finally by further processes such as inferences and judgments (Bless, Fiedler, and Strack 2004: 19).

1 There are slightly different approaches to social cognition: a cognitively and a socially oriented framework (Nye and Brower 1996).

4   Hanna Pishwa When people are exposed to social stimuli, their minds start comparing their properties with already existing information in order to understand others’ behavior (Fiske 2004: 16). The first stage of this process is perception steered by attention in order to make relevant choices between the stimuli. It is generally assumed that what is attended well is processed in a more in-depth manner and memorized better. The limited attention is directed to objects and events compatible with interactants’ goals (Bless, Fiedler, and Strack 2004: 32). It is also caught by salience, for instance, unexpected social events. An example of this is people’s trust in others with the expectation that other people are good. When this trust is not met urgent attention is needed (Fiske 2004: 126). The general cognitive pattern is that negative properties not matching expectations are processed in more detail and stored more firmly in memory than positive ones. Attention results in the categorization of the stimulus in accordance with available information, which consists of social categories and schemas, e.g., the self, roles, and events as well as content-free schemas such as causal schemas (Fiske and Taylor 1991: 118–121) that may become targets of the initial categorization process. The perceived and attended information is added to the appropriate category or schema, i.e., organized knowledge patterns, and can also be retrieved from there on a later occasion. Most knowledge patterns are constructed hierarchically, allowing the activation of the whole network by means of subordinate nodes at a lower hierarchical level. The higher levels rarely serve as fruitful topics for interaction because of their broad coverage and high abstractness (see below). Social information may even be packaged in smaller chunks, for instance, mental models, which are more variable than schemas (see further Schank 1982; Graesser and Clark 1985; Wyer and Srull 1989). They are frequently used jointly in mental spaces (see Oakley and Hougaard 2008) and in situated concepts (Langlotz, this volume) to fill actual communication needs. Chapters in this volume illustrate people’s amazing ability to activate the right kind and amount of information for their purposes. In addition to their function as background knowledge communication, categories and schemas form the foundation for social concepts, which allow references even to the abstract (top) parts of information patterns (see Bridges and Bartlett). Some of these knowledge representations are culture-specific. Their availability to a culture, the topic of the first section of the volume, provides evidence for the collective nature of memory among the members of a certain group (Sharifian). In their description of the introduction of ‘privacy’ into two cultures, Kuha and Bolgün (this volume) illustrate the emergence of a social concept along with induced changes in society, indicating that the development of concepts proves humans’ social orientedness and constant interaction with

Linguistic structures as cues for social cognitive functions   5

the environment. The notion of collective memory is also dealt with in terms of “distributed cognition” by Patent, Ritchie, Sharifian, and Schulze. Patent holds that social concepts are available to all cultures on an abstract level, but that differences in their lexicalization occur on a lower level (see also Chilton). The self schema contains highly specified information and belongs to the most important knowledge structures available to us. It contains representations of past behaviors, values, beliefs, etc. (Moskowitz 2005: 158–159) and influences our perception of others as well as our own behavior. Some of its aspects are referred to in terms of ‘personhood’ (Patent) or ‘privacy’ (Kuha and Bolgün) in the volume. Some other chapters show that the self can be negotiated in discourse by enhancing one’s (even of companies and organizations) self-image (Koller) or be supported by reporting on success (Bridges and Bartlett). Self-enhancing is one of the major social motives for people, who “like to feel good about themselves” (Fiske 2004: 22); it also increases accuracy (Fiske and Taylor 1991: 178). Or the way of talking about the self may be adapted to cultural thinking, so that the display of a positive self image becomes stigmatized in order to avoid a face threatening act (Sharifian). The self becomes particularly significant when opposed to other person categories in communication, when people seek to discover each others’ feelings, wants, intentions and goals with the objective of sense-making for proper responses (Infante, Rancer, and Womack 1997: 198–291; Fiske 2004: 124). The processes employed may be heuristic, i.e. effortless, automatic and top-down. People are likely to choose this way if they trust their quick judgments (Moskowitz 2005: 231; Kreuz and Caucci). Attributions, a prominent topic of social psychology/cognition, are automatic and effortless. They refer to a mental process that helps us to make sense of the world by searching for causal explanations for others’ and our own behavior. The causes may lie in the actor himself (disposition) or in the circumstances (Fiedler and Freytag; Holtgraves and Anderson). Attribution theory has revealed biased behaviors as mentioned above. An example is the “self-serving bias”, which suggests that we consider ourselves particularly competent by attributing failures to the circumstances and success, to our disposition. Fiedler and Freytag (see also Holtgraves and Anderson) provide a detailed account of this process and other biases. Many situations, however, demand more effortful, systematic top-down and bottom-up processing (Moskowitz 2005: 202). This is the case, e.g., when perceived information is inconsistent with expectancies and the perceiver cannot rely on rules-of-thumb, because she believes that they may cause erroneous interpretations. Langlotz provides an accurate description of cognitive processing required for sense-making in a problematic situation

6   Hanna Pishwa due to linguistic difficulties, when it is not possible to arrive at the intended meaning only by activating existing information or using heuristics. People employ various strategies in order to cope with unexpected and difficult situations and may use language play (Ritchie), humor (Langlotz), or ironical statements (Kreuz and Caucci). The latter are described to offer the advantage of covering several highly varying functions, for instance, face-saving, and can be employed in varying situations (see below). Kreuz and Caucci argue that ironical statements are employed in accordance with the “principle of inferability”, i.e., when they are “likely to be understood”, a principle that involves heuristics. Social relations and roles can also be negotiated by means of metaphors (Chilton, Langlotz). Generally people strive for cooperation in communication. Accordingly, interaction is considered collaborative. Hougaard and Hougaard (this volume) argue that “sense is achieved socially by knowledgeable, interacting bodies.” This is in accordance with assumptions in communication theories that “in conversation, partners come to share, and often to create, jointly understood meaning” (Littlejohn and Foss 2005: 178). Furthermore, Ritchie’s paper demonstrates that collaboration contributes to common creativity. Other instances of accommodation are justification of behaviors or emotions in communication (Bednarek) as well as modification of earlier beliefs (Schulze) in order to accommodate to the situation. Before concluding this section, I would like to provide some reasons for why people are considered “cognitive misers” (Fiske 2004: 124; Barone, Maddux, and Schneider 1997: 130–139). One reason is the limitedness of attention and processing capacity, which influences the degree of accuracy in perception. Cognitive psychologists have recognized a “bottle neck” that prevents parallel processing, however, without being able to identify its location. Therefore, people weigh between the degree of accuracy and cognitive economy, still mostly achieving cognitions that are relatively accurate for the task at hand (“motivated tacticians”). In addition, the quantity of information leads to the title of “cognitive mizer”: the large body of information that people have about themselves overweighs that of others (see “attribution theory” below) and causes biased cognition. Accuracy is also dependent on the channel of perception: face and body are the channels in communication that cause the highest degree of accuracy (Fiske 2004: 130; Hougaard and Hougaard). Even some social dimensions such as extroversion and agreeableness are observed more accurately than others (Fiske 2004: 131). A further reason for poor processing is the dynamicity of the memory (see Pishwa 2006; Schank 1982), which constantly updates already stored information. Recall also functions best in the same kind of mood that the information

Linguistic structures as cues for social cognitive functions   7

was stored in. These few examples suggest that the available representations are not really “stored” but keep changing constantly and are also retrieved in dependence on moods. This is the cause for biased memorizing, as in the “hindsight bias”, which predicts that it is difficult to “reconstruct the way we had understood events before we had known about their eventual outcomes” (Kunda 1999: 182; see also Schulze). These facts indicate that social knowledge serves as an autopilot in interaction and might not be verbalized at all. The chapters provide a contribution for discovering whether, how, and in what cases this key information is verbalized, and what is more, whether it is ever an entrenched part of language, i.e. grammaticalized. In the following, the discussion will emphasize the verbalization of social cognitive topics considered in the papers. The most frequently discussed issue is social knowledge clustered in schemas2 and categories around prototypes, which enable social interaction by serving both as background information and as verbalized concepts. 3. Language and social cognitive aspects: Findings Since the “social turn” in linguistics (Gontier), there has been common consent that language has some kind of affinity to the social world. Hence, we can argue that, given the task of social cognition is to capture “the processes through which people come to know their social world” (Holtgraves and Anderson), language use can be assumed to be part of this process due to its role as a “collective memory bank” (Sharifian). However, as several papers in this volume (e.g., Ritchie; Hougaard and Hougaard) show, language might not be the primary or only means of communication. Ritchie, relying on Dunbar (1996), holds that language is “a byproduct of the development of cognitive abilities…” and basically inaccurate. In line with this is the assumption by Hougaard and Hougaard that its function is to coordinate the other, non-verbal resources and that sense making comprises the whole body without being restricted to minds or brains; hence, the framework is called “fused 2 The notion of schema covers varying contents in cognitive psychology. A broad conception is that schemas are more abstract and larger than categories and that they contain relationships between features and rules governing them, while categories classify things and persons. A more restricted view of schemas holds that they represent only event chains and their relationships with goals. For the following discussion, we adopt the broader definition in order to cover variable use in the papers; schemas may also be called ‘scripts’ or ‘frames’.

8   Hanna Pishwa bodies”. Also communication theorists attribute the primacy to nonverbal communication; however, measurements such as only 7 per cent of verbal aspects influencing the meaning are not considered reliable any longer (Hargie and Dickson 2004: 47). Obviously, the various elements of interaction have their specific functions as, for example, Forgas (1985: 140) holds that nonverbal messages are more efficient in conveying emotions and attitudes than language; their verbalization is shown to carry other functions (Bednarek). Kreuz and Caucci argue that paralinguistic cues should be taken into account beside linguistic cues in decoding ironical language. These facts corroborate the assumption that language has not developed with the goal of promoting social adaptation, because human nature is inherently social (Gontier). Kitis is, however, skeptical about the claim that language is socially constructed from the very beginning: she demonstrates that the development of the verb ‘fear’ both in Greek and English proceeds from a referential meaning to an intersubjective meaning through a textual and an expressive meaning. This is the developmental path in emerging languages such as second languages and pidgins as well. Having discussed the role of language in interaction, we now turn to the main findings concerning language in the present volume, starting with concepts. We have to keep in mind, however, that the present enterprise is limited as the number of possible topics in social cognition is large, in particular, when linked with language. Social (cognitive) concepts seem to belong to the few direct ways of mapping social contents onto language. As mentioned above, concepts can be utilized in two ways: (1) as background information to guide interaction, and (2) as topics in interaction. While the first role of concepts is ubiquitous, the second function is relatively rare. It is remarkable that, despite a large number of available social concepts, those chapters dealing with concepts address one single topic in this volume: the self schema. This fact lends support to the assumption about its high relevance for social interaction. As we already discussed this issue in detail above, we will turn to a closely linked phenomenon, the self as communicator. The self (the speaker) is the “origo” in all interaction. Therefore, deictic expressions serve as pointers toward other participants and can be regarded as markers for social functions. Consequently, tense marking can be used to create distance of all kinds (Koller); according to Wallace (1982), tense is used for anything but encoding time. In contrast, aspect use does not have this implication; it is rather related to perception and tends to express affects (Pishwa 2006a). The self can also be related to others by means of metaphorical use of spatial prepositions (Chilton), which cause differing social effects in dependence on the axe they occupy and the distance to the self.

Linguistic structures as cues for social cognitive functions   9

Pronouns as deictic elements exhibit true social cognitive functions in all languages, in that they can be used to categorize participants in interaction situations in their relation to the speaker, thus contributing to grouping people. This is in particular the case with the first person plural (Koller). Pronouns can also refer to those outside of the situation. The complexity of pronoun systems is culturally dependent. So is the employment of single pronouns as honorifics (Sharifian), whose main function is to create distance. This usage is similar to what can be achieved by means of prepositions along the axes surrounding the self (Chilton). Among syntactic means covering social contents we find subject and object properties attributed by the verb or adjectives. The interplay between verb/ adjective semantics and clause subject is linked with further dimensions such as controllability of events by the subject and the stability of certain properties over time (Fiedler and Freytag). Controllability has been captured in linguistics in frame semantics as initiated by Fillmore (1968), in the notion of transitivity (Hopper and Thompson 1980) as well in cognitive grammar in the ‘billiard ball model’ (Langacker 1991: 395–313). Adjectives, being more restricted in meaning than nouns in relation to a whole schema (Kuha and Bolgün), are appropriately used for evaluating people and are therefore frequently found in advertising (Koller; Janich 2001: 103), where they tend to appear in comparative constructions. Adjectives as well as adverbs can also contribute to irony (Kreuz and Caucci). Although emotions are an inherent part of our cognition and communication, they are not necessarily expressed by verbal means; instead there are other (facial, vocal, physiological) cues that we use to communicate our own emotions (Planalp 1999: 48; Bednarek). Their communication is culture-dependent, though all cultures offer rituals for their expression, for instance, the burial. This does not mean that we never refer to them by using emotion words; Bednarek illustrates the difference between “authorial” (the emoter’s own) and “non-authorial” (someone else’s) affect expressions. Communication of emotions is in any case a social act and “an effective way to establish social connections” (Planalp 1999: 138). This applies to the findings presented in this volume: the chapters addressing emotion expressions reveal that they are used to justify behaviors, decisions and other affects (Bednarek). Kitis’ investigation of the verbs ‘fear/be afraid’ in Greek and English shows that the emotive meaning developed metonymically out of verbs of movement (in Greek ‘flee’). In modern languages, these verbs have developed a meaning expressing ‘regret’ with regard to the proposition in an interactional function. This finding has parallels with Bednarek’s findings in that emotion words acquire further meanings, probably because speakers rarely use them to describe their own

10   Hanna Pishwa affects. It should be added here that regret can also be expressed by means of patterns with hindsight, which seems to be concerned only with a view on prior information or events (Schulze), so that we get a situation where emotion words are used for interactive purposes, and non-emotion words cover affects, however, as an additional cue. The fact that metaphors can be added to this account is not surprising (Chilton). Metaphors, being based on similarity of structure (Schulze 2006), are useful for the description of social cognitive aspects because they are adaptable to the situation by “allow[ing] for the emergence of properties that are not obviously part of either the topic or the vehicle” (Finke, Ward, and Smith 1992: 105). This is exactly what is required for the encoding of hidden social messages: an anchoring point or a structure with a freedom to either choose some or all of its properties or to add a further feature compatible with the structure. A similar function is provided by metonymic processes as shown by Kitis and by Schulze. For instance, the use of hindsight activates not only the review of the event or the decision made in the past, but also its emotional consequences. A further, important finding (Fiedler and Freytag) is that the degree of abstractness of linguistic expressions can be adapted by speakers to match with expectations. For instance, terms of high abstractness are used when the information is expected, i.e., when the speaker expects the hearer to possess previous knowledge of it. Abstract structures can also be utilized for manipulation by attributing certain characteristics to a person. Fiedler and Freytag show that abstract words are more likely to prime particular properties than specific words. The chapters also illustrate the social functions of irony, language play, and humor. 4. Conclusions Although the present volume presents only a fraction of the potential linguistic expressions for social cognitive aspects, some statements can be made. Apart from concepts, we can observe implicitness as a dominating property of their verbalization and conclude that linguistic devices used to express social functions show multifunctionality. In the following, I would like to propose an explanation for the choice of structures covering social cognitive contents and for the small number of direct mappings for them as far as the findings allow. It appears that we choose linguistic elements referring to single knowledge representations stored in collective memory for the encoding of social contents, as is the case, for example, with metaphors. This means that all we have to do is to evoke a knowledge representation by using a linguistic element as a cue linked to

Linguistic structures as cues for social cognitive functions   11

that information, and so retrieve the contextually relevant social function attached to it. As the above account illustrates, this process is an economical way of expressing and decoding social cognitive messages, since they must be understood quickly, primarily because of their structuring task in communication; the chapters provide evidence for the assumption that communicators are geared towards discovering social messages in the first place. An easy access is vital also because of the high complexity of social contents, whose comprehension requires flexible tools. Rapid comprehension is not possible if the message has to be decoded by means of problem solving strategies as they are far too effortful for this purpose. The linguistic devices, including concepts, presented in this volume are compatible with the idea of rapid, automatized meaning identification. Among the structures with hidden functions, the clause subject is an outstanding device to be used for the encoding of various social aspects. The subject properties, which are attributed by the verb or an adjective, discussed by Fiedler and Freytag, are so deeply entrenched in our collective minds that they are activated automatically in interaction. This applies, for instance, to agentivity and control; the latter is viewed as one of the core social motives for humans (Fiske 2004: 16, 19–22). Similar to this structure are constructions with hindsight (Schulze), deictics, and emotion expressions. Irony (Kreuz and Caucci), which sometimes supersedes automaticity effects, appears to be an exception as its comprehension may require additional mental effort in some cases. However, the authors show that people use heuristics, i.e., automatized rules of thumb, for its decoding. If we weigh between its advantages and the error probability, we discover that its use is economical in that it can be employed to fulfill a number of various tasks on the one hand, and that it heightens the memorability of the content on the other (Kreuz and Ashley 2006). In comparison, language play does not seem to lead to problems in interpretation. A real exception is the employment of problemsolving strategies described in the chapter by Langlotz occurring in situations deviating from the default case which excludes extra effects. The decoding process of the social meanings of the above structures is usually dealt with in terms of pragmatics. However, the findings indicate that we have to make more fine-grained distinctions. The default activation of social contents is clearly involved with priming and is an instance of “conventional implicature” (Levinson 1983: 127–128). Hence, we can assume that social contents are already stored with the linguistic cues and that these knowledge representations contain both world knowledge and collective social experiences, so called ‘experiental’ meanings, as suggested by Schulze. This assumption is in line with the developmental sequence of the functions put forward by Kitis relying on Traugott: propositional > textual > expressive > interpersonal, according to which the first contents to be expressed relate to the world and not to

12   Hanna Pishwa the speaker’s mental state or to social aspects. This developmental path is compatible with rapid meaning discovery: conventional implicature takes an intermediate position in the entrenchment scale linguistically, which starts with conversational implicature and ends with lexicalization or grammaticalization. Our ability to rapidly recognize social contents lends support to the assumption that people follow the cooperative principle and interpersonal rules stored in social schemas in communication (Holtgraves and Anderson). The ability to communicate in a relevant way implied that the cooperative principle is probably not inborn, but is acquired within each culture, based on social experiences and the human tendency to be a “cognitive miser” and a “motivated tactician” at the same time. In this sense, Ritchie argues that relational and social schemas influence “social behavior and thereby help to shape their relationships and interactions”. Gibbs (1999: 142) advocates the primacy of discovering others’ intentions arguing that this “isn’t grounded in the meanings of words or utterances themselves”, but that “speakers and listeners actively collaborate and coordinate their beliefs and knowledge to achieve mutual understandings”. The ability to reach this is based on extensive experience according to Gibbs (1999: 142), that is, previous knowledge, which becomes automatized and, hence, procedural. Kitis argues in line with this that these “pragmatic uses … may exceed semantic memory capacity and be stored as a type of procedural memory in social cognition”. Social knowledge can, hence, be considered a kind of autopilot steering behaviors subconsciously. Obviously, this process implies even more than the decoding of social aspects by using linguistic cues as proposed by Hougaard and Hougaard. With these facts in mind, we can argue that the implicitness of social messages is due to the ubiquity of social aspects for humans, as convincingly shown in many chapters: communication is rather a social act than sending and receiving messages. A further reason is the considerable complexity and volatility of social objects and their relations in interaction. How could speakers be able to develop fixed structures for such intricate phenomena that can be both semantic and procedural? If languages offered overt expressions for all social functions, their learnability would be radically reduced, and it would be a disadvantage more severe than the multifunctionality of the structures used to express them. The fact that the mind and the behavior of the participants are steered by variable abstract knowledge, difficult to categorize, may have been the reason for the neglect of social cognitive issues in linguistic approaches. However, if social aspects are so basic for human interaction and even for their existence, we can take for granted that language offers a whole lot of insights into their nature. Hence, the foundation laid in this volume has to be complemented by further studies.

Linguistic structures as cues for social cognitive functions   13

5. The chapters The interdisciplinary papers written by linguists, social psychologists, communication, and education students provide a multifaceted view on two issues: (a) the degree to which social cognitive information is entrenched in linguistic structure and, hence, found in different societies, and (b) the way social cognitive aspects are expressed in communication. The volume, consisting of three sections, is structured according to the principle ‘from linguistically broader to narrower issues’. Within each section, the chapters are organized according to their social cognitive contents. The first section addresses overarching issues such as the social nature of language and its cultural dependency. In part two, the scope of aspects dealt with is limited to the question of how social cognitive topics can be traced in discourse, or the reverse, i.e., how discourse can reveal these. The chapters in the third part focus on the expression of social communicative aspects in communication by means of particular linguistic devices. Section I: Social cognition and language Most of the papers in this part consider language to be a device for the distribution of collective cognition. The first chapter provides an historical overview of the social origin of language, while the second chapter presents a novel interaction model arguing that language is a purely social phenomenon in that cognition or meaning do not reside within individuals but is negotiated by interactants. The basically social view of language is corroborated in the next chapter, which provides a cultural view on the conceptualization of ‘personhood’ describing the emergence of the concept of ‘privacy’ in two different cultures and their languages. The last paper illustrates linguistic phenomena in various languages in order to lend support to a view of language as a “memory bank” for cultural cognition. The first chapter in this section (Chapter 2) provides an historical overview of the interpretation of the functions of language (Gontier) with a particular focus on its social properties. It starts with a description of its role for referential purposes as a knowledge-retrieving device and then proceeds to illustrate the “social turn”, which is shown to have existed a long time ago: already for Artistotle, spoken sounds were symbols for affections of the soul. From the Renaissance onwards, language was “understood as a communicative device” and considered to contribute to the creation of a political community. The conception that language represents the cultural nature of humans in opposition to

14   Hanna Pishwa the “natural” switched the object of linguistic study from word to syntax in order to change its focus in the 20th century to discourse and cognitive linguistics. The result of the review is that a social message is inherent in language since humans are basically social creatures. The next chapter (3) introduces a holistic, thought-provoking model of interaction, “fused bodies” (Hougaard and Hougaard). The social aspect is not viewed as one of the elements of a communication situation; it is rather the foundation of everything since humans are socially oriented. Intersubjectivity is also assumed to be the basis for individualism, and not vice versa. This means that we see the world from a social perspective and acquire our knowledge from this angle as well. The main idea from a linguistic point of view is that sense-making in a face-to-face interaction is not based on language alone, but that language is only one of the many aspects contributing to communication. It takes a role of integrating the non-verbal components such as gaze, gestures, movements of the body as well as cognitive and psychological effects into a coherent, multifaceted whole, suggesting that it is not enough to explain interaction by relying on cognition, psychology, or language alone. In the same vein, Patent’s chapter (4) advocates the idea of social orientation among humans from a cultural point of view. The author proposes that all cultures have access to universal supracultural models at a meta level but that single cultures may prefer different foundational schemas at a lower level. This assumption is based on his study on the conceptualization of ‘personhood’ among Chinese and American students. The results suggest that this category is available even to the Chinese despite their collectively structured society; the differences arise from focus put on different aspects of this cultural model. Evidence for the presence or absence of a particular cultural model is found in the use of particular linguistic forms such as pauses and false starts as well as viewpoints. Cultural conceptualization is also the topic of the next chapter (5) by Kuha and Bolgün addressing the emergence of the social concept ‘privacy’ in Turkish and Finnish. This notion is shown to be related to ‘self’ when contrasted to others. The authors demonstrate that such concepts appear in the two languages along with social changes such as urbanization and economic changes, for example, increasing wealth among the population, as these circumstances allow people more privacy. The material used for the investigation consists of dictionaries from different periods as well as of present-day corpora. An interesting finding is the difference in the scope of adjectives and nouns: while the use of adjectives binds the speakers to particular aspects of the concept, a noun comprises the whole spectrum of senses. So far, Turkish has developed only adjectives, while Finnish offers both categories.

Linguistic structures as cues for social cognitive functions   15

The perspective taken by Sharifian (Chapter 6) is also concerned with cultural conceptualization, i.e., cultural cognition. The author argues that both knowledge and language are socially grounded and form the foundation for collective memory. This kind of collective and cultural cognition is assumed to be holistic in being more than the sum of individual minds as it emerges from interaction between the members of the community. Sharifian holds that cultural patterns are distributed heterogeneously among the members of a culture; hence, he views cognition as a “complex, adaptive system” with a self-organizing ability. The task of language is to serve as a “collective memory bank” by providing devices for the expression of social and cultural aspects. These assumptions are underpinned by linguistic evidence, partly from conceptualization, partly from grammar. The author describes the classification of concepts in the languages of Australian Aborigines and compares metaphorical expressions for cultural concepts in English and Chinese. Among the grammatical phenomena discussed we find honorifics in the Persian pronoun system and their correlation with verbal inflections. A detailed account of various realizations of the pronoun ‘self’ in Persian finally serves as additional evidence for the relation of language and culturally dependent collective cognition. Section II: Social cognition in discourse The second section begins with a paper on assumptions that people make of messages in interaction. It is shown that they follow Gricean maxims, which are not, however, able to explain all behavior, as it is also considered to be steered by other principles. This paper is followed by a comprehensive and detailed account of knowledge structures and processes required to make sense of the use of non-native English to native speakers. The next two chapters address social schemas from various points of view. While one of them reports on the utilization of the top level of a schema for structuring an interview, the topic of the second one is the enhancement of the self schema. The last two papers in this section discuss language play and irony as tools for achieving discourse goals. The chapter by Holtgraves and Anderson (Chapter 7) provides evidence for the social nature of language. They show that addresses tend to follow interpersonal rules, i.e., Gricean maxims and conversational implicature, assuming that the information provided is always relevant. This orientation is used as an explanation for erroneous reactions by subjects partici­pating in psychological experiments, where their judgments can be controlled by varying the kind, the relevance, or the amount of information. The authors add that this behavior is

16   Hanna Pishwa not restricted to laboratory experiments but can also be observed in real-life situations. The Gricean theory is also criticized for being restricted only to western cultures. The authors argue that conversational maxims are probably not the only cause for the described behaviour but that they interact with interpersonal concerns. Langlotz (Chapter 8) provides a detailed account of social cognitive processes taking place in difficult interaction situations. The author presents the strategies employed by non-native speakers of English in a Swiss tourist office when delivering information to non-German speakers. The analysis of the decoding process of these strategies is the main objective of the paper. For this purpose, the author presents cognitive models as social representations, which are used to describe the cognitive elements of the interaction situation. The next step, the processing of the evoked representations, is illustrated in terms of mental spaces, then matched with the model of situated conceptualization, a dynamic view on categorization. Finally, the activated knowledge representations are coupled with the linguistic cues in terms of a ‘semantic map connectivity hypothesis’. The analyses of the three conversations achieved by means of this accumulated framework yields that the Swiss officers try to manage the situations by employing humor and allusions because of their deficient language skills. The important finding is that “the relationship between language and social cognition is bound to an intricate discursive interplay between linguistic cues, cognitive representations and processes, and motivated social action.” The following chapter (9) by Bridges and Bartlett illustrates the function of the top layer of a social schema for the structuring of a report on success. The paper discusses the social and cognitive elements underpinning the administrator’s belief schema for her success concerning methodology. The material is an interview based on a concept map completed by a second language administrator earlier for “stimulated recall”. The interview method, the use of a prefabricated concept map, is to provide the interviewee with an opportunity to check “with respondents on the accuracy and completeness of their accounts”. The analysis shows that effects are preferred to causes in the causally structured explanation of the concept map. The social cognitive aspect is referred to as the perspective taken by the interviewee, which functions as a kind of filter for the content and the way it is communicated. The reference to the top-level structure serves as evidence for the authors that humans are likely to choose the most relevant contents and processes for each situation, and hence, to adapt themselves to the addressee. The next chapter addresses the negotiation, i.e., the enhancement, of the self, the most important schema for people. The contribution by Koller (Chapter 10), delivered within the framework of Critical Discourse Analysis, shows

Linguistic structures as cues for social cognitive functions   17

by means of corpus data how companies and communities create a positive image in reports on the self by means of expressions for self-centredness and favourable self-evaluation. When contrasting this finding to references made to others, e.g. customers and investors, in discourse by the same contributors, the author discovers biased behavior in terms of detachment from them. The linguistic devices used to bolster the favorable effects are comparisons of character traits, degree of agency and personal pronouns for personification, and modality and tense for the ideal and represented self and transformations between these. Metaphor and metonymy are also used as discourse tools to tell who is represented in terms of what. Ritchie (Chapter 11) examines language play in interaction embedded in a scientific setting showing that language use is a social and holistic process involving numerous other aspects. The study comprises a comparison of scientific reports on the discovery process of the double helix structure (DNA). Word play is considered an instance of distributed and extended cognition because – in line with the findings of the previous papers – it satisfies discourse goals in interaction, increasing creativity by promoting “collaborative idea-generation and connection”. A related phenomenon, irony, the topic of Chapter 12 by Kreuz and Caucci, viewed from multiple perspectives, is found to carry several functions in interaction. Drawing on findings from previous experiments, the authors examine verbal irony from an interactional point of view and come to the conclusion that the processes involved in its production and understanding are highly complex because the phenomenon is dependent on factors such as the communication goal as well as the characteristics of the speaker and the situation irony is used in. The linguistic means used to express irony include exaggeration, lexical means such as adjectives and adverbs, interjections, as well as paralinguistic cues. The chapter also refers to the decoding process, which may end up in a communication failure because of the speaker’s inability to calculate the degree of inferability among the addressee(s). Section III: Social cognitive functions of single structures In the final section, the linguistic elements carrying social cognitive functions are single words or structures. Chapters (13–16) address structures used to reflect various social cognitive processes in interaction. Chapter 16 also illustrates a developmental path for a verb that has gone through a chain of changes. The last chapter discusses the metaphoric uses of prepositions for the expression of social cognitive phenomena.

18   Hanna Pishwa Fiedler and Freytag (Chapter 13) address sense making in interaction by examining the linguistic encoding of attribution. As already indicated, attribution theory is concerned with the ways people explain the causes of behaviors (see also Holtgraves and Anderson). In this sense, it is a cognitive process that activates social knowledge, specifically about control, and involves a decision process. The authors present the “linguistic category model” comprising verbs and adjectives which describe interpersonal behaviors with varying degrees of abstractness. The attributional dimensions attached to these linguistic categories are internal factors (disposition) vs. external conditions (situation), controllability and stability over time and situations. These properties are inherent in the subject and object, of course, in dependence on the verb, so that, for example, the degree of informativeness concerning the subject interacts with the degree of the abstractness of the sentence. The authors demonstrate that these linguistic categories may be used both in a cooperative and in a manipulative way as they exert influence on people’s judgments and therefore have psychological consequences; their use may also reveal unintended contents. Moreover, the authors argue that the various models of attribution theory are not deficient but cover distinct functions. This contribution corroborates the overall finding of this volume that social cognitive aspects tend to be expressed indexically in languages. The chapter by Schulze (14) examines linguistic patterns with hindsight for the revaluation of stored information in written communication, when the writer considers prior events or decisions. The material used for the analysis comprises “frozen” linguistic patterns of two to five words. These are assumed to reflect collective cultural cognition in that the speakers agree on the combination of the elements of the strings as well as on their meanings. The use of a corpus (BNC) allows not only analyses of single words but also the consideration of their linguistic contexts, i.e. collocates and the position in the clause. The analysis reveals that that hindsight does not only refer to the notion of knowledge but that each of these patterns carries “hidden messages” that can be related to ‘frames’ such as vision, history, and regret. This means that they exhibit experiential and expressive readings in addition to revaluation of prior knowledge. The additional readings arise partly due to the different collocational and syntactic patterns of the strings and partly also due to the context. Such linguistic patterns are shown to be conventionalized due to “collaboration of speakers’ and writers’ memories of all sentences […] giving rise to the detection of hidden meanings”. The objective of Bednarek’s chapter (15) is to discover cognitive-psychological functions for expressions of affect, emotion talk (direct expression) and emotional talk (implicit expression) in discourse. The analyses are based on corpus data and appraisal theory as well as cognitive theories of emotions.

Linguistic structures as cues for social cognitive functions   19

The results show that affective language frequently appears in causal chains for the justification of behaviors, subjective judgments or even other affects. This means that affect expressions function as attributions (not identical with Attribution Theory) and hence contribute to the question of how people understand themselves and others. The author notes that affects may be found in different positions in a causal chain and may therefore also take the role of an explanandum. The chapter by Kitis (16) also addresses emotion verbs as expressions for interpersonal concerns by providing a developmental view of ‘fear’ verbs in Greek and English, where it is used in the meaning of ‘regret’ concerning the content of the proposition. Kitis illustrates parallel developments of these two verbs on their way from motion to emotion verbs and finally to speech act verbs in these two languages. The semantic process responsible for this development is metonymy. This path is shown to follow the chain proposed by Traugott: propositional > textual > expressive > interpersonal. It is suggested that such shifts lead from one domain to another, with the last stage being the interpersonal level of language. Consequently, the author argues that emotions must be “social constructs played out in the real world”, but that this stage is preceded by personal emotions. Drawing on emotion theories, Kitis suggests that it would be advisable for psychologists to consult linguistic developments in order to avoid extreme proposals such as the claim that emotions are purely interpersonal from the very beginning. Chilton (Chapter 17) illustrates the metaphorical use of spatial prepositions as cognitive-linguistic devices for social contents, when considered in the axes surrounding the self. The axes are realized linguistically by means of prepositions for location and time. The author illustrates the motivated use of spatial expressions for social relations in terms of their complexity and the axes they occupy. The vertical axes conceptualize status, power and control, the frontback axis, order and priority in accordance with the path schema, and the lateral axis, support and protection. Distance and direction are shown to be the most productive sources for mapping social relationships of various kinds. Examples of distance metaphors are supporting others vs. distancing oneself from other people; this dimension also expresses acceptability.

References Barone, David F., James E. Maddux, and C.R. Schneider Social Cognitive Psychology: History and Current Domains. New York 1997 and London: Plenum Press.

20   Hanna Pishwa Bless, Herbert, Klaus Fiedler, and Fritz Strack 2004 Social Cognition: How Individuals Construct Social Reality. Hove: Psychology Press. Dunbar, R. 1996 Grooming, Gossip, and the Evolution of Language. Cambridge, MA: Harvard University Press. Fiedler, Klaus, Gün R. Semin, and C. Koppetsch 1991 Language use and attributional biases in close personal relationships. Personality and Social Psychology Bulletin 17: 147–156. Fillmore, Charles 1968 The case for case. In Universals in Linguistic Theory, E. Bach and R. Harms (eds.), 1–81. New York: Holt, Reinhart and Winston. Finke, Ronald A, Thomas B. Ward, and Steven M. Smith Creative Cognition: Theory, Research, and Applications. Cambridge, 1992 MA: The MIT Press. Fiske, Susan T. 2004 Social Beings: A Core Motives Approach to Social Psychology. Hoboken, NJ: Wiley. Fiske, Susan T., and E. Taylor Shelley Social Cognition. 2d ed. New York: McGraw-Hill. 1991 Forgas, Joseph Interpersonal Behaviour: The Psychology of Social Interaction. Syd1985 ney: Pergamon Press. 2001 Affect and the “social mind”: Affective influences on strategic interpersonal behaviors. In The Social Mind: Cognitive and Motivational Aspects of Interpersonal Behavior, Joseph Forgas, Williams Kipling, and Ladd Wheeler (eds.), 46–71. Cambridge: Cambridge University Press. Fussell, Susan R., and Mallie M. Moss Figurative language in emotional communication. In Social and Cogni1997 tive Approaches to Interpersonal Communication, Susan Fussell and Roger Kreuz (eds.), 113–141. Mahwah, NJ: Erlbaum. Gibbs, Raymond W. 1999 Intentions in the Experience of Meaning. Cambridge: Cambridge University Press Intentions as emergent products of social interactions. In Intentions and 2001 Intentionality: Foundations of Social Cognition, Bertrand Malle (ed.), 105–122. Cambridge, MA: MIT Press. Graesser, Arthur, and Leslie F. Clark 1982 Structures and Procedures of Implicit Knowledge. Norwood, NJ: Ablex. Hargie, Owen, and David Dickson Skilled Interpersonal Communication: Research, Theory and Practice. 2004 4th ed. London and New York: Routledge Holtgraves, Thomas M. 2002 Language as Social Action. Mahwah, NJ: Erlbaum.

Linguistic structures as cues for social cognitive functions   21 Hopper, Paul J., and Sandra A. Thompson 1980 Transitivity in grammar and discourse. Language 56: 251–299. Infante, Dominic, Andrew Rancer, and Deanna Womack Building Communication Theory. 3d ed. Prospect Heights, Ill.: Wave1997 land Press. Janich, Nina 2001 Werbesprache: Ein Arbeitsbuch. Tübingen: Gunter Narr. Kreuz, Roger J., and Aaron Ashley 2006 Nonliteral language, persuasion, and memory. In Language and Memory: Aspects of Knowledge Representation, H. Pishwa (ed.), 425–443. Berlin: Mouton de Gruyter. Kunda, Ziva Social Cognition. Cambridge, MA: MIT Press. 1999 Langacker, Ronald W. 1991 Foundations of Cognitive Grammar. Vol. II. Descriptive Applications. Stanford: Stanford University Press. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University. Littlejohn, Stephen W., and Karen A. Foss 2005 Theories of Human Communication. 8th ed. Belmont, CA: Thomason Wadsworth. Malle, Bertram F. 2002 Verbs of interpersonal causality and the folk theory of mind and behavior. In The Grammar of Causation and Interpersonal Manipulation, M. Shibatani (ed.), 57–83. Amsterdam: Benjamins. Nye, Judith L., and Aaron M. Brower What is social about social cognition research? In What’s Social about 1996 Social Cognition?: Research on Socially Shared Cognition in Small Groups, Judith L. Nye and Aaron M. Brower (eds.), 311–323. Thousand Oaks, CA: Sage Publications. Oakly, Todd, and Anders Hougaard (eds.) 2008 Mental Spaces in Discourse and Interaction. Amsterdam: Benjamins. Pishwa, Hanna Memory and language: Introduction. In Language and Memory: As2006 pects of Knowledge Representation, Hanna Pishwa (ed.), 1–34. Berlin: Mouton de Gruyter. 2006a Tense and aspect: Source of information. In Information distribution in English Grammar and Discourse and Other Topics in Linguistics, See-Young Cho and Erich Steiner (eds.), 231–254. Frankfurt am Main: Peter Lang. Planalp, Sally 1999 Communicating Emotion: Social, Moral, and Cultural Processes. Cambridge: Cambridge University Press. Schank, Roger Dynamic Memory. Cambridge: Cambridge University Press. 1982

22   Hanna Pishwa Schulze, Rainer On the reduction of complexity: Some thoughts on the interaction be2006 tween conceptual metaphor, conceptual metonymy, and human memory. In Language and Memory: Aspects of Knowledge Representation, Hanna Pishwa, (ed.), 143–162. Berlin and New York: Mouton de Gruyter. Schütz, Astrid 2001 Self-esteem and interpersonal strategies. In The Social Mind: Cognitive and Motivational Aspects of Interpersonal Behavior, Joseph Forgas, Williams Kipling, and Ladd Wheeler (eds.), 157–176. Cambridge: Cambridge University Press. Smith, Eliot R, and Gün R. Semin 2004 Socially situated cognition: Cognition in its social context. Advances in Experimental Social Psychology 36: 53–117. Van Dijk, Teun Social cognition, social power and social discourse. Text 8: 129–157. 1988 Wallace, Stephen Figure and ground: The interrelationships of linguistic categories. In 1982 Tense-Aspect: Between Semantics and Pragmatics, Paul J. Hopper (ed.), 201–223. Amsterdam and Philadelphia: John Benjamins. Wyer, Robert, and Thomas Srull Memory and Cognition in Its Social Context. Hillsdale, NJ: Lawrence 1989 Erlbaum.

Section I Social cogniton and language

Chapter 2 The origin of the social approach in language and cognitive research exemplified by studies into the origin of language Nathalie Gontier

1. Introduction During the last years, an increasing interest can be detected in social cognition and the latter’s relation to language, as well as the origin of language. In this introductory chapter, we trace the origin of this social approach. It will be demonstrated that the overall social approach in cognitive and linguistic sciences stems from the shift in philosophy from the “referential approach” to the “social turn”. Both paradigms endorse completely different ideas on what cognition is and what role language plays in general cognizing. This paradigm shift will be exemplified by an analysis of past and current research into the origin of language. Origins of language studies carry in their theoretical assumptions the shift from a transcendental to a secular world view. In the former, it is argued that language is referential for it bears true, factual knowledge of the outer, transcendental or physical world. In the latter, it is argued that language is social, and rather than provide truth about the world, it provides meaning to members of a certain linguistic community. In the referential approach, language and cognition become intertwined. And with the social turn, social cognition and language become related to one another. How exactly this happened is exemplified by studies into the origin of language. 2. The referential approach to language and cognition That cognition and language are two intertwined capacities is nowadays an unquestioned truism. However, this has not always been the case. Why then, and

26   Nathalie Gontier since when, did the study of cognition become related to the study of language? In this section the answer to this question is provided. The term “cognition” is nowadays by and large a container term that is used to denote the various (thought) processes of the mind. The mind, in turn, is a concept that is used to refer to the workings of the neurons in the brain. And “cognizing” has traditionally been a synonym for thinking rationally, i.e. for understanding something by making use of human reason or the rational intellect. Throughout Western history, the human ratio has always been interpreted to be of a linguistic nature. For many ages, both our thought processes as well as our mental concepts have been assumed to be structured linguistically. In fact, neurology (Damasio 1999; Ledoux 1996) only recently demonstrated that our thinking or cognizing can occur silently, and that mental categories cannot straightforwardly be associated with lexical categories. Historically, the goal of the ratio or of thinking was argued to be the acquisition of knowledge, more specifically, knowledge of the physical world. Cognition, language and knowledge have thus traditionally formed a tripartite and this tripartite forms the foundation of the referential approach. The referential approach to cognition, language and knowledge is as old as written history and can thus be traced to at least 7,000 years ago. The basic tenet of this paradigm is that human cognition is understood as a neutral knowledgeretrieving device. Human cognition enables the retrieval of factual knowledge about the physical world. Moreover, this factual knowledge is assumed to always be of a linguistic nature because thoughts are assumed to be structured linguistically. Language therefore is understood to objectively refer to the outer world. For example, the word “cat”, is argued to be the verbal expression of a mental representation of a certain animal in the world. In other words, the word “cat” refers to the animal cat, as well as to the mental concept (image or sensation) of a cat. All three, the mental concept, the word and the animal are traditionally assumed to share some basic properties. Stated differently, in ancient times it was assumed that there existed a 1-to-1 correspondence between thoughts, words and objects in the world. This position is called realism. Words were argued not to be arbitrary signs of mental categories or physical objects, rather, it was assumed that words actually provide objective knowledge of the objects they refer to. Exactly because of this idea, such an isomorphism between words and things could be endorsed. To understand this view more fully, an overview is given on how language, cognition and knowledge became related during the course of human history.

The Origin of the Social Approach to Language and Cognitive Research   27

2.1. Ancient Eurasian cultures From at least 7,000 years ago onwards, we can find traces in written history of the existence of cultures that develop explanations for the presence of language. From this time onwards, there is evidence that people believed that spoken and written languages have sacred or divine characteristics because they bring knowledge in and of the life of men, if not the whole of the universe. A Mesopotamian religion called Zoroastrianism would introduce the idea of a Sun-God. This is a creating god that stands above and outside of the world. And in India, the idea would develop that Sanskrit, the language of the RigVeda, is a perfect, sacred and divine language, as well as that it is the “language of the cosmos” (Hewes 1999: 573). The idea of a Sun-God as well as the idea of a sacred language of the cosmos would set the stage for all future theorizing on language in (at least) Indo-European countries up until the present. Descendents of these first scriptural cultures would populate the Middle East, ancient Egypt, ancient Greece, Rome and eventually they would colonize all of Europe, and later on also Asia, America, Australia and their original human populations. The cultural intuitions underlying these original cultures would give way to Ancient Egyptian mythology, Milethic (Ionic) and Hellenic philosophical thought, the three Semitic religions (Judaism, Christianity and Islam), Western secularism, humanist ideology and western science. More specifically, the cultural intuition (Pinxten 1997: 87) developed that the world we live in is an ordered rather than a chaotic one, and that this order is of a linguistic nature. In ancient Egyptian mythology, the order in the world is assumed to be the result of the “living creator of the life of the world” (Derrida 1981: 87), namely, the “enlightened” Sun-God Ra. The latter is also called Re or Ammon-Ra, Ammon stands for “hidden”. His firstborn son is the god Theut or Thoth. Thoth is the most powerful god in ancient Egyptian religion for he is acknowledged as “the master of divine words” (Derrida 1981: 91). More specifically, he is associated with the thoughts of his father, Ra. Ra’s thoughts are “hidden”. Thoth is the one who is able to speak his father’s thoughts and thereby bring them into the light (the light in turn is again associated with Ra). He can do this because Thoth is the god of speech, and through speech he can make the hidden truth and wisdom clairvoyant. Because of this, he is also the god of wisdom, and knowledge is understood to be of a linguistic nature. It is thus in ancient Egyptian times that the cultural intuition emerges that speech is both a vehicle for thoughts as well as an instrument to refer to (the truth about) the world. Both thought and knowledge are therefore language-based.

28   Nathalie Gontier 2.2. The origin of the logos-theory Zoroastrianism and ancient Egyptian mythology inspired Milethic (Ionic) and Hellenic philosophical thought. In Egyptian mythology, the sun shined upon the whole of its creation, thereby emanating a hidden truth or order which was made linguistic and thus literally audible and comprehensible by Thoth. This order, and eventually the capacities associated with the god Thoth, would be called the logos by the ancient Greek philosophers. In the 6th century BC, pre-Socratic philosophers, such as Heraclites and Parmenides who lived in the Ionian city of Milethe (presently in Turkey), would start a philosophical search for this hidden logos or world-order (later also interpreted as world-soul). Logos is a concept that translates as order, language, reason, thought, doctrine, account, statement, and the laying open of a relation (Held and Kirkland 2002: 83; Hillar 1998: 22). All these concepts are thus assumed to be synonymous (Coseriu 2003: 24–25). Steadily, the following cultural intuition would grow: since the logos is linguistic (logos is synonymous for thinking, knowing and speaking), and since humans are linguistic creatures, humans carry a part of the logos in themselves. Humans are thus actually part of the logos. This is also one of the reasons that humans can get to know the logos in the first place, as well as that they can lay open the relation between the words and the things. The nature of the relation between thinking, speech and knowledge would also remain the main topic of investigation in Hellenistic philosophy of the 4th century BC. Hellenistic philosophers raised the following discussion: since names are given to true and existing things in the world, are the names themselves also true and correct names that connote these things? Associated with this query is whether the names are given with insight into the true nature or essence of these things, and whether it are humans or gods that give the things their proper names. This discussion can especially be found in Plato’s dialogue entitled Cratylus (Plato 1921: 383–440) that was written in 360 BC. In the text, Hermogenes, Cratylus and Socrates discuss the origin and nature of words as well as their relation to the things in the world. The dialogue begins with Hermogenes who summarizes the views defended by Cratylus so that Socrates can join the debate. Cratylus is said to argue that names are “natural” rather than “conventional”, which means that they have an inherent rightness and correctness or truth in them. Moreover, all things have but one true name and these true names are the same for all humans, Hellenes or foreigners.

The Origin of the Social Approach to Language and Cognitive Research   29

Hermogenes (Plato’s Cratylus 1921: 384c–5a) on the other hand, argues that the correctness of names, or the truth-value of names, lies in their conventional use. Rather than a divine entity or a certain class of humans, all humans agree on the words they give to certain things. Therefore, the given name is always true and correct, even if one would choose to give the same thing a different name in the future. As such, there is no naturalness of names, for names are the result of human habits and customs. Having heard the two positions, Socrates (Plato’s Cratylus 1921: 385b–d) takes over the dialogue. He reasons as follows. Suppose I would give the name “human” to a horse and the name “horse” to a human. In this case, in my private language, horses would be called humans and humans would be called horses. However, if one accepts that one can speak the truth as well as that one can lie, then there must exist a true and a false language. The true language says it like it is – that horses (the name) are horses (the thing) – and speaks of the things as they are. The false language, on the contrary, does not say things like they are – but argues that a human (the name) is a horse (the thing). Agreeing with this line of reasoning, according to Socrates, equals agreeing that with language one can say what is the case (the truth) as well as what is not the case (falsity). If a language is true, so are the names that make up the language. Therefore, a thing can only have one true name. All the other names that are used to talk about the same thing are false. That a particular thing can only have one true name is explained by Socrates by arguing that the true name says something about the essence of the thing involved. More specifically, the essence of a thing is fixed, and since the name says something about this unchanging essence, the thing can similarly only have one fixed, true name. The word horse, for example, says something about the what-ness or essence of a horse: it defines what a horse is. Moreover, this definition remains equal for all the horses that have ever existed or shall exist in the future. Now the true name of a thing can only be given in agreement with the essence of that thing if the name-giver has named the thing with insight and knowledge into its true essence. Who then is this name-giver that can name with insight, and how is he able to do this? Not just any man can name the things by their right name. Only trained humans (especially philosophers) and gods can name the things correctly. They can name the things by their correct names because they are lawgivers who are able to contemplate on, or actually see the true (Platonic) model of both the name and the thing. When the name given to a certain thing is the true name, the name gives knowledge (epistème) of the thing. If the name is not true and thus not given in accordance to the essence of a thing, the name is false, or in

30   Nathalie Gontier accordance with the mere opinion (doxa) of the human name-giver. Gods are assumed to always use the right names because they naturally name with insight and knowledge. Humans however, more often than not, name according to their opinion (Plato’s Cratylus 1921: 400d, 401a–b), without caring for the truth. Thus, Socrates argues that the name can differ from the thing it names, because humans often have false opinions on the essence of things. The true name is that name that resembles the thing, in the same sense as a painting of a man, for example, resembles that man, although the painting is not the same as the man. An untrue, false name is that name that does not resemble the thing, in the same sense that one would argue that the portrait of a female is the portrait of the man. As such, names can be untrue and the result of mere opinions, rather than that names are always true (Cratylus’ idea). Socrates (Plato’s Cratylus 1921: 430d): “I call that kind of assignment in the case of both imitations paintings and names – correct, and in the case of names not only correct, but true; and the other kind, which gives and applies the unlike imitation, I call incorrect and, in the case of names, false.” Socrates concludes that there is still room for convention, since the name and the thing do not immediately coincide and false opinions can blur the true essence of a thing that is normally part of its name. In sum, it is in Plato’s dialogue that one can find the origin of realism as well as nominalism. Within a realistic doctrine one argues that language can provide knowledge about the world because words form a one-to-one correspondence with the essence of the things. In an ultra-realistic position, the study of language therefore suffices to study the world. Within a nominalist doctrine one argues that names do not provide insight into the nature of the world or the essence of things. Rather things receive their names by convention. This position does not exclude the possibility that knowledge about the world is nonetheless expressed linguistically. Rather, words arbitrarily refer to certain things of the world rather than that the words coincide with these things in the world and their essence. 2.3. The Word in Judeo-Christian traditions From at least the 3rd if not the 6th century BC onwards, the polytheistic religions of Egypt, Greece and Rome were countered by Middle Eastern, monotheistic belief systems such as Judaism, Christianity and Islam. In all three traditions, God created the universe as well as mankind and language becomes a gift from God. The first book of the Torah, Genesis, describes YHWH’s creation of the world. Out of nothing, this eternal entity creates through speaking. The first

The Origin of the Social Approach to Language and Cognitive Research   31

thing he says is “Let there be light” (Genesis, I, 1–3). This verse shows that in this monotheistic religion also the sun (Ra) becomes subjected to God’s powers, because now the light is created as well. But most of all, this passage demonstrates that creation becomes understood as a speech act. Moreover, Genesis describes how God creates man out of clay (in parallel with Plato’s Demiurge), according to his image. He blows Adam’s soul into his nose (Genesis 1–2) and this is what brings him to live. Besides God’s speech, God’s breath can therefore also create. Having been moulded out of clay in God’s image, man becomes physically similar to God and man too can speak, name and thus rule over creation. This ruling over creation involves subordinating and naming it. “Naming is knowing and subjecting” (Pombo 1987: 39). In Genesis, II: 19–20, it is told how God brought all the animals and birds that he had created to Adam to see how the latter would name them. In this regard Adam is similar to Plato’s onomatourgou or name-giver (see also Ecco 1995: 7–8 and Pombo 1987: 34–40). “Adam comes close, then, to the wise legislator and onomaturge of Plato’s Cratylus, who likewise […] determines the name on the basis of his knowledge of the essential nature of the object.” (Pombo 1987: 39). In fact, Adam becomes as God in this respect. As shall become obvious in what follows, this part of Genesis would become one of the most important texts in Western thinking, including the Western study of language. In general, it would become interpreted as follows: man received his language, which is a part of the soul, from God when he breathed life into Adam. The language spoken by Adam is the language that God speaks. This is why Adam (and all subsequent humans such as Moses, Job, etc.) can have actual conversations with God to begin with. In later writings, this language would be called the Adamic language. Jewish scholars as well as Christians would, in agreement with the story of Genesis, argue that at least until the flood (the story of Noah’s ark), the language shared by God and his creatures was one. Moreover, it was assessed that this language was invariant until the confusion of tongues that occurred at Babel. Regarding the Christian New Testament, especially the Gospel of John (1,2– 1,5) is interesting to place cultural intuitions about language (origin studies), for it begins with the verses: In the beginning was the Word, and the Word was with God, and the Word was God. The same was in the beginning with God. All things were made by him; and without him was not anything made that hath been made. In him was life; and the life was the light of men. And the light shineth in the darkness; and the darkness apprehended it not. (King James Version 1997)

32   Nathalie Gontier In this passage, the “Word” is to be interpreted as Jesus, the son of God (see also Apocalypse 19,13 where Jesus is again called the word of God). In the Gospel of John, Jesus thus bears striking resemblances with Thoth who is both the word of Ra as well as Ra’s substitute. Moreover, in the early Greek versions, this Word and thus also Jesus, is called the logos. John can therefore also be read as follows: “In the beginning was the logos, and the logos was with God and the logos was God.” As the speaking representative of God, Jesus has an important linguistic message for God’s creation: namely that they are freed from their sins if they believe in God’s name (John, 1,12). Believing in God’s name needs to be interpreted from within Platonic tradition: if they belief in the truth and correctness of his name, then they believe that God exists (and this name also says something about God’s essence). Moreover, the freeing of the sins is, just as creation was, a speech act or a linguistic event. Speaking therefore equals enacting and creating. God as well as his (linguistic) thinking are placed outside this earthly world, in analogy to the sun. Human beings, created in the image of God, are able to transcend their earthly body (language is considered to be part of the soul) as well as the world. They are able to see the whole and name and thereby structure its parts in an orderly fashion from a “God’s eye view”. Men can, in other words, become as God. With language, humans can be “objective” about the world which means that they can obtain and posses true knowledge of the world, as well as “objectify” the world, i.e. take and outsider position and look at the world as a whole that can be divided and ordered into different parts. This line of thought would found the correspondence idea that is typical of realism. More than anything, language would therefore become understood as a knowledge device, an objective instrument that orders our thoughts and refers to the outer world as well as structures this world. That language is a communicative device amongst human beings was secondary and at best understood as an act of charity by God. Language is understood to be primarily referential rather than social or cultural. A result of this intuition was the Medieval universalia debate. This is a debate wherein the ontological status of particular and universal (general) names is investigated (Coseriu 2003: 148–169; De Libera 1995: 319–339). Besides realistic positions, also nominalist positions would be endorsed. Realists would argue that universal terms, e.g. general terms such as ‘human’, ‘cat’, etc., are truly existing entities. They exist in a transcendental, ideational world and bring forward the particular, e.g. the ‘specific human’ named ‘Socrates’. Moreover, that Socrates is human says something about his essence. Universal terms are a precondition for particular terms and for the objects and

The Origin of the Social Approach to Language and Cognitive Research   33

subjects that correspond to them. In this line of thought, the word creates once more. Realists therefore assume that there is a direct isomorphism between universalia, particulars and the things they connote. Nominalists on the contrary would argue that only particular things truly exist and that these things exist independently of the words that denote them. Both particular (e.g. ‘Socrates’) as well as universal names (e.g. ‘human’) are abstractions of the mind. The shared essence that is presumed to be common to all human beings is again an abstraction of the mind. Universalia therefore do not have an existence beyond the human intellect. Besides an objective instrument, language would also become an occult instrument whereby one can, through language, invoke or call on God (e.g. by saying a prayer) or become God (by saying magical spells such as Abracadabra which literally means “I create as I speak”). 2.4. The Renaissance and the shift in cultural intuitions concerning the nature of language By the onset of the Renaissance it had become a given that God and Adam spoke the same, divine language during their communications. This “Adamic language” was regarded as transparent, perfect, unique and also as universal (Pombo 1987: 38–9). That is, the Adamic language was understood to be a perfect instrument of knowledge as well as a universal means of communication. It was also endorsed that this language got lost after Babel. The encounter of different cultures and different knowledge systems through trade and colonization would make Christian scholars realize that, contrary to the Adamic language, their many natural languages are all but perfect instruments to either obtain objective knowledge or to communicate fluently. This situation would become an impulse for the utopian thinkers of the late Medieval and Renaissance times to create a sense of nostalgia towards the period that existed prior to Babel, namely paradise where all men were united into one folk with one language. In this regard, the search for the Adamic language would eventually become redefined into two distinct problems: (1) the search for a universal language of knowledge understood to be of the highest cultural value; and (2) the search for the original mother tongue as well as the natural, primordial condition of man. A rupture would thus emerge between the scholars involved in the search for the original language. On the one hand, philosophers emphasized the heuristic value of the Adamic language. The Adamic language provides insight into the true nature of things. Language, in their account, can be defined primarily as

34   Nathalie Gontier an a-historic, a-social, objective instrument of knowledge. Its communicative value is only secondary. In so far as the Adamic language was lost, rationalist philosophers would engage in three things: they would search for natural languages that could take on the role of the Adamic language (e.g. Greek or Latin); they would try and “purify” natural languages (especially German) so that these languages could function as universal languages of knowledge; and they would construct artificial, logical languages that would enable to gain objective knowledge (Pombo 1987: 23). Especially the latter two rationalist endeavours would result in the onset of research into the grammatical structure of language as well as mental (equalled with lexical) categories of the mind. On the other hand, philosophers of law, philosophers of economics and eventually also philologists would understand the Adamic language to be a means to establish a social bond between all of mankind. They therefore defined language as a social, historical means of communication between men. This would eventually give rise to the social turn. 3. The social turn During the past 200 years, the validity of the referential approach to knowledge and cognition has been called into question by “the social turn”. Within the social turn, human cognition is not understood to be a neutral, linguistic device that is fine-tuned to retrieve factual knowledge of the physical world. Rather, both language and human cognition are understood to be an outcome of biological and social enculturation processes. Biologically, many aspects of cognition are currently recognized as “silent” or non-linguistically structured, and thus cognizing does not always make use of the linguistic medium. Socioculturally, a large part of cognition is currently recognized as the outcome of social enculturation. Moreover, it is endorsed that linguistic signs are arbitrary and that they find their origin in society through human convention. Why then, and since when, did the study of cognition and language become related to the study of social life? In this section the answer to this question is provided. Ideas on the arbitrariness of the sign can be traced back as early as the 5th century BC, especially in the works of Aristotle. These ideas would mainly be reintroduced in Humanist circles by (crypto-)Jewish scholars. But it is especially during the Enlightenment that one can find the foundations of the social turn that would characterize the theorizing on language in the 19th and 20th century. We now turn to these positions one by one and investigate how language became associated with social cognition.

The Origin of the Social Approach to Language and Cognitive Research   35

3.1. The arbitrariness of the sign The idea that the words or verbs are signs that arbitrarily denote different things or events, and that these signs are invented by men and agreed upon by convention is not merely a modern, Saussurian discovery. Earlier versions of this idea can already be found in ancient philosophical texts. As explained above, this idea was already partly discussed by Hermogenes in Plato’s Cratylus. But especially Aristotle, a student of Plato, elaborated more fully upon the idea. In his De Interpretatione, Aristotle (1995: 16a) writes that the alphabet consists of written symbols of spoken sounds, i.e. letters, while the spoken sounds are symbols of affections of the soul. They do not correspond to the affections, rather, they symbolize them. Both the written as well as the spoken symbols can differ in mankind, but the affections of the soul are the same. The idea that the affections of the soul are similar to all humans is essential to empiricist ideas as well as the idea of a psychic unity. The latter doctrine says that all men are equipped with the same sense apparatus, and when they are put in the same environment and stimulated by the same things, they will feel the same sensations. Taken on their own, these sensations are neither true nor false, they just are. Subsequently, these sensations and mental images are labelled verbally by convention (which explains why symbolic signs differ within different communities). “A name is a spoken sound significant by convention, without time, none of whose parts is significant in separation” (Aristotle 1995: 16a). Verbs are conventional signs that, additionally to words, also signify time. Similar to sensations, both words and verbs are, taken on their own, neither true nor false. They are a communal given. Words and verbs are only true or false in combination, for example in sentences such as “All cats are Persian” which is false and “Not all cats are Persian”, which is true. The endorsement that signs are arbitrary therefore does not necessarily question the possibility to obtain true and objective knowledge of the world. Rather, truth claims result from the combination of words and verbs into sentences. During the Christian Middle Ages many ideas of Aristotle were condemned by the Catholic Church in favour of Plato’s teachings. As a consequence, realism gained more foothold than nominalism. In Jewish traditions as well as in Arabic milieus, on the contrary, Aristotle’s work remained influential. Eventually, it were mainly Jewish scholars that reintroduced Aristotle’s ideas into Humanist, Protestant and eventually Catholic environments. An interested figure in this regard is Juan Luis Vives (1492–1540). Vives (Coseriu, 2003: 170–171) is famous for countering scholastic (Platonic) thoughts and (re)introducing Aristotelian ideas on language. Vives was one of the pio-

36   Nathalie Gontier neering scholars in semiotics and grammar. But most importantly, he argued against universalism in favour of historic particularism, and against language as an objective tool in favour of it being of an inter-subjective nature (Coseriu 2003: 174–176). Vives endorsed the view that every language follows its own grammatical rules and that words receive their content by convention rather than in accordance with the essence of a thing. Moreover, he primarily understood language to be a means to establish or deter social cohesion. In fact, it was language that turned humans into social and cultivated beings, and in this respect they also differ from other animals (Coseriu 2003: 175). Language enables reason and since animals lack language, they are argued to lack reason as well. Although language is still an instrument for knowledge, it is not exclusively understood to be an instrument of knowledge. Rather, in humanist thought language becomes associated with the highest form of culture. It is a tool that enables general cultivation. According to humanists, language is what makes humans educated and therefore civilized creatures rather than beasts (Coseriu 2003: 292). 3.2. The rise of philology As already described, during the Renaissance, trade and colonisation resulted in the encounter of different unchristian cultures. These cultures as well as their languages often outdated the European Christian culture. A first consequence was that other cultures became understood as preAdamic or, in other words, as “primitive” or even “primordial”. As such, they differed from the Europeans who had become “civilized” by the word of God, written down in The Bible. Other cultures and their members therefore became understood as “uneducated” “children”, “barbaric” “uncivilized” creatures, or plain “beasts” that lacked reason altogether. In sum, they were considered to live in a “natural” state rather than a “cultural” one. Secondly, the encounter of other cultures made utopian scholars wonder about a “natural” condition of man and how it differed from a “cultural” state. Is a “savage” man in a “natural”, “uncivilized” condition, and thus without the word of God, capable of distinguishing between good and evil? Is such a man capable of reason? Can his knowledge be used to gain truths about the world? And is the “cultural” state that “civilized” men find themselves in the best state to live in? Are there cultural alternatives? Can one reach a “higher” “civilized” form? Thirdly, many Protestant as well as utopian scholars would argue that paradise does not lie at the beginning of time, but at the end. Man can “perfect”

The Origin of the Social Approach to Language and Cognitive Research   37

himself through education. This first of all entailed learning to read and write so one can learn (about) the word of God. These changes in cultural intuitions would set the following cascade of events in motion. First, Europeans started to wonder whether all these creatures are the children of God and thus whether all the world’s languages can be united into one single mother tongue that was lost at Babel. Those in favour of a unity of mankind would argue that there is a universal reason (the rationalists) or a common sense (the empiricists) that unites all men. They would therefore also endorse single origin theories of language. Those against such a unity would pave the way for racism and multiple origin theories of language. Secondly, the thought-experiments laid open by the utopian scholars concerning “natural” and “civilized” conditions, as well as the Christian ideas of the existence of “primordial”, “savage” people would introduce the nature/ culture divide. The mind/body divide had already been a cultural intuition of many ages. The mind was rational and the body passionate. The extension of this dichotomy to the whole of nature and the whole of human culture only emerged from the Renaissance onwards. Not just language and reason, but the whole of culture (religion, rituals, customs, law, economics, education, etc.) were interpreted as completely different from nature. Culture was argued to separate white European men from “beasts” (to be interpreted as animals as well as members of different cultures). Finally, Protestant theologians would argue that the Adamic language is lost. It is the God-given human condition if not plain punishment that there exists a confusion of tongues. Thus, it is contra God’s will to search for the unifying Adamic language. In this regard, the translation of the Bible into many different languages by Protestant theologians such as Martin Luther (1483–1546) and Jean Cauvin (1509–1564) is to be considered as a pure act of charity (Pombo 1987: 36–37). The many languages became considered as the main obstacle that prevented an individual’s salvation. The Biblical translations enabled to evangelize all humans and their belief in God might allow the latter to have mercy on their souls. Evidently, to adequately translate the Bible into many tongues one must speak several of them fluently, as well as have sufficient knowledge of the Biblical textures. It were therefore mostly Protestant theologians that would engage in the effort to understand other languages as well as to translate the Bible into many tongues. It is in this tradition that textual philology would originate (textual analyses of different versions of the Biblical text in order to reconstruct the original version) as well as comparative philology (the study of the structure and relation of different languages). Especially the lat-

38   Nathalie Gontier ter would emancipate from Protestant beliefs and evolve into comparative linguistics. Moreover, Protestant scholars, again inspired by utopian thinkers (in turn inspired by the Bible its eschatological texts), endorsed that perfection did not lie at the beginning of time but at the end of time. And this perfection was argued to be reachable through civilization. This intuition would eventually enable historicist progressive ideas and evolutionary thought altogether. Regarding language, the Bible says that God brought the animals to Adam to see how Adam named them. According to Richard Simon and Grotius (both of the 17th century) language was thus an imperfect, autonomous human invention (Ecco 1995: 86). In the 18th century, Court de Gébelin would therefore argue that God, when speaking to his creatures, adapted himself to the imperfect language that this creature spoke. If perfection of mankind lies at the end of time, it follows that the beginning of time can be characterized as imperfect. How then can one reach perfection? It would especially be philosophers of law and economics that would tackle this question. These intuitions would eventually give rise to the many different academic fields that exist today: politics, economics, biology, psychology, anthropology, sociology, and linguistics. It would be these, at first philosophical scholars, that would introduce the idea that language first and foremost is an instrument of social life, rather than that it is an instrument of knowledge. 3.3. The search for the social origin of language Currently, scholars are inclined to understand language as an instrument that facilitates social cohesion. This modern notion on language first arises in the works of social contract theoreticians such as Hobbes, Locke, de Condillac, Rousseau, Smith and Herder. These philosophers are famous for their political and moral theories wherein they distinguish between a natural condition of man and an artificial, cultural one. This distinction related to the onset of the nature/culture divide. In the natural condition, human beings are considered to be a-moral: they are neither immoral nor moral, rather they live according to their natural instincts. The main point of discussion was whether humans are good or bad natured. Hobbes, for example, would argue that in a natural condition men do not socially enjoy each others company. Rather, they “[…] are in that condition which is called Warre; and such a warre, as is of every man, against every man” (Hobbes 1909: 96, p. 63 in the original).

The Origin of the Social Approach to Language and Cognitive Research   39

It is only when humans are united into an artificial society wherein they give up part of their natural freedom in order to live according to the law of the Common Wealth that they become moral and civilized. The origin of such a society wherein man becomes civilized and moral is characterized by Rousseau (1972) as a situation wherein men engage in a social contract. What is interesting, but unfortunately for a large part neglected by current evolutionary linguists, is that all these political and moral philosophers also developed theories on the origin of language that were associated with their ideas on the rise of a political society. In the same fashion as they emancipated laws and ethics from religious thought, they emancipated the origin of language. That is, no longer was language understood to be a present from God, rather, language became characterized as a human and social (rather than natural) invention as well as convention. In his Leviathan, Hobbes (1909) would write that language is a human invention rather than a divine gift. Language enables social life. Without language “[…]there had been amongst men, neither Common-wealth, nor Society, nor Contract, nor Peace, no more than amongst Lyons, Bears, and Wolves.” (Hobbes 1909: 24, p. 12 in the original). The cultural intuition of this era is that, more than anything, language enables and facilitates culture. As a consequence, language is not understood as part of the domain of nature. Rather, language is what enables social cohesion within the common wealth. In other words, language is something whereby the common wealth can “[…] signifie to another, what he thinks expedient for the common benefit” (Hobbes 1909: 130, p. 88 in the original). This theme would also be repeated by Locke and Rousseau. In his An Essay concerning Human Understanding, the empiricist Locke (1942: 201) would argue that “GOD, having designed man for a sociable creature, made him not only with an inclination and under a necessity to have fellowship with those of his own kind, but furnished him also with language, which was to be the great instrument and common tie of society.” And in his posthumously published Essai sur l’origine des Langues, Rousseau would declare that “[…] la parole étant la prémiére institution sociale […] [Language was the first social institution].” (Rousseau 1970: 27, p. 1 in the original). Language thus becomes a social and political instrument. It is a communicative device that bonds humans in social life, rather than that it provides true and objective knowledge of the physical world. Language also remained associated with thoughts, especially moral thoughts that distinguish man from other animals. That is also the reason why language cultivates and civilizes. It brings humans outside of the natural domain and into the cultural realm. Although languages originate as natural human inventions,

40   Nathalie Gontier the subsequent development of languages is first and foremost interpreted as a social and cultural event rather than a biological one. De Condillac, for example, would argue that language originated in a natural condition as well as from a natural condition. More specifically, language originated with the beginning of our species. In the natural condition, all men would share a same “language of actions”. This is a communication system that is based upon emotional states such as “cries of passion” as Condillac (1746: part II, 5–6) calls them. The more language “progresses” and “cultivates”, the more such cries are replaced by arbitrary spoken signs. Because of this, language is argued to become more abstract and artificial and as such, it becomes a representative of culture. In other words, it becomes an instrument to measure the “level of civilization” one is in. And also Rousseau (1970: 29, p. 3 in the original) would argue that language “progressed” (the word evolution was only introduced later in time) from a more natural and “simple, unlearned” state to a more artificial and “complex, educated” state. The original, natural languages were therefore assumed to be gestural and iconic, while the later emerging, social languages were vocal and abstract. Rousseau argued that in a natural condition, this unlearned pantomimic language would suffice for the whole of mankind. In a natural condition, all humans express their physical needs in the same, universal gestural language. It is our social (second) nature that requires the origin of vocal languages. Moreover, it is this social nature of human beings that makes vocal languages differ in the first place (because of the convention of the sign), rather than that vocal languages too are spoken universally. 3.4. Language and historicism, evolutionism, anthropology and linguistics By the 19th century, it had become a truism that vocal languages are learned through enculturation in a social community. This idea had two consequences. To begin with, it became custom to assume that languages can be used as instruments to measure the level of cultivation the nation finds itself in. Secondly, it would be argued that the state of the nation can be changed by the apprenticeship of a different language and its accompanying governmental organization. Here lies the beginning of the cultural intuitions that there exists a close interrelation between nations (ethnié), languages and a governmental style of organization. This tripartite would be poured into a developmental, historicist and evolutionist “scale of progress”. The latter would ground nationalism and racism, imperialism and western hegemony, but also linguistic determinism and linguistic relativism.

The Origin of the Social Approach to Language and Cognitive Research   41

Within historicism and evolutionism, human history was presumed to run through certain qualitatively distinct phases and to “develop” according to well-defined historical and evolutionist laws. In practice, this meant that certain languages or cultural groups were falsely classified as “degenerated”, “less evolved”, or “underdeveloped” in comparison to “enlightened” and “civilized” westerners. In other words, these historicist and evolutionist scholars used culture and biology to argue against a psychic unity of mankind, and in favour of a division of mankind into different “races”. Anthropologists, on the other hand, would argue in favour of a psychic unity. Beginning with Herder, it was argued that language is a universally shared human trait that distinguishes us from other animals, while on the other hand, language also allows for cultural particularism and as such causes for differentiation amongst human beings. As Pan (2004: 14) points out, Herder’s ideas imply that different languages will result in different forms of reasoning. Eventually, this idea also allows for the assumption that language actually creates a world. Although the idea of linguistic determinism (the view that language creates a worldview) is generally subscribed to Wilhelm von Humboldt, it was actually formulated 50 years earlier in time by Herder. Anthropologists such as Boas would combine the social approach to language with non-racial evolutionary biology. Boas distinguished between habits that are learned and that become automated, and instincts that are innate (and therefore part of the organic). According to Boas (1962: 139), a particular language is learned as a habit, while the faculty of language is innate (instinctive) and thus the result of our biological endowment. Namely, a habit (or faculty, e.g. the language faculty) is biologically (organically, instinctively) determined, but the specific content of the habit (or faculty) is acquired. More specifically, the content is acquired from society. The habitual use of a particular language subsequently becomes automated. In sum, Boas assesses that a distinction needs to be made between, on the one hand, the faculty of language that is innate and biologically determined, and on the other hand, a particular language, that is learned in society and that subsequently becomes habitually used and automated. This would inspire Kroeber (1963: 3 – §  2), who argued that culture and language need to be interpreted as extra-individual entities that take on the form of “more-than-organic phenomena”. The latter is alternatively called a superorganic structure. The existence of a superorganic structure would again be countered by Sapir (1917). Nonetheless, the latter, together with one of his students, Benjamin Lee Whorf (1956), would argue that language is influenced by culture and vice versa, culture by language, in such a manner that language

42   Nathalie Gontier determines an individual’s and a society’s thinking. Both cognition and language are therefore immediately social. Moreover, the distinction between a universal biological faculty of language on the one hand, and on the other hand, the existence of particular languages that are interpreted as social institutions would remain a recurrent theme in linguistics, from de Saussure (1972) to Chomsky (1965) to the present. Eventually, the social turn would also enter the philosophical realm (Gontier 2006). Wittgenstein (1989) would argue that one cannot demonstrate that language refers to an outer world. Rather, language originates in the social realm. Similar to the case of language, knowledge is subsequently first and foremost interpreted as a social endeavour, and scientific theories become “languages” of society. In the same manner as language is the result of a social contract, so scientists endorse in a social contract that determines the rules and politics of science. In this post-modern account, convention rather than truth unites scientific scholars. 3.5. Evolutionary linguistics The shift from the referential to the social turn eventually also entered evolutionary linguistics. Not only is it customary to explain language as a social phenomenon, nowadays also its evolutionary biological origin is explained from within social selection pressures. That language arose for social reasons is subscribed to by scholars working within artificial intelligence, primatology, psychology, anthropology, neurology, palaeo-anthropology and archaeology (see Gontier 2007 for an overview). Hurford, Studdert-Kennedy and Knight (1998), for example, note on the back cover of their 1998 volume: For the past two centuries, scientists, as children of societies preoccupied with technology, have tended to see language function as largely concerned with the exchange of practical information about the mechanisms of the physical world: tool making, hunting, and so forth. By contrast, this volume (a product of the age of mass democracies) takes as its starting point the view of human intelligence as social, concerned with one’s own and others’ desires and motives, and of language as a device for forming alliances, making friends, and thus achieving successful feeding and mating through a complex social network.

Evolutionary linguists however are scarcely aware that a 7–8,000 year old history precedes their current theorizing on the origin and evolution of language. This history nonetheless needs to be taken into account because it explains why questions are framed the way they are and why we examine the evolution of language the way we do today. Put negatively, because scholars are largely

The Origin of the Social Approach to Language and Cognitive Research   43

unaware that the social approach emerged to counter the referential approach, potential biases or cultural intuitions such as the nature/culture divide remain largely undetected. Moreover, certain intuitions, although they have lead to new avenues of research, currently remain ungrounded. That words for example arise as a result of social convention rather than that they are the outcome of private thought, is an intuition that has become such a truism that scholars mostly do not focus on grounding this intuition scientifically. With this article some of these historical biases have been exposed and the hope is cherished that in the future, the social approach will become the subject of reflexivity as well as firmer scientific grounding. 4. Conclusion As far as we can trace written history of certain cultural ideas back in time, language has been the subject of investigation from at least 7,000 BP onwards. Language has traditionally been understood as a trait of a divine entity, a divine entity tout court (e.g. in the idea that the word has become flesh), or a gift from God to humans. It is a shared device that allows God and men to abstract hidden knowledge (i.e. the underlying structure of the physical or transcendental world) and as such to create order. Since language leads to the eternal truth, and truth is one, there can only be one true language that says it like it is. It follows that all other languages are false derivatives of the first, true language. The origin of ideas on purification of existing languages, the search for the Adamic language, or the creation of universal languages is not that hard to understand from within this cultural background. It additionally explains why thoughts are associated with language, why language is reduced to the study of the word, why words and thoughts are argued to be re-presentations of things, and why language is only secondarily a communicative device. If it is the goal to reach individual salvation or happiness, it is merely an act of charity to instruct others in the true word. The search for the universal divine language actually triggered the idea that there was a primordial, natural condition. The latter, so it is argued, is distinct form the current, cultural “perverted” condition. Beginning from the Renaissance onwards, language becomes primarily understood as a communicative device. Rather than transmit knowledge, language enables the transmission of social information between members of the same community. Language enables the creation of a political community because it facilitates social coherence. In this regard, language becomes a representative of all that is cultural which in turn becomes opposed to all that is natural. Language becomes a measure

44   Nathalie Gontier as well as a representative, not of truth (as e.g. Thoth or Jesus are), but of social life, political coherence, and culture altogether. Besides being a measurement of the degree of civilization one is in, it enables education and language can therefore lead to patriotism or nationalism. Language enables cultivation into social, political and cultural life. Rather than being a device of knowledge, it becomes a device for the transmission of culture. It is in this climate that the emphasis on the word is loosened to include the study of syntax, and that the question on the biological origin of language is first posed. Originally, language is argued to have originated out of moral needs, a characteristic that enables men to be lifted up from the condition of animals. Passionate cries and gestures are argued to become combined into a pantomimic language and eventually spoken language emerged. Language, in this view, not only enables social cohesion and the possibility to transcend a natural state, language actually influences the type of cultural environment one creates and vice versa, the cultural environment influences the language that is spoken. These ideas paved the way for the doctrine that language creates an idiosyncratic worldview rather than that it objectively lends insight into the outer world that exists independently of human beings. Moreover, in post-modern times, politics, science, culture etc. become nothing more than languages or regimes that determine, from within, the rules according to which one has to play to be political, scientific, etc. And within Post-Neodarwinian traditions, it is argued that language evolved because of this social aspect that language enables. Given the long historical background, the latter however should become the main topic of scientific investigation rather than that it is taken as a truism that guides current theorizing.

References Aristotle 1995 Boas, Franz 1962

Complete Works of Aristotle, the Revised Oxford Translation, Vol. I and II, Jonathan Barnes (ed.), Princeton, NJ: Princeton University Press.

Reprint. Anthropology and Modern life. New York: W.W. Norton and Company, 1928. Chomsky, Noam 1965 Aspects of the Theory of Syntax. Cambridge, MA: The MIT Press. Condillac, Etienne Bonnett de 1746 Essai sur l’Origine des Connoissances Humaines: Ouvrage où l’on Réduit à un Seul Principe Tout ce qui concerne l’Entendement Humain.

The Origin of the Social Approach to Language and Cognitive Research   45 Institut National de la Langue Française, CNRS: http://gallica.bnf.fr/ ark:/12148/bpt6k87990k.item. Coseriu, Eugenio Geschichte der Sprachphilosophie von der Anfängen bis Rousseau: Neu 2003 bearbeitet und erweitert von Jörn Albrecht. Tübingen/Basel: A. Francke Verlag. Damasio, Antonio The Feeling of What Happens: Body and Emotion in the Making of 1999 Consciousness. New York: Harcourt Brace and Company. De Libera, Alain La Philosophie Médiévale. Paris: Presses Universitaires de France. 1995 Derrida, Jean Jacques Dissemination. Chicago, IL: University of Chicago Press. [Translated 1981 by Johnson, B.] de Saussure, Ferdinand Reprint. Cours de Linguistique Générale. Paris: Editions Payot, 1916. 1972 Ecco, Umberto The Search for the Perfect Language. [Ricerca della lingua perfetta 1995 nella cultura europa.] Oxford/Cambridge: Blackwell Publishers, 1993. Gontier, Nathalie Introduction to evolutionary epistemology, language and culture. In Evo2006 lutionary Epistemology, Language and Culture, Nathalie Gontier, Jean Paul Van Bendegem and Diederik Aerts (eds.), 1–29. Dordrecht: Springer. 2007 The dynamics of language activity: An evolutionary reconstruction. Ph.D. diss., Department of Philosophy, Vrije Universiteit Brussel. Held, Klaus, and Sean Kirkland The origin of Europe with the Greek discovery of the world. Epoché 7 2002 (1): 81–105. Hewes, Gordon W. 1999 A history of the study of language origins and the gestural primacy hypothesis. In Handbook of Human Symbolic Evolution, Andrew Lock and Charles Peters (eds.), 571–595. Oregon/New Jersey/Oxford: Blackwell Publishers. Hillar, Marian The Logos and its function in the writings of Philo of Alexandria: Greek 1998 interpretation of the Hebrew myth and foundations of Christianity. Part I. A Journal from the Radical Reformation: A Testimony to Biblical Unitarianism 7 (3): 22–37. Hobbes, Thomas Reprint. Leviathan. [Leviathan or the Matter, Form and Power of A 1909 Common Wealth, Ecclesiasticall and Civil.] Antwerpen: Boom, Original edition, Oxford: Clarendon Press, 1651.

46   Nathalie Gontier Hurford, James, Michael Studdert-Kennedy, and Chris Knight (eds.) 1998 Approaches to the Evolution of Language. Cambridge, MA: Cambridge University Press. Kroeber, Alfred L. 1963 Reprint. Anthropology: Culture Patterns and Processes. New York: Harbinger Books, 1923. Ledoux, Joseph The Emotional Brain: the Mysterious Underpinnings of Emotional Life. 1996 New York: Touchstone. Locke, John 1942 Reprint. An Essay Concerning Human Understanding. London: J.M. Dent and Sons LTD, 1690. McKirahan, Richard 2003 Presocratic philosophy. In The Blackwell Guide to Ancient Philosophy, Christopher Shields (ed.), 5–25. Oregon/New Jersey/Oxford: Blackwell Publishers. Pan, D. 2004 J.G. Herder, the origin of language, and the possibility of transcultural narratives. Language and Intercultural Communication 4 (1–2): 10–20. Pinxten, Rik When the Day Breaks: Essays in Anthropology and Philosophy. Frank1997 furt am Main: Peter Lang. Plato 1921 Plato in twelve volumes. Cambridge, MA: Harvard University Press. Eversion of the Persues Digital Library at Tufts University: http://www.perseus.tufts.edu/cgi-bin/ptext?doc=Perseus%3Atext% 3A1999. 01.0172. Pombo, Olga 1987 Leibniz and the Problem of a Universal Language. Münster: Nodus Publikationen. Rousseau, Jean-Jacques 1970 Reprint. Essai sur l’Origine des Langues: Ou il le parle de la Mélodie et de l’Imitation Musicale. Bordeaux, Ducros, 1781. 1972 Reprint. Du Contrat Social. Oxford: Clarendon Press, 1762. Sapir, Ernst 1917 Do we need a superorganic? American Anthropologist 19: 441–447. The Bible 1997 Authorized King James Version with Apocrypha. Oxford: Oxford University Press. Whorf, Benjamin Lee Reprint. Language, mind, and reality. In Language, Thought, and Re1956 ality: Selected Writings of Benjamin Lee Whorf, John B. Carroll (ed.), 246–270. Cambridge, MA: MIT Press, 1941. Wittgenstein, Ludwig Tractatus Logico-philosophicus, Tagebücher 1914–1916, Philoso1989 phische Untersuchungen. Frankfurt am Main: Suhrkamp.

Chapter 3 Fused bodies: Sense-making as a phenomenon of interacting, knowledgeable, social bodies Anders R. Hougaard and Gitte R. Hougaard

Introduction The present volume shows how language users in and through their choices of wording display cognitive dispositions to social aspects of the discourse and how sensitivity to this enriches linguistic description in ways that are only beginning – with initiatives such as this volume – to be fully recognized. An obvious partner in this endeavor is cognitive linguistics, which however, despite its new ambitions (see for instance Croft, to appear) to embrace the social aspects of cognition and language use, has not yet explored the potentials of social cognition. For the most part the “social” turn in cognitive linguistics, which is a very heterogeneous development, has resorted either to macro-analyses in the vein of sociolinguistics (Geeraerts 2005) or to notions from the field of pragmatics such as speech act theory (Pascual 2008), discourse analysis (Williams 2005) or Herbert Clark’s (1996) ‘common ground’ (Croft 2005). Since both disciplines have roots in cognitive psychology, a wedding between social cognition and cognitive linguistics seems very timely. Of course a cognitive approach that wishes to avoid solipsism and disembodiment must take into account how language use reflects social cognitive aspects of discourse and discourse situations. However, in this contribution we take issue with all approaches to such a wedding which either start with concerns for the “inner life” of the individuals that make up discourse situations or which “isolate” language as the carrier of sense. We aim at introducing a truly social approach, i.e., a sociological approach, not to language and (social) cognition but to sense achieved interactionally in and through the use of all embodied resources (including language) in face-to-face-interaction. This approach does not presuppose particular individual cognitive processing of the social situation or of non-social things. What it assumes at the outset is the reality, independent of any single human (but done by human beings), of

48   Anders R. Hougaard and Gitte R. Hougaard social processes. Through a particular approach to these social processes we aim at exploring the cognitive aspects of interaction as these are oriented to, constructed, developed, assumed and relied upon by the participants of social interaction themselves. With this approach, we do not assume a body of cognitive psychology terminology or any other cognitive terminology. What we assume is that social interaction is a socially ordered process in which sense is achieved socially by knowledgable, interacting bodies (as opposed to simply “minds” or “brains”). This approach is new both to cognitive approaches to sense-making and to traditional fields of discourse studies including ethnomethodology (Garfinkel 1967), conversation analysis (Heritage 1984; Hutchby and Wooffitt 1998) and discursive psychology (Edwards and Potter 1992). Mainstream cognitive science (here used as a cover term for disciplines such as cognitive psychology, cognitive linguistics, social cognition, cognitive and neuroscience and social neuroscience) is firmly wedded to the idea that any social process must presuppose the existence of individual minds or individual cognition. However the fact that social organization can only be achieved by the collaboration of individuals does not necessarily entail that ‘thinking’ and ‘mind’ are also rooted first and foremost in the individual. Durkheim (1966: xvii) anticipates this argument in the preface to The Rules of Sociological Method: …because society is composed only of individuals, the common-sense view still holds that sociology is a superstructure built upon the substratum on the individual consciousness and that otherwise it would be suspended in a social vacuum.   What is so readily judged inadmissible in the matter of social facts is freely admitted in the other realms of nature. Whenever certain elements combine and thereby produce, by the fact of their combination, new phenomena, it is plain that these new phenomena reside not in the original elements but in the totality formed by their union. The living cell contains nothing but mineral particles, as society contains nothing but individuals. Yet, it is patently impossible for the phenomena characteristic of life to reside in the atoms of hydrogen, oxygen, carbon, and nitrogen. (…) Let us apply this principle to sociology. If, as we may say, this synthesis constituting every society yields new phenomena, differing from those which take place in individual consciousness, we must, indeed, admit that these facts reside exclusively in the very society itself which produces them, and not in its parts, i.e., its members.

Durkheim’s focus here remains on “social facts”, which have a coercive power over any individual. Our argument extends Durkheim’s insights to include the nature also of ‘thinking’ and ‘mind’. When a child emerges that has been brought up or lived most of its life with animals we readily accept that it has learned to see the world as an animal, we do not ascribe to it an ordinary human mind. However, when it comes to ordinary members of society we tend to insist that the individual must somehow be pre-equipped cognitively to enter

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   49

social relations. We turn this common sense view upside down and instead start with the proposition that for the scientist as for the participants themselves “thinking” or having the “mind” of a human being starts with the concrete facts of social interaction, which is initiated at the very moment of birth, or perhaps even before. It is our hope that by introducing our approach in this forum, our ideas will not simply be perceived as alien but as sources of fruitful debate between two major currents of research in sense-making that have existed for too long in mutual agnosticism: what we may tentatively call the “cognitive” approach (including social cognition) and the “sociological” approach. 1. The concepts of behavior and cognition in the study of sense-making in interaction Face-to-face interaction is a seminal human activity. It fills and shapes us and our lives from the minute we are born to the moment we die. During face-to-face interaction, an extremely complicated process of shared sense-making is taking place. Since the rise of modern linguistics in the early 20th century, attempts at describing and understanding what is typically labeled “discourse” or “communication” have been defined by scholars in many different disciplines which acquire their profile from their specific focus on aspects of the verbal communication. This means that the development of human sciences have divided the – for the participants themselves – integrated communication process into numerous more or less independent fields of inquiry. This process of segregation and specialization in aspects of verbal communication has created a permanent disintegration of naturally coherent phenomena which researchers coming from many different fields now paradoxically strive to remedy to understand their own specific focus better. To mention just a few prominent ones, disintegrations include propositional meaning from context (of all kinds), language from society, cognition from situated behavior, language from other bodily movements (such as gesture), and the individual from the social group. In this paper we propose an approach to face-to-face sense-making which makes no a priori divisions or hierarchical organizations of resources of face-to-face-interaction, nor does it a priori exclude any kind of bodily movement as a resource of sense-making. Our approach is firmly rooted in the behavioral, micro-analytic frameworks of ethnomethodology (EM) and conversation analysis (CA) but, in contrast to these current approaches (discursive psychology (DP), included) we do not exclude (some notion of) cognition as a describable feature of sense-making in face-toface interaction. In fact, a central part of our scheme is to be able to explore the nature of intersections between social behavior and cognition (again without

50   Anders R. Hougaard and Gitte R. Hougaard being committed to particular concepts and terminology from the cognitive science fields). That being said, in line with EM and CA thinking we do not make a priori assumptions about the nature of the resources of face-to-face-interaction, we do not make specific a priori assumptions about the specific role and nature of cognition either, and in the vein of EM and CA we exclude cognitive notions as analytic categories or points of departure. Instead we see them as possible upshots of our studies, although possibly and likely conceived of in ways that are radically different from mainstream cognitive science. The nonspecific assumptions we make are then that without assuming any directionality social behavior is essentially and inextricably connected to some sort of “cognition” in and through the actions, physiology and neurology of the interlocutors. It is evident that interaction is not just sets of behavior; it is concerted, senseful actions produced by whole, mutually oriented, knowledgeable bodies. Face-to-face interaction literally connects and integrates bodies in and through orientation, posture, coordination, touches, vision, sound waves, artifacts and smell, and these physical conditions are complemented by an inescapable consciousness of each other. These are bare facts. There is nothing controversial about them. What we attempt to do is to take them seriously and to set ourselves free from the mainstream cognitive science and psychology notion that there has to be individual cognition before interaction and the mainstream social interaction notion that cognition has no place in empirical, sociologically based and procedurally focused studies of sense-making. We consider “fused bodies” as an appropriate framework for this endeavor. In this paper we develop the fused bodies approach programmatically by discussing basic assumptions and hypotheses and the approach’s position on the map of approaches to sense-making and analyses of data. The paper comes in two main parts: a part that introduces the fused bodies notion of sensemaking (section 2) and a part that discusses how that notion is related to main stream notions of cognitive and social science (section 3). First (section 2.1) we discuss the historical, scientific disintegration of communication into fields specialized in its constitutive features. Then (in section 2.2) we show how faceto-face sense-making involves many more resources than are typically focused on in semantic, pragmatic and discourse analyses. In section 2.3 meeting the likely critique that the fused bodies approach falls into the trap of just including “everything” we show how interactional situations contain a set of attended-to (by the participants) systematic actions, which are the ones that our analysis focuses on. “Everything” is potentially of relevance. However, only when given observable attention by the participants themselves, do actions and things or circumstance in the milieu become objects of analysis. Section 2.4 discusses the phenomenology of social action and basic, rudimentary fused bodies con-

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   51

cepts of knowledge and mind. The analysis of a fragment of interaction between an aphasiac patient and her therapist (section 2.5) provides an example of how co-participants orient to and co-construct moments of “falling out of interaction”. Hence the fragment provides empirical support for a basic fused bodies hypothesis that people fundamentally orient towards other people and their being or not being potentially attentive to oneself and/or to interactional activity. With this we emphasize the special orientation that people have towards people, which is one precondition for the fused bodies hypothesis that our way of having a world, having knowledge and being minded is based on our social orientation. Section 2.6 concludes the first main part of the paper by providing a first fused bodies definition of sense-making as a feature of “units” in time and space constructed by and composed of whole, interacting, knowledgeable bodies. In the second part of the paper (section 3) – Fused Bodies in Context – we very briefly discuss the position of fused bodies with respect to the notions of ‘embodiment’, ‘individualism’ and ‘dualism’ in major approaches to mind and sense-making. A glossary of transcript conventions is appended at the back of the chapter. Describing the ideas and assumptions of the fused bodies approach and situating it on the vast map of philosophies and sciences of sense-making is a huge task which will eventually require a whole book. Our purpose in this article is neither to offer an exhaustive or practical introduction to the fused bodies approach nor to present fully developed ideas and thoughts. We consider ourselves to be on the way towards an approach and the work that will maturate and test our approach still lies ahead of us. Yet, having introduced and discussed our approach at several conferences and workshops over the last couple of years we think the time has come to present our program in writing. 2. Sense-making 2.1. The disintegration of language and social interaction The study of aspects of verbal communication1 is – as the chapters in the present volume testify to – spread out across numerous fields and subfields. 1 When we refer to the program of the fused bodies approach, we use the term ”social interaction” to indicate our focus on naturally occurring face-to-face-interaction. The term ”(verbal) communication” is meant as a cover term for all studies which deals with some aspect of communication, but does not necessarily focus on naturally occurring communication or face-to-face-interaction.

52   Anders R. Hougaard and Gitte R. Hougaard Like other modern sciences, the study of social human sense-making has seen an ever increasing specialization. Since the time of Sapir and Whorf (Sapir 1929; Whorf 1956) dominant schools in linguistics have extracted systems of language structures from communicative situations and studied these in their own right. And within linguistics, language is split into the study of sound representations (phonology), sound production (phonetics), words (for instance lexicosemantics), word composition (morphology), word combination (syntax), sentence combination (discourse) and more recently neural correlates of linguistic phenomena (as in neurolinguistics and cognitive neuroscience). In particular, generative linguistics and its sisters in formal semantics (Montague 1974; Davidson 1967, and others) and the modular philosophy of mind (e.g., Fodor 1983) have shaped this development. Language acquisition studies have focused on these issues developmentally. Following the rise of pragmatics with Austin’s lectures (1962) and Searle’s speech act theory (1962), many linguistics schools of the latter half of the 20th century have partially integrated the study of the language system with its many contexts. Discourse analysis has gone beyond the single sentence to consider how language is organized in larger chunks in a social context. Sociolinguistics examines social variations in language. Functional linguistics studies language as a function of communicative needs and purposes. Cognitive linguistics studies language in terms of its embedding in general cognitive and perceptual systems. And most recently several cognitive linguists have called for at social-cognitive agenda embedding the study of language in both a cognitive and a social framework (for instance, Sinha 1999 and 2005; Croft 2005; Geerearts 2005; Bernárdez 2005). However, none of these typically analyze empirical instances of social interaction. They tend to interpret “social” as “pragmatics” or as “social variation” at a macro level. Finally, the field of social cognition has focused on how individual participants understand themselves and others and whether such social cognitive processes belong to cognition in general or are of a specific kind. Other disciplines have focused almost entirely on the social (“outer”) and contextual aspects of social interaction, being agnostic about or even denying the role of “mental” phenomena in social interaction. As mentioned above, EM and CA and daughter disciplines such as DP, focus on the details of naturally occurring everyday social organization and sense-making without reference to anything but observable, non-mental, systematic patterns of behavior. These disciplines do not only focus on language, but in principle on all sorts of resources of sense-making, including such things as gesture (Goodwin 2000; Kendon 1980, 1983, 1995, 2004; Laursen 2002; Müller 1998), laughter (Jefferson 2004), body posture (Kendon 2002) and crying (Hepburn and Potter, in press). These studies are then examples of technical specialization in one resource of sense-

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   53

making, and though they shed light on important aspects of social interaction which tend to be overlooked in the general, traditional focus on ‘language’ (as words) as the primary communicative device, and though they often emphasize that their division of interactional resources is only analytic, they still tend to disintegrate the social interaction in and through their technical specialization. It is not the purpose of this section to question the importance or insights of the work of the people and schools that we have mentioned. As already mentioned above and as will become even clearer in the sections to follow, we are heavily indebted to much previous work in EM and CA as well as their philosophical and sociological roots in phenomenology: Schutz (1962, 1964) who, among other things, discusses the necessary common sense assumptions that participants must make in order to engage in senseful social interaction, Wittgenstein – who (according to Williams 1999) presents a more truly social approach to intersubjectivity (compared to Schutz), Durkheim, whose work on “social facts” is presented as a main inspiration in the very title of the recent republication of Garfinkel’s founding Studies in Ethnomethdology: Working out Durkheim’s Aphorism (2002, [1967]). As mentioned in the introduction, we would like to point out the historical irony in the development of disciplines that focus on aspects of social meaning construction. A strong current tendency in disciplines that have to do with human sense-making broadly construed (including linguistics, cognitive science, ethnography, psychology, sociology, anthropology and more) is to develop so-called “interdisciplinary” approaches which connect one’s specific focus with other specific foci in order to achieve a more complete picture of the phenomena that one is studying. A few examples will suffice: linguistic anthropology, social neuroscience, social cognitive linguistics, cognitive science and interactional linguistics. In addition to these established fields “interdisciplinary” centers, labs, conferences and collaborations crop up everywhere with an increasing frequency. Hence, while initially anthropologists and others seem to have concluded that social life and sense-making is so complex and rich as to require specialization, there now seems to be a general realization that the decontextualization that follows from much specialization in fact artificially isolates resources of social interaction from the unit of sense-making. It splits up the natural, phenomenological unit that the sense-makers themselves produce and experience in sense-making situations. The historical irony is that decades of specialized focus has generated specific methodologies and specific vocabulary for isolated aspects of social interaction so that it is now a great challenge (perhaps insurmountable) to integrate these. Of course specialization is at various points necessary and there are indeed foci which do not naturally come together but which due to the very nature of the phenomena under scrutiny require much work to integrate.

54   Anders R. Hougaard and Gitte R. Hougaard The latter include for instance behavioral studies and studies of the neural underpinnings of behavior. Neural substrates do not lend themselves to direct translation into behavioral patterns or vice versa. Often, however, it is one’s philosophical stance that determines whether specialization or an integrated approach is preferable. A famous example is the generativist modular view on mind and language and its specialized focus on syntactic structure as isolated completely from meaning and context. We propose an approach that studies face-to-face interaction without a priori analytic divisions of aspects or resources of sense-making and without a priori exclusion of modes of description and explanation. Hence though our studies are based on a micro-analytic, behavioral basis, we do not see social sense-making as ultimately a purely socio-behavioral phenomenon. Thus in contrast to mainstream EM and CA, we consider sense as an achievement that participants accomplish in and through their social behavior and cognitive processes inherent therein. It is very important to emphasize that by the terms “social behavior” and “cognitive processes” we mean the entire set of behaviors and resources that are brought to bear on the interaction and which are understood by the participants as being “interactionally set” (we will return to that notion in section 2.5.1). “Social behavior” in this approach is not (just) language, a social “face” or role play (Goffman 1967) or gesture, nor are “psychological processes” merely phenomena of individual consciousness that are biologically grounded in activity in the individual central nervous system. The “social” and the “cognitive” are to us inseparable aspects of entire bodies which perform a socio-phenomenological, systematic fusion at every moment of sensible human contact. With this we aim at introducing a sense of embodied sense making, which is different from many mainstream embodied approaches to sense-making. In section 3, we discuss fused bodies in connection to other influential notions of “embodiment”. In section 2.3 we discuss how this approach is constrained and how it avoids becoming a study of “everything”. 2.2. Resources of sense-making: how the body has a social world So far we have only indicated in more general or theoretical terms how sensemaking in face-to-face-interaction relies on a great number of resources. In this section, we illustrate this point in more detail. In the early 1960s anthropologist Margaret Mead (1964) introduced the notion of ‘total communication’ as a method for finding a shared and equal basis for interacting with cultures whose languages she was not familiar with. The idea was to use any alternative resource of communication to avoid the possible cultural dominance following

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   55

from using one part’s language. Later the notion was adopted as a basis for teaching deaf people. They were supposed to be taught their own “language”, sign language, and not be a communicatively deprived minority. ‘Total communication’ may of course in some contexts be an ideologically loaded notion, yet its success testifies to a central fused bodies point: though languageas-words is indeed established (through communicative practices, laws and regulations and folk myths) as the central organizing mode of communication and as the carrier of meaning, there is nothing inherent in language-as-words which makes it inescapably so. Linguists and psychologists have long entertained the idea that language-as-words is what separates humans from other primates. Language-as-words is thus seen as quintessential to what it means to be a human being and in communication everything else is considered as a context of language. This, we believe, is a wrongheaded a priori assumption that cannot have taken into consideration the actual richness of devices of sense-making in face-to-face-interaction or their actual complex and dynamic, locally constructed interrelationships. Face-to-face interaction is simply naturally occurring total communication between natives. Consider the following example, for which we have provided four still pictures (the numbers under the picture correspond to the numbers in the transcript): Excerpt (1)



1

2

3

4

56   Anders R. Hougaard and Gitte R. Hougaard (1) (2) (3) (4)

*SEI:=°je’ me:ner°, ·hh[1] Ritt Bjerregaard sender os et manuskript ‘I mean Ritt Bjerregaard sends us a manuscript’ fuldstændig frivilligt¿ [2] (0.2) ·hh ø:: stiller det til ‘completely of her own free will ehm puts it at’ rå:dighed for os=i øvrigt [3] oss’ til andre avisredaktio:ner; ‘our disposal and by the way at other newpapers’ disposal’ = ønsker den[4] maksima:le ekspone:ring af det¿= ‘wants maximum exposure of it’

When one hears these sounds – especially when in a foreign language, you realize that these are not words that carry meaning, but sounds produced by a body in interaction. Sounds are, however, just one thing that the body produces in interaction and which create bodily effects in the recipients – with the sounds go gestures, facial expressions, gazes and bodies draw on materials from the surroundings. To understand all of these ‘movements’, and thus what the ‘other’ is doing, you need specific techniques in and through which you systematize them. Many studies have been and are carried out on the ways in which gesture is organized as an activity, of how it is organized in relation to speaking and how it contributes to the total meaning of utterances of which they are a part (see, for instance, Goodwin 2000; Kendon 1980, 1993, 1994, 2000, 2004; Laursen 2002; McNeill 1985; McNeill and Duncan 2000; Schegloff 1984; and Streeck 1993, 2002). In most of this work, speakers are seen as using “gesture as an integral part of the act of producing an utterance. An utterance is looked at as an ‘object’ constructed for others from components fashioned from both spoken language and gesture.” (Kendon 2004: 5). Another study by Kendon (2002) shows that a headshake is not to be understood simply as the kinesthetic equivalent of a unit of verbal expression. Rather, a headshake is an expression in its own right, a modality of expression available and deployed by the speaker in the course of building a unit of expression. Streeck (1993) shows a pattern in the coordination of gaze and gesture: speakers turn their gaze at their gesture (iconic gesture) before uttering a specific central word that is semantically coherent with the meaning of the gesture. Bavelas, Coates and Johnson (2002) have studied how speaker’s gaze coordinates the interactional work between speaker and recipient and Goodwin (1981) shows how speaker while speaking orients towards receiving recipient’s gaze (see also Kendon 2004). As mentioned above, bodies also draw on materials from the surroundings to make sense. Streeck (2002) studies the gestures made by a car-mechanic during a single extended episode of work. The gestures are recurrent and are parts of the repertoire of communicative forms with which the car mechanic responds to the understanding of the tasks

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   57

that routinely occur in his work. The gestures “reveal the specific stance and knowledge that the body making them has acquired in its day-to-day coping, the body’s being-in-the-world.” Thus, Streeck in his study focuses on the gestures that articulate the “grasp that hands have of their life-worlds.” (2002: 19). Following LeBaron and Streeck (2000) and echoing with the phenomenologist Merleau-Ponty (1945), we consider the “grasp” of the hands as worth studying in its own right as these are resources with which bodies understand in the first place. Or to put it in Streecks (2000) words: “It [the word] is known in the first place through acts of physical manipulation (Hastrup 1995), and in this regard the knowledge that a person possesses about a given life-word is “embodied knowledge” whose criteria do not reside in the ability to adequately represent it, but in the skillfulness with which it is handled.” (2000: 27). He continues: “Our gestures show the world not as it is, but as we understand it.” This understanding is acquired, according to Streeck, through handling, exploring and doing things in the world. We could go on and discuss ever more subtle resources that bodies may use and make sense of in interaction, but the point is that the body makes use of many different ‘movements’–sounds included. The research field of multi-modal communication in many cases focuses on one – for the purposes isolated – modality apart from language, gesture for instance, which is then defined in relation to or in terms of language-as-words. Hence, gesture is in some studies regarded as identical to language (Armstrong, Stokoe and Wilcox 1995), as a part of language or as a form of language in its own right (for a discussion, see Kendon 2000). Kendon (2000) regards gesture as a mode of symbolic representation, as spoken language, and argues “that it is through the partnership between gesture and speech that we see so often is co-present in conversation, that utterance meaning is achieved.” (2000: 50). In contrast to studies that locate gesture in the process of an utterance, LeBaron and Streeck (2000), again echoing Merleau-Ponty, claim that their data demonstrate that conversational hand gestures ascend from ordinary, non-symbolic exploratory and instrumental manipulation in the world of matter and things, and that the knowledge that the human hands acquire (both individually and historically) in these manipulations is realized through and brought to bear upon the symbolic tasks of gestural representation. Ultimately, it is through these indexical ties of gestures to the material world that gestural signifiers can be seen and recognized. (2000: 119).

Though methodologically and phenomenologically in contrast to these studies that are mostly interactionally informed or, as McNeill (2000) puts it, has an “outside” view on linguistic code and gesture, research in the vein of cognitive

58   Anders R. Hougaard and Gitte R. Hougaard psychology is also carried out on multi-modal communication. Here emphasis is on the processes “inside” individual speakers and listeners, and focus is mostly on theories for explaining the dynamics of utterance formation. Also within this ‘inside’ approach two views can be sketched out, namely one that regards gesture and speech as united into “single idea units” (McNeill 2000: 140) and one that conceptualizes gesture and speech as parallel but basically separate streams of cognition (Kita 2000). As already argued in the introduction and as will be discussed further in the following sections we do not support a dualistic or an ‘inside/outside’ view on communication or sense-making. Further, we attempt to avoid the common sense view that sense-making is achieved through an understanding of isolated modalities like speech, gesture, headshake and so on. Neither do we assume again that some ‘movements’ are more central than others. Instead, we suggest that understanding or sense is a social, integrated whole. Ultimately, our view on interactional movements as not a priori hierarchically divided, but as integrated wholes, lead to a reconsideration of the very notion of multi-modality. Multi-modality emerges from analysis; it is not a prerequisite of participants’ sense-making, nor ought it be a guideline for analysis. Due to length limitations, we will not dig deeper into this issue here. We consider the resources of sense-making as ways in which the body has a social world, and language is but one of them. But, more than this our hypothesis is that the shared social being in and having a world precedes Husserl’s phenomenological subject’s having a world. In other words, intersubjectivity, we propose, is not a superstructure on individual reflection on the world, but instead a fundamental influence on and basis for the individual. We do not consider this a defining assumption of the fused bodies approach, but a basic hypothesis which of course follows specifically from programmatically “situating” sense-making in the unit produced by interaction bodies. 2.3. The systematics of body-movements It could be argued that an approach which incorporates entire bodies will fall into the trap of incorporating everything. However, sense-making in the framework that we suggest is dependent on recognizability. To achieve recognizability the ‘movements’ have to be organized in specific ways. One central principle in the organization hereof is the place or the position in which the ‘movement’ occurs in relation to other ‘movements’, that is in relation to its context. In this context the ‘movement’ is understood as a social action. (Language is of interest in so far as it is viewed as playing an integral part in the

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   59

enactment of social action and communication.) In EMCA the relation between social actions on one hand and the contexts in which they occur, is viewed as being characterized by reflexivity: by this is meant that social contexts are constructed and understood through social actions on one hand while on the other hand the actions are understood through these contexts. By context we then mean local, sequential context. Local context is verbal and/or nonverbal actions that have been produced previous to some current action as for instance exemplified in the following excerpt: Excerpt (2) (1) S: (2) D: (3) S: (4) D: (5) S: (6) D: (Blood4b2)

kommer du li fra arbejde af nu her ‘are you coming here directly from work?’ ja du ka måske lugt at je lugter lidt af gris hehe ‘yes, perhaps you can smell that I smell like pig’ he nej det ka je godt nok ik ‘he, no I actually cannot’ nå hehe ‘ok, hehe’ hehe er du landmand da ‘hehe, are you are farmer then’ ja ‘yes’



Line 3 is only understandable as an action in the context of line 2 which again is understandable in the context of 1. So, every action is shaped by its local context as it itself renews that context and thus becomes the context of a subsequent action. The placement/organization of actions in sequences is considered as one technique among many others that participants in interaction make use of to make sense. These methods and techniques are especially interesting in so far as they are found to occur systematically in interaction. This means that the techniques are found in a variety of interactions organized by different participants at different times.2 The reason for this is that we are interested in actions, and techniques that are recognizable to the participants as specific techniques 2 Actually CA is interested in the generic structure of interaction (the “machinery”) and therefore interested in techniques that are not only found in a variety of interactions organized by different participants at different times, but also organized across gender, age, culture and so on.

60   Anders R. Hougaard and Gitte R. Hougaard in and through which a specific understanding is achieved. And recognition is achieved through systematically used actions and techniques in social encounters. The boundaries of these encounters are constantly (re)established by the participants in accordance with the contingencies of any given moment in the interaction.3 In order for participants to construct actions and techniques (systematically) that they take to be recognizable to their coparticipants, they must, according to Schutzian social phenomenology, make specific everyday life assumptions. The participants in interaction – described by Alfred Schutz as practical theorists – assume among other things that: 1. Objects of the world are what they appear to be. The person coping with everyday affairs could as a matter of fact exercise a rule of doubt about any object of the world that they are as they appear to be. Instead the person assumes that things are what they are – to him. He expects that the coparticipant employs the same expectancy in a more or less identical fashion and expects that, just as he expects that relationship to hold for the other person, the other person expects it to hold for him. 2. An intended object is the same object now as it was intended in the past and can be intended again in the future despite the facts of time sampling, and changes of context, circumstances, and actual appearances. If each participant were to exchange places with the other, each would recognize the scene in a manner that was for all practical purposes of their interaction more or less similar. 3. Differences in perspective that the person knows originate in his own and in the other persons particular biographical situations are irrelevant for the purposes at hand of either, and that both have selected and interpreted the actually and potentially common objects and their features in an “empirically identical” manner that is sufficient for all their practical purposes. It goes for all the participants’ assumptions that whatever is assumed the participant assumes that the other participant assumes the same and that the other participant assumes the same for him. Phenomenology is traditionally concerned with psychological and cognitive issues. However, such mental issues are in EM and CA translated into a phenom3 The establishment of boundaries of encounters are also described in terms of “frame attunement” (Kendon 1990). Frame attunement is seen as the mutual coordination of locally relevant situated identities and the conjoint establishment of a local space that makes it possible for an encounter to take place. Our understanding, following Hutchby 1999, is that “interactants constantly attune and reattune their frames according to the contingencies of the moment.” (1999: 42)

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   61

enology of social, collaborative activity. What is at stake is not how the individual Husserlian subject has a world, but how a world is had in concert by coparticipants in and through mutually directed actions and joint sense-making. Accordingly, our framework suggests that not only talk but all body-movements (including talk) are described in terms of such interactional procedures. However, the framework suggests that these procedures are not only a social but also at the same time a psychological and cognitive achievement. The psychological and cognitive issues are though not re-embedded in the individual but are, as we will argue in more detail below, part of a larger situational unit of joint sense. 2.4. Body movements, knowledge and mind The systematics of body movements suggests an understanding of the movements as based on knowledge. That knowledge may be mere knowledge by familiarity, i.e. practical knowledge (Kennen or Connâitre) or theoretic or formulated knowledge (Wissen or Savoir). As is the case with regard to the former kind of knowledge, an actor does not in the latter case merely “act in a context” (as does the child who knows how to behave) but “links the act to its context” (that is knows why he does it) (McHugh 1970). Still, also acting in a context is knowledge in that, for instance, the child is aware of his milieu and the activity. He is thus, as ethnomethodologist McHugh formulates it not “a robot or a dope”. McHugh goes on: “A practical actor, as with children, can attend to his circumstance in the sense that he could record his experience, write it in a letter, tell it, draw a picture of it, or the like. He can even do these in a way that satisfies him and in a way that could be made intelligible to others. (1970: 76).” The practical knowledge is in other words the knowledge that is acquired in and through our bodily engagement with the material world. In contrast to a third kind of knowledge, innate bodily knowledge, described by Damasio (1994) which includes such things as metabolism and instincts, the former is a kind of knowledge that we are not born with, but which is acquired through experiences. All kinds of knowledge (innate and acquired) are according to Damasio (1994: 122) interrelated in the sense that the neural circuits that we are born with regulate and are regulated by new neural systems that are developed with new (social) experiences. The interaction between the old and the new circuits helps us survive in the world that we are born into. There has been (and is) a tradition within anti-Cartesian cognitive science for subsuming everything from basic perceptual processes to so-called “high-level” cognitive operations as reasoning and logic under the same heading as “cognition” or “mind” (see, for instance, Ulric Neisser’s all-encompassing definition of cogni-

62   Anders R. Hougaard and Gitte R. Hougaard tion in his classic 1967 book Cognitive Psychology), and indeed to see the basic bodily/innate processes as somehow responsible for the nature of the higher-order processes (see Lakoff and Johnson 1999 for a very prominent example of this line of thinking). However, we see a fundamental problem in drawing a straight line from inescapable perceptions of the world to human concepts, with the former as some form of originator of the latter and as if this process took place more or less automatically, transforming perception directly into conception. Not only do phenomena of different orders thus get mixed in a way that does not explain anything but simply deletes the problem (of relating body and mind), it overlooks the basic fact that most of our knowledge as individuals is not individual but acquired in and through social interaction with and about the world we are a part of. Thus sharing the general sentiment (but not necessarily agreeing on the details) of precursors such as the later Wittgenstein (1953) and Coulter (1979) we propose a conception of mind which is fundamentally social. Innate knowledge or inescapable perception does not just become conception or theoretic or practical knowledge as a consequence of how our cognitive “system” works. Inescapable perception and innate knowledge become conscious reflections in and through the social determination of them, and thus cognition is social and mind is social. 2.5. Bodies’ social mind Our hypothesis rests on the premise that human beings have a particular orientation towards fellow human beings as opposed to other potential resources of understanding the world or having a world. It presupposes that interactional availability or potential availability is a “thing” to people as opposed to just being with other people and not being oriented towards their potential orientation towards oneself or the (potential) implications of their presence at a moment for one’s being in that moment too. Furthermore, our argument presupposes that we can also know (that is learn and then recognize) when others orient towards us or a shared activity that we conduct with them. As we will show in the next section, empirical data shows that this is the case. Co-participants do orient towards and co-define each other as interactionally set or interactionally out of the loop. The entire body plays a role in such basic concerted acts, thereby underscoring the importance of seeing sense-making as a matter of the whole body. 2.5.1. Is there any contact? Sense-making human beings know when and how body movements are normative or ‘practical’ resources for the construction of meaning. In other words,

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   63

they know when they are interactionally set. They also know when co-participants no longer work on making sense, when they “fall out“ so to speak or when there “is no contact“. In our daily work we have observed this in data with people suffering from schizophrenia interacting with nurses and phychiatrists, with people suffering from dementia interacting with nurses and people suffering from aphasia interacting with speech therapists. The following fragment exemplifies the phenomenon.4 The co-participants are a woman suffering from aphasia and a speech therapist. The woman in the clip is typically very lively when interacting in that she constantly makes use of head movements, facial expressions, gestures, talk and other body movements. She suffers from a “light“ kind of aphasia which means that her problems with expressing herself are clear enough to be noticed by co-participants. This, according to the woman herself, is a social problem as co-participants recognize that there is something slightly strange about her behavior – something that they cannot explain. When looking closely at the data, one can observe that the woman makes use of different strategies when she has problems expressing herself. These strategies are very conventional: she indicates that she searches for words through so-called self-repairs (Brouwer 2003; Schegloff, Jefferson, and Sacks 1977) and/or she invites her co-participant to help her find a word. This is shown in fragment 3, for which we provide a series of stills showing some of her movements (again the numbers by the pictures correspond to the bracketed numbers in the transcript): Excerpt (3) 1. right hand gesture:

2. head movem.:

4 The data has been collected by Mette Nedergaard Paulin, Tine Eltang Petersen and Tine Kappel Petersen.

64   Anders R. Hougaard and Gitte R. Hougaard 3. right hand gesture:

(1) A: Com (2) (3)



Com (4) Ps Com (5) A: Com (6) A: Com (7) A: (8)

B: Com

det er jo os lidt [1] det der er probl°m:et n:år når° telefonen ringer ‘there’s also this problem when when the phone rings’ puts out her hand, palm upwards, keeps it there derhjemme ik’ at jeg ik ka se (.) hvem der ringer (.) ‘at home right that I cannot see who is calling’ jeg har faktisk overvejet at få sån en øh ‘I have actually thought about getting one of those ehm’ hand back in lap, wringles forehead, gazing to the left (3.4) head moves to the right, turned down, gazing to the left [2] sån en hvor man ka se hvem ‘the kind of thing where you can see who’ puts out her hand pointing upwards, palm up [3] (.) det er der ‘it is that’s’ turns the head and looks down in front of her [ringer hvad hed- ] [‘calling what’s it cal-‘] [ja sån en nummerviser] [‘well a number display’] A directs her gaze at B

In lines 1–5 A repeats lexical items, makes several macro and micro pauses and reformulations. Her various activities are understood and interactionally co-defined by B as a search for a word. So, B delivers that word in line 8 which occurs in overlap with line 7 in which A invites (with “what’s it cal-…) B to help her finding that word. As can been observed in the fragment, A is very “active”, moving even during the pauses and gesturing. That is, in our understanding, she is still orienting to the interaction and the task of making sense intersubjectively. However, this orientation is absent in the subsequent sequence in which A in the pause in line 11 stops all movement and for a short time appears “absent”:

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   65

Excerpt (4)

(9)

A:

(10) Ps Com (11) A:

Com

(12) A: Com (13) A: Com (14) A: Com (15) A: Com (16) A: Com

ja ‘yes’ (0.6) A sniffs and looks down ø:hm (2.6) eehm Gazing away from B, down to the left, no wrinkles in fore-head, face non-expressive, “empty” look. Then the head is moved to the left and again down to the left fordi bare det ‘because if only’ looks at B at jeg så ka nå og indstille ‘I have time to adjust’ index finger moved back at the side of the head det ka jeg jo se på min: I can see that on my both hands taken to the left, gazing shortly to the left mobiltelefon (.) /der generer mig det jo ik’ ‘mobile phone there it doesn’t bother me’ ‘B nods, A puts out hands in front of her, one over the other, palms up’ (0.5) B shakes his head

The absence (see the pause in line 11) that is embedded in the interactional searches for words is, we assume, probably what people, as reported by A, find strange. Her co-participant also orients to that absence. Though this is not made transparent in the transcript, B is moving a) when he is talking himself and b) when A is talking, though to a lesser degree. However, when A becomes “absent”, B is almost completely motionless, obviously as a speech therapist waiting for A to find a way to solve her problems. Hence, though still otheroriented, that is, though still alert to the potentially interactive nature of the situation shown by the speech therapist by his continued gazing at the patient, the bodies have for that moment in coordination been put in a non-interactional

66   Anders R. Hougaard and Gitte R. Hougaard mode. This indicates that the bodies understand and thus know when something is done in relation to the world and when not. That something is, in our approach, considered as potential resources of sense-making in interaction. 2.5.2. Sense-making – a product of interacting knowledgeable bodies Sense is a social accomplishment. Usually, as ordinary competent participants in interaction sense just comes to us, we tend not to take notice of the fact that for that to happen a lot of work is required by our knowledgeable bodies in and through interaction. In that sense, intersubjectively recognized meaning is truly an achievement. As is the case with EM and CA, we are interested in capturing the participants’ intersubjective understanding of their interactional actions. To take an obvious example, let us look at line 2) (Excerpt 1) again. This action is understood by S as a question which S displays in the subsequent turn (l. 3) in which he produces an action that is understood as an answer. Notice the method used here (one among others within EMCA) for capturing the participants’ understanding of what is going on, This method is called next turn proof procedure. Understanding is thus achieved in and through the placement of a particular movement (sounds, expressions, gestures etc.) in a specific position (context). Furthermore, the socially achieved understanding is the product of the work of interacting knowledgeable bodies. Knowledge is then embedded in social, bodily action and is constantly being (re)constructed, (re)run, (re)done, and modified along with the experiences that the bodies have. Experiencing is what bodies do – when they interact. Streeck argues, as discussed above, that the single body grasps the world with its hands (Streeck 2002: 19) and that gestures rely on such grasping. We argue that bodies also experience and grasp the world when they make sense in interaction. That is they grasp the interactional lifeworld. Often scholars ask “where is sense then made?”. This question, however, is not equivalent to our conceptualization of sense-making. The question of where is simply, we believe, misguided, perhaps informed by wrong analogy to other sciences. Sciences such as biology, medicine, mathematics and chemistry all consider the question of where as one of their central obligatory challenges. Biology can map the circuits of hormonal processes, medicine can locate organs and sources of disease, mathematics places great emphasis on chronology and linearity and chemistry can locate the presence or absence of certain molecules. But, sense is not a single, concrete or abstract object in space, and therefore it is of course misleading to think of it as such. Thus, instead of asking where sense is, we propose to ask what sense is and how it is achieved. A short answer to the former question is that sense-making is the total unit at any given moment of mutually oriented, interacting (in terms of made-relevant move-

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   67

ments), neural and physical, knowledgeable bodies that – to answer the latter question – in concert systematize their movements. Sense then, to us, does not have a location as such, but it is observable. Think of a pair of skilful dancers. Think for instance of an experienced tango (Figure 1). Each dancer is constantly “tuned in” on his or her partner and every movement is done in precise coordination with the movements of the other. In and through their coordination and timing they become one whole, a gracefully moving unit which is two and one at the same time. A skilful tango is not different in kind from everyday interaction, only in quality. The tango performs specific sorts of beautiful movements which it takes time to learn and it focuses on a passionate contact between the dancers. Everyday interaction shares with it the human talent for social timing and coordination, the ability to act as two in a graceful unit. Some mistakenly think that communication is only a joint, timed and coordinated activity when the participants are in agreement with or like each other. That is wrong. Arguments, fights, and quarrels are performed in the same concerted way displaying the shared sense of being against each other and acting in accordance therewith. Otherwise, the situation simply would not make sense.

Figure 1. Bodies fused in tango (Courtesy of: http://www.easybuenosairescity.com)

68   Anders R. Hougaard and Gitte R. Hougaard We propose to think of and analyze sense-making in terms of mutually oriented-to and coordinated actions of knowledgeable bodies in space and time which can be analyzed systematically. These fused bodies sense-makings we hypothesize to be responsible for all that is characteristic to us as social human beings: mind, cognition, knowledge, society, language, norms, and so on. They are what makes us fundamentally social and provides for us an intersubjective way of having and being in the world. Contrary to Durkheim, but with EM and CA we do not assume a static common mind or shared knowledge. We instead assume that co-participants in sense-making are knowledgeable and thereby mindable, that is able to act knowledgeably and mindedly in time and space. This knowledge concerns all that may be made relevant to and recognized during sense-making, including of course knowledge of the very procedures of sense-making themselves. Hence, we also disagree with any cognitive account of knowledge and mind that presupposes static inventories “inside” individuals. Such accounts are purely speculative and non-beneficial to our understanding of sense-making. We also disagree with any cognitive account that studies behavior then only to try to explain it by reducing it to “abilities”. We do not rule out the notion of cognitive or psychological abilities. It is evident from our discussion of knowledgeability that we assume abilities of this kind (which does EMCA too). Instead, what we do is simply to repeat the phenomenological dictum of staying with the thing itself. Most cognitive science is based upon the common sense understanding that human behavior is generated by abilities or competences. If people make a metaphor, it is assumed that there must be a metaphor competence. Chomsky insisted that language behavior must be rooted in language competence (1957). Intersubjectivity is by psychologists seen as an indication that children develop capacities for intersubjectivity; for instance, Tomasello (1999) has argued that human beings must have a unique ability to establish joint attention. And one could go on. The point is not that all these claims are wrong, but that it seems to be impossible for much common sense cognitive or psychological thinking to see the situation of sense-making itself as the cause of the behavior observed as well as its products. This may be due to the fact that cognitive science and psychology since the 1950s have done all they can to wipe out any trace of behaviorism. But as a consequence they have also tended to wipe out a crucial insight which is at least as old as the writings of Emile Durkheim and with which we began this paper: that there exists an irreducible and crucial social dimension of human existence which simply cannot be accounted for by cognitive or psychological reductionism.

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   69

3. Fused Bodies in Context In this last part of the paper, we briefly address three key issues in mainstream approaches to sense-making: embodiment, individualism and dualism. 3.1. Embodiment The term “embodiment” has a wealth of uses in cognitive science, psychology, discourse studies, sociology and other fields. Here we situate the fused bodies approach with respect to two main stream uses of the term. EMCA typically uses the term to highlight the live reality of interaction by people who collaborate to make sense. In cognitive science the term has been used to indicate opposition to Cartesian dualism, and highlight the role of bodily experience as the basis of thought. While we are inspired by the first notion we do, as discussed at the end of section 2.2, take issue with the second one. Though making a great effort to open up the solipsist Cartesian mind to its surroundings and despite spending much time on discussing the societal context and impact of embodied thinking (such as for instance metaphor), embodiment in cognitive science has a tendency to end up becoming the study of processes internal to the individual. So for instance neural processes are focused on as the key to understanding metaphor and conceptualization (Lakoff and Gallese 2005), connections in the individual between perception and conception are hypothesized to shape our thinking, and individual judgment is seen as another source of metaphor (Lakoff and Johnson 1999). Hence, though programmatically emphasizing the environment, cognitive science embodiment in this vein tends to become individualistic, solipsistic and mentalist. To us embodiment emphasizes the basic “other-orientation” that people have. We have said that the participants “tune in” on each other. By this we mean that participants attend to each other’s communicative movements (language, gesture, facial expressions) and orientations (gaze, posture) in and through their own movements and orientations. “Tuning in” on each other seems to be a human instinct which is active right after birth. For instance, psycholinguistic studies show that babies attend to their parents’ native tongue, rather than to other languages (sounds) (Moon, Cooper and Fifer 1993). Hence, the baby tunes in on familiar sounds from its surroundings rather than just any sound. Though the baby may not realize that these sounds are communicative, directed towards the baby, these sounds ”stand out” as figures against the background of other sounds. Another instinct experience is reported by one of the authors of the present article. She claims that all of the four kids that she

70   Anders R. Hougaard and Gitte R. Hougaard has given birth to have gazed directly into her eyes shortly after being put in her arms after birth. She claims that their gazing was not just “looking” but “looking into” her eyes and thereby achieving some sort of primary intersubjectivity. This argument can be taken even further. One of the authors of this paper often asks her students to think of the difference between being at home when there is someone else in the house as opposed to being home all by your self, even if the other person is in a different room or on a different floor. The situation, the students report, is not the same to them. We are somehow constantly oriented to the other persons’ being there and the potential of interaction with them. All of this shows that human beings have a very special orientation towards each other. We are more or less constantly other-oriented, either directly in and through our actions or just in and through our conscious state. Thus we are bound to have a deep impact on each other. Our use of the term “embodiment” emphasizes not only the live reality of interaction, but also the constant orientation of bodies towards each other. 3.2. Individualism Scholars who emphasize the role of intersubjectivity in human thought are often met by counterarguments such as ‘no two individuals are exactly the same’, ‘people often disagree or see the world differently’ or ‘much creative work (for instance art, novels, letters and articles) is done by individuals alone’. The fused bodies approach does not deny individualism or individuals. EMCA as well as its philosophical precursor Alfred Schutz (1962 and 1964) acknowledge that co-participants have individual biographies. In fused bodies terms we would say that each individual acquires its personal biography through the myriad of specific sense making interactions she engages in throughout her life. Therefore, because each individual has a different history of interactions, people can have different experiences, differences of opinion, belong to different cultures and so on. And they can also have a sense of being an individual because of their unique history. However, crucially all these things still come from each individual’s being from moment to moment engaged in specific social encounters. Hence, individualism in fact follows from intersubjectivity. Also, when we act sensefully in solitude we are still other-oriented. What we do is for instance meant for others, done as learned from others, as others would do it or definitely not do it, about others, caused by others, and so on. Even when we do things that we intend to be something we “do for ourselves”, this is done within a socially defined frame of practice: going to bed early, getting a good massage, cooking a good meal for one self,

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   71

buying oneself a new dress, and so on. Social life is indeed constituted by people who are individuals because they have different stories but individualism understood as a feature that comes from “within the individual” is but a common sense construal of these biographies. Furthermore, the fact that most of our interactions are with the same people, who in turn have most of their interactions with the same people, who are also the same people as we ourselves have most of our interactions with, means that there is a great deal of overlap in our personal biographies, in the form of ideas, topics, norms, values, etc. The latter we may also call society. 3.3. Dualism and monism Dualist thinking takes many specific forms. What we understand as dualism is not just the Cartesian programmatic division of body/surroundings and mind as separate and irreconcilable matters. In relation to sense-making, we see as dualism any approach which ends up conceptualizing or treating behavior as outer signs or prompts of inner processes (as seen in some discursive treatments of cognition (see, for instance, te Molder and Potter 2005) and in many cognitive treatments of discourse (see, for instance, Fauconnier 1985). Such a conceptualization allows researchers to insist on the appropriateness of studying one aspect of sense-making in its own right, treating other aspects as merely a means to reach the first or simply as irrelevant. In the case of the latter position, the argument may be that scientific accuracy is preferred for “messy”, all-inclusive approaches. However, there is a tendency in for instance CA for such a position to become outright scientific solipsism or eliminativism, in the sense that focus on observable systematic actions goes from being a matter of focus to a matter of believing that sense-making to co-participants consists only in the actions themselves. The fused bodies approach is a sort of attributive monist stance in the sense that we argue that to the participants themselves a sense-making situation consists of many different aspects that are all part of the same total sense achieved at a particular moment. 4. Conclusion In this article we have argued that meaning and understanding are not situated in language (words), no matter whether it is described in terms of ways in which the social world is reflected in it or not. Language together with other resources like gestures, facial expressions and so on is the vehicle of sense-

72   Anders R. Hougaard and Gitte R. Hougaard making. Understanding thus does not emerge from language alone but from many different resources used in a context. However, the context is created in and through the very same resources. The resources are produced in time and space, systematized and thus recognized not by ‘minds’ nor by ‘mindless’ behaving human beings but by acting, knowledgeable and mindable beings that draw upon their entire bodily inventories of movement when making sense. The sense made is the product of the work that the interacting people do when producing and coordinating actions, when ‘tuning in’ on each other. Making sense is thus truly a social achievement. Like sense, knowledge is the product of social interaction. Knowledge of and about the world (including knowledge about how to make sense) is constantly (re)achieved in interaction. Knowledge is thus in all its aspects based upon an intersubjective social understanding. And it follows from this that the human mind is fundamentally social too. However, by “intersubjectivity” we do not mean ‘between subjects’, i.e. individuals. Individuals, or rather individual understandings, do not exist prior to ‘intersubjectivity’. On the contrary, the concept of ‘individual’ follows from ‘intersubjectivity’. The smallest unit of the social world we take to be moments of face-to-faceinteraction. Thus, this is the unit we take as our starting point when investigating how human beings manage to make sense. The work of making sense in such interactions we argue may well be understood in terms of a dancing couple. The dance follows from the social timing and coordination of every single movement the couple makes, from the ability of the two to act as one. We hope the “fused bodies” approach will shed new light on and provide new insights for studies of sense-making, meaning construction, knowledge, mind, language, communication, cognition and social cognition. References Alac, Morana Working with brain scans: Digital images and gestural interaction in 2008 fMRI laboratory. Social Studies of Science 38 (4). Austin, John How to Do Things with Words. The William James Lectures delivered 1962 at Harvard University in 1955, J.O. Urmson (ed.). Oxford: Clarendon. Bavelas, J., N. Chovil, D. Lawrie, and A. Wade Interactive gestures. Discourse Prozesses 15: 469–489. 1992 Bavelas, J., L. Coates, and T. Johnson Listener Responses as a Collaborative Process: The role of gaze. Jour2002 nal of Communication 52 (3): 566–579.

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   73 Bernárdez, Enrique 2005 Social cognition: Variation, language, and culture in a cognitive linguistic typology. In Cognitive Linguistics: Internal Dynamics and Interdisciplinary Interactions, Francesco Ruiz de Mendoza Ibañez and Sandra Peña Cervel (eds.), 191–222. Berlin: Mouton de Gruyter. Billig Michael A psychoanalytic discursive psychology: From consciousness to uncon2006 sciousness. Discourse Studies 8 (1): 17–24. Chomsky, Noam 1957 Syntactic Structures. New York: de Gruyter. Clark, Herbert Using Language. Cambridge: Cambridge University Press. 1996 Coulter, Jeff The Social Construction of Mind. London: Macmillan 1979 Coulter, Jeff 2005 Language without mind. In Conversation and Cognition, Hedwig Molder and Jonathan Potter (eds.), 79–93. Cambridge: Cambridge University Press. Croft, William (To appear) Toward a social cognitive linguistics. In New directions in cognitive linguistics, Vyvyan Evans and Stéphanie Pourcel (ed.). Amsterdam: John Benjamins. Damasio, A. 1994 Descartes’ Error: Emotion, Reason, and the Human Brain. New York: Putnam’s Sons. Davidson, Donald Truth and Meaning. Synthese 17. 1967 Durkheim, Emile The Rules of Sociological Method. New York: The Free Press. 1966 Edwards, Derek, and Jonathan Potter 1992 Discursive Psychology. London: Sage. Fodor, Jerry 1983 The Modularity of Mind: An Essay on Faculty Psychology. Cambridge, Mass.: The MIT Press. Gallese, Vittorio Intentional Attunement: A neurophysiological perspective on social 2006 cognition and its disruption in autism. Brain Research 1079: 15–24. Gallese, Vittorio, and George Lakoff 2005 The brain’s concepts: The role of the sensory-motor system in reason and language. Cognitive Neuropsychology 22: 455–479. Garfinkel, H. 1963 A conception of, and experiments with ‘trust’ as a condition of stable concerted actions. In Motivation and Social Interaction, O.J. Harvey (ed.), 187–238. New York: Ronald Press.

74   Anders R. Hougaard and Gitte R. Hougaard Studies in Ethnomethodology. Englewood Cliffs, N.J.: Prentice-Hall. 1967 Geerearts, Dirk Lectal variation and empirical data in cognitive linguistics. In Cogni2005 tive Linguistics: Internal Dynamics and Interdisciplinary Interactions, Francesco Ruiz de Mendoza Ibañez and Sandra Peña Cervel (eds.), 163– 189. Berlin: Mouton de Gruyter. Goffman, Erving Reprint. Interaction Ritual: Essays on Face-to-Face Behavior. New 1967 York: Doubleday. Original edition, Anchor, 1895. Goodwin, Charles Restarts, pauses, and the achievement of mutual gaze at turn-beginning. 1980 Sociological Inquiry 50 (3–4): 272–302. Special Double Issue on Language and Social Interaction, Don Zimmerman and Candace West (eds.). 1981 Conversational Organisation – Interaction between Speakers and Hearers. Orlando: Academic Press 1987 Unilateral departure. In Talk and Social Organization, G.J. Button and R.E. Lee (eds.), 206–218. Clevedon: Multilingual Matters. 2000 Gesture, aphasia, and interaction. In Language and Gesture, D. McNeill (ed.), 84–98. Cambridge: Cambridge University Press. Heath, Christian 1986 Body movement and speech in medical interaction. Studies in Emotion and Social Interaction. Cambridge University Press. Hepburn, Alexa 2005 Crying: Notes on description, transcription and interaction. Research on Language and Social Interaction 37 (3): 251–290. Hepburn, Alexa, and Potter, Jonathan Crying receipts: Time, empathy and institutional practice. Research on (in press) Language and Social Interaction 39. Heritage, John Garfinkel and Ethnomethodology. Cambridge: Polity Press. 1984 Husserl, Edmund 1973 Logical Investigations. London: Routledge. Hutchby, Ian, and Robin Wooffitt 1998 Conversation Analysis. Cambridge: Polity Press. Jefferson, Gail A technique for inviting laughter and its subsequent acceptance/declina1979 tion. In Everyday Language: Studies in Ethnomethodology, G. Psathas (ed.), 79–96. New York: Irvington Publishers. Jefferson, Gail 2004 A note on laughter in ‘male-female’ interaction. In Discourse Studies 6 (1): 117–133. Johnson, Mark, and George Lakoff Philosophy in the Flesh: The Embodied Mind and Its Challenge to 1999 Western Thought. New York: Basic Books.

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   75 Kendon, Adam Gesticulation and speech: Two aspects of the process of utterance. In 1980 The Relationship of Verbal and Nonverbal Communication, M.R. Key (ed.), 207–227. The Hague: Mouton. 1983 Gesture and speech: How they interact. In Nonverbal Interaction, J.M. Wieman and R.P. Harrison (eds.), 13–45. Beverly Hills: Sage Publications. 1995 Human gesture. In Tools, Language and Cognition in Human Evolution, T. Ingold and K.R. Gibson (eds.), 43–62. Cambridge: Cambridge University Press 2000 Language and gesture: Unity or duality? In Language and Gesture, D. McNeill (ed.), 84–98. Cambridge: Cambridge University Press. 2002 Some uses of head shake. Gesture 2: 147–183. 2004 Gesture – Visible Action as Utterance. Cambridge: Cambridge University Press. Laursen, Lone 2002 Kodeskift, gestik og sproglig identitet på internationale flersprogede arbejdspladser. Unpublished Ph D Dissertation, University of Southern Denmark. LeBaron, C., and J. Streeck 2000 Gestures, knowledge and the world. In Language and Gesture, D. McNeill (ed.), 118–138. Cambridge: Cambridge University Press. Maynard, Douglas 2003 Bad News, Good News: Conversational Order in Everyday Talk and Clinical Settings. Chicago: University of Chicago Press. McNeil, David So you think gestures are non-verbal? Psychological Review 92: 350– 1985 371. McNeill, David and Susan Duncan Growth points in thinking-for-speaking. In Language and Gesture, D. 2000 McNeill (ed.), 141–161. Cambridge: Cambridge University Press. Mead, Margaret Vicissitudes of the study of the total communication process. In Ap1964 proaches to Semiotics: Cultural Anthropology, Education, Linguistics, Psychiatry, Psychology (Transactions of the Indiana University Conference on Paralinguistics and Kinesics). Thomas A. Sebeok, Alfred S. Hayes, and Mary Catherine Bateson (eds.), 277–287. The Hague: Mouton. Merleau-Ponty, Maurice 1945 Phénoménologie de la Perception. Paris: Gallimard. Montague, Richard Formal philosophy. In Selected papers of Richard Montague, Richmond 1974 H. Thomason (ed.), 193–207. New Haven: Yale University Press. Moon, C., R.P. Cooper, and W.P. Fifer Two-days-olds prefer their native language. Infant Behavior and Devel1993 opment 16 (4): 495–500.

76   Anders R. Hougaard and Gitte R. Hougaard Müller, Cornelia Redebegleitende Gesten: Kulturgeschichte, Theorie, Sprachvergleich. 1998 Berlin: Berlin Verlag. Neisser, Ulric Cognitive Psychology. New York: Appleton-Century-Crofts. 1967 Pasqual, Esther 2008 Fictive interaction blends in everyday life and courtroom settings. In Mental Spaces Approaches to Discourse and Interaction, A. Hougaard and T. Oakley (eds.) 79–107. Amsterdam: John Benjamins. Sinha, C., and C. Rodriguez Language and the signifying object: From convention to imagination. In 2008 The Shared Mind: Perspectives on Intersubjectivity, J. Zlatev, T. Racine, C. Sinha, and E. Itkonen (eds.), 357–378. Amsterdam: John Benjamins. Sacks, Harvey 1992 Lectures on Conversation, G. Jefferson (ed.), 386–387. Oxford: Blackwell. Sacks, H., E.A. Schegloff, and G. Jefferson 1974 A simplest systematics for the organisation of turn-taking for conversation. Language 50: 696–735. Sapir, Edward 1958 Reprint. The status of linguistics as a science. In Culture, Language and Personality, D.G. Mandelbaum (ed.). Berkeley, CA: University of California Press. Original edition, 1929. Schegloff, Emanual 1984 On some gestures’ relation to talk. In Structures of Social Actions: Studies in Conversation Analysis, J.M. Atkinson and J. Heritage (eds.), 266–296. Cambridge: Cambridge University Press. Schutz, Alfred The Problem of Social Reality: Collected Papers I. Den Haag: Martinus 1962 Nijhoff. 1964 Studies in Social Theory: Collected Papers I. Den Haag: Martinus Nij­ hoff. Searle, John 1962 Meaning and speech acts. In Philosophical Review 71 (4): 423–432. Sinha, Chris Grounding, mapping and acts of meaning. In Cognitive Linguistics: 1999 Foundations, Scope and Methodology, T. Janssen and G. Redeker (eds.), 223–255. Berlin, Mouton de Gruyter. 2005 Blending out of the background: Play, props and staging in the material world. Journal of Pragmatics 37: 1537–1554. Streeck, Jürgen Gesture as communication I: Its coordination with gaze and speech. 1993 Communication Monographs 60: 275–299. 2002 A body and its gestures. Gesture 2 (1): 19–44.

Sense-making as a phenomenon of interacting, knowledgeable, social bodies   77 2003

The body taken for granted: Lingering dualism in research on social interaction. In Studies in Language and Social Interaction: In Honor of Robert Hopper, Phillip Glenn, LeBaron, and Jenny Mandelbaum (eds.), 427–440. Mahwah, N.J. Lawrence Erlbaum. Te Molder, Hedwig, and Jonathan Potter (eds.) Conversation and Cognition. Cambridge, Cambridge University Press. 2005 Tomasello, Michael The Cultural Origins of Human Cognition. Cambridge, Mass.: Harvard 1999 University Press. Whorf, B. 1956 Language, Thought and Reality. Cambridge, Mass.: The MIT Press. Williams, Meredith 1999 Wittgenstein, Mind and Meaning: Towards a Social Conception of Mind. New York: Routledge. Williams, Robert 2005 Material anchors and conceptual blends in time-telling. Unpublished Ph D Dissertation, University of California, San Diego. Williams, Robert 2008 Guided conceptualization: Mental spaces in instructional discourse. In Mental Spaces in Discourse and Interaction, T. Oakley and A. Hougaard (eds.), 209–234. Amsterdam, Benjamins. Wittgenstein, Ludwig Reprint. Philosophical Investigations. Oxford: Blackwell Publishing. 2001 Original edition, 1953.

78   Anders R. Hougaard and Gitte R. Hougaard Appendix Transcript notations: [first name] Relevant contextual information Unrecoverable speech [X] Transcriptionist doubt ( ) Numbers in brackets indicate the number of lines in the transcription. (1) [ Left-hand brackets link an ongoing with an overlapping Utterance at the point where overlap begins. = The equal sign links two adjacent utterences when there is no perceptible interval between them O R the equal sign links different parts of a single speakers continous flow of speech that has been carried over to another line to accommodate an intervening interruption. (0.1) Inter- and mid-turn silences represented in tenths of a second A dot in parentheses indicates a pause less than 2/10 of a second (.) A comma indicates a rising intonation, weaker than? , Rising intonation ? Falling intonation contour, less than final ¿ . Falling, or final, intonation contour Laughter he ‘smiley voice’ (gut) Audible inhalations *h Audible exhalations h* Colons represent lengthened vowels and consonants. The number of co: lons shows the relative stretch of sound. she Stress or emphasis: the more underlining, the greater the emphasis Loud talk: the louder, the more letters in upper case She The talk following the degree sign was markedly quiet or soft ° Sharp rise, a shift or resetting of the pitch register  Sharp fall, a shift or resetting of the pitch register 

Chapter 4 Supracultural models, universalism and relativism: The language of personhood in Chinese and American cultures Jason D. Patent

1. Introduction This paper reports findings on differences in the language and conceptualization surrounding personhood in two different cultural communities:1 native speakers of Chinese raised in the People’s Republic of China, and native speakers of English raised in the United States. In delving into these issues, this research is partial heir to many traditions of linguistic, anthropological and psychological research, but direct heir to nothing that could legitimately be called a “tradition.” In other words, this research aims to fill a methodological and theoretical hole lying just outside of work that has been done in other fields: linguistics (especially cognitive linguistics, as well as discourse analysis, to the extent that this can be considered a branch of linguistics), anthropology (cultural, cognitive and psychological)2, and psychology (cultural and cross-cultural). 1 I deliberately eschew here the term speech community, as recent sociolinguistic research has provided convincing arguments against its utility. See Bucholtz (1999) for a thorough critique. Bucholtz herself prefers the term community of practice, but for present purposes such a term is too general. Cultural community, my chosen term, is meant to convey that there are groups whose internal homogeneity (hence the term community) and external heterogeneity may be thought of largely in cultural terms. The precise degree of internal homogeneity in the communities under discussion is not the focus of this study. Still, because conflict between and among cultural models is the focus, internal homo- and heterogeneity will at times be discussed. 2 Anthropological linguistics and linguistic anthropology would at first blush seem to be important sites for this research as well. However, both of these fields focus more on conversational interaction as a cultural phenomenon than they do on the cultural models being conversed about. While such questions are fascinating, and while my

80   Jason D. Patent In this paper I adopt the view from cognitive linguistics that language is reflective of conceptual structure. I adopt the view from recent work in anthropology (see below) that culture is shared conceptual structure. Taken together, these views implicate language in any discussion of culture. All the data I present below are simultaneously cultural, conceptual and linguistic data. Using the cultural model (Holland and Quinn (1987), D’Andrade and Strauss (1992), Shore (1996) as the primary unit of inquiry, I present two contrasting views of personhood, which, for reasons of convenience, I will refer to as “Chinese” and as “American”.3 A person – expectations of what a human life will look like, including how an individual relates to family and society – is conceived of differently in the two systems composed of differing cultural models. Conceiving of the relevant cultural systems in terms of models enables us to look at culture at varying levels of granularity: from very specific human interactions to broader society-structuring principles.4 At the same time, I will argue, the wording I just used – “differing cultural models” – is a bit deceptive. I will show how the relevant differences, rather than resulting from two mutually exclusive, or even meaningfully different, sets of cultural models, instead result from differing choices (mostly unconscious, of course) among many shared and available cultural models. The findings reported here, while interesting in their own right, are also pressed into service for a theoretical end: the proposal of the notion of “supracultural models” as an updated, nuanced approach to questions of cultural universalism and relativism. Below I present first a brief theoretical discussion, followed by data, and ending with more on the theoretical side. 2. Theoretical context The present study is intended as a preliminary re-investigation of issues of universalism and relativism, as examined through some selected cultural models data would surely be amenable to such an analysis, my primary concern is the latter rather than the former. For a helpful discussion about the differences between anthropological linguistics and linguistic anthropology, see Duranti (2001: 2–5). 3 The cultural communities that could fairly, if problematically, be referred to as “Chinese” and “American” are of course vast and heterogeneous. Yet despite regional and political differences among the interviewees in both of these communities, a striking degree of homogeneity is still apparent. See below. 4 Why the cultural model is a particularly useful unit of inquiry is addressed by Quinn and Holland (1987); Shore (1996: 42–71); DiMaggio (1997: 266–274), who, following psychologists, uses the term schema; and Kiesling (2005: 98–99). For a general history of cultural models/schemas in anthropology, see D’Andrade (1995: 122–181).

Supracultural Models   81

in two different cultural communities. Given the centrality and long history of universalism and relativism in questions of language and culture, even a book-length treatment of the issues would hardly suffice. The findings here are intended to be suggestive rather than conclusive. Here I sketch in brief what I see as this paper’s main contributions to this area. As we will see in the data analysis section below, I sketch a view of culture that is dynamic rather than static, heterogeneous rather than monolithic. A culture is not an object with a fixed set of properties. It is, rather, a set of tendencies, or of inclinations, toward resolving conflicts in one way or another. Thus from the start we must be wary of asking questions such as, “Is X a part of culture Y?” or “Is Z a part of every culture in the world?” The answer is, more likely than not, both yes and no. In this emerging view, culture is anything but fixed. Rather, a culture provides a set of resources, or tools, to be drawn on by members of a cultural community when engaging in conversing, reasoning or in any other culturally mediated activity. Such reasoning is local and highly context-dependent.5 Culture is thereby revealed as analogous to language as understood by sociolinguists and linguistic anthropologists: as a tool to be employed in specific situations for the accomplishment of local purposes. It is dynamic and flexible. This view of culture is not novel,6 but I know of no previous cross-cultural study that has employed this view of culture along with cultural models as the key units of inquiry. What emerges from the present study is a small (for now) list of “supracultural models”: cultural models which are shared across different cultural communities, and which are differentially available to members of these communities. “Differentially” is meant here on two levels: any given supracultural model will be available to a greater or lesser degree to one cultural community than to another, and will be available to a greater or lesser degree to any given person within a cultural community than to another. Of all the viewpoints on culture that have been articulated by previous scholars, and with which I am familiar, the two that resonate most with the approach taken here are those of Shweder and Sullivan (1993) and Shore (1996). My view can be seen as a blend of these two. First is Shweder and Sullivan’s “universalism without the uniformity.” They write (1993: 517): 5 See Patent (2003: 264–280) for a discussion of how metaphor may be conceived of in similar terms. 6 See the concluding section of this essay for a brief review of previous research supporting a similar view of culture.

82   Jason D. Patent There are undoubtedly many ways to reconcile human variety with our common humanity. One way is to argue that what everyone has in common, what unifies and in a sense universalizes us is itself a heterogeneous complex of inherited psychological processes and forms. These processes and forms are activated, institutionalized, and rationalized by various cultures selectively and differentially, but considered as a complex whole and examined theoretically as an etic grid, make the study of cultural psychology possible.

Shore, while agreeing in broad strokes with Shweder and Sullivan’s views, takes Shweder, specifically, to task (1996: 34–39) for falling into the old dichotomy of predictability versus arbitrariness. He advocates instead the idea of constraints over strict determination. (This is the notion that cognitive linguists, for instance, have termed motivation.) Shore’s criticism seems to be aimed at Shweder’s earlier (1984) work, not taking into account his more recent work, cited above: it appears that by 1993 Shweder had adopted a view closer to Shore’s. Regardless, Shweder and Sullivan’s idea of “a heterogeneous complex of inherited psychological processes and forms” manifesting differently in different cultures fits well with my findings. Despite Shweder and Sullivan’s good intentions, however, little work has been done since their 1993 paper on forging a methodological path leading to a delineation of just what this “heterogeneous complex” might look like, and how various cultures might adopt various aspects of this complex. Little work, that is, except for that of linguist Anna Wierzbicka (1992, 1993, 1996, 1997), who deserves special mention here, since she is the only scholar of whom I am aware who is asking similar questions and has attempted to develop methods of investigation that might get us some answers. And while I suspect that she and I would have some disagreements about the specifics of our methods, I consider the present work to be complementary in spirit to Wierzbicka’s. 3. Previous research Precisely how the present study breaks new ground will become apparent throughout the paper; in the meantime, it is worth noting from which intellectual traditions this study has sprung. In spirit, this research was born of training in cognitive linguistics, yet most of the core theoretical notions of cognitive linguistics – metaphor, metonymy, frame semantics, image schemas, mental spaces – are not explicitly invoked. Methodologically this study is closest to discourse analysis, but (a) this specific subject matter has not been examined before within discourse analysis; (b) this study’s emphasis is on generalizations across several bodies of discourse and across cultural communities; and (c) again, the focus on cultural models sets this study

Supracultural Models   83

apart.7 This study shares with anthropology a central concern with culture, yet the different conception of “the field” (see the discussion on methods below), the centrality of cultural models, and the emphasis on linguistic data differentiate this study from more prototypical anthropological studies. The emphasis on culture is also shared by cultural psychology, yet again this study’s focus on language, and on the cultural model as the relevant unit of culture, sets it apart. By focusing on personhood, this study also becomes heir to certain other conversations. There is a vast literature on the “self” and the “person.” Not all scholars are convinced of the usefulness of distinguishing between the two terms. Rosaldo (1984, especially 145–150) is most frequently cited as the most eager skeptic of the distinction, while Mühlhäusler and Harré (1990) – following Mauss (1985/1938) – and Cohen (1994) are defenders of the distinction. The ins and outs of the debate need not concern us here.8 To the extent that there ever has been an intuitively satisfying distinction to be made, self may be thought of as “from the inside” (i.e., the subjective sense of separateness from others) and person as “from the outside” (i.e., the atomic units which together make up social groupings and institutions). Since this study involves mostly human beings as functioning within social and political groupings, person is the term I have chosen to use, though literature on the self will also be cited. In focusing on personhood, the present study explores similarities and differences in how two cultural communities wrestle with: expectations of how human beings relate to one another; to what extent decisions about one’s actions are and/or should be based on one’s relationships with other human beings; how people may expect to be treated by others; and, in general, what sorts of concerns people have and should take into consideration when making decisions about how to act. In philosophy, questions of self- and personhood have been frequent topics of investigation. I will not address these here,9 since my concern is with situating my study within the social sciences, where a great deal of attention has been received from psychologists and anthropologists. Studies have ranged 7 Two previous studies – Gee (1999) and Kiesling (2005) – have examined the relationship between discourse analysis and cultural models. However, the specific content addressed has been quite different, and neither study has been cross-cultural in nature. 8 See Foley (1997: 262–264) for a concise summary of the debate. 9 For an excellent summary of “Western” notions of self from a philosophical perspective, see Taylor (1989); for views from Chinese philosophical traditions see Hall and Ames (1998: 3–99), Tu (1985), and Elvin (1985); for a comparison between “Eastern” and “Western” views, see Northrop (1946) and Hsu (1981).

84   Jason D. Patent from general theorizing about self- and personhood (Kohut 1977; McCall 1987; Gergen 1990, 1991; Wierzbicka 1993); to anthologies of general (mostly nonethnographic) studies of self- and personhood (Gergen and Davis 1985; Bond 1988); to summaries and expansions upon large bodies of specific ethnographic research (Markus and Kitayama 1991; Markus, Kitayama, and Heiman 1996); to essay-length ethnographies (Read 1955; the essays in Carrithers, Collins and Lukes 1985; and in White and Kirkpatrick 1985); to monograph-length ethnographies (Rosaldo 1980; Doi 1985; Myers 1986; Kondo 1990). Virtually all of the research that has been done on self- and personhood is from outside my field of training, linguistics. This research varies widely in its degree of attention to language. As a linguist I am primarily concerned with personhood as reflected or revealed in language. I will not deal in any one place with studies specifically addressing personhood in Chinese. There is yet another body of literature reporting findings that are suggestively convergent with mine, mostly from psychology, but also from philosophy, anthropology and sociolinguistics. (For a small but representative sampling, see Hsu 1985; Bond and Hwang 1986; Markus and Kitayama 1991; Gardner, Gabriel and Lee 1999; Han, Leichtman, and Wang 1998; Ji, Schwartz and Nisbett 2000; Ji, Nisbett and Su 2001; Ji, Peng and Nisbett 2000; Nisbett 2003; Young 1994: 1–65). 4. The study The data for this study consist of conversations recorded during autumn of 2001 on the campus of the University of California, Berkeley. Each conversation was between either two native speakers of Chinese10 raised in the People’s Republic of China or two native of speakers of English raised in the United States. During recruitment, potential participants were asked to bring with them a friend or well-known family member to serve as conversational partner. The intention was to create as “natural” a conversational setting as possible. 10 It is generally irresponsible for a linguist to refer to a “native speaker of Chinese,” as there are so many varieties of Chinese. Material constraints – time and money – prevented me from limiting my search to native speakers of standard Northern Mandarin. For college-educated people from China, Mandarin serves as a lingua franca: one can always expect, at a Chinese university, that students can converse comfortably and fluently in Mandarin on any topic. Since all the Chinese participants in my study were highly educated, they were also comfortable speaking Mandarin with each other.

Supracultural Models   85

In all, six Chinese and thirteen English conversations were recorded. Englishspeaking participants were all University of California Berkeley undergraduates; Chinese-speaking participants were graduate students.11 Participants sat across a small table from each other, while I sat at a desk a few feet away, with my back to them. Twelve printed questions were in a stack on the table. Participants were instructed to take a question, read it aloud, discuss it for a few minutes, and then take the next question, until they were finished. After they were finished, I moved over to the table to ask follow-up questions based on what they had said.12 Interviews, including follow-up questions, generally took 30 to 60 minutes.13 The questions had to do, broadly speaking, with personhood. As we have seen, this has been a topic of enormous interest within anthropology and psychology. I take as a working understanding of personhood Kenneth Gergen’s (1990) notion of “inscription”: a person may be understood only via her social relationships with other people, social groups and institutions. As Gergen writes (1990: 584–585): The individual’s well-being cannot be extricated from the web of relationships in which he/she is engaged. The character of the relationship depends, in turn, on the process of adjusting and readjusting actions. In effect, forms of relationship depend on the mutual coordination of actions. As this coordination is progressively realized, it may be said that a relational nucleus is formed. A relational nucleus may be viewed as a self-sustaining system of coordinated actions in which two or more persons are engaged.…As this coordination takes place – taking turns in conversation, adjusting the tone of voice so that the other may hear without discomfort, speaking in mutually acceptable patterns of words, walking together at similar speeds, and so on – we come to form a relational nucleus.

I believe that Gergen’s overall position is somewhat overstated, as he is willing to accord nothing to individual mind states, seeing all meaning as social fact. 11 I had minimal control over the demographics of the participants. Ideally ages and educational levels would have matched better across the two populations. 12 I am grateful to Alan Cienki (personal communication) for describing and suggesting this method. I am of course responsible for any shortcomings. 13 For more on methods of investigating culture through language, see the essays in Quinn (2005), especially those by Quinn, by D’Andrade and by Strauss. These essays discuss why interview methods, when thoughtfully formulated, can yield excellent cultural data. Put simply, the way I see it is that the researcher wants to create as natural and comfortable an interview setting as possible, so that participants will be more likely to adopt forms of speech (“social languages” – Gee 1996: ch. 4) closer to those they would tend to use with family and friends, and less like they would tend to use in formal interviews. Ideally, of course, we would be able to gather language, to use Hutchins’ (1995) terminology, “in the wild,” but short of that the best we can hope for is a “feral” look at language, cognition and culture.

86   Jason D. Patent Strauss and Quinn’s (1997) balancing of the intrapersonal and the extrapersonal is, to give one example, a fairer account of how meaning is created and interpreted. Yet Gergen’s account has to date been the most articulate and convincing as locating self and person at the nexus of various social interactions, through the notion of the “relational nucleus.” It is for this reason that I cite his conceptualization here, for as will become clear later in this essay, people understand themselves and others only through such relational nuclei, mediated by institutions such as the family and the state. In the present study, cultural models of these relational nuclei were probed for by proposing scenarios, and asking the participants to discuss various aspects of the scenarios.14 For instance, the question below (heretofore referred to as the “Rock Band” question) posited a scenario having to do with nuclear family relationships, and with expectations for how such relationships shape an individual’s aspirations. 15 Below are both versions of the question: 16 Wáng’èr kuàiyào gãozhõng bìyè le. Suïrán tã de fùmû xïwàng tã shàng dàxué, dànshi tã bù xiâng shàng dàxué, xiâng zû yï ge yáogûn yuètuán. Nî juéde Wáng’èr huì zênmeyàng hé tã de fùmû shuõ? Tãde fùmû yòu huì zênyàng fânyìng? Zuìhòu tãmen huì juédìng zênme zuò? Nî huì zhïchí nâ yï biãn? ‘Tom is about to graduate from high school. He decides he doesn’t want to go to college, despite his parents’ wishes. Instead, he wants to join a rock band. What

14 Peng, Nisbett and Wong (1997) present experimental evidence that scenario-based investigations of cross-cultural differences in values are more valid than the traditional ranking and rating methods used in psychology. Holland and Skinner (1987) demonstrate that interviewees provide more robust cultural data when describing specific scenarios than when listing abstract category attributes. Together these findings provide empirical support for the intuition, expressed in Patent (2003: 33), that “…specific scenarios allow us to project ourselves imaginatively into specific frameslots and reason through the temporal unfolding of a frame by drawing on our expectations about the behavior of the frame participants. If instead we are asked a very general question meant to cover a wide range of frames, we will be unable to ground ourselves in any particular frame structure, and will likely fall back into (possibly made-up) generalizations that will be of little or no use to the researcher.” 15 It is quite impossible, try as one might, to use language that is independent of cultural models. In using terms such as “individual’s aspirations” I am invoking a set of cultural models – to be described later in this essay – that belongs specifically to American culture, if not more generally to the cultures of “Western” “industrialized” societies. 16 Interlinear glosses in this forum would simply take up too much space, and given the coarse granularity of this study with respect to linguistic structure, they are not necessary.

Supracultural Models   87 will the family members all say to one another? What will happen in the end? Who is right?’

The hope – largely realized – was that this sort of question would yield a discussion that offered up cultural models of how Tom and his parents are inscribed mutually within their family, and more broadly within society at large. 5. Findings In mining the data, what I was primarily keeping an eye out for was linguistic evidence for cultural models. After identifying the relevant cultural models, I examined how they fit together into two different systems. As we will see, two types of “conflict” among differing cultural models are evident, belonging to what Strauss and Quinn (1997) refer to as the realms of the intrapersonal and the extrapersonal. Intrapersonally, a given speaker (along with his conversational partner) will often be conflicted about which cultural model to prioritize in discussing a given scenario. We could say that each speaker is internally conflicted about which models to prioritize.17 Extrapersonally, certain models are invoked and privileged more frequently in each participant group than are certain other models. Thus there are both individual preferences and culturalcommunity “preferences,” or tendencies, in the cultural models that appear in the data.18 The upshot of this distinction, which will become more apparent below, is to point the way toward making more accurate statements about the characteristics of cultural communities.19 Here is how this works: If a certain model “wins out” over another to a significant degree extrapersonally – that is, if a certain model is privileged more often than another among a significant number of speakers belonging to a given cultural community – we may be tempted to draw generalizations about that cultural community and to make essentializing statements about a given culture, such as the commonly heard statement that “there 17 Over the course of several essays, Strauss (1990, 1997, 2005) has developed a model incorporating three types of intrapersonal model conflict: compartmentalization, ambivalence and integration. The examples examined here are all cases of ambivalence. 18 This is actually not quite what Strauss and Quinn mean by extrapersonal. The term more properly refers to the counterpart to the intrapersonal – the publiclyavailable cultural forms that help construct, and are reconstructed by, the intrapersonal. Sharedness is thus but one a aspect of the extrapersonal. 19 For an analysis of how the interpersonal v. extrapersonal distinction can help linguists make more accurate statements about category structure, see Patent (2001).

88   Jason D. Patent is no concept of human rights in China.” As we will see, though, a broadly similar range of models is available to both cultural communities investigated here. The differences between the two cultures can be traced not so much to a “lack” of cultural and linguistic resources in a given community, but rather to different (mostly unconscious) choices among available cultural models.20 Of crucial importance in the present study is what this means for our conceptualization of culture. It is in the moments of (both intrapersonal and extrapersonal) conflict that we see the full range of available models, and that in many cases this range of models is available to both sets of participants. What differentiates the two groups is their differing prioritizations of the competing models. Furthermore, the absence of model conflict in response to certain questions may point us toward cultural models that hold particular salience for conceptualizing certain situations for a given cultural community. Overall, then, by keeping one’s eye trained on intra- and extrapersonal model conflict, the researcher is given particularly useful insights into culture. The bulk of the remainder of this essay will consist of supporting examples. Some examples will involve mostly intrapersonal conflict, some mostly extrapersonal conflict, some a bit of both. The overall point, though, remains the same: conflict between and among cultural models – or its lack – provides us with important evidence about the cultural and linguistic resources of a given cultural community. 5.1. “Success” versus “Follow Your Dreams” Let us begin with an answer to the Rock Band question, presented above. Speakers from both groups offer up models I have dubbed Success and Follow Your Dreams. Consider the following English example:21

20 With China (a prototypical representative of what is “Eastern”) and the U.S. (a prototypical representative of what is “Western”) as foci of this study, it would be irresponsible not at least to mention that these issues are inevitably bound up with colonialism and imperialism. For good summaries of these issues as they relate specifically to language and culture, see Liu (1995, 1999). 21 Participants are identified via a combination of a letter (“E” for “English” or “C” for “Chinese”) and a number which indicates the order in which they participated in the study. The larger of the two Chinese numbers in each pair is odd because the initial pilot interview I carried out had only one participant. Bold type indicates strong stress; ellipses indicate pauses. The transcriptions are fairly broad. This is because, for the type of data most relevant to this study, it is sufficient. A narrow

Supracultural Models   89

(1)

E-14 E-13 E-14 E-13 E-14 E-13 E-14 E-13

E-14 E-13 E-14 E-13

Well if he really wants to join a rock band he’s gonna do it. I think Tom will end up working at McDonald’s. And he won’t be able to…I mean… You don’t have any faith in Tom. No it’s like…there’s so many bands out there, and a lot of them…yeah, a lot of them hit it big, but then they’re usually one-hit wonders… Yeah. …and then they’re just like, plop, you know? Chances are he’s not gonna make it. He’s just gonna end up… It’s just the statistics, you know, I don’t wanna have like a negative outlook, but I think in the end he’ll just run back to his parents and go, okay I’ll go back to college, can you guys pay for it, you know? Yeah. Something like that. Yeah. Chances are that would happen. But I mean if that’s really what he wants to do, shouldn’t he do it? Yeah, I think you should totally go after your dreams and do what you want, but you also have to keep in mind that there are consequences for what you choose, and you have to weigh those.

A brief sketch of the relevant models: Success: It is important to gain employment that is economically viable.22 Follow Your Dreams: A person should do with her life what she most loves doing.

transcription would yield visual clutter with little added utility. See D’Andrade (2005: 91) for a brief discussion of issues related to granularity of transcription. 22 The Success model discovered here is importantly different from that discussed by D’Andrade (1984) and by Strauss (1992). That model could perhaps also be called “Up by the Bootstraps”: the old Horatio Alger myth of overcoming long odds to reach the pinnacle of one’s profession. The model here is somewhat more pedestrian, having to do simply with making sure one has a stable economic foundation. This model is, in turn, somewhat different from its nearest Chinese counterpart; see below.

90   Jason D. Patent It is clear from this example that E-13 and E-14 are conflicted about which of the models to prioritize. There are a fortunate few in this world for whom these models are not in conflict, but for “the rest of us,” these two models often conflict. Both models are available to these speakers, and they are not sure which to privilege. In the following example, however, the model clash receives something approaching resolution: (2)

E-9

E-10

E-9 E-10

I have a friend, his name is J___, he did exactly this, and like his parents wanted him to go to college, and the family members at first were like, don’t do it, don’t do it, and they were all angry about it, and then they sort of uh, and they definitely to one another speak badly of the idea, they think it’s bad, but I mean eventually, eventually in the end, if, if, what do I mean in the end, but at the moment, they’ve sort of resolved it with saying, well this is what he wants to do, and like you know he’ll learn or he won’t, or whatever, but this is his path, and I… I think that’s right too, like conventionally, you know what I’m saying it’s not a safe thing to do if you want to be financially secure all your life or something, but if that’s not your big important thing… …which it shouldn’t be… Yeah, I definitely agree with you on that, but then I think you should just, I don’t know, I think you should pursue, what you’re interested in.

Here, E-9 and E-10, after some consideration, settle on Follow Your Dreams. E-19 and E-20 are also conflicted. First there is a fairly strong statement in support of Follow Your Dreams: (3)

E-20 E-19 E-20

I guess if the guy’s got a lot of talent, and he really really wants to do it… I think he should do it. I think he should do it. ‘Cause I mean, you’re only one life, right? Do what you want.

But then, after a bit more discussion, E-20, adopting the parents’ viewpoint begins:

Supracultural Models   91

(4)

E-20 E-19 E-20

Let’s see…let’s see, if it was my kid…seeing as how I am at a university, working my butt off, trying to get my degree, I’d probably be kind of upset at him. Probably. No, I know I would be very upset with him. [laughter] I would want him to go to college. ‘Cause chances are… it’s not very good. But, I mean…and yet I just said that he should go and pursue his dreams, okay anyways I’m full of it. I don’t know. Well…I’d probably let him do it though. In spite of my wishes. ‘Cause I don’t want an unhappy kid. And I don’t want him to hate me for the rest of my life. ‘Cause I didn’t just spend like you know eighteen years raising this kid, to have him like hate me. [laughter]

Likewise, E-15 becomes convinced that Tom should join the rock band: (5)

E-15

I don’t know. I think that…you have to do what you want to do.

At this point in the conversation, to E-15 the importance of following one’s dreams is such that it is not even a choice any longer, but rather an obligation. However, after some convincing by E-16, who adopts a more Success-oriented approach, she relents a bit: (6)

E-15

I still think he should pursue his own interests, but he has to be…um, he has to have common sense.

The overall tendency in the English data is support for Follow Your Dreams, tempered by a cautionary note informed by Success. Turning now to the Chinese data, we will see again that both models are available. The following example lays out the basics of Success: 23,24 23 Ideas of success in China and the U.S. are obviously not identical. For instance, in China an important aspect of one’s success is the honor it brings to family members. While this is true to some extent in the U.S., it is less true than in China. “Reputable” occupations also differ: teachers, for example, are accorded higher status in China than in the U.S. So it is risky to claim there is one Success model shared by the U.S. and China. However, on a schematic level, there is enough similarity to warrant positing a single, shared model: there is a highly general model in both cultures that privileges money, reputability and stability. See below for further discussion of “equivalence.” 24 In the English translations, I have tended more toward fidelity than elegance, in order to bring readers not familiar with Chinese closer to the Chinese discourse.

92   Jason D. Patent (7)

C-18

Wô juéde zài xiànjïn Zhõngguó shèhui, bâifënzhï jiûshíwû de fùmû shì bù huì tóngyì tã zhèi yàng zuò ne. Yïnwèi dà… jiùshi…zài Zhõngguó de fùmû de pûbiàn xïnlî jiùshi xïwàng zìjî de háizi kâoshàng yï ge míngpái dàxué, ránhòu jiãnglái yôu yï ge fëicháng tîmiàn de gõngzuò. ‘I think that these days in Chinese society, ninety-five percent of parents would not want Wang Er to do this. Because…that is…it is the common psychology of Chinese parents to hope for their children to get into a famous university, and then in the future get a creditable job.’

The emphasis here is on the aspects of Success that have more to do with public perceptions than with economics. Economic considerations take center stage in the following example: (8)

C-11

Wô juéde dú dàxué de hua, qîmâ tã jiù hái kêyî zhâo fènr xiãngduì lái shuõ hâo diânr de gõngzuò yânghuo zìjî ba. Nî yàoshi zûzhi yáo…yáogûn yuètuán, nî xiànzài yê bù dúshø le, nî wèilái huì…dàodî huì zênmeyang? ‘I feel that if he goes to college at least he’ll be able to find a relatively good job to take care of himself. If you want to start a rock band and you don’t study now, what will your future ultimately be like?’

Success and Follow Your Dreams25 also come into conflict. As C-18 continues his ruminations, though, he privileges Success: (9)

C-18

Wô juéde wô hái huì zhïchí tã de fùmû. Yïnwèi…qíshí tã dú le dàxué yîhòu yê tóngyàng huì yôu gè yàng de xìngqu hé àihao kêyî dédào gè zhông gè yàng de fãzhân. Dàn wô juéde zhèi ge zhïshi shuîpíng shì juédìng tã zhèi ge rén de

25 As with Success, the Follow Your Dreams models in the U.S. and China are not identical. Where in the U.S. the focus is on passion, and people’s very identities are bound up with their “dreams,” in China it has more to do with “hobbies” and “interests,” as can be seen in this example. “Hobbies” and “interests” don’t define a person, and a life, in the same way “dreams” do. See below for further discussion of “equivalence.”

Supracultural Models   93

yï ge sùzhi de yï ge zhòngyào de yï ge fãngmiàn. Suôyî tã wánquán…tã zû zhèi ge yáogûn yuètuán wánquán kêyî zài tã dú le dàxué yîhòu, tã yôu zìjî yídìng de fënxï wèntí, jiêjué wèntí yídìng nénglì yîhòu, ránhòu zài…zài chóngxïn kâolý tã zhèi zhông xiângfâ. ‘I think I would support his parents. Because…in fact after he has finished college he can still have all sorts of interests and hobbies to be developed in all sorts of ways. But I think knowledge level is an important aspect in determining this person’s quality. So it is no problem for him…it is no problem for him to form this rock band after finishing college, after he has the definite ability to analyze problems and solve problems. Then he can…once again consider this plan of his.’ Follow Your Dreams is never explicitly advocated by any of the Chinese speakers. However, it is available as a model,26 as we can see in the following exchange: (10) C-14

C-15 C-14 C-15 C-14

Zuìhòu wô xiâng tã de fùmû kêndìng huì qiângpò tã qù shàng dàxué. Dànshi wô yàoshi zuòwéi Wáng Èr de fùmû wô kêndìng huì shuõ de bîjiào cõngming yï diân, jiùshi shuõ, nî xøyào qù shàng dàxué, dànshi nî de yáogûn yuètuán zhèige mèngxiâng yê bù yïdìng jiù huì pòmiè ne. […] … wô shì zhïchí tã yìbiãn shàng dàxué ránhòu yìbiãn jiàn yáogûn yuètuán. […] Tã jiù wán le, tã zhèi ge tiãncái jiù huî le. […] Mêiguó hâo duõ dàxuéshëng, tãmen yê shì yìbiãn shàngxué yìbiãn yê shíxiàn le zìjî…jiùshi xuéyè zhï wài de hên duõ mèngxiâng, wô juéde zhèi liâng ge búshi tèbié chõngtø de. Duì, kêyî. Wô yê tóngyì nî de guãndiân. ‘In the end I think his parents will definitely force him to go to college. But if I were Wang Er’s parents I’d be smarter

26 See below for a discussion of how “availability” may be assessed.

94   Jason D. Patent

C-15 C-14

C-15

about it. I’d say, you need to go to college, but your dream of starting a rock band needn’t be shattered.’ […] ‘…I support that he goes to college and forms a rock band at the same time.’ […] ‘Then he’s finished, his talent will be destroyed.’ […] ‘Many college students in America, they also go to college and at the same time realize…a lot of dreams outside of their studies. I think these two are not especially in contradiction.’ ‘Yes, okay. I agree with your point.’

While neither speaker here advocates prioritizing Follow Your Dreams over Success, they clearly see the importance of Follow Your Dreams in reaching a satisfactory resolution. Overall in the Chinese data the strong tendency is to come down on the side of Success. This is sometimes done explicitly, as we have seen, and sometimes implicitly, by adopting the parents’ viewpoint – that is, talking through what the parents, as opposed to Wang Er, would think (and assuming that the parents would privilege the Success model). This is in contrast to the Americans, who, while they acknowledge the importance of the practical considerations attendant upon the Success model, tend not only to adopt and maintain Tom’s viewpoint, but to support him explicitly as well. We can also see both intrapersonal and extrapersonal conflict in these examples. Intrapersonal conflict is evident in speakers’ uncertainty about which model to privilege. Extrapersonal conflict manifests in two ways. Sometimes speakers disagree with each other about which model to privilege. And sometimes one model “wins out” by being invoked and advocated more frequently across the speakers in a cultural community. The lesson to be drawn from these examples of intrapersonal and extrapersonal conflict is that both the Success and Follow Your Dreams models are available to speakers from both cultural communities, but are differentially privileged by each of these communities. 5.2. “Live And Learn”, “You Can’t Change Me” and “Independence” Two models that appear frequently in the American data are what I have called Live And Learn, and You Can’t Change Me:

Supracultural Models   95 Live And Learn: People, especially young people, should be allowed to make their own decisions. They will inevitably make mistakes, but from the negative consequences of these mistakes they will learn to make better decisions in the future. You Can’t Change Me: (Closely related) If a person is put in a position of doing something he does not wish to do, then he will do it poorly precisely because of his lack of desire to do it.

A clear invocation of Live and Learn, from E-19 and E-20, comes soon after (3) above: (11) E-19 E-20 E-19 E-20 E-19 E-20

You could let him go with his rock band thing, and if it takes off great, and if it doesn’t, then he’ll realize that he needs to do other things. The hard way? The hard way, yeah. But at least he’ll learn, right? Yeah. And he’ll know.

You Can’t Change Me also makes frequent appearances, but is perhaps given its clearest articulation by E-12: (12) E-12

I think he’s right too, I mean because if his parents force him to go to high school…I mean to go to college, and like…and so he ended up going to college instead of joining the rock band, he wouldn’t try hard, he wouldn’t study, he wouldn’t do any of that, because that’s not where he wants to be, so it’d be a waste of like…of his time and of like the parents’ money and of just…it’d be a waste of everything ‘cause like he won’t be trying hard because he doesn’t wanna be there, so it’d be better for him to do what he wants to do, because then he’ll…put in a lot of effort to what he wants to do. And like if he wants to be a rock star, then he’s gonna try really hard, and he’s gonna be happy doing what he’s doing, instead of like listening to his parents and being miserable.

Both Live and Learn and You Can’t Change Me are connected with a developmental model of Independence: at or around a certain age, a person assumes responsibility (which had before lain primarily with the person’s guardians) for making decisions of major import for her future. (For the remainder of this essay, in appropriate contexts lower-case independence will refer to this

96   Jason D. Patent particular form of “life-decision-making” independence.) This model has been institutionalized in the United States via, for example, the voting age and the drinking age. E-22 makes this explicit: (13) E-22

I’m thinking in the context of the U.S. domestic society, and once you’re eighteen you can do whatever you’d like. Uh, you know and uh…or you could…I mean basically you… you’re almost a legal adult, in some sense, and if he can support himself without his family means, then go ahead…there might be ways he could argue it, like you had said.

E-10 brings up this model as well, though not explicitly its institutionalization: (14) E-10

The only thing I think that would be wrong is I think at that point in time it’s probably time to let your kid decide on his or her own, and I don’t know be confident in your job of parenting, you know, and don’t expect them to follow your exact path if you teach them to be independent.

In six of the nine English conversations at least one of these three models is considered, usually strongly. In contrast, the Chinese data show only one instance of any of the three models, Live And Learn: (15) C-11 C-10 C-11 C-10

Juéduì bù xiàng nî shuõ huà shuõ de nènme qïngsõng. Nà kêyî ràng tã qû ge jiàoxùn ma. Chï yï qiàn, zhâng yï zhì ma. ‘There’s no way you’d be as cavalier as you’re saying now.’ ‘Let it be a lesson for him. Learn from his mistakes.’

That Live and Learn is available as a model for Chinese speakers is evident in how easy it is for C-10 to express this: he uses the following set phrase, or chéngyû, which is worth looking at in a bit of interlinear detail: (16) Chï yï qiàn, zhâng yï Eat one pit, grow one ‘Fall into a pit and your wisdom grows.’

zhì. wisdom.

If Live And Learn were not an available model, it is doubtful that such a set phrase would exist. Yet it is striking that in all six discussions of the Rock Band question in the Chinese conversations, this is the only invocation of Live and Learn. You Can’t Change Me and Independence do not appear at all.

Supracultural Models   97

This exemplifies a striking difference, this time from a mostly extrapersonal perspective: While Live and Learn, You Can’t Change Me and Independence form the core of the English discussions, these models seem barely available in the Chinese conversations. This statement requires a caveat. In (9) above, we saw C-18 invoke a model that looks much like Independence, in that he claims that Wang Er can decide for himself once “he has the definite ability to analyze problems and solve problems.” This indicates that there is indeed a corresponding Independence model – as we would surely expect – but it does not appear in relation to the Rock Band question because the timing in the two models is different: in the U.S., people reach independence around the age of 18, coinciding with the start of college; according to C-18, independence comes after college. The absence of the Independence model in any of the other Chinese responses to the Rock Band question may have resulted from Wang Er’s young age with respect to when independence is thought to begin in earnest. Thus Independence is indeed an available model; it is just more available in this scenario for the Americans than it is for the Chinese, due to the earlier onset of independence in American culture. Similarly, You Can’t Change Me is available as well for the Chinese participants. It makes an appearance in an answer to a different question, the “Draft” question: Zhèngfû juédìng yào dâzhàng. Zhãngsãn suïrán bèi yìngzhëng rùwû, dànshi tã xïn lî qiángliè rènwéi zhànzhëng shì bù duì de. Nî juéde, zài zhèi zhông qíngkuàng xià, Zhãngsãn huì zênmebàn? Tã yïnggãi zênmebàn? ‘The government decides to go to war. Chris is drafted into the army to fight, but he feels strongly that war is wrong. What will Chris say and do? What should he do?’

C-14, as part of her answer, offers: (17) C-14

Jiùshi shuõ kêndìng bù huì shì yï ge hâo zhànshì. […] Tã huì yï biãn qù shã rén, yï biãn juéde zhèi yàng de shã rén shì bù duì de. ‘That is to say he definitely won’t be a good soldier. […] At the same time he’s killing people he’ll be thinking that war is wrong.’

Both she and C-15 are conflicted. Immediately after (16) above, they say: (18) C-15 Zhèi ge wèntí duì nî lái tài kùnnán le. Suôyî…wô yê bù yídìng jiâng de hâo, yïnwèi…zhèi ge bù zhïdào zênme huídá.

98   Jason D. Patent C-14 C-15

Zhèi ge hên nán, zhèi…tã yïnggãi zênme bàn? Zhèi shì kàn nî bâ shénme fàng zài zuì qiánmiàn de, tã qiánmiàn tã yôu shùnxù de. Zhõngguó chuántông de shì guójiã shì fàng zài zìjî de qiánmiàn de, dàn wô xiãngxìn Mêiguó yê shì zhèi yàng de, suôyî fëicháng jiândãn. Tã jiù huì qù dâ zhàng, ránhòu tã huì zuò yï ge hên hâo de zhànshì qù dâ zhàng.

C-15

‘This question is too difficult for you. So…I also won’t necessarily be able to address this well, because…I don’t know how to answer this.’ ‘This is hard, this…what should he do?’ ‘That depends on what you put first. There’s an order to what is before him. Traditionally in China the country comes first, but I believe it’s also this way in America, so it’s really simple. He will go fight, and he will be a very good soldier in the war.’

C-14 C-15

In all of the Chinese discussions of all of the questions, this is the only instance of You Can’t Change Me (and it is promptly dismissed). As with Live And Learn, this tells us two things. First, as we have seen, You Can’t Change me is significantly more favored by American speakers than by Chinese speakers in response to the Rock Band question. But second, and more germane to this essay, You Can’t Change Me is available to the Chinese cultural community; it is in no way “lacking” as a cultural resource. It is simply not drawn upon by most speakers in this particular scenario. 5.3. “Surprise arrest”: “Good Reason” and “Disclosure” What was just presented was perhaps the starkest example in all the data of different models being invoked by the two cultural communities. In this section we will examine its diametric opposite: an instance of almost complete overlap. The overlap occurs in responses to “Surprise Arrest”: Suppose a citizen is walking down the street one day when the police arrest him.27 They don’t tell him why; they simply arrest him and keep him locked up for three 27 Except in cases in which gender specification was necessary for a question to be comprehensible, for all generic individuals in the interview questions I used males, for two reasons. First, I used males instead of females because I wanted the scenario to be as “unmarked” as possible. That is, given the markedness of gender,

Supracultural Models   99 days before letting him go. He has done nothing illegal. What will this person think? What will this person do? Jiârú yôu yï gè gõngmín, shénme zuì dõu méiyôu fàn, zài guàng jië de shíhou bèi jîngchá dàibû. Jîngchá bù gàosu tã wèishénme bèi dàibû, jiù bâ tã suô zài jiãnyù lî, yïzhí dào sãn tiãn yîhòu cái fàng tã zôu. Nî juéde zhèi ge rén huì zênme xiâng? Tã huì rúhé fânyìng? Câiqû shénme xíngdòng?

Before discussing the cultural models that appear in the responses, let us first take a model that must be presupposed in order for the scenario presented in Surprise Arrest to be something interpreted as negative: Well-Being. Surprise Arrest is problematic because the Well-Being of the protagonist is negatively affected.28 The first model to be discussed is the Good Reason model: if the police are going to inconvenience someone by arresting him, there must be a good reason for it. First a Chinese response: (19) C-11

Wô kêndìng huì yào nòng qïngchu tã wèishenme yào zhèi yàng zuò. Rúguô tã méiyôu rènhé dàoli dehuà, tã yïnggãi câiqû…tã yïnggãi duì wô câiqû yìxië shénme yàngzi de bûcháng, dàoqiàn. ‘I would definitely want to get it clear why he did this. If he didn’t have any good reason29 for what he did, he should take…he should take measures to compensate me, and apologize.’30

interviewees, in noting (consciously or unconsciously) the presence of a female “character,” might then “gender” the question and seek gender-based explanations, which would distract from the individual-versus-state type of discussion I was seeking. Second, I used exclusively males because I did not wish to introduce another parameter of variation. 28 There is much to say about Well-Being as a cultural priority (see, e.g., Lakoff 2002), but here we will leave this at a highly schematic level: the putative citizen’s Well-Being has been harmed. 29 What I have translated as reason is the Mandarin term dàoli. This would translate more accurately as reasonableness, but this seems to push the fidelity/elegance line too far toward fidelity. The point is the same: what the police officer did is not reasonable, which is another way of saying that there was no good reason. What exactly this model is about will be discussed below. 30 The Mandarin term, dàoqiàn, is the most severe sort of apology, causing a complete loss of face to the one doing the apologizing, the kind that the Chinese government demanded of the U.S. government after the downing of the American E-P3 surveillance airplane in April 2001. For a fascinating cross-cultural (China– U.S.) study of apology based on this event, see Gries and Peng (2002).

100   Jason D. Patent This is highly typical of the Chinese responses. A natural question to ask is: what constitutes a good reason? The following American response sheds some light: (20) E-10

I think, my personal opinion is it’s understandable that if, if, like say there’s some crazy person going around and like shooting everybody, and you like match up with his stuff or something and they see you and they arrest you, I mean, I mean granted I think they should…check the possibility, of you being that person, if you like match them up identically, but they should tell you that.

Here, the Good Reason model is revealed as based on a tradeoff of well-being. The default case is that people’s well-being should be respected. If well-being is to be deliberately undermined, then there must be a tradeoff with someone else’s well-being. For instance, here E-10 is willing to forgive the officer if he believes that arresting our protagonist may actually positively affect the wellbeing of others, by protecting them from a possibly dangerous criminal. Note also the end of E-10’s passage: “…they should tell you that.” C-11 similarly says, “I would definitely want to get it clear why he did this.” These statements exemplify the Disclosure model, which is closely related to Good Reason: People, in China as in the U.S., are expected to behave in a “reasonable” fashion, that is, to have good reasons for their actions. This is especially true of people who have the power to affect our well-being negatively, such as the police. But people cannot always be trusted to evaluate the reasonableness of their own actions, and so Disclosure is important: in cases where one’s well-being is violated, the person undergoing the violation can help assess the degree of reasonableness. If the arresting officer doesn’t have good reasons for his actions, then what he did was wrong. Or, if he has good reasons but doesn’t disclose them, then what he did was wrong. The upshot up the interaction between Good Reason and Disclosure is that the person whose well-being is violated must somehow be allowed to consent to the action. The person may not have any practical means of resisting, should he not consent, but at the very least the person will not have to waste precious energy wondering why he has been arrested. The Chinese and American responses show surprising uniformity. Some more examples: (21) C-17

Jîngchá méiyôu gàosu tã gàn shá, jiù báibái dàibû tã sãn tiãn ránhòu gêi fàng le? Zhèige shìqing bù kênéng fãshëng ba? Fânzhèng huì juéde hên mòmíngqímiào, zênme gâo de?

Supracultural Models   101

Wô kêndìng yào gâo qïngchu nî wèishenme nèi yàng zhuã wô. ‘The police didn’t tell him what he did, just locked him up for no reason for three days and then let him go? This sort of thing couldn’t happen, could it? Regardless he would feel extremely confused, how could this happen? I would definitely want to clarify why you arrested me like that.’ (22) C-18

C-19 C-18 C-18 C-19 C-18 (23) E-19

(24) E-23

E-24

Wô juéde rúguô wô shì zhèi yàng de rén dehuà, wô huì juéde fëicháng bùkêsïyì, wô jìrán méi fàn rènhé de cuòwu, wèishenme yào bâ wô guãn qilái? Wèishenme bâ wô suô zài jiãnyù lî? Érqie méiyôu rènhé jiêshì. Duì. ‘I think that if I were this sort of person, I would feel it was inconceivable. Since I didn’t make any mistake, why lock me up? Why lock me in jail?’ ‘Plus there’s no explanation at all.’ ‘Right.’ If that was me, I’d be angry. Because, they didn’t even tell me what I did. And if I hadn’t done anything wrong then there’s no reason to arrest me. You should at least know, like, the reason. Obviously, I would…he would be really angry, and…I think would start to become very bitter about the government. What kind of system is this, that he didn’t do anything, they don’t tell him his rights, or why he’s…he’s being arrested and he’s, um, illegally detained. Yeah. I’d be really pissed off.

E-23 and E-24 specifically mention anger, which also shows up consistently in the Chinese responses. So for Surprise Arrest, responses are similar across both cultural communities, in contrast to Rock Band. And within each cultural community, model conflict, both intra- and extrapersonal, is virtually absent. To round out this tour of similar and differing responses, we turn to another question.

102   Jason D. Patent 5.4. “Tax Hike” The “Tax Hike” question asks: The government passes a law doubling the income tax without consulting the citizens. Is the government right to do this? What would citizens say? What would they do? What should they do? Rúguô zhèngfû méiyôu dédào rénmín de tóngyì, jiù zìxíng juédìng bâ suôdé shuì zëngjiã yï bèi, nî juéde zhèi yàng duì bù duì? Rénmín huì yôu shénme fânyìng? Câiqû shénme yàng de xíngdòng? Nî juéde rénmín yïnggãi rúhé huíyìng zhèi yàng de zhuàngkuàng?

The Chinese responses are noticeably milder than the responses to Surprise Arrest. Overall, the general sense is that while the government’s actions are indeed wrong, people would not feel very angry or feel inclined to take strong measures in response. Consider the following: (25) C-21

Jînjîn shì yïnwèi suôdé shuì zëngjiã yï bèi dehuà bù huì yînqî hên dà de qìfèn. Dànshi jiùshi shuõ yìbãn de zhèi ge zhêng gè shèhui dàjiã dõu huì fânduì zhèi jiàn shìqing, nà kênéng shì, tã yôu hên duõ hên duõ wèntí, dàjiã fëicháng bù mânyì, ránhòu kênéng huì jïjí qîlái fânduì. Rúguô…dàn biéde dõu hên hâo, zhî shì jiù tí chø zëngjiã yï bèi bù huì yôu hên dà de nèi ge fânyìng. ‘Just doubling the income tax wouldn’t cause much anger. But, but the whole society, in general, would oppose this if, if possibly, [society] many, many problems, everyone is extremely dissatisfied, and then may actively stand up and oppose it. If…but if everything else is good, and only the matter of doubling the tax is raised, there won’t be a very big reaction.’

If the tax hike doesn’t affect the citizens’ living conditions in a profoundly negative way, and if society as a whole is in good shape, then this sort of action by the government will not be responded to with much vigor. This theme is sounded throughout the Chinese data. The Disclosure model is completely absent, and Good Reason only appears once: (26) C-17

Nî dêi dédào rén…jué dà duõshù rén tóngyì. Yêxû wô bù tóngyì, dànshi rúguô…bîrúshuõ wô bù tóngyì dàn bâifënzhï bãshí rén tóngyì, wô yê méiyou yìjian. Wô jiù huì fúcóng dà

Supracultural Models   103

duõshù rén de yìjian. Dànshi xiànzài shéi zhïdào nî zhèi ge zhëng shuì nî yuánlái mùdi shì gàn shénme. ‘You must obtain the people’s…the consent of the vast majority of the people. Perhaps I don’t agree, but if…for example I don’t agree but eighty percent of the people agree, then I have nothing to say. I’ll just follow the opinion of the majority of the people. But now who knows what the original goal of your collecting these taxes is?’ Unfortunately there is no further elaboration. Still, at least C-17 thinks it somehow matters why the government is raising taxes. Overall, the sense in the Chinese responses is one of resignation.31 In terms of cultural models, they tend to invoke a model I have elsewhere32 dubbed Nothing To Be Done. This model comes into play when it is determined that a problem cannot be solved, and so no further energy should be expended even on contemplating it, let alone trying to solve it. The American responses are different. E-19 and E-20, for instance, are upset: (27) E-19 E-20 E-19 E-20 E-19 E-20 E-19

What would the citizens say? They’d be very pissed off. Like, really pissed. But then like, governments do stuff all the time that don’t represent like what all your citizens want. But how many people really want their income tax raised? Actually yeah, most people would respond to that. Negatively. People…yeah, ’cause that would be like a…a united opinion, like everybody would be, like, no.

Other participants suggest that the people would rebel, perhaps even violently. Still, most are willing to allow for the possibility of there being a Good Reason for the tax hike: (28) E-22

To double the income tax, in certain situations, you know maybe, I mean I…I think it’d also have to be in the context of like a situation, is…is it a temporary income tax, or is it

31 Surely the political realities of the PRC play a role in this sense. See Patent 2003 (168–171) for a discussion of this and of broader issues of how politics and culture interact. 32 See Patent (2003: 168).

104   Jason D. Patent permanent…on a permanent basis? Um…is it just in a time of war, that they need more funds? (29) E-23 E-24 E-23 E-24

Where is the money going? …maybe there’s a…maybe there’s a r…emergency or something like a war or like some… Economic crisis. Yeah.

In other words, if the extra money from the tax hike were going to some important cause,33 then it might be acceptable. Disclosure is also mentioned in this connection: (30) E-26

But, I mean just in case there is a reason for the citizens should demand an explanation [sic]. I mean, perhaps it’s wartime, and the g…you know, country’s about to be taken over, then maybe a temporary change would be…warranted.

Just as in Surprise Arrest, the Americans see the actions of the offending party – here, the government – in moral terms, hence their invocation of Good Reason and Disclosure, which revolve around the moral calculus of informed acquiescence.34 Chinese responses, in contrast, are framed in more “practical”35 terms: if the problems brought about by the tax hike are not severe, then people will not respond vigorously. What seems to have happened, then, with Surprise Arrest and Tax Hike is that Chinese and American respondents have categorized these two scenarios in different ways. The two scenarios are seen by the Chinese participants as different sorts of events, whereas to the Americans the scenarios are seen broadly as members of the same category of event. And, crucially, our evidence 33 The fact that people in five of the thirteen interviews mention war as a good reason is perhaps due to timing: all interviews were carried out in autumn of 2001, when memories of September 11 were fresh, and many in the U.S. felt more bellicose than at other times. 34 For a particularly interesting study contrasting a “Western,” context-independent, “justice-based” view of morality with “non-Western” context-dependent, particularistic notions of morality, see Shweder, Mahapatra and Miller (1990). 35 The quotation marks stem from a long-standing essentialization of “Chinese culture” as “more practical” than “Western culture.” One of the projects of Patent 2003 was to problematize this essentialization. In the present instance, at least, the description seems apt, and not likely to do harm when scare-quoted.

Supracultural Models   105

for this is the cultural models that do or do not get invoked in people’s answers to the questions. For the Tax Hike question – as with Surprise Arrest – both intra- and extrapersonal model conflict are virtually absent: each cultural community is fairly unanimous in its responses. 6. Foundational schemas Each of the preceding data sections has presented a small piece of the larger cultural puzzles at hand. In this section I endeavor to sketch out how the pieces fit together, and thereby to give a more complete picture of the differences we have seen in the Chinese and American responses. In his 1996 book Culture in Mind, Bradd Shore attempts a taxonomy of cultural models. One of his key categories is the foundational schema. Foundational schemas “organize or link up a ‘family’ of related models.” (1996: 53). Shore sums up nicely the import of foundational schemas to studies like the present one: Not all cultural models have an encompassing foundational schema. Many cultural models are special-purpose models with no family resemblance to other models. On the other hand, those models linked through foundational schemas have a special status in any community, contributing to the sense that its members live in a world populated by culturally typical practices and a common worldview.36 (53)

In Shore’s terms, the relationship of model to (foundational) schema is that of specific to general: a schema consists of highly underspecified knowledge generalizable over many specific models. An example of a foundational schema in the United States might be what Deborah Tannen (1998) found to be America’s fascination with argumentation as a means of resolving disputes. On the level of Shore’s “cultural model” we have: lawyer-versus-lawyer courtroom 36 As we well know, the term worldview (along with terms such as national character) is laden with difficult questions of colonialism and imperialism, as colonizers have time and again essentialized “the other” through recourse to such terms. This has led many anthropologists to abandon culture, to the extent that it has been identified with worldview, as a useful category of inquiry. One of Shore’s projects is to resurrect culture as the central focus of anthropology. I refer the reader directly to Shore (1996), specifically his discussion on pages 8–11. Within psychology, research on worldview as a central psychological construct has recently been revived. For a history of the field, and for a proposed theory of worldview, see Koltko-Rivera (2004).

106   Jason D. Patent argumentation, television programs such as CNN’s Crossfire, presidential debates, and so on. The foundational schema is what captures the structure that is shared among all these models: the existence of verbal conflict, along with the view that such conflict is a good way to resolve disputes, or at least to inform people about issues.37 Shore argues that foundational schemas are a crucial factor in understanding the systematicity underlying what might superficially appear to be disparate phenomena. It is in foundational schemas that intuitive notions of cultural difference may be characterized in an empirically responsible way.38 I have identified two different foundational schemas among the Chinese and American respondents. They are, respectively, Problem Management and Moral Individual.39 According to the Problem Management schema, to be a person means to be engaged in a struggle to minimize the trouble one has to deal with in life. Particulars of given situations are given priority over generalizations meant to cover all possible situations.40 One problem a great many people must solve is 37 This is surely connected as well to other models of competition in American culture. The foundational schema that captures the commonalities of all competitionbased cultural models would be yet more schematic than the one just discussed here. 38 Shore points out (1996: 69–70, fn. 12) a difficult issue confronting researchers making claims about the psychological reality of foundational schemas: how do we know the researcher is not simply imagining the existence of a schema? The short, nihilistic answer is: there is no way to know anything at all. Shore’s longer, more nuanced answer is: there exists a great deal of evidence that schema induction is something people often do in fact engage in when reasoning analogically. Plus there is a strong subjective sense, not only among ethnographers but among anyone involved in understanding cultures other than one’s own, that what we experience as “cultural difference” must result from more than a haphazard collection of specific models, stemming instead from what is shared among those models. To this we might also add Langacker’s (1987) claim that much of our knowledge of the world is stored at a relatively schematic level, and that one thing humans do simply by dint of being human is extract schemas from specific-level information. 39 I cannot do these schemas justice here in all their complexity; for a fuller account, see Patent (2003: 240–264). 40 While it is risky to speculate on the relationship between philosophical traditions and present-day cultural models, two such attempts are worth noting. First, F.S.C. Northrop (1946: 375–404) writes of the valuing in “Oriental” philosophy of the (context-specific) “indeterminate” over the (eternal, all-encompassing) “determinate.” He describes the importance of “treat[ing] the determinate portion of one’s introspected convictions and the determinate portion of other persons and things for what they are, namely, factors which do not hold for everybody under all circum-

Supracultural Models   107

establishing an economic foundation for oneself and one’s family. Privileging the Success model provides one with the tools to solve this problem: do what you need to do in order to make sure your economic needs are met. Yes, we may have desires to do certain things with our lives (Follow Your Dreams), but if these desires are leading us down a path that will result in not having a solid economic foundation, then pursuing one’s dreams is not worth the risk of not having enough money to pay the bills or to eat. It is also possible to make a mistake and learn from it (Live And Learn), but why bother if it is obvious, from generations of experience and accumulated wisdom, which path will lead most quickly to success and allow us to solve perhaps the most important problem of our entire lives? The Problem Management schema affects reasoning about the Surprise Arrest and Tax Hike questions as follows: A typical Chinese interviewee sees the scenario presented in Surprise Arrest as a gigantic personal problem that requires solving. A tremendous amount of inconvenience is created for the citizen, resulting in anger. Still, it is a relatively small-scale, tractable problem: specific, identifiable individuals are responsible. It is possible to deal with these specific individuals both by reporting their misdeeds to their superiors and by suing them in court. Concrete action on the part of the individual citizen is enough to redress the problem. From this problem-solving perspective, the Tax Hike is completely different. It may be annoying to pay more taxes, and it is not right of the government to raise taxes without consulting the citizens, but the ultimate effect on any individual is relatively minor. Furthermore, the problem is society-wide, and to redress it would simply be impractical, even potentially dangerous. For both the Surprise Arrest and Tax Hike questions, the Chinese perspective is one of problem-solving. Americans, on the other hand, tend to reason in terms of the Moral Individual schema. This schema begins with the assumption that human beings are born (or, in some circles such as Fundamentalist Christianity, conceived) into a community of individuated beings who have a purpose. The purpose is to bestances, and hence as factors which one has no right to elevate into immortal commandments, giving one the supposed moral and religious duty to force them upon other people…” (387) Second, Linda Young’s (1994) findings on differing Chinese and American communicative styles, which she ties together with Chinese philosophy and poetry, also dovetail with my findings of a more particularistic, contextbased emphasis in Chinese culture, in contrast to Americans’ abiding universalism. For other, convergent findings from philosophy, see Nakamura (1985/1964) and Munro (1969). For a survey of philosophical differences between “East” and “West” in the service of cross-cultural psychology, see Nisbett (2003: 1–28).

108   Jason D. Patent have in a moral fashion, or to do what is right. Naturally, in the course of one’s life, morality will come into conflict with other desiderata, but our primary purpose in this life is to be “good people.”41 A key aspect of Moral Individual is each person brings something unique to the world. Part of our uniqueness resides in our talents and skills, and our uniqueness is wasted if we don’t develop our talents and skills as thoroughly as we can. An important aspect of developing our uniqueness is that this development should be undertaken for its own sake. There are no ulterior motives; it is its own reward. But it also may make a person happy to develop her talents, and when it does, it is important to Follow Your Dreams: when we have unique talents and developing these talents makes us happy, we should make developing these talents our top priority. Live And Learn reveals another, related aspect of self-definition: that the best way to learn to make good decisions is to allow a person to decide things for herself. The risk is that she might make a bad decision, but she will learn this lesson well because it was she who made the bad decision, and she who suffered the consequences. This gives her a deeper understanding of good decision making because she has tried an alternative path and discovered that it wasn’t what she had hoped it would be. This idea of the “road not taken” is key: if instead someone else had made the “right” decision for her, she herself wouldn’t know that the decision was “right” not only because she would resist (You Can’t Change Me), but because she never would have had the opportunity to observe the “wrongness” of the other path and to experience the consequences of having taken that path. She would have less data on which to base future decisions. Where Problem Management privileges the particular, Moral Individual favors the general. Subscribers to the Moral Individual schema are engaged in a never-ending search for principles that govern all possible situations. Specifics of a given situation are important mostly to the extent that they instantiate, or help us learn or draw conclusions about, general principles governing human behavior.42 Now that we have sketched these foundational schemas, and have seen how the various models fit into them, we are in a position to revisit the data and see it in terms of a view of culture that simultaneously avoids the pitfalls of overly essentializing difference, while at the same time offering an empirically responsible account of such difference. 41 What exactly is considered moral behavior varies a great deal within American culture, as most thoroughly analyzed by Lakoff (2002). 42 There is of course a vast philosophical literature spanning several centuries addressing issues of the individual, agency and “the good.” For a fascinating summary of the history of some of these ideas, see Taylor (1989).

Supracultural Models   109

7. “Equivalence” Seven specific cultural models have been introduced, and claimed to be available to both cultural communities: 1. 2. 3. 4. 5. 6. 7.

Success Follow Your Dreams Live And Learn You Can’t Change Me Independence Good Reason Disclosure

To this point I have been acting as if it is more or less unproblematic to assume that each of these models’ nearest counterpart in the other culture can be called by the same name. This is, however, not at all unproblematic. Addressing this concern will be done through a series of smaller tasks: (1) assessing the validity of the level of specificity at which the cultural models have been presented; (2) resolving some concerns about what constitutes evidence for the presence of a given cultural model; (3) addressing methodological issues that arise when the investigator is not a native of at least one of the cultures under study; (4) making precise what is meant by “equivalence”; (5) making precise what is meant by “availability.” 7.1. Level of specificity Positing the above models encourages adopting the convenient fiction that each model is somehow discrete. As we have seen, though, models are anything but discrete, relating to one another in complex ways, and in many cases linked to one another via a foundational schema. Still, positing separate models is investigatively crucial in allowing us to give names to what are felt intuitively to be pieces of a larger whole. For instance, Follow Your Dreams refers specifically to the notion that people should pursue what they most enjoy doing. The model can only be fully understood with respect to the Moral Individual foundational schema, even as it has an intuitive life of its own. My use of the word intuitive is intentional: this sort of research simply does not lend itself to laboratory verification or to precise analyses of category structures. Still, there may be hope for salvaging a more empirically solid basis for this intuition. For instance, D’Andrade (1987: 112), citing earlier work in psychology, points out that cognitive schemas – a term closely related to what we

110   Jason D. Patent are calling cultural models – tend to be limited to roughly seven in the number of entities they contain, due to the constraints of working memory. It could be that the cultural models enumerated above are of a particular cognitive “size,” feeding the intuition that each model is a cognitively useful “chunk.” In this view, each relevant foundational schema is simply too complex, and meant to cover too wide a range of scenarios, to be felt to be useful on the same level of specificity as are the particular cultural models. Overall there is plenty of research to support the notion that cultural models exist at different levels of specificity.43 I will therefore proceed under the assumption that the distinction between model and (foundational) schema is a useful one. 7.2. Evidence What can be counted as evidence for a given cultural model in all its complexity? This is difficult, especially when much is left implicit. The investigator is forced to look for as many clues as possible as to what background knowledge and understandings must be assumed in order for a given bit of discourse to be expected, by the speaker, to make sense to the hearer. The investigator must bring to bear every last bit of evidence he can, including, in the present case, the systematic way in which the various models relate to one another and to the overarching foundational schemas. Naomi Quinn, in her study of cultural models of American marriage (1987: 190–191), put it this way: Thus, only by deciphering certain American cultural understandings of the self can we fathom the connection in Americans’ thinking about marriage between its benefits and its difficulties. The passages we analyze in this paper give only a sampling of that folk psychology; discussions of needs and their fulfillment arise naturally in the course of talk about marriage and are scattered throughout the entire body of discourse under study. This sporadic evidence must be drawn together to permit reconstruction of the cultural model of needs and their fulfillment sketched here – so that the application of this folk psychology to marriage can be appreciated.

Thus, while there is no definitive way to delineate what does and does not count as evidence for cultural models, at least the following range of clues may be sought and used: explicit statements containing particular linguistic forms; extralinguistic events such as pauses and false starts; commonalities

43 See, e.g., Quinn and Holland (1987: 32–25), D’Andrade (1987, 1995: 150–181), Quinn (1987: 189–191), Lutz (1987: 299–302), and of course Goffman (1974).

Supracultural Models   111

and differences across a range of topics; viewpoint and viewpoint shifts.44 To the extent that the evidence converges on a single conclusion, it tends toward validity.45 As is already apparent, these types of evidence have all been used in the present study, and have converged on a particular picture of personhood in Chinese and American culture. 7.3. The non-native investigator Evidentiary problems are of course compounded when native intuition – a form of evidence used, e.g., by Sweetser (1987), by Lakoff and Kövecses (1987) and by Kay (1987) – is not available, as is the case for the present investigator visà-vis Chinese language and culture. My having spent, at the time this analysis began, around 15 years studying the Chinese language, and approximately four years living in China – speaking mostly Chinese and interacting extensively with locals in many parts of China and in many different settings – should, I think, partially offset this disadvantage. A further offsetting is my having checked these findings extensively with native Chinese. 46 My hope is that the evidence presented here, while not definitive, will nonetheless be convincing to the reader. 7.4. Principles for establishing equivalence Identifying specific models within one cultural community is of course but one small part of the picture. This study is cross-cultural, necessitating comparisons between models in two different cultural communities. This in turn 44 For a fuller discussion of method and evidence, I again refer the reader to Quinn (2005). 45 For more on convergent evidence, see Gee (1999: 94–96). 46 Muriel Saville-Troike (2003: 88–92) provides a nice summary of these issues. Here is the most relevant portion of her discussion (90): “When ethnographers choose to work in other cultures, the need for extensive background study of the community is critical, and a variety of field methods must be employed to minimize imposition of their own cultural categories and perceptions on recording the interpretation of another system. In some cases ‘outsiders’ may notice behaviors that are not readily apparent to natives of the community, for whom they may be unconscious, but conversely no outsider can really understand the meaning of interaction of various types within the community without eliciting the intuitions of its members.”

112   Jason D. Patent necessitates identifying “counterpart” models: models that are deemed to be roughly equivalent in two different cultural communities. Up to this point, for the sake of easy exposition, I have not addressed (apart from some footnotes) the problems inherent in identifying counterparts. I now turn to this task. All seven of the specific models addressed above have been claimed to have counterparts in each culture under investigation. What is the basis for the proposed equivalence? Is it justifiable to claim, for instance, that the American Follow Your Dreams Model is meaningfully “the same” as the Chinese Follow Your Dreams model? My claim is that, while the investigator needs to be careful in assigning equivalence, meaningful equivalence is nonetheless discoverable, and indeed holds of all seven models addressed in this essay. The task is inevitably subjective, but no more subjective than qualitative analysis generally. A useful starting point is provided by Boster (2005: 213–218). In his study of Waorani emotion terms, one of his tasks was to establish cross-linguistic equivalents for various English language emotion terms. Based on an analysis of facial expressions, Boster identified the closest Waorani equivalent of English frightened or scared to be ankai giñente. A question was then asked: “What causes you to feel ankai giñente?” Two frequent responses are “I hear a jaguar howl in the forest” and “seeing a poisonous snake on the path.” Another emotion, the equivalent of English happy, is tote, and a common response to “What causes you to feel tote?” is “I spear a wild pig with a lance.” Boster then summarizes (2005: 214): I imagine that I could ask 10,000 Americans what causes them to feel happy, and not one would offer I spear a wild pig with a lance. Nevertheless, Americans would be able to infer that it might constitute “making reasonable progress toward the realization of a goal” for the Waorani. Similarly, although the same 10,000 Americans might never offer I hear a jaguar howl in the forest as a reason to feel fright, they would be able to infer that it might constitute “facing an immediate, concrete, and overwhelming physical danger” for the Waorani.

Boster then goes on to investigate American responses to the emotion-inducing scenarios proposed by the Waorani, finding, not surprisingly, that equivalence is detectable, even if only partial. The point of presenting Boster’s research here is to show that while any given cultural model in all its specificity may not have an exact equivalent in another culture, there may well still be shared structure at a more schematic level, and this shared structure can motivate and justify positing some degree of equivalency. This is the approach I follow here, as we look at each of the seven cultural models discussed previously.

Supracultural Models   113 Success Shared structure: the necessity of establishing minimal economic self-sufficiency. Differences: While the American model stops at economic self-sufficiency, the Chinese model also focuses on public perceptions, specifically on finding a “creditable” profession. Follow Your Dreams Shared structure: Desire is important. If someone has something they particularly want to do with their lives, then this must be considered in any assessment of potential careers or other life-long public pursuits. Differences: The American model, informed as it is by the Moral Individual foundational schema, is inevitably bound up with an individual’s identity and uniqueness. No evidence of these aspects has yet been found in the Chinese model. Live and Learn Shared structure: Valuable lessons can be learned through making mistakes. Differences: We may surmise, based on the Problem Management schema, that the Chinese model is based mostly on problem-solving: we can solve problems better if we have previously made mistakes trying to solve them in other ways. In contrast, the Moral Individual schema dictates that Live And Learn is about learning to be a particular kind of independent and autonomous adult, not dependent on others for instruction. You Can’t Change Me Shared structure: A person cannot be forced to do something well against her will. If forced, a person will not perform the desired task well. Differences: Based on the available data, I have not yet found any differences.47 Independence Shared structure: As a person ages from infancy through childhood and into adulthood, he takes on more and more responsibility for making decisions about his actions. 47 You Can’t Change Me requires some notion of intentions leading to actions and results. If indeed an individual’s will, including intentions, is believed to affect one’s course of action in American culture, then it makes sense that You Can’t Change Me would be strong. Contrariwise, to the extent that individual will is subordinated to other factors in Chinese culture, or to the extent that will and desire are not seen as linked to action, You Can’t Change Me would not be as salient. One place to look for differences between the Chinese and American versions of You Can’t Change Me would be in models of the mind, specifically how states of intention and desire relate to action. D’Andrade (1987: 115–127) posits a common American model of the mind that is surely different from any common Chinese model. For now, however, this must remain speculative.

114   Jason D. Patent Differences: Minimally, the American model puts putative independence earlier (post-high-school) than does the Chinese model (post-college). Additionally, in the Chinese model one maintains deeper connections and obligations to one’s parents for quite a bit longer than in the American model. Good Reason Shared structure: When one person’s well-being is intentionally violated, there must be a concomitant improvement of other people’s well-being. Differences: Based on the available data, I have not yet found any differences. Disclosure Shared structure: When a good reason is present for violating someone’s well-being, that good reason must be communicated to the victim, so that he may at the very least stop wasting energy on wondering why his well-being has been violated, and at the very most so that he may opt out of the well-being violation, should there be an option. Differences: Based on the available data, I have not yet found any differences.

We can see from this discussion that even when two models may be seen as rough equivalents, we must not lose sight of the foundational schemas connected to the specific models, as these foundational schemas inform the specific models in a deep way. For instance, Live And Learn may well be fundamentally about different things in the cultural contexts characterized by the Problem Management and Moral Individual schemas: the former is about making mistakes so that one can better solve future problems; the latter is about a person developing into an independent, heuristically sound decision-maker. Such differences do not invalidate equivalence; rather, they help us better to problematize the limitations of equivalence. Having analyzed just how the relevant cultural models in the two cultural communities relate both to one another within a cultural community and across cultural communities, we may now take up the final topic of this study: availability. 8. “Availability” Throughout this essay I have used the term available somewhat loosely, trusting in an intuitive notion of what the word means. In this section I endeavor to delineate more precisely a somewhat technical meaning of the term, and in the process to point the way toward a new conceptualization of how cultures dif-

Supracultural Models   115

fer. We must address two questions: (1) What is meant by availability? (2) How does the researcher recognize it? Below I will handle these questions together rather than sequentially. Most simply stated, a cultural model is available to a member of a cultural community if it is among the models that person may choose (consciously or, more likely, unconsciously) to use in reasoning about any given situation. In this simplest sense, then, a vast range of cultural models is available to anyone in any situation: we have seen in this paper, through all the examples of intrapersonal and extrapersonal conflict, that indeed many models are available in any given situation. Would we want to say, however, that all models are equally available? For instance, it seems clear that Live And Learn is available to the Chinese participants. There is a fixed expression in the language for this model, and it is actually invoked by an actual member of this cultural community. Yet it is only invoked by this one speaker, and is not advocated for very strongly. It seems dubious to claim that, relative to the Rock Band question, Live And Learn is as available as are other models to the Chinese cultural community.48 Availability, then, is not a binary phenomenon. It admits of degrees. To cite further examples: we would not want to say that Follow Your Dreams and Success are equally available to the Chinese participants in discussions of the Rock Band question. The strong tendencies within the Chinese cultural community, when addressing the Rock Band question, are: (1) to invoke Success as the first model to be considered, and (2) to privilege Success when it comes into conflict with other models, most notably Follow Your Dreams. Likewise, we would not want to say that Nothing To Be Done is as available to Americans as is Good Reason when it comes to the Tax Hike question. What is meant, then, by this more technical sense of availability is that a given cultural model is available to the extent that it is likely to be (1) invoked, and (2) prioritized. A researcher may identify a given model as more available to the extent that she finds evidence for the model, and to the extent that the model is (a) invoked by a significant proportion of members of a cultural community, and (b) argued for as the model that should get top priority in cases of conflict.

48 Farzad Sharifian (2003, Forthcoming) has examined what I am calling availability in the service of a separate but related project: investigating the degree of sharedness, within a cultural community, of cultural models and other “cultural conceptualizations.” Specifically he writes of the heterogeneous distribution of such conceptualizations.

116   Jason D. Patent 9. Theoretical implications revisited Having introduced the technical notions of equivalence and availability, we are now in a position to explicate more precisely what is meant by “supra­ cultural model.” A supracultural model is the schematic structure shared by any two (or, in principle, more) cross-culturally equivalent cultural models that are available to members of the relevant cultural communities. This obviously encompasses a vast number of cultural models, which is exactly as it should be: as I have argued throughout this essay, superficially divergent cultural data may, upon second look, contain more commonality than had been immediately apparent. While cultural difference rightly alerts us to the ways in which the world’s cultures are unique, and thus inclines us toward relativism, it may also be the case that universality has been hiding in places where we had not looked for it before. The present study only deals with two cultural communities, and with only a rather circumscribed set of cultural models within those communities. I would therefore want to stop far short of claiming even the mildest universality for any of the cultural models discussed here. Yet at the same time all these cultural models seem at least to deal with human experiences that are in some sense “basic,” and therefore likely to be shared widely, if not universally, among different cultural communities (in other words, likely to be part of Shweder and Sullivan’s “heterogeneous complex”): getting enough to eat, learning through trial and error, linking desire and intention to action, expecting (in unmarked cases) to be treated well. In this view the particular manifestations of these highly schematic, potentially universal notions are “free” to be instantiated differently in different cultures (in ways that are likely to be guided by foundational schemas), but will always share some important schematic structure. This is exactly the view of culture that is borne out by my findings. Additionally, from the foregoing discussion it should be clear that culture is anything but static and monolithic. It is, rather, fluid, flexible, and dynamic. In this vein my findings resonate with those of a handful of previous scholars from a number of disciplines. Sociologist Ann Swidler (1986) uses the term toolkit to refer to what we would call an inventory of cultural models. Anthropologist Ulf Hannerz (1969: 177–195), uses the term repertoire. Sociologist Charles Tilly (1992) also employs the term repertoire, in his brilliant piece of historical–political ethnography of French and British “contentious gatherings.” Linguist Paul Kay (1996: 110), uses both tool box and repertoire. These notions are present in what anthropologist and sociologist Pierre Bourdieu (1977: 82–83) terms the habitus, which he defines as “a system of lasting, transposable dispositions which, inte-

Supracultural Models   117

grating past experiences, functions at every moment as a matrix of perceptions, appreciations, and actions and makes possible the achievement of infinitely diversified tasks….” (emphasis in original). In Bourdieu’s slightly later work (1990/1980), which elaborates on the habitus as intermediate between determinism and freedom, he writes that the habitus “is what makes it possible to produce an infinite number of practices that are relatively unpredictable…but also limited in their diversity.” (55). And finally, psychologists Donald Briley, Michael Morris, and Itamar Simonson posit a model of culture as a diverse set of cognitive structures49 that may be brought into action differently in different contexts – see Briley, Morris, and Simonson (2000). The view of culture sketched in this essay cautions us against adopting dualistic approaches to describing cultural systems. In the present case this means we should treat with caution, and address with rigor, attempts to break down conceptions of the self and of the person into “Western” and “non-Western.”50 As Melford Spiro (1993: 144–145) aptly puts it: …[A] typology of the self (or of personality) that consists of only two types – a Western and a non-Western – is much too restrictive to accurately describe either, and only serves to distort both. Thus there is evidence for the proposition that many putative characteristics of the Western self, which allegedly make it “peculiar,” are to a greater or lesser degree also found in the non-Western self; conversely, many putative characteristics of the non-Western self, which allegedly are its distinguishing features, are to a greater or lesser degree also found in the Western self.

A similar statement comes from Markus, Kitayama and Heiman (1996: 862): “An analysis of other cultural contexts may well reveal the unknown and the unshared, but it may also illuminate the obscured, the denied, or the subterranean features of the researchers’ own backyard.” It is the hope of this author that the present study may draw our attention more to the “subterranean” in our own and in other cultures, and that we may thus begin to unearth – from the

49 There is a long history of debate in anthropology (as well as in psychology) about whether culture is properly viewed as knowledge or as publicly available forms, with Ward Goodenough (1957) as the archetypical proponent of the former, and Clifford Geertz (1973) of the latter. See Bradd Shore (1996: 15–41) for a history of, and proposed resolution to, this debate. 50 This is not to say that there are not legitimately different generalizations to be drawn. It simply means that we must be careful in how we characterize and problematize these purported general characteristics. For some scholarly debate on the matter, see Murray’s (1993) criticism of Shweder and Bourne (1984) and of Lutz (1988), and Spiro’s (1993) criticism of Geertz (1984) and of Markus and Kitayama (1991).

118   Jason D. Patent enormous differences apparent to anyone who has experience with a culture other than her own – our shared humanity. References Bond, Michael H. (ed.) The Cross-Cultural Challenge to Social Psychology. Beverly Hills, CA: 1988 Sage. Bond, Michael H., and Kwang-kuo Hwang The social psychology of Chinese people. In The Psychology of the Chi1986 nese People, Michael H. Bond (ed.), 213–266. New York: Oxford University Press. Boster, James S. 2005 Emotion categories across languages. In Handbook of Categorization in Cognitive Science, Henri Cohen and Claire Lefebvre (eds.), 187–222. Oxford: Elsevier. Bourdieu, Pierre Outline of a Theory of Practice. Cambridge: Cambridge University 1977 Press. 1990 The Logic of Practice. Stanford: Stanford University Press. 1980. Briley, Donnel A., Michael W. Morris, and Itamar Simonson 2000 Reasons as carriers of culture: Dynamic versus dispositional models of cultural influence on decision making. Journal of Consumer Research 27 (2): 157–178. Bucholtz, Mary “Why be normal?” Language and identity practices in a community of 1999 nerd girls. Language in Society 28 (2): 203–223. Carrithers, Michael, Steven Collins, and Steven Lukes (eds.) The Category of the Person: Anthropology, Philosophy, History. Cam1985 bridge: Cambridge University Press. Cohen, Anthony 1994 Self Consciousness: An Alternative Anthropology of Identity. London: Routledge. D’Andrade, Roy Cultural meaning systems. In Culture Theory: Essays on Mind, Self, 1984 and Emotion, Richard A. Shweder and Robert A. LeVine (eds.), 88–119. Cambridge: Cambridge University Press. 1987 A folk model of the mind. In Cultural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 112–148. Chicago: University of Chicago Press. 1995 The Development of Cognitive Anthropology. Cambridge: Cambridge University Press.

Supracultural Models   119 2005

Some methods for studying cultural cognitive structures. In Finding Culture in Talk: A Collection of Methods, Naomi Quinn (ed.), 83–104. New York: Palgrave MacMillan. D’Andrade, Roy, and Claudia Strauss (eds.) 1992 Human Motives and Cultural Models. Cambridge: Cambridge University Press. DiMaggio, Paul 1997 Culture and cognition. Annual Review of Sociology 23: 263–287. Doi, Takeo The Anatomy of Self: The Individual Versus Society. Tokyo: Kodansha. 1985 Duranti, Alessandro 2001 Linguistic anthropology: history, ideas, and issues. In Linguistic Anthropology: A Reader, Alessandro Duranti (ed.), 1–38. Oxford: Blackwell. Elvin, Mark Between the earth and heaven: Conceptions of the self in China. In The 1985 Category of the Person, Michael Carrithers, Steven Collins and Steven Lukes (eds.), 156–189. Cambridge: Cambridge University Press. Foley, William A. 1997 Anthropological Linguistics: An Introduction. Oxford: Blackwell. Gardner, Wendi L., Shira Gabriel, and Angela Y. Lee “I” value freedom, but “we” value relationships: Self-construal priming 1999 mirrors cultural differences in judgment. Psychological Science 10 (4): 321–326. Gee, James Paul 1996 Social Linguistics and Literacies: Ideology in Discourses. 2d edition. London: Taylor & Francis. 1999 An Introduction to Discourse Analysis: Theory and Method. Oxford: Routledge. Geertz, Clifford Thick description: Toward an interpretive theory of culture. In The In1973 terpretation of Cultures: Selected Essays by Clifford Geertz, Clifford Geertz (ed.), 3–30. New York: Basic Books. 1984 ”From the native’s point of view“: On the nature of anthropological understanding. In Culture Theory: Essays on Mind, Self and Emotion, Richard A. Shweder and Robert A. LeVine (eds.), 123–136. Cambridge: Cambridge University Press. Gergen, Kenneth J. 1990 Social understanding and the inscription of self. In Cultural Psychology: Essays on Comparative Human Development, James W. Stigler, Richard A. Shweder and Gilbert S. Herdt (eds.), 569–606. Cambridge: Cambridge University Press. 1991 The Saturated Self: Dilemmas of Identity in Contemporary Life. New York: Basic Books.

120   Jason D. Patent Gergen, Kenneth J., and Keith E. Davis 1985 The Social Construction of the Person. New York: Springer-Verlag. Goffman, Erving Frame Analysis: An Essay on the Organization of Experience. Boston: 1974 Northeastern University Press. Goodenough, Ward 1957 Cultural anthropology and linguistics. In Report of the Seventh Annual Round Table Meeting on Linguistics and Language Study, Paul L. Garvin (ed.), 167–173. Washington, D.C.: Georgetown University. Gries, Peter Hays, and Kaiping Peng 2002 Culture Clash? Apologies East and West. Journal of Contemporary China 11 (30): 173–178. Hall, David L. 1998 Thinking from the Han: Self, Truth and Transcendence in Chinese and Western Culture. Albany: State University of New York Press. Han, Jessica Jungsook, Michelle D. Leichtman, and Qi Wang 1998 Autobiographical memory in Korean, Chinese, and American Children. Developmental Psychology 34 (4): 701–713. Hannerz, Ulf Soulside: Inquiries into Ghetto Culture and Community. New York: Co1969 lumbia University Press. Holland, Dorothy, and Debra Skinner Prestige and intimacy: The cultural models behind Americans’ talk 1987 about gender types. In Cultural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 78–111. Chicago: University of Chicago Press. Holland, Dorothy, and Naomi Quinn (eds.) Cultural Models in Language and Thought. Chicago: University of Chi1987 cago Press. Hsu, Francis L.K. 1981 Americans and Chinese: Passage to Differences. Honolulu: University of Hawaii Press. 1985 The self in cross-cultural perspective. In Culture and Self: Asian and Western Perspectives, Anthony J. Marsella, George A. De Vos, and Francis L.K. Hsu (eds.), 24–55. New York: Tavistock Publications. Hutchins, Edwin Cognition in the Wild. Cambridge, MA and London: MIT Press. 1995 Ji, Li-Jun, Kaiping Peng, and Richard E. Nisbett Culture, control, and perception of relationships in the environment. 2000 Journal of Personality and Social Psychology 78 (5): 943–955. Ji, Li-Jun, Norbert Schwarz, and Richard E. Nisbett 2000 Culture, autobiographical memory, and behavioral frequency reports: Measurement issues in cross-cultural studies. Personality and Social Psychology Bulletin 26 (5): 585–593.

Supracultural Models   121 Ji, Li-Jun, Richard E. Nisbett, and Yanjie Su 2001 Culture, change and prediction. Psychological Science 12 (6): 450–456. Kay, Paul 1987 Linguistic competence and folk theories of language: Two English hedges. In Cultural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 67–77. Chicago: University of Chicago Press. 1996 Intra-speaker relativity. In Rethinking Linguistic Relativity, John J. Gumperz and Stephen C. Levinson (eds.), 97–144. Cambidge: Cambridge University Press. Kiesling, Scott F. 2005 Norms of sociocultural meaning in language: Indexicality, stance, and cultural models. In Intercultural Discourse and Communication, Scott F. Kiesling and Christina Bratt Paulson (eds.), 92–104. Oxford: Blackwell. Kohut, Heinz The Restoration of the Self. New York: International Universities Press. 1977 Koltko-Rivera, Mark E. 2004 The psychology of worldviews. Review of General Psychology 8 (1): 3–58. Kondo, Dorinne K Crafting Selves: Power, Gender, and Discourses of Identity in a Japa1990 nese Workplace. Chicago: University of Chicago Press. Lakoff, George 2002 Moral Politics: How Liberals and Conservatives Think. 2nd ed. Chicago and London: University of Chicago Press. Lakoff, George, and Zoltán Kövecses 1987 The cognitive model of anger inherent in American English. In Cultural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 195–221. Chicago: University of Chicago Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar, Volume I: Theoretical Prerequisites. Stanford: Stanford University Press. Liu, Lydia H. 1995 Translingual Practice: Literature, National Culture, and Translated Modernity, China, 1900–1937. Stanford, CA: Stanford University Press. 1999 Tokens of Exchange: The Problem of Translation in Global Circulations. Durham and London: Duke University Press. Lutz, Catherine Goals, events, and understanding in Ifaluk emotion theory. In Cultural 1987 Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 290–312. Chicago: University of Chicago Press. 1988 Unnatural Emotions: Everyday Sentiments of a Micronesian Atoll and Their Challenge to Western Theory. Chicago: University of Chicago Press.

122   Jason D. Patent Markus, Hazel Rose, and Shinobu Kitayama Culture and the self: Implications for cognition, emotion, and motiva1991 tion. Psychological Review 98 (2): 224–253. Markus, Hazel Rose, Shinobu Kitayama, and Rachel J. Heiman 1996 Culture and ‘basic’ psychological principles. In Social Psychology: Handbook of Basic Principles, E. Tory Higgins and Arie W. Kruglanski (eds.), 857–913. New York: Guilford. Mauss, Marcel (translated by W.D. Halls) 1985/1938 A category of the human mind: the notion of person; the notion of self. In The Category of the Person: Anthropology, Philosophy, History, Michael Carrithers, Steven Collins, Steven Lukes (eds), 1–25. Cambridge: Cambridge University Press. McCall, George J. 1987 The structure, content, and dynamics of self: Continuities in the study of role-identities. In Self and Identity: Psychosocial Perspectives, Krysia Yardley and Terry Honess (eds.), 133–145. New York: John Wiley. Munro, Donald 1969 The Concept of Man in Early China. Stanford: Stanford University Press. Murray, D.W. 1993 What is the Western concept of the self? On forgetting David Hume. Ethos 21 (1): 3–23. Myers, Fred R. 1986 Pintupi Country, Pintupi Self. Washington, D.C.: Smithsonian Institute Press. Mühlhäusler, Peter, and Rom Harré 1990 Pronouns and People: The Linguistic Construction of Social and Personal Identity. Oxford: Basil Blackwell. Nakamura, Hajime 1985/1964 Ways of Thinking of Eastern Peoples. Honolulu: University of Hawaii Press. Nisbett, Richard 2003 The Geography of Thought: How Asians and Westerners Think Differently…and Why. New York: Free Press. Northrop, F.S.C. 1946 The Meeting of East and West. New York: Macmillan. Patent, Jason D. 2001 A Unified Account of Essentially Contested Concepts. In Proceedings of the Twenty-Seventh Annual Meeting of the Berkeley Linguistics Society, 205–213. Berkeley, CA: Berkeley Linguistics Society.. 2003 Are These Truths Self-Evident? Language, Culture and Human Rights in the U.S. and China. Linguistics. UC Berkeley: Dissertation. Peng, Kaiping, Richard E. Nisbett, and Nancy Y.C. Wong 1997 Validity problems comparing values across cultures and possible solutions. Psychological Methods 2 (4): 329–344.

Supracultural Models   123 Quinn, Naomi Convergent evidence for a cultural model of American marriage. In Cul1987 tural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 173–192. Chicago: University of Chicago Press. 2005 Finding Culture in Talk: A Collection of Methods. New York: Palgrave MacMillan. 2005 How to reconstruct schemas people share, from what they say. In Finding Culture in Talk: A Collection of Methods, Naomi Quinn (ed.), 35–81. New York: Palgrave MacMillan. 2005 Introduction. In Finding Culture in Talk: A Collection of Methods, Naomi Quinn (ed.), 1–34. New York: Palgrave MacMillan. Quinn, Naomi, and Dorothy Holland Culture and cognition. In Cultural Models in Language and Thought, 1987 Dorothy Holland and Naomi Quinn (eds.), 3–43. Chicago: University of Chicago Press. Read, Kenneth 1955 Morality and the concept of the person among the Gahuku-Gama. Oceania 25: 234–282. Rosaldo, Michelle Z. 1980 Knowledge and Passion: Ilongot Notions of Self and Social Life. Cambridge: Cambridge University Press. 1984 Toward an anthropology of self and feeling. In Culture Theory: Essays on Mind, Self and Emotion, Richard A. Shweder and Robert A. LeVine (eds.), 137–157. Cambridge: Cambridge University Press. Saville-Troike, Muriel 2003 The Ethnography of Communication: An Introduction. 3rd ed. Oxford: Blackwell. Sharifian, Farzad On cultural conceptualizations. Journal of Cognition and Culture 3 (3): 2003 187–207. Forthcoming Distributed, emergent cultural cognition, conceptualisation and language. In Body, Language and Mind: Sociocultural Situatedness, Roslyn M. Frank, René Dirven and Tom Ziemke (eds.). Berlin: Mouton de Gruyter. Shore, Bradd Culture in Mind: Cognition, Culture, and the Problem of Meaning. New 1996 York: Oxford University Press. Shweder, Richard A. 1984 Anthropology’s romantic rebellion against the enlightenment, or there’s more to thinking than reason and evidence. In Culture Theory: Essays on Mind, Self and Emotion, Richard A. Shweder and Robert A. LeVine (eds.), 27–66. Cambridge: Cambridge University Press. Shweder, Richard A., and Edmund J. Bourne 1984 Does the concept of the person vary cross-culturally? In Culture Theory: Essays on Mind, Self and Emotion, Richard A. Shweder and Robert A. LeVine (eds.), 158–199. Cambridge: Cambridge University Press.

124   Jason D. Patent Shweder, Richard A., and Maria A. Sullivan 1993 Cultural psychology: Who needs it? Annual Review of Psychology 44: 497–523. Shweder, Richard A., Manamohan Mahapatra and Joan G. Miller 1990 Culture and moral development. In Cultural Psychology: Essays on Comparative Human Development, James W. Stigler, Richard A. Shweder and Gilbert S. Herdt (eds.), 130–204. Cambridge: Cambridge University Press. Spiro, Melford E. 1993 Is the Western conception of the self “peculiar” within the context of the world cultures? Ethos 21 (2): 107–153. Strauss, Claudia 1990 Who gets ahead? Cognitive responses to heteroglossia in American political culture. American Ethnologist 17 (2): 312–328. 1992 What makes Tony run? Schemas as motives reconsidered. In Human Motives and Cultural Models, Roy D’Andrade and Claudia Strauss (eds.), 197–224. Cambridge: Cambridge University Press. 1997 Research on cultural discontinuities. In A Cognitive Theory of Cultural Meaning, Claudia Strauss and Naomi Quinn (eds.), 210–251. Cambridge: Cambridge University Press. 2005 Analyzing discourse for cultural complexity. In Finding Culture in Talk: A Collection of Methods, Naomi Quinn (ed.), 203–242. New York: Palgrave MacMillan. Strauss, Claudia, and Naomi Quinn A Cognitive Theory of Cultural Meaning. Cambridge: Cambridge Uni1997 versity Press. Sweetser, Eve E. 1987 The definition of lie: An examination of the folk models underlying a semantic prototype. In Cultural Models in Language and Thought, Dorothy Holland and Naomi Quinn (eds.), 43–66. Chicago: University of Chicago Press. Swidler, Ann 1986 Culture in action: Symbols and strategies. American Sociological Review: 273–286. Tannen, Deborah 1998 The Argument Culture: Moving from Debate to Dialogue. New York: Random House. Taylor, Charles 1989 Sources of the Self: The Making of the Modern Identity. Cambridge, MA: Harvard University Press. Tilly, Charles 1992 How to detect, describe, and explain repertoires of contention. Working Paper No. 150. Center for the Study of Social Change. New School for Social Research.

Supracultural Models   125 Tu, Wei-ming Selfhood and otherness in Confucian thought. In Culture and Self: 1985 Asian and Western Perspectives, Anthony J. Marsella, George A. De Vos and Francis L.K. Hsu (eds.), 231–251. New York: Tavistock Publications. White, Geoffrey M., and John Kirkpatrick (eds.) Person, Self and Experience: Exploring Pacific Ethnopsychologies. 1985 Berkeley: University of California Press. Wierzbicka, Anna Semantics, Culture, and Cognition: Universal Human Concepts in Cul1992 ture-Specific Configurations. New York: Oxford University Press. 1993 A conceptual basis for cultural psychology. Ethos 21: 205–231. 1996 Semantics: Primes and Universals. Oxford: Oxford University Press. 1997 Understanding Cultures Through their Key Words: English, Russian, Polish, German, and Japanese. New York: Oxford University Press. Young, Linda W.L. 1994 Crosstalk and Culture in Sino-American Communication. Cambridge: Cambridge University Press.

Chapter 5 The development of Turkish and Finnish words related to privacy Mai Kuha and M. Ali Bolgün

1. Introduction: Privacy and social cognition Imagine that you are boarding an airplane. You locate your window seat and discover a fellow traveler already occupying the aisle seat next to yours. Consider the social information you have to recall and the elements of the situation that you have to process in order to uphold your standards of privacy in this tight fit. In the actions you take towards the goal of getting into your seat, personal space and freedom to act are an immediate concern. If you are traveling in the U.S., you may refrain from squeezing past your seat-mate, but perform an indirect speech act instead, such as musing to yourself: “Is this row 8? Looks like that’s my seat.” The other person gets up to let you pass. Settling in, you squirm to reach your seat belt without making contact with your fellow traveler’s thigh, but it’s no use: he just sat down on your buckle. Faced with the impossibility of respecting both his autonomy and his personal space, you perform remedial face work: “Excuse me, is that yours or mine?” With reduced personal spaces successfully established, the privacy of information becomes the next concern. You try to behave as if you could not hear your fellow traveler’s cell phone conversation or see the documents displayed on his laptop. Conversation seems to be optional; if the other person offers any encouragement to opening conversational moves about the quality of the airline or weather conditions, it is permissible to then ask: “Are you going home?” This question is not very intrusive, as it allows a great deal of latitude in terms of how much information to offer about the circumstances of the trip. Consider, also, how much depends on factors such as the travelers’ gender and the social distance between them. In the scenario above, the two travelers are strangers. If the other person is your colleague, the boundaries are quite different: moderate eye contact becomes permissible, and some social talk is

128   Mai Kuha and M. Ali Bolgün required. At the far end of the social distance continuum, couples on unusually good terms may even find room for both elbows on the arm rest. The airplane scenario illustrates the general principle that certain areas (such as personal space and information) are relevant to privacy, subject to variation according to social and situational factors, and also subject to cultural variation. The way in which privacy is constructed in different cultures must play an important role in our ability to process complex social information in order to meet the expectations of appropriate behavior in a variety of situations. 1.1. Interpersonal schemata A schema is “a cognitive structure that represents knowledge about a concept or type of stimulus, including its attributes and the relations among those attributes”  (Fiske and Taylor 1991: 98). The concept of a schema as a way to organize information is usually traced back to Bartlett’s (1932) research on memory (see, for example, Fiske and Taylor 1991: 103–105). In the 1980s, other researchers in cognitive psychology continued to work on memory, and also used schema theory to understand problem solving and other phenomena (Stein 1992: 49). When social cognition emerged as a field of study, one of its basic assumptions was that information processing is facilitated by structures such as schemata (Linville and Carlston 1994: 146). It was proposed that people use schemata that organize their understandings and expectations of themselves, other individuals, social roles, and types of events (Fiske and Taylor 1991: 118–120). These schemata about the self and others became useful in clinical psychology, starting with Beck’s (1967) model of depression. Later, dysfunctional schemata were seen as causes for the development of other emotional disorders as well (see, for example, Riso et al. 2007). Schema theory also provided a framework for understanding a number of phenomena of interest in cognitive anthropology; for a review, see Casson (1983). The concept of schema was introduced in cognitive linguistics as well. Fillmore’s related concept of frame gave us insight into semantic issues such as the relationship between the verbs buy, sell, and pay (1977: 58–59) or the boundary of the meaning of bachelor (1977: 68–69). Fillmore’s earlier ideas were further developed as part of Langacker’s (1987) cognitive grammar, which is also concerned with integrating meaning and structure. For a review of the role of schema theory in cognitive linguistics, see Kövecses (2006). In other areas of linguistics, frames have been used to describe how discourse is organized and processed (see, for example, van Dijk 1977).

The development of words for privacy   129

How, then, can we use the concept of schema to describe privacy? The person, role, and event schemata from social cognition give us a starting point. More specifically, a relational schema represents knowledge relevant to “patterns of interpersonal relatedness”, either in a particular social situation or in a particular relationship (Baldwin and Dandeneau 2005: 33). Similarly, interpersonal schema has been defined as “a generic cognitive representation of interpersonal events” (Safran 1990: 89). In clinical psychology, the elements of interpersonal schemata are often left relatively unspecified, although Safran suggests that they include beliefs about self and others (1990: 93). Widdowson (1984) outlines ideas that can help us further integrate systematic linguistic research and the interpersonal, social aspect of schemata; he defines interpersonal schemata as those that “service interpersonal knowledge” (1984: 105) and gives examples of such schemata: the sets of conditions that allow people to determine whether their interlocutors have made a request or what terms of address to use with them. Another use of schema theory that overlaps with our purposes in this chapter is found in Palmer’s (1996: 173–175) view of speaker and listener as part of a participation schema. If we view the concept of privacy as an interpersonal schema (see 2.2), it seems possible to retain the broad sweep and overt emphasis on the social in clinical psychology and combine them with the greater attention to detail characteristic to cognitive linguistics. In clinical psychology, the concept of interpersonal schema often seems to be relevant to one specific relationship, but privacy is not like this. Rather, elements of the privacy schema apply to people in a range of social roles and in a range of situations. Therefore, our discussion of privacy will be an opportunity to illustrate how people can organize procedural knowledge in schemata to model rights and responsibilities between social actors more generally. The interpersonal schema we will propose for privacy is somewhat similar to frames such as Fillmore’s well-known commercial event frame, but emphasizes social information even more. 1.2. Privacy Key elements of definitions of the concept of privacy in previous research can be usefully represented in a diagram of privacy as an interpersonal schema (see Figure 1). Privacy has been described as “a dynamic, interpersonal boundary control process regulating access to the self” (Altman 1976). We can further specify areas of the self relevant to privacy: personal space, personal property, personal information, and the other items listed next to Person 1. Privacy, then, is about restricting others’ access to these areas; the actors whose access is

130   Mai Kuha and M. Ali Bolgün restricted may be other individuals (Person 2) or institutions, such as the government. Another key point is that the boundary is controlled by the individual (Nock 1998; Westin 2003); that is, Person 1 has, at least to some extent, a right to regulate this interface according to social norm, or even according to the law, as indicated in Figure 1. One aspect of privacy is the autonomy of Person 1: the right to choose one’s actions (Margulis 2003: 245). However, as shown, illegal or unacceptable actions are not protected by privacy (Westin 2003: 433). The rectangle in the center of the diagram, representing the privacy boundary, specifies some variables which may affect the level of privacy that is considered appropriate: the constraints of privacy may be reduced in the case of children, in-group members, and in emergency situations; on the other hand, greater privacy may be required in mixed-sex interaction in some situations.

Figure 1. Privacy as an interpersonal schema.

The development of words for privacy   131

Note that the existence of a noun meaning privacy allows social actors to conceptualize the entire complex system in Figure  1 as an object. We will return to this point. Researchers seem to agree on the universality of privacy, whether they consider it a universal process (Altman 1977: 66), need (Newell 1994: 66), or desire (Moore 1984: 276; Kasper 2005: 70). Although privacy research has suffered from cultural bias (Kasper 2005: 74), taking an excessively North American and European perspective (Margulis 2003: 256), cultural variation in the concept of privacy clearly exists, even if the number of studies documenting it is relatively small. As Moore (1984: 49) points out, any human society has “disagreeable pressures or obligations”, but “the actual and perceived weight of these obligations, and the real and socially defined opportunities to do something about them, all vary enormously”. Both variation and universality are illustrated in Altman’s (1977) ethnographic survey of cultures around the world. Altman identifies societies with “minimal privacy”, such as the Mehinacu of central Brazil (72–74), and societies with “maximum privacy”, such as the Tuareg (76–77), and concludes that culturally specific mechanisms for regulating privacy emerge in societies in which close living arrangements make privacy less attainable. In a study of perceptions of privacy in Ireland, Senegal, and the U.S., Newell (1998) found that the largest proportion of respondents in all three countries ranked being undisturbed as the most important aspect of privacy and identified being relaxed as an aftereffect of privacy. A difference was that Senegalese participants reported being able to get privacy less frequently than the other participants did. It has been noted that privacy is not a concept indigenous to Japan (Iwata 1988: 199) or China (Chan 2000: 2). However, Iwata’s participants’ ratings suggested that four indigenous Japanese concepts, ‘private life’, ‘freedom’, ‘solitude’, and ‘secrecy’ are all “closely related with the foreign concept of privacy” (1988: 204). Similarly, the Chinese value ‘solitude’, ‘reserve’, and ‘intimacy’ (Chan 2000: 2). Two aspects of the Chinese view of privacy contrast sharply with the Western view: the desirability of closeness with the family group even while pursuing ‘solitude’ and ‘reserve’ (3), and the traditional view that “[t]he dominant in a dyad relationship can acquire the information he/she wants from the other party because it is considered justifiable and natural to do so”, leading to asymmetry in rights to privacy (5). Frames are “open to cross-cultural variation” because they often construct concepts, rather than simply reflecting some objective reality in the natural environment; a given frame may or may not exist in a given culture (Kövecses 2006: 65). It seems likely that elements in the schema in Figure 1 vary from

132   Mai Kuha and M. Ali Bolgün one culture to another: which areas (property, information, personal space, etc.) are considered to be restricted by privacy, whether privacy legislation exists, whether privacy is viewed primarily from the standpoint of Person 1 (privacy as a right) or Person 2 (privacy as a responsibility), and to what extent variables such as shared group membership influence the degree of privacy granted. In particular, less privacy may be called for among in-group members in group-oriented cultures. In Greece, for example, “distance among closely related people and family members is assumed to be minimal” (Sifianou and Antonopoulou 2005: 267). Cultures may also vary in terms of which areas are subject to greatest privacy demands; for example, violations of personal space may be considered a more serious offense than violations of the privacy of information. No previous linguistic research on words related to privacy could be found, although Hollander (2001) outlines the etymology of the English word privacy, and Schauer (2001) points out in passing that “the law’s use of the language of privacy has become an influential source of popular ideas about privacy and thus about the concept of privacy itself” (52). 2. Procedures Etymological information about Turkish words related to privacy was obtained from five bilingual dictionaries and three monolingual dictionaries. These sources include online dictionaries and were published between 1890 and 2007. Two of the online monolingual dictionaries focus on meanings as they are perceived by native speakers, which may or may not be compatible with what is given in print dictionaries. These were utilized because they reflect what the speakers think the words mean rather than what the dictionaries claim they mean. The main Finnish-English bilingual dictionaries were also used, as well as two kinds of monolingual dictionaries: general dictionaries of modern Finnish and etymological dictionaries of loanwords used in formal registers. These reference works were published between 1973 and 2002; they provide fairly comprehensive information about the prescribed meanings of privacy-related words in literary Finnish, as there may be as few as five general monolingual dictionaries in Finnish (although there are several versions and editions, of course). The Turkish data come from the currently 2-million-word METU Turkish corpus (Say et al. 2002). It has 2000-word written Turkish samples collected post-1990 from various sources, balanced across genres (see Table 1), although

The development of words for privacy   133

“no statistical representativeness scheme was applied” (Atalay, Oflazer, and Say 2003). Table 1. Genres in the METU Turkish Corpus Genre News

Words

Samples

848,378

419

Novel

251,043

Story

224,874

Column

Sources

Percent (%)

issues of 3 dailies

42

125

63 books

13

112

56 books

11

167,013

83

issues of 3 dailies

8

Article

155,561

84

71 journal issues

8

Essay

138,612

69

35 books

7

49 26 books 19 3 books and 13 journal issues

5 2

Research Monograph Travel

98,447 35,932

Interview

14,054

7

1 book and 3 journal issues

1

Other

64,253

32

17 books

3

Total

1,998,167

999

100

Turkish data on özel and mahrem (and other related words such as kişisel ‘personal’, and hususi ‘special’) were found using MonoConc Pro concordancer, and were analyzed manually and grouped according to their meanings. Data about the actual use of privacy-related words was obtained from the Finnish Language Text Collection, the written Finnish portion of the Kielipankki corpus.1 At the time of our analysis, this portion of the corpus consisted of 94 sub-corpora totaling nearly 180 million words from newspapers (primarily Helsingin Sanomat, Turun Sanomat, and Aamulehti) published between 1990 and 2000; approximately 5 % of the text was from various sources other than newspapers. The Kielipankki materials were compiled by the Research Institute for the Languages in Finland, the Department of General Linguistics of the University of Helsinki, the Foreign Languages Department of the University of Joensuu. We identified relevant word frequencies, concordances, and collocations using WWW-Lemmie 2.0, a web-based corpus analysis tool developed by CSC – Scientific Computing Ltd.2 This tool allows users to identify all occurrences 1 http://www.csc.fi/kielipankki/ 2 http://www.csc.fi/kielipankki/ohjelmistot/lemmie.phtml

134   Mai Kuha and M. Ali Bolgün of the base form of lexical items of interest, without having to search for each inflected form separately. 3. Results 3.1. Turkish dictionary entries The earliest word for private (or for one of its senses) seems to have come from Arabic. For example, A Turkish and English Lexicon (1890) defines mahrem still in use–though rare–in contemporary Turkish, as: “a. 1. Prohibited; especially, … within the forbidden degrees of relationship for marriage; such, though of different sexes, may meet on intimate terms. 2) Intimate. 3) A forbidden place for certain acts; a sacred place.”

The word often used to translate private in Turkish is özel. It is important to note that in A Turkish and English Lexicon (1890) özel is not found. Its root öz is found as özlük,3 however. It is defined as “1) [s]elf; identity. 2) [e]gotism; selfishness. 3) [t]he essential quality, the pith, marrow, cream, or substance of a thing,” among others. It is the first and, to a certain extent, the second sense of özlük that özel is based on, and nowadays means ‘personal’ with some instances meaning ‘private’. There is another word for ‘personal’ in Turkish: kişisel, with no instances that have the meaning ‘privacy’. An English-Turkish Dictionary (1952) defines ‘privacy’ as: mahremlik, (for the English meaning, see the indented quote above), yalnızlık (‘loneliness’), and inziva (‘seclusion’). And ‘a desire for privacy’ is defined as “the desire to live in one’s own world / universe and as the desire to stay away from others’ looks / attention”.4 A more recent dictionary, The Oxford Turkish-English Dictionary (1992),5 also defines mahrem as: “Confidential; secret; intimate; within the relationships forbidden for marriage (thus having access to the harem).” The same dictionary defines privacy exactly as it was defined in An EnglishTurkish Dictionary (1952) 30 years earlier. The only addition to the definition is the word halvet (‘solitude’). 3 The suffix -lik (and its allomorphs -lık, -luk, and -lük) is a derivational suffix that derives a noun from an adjective. 4 The exact Turkish definitions for these two are: kendi aleminde yaşama arzusu; yabancı nazarlardan uzak kalma arzusu. 5 By the same lexicographers as that of An English-Turkish Dictionary (1952).

The development of words for privacy   135

The Oxford Turkish-English Dictionary (1992) defines özel as “[p]ersonal; private; distinctive; particular; special; specific”. (Seslisozluk (2007), in addition, lists ‘exclusive,’ ‘individual,’ ‘intimate,’ ‘proprietary,’ ‘sole,’ ‘self’). Aside from this, there are many noun phrases that are made up with özel, such as özel ad (‘proper noun’), özel düşünce (‘personal opinion’), özel haber (‘exclusive story, scoop’), and özel okul (‘private school’), etc. This proliferation of phrases with özel may perhaps be due to the fact that the word that used to be used in these phrases was the Arabic word hususi (‘special, specific, personal’). Therefore, while the word privacy is one of the words listed in the definition of özel, it is the meaning of hususi that özel mainly took over, not of ‘privacy’. Langenscheidt Standard English Dictionary: English-Turkish, TurkishEnglish (1985) defines ‘privacy’ as özellik (see footnotes 2 and 3, and the definition of özel above); gizlilik (‘secrecy’); mahremiyet (the noun form of mahrem); kişisel dokunulmazlık (‘personal untouchable-ness’). It should be noted that while it is possible to derive a noun from another noun or from an adjective using the suffix -lIk (-lık, -lik, -luk, -lük), özellik (özel+lik) has never been (and perhaps will never be) used as a noun for ‘privacy’ the reason being that özellik is a commonly used word which means ‘peculiarity’ and ‘characteristic’. As we will see in our discussion of the Turkish corpus, there was no instance of özellik in the sense of ‘privacy’. It seems that there is no noun form of the word privacy; it exhibits itself, if any, only as an adjective in Turkish. The word özellik (when used to define privacy) has an overwhelmingly more common meaning of ‘characteristic’ and as such, when used, özellik is not thought of as meaning ‘privacy’ but ‘characteristic’. The Arabic mahremiyet, which refers to privacy in special contexts, is the only noun used for privacy. Özel is defined as hususi (the Arabic word that is used somewhat synonymously with özel) and hususi, in turn, is defined as “1) particular, special, distinctive, characteristic; 2) personal, private” and also as “3) reserved (seat); 4) privately owned car” (in colloquial language). In Türkçe Sözlük (1995), a monolingual Turkish dictionary, özel is defined in similar lines as #1 and #2 senses of hususi (see previous paragraph). The third and fourth senses are given as “that which does not belong to the public or state/government (but) that belongs to the individual, e.g. özel arabalar çoğaldı ‘private cars have increased’” and “exclusive to the individual; that which does not concern everybody; personal; e.g. özel mektuplar açılmaz ‘private letters are not (supposed to be) opened’”. Hususi is defined as “see özel” and as the adverb form of özel.

136   Mai Kuha and M. Ali Bolgün The use of gizlilik ‘secrecy’ and inziva ‘seclusion’ in the definition of privacy is motivated by the definition of the word privacy in monolingual English dictionaries. For example, The American Heritage Dictionary uses simply these two words (secrecy) and (seclusion) to define privacy. 3.2. The Turkish corpus The total number of instances of nouns with the approximate meaning ‘privacy’ and adjectives with the approximate meaning ‘private’ in the METU Turkish corpus is shown in Table 2. Starting with the Turkish language revolution in 1928, most borrowings were consolidated into fewer, newly coined words (Lewis 2002). The two adjectives that have the highest number of hits (özel and kişisel) are Turkish and the rest are borrowings. It seems that kişisel ‘personal’ has taken over şahsi; all 56 uses of kişisel ‘personal’ were deemed to mean ‘personal’ or ‘individual’. Özel, on the other hand, seems to have taken over the meanings of the other borrowings; namely, mahrem, hususi, müstakil, and has. There is more to say on özel, but first it should be stated that the data analysis shows that all six instances of hususi are used to mean ‘special’ or ‘specialized for’. Four of the ten instances of müstakil are used with ‘housing’ of some sort, as in private villa, to mean it is a separate unit. This is to separate it from high-rise apartment style housing. Five instances are used when talking about historical events, and to mean ‘independent’. One is used with bölüm ‘section’ and to mean ‘special’. In nine of the total 20 instances, has means ‘peculiar to’, ‘representative of’. In two instances, it means ‘only’. In two instances, it means ‘complete’ or ‘total’. In six instances, it is used as part of a company name (as a proper noun), and in one instance, it has a historical use, as a noun to mean ‘tax/privilege’. All four instances of şahsi mean ‘personal’. Table 2. Frequency of Turkish words used in dictionaries to define ‘privacy’ and ‘private’ base form of noun özellik6 mahremiyet özel ‘personal, special, private, peculiar’

number of instances 0 7 2997

6 In the sense of ‘the state of being private’ (not of ‘characteristic’). 7 Of the 303 instances that were found using MonoConc Pro concordancer, 4 hits were duplicates and as such were excluded.

The development of words for privacy   137 base form of noun

number of instances

kişisel ‘personal’

56

mahrem ‘private’

4

şahsi ‘personal’

4

hususi ‘special’

6

müstakil ‘private, separate, independent’

10

has ‘peculiar to’

20

adverbial phrase özel olarak ‘privately’

number of instances 5

3.2.1. Özel Table 3. Breakdown of the meanings of özel meaning ‘special’, ‘peculiar’, ‘specific’, ‘unique’, ‘privileged’ (including the 5 instances of özel in adverbial phrases) ‘non-government’, ‘public’ ‘private’ ‘personal’, or ‘personalized’ other (as a proper noun, as in last names, and company names)

147 (49.1 %) 61 (20.7 %) 48 (16.0 %) 38 (12.7 %) 5 (1.6 %)

As Table 3 shows, the majority (49.1 %) of the instances of özel found in the corpus center around the meanings of ‘special’, ‘peculiar’, ‘specific’, ‘unique’, and ‘privileged’. For example: (1)

Eğer  sıradan bir If    ordinary a

vatandaş-sa-nız,   demokrasi-yle citizen-IF-2PL   democracy-WITH

ilgi-niz ol-ma-z. Eğer concern-2PL- NEG-AOR If o zaman en most then

özel bir vatandaş-sa-nız, special a citizen-IF-2PL

tehlikeli    iş-ler-e   de giriş-se-niz dangerous  job-PL-DAT  also venture-IF-2PL

demokrasi   sizin democracy   your

için for

çalış-ır. work-AOR

‘If you are an ordinary citizen, you would have nothing to do with democracy. If you are a special citizen, then democracy works for you even if you undertake dangerous businesses.’

138   Mai Kuha and M. Ali Bolgün It is clear that by özel bir vatandaş, what is meant is ‘a privileged citizen’ and not ‘a private citizen’. This is made clear by the adjective sıradan ‘ordinary’ that is used in the first sentence. The second sentence contrasts such a citizen (an ordinary one) with a non-ordinary one. In 20.7 % of the instances özel indicates that the noun phrase it modifies is a ‘non-government’ entity. This could be a private school, private TV station, or private sector, as the example below shows: (2)

Hangisi   zor,    kamu    görev-i   mi,  özel  sektör  mü? Which one  difficult  government  job-CM  Q   private sector Q ‘Which one is more difficult: government job, (or) private sector?’

In 10.7 % of the instances özel simply indicates that the noun phrase it modifies belongs to the person; it means personal. There is nothing necessarily private about it. For example: (3)

Bu   da   Cumhurbaşkanı-nın  özel    merak-ı-nı this   also  president-GEN   personal  curiosity-POSS-ACC gider-mek   için   yap-tığ-ı quench-INF  for   do-PTCL-POSS olarak as

gezinti-ler-in tour-PL-GEN

inceleme survey

nitelendir-il-me-si-dir. characterize-PASS-NML-POSS-COP

‘And this is characterizing the trips that the president goes on (just) to satisfy his personal curiosity as (official) survey.’ In 1.6 % of the instances, özel is simply a last name, a company name, or part of a book title. Of the remaining 48 instances (or 16.0 %) of özel (to mean ‘private’), 27 (or 56.2 %) collocate with yaşam, yaşantı, or hayat all of which mean ‘life’. The collocates of these instances of özel are given in Table 4. Table 4. Collocations used with the ‘private’ sense of özel yaşam, yaşantı, and hayat ‘life’ alan ‘sphere’, ‘space’ görüşme ‘meeting’ konuşma ‘communication’ dünya ‘world’ ilişki ‘relationship’ konu ‘subject; matter’ toplumsal bütün ‘social sum’

27 11 3 2 1 1 1 1

The development of words for privacy   139

3.2.2. Mahrem and mahremiyet All 11 instances of mahrem and mahremiyet mean ‘private’ or ‘privacy’ in the most intimate way. It seems that although özel has taken over a number of borrowings (such as hususi ‘special’) to a large extent, there are some areas where özel does not quite express the intended meaning. Even when özel is used with yaşam ‘life’, it often means ‘non-job related life’ and not ‘private life’ in its strictest sense to mean ‘no access’ to that characterized as özel. For example: (4)

Fakat Cumhurbaşkanı  da benim gibi like however president   also me

bir a

insan. person

Konuşmayı   sev-en   bir  insan.  Duy-duğ-u-m-a talk-ACC    like-PTCL  a   person  hear-PTCL-POSS-1SG-DAT göre   özel   yaşam-ı-nda   çok   hoşsohbet-miş. according   private   life-POSS-LOC  very  sociable-PAST ‘However, the president, is a human being too; just like me. A human being that likes to talk. As far as I have heard, he is very sociable in his private/personal life.’ Similarly, the following contrasts ‘private/personal life’ with ‘job life’: (5)

yaşam-ı-yla Şekerim, insan özel honey person private life-POSS-WITH

iş-i-ni job-POSS-ACC

Ne ev-e ayır-malı. separate-OBLIG neither home-DAT

taşı-malı-sın, carry-OBLIG-2SG

ne nor

iş-in-e. job-POSS-DAT

de also

ev-in-i home-POSS-ACC

‘Honey, one should separate one’s personal/private life from one’s job. You should not bring your job home; neither should you bring your home to your job.’ Mahrem, on the other hand, is stricter in that access or penetration to that characterized by mahrem constitutes a bigger violation of the norm. This is not surprising since mahrem shares the same root with the words that mean ‘forbidden,’ ‘prohibited,’ ‘unlawful,’ ‘sacred,’ and ‘sin,’ among others (Al-Mawrid

140   Mai Kuha and M. Ali Bolgün 1996; Arabic-English Dictionary 1994). Compare the two examples above with the following two examples: (6)

Vücud-u-nu şehvet düşkünlüğ-ü-yle öylesine body-POSS-ACC lust addiction-POSS-WITH such kötü-ye kullan-mış-tı    ki, bad-DAT use-PPTCL-PAST   CONJ

mahrem  yer-ler-i private   part-PL-POSS

başka other

doğa-nın nature-GEN

kadın-lar-ın-ki gibi woman-PL-GEN-REL as

belirle-diğ-i yerde    değildi ve determine-PTCL-POSS place-LOC  not-PAST and

sanki as if

yüz-ü-ne vur-muş-tu face-POSS-DAT reflect-PPTCL-PAST ‘She had used her body with lust so badly that her private areas were not where the nature intended for them to be, like in other women, and it is as if (her lust) was reflected in her face.’ In the example above, what is meant by private areas are clearly sexual organs. In such a context, mahrem is picked over özel. Similarly, in (7), the context is lovemaking and the adjective used in this context is mahrem. (7)

Para-sı-nı    ver-ip  sokak-lar-dan  sahip-siz money-POSS-ACC  give   street-PL-ABL  owner-less

beden-ler body-PL

topla-mak,   onlar-la    bu collect-INF  they-WITH  this

mağara cave

boş ev-in empty house-GEN

kovuk-lar-ı-na   benze-yen    sessiz  oda-lar-ı-nda hole-PL-CM-DAT  resemble-PTCL  quiet  room-PL-POSS-LOC mahrem oyun-lar private game-PL

oyna-mak play-INF

bir a

mucize gibi miracle as

gel-iyor-du    ban-a;  orospu-lar-la   yaşa-dığ-ım come-PROG-PAST   I-DAT   prostitute-PL-WITH  live-PTCL-1SG

The development of words for privacy   141

her every

parçala-n-mış    sevişme-den break-PASS-PTCL  lovemaking-ABL

bir  huzur a    peace

ve and

ferahlık contentment

sonra büyük after big

duy-uyor-du-m. feel-PROG-PAST-1SG

‘Paying for and collecting ownerless bodies from the streets and playing with them private games in this house’s rooms that looked like hollows of caves seemed like a miracle to me; I was feeling a sense of peace and contentment after every shattered lovemaking that I had with prostitutes.’ The high number of instances of özel being used with ‘life’ could be due to the lack of a Turkish noun that means privacy. Mahremiyet, the Arabic borrowing meaning ‘privacy’, does not always satisfy the current need since it refers to a specialized form of privacy. It appears that the recent özel yaşam and the older mahremiyet are both needed to compensate for privacy. For example, özel yaşam ‘private life’ is too general in the context below and cannot substitute for mahremiyet ‘privacy’. Consider: (8)

Verimliliğ-i   önemse-yen kimi productivity-ACC  value-PTCL some

şirket-ler-in company-PL-GEN

ofisleri adeta kusursuz-du. Ama office-PL-POSS almost flawless-PAST but çalışan-lar-ın mahremiyet-i-ne worker-PL-GEN privacy-POSS-DAT bir a

tasarım design

hiçbir-i-nde none-POSS-LOC

önem     ver-en importance  give-PTCL

yoktu. nonexistent-PAST

‘The offices of some companies that value productivity were almost flawless. However, in none of them was there a design that values the workers’ privacy.’ In the above example, the focus is on the immediate privacy of the workers and not their private life in general. As such, özel yaşam ‘private life’ would not quite designate the intended meaning, which is most likely to be ‘hidden from the undesired gaze’.

142   Mai Kuha and M. Ali Bolgün Perhaps privacy should be considered as having a number of levels rather than being taken to be a single-level concept. In Turkish, it is clear that it has at least two levels: one that is described with özel (the özel-level) and one that is described with mahrem (the mahrem-level). The özel-level indicates a higher level of penetrability. In this level, the use of özel is mainly to separate entities or processes from ‘government’ entities or processes because the use of özel in some cases indicates ‘nongovernment’ entity or process, and this entails a deviation from the default. For example, okul ‘school’ would not be labeled özel okul ‘private (nongovernment) school’ if it were a government-operated school. Similarly, özel can be used to mean ‘non-public’, and closer in meaning to ‘personal’. On the mahrem-level, on the other hand, penetrability becomes more difficult and undesirable. This is an area where people do not normally want to be seen, or heard. In Lugat-it Turk (2005), one of the contributors to the website defines it as: “cinsel uzuvlar söz konusu olduğunda ‘saklanması gereken gizli yer’ anlamında uzuvlar için kullanılır. yani mahrem yeridir onlar.” [when it comes to sexual organs, it is used in reference to the [sexual] organs to mean “a secret area that needs to be hidden.” they are the ‘mahrem’ areas.]

Perhaps because mahrem is about things people are not supposed to talk about, we do not see as many instances of it as we do with özel. In both (9) and (10), there is a reference to ‘non-public’; however, in (11), as well as (6) and (7) above, there is a reference to personal life and seclusion in its absolute sense, often having a connotation of sexuality. (9)

Özel   konuşma-lar-a     kulak private  conversation-PL-DAT   ear

kabart-an-lar, kapı raise-PTCL-PL door

dinle-yen-ler varsa,   iş-i-miz   iş. listen-PTCL-PL exist-IF  job-POSS-1PL   job ‘If there are people eavesdropping, and listening to private conversations (through) the door, (then) we are in trouble.’ (10) Özel private ve and

hayatını, karısını,   kızını life-POSS-ACC wife-POSS-ACC  daughter-POSS-ACC oğlunu hiçbir zaman ikinci plana son-POSS-ACC no time second plan-DAT

The development of words for privacy   143

atmamış,    ancak throw-NEG-PPTCL   however

işi job-POSS

sorun-lar-ı onlarla problem-PL-ACC they-WITH

paylaş-ma-mış-tı. share-NEG-PPTCL-PAST

ile with

ilgili related

‘He had not put his private life, his wife and his son on the back burner, but he had not shared with them the problems related to his job (either).’ (11) İtaliğ-i   çağrıştır-an   bir  yatıklık-la,  yuvarlak,  süslü  ama italic-ACC  remind-PTCL  a   slant-WITH  round  decorative  but rahat  oku-n-an harf-ler-le   yazılmış  mektub-u, easily  read-PASS-PTCL letter-PL-WITH   written   letter-ACC dudak-lar-ı-nı    ısır-arak,  mahremiyet-e gir-me-nin lip-PL-POSS-ACC  bite-ADV  privacy-DAT   enter-NML-GEN o    günahsı ama that   sinful but

heyecanlı exciting

his-si-yle feeling-POSS-WITH

okuyuver-di. read-PAST

ve and

tatlı sweet

‘[She] read the letter, which was written with round, decorative but comfortably read letters that had a slant that resembled italic, biting her lips, and with a sin-like but exciting and sweet feeling of entering [violating] privacy.’ 3.3. Finnish dictionary entries As we will see in our corpus analysis below, the primary lexical item for privacy in present-day Finnish is the noun yksityisyys. However, the 1978 Dictionary of Modern Finnish (Nykysuomen sanakirja) has no entry for this word. How, then, was privacy translated from other languages? The 1973 edition of the main bilingual dictionary, the English-Finnish General Dictionary (Englanti-Suomi suur-sanakirja), defines the English word privacy using yksinäisyys (‘loneliness’) and eristyneisyys (‘isolation’), which have negative con-

144   Mai Kuha and M. Ali Bolgün notations. A secondary definition for privacy is vaiteliaisuus, salamyhkäisyys (‘not speaking, being secretive’). The entry for privacy also includes an element with a more neutral meaning: yksityiselämä (‘private life’), formed from the bound adjectival yksityis- and elämä, ‘life’. The expression oma rauha (oma ‘own’ and rauha ‘peace’, lit. ‘one’s own peace’) is also used; it has the highly positive meaning of being autonomous, unimpeded, free of disturbance. Some translations of common English expressions in this entry underscore tranquility, quiet, and the absence of disturbance as desirable aspects of privacy. (12) kotirauhan häirintä home-peace-GEN.SG disturbance ‘invasion of privacy’ (lit. ‘disturbance of domestic tranquility’) The English expression in privacy is described using words meaning ‘in peace, in silence’ (rauhassa, hiljaisuudessa) and ‘by oneself’ (itsekseen). The centrality of peace and quiet in the concept of ‘privacy’ is also evident in the entries for the adjective private in this same dictionary: (13) hautaus toimitettiin hiljaisuudessa burial carry out-PASS.PAST silence-INE.SG ‘the funeral was private’ (lit. ‘the funeral was held in silence’) The importance of the concept of privacy is evident in the inclusion of more than two pages of compounds in the Dictionary of Modern Finnish (Nykysuomen sanakirja) formed with the bound adjectival form yksityis-, including such commonly used lexical items as yksityistalo (‘privately owned house’). The adjectival form also occurs as the free morpheme yksityinen, ‘private’. Its meaning is linked to ‘personal’, ‘individual’, and contrasted with ‘common/general’, ‘public’, ‘official’, and ‘shared’. The meaning of ‘individual’ seems to be central, as the morphological similarity between yksityinen and the forms yksi (‘one’) and yksin (‘alone’) suggests. Yksityinen is also the primary element in the definition of the English word ‘private’ in the 1973 edition of the EnglishFinnish General Dictionary (Englanti-Suomi suur-sanakirja). Even in the absence of a noun meaning ‘privacy’ in these early dictionary entries, the concept of privacy clearly exists; as the expressions above show, the components of the interpersonal schema of privacy that are articulated more or less explicitly include the idea of restricting others’ control over one’s property, freedom to act, and personal space. However, some of these elements emerge as more central, while others are backgrounded. The elements of the schema also vary in terms of the positive or negative connotations of expressions used

The development of words for privacy   145

to articulate them. Notably, where the idea of protecting personal information is mentioned at all, it is in the negative sense of ‘being secretive’. Since legitimate activities are private, in contrast to questionable activities, which are secret (Westin 2003), these negative connotations suggest that the right to restrict access to personal information may not have been part of the privacy schema at the time. Less than two decades later, the status of nouns meaning ‘privacy’ in Finnish seems quite different. The 1990 Basic Dictionary of the Finnish Language (Suomen kielen perussanakirja) now features an entry for the noun yksityisyys, defined as ‘private life’ (yksityiselämä). In addition, intiimiys is given as a synonym for yksityisyys. Privacy is also described more positively than earlier. Sample expressions such as ‘right to privacy’ (oikeus yksityisyyteen) and ‘violation of privacy’ (yksityisyyden loukkaus) present privacy as a basic right, and the absence of privacy as a problem. A similar shift can be seen in the 1990 edition of the main bilingual dictionary, the English-Finnish General Dictionary (Englanti-Suomi suur-sanakirja). Here, the order of elements in the entry for privacy has changed, so that a positive meaning appears first: ‘one’s own peace’ (oma rauha). In second place is the new element: yksityisyys. ‘Private life’ is still part of the definition, and ‘loneliness’ and ‘isolation’ are now the last elements in the entry. The English-Finnish General Dictionary explains the English expression ‘in privacy’ as follows: (14) [kaikessa] rauhassa, omissa oloissaan all-INE.SG peace-INE.SG own-INE.PL being-INE.PL ‘in (complete) peace’, ‘undisturbed in one’s thoughts and actions’ Interestingly, the first part of this definition, kaikessa rauhassa, can mean ‘unhurried’ or ‘unimpeded’; this expression can easily describe activities carried out in a public place. Taken together, these entries in the 1990 editions of the main dictionaries suggest that tranquility and being unimpeded are still central elements in the Finnish concept of privacy. The 1990 Basic Dictionary of the Finnish Language features a novelty: an entry for intimiteettisuoja, defined as protection (including legal protection) regarding private life. This noun is a compound formed from the loanword intimiteetti and suoja, ‘shelter, protection’. The 2002 Dictionary of Foreign Words defines intimiteetti as ‘the property of being inside, closeness’ and ‘of a personal nature’; yksityisyys is given as a synonym. In summary, these dictionary entries suggest that one or two nouns that overlap considerably with English privacy have emerged quickly, within 30 years.

146   Mai Kuha and M. Ali Bolgün Earlier, various expressions that foreground tranquility, quiet, and freedom from social obligation could be used to refer to aspects of privacy, but there was no noun corresponding to privacy. 3.4. The Finnish corpus 3.4.1. Frequency The number of instances of nouns and adjectives in the Kielipankki corpus that overlap in meaning with privacy and private in English is shown in Table 5. As we can see, the distribution of these words in the corpus is consistent with the picture that emerged from the dictionary entries: lexical items that are native to Finnish, rather than borrowed, include a well-established adjective, yksityinen, and the related noun, yksityisyys, which has become far more frequent than the alternatives, including the native oma rauha (at least in the news discourse that this corpus represents). Table 5. Frequency of Finnish nouns and adjectives related to privacy and private. base form of noun yksityisyys oma rauha intimiteettisuoja intiimiys intiimisyys base form of adjective yksityinen / yksityisprivaatti

number of instances 1,006 259 108 49 1 number of instances 21,685 42

relative frequency 0.0000056 0.0000014 0.0000006 0.0000003 0.0000000 relative frequency 0.0001205 0.0000002

3.4.2. Paraphrases and some other sentence contexts Paraphrases for the three main nouns related to privacy illustrate their meanings and connotations by means of synonyms and characteristics in sentence context. Consistent with the frequencies we saw in Table 5, yksityisyys emerges as the most unmarked way to reify the concept of privacy, while oma rauha and intimiteettisuoja are reserved for more specific meanings. A search for yksityisyys on... (‘privacy is...’) yielded 25 hits. Six of these paraphrase yksityisyys with a noun phrase, and the rest articulate specific attributes of the concept of privacy by means of adjectives or other material. As we see in (15) – (17), privacy is characterized as an object in these paraphrases.

The development of words for privacy   147

More specifically, it is presented as a commodity that can be owned and has great value: yksityisyys is explicitly equated with ‘wealth’ and ‘property’. (15) Yksityisyys   on    tarkoin privacy-NOM.SG  be-3SG  carefully

vartioitua guarded-PART.SG

vaurautta. wealth/asset-PART.SG ‘Privacy is a carefully guarded asset.’ In the next example, the writer expresses a wish that two public figures avoid excessive media attention. Here, other elements (‘valuable’, ‘look after’) also show that privacy is a type of property that is desirable and highly valued. (16) ... suhteen    yksityisyys on heidän    ainut relationship-GEN.SG  privacy-NOM.SG be-3SG 3PL.GEN  only arvokas   yhteinen valuable-NOM.SG  shared-NOM.SG

omaisuutensa, property-POSS.3PL

jota pitäisi ymmärtää   huolella which-PART.SG should-3SG-IF understand-INF   care-ADE.SG vaalia. look after-INF ‘... the privacy of the relationship is their only valuable shared property, which one should have the sense to look after with care.’ Nominalization makes it possible to refer to privacy as a commodity (‘target of commerce’) that has a price and can be bought and sold for money, much in the same way as a physical object could. This is particularly clear in (17). (17) Kansalaisen yksityisyys    on    kaupankäynnin citizen-GEN.SG privacy-NOM.SG  be-3SG  commerce-GEN.SG kohde, yksityisyydellä   on target-NOM.SG privacy-ADE.SG  have

148   Mai Kuha and M. Ali Bolgün eräänlainen   tekijänoikeus, josta sen certain-NOM.SG   copyright-NOM.SG which-ELA.SG it-GEN haltija voi owner-NOM.SG can-3SG

sopivaa suitable-PART.SG

korvausta vasten luopua. compensation-PART.SG against give up-INF ‘A citizen’s privacy is a target of commerce; privacy has a kind of copyright, which its holder can give up for a suitable compensation.’ Although nominalization opens up such possibilities by making the privacyas-object metaphor easier to access, the adjectival forms yksityinen and yksityis- provided a way to talk about private property, even before the noun yksityisyys became established (as we will see in 3.4.3). In contrast, using the adjective to express the idea of ‘private thoughts’ might have been awkward or even semantically anomalous. The development of the noun yksityisyys widens the domain of privacy to include inner mental states, as illustrated in (18). (18) Yksityisyys   on   pään   sisässä, privacy-NOM.SG  be-3SG  head-GEN.SG  inside-INE.SG ajatukset ovat yksityiselämää. thought-NOM.PL be-3PL private-life-PART.SG ‘Privacy is inside the head; thoughts are (part of) private life.’ The paraphrases also illustrate three areas of concern related to privacy. The first, illustrated in (19), is the potential conflict between an individual’s wishes and those of others in one’s social group as an essential characteristic of privacy. (19) Yksityisyys   on   suhteessa siihen, että privacy-NOM.SG  be-3SG  relationship-INE.SG it-ILL that on myös be-3SG also

muita ihmisiä. other-PART.PL person-PART.PL

‘Privacy is relative to (the fact) that there are other people too.’

The development of words for privacy   149

The special needs of vulnerable individuals are another area of concern. Examples (20) and (21) articulate this with respect to children and nursing home residents. We see here the theme of the positive value (‘important’ and ‘good’) of privacy combined with the suggestion that those in authority may be responsible for caring for the privacy needs of those less able to assert themselves. (20) Yksityisyys on myös privacy-NOM.SG be-3SG also

tärkeää, important-PART.SG

niinpä so-PARTICLE

jokaisella lapsella on every-ADE.SG child-ADE.SG have

oma own-NOM.SG

kaappi ja closet-NOM.SG and

sänky, bed-NOM.SG

kertoo Pelkonen. tell-3SG.PRES Pelkonen

ja and

oma own-NOM.SG

‘Privacy is also important, and so every child has his or her own closet and bed, Pelkonen says.’ yksityisyys    on    hoivakodissa (21) Asukkaiden resident-GEN.PL privacy-NOM.SG  be-3SG  nursing home-INE.SG hyvä good-NOM.SG

verrattuna entiseen compare-PPTCL.PASS-ESS.SG earlier-ILL.SG

rakennukseen. building-ILL.SG ‘Residents’ privacy is good in the nursing home, compared to the earlier building.’ A third area of concern is that the meaning or availability of privacy is perceived to change. Some writers complain that there is now too much privacy. (22) Yksityisyys   on privacy-NOM.SG   be-3SG ‘Privacy has been taken far.’

viety take-PPTCL.PASS

pitkälle. far

150   Mai Kuha and M. Ali Bolgün (23) Perheen yksityisyys family-GEN.SG privacy-NOM.SG

on kohtalaisen be-3SG relatively

uusi      ilmiö,     mutta  se   vietiin new-NOM.SG  phenomenon-NOM.SG  but   it  take-PAST.PASS välillä    jopa    niin  pitkälle,  että  perheessä sometimes  even-PARTICLE  so   far     that  family-INE.SG sai      tapahtua   melkein  mitä vain  eikä siihen may-3SG.PAST  happen-INF  almost  anything  and-NEG  it-ILL saanut may-3SG.PAST.NEG

puuttua. interfere-INF

‘The privacy of the family is a relatively new phenomenon, but it was sometimes taken so far that almost anything could happen in a family, and one could not get involved in it.’ Finally, yksityisyys seems to extend to autonomy and freedom from social control as well: (24) Yksityisyrittäjän    pöydällä     ja   työhuoneessa entrepreneur-GEN.SG  table-ADE.SG  and  work-room-INE.SG yksityisyys on vahvasti  läsnä. privacy-NOM.SG be-3SG strongly  present ‘On an entrepreneur’s worktable and in his/her working space, privacy is strongly present.’ In this part of its meaning, yksityisyys overlaps with oma rauha, which seems to have the more specific meaning of tranquility and freedom from social control. This is clear in statements such as (25), where oma rauha is contrasted with situations that potentially threaten it: close neighbors and frequent visitors. That is, social obligations and the close proximity of other people are at odds with oma rauha. (25) Kerrostalossa apartment-building-INE.SG

asunut ja live-PPTCL.NOM.SG and

The development of words for privacy   151

anonyymiin    elämään    tottunut anonymous-ILL.SG  life-ILL.SG  accustom-PPTCL.NOM.SG kaupunkilainen city-dweller-NOM.SG

etsii seek-3SG

omaa rauhaa tranquility/autonomy-PART.SG kun when

ja and

omakotitalosta own-home-ELA.SG pettyy, be-disappointed-3SG

tiiviillä omakotitaloalueella   törmää dense-ADE.SG own-home-area-ADE.SG  collide-3SG

sosiaaliseen social-ILL.SG

kontrolliin. control-ILL.SG

‘A city-dweller who has lived in an apartment building and has become accustomed to anonymous life seeks tranquility/autonomy in a house and is disappointed upon colliding against social control in a densely populated suburb.’ You can have oma rauha in plain view, in a public place–for example, by wearing headphones or riding a motorbike: (26) Kun  pistää   ajovehkeet     päälle  ja   kypärän when  put-3SG  riding-gear-ACC.PL  on    and  helmet-ACC.SG päähän, head-ILL.SG

on oma rauha ja be-3SG tranquility/autonomy -NOM.SG and

vapaus   mennä, freedom-NOM.SG  go-INF

kuvailee describe-3SG

motoristin biker-GEN.SG

tunteita espoolainen Jussi Järvelä. feeling-PART.PL Espoo-person J.J. ‘When one puts on riding gear and a helmet, one has tranquility/autonomy and freedom to go, says Jussi Järvelä from Espoo, describing the feelings of a biker.’ In contrast with both yksityisyys and oma rauha, intimiteettisuoja seems to have more to do with solitude for intimate, personal matters. In (27), the absence of shower curtains threatens this aspect of privacy.

152   Mai Kuha and M. Ali Bolgün (27) Tyttöjen girl-GEN.PL

intimiteettisuoja   on privacy-NOM.SG  be-3SG

kun     suihkuissa because/when  shower-INE.PL

ei NEG

vaarassa, danger-INE.SG

ole verhoja. be-3SG curtain-PART.PL

‘The girls’ privacy is threatened when the showers have no curtains.’ 3.4.3. Collocations As we have seen, yksityisyys occurs much more frequently than the other Finnish nouns that overlap in meaning with privacy. It also has a much larger number of collocates than intimiteettisuoja and oma rauha. The most salient group of collocates for yksityisyys consists of various forms meaning ‘protect’ or ‘protection’ (suojella, suojata, suojaaminen, varjella, turvata), ‘respect’ (kunnioittaa, kunnioittaminen) and ‘preserve’ (säilyttää). Together, this group accounts for 10 % of collocations. Other frequently occurring collocates include forms of ‘threaten’ (loukata, loukkaaminen, loukkaus, loukkaavasti, loukkaamisjuttu), ‘break’ (rikkoa, rikko, rikkominen), and other words with related meanings. Other collocations show that yksityisyys comes up in discussions of legal concerns, telecommunications, and information, as well as public performance or art. They also suggest that it contrasts with ‘public’ (julkinen, julkkis, julkistaminen, julkisuus), and is associated with domestic life: ‘home’ (koti), ‘domestic tranquility’ (kotirauha), ‘family’ (perhe), ‘marriage’ (avioliitto). Yksityisyys also collocates with ‘pleasant’ (mieluisa) and ‘want’ (kaivata, haluta). We see some similarities in the collocates for intimiteettisuoja: they tend to have the meaning of ‘protect’ (suojella), ‘take care of’ (huolehtia), as well as ‘violate’ (loukata) and ‘appeal to’ (vedota). Intimiteettisuoja is also information-related: ‘information’ (tieto), ‘conversation’ (keskustelu), ‘name’ (nimi). These collocates reveal a considerable similarity between yksityisyys and intimiteettisuoja. It seems that both are contrasted with the public; both are at risk of being violated and must be protected. Compared with intimiteettisuoja, yksityisyys collocates with a larger number of nouns, with referents related to a wide range of concerns: law, media, information, but also personal life. In contrast with the other privacy-related nouns (yksityisyys and oma rauha), collocates for intimiteettisuoja include many nouns with human referents: ihminen (‘person’), kuljettaja (‘driver’), asiakas (‘customer’) potilas (‘patient’), mies (‘man’), nuori (‘young person’), me (‘we’); these account for more than 16 % of collocations.

The development of words for privacy   153

The collocates of oma rauha, on the other hand, paint a strikingly different picture related to positive personal feelings: oma rauha is something people want (haluta, kaivata), need (tarvita), enjoy (nauttia), and value (arvostaa); these types of meanings account for 15 % of the collocations. In contrast with the other two nouns, almost the only type of person that oma rauha collocates with is ‘I’ (minä) or ‘self’ (itse): even in the journalistic genre from which the corpus is drawn, oma rauha is related to the writer himself or herself, writing in the first person. On a related note, oma rauha occurs in a possessive form much more frequently than the other nouns: 74 of 259 instances, or 29 % (compared with 10 % of instances of yksityisyys occurring as possessives). There is also a suggestion that oma rauha, while highly desired, may not always be easily available, given the significant frequency of ‘get’ (saada), ‘if’ (jos), ‘when’ (kun), and ‘but’ (mutta). In Table 6, the words that occur immediately to the right of yksityinen/ yksityis- (almost all nouns modified by either of these two adjectival forms) are given in order of descending statistical significance relative to the frequency of the collocate in the corpus. As this table shows, the referents of collocates of the adjectival forms meaning ‘private’ are generic people (‘person’, ‘citizen’) and also various kinds of businesspeople (‘entrepreneur’, ‘investor’), as well as nominalizations and other abstract nouns (‘care’, ‘service’). These collocates suggest that these adjectival forms are generally used to highlight the private nature of property (or of activities or processes construed as property), and are more strongly related to the world of business and finances than the noun yksityisyys is. The adjectival forms modify nouns to make their referents into private property; but when the adjectival yksityinen becomes the noun yksityisyys, it is property. Table 6. Collocates immediately following the adjectival forms yksityinen / yksityis(‘private’)

collocate sektori lääkäriasema kulutus liikennöitsijä palveluala päiväkoti yritys

‘sector’ ‘health center’ ‘consumption’ ‘traffic contractor’ ‘service sector’ ‘day care’ ‘enterprise’

chi square 882501 185571 119105 55010 45979 33613 33550

absolute frequency (collocation) 604 69 221 46 76 130 423

absolute frequency (collocate) 7633 474 7554 710 2315 9225 96243

154   Mai Kuha and M. Ali Bolgün

collocate maanomistaja päivähoito sijoittaja puhelinyhtiö ihminen kansalainen henkilö hoito rahoitus omistus työnantaja palvelu yrittäjä metsänomistaja omistaa puoli omistaja taho yhtiö raha ala koulu maa

‘land owner’ ‘day care’ ‘investor’ ‘phone company’ ‘person’ ‘citizen’ ‘person’ ‘care’ ‘funding’ ‘ownership’ ‘employer’ ‘service’ ‘entrepreneur’ ‘forest owner’ ‘own’ ‘side’ ‘owner’ ‘side’ ‘company’ ‘money’ ‘field’ ‘school’ ‘land’

chi square 26513 26482 21488 17520 16833 16066 13131 11823 10539 10140 8332 8313 7931 6333 3646 3617 2548 2463 1167 1107 780 760 281

absolute frequency (collocation) 78 88 85 39 405 160 169 141 105 75 85 1 108 40 68 161 56 38 71 72 49 59 70

absolute frequency (collocate) 4218 5371 6168 1598 172127 28892 39220 30374 18967 10108 15714 42786 26477 4613 22611 121945 21802 10518 71382 76868 50694 73645 221516

4. Discussion and conclusions We have seen that Finnish and Turkish used to offer more limited ways of talking about privacy. In recent decades, nouns related to privacy have begun to emerge, though not in identical ways in the two languages. These linguistic changes can be better understood in the context of social changes. A key process that transformed both Finland and Turkey is urbanization. In Finland, the proportion of the population earning a living through farming or forestry went from 42 % in 1950 to 2 % in 2006 (Kirby 2006: 296). In the second half of the twentieth century, the southern towns saw such a rapid influx from rural areas, particularly from the northern and eastern areas near the border, that “[f]ewer than one in three Finns lived in a town fifty years ago; now, fewer than one in three live in the countryside” (Kirby 2006: 286–288).

The development of words for privacy   155

Density of occupancy and type of dwelling are also relevant to privacy: in 1960, households in some rural communities in the north and northeast averaged more than 1.8 people per room, including the kitchen. Nation-wide, the proportion of households with more than two people per room went from 34.8 % in 1950 to 24 % in 1960 (The Ministry of Housing, Copenhagen, 1968: 101). It is easy to see how these changes may have influenced concerns about privacy: prior to 1950, the typical experience for a Finn would have been to live in fairly crowded quarters with other members of the same household (in a climate which limits opportunities to seek solitude outdoors), but often quite isolated from people beyond the household. In modern Finland, urban dwellers have more space in their homes, but live with the constant and close presence of strangers or near-strangers everywhere: in their apartment building, on the street, and on the bus. Both types of living conditions seem consistent with the value placed in Finnish culture on “quietude”, a time when people are “(thoughtfully engaged in their own thoughts and should not be disturbed” (Carbaugh, Berry, and Nurmikari-Berry 2006: 216). Moving to the corporate and governmental sphere, the issue of protecting information has obviously increased in importance in recent decades, leading to national data protection laws in Europe. In the 1990s, a directive of the European Union established procedures to protect consumers’ and employees’ personal information (Westin 2003: 442–443). These cultural factors lead us to speculate that moments of freedom from social control and from social interaction have long been the most crucial component of privacy in Finland, as the salience of the indigenous term oma rauha also suggests, and that urbanization and the information age have brought concerns about the privacy of property and information to the foreground. There have been significant social, economic, and political changes in Turkey especially after the military coup in 1980: Turkey has had to deal with high rates of unemployment and inflation, economic instability due to new economic policies, and terrorism and the Kurdish issue in southeastern part of the country. The education system was affected by the opening in large numbers of schools shaped by religious ideology, and by the assertion that anyone should be able to wear a headscarf in schools and government businesses. In fact, secularism (defined as the separation of religion and state) is still a hotly debated issue and very recently was in the forefront of discussions and demonstrations in Turkey. For a significant period of time (about 10 years from 1980–1990) there was a lack of serious political opposition, which resulted in political instability and intensified the rumors of corruption. These and many other changes have helped to reshape Turkish society and the way it looks at things in general. For example, in a study aimed at looking at changes in religious, social and politi-

156   Mai Kuha and M. Ali Bolgün cal arenas in Turkey, Çarkoğlu and Toprak (2006) found that 66.2 % of those surveyed would not want to have homosexual couples as their neighbors. 49 % stated they would not want irreligious families as neighbors. 42.9 % would not want to see a Greek family as their neighbors, 42 % an Armenian family, 39.1 % a Jewish family, 28.2 % a Kurdish family, and 24.4 % a family of a different sect. These numbers are especially striking because the likelihood of the responders’ living next to the mentioned group is extremely low. Also found was the fact that the mid to high level of intolerance among people is at 53.5 %. Urbanization statistics in Turkey are strikingly similar to those in Finland. In 1927, 75.78 % of the population lived in villages and 24.22 % in cities. In 2000, 64.90 % of the population lived in cities and 35.10 % in villages (TÜİK 2007). It may very well be that Turkish people have had a substantial need for privacy and a new way of expressing it. This should be normal since, for example, if people are busier (due, for instance, to economic reasons), they might feel a bigger need for privacy. Similarly, if people do not trust each other, they might tend to seek more privacy. In an interesting study, Turkey ranked the lowest after Brazil among 40 countries in terms of people trusting each other. The percentage of those who say “yes, I trust most people” was 6.5 % in 1997, down from 10 % in 1990 (Fukuyama 1997). In a recent interview in the Turkish newspaper Milliyet, sociologist Prof. Dr. Ayata claims that a new middle class different from the traditional middle class has emerged in Turkey in recent times (Sevimay 2007). According to Ayata, this middle class does not like the traditional concepts of religious sects and the agha system, and is characterized by reaching this level of middle class through education. One important characteristic of this class, according to Ayata, is that the members of this class do not like other people (including family members, relatives, neighbors, and other communities) to be involved in or interfere with their way of life, including their decision-making on any issues. All these might explain the increased use of özel yaşam ‘private life’ since mahrem and mahremiyet, with their more specialized meanings, could not satisfy the new need. The connection between these changes in Finland and Turkey and the development of new ways of talking about privacy can be seen as potential evidence against Sapir’s and Whorf’s often-cited ideas about language shaping our perceptions. While we cannot know to what extent the absence of a noun for ‘privacy’ may have influenced Finnish and Turkish speakers’ attitudes towards privacy, ultimately it seems that we have here a case of language adapting to meet new demands, a normal occurrence (Steinberg et al. 2001: 256). For comparison, it is interesting to note that the frequency of the word privacy in the 1990s in the TIME Magazine Corpus was 30.6 per million, or approximately

The development of words for privacy   157

0.000031 (Davies 2007); this is more than five times the frequency of yksityisyys in the Finnish corpus, not to mention mahremiyet in the Turkish corpus. In summary, despite its fairly recent appearance, Finnish yksityisyys is the most frequently occurring noun for ‘privacy’ and has the widest distribution, which suggests that it is the unmarked variant. The two other nouns have narrower meanings: the traditional oma rauha foregrounds autonomy and tranquility, while the lexical item of foreign origin, intimiteettisuoja, is more likely to relate to the privacy of information or privacy in an intimate sense. In contrast, Turkish does not have a noun for ‘privacy’ in common use. The potential candidate özellik is blocked, due to an existing meaning for it. Therefore, a noun borrowed from Arabic, mahremiyet, is sometimes pressed into service, but it has a somewhat specialized meaning, often relating to taboo subjects. Turkish adjectives used to mean ‘private’ also seem somewhat restricted; ‘private’ is not the primary meaning of özel (that is, özel is more frequently used to express other meanings); when it does mean ‘private’, it tends to be used with yaşam, hayat, or yaşantı, all of which mean ‘life.’ In other cases, it contrasts with ‘governmental’ or ‘public’. The information we have presented from dictionaries and corpora suggests some differences in the Turkish and Finnish privacy schemata. Finnish culture seems to have always valued greatly the occasional escape to a peaceful state in which one is free to choose one’s actions. In recent decades, emphasis on the privacy of information has increased. While the Turkish özel yaşam (‘private life’) is a fairly general term, mahremiyet (‘privacy’) foregrounds the body and personal space as areas in which privacy is a concern. Also, gender has a larger influence on the degree of privacy considered appropriate. While an adjective meaning ‘private’ allows speakers to refer to specific aspects of privacy (such as private property or private information), having a noun for ‘privacy’ frees them from having to attend to the particulars of the privacy schema. They can then think and talk about the entire complex interrelated set of rights and responsibilities included in the privacy schema as if it were an object, so that it becomes more natural to construct privacy as a thing, a right, a possession that can be protected by law. The availability of a noun may also naturalize the status of privacy as a right, making claims to privacy even less open to debate. Abbreviations 1 2 ABL ACC

first person second person ablative accusative

158   Mai Kuha and M. Ali Bolgün ADE AOR CM COP CONJ DAT ELA ESS GEN IF ILL INE INF LOC NEG NML NOM OBLIG PART PL PASS PAST PTCL POSS PPTCL PRES Q SG

adessive aorist compound marker copula conjunction dative elative essive genitive conditional illative inessive infinitive locative negative nominalization nominative obligation partitive plural passive past tense participle possessive past participle present tense question particle singular

References Altman, Irwin 1976 Privacy. Environment and Behavior 8 (1): 7–29. 1977 Privacy regulation: Culturally universal or culturally specific? Journal of Social Issues 33 (3): 66–84. Atalay, Nart B., Kemal Oflazer, and Bilge Say 2003 The annotation process in the Turkish treebank. Proceedings of 11th Conference of the EACL – 4th Linguistically Interpreted Corpora Workshop-LINC, Hungary. Baldwin, Mark W., and Stéphane D. Dandeneau 2005 Understanding and modifying the relational schemas underlying insecurity. In Interpersonal Cognition, Mark W. Baldwin (ed.), 33–61. New York and London: The Guilford Press.

The development of words for privacy   159 Bartlett, Frederic C. 1932 Remembering: A Study in Experimental and Social Psychology. Cambridge: Cambridge University Press. Beck, Aaron T. 1967 Depression: Clinical, Experimental, and Theoretical Aspects. New York: Harper and Row. Carbaugh, Donal, Michael Berry, and Marjatta Nurmikari-Berry 2006 Coding personhood through cultural terms and practices: Silence and quietude as a Finnish ‘natural way of being’. Journal of Language and Social Psychology 25 (3): 203–220. Casson, Ronald W. 1983 Schemata in cognitive anthropology. Annual Review of Anthropology 12: 429–462. Çarkoğlu, Ali, and Binnaz Toprak Değişen Türkiye’de Din, Toplum ve Siyaset [Religion, Society and Poli2006 tics in changing Turkey]. TESEV Yayınları. Chan, Ying-Keung 2000 Privacy in the family: Its hierarchical and asymmetric nature. Journal of Comparative Family Studies 31 (1): 1–17. Davies, Mark. 2007 TIME Magazine Corpus (100 million words, 1920s-2000s). Available online at http://corpus.byu.edu/time. Fillmore, Charles J. 1977 Scenes-and-frames semantics. In Linguistic Structures Processing, Antonio Zampolli (ed.), 55–81. Amsterdam: North-Holland. Fiske, Susan T., and Shelley E. Taylor Social Cognition, 2nd ed. New York: McGraw-Hill. 1991 Fukuyama, Francis 1997 Tarihin sonu mu? [The End of History?] Rey Yayınları. Kayseri. Hollander, John The language of privacy. Social Research 68 (1): 5–28. 2001 Iwata, Osamu 1988 Similarities and dissimilarities in the Japanese semantic structure of privacy and its associated concepts. Psychologia 31 (4): 198–206. Kasper, Debbie The evolution (or devolution) of privacy. Sociological Forum 20(1): 69–92. 2005 Kirby, David A Concise History of Finland. Cambridge: Cambridge University Press. 2006 Kövecses, Zoltán 2006 Language, Mind, and Culture: A Practical Introduction. Oxford University Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar, Vol. I: Theoretical Prerequisites. Stanford: Stanford University Press.

160   Mai Kuha and M. Ali Bolgün Lewis, Geoffrey The Turkish Language Reform: A Catastrophic Success (new edition). 2002 Oxford University Press. Linville, Patricia W., and Donal E. Carlston Social cognition of the self. In Social Cognition: Impact on Social Psy1994 chology, Patricia G. Devine, David L. Hamilton, and Thomas M. Ostrom (eds.), 143–193. San Diego: Academic Press. Margulis, Stephen T. 2003 Privacy as a social issue and behavioral concept. Journal of Social Issues 59(2): 243–261. The Ministry of Housing, Copenhagen 1968 Housing in the Nordic Countries. Copenhagen: S.L. Møllers Bogtrykken. Moore, Barrington 1984 Privacy: Studies in Social and Cultural History. Armonk, NY: M.E. Sharpe. Newell, Patricia B. 1994 A systems model of privacy. Journal of Environmental Psychology 14: 65–78. 1998 A cross-cultural comparison of privacy definitions and functions: A systems approach. Journal of Environmental Psychology 18: 357–371. Nock, Steven L. 1998 Too much privacy? Journal of Family Issues 19 (1): 101–118. Palmer, Gary B. 1996 Toward a Theory of Cultural Linguistics. Austin: University of Texas Press. Rice, G. Elizabeth 1980 On cultural schemata. American Ethnologist 7 (1): 152–171. Riso, Lawrence P., Pieter L. du Toit, Dan J. Stein, and Jeffrey E. Young (eds.) Cognitive Schemas and Core Beliefs in Psychologal Problems: A Sci2007 entist-Practitioner Guide. Washington, DC: American Psychological Association. Safran, Jeremy D. 1990 Towards a refinement of cognitive therapy in light of interpersonal theory: I. Theory. Clinical Psychology Review 10: 87–105. Say, Bilge, Deniz Zeyrek, Kemal Oflazer, and Umut Özge Development of a corpus and a Treebank for present-day written Turkish. 2002 In Current Research in Turkish Linguistics, Kamile İmer and Gürkan Doğan (eds.), 183–192, Proceedings of 11th International Conference on Turkish Linguistics. Eastern Mediterranean University, Northern Cyprus. August, 2002. Schauer, Frederick 2001 Introduction: The legal construction of privacy. Social Research 68 (1): 51–53.

The development of words for privacy   161 Sevimay, Devrim Meydandakiler ‘yeni orta sınıf’tır [Those in the public square are ‘the 2007 new middle class’]. Soru-Cevap [Question-Answer] In Milliyet.com. Retrieved on 21 May 2007 from http://www.milliyet. com.tr/ 2007/05/21/ siyaset/asiy.html. Sifianou, Maria, and Eleni Antonopoulou Politeness in Greece: The politeness of involvement. In Politeness in 2005 Europe, Leo Hickey and Miranda Stewart (eds.), 263–276. Clevedon: Multilingual Matters. Stein, Dan J. 1992 Schemas in the cognitive and clinical sciences: An integrative construct. Journal of Psychotherapy Integration 2 (1): 45–63. Steinberg, Danny D., Hiroshi Nagata, and David P. Aline 2001 Psycholinguistics: Language, Mind and World. 2d edition. London and New York: Longman. TÜİK: Türkiye İstatistik Kurumu [Statistics Institute of Turkey] 2007 http://www.tuik.gov.tr/Start.do. The particular statistical chart was retrieved 28 May 2007 from http://www.tuik.gov.tr/PreIstatistik Tablo. do?istab_id=202). van Dijk, Teun A. 1977 Text and Context: Explorations in the Semantics and Pragmatics of Discourse. London: Longman. Westin, Alan F. 2003 Social and political dimensions of privacy. Journal of Social Issues 59 (2): 431–453. Widdowson, Henry G. 1984 Explorations in Applied Linguistics 2. Oxford: Oxford University Press.

Reference works consulted 1890 1952 1973 1978 1985 1985

A Turkish and English Lexicon by Sir James W. Redhouse. Constantinople (İstanbul). An English-Turkish Dictionary by Fahir İz and H.C. Hony. Oxford at the Clarendon Press. Englanti-Suomi suur-sanakirja [English-Finnish General Dictionary] by Raija Hurme and Maritta Pesonen. Helsinki: WSOY. Nykysuomen sanakirja: lyhentämätön kansanpainos [Dictionary of Modern Finnish: Unabridged Edition] by Matti Sadeniemi. Porvoo: WSOY. Langenscheidt Standard English Dictionary: English-Turkish / Turkish-English. İnkılap Kitabevi. Uusi sivistyssanakirja [New Dictionary of Foreign Words] by Annukka Aikio and Rauni Vornanen. Helsinki: Otava.

162   Mai Kuha and M. Ali Bolgün 1990 1990 1992 1994 1994 1994 1995 1996 2002 2005 2007

Englanti-Suomi suur-sanakirja [English-Finnish General Dictionary] by Raija Hurme, Maritta Pesonen, and Olli Syväoja. Helsinki: WSOY. Suomen kielen perussanakirja [Basic Dictionary of the Finnish Language] by Risto Haarala. Helsinki: Kotimaisten kielten tutkimus-keskus. The Oxford Turkish-English Dictionary by H.C. Hony and Fahir İz (Third Edition: A.D. Alderson and Fahir İz). Oxford: Oxford University Press. The American Heritage Dictionary (Third edition). Houghton Mifflin Company. Uusi sivistyssanakirja [New Dictionary of Foreign Words] by Annukka Aikio and Rauni Vornanen. Helsinki: Otava. Arabic-English Dictionary: The Hans Wehr Dictionary of Modern Written Arabic (Fourth edition). Edited by J.M. Cowan. Ithaca, NY: Spoken Language Services, Inc. Türkçe Sözlük [Turkish Dictionary] by Ali Püsküllüoğlu. İstanbul: Yapı Kredi Yayınları Ltd. Şti. Al-Mawrid: A Modern English-Arabic Dictionary by Munir Ba‘albaki. Beirut, Lebanon: Dar El-Ilm Lil-Malayen. Sivistyssanakirja: nykysuomen opas [Dictionary of Foreign Words: Guide to Modern Finnish] by Kalevi Koukkunen, Vuokko Hosia, and Jukka Keränen. Helsinki: WSOY. Lugat-it Türk: http://www.lugatitturk.com/ Seslisozluk: English-Turkish-German dictionary (version 3.5): http:// www.seslisozluk.com

Chapter 6 On collective cognition and language Farzad Sharifian

1. Introduction In the “call for paper” that was prepared for this volume, the editor asked if it would be possible to discover a “social or collective memory” in language (e.g., Climo and Cattell 2002; Gedi and Elam 1996; Halbwachs [1950] 1992). This chapter takes up this topic, discussing one form of collective memory, or more precisely collective cognition. Western psychology has mainly focused on cognition from the perspective of the individual. Even scholars in the area of social cognition have been interested primarily in perception, processing, and representation of social information by individuals. However, a number of scholars have viewed cognition to be a property of groups and not just individual minds (e.g., Clark and Chalmers 1998; Sutton 2005, 2006; Wilson 2005). Group-level cognitions are “at once grounded in and yet transcending the underlying mental states of the interacting agents to which they are collectively ascribed” (Panzarasa and Jennings 2006: 402). The particular form of group-level collective cognition that concerns us in this chapter emerges from the interactions between the members of a cultural group across time and space, which I refer to as cultural cognition (Sharifian 2008b). It should be added that my definition of cultural cognition is not limited to capturing culture-specific cognition, rather it also applies to the fact that different cultural groups may develop similar or partly similar or very different group-level cognition. In fact, the interactions between cultural cognitions of two or more cultural groups may lead to the emergence of new blended cultural cognitions. Cultural cognition is emergent in the sense that it is a gestalt that is more than the sum of its parts and cannot be reduced to the cognition of a single individual in the group (Beckermann, Flohr, and Kim 1992). According to the Dictionary of the Philosophy of Mind “Properties of a complex physical system

164   Farzad Sharifian are emergent just in case they are neither (i) properties had by any parts of the system taken in isolation nor (ii) resultant of a mere summation of properties of parts of the system”1. Thus, cultural cognition is neither totally captured by the cognition of an individual member of a group, nor is it the result of a mere summation of the minds in a group. It is the constant communicative interaction taking place between the members of a group that leads to the emergence of a collective, cultural cognition. As Panzarasa and Jennings (2006: 404) put it, “it is the move from agents’ cognition to a social cognitive structure via social interaction [italics original] that brings about a new form of collective cognition”. Their view of collective cognition is similar to the view of cultural cognition presented here in that they maintain “collective cognition is holistic in the sense of being essentially macroscopic rather than a mere summation of microscopic local properties” (2004: 405). The notion of emergence has been employed in describing a wide array of phenomena such as hurricanes, ant colonies, climates, stock markets, etc. (e.g., Johnson 2001). It has also been used to describe human cognition as emerging from the neural activity in the brain. I find the notion of emergence useful in describing the patterns of cognitive and behavioural life that are often attributed to groups. It should be noted, however, that “emergence” is still very much a descriptive term and, as such, still in its infancy. Research in the area of computational social simulations seems promising in shedding light on the nature of the emergence of macro-level phenomena from the interactions between micro-level, local agents (e.g., Panzarasa, Jennings, and Norman 2001). Cultural cognition appears to be a form of distributed cognition (Hutchins 1994) in the sense that the cognitive structures whose interactions lead to emergent properties are distributed (albeit heterogeneously) across the minds of the members in a cultural group, across time and space (see also Hutchins 2000). The following section now turns to discuss cultural cognition from a complex adaptive system perspective. 2. Cultural cognition as a complex adaptive system The above properties of cultural cognition make it compatible with the descriptive approach of “complex adaptive systems” thinking. Complex adaptive systems are studied through complexity science (e.g., Holland 1995; Waldrop 1992), a multidisciplinary science which seeks to explain how relationships between parts, or agents, give rise to the collective behaviours of a system or 1 http://philosophy.uwaterloo.ca/MindDict/emergence.html

On collective cognition and language   165

group. This approach to science has provided a useful frame of thinking about many aspects of the universe (e.g., Bak 1996). In complex adaptive systems the order is emergent and self-organizing (Allen 1997). As explained above, “emergence” in this sense refers to properties that result from the interactions between the agents in the system. “Self-organization” here means that there is no central control over the behaviour of the individual agents. Complex adaptive systems are also nested and adaptive. Being “nested” means components of complex adaptive systems are themselves complex systems and “adaptive” refers to the ability of the system to learn from experience, and therefore to evolve. Cultural cognition, as mentioned earlier, is an emergent system in that it results from the interactions between the members of a cultural group across time and space. Thus, the emergent properties of cultural cognition as a system at the macro level are not mirror images of those that characterize the cognition of each individual within the group. A closely related property of complex adaptive systems is that the agents constituting the system cannot contain the whole. Again, in this sense cultural cognition is also a complex system in that an individual’s cognition does not comprise the cultural group’s collective and emergent cognition as a whole. Cultural cognition is also nested in that members of a cultural group, as agents of the system, are themselves complex systems, controlled by nervous systems, endocrine systems, etc. As in the case of other complex systems, cultural cognitions have their own unique history of interactions that constantly construct and reconstruct the system. Often changes in the interactions of cultural groups that may initially be viewed as insignificant have a remarkable influence on the future direction of their cultural cognition. This view is largely reflected in the writings of Vygotsky (e.g., 1978), who viewed cognitive phenomena as embodying the characteristics of historically bound sociocultural relations. Another characteristic of complex systems is that they are “open” systems, that is, it is difficult to determine their boundaries. They are also open in the sense of their openness to inputs from individual agents, who have a two-fold role in the complex system. On the one hand, the individual is the locus of cognition and can have a causal role in the development, dissemination and/ or reinforcement of group-level cognition. On the other hand, an individual’s thought and behaviour can be influenced or determined to a varying degree by the cultural cognition that characterizes the cultural group. Thus, the role of individuals in a cultural group may best be described in terms of a circular pattern of cause and effect. Panzarasa and Jennings (2006: 402) maintain that “individual cognition is necessary for collective cognition to come into existence: thus the latter is nomologically dependent on the former”. However, they

166   Farzad Sharifian observe that collective cognition is “ontologically autonomous”, that is, ontologically speaking it has an existence beyond the level of individual cognition. Cultural cognitions are dynamic in that they are constantly being negotiated and renegotiated across generations, and across time and space by members of a cultural group. In this sense, the actions of members of a group constitute a macro-cognitive network that functions as the base for the emergent collective cognition. Because the interactions between the members of a group are not mirror images of each other, the emergent cognition is constantly evolving, making the system adaptive. 3. Cultural cognition and cultural conceptualizations Two intrinsic aspects of cultural cognition are cultural conceptualizations and language. Cultural conceptualizations are the ways in which people across different cultural groups construe various aspects of the world and their experiences (Sharifian 2003). These include people’s view of the world, thoughts, and feelings. For example, different cultural groups may conceptualize the origin of the world and their relationships to each other and to nature quite differently. Also, research in cognitive linguistics has shown how the ways in which people “think about” their thinking and their emotional experiences may differ from one cultural and linguistic group to another (e.g., Palmer, Goddard, and Lee 2003; Enfield and Wierzbicka 2002). Traditionally, cognitive scientists interested in cultural differences in cognition, such as cultural psychologists and cognitive anthropologists, have used analytical tools such as “schema” (e.g., Rice 1980; Shore 1996; Strauss and Quinn 1997), “category” (e.g., Rosch 1978), and more recently “metaphor” (Kövecses 2005) in accounting for the abovementioned cultural conceptualizations. However, the focus of most research in this area has been the cognition of an individual, rather than the group. Studies in social cognition have also largely employed the notions of “schema” and “category” in accounting for people’s perception and processing of social information. I maintain that within the framework of cultural cognition, as sketched out in this chapter, it is equally feasible for us to view schemas, categories, metaphors, etc. as components of cultural cognition. In this sense a cultural schema is an emergent property resulting from the interactions between the members of a cultural group. The cognitive structures (which are themselves cognitive schemas) that give rise to an emergent schema are heterogeneously distributed across the minds in a cultural group. That is, members of a cultural group may share some but not all elements of an emergent cultural schema with each

On collective cognition and language   167

other, and what is shared between two members may not be exactly the same as what is shared by two others in the group (see more in Sharifian 2003). Also, an individual’s cognitive repertoire often includes elements from cultural conceptualizations of different cultural groups, depending on the nature of the person’s social interactions. In today’s globalized world, most people move inevitably between cultural groups and as such internalize cultural conceptualizations from more than one group. At this stage, it should be made clear that the interactions between the group members which give rise to emergent conceptualizations do not take place in an experiential vacuum, but are in fact embedded within the context of their physical experiences. Often people view various aspects of the environment as cognitive “anchors” for their conceptualizations. For example, Aboriginal Australians have conceptually associated their totemic and cultural stories with aspects of their environment. This has often acted as a basis for considering various aspects of the environment, such as a rock, as “sacred”. An important prime source for conceptualizing various aspects of experience is the human body. Studies of embodiment within cognitive linguistics have increasingly shown how different cultures may conceptualize their thoughts and feelings differently in relation to their body parts. Although the notion of the body itself has been alleged to be universal (Wierzbikca 2007), one cultural group may conceptualize emotions in terms of the heart and another group in terms of the belly (Gaby 2008). Often these cultural conceptualizations have been observed to originate from ethnomedical or religious traditions (see further Sharifian et al. 2008). An important class of cultural conceptualizations is that of cultural categories. Although categorization seems to be a universal human faculty, the ways in which people across different cultural groups categorize their experiences may differ. We tend to categorize every single entity around us, perhaps to achieve cognitive economy for more efficient cognitive processing. But we also develop cognitive categories that may not have tangible referents in the external world. These include categories such as “time”. Different cultural groups may not only conceptualize such categories differently but may also use different “concrete experiences” as the basis for conceptualizing these categories. For example, in many industrialized cultures people conceptualize “time” as a “commodity”, which can be “saved”, “spent”, and “budgeted”. This phenomenon has been referred to as conceptual mapping or conceptual metaphor in cognitive linguistics. It should be emphasized again that from a complex systems perspective, cultural conceptualizations operate at two levels: they have a macro-level and a micro-level. They have a macro-level (global level) existence and structure when the unit of analysis is the group, and at the same time they can be viewed

168   Farzad Sharifian as having a micro-level (local level) when the unit of analysis is the individual. This ontological distinction entails different research approaches and methodologies depending on what level is aimed at. 4. Cultural conceptualizations and language Cultural conceptualizations discussed so far in this paper have conceptual existence as well as linguistic encoding. Language is a central aspect of cultural cognition in that it serves as a “collective memory bank” (Frank 2003, 2005; wa Thiong’o 1986) for cultural conceptualizations, past and present. It is shaped by the cultural conceptualizations that have prevailed at different stages in the history of a speech community and these can leave their traces in current linguistic practice. In this sense language can be viewed as one of the primary mechanisms for storing and communicating cultural conceptualizations. It acts as both a memory bank and a fluid vehicle for the (re-)transmission of these socioculturally embodied cultural conceptualizations. Like cultural cognition, language can also be viewed as a complex adaptive system (e.g., Frank 2008; Steels 2000; Sharifian 2008b). The lexicon of a language is perhaps the most direct link with cultural conceptualizations in the sense that lexical items largely act as labels, and hence “memory banks”, for conceptualizations that are culturally constructed. At least within the circle of cognitive linguistics it is agreed that meaning is conceptualization, and within the subdiscipline of cultural linguistics (Palmer 1996; Sharifian and Palmer 2007) it is further emphasized that conceptualizations are largely culturally constructed. “Cultural construction” of conceptualizations refers to the emergent nature of these conceptualizations at the level of cultural cognition. In short, the lexical items of human languages need to be viewed as capturing and storing cultural conceptualizations such as cultural schemas and categories. The author’s research on Aboriginal English has revealed how even everyday words of this indigenized variety of English, such as ‘family’ and ‘home’, instantiate Aboriginal cultural schemas and categories that are largely distinct from those associated with Standard Australian English (Sharifian 2005a, 2006). For instance, the word ‘mum’ in Aboriginal English may refer to someone who is culturally in the same category as one’s biological mother, such as one’s mother’s sister (e.g, Sharifian 2007). On the other hand, ‘home’ may be used by Aboriginal English speakers to refer to wherever their “extended” family members live (e.g., Sharifian 2008c). Then we find that at the level of grammar, some languages reveal interactions between certain syntactic devices and cultural conceptualizations such as those of politeness and kinship. Murrinh-

On collective cognition and language   169

Patha, an Australian Aboriginal language, uses ten noun classes, which are reflective of Murrinh-Patha cultural classification (Walsh 1993; Street 1987). These classes are identified through noun class markers appearing before the noun. The following list includes the class markers and the definition of each category (Walsh 1993: 110): 1. kardu: Aboriginal people and human spirits. 2. ku: Non-Aboriginal people and all other animates and their products. 3. kura: potable fluid (e.g., ‘fresh water’) and collective terms for fresh water (e.g., ‘rain’, ‘river’). 4. mi: flowers and fruits of plants and any vegetable foods. Also faeces. 5. thamul: spears. 6. thu: offensive weapons (defensive weapons belong to nanthi), thunder and lightning, playing cards. 7. thungku: Fire and things associated with fire. 8. da: place and season (e.g., dry grass time). 9. murrinh: speech and language and associated concepts such as song and news. 10. nanthi: a residual category including whatever does not fit into the other nine categories. The above categorization also allows for multiple memberships, that is, depending on the function of an entity at the time, it may be categorized into one or another class. For instance, a boomerang may be categorized as nanthi when it is used as a back-scratcher and thu when it is used as an offensive weapon (Walsh 1993). Also in the Dreamtime Creation stories, when the Ancestor beings turn into animals in their journey of creating nature, this is signalled by a switch from one noun class to another. This system of noun classification is obviously entrenched in Murrinh-Patha cultural conceptualizations. For instance, as Walsh argues, the fact that fresh water, fire, and language have separate classes is an indication that each holds a prominent place in the culture of the Murrinh-Patha. This is a revealing case of how language has acted as a collective memory bank for certain cultural conceptualizations, which may or may not be currently active in the cultural cognition of the group of speakers. Even if they are active at the level of cultural cognition, they are likely to be heterogeneously distributed across the minds in the group, rather than equally shared between them. As an example of the instantiation of cultural conceptualizations in the pronoun system, in Arabana, another Aboriginal language, the pronoun arnanthara, which may be glossed into English as ‘kinship-we’, captures the following complex category (Hercus 1994: 117):

170   Farzad Sharifian Arnanthara = we, who belong to the same matrilineal moiety, adjacent generation levels, and who are in the basic relationship of mother, or mothers’ brother and child.

In Arabana, this cultural categorization of kin groups is also marked on the second plural kinship pronoun aranthara and third person plural kinship pronoun karananthara. It can be seen here that for an understanding of the meaning of this pronoun the reader would need to know what ‘matrilineal moiety’ means, which itself captures a cultural system of categorization of people in the Aboriginal community. Another example of the link between grammar and cultural conceptualizations is found in an entirely different language, namely Persian, and specifically, in the case of the second-person plural pronoun shomaa2 . This pronoun is used as a second person singular honorific and the third person plural pronoun ishaan is also used as an honorific for the third person singular. Plurality as a marker of respect is not only marked in the pronoun system but can also be optionally marked by the verb ending. In fact, the interaction between the choice of pronoun, verb ending and the verb can yield a hierarchical system in terms of the degree of respect that each sentence conveys. Consider the following examples: (1)

a. b.

c.

d.

in nokteh raa    ‘u    beh man goft. this point DO.marker  he/she  to me said-SG in nokteh raa     ishaan    beh   man this point DO.marker  he/she(respect)  to   me goft. said-SG in nokteh raa    ishaan    beh  man this point DO.marker  he/she(respect)   to   me goft-and. said-PL in nokteh raa    ishaan    beh  man this point DO.marker  he/she(respect)  to   me farmud-and. said-PL ‘He told me this point.’

The above set of sentences differ in terms of the degree of respect and esteem that one holds for the person being talked about, whether or not the person is 2 In Persian transcriptions, the letter “a” symbolises a low front vowel which is close to the sound of “a” in the word “cat”. The “aa” sequence, on the other hand, stands for a low back vowel which is close to the sound of “a” in the word “father”.

On collective cognition and language   171

physically present when the conversation is being conducted. The degree of respect increases from (a) to (d). Sentence (a) is the most neutral in terms of respect. In (b) the degree of respect is increased by the choice of a plural pronoun for a third person singular case. Sentence (c) conveys a higher degree of respect by adding a plural verb ending while (d) shows the highest degree of respect by choosing the verb farmud, which is considered more respectful than goft. In addition it brings into play the plural verb ending and pronoun. Thus, in cases such as the above, which abound in Persian, the cultural conceptualizations of politeness are marked in the choice of pronoun, verb, and verb ending. It should be added that the choice between the three versions is not a straightforward rule that can be explicated in one sentence but in fact requires familiarity with cultural schemas governing communicative interactions between speakers of Persian. The examples discussed so far should suffice to show how grammatical features of a language may be entrenched in cultural conceptualizations. Another aspect of language that embodies cultural conceptualizations is the use of expressions which include a body part and appear to be metaphoric. In English one finds expressions such as you broke my heart, which suggests a conceptualization of heart as the seat of emotions. As mentioned earlier, many studies have revealed cultural differences in conceptualizations of body parts, in the sense that different body parts may be conceptualized as the seat/centre of thoughts, feelings, courage, etc. (e.g., Sharifian et al. 2008). These conceptualizations are usually encoded in linguistic expressions that appear to be figurative, particularly if the original sources of the conceptualizations, such as ethnomedical traditions, are not consciously accessible to the speakers. An example from Chinese comes from Yu (2007) who explores the Chinese cultural conceptualizations of the heart. These give rise to metaphors that profile this internal body organ as a physical entity (e.g., the heart as a container), a part of the body (e.g., the heart as the ruler of the body), and the locus of affective and cognitive activities (e.g., the heart as the house of all emotional and mental processes). Yu observes that the Chinese word xin refers to faculties that are covered by the ‘heart’ and the ‘mind’ in English. He attributes this to ancient Chinese philosophy in which the heart was conceptualized as the organ for thinking, feeling, will, reason, and intuition (Yu 2008). He further attributes the conceptualization of the heart as the monarch of the body to traditional Chinese medicine, which is based on the categorization of five elements. In traditional Chinese medicine the heart is the master of the body and governs various emotional and intellectual activities. As examples of the encoding of this conceptualization of the heart in the Chinese language, Yu (2007: 67) provides the following:

172   Farzad Sharifian (2) 怡悦荡心房。 Yiyue dang xin-fang. joy wave (in) heart-house/room ‘Joy rippled in the heart.’ (3) 进城几年了，乡亲们的嘱托他一直记在心间。 Jin cheng ji nian le, xiangqin-men de enter city several years per fellow-villagers mod zhutuo ta yizhi ji zai xin-jian. advice he always remember in heart-room/inside ‘Having lived in the city for several years, he always bears in mind (lit. in the heart room or inside his heart) the fellow villagers’ advice.’ The above examples clearly reflect conceptualization of xin as the seat of both memory and feelings. It is to be added here that although the encoding of the cultural conceptualizations under discussion have remained relatively constant linguistically speaking – with language acting as a memory bank – at the cultural level of cognition their representation is very likely to be heterogeneously distributed. That is, at the level of the individual language agent, Chinese speakers are likely to reveal individual differences in terms of the extent to which they consider the heart as the real seat of thinking and feeling. Some may consider such expressions as merely a matter of figurative language. From the perspective of cultural cognition, what is important is the view that although these conceptualizations originated from traditional medical/philosophical traditions, they have developed an emergent (macro-level) existence, which is the result of the “negotiation” and “renegotiation” of these conceptualizations by Chinese speakers in their communicative interactions across generations and thus across time and space. I now return to the link between cultural conceptualizations and pragmatics. Cultural conceptualizations closely govern pragmatic meanings and the ways in which speech acts are interpreted. Scholars engaged in research in the area of pragmatics view pragmatic meaning as residing in the knowledge shared between speakers. Moreover, researchers are well aware that pragmatic meanings are subject to cross-cultural differences as well as similarities (e.g., Blum-Kulka, House, and Kasper 1989; Gass and Neu 1995; Wierzbicka 1991; Wolfson 1981). From the theoretical discussion presented thus far in this chapter, it should be clear that the notion of cultural conceptualizations is partly an attempt to provide an account of this supposedly “shared cultural knowledge” that provides a basis for understanding pragmatic meanings across different languages. Within this approach, the notion of “sharedness” is more precise-

On collective cognition and language   173

ly viewed as “heterogeneously distributed”, while “knowledge” is viewed as largely a matter of “conceptualization”. I maintain that people within a speech community understand implicatures or illocutionary forces of each other’s communicative acts in the light of the cultural schemas and categories that characterize the cultural cognition of the community in question. Naturally when it comes to intercultural communication, differences and similarities between the cultural cognitions of the cultural groups involved may facilitate or debilitate the understanding of pragmatic meanings. 5. Cultural schemas of ‘Self’ in Persian In the rest of this chapter I focus on providing an example of the link between cultural schemas and pragmatic levels of language by considering the case of a ‘self’ schema in Persian, which should be of particular interest to scholars working in the area of social cognition. One of the significant cultural schemas affecting communicative interactions in Persian is that of shekasteh-nafsi. The phrase itself is composed of two morphemes: shekasteh ‘broken’ and nafs, which may roughly be glossed into English as ‘self, ego, soul, inner self, psyche’. As a whole this expression refers to ‘humbleness’ through somehow suppressing or “breaking” one’s “self” and “ego”. This schema captures norms about the ways in which a person should place their nafs in relation to those of others in the wider society. Nafs has a pivotal role in the spiritual tradition of Sufism in which at least three stages of “self” have been conceptualized (Nurbaksh 1992). The first stage, as part of the purification of the soul, is combating nafs amaareh ‘carnal’ nafs. This aspect of the “self” reveals a tendency for pride, greed, and selfishness, which are considered as “evil” thoughts and acts. The second stage is what is called nafs lavaameh, or the ‘reproaching’ nafs, which admonishes a person for their “evil” acts and impels them to perform “good” deeds. The last stage is nafs motma’eneh, or the ‘satisfied’ nafs, in which the Sufi is free from all materialism and earthly problems and would be satisfied with the will of God. It is the Sufi’s battle with their materialistic and carnal nafs, in particular with traits such as selfishness and pride, which is the core value captured by the cultural schema of shekasteh-nafsi. While Sufism is not a strong spiritual tradition in contemporary Iranian society, the cultural schema of shekasteh-nafsi is a significant part of contemporary Persian cultural cognition. This appears to be due to the fact that many Persian-speaking literary figures, particularly poets, have followed the Sufi tradition and their works embody values and principles cherished by Sufism. Persian speakers pride themselves highly on their

174   Farzad Sharifian well-known works of literature, including Sufi literature, and urge the younger generations to model the ethos incorporated in such Persian literature. Ahmadi and Ahmadi (1998: 104) observe that “Persian literature is full of texts urging everyone to pay respect to others, to be extremely polite in front of others, not to speak of one’s ‘I’ and one’s achievements”. Thus it is not surprising to see the link between the cultural schema of shekasteh-nafsi and the Sufi tradition. This is a good example of an emergent conceptualization, where a value system originally part of a spiritual tradition finds its way into the literary works of a speech community and then into the cultural cognition of a group through sustained communicative interactions taking place between the members of the group across time and space about the value system. Often people whose interactions reflect such cultural conceptualizations are not consciously aware of the original root/source of them. Returning to the link between cultural conceptualizations and language, as an example, a significant case of the reflection of the cultural schema of shekasteh-nafsi in communicative interactions between speakers of Persian is in compliment responses (see more in Sharifian 2005b, 2008a). In line with the conceptualization of shekasteh-nafsi, when receiving compliments and praise, speakers tend to highly disagree with a compliment, play it down, return it to the complimenter, or reassign it to an interlocutor, a family member, and/or God. Often the interlocutors in the communicative event engage in extended chains of returning the compliment back and forth, in the sense that the receiver of the compliment returns the compliment to the complimenter, who in turn returns the compliment to the complimentee and this may continue for some time. Another instantiation of the cultural schema shekasteh-nafsi in the Persian language comes about as a result of the functioning of schemas associated with another common morpheme referring to “self”, namely, khod. This morpheme forms part of many Persian words, which are usually composed of a noun and adjective. The following are some common compounds that include this morpheme (translations are from Aryanpur Persian-English Dictionary, 1984): Khodbeen Khodparast Khodpasand Khodkhaah Khoddaar Khodra’i Khodsaakhteh Khodsar Khodkaameh

‘khod+see’ (conceited) ‘khod+worship’ (selfish, egoist) ‘khod+choose’ (self-admire) ‘khod+desire’ (egoist, self-centred) ‘khod+have’ (reserved) ‘khod+opinion’ (obstinate) ‘khod+built’ (self-made) ‘khod+head’ (opinionated) ‘khod+desire’ (self-centred)

On collective cognition and language   175

Khodmokhtaar ‘khod+govern’ (autonomous, self-governed) Khodnamaa ‘khod+show’ (showy) From the above list of words, only the word khodsaakhteh ‘self-made’ captures a positive and approved-of trait, from a Persian perspective. All the other words in the list capture a negative trait, except the word khoddaar, which refers to someone who keeps their thoughts and feelings to themslves. This term can be used in a positive or a negative sense. Interestingly, the word khodmokhtaar, which refers to a wilful person, who usually acts and makes decisions according to their own will or desire, is largely a negative trait and can be used to refer to a child who does not follow its parents’ guidance. However, the English translation provided by the Aryanpur Persian-English Dictionary (1984) equates this word with ‘autonomous’, which is a rather positive term in English. Overall and in line with the cultural schema of shekasteh-nafsi any form of self-endearment is negatively valued in Persian. It should be added that while it seems English has similar values attached to such traits, two points of caution need to be kept in mind. First, in English, words such as ‘self-centred’ and ‘egoist’ are not very commonly used while their counterparts in Persian are very common words. Also what is categorized as khodkhaahi in Persian is not necessarily viewed as ‘self-centredness’ in English. I maintain that often the lexical items regarded as equivalent by bilingual dictionaries may in fact be equating cultural conceptualizations with each other whose fields of meaning or referentiality may overlap to some extent but which are structured by the different schemas and categories. One such thorny expression is the English word self-love, which can be a positive trait that is viewed approvingly in English. The Persian term whose meaning is closest to that of this English expression of ‘self-love’ would be khodsheefteh ‘khod-lured’, which captures a very negative trait, one that is viewed with strong disapproval in Persian. In short, the examples provided here should make it clear how entrenched cultural conceptualizations can be brought into focus, stored and recirculated through the use of linguistic resources. This, in turn, supports the view that lexicon, and language in general, may act as an “archival site” or “memory bank” for heterogeneously distributed collective cognition. 6. Concluding remarks This chapter elaborates on the author’s thinking about the relationship between cognition, culture, and language by exploring the notion of cognition as a property of groups, and not just individuals. This group-level, collective cognition is

176   Farzad Sharifian an emergent property of the interactions that take place between members of a cultural group. Two intrinsic aspects of cultural cognition are cultural conceptualizations and language, aspects that are deeply intertwined. Cultural conceptualizations are group-level conceptualizations that are constantly negotiated and renegotiated across time and space by members of a cultural group. These conceptualizations have a micro-level (local) and a macro-level (global) existence. The micro-level cognitive structures are those that characterize the cognition of the individual while the macro-level ones are those which emerge, cumulatively, from the effects of the micro-level cognitions during such communicative interactions. The properties of what is viewed as cultural cognition seem to be in consonance with complex systems thinking in that they reveal emergent properties, they are nested and “open”, and they are also dynamic and self-organizing. Language is a central aspect of cultural cognition in that it serves as a “collective memory bank” for cultural conceptualizations that have prevailed at different stages in the history of a speech community. Language may best be viewed as a primary mechanism, but surely not the only one, for storing and communicating cultural conceptualizations. This chapter has provided examples from various levels of language and from several languages where different linguistic features and devices appear to be entrenched in the cultural conceptualizations of their speakers. The observations presented in this chapter are meant to provide some preliminary thoughts for further theoretical as well as empirical work in cognitive science that in turn hopefully will allow for fresh insights into the complex relationship between culture and language. It seems that the analytical tools of cognitive science, such as schemas and categories, as well as recent developments in the area of complex adaptive systems can facilitate our understanding of the relationship between culture, cognition, and language.

Acknowledgements The author wishes to thank Professor Roslyn M. Frank for her great support and encouragement as well as very helpful comments on the earlier versions of this chapter. Hanna Pishwa also deserves a special word of thanks for her helpful comments.

References Ahmadi, Nader, and Fereshteh Ahmadi Iranian Islam: The Concept of the Individual. London: MacMillan 1998 Press.

On collective cognition and language   177 Allen, Peter M. 1997 Cities and Regions as Self-organizing Systems: Models of Complexity. London: Gordon and Breach Science. Aryanpur Kashani, Abbas, and Manouchehr Aryanpur Kashani 1984 Aryanpur Persian-English Dictionary. Tehran: Amir-Kabir Publications. Bak, Per 1996 How Nature Works: The Science of Self-Organized Criticality. New York: Copernicus. Beckermann, Ansgar, Hans Flohr, and Jaegwon Kim 1992 Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism. Berlin: Walter de Gruyter. Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper (eds.) 1989 Cross-cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Clark, Andy, and David Chalmers 1998 The extended mind. Analysis 58: 10–23. Climo, Jacob J., and Maria G. Cattell (eds.) Social Memory and History: Anthropological Perspectives. Walnut 2002 Creek, CA: AltaMira Press. Enfield, Nick, and Wierzbicka, Anna (eds.) 2002 The Body in the Description of Emotion. Special issue of Pragmatics and Cognition 10 (1). Frank, Roslyn M. 2003 Shifting identities: The metaphorics of nature-culture dualism in Western and Basque models of self. Metaphorik.de O4/2003 (electronic journal). http://www.metaphorik.de/04/frank.pdf Shifting identities: A comparative study of Basque and Western cul2005 tural conceptualizations. Cahiers of the Association for French Language Studies 11 (2): 1–54. http://www.afls.net/cahiers.html [accessed 01.22.07] The language-species-organism analogy: A complex adaptive systems 2008 approach to shifting perspectives on ‘language’. In Body, Language, and Mind. Vol. 2. Sociocultural situatedness, Roslyn M. Frank, René Dirven, Tom Ziemke, and Enrique Bernárdez (eds.), 215–264. Berlin/ New York: Mouton de Gruyter. Gaby, Alice 2008 Gut feelings: Locating intellect, emotion and life force in the Thaayorre body. In Culture, Body, and Language: Conceptualizations of Internal Body Organs across Cultures and Languages, Farzad Sharifian, René Dirven, Ning Yu, and Susanne Niemeier (eds.), 27–44. Berlin/New York: Mouton de Gruyter. Gass, Susan M., and Joyce Neu 1995 Speech Acts across Cultures: Challenges to Communication in a Second Language. Berlin: Mouton de Gruyter.

178   Farzad Sharifian Gedi, Noa, and Yigal Elam 1996 Collective memory: What is it? History and Memory 8 (1): 30–50. Halbwachs, Maurice The Collective Memory, Lewis A. Coser (ed. and trans.). Chicago: Chi1992 cago University Press. Original French edition, 1950. Hercus, Luise A. 1994 A Grammar of the Arabana-Wangkangurru Language Lake Eyre Basin, South Australia. Pacific Linguistics Series C – 128. Australian National University: Canberra. Holland, John H. 1995 Hidden Order: How Adaptation Builds Complexity. Reading, MA: Addison Wesley. Hutchins, Edwin Cognition in the Wild. Cambridge, MA: MIT Press. 1994 2000 Distributed cognition. http://eclectic.ss.uci.edu/~drwhite/Anthro179a/ DistributedCognition.pdf 2005 Material anchors for conceptual blends. Journal of Pragmatics 37 (10): 1555–1577. Johnson, Steven 2001 Emergence: The Connected Lives of Ants, Brains, Cities and Software. New York: Scribner. Kövecses, Zoltán Metaphor in Culture: Universality and Variation. Cambridge: Cam2005 bridge University Press. Nurbakhsh, Javad 1992 The Psychology of Sufism (Del va Nafs). London/New York: KhaniqahiNimatullahi Publications. Palmer, Gary B. 1996 Toward a Theory of Cultural Linguistics. Austin: University of Texas Press. Palmer, Gary B., Cliff Goddard, and Penny Lee (eds.) Talking about Thinking across Languages. Special issue of Cognitive 2003 Linguistics 14 (2–3). Panzarasa, Pierto, and Nicholas R. Jennings 2006 Collective cognition and emergence in multi-agent systems. In Cognition and Multi-agent Interaction, Run Sun (ed.), 401-408. Cambridge: Cambridge University Press. Panzarasa, Pierto, Nicholas R. Jennings, and Timothy J. Norman 2001 Social mental shaping: Modelling the impact of sociality on the mental states of autonomous agents. Computational Intelligence 17 (3): 738–782. Rice, G. Elizabeth On cultural schemata. American Ethnologist 7: 152–171. 1980 Rosch, Eleanor 1978 Principles of categorization. In Cognition and Categorization, Eleanor Rosch, and Barbara B. Lloyd (eds.), 27–48. Hillsdale, NJ: Lawrence Erlbaum.

On collective cognition and language   179 Sharifian, Farzad On cultural conceptualisations. Journal of Cognition and Culture 3 (3): 2003 187–207. 2005a Cultural conceptualisations in English words: A study of Aboriginal children in Perth. Language and Education 19 (1): 74–88. 2005b The Persian cultural schema of shekasteh-nafsi: A study of compliment responses in Persian and Anglo-Australian speakers. Pragmatics and Cognition 13 (2): 337–361. 2006 A cultural-conceptual approach to the study of World Englishes: The case of Aboriginal English. World Englishes 25 (1): 11–22. 2007 Aboriginal language habitat and cultural continuity. In The Language Habitat of Aboriginal Australia, Gerhard Leitner, and Ian G. Malcolm (eds.), 181–196. Berlin/New York: Mouton de Gruyter. 2008a Cultural schemas in L1 and L2 compliment responses: A study of Persianspeaking learners of English. Journal of Politeness Research 4 (1): 55–80. 2008b Distributed, emergent cognition, conceptualisation, and language. In Body, Language, and Mind. Vol. 2: Sociocultural situatedness, Roslyn M. Frank, René Dirven, Tom Ziemke, and Enrique Bernárdez (eds.), 109–136. Berlin/New York: Mouton de Gruyter. 2008c Cultural model of Home in Aboriginal children’s English. In Cognitive Sociolinguistics, Gitte Kristiansen, and René Dirven (eds.), 333–352. Berlin/New York: Mouton de Gruyter. Sharifian, Farzad, and Gary B. Palmer (eds.) Applied Cultural Linguistics: Implications for Second Language Learn2007 ing and Intercultural Communication. Amsterdam/Philadelphia: John Benjamins. Sharifian, Farzad, René Dirven, Ning Yu, and Susanne Niemeier, (eds.) Culture, Body, and Language: Conceptualizations of Internal Body 2008 Organs across Cultures and Languages. Berlin/New York: Mouton de Gruyter. Shore, Bradd 1996 Culture in Mind: Cognition, Culture, and the Problem of Meaning. Oxford: Oxford University Press. Steels, Luc Language as a complex adaptive system. In Proceedings of the Sixth 2000 Conference on Parallel Problem Solving from Nature, Marc Schoen­ auer, et al. (eds.), 17–26. Berlin: Springer-Verlag. Strauss, Claudia, and Naomi Quinn A Cognitive Theory of Cultural Meaning. New York: Cambridge Uni1997 versity Press. Street, Chester S. 1987 An Introduction to the Language and Culture of the Murrinh-Patha. Darwin: Summer Institute of Linguistics. Sutton, John Memory and the extended mind: Embodiment, cognition, and culture. Cognitive Processing 6 (4): 223–226.

180   Farzad Sharifian Sutton, John (ed.) Memory, Embodied Cognition, and the Extended Mind. Special issue of 2006 Philosophical Psychology 19 (3). wa Thiong’o, Ngugi 1986 Decolonising the Mind: The Politics of Language in African Literature. London: Heinemann. Vygotsky, Lev Semenovich 1978 Mind in Society: The Development of Higher Psychological Processes. Cambridge: Harvard University Press. Waldrop, M. Mitchell 1992 Complexity: The Emerging Science at the Edge of Order and Chaos. New York: Simon and Schuster. Walsh, Michael 1993 Classifying the world in an Aboriginal language. In Language and Culture in Aboriginal Australia, Michael Walsh and Colin Yallop (eds.), 107–122. Canberra: Aboriginal Studies Press. Wierzbicka, Anna Cross-cultural Pragmatics. Berlin: Mouton de Gruyter. 1991 2007 Bodies and their parts: An NSM approach to semantic typology. Language Sciences 29: 14–65. Wilson, Robert A. 2005 Collective memory, group minds, and the extended mind thesis. Cognitive Processing 6 (4): 227–236. Wolfson, Nessa Compliments in cross-cultural perspectives. TESOL Quarterly 15 (2): 1981 117–123. Yu, Ning The Chinese conceptualization of the heart and its cultural context: Im2007 plications for second language learning. In Applied Cultural Linguistics: Implications for Second Language Learning and Intercultural Communication, Farzad Sharifian, and Gary B. Palmer (eds.), 65–85. Amsterdam/Philadelphia: John Benjamins. 2008 The Chinese heart as the central faculty of cognition. In Culture, Body, and Language: Conceptualizations of Internal Body Organs across Cultures and Languages, Farzad Sharifian, René Dirven, Ning Yu, and Susanne Niemeier, (eds.), 131–168. Berlin/New York: Mouton de ­Gruyter.

Section II Social cognition in discourse

Chapter 7 Conversational pragmatics and social cognition Thomas Holtgraves and Blake M. Anderson 1. Introduction A fundamental task for social cognition is to describe the processes through which people come to know their social world. To that end, mainstream social cognition researchers have explored in detail the attributions that people make for their behavior and that of others, what they remember about others, their perceptions and judgments of people, and so on. Not surprisingly, the emphasis in this research has been on intraindividual processes, the internal memory, reasoning, and judgment processes that are assumed to underlie these effects. But the assessment of these processes usually involves language in some manner – people are asked to report their attributions, attitudes, memories, judgments, and so on. There have now been numerous demonstrations that assumptions regarding conversational cooperativeness can influence the nature of what participants report in these situations (Hilton 1995; Schwarz 1996). In this way many social cognitive phenomena reflect, at least in part, participants’ adherence to certain interpersonal rules for using language. The purpose of this chapter is to elaborate on this claim by interpreting a variety of empirical findings in terms of conversational pragmatics. To do this, we first summarize Grice’s (1975) theory of conversational implicature; this theory has served as the major framework for examining the communicative foundation of social cognition, and it will act as the framework for the present chapter. We then consider how the operation of these pragmatic rules can play an important role in social cognition experiments. Finally, we consider some possible extensions of this approach as well as some potential limitations. 1.1. Grice’s theory of conversational implicature Grice’s (1975) theory of conversational implicature is essentially concerned with how it is possible for people to convey nonliteral meaning. He proposed

184   Thomas Holtgraves and Blake M. Anderson that this is possible because interlocutors abide by what he termed the cooperative principle (CP), a rule stating simply that a person should “make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged” (1975: 45). This very general requirement is further specified in terms of the following four conversational maxims: 1. Quantity Make your contribution as informative as required (i.e., do not be either overinformative or underinformative). 2. Quality Try to make your contribution true; one for which you have evidence. 3. Manner Be clear. That is, avoid ambiguity, obscurity, etc. 4. Relation Make your contribution relevant for the exchange. To abide by the CP when conversing, then, a person’s utterances should be clear (manner), truthful (quality), relevant for the topic (relation), and the right size (quantity). However, people do not always abide by these maxims. Sometimes people make an irrelevant statement, other times they may say too much or too little, and so on. It is usually the case, however, that people will mutually assume adherence to the CP and its maxims, and this assumption serves as a frame for interpreting a speaker’s utterances. In other words, a speaker’s utterances will be interpreted as if he were clear, relevant, truthful, and informative, and recipients will generate a conversational implicature in order to make it adhere to the CP and/or its maxims. For example, a person who asks, “Can you pass the salt?” at the dinner table violates the relation maxim (i.e., in this context, inquiring into another’s ability to pass the salt would not be a relevant contribution). However, because the hearer assumes the speaker is being relevant, she will search for an ulterior meaning (and make a conversational implicature); she will interpret the utterance in such a way so as to maintain adherence to the conversational maxims. The “pass the salt” example is an instance of an implicature that arises to preserve adherence to a specific conversational maxim. Additionally, speakers will sometimes intentionally flout or violate a maxim, in which case it is simply not possible for the hearer to assume the speaker is adhering to the maxims. For example, a person who abruptly changes the topic by saying, “I hope it stops raining soon” in response to a personal question such as “How did you do on that Chemistry exam?” is violating the relation maxim and will convey much more than a strict literal meaning of the remark (possible gloss: I didn’t do well on the exam). In this case, it is obvious the speaker is not complying with the relation maxim. Despite this apparent noncom-

Pragmatics and social cognition   185

pliance, the hearer will usually still assume overall cooperativeness on the part of the speaker, and as a result, generate a conversational implicature that makes sense of the violation. Many types of figurative language can be explained in this manner. Consider, for example, tautologies such as “Boys will be boys” or “War is war.” These statements clearly violate the quantity maxim; they are essentially uninformative. Yet these utterances will in most contexts convey clear nonliteral meanings. When a person utters a redundant statement, people search to remove that redundancy in order to make sense of the utterance. According to Grice, then, all communication comes with a presumption of relevance (via the cooperative principle) (see also Sperber and Wilson 1986, 1995). People automatically assume a speaker’s utterance makes a relevant contribution, and if its relevance is not immediately apparent, they will generate an inference to make it so. In addition, and more importantly for the present concern, people typically assume that all of a person’s communications are relevant (i.e., the relation maxim). In this regard, social cognition experiments represent a unique situation because experimenters may assume these maxims are not applicable, a view that is not shared by the participants. More specifically, experimenters may provide participants with information in order to see if they use it or not. However, they often fail to realize that participants assume that all information is relevant and so they use it. Somewhat more subtly, participants may generate implicatures that the experimenter did not intend, implicatures designed to enhance the relevance of the experimenter’s communications (e.g., the experimental instructions). The experiments that we will discuss can be divided into two categories. First, we will review experiments examining the effects of explicit irrelevant information; in other words, information that the experimenter acknowledges and believes to be irrelevant in the experiment. Second, we review experiments that may contain irrelevant implicit information; information that is not explicitly conveyed but is implicit and generated by participants. 3. Explicit irrelevant information 3.1. Representativeness heuristic One of the earliest studies to be interpreted within the Gricean framework was Kahneman and Tversky’s (1973) classic study of the use of the representativeness heuristic. In this research, Kahneman and Tversky (1973) provided participants with descriptions of people such as the following:

186   Thomas Holtgraves and Blake M. Anderson Jack is a 45-year old man. He is married and has four children. He is generally conservative, careful, and ambitious. He shows no interest in political and social issues and spends most of his free time on his many hobbies, which include home carpentry, sailing, and mathematical puzzles.

After reading the description, they were given one version of the following instructions: A panel of psychologists have interviewed and administered personality tests to 30 (70) engineers and 70 (30) lawyers, all successful in their respective fields. On the basis of this information, thumbnail descriptions of the 30 (70) engineers and 70 (30) lawyers have been written. You will find on your forms five descriptions, chosen at random from the 100 available descriptions. For each description, please indicate your probability that the person described is an engineer, on a scale from 0 to 100.

In this study, participants were far more likely to predict that Jack was an engineer rather than a lawyer, even when the base-rate for lawyers was far greater than that for engineers. The usual interpretation of this finding is that participants’ judgments were driven by the representativeness heuristic. In other words, participants’ judgments were based on the degree of similarity between a sample and a population (the description above matches most closely the stereotype of an engineer), and they failed to appropriately use the provided base-rate information. The experimental instructions appear to be crucial here. Participants were told the description was based on personality tests administered by a panel of psychologists. Personality tests administered by psychologists clearly represent individuating information rather than base-rate information. Because the experimenter gives this information to participants as part of the experimental procedure, it seems reasonable that participants will assume that they should use this information in making their judgments. They assume the information is relevant (in Grice’s sense). And so they use it; their responses reflect this information and hence the putative operation of the representativeness heuristic. Importantly, several studies have demonstrated that if this task is changed in various ways, then the likelihood that participants will “inappropriately” use individuating information is decreased. For example, if the presumed relevance of the individuating information is undermined, people are far less likely to use it. Schwarz et al. (1991) tested this by telling some participants that the personality description, rather than being written by a psychologist based on a personality test, was compiled by a computer that randomly drew descriptive sentences bearing on the target person. Under these circumstances, participants were far less likely to use the individuating information, and instead they placed greater weight on the base rate information.

Pragmatics and social cognition   187

Other research has demonstrated how various features of the Kahneman and Tversky task contribute to the assumed relevance and hence use of the personality descriptions. In the Kahneman and Tversky task, participants are usually given five different descriptions of the target (hence the individuating information changes) but always the same base-rate information. Because of this, participants may assume that it is the descriptions rather than the base rate information that is most relevant for the task. If this is the case, then varying the base-rate information but leaving the description constant should result in greater use of base-rate information. Research suggests that it does (Schwarz et al. 1991, Experiment 2). The presentation order of the information in the Kahneman and Tversky study appears to be important as well. Participants first received the base-rate information followed by the individuating information. In general, the information a speaker considers the most important is presented last (e.g., the givennew contract, Clark 1985). If individuating information is provided after the base-rate information, then base-rate information must not be sufficient for performing the task; otherwise the experimenter would not have presented it. Consistent with this reasoning, Krosnick, Li, and Lehman (1990) found that participants were more likely to use base-rate information if it was presented after, rather than before, individuating information. This effect appears to be mediated by the assumed relevance of information that is presented last. The finding that information presented last is weighed most heavily disappears when participants are told the order with which the information was presented was randomly determined. Similar order effects have been reported with persuasive communications (Igou and Bless 2003). 3.2. Dilution effect Another well-documented judgmental phenomenon that may reflect the operation of conversational pragmatics is the dilution effect. The dilution effect refers to the reduction in the weight given to diagnostic information when that information is diluted with nondiagnostic information. The inclusion of nondiagnostic information obviously should not have an effect on judgments – yet it frequently does (e.g., Nisbett, Zukier, and Lemley 1981; Zukier 1982). For example, participants generated strong inferences about a student’s GPA when told he studied 3 hours versus 31 hours per week. However, participants modified these inferences when given information that was not relevant for the judgment (e.g., that the student played tennis three times a week). Obviously this information should not be used in making an inference about a student’s GPA.

188   Thomas Holtgraves and Blake M. Anderson Similar to demonstrations of the representativeness heuristic, participants in these studies are provided with irrelevant information in order to see if they inappropriately use it. Participants, on the other hand, assume the information given to them is relevant – and so they use it. As with the representativeness heuristic, there is experimental evidence demonstrating that mentioning the possibility that some of the information may not be relevant (and hence suspending the conversation maxim of relation) reduces the dilution effect (Igou and Bless 2005; Tetlock et al. 1996). However, the extent and nature of these effects is not entirely clear (see Kemmelmeier 2007; Igou 2007). There are other judgment effects that are quite similar to the dilution effect for which conversational pragmatics may play a role. These effects differ from the dilution effect in that the “irrelevant” information applies to stimuli other than the to-be-judged target. For example, Tormala and Petty (2007) investigated the effects of perceived knowledge on persuasion. In their studies, participants received information about two stimuli (e.g., two department stores). One of the two stimuli served as the attitude target, and they manipulated the amount of information participants received about the non-target stimulus. The more information that participants received about the non-target stimulus, the less they believed they knew about the target stimulus, and the less favorable was their attitude toward that target. Again, participants may have assumed the information provided about the non-target stimulus was relevant and so they used it in a comparison sense – reasonably believing the more they knew about the non-target stimulus, the less they knew about the target (relatively speaking). These effects extend to other features of a message such as source credibility. Tormala and Clarkson (2007) manipulated the credibility of the source of a message that was presented before a target message. Participants were more persuaded by the target message when it was preceded by a message from a low credibility source than when it was preceded by a message from a high credibility source. Again, participants assume all of the presented information is relevant (including the credibility of the source of the nontarget messgage) and so they attempt to incorporate the material into their judgments. The asymmetric dominance effect (more commonly known as the attraction effect) is also similar to the dilution effect in that information provided about one stimulus influences judgments about a different stimulus. For example, Kim and Hasher (2005) provided participants with a choice between a target brand and a competitor’s brand. The discount of the target was 25 % with a price of $ 45; the discount of the competitor was 15 % with a price of $ 20. In this situation approximately 50 % of the participants preferred the target and 50 % preferred the competitor. However, when a third brand was added to the

Pragmatics and social cognition   189

choice set, one that had a slightly higher discount (26 %) but a dramatically higher price ($ 100) than the target, then 70–80 % preferred the target. Participants are likely to believe the experimenter is asking them to compare the third option with the other two options, and because the third option is similar to the target in terms of the discount but much more costly, it is the target (rather than the competitor) that benefits. 3.3. Attibutions The study of attributions typically involves the presentation of information about a person and/or situation, followed by questions regarding the perceived cause of a particular outcome. As discussed in chapter 13 by Klaus Fiedler and Peter Freytag, language is commonly seen as having an effect on attributions. While they argue that these may be wired into linguistic categories, conversation pragmatics offers another way that language may play a role in the attribution process. As with the other effects discussed in this section, participants will tend to assume that any information provided by the experimenter is relevant for the task and should be used, even though it is precisely the experimenter’s hypothesis that such information is not relevant and hence should not be used. For example, the fundamental attribution error, or the tendency to “erroneously” infer internal states based on nondiagnostic information, can be interpreted in this way. Participants in the classic Jones and Harris (1967) study read an essay presumably written by another student and were then asked to judge the essay-writer’s true opinion regarding the issue. Some participants were told the essay writer had been instructed by the experimenter to adopt the position conveyed (forced-choice condition); other participants were told the writer’s position had been freely chosen (free-choice condition). The major finding was that participants inferred the author’s position to be in line with the position advocated in the essay, even when that position had been assigned and hence was not diagnostic of the writer’s true opinion. Note the similarity here to the Kahneman and Tversky (1973) study. Participants are given information to use in performing a specific task (judging the essay writer’s opinion), and they could reasonably assume that such information should be used (why else was it provided?). In other words, the experimenter’s contribution is assumed to be relevant. When the relevance of the information is lessened, people are less likely to use it. Wright and Wells (1988) replicated the Jones and Harris (1967) study but told some participants that the information given to them was randomly selected, and hence, it might not be relevant

190   Thomas Holtgraves and Blake M. Anderson for the judgment task they were to perform. In this condition, the fundamental attribution error was dramatically reduced. One of the most well-known findings in the attribution literature is the actor-observer difference. In a recent meta-analysis, Malle (2006) reported that, contrary to popular belief, the actor-observer difference is not as strong as commonly assumed. However, one experimental situation for which the actorobserver difference reliably occurs is when participants are given base-rate information that suggests the actor’s behavior is highly unusual. Observer participants readily use this information (because it is available and hence presumably relevant) and consequently they make actor attributions. Actors, of course, are less influenced by this information because it may be at odds with their phenomenal experience. One frequently cited explanation for the actor-observer difference is that actors and observers have different perspectives and this effect may also reflect conversation pragmatics. Taylor and Fiske (1975, Experiment 2) had participants attend specifically to a particular person during a ten-minute conversation between two people shown on a television. The results show that there was no significant relationship between the amount of information a participant could recall and the person they were assigned to watch, nor were there any significant differences in dispositional attributions as a function of perspective. However, participants reported the person that they were assigned to watch had a greater causal role in the conversation. It is quite likely that participants saw the person the experimenter assigned them to observe as the causal focus. 3.4. Eyewitness testimony Studies of eyewitness testimony have also yielded findings that may reflect the operation of conversational pragmatics. Specifically, Bell and Loftus (1988, 1989) have demonstrated that eyewitness testimony containing irrelevant trivial evidence influences judgments of witness credibility, along with guilty verdicts. In these experiments, participants read a court case that included eyewitness testimony, some of which contained trivial details. The results demonstrate that trivial evidence increases credibility and (when it is a prosecution witness) increases guilty verdicts. It is likely that participants assumed the trivial and irrelevant details to be relevant (why else was it provided in this context?) and important for their judgment and decisions. In terms of eyewitness identification, Wells and Bradfield (1998, 1999) argue that feedback given to eyewitnesses during the identification process can dis-

Pragmatics and social cognition   191

tort the eyewitnesses report regarding their experience witnessing the crime. In the experiments, participants watched a video with poor resolution that depicted an individual who would later commit a crime. After the video, participants received a sheet with five pictures of different people who may or may not have been the culprit. After the participant identified who he or she thought was the criminal, the experimenter responded “Good. You identified the actual suspect in the case.” or “Oh, You identified number __. The actual suspect is number __.” The participant then received a form that asked various questions, among these were different measures: participants’ certainty of their identification, view of the suspect, length of time seeing the suspect in the video, difficulty to identify, willingness to testify, eyewitness trustworthiness, and good basis for identification. The results show that participants who received positive feedback rated themselves higher on each of these measures. Wells and Bradfield (1998, 1999) argue that feedback causes eyewitnesses to reconstruct their experience. On the other hand, participants are active interlocutors in the conversation with the experimenter and they are employing the cooperative principle. If a participant is told that he or she did well in identifying the suspect, then the participant may assume that he or she had a good view, saw the face long enough to identify it, and so on. Wells and Bradfield (1998) even collected data before feedback in one experiment, and this data shows no difference. Therefore, the difference comes to play during conversation between participant and experimenter. Participants assume the experimenter is being relevant and an active participant in the conversation with his or her comment regarding the participant’s performance. 4. Implicit irrelevant information In the experiments described above, participants and experimenters differed in their pragmatic assumptions; participants assumed all information provided by the experimenter was relevant, an assumption that was not made by the experimenter. This type of divergence can occur in other ways as well. Much communication is implicit and will involve inferential processing on the part of the recipient; in Grice’s (1975) terms, recipients generate conversational implicatures. In this way, people actively construct interpretations that are consistent with the maxims of relation, quantity, and so on. One consequence of this process is that interpretations may be generated by recipients that were not necessarily intended by a speaker. This can occur in psychology experiments as participants interpret the experimenter’s communications and the experimental materials within the framework of Grice’s maxims.

192   Thomas Holtgraves and Blake M. Anderson 4.1. Question wording Questions can presuppose a certain state of affairs, and such presuppositions may be taken as relevant by a recipient and be used in formulating a response. For example, in a classic study, Loftus and Palmer (1974) examined the effects of leading questions (e.g., How fast was the car going when it smashed/hit the truck?) on memory for a crash that had been viewed earlier. Participants remembered the car traveling at a faster speed when the verb was “smashed” than when the verb was “hit.” Rather than reflecting a memory effect, this difference may reflect participants’ beliefs in the relevance of the verb (smashed vs. hit) in the question. In a very similar way, research participants may interpret self-report questions (e.g., personality measures, attitude measures) so as to maximize their presumed relevance. For example, Strack, Schwarz, and Wanke (1991) asked respondents to rate their happiness and their satisfaction with life, two questions that are obviously highly related. For some participants, the two items were presented together at the end of a questionnaire; for others, the life satisfaction item was presented as the first item in an ostensibly unrelated questionnaire. In the first situation the second question would appear to be redundant; it violates Grice’s maxim of quantity. Because of this, participants might have reasoned that the second question must mean something different from the first. Why else would it be asked? And responses to the two questions differed significantly when they were part of the same questionnaire. Such a concern does not arise when the two items are part of two separate questionnaires, and responses to the two items were almost identical in this situation. This finding has obvious implications for the common research practice of using multiple, closely-worded, items as a means of achieving high internal consistency. Respondents’ answers to these items might vary due to their attempt to maximize the relevance of each item. 4.2. Response scales There are similar implications for the construction of response scales. Respondents assume that response scales are relevant in the Gricean sense (i.e., they are a meaningful component of the questionnaire). Hence, the numerical value of rating scales may be used by respondents to interpret the meaning of the scale’s response options. For example, in response to questions about how successful they have been in life, respondents are more likely to endorse a response option on the bottom half of the scale if the scale values range from

Pragmatics and social cognition   193

0 to 10 than if they range from –5 to +5. In the former case the pairing of the endpoint label “not at all successful” is interpreted as the absence of success when paired with 0, but as the presence of failure when paired with -5 (Schwarz et al. 1991). Questions designed to assess frequency estimates are also susceptible to these types of effects; respondents use the provided response alternatives as a frame of reference. For example, respondents were far more likely (37.5 % vs. 16.7 %) to indicate watching television for 2.5 hours or more a day when high frequency alternatives were provided (.5 hour increments starting at 2.5) than when low frequency alternatives were provided (.5 hour increments starting at .5) (Schwarz et al. 1985). Even open-ended questions may be effected by conversational maxims as respondents are not likely to report information that is self-evident or that they do not perceive as relevant (Schwarz 1996). For certain content domains the importance of response formats is substantial. For example, the results of research on gender differences in jealousy consistently vary as a function of response format. The predicted gender differences (males are more bothered by sexual infidelity and females are more bothered by emotional infidelity) consistently occur when a forced choice measure (choose either sexual or emotional infidelity as the most distressing) is used, but not when continuous rating scales are used (e.g., DeSteno et al. 2002). Continuous scales, unlike the forced choice format, legitimize the possibility of neither type of infidelity being more distressing. An easy and interesting way to test this would be to use a forced choice format but provide a “neither” or “equal” option for respondents. 4.3. Shifting standards model The shifting standards model is an explanation for the shifts in ratings that occur when participants rate individuals of stereotyped social groups on different types of scales, particularly subjective and objective scales. Biernet, Manis, and Nelson (1991) argue that participants perceive the different scales as asking them to rate stimuli according to different criteria, or against the backdrop of different groups. In a series of experiments, participants were presented a set of photographs of different people with three scales (one objective and two subjective) following each picture. When rating on an objective scale (e.g., reporting weight in pounds or height in feet and inches), participants did not appear to use stereotypic information. However, when rating on the subjective scales (e.g., rating weight or height on a Likert

194   Thomas Holtgraves and Blake M. Anderson scale), participants did use stereotypic information. With these scales different ratings were obtained with anchors showing overall average (e.g., average height) and group average (e.g., average women height). As previously mentioned, Strack et al. (1991) showed that similar questions presented next to each other will be interpreted as asking about different things to avoid redundancy. Biernet et al. (1991) presented participants with three scales that appeared together, with each scale rating the same phenomena (i.e., height, weight, etc.). Participants may have interpreted each scale as requesting different information not only because the scales differed, but to avoid reporting redundant information. 4.4. Conjunction fallacy There is another classic judgmental heuristic documented by Tversky and Kahneman (1983) – also assumed to reflect the operation of representativeness – that might reflect conversational inferencing. The basic situation involves the presentation of information about a person followed by questions regarding the probability of certain outcomes. The most well-known example is the Linda problem which (in some versions) is presented as follows: Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student she was deeply concerned with issues of discrimination and social justice and also participated in antinuclear demonstrations.

Participants are then asked which of the following two alternatives is more probable: Linda is a bank teller. Linda is a bank teller and is active in the feminist movement.

Participants are far more likely to endorse the second alternative, thereby demonstrating a fallacy in so far as the probability of a conjunction can not exceed the probability of any of its constituents. Note, however, that there is a potential implicature here. The alternative “Linda is a bank teller” could be viewed as implicating that Linda is a bank teller and not a feminist. And in fact when participants do make this interpretation they are more likely to judge the conjunction more probable (Dulany and Hilton 1991). Moreover, this interpretation is more likely when participants are provided with a more extensive description of the target (Dulany and Hilton 1991). In this case the likely inference is that Linda is a bank teller and not a feminist because if she was a feminist, that information would have been provided.

Pragmatics and social cognition   195

5. Pragmatics and judgments outside the lab So far we have focused on the reasoning of participants in laboratory experiments. Our fundamental assumption has been that participants view the information provided by an experimenter to be task relevant (erroneously from the experimenter’s view). It is important to note that although this effect is quite pronounced in experimental settings, it is not unique to them. In fact, many real-life (non-laboratory) judgments may be influenced in a similar way. That is, attitudes, attributions, perceptions, and so on are often elicited in a conversational context. People are asked their opinions, and whether they believe something is likely to occur, and so on. The manner in which the question is asked, and the speaker’s beliefs about the state of the questioners knowledge, can influence the answers that are given. For example, the attributions that people give in a conversational setting are sometimes based on assumptions regarding the recipient’s knowledge (Slugoski et al. 1993); people tend to focus on and identify those causes of which the recipient is probably not aware. People do not usually ask for information they already possess. Similarly, people are more likely to convey stereotype-consistent information if they believe the recipient does not possess that particular information (Lyons and Kashima 2003). Everyday judgments then, are not simply the result of intrapersonal cognitive processes, but frequently the result of an interaction between those processes and certain relevant, pragmatic principles regarding communication. This point can be taken further because Grice’s conversational maxims are very basic principles that apply to actions in general, and not just conversational interaction. Even when our thoughts are not elicited in a conversational context, the manner in which they are formulated, the direction that they take, may be influenced by the principles of relevance, quality, quantity, and so on. For example, the intrapersonal process of repression may be viewed as changing the topic of one’s internal thought processes; in effect, a relevance violation (cf. Billig 1999). Notably, Grice’s theory applies to all action, not just linguistic communication. That is, the cooperative principle and corresponding maxims are assumed to underlie rational interaction of any sort, verbal and nonverbal. Imagine, for example, two people cooking dinner together. In order to do this successfully they need to cooperate, and their cooperation involves making their contributions to the joint endeavor relevant (e.g., if asked to stir the sauce one should do it immediately, not later), truthful, and so on. This raises the interesting possibility that reports of perceptions of objects might be influenced by assumptions regarding their relevant functions. For example, research on functional fixed-

196   Thomas Holtgraves and Blake M. Anderson ness has demonstrated that participants are typically unable to perceive objects in novel ways (Maier 1931; Duncker 1945; Chrysikou and Weisberg 2005). When participants are asked to solve a problem and are given certain objects in this context they view those objects in terms of their assumed “typical” uses. Consistent with this reasoning, Glucksberg and Weisberg (1966) showed that when target objects are given verbal labels, people had more trouble overcoming functional fixedness if the label was specific rather than general (e.g., a matchbox labeled as a container versus as a box). 6. Summary and conclusions Over the past 20 years, an area of research has developed regarding the role of pragmatics, primarily conversational implicature, in reasoning. Prior to this development, most research on judgment and reasoning centered on internal, intrapersonal processes (inference, attention, memory, etc.). A common theme running through this literature was that human judgment and reasoning are often nonoptimal. For example, people fail to use base rate information (Kahneman and Tversky 1973), they ignore the effects of situations on people’s behavior (Nisbett and Ross 1980), they use nondiagnostic information in their decisions (the dilution effect; Nisbett et al. 1981), they are susceptible to leading questions (Loftus and Palmer 1974), and so on. However, it is clear that these biases reflect, at least partly, participants reasoning about the materials and the instructions. This does not mean that all or even most social cognitive phenomena are the result of conversation rules. And in fact at least in some cases the potential influence of these principles might be overstated (e.g., the dilution effect; see Kemmelmeier 2007). But clearly the responses provided by participants in experimental contexts are sometimes influenced by their beliefs about how conversations work. The research demonstrating the role of conversation pragmatics in research contexts obviously has methodological implications. At a very general level, researchers should try to consider the pragmatic assumptions that participants might make. Pretesting could be conducted to test for the presence of such assumptions. If the experiment will contain irrelevant information, the experimenter should take extra measures to assure that this information is perceived as irrelevant. Otherwise, the information can hardly be called irrelevant. An alternative would be to include a condition that attempts to manipulate the perceived relevance of the critical information (e.g., Schwarz et al. 1991). Some of the results have specific methodological implications. For example, research suggests that the practice of including multiple, closely-worded items

Pragmatics and social cognition   197

to create a reliable measure of a construct might result in participants varying their responses to those items because of their perceived redundancy. Similar to most research in this area, we have used Grice’s (1975) model of conversational cooperativeness as the framework for understanding conversational reasoning. However, there are two potential problems with the use of Grice’s model. First, there is concern about whether Grice’s maxims are crossculturally valid, a point never raised by Grice, although he seems to have pitched his theory, implicitly at least, as being crossculturally valid. Some researchers, however, have noted that the maxims may not be valid in certain cultures. In a wellknown paper, Keenan (1976) argued that people in Malagasy routinely withhold information from one another (information is a culturally prized commodity), an action that is in clear violation of the quantity maxim. As a result, violations of this maxim do not usually result in conversational implicatures (as presumably would be the case in most other cultures). There is a more general criticism here. It has been suggested that Grice’s view, with its emphasis on individual autonomy (i.e., the individual is the source of his or her utterances) has relevance only in Western cultures (Fitch and Sanders 1994; see also Patent, this volume). It is possible that any maxims people follow in communicating cooperatively will vary over cultures; reflecting, perhaps, differences in what is regarded as rational interaction. Hence, it is possible that Grice’s framework is relevant only in Western cultures. Second, conversation inferencing is guided not only by conversation maxims but also by certain interpersonal considerations, interpersonal motives that influence when maxims are violated and the types of inferences that are generated (Holtgraves, 1998; 1999). It seems likely conversation maxims and interpersonal concerns will operate in conjunction with one another. For example, Hodges and Geyer (2006) recently articulated a values-pragmatic account of the “conformity” experiments of Asch. On this view, participants in the Asch experiments are engaged in a conversation involving the other participants and the experimenter. Because it is a conversation, there are demands for truth (the Gricean maxim of quality) as well as pressure to be attentive to one another’s face (Goffman 1967). Participants must balance these conflicting pressures in order to stay in the conversation (Holtgraves 2002). And they do this by varying their responses over trials – most often giving their own view (truth) but occasionally going along with the majority – a symbolic nod to their presence and importance. What is crucial here is the temporal dimension – crucial because there are multiple trials and so participants can balance their competing concerns – a possibility that would not exist with a single trial.

198   Thomas Holtgraves and Blake M. Anderson References Bell, Brad E., and Elizabeth F. Loftus Degree of detail of eyewitness testimony and mock juror judgments. 1988 Journal of Applied Social Psychology 18: 1171–1192. 1989 Trivial persuasion in the courtroom: The power of (a few) minor details. Journal of Personality and Social Psychology 56: 669–679. Biernet, Monica, Melvin Manis, and Thomas E. Nelson. 1991 Stereotypes and standards of judgment. Journal of Personality and Social Psychology 60: 485–499. Billig, Michael Freudian Repression: Conversion Creating the Unconscious. Cam1999 bridge, UK: Cambridge University Press. Chrysikou, Evangelia G., and Robert W. Weisberg Following the wrong footsteps: Fixation effects of pictorial examples 2005 in a design problem-solving task. Journal of Experimental Psychology: Learning, Memory, and Cognition 31: 1134–1148. Clark, Herbert H. 1985 Language use and language users. In The Handbook of Social Psychology. Vol. 2, 3d ed., G. Lindzey and E. Aronson (eds.), 179–232. Reading, MA: Addison-Wesley. DeSteno, David, Monica Y. Bartlett, Julia Braverman, and Peter Salovey 2002 Sex differences in jealousy: Evolutionary mechanism or artifact of measurement? Journal of Personality and Social Psychology 83: 1103– 1116. Dulany, Don E., and Denis J. Hilton Conversational implicature, conscious representation, and the conjunc1991 tion fallacy. Social Cognition 9: 85–110. Duncker, Karl On problem-solving (translated by L.S. Lees). Psychological Mono1945 graphs 58, No. 270. Fitch, Kristine L., and Robert E. Sanders 1994 Culture, communication, and preferences for directness in expression of directives. Communication Theory 4: 219–245. Glucksberg, Sam, and Robert W. Weisberg 1966 Verbal behavior and problem solving: Some effects of labeling in a functional fixedness problem. Journal of Experimental Psychology 71: 659–664. Goffman, Ervin 1967 Interaction Ritual: Essays on Face to Face Behavior. Garden City, NY: Doubleday. Grice, Herbert Paul Logic and conversation. In Syntax and Semantics 3: Speech Acts, P. 1975 Cole and J. Morgan (eds.), 41–58. New York: Academy Press.

Pragmatics and social cognition   199 Hilton, Denis J. 1995 The social context of reasoning: Conversational inference and rational judgment. Psychological Bulletin 118: 248–271. Hodges, Bert H., and Anne L. Geyer A nonconformist account of the Asche experiments: Values, pragmatics, 2006 and moral dilemmas. Personality and Social Psychology Review 10: 2–19. Holtgraves, Thomas Interpreting indirect replies. Cognitive Psychology 37: 1–27. 1998 1999 Comprehending indirect replies: When and how are their conveyed meanings activated? Journal of Memory and Language 41: 519–544. 2002 Language as Social Action: Social Psychology and Language Use. Mahwah, NJ: Erlbaum. Igou, Eric R. 2007 Additional thoughts on conversational and motivational sources of the dilution effect. Journal of Language and Social Psychology 26: 61–68. Igou, Eric R., and Herbert Bless 2003 Inferring the importance of arguments: Order effects and conversational rules. Journal of Experimental Social Psychology 39: 91–99. 2005 The conversational basis for the dilution effect. Journal of Language and Social Psychology 24: 25–35. Jones, Edward Ellsworth, and Victor A. Harris The attribution of attitudes. Journal of Experimental Social Psychology 1967 3: 1–24. Kahneman, Daniel, and Amos Tversky On the psychology of prediction. Psychological Review 80: 237–251. 1973 Keenan, E.L. 1976 Towards a universal definition of subject. In Subject and topic, Charles N. Li (ed.), 303–333. New York: Academic Press. Kemmelmeier, Markus 2007 Does the dilution effect have a conversational basis? Journal of language and Social Psychology 26: 48–60. Kim, Sunghan, and Lynn Hasher 2005 The attraction effect in decision making: Superior performance by older adults. The Quarterly Journal of Experimental Psychology 58A: 120– 133. Krosnick, Jon A., Fan Li, and Darrin R. Lehman 1990 Conversational conventions, order of information acquisition, and the effect of base rates and individuating information on social judgment. Journal of Personality and Social Psychology 59: 1140–1152. Loftus, Elizabeth F., and J.C. Palmer 1974 Reconstruction of automobile destruction: An example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behavior 13: 585–589.

200   Thomas Holtgraves and Blake M. Anderson Lyons, Anthony, and Yoshihisa Kashima 2003 How are stereotypes maintained through communication? The influence of stereotype sharedness. Journal of Personality and Social Psychology 85: 989–1005. Maier, Norman R.F. 1931 Reasoning in Humans: II. The solution of a problem and is appearance in consciousness. Journal of Comparative Psychology 12: 181–194. Malle, Bertrand The actor-observer asymmetry in attribution: A (surprising) meta-anal2006 ysis. Psychological Bulletin 132: 895–919. Nisbett, Richard E., and Lee Ross Human Inference: Strategies and Shortcomings of Social Judgment. 1980 New York: Prentice Hall. Nisbett, Richard E., Henry Zukier, and Ronald E. Lemley The dilution effect: Nondiagnostic information weakens the impact of 1981 diagnostic information. Cognitive Psychology 13: 248–277. Ochs Keenan, Elinor 1976 The universality of conversational implicature. Language in Society 5: 67–80. Schwarz, Norbert 1996 Cognition and Communication: Judgmental Biases, Research Methods, and the Logic of Conversation. Mahwah, NJ: Erlbaum. Schwarz, Norbert, Fritz Strack, Denis Hilton, and Gabi Naderer Base rates, representativeness, and the logic of conversation: The con1991 textual relevance of “irrelevant” information. Social Cognition 9: 67– 84. Schwarz, Norbert, Hans J. Hippler, Brigitte Deutsch, and Fritz Strack Response scales: Effects of category range on reported behavior and 1985 subsequent judgments. Public Opinion Quarterly 49: 388–395. Slugoski, Ben R., Mansur Lalljee, Roger Lamb, and Gerald P. Ginsburg 1993 Attributions in conversational context: Effects of mutual knowledge on explanation giving. European Journal of Social Psychology 23: 219– 238. Sperber, Daniel, and Dierdre Wilson 1986 Relevance. Cambridge, MA: Harvard University Press. 1995 Relevance. 2d ed. Cambridge, MA: Blackwell Publishers. Strack, Fritz, Norbert Schwarz, and Michaela Wanke Semantic and pragmatic aspects of context effects in social and psycho1991 logical research. Social Cognition 9: 111–125. Taylor, Shelly E., and Susan T. Fiske Point of view and perceptions of causality. Journal of Personality and 1975 Social Psychology 32: 439–445. Tetlock, Philip E., Jennifer S. Lerner, and Richard Boettger 1996 The dilution effect: Judgmental bias, conversational convention, or a bit of both? European Journal of Social Psychology 26: 915–934.

Pragmatics and social cognition   201 Tormala, Zakary L., and J.J. Clarkson Assimilation and contrast in persuasion: The effects of source credibil2007 ity in multiple message situations. Personality and Social Psychology Bulletin 33: 559–571. Tormala, Zakary L., and Richard E. Petty Contextual contract and perceived knowledge: Exploring the implica2007 tions for persuasion. Journal of Experimental Social Psychology 43: 17–30. Tversky, Amos, and Kahneman, Daniel Extensional versus intuitive reasoning: The conjunction fallacy in prob1983 ability judgment. Psychological Review 90: 293–315. Wells, Gary L., and Amy L. Bradfield “Good, you identified the suspect”: Feedback to eyewitness distorts 1998 their reports of the witnessing experience. Journal of Applied Psychology 83: 360–376. 1999 Distortions in eyewitness’ recollections: Can the postidentificationfeedback be moderated? Psychological Science 10: 138–143. Wright, Edward F., and Gary L. Wells 1988 Is the attitude-attribution paradigm suitable for investigating the dispositional bias? Personality and Social Psychology Bulletin 14: 183–190. Zukier, Henry 1982 The dilution effect: The role of the correlation and the dispersion of predictor variables in the use of nondiagnostic information. Journal of Personality and Social Psychology 41: 1163–1174.

Chapter 8 The creative construction of social orientation: Situated positioning with English as a lingua franca Andreas Langlotz

1. Introduction It is a truism that English has become the uncontested global lingua franca of the 21st century (Crystal 2003: 1−28; Meierkord 2006: 165; Jenkins 2003: 14−21). For non-native English-speaking professionals this global presence of English affects their options for communicating social meaning deeply. When using English as a lingua franca, they have to accommodate the linguistic construction of social information to a code that does not reflect their language-specific, local linguistic cues for identity expression (Agha 2007: Ch. 3), nor convey the culturespecific discursive practices for social positioning (Davies and Harré 1990). Even worse, EFL speakers tend to control the lingua franca merely at a level that is considerably below native linguistic and pragmatic competence (Meierkord 2006: 166−167; see also Meierkord and Knapp 2002). As a result, their registers and cues for conveying a positive social image and for engaging in authentic social action become de-codified, de-localized, and fairly limited (Blommaert 2004: Ch. 4). Hence, the linguistic management of social categorization in lingua-franca interaction becomes a considerable processing challenge. The challenges of employing English as a foreign language today are not only related to lingua franca usage. International English has been increasingly diversifying into novel communicative niches of cross-cultural business communication (Crystal 2003: Ch. 4). One of these recent business communication genres is international tourist information. Tourism is a prime example of a globalized service industry in the 21st century. In front-desk interactions (among other forms of communication: call centres, websites, brochures, etc.) the notion of service is performed and negotiated as a joint discursive activity (Chappell 2005). Obviously, tourist information service, apart from its transactional function of providing orientation, is also very centrally characterized by

204   Andreas Langlotz conveying a positive social impression. Tourists should feel comfortable; they expect to be helped and catered for in a friendly and competent way. Thus, the marketed image of a place or country is at stake when values such as hospitality, authenticity, friendliness, or quality become subject to immediate experience in tourist information encounters. However, to establish social orientation in this communicative context, tourists and tourist information officers, of course, cannot rely on previous interpersonal experience. Also, the institutionalized nature of tourist information transactions limits the space for relational talk considerably (Koester 2004: 1406–1407). For the sake of communicating an authentic and original local self-image, tourist information officers are therefore forced to manipulate the discursive landscape of information transaction creatively to develop contextually grounded relational meaning. Hence, the interplay of language and social cognition in this challenging lingua franca context becomes intriguingly complex. This article focuses on the creative socio-cognitive processing strategies for the communication of social orientation by tourist information officers who use English as a lingua franca. The cognitive contextualization strategies that allow the interlocutors to position themselves in a local role and to convey a situated identity in the absence of a rich repertoire of corresponding linguistic cues are exemplified by means of a qualitative analysis of two striking examples from front-desk interactions at a Swiss tourism board.1 The data are primarily employed to illustrate strategies of dealing with the social challenges in cross-cultural tourist information service. To theorize the complex relationship between language, social cognition, and social action, the central interest of this paper is to develop a socio-cognitive model of situated social sense-making. It is an attempt to explain how entrenched social representations can be creatively adapted to create contextspecific social categories; these situated social categories are based on stable social models. But these stable categories can be played with to construct an ad hoc social image (see also Langlotz 2008). The present model primarily combines conversation analytical methods with Fauconnier’s mental space theory (1994, 1997) and Barsalou’s (2005) model of situated conceptualization; it also shares close affinity with the work of Clark (1996). Accordingly, it is argued 1 The data for the present analysis were collected for the research project on Languages, identities and tourism: Towards an understanding of social and linguistic challenges in Switzerland in the context of globalization (2005–2008) funded by the Swiss National Fund (project number 108608) as part of a national research program on Language Diversity and Linguistic Competence in Switzerland (http:// www.nfp56.ch/). The grant is jointly held by Ingrid Piller and Alexandre Duchêne.

Situated positioning with English as a lingua franca   205

that the creative linguistic construction of social orientation can only be fruitfully modelled by closely combining the social with the cognitive dimension of sense-making (Clark 1996: 24). Creative social sense-making thus becomes a situated socio-cognitive practice that is based on conventionally encoded social concepts, but modifies them through linguistic variation in context. As will become obvious in the analysis, the corresponding linguistic strategies centrally involve irony and indirectness. Therefore, this paper also shares a number of central ideas with Kreuz and Caucci’s article on verbal irony in this volume. The article is divided into two sections. First, the theoretical model for the linguistic construction of social orientation is established. In the second part, this model is then exemplified by applying it to the tourism data. 2. The linguistic construction of social orientation 2.1. Social orientation and language Social cognition describes the human cognitive capacity of perceiving, representing, and processing social information, i.e. information about ourselves, the people and groups that we encounter in our daily life, about the social contexts in which we encounter them, as well as our attitudes towards them (Kunda 1999: 1). It is a core premise of research in social cognition that human beings make sense of their social environments by categorizing and conceptualizing them (Kunda 1999: Ch. 2; Bless et al. 2004: Ch. 2). To construct a coherent social reality from the social cues that they perceive in their environment, social cognizers activate entrenched social representations that are stored in long-term memory (Smith and Mackie 2005: 14–15). Such representations include concepts for personality traits, categories of people and groups such as students or linguists, as well as social events and constellations including, for instance, parties, funerals, or romantic relationships. These cognitive representations work as cognitive models for social information that are used like maps to understand the social environment. A cognitive model works as a complex “structured representation of conceptual categories” (Kövecses 2006: 64). For instance, to understand romantic affection between human beings, social cognizers can employ the love model.2 The model specifies two social roles, the lovers, as well as their intimate love relationship. Apart from this crude skeletal structure, the model for this complex social constellation also contains knowledge about the typical behaviour of lovers, the 2 In line with the general cognitive-linguistic convention, I use small capitals to refer to conceptual structures.

206   Andreas Langlotz ways they interact, and an evaluative judgement of love as a positive social relationship. Obviously, the social constellations that serve as input to such social cognitive models and corresponding categorization events are social schemas themselves, i.e. products of social engagement. They consist of culture-specific social events, such as cocktail parties, family dinners, business meetings, and many more. Moreover, they include the conventionalized role (stereo)types and social practices that people establish to fellow social players in these encounters, e.g. acting as hosts, friends, newcomers, and so forth. Therefore, our cognitive construct of surrounding social reality is itself primarily based on historically grown social inventions and conventions (Blommaert 2004: Ch. 6). For two reasons, the social nature of these stimuli therefore makes it very difficult to produce clear-cut social judgments. This renders social cognition different from the representation of general world knowledge for objects, spatial relations, and so forth. First, people, relationships, and groups may change very rapidly. Therefore, unlike objects, their categorization and conceptualization must also be subject to dynamic adaptations. Second, many social attributes are not visible and cannot be directly manipulated. Thus, while we can see the colour, size, or form of an object, social attributes such as trustworthy, humorous, loyal, etc. cannot be directly perceived but must be inferred from associated social cues. Due to this fact, social orientation often cannot be created through fast categorization. Rather, it depends on a great deal of constructive effort. In Bless et al.’s (2004: 9) words: To make social judgements, individuals must go beyond the information given […] far more than when judging inanimate objects. […], it is the social judgements that require far more inferential processes, and it is precisely this large proportion of constructive processes that makes social cognition a specific and unique process.

In accord with these core assumptions, social orientation is a constructive cognitive effort that applies to dynamically changing and socially constructed input. From an interactional sociolinguistic perspective such as conversation analysis, this skeletal account of social sense-making still is not fully satisfactory because it lacks one central element – language (see, e.g., Drew 2005; Edwards 2005; Schegloff 1999). Rather than merely categorizing and processing social input cognitively, human beings are often deeply engaged in socio-communicative practices in the very process of constructing and manipulating their social environment (Clark 1996: Ch. 1). In specific social encounters, a cocktail party for instance, human beings perform social activities and actions that change the social environment immediately. Very centrally, people perform linguistic actions such as greetings, small talk, flirting, flattering, joking, or winding each other up. Thus, they produce utterances and speech acts, and interact in conversations. It is this sociolinguistic engagement at the micro-level

Situated positioning with English as a lingua franca   207

of communicative interaction that turns our arenas of social encounter into highly dynamic and potentially elusive environments. To find orientation in them, social cognizers must produce, attend to, and process linguistic signals as social cues that may have an immediate impact on their construct of social reality (Pomerantz and Mandelbaum 2005: 151–152). In other words, linguistic signals and actions work as the central linking element between social cognition, social convention, and social process. Through the ingenious manipulation of linguistic structures, interactors become able to creatively adapt to the social environment as well as to adapt this environment to their purposes (Langlotz 2008). Tajfel (1969: 81, quoted in Bless et al. 2004: 8) regards this ability as an enormous evolutionary advantage: “[T]he greatest adaptive advantage of man [sic!] is his capacity to modify his behaviour as a function of the way in which he perceives and understands a situation”. Since linguistic action is both cognitive action (linguistic structures must be categorized and processed to be understood) and social action (utterances are produced to influence interactors, including ourselves, communicatively), the creative linguistic construction of social orientation must be understood as a socio-cognitive process by definition (Figure 1). In short, the fundamental role of language for social sensemaking must be scrutinized with a view to both its social and cognitive nature. LANGUAGE: linguistic cues and actions SOCIAL COGNITION

SOCIAL BASIS

conventionalized social practices and processes to interact with others in specific social events

SOCIAL ENVIRONMENT SOCIAL INPUT

social cues about: ourselves, others, groups, social events, etc.

categorization and conceptualization of social input on the basis of cognitive models for

ourselves, others, groups, social events, etc.

SOCIAL ORIENTATION: adaptation of and adaptation to social environment

Figure 1. Social orientation

Language, unquestionably, is one of the prime means to communicate social meaning. Nevertheless, it is extremely difficult to specify and categorize precisely the linguistic structures and processes that underlie the socio-cognitive

208   Andreas Langlotz construction of social meaning (a very good and comprehensive model is offered by Agha 2007). One central problem for social sense-making is the context-dependent nature of meaning generation (see, e.g., Verschueren 1999: Ch.  2). Rather than triggering social meaning directly, the same linguistic structures may evoke alternative social implicatures depending on their specific contextual placement and interpretation. (1)

I love you so much, darling.

(1) may evoke very different social meanings depending on who says it to whom in which context. Of course, this utterance can be interpreted as a romantic speech act that is addressed from one lover to the other. In this context, the verb love and the quantitative adverbial so much are intended to be understood literally, which is also true for the endearing address form darling. Note that these lexical cues are of course strongly supported by phonological and supra-segmental features as well as non-verbal cues in spoken interactions. Along the lines of lover discourse and Davies and Harré’s model of positioning, the relational meaning of (1) thus other-positions the interlocutors relative to the cognitive model love (Figure 2): Positioning, … , is the discursive process whereby selves are located in conversations as observably and subjectively coherent participants in jointly produced story lines. There can be interactive positioning in which what one person says positions another. And there can be reflexive positioning in which one positions oneself. … . One lives one’s life in terms of one’s ongoingly produced self, whoever might be responsible for its production. (Davies and Harré 1990, quoted according to Wetherell et al. 2001: 264)

In accord with this social placement relative to the “story-line” of love, (1a) can therefore be expected as a conversational sequence. (1a) I adore you too, honey. Alternatively however, (1) could also be read as an ironic provocation. For example, one could imagine it to be uttered by a football player to the referee after some questionable arbitration. In such a situation, love would obviously adopt an opposite semantic value and darling would not work as an endearing form of address but rather be intended to belittle the referee and his/her competence. Here, the social cognitive model of love would be evoked as a counter-world to the actual reality on the football pitch. By interpreting the actual constellation on the pitch – to be construed according to the cognitive model of disputed arbitration – relative to this possible world of love, the ironic semantic effect emerges; it positions the player as a provoker and the arbiter as an incom-

Situated positioning with English as a lingua franca   209

petent wretch. Accordingly, the sequence in (1b) could represent a likely reaction by the referee. (1b) Do you want to take an early shower?

Categorize, construe and position

COGNITIVE MODEL FOR SOCIAL CONSTELLATION: LOVE Coupled with linguistic cues and actions  conventionally worded cognitive model of lovers’ social relationship

Categorize, construe and position

Linguistic cues activate

Lover A

I love you so much darling

Lover B

SOCIO-COMMUNICATIVE CONTEXT

Figure 2. Languaging social orientation

The generation of social orientation thus depends on the construction of cognitive models (love, disputed arbitration) relative to which social encounters are categorized and assessed. These cognitive models are evoked by linguistic cues (love, so much, darling), which are embedded in and constitute linguistic actions (the speech act: I love you so much, darling), (Figure 2). In what follows, this cognitive-linguistic dimension of social orientation is scrutinized. 2.2. Dynamic and situated social conceptualization through cognitive simulation across mental spaces To refine the cognitive dimension of the linguistic construction of social orientation, the generation of social meaning can be further elaborated by applying Fauconnier’s mental space theory to the processing of social categories (Fauconnier 1994, 1997).3 The central assumption of this theory is that linguistic 3 Note that Fauconnier’s theory was originally developed as a cognitive linguistic counter-model to truth-conditional semantics. Therefore, this heuristics was not

210   Andreas Langlotz units and structures do not contain meaning. Rather, they provide cues for the context-specific cognitive generation of meaning: A linguistic expression E does not have a meaning in itself; rather it has a meaning potential, and it is only within a complete discourse and in context that meaning will actually be produced. The unfolding of discourse brings into play complex cognitive constructions. They include the setting up of internally structured domains linked to each other by connectors; this is effected on the basis of linguistic, contextual, and situational clues. (Fauconnier 1997: 37)

In accord with this core assumption (see also Croft and Cruse 2004: 97–101), linguistic meaning becomes a dynamic phenomenon by definition. It is only fully developed in actual discourse and resides in conceptual representations, mental spaces, and the connections or mappings between them that are mentally constructed when speaking and hearing. Mental spaces are defined as “partial structures that proliferate when we think and talk, allowing a finegrained partitioning of our discourse and knowledge structures” (Fauconnier 1997: 11). More specifically, mental spaces must be seen as abstract conceptual packages that are evoked mentally to function as cognitive contexts relative to which cognizers become able to track reference. Imagine that the referee in our fictional conversation uttered (2) as a reaction to the football player’s provocation: (2)

If I were your darling, I would teach you some manners!

According to Fauconnier (1997: Ch. 2), we can only make sense of such counterfactuals by opening two mental spaces: a) the base space: the actual reality space of the conversation – structured by the mental model of the football game – and b) the target space: the hypothetical love-scenario in which the referee adopts the role of the player’s darling. This target space is evoked by the space builder if, the linguistic element that triggers the construction of the hypothetical world (Fauconnier 1997: 40, 106–113). Both spaces are internally structured by cognitive models (for a comprehensive overview see Croft and Cruse 2004: Ch. 2). The base space recruits conceptual knowledge about interactors, processes, roles and relations via the model football match, while the target space is structured by the love-model (Figure 3). This knowledge is evoked by the address term darling. With regard to social meaning, both models propose different role relationships. In the base space the referee acts as the superior arbiter of the match originally designed to explain social categorization but to solve a number of fundamental problems at the semantics-pragmatics interface.

Situated positioning with English as a lingua franca   211

whose decisions must not be questioned. In the love model, the default slots are characterized by the two members of the loving couple. To keep track of the hypothetical shift of reference from the actual conversation to the hypothetical context, the cognizers must build mental connections between these two spaces. Thus, the speaking referent referee (activated through the person deictic I) must be mapped onto the darling-slot (your darling) in the hypothetical space. However, unlike the default social relationship in the loving couplemodel, in the hypothetically evoked scenario the referee-as-darling adopts the didactic role of a teacher (I would teach you some manners). Obviously, this ironically constructed anti-lover-scenario is created by mapping the role relationship that is implied in the actual conversational action of (2) onto the target space: the referee reprimands the football player in the same way as he would do if he were his lover. It is this creative play with mentally evoked role models and stereotypes that leads to the creative social self-positioning and other-positioning by the speaker. The arbiter re-positions himself as being superior to the football player by rejecting the provocatively suggested role model contained in the love-scenario. role mapping

A

B

Base space: FOOTBALL MATCH-model A = ref B = player → = reprimand

* A

B

Target space: LOVE –model Triggered by space builder: if A = darling B = player default → = love evoked →* = reprimand

Figure 3. Creative role construction across mental spaces

2.2.1. Situated social conceptualization This mental space account of creative social categorization is compatible with Barsalou’s (1999, 2005) theory of situated conceptualization. Rather than seeing concepts as merely stored in a relatively autonomous semantic memory,

212   Andreas Langlotz Barsalou (2005: 626) also conceives them as dynamic constructs that “can be viewed as an agent-dependent instruction manual that delivers specialized packages of inferences to guide an agent’s interactions with particular category members in specific situations”. Thus, conceptual categories are not only represented as static pieces of memorized knowledge but can also emerge as specialized goal-specific and situation-specific constructs that help cognizers derive ad hoc inferences about the people and phenomena they are dealing with. In this sense, Barsalou’s dynamic model of conceptual representation and categorization differs from more traditional, static views of concepts (see Smith and Samuelson 1997). In accord with Barsalou’s dynamic model of categorization, the mapping of social roles and relations across different mental spaces in (2) involves the speaker’s attempt to construct a situated social conceptualization. This ad hoc conceptualization does not only derive its meaning from the entrenched underlying cognitive models and the corresponding social representations (football arbitration, love); rather, the actual categorization of the relationship between the two opponents is derived from simulating a hypothetical social world through the active and “online” mapping of information between these cognitive models. Following Barsalou’s words, the creative play with the hypothetical social roles (darling, didactic master) thus serves as an “agentdependent instruction manual” that evokes “specialized packages” of social inferences and thus guides the referee’s engagement with the provocative football player. The dynamic and context-specific cognitive simulations that constitute a situated conceptualization depend on the recruitment of various stable elements that Barsalou calls simulators: “[s]imulators integrate information across a category’s instances, whereas simulations are specific conceptualizations of the category” (Barsalou 2005: 624). In other words, a simulator works as a category-type, whereas a simulation is a context-specific instantiation or reenactment of this type. Accordingly, a simulator is regarded as a relatively stable mental representation of perceptual attributes that are structured into a frame (Barsalou 1999: 590–592). The frame integrates this perceptual information into a coherent cognitive model (see Evans and Green 2006: 204). Thus, a simulator for a basic concept such as cat is a multimodal representation of its recurrent characteristics: Consider the simulator for the category of [CATS]. Across learning, visual information about how cats look becomes integrated in the simulator, along with auditory information about how they sound, somatosensory information about how they feel, motor programs for interacting with them, emotional responses to experiencing them, and so forth. The result is a distributed system throughout the brain’s

Situated positioning with English as a lingua franca   213 feature and association areas that accumulates conceptual content for the category. (Barsalou 2005: 624)

Similarly, social cognizers must possess relatively stable cognitive models of themselves, others, groups, social encounters, i.e. distributed systems of multimodal social knowledge. Thus, in (2) the cognitive models football arbitration and love can be regarded as working as simulators; they function as conceptual types that evoke the default role-models referee, player, and lovers in association with the corresponding stereotypical role relationships. The actual simulation underlying (2), however, only “re-enacts small subsets of [their] content” (Barsalou 2005: 625). This conceptual simulation therefore functions as a creative and novel role model that suits the purpose of channelling the social relationship in the given context. 2.2.2. Simulating social relationships linguistically – A semantic-map model for address formulae In accordance with Barsalou’s theory of conceptualization in particular and cognitive-linguistic theory in general, simulators of social knowledge must be seen as embodied (see Evans and Green 2006: 163–165); they are derived from experiencing the social world and interacting with it in concrete instances. Obviously, these social experiences are also communicative experiences in most cases. Therefore, simulators for social knowledge must also integrate linguistic knowledge (see also Agha 2007: Ch. 2). In other words, cognitive models of social encounters are coupled with the linguistic features, structures, and processes that recur in them. This makes it possible for the linguistic cues to work as space builders (Langlotz 2008). To construct a situated conceptualization linguistically, a social cognizer can therefore rely on entrenched knowledge from cognitive models that can be triggered by linguistic cues (see Figure 2). Being space builders, the linguistic cues work as cognitive indexicals (Agha 2007: 14). On the one hand, they point to cognitive models and therefore support the activation of the simulators. On the other, they also guide the mapping of connections between the cognitive models/simulators for the purpose of building a situated simulation (A. Clark 1997: Ch. 10). To model the coupling of linguistic structures with cognitive models, Croft’s semantic map connectivity hypothesis offers a good heuristics. According to this model “any relevant language-specific and construction-specific category should map onto a connected region in conceptual space” (Croft 2001: 96). Thus, the meaning and discourse-functions of different linguistic constructions and features, which constitute the register of coding alternatives in a given lan-

214   Andreas Langlotz guage, can be distinguished on the basis of how they chart the conceptual and discursive spaces or models with which they are coupled (Croft 2001: 92–98). For illustration, the distribution of English address formulas can be analyzed (see also Laver 1981). English address formulas can be seen as mapping the conceptual space established between the two dimensions of familiarity and hierarchy. Both dimensions constitute continua defined by the poles of intimacy vs. distance and superiority vs. inferiority, respectively. The different address terms chart this socio-conceptual space by mapping a connected region relative to these dimensions (Figure 4). superior

HIERARCHY

Andi!

Mr. Langlotz!

Hey you!

Hi Andi

Hello Mr. Langlotz

Hello

Hi Mr. Langlotz

Dear Mr. Langlotz

Hello Sir

inferior FAMILIARITY

intimate

foreign

Figure 4. Semantic-map model for address formulae

By using one of these address forms, the speaker and hearer can activate and thus simulate the default social relationship that underlies the conventional coupling of address form with the corresponding semantic and discoursefunctional space. This makes it possible to trigger the cognitive simulation of a social relationship by merely evoking the underlying cognitive model via the corresponding linguistic cues. In the same way, words such as darling, honey, etc. can be assumed to be coupled with the cognitive model of love, which implies a corresponding role relationship. It is this cognitive-linguistic coupling of linguistic forms with cognitive simulators as well as the creative play with these associations that constitute the link between the cognitive and the linguistic dimension of social meaning. In what follows, the sociolinguistic dimension of conversational interaction has to be integrated.

Situated positioning with English as a lingua franca   215

2.3. The joint linguistic construction of social orientation To this point, our model of social orientation has focused on the cognitivelinguistic dimension of constructing situated social conceptualizations. So far, it has been neglected that these cognitive-linguistic representations and sensemaking processes are embedded in the sociolinguistic practices of producing speech acts and conversations. Thus, the linguistic cues are part and parcel of the dynamic socio-communicative engagement between two interlocutors, such as the fictive conversations in (1)–(1a) and (1)–(1b). In conversations, situated conceptualizations are constantly modified, extended, rejected, and reassessed. Thus, the conversational move of lover B in (1a) is designed to verify the meaning ‘I love you’ as common ground to be shared by the two interactors (Clark 1996: 93). By contrast, the referee’s reply in (1b) or (2) is a clear rejection of the possible world of love. The referee’s contribution can be considered a conversational extension and modification of the model of provocation; with his turn, the referee becomes a provoker himself in order to re-establish his social position of power. It is important to note that the conversational turns (1a) and (1b) differ considerably in terms of their conventional expectation. While (1) is a predictable speech act between two lovers, it is certainly not a default utterance on a football pitch. This shows how specific social encounters create different discursive expectations, i.e. the presence of specific linguistic cues and corresponding conventional semantic maps. In other words, the discursive expectation of encountering some conventional linguistic cues and communicative moves links the social with the cognitive dimension of situated sense-making. Specific social contexts prime the social cognizers for activating conventional social simulators and cues rather than unexpected ones (Bless et al. 2004: 39–42). Thus, with regard to our model of the coupling of cognitive models with linguistic cues, it can be claimed that these conversational expectations are effected by the entrenched association of linguistic cues and moves with cognitive simulators for typical social practices. The conversational evocations and adaptations of cognitive models in (1a) and (1b) are themselves embedded in and linked to conventional discursive genres, i.e. the discourses of love and football arbitration. It is relative to these social, discursive environments that the categorization of the linguistic cues causes processes of social sense making. Thus, the internalized mental representations of these conversational practices also act as central simulators for the evocation of social meaning (Figure 5). Following Verschueren (1999: 151): “the generation of meaning does not take place in a vacuum”. Utterances and their structural constituents are embedded in comprehensive higher-level discursive complexes.

216   Andreas Langlotz Such speech events are fundamentally social phenomena. They are associated with social activities or encounters, e.g. classroom teaching, dinner conversations, wedding ceremonies, storytelling, etc. As social practices they add a socio-communicative dimension to the cognitive simulation of meaning.

CONVERSATION: JOINT SOCIO-COMMUNICATIVE ACTIVITY (DISCOURSE GENRE triggers conventional expectations: linguistic cues, cognitive models, turns) Categorize

COGNITIVE MODEL OF SOCIAL RELATIONSHIP

Categorize

LOVE

Cues activate

Lover A

I love you so much darling

Lover B

MOVE 1 Cues verify

I adore you too

MOVE 2

Figure 5. The dynamics of languaging social orientation in conversation

Situated positioning with English as a lingua franca   217

Communicative genres are conventionalized but allow for infinite variability and change. A genre … is a recognizable communicative event characterized by a set of communicative purpose(s) identified and understood by the members of the … community in which it regularly occurs. Most often it is highly structured and conventionalized with constraints on allowable contributions in terms of their intent, positioning, form and functional value. These constraints, however, are often exploited by the expert members of the discourse community to achieve private intentions within the framework of socially recognized purpose(s). (Bhatia 1993: 13)

In other words, discursive genres can be regarded as entrenched and conventionalized socio-cognitive sense-making systems. As activity or event simulators they are always meaningful in the sense that, like any other form of social action, they are directly interpreted by the actors involved (Verschueren 1999: 151–152). Specific linguistic cues and actions that can be expected in these discourses trigger conventionalized cognitive models to make sense of a given situation. It is these discursive constellations which link the linguistic and cognitive with the social dimensions of sense-making. And, with regard to our examples, it is the conventional discourses of being in love and playing football that create strong interpretive frames relative to which linguistic cues are assessed for their potential social implications (Cicourel 1973: Ch. 1). As Bhatia highlights, meaning in general and social meaning in particular can also be created by creatively playing with and violating the discursive expectations of speech genres, as in (1b). Deviation from the socio-communicative norms and cognitive expectations creates semantic effects that force the interpreter to deviate from the default categorization of linguistic cues and to invest mental energy in their deep processing (Smith and Mackie 2005: 18; Bless et al. 2004: 42–44). In other words, the social cognitive distinction between deep and superficial processing of social input has its correlates on the socio-communicative level of constructing and negotiating social meaning. The construction of social meaning thus becomes bound to the intricate interplay of convention and conversational variability (Agha 2007: 1). The distinction can be correlated with Grice’s famous terminological dichotomy between conventional implicature and conversational implicature (Grice 1989). While conventional implicatures reside in the activation of expected semantic values, conversational implicatures must be derived ad hoc in the context of a particular conversation. This dialectic between convention and variation correlates with the depth of meaning construction (Figure 6). The tension between convention and creative conversational construction becomes particularly important for the creative positioning strategies by lingua franca speakers in tourist-information discourse.

218   Andreas Langlotz

CONVENTION

VARIATION

Encoded social information

Inferred social information

Conventionalized implicatures – – – –

Language choice Social “lects” Social lexis and deixis Speech Acts Superficial categorization of linguistic cues

Identity Relations

Situated sense-making processes (Conversational implicatures)

Roles Attitudes Deeply processed constructs based on plays with cues

LANGUAGING OF SOCIAL MEANING DIALECTIC

Figure 6. The dialectic between conventionality and variability in the process of social sense-making through language

3. The linguistic construction of a local identity at a tourism board Social orientation in tourist information discourse raises the following central question: How do officers position themselves as grounded in a specific local role when using English as a non-native language, a lingua franca, in this socially challenging communicative context? To answer this question, our theoretical apparatus is applied to two salient conversational examples of creative social positioning at a Swiss tourist information office. These extracts are drawn from an overall data set of 8 hours of recorded front-desk interactions, including approximately 100 front-desk service encounters. While most of these encounters are highly routinized and follow institutionalized conversational sequences (see Drew and Heritage 1992: 29–53), the selected examples provide invaluable insight into the socio-cognitive practice of creative social positioning. In what follows, the specific communicative challenges of tourist information discourse are discussed. In a next step, typical transactional routines are portrayed. It is relative to these routines that the creative construction of a local role and self-image can then be scrutinized.

Situated positioning with English as a lingua franca   219

3.1. Challenges to social communication at tourist information Lingua franca communication is a bridge to overcome the communicative gap between people from different linguistic and cultural backgrounds: At the international level, interaction in a lingua franca often implies communication between speakers of different varieties of the lingua franca: indigenized forms and nonnative or learner varieties potentially meet, and as a result interactions in a lingua franca are highly heterogeneous. (Meierkord 2006: 168)

The lingua franca competence of one or both speakers may be less than perfect.4 Thus, lingua franca communication may be thwarted by interferences from the linguistic features as well as the discursive systems of a given culture, subculture, or community of practice (see Scollon and Scollon 1995: Ch. 7). Accordingly, the interactors must be expected to follow culturally biased perceptions of these linguistic and conversational cues. Meierkord and Knapp (2002) therefore suspect that lingua franca communication is problematic due to its very nature. Others have even claimed that mutual understanding in lingua franca communication is a myth (House 1999). Consequently, the relative absence of conventional common ground between the communicative partners turns lingua franca interactions into minefields where they are very likely to fail (Clark 1996: 100–106). By contrast, Firth (1996) adopts a more positive evaluation. Arguing that lingua franca speakers are aware of these problems and risks, he claims that they strive actively for mutual understanding and develop a mitigating attitude. Obviously, tourist information services are centrally shaped by the very challenges of using English as a lingua franca. Although tourist destinations sell locations, sightseeing tours, and attractions rather than wishing to provide a basis for personal contact between the local population and the guests, the close and immediate contact between the tourists and the information officers nevertheless constitutes a very important platform for the construction of a positive image. Thus, the international customers, who confront the tourist information officer with highly diverse personal needs, expect to experience a glimpse of local hospitality and an authentic interest in their presence. As a result – and in the light of the negative evaluation of successful lingua franca interaction by some scholars – the pressure on the officers to turn the immedi4 With regard to the data analyzed, I consider interactions between a native speaker (the tourist) and a non-native speaker of English (the officer) as lingua franca interactions. For the non-native participant, the English language is clearly the tool to establish a communicative bridge although his/her command of the language is less than perfect.

220   Andreas Langlotz ate front-desk constitution into a positive service experience is intense. Indeed, in corresponding ethnographic interviews tourist information officers tend to regard their jobs as highly demanding.5 Having to convey the hospitality of the tourist destination and therefore striving to cater for the customers’ service and information needs in the best and most friendly way, the front-desk interactions can be expected to reflect a high degree of accommodation to the tourists. Mutual understanding is a must and the customers’ well-being a core business category that goes well beyond the level of a “nice-to-have” side-effect. Thus, the relational dimensions of front-desk interactions are also heavily influenced by the marketing design of tourism. Table 1. Communicative challenges to using English as a lingua franca in tourist information

Tourist information

English as lingua franca

Communicative Characteristics

Problems and Challenges

immediate presence

cross-cultural communication

foreign customers

unfamiliarity

standardization and routinization of interaction

limits of business transaction define social roles limited communication registers limited expression of a specific local identity through linguistic cues

EFL competence non-native, non-local

With regard to the notion of language work in general, Heller (2005) also highlights a number of tensions and dilemmas that communication workers have to face in a global context. On the one hand, business communication genres become increasingly subject to incentives of styling and standardization to guarantee their measurability and to control their efficiency (see also Cameron 2000). Communicative styling, however, runs counter to the central demand of caring flexibly for the personal demands of individual tourists. In addition, communicative styling counteracts the tourists’ expectation of experiencing the destination as being culturally authentic. With a made-to-measure communicative service the central demand of caring for the guest in a very hospitable 5 These spontaneous interviews were conducted when the tourist information data were recorded.

Situated positioning with English as a lingua franca   221

way therefore cannot be fulfilled. On the level of language choice, the limits of constructing local authenticity are also severely challenged by the lingua franca context. While offering a multi-lingual information service – most of the tourist information officers at the site of investigation speak up to five languages fluently – the impression of linguistic authenticity obviously cannot be rooted in the use of the local vernaculars. Thus, key features of genuineness cannot be directly expressed by means of the local linguistic cues. In the light of these many communicative challenges (summarized in Table 1), social positioning therefore becomes increasingly subject to the creative use of situated social conceptualization strategies. 3.2. Typical transaction genres Being forms of standardized institutionalized talk, tourist information exchanges are obviously subject to a routinized transactional order (Drew and Heritage 1992). As a specific genre of business communication, these interactions are clearly goal-oriented with a conventional distribution of the relevant institutional participant roles. Their talk structure is characterized by constraints on allowable contributions, which involve specific inferences that are particular to our perception of the talk structure as a form of tourist interaction (see Heritage 2005: 106). On the basis of the recorded data, three basic interactional routines can be distinguished. The first routine involves providing information. That is, the officer helps the tourist to find a location or he/she indicates the directions to a given venue by using or handing out a map. The second routine is concerned with booking, i.e. booking accommodation, a city tour, or a trip around or outside the tourist destination. This booking routine is complex because it is based on very personal tourist needs and expectations that have to become common ground between the tourist and the officer. Also, booking accommodation can become a very stressful affair in times when the city of destination is overbooked (as will be seen below). Obviously, social images are very much at stake in such a situation. The third distinctive routine is a seller-buyer transaction that occurs whenever the tourist wishes to purchase an article, such as a souvenir, a brochure, programme, or ticket. But this routine is irrelevant for the subsequent analyses. The following interaction reflects a typical “providing orientation” transaction. It involves two American tourists who ask for some interesting sightseeing spots in the city. (Note that only one of them is speaking). The following conversational transcription divides the transaction into its salient sequences

222   Andreas Langlotz and corresponding turns. The details of turn taking (overlaps, etc.) are not accounted for because they are not relevant for the description of the default organization of the genre. Transcript 1.  “Providing orientation” routine I. Negotiation of language choice, language alignment 1

T

do you speak English?

2

Off

yes

3 T yes OK [laughs] II. Greeting address, opening business 4 T III. Request

hi, how are you

5

mmm … we’re, we’re just here for … in [CITY NAME] today … well, we wanna see a few sights … where should we go?

6 Off IV. Mobility checking

you can go to the city centre…

7

Off

do you have tram pass … or do you?

8

T

yes

9 Off yes V. Providing information and orientation 10

Off

[using map for reference] you’re here … you take outside … mmm … you can take the number two tram that’s the platform number … platform three … get off at the third stop …

11

that’s the Kunstmuseeum Art museeum and there you can walk to the left and you can enter the old town

12

that’s the cathedral and from there you see all over the Rhine … you have beautiful view … from the back of the cathedral …

13

and then all down here this is all old city … historical parts … to the market square city hall …

14

up here this is also hill on the opposite side of the town hall … a city gate 14th century …

Situated positioning with English as a lingua franca   223 15

that’s all within the city very nice to walk …

15

and at the same time you have here the shopping area [?, unclear]

16

[…]

17

across the Rhine you can walk would follow the Rhine … you will see its a scenic tour [?, unclear] VI. Appreciation, closing sequence 18

T

OK, well, thank you

19

Off

you’re welcome

20 T that helped so much, thanks VII. Providing further information 21

Off

you can also have a description about the old town its there on the wall in second row on the top its [?] in [CITY NAME]

The interaction reflects seven salient stages (I–VII). With regard to social orientation and service, the opening and closing sequences are of central importance. The conversational opening involves the language negotiation and greeting phases (sequences I and II). In these phases the service relationship is established. The phase of language alignment pulls both interactors into a joint social space. While the tourist actively searches for the construction of common linguistic ground by uttering do you speak English? (Clark 1996: 107–108), the officer shows that she is ready to offer the information service by tuning into the language of the tourist (yes). Cognitively speaking, these first turns thus work to establish a joint mental space of service which is marked for the code English. As mentioned above, the choice of this shared cognitive and linguistic world however also restricts the cues for social expression for the officer considerably because to elaborate the social dimension of the service-model she can no longer depend on her Swiss German semantic maps and the corresponding cognitive models to manage the social relationship. In the context of the globalized service industry of tourism, the phase of language negotiation thus constitutes a decisive step for the potential identities that can be expressed in the interaction. Language negotiation prolongs the phase of establishing the relational dimensions of the service relationship. Therefore, it also plays an important role in relation to the greeting phase (as reflected in turn 4). Since the language negotiation phase already functions to establish contact and to position the interactors rela-

224   Andreas Langlotz tive to the service-model with the corresponding roles service provider and service recipient, the address turn hi how are you seems socially redundant. This, however, is not the case; rather, turn 4 confirms and reinforces the tourist’s recognition of the service relationship as being common ground. Moreover, the conventionalized reference to the addressee’s well-being (how are you) makes explicit that the officer is appreciated as a person rather than a mere provider of services. Thus, relative to the semantic map of address formulas (Figure 4), this form of address places the interactants as peers and reduces their social distance. The address turn 4 also serves a transactional function. It marks the transition to the actual business of requesting and providing information about the sightseeing options (turns 5–6) and the subsequent explanations (turns 7–17), (see Duranti 1997: 68). Note that all of these turns are clearly situated in the genre of tourist information discourse. On the level of content, they are predictable linguistic moves and reflect the default cognitive steps that are to be taken in a tourist information context. In other words, this sequence of task-related linguistic cues does not modify the mental space of service but merely elaborates the cognitive model of what to see in this city that is evoked to differentiate this space linguistically and conceptually. Therefore, the social roles that are implied in the service-model are not modified or re-negotiated; they are performed relative to the default-genre. The service expectation of catering and caring for the needs of the tourist is fulfilled by providing comprehensive and detailed information about a potential city tour. The social roles of the officer and the tourist only become part of the central focus again in the appreciation and closing sequence (step VI, turns 18–20). Here the tourist thanks for the service provided (well, thank you) and makes explicit her gratefulness for the service (that helped so much). Thus, the professional identity of the officer as a good and competent service provider is addressed. Obviously, these moves can still be regarded as default steps in the complex discursive model of service. Thus, both the tourist and the officer maintain the institutional role relationship. They do not attempt to simulate any alternative role model. The additional service turn 21 can also be interpreted in this sense. By providing further information, the officer further enhances her role as a service agent. With regard to the linguistic management of social cognition this typical transactional genre thus offers the following insights. The institutional social roles of the interactors are evoked by following the general semantic map of the tourist-information genre. According to this semantic map, the relational dimensions of the encounter are allocated to the opening and the closing sequences of the interaction. By performing these sequences in default fashion, the entrenched cognitive simulators of the service-model and the corresponding institutional role models are evoked to structure the situated simulation of

Situated positioning with English as a lingua franca   225

the tourist-information encounter. All the simulated transactional steps can be expected. Therefore the interactants’ roles also have to be interpreted relative to the default models service provider and tourist. Interestingly, the jointly negotiated code-switch to the lingua franca English is part of this default simulation. While the code-switch simulates service, it restricts the repertoire for simulating the officer’s authentic and local personality considerably. But, officers can deviate from the standard semantic map creatively to create a strong impression of being grounded in a local role. This will become obvious in the subsequent interactions. 3.3. “No way” – constructing the local service provider The following transaction took place between a Swiss tourist information officer and an English tourist – more specifically, a football fan. The context for this interaction was a UEFA-cup football match that was about to be played later in the evening. To kill the time in between, the tourist requested information about the shopping options in the city centre. The transcript starts at the end of the officer’s explanations. Transcript 2.  “No way” I. End of providing orientation 1

Off

all over here in this part that’s all shopping area and old city at the same time also the historical part II. Appreciation, closing sequence 2 T that’s great III. Request for further information 3

T

could you tell me also where the football stadium is? For later

IV. Jocular resistance 4

Off

no way

5 T [LAUGHS] V. Providing orientation/information 6

Off

you take number 14 tram in the city centre …

Obviously, the turns 1–3 follow the previous default model very closely. The speech act could you tell me also where the football stadium is for later opens a further request that is fully integrated within the discursive model of providing information. Obviously, this speech act also perpetuates the default social

226   Andreas Langlotz role-models that underlie the service encounter. The striking and notable turn occurs in 4, when the officer utters no way. Obviously, this turn is designed by linguistic cues that do not belong to the default semantic map of reacting to a request in the tourist information genre. Thus, relative to the default-model, these cues have to be categorized as an act of resistance and role rejection. Thus, when incorporating the meaning of these cues relative to the standard interpretive procedures (Cicourel 1973: 84–87), the officer implies: no, I certainly won’t help you. Obviously, this role rejection is not acceptable in the tourist information context. Therefore, the tourist is forced to engage in the active and deep processing of these linguistic cues to derive and mentally construct a more appropriate situated conceptualization. To do so, he must connect the cues to another potential cognitive model relative to which they make more sense. This non-default mental space is present in the very context of the request that is at stake: the way to the football stadium. By opening this football matchmodel, the cues no way can be re-contextualized by mapping them from the service-encounter to the football match-simulation. In this alternative mental world, the tourist and the officer may adopt alternative – and with regard to the institutional setting non-default – social roles. In the same way as the tourist, the information officer can act as a football supporter in the mentally evoked football-context. As fans of the two competing teams, the two interactors become rivals, of course. It is relative to this mental simulation of rivalry that the jocular act of resistance (no way) can be fully integrated.

Could you tell me also where the football stadium is?

No way ...

MODEL 1a INFORMATION service provider (English is part of this service!)

MODEL 1b RESISTANCE role rejection I don‘t help you

Figure 7. Mental space model for jocular resistance

BLEND JOCULAR RESISTANCE Local, authentic service provider

MODEL 2 FOOTBALL MATCH rivalry local team supporter

Situated positioning with English as a lingua franca   227

With regard to the construction of a situated local identity, the simple two-word conversational turn is fascinating. While neither the use of English as a lingua franca nor the institutionalized discursive sequences of the tourist information genre allow the officer to express her personality and authentic local role, her simple, creative and jocular act of resisting the conventional procedure of informing the tourist makes it possible for her to trigger the complex construction of a more personal and more local social identity. To construct and to understand this novel social conceptualization, both the officer and the tourist must map and blend the roles of tourist provider from the default-space of service with the mentally simulated role of supporter of local team in the football-space. (The process of blending is explained in Fauconnier 1997: Ch. 6). The resulting situated conceptualization is a blend of the two roles and contexts: the officer emerges as service provider, but as a service provider who is a local football fan, i.e. one with a strong and authentic local basis. Obviously, this blend of the two roles is not pre-given in the conceptual simulators of either the service-model or the football-model. It is constructed “online” as an agent-dependent instruction manual for context-specific social positioning. As we can judge from the tourist’s laughter, he both understands this social joke and also appreciates it. The example thus reflects more closely Firth’s positive evaluation of lingua franca interactions. The officer clearly aims to foster mutual understanding and mutual acceptance. By her jocular resistance, she manages to give both interactors a “momentary mental vacation” (Eastman 1939, quoted according to Holmes and Stubbe 2003: 109). Her communicative strategy of creating a sudden and momentary switch of social role and power relationship further reflects the more general pattern of jokes: … characteristic techniques recur in all forms of verbal humor, namely production of incongruity based on linguistic constructions or on the events described. Such incongruity can be represented as a clash between opposed semantic scripts. … . Recognition of the sudden script switch releases psychic energy as laughter … . (Norrick 2006: 425)

Obviously, the clash between the service-model and the football-model also creates a sudden switch between two incongruent semantic scripts. The joint release of psychic energy between the tourist and the officer modifies their social roles and brings them closer to a more authentic local context than the institutionalized procedures of the information service. This makes this break of convention also highly interesting with regard to communicative styling in business discourse. The officer’s no way does not only resist the contextual demand of the tourist, it is also a momentary escape from the pressure of a fairly standardized way of communicating. Interestingly, it is through this deviation

228   Andreas Langlotz from the norm that she also manages to communicate some of the core values of tourism discourse – authenticity and personal support. However, as we will see in the next example, such norm-breaking strategies of creating authenticity do not always succeed. 3.4. “Outside” – Ironic resistance The following interaction also occurred before the UEFA-cup football match. In addition, it is also essential to know that the city was hosting a very prestigious international trade fair at the same time. Therefore, the destination was hopelessly overbooked. Obviously, this situation placed the tourist information officers under additional stress. The following interaction has to be seen in this light, exactly. Transcript 3.  “Outside” I. Address, opening 1 T II. Booking request

hello there [?]

2

what is the cheapest room tonight, do you have much accommodation?

T

III. Ironic resistance 3

Off

the cheapest one is somewhere outside because there is no … room available

4

T

outside that’s in … outside the city, you mean

5

Off

6

Off

maybe you get an accommodation at the YMCA

7

T

yeah or a youth hostel

8

Off

or at the youth hostel but I am afraid it’s booked

9

T

the youth hostel is booked?

10

Off

yes it is[?] I can try

no, [LAUGHS]… outside of the rooms outside … there is no [much] room available we have only IV. Getting back to business, negotiating alternatives

As one can see from turn 1, the interaction starts with the default opening sequences. This opening differs somewhat from the typical international

Situated positioning with English as a lingua franca   229

transactions because it does not contain an explicit language negotiation phase. The language choice is direct and immediately places the officer in her service provider-role. This social position becomes even more explicit with the tourist’s unconventional request for booking the cheapest room tonight in turn 2. This direct and somewhat impolite enquiry strengthens the role of the officer as a mere provider of services. Given the context of an overbooked city with the corresponding pressure for the information agents – a situation that the tourist does not seem to be aware of – this direct and strong positioning triggers an equally strong, ironic reaction from the officer. And it is this reaction that is of central interest for the subsequent socio-cognitive analysis. The officer’s turn the cheapest one is somewhere outside because there is no … room available is ambiguous. At the core of this ambiguity is the lexical choice outside. The officer uses this lexical cue to simulate a scenario of sleeping-in-the-street (Figure 8). Since this scenario, obviously, is not viable for a city tourist, the suggested evocation of this absurd mental counter-world shows her to be an uncooperative service-provider. Obviously, her irony can only be understood relative to the further mental context of the overbooked city. Relative to this space, the tourist’s request could be revealed as being naïve, while the role of the provider turns from being a conventional service provider to becoming a more knowledgeable, sarcastic service provider.

BLEND IRONIC RESISTANCE Local, authentic service provider

MODEL 1a

What‘s the cheapest room tonight – do you have much accommodation?

The cheapest one is somewhere outside

INFORMATION

service provider (English is part of this service!)

MODEL 1b INFO Cooperative provider

MODEL 2 SLEEPING OUTDOORS Uncooperative provider

Figure 8. Mental space model for jocular resistance

MODEL 3 OVERBOOKED Naïve tourist vs. Sarcastic local provider

230   Andreas Langlotz By winding the tourist up in this way, the officer also stresses her authentic personality. Rather than being a source for cheap rooms, she constructs herself as a strong person that can mentally simulate unpleasant alternatives for the overly direct tourist. This situated conceptualization of the officer’s self image emerges by blending her default-role of the service provider with her suggested alternative that she is a strong, independent and therefore authentic local personality. With regard to Heller’s mentioned dilemmas and tensions of language work (as mentioned in section above), we may claim that the officer again tries to escape from the standards of having to follow the conventions of the touristinformation genre. However, unlike in the previous interaction, the officer here does not aim to strengthen the social bonds with the tourist. Rather, she seems to look for a momentary release of stress and to vent her amazement at the tourist’s naïve request. Her creative mental construct of offering accommodation in the street is a way to highlight personality without having to abandon one’s professional identity as a competent service provider. It is important to emphasize that this role play is bound to the situated mental simulation of an alternative self conceptualization, rather than an active and real departure from the professional identity. Due to the ambiguity of the cue outside, however, the tourist does not join into this short and momentary situated re-conceptualization of their service relationship. Rather, he seems to interpret this cue by trying to integrate it into the default mental-space of the service transaction: Outside that’s in … outside the city, you mean. Thus, by trying to make sense of the officer’s reaction relative to the default service-encounter simulation, he still positions the officer as being a cooperative and subservient information provider. In other words, he does not understand and therefore cannot follow the officer’s irony. As a consequence, the alternatively suggested role conceptualization does not become part of his mental simulation of their social relationship. Interestingly, the officer’s use of English as a lingua franca seems to lie at the very heart of this failure to construct the situated social conceptualization as a shared category. Obviously, the use of outside is not a native-like lexical choice to trigger the desired meaning of ‘outdoors’. In this sense, this lexical choice supports Meierkord and Knapp’s doubts concerning the efficiency of English lingua franca communication. Clearly, the suboptimal lexical choice prevents the officer from communicating her desired act of role resistance in a satisfactory way. Nevertheless, given the high risk of face threat for the tourist, her imperfect command of English possibly prevents her from offending the tourist herself. In this sense, the repair strategy of returning to business quickly in turn 5 and of offering a potential alternative to sleeping “outside” (maybe you get an accommodation

Situated positioning with English as a lingua franca   231

at the YMCA) in turn 6 indicates the officer’s attempt to abandon her alternative mental evocation of the service relationship. Thus, although the situated social conceptualization in this second example does not become a mutually shared conceptualization, it nevertheless illustrates nicely the previously outlined theoretical model of the creative construction of social orientation. 4. Conclusions Despite the many limitations for the construction of social orientation in tourist information by using English as a lingua franca, speakers are able to construct an authentic and personal local role or identity by means of creative communication strategies. Instead of the local cues for the linguistic marking of social meaning, the speakers use humour and allusions as creative departures from the conventionalized transaction routines. In doing so, they carve out their own local role and position. They invite their interlocutors to accept these creatively constructed social positions by joining into this creative “role play”. The analysis suggests that the relationship between language and social cognition is bound to an intricate discursive interplay between linguistic cues, cognitive representations and processes, and motivated social action. Linguistic structures are used strategically to trigger mental simulations of alternative social relationships. These creative plays make use of two central strategies. First, they attempt to manipulate the sequences of linguistic cues in a way that suggests or implies a departure from the default expectations of the tourist information genre. On the basis of this creative break with convention, the cues must be actively embedded in alternative mental contexts or scenarios. That is, the cues must be reallocated in an alternative mental space and conceptual simulation relative to which they become consistent. This cognitive process of opening alternative mental scenarios is also accompanied with an active evocation of the alternative social relationships and role models that they imply. To reconcile the conventional and institutionalized role models of the service encounter with the incongruent linguistic cues, the interactors are forced to cognitively construct connections with the alternative mental backgrounds that the cues point to. I doing so, they create a situated social conceptualization that offers a momentary release from the institutionalized default relationship. However, the actual implementation of this alternative, situated role conceptualization is more than a cognitive phenomenon; it is socio-cognitive. In order to become a joint “reality”, the cognitively constructed role-alternative must be shared and mutually accepted by the interactors. Otherwise, it remains a conceptual potential only.

232   Andreas Langlotz The qualitative analysis suggests that the communication of rich social meaning with English as a lingua franca is possible despite the given limitations. However, the creativity mirrored in these interactions is dependent on a solid communicative competence in English. Therefore, the awareness of people who have to use English as a lingua franca in the complex and diverse communicative niches of the 21st century has to be sharpened accordingly. Otherwise, they will find it difficult to adapt their personal communication styles to the constraints of the demanding communicative situations. References Agha, Asif 2007

Language and Social Relations. Cambridge: Cambridge University Press. Barsalou, Lawrence W. 1999 Perceptual symbol systems. Behavioral and Brain Sciences 22: 577– 660. 2005 Situated conceptualization. In Handbook of Categorization in Cognitive Science, Henri Cohen and Claire Lefebvre (eds.), 619–650. Amsterdam: Elsevier. Bhatia, Vijay K. 1992 Analysing Genre: Language Use in Professional Settings. London: Longman. Bless, Herbert, Klaus Fiedler, and Fritz Strack Social Cognition: How Individuals Construct Social Reality. Hove: 2004 Psychology Press. Blommaert, Jan 2004 Discourse: A Critical Introduction. Cambridge: Cambridge University Press. Cameron, Deborah 2000 Good to Talk? Living and Working in a Communication Culture. London: Sage. Chappell, Denise 2005 Opportunities for negotiation at the interface of phone calls and servicecounter interaction: a case study. In Calling for Help: Language and Social Interaction in Telephone Helplines, Carolyn D. Baker, Michael Emmison, and Alan Firth (eds.), 237–256. Amsterdam and Philadelphia: John Benjamins. Cicourel, Aaron V. 1973 Cognitive Sociology. London: Penguin. Clark, Andy Being There: Putting Brain, Body, and World Together Again. Cam1997 bridge, Mass.: MIT Press.

Situated positioning with English as a lingua franca   233 Clark, Herbert H. 1996 Using Language. Cambridge: Cambridge University Press. Croft, William 2001 Radical Construction Grammar: Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. Croft, William, and D. Alan Cruse Cognitive Linguistics. Cambridge: Cambridge University Press. 2004 Crystal, David English as a Global Language. Cambridge: Cambridge University Press. 2003 Davies, Bronwyn, and Rom Harré Positioning: the discursive production of selves. Journal of the Theory 1990 of Social Behaviour 20: 43–65. Drew, Paul 2005 Conversation analysis. In Handbook of Language and Social Interaction, Kristine L. Fitch and Robert E. Sanders (eds.), 71–102. Mahwah/ London: Lawrence Erlbaum. Drew, Paul, and John Heritage 1992 Analyzing talk at work: an introduction. In Talk at Work: Interaction in Institutional Settings, Paul Drew and John Heritage (eds.), 3–65. Cambridge: Cambridge University Press. Duranti, Alessandro 1997 Universal and culture-specific properties of greetings. Journal of Linguistic Anthropology 7 (1): 63–97. Eastman, Max Enjoyment of Laughter. London: Hamish Hamilton. 1936 Edwards, Derek Discursive psychology. In Handbook of Language and Social Interac2005 tion, Kristine L. Fitch and Robert E. Sanders (eds.), 257–273. Mahwah/ London: Lawrence Erlbaum. Evans, Vyvyan, and Melanie Green 2006 Cognitive Linguistics: An Introduction. Edinburgh: Edinburgh University Press. Fauconnier, Gilles 1994 Mental Spaces. Cambridge: Cambridge University Press. 1997 Mappings in Thought and Language. Cambridge: Cambridge University Press. Firth, Alan 1996 The discursive accomplishment of normality: On “lingua franca” English and conversation analysis. Journal of Pragmatics 26 (2): 237–260. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge, Mass.: Harvard University Press. Heller, Monica 2005 Language, skill and authenticity in the globalised new economy. Noves SL. Revista de Sociolinguistica [electronic version]. Retrieved April 24 from http://www.gencat.net/presidencia/llengcat.

234   Andreas Langlotz Heritage, John 2005 Conversation analysis and institutional talk. In Handbook of Language and Social Interaction, Kristine L. Fitch and Robert E. Sanders (eds.), 103–147. Mahwah/London: Lawrence Erlbaum. Holmes, Janet, and Maria Stubbe Power and Politeness in the Workplace: A Sociolinguistic Analysis of 2003 Talk at Work Harlow: Pearson. House, Juliane Misunderstanding in intercultural communication: Interactions in Eng1999 lish as a lingua franca and the myth of mutual intelligibility. In Teaching and Learning English as a Global Language, C. Gnutzmann (ed.), 73–93. Tübingen: Staufenberg. Jenkins, Jennifer 2003 World Englishes. London: Routledge. Knapp, Karlfried, and Christiane Meierkord (eds.) Lingua Franca Communication. Frankfurt a.M.: Peter Lang. 2002 Koester, Almut Josepha 2004 Relational sequences in workplace genres. Journal of Pragmatics 36: 1405–1428. Kövecses, Zoltàn 2006 Language, Mind, and Culture: A Practical Introduction. Oxford: Oxford University Press. Kunda, Ziva Social Cognition: Making Sense of People. Cambridge, Mass: MIT 1999 Press. Langlotz, Andreas 2008 Contextualisation cues as mental-space builders. In Du Fait Grammatical au Fait Cognitif – From Gram to Mind: Grammar as Cognition, Jean-Rémi Lapaire, Guillaume Desagulier, and Jean-Baptiste Guignard (eds.), 347–366. Bordeaux: PUB – Presses Universitaires de Bordeaux. Laver, John D. 1981 Linguistic routines and politeness in greeting and parting. In Conversational Routines: Explorations in Standardized Communication Situations and Prepatterned Speech, Florian Coulmas (ed.), 289–304. The Hague: Mouton. Meierkord, Christiane Lingua francas as second languages. In Encyclopedia of Language and 2006 Linguistics. Keith Brown (ed.), 163–171. Amsterdam: Elsevier. Meierkord, Christiane, and Karlfried Knapp 2002 Approaching lingua franca communication. In Lingua Franca Communication, Karlfried Knapp and Christiane Meierkord (eds.), 9–28. Frankfurt am Main: Peter Lang. Norrick, Neal R. 2006 Humor in language. In Encyclopedia of Language and Linguistics. Keith Brown (ed.), 425. Amsterdam: Elsevier.

Situated positioning with English as a lingua franca   235 Pomerantz, Anita, and Jenny Mandelbaum Conversation analytic approaches to the relevance and uses of relation2005 ship categories in interaction. In Handbook of Language and Social Interaction, Kristine L. Fitch and Robert E. Sanders (eds.), 149–171. Mahwah/London: Lawrence Erlbaum. Schegloff, Emanuel A. 1999 Discourse, pragmatics, conversation analysis. Discourse Studies 1: 405–435. Scollon, Ron, and Suzanne Wong Scollon Intercultural Communication: A Discourse Approach. Oxford: Black1995 well. Smith, Eliot R., and Diane M. Mackie Social Psychology. Hove: Psychology Press. 2000 Smith, Linda B., and Larissa K. Samuelson Perceiving and remembering: Category stability, variability, and devel1997 opment. In Knowledge, Concepts, and Categories, K. Lamberts and D. Shanks (eds.), 161–195. Hove: Psychology Press. Tajfel, Henri 1969 Cognitive aspects of prejudice. Journal of Social Issues 25: 79–97. Verschueren, Jef 1999 Understanding Pragmatics. London: Arnold. Wetherell, Margaret, Stephanie Taylor, and Simeon J. Yates 2001 Discourse Theory and Practice: A Reader. London/Thousand Oaks: Sage.

Chapter 9 Constructing knowledge schemas in the workplace: A microanalysis Susan Bridges and Brendan Bartlett 1. Introduction TESOL institutions in Hong Kong and around the world have competed in recent times to provide language training and education packages for secondary and primary teachers of English. Driven by the Hong Kong Language Proficiency Requirement (LPR) policy (Standing Committee on Language and Education Research 2003), Hong Kong’s language teachers have sought to upgrade and demonstrate their own language proficiency. Some research has been conducted on learner-related aspects of their effort and of the programs through which they undertook further study (Bridges 2007; Coniam and Falvey 2003). However, little has explored the enterprise from a provider’s perspective. In this chapter, the authors attempt to do so. We present a microanalysis of what one administrator has told us of her work. The analysis centres on the social-cognitive dimensions of the administrator’s conceptualisation of “success”. This is framed in a synthesis of two layers of analysis from previously distinct ontological and epistemological perspectives – interactionism and cognitive psychology. The first is from the field of social interaction (Baker 2004; Freebody 2003; Silverman 1998) and works from ethnomethodological perspectives founded by Harvey Sacks. It assumes that communication is an intersubjective, co-constructed accomplishment that can be examined in talk data. The second layer of analysis draws on a theory of mind (Bartlett 1978, 2003; Meyer, Young, and Bartlett, 1989, 1993) that accounts for the organisation, prioritisation and communication of schematic content in terms of levels of idea structure. Specifically, the theory of top-level structuring (TLS) reflects an associative view of how memories are constructed and retrieved and meaning generated at the mostintegrated or “top” level of the idea hierarchy underpinning a communication (Bartlett 2003). Thus, when looking at what the administrator said, we added

238   Susan Bridges and Brendan Bartlett a technical analysis of how she said it. We did so by identifying links within and between clusters of related ideas in her account (through role and propositional analysis) and finding the integrating theme throughout what she said (through prioritization analysis). We have synthesised the interactional analysis of her co-construction of the interview text with a TLS analysis of the administrator’s data as a means to examine a social, cognitive and metacognitive perspective of “success” in a language enhancement program. This synthesis of interactional and cognitive psychological perspectives provides a means of examining both the context and the content of what is said. 2. Context This research originated from a case study of stakeholders’ perceptions of learning and provision during a specific English language program (Bridges 2007). The pedagogical context of the program was clearly defined. English teachers from Hong Kong who had either Cantonese or Mandarin as their first language (L1) came to Australia for intensive language proficiency training and assessment. The Hong Kong government determined the program’s syllabus, including assessment instruments and criteria in the Syllabus Specifications for the Language Proficiency Assessment for Teachers (English Language) (LPATE) (Government of the Hong Kong Special Administrative Region (HKSAR) 2000). The Australian provider had created the program from the specifications and had developed appropriate teaching and assessment materials for its implementation in all syllabus components. Additionally, the provider was responsible for administering and marketing the program. Delivery was in immersion mode with the Hong Kong teachers travelling to Australia and residing with Australian ‘homestay’ families for the 6-week program. The ‘guiding issue/question’ for the case study (Stake 1995, 2000) was: How did the multiple stakeholders perceive learning and provision? Of specific interest is research undertaken with an administrator who also held a wider role in the general marketing of TESOL in-service and education training programs (INSET) (Bolam 1986). She was a senior course administrator and had been part of the administrative team for this LPATE immersion since its first delivery by this Australian university. A two-pronged approach was taken in the research design with application of a visual stimulus and a verbal reporting technique to gather stimulated response data. Working from the position that meaning is constructed in social contexts, these data gathering tools were used to foreground her viewpoint.

Constructing knowledge schemas in the workplace   239

3. Concept mapping The visual stimulus was a concept map drawn by the senior course administrator (Participant M) relating to a role the participant had filled in the INSET. She was asked to draw a concept map of successful INSET provision that focussed on “what made for a successful immersion LPATE delivery”. This was used as the visual stimulus for the stimulated recall session conducted as an open-ended interview. In the session she explicated her concept map labels and design with the researcher. Her account was audio taped and transcribed.1 Use of concept maps as stimuli originated in cognitive process research (Novak and Musonda 1991), and the stimulated recall approach has been applied in education since the 1970s in order to understand teachers’ judgements in classrooms (Cooke 1999). Cooke saw recent changes in working conditions across many fields as a justification for “knowledge elicitation” methodology. She believed that such a methodology had a “stronger focus on the cognitive complexity of job performance” and that it measured “the connection between knowledge and performance” (Cooke 1999: 8). She defined elicitation of knowledge as a process which constructs “a model of the expert’s knowledge – the outcome of which may reflect reality in varying degrees” (Cooke 1999: 3). Gass and Mackey (2000) described this approach as an “introspective paradigm” of cognitive psychology. In this study, concept mapping and stimulated recall were applied as introspective tools to elicit knowledge. 4. Stimulated recall Stimulated recall is a verbal reporting technique which Gass and Mackey (2000: 11–12) saw as evolving from earlier “think aloud” protocols that had subjects vocalise their “silent thoughts” whilst completing a task. They justified stimulated recall as an “information processing approach whereby the use of and access to memory structures is enhanced, if not guaranteed, by a prompt that aids the recall of information” (Gass and Mackey 2000: 17). They justified its application by asserting it helps participants relive the situation under investigation. The aim of the two-pronged approach to knowledge elicitation was to explore this key stakeholder’s perception of a successful program delivery. An 1 All qualitative data quoted in this study are expressed in italics with interview transcription conventions based on the work of Gail Jefferson and turn numbering based on Baker (2004). The content is authentic reproduction with uncorrected instances of expression.

240   Susan Bridges and Brendan Bartlett interpretive analysis of the relationship between her curriculum knowledge and her administrative knowledge was made in investigating her account of factors contributing to successful INSET. Analysis is initially framed by the interactional features of the recall session in order to establish how the ‘social’ contributed to and impacted upon the knowledge elicitation process (Holstein and Gubrium 2004). This is followed by top-level structural analysis to reveal how the program administrator clustered her knowledge into a coherent account (Bartlett 1978, 2003; Meyer et al. 1989, 1993). This approach provided insights into the participant’s conceptualisation of the elements of INSET provision and also into the impact of the research tools upon the data-gathering process. One advantage in having the administrator’s concept map was that her thought processes in schematising, organising and depicting the INSET provision had already been depicted and checked with her as accurate and complete prior to the interview with the researcher. The narrative provided by the interview, therefore, was highly structured and coherent. 5. Language and interaction Interviews have been questioned as an objective form of data gathering with qualitative researchers (Baker 2004; Freebody 2003; Holstein and Gubrium 2004; Potter and Hepburn 2005) arguing that one cannot view the interviewee’s account as an objective reality but rather as the intersubjective, co-construction of formulated accounts of an interaction. The research interview, therefore, is seen as an interactive accomplishment between the researcher and the research participant drawing upon cultural knowledges in a manner similar to naturally-occurring talk. Interactional issues to be explored in the following section include: 1. the intersubjective accomplishment of a recall interview as socially constructed talk-in-action; and 2. the sequential features of the spoken text. Potter and Hepburn (2005: 285) identified “contingent problems” in qualitative interviews in psychology such as: –– –– –– –– ––

“the deletion of the interviewer; the conventions of representation of interaction; the specificity of observations; the unavailability of the interview set-up; and the failure to consider interviews as interaction”.

Constructing knowledge schemas in the workplace   241

In what follows, we call upon analytic perspectives from ethnomethodology to address these contingent problems by a) describing those interacting, their roles and relationships; b) describing the interview set-up; and, c) examining the transcribed text of the interview as interaction. An interactionist perspective on stimulated recall data produced in an openended context may be of greater interest to social cognitivists than data framed by scripted questions. The former bring the interviewer to life in the data gathering process from the virtual reality of predetermining questions and direction to one of participating at point of demand in helping an interviewee – from such practical matters as rewinding a tape to more interpersonally cognitive ones such as sorting or clarifying a recall or what it means. 5.1. The intersubjective accomplishment of a recall interview as socially constructed talk-in-action Analysis of interactional dynamics of the interview with the administrator (Participant M) provides a view of the social context of the recall session. 5.1.1. The interactants The interviewer and Participant M were native speakers of English who interacted socially as well as professionally. They shared a background and interest in working in international education, in both onshore and offshore contexts. Data-gathering involved the use of social time and home spaces and reflected the high level of comfort in the interaction. Table 1. Interactional characteristics of interview Characteristics

Interviewer (I)

Interviewee (Participant M)

Role in the study

Researcher/teacher

Research participant/ administrator

Role in the INSET

Language instructor

Course administrator/ marketer

English proficiency

English L1

English L1

Professional ­background

Secondary teaching and tertiary TESOL teacher education

International education administration

Gender

Female

Female

Age

Early 40s

Mid 30s

242   Susan Bridges and Brendan Bartlett Participant M had a high level of industry experience with this and similar ESL/EFL immersion programs. As a senior administrative team member, she had worked on most INSET run by this provider, including the LPATE. As an administrator since its first delivery three years before, she had a thorough knowledge of LPATE policy and of this ELICOS provider. Her prior experience included course marketing and management in all phases from conceptualisation through implementation, evaluation and replanning, to final reimplementation. She had represented the institution’s LPATE involvement in: 1. 2. 3. 4.

Liaising with officials from the Hong Kong government; Marketing and liaising with the Hong Kong agent; Managing client relations and administrative support staff; and Reporting to senior management staff.

Her experiential knowledge of the administrative processes involved in INSET and the LPATE program in particular provided a relevant background as an administrative stakeholder. Additionally, this extensive experience in delivery may have afforded her opportunities to understand other positions and stakeholders and their positions within the institution. This included notions of program design and learners’ needs. The interface between administrative knowledge and curriculum knowledge as displayed by the administrators/marketers of INSET delivery and perceptions of clients’ needs as learners were of particular relevance to the wider study of the INSET phenomenon in the internationalisation of higher education in Australia (Bridges 2007; Bridges and Bartlett 2007, 2008). The interviewer was a language instructor undertaking contract work on the same program. She taught two of the three classes offered in this Australian immersion program and had prior offshore LPATE teacher education experience. She was also a student undertaking doctoral studies at the time. The two shared social time with families as well as collegial time at work. 5.1.2. The interview set-up The concept map and interview process for Participant M involved three steps. First, the researcher explained and modelled the concept during a shared car journey. Two nights later, Participant M completed the concept map in a 30-minute session in her home. She then telephoned and was visited immediately at home for an audio taping of her account. This process reflected the realities of Participant M’s availability. She was in Australia only briefly between international marketing trips. Gass and Mackey (2000) noted Bloom’s (1954)

Constructing knowledge schemas in the workplace   243

argument that “recall accuracy” diminishes significantly in proportion to the length of time that passes between the event and the stimulated recall – “generally 48 hours”. This warning informed data gathering for this study in that Participant M was interviewed immediately after completing her concept map. The social context of the recall session allowed for use of personal time and home environments as well as immediate access. The respondent produced a coherent concept map (see Figure 1) which she explicated well in interview conditions, showing through some minor adjustments as she spoke that she was attending to the task and rethinking the depiction represented in her map. The work of this first section has been to indicate some of the intersubjectivity that may be at play in the interaction by exploring the roles and relationships of the interactants. Secondly, it has provided a detailed account of the set-up of the interview indicating how the recall tools were employed. In what follows, we examine more closely the talk data itself to explore how the use of a recall item may have influenced the interactional accomplishment of the interview. In particular, we draw briefly upon the ethnomethodological traditions to examine the structure of the text in terms of the turn-taking pattern employed by the interactants. 5.2. The sequential features of the spoken text The transcript examined here provides this administrator’s reflection on her conceptual framing of a successful INSET stimulated by her attention to a freshly completed concept map of that framing (see Appendix 1). The following analysis draws upon the tradition of Ethnomethodology (EM), specifically Conversation Analysis (CA) (see Silverman 1998). The sequential organisation of turns is of special interactional interest in these talk data. Ethnomethodology looks at turns at talk as sequential organisation within a three part structure: 1) relation of a turn to the prior one; 2) what is occupying the turn; and 3) relation of the turn to a succeeding one (Sacks, Schegloff and Jefferson 1974). That is, the analysis explores how a turn is “heard” and “understood” by the hearer and then how this understanding is “heard” by the next interlocutor/s. In his interpretation of Sacks’ work, Silverman (1998) reasoned that turns have three consequences: 1) needing to listen; 2) understanding, and 3) displaying understanding. Analysis of talk data that focuses on turns is therefore seated in the understanding that talk is social action and that analysis of their sequential organisation can provide the apparatus whereby the analyst may examine how the social is co-constructed (Freebody 2003). The interview text in this study holds specific points of interest in terms of the structure of interview talk-in-action, specifically in terms of the turn-tak-

244   Susan Bridges and Brendan Bartlett ing structure. The conversation analysis literature has established the pattern of initiation-response sequences with a) an interviewer prompt in the form of a question and statement followed by b) an interviewee response. Conversation analysts have established the highly complex nature of the interactive work undertaken in this simple pattern in terms of the formulation of the turns and how they are heard by the interactants (Silverman 1993). While a full conversation analysis of the talk data is beyond the scope of this paper, examination of some of the fundamental aspects of sequence and talk provides key insights this particular text and its interactional accomplishment. In the text shown in Appendix 1, the overall pattern of interview talk is clear as initiation-response sequences. Indeed, the interaction is quite rapid and freeflowing as indicated by frequent overlapping and latched speech and with few pauses or reformulations. However, it has particular features of interest in the turn-taking organisation, particularly relating to the interviewer’s role in the interaction. The first is in the absence of a traditional question-answer pattern. The second is a predominance of response (or receipt) tokens (Gardner 1998; McCarthy 2003). 5.2.1. Turn-taking As an open-ended interview, the talk consists of 116 turns. There is a total absence of questions in the turn-taking exchanges. The interviewer begins with an open-ended prompt in Turn 1 tell me about you concept map and uses a second prompt at the end of the interview in Turn 102 mm You‘ve got this one=. There is little interviewer talk, one possible cause being that the nature and presence of the recall stimulus had already reconfigured the interactional dynamics of the interview in a way that was mutually acceptable and that needed no additional negotiation. The interviewee had already formed her cognitive schemata and the work of the talk was to explicate its visual representation in the form of a concept map. As such, the interviewee was able to substantively lead the interaction with only two prompts from the interviewer that sought content or clarification – one to start (Turn 1), the other to explicate a label that had not been addressed (Turn 102). This second prompt was acknowledged immediately by the interviewee who addressed it in Turn 103 as indicated by the latching symbol (=) at the end of Turn 102. The latching reflects rapidity of response. The physical presence of the concept map stimulus may have contributed to this rapidity. Both interactants were seated in close proximity reading the map together and physically indicating labels and links. In Turn 102 when the interviewer indicated a label that had not been explicated, the interviewee responded immediately.

Constructing knowledge schemas in the workplace   245

These turns indicate the interviewer’s work at controlling the opening sequence and correcting a possible omission. In addition the interviewer takes on the work of closing the interaction from her first overlapping talk at Turn 112 miminal curriculum changes to her summary statement at Turn 114 mm Tha::t is a successful in-service delivery! to her closing thanks at Turn 116. Besides this work, the main interactive task for the interviewer is as an active listener and this is accomplished verbally through the use of response tokens (McCarthy 2003). The role of active listener affected the structure of the talk in terms of length of turn as the interviewee’s long turns and the interviewer’s short ones dominated the pattern of this text (see Appendix 1) thus departing from more established patterns. 5.2.2. Response tokens Gardner (1998: 220) identified the “core functional meaning” of tokens as “continuation of current speaker, of aligning/agreeing acknowledgement and of minimal acknowledgement”. Part of the work of these tokens is “back channelling” whereby the listener maintains active participation in the action of the talk. Besides the work of the few turns indicated above, in the remainder of her turns, the interviewer provides multiple response tokens such as mm, yeah, yep, ahah, OK, right, alright as well as laughter. These turns do interactive work in signalling active listening and participation as well as indicating agreement and encouraging further talk. As such, they reflect the first two of Gardner’s (1998) core functional meanings. Indeed, the high usage of response tokens indicates that the majority of the interviewer’s talk was used as encouragement. Additionally, there are instances of acknowledgement such as in the turns below: 32. 33.

M: A little bit circular I: no no that’s great

The work of this sequence is affirmation in response to the heard hesitation from the interviewee and as such it continues the work of active listening and encouraging. This brief discussion of the organisational structure of the interview has highlighted two specific interactive features. First, the interviewee’s explanatory talk dominates the interaction with the interviewer providing relatively few prompts and doing a large amount of agreeing and encouraging through the use of response tokens. Second, the turn taking is rapid and tightly structured. These features may be attributable to the recall task of concept map construction and to the map’s role in providing stimulus and structure for the interaction.

246   Susan Bridges and Brendan Bartlett 6. Language and cognition Interactional analysis indicated how social context and the stimulus prompt may have influenced interactional patterns within the interview. Of added interest is an exploration of the content of the interviewee’s recall interview. We now draw upon the top-level structural analysis to explore the structure of procedural knowledge. 6.1. The structure of procedural knowledge Communicated knowledge is a representation of the apposite realities of those involved in its communication as much as it is of the content that communicators leave for each other. Its vehicle typically is language and in this study, Participant M’s advice on the content and structure of her thinking was moderated by her rendering of them as first a concept map and then as a recall of its representation. As our account of this process shows, Participant M was an active and engaged respondent. She was stimulated not only to recall what the map had meant to her, but also to correct it along the way so as to better represent her views on looking at (.) elements of successful (0.3) provision of the ((course name)) or a teacher professional training course=in-service training course. (Turn 2). As a finished product, the respondent’s concept map had been constructed with labels stuck onto a blank manila folder with hand-drawn arrows and connecting phrases joining them to show where clusters of information were formed. This map was then reproduced electronically as an organisational chart (Figure 1). The reproduction however shows straight arrows rather than the looped ones that Participant M had drawn to indicate relations between sub-clusters of information on the original. The relations fell generally into two forms of rhetorical predicates – lists and cause-and-effect structures and are described in this way throughout the chapter. Participant M’s approach to the task cognitively is analysed here in terms of top-level structure theory – a conceptualisation of language-evidenced cognition that assumes people schematically organise and prioritise ideas into more or less coherent schemas that may be identified by a finite number of rhetorical predicates (Bartlett 1978, 2003, 2008; Meyer, Young and Bartlett 1989, 1993; Fletcher et al. 2008). These predicates have specific functions in arranging information into formats such as lists, comparisons, causal chains or problems and solutions. They do so with progressively larger clusters of information as a communication is constructed so that one or

Constructing knowledge schemas in the workplace   247

another of the predicates will provide the macrostructure – the top-level structure of a completed text, while any number of them will be present at lesser levels. Participant M had three definable moments in producing her map. First, she itemised key features of successful INSET delivery as she saw them, and generated labels for each. At this point, the top-level structure of her record was a list of attributes. However she moved on, sorting and reclustering the items into two differentiated lists of content shown in Figure 1 as “Maximum enrolments” and “Successful delivery”. She added detail during this second-stage review. For example she now included content about “positive evaluations” and “school visits” when talking through her item on “student satisfaction”. At this second point, she had a list of lists. But again she changed; this time rearranging the two lists in causal relation to the issue at hand, i.e. indicating in description of the connecting lines that a successful INSET was an outcome of “Maximum enrolments” and “Successful delivery”. Her label as shown in Figure 1, “Successful outcomes”, underscores the “effect” of the two causal conditions. The importance levels for the respondent of each piece of information are indicated by their positions on the seven levels represented in Figure 1. Most important is the cause (“Successful outcomes”) – effect (List of two: “Maximum enrolments” and “Successful delivery”) cluster shown on the first and second levels of the array. They constitute the top-level structure of her account as a cause-and-effect rhetorical predicate of structural and content elements. Her key message is that a successful INSET relies on maximum enrolments and successful delivery. The specific top-level structuring Participant M used indicates two features in her message about success. First, her social cognition in relation to the matter at hand is heavily causal. What matters is getting it right. Second, in informational terms, what is needed to do this is contained within enrolment and delivery issues. Participant M’s opinion is structured around success for the institution providing the course, as might be expected from an administrator of that institution. “Maximum enrolments” unfolds in her account as information associated with effective promotion and cost-efficiencies – such as communicating well with agents and having promotional seminars well-attended. Crossovers indicating connections with details of the second cluster occur, such as between cost-efficiencies and minimal dramas. “Successful delivery” is her second cluster. She weighted it generally at the same level as the first. Participant M thinks the information this schema contains is generally as important in its causal function as that for “Maxi-

248   Susan Bridges and Brendan Bartlett mum enrolments”. It does have more detail and hence its depiction descends further. It loads strongly at each level on institutional information. The view includes items on student satisfaction, timely assessment and notification of results, positive moderation visits, all students passing and the Hong Kong Government being happy. Such clustering is listed rightly under a “delivery” label rather than a “transaction” or “reception” one. Its emphasis is on how the composite accomplishments of a successful delivery connect with future enrolments (strong) and minimal curriculum changes. There is no explicit message about the nature and quality of learning or socialisation experiences of students that matches an emphasis Bridges (2005) found in their reports. Together, the two clusters cohere as causal explanations of “Successful outcomes”, rather than other forms of organisation such as an explanation highlighting solutions for problematic issues expected along the way to being successful, or one presenting alternative, comparative or contrasting arguments for success. The interview highlighted Participant M’s cognitive processing as she explained her change in rhetorical predicates when making decisions about how best to present her information. She advised that she had used categories generated in the original list to instantiate content into causal relationships as the map. Her map design and her explication of the causality among categories followed a strongly hierarchical structure as Bartlett (1978) and others (Meyer et al. 1989, 1993) indicated it should. It followed logically from her focus on outcomes, as did the transcription protocols of student-learners in the same INSET that Participant M had just administered. There were two important differences, however, in the account from Participant M and those of student-learners that are reported elsewhere (Bridges 2005, 2007). First, student-learners tended to organise their accounts of “success” in a problem-solution top-level structure rather than the causal one our respondent had used. They talked through what they needed to do in order to address problematic issues such as passing examinations of their English language competency, or how better to benefit from immersion experiences in the Australian context of their studies, or how to transfer learning on return to Hong Kong. These centred on their learning and benefits that would follow – again reflecting the different orientations and social press of the parties. The absence of detail in Participant M’s accounts of “Student satisfaction – positive evaluations”, “Assessment/Results: Provision on time” and “Student results: All pass” suggests a concept of benefit imposed as a students’ perspective rather than one derived from it.

Homestay arrangement successful

Figure 1. Concept map of successful INSET provision

Promotional seminars well attended

Cost effective delivery

Good communication with agents

Promotional material accurate and available

void

Effective promotion

Maximum enrolments

Good PR future enrolments strong

Student satisfaction positive evaluations

Minimal dramas

Successful Outcomes

Useful/positive school visits

Minimal curriculum changes

Student Results: all pass

Hong Kong Government happy

Positive moderation visits

Experienced staff Assessments / results provision on time

Successful delivery

Constructing knowledge schemas in the workplace   249

250   Susan Bridges and Brendan Bartlett 6.2. Successful outcomes In explicating the construction of the map, Participant M discussed her organising framework. Initially, she had started with a ‘top three’ – effective promotion, successful delivery and assessment and results. Her justification for this was that this was a chronological framing with initial course promotion followed by delivery and then results. Results were viewed as follow-up including post-course assessment. However, she reconsidered this structure before pasting down the labels. Her decision to change to a cause-and-effect top-level structure by including successful outcomes as part of the overarching issue was that it was the better way to tell her story. With her new organisational format, she decided that the most important labels under the header of successful outcomes were maximum enrolments and successful delivery. These had strongest causal links to successful outcomes and so the cause-and-effect top-level structure was instantiated with successful outcomes (follow) maximum enrolments and successful delivery as specific content. Participant M also argued that these two “causes” were fairly equal – placing them at the same level immediately below the effect they created in combination. In teasing out these two major factors as sub-headings and causal organisers, she recognised that maximum enrolments were not as quantifiable as successful delivery. She accounted for the visually unbalanced nature of the map’s branches by stating that successful delivery was very broad and difficult to measure, which is why there are more stickers under it. Participant M’s explication of these two major sub-headings is addressed further in the two sections that follow. 6.3. Maximum enrolments The strong causal format of Participant M’s thinking on the issue of a successful INSET is reflected not only at the top-level of her structural information, but throughout middle-level and lower levels of ideational importance for each of the two causal factors. In explicating maximum enrolments, she linked various exemplifications of the ideas effective promotion and cost-effective delivery back through a causal chain stretching from accurate and available promotional material and well-attended promotional seminars through good communication with agents. The dominance of the cause-effect predicate in her thinking is revealing also in terms of the greater weighting given to ‘effect’ content throughout the

Constructing knowledge schemas in the workplace   251

causal chain in the cluster. If a distinction is to be made, Participant M is more interested in outcomes than in their antecedents – she has prioritised effects over causes across her whole communication. The final representation of the hierarchy reflects her thinking that the action behind the more detailed information at any lower level exemplifies and leads to that of levels above it. This was not how she presented at the beginning of the data-gathering exercise. Originally, she had positioned the various items in almost reverse order to that shown in Figure 1 and reflecting the chronological framing with which she was familiar of INSET – promotion followed by delivery and then results that we outlined previously. However, in reconstructing the map she rearranged the items without any prompting, arguing a need to re-prioritise them. 75. 76. 77. 78.

M: and – I don‘t know if you want me to go across here now- but ‘maximum enrolments’ I put will also help to ensure ‘cost effective delivery’ I: Yeah M: because if you’re delivering to less that your potential obviously it’s going to cost you more. I: mm

Her connecting maximum enrolments to cost effective delivery was not delineated further in her account nor was it isolated to maximum enrolments. In contrast with effective promotion, she saw it as having direct links to subbranches of successful delivery. A perfect binary with links at each level was not established between these two major organizers. Rather, a synergy was woven between them. Direct links that facilitated this synergy were minimal dramas and successful delivery with Participant M adding minimal curriculum changes as an afterthought in the interview. She accounted for this with the argument that dramas are what cost you money and that it’s curriculum development …that costs you a lot of extra money (Turns 95–97). These links provide in number and content the strongest account of financial gain as a key factor in “success” from her viewpoint. All the concepts under maximum enrolments were financial in terms of marketing – ‘selling’ the program and gaining revenue from enrolments. These naturally evolved from her administrative role and are consistent with expectations of this perspective. We have noted already that the concept label cost-effective delivery ties the content maximum enrolments and successful delivery. When explicating successful delivery, Participant M added the additional connector that minimal dramas contribute to student satisfaction. The causality here was that a little drama it can cause a really big issue and rea:lly impact badly on your student (.) perceptions (Turn 71). She later related student satisfaction to adding value

252   Susan Bridges and Brendan Bartlett in terms of good PR for the course which should result in strong future enrolments. The logical connection between the two sets of linked information was to maximum enrolments – a connection illustrating close relation in Participant M’s cognitive framing between business and costs in curriculum elements. She positioned satisfied students as “satisfied clients” able to promote the educational product to peers. To have them capable of doing so increased the institution’s profile and prospective market share (Turn 89). The financial aspect of perceived success is an important thread in the tapestry of INSET in the international ESL marketplace. It highlights a difference in focus between administrators/marketers and course designers/language instructors and the possible influence this can have on stakeholder interactions. Course designers for INSET generally adopt a reflective, cyclical approach to program development and evaluation and operate under the premise that programs need constant revising, replanning and renewing of materials based on reflections of what did and did not work in the classroom for each particular group. This stance might create tension with administrators. While recognising a need to invest in curriculum change as a means of ensuring client satisfaction, administrators are cognizant also of its high cost. A balance is required between reducing investment in curriculum development for cost-effective delivery and replanning for greater success in the classroom. Finding the balance between curriculum change and cost-effectiveness may become a major source of negotiation and tension between the two stakeholders. 6.4. Successful delivery What Participant M was thinking in relation to Successful delivery was explicated in terms of a hierarchical organisational structure centred on a chain of causal relationships. This replicated her organisation in the previous cluster (Maximum enrolments). Major organisers were the concept labels, minimal dramas, and, experienced staff. In her account that follows, “staff” meant “teaching staff” (course developers and language instructors). 34. 35. 36.

M: um (0.1) so over on the other side. ‘Successful delivery’. I had these little bits that didn’t necessarily fit in anywhere else but I felt were really important elements to include. I: mm M: One was ‘experienced staff’ which are vital to ‘successful delivery’ but I put your ‘successful delivery’ relies on ‘experienced staff’. You couldn’t possibly have it with a bunch of (.) newcomers or amateurs.

Constructing knowledge schemas in the workplace   253

37. 38. 39. 40. 41. 42.

I: mm M: And that ‘minimal dramas’ would contribute to your ‘successful delivery’ I: ((laugh)) M: Those are (.) these outlying issues (.) but they do:: I: mm M: link up with other things. But rea:lly then I broke ‘successful delivery’ down into four elements

Participant M recognised that really important elements of successful INSET delivery were experienced staff and minimal dramas. She argued that experienced staff were vital (original emphasis) to successful delivery and appreciated the key role of experienced ESL teacher educators who were familiar with the context. Minimising dramas was important because dramas are what cost you money (Turn 73). Her initial difficulty in placing these elements in the concept map is reflected in the excerpt above. It illustrates her indecision when she had these little bits that didn’t necessarily fit in anywhere else but I felt were really important elements to include (Turn 34) even though they were outlying issues (Turn 40) within the concept map. She recognised the importance of curriculum design in the enacted and experienced curricula. However, she also was confused about how to integrate and represent them so that the diagrammatic exemplar of her cognitive schema was accurate. While our respondent displayed some understanding of curriculum components, she lacked clarity about their adequacy or perhaps how they fitted best alongside other elements of a successful INSET. This hesitation would be understandable given that curriculum design is outside her area of expertise. Participant M listed a cause-and-effect top-level structure as she explicated four elements of successful delivery. The first was cost-effective (delivery). This was a financial consideration that she recognised in her account as coming from my perspective in the admin ro:le (Turn 46). While she qualified the importance of cost-effective delivery in the branch of successful delivery by stating that it was pa:rtly measured by its cost effectiveness (Turn 58), she perceived it as a causative thread linking her cognitive schema. The second element was student satisfaction. Participant M explained this as positive evaluations by the students. The student satisfaction survey was delivered at the end of the program and is used institutionally as a data-gathering tool to measure that satisfaction. Participant M’s account of program evaluations has agency for students alone and was based on data generated in response to the provider’s questions. In this account, she did not include provision for what course designers or language instructors may have had to relate from students’ interactions with them, or from their own perspectives.

254   Susan Bridges and Brendan Bartlett Her third element of successful delivery was assessment and results being provided on time. Participant M’s rationale for its inclusion was that it was there because a lot of that takes place after the course finishes (Turn 54). Yet, she did not detail or exemplify the item identity and timing issues involved in her statement. Her fourth element was positive moderation visits. While a small element, she argued that they’re very important, I think, because it’s a government course (Turn 56). She believed that if you were to measure ‘successful delivery’ part of it would be that those visits went well (Turn 62). External relations are important to Participant M’s success in marketing and administrating INSET. Therefore, the positive evaluations and reports of moderation visits by the commissioning body hold a significant place in her concept map. After justifying the placement of the four elements, Participant M then drew links between them. She argued that these little bits like ‘experienced staff’ will contribute to good moderation visits. They will help to get ‘on time provision of results’ and will lead to:: your ‘student satisfaction’ (Turn 64). Strong causality is drawn between results and satisfaction as ‘student results’…i.e. them passing, will lead to ‘student satisfaction’ (Turns 83–85). It was only after explaining this subordinate level that Participant M returned to the level above and the missing sub-branch of successful delivery. This was minimal dramas. The label proved to be an outlier in form. As a descriptor it was the only instance where she used informal language. The label was afforded great importance because it’s one of those things if you have a little drama it can cause a really big issue and rea:lly impact badly on your student (.) perceptions (Turn 71). While the nature of the dramas was not elaborated, their importance was recognised in her statement that ‘minimal dramas’ (.) contribute to ‘student satisfaction’ (Turn 69) and connect with the concept map label, student evaluations (Level 4). Dramas related to the enacted and experienced curricula with which, as an administrator, she was not familiar. While her understanding of the nature of the dramas may have been limited, Participant M did recognise they linked to the other, more financial side of her map. This is apparent in her argument that ‘minimal dramas’ will also assist with cost-effective delivery because dramas are what cost you money (Turn 73). Participant M then moved her discussion to kind of a different layer (Turn 89) to explicate the more detailed levels in her thinking. She was consistent with her application of the cause-and-effect structure noting that ‘student satisfaction’ will lead to ‘Good PR’ (.) for the course which will lead to:: ‘strong future enrolments’ (Turns 89–91). She also saw a causal relationship between the ‘Hong Government being happy’ and a ‘positive moderation visit’ (Turn 103).

Constructing knowledge schemas in the workplace   255

103. M: =Over here you’ve got the ‘Hong Kong Government being happy’ umm 0That’s meant to come in here too 0. Really what you need to keep the ‘Hong Kong Government happy’ is that they have a ‘positive moderation visit’. They like what they see= 104. I: [mm] 105. M: =They get the results on time= 106. I: [mmm] 107. M: =and the results are good. 108. I: mm 109. M: If you have those three things 110. I: [mm] 111. M: you’re gonna have a ‘happy government’ and tha::t will also contribute to= 112. I: = ‘minimal curriculum changes’ 113. M: Because they can demand that things (.) be different. Her conclusion in the excerpt indicates that if positive elements fill the causeand-effect relationship, then a successful in-service delivery will result and in turn contribute effectively as one of two causes of a successful INSET. Cost-effective delivery as a category is unique in Participant M’s cognitive map. Its placement creates coherence between administrative notions of financial viability and maximum enrolments on one side of the map, and program design elements of a successful enacted curriculum and future enrolments on the other. For an administrator, cost-effective delivery of INSET is a bottom line issue and its position as a unifying element in her cognitive map is both descriptive and predictable. Participant M values good curriculum delivery. Her marketing background and administrative role dominate her perceptions of it and accentuate its worth. This is apparent in her map where curriculum achievements were represented in terms of profitable marketing. She administers and promotes INSET as a financially-viable training product for her institution. Her concept map reflects these administrative and promotional priorities and so has little to say about education or the learning that may accrue for students and teaching staff. 7. Summary Our account of Participant M’s opinion of what makes for a successful INSET drew heavily from theory that suggests knowledge and its communication are affected by social-cognitive dimensions of language. The two dimensions we

256   Susan Bridges and Brendan Bartlett have emphasised are perspective and top-level structure. The former we contend acts to filter both perception of the task of communicating one’s thoughts on an issue, and sequential organisation of what is communicated. Thus it introduces into communication particular bias that is reflected in what is co-constructed as interview talk. In this instance, we saw the interviewee dominating the turns at talk. This reconfiguration of the social and organisational structure of a research interview may be attributable to the cognitive application of concept mapping as a stimulus. The latter suggests that there are identifiable ways in which we organise what we communicate as big-picture information. These ways help us to predicate content so as to highlight the relative importance of its elements and to define in structural terms how the keypoint of our message is constructed. Participant M is a senior manager of a provider institution whose training program through INSET for teachers of English in Hong Kong needed to be financially viable as well as educationally responsive. Her position in the institution was to ensure success in these terms. She was committed to the goal of successful provision of the LPATE. Her constructs of how provision should be scaffolded were driven by this goal. Given her perspective, predictions could be made that what she would communicate to us as a “successful” program would be loaded with information about viability. And it was. It may seem that her relatively minor detailing of educational content in learning terms or in relation to what students and their teachers considered important was surprising, given her prior related experiences and opportunity she might have made to speak with students in this specific program. However, she knew the first author both as a colleague who worked as a teacher-programmer in INSET and as the researcher who had set up the mapping and interview task. She may have assumed that we would fill in the pieces more pertinent to educational viability while she emphasised the business elements. If so, it would be a piece of meta-social cognition familiar to all who have made assumptions about what audiences or readers know and how they act in communication events. What we have added to what might otherwise have been predictable in Participant M’s account, is a characterisation of how she verbalised her thoughts so that they might accurately and effectively represent her thinking. We did this by using top-level structuring theory. Her attention to getting labels positioned just right was evidence of the socialisation pressures on her for precision in her communication. This precision manifested in applied linguistics terms as a very strong causal top-level structuring of what she had to tell us. She brought her thinking to consciousness first as a simple list of ideas that might address the issue and that followed a sequence she had become accustomed to as she

Constructing knowledge schemas in the workplace   257

had moved through various roles in her institution. She then sorted and sifted her information, interrelated and prioritised it until she had communicated the message that maximum enrolments and successful delivery were the antecedents of a successful INSET. Actually, Participant M’s account was dominated throughout by a causal predication of information. Other than with her main idea, she used it also to link content that she felt explained what predisposed a course to be successfully delivered and what factors led to maximum enrolments. It may not be a common practice for Participant M or for people generally to rely so heavily on one type of predicate throughout the cognitive and communicative processing of content. Typically, our written and spoken texts are more diverse. Participant M’s use of the cause-effect predicate not only at the top-level of her message, but also in its supporting detail is suggestive of a constraint in her reporting to ensure she had tied down all that she said as logical and iterative. A second suggestion is that she felt this was an appropriate form of address given her prior knowledge of the interviewer/researcher. If this is the case, it is reason to support the virtues of an interactionist perspective that situates a research interview as a co-constructed accomplishment. The interviewee’s account is therefore not examined as an objective truth but rather as an intersubjective, co-construction of an account that is formulated in-situ. It is uncertain at what level of consciousness Participant M operated during and after the research concerning the bias in her account toward “success” in business terms. It is also unclear whether and how a wider view of benefits and beneficiaries in a successful INSET to include learning, students and teaching staff alongside her institutional concerns would affect ways in which she does her work and verbalises it. An important methodological outcome of a socialcognitive approach to exploration of how people communicate their thinking is that it opens possibilities for checking with respondents on the accuracy and completeness of their accounts. For example, pointing out the structure at the top-level of presented information might then be followed by inviting respondents to reconfigure their thinking using alternative predication. In Participant M’s case, her revisit might prompt alternative constructions of success if she were to access that part of her knowledge schema that had been co-constructed with her staff and or students in the INSET. In validating the respondent’s data against the given perspective, the interviewer-researcher might help Participant M to see legitimate bias as a starting point for a more fulsome account of “success” that incorporates education as what is learned and voices other than her own.

258   Susan Bridges and Brendan Bartlett 8. Implications for method In the recall session, further analysis could be done on factors such as task completion style in terms of time, work habits and the actual process. In the concept map, Participant M was loath to place stickers down before she had totally finished the process mentally and said she had conceptualised it in her head before starting, as she didn‘t want to ‘make a mess’. She even suggested creating it straight into a computerised organisational chart as she felt that she thought better on a screen. This reflected a different approach to the representation of knowledge. From a social-cognitive perspective, stimulated recall provided a principled account from an important stakeholder in INSET. An insightful picture of the provider’s account was drawn by facilitating participants’ reflection and metacognition on the administrative perspectives of INSET.

Constructing knowledge schemas in the workplace   259

Appendix: Interview transcript I = interviewer M = senior administrator 1. I: ((M)), tell me about your concept map. 2. M: OK. The concept map is looking at (.) elements of successful (0.3) provision of the ((course name)) or a teacher professional training course=inservice training course. u::m So I started, I wrote all my bits and pieces and then I didn’t stick them do:wn 3. I: [mm] 4. M: because I sta:rted wi::th (.) I actually started with ‘maximum enrolments’, ‘successful delivery’ and ‘assessments/results’ (.) as my= 5. I: [mmm] 6. M: =No, sorry I tell a lie. My three=I started with my top three as (0.2) ‘effective promotion’ u::m ‘successful delivery’ and ‘assessment and results’= 7. I: [mm] 8. M: =thinking that, there were three, breaking it down kind of chronologically. >You promote the course, you deliver the course< and then you have to do all that (.) results follow-up= 9. I: [Yeah] 10. M: =and assessing videos etcetera (.) and that those would be the three and I had it all laid out but the:n in the end I decided the overarching issue at the ve:ry top was ‘successful outcomes’ and I’d already written that one. I just mo:ved it up. 11. I: mm 12. M: But it took me a while to get to that (.) And then I decided that rea:lly in terms of ‘successful outcomes’, the most important things we:re (.) ‘maximum enrolments’= 13. I: [mm] 14. M: =for what we’re able to take and ‘successful delivery’ which is ve::ry broad. It’s unfortunately not as measurable and specific as ‘maximum enrolments’ which is why there are more stickers under it ((jokingly)) 15. I: ((laugh)) 16. M: So I’ll start with ‘maximum enrolment’ = 17. I: [OK] 18. M: =because it’s easier. So I (.) actually I went here. I said I decided these were fairly equal ‘successful outcomes’ a:re ‘maximum enrolments’ and are ‘successful delivery’ 19. I: [mm] 20. M: Now, ‘maximum enrolments’ (.) actually started at the very bottom (.) of this. If you have this and this and this you will get maximum enrolments but then I decided as the outco::me (0.1) 21. I: mm

260   Susan Bridges and Brendan Bartlett 22. M: It had to come up the other wa:y because I started with outcomes. That ‘maximum enrolments’ – but then I didn’t apply that to the rest of this this = 23. I: [that’s alright] 24. M: = 0then moves back the other direction0 – will result from (.) ‘effective promotion’ (.) and ‘effective promotion’ (.) will be assisted by ‘good communication with agents’ so the people promoting the course and representing us in Hong Kong 25. I: [mm] 26. M: are rea:lly the people we rely on to promote the course. And ‘good communication with the agents’ will lea:d to:: ‘accurate and available promotional material’ which will (.) and will which will also lead to ‘promotional seminars being well attended’ 27. I: [mm] 28. M: and that your communication with your agents would contribute to that as we:ll= 29. I: [Yeah] 30. M: =because they’re promoting the course and that ultimately that would assist to achie:ve ‘maximum enrolments’ 31. I: ahah 32. M: A little bit circular 33. I: no no that’s great 34. M: um (0.1) so over on the other side. ‘Successful delivery’. I had these little bits that didn’t necessarily fit in anywhere else but I felt were really important elements to include. 35. I: mm 36. M: One was ‘experienced staff’ which are vital to ‘successful delivery’ but I put your ‘successful delivery’ relies on ‘experienced staff’. You couldn’t possibly have it with a bunch of (.) newcomers or amateurs. 37. I: mm 38. M: And that ‘minimal dramas’ would contribute to your ‘successful delivery’ 39. I: ((laugh)) 40. M: Those are (.) these outlying issues (.) but they do:: 41. I: mm 42. M: link up with other things. But rea:lly then I broke ‘successful delivery’ down into four elements 43. I: mm 44. M: ‘cost effective’ 45. I: Yep 46. M: and that’s from my perspective in the admin ro:le 47. I: mm 48. M: ‘student satisfaction’ and then I put in brackets ‘positive evaluations by the students’= 49. I: [mm] 50. M: =and that’s how you’d measure that satisfaction. 51. I: mm

Constructing knowledge schemas in the workplace   261 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64.

M: The ‘assessment and results being provided on time’ I: mm M: because a lot of that takes place after the course finishes I: mm M: and ‘positive moderation visits’. Although they’re really a small element, they’re very important, I think, because it’s a government course. I: mm M: So we’ve got ‘successful delivery’ pa:rtly measured by its ‘cost effectiveness’ will lead to ‘student satisfaction’ I: [mm] M: requires the ‘on time provision’ of these things and inclu::des (.) I: positive = M: = If you were to measure ‘successful delivery’ part of it would be that those visits went well. I: mm M: then you have these little bits like ‘experienced staff’ will contribute to good moderation visits. They will help to get ‘on time provision of results’ and will lead to:: your ‘student satisfaction’.

((0.12 break to tend to a boiling pot)) 65. M: a:nd (.02) and again your ‘minimal dramas’ really that mainly,>oh I should have, I‘m missing one‘cost effective’ I didn’t break down any further< but ‘student satisfaction’ and ‘positive evaluations’. You have your ‘homestay arrangements’ being successful and I put that ‘plays a large role in student satisfaction’ 80. I: Yeah

262   Susan Bridges and Brendan Bartlett 81. M: ‘Useful/positively evaluated’, really ‘school visits’ contribute to ‘student satisfaction’ 82. I: mm 83. M: ‘Student results’ 84. I: [mm] 85. M: i.e them passing, will lead to ‘student satisfaction’ 86. I: mm 87. M: u::m (.) and then I kind of jumped beyond those= 88. I: [alright, yep] 89. M: =again to kind of a different layer. That ‘student satisfaction’ will lead to ‘Good PR’ (.) for the course which will lead to:: ‘strong future enrolments’. 90. I: Right 91. M: And will also contribute to ‘minimal curriculum changes’ = 92. I: [mm] 93. M:= which I saw as a really positive thing. 94. I: mm 95. M: In fact tha::t should link up to ‘cost effective delivery’ because it’s ‘curriculum development’ 96. I: [Yeah] 97. M: that costs you a lot of your extra money. 98. I: Yeah 99. M: So if you have ‘minimal curriculum changes’ 100. I: [mm] 101. M: it helps to (.) achieve ‘cost effective delivery’. 102. I: mm You’ve got this one= 103. M: =Over here you’ve got the ‘Hong Kong Government being happy’ umm 0 That’s meant to come in here too0. Really what you need to keep the ‘Hong Kong Government happy’ is that they have a ‘positive moderation visit’. They like what they see= 104. I: [mm] 105. M: =They get the results on time= 106. I: [mmm] 107. M: =and the results are good. 108. I: mm 109. M: If you have those three things 110. I: [mm] 111. M: you’re gonna have a ‘happy government’ and tha::t will also contribute to= 112. I: = ‘minimal curriculum changes’ 113. M: Because they can demand that things (.) be different. 114. I: mm Tha::t is a successful in-service delivery! 115. M: ((laugh)) 116. I: Thank you ((M))!

Constructing knowledge schemas in the workplace   263

References Baker, Carolyn Membership categorization and interview accounts. In Qualitative Re2004 search: Theory, Method and Practice, David Silverman (ed), 162–176. London: Sage. Bartlett, Brendan J. 1978 Top-level Structure as an Organizational Strategy for Recall of Classroom Text. Tempe: Arizona State University. 2008 Action research in changing workplace climate. In Evaluating Action Research, E. Piggot-Irvine, and B.J. Bartlett (eds.) Auckland: New Zealand Research Council. 2003 Valuing the situation: A referential outcome for top-level structurers. Paper presented at the 1st International Conference on Cognition. Language and Special Education, Gold Coast, Australia. 2008 I’ve Been Workin’ on the Railroad: Action Research in Changing Workplace Climate. In Evaluating Action Research, E. Piggot-Irvine, and B.J. Bartlett (eds.), 167–190. Auckland: New Zealand Research Council. Bridges, Susan English Language Immersion: Theorising from Stakeholders’ Ac2005 counts. Unpublished Doctoral thesis. Brisbane: Griffith University. 2007 Mapping learner perceptions of an English immersion INSET. Prospect 22 (2): 39–60. Bridges, Susan, and Brendan J. Bartlett 2007 Prospect or promise – Internationalisation in Australia. In Language and Languages: Global and Local Tensions, Christina Gitsaki (ed), 15–39. Newcastle, UK: Cambridge Scholars Publishing. 2008 Global academics: Wandering scholars or tradable commodities? Journal of the World University Forum I (6): 91–98. Bloom, B. 1954 The thought processes of students in discussion. In, Accent on Teaching: Experiments in General Education, S.J. French (ed.), 23–46. New York: Harper. Bolam, R. 1986 Conceptualising inservice. In Inservice Training and Development: An International Survey, D. Hopkins (ed.), 14–26. London: Croom Helm. Coniam, David, and Peter Falvey Benchmarking the benchmark: Assessing the fit of a new test with its 2003 target population of teachers of English in Hong Kong. Hong Kong Journal of Applied Linguistics, 8 (1): 1–15. Cooke, Nancy 1999 Knowledge elicitation. In Handbook of Applied Cognition, F.T. Durso (ed.), 479–509. New York: Wiley.

264   Susan Bridges and Brendan Bartlett Fletcher, M., O. Zuber-Skerritt, E. Piggot-Irvine, and B.J. Bartlett Data collection methods for evaluation. In Evaluating Action Research, 2008 E. Piggot-Irvine, and B.J. Bartlett (eds.), 53–90. Auckland: New Zealand Research Council. Freebody, Peter Qualitative Research in Education: Interaction and Practice. London: 2003 Sage. Gardner, Rod Between speaking and listening: The vocalisation of understandings. 1998 Applied Linguistics 19 (2): 204–224. Gass, Susan, and Alison Mackey Stimulated Recall Methodology in Second Language Research. 2000 Mahwah, N.J: Lawrence Erlbaum. Government of the Hong Kong Special Administrative Region (HKSAR) 2000 Syllabus Specifications for the Language Proficiency Assessment for Teachers (English Language) (Syllabus). Hong Kong: Government of the HKSAR. Holstein, James A., and Jaber F. Gubrium 2004 The active interview. In Qualitative Research: Theory, Method and Practice, D. Silverman (ed.), 140–161. London: Sage. McCarthy, Michael 2003 Talking back: “Small” interactional response tokens in everyday conversation. Research on Language and Social Interaction 36 (1): 33–63. Meyer, B.J.F., C.J. Young, and B.J. Bartlett 1989 Memory Improved: Reading and Memory Enhancement across the Life Span through Strategic Text Structures. Hillsdale, N.J: Lawrence Erlbaum Associates. 1993 Reading comprehension and the use of text structure across the adult lifespan. In Reading across the lifespan, S.R. Yussen and M.C. Smith (eds.), 165–192. New Jersey: Springer-Verlag. Novak, J.D., and D. Musonda A twelve-year longitudinal study of science concept learning. American 1991 Educational Research Journal 28 (1): 117–153. Potter, Johnathan, and Alexa Hepburn Qualitative interviews in psychology: Problems and possibilities. Quali2005 tative Research in Psychology 2: 281–307. Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson A simplest systematics for the organisation of turn-taking in conversa1974 tion. Language 50 (4): 696–735. Stake, Robert 1995 The Art of Case Study Research. Thousand Oaks, CA: Sage. 2000 Case studies. In Handbook of Qualitative Research. 2d ed. Norman K. Denzin and Yvonna S. Lincoln (eds.), 435–454. Thousand Oaks: Sage.

Constructing knowledge schemas in the workplace   265 Standing Committee on Language and Education Research 2003 Action Plan to Raise Language Standards in Hong Kong (Final report of language education review). Hong Kong: Government of the HKSAR. Silverman, David 1993 Interpreting Qualitative Data: Strategies for Analysing Talk, Text and Interaction. London: Sage. 1998 Harvey Sacks: Social Science and Doing Conversation Analysis. New York: Oxford University Press.

Chapter 10 Corporate self-presentation and self-centredness: A case for cognitive Critical Discourse Analysis1 Veronika Koller

1. Introduction: Corporate branding discourse This chapter focuses on the way large corporations conceptualize themselves as corporate brands and try to establish this ideal self in the minds of stakeholders such as employees, investors or customers. As such it seeks to describe the structure of the socio-cognitive representation that is the corporate brand, and to elaborate on what linguistic and visual features discourse producers use to communicate these brands. In a further step, the chapter presents evidence of how corporate branding as a cognitive and discursive phenomenon has spread to other social fields, here local government. To this end, corporate misson statements and multimodal city branding texts are analyzed and the results compared. In doing so, the chapter combines notions from social cognition, notably the idea of possible selves (Markus and Nurius 1986), with a critical approach to discourse. Companies can be seen as social agents, engaging in relational behaviour with stakeholders with the aim of maximizing profit and shareholder value, and thereby constituting and projecting a sense of the corporate self. Companies are also paradoxical creatures: On the one hand, they are increasingly forcing their way into the private sphere of consumers and employees alike, as evidenced by aggressive direct marketing aimed at so-called target groups2 and demands for complete commitment and identification on part of the workforce (Gee, Hull 1 I am grateful to Ann Mitchell and Graham Pinfield for providing me with useful information on the branding processes for Oldham and Blackpool, respectively. 2 Direct marketing comprises both telemarketing in the form of phone calls and customized email offers as well as personal selling. An extreme form of the latter is viral marketing, where private individuals are commissioned by companies to spread word-of-mouth messages about, and give out samples of, products in their respective peer groups (Grauel 2004).

268   Veronika Koller and Lankshear 1996: 31). Other outward orientation involves the “courting” of shareholders (to use a favourite metaphoric expression from corporate discourse). Companies further impact on the communities where they do business, often in positive ways through what has come to be known as corporate social responsibility. And as we shall see, corporate discourse has also spread to social fields such as higher education and local government. On the other hand, however, corporations are highly introspective and concerned about their identity. Such soul-searching has given rise to a number of writings on Corporate Identity (CI), defined as the company’s actual self, i.e. a collective model of the company’s specific features, its past and goals, its value and behaviour that originates from, but ultimately transcends employees’ notion of the company. CI constitutes a form of collective identity in that employees may construct their social self as “derive[d] from [their] membership in larger, more impersonal collectives” (Brewer and Gardner 1996: 83) such as the company they work for. Although the two trajectories of corporate reach seem to point in different directions, they do meet in the notion of the corporate brand. Defined as the “outward oriented (i.e., projected) organizational identity” (Kapferer 2002: 185), a corporate brand is the image a company wishes to convey to it stakeholders, its ideal or “ought self” (Kunda 1999: 472). The present paper sets out to look at corporate brands in terms of their sociocognitive structure. An important strand in social cognition is represented by the theory of social representations (Moscovici 2000). Social representations, i.e. the cognitive structures jointly held by members of a particular group, are theorized to be the subject of “continual renegotiation […] during the course of social interaction and communication” (Augoustinos and Walker 1995: 178). Further, social representations are said to provide a shared frame of reference so that communication can take place at all, i.e. make available the assumed shared background knowledge that tends to become naturalized as “common sense” in hegemonic discourses. Finally, social representations can be seen as establishing social identies and relationships by being communicated (Augoustinos and Walker 1995: 178, 180). In short, they fulfill both an ideational and an interpersonal function in the Hallidayan sense and thus lend themselves very well to an analysis of the links between language and social cognition. In our case, the concept of social representation is used interchangeably with that of the socio-cognitive model. As such, it can be transferred to the cognitive structure that is a brand, the central currency and defining feature of brand communities (Muniz and O’Guinn 2001). In socio-cognitive terms then, a corporate brand constitutes the ideal self a company wishes to communicate to others. Ultimately, this ideal self is meant to converge with the “epistemic social schema” (Kristiansen 2008), i.e. the

A case for cognitive Critical Discourse Analysis   269

beliefs a social group, in the present case, stakeholders, has about the nature of another social group or entity, here, the company. Moreover, the chapter investigates in how far corporate brands can be seen as a form of individual, rather than collective, identity. Before gathering the textual evidence for or against this, the next section will discuss brands and branding in the context of discourse and social cognition. 2. Context: Brands as socio-cognitive representations in discourse Brands represent the socio-cognitive representations, including their affective components, that stakeholders maintain about a particular product, service, company or, in the case city branding, place. To make these abstract models graspable, companies seek to endow their brands with a quasi-human “personality” (Christensen and Askegaard 2001), up to the point where brands become metaphorical persons interacting with stakeholders.3 This brand personality can again be seen as a “possible self” (Markus and Nurius 1986), i.e. the communicated ideal self as a specific model of the self in future states. Another version of the possible self is the ought self, i.e. the self as it is supposed to develop. Such models are sociocognitive because they are not merely “an idealized set of cognitive objects, but a range of socially situated discursive practices” (Augoustinos and Walker 1995: 306). According to van Dijk (2003), it is particular models only that will be distributed in discourse, and they are disseminated at the expense of other models which are equally cognitively available but less socially motivated. Given the growing importance of brands in late capitalist consumer culture,4 it has become a central task of corporate discourse to communicate and reinforce particular concepts through various texts. One site at which such impression management can be witnessed is the corporate mission statement. As the one genre in corporate discourse that acts most as “[carrier] of ideologies and institutional cultures” (Swales and Rogers 1995: 225), mission statements have the dual purpose of communicating executives’ image of the corporate self to employees, thus achieving identification and internal cohesion, and of persuading a multitude of external stakeholders of the inherent value of the company in question. As part of a wider discourse, the individual texts that fall within the boundaries of the mission statement genre draw on specific models to com3 For an explanation in terms of metonymy, see Davidse (2002: 154). 4 The importance of brands is reflected in the fact that even in 2000, already more than 60 per cent of the value of fast-moving consumer goods (such as cosmetics or soft drinks) were accounted for by their brand value (Grauel 2000: 18).

270   Veronika Koller municate a particular impression of a company into the public sphere. Texts as both the process and the product of discursive behaviour tend to draw on particular socio-cognitive representations selectively, often in order to sustain or subvert power relations between discourse participants. By engaging in relational behaviour, discourse participants (re-)produce “a set of intercategory relationships, a set of […] conditions […] that are perceived to affect people because of their shared or different social category memberships” (Abrams 1999: 198). Moreover, by doing so, they also engender their identities as participants on that set. Vice versa, these identities also motivate relational behaviour. Applying these basic notions to the corporate sector, we find “companies” – the metonymic shorthand for the identities of, and relations between, executives as decision-makers and other employees implementing these decisions – engaging in relational behaviour with “stakeholders”, an umbrella term for investors, clients, competitors, government agencies, NGOs and other groups and individuals. The salient conditions at work are the economic systems of late capitalism in which the ultimate aim for corporations is to increase profits and thus shareholder value by promoting consumption of their goods and services, widening their market share at the expense of competitors and keeping costs, including labour costs, low. By pursuing this aim, corporations both actualize and constitute their sense of self. As far as the same, or similar, linguistic and visual strategies are employed routinely, the corporate self, while still dynamic in principle, stabilizes into a highly accessible sociocognitive model that will be used in future texts. Corporate behaviour is manifold, ranging from obvious acts of buying and selling to lobbying in the political sphere and, increasingly, as the welfare state is being dismantled, philanthropy or corporate social responisbility. For present purposes, the focus is on corporations’ discursive behaviour, i.e. their textually mediated relational actions. Corporate texts are distributed through various channels to, and interpreted by, different categories of stakeholders. Seen in its entirety, this highly complex process of text production, distribution and interpretation constitutes corporate discourse as a social practice (Fairclough 1995: 135). Discourse is linked to cognition in that socio-cognitive models underlie both meaning-making processes and other discursive behaviour. These models capture notions of the self, including scripts for its relational behaviour, its historical past and its goals, as well as its inherent value, and also compromise affective components in that they can trigger emotional responses when communicated. Linking these theoretical considerations back to the case at hand, models of the corporate self are posited as a basis for corporate discourse. In the highly complex formation that is corporate discourse, mission statements stand out as the genre that is perhaps most endowed with traces of the

A case for cognitive Critical Discourse Analysis   271

ideal corporate self, or corporate brand. Given the variety of audiences, they have to reconcile the needs and views of different target groups with the necessity to communicate a positive and credible company image (Gurau and McLaren 2003). Mission statements are ideally suited to act as branding propositions to various audiences as they are located at the interface between CI and corporate brand, actual and ideal self, internal and external corporate discourse. Indeed, internal and external audiences overlap to a considerable degree. As a consequence, much market-related communication can be understood as communication through which a company confirms and reinforces itself (Christensen and Askegaard 2001). In that sense, mission statements are among the vehicles by which autopoietic systems (Luhmann 1984: 59–63) refer to and reproduce themselves. The subsequent analysis will show in how far this self-referential nature of mission statements is likely to meet – or fail to meet – with its purpose of persuading stakeholders of the company’s inherent value. As mission statements originate within the social system they help to reproduce, they selectively draw on particular socio-cognitive representations that seem best suited to sustain or enhance that system and the identities and power relations within it. Branding can be seen as an essentially discursive process (Flowerdew 2004: 585) in which a set of core values or list of attributes is, by means of semiosis, translated into multimodal texts such as mission or vision statements, logos and brand claims. The ultimate aim of this process is to persuade people to buy a particular product or service or, in the case of corporate and city branding, endorse the brand and become loyal to it. With regard to the study at hand, these considerations give rise to the following research questions: What is the structure of the socio-cognitive models that are brands, and how are these encoded at the textual level? Further, how does the discourse of corporate branding spread to other social fields, such as local government? The following section will outline the data and methods that were employed to answer those questions. 3. Corpus and methods 3.1. The mission statements corpus The first case study is based on a corpus of corporate mission statements as representing a company’s ideal self. The mission statements in question were taken from the websites of the 2003 Fortune Global 500, a list of the largest companies worldwide that is ascribed great importance by the business community. In particular, the texts comprise the top 50 and the bottom 50

272   Veronika Koller companies, thus building a corpus of 29,925 words. Interestingly, the upper tier (i.e., numbers 1–50 of the Fortune list) shows a shorter average text length (286 words) than the lower tier (i.e., numbers 451–500), the texts in which are thus, at 401 words on average, more verbose. However, upper-tier companies are more likely to have self-presentation texts – such as a mission/vision/value statement – on their websites in the first place than are their counterparts from the lower tier (88 per cent compared to 78 per cent). As indicated by the slashes, mission/vision/value statements encompass a greater variety of texts, being generically heterogeneous in several regards. As for format, they range from simple no-frills text documents to audio files and sophisticated video presentations, with the former showing considerable variation in length by ranging from 28 to 1,358 words across the corpus. In terms of text type, mission statements come under the headings of “Our Vision”, “Message from the CEO” (a feature that is particularly popular with Japanese companies), “The essence of our company”, “Strategy” or, indeed, “Our Mission” (see Swales and Rogers 1995: 226). By using these headings interchangeably, companies corroborate Collins and Porras’s observation of mission and vision statements as “a muddled stew of values, goals, purposes, philosophies, beliefs, aspirations, norms, strategies, practices, and descriptions” (1996: 77). Importantly, using mission and vision synonymously, texts in the genre also blur the boundaries between actual and possible self. 3.2. The city branding corpus The corpus for the second case study consists of the community strategies and any community visions, logos and brand claims of 54 mid-sized UK cities and towns, i.e. those with a population of between 100,000 and 300,000 according to the 2001 Census.5 All councils have a community strategy and often also a vision, the latter of which is usually commissioned by the Council and drawn up by representatives of the Local Strategic Partnership6 together with external 5 The information pertains to cities and towns, not boroughs, and was compiled from the following three websites (all accessed 18 May 2007): www.statistics.gov.uk/ StatBase/Expodata/spreadsheets/D8271.xls (England and Wales), www.gro-scotland.gov.uk/files/setloc-ks01.xls (Scotland), www.nisranew.nisra.gov.uk/census/ Excel/ks_settlements/ ks01_com_st.xls (Northern Ireland). 6 The UK Local Government Act of 2000 requires all councils to have a Local Strategic Partnership with representatives from the police authority, the Primary Care Trust (an administrative unit of the National Health Service) and representatives of the private sector.

A case for cognitive Critical Discourse Analysis   273

consultants. That working pary will draft a document that then undergoes a process of public consultation, the results of which ideally feed into an adapted version. (In this sense, the process of drafting a mission/vision statement may be more important than the actual product.) Community strategies and visions can be see as the nodal points of “genre colonies” (Bhatia 2004: 57–84), a notion similar to Faircloughs’s “genre chains” (2003: 31–32) but not necessarily involving a temporal sequence. The dissemination of these strategic branding documents is twofold, translating into multimodal texts such as logos and brand claims and other genres of governance (Fairclough 2003: 32–34) that can be classified as implementation documents. An example of the latter would be a call to bid for the refurbishment of a neighbourhood. This shows how city branding has conrete material implications, which are very likely the most important aspect of the process to residents. The present case study focuses on logos and brand claims, as well as on the vision document for the metropolitan borough of Oldham in Greater Manchester. 3.3. Parameters for analysis The methods used to investigate the two corpora represent functional qualitative analysis of selected sample texts. The parameters selected are first and foremost attribution of the company or city, as well as its parts and people. This is intended to show the “character traits” of the brand personality. Formally, this part of the analysis includes comparative and superlative forms as well as relational process types that provide information about the qualities of the actor (Halliday and Matthiessen 2004: 170–175), e.g. “We are the industry benchmark for distribution services” or “Oldham has the poorest rail service”. In order to see in how far the brand is actually structured metaphorically through personification, the analysis further addresses individual and collectivized actors (van Leeuwen 1996), investigating to what extent the brand can grammatically be seen as an entity endowed with agency. This step also comprises a look at the use of personal pronouns. Thirdly, the analysis of modality and tense helps ascertain how the ought self is represented (deontic modality) and how likely it is seen to be that a transformation from actual to ideal self will take place (epistemic modality). The use of tenses, especially present perfect and present continuous, encodes the transition from one to the other. The parameters of metonymy and metaphor relate to metaphoric and metonymic expressions as surface-level instantiations of underlying conceptual metaphors. This part of the analysis is summative, intended to infer from the linguistic evidence if the socio-cognitive model that is the brand is structured

274   Veronika Koller metonymically, with the brand standing in for its stakeholders, or metaphorically, with the brand being conceptualized as a person in its own right. Finally, the analysis also addresses the visual aspect of the texts in the corpus, looking at the layout of documents and at logo composition to see how the brand model is transferred from one semiotic mode to the other. In total, the qualitative investigation of the sample texts is intended to infer the socio-cognitive representations prevailing in branding discourse, show the way in which companies instantiate these to project themselves as corporate brands, and in how far the discursive strategies of corporate branding have been transferred to discourses in the public sector. 4. Case study 1: Corporate mission statements In the following, two mission statements will be analyzed and compared, one from the top 50 group of the Fortune list and one from the bottom 50. The two companies in question are IBM, a provider of software consultancy services, and TPG (now TNT), a logistics and distribution company. Both texts date back to 2004. The excerpt chosen from IBM’s value statement is 603 words long, while the excerpt from TPG’s mission statement accounts for 701 words.7 Starting with attribution, the IBM text shows 17 adjectives describing the company and its employees, twelve of which are positive (e.g. “professional”, “smart”, “crystal clear”), while five are negative (e.g. “uncertain”, “bureaucratic”, “dysfunctional”). By contrast, eight of the ten attributes in the TPG text are positive (e.g. “successful”, “reliable”, “forward-looking”), while two are arguably positively connoted in a corporate context (“large”, “global”). Given the promotional nature of mission statements, it is somewhat surprising to find negative evaluation in the IBM text; however, this may point to a more differentiated self-image and thereby increase credibility, perhaps indicating a trend towards “anti-spin”. Notably, however, the negative attributes cluster at the beginning of narratives and are thus firmly lodged in the past. Both texts are similar in their use of hyperbole, encoded in superlatives (e.g. “most advanced”) and intensifiers (e.g. “highly trained”). Attribution can be realized by various means and it is here where the two texts differ again: While the IBM text uses lists of adjectives joined by conjunctions (e.g. “uncertain, conflicted 7 The two texts are or were available at http://www.ibm.com/ibm/values/us/ (accessed 18 August 2008) and http://www.tpg.com/abouttpg/tpgataglance/index.asp (accessed 4 November 2004), respectively.

A case for cognitive Critical Discourse Analysis   275

or hesitant”), the TPG text tends to use relational process types (e.g. “We are the industry benchmark”). This feature has the effect of driving up the number of occurrences of the corporate self as actor in the latter text. By far the most frequent actor in both texts is the first person plural, “we”. This is even more pronounced in the TPG text with its 46 occurrences of the pronoun, giving a density of 6.56 per 100 words, in contrast to the 17 occurrences in the IBM text accounting for a density of 2.82. Furthermore, the first person plural is almost always found in the actor position, only the TPG text shows six instances of the “we” as a goal of an action (e.g. “our customers entrust us to store, move and deliver their goods”). All other actors pale in comparison, pointing towards a narcissistic focus on the actions of the corporate self. Actor roles are ambiguous in that the predominant “we” refers to collectivized actors like “IBMers” but is backgrounded elsewhere. Thus, IBMers are later allocated third-person status, just as “we” is set apart from “our employees” in the TPG text. This leaves the reader to infer that “we” denotes another collectivized actor, namely the “board of management” mentioned in the TPG text. It seems that the principal of the text, i.e., senior management, avoids fully identifying themselves, making “we” a paradoxical actor that is simultaneously inclusive and exclusive (Dieltjen and Heynderickx 2001). One particularity of the IBM text is its use of direct address and first person singular (e.g. “I must tell you, this process has been very meaningful to me”). As the linguistic equivalent of significant eye contact, these phrases are meant to engage the reader and lend further credibility to the text. The first person singular is seen as engaging in mental-affective and verbal processes of thinking, feeling and talking to the reader in a pseudo-conversational interaction. This is mirrored in the signature to the text, which gives the name and function of IBM’s CEO, thus interdiscursively drawing on the letter genre to further personalize the text. Collectivized actors other than “IBMers” and “TPG people” are “our people” in general and more specifically “our customers”, “our employees” and “our investors”, i.e. the three main audiences at which the texts are aimed. The corporate brand is reflected in the companies’ names featuring as actors. While IBM’s actions are restricted to relational processes (e.g. “what IBM has always been about”), TPG can be seen to be engaged in a wider range of processes, including relational, behavioural and material ones. Crucially, however, TPG as an actor cannot think, feel or communicate, making it less likely that the socio-cognitive model is structured metaphorically. The excerpts are notable for their almost complete lack of epistemic modality, avoiding likelihood in favour of certainty. This is further conveyed by the pronounced use of deontic modality (including “we needed to affirm” [IBM]

276   Veronika Koller and “we must build trust” [TPG]). What epistemic modality there is is “typically of the unhedged variety” (Swales and Rogers 1995: 227), e.g. “delivering on the confidence they place in us […] will produce that trust” in the TPG text. Just like it had a differentiated use of evaluative attributes, the IBM text also shows a more complex use of both modal verbs – the TPG text restricting itself to “must” – and tenses: the IBM excerpt contains a meta-narrative about the processes involved in drawing up the values statement, which brings in simple past (e.g. “we needed to engage everyone”), next to future, present perfect and present continuous (e.g. “we are getting closer to what IBM has always been about”). The TPG text by contrast shows almost exclusive use of the simple present, often in repetitive paratactic sentences of the “We do X” formula. This static use of tenses is at odds with TPG’s proclaimed “aims” and “intentions”, which could have been realized using more dynamic tenses to show that the company is approaching its ideal self. The reason that TPG mentions its aims has to do with its position in the lower tier of the Fortune list, but its undifferentiated use of all linguistic parameters investigated also points to a lower quality of text production. At this point, it seems that the corporate brands oscillate between the metonymic structure of a collectivized actor and the metaphoric structure of an actor engaging in a number of processes, although crucially not mental-affective or verbal ones. On the level of language and layout, the companies are represented as living organisms, as indexed by metaphoric expressions such as “create”, “grow/growth” and “healthier”. Visually, this metaphor is reflected by the fact that the IBM value statement used to feature the stylized image of a double helix, implying that values are part of the company’s DNA. To sum up, the two texts are persuasive in intent due to the genre they represent, and thus employ a number of relevant devices, notably direct address and deontic modality. Nevertheless, these devices clash with the texts’ self-centred nature, evidenced by the dominance of the first person plural in actor position, and a detached managerial class unwilling to identify themselves. While a focus on the corporate self may be necessary to self-reproduce the discourse community in which the texts originated, it makes them meet only half of their communicative function. Also, both examples show a semantic field of emotion, centring on words such as “responsibility”, “trust”, “passion(ate)” and “pride”. This feature may not only compensate for the scarcity of affective process types in relation to the corporate brand, but also further acts in concert with the pseudo-personal interaction witnessed in the IBM text, blending the public and the private sphere. In as far as companies endow themselves with such additional emotional meaning, they strengthen their position in the social system of late capitalism.

A case for cognitive Critical Discourse Analysis   277

The current predominance of the corporate sphere backgrounds the formerly more powerful sector of government and politics. It stands to reason, then, that the latter should try and regain ground by adapting discursive strategies employed in corporate discourse. The next section looks at how this is happening in the phenomenon of city branding. 5. Case study 2: City branding Conceiving of cities as “corporate brands” throws into relief the changing uses of the word “corporation”. An etymological body metaphor, the word was originally used to refer to a municipal entity, before being transferred to business entities. It has now come full circle by being re-appropriated in the context of local government. Bourdieu (1994: 7) states that bureaucratic processes, educational structures and social rituals shape the cognitive models citizens hold about themselves, about the government and the relation between the two. Pagani (2007: 6) adds to this that “official discourses […] define the meaning of citizen, resident etc. and state”. For post-bureaucratic structures of governance, we can note that these adapt branding as a discursive process from the corporate sector for the same reason. City branding embodies two major trends in contemporary consumer capitalism. Firstly, it exemplifies the shift from material to semiotic and cognitive production: Companies these days do no longer primarily provide goods and services but rather use discourse to construct and disseminate the socio-cognitive representations that are brands. This trend has led to an almost complete reversal, where products are now seen by some as the mere materialization of brands (Askegaard 2006: 100). This involves the brand “personality” being reduced to a set of core values that are expressed by various semiotic means. Secondly, city branding is an instance of how corporate discourse and corporate genres have spread to the public sector (Mautner forthcoming), so that political entities such as nation states (de Michelis 2008) and cities (Flowerdew 2004) are branded, also pointing towards a commodification of public goods and services. Any brand communication is directed at multiple audiences, with the employees, customers and shareholders of corporate discourse mapping onto the residents, tourists and investors involved as target groups of city branding. The question arises whether one brand can reach both internal and external groups. While uniformity is probably unrealistic and even counter-productive (Parkerson and Saunders 2005: 246–247), over-differentiation of the brand can descend into chaos: Parkerson and Saunders’ (2005) case study of Bir-

278   Veronika Koller mingham shows how the three different organizations responsible for branding to different target groups are at odds with each other. However, a measure of brand complexity is unavoidable given that cities are not organizations but networks of organizations and individuals. This entails that brand uniformity cannot be enforced in the same way as it can in a company; rather, the city brand can be adopted by anyone who wants to be associated with it – or not, as the example of Blackpool shows, where the brand claim drawn up together with the logo has to compete with another claim preferred by the chief executive of the council. Brand claims are not a common feature to go with a city’s logo. However, examples include the following: (1) (2) (3) (4) (5) (6)

“Preston: England’s newest city” “Dundee: changing for the future” “Middlesbrough: moving forward” (used with a logo that includes two arrows pointing upward) “Oxford: building pride in our city” “Growing the right way for a bigger, better Peterborough” “A safer, cleaner, ambitious Nottingham: a city we’re all proud of”.

These claims share a number of features with the mission statements, such as the words “pride” and “proud”, an ambiguous “we”, the use of present continuous to indicate dynamic, ongoing change, and comparative and superlative forms. Comparatives are particularly favoured by companies in the lower tier of the corpus, but can be counter-productive to the promotional and persuasive purpose of the genre: After all, what does example (6) say about the present state of Nottingham? Moving on to city logos, these can be classified into five categories: –– logos with visual, but no pictorial elements (six instances, e.g. the Bolton logo using only colour but no image) –– landmark logos incorporating a well-known feature of the city (eleven instances, e.g. the Cambridge logo including a stylized picture of the cathedral) –– heraldic logos; these contextualize the city in its history (eleven instances, e.g. Blackburn) –– hybrid logos combining heraldic and abstract elements; these recontextualize the historic element (six instances, see Figure 1) –– abstract logos: these represent decontextualized brands (eleven instances; see Figure 2).

A case for cognitive Critical Discourse Analysis   279

Figure 1. Hybrid city logo

Figure 2. Abstract city logos

Abstract logos are the most brand-like, akin to the decontextualized image bank images discussed by Machin (2004), which are so generic as to be easily inserted into different contexts. As such, decontextualized images and logos represent a shift from denoting concrete entities and objects, e.g. in relation to landmarks or to history, to connoting a limited set of brand values. Thus, the swoosh-like elements of some hybrid and abstract city logos are likely to be intended to convey dynamism, while the equally popular composite elements, seen e.g.in the Sandwell logo, visually encode the idea of unity in diversity. Moving on to policy documents that contribute towards the construction and communication of the city brand, the remainder of this second case study will concentrate on the vision document for the borough of Oldham. This complements the community strategy, which not only has the generic title “Oldham: a place where everyone is proud to be belong” but also features pieces of a jigsaw puzzle as design elements, not unlike the Sandwell logo. The vision document, published in April 2004, is entitled “Oldham beyond”, pointing toward the ideal self of the city that is supposed to evolve out of its actual self. Available at http:// www.oldham. gov.uk/oldham_beyond_vision.pdf (accessed 18 December 2008), it was commissioned by the Local Strategic Partnership together with the Northwest Development Agency and drawn up by a team of consultants and external experts. It underwent quite a rigorous consultation process, involving 2,000 people and various techniques such as workshops, focus groups, work with schools, a touring discussion space and a postcard campaign. Voices of

280   Veronika Koller the consulted residents are incorporated into the document in the form of indirect speech. This comprehensive consultation process corroborates the claim that successful city branding and “successful regeneration can only be achieved trough genuine community engagement and local ‘ownership’ of the […] process” (Jones and Wilks-Heeg 2004: 342–343). With regard to the branding of Hong Kong, Flowerdew (2004: 596) further notes that “the objectives of the consultation are not only to solicit views, but also […] to develop awareness and consensus, promote ownership and secure acceptance”. The more comprehensive the consultation process, and the more its results feed back into the documents, the more likely the brand is to be a cognitive structure jointly held by members of the community, i.e. a socio-cognitive representation. In the vision document, different sections are dedicated to the historical Oldham, its actual and ideal self. In terms of attribution, we can see a rise in the number of different attributes, from 14 for the historical self, to 37 for the actual and 51 for the ideal self. This raises the possibility that the ideal self could be the most differentiated of the three. However, there is also a shift in the evaluative load of the attributes: While neutral attributes such as “Northern” stay largely the same over the three selves (21, 16 and 18 per cent), negative attributes, e.g. “weak”, account for roughly the same percentages (43 and 46, respectively) in the sections on Oldham’s historical and actual selves, but do not feature at all in its ideal self. Consequently, positive attributes rocket to 82 per cent there. This is unsurprising, given that an ideal self is the best possible self in the future, but begs the question of how differentiated and hence credible a brand can ever be. There is some overlap between the actual and ideal selves; notably, the attributes “ambitious”, “independent”, “strong” and “proud” are all repeated in higher numbers in the section on the ideal self, indicating that the actual self has the potential to transform into the ideal self. This future ideal is further reinforced by the positive attributes being realized in word classes other than adjectives alone, a feature that is unique to the ideal self. Thus, the seven instances of “attractive” are matched by a combined 33 instances of the verb (“attract”) and noun forms (“attraction[s]”), “proud” (six times) finds its counterpart in the nine occurrences of “pride”, “(self-)confident” (five times) is mirrored by 15 instances of the noun, and the eight tokens of “intercultural” are transformed into the neologism “interculturalism”, which is explicitly introduced as a new term and set in italics. The prominence of this last concept has also been observed for the rebranding of Leicester (Machin and Mayr 2007) and Liverpool (Jones and Wilks-Heeg 2004: 352). In both cases, and indeed for Oldham, multiculturalism stands in stark contrast to racial segregation, racially based social exclusion and, for Oldham, the race riots of May 2001. Here, the actual self is nowhere near the ideal self.

A case for cognitive Critical Discourse Analysis   281

It is striking that almost all actors in the document are collectivized, with only the three external experts involved in the branding process being mentioned by name. Other than that, actors are “sector/s” (21 times), “council” (23), “community/ies” (92), “town” (136), “people” (137), “borough” (294) and of course “Oldham” (376). These collectivized actors are represented with two opposing linguistic strategies; in examples such as “the Council were successful in promoting a range of building projects”, the capitalization individualizes the actor while the plural form points towards the group of people constituting it. Crucially, all collectivized actors engage in a range of process types; however, these are not distributed evenly. While the council engages mostly in relational and material actionns, the people can be seen as actors in relational and affective processes. “Oldham” has the widest range, but, like the companies of the first case study, cannot think or feel. Those actions are reserved for human actors. Modality also mirrors the corporate genre, in that we find high-affinity deontic and, to a lesser extent, epistemic modality, encoded in the modal verbs “must”, “should”, “can” and “will”. The latter crosses over into tense, with the use of tenses in the document being in line with the three selves. Thus, the historical self is related to simple past and present continuous (e.g. “Oldham has always been an industrial town”), the actual self is expressed through simple and continuous present as well as present perfect again (e.g. “Bus services have recently been improved”), and the ideal self, being a presentation of the self in a future state, is linked to future and, interestingly enough, future perfect. Again, present perfect and present continuous indicate change and transition from the actual to the ideal self and thus reinforce the shift observed in the evaluative load of attributes. Another device to the same end is suffixation of attributes; while the actual self is described with a number of de- attributes (“declining”, “degraded”, “derelict”), the ideal self is encoded with re- attributes such as “refurbished”, “revived”, “reborn”, the latter of which is again reinforced through its etymological noun equivalent “renaissance” with no fewer than 22 occurrences. As for the layout of the document, suffice it to say that it is typical of its genre, with large computer-generated images of Oldham in 15 years’ time at the top of pages (the “ideal” space according to Kress and van Leeuwen 2006: 186), a number of diagrams and charts, bullet point lists and colour used for highlighting. The dominant colour of the vision document is dark blue, which shows a lack of coherence with the colours of the logo that were used at the time that the vision document was published.8 It was composed of the 8 Oldham underwent a rebranding exercise in 2007/2008 in the course of which the logo was changed to the one now available at http://www.oldham.gov.uk/. The discussion in this paper pertains to the previous logo, which is no longer used.

282   Veronika Koller light blue text, set in one line, “Oldham together”, with the borough name in bold type and the two words separated by an orange swoosh pointing to the right. Underneath the word “together”, the then slogan “the outlook’s bright” could be found in a smaller, orange type. Meant to represent an owl and thus integrate Oldham’s historical heraldic animal, the logo is nevertheless more abstract than hybrid (see the logo of Preston for contrast, Figure 1), containing a “swoosh” as a deconextualized element connoting dynamism. The arch further reinforces the vision document’s future orientation by pointing toward the right, i.e. the “new” space on the page in Western cultures (Kress and van Leeuwen 2006: 179–185). This future orientation is further underscored by the metaphorically used “outlook”, while the equally metaphorical “bright” is literalized through the orange colour. Speaking of orange, the mobile network provider of the same name uses the brand claim “the future’s bright”. This may be coincidence but shows how generic and therefore similar corporate and city brand claims have become. Comparing corporate and city brands, there are significant overlaps in the use of attribution, modality and collectivized actors. Further similarities can be found in layout and logo composition. Partial differences, on the other hand, concern the use of the first person plural; this is unspecified in the mission statements and the city brand claims, but specified in the Oldham vision, where it refers to the authors of the document. The difference in the use of tenses is mainly due to the unimaginative and repetitive use of the present simple in the TPG mission statement. It has been noted that material implications such as sports and cultural facilities, quality of the housing stock, pedestrian zones and rubbish collection are the most salient aspect of city branding for residents. Indeed, as Fairclough, Pardoe and Szerszynski (2006) have noted, material objects, along with discursive and other social practices (e.g. consultations, voting) contribute towards citizenship, understood in this paper as a structured socio-cognitive model of the self in relation to the abstract entity of a city or nation state. With regard to Birmingham (Parkerson and Saunders 2005: 258–259) and Bradford (Trueman, Klemm and Giroud 2004), urban regeneration has been discussed as improving brand perception and increasing brand equity. A more critical study of Liverpool, however, found that material implications, while improving tourists’ and investors’ attitude, can also work against residents, e.g. by restricting leisure time activities and stepping up security to cordon off the most tourist-intensive areas, while allowing non-core residential areas to degenerate (Jones and Wilks-Heeg 2004: 356–357). Whatever their advantages and disadvantages for different stakeholder groups, material implications of branding matter and are indeed given much space in the Oldham vision. This

A case for cognitive Critical Discourse Analysis   283

is in clear contrast to the limited set of intangible and virtually interchangeable corporate brands. On the whole then, the case study suggests that cities have been reconceptualized as corporate brands as far as their nature as networks allows, but no further. 6. Implications for cognitive Critical Discourse Analysis The implications of the two case studies are twofold, concerning the cognitive structure of brands and the role of socio-cognitive representations of the self in discourse. As mentioned above, brands as socio-cognitive models seem to be both metonymically and metaphorically structured. As far as they are a collectivized actor on par with others, they are in a contiguity relation with their stakeholders. Evidence for this is the shared attributes, process types and modality between the two. However, in that brands are linguistically represented as individual actors, they are metaphoric, being in a similarity relation with stakeholders and reflecting the “model citizen” (Machin and Mayr 2007), or indeed employee or customer. Evidence includes the fact that some attributes are reserved for non-human actors (e.g. “attractive”, “successful”) and that brands are actors in not only relational, but also material (if not mentalaffective) process types. On a related note, brands are in both cases conceptualized as living organisms (e.g. “Oldham greew up as a satellite of Manchester”). In view of the evidence it seems that the cognitive structure of brands shows features of both metonymy and metaphor, with the first being slightly more pronounced. Cognitive Critical Discourse Analysis is an approach to the interpretation of results as much as it informs the parameters of analysis: If the socio-cognitive models that are brands are assumed to embody traits of the self, scripts for its interaction with others and beliefs about likelihood and obligation, the analysis will address attribution, actors and process types, and modality. A cyclical approach starts with the textual evidence, infers the underlying cognitive model and then seeks to corroborate this model through further text analysis. In general, cognitive CDA rests on the notion that people live in particular material realities, engage in certain practices and act according to specific socio-cognitive representations about themselves, others and the world in general. Some of these people, by dint of their social roles, are in the position to produce and distribute more or less influential texts. Texts are instances of discourse which are realized through a set of linguistic and semiotic devices that stabilize into genres. Texts are also vehicles for their producers’ cognitive models. Other people receive these texts and doing so repeatedly under similar conditions of reception is likely to impact on their practices, material reality and indeed

284   Veronika Koller cognitive models. From a cognitive, discourse analytical viewpoint, such renegotiation of socio-cognitive representations is effected through intertextual chains, in which the respective models are recontextualized and possibly reinforced in discourse. This impact may be one of the producers’ intentions with the text, preferably to align their cognitive models with those of the recipients in order to secure compliance. This chapter has investigated the discursive construction of ideal selves in different contexts under the conditions of consumer capitalism to uncover self-serving bias and self-centredness in language. However, the approach can equally be applied to discuss other social issues on the basis of linguistic evidence, and it has been one of the aims for this article to foster further research into relevant areas.

References Abrams, Dominic 1999 Social identity, social cognition, and the self: the flexibility and stability of self-categorization. In Social Identity and Social Cognition, Dominic Abrams and Michael A. Hogg (eds.), 197–229. Oxford: Blackwell. Askegaard, Søren 2006 Brands as a global ideoscape. In Brand Culture, Jonathan E. Schroeder and Miriam Salzer-Mörling (eds.), 91–102. London: Routledge. Augoustinos, Martha, and Iain Walker 1995 Social Cognition: An Integrated Introduction. London: Sage. Bhatia, Vijay K. 2004 Worlds of Written Discourse. London: Continuum. Bourdieu, Pierre 1994 Rethinking the state: Genesis and structure of the bureaucratic field. Sociological Theory 12 (1): 1–18. Brewer, Marilynn B., and Wendi Gardner 1996 Who is this “we”? Levels of collective identity and self representations. Journal of Personality and Social Psychology 71 (7): 83–93. Christensen, Lars Thoger, and Søren Askegaard 2001 Corporate identity and corporate image revisited: A semiotic perspective. European Journal of Marketing 35 (3–4): 292–315. Collins, James C., and Jerry I. Porras 1996 Building your company’s vision. Harvard Business Review 74: 65–77. Davidse, Kristin 2002 Nominative and oblique in English: reflexive clauses as a test case for distinct Agent-Patient models. In The Nominative and Accusative and Their Counterpart, Kristin Davidse and Béatrice Lamiroy (eds.), 131–

A case for cognitive Critical Discourse Analysis   285 173. (Case and Grammatical Relations Across Languages 4) Amsterdam: Benjamins. De Michelis, Lidia 2008 Britain™ and “corporate” national identity. In Language and Communication in the Public Sphere, Ruth Wodak and Veronika Koller (eds.), 203–222. (Handbook of Applied Linguistics, Vol. 4) Berlin: de Gruyter. Dieltjen, Sylvain, and Priscilla Heynderickx Evasive actions in top-down communication: Strategies for avoiding di2001 rect sender and receiver references. Document Design 2 (2): 210–219. Fairclough, Norman 1995 Critical Discourse Analysis: The Critical Study of Language. London: Longman. 2003 Analysing Discourse: Textual Analysis for Social Research. London: Routledge. Fairclough, Norman, Simon Pardoe, and Bronislaw Szerszynski 2006 Critical discourse analysis and citizenship. In Analysing Citizenship Talk, Heiko Hausendorf and Alfons Bora (eds.), 98–123. (Discourse Approaches to Politics, Society and Culture 19) Amsterdam: Benjamins. Flowerdew, John 2004 The discursive construction of a world-class city. Discourse and Society 15 (5): 579–605. Gee, James P., Glynda Hull, and Colin Lankshear The New Work Order: Behind the Language of New Capitalism. St. Le1996 onards, NSW: Allen and Unwin. Grauel, Ralf 2000 Die Macht der Marke. Brand eins, March: 16–24. Mitarbeiter des Monats: Der Kunde. Brand eins, November: 18–20. 2004 Gurau, Calin, and Yvonne McLaren 2003 Corporate reputations in UK biotechnology: An analysis of on-line “company profile” texts. Journal of Marketing Communication 9 (4): 241–256. Halliday, Michael A.K., and Christian M.I.M. Matthiesen An Introduction to Functional Grammar. 3d ed. London: Arnold. 2004 Jones, Paul, and Stuart Wilks-Heeg Capitalising culture: Liverpool 2008. Local Economy 19 (4): 341–360. 2004 Kapferer, Jean-Noël 2002 Corporate brand and organizational identity. In Corporate and Organizational Identities: Integrating Strategy, Marketing, Communication and Organisational Perspectives, Bertrand Moingeon and Guillaume Soenen (eds.), 175–193. London: Routledge. Kress, Gunther, and Theo van Leeuwen Reading Images: The Grammar of Visual Design. 2d ed. New York: 2006 Routledge

286   Veronika Koller Kristiansen, Gitte 2008 The group as a variable in the development of cognitive schemata. In Body, Language and Mind, Vol. II, René Dirven, Roslyn Frank, Tom Ziemke, and Jordan Zlatev (eds.), 409–432. Berlin/New York: Mouton de Gruyter. Kunda, Ziva Social Cognition: Making Sense of People. Cambridge, MA: MIT Press. 1999 Luhmann, Niklas 1984 Soziale Systeme: Grundriß einer allgemeinen Theorie. Frankfurt a.M.: Suhrkamp. Machin, David 2004 Building the world’s visual language: The increasing global importance of image banks in corporate media. Visual Communication 3 (3): 316–336. Machin, David, and Andrea Mayr Antiracism in the British government’s model regional newspaper: The 2007 “talking cure”. Discourse and Society 18 (4): 453–477. Markus, Hazel R., and Paula S. Nurius 1986 Possible selves. American Psychologist 41: 954–969. Mautner, Gerlinde forthcoming The spread of corporate discourse to other social domains. In Language and the Market, Helen Kelly-Homes and Gerlinde Mautner (eds.). Basingstoke: Palgrave. Moscovici, Serge 2000 Social Representations: Explorations in Social Psychology. Cambridge: Polity Press. Muniz, Albert M., and Thomas C. O’Guinn Brand community. Journal of Consumer Research 27: 412–432. 2001 Pagani, Giulio Expressions/representations of citizenship between the “state” and the 2007 “citizen”: Register analysis of local government discourse. Critical Approaches to Discourse Analysis Across Disciplines 1 (1): 1–18. http:// www.cadaad.org/ejournal/2007/1/pagani Parkerson, Belinda, and John Saunders 2005 City branding: can goods and services branding models be used to brand cities? Place Branding 1 (3): 242–264. Swales, John, and Priscilla S. Rogers Discourse and the projection of corporate culture: The mission state1995 ment. Discourse and Society 6 (2): 223–242. Trueman, M., M. Klemm, and A. Giroud 2004 Can a city communicate? Bradford as a corporate brand. Corporate Communications 9 (4): 317–330. van Dijk, Teun A. 2003 The discourse-knowledge interface. In Critical Discourse Analysis: Theory and Interdisciplinarity, Gilbert Weiss and Ruth Wodak (eds.), 85–109. Basingstoke: Palgrave.

A case for cognitive Critical Discourse Analysis   287 van Leeuwen, Theo The representation of social actors. In Texts and Practices: Readings 1996 in Critical Discourse Analysis, Carmen Caldas-Coulthard and R. Malcolm Coulthard (eds.), 32–71. London: Routledge.

Chapter 11 Distributed cognition and play in the quest for the double helix L. David Ritchie

1. Introduction I have never seen Francis Crick in a modest mood. Perhaps in other company he is that way, but I have never had reason so to judge him. It has nothing to do with his present fame. Already he is much talked about, usually with reverence, and someday he may be considered in the category of Rutherford or Bohr. But this was not true when, in the fall of 1951, I came to the Cavendish Laboratory of Cambridge University to join a small group of physicists and chemists working on the threedimensional structures of proteins (Watson 1968: 16).

The discovery of the double helix structure of the DNA molecule is a prime example of both extended cognition (Clark 1997) and distributed cognition (Hutchins 1995). In large part due to the refreshingly candid (inappropriately candid according to some critics) autobiographical reminiscences of James Watson (1968), one of the discoverers, it also provides an unusual insight into the social and emotional experience of science. Watson’s account stimulated several others to write their own accounts, the cumulative result of which is a fascinating study in the retrospective reconstruction of culturally significant events. In his controversial account of the events leading up to the discovery, Watson shuns the more conventionally heroic reconstruction of a great discovery as an orderly process of reasoning from known facts to new observations, then to logical conclusions. Watson chooses instead to focus on the “very human events in which personalities and cultural traditions play major roles,” and to “convey the spirit of an adventure characterized both by youthful arrogance and by the belief that the truth, once found, would be simple as well as pretty” (p. ix). Watson approaches the telling of the story just as he would have us believe he approached the discovery itself, in a grand spirit of play – heedless at times, often thoughtlessly cruel, but always true to the spirit of adventure for its own sake.

290   L. David Ritchie As such, Watson’s account affords a singularly valuable opportunity to investigate the interaction of two senses of the phrase, “social cognition.” In the sense of cognition that is accomplished in part by way of communication within a social network, social cognition is very similar to what Hutchins (1995) and Clark (1997) call “distributed cognition.” In the more conventional sense of cognition about social relationships and processes, social cognition is essential to the constitution and maintenance of the social network itself. In this essay I use Watson’s biographical account of the discovery of the double helix to argue that these two senses, social cognition and social cognition, are complementary aspects of a single complex process, and to explore how play (in both the sense of competitive play and the whimsical or atelic sense of delight in an activity for its own sake) can contribute to both aspects of social cognition. In the following sections I will provide a brief discussion of the concepts of extended and distributed cognition. I will then provide a brief explanation of the multiple concepts of play and their place in human social and intellectual life. Because Watson’s account was written in what appears to be a spirit of playfulness – but was not universally received in the same spirit (see, for example, Sayre 1975; Wilkins 2005), the language he used, both his phrasing and his choice of a mock-heroic presentation, will also require attention throughout. I will also examine the social cognitive processes involved in the discovery of the double helix, drawing on James Watson’s account, with frequent comparisons to other accounts, which have been subsequently written from other perspectives. 2. The madness to my method Before turning to theoretical and conceptual issues, a methodological caveat is in order. Research with human beings and their relational interactions is like peering through a warped lens into an unknown and distant landscape. We can never know exactly how the lens is warped, so we can never know exactly how the view it affords us has been distorted. Part of the game is to attempt to understand the lens even as we attempt to understand what we believe we see through it. Thus, if we are to rely on an account which Watson himself (1968: ix) admits to be, at least in places, “one-sided and unfair,” then we must devote some attention to the account itself, even as we attempt to understand what it reveals to us. Fortunately, we have other lenses, each distorted in a different way, which we can compare with Watson’s account. We can also examine the form and language in Watson’s own account from a more general theoretical perspective. The objective is not to construct anything like a “true” account of what actually

Play in the Quest for the Double Helix   291

happened, but rather to gain insight into the cognitive and social processes of people, both as they engaged in an extremely difficult cognitive endeavor, and as they retrospectively reconstructed their social and cognitive processes. 3. The Double Helix: Background In 1951 James Watson, a brash young biologist from the American Midwest, sporting a newly-minted PhD and a combination of playfulness, open curiosity and burning ambition, invaded the relatively staid world of post-war British academic science, joined forces with Frances Crick, an older and more experienced, but equally brash English physicist-turning-biologist, and set out to capture the (figurative) prize of discovering the structure of DNA and the (literal) prize subsequently handed out by the Nobel Foundation to the discoverers. Sixteen years later, in response to the urging of several colleagues, Dr. Watson set forth a lively, very unconventional, and at times scandalous account of the discovery and various events leading up to it. Unlike many accounts of scientific discovery, Watson’s account makes no attempt to depict scientists, their activities, or their relationships in lofty or heroic terms, or to soften the edges of his own youthful prejudices and indiscretions. Indeed, Watson’s account has been severely criticized for the excessive candor with which he recalls both his own foibles and that of his fellow adventurers – and for the careless cruelty with which he describes the male scientists’ (and particularly his own) treatment of Rosalind Franklin. The Double Helix combines elements of a picaresque novel with elements of a buddy novel; on another level it reads like a good mystery (Crick 1988). The author disarmingly confesses his own laziness about learning basics such as organic chemistry (and the technical ineptitude that, he tells us, got him barred from the organic chemistry lab). He also confesses to being distracted by thoughts of Rosalind Franklin’s physical appearance when he should have been taking notes on her lecture, forgetting crucial details because he had not learned enough of the underlying physics and mathematics to recognize their importance, and neglecting the research specified by his fellowship grant while pursuing a completely different line of inquiry. These confessional asides, sprinkled throughout the book, can be interpreted as modesty – or as a form of bragging-by-understatement. Either way, they are quite contrary to our view of the scientist as a genius devoted solely to amassing facts and forming them into brilliant theories, who rarely if ever stumbles or falters. On another level, Watson’s racy account is wholly consistent with Clark’s (1997) contentions that the human brain excels not in its ability to solve problems

292   L. David Ritchie but in its ability to organize the world into the kind of problem it can solve. Even if we allow for some degree of post-adolescent bragging, it still supports a remarkably original account of intellectual work and its relationship to everyday life. 4. Extended and distributed cognition In recent years there has been increasing criticism from within cognitive science of the classic view of mind as essentially disembodied and communication as essentially a process of encoding, transmitting, and decoding determinate meanings. Until recently the standard view, at least within the cognitive sciences, represented human cognition as a process that is distinct from and logically separable from both the human body and the surrounding environment. The obvious metaphoric comparison is with a computer CPU, which receives input from sources such as keyboard and mouse, retrieves data from a “memory” device, performs certain operations, and transmits “output” to the memory device or to a destination such as a CRT or printer. Pursuing this metaphor, the logical operations conducted in the human brain are thought to be independent of the actual biology of the brain, with the implication that the cognitive essence of humans, their minds and personalities, could be “downloaded” into a more durable and less vulnerable medium with no difference in functionality. Similarly, social interactions are metaphorically mapped, within this model, onto “networked” computers, each performing its computations independently of the others, then transmitting and receiving only the completed products of these computations. This model has recently come under criticism from several sources. Barsalou (1999, 2007) argues forcefully against a primarily amodal theory of human cognition, and proposes in its place a model based on a conceptual neural system that parallels and partially simulates the activities of the perceptual neural system. Barsalou’s model of cognition as perceptual simulation is supported by extensive investigation of neural activation during language processing (Gibbs 2006). Research in cognitive linguistics has raised serious doubts about “code” and “conduit” metaphors for language (Reddy 1993), and produced evidence that much of language is itself fundamentally ambiguous. What is emerging from the rapidly accumulating body of evidence is a model of language and communication generally as ambiguous, biological and embodied. Andy Clark (1997) argues against the implicit isolation of the brain from its physical environment, and proposes instead that human cognitive activities integrate at every step not only the extended physical body but also the physical features of the external environment. Clark observes that the human brain is in

Play in the Quest for the Double Helix   293

fact not very good at computation, and suggests that what the brain does best is rearranging the external environment in a way that is suited to the its own limited capacities, in a process of extended cognition. An innovative study by Hutchins (1995) lays a foundation for extending these concepts into the realm of social interaction. Based on observation of the bridge navigation team on a ship of the U.S. Navy, Hutchins showed that the knowledge and cognitive processes required both for routine computation of the ship’s position and for recovering from a lost gyrocompass crisis were distributed across the entire team. Building on Hutchins’s work as well as on his own observation of complex motor control processes, including those of a human infant learning to walk, Clark (1997) proposes that human cognition be defined as both extended beyond the biological individuals, to include our cognitively-motivated re-organizing of the physical environment, and distributed across collaborative work groups. The phrase, social cognition, is ordinarily taken to mean something like individual cognition about social relationships, but the work of Clark, Hutchins, and others in the forefront of contemporary cognitive thinking suggests that it can also be meaningfully taken to mean something like cognition that occurs across and by means of social interactions. If cognition does occur, at least in some circumstances, by means of communication within social relationships, then the cognitive processes through which these social relationships are established and maintained are integral to distributed cognition. Thus, the two senses of the phrase, social cognition, must be understood as parts of a single complex concept, and understanding the cognitive or “computational” processes requires equal attention to the social and communicative processes. 5. Social cognition in maintenance of extended cognition Dunbar (1996) proposes a view of language, and of the evolution of language, that turns conventional views upside-down, suggesting that the exchange of information is secondary, and the maintenance of social structure is primary. Dunbar begins with grooming behavior in other primates, particularly chimpanzees. He notes that chimpanzees and baboons, for example, use grooming to build and maintain social coalitions, and that this one-on-one physical grooming supports primary groups with a maximum size of around 50. Citing findings that 60–70 % of all conversations, among men as well as women, is centered around “gossip” (conversation about relationships and about other people), and that the size of primary groups among humans, in virtually all cultures, is around 150, Dunbar suggests that a primary function of language is to allow humans to engage in grooming-like behavior with up to three other

294   L. David Ritchie people at once and, equally important, to acquire information about social structure within an extended group through gossip – conversation about others’ relationships. This conclusion, Dunbar claims, is also supported by the fact that the ratio of neo-cortex to body mass in humans is about three times the ratio among chimpanzees and baboons. Dunbar claims that language evolved primarily for its social function, its ability to extend the social-facilitative and organizing function of grooming in support of much larger primary groups. However, Dunbar’s evidence is equally consistent with a number of other accounts of language evolution, including the idea that language evolved as a by-product of the development of cognitive abilities including abstract conceptualization and a “theory of mind,” and was pressed into service as an extension of physical grooming behaviors. The fossil record affords at best only indirect evidence to support or contradict these various accounts of cognitive evolution, and for our purposes what matters most is that language – talk – apparently does support a complex array of processes, including both social organization and the transmission and exchange of data. 5.1. Play Of no less importance, language also supports an extension and elaboration of a cognitive function that is observed in all mammals as well as many other species: play. Watson describes many instances of what seems very much like play, and the tone of his account is playful throughout. Play is itself a complex concept. Playful behaviors, like grooming in other primates and gossip among humans, often serve social facilitative and social organizing functions. Among other mammal species, mock combat not only prepares the juvenile for the more serious combat in which adults establish and maintain a position in the group’s social hierarchy (and secure mating opportunities); it may also establish the juvenile animal’s position in an emergent hierarchy that will carry over into adulthood (Fagen 1995; Bateson 2005). Among humans, competitive games, particularly competitive team games, help the child develop and practice a behavioral repertoire that will be of obvious use in the modern capitalist economy. By turns, the child may practice being both a “team player” and, perhaps, a “leader” (Cook 2000). But play, like grooming, is also intrinsically rewarding. Just as grooming behavior among other primates releases pleasurable opioids that reinforce the social bonding effect, playful behavior among all animal species also release opioids: the chance to play is itself a motivation that can be as effective as food in conditioning experiments among rats and other laboratory species (Fagen

Play in the Quest for the Double Helix   295

1995). Among humans, smiling and laughing also release reinforcing opioids, and both joking and wordplay figure prominently in the social-facilitative talk Dunbar calls “gossip.” Theories of play have tended to focus on young children (e.g., Bateson 2005), and emphasize the extrinsic, survival-related purposes of play (e.g., Fagen 1995; Bateson 2005). Among human juveniles, we can add to this list of skills built by play the rehearsal of language and other communicative skills, and the honing of cognitive skills such as object constancy and perspectivetaking (Cook 2000). Discussions of adult play have tended to emphasize alea (games of chance) and agon (struggle), the competitive elements in play (e.g., Huizinga 1955; Malaby 2006, 2007) at the expense of whimsical, “just for fun” elements (whimsy). Cook (2000) identifies several other categories of play, including mimicry and ilinx, the feeling of vertigo or giddiness that accompanies, for example, riding a merry-go-round or sky-diving, as well as the feeling of “flow” theorized by Csikszentmihalyi (1980), that comes from engaging wellhoned talents and skills. At least among humans, both children and adults play alone as well as socially; solitary play includes elaborate fantasizing, solving puzzles, playing with objects, and even engaging in word-play (Apter 1991). Other primates have also been observed engaging in solitary play, for example enacting nurturing type play with stones, sticks, and other inanimate objects, manipulating both natural and (in zoos and animal research centers) human-provided objects. Humans are by no means the only species to use toys for the intrinsic pleasure they provide, but we excel at devising elaborate toys, and we have extended the invention of toys to include the playful distortion of language. Language play, by adults as well as infants, includes exploitation and distortion of every feature of language, phonology, lexis, and grammar (Carter 2004; Cook 2000). Poetry makes use of phonology in its use of rhyme, alliteration, and rhythm, of lexis and grammar in double entendres as well as in metaphors, metonyms, and irony. Each of these elements is also apparent in the nonsense rhymes, nursery rhymes, and other word play heard on playgrounds everywhere, and in the playful talk of adults. Puns, like poetry, require that words be chosen, not so much for their meaning as for their formal qualities, often independent of meaning. In nonsense rhymes such as “The Owl and the Pussycat,” discussed at length by Cook the entire “story” is developed according to the requirement of the rhyme scheme rather than following any objective sense. “The fiction thus created… seems to incorporate a wild and random element, to be controlled by language itself rather than by reality or the will of the writer” (Cook 2000). Adoption of a rhyme or alliteration scheme, a meter, or any other formal “rules” for composing a narrative, poem, or even a joke inverts the usual

296   L. David Ritchie relationship of language to reality and creates a new realm of possibilities and imagination. Cook suggests that, as language is more fully mastered, its possibilities become more constrained by meaning, and claims that this explains at least in part the sense of adulthood as a time when the magic of childhood is lost. But this is perhaps an overly dreary view of adulthood: It is equally plausible to view adulthood as a time when we constrain and elaborate the “magical” subversion of language and logic by inventing ever-more complex rules for our subversions of ordinary meaning and sense, and use these rules not to suppress but to enhance the “magic” of playful reality-subversion (Ritchie 2005; Ritchie and Dyhouse 2008). Extending Cook’s idea about the subversion of ordinary reality through play in a different direction suggests that playfulness – not merely competition, but also whimsy – may contribute in important ways to creativity. This is immediately apparent in language, but it also applies to other forms of play, such as puzzles and other object play. Lakoff and Johnson (1980) argue that complex concepts are often built up by combining simple conceptual metaphors that are themselves based on experienced correlations between different kinds of sensory perceptions. The kinds of language play Cook describes, in which ordinary reality is deliberately subverted and new realities created through application of incidental rules (rhyme in the case of “The Owl and the Pussycat,” sound in the case of alliteration or puns), can be extended to the other arts, such as dance or architecture. But I will argue in the following that a similar kind of subversive element may be present in many forms of “intellectual play” beyond language – including the “intellectual play” of mathematicians and theoretical scientists (see for example Byers 2007). Play with objects provides a missing impetus for Andy Clark’s ideas about reconstructing the world in a form our brains can compute – simultaneous play with objects and with ideas, play for its intrinsic pleasure, provides a mechanism that dispenses with teleology. Object play is probably not coincidental to, but rather integral to creative thinking. There has been a tendency among play and humor theorists to emphasize the competitive, even combative, elements in adult play, as if somehow that makes play activities more acceptable as an occupation for adults (Malaby 2006, 2007). Adult play often does have a strongly competitive element, as demonstrated by card games, board games, and organized sports. But Kohn (1986) cites extensive evidence that humans, including adult humans, often prefer forms of play in which the competitive element is minimized or suppressed altogether. Moreover, humans also play with competition itself, and indeed with every aspect of our life, including relationships, religion, and, I will argue, science.

Play in the Quest for the Double Helix   297

One final caveat: The word, play, has a benign sound and evokes a sense of joy and pleasure, but play is neither necessarily nor exclusively benign. The play of a cat with a bird or mouse has a decidedly cruel edge to it, and we humans are also equally capable of playing with other people’s lives and, much more frequently, with their emotions and their social reputations. Watson’s use of the name Rosie for Rosalind Franklin seems to be motivate far more by a spirit of playfulness than of spite, but the play is within his own team, at the expense of Dr. Franklin, and for all its playfulness is nonetheless potentially quite cruel in its effect (Sayre 1975). 5.2. Perceptual Simulators and Play Barsalou (1999, 2007) proposes that thinking is accomplished by partial simulations of actual perceptions; for example, a person thinks of a cat by simulating perceptual features strongly associated with the concept of cat, based on past experiences with cats. Perceptual simulators are organized into schemas on the basis of correlations among underlying perceptual experiences; for example, a cat schema includes links to simulators of sound (purring and so forth), the feeling of a cat’s fur, claws, weight, and body warmth, emotional responses to previously encountered cats, and narratives involving cats. The cat schema also includes links to the word “cat,” and weaker links to associated words and concepts, such as “feline,” “pet,” and “carnivore.” Hearing or glimpsing a cat partially activates the schema, including the word “cat.” Conversely, recognition of an animal as a cat or not a cat involves partially activating perceptual simulators linked to the cat schema and comparing the ensuing simulations to the actual perceptions. Language use routinely involves many areas of the brain, and there is substantial evidence that simulations of muscular action and internal states, along with perceptions of pressure, vision, and so on, are activated as relevant language is processed (Gibbs 2006). Conversely, these perceptual sensations are also interconnected in elaborate networks or schemas with each other and with related language (Barsalou 2007). Schemas for commonly-encountered objects and experiences include simulators for associated qualities, emotional and physiological responses, aspects of perception, and cultural associations. When a word or concept is encountered, many if not all of the associated simulators, including more or less weakly connected words, are briefly activated; those that are not relevant in the current context (do not readily connect to other schemas currently activated in working memory) are suppressed before we can even become conscious of them (Gernsbacher et al. 2001; Kintsch 1998).

298   L. David Ritchie Expressive language, including word-play and humor as well as metaphors, may activate a multitude of emotional and introspective simulations – often these may figure more importantly in the way people interpret utterances than the primary meanings (Ritchie 2006; 2008). Words, phrases, and other language elements may be chosen because of their ability to activate one set of simulations, but they may activate a different and unanticipated set of simulations in some hearers and readers, often with consequences that may not have been either intended or desired by the speaker or writer. A key aspect of figurative language, such as metaphor, is the enhanced activation of secondary associations, which then connect with other concepts in working memory in novel ways (Ritchie 2006; 2008). 6. The Double Helix: Science as play As Watson admits in his preface, The Double Helix is incomplete and onesided, but it nevertheless provides a surprisingly detailed account of the intricate warp and woof of social and cognitive processes through which the intricate puzzle of the double helix was solved. In this section, I will consider the evidence furnished by Watson’s account in detail. In following sections, I will review the evidence from retrospective accounts by other participants, some of which supports and some of which contradicts various aspects of Watson’s narrative. I will then return to the question of how social cognition, in the sense of cognition about social relationships, interacts with social cognition in the sense of socially distributed information processing. 6.1. The narrative The spirit of play – including Watson’s playful self-deprecation as well as his manner of mixing playful insults with compliments in his descriptions of others – is apparent from the very beginning of The Double Helix. In the Preface, apparently written after initial drafts of the book had elicited a storm of criticism that nearly led to cancellation of the book, Watson describes the narrative as an attempt “to re-create my first impressions,” and admits that the result is neither as objective nor as fair as a more balanced narrative might be, but justifies his approach as conveying “the spirit of an adventure characterized both by youthful arrogance and by the belief that the truth, once found, would be simple as well as pretty” (1968: ix [emphasis added]). The spirit of adventure pervades the entire account, and in many ways places it more in the genre of Treasure Island or Huckleberry

Play in the Quest for the Double Helix   299

Finn than of conventional science writing. The resulting blend of science biography with adolescent adventure yarn is itself an example of language play, play with both literary and intellectual forms, play that deliberately violates standard assumptions about intellectual biography and, as a consequence, also violates standard assumptions about intellectual work. Indeed, it seems very likely that Watson’s inversion of the ordinary assumptions about scientific autobiographical writing is at least partly responsible for the negative reactions to the book. Repeatedly, Watson frames the enterprise as an extended game. He introduces and emphasizes a spirit of competition, for example on page 37, “Within a few days after my arrival, we knew what to do: Imitate Linus Pauling and beat him at his own game.” But he also freely admits that other players did not seem to be nearly as obsessed with the competitive spirit of the endeavor: “Rosy did not give a hoot about the priority of the creation of the helical theory and, as Francis prattled on, she displayed increasing irritation” (1968: 66). Franklin’s lack of interest in the boyish competition seems to have fed into Watson’s casually sexist remarks about her inasmuch as it put her outside the “game,” outside the “grand adventure” and into the “serious world” of conventional science. In their subsequent accounts, both Crick (1988) and Wilkins (2005) down-play the competitive aspects of the endeavor in favor of the more conventional “for its own sake” account of scientific inquiry. The description with which Watson introduces Frances Crick, his collaborator (quoted at the beginning of this essay) is itself telling: Here and in a number of other places throughout the book he portrays Crick in Rabelaisian terms, a rebellious genius, a kind of intellectual Huck Finn to Watson’s Tom Sawyer. Watson complements his anti-heroic portrait of Crick by constructing a different sort of anti-hero image of himself as a naïve Midwestern bumpkin, describing with frankness that verges on relish his own ineptness with girls, his bouts of intellectual laziness, his failure to remember crucial information from Franklin’s lecture. That these reminiscences are committed to print after winning the Nobel Prize on Physiology or Medicine lends them an air of irony that seems deliberate (but may not be). Again, it is instructive to compare this with Crick’s self-characterization, in which at several points he asserts his own ignorance, both as explanation for his ability to make a radical change in career focus (Crick 1988: 6) and as an explanation for crucial mistakes (1988: 65). 6.2. The structure of the game Throughout Watson’s account, it appears that several elements of play are intermingled. The “just for its own sake” of what I call whimsy, the pure

300   L. David Ritchie delight in the activity of science, and the closely-related intellectual play of solving a very difficult puzzle, which are present throughout, neatly fit Apter’s (1991) description of play as paratelic. In parallel with these themes, Watson repeatedly interjects the sense of competition, of a race for a prize, that (Watson claims) was shared by Crick, and as nearly as we can tell from Watson’s account, by Linus Pauling. Wilkins (2005), on the other hand, flatly denies the role of competition, and Crick (1988) downplays the spirit of competition – and denies that either he or Watson thought much if at all about the possibility of winning the Nobel Prize. Sir Lawrence Bragg, Director of the Cavendish Laboratory where Crick and Watson did much of their work, did not appear to give the competition for the structure of DNA much importance, apparently in part because he regarded it as a distraction from other work and in part because he did not take seriously the possibility that they might succeed. (It must also be noted that the consensus at the time did not favor DNA as the basic stuff of heredity; that insight followed rather than preceded the discovery of its structure.) Rosalind Franklin, also at King’s College, gave the “race” no importance at all. Both from Watson’s account and from other evidence (Crick 1988; Sayre 1975; Wilkins 2003), it appears that Wilkins, and possibly Franklin, were at least in part distracted by their implicit competition for control of DNA research in the laboratory at Kings College. This was in part due to ambiguities created by John Randall, who hired Franklin as Wilkins’s equal but worded his directive to Wilkins in a way that allowed Wilkins to conclude that she had been hired as his assistant (Sayre 1975; Wilkins 2003). Watson heard of this dispute from Wilkins before he had met Franklin; it seems likely that his ready acceptance of Wilkins’s point of view was abetted by his own tendency to view the quest for the structure of DNA in terms of a team sport metaphor and his consequent desire to recruit Wilkins onto their team as much as by the casual sexism that was rampant in the 1950s. It appears that Watson and Crick hit it off from the first, and their ready amity was reinforced by their mutual interest in solving the puzzle posed by DNA – and in capturing the prize of discovery (again, Crick down-plays the importance of the “prize,” emphasizing instead the more conventionally acceptable motivation of curiosity). The boyish quality of this relationship is evident throughout The Double Helix in Watson’s descriptions of their frequent forays to pubs and parties where they could meet au pair girls, and in his accounts of Crick’s enthusiastic bragging about their impending conquest of DNA in coffee-rooms, at dinner parties, and wherever else they could find an audience. Wilkins’s ambiguous social position with respect to this team is somewhat ironically captured by the title of his (2005) autobiography, The

Play in the Quest for the Double Helix   301

third man of DNA1 Working in London, he was geographically separated from the two; it also appears that he never did share Watson’s view of the project as either a grand adventure or as a sport-like competition. To the contrary, in his own autobiography, discussed in the next section, Wilkins (2005) repeatedly emphasizes his own view that science should be fully cooperative and his distress over the competitive elements that are celebrated in Watson’s account. Franklin was cast by Watson (1968) in the role of an outsider whose stubborn opposition was to be overcome rather than won over. This was partly because of her conflict with Wilkins, partly because of her gender – a female in what was still very much a male club, as Watson acknowledges in a couple of passages. Both Crick and Wilkins deny this allegation, pointing to the presence and complete acceptance of many women scientists in the King’s College lab and down-playing the importance of the men-only break room at King’s. But Crick (1988) does acknowledge that Franklin’s lack of interest in model building may have been motivated in part by a reasonable concern, because of her gender, for her scientific reputation. It also suits the agonistic and mock-heroic form of Watson’s narrative to have the path blocked by a powerful feminine opponent. As for Franklin, she had little use for the competition, and seems to have viewed the attempt to solve the structure of DNA not so much as a prize, a sort of “holy grail” of science, but rather as part of an overall extended project of using X-ray diffraction to understand the structure of biological molecules. Watson’s boyish enthusiasm for the competitive chase, for the grand adventure, seems to have struck Franklin as a distracting nuisance. “Seems to,” because we have primarily Watson’s word for it, and Franklin left no commentary on the relevant events (beyond the deliberately playful irony of a hand-written invitation to a colloquium in which she and her assistant prematurely declared the “death” of the helical model for DNA). From Watson’s (1968) epilogue, it is apparent that some if not most of the antagonistic division of the scientists involved in the process was a product of his own imagination, both at the time and later, when he reconstructed the events in his narrative. This impression is confirmed by both Crick and Wilkins. What remains fairly clear, though, is that Watson placed much more emphasis on the competitive aspects of the enterprise than did the other participants, both at the time and in retrospect, and that at least some of his and Crick’s actions were influenced by a playfully agonistic view of the enterprise. At the same time, it is also apparent that Watson and Crick may also have placed much more emphasis than any of the others did on the more whimsical aspects, including the 1 According to Wilkins (2003), the title was not of his choosing, but rather was more or less forced on him by his publisher.

302   L. David Ritchie sheer fun of manipulating their molecular models in Crick’s office: As Crick put it, in his rather more sober description of one of Watson’s serendipitous discoveries, “play is often important in research” (1988: 66). 6.3. Extended cognition Andy Clark (1997: 61) cites the routine use of paper and pencil to do long multiplication and division as an example of the external environment becoming an extension of mind.2 Clark argues that much of human intelligence and creativity is accomplished by thus constructing physical systems that organize the environment in such a way that it is optimally suited to our limited cognitive capacity. Watson and Crick’s approach to solving the structure of DNA involved numerous such externalized extensions and amplifications of their cognitive processes. The most obvious of the externalized extensions of cognition is the device of constructing three-dimensional representation of the four bases known to make up DNA, which Crick and Watson copied with considerable relish (according to Watson) from the approach used by Linus Pauling in solving the similar puzzle of the structure of the polypeptide chain. Long before the duo had assembled the molecular information they would eventually need, they assured each other that “perhaps a week of solid fiddling with the molecular models would be necessary to make us absolutely sure we had the right answer” (Watson 1968: 55). As in many other instances, Crick gives a more conventionally “serious” explanation for the importance of models: “What Pauling did show us was that exact and careful model building could embody constraints that the final answer had in any case to satisfy. Sometimes this could lead to the correct structure, using only a minimum of the direct experimental evidence” (1988: 60). If Clark’s argument for the role of embodied interactions with the physical environment is valid, as the extensive research reviewed by Gibbs (2006) suggests that it is, then the “fiddling” with models provides not merely a way of testing insights against the constraints represented in the models, but also a separate perceptual experience, muscular as well as visual, independent of and interacting with their linguistic and mathematical formulations. It is one thing to describe a helix; quite another thing to see one and manipulate it. 2 We sometimes construct mental models of paper-and-pencil algorithms, which we manipulate in roughly the same way that we originally manipulate the physical objects, and thus learn to do large calculations in our heads.

Play in the Quest for the Double Helix   303

Watson and Crick attempted to interest the team at King’s College in the model-building approach, but Wilkins did not seem to grasp the possibilities, and Franklin seemed outright hostile: “the idea of using tinker-toy-like models to solve biological structures was clearly a last resort… only a genius of his (Pauling’s) stature could play like a ten-year-old boy and still get the right answer” (Watson 1968: 51). Here as elsewhere it is not evident to what extent this passage accurately represents Franklin’s views, but it is apparent from other sources that both she and Wilkins believed that it was necessary to solve the underlying structure mathematically before any attempt at constructing a physical model could possibly succeed (Sayre 1975; Wilkins 2005). Franklin’s assessment of the model-building enterprise could only have been reinforced by the results of Watson and Crick’s first attempt, in which they built a model based on an inaccurate representation of the shape of the bases and, much more seriously, on a radical underestimation of the amount of water present in the molecule (Watson 1968: 54). Crick (1988: 68) gives a slightly different explanation of Franklin’s disdain for model-building: “Rosalind… wanted to use her experimental data as fully as possible. I think she thought that to guess the structure by trying various models, using a minimum of experimental facts, was too flashy.” After denying that Franklin was handicapped by sexism in the laboratory, Crick goes on to suggest that her insistence on a meticulous, data-based approach to solving the structure may in fact have been in part due to a desire “to be treated as a serious scientist” (1988: 69) and to acknowledge that gender differences may indeed have played a role. “She lacked Pauling’s panache. And I believe that one reason for this, apart from the marked difference in temperament, was because she felt that a woman must show herself to be fully professional. Jim had no such anxieties about his abilities. He just wanted the answer, and whether he got it by sound methods or flashy ones did not bother him one bit” (1988: 69). Watson and Crick’s initially disastrous mistake with respect to the water content of the molecule, due in large part to Watson’s failure to take notes at Franklin’s lecture on her work thus far, underscores Clark’s point about paper-and-pencil as an extension of cognitive processes. Perhaps learning from his initial mistake, on the train back to Cambridge from a later visit to the King’s College lab where he had got a glimpse of Franklin’s best photograph of the molecule, Watson sketched the image on the margin of a newspaper to be certain he did not forget crucial details. The X-ray diffraction technique (including the mathematics required to interpret the diffraction photos) is itself an excellent example of the human ability to re-organize the physical environment in a way that renders it susceptible to our limited cognitive capacity (Clark 1997). Like the three dimensional model Watson

304   L. David Ritchie and Crick built, the X-ray diffraction technique draws visual perception into the cognitive realm, along with a rich array of simulators activated (for a scientist with sufficient experience) by the pattern of light and dark spots on the photographic plate. 6.4. Distributed cognition “Advanced cognition depends crucially on our abilities to dissipate reasoning: to diffuse achieved knowledge and practical wisdom through complex social structures, and to reduce the loads on individual brains by locating those brains in complex webs of linguistic, social, political, and institutional constraints.” (Clark 1997: 180). The most obvious example of distributed cognition is the basic duo of Crick and Watson, represented in Watson’s (1968) account as if they formed a cognitive unit throughout. Sometimes they feed each other’s mania; sometimes one acts as a brake on the other’s enthusiasm: “Francis, however, remained lukewarm, and in the absence of any hard facts I knew it was futile to try to bring him around” (Watson 1968: 77). Crick puts it slightly differently, emphasizing the mutual criticism: “If either of us suggested a new idea the other, while taking it seriously, would attempt to demolish it in a candid but nonhostile manner…. The advantage of intellectual collaboration is that it helps jolt one out of false assumptions” (1988: 70). The basic two-person team worked within a looser and more extended structure that included some scientists who were frequently consulted, and others who introduced key information (or in some cases, scotched distracting ideas) at crucial points during the process. For example, when Watson came across Chargaff’s report that adenine and thymine molecules are present in DNA in approximately equal numbers, as are the guanine and cytosine molecules, a key piece of evidence was introduced into their cognitive system. But, “When I first reported them to Francis they did not ring a bell, and he went on thinking about other matters” (Watson 1968: 83). Only after a series of conversations with a theoretical chemist, John Griffith, including a speculation about the possibility of “a perfect biological principle,” did Crick begin to think about the possible importance of Chargaff’s findings. Griffith was committed to an alternative mechanism for DNA replication, but after extensive discussion (over what seems to have been a substantial quantity of ale), he did the requisite calculations and, a few days later reported the possibility that, indeed, adenine and thymine molecules might stick together along their flat surfaces, as might guanine and cytosine molecules (Watson 1968: 85). Even with key information in hand, several mechanisms were still missing, but this episode (which extended over several weeks) exemplifies

Play in the Quest for the Double Helix   305

the complex way in which the extended social network both brought in key information and distributed the work of conducting highly specialized computations. In addition to Griffith, several other scientists, for example Bill Cochran, served as sounding boards for Crick and Watson, listening to their theories and pointing out obvious flaws or bringing additional useful facts to their attention. Here also the difference between Watson’s “subversive” account and Crick’s more conventional account is evident. Both emphasize the limitations of their own knowledge, and at times seem to imply that this fundamental ignorance was important if only because it allowed them to remain open to new possibilities that a more knowledgeable person might dismiss out of hand. But Crick acknowledges that this ignorance about key scientific facts is contrary to conventional expectations: “We have also been criticized because we had not perfectly mastered all the very diverse fields of knowledge needed to guess the double helix, but at least we were trying to master them all, which is more than can be said for some of our critics” (1988: 74). Given the number of very diverse disciplines necessary for the discovery of the double helix, much less for integrating it into a full account of genetic transfer, this bit of defensiveness seems odd, to say the least. Watson’s frank admission that the duo of necessity drew on an extensive network of expertise seems much more realistic. The intellectual network within which Crick and Watson worked was structured in a complex way. The core duo, Crick and Watson, seem to have been drawn together not only by a shared ambition but also by an almost instantaneous friendship, and by a shared spirit of adventure and playfulness. The intellectual play of puzzle-solving and the associated scavenger-hunt-like search for key facts and relationships was of course central – but a mutual delight in frivolous parties, pub visits, and chasing “popsies,” in short for all the nonintellectual attractions Cambridge had to offer, frequently shared by Crick’s wife Odile, also provided part of the glue that held the relationship together. Relationships with others beyond the core duo were more complex. Some, like Griffith and Bill Cochran, who also served as a frequent sounding board, seem to have been drawn into their pursuits through a similar enjoyment of the puzzle for its own sake (and a shared delight in the less abstract pleasures of pub and table). Both Crick and Watson attempted to bring Wilkins into the game, apparently in part because of a genuine friendship between Crick and Wilkins but also in part because, under the “gentleman’s rules” of British science, Wilkins was regarded as having a certain proprietary interest in pursuing the structure of DNA, having begun serious investigations of it first. Perhaps more important, Wilkins (and his King’s College colleague, Franklin) had possession of the needed X-ray photographs: “The painful fact that the pictures belonged to Maurice could not be avoided” (Watson 1968: 41).

306   L. David Ritchie Neither Wilkins nor Franklin seemed to share Watson and Crick’s playful approach to the problem, and Franklin seemed, in Watson’s account, actually disdainful both of their adventurous attitude and of their preoccupation with model-building. These complex interpersonal factors contributed to the development of a relationship that was partly built on Watson’s (and apparently Crick’s) commiseration with Wilkins over his difficulties with “Rosie,” partly on an understated competition between the two groups – understated if only because it is not clear that the King’s College group ever really saw it as a race. It is unclear, but seems likely that this boyish division of the interested scientists into “teams,” with its implications of aggressive competition, may have justified in Watson’s own mind his act of what he describes as espionage, when he took advantage of Wilkins’s unauthorized action in showing him a recent high-quality photograph taken by Franklin (Watson 1968: 107) and, on the way back to Cambridge copied the details down on the edge of his newspaper. It does not appear that Franklin was ever told that Wilkins had given away the results of her labors without her knowledge (Sayre 1975). In a short passage between the preface and Chapter one, Watson (1968) recounts a little story about encountering Willy Seeds, a scientist who had worked with Maurice Wilkins at King’s College, on a hike up the Obergabelhorn. “All he said was, ‘How’s Honest Jim?’ and quickly increasing his pace was soon below me on the path.” Watson makes no further comment about this encounter, and Seeds has no further part in the story. Although other explanations are possible, this is an odd bit of narrative with which to introduce a tale of scientific discovery, and it seems plausible that Watson remained uncertain about the moral basis of his action in taking advantage of Wilkins’s lapse. The relationship with Linus Pauling, at Cal Tech, was more unambiguously competitive, even when Linus’s son Peter joined Crick and Watson’s circle of friends and began to serve as an unofficial conduit of information from his father’s lab. Perhaps because the spirit of friendly competition was shared, the relationship with Pauling was less ambiguous than that with the King’s College group. Even so, news from Pauling’s lab sometimes acted as a spur to greater effort, but sometimes added crucial bits of knowledge to the puzzle. Consulting Pauling’s published work, and imitating his (also published) model-building approach are both generally accepted within the idea of “fair play,” and it does not appear that the relationship with Peter was ever abused to gain privileged information, although it did allow Crick and Watson to learn of new developments in Pasadena long before they were published (a role now played by the internet).

Play in the Quest for the Double Helix   307

It is becoming increasingly apparent that suppression is as important as expression in cognitive processes (Clark 1997; Feldman 2006). As Crick put it, “to obtain the correct solution of a problem, unless it is transparently easy, usually requires a sequence of logical steps. If one of these is a mistake, the answer is often hidden, since the error usually puts one on completely the wrong track. It is therefore extremely important not to be trapped by one’s mistaken ideas” (1988: 70). Several members of Watson and Crick’s extended social network played this role on various occasions, pointing out errors in their thinking or providing new evidence that contradicted a hypothesis, and as Crick points out, the two continually played this role for each other. Franklin herself performed this role at an early stage, providing information that proved Crick and Watson’s first attempt at a model impossible. When they had finally found the right model, Franklin performed an equally valuable function of confirmation. (Incidentally, the magnanimity of her demeanor on this occasion seems to have given Watson his first inkling that he might have previously treated her unfairly.) 6.5. Managing the assembly and flow of information Examining Watson’s account of the discovery of the double helix as an instance of distributed cognition leads to an interesting view of successful science as a process of information management. Certainly Watson contributed some key insights, although he attributes many of the most important insights to Crick. What both Watson and Crick seem to have been very good at was the combination of bringing together needed data and needed computational resources and the all-important prior step of recognizing what is needed and what is not needed. This process of knowledge management is accomplished in many different ways – pubs and parties with au pair girls may be more fun than e-mail and (for the most part) academic conferences, but even here I would argue that the spirit of adventure and fun may play a more central role than we have been willing to recognize. The spirit of adventure and play, fore-grounded by Watson’s language, by his racy narrative style, and by the structure of the narrative, may not be a necessary component of this process of managing knowledge, at least not in quite the extreme form suggested by this book. It does seem evident that the shared spirit of adventure and play is at least part of what brought Crick and Watson together and kept them working together, and that it contributed to the way the social network of scientists was constructed. As will become evident in the next section, the absence of a shared spirit of adventure and play seems to have had a deleterious effect on the work at King’s College.

308   L. David Ritchie 7. The view from King’s College Watson’s account is written in a spirit of grand adventure, and describes an extended social and cognitive network, with at least three levels of involvement, all organized around the central duo. Where opposition is encountered, it is overcome in a manner consistent with the underlying spirit of play, competition, and adventure. The limitations of Watson’s grant are overcome first by ignoring them, then by renegotiating them, and again by ignoring the limits apparently imposed by the renegotiated grant. Rosalind Franklin’s opposition to the model-building enterprise and to any real collaboration with either Wilkins or the Cambridge team is overcome by an act of espionage, abetted by cultivating Wilkins’s sympathy and appealing to Wilkins’s sense of collaboration. Wilkins’s (2005) account also describes an extensive external network, but what should have been the crucial internal collaborative link, with Franklin, is blocked by conflict. Where Watson describes a playful adventure, Wilkins describes a frustrating struggle. In his account, Watson almost revels in his own boyish displays of rebelliousness, insensitivity, and even obtuseness, Wilkins tells his “side” of the story in a spirit that combines defensiveness with selfjustification and an overall lack of humor. It appears from various passages that Wilkins’s (2005) autobiography was written in part as a response to Sayer’s (1975) biography of Rosalind Franklin, in which Wilkins was implicitly accused of sexism because of his conflict with Franklin. In several passages, Wilkins emphasizes his family’s involvement in the suffrage movement, the large number of women working in the King’s College lab, both as scientists and as technicians, and his own collegial relationships with several of these women. Although there is an air of defensiveness about many of these passages, Wilkins’s explanation of the role played by Randall in precipitating a hostile relationship between Wilkins and Franklin is consistent with other accounts, including Sayre (1975). In a letter to Franklin, which Wilkins did not see until much later, Randall implied that both Wilkins and Stokes would cease to work on DNA, leaving Franklin, assisted by Gosling, with primary responsibility for the study of DNA. This impression was reinforced by the fact that Wilkins, away on vacation, was absent from the first staff meeting at which Randall and Gosling briefed Franklin on the project (Wilkins 2005: 144–145). Although Wilkins protests that he had subsequently made his continued interest in DNA clear to Randall, he describes his own attitudes as lukewarm toward the project as late as 1949 and 1950, up until just before Franklin arrived. For example, “Having given up the somewhat inhuman study of physics, we were disinclined to make physical study of the seemingly non-living, static structure of DNA. I was dis-

Play in the Quest for the Double Helix   309

tracted by making new kinds of microscopes to study living cells.” (Wilkins 2005: 115). It was only after Franklin’s appointment, during the holiday that prevented his attendance at the crucial first staff meeting with Franklin, that Wilkins gained the insight that concentrating on X-ray study of DNA would provide the best avenue for pursuing his original interest in genes. Certainly Randall would have done better to have given Wilkins a copy of the letter to Franklin, but Randall’s assumption that Wilkins was not interested in pursuing the DNA problem does not appear to have been unreasonable, given Wilkins’s frequent negative comments. Wilkins downplays the possible role of his own previous “cold attitude” toward X-ray analysis in Randall’s decisions, and emphasizes instead the role of Randall’s desire to renew his own research program through a closer involvement in the DNA work: “Randall may have hoped to achieve that closer link with Rosalind since she was directly responsible to him” (2005: 145). Whether Randall indeed “intended to manoeuvre me out of the work” (2005: 146) or was simply responding to an understanding of Wilkins’s ambivalent attitude toward the project that he had gleaned from previous casual comments is uncertain. From his own account, it seems possible that Wilkins became genuinely interested in DNA only after he had all but given the problem away, first by his apparent lukewarm attitude toward it and second by his suggestion to Randall that the newly-hired Franklin be assigned to the project. Wilkins admits the possibility that his own remarks led Randall to conclude that he was ready to leave the study of DNA. “It is true that I had often said that I found more joy in microscopes and watching living cells, whereas X-ray study of static structures was rather a bore. But I certainly never said to Randall that I was going to give up the DNA X-ray work…. Whether I made my feeling about ‘dynamic DNA’ clear to Randall I do not know, but I always assumed I had” (2005: 148). Nonetheless, he accuses Randall of ruthlessness, and implies that Randall may have been responsible for the conflict that barred Wilkins and Franklin from working together as a team: “(Randall) may have kidded himself that he could persuade me to drop out of DNA X-ray work: what would he lose by trying? But how much did Rosalind and I lose by the ‘Wilkins and Stokes are leaving’ story? If Randall had not barged in it might even been that Rosalind could have worked happily alongside Stokes and me…” (2005: 149). The language of this passage is very telling: “barged in” activates a sense of uninvited intrusion (an interesting contradiction with Randall’s administrative role). In the paragraph that follows, Wilkins accuses Randall of modeling himself after Napoleon (echoing a characterization that also appears on p. 101), a metaphor that carries implications of strategic brilliance and conquest as well as tyranny and destruction, only to contradict many of these implications

310   L. David Ritchie by relating a story of an artist commissioned to paint a portrait of Randall who characterized him as “frail,” “uncertain,” and “anxious.” The “Napoleon” characterization also seems to contradict Wilkins’s earlier description of Randall’s lab, “especially in its early years,” as “rather casual about organizational boundaries” (2005: 147), a description that seems intended to justify Wilkins’s own subsequent actions in “barging in” on Franklin’s work with DNA. Early in his description of the working conditions at King’s College, Wilkins relates that he and Randall, every year or so, “had a stand-up row, and nearly always he would give in…. When we had such a row we were open – we said what we thought and we heard what was said” (2005: 103).3 Yet in his explanation of the on-going conflict with Franklin and the underlying ambiguity of his own role with respect to the DNA project, Wilkins expresses a reluctance toward direct confrontation and admits that the issue might have been resolved, had he approached either Randall or Franklin directly. Wilkins characterizes Franklin’s reaction to his own continued interest in the DNA project in terms of annoyance and even outright hostility. As he was leaving the building in July 1951 (six months after Franklin arrived), “Rosalind Franklin came up to me and announced, quietly and firmly, that I should stop doing X-ray work. She concluded with the instruction: ‘Go back to your microscopes!’” (2005: 142). In the same paragraph he compares his reaction to this to a previous incident when a professor who was doing an illegal experiment on a live cat abruptly kicked him out of a lab, then goes on to ascribe motives of jealousy to Franklin. “She had been in our lab for more than six months, and I had found her friendly enough, apart from sometimes being a little sharp with her tongue. She seemed a high-principled, civilized person. I felt that there might be something about my talk (possibly its success) that had disturbed her…” (2005: 143). Throughout this long passage about the ambiguity in Wilkins’s relationship to the DNA project subsequent to Franklin’s arrival, he seems to move back and forth between hints, and even outright allegations, of scheming against himself (Randall wants him out of the way; Franklin is jealous of his success) and admissions of his own earlier ambivalence about the project. It is evident that Wilkins never asserted his own sense that he had a claim to a part in the DNA project, or even made clear his continued interest, either to Franklin or to Randall. It is also evident that communication within the King’s College group was far less open and free than it was within the Cambridge group. Indeed, 3 From the extent of the misunderstanding over Franklin’s role it is apparent that the two men very likely did not always hear the same things, but rather quite often heard what they expected or wanted to hear.

Play in the Quest for the Double Helix   311

it appears that communication was better between King’s College and Cambridge than it was within the King’s College lab – Wilkins talked more freely about DNA with both Crick and Watson than with either Franklin or Randall. Wilkins’s account gives several hints of an undertone of sexual tension in his relationship with Franklin that may have exacerbated the sense of conflict between them. When he first introduces her Wilkins describes Franklin as “quietly handsome with steady, watchful, dark eyes,” and relates that she “did one thing I had not noticed other women do in our lab: she placed a small mirror on the wall so that it faced her as she sat at her desk. It seemed too small to allow her to see who was behind her at the office door, and at the time I wondered whether she was anxious about her appearance.” The observation that Franklin may have been anxious about her appearance contrasts interestingly with Watson’s ruminations about how Franklin might look if she would “do something interesting with her hair” (Watson 1968: p. 51) – but it scarcely constitutes a convincing defense against Sayre’s (1975) accusations of sexism! During Franklin’s early months at King’s, Wilkins and Franklin often lunched together on Saturday, often with others from the lab but sometimes alone. “I found Rosalind pleasant to talk to but occasionally a small spikiness would show. For example, after we had eaten fruit salad and cream, I remarked that the cream had been good. Rosalind replied coldly, ‘But it was not real cream.” I was amused but said nothing, not having had the advantage of living in post-war Paris where food was not rationed as it had been in Britain” (Wilkins 2005: 132). “I was amused but said nothing…” Obviously Wilkins was amused by but not with Franklin: the turn of phrase carries an implication of superiority and judgment. “Coldly” is also interesting here: Without the metaphor the sentence would still convey the intended contrast between conditions in post-war Paris and post-war London, and adequately illustrate Franklin’s alleged “spikiness.” “Coldly” activates a set of social as well as visceral schemas that contrast sharply with the characterization of Franklin, at the beginning of this passage, as “pleasant to talk to.” Consider how differently this passage would read if “playfully” or “thoughtfully” were substituted for “coldly.” Implications of an underlying sexual tension are more explicit in Wilkins’s ruminations about the possibility that he could have had a romantic relationship with Franklin, based on similar family backgrounds and interests. “My rather juvenile attitude towards women caused me to seek special liveliness in beauty, spirit or artistic interests, especially in ‘shy young women’ (as Eithne called them), and that for me ruled out any romantic interest in Rosalind” (2005: 133). Franklin apparently did have both spirit and artistic interests (Sayre 1975), but she was certainly no “shy young woman.”

312   L. David Ritchie A while after the “Go back to your microscopes” encounter (and after a long digression about Randall’s apparent “scheming” to take over the DNA problem) Wilkins relates that his weekly lunches with Franklin had already petered out. (They apparently didn’t last long.) I was still, after several years, seeing my Jungian analyst… and he suggested I should invite Rosalind to dinner so that we could get to know each other better…. I went to see Rosalind on a very warm afternoon, and found her in a large lab…. She was sitting on the floor in a labcoat and seemed quite willing to talk. The work must have been hard, for she was sweating in the heat, but she did not seem to mind the very close atmosphere in the lab. In those days before deodorants we were all used to smelling rather bad after some physical exertion, but in the stifling lab I found myself quite unable to imagine sitting down to dinner with Rosalind that day. I very much admired her hands-on approach to the work, and respected the effort she was making, but I could no longer face the challenge of a sociable evening with her. I seemed to forget that our dinner was meant to be the means to a very worthwhile end: that of developing a better relationship about our research. Instead, I drifted away. (Wilkins 2005: 150–151).

In the long first section of his autobiography, Wilkins confesses to an early awkwardness around women, an awkwardness that extended well into his twenties. The rather strange passage about the encounter in the laboratory suggests that the awkwardness was still very much a factor in Wilkins’s response to Franklin. She “seemed quite willing to talk,” but her no doubt earthy, animal odor, after an extended period of physical exertion on a warm day, presented an insurmountable barrier. “In those days before deodorants,” it seems unlikely that Wilkins would have felt nearly so strong a wave of disgust in the presence of a sweaty male scientist under similar circumstances. So the picture we have from Wilkins is one of a situation in which key members of the team (Wilkins, Franklin, and Randall; we don’t know about other team members) were distrustful of each others’ motives, relationships among team members were fraught with sexual tensions and resentments, and attempts at socializing were at best awkward. There is little of the sense of exuberant playfulness that Watson describes in Cambridge. The occasional effort at “grooming” (Dunbar 1996) and playful casual interaction misfires or leads nowhere, and the mutual mistrust completely blocks the kind of frank mutual criticism that Crick found so vital to his collaboration with Watson. Wilkins’s more serious and less playful approach to science may also have contributed to his failure to see the importance of the one-to-one ratios between guanine and cytosine, and between adenine and thymine, which proved crucial for the Cambridge team. Wilkins tells us that “my considerable respect

Play in the Quest for the Double Helix   313

for Chargaff may have led me to misinterpret his caution about base-pairing” (2005: 152). The implication is that, if Chargaff did not consider the ratios to be adequate evidence that these bases were actually paired, there was no reason for Wilkins to pursue the possibility further. Nonetheless, Wilkins must have had some sense of the importance of these ratios: back in London, he tells us that he wanted to tell Franklin about Chargaff’s evidence, but she also wanted to tell him about her new X-ray results. “I tried to finish what I was saying, because Chargaff’s evidence seemed so basic to our work, but she stopped me again by remarking, in a cool, amused way, that it seemed that I did not want to hear what she had to tell me…. Her quiet air of confident superiority disturbed me. I stopped talking and asked her to tell me what she had done” (2005: 154). Again, Franklin is “cool”; here it is she, not Wilkins, who is “amused.” “Her air of cool superiority – a look I have never forgotten – temporarily undermined my self-confidence, and gave me a brief feeling of panic” (2005: 155). Echoing Watson’s more humorous descriptions of encounters with Franklin, Wilkins concludes this passage with the comment that “Rosalind could be terrifying” (2005: 156). In following passages, Wilkins describes his own retreat from confrontations with Franklin that might have led to a more candid sharing of information and insights, and admits that a more forthright assertion of his own interests might have led to clarification of the situation and a more open communication within the lab. It is clear that the interpersonal tensions, originating in the ambiguity about work assignments but probably reinforced by Wilkins’s ambivalent and often quite awkward response to Franklin as a woman, prevented either of them from engaging each other in anything like the spirit of “extended cognition” that characterized the relationship between Watson and Crick, or for that matter, the relationship of both Watson and Crick to other scientists at Cambridge that they consulted and drew upon. 8. The Double Helix and its consequences The three autobiographical accounts of the discovery of the double helix differ in several ways. Watson’s account is fictionalized in the sense that it follows the conventions of popular fiction, including a relatively rapid pace, an emphasis on the hero’s quest, initial setbacks, opposition, and eventual triumph. Watson manages to get in a good deal of the science (as both Crick and Wilkins somewhat grudgingly admit), but he does so in a way that never interferes with (and indeed often amplifies) the suspense – and the fun. Both Crick and Wilkins acknowledge that Watson’s account is accurate in most respects,

314   L. David Ritchie but Crick in particular seems to object to the emphasis on the chase, and on the drama of the pursuit: “It is the molecule that has the glamour, not the scientists” (Crick 1988: 67). Consistent with this preference, Crick structures his book in a way that gives far more weight to the way the structure of DNA, once it was discovered, fit within the continuing attempt to understand the basis for biological heredity. In a mostly positive review of the BBC docu-drama about the discovery of the double helix, Life Story, Crick complains that the movie dwells on the moment of discovery as one of great illumination and finality. In contrast, he points out, neither the helical structure nor its biological significance were fully established until many years later, and Watson was so uncertain at the time that he resisted even mentioning the potential biological significance for fear that they would be made to look foolish if some crucial mistake were discovered. Watson’s casual sexism in his treatment of Rosalind Franklin (“Rosie”), along with the implication that it was shared by other male participants precipitated critical responses including Sayre’s (1975) feminist biography of Franklin, which in turn led to an extended defense in Wilkins’s (2005) autobiography and a somewhat shorter defense in Crick’s (1988) account. More important for the present purposes, Watson’s account stands more or less unique within the genre of scientific autobiography in its candor about the human weaknesses and foibles of scientists and, most important of all, in its exuberant celebration of the spirit of play that infuses scientific work. 9. Cognition as a social activity An emerging theme through all of this is the contrast of myth and reality. The myth of scientific discovery is the lone scientist, alone in the lab, late at night, painstakingly amassing facts, connecting them through the sheer power of his4 brilliant logic until the truth reveals itself, beyond mistake, doubt, or challenge. This may be true or may have been true of some scientists making some discoveries under some circumstances, but The Double Helix, along with all the subsequent criticism, commentary, and alternative accounts, paints a different and, frankly, more interesting picture.

4 The gender stereotype implied by the pronoun, his, is itself part of the myth. As I discuss in more detail in a later section, gender may have played more of a role than merely the segregated break room and the apparent sexual tension between Maurice Wilkins and Rosalind Franklin.

Play in the Quest for the Double Helix   315

If we are to believe Watson, he and his good buddy Crick did not actually do any experiments on DNA. They each did experiments on other molecules but with respect to DNA they scarcely fit the image of lone scientists hard at work at their lab bench. They were part of a network of scientific information, distributed cognition in a very broad sense. Given the complexity of the problem they tackled, there were many facts to be assembled and analyzed, facts that required an extensive array of skills as well as background knowledge even to understand, much less to gather. Books were not sufficient – at a crucial point it turned out that even the most current organic chemistry textbooks contained a crucial error that required very specialized expertise to detect and correct. Obtaining the DNA samples for analysis required yet other skills. Crick (1988) was a little defensive about this point but it would seem wastefully time-consuming for one person to try to master all the necessary knowledge, the facts and the skills needed to make sense of those facts, when there are other people available, who already have these skills. A single scientist working alone might in time have succeeded in mastering all of the specialized knowledge needed for the discovery – but it seems evident that a network of scientists with complementary knowledge and skills is a much more efficient way to get the work done. The myth has it that the scientific thinker reasons his way to a solution by pure force of logic. But as Crick (1988) points out, when reasoning about a complex phenomenon such as DNA, there are multiple opportunities to make errors. So even in the process of reasoning, the discovery of the double helix required extended cognition, the two principle players testing and correcting each others’ logic and, on several occasions, bringing in outsiders to test and correct their logic. Here again, it is possible for a scientist, working alone, to critique his or her own logic, find and correct the errors. But it can be done much more efficiently by multiple scientists, interacting in a spirit that playfully combines competition with cooperation. Restrictions in the power of purely internalized cognitive processing are also evident in the important role of model-building. In some circumstances, drawing diagrams, building models, even doodling on a cocktail napkin may simply be a way of stimulating imaginative processes. But human beings have a cognitive capacity that is very restricted, relative to what we often attempt to accomplish with it. As Crick (1988) explains, the models he and Watson built captured key constraints in a form that could be (physically as well as conceptually – see Gibbs 2006) grasped. In principle, this could have been accomplished without the aid of physical models, but to do so is much more intellectually demanding, and in any event would almost certainly require an alternative form of extended cognition, such as a pencil and notepad.

316   L. David Ritchie 9.1. Science is work Several years ago, at a workshop on the topic of motivating students, I was seated at a table with a young mathematician who, in response to my query about his research, described a rather elegant set of ideas in tones of obvious relish. “It sounds like fun,” I observed. He agreed that he enjoyed it very much. Then I asked if he tells his students how much fun it is. His reply: “But mathematics is work.” It didn’t help when I pointed out that skiing, sail-boarding, mountain-climbing, and crossword puzzles are all work. I never could convince him that it might be helpful to tell his beginning math students how much fun mathematics can be. Math is indeed work. Science is work. But, as taxing as science is, depending as it does on skills that take years to develop, at its best science is jolly good fun. That’s the dirty little secret Watson revealed to the world: Don’t feel sorry for scientists in their cold, drafty labs – they’re having fun! (And not all of their work is done in cold, drafty labs. Some of it is done in pubs and restaurants, at parties, on hiking trails and ski slopes.) At first it might seem like a motivational thing – of course science is rewarding or people wouldn’t keep doing it. But Watson’s account suggests more than that. The fun, the play, may be integral to the work of science. And indeed, given that playful behavior is so ubiquitous among the higher animals, and whimsical play among all human cultures, playful behavior must have functions beyond merely rehearsing for adult roles. Three possible social-cognitive functions of play are apparent in Watson’s account. Most obviously, play is an important part of imagination and exploration, as evidenced by commonplace phrases such as “play with ideas” (Byers 2007). Often when the playful person inverts reality all that is discovered is nonsense – but once in a while, what is discovered is a new and interesting, sometimes even useful, alternative reality. So, yes, as Crick insists, the models of the four bases were very serious business, a way of representing the constraints of chemical bonds. But the human predilection for manipulating things, moving things around, playing with things is itself crucial. Handling the models provided a direct access, by way of muscles and joints as well as visually, to areas of the brain needed to assemble the completed abstract simulation of the DNA molecule. The play was a nearly-indispensable part of the work. Play is also an important part of what Dunbar (1996) calls grooming. Playful behavior gives pleasure to self and others, and shared pleasure helps build and maintain social bonds. Building models together, punting on the River Cam, a bit of horseplay, flirting with the au pair girls at one of Odile’s parties, having a couple of pints at the local pub, trading jokes and outrageous puns: All of these contribute to building and maintaining the social network so it can also

Play in the Quest for the Double Helix   317

function as an extended cognition network (see for example Plester and Sayers 2007; Terrion and Ashford 2002). Third, the competitive form of play interacts with the whimsical form in a way that both stimulates and softens the sting of criticism. Crick makes a point, rightly I think, of the crucial role of mutual criticism in testing and refining an argument. But Watson’s account makes it clear that this was, at least between the two of them and usually between them and others in their cognitive network, stimulated in part by competitive play and softened by more whimsical forms of play (Norrick 1993, makes a similar point about the role of playfully aggressive humor in close relationships). 9.2. “No gurlz aloud.” Watson’s candor about the treatment of Rosalind Franklin (“Rosie”) makes special sense in the context of the crucial role of grooming – gossip and play in its various forms – in building and maintaining “coalitions” (Dunbar 1996), the extended cognition networks that so effectively amplified the cognitive reach and powers of their members. As the subheading of this section is intended to imply, there is something in Watson’s description that is reminiscent of Ben Waterson’s Calvin and Hobbes, and of course many other depictions of the pre-adolescent “war between the sexes.” Watson describes the competitive aspects of the “race” for the double helix (downplayed in both Crick’s and Wilkins’s accounts) as a kind of “team sport,” and “Rosie” is almost reflexively assigned to the other team. As Crick (1988) protested regarding the movie, Life Story (Jackson 1987), this and other elements of the account are apparently emphasized for dramatic purposes – but the underlying pattern is consistent with both Crick’s and Wilkins’s accounts. It is possible that a male scientist, introduced into the lab at King’s College under similar circumstances, might also have become engaged in a conflictridden relationship with Wilkins, and it is also possible that Wilkins’s attempts at “grooming” with a male colleague might have misfired as badly as his actual attempts with Franklin did. But it seems apparent that the sexual tensions Wilkins experienced aggravated the relationship and may have precluded the kind of playful behavior – joking, teasing, mock aggression – that might have defused the competitive tensions between two male colleagues under comparable circumstances. Sayre (1975) takes pains to portray a playful side to Franklin’s personality, but she admits that it did not seem to come out in Franklin’s relationship with her colleagues at King’s, and it appears from Crick’s and Wilkins’s accounts as well as from Watson’s account that she held herself aloof

318   L. David Ritchie from any kind of playfulness or casual gossip. The awkwardness and constraint was apparently mutual. From recent commentary, it is apparent that women engaging in traditionally male endeavors such as politics and science may continue to face an unexpected obstacle, not in perceptions of job-related competence, but in expectations related to playfulness. In a recent commentary on National Public Radio, Norris (2006) pointed out that, in the United States, presidential candidates show their “human” side by being photographed in such activities as jogging in warmups, doing physical labor around the yard (or for candidates from Texas or California, around the ranch) in blue jeans, and playing golf – but voters give high credibility ratings to women only when they are dressed in business attire and photographed in a serious, power-related setting. Crick (1988) speculates that Franklin’s reluctance to engage in model building may have stemmed in part from a desire to be taken seriously as a scientist and a fear that “playing around” would distract from her professionalism – a fear that would not even have occurred to either Crick or Watson. Wilkins’s attempt at reconciliation with Franklin was derailed by the earthy smell of her sweaty body after a long afternoon of exertion in an overly-warm laboratory. Watson was distracted during Franklin’s seminar by ruminations about her physical experiences. Inclusion of Franklin in the “extended cognition” network under these circumstances would not have been necessarily impossible, even during the early 1950s. Nor would it have been necessarily impossible for her to join in the play, the spirit of fun that helped to motivate Watson, Crick, and several of the other players, and to bind them together as a cognitive unit. But it does seem plausible that the undertone of interpersonal discomfort introduced by the unacknowledged sexual tensions and gender stereotypes may have interfered with this kind of inclusion. Even more immediately, it appears that Franklin was at least in part in thrall to the “myths of science” discussed in the previous section. It seems that, for whatever reason (the subtle opposition of her family, as Crick speculates, her uncertainty about the decision to leave Paris and return to London, which Sayre relates), Franklin felt a need to establish and maintain her credentials as a serious scientist. Joining into an apparently juvenile, playful approach to the enterprise would have seemed quite at odds with the image of professional competence she wished to project. 10. Conclusion Conventionally social cognition refers to how people think about social relationships and interactions, their social schemas, and conversely how people’s

Play in the Quest for the Double Helix   319

social schemas influence their social behavior and thereby help to shape their relationships and interactions. Clark (1997) and Hutchins (1995) argue that cognition, information processing in a general sense, is itself an embodied behavior, and is often distributed across members of an interacting group or team. I have argued elsewhere (Ritchie 2006) that schemas of relationship and of conversational interactions in general are activated as part of what Sperber and Wilson (1986) call the cognitive context, along with schemas of the current interaction and the relationship within which that interaction takes place. A communicative act may be relevant to, hence connect with and alter, relationship and interaction schemas as well as any other activated schemas. In brief, the relational and interactive schemas of social cognition are embedded in people’s “cognitive environments” at all times, and may be reshaped even as they shape individuals’ communicative actions and interpretations. The independent relationship and conversation schemas of each participant in a particular interaction activate partial simulators that influence how that person interprets others’ communicative acts and shape that person’s responses. Thus, independent relationship and conversation schemas interact by way of the conversation itself to shape a collectively experienced “social reality,” social cognition at the extended-cognitive level of an interaction group. “Myths of science,” widely-shared sets of ideas and expectations about how science is conducted and about the role of a scientist, are part of the cultural background that contributes to and shapes scientists’ (and others’) ideas about science and about science communication. Elements of these commonplace ideas about science are available to be taken up into the social relationship and conversation schemas of each individual, and thus to shape communication among scientists, including the degree to which they participate in extended cognitive networks or strive to “go it alone.” The events and interactions involved in the discovery of the double helix nicely illustrate the way various participants’ ideas about science and about science communication shaped their science-specific social schemas and consequently their communicative behavior. An implicit part of every conversation is maintaining the conversation itself, as well as the relationship within which it takes place (Clark 1996). We often tend to think of “phatic” communication as separate from the “serious” parts of a conversation, particularly when the conversation is about something really serious, like philosophy or science. But the evidence from our examination of the discovery of the double helix suggests that the non-serious parts of conversation serve a very important function, in the guise of what Dunbar (1996) calls “grooming,” by maintaining a communicative relationship conducive to a free interchange not only of facts but more importantly of mutual critique and

320   L. David Ritchie collaborative idea-generation and connection. Indeed, it seems likely that the play itself may often be an important part of the process through which the “serious” business of conversation is accomplished (Fazioni 2008). This study is of course limited by its unavoidable reliance on retrospective accounts by some but not all participants, each of whom constructed the events through a particular set of ideas about science – and with a certain self-image to construct and project. The differences among the various perspectives, so evident in examining the tone as well as the focus of the separate accounts by Watson (1968), Crick (1988), then Wilkins (2005), are very consistent with the differences among the three scientists’ roles in the communicative process that led to the discovery, as deduced from evidence in all three accounts. Rosalind Franklin did not leave an account, but the convergence among the three accounts we do have, not only with respect to her role but also with respect to their own roles and relationships, allows us some limited confidence in attributing to her a certain set of attitudes toward science in general and this series of events in particular. In spite of its limitations, this study suggests some interesting avenues for future research and theorizing. The interaction of playful and figurative with “serious,” topic-related elements in goal-directed conversation, and the role of playful and figurative language in maintaining social structure and advancing the purpose of conversation would seem to merit more attention in future research. The interaction of different, even contradictory, ideas about the topic of a conversation – or about the conversation itself – as exemplified by the very different expectations and understandings of Franklin, Wilkins, and the Watson-Crick team also merits detailed study in future research. A final caveat: The foregoing analysis could be interpreted as support for a strong constructionist view of science. Nothing of the sort is intended. As Crick points out, the double helix would almost certainly have been discovered sooner or later. If it had been discovered and explained in a very different style by some other research team (or lone polymath researcher), its early exposition and development would probably have been different. Different errors would have been made and required correction, connections with other topics in biological sciences might have become apparent in a different order, and so on – but the different possible tracks are all constrained by the same set of observable facts, and would gradually have converged on what will, fifty years hence, be the same canonical understanding of the constitution, shape, and function of the DNA molecule. So, with respect to the long-range concerns of biological science, I must concur with Frances Crick: “The molecule gets all the glory.” However, with respect to the continuing attempt to understand human communication, including communication among humans engaged in scientific dis-

Play in the Quest for the Double Helix   321

covery, exactly the opposite is true: The relationships and interactions among fascinating, flawed, brilliant human beings get all the glory.

References Apter, Michael J. 1991 A structural phenomenology of play. In Adult Play: A Reversal Theory Approach, John A. Kerr and Michael Apter (eds.), 13–30. Amsterdam: Swets and Zeitlinger. Barsalou, Lawrence 1999 Perceptual symbol systems. Behavioral and Brain Sciences 22: 577– 609. 2007 Grounded cognition. Annual Review of Psychology 59: 617–645. Bateson, Patrick The role of play in the evolution of great apes and humans. In The Na2005 ture of Play: Great Apes and Humans, Anthony D. Pellegrini and Peter K. Smith (eds.), 13–26. New York: The Guilford Press. Byers, William How Mathematicians Think: Using Ambiguity, Contradiction, and Par2007 adox to Create Mathematics. Princeton, NJ: Princeton University Press. Carter, Ronald Language and Creativity: The Art of Common Talk. New York: 2004 Routledge. Clark, Andy Being There: Putting Brain, Body, and World Together Again. Cam1997 bridge, MA: The MIT Press. Clark, Herbert H. 1996 Using Language. Cambridge: Cambridge University Press. Cook, Guy Language Play, Language Learning. Oxford: Oxford University Press. 2000 Crick, Francis What Mad Pursuit: A Personal View of Scientific Discovery. New York: 1988 Basic Books. Csikszentmihalyi, Mihaly Flow: The Psychology of Optimal Experience. New York: Harper and 1980 Row. Dunbar, Robin Grooming, Gossip, and the Evolution of Language. Cambridge, MA: 1996 Harvard University Press. Fagen, Robert Animal play, games of angels, biology, and Brian. In The Future of Play 1995 Theory: A Multidisciplinary Inquiry into the Contributions of Brian

322   L. David Ritchie Sutton-Smith, Anthony D. Pellegrini (ed.), 23–44. Albany, NY: State University of New York Press. Fazioni, Odilia The role of humor and play in a workplace interaction. A thesis submitted 2008 in partial fulfillment of the requirements for the degree of Master of Arts in Communication Studies. Portland, OR: Portland State University. Feldman, Jerome A. 2006 From Molecule to Metaphor: A Neural Theory of Language. Cambridge, MA: MIT Press. Gernsbacher, M.A., B. Keysar, R.W. Robertson, and N.K. Werner 2001 The role of suppression and enhancement in understanding metaphors. Journal of Memory and Language 45: 433–450. Gibbs, Raymond W. Jr. 1994 The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge: Cambridge University Press. 2006 Embodiment and Cognitive Science. Cambridge: Cambridge University Press. Huizinga, Johann 1955 Homo Ludens: A Study Of The Play-Element In Culture. Boston, MA: Beacon Press. Hutchins, Edwin 1995 Cognition in the Wild. Cambridge, MA: MIT Press. Jackson, Mick 1987 Life Story. London: BBC. Keysar, Boaz, and Briget Bly 1999 Swimming against the current: Do idioms reflect conceptual structure? Journal of Pragmatics 31: 1559–1578. Kintsch, Walter 1998 Comprehension: A Paradigm for Cognition. Cambridge:  Cambridge University Press. Kohn, Alfie 1986 No Contest: The Case Against Competition. Boston: Houghton Mifflin. Lakoff, George, and Mark Johnson 1980 Metaphors We Live by. Chicago, IL: University of Chicago Press. 1999 Philosophy in the Flesh: The Embodied Mind and its Challenge to Western Thought. New York: Basic Books. Malaby, Thomas M. 2006 Parlaying value: Capital in and beyond virtual worlds. Games and Culture 1: 141–162. 2007 Beyond play: A new approach to games. Social Science Research Network, id922456. Norrick, Neal R. 1993 Conversational Joking: Humor in Everyday Talk. Bloomington, IN: Indiana University Press.

Play in the Quest for the Double Helix   323 Norris, Michele Are U.S. voters ready to elect a woman president? National Public Ra2006 dio. Morning Edition Dec. 20. Plester, Barbara A., and Sayers, Janet “Taking the piss”: Functions of banter in the IT industry. Humor 20: 2007 157–187. Reddy, Michael J. 1993 The conduit metaphor: A case of frame conflict in our language about language. In Metaphor and Thought, Andrew Ortony (ed.) 164–201. 2d ed. Cambridge, UK: Cambridge University Press. Ritchie, L. David Frame-shifting in humor, irony and metaphor. Metaphor and Symbol 2005 20: 275–294. 2006 Context and Connection in Metaphor. Basingstoke, UK: Palgrave Macmillan Ltd. 2008 X IS A JOURNEY: Embodied simulation in metaphor interpretation. Metaphor and Symbol 23: 174–199. Ritchie, L. David, and Valrie Dyhouse 2008 FINE AS FROG’S HAIR: Three Models for the Development of Meaning in Figurative Language. Metaphor and Symbol 23: 85–107. Sayre, Anne 1975 Rosalind Franklin and DNA. New York: W.W. Norton. Terrion, Jenepher Lennox, and Blake E. Ashforth 2002 From ‘I’ to ‘we’: The role of putdown humor and identity in the development of a temporary group. Human Relations 55: 55–88. Watson, James D. 1968 The Double Helix. New York: Penguin. Wilkins, Maurice 2005 The Third Man of the Double Helix: The Autobiography of Maurice Wilkins. Oxford: Oxford University Press.

Chapter 12 Social aspects of verbal irony use Roger J. Kreuz and Gina M. Caucci

1. Introduction This chapter will provide a summary of psychological theories and research on how people use and comprehend verbal irony and sarcasm. One line of this research has been to delineate the mental processing of such remarks, and several predominantly cognitive theories of verbal irony exist (for reviews, see Gibbs 1994, 2003; Gibbs and O’Brien 1991; Giora 2003). However, this chapter will focus on theories and research which have emphasized the social and affective aspects of using such language. In order to do so, however, the first task is to address exactly what the terms “irony” and “sarcasm” refer to. 2. Definitions of “irony” and “sarcasm” 2.1. The concept of irony For better or worse, psychologists are heirs to several centuries of discussion about the concept of irony by philosophers and students of rhetoric and literary criticism (for reviews, see Preminger and Brogan 1993; Wiener 1973). This term has been used to refer to a broad range of phenomena, such as a pretension of ignorance (Socratic irony; see Booth 1974) and the actual ignorance of a character in a drama, such as Oedipus (dramatic irony, see Fowler 1965). More commonly, irony has been used to refer to states or events that seem odd or noteworthy when juxtaposed, such as the swindler who is himself swindled (situational irony; see Lucariello 1994). Another subtype of irony is verbal irony, in which one means something different (and frequently the opposite) of what one says. This label would apply to utterances like “What lovely weather we’re having!”, if spoken dur-

326   Roger J. Kreuz and Gina M. Caucci ing a downpour. Grice (1975, 1978) proposed that such statements represent violations of the Maxim of Quality (say what you believe to be true). When such violations occur, the listener will assume that the speaker is still being cooperative, and search for a related meaning (e.g., that the weather is not, in fact, lovely). 2.2. Sarcasm Sarcasm is viewed by many researchers as a subtype of verbal irony. Sarcasm, however, is irony with an attitude. This can be seen in the etymology of the term, since it comes from the Greek, meaning “to tear flesh” (Trumble and Stevenson 2003). Definitions of sarcasm frequently conflate it with satire (e.g., Mish 1993), although the two terms are not interchangeable (Kreuz and Roberts 1993). Most typically, sarcasm has been described as verbal irony which has a victim; that is, someone who serves as a target for the remark (Fowler 1965; Kreuz and Glucksberg 1989; Lee and Katz 1998). Unlike the complaint about the weather cited above, a sarcastic comment like “You’re a real genius!” makes clear that the remark is directed at a particular individual. It has been claimed that the distinction between irony and sarcasm in contemporary speech is being obscured (Nunberg 2001), and that experimental participants don’t always agree about what constitutes verbal irony (Wasserman and Schober 2006). Other research, however, suggests that participants do distinguish between these two terms (Kreuz and Glucksberg 1989), and that widespread agreement exists about how to define them (Dress et al. 2008). Specifically, college students defined irony as having counterfactual and unexpected elements, and sarcasm as a verbal form that involves humor and negativity. Humor and negativity were rarely mentioned in definitions of irony, and unexpectedness was almost never mentioned in definitions of sarcasm (Dress et al. 2008). So it seems that, at least under some circumstances, people can make reliable distinctions between these concepts. Although definitional issues may seem rather tedious, they are important to keep in mind, because not all researchers have defined irony and sarcasm in the same way. As a result, the construction of stimulus materials by experimenters may reflect their own beliefs about nonliteral language (see Katz 2005, and Toplak and Katz 2000 for related concerns). With these caveats in mind, for the remainder of this chapter we will use the term “irony” as a generic term to refer to both verbal irony and sarcasm.

Social aspects of verbal irony use   327

3. Ubiquity of verbal irony How often do speakers employ verbal irony? If it is relatively uncommon, then it may not be a particularly important phenomenon to explore and understand. If it is relatively common, however, it poses an interesting problem for language researchers. Why, for example, should speakers go out of their way to say the opposite of what they mean? It should be noted, however, the task of identifying ironic statements in context is more difficult than identifying metaphors, for example, since most common metaphors have an obvious literal connection between topic and vehicle (e.g., in the case of “The road was a snake,” “curvy” would be the literal connection between “road” and “snake”). In verbal irony, no such literal connection exists. Another difficulty in estimating how frequently people use irony lies in determining the nonliteral intentions of the speaker. Once again, this is not a problem in the case of metaphor: any identity statement of the form “An A is a B” is a potential metaphor, and context is not needed to disambiguate this. As a result, it is possible to harvest simple nominative metaphors (e.g., “Life is a journey”) for study from any suitably large corpus of text, such as the Internet (Roncero, Kennedy, and Smith 2006). In the case of verbal irony, however, both the literal and nonliteral interpretations of an utterance are at least plausible. A road cannot literally be a snake, but when someone says “You’re a genius!”, a literal interpretation is available, as well as its opposite. Therefore, without additional information, one cannot pluck ironic statements from a corpus for study. Instead, they must be inferred from the context, which provides clues about the intentions of the speaker. Here is an example from the Santa Barbara corpus (DuBois et al. 2000): Alina:  So there’s – you know, these goofy little white spots, and everything, and they’re going, oh, Alina, you look so chic Lenore:  Huh Alina:  in your suede pants. And I’m going, oh, yeah, real chic. You know, my hair’s dirty, I have it pulled back in a pony tail, I know I look lovely. So I’m sitting there talking to these idiots for a while, then this new wave of people comes in. One can infer from the context that Alina’s use of “real chic” and “lovely” are the opposite of what she intends, so the interpretation is relatively straightforward. However, in many other cases, speakers do not explicitly contrast their spoken and intended meanings, and therefore nonliteral intent can only be inferred indirectly.

328   Roger J. Kreuz and Gina M. Caucci 3.1. Estimates of use Despite these difficulties, a few studies exist in which researchers have attempted to determine the frequency of verbal irony use. Tannen (1984) tape-recorded a group of friends engaged in casual conversation, and found relatively few instances of sarcasm. In contrast, Dews and Winner (1997) analyzed utterances in thirty-minute situation comedies from American television, and found that ironic statements were made, on average, between four and five times per episode. The vast majority of these statements were ironic insults (e.g., “Oh that was very impressive”), but a few ironic compliments were noted as well (e.g., referring to a mansion as a “ghetto shack;” Dews and Winner 1997: 389). Such results suggest that television viewers would be exposed to several thousand such statements in a year of viewing. Although television programs are watched by virtually everyone, the scripted discourse they contain cannot be equated to the spontaneous speech that people engage in every day. In an attempt to obtain a more ecologically valid sample, Gibbs (2000) collected over sixty 5–10 minute conversations between college students and their friends. Gibbs found a much higher proportion of verbal irony: 4.7 instances per conversation, or about 8 % of the conversational turns. However, Gibbs’ results are difficult to interpret because he chose to define verbal irony very broadly: he coded all instances of jocularity, sarcasm, hyperbole, rhetorical questions, and understatements as instances of irony. More typically, these forms of nonliteral language are considered separately, since they are not all counterfactual and may have different discourse goals (Roberts and Kreuz 1994). So although it is difficult to estimate how often people employ verbal irony, it seems clear that such language is used rather frequently. 4. Theoretical approaches A number of researchers have proposed theories to explain the production and processing of verbal irony. Although these theories have many similarities, they differ in emphasis and degree of empirical support. 4.1. The Tinge Hypothesis Dews and her colleagues (Dews, Kaplan, and Winner 1995; Dews and Winner 1995) have proposed that an important social function of verbal irony is to mute or attenuate the meaning of the utterance. These researchers have dem-

Social aspects of verbal irony use   329

onstrated that ironic criticisms (e.g., “You’re a real genius!”) are perceived as less aggressive than literal criticisms (e.g., “I think you are very stupid”), and that ironic compliments (e.g., “It’s too bad you have no talent as a runner”) are perceived as more aggressive than literal compliments (e.g., “You’re a very talented runner”). A similar conclusion can be found in the results of Jorgensen (1996). Viewed in this way, verbal irony serves the purpose of adjusting the affective content of an utterance. 4.2. Discourse goals approach Kreuz, Long, and Church (1991) asked experimental participants to read scenarios which concluded with ironic or literal statements. Next, they indicated, via a discourse goals checklist, which goals were intended by these statements (e.g., to criticize, to boast, or to be funny). Participants endorsed a larger number of discourse goals for the ironic than for the literal statements, suggesting that verbal irony might fulfill a larger number of discourse goals than literal language. Roberts and Kreuz (1994) expanded on this insight by proposing that various forms of nonliteral language differ in the discourse goals that they fulfill. In this study, participants were assigned to one of eight nonliteral language conditions (hyperbole, idiom, indirect requests, irony, metaphor, rhetorical questions, simile, and understatement). They were shown several examples of each form and were asked to generate additional examples. Finally, they were asked to provide the reasons why a person might choose to speak in this way. A discourse goal taxonomy was generated from the participants’ responses, and as expected, there were large differences between the nonliteral language forms. For example, “to be polite” was commonly listed as a discourse goal for indirect requests, but almost never for the other nonliteral forms. In the case of irony, the most frequent responses were “to show negative emotion” and “to be humorous.” These responses highlight the utility of verbal irony: such remarks allow speakers to express hostility in a socially acceptable way (see also Kreuz 2000). 4.3. Goal oriented theory In a similar vein, Attardo (2000) has proposed a number of related and additional goals of ironic communication, including group affiliation, sophistication, retractability, evaluation, rhetorical functions, and its role as a politeness strategy. Several of these goals will be explored in Section 5.

330   Roger J. Kreuz and Gina M. Caucci 4.4. Heuristics approach Another way of thinking about verbal irony focuses on the interplay between speaker and listener. Can speakers be sure that the listener will interpret their remarks as intended? In order to increase this likelihood, speakers may exploit the cues and constraints associated with verbal irony, and which will be explored in Sections 6 through 8. The listeners’ job is to make use of these cues and constraints, and to determine, as best they can, whether the speaker is being literal or nonliteral. The listener can make use of certain rules of thumb, and while these heuristics offer a good chance of a correct interpretation, there is always the possibility of error (Kreuz 1996, 2000). Seen in this way, the understanding of ironic statements can be thought of as probabilistic, and not absolute. 4.5. Implicit display theory Utsumi (2000) has fleshed out a heuristics-based approach to explain how ironic statements are distinguished from those not intended as ironic. Specifically, he proposes that ironists make use of an ironic environment, and that the utterance is displayed implicitly (i.e., the ironist does not explicitly announce his or her ironic intent). He also suggests that listeners compare ironic utterances with prototypes of irony, and if the match is high, then the listener can be more certain of a correct interpretation. As with the heuristic-based approaches in general, this theory predicts that listeners will sometimes fail to interpret a speaker’s intentions correctly. 5. Goals of ironic communication In this section, we will briefly review the reasons why speakers choose to express themselves via verbal irony. Making sense of this literature is not easy, because researchers have approached this task from a variety of perspectives, and have made use of a diverse number of research paradigms. Perhaps not surprisingly, therefore, the list of goals is large and even contradictory. 5.1. To be polite Brown and Levinson (1987), in their influential analysis of linguistic politeness, classified verbal irony as an off-record strategy that allows face-threatening

Social aspects of verbal irony use   331

statements to be expressed via violations of Grice’s (1975) Maxim of Quality. Since verbal irony is often used to express negative emotion (see Section 4.2), verbal irony can be characterized as a face-saving technique. Politeness is also central to Barbe’s (1995) characterization of verbal irony. 5.2. To be humorous Many accounts of humor have highlighted the role played by verbal irony (e.g., Long and Graesser 1988), and this observation has been verified empirically. The participants in the Kreuz, Long, and Church (1991) study described in Section 4.2 were almost seven times more likely to endorse the discourse goal “to be funny or witty” as an attribute of ironic statements in comparison to equivalent literal statements. Similarly, humor was the second most often mentioned discourse goal by the participants in the Roberts and Kreuz (1994) study. 5.3. To express negative emotion Nonliteral language has frequently been cited as a way to verbalize ideas that are difficult to express via literal language, such as emotional states (e.g., Fainsilber and Ortony 1987; Ortony 1975), although this idea has primarily been studied in terms of metaphor usage. However, “to show negative emotion” was the most frequently mentioned discourse goal for irony in the Roberts and Kreuz (1994) study, with 94 % of the participants mentioning it in one form or another. The relation between negative emotion and sarcasm also appears in experiments of Jorgensen (1996), and those of Leggitt and Gibbs (2000), who found that speakers using sarcasm were perceived as angry, disgusted, or scornful. Irony has also been shown to be a common way of expressing surprise (Colston 1997a). 5.4. To criticize As described in Section 4.1, one function of verbal irony may be to mute criticism (Dews, Kaplan, and Winner 1995; Dews and Winner 1995), although other researchers have found the opposite. The critical aspect of sarcasm was explored by Jorgensen (1996), who found that sarcasm was typically employed by speakers to criticize or to lodge complaints against the listener. Colston (1997b) found that ironic criticism seems to enhance rather than dilute condemnation, and Toplak

332   Roger J. Kreuz and Gina M. Caucci and Katz (2000) obtained similar results. These inconsistent findings may have been resolved by Pexman and Olineck (2002a), who found that the perspective taken in interpretation was a key factor. When speaker intent is assessed, irony may be perceived as more critical than literal statements. However, when the general social impression created by the remark is assessed, ironic statements may be perceived as less critical than literal statements (see also Pexman 2005). 5.5. To persuade Other forms of nonliteral language, such as metaphor and analogy, have been characterized as more persuasive than literal language (for a review, see Kreuz and Ashley 2006). Gibbs and Izett (2005) make a similar claim for irony, arguing that such statements require listeners to carefully reflect upon the speaker’s message in order to make sense of the incongruity between what is said and what is intended. 5.6. To increase cohesiveness and create exclusiveness Fowler (1965), in his classic analysis of the concept of humor, claimed that the motive or aim of irony is to create exclusiveness. The audience for such remarks is an “inner circle” that can appreciate the discrepancy between what is said and what is intended, and this serves to exclude anyone who cannot appreciate this disparity. Similar ideas appear in the writing of Myers Roy (1981) and in Clark and Gerrig’s (1984) pretense theory of verbal irony. In considering the risks and rewards of ironic communication, Gibbs and Colston (2001) have proposed that irony may bond or distance speakers from listeners. As can be seen, this diversity of goals suggests that verbal irony can be employed for a wide variety of reasons. It also suggests why the ironist would be willing to run the risk of misunderstandings: verbal irony can accomplish discourse goals that would be, in many cases, difficult to accomplish via literal language. 6. Speaker characteristics It would be a mistake to suppose that everyone uses verbal irony in the same way, or to the same degree. This section will briefly review the ways in which the use or interpretation of irony is affected by the characteristics of those who employ it.

Social aspects of verbal irony use   333

6.1. Individual differences Ivanko, Pexman, and Olineck (2004) created a sarcasm self-report scale (SSS) in order to assess the degree to which individuals use verbal irony. They found that SSS scores predicted participants’ selections of ironic statements in a production task, and that the scores were related to the reading times for such statements. These results are important because they provide evidence for the intuitive notion that people differ in their propensity to use verbal irony, and perhaps also in their ability to interpret it as intended. 6.2. Regional variation There may also be differences in verbal irony use within the same linguistic community. Dress et al. (2008) gave part of the SSS to participants in Northern and Southern dialect regions in the United States, along with a variety of production tasks. We found that Northern speakers self-reported using sarcasm more than Southern speakers, and that their self-report scores correlated moderately with their use of sarcasm. Northern participants generated sarcastic completions to scenarios at a rate that was significantly higher than the Southern participants. We also found that the two groups defined sarcasm differently: Northern participants were more likely than the Southern participants to mention the role of humor in sarcasm. This may, in fact, explain the difference in sarcasm use between the two groups: the Northern participants may use sarcasm more because they perceive it as more humorous, and therefore less threatening. 6.3. Gender Several studies have been conducted to determine whether men or women are more likely to use verbal irony. In general, the results suggest that men use irony more than women (Colston and Lee 2004; Gibbs 2000; Jorgensen 1996; Katz, Piasecka, and Toplak 2001). However, this simple statement may need to be qualified in a number of ways. Dress et al. (2008) found that, although men had higher sarcasm self-report scores than women, the two groups did not differ in terms of linguistic behavior (i.e., generating sarcastic completions to scenarios). In a study of emotions and nonliteral language, Link and Kreuz (2005) found no main effect of gender, but an interaction with emotional valence: men were more likely than women to use nonliteral

334   Roger J. Kreuz and Gina M. Caucci language to describe negative emotions. Although the Link and Kreuz (2005) study primarily involved the use of metaphor, these results suggest that other factors may play a role in the relation between gender and verbal irony use. 6.4. Speaker occupation Just as we might expect a young male to be more sarcastic than an elderly female, would we expect a cab driver to be more sarcastic than a clergyman? Speaker occupation does play an important role in how potentially sarcastic remarks are interpreted (Katz and Pexman 1997; Pexman 2005; Pexman and Olineck 2002b). More specifically, Pexman and Olineck (2002b) determined that particular traits (the tendencies to be humorous, to criticize, to be sincere, and an occupation’s educational level) significantly accounted for the variance in participants’ sarcasm ratings. 7. Situation characteristics Over and above the characteristics of particular speakers, there are elements of situations themselves which affect the production and processing of verbal irony. 7.1. Asymmetry of affect Verbal irony typically involves making positive evaluations of negative situations (e.g., “What lovely weather!” spoken during a downpour). In contrast, negative evaluations of positive situations (e.g., “What terrible weather!” spoken on a beautiful sunny day) are perceived as either weakly sarcastic, or indicative of some other state of affairs (e.g., irritation on the part of the speaker, or the mocking of a failed prediction; Kreuz and Glucksberg 1989). Clark and Gerrig (1984) refer to this phenomenon as the asymmetry of affect. Online reading measures have shown that negative evaluations of positive situations take longer to comprehend than the reverse (Kreuz and Link 2002). These results suggest that sarcastic statements are more interpretable when they violate the default expectations that people have about events (for example, that the weather should be pleasant; Matlin and Stang 1978; see also Matthews, Hancock, and Dunham 2006 and Kumon-Nakamura, Glucksberg, and Brown 1995).

Social aspects of verbal irony use   335

7.2. Shared common ground Conversational participants also differ in terms of the knowledge and beliefs they share, and know that they share. This common ground (Clark 1996) exerts pervasive effects in communication, and also affects how verbal irony is interpreted. For example, participants perceive ironic remarks as more appropriate when they are exchanged between high common ground dyads in scenarios, in comparison to low common ground dyads (Kreuz, Kassler, Coppenrath, and McLain Allen 1999). In addition, ironic remarks directed at high common ground targets are perceived as more ironic, and read more quickly (Kreuz and Link 2002). 7.3. Perceived closeness Another interpersonal dimension of communication is the perceived closeness that two people share (e.g., Slugoski and Turnbull 1988). This is not the same as common ground, since an estranged couple might share a lot of common ground, yet feel quite distant from each other. Jorgensen (1996) found that appropriateness ratings for the use of sarcasm differed significantly depending on the degree of social distance or intimacy (e.g., a sibling versus a professor). In a similar vein, Kreuz (1996) asked participants to rate perceived closeness and use of sarcasm for a variety of relationships (e.g., best friend, distant relatives, and classmates), and found a robust positive correlation. 7.4. Interaction modality Technology allows people to communicate through many different modalities (e.g., by letter, telephone, e-mail, or text messaging). These forms differ greatly in their richness, and one might expect that the presence or absence of paralinguistic cues (see section 8) would affect whether people make use of verbal irony. It might be that individuals avoid using irony in impoverished mediums to lessen the chances of miscommunication. Somewhat surprisingly, this does not seem to be the case. Hancock (2004) found that participants used more verbal irony in computer-mediated conversations than when conversing face-to-face. Instead of accepting the limitations of a conversational medium, people seem to find ways to adapt them to their needs. This has occurred in the case of e-mail, for example: many people use emoticons, or “smileys” to indicate nonserious intent (Eisenberg 1994).

336   Roger J. Kreuz and Gina M. Caucci 8. Cues of ironic intent How do speakers signal their ironic intentions to their listeners? It seems that speakers have a variety of cues at their disposal, which range from the words they use to the way they speak them. In this section, we will review some of the linguistic and paralinguistic cues that signal ironic intent. 8.1. Exaggeration Kreuz and Roberts (1995) suggested that exaggeration, or hyperbole, and verbal irony are frequently intertwined. Consistent with this view, Leggitt and Gibbs (2000) found that emotional reactions to hyperbole and sarcasm are very similar. It may be that the extreme exaggeration often associated with irony helps to signal the nonliteral intentions of the speaker. For example, someone who describes a good meal as being “The best meal I ever had in my entire life” would be speaking hyperbolically. On the other hand, someone who describes a bad meal as “The best meal I ever had in my entire life” is being ironic. In both cases, the extreme nature of the words used (“ever,” “entire”) lets the listener know that the speaker is most likely not being literal. This cue may not be sufficient, however, which is why speakers tend to combine cues, such as rolling their eyes, or changing their tone of voice, when they deliver an ironic utterance. Roberts and Kreuz (1994) noted that irony and hyperbole share similar discourse goals, such as “to be humorous” and “to emphasize.” These researchers also found that hyperbolic ironic statements were judged as more ironic than nonhyperbolic ironic statements (Kreuz and Roberts 1995). In addition, irony performs these discourse functions better than hyperbole (Colston and O’Brien 2000). However, depending on the explicitness of expectations, irony or hyperbole can be more effective (Colston 1997a). 8.2. Lexical factors Much of the exaggeration in hyperbole and irony can be attributed to certain extreme adjective-adverb collocations (e.g., “absolutely amazing,” “simply magnificent”; Kreuz and Roberts 1995). Utsumi (2000) has suggested that these lexical factors are used to implicitly express a negative attitude with a positive evaluation. It is by this discordance that the listener is able to understand that the speaker is being ironic or at least nonliteral.

Social aspects of verbal irony use   337

8.2.1. Adjectives and adverbs Kreuz and Caucci (2007) attempted to empirically test the notion that extreme adjective-adverb collocations cue readers to ironic intent. Excerpts were taken from published works that originally contained the phrase “said sarcastically.” The word “sarcastically” was removed from the excerpts and shown to participants. There were 100 excerpts that had the word “sarcastically” deleted, as well as 15 similar excerpts used as fillers. The participants were asked to rate, on a 7 point Likert-type scale, how sarcastic they thought the speaker was. As expected, the sarcastic items were judged as being more sarcastic than the filler items. However, a regression analysis determined that the adjective-adverb collocation variable failed to account for a significant amount of the variance in participants’ ratings. Unfortunately, the excerpts were coded rather coarsely, in a binary fashion, with 0 indicating no adjectives or adverbs and 1 indicating the presence of one or more adjectives or adverbs. When the excerpts were recoded and only extreme adjective or adverbs were coded, this factor approached significance as a predictor of the sarcasm ratings. 8.2.2. Interjections Kreuz and Caucci (2007) also coded the excerpts for interjections to determine if they might act as a cue for ironic intent. Interjections were defined as abrupt remarks made as asides or interruptions (e.g., “gee,” “gosh,” and “oh”). Kreuz and Roberts (1994) had first proposed that interjections might be an important lexical cue for irony, but this possibility had not been tested previously. Results from the regression analysis indicated that interjections did account for a significant amount of the variance, although the total amount of variance accounted for was rather low (about 5 %). In other words, the sarcastic utterances that contained interjections were rated as more sarcastic than those that did not contain any interjections. Even though interjections represent a small class of words, they seem to play an important role in cuing ironic intent. 8.3. The ironic tone of voice Although the expression of sarcasm or irony does not depend on verbal cues, speakers often use them in order to help the addressee realize that what is being said is not what is meant. Researchers have proposed a variety of bodily and paralinguistic cues that might be used to signal ironic intent (e.g., eye-rolling, pointing, laughing, eyebrow raising, and tone of voice; see Attardo et al. 2003;

338   Roger J. Kreuz and Gina M. Caucci Cutler 1974; Haiman 1998; Kreuz 1996). It is to one of these cues, tone of voice, that we now turn. Cutler (1974) described what is referred to as the ironic tone of voice (ITOV) as possessing three factors: a slow speaking rate, heavy stress, and nasalization (see also Attardo 2000). Other researchers noted that sarcasm can be characterized by intonational “misfits,” in that words with positive meanings are spoken with negative prosody (Haiman 1998). However, it has also been suggested that tone of voice is not a reliable cue for irony and may be conflated with other factors, such as hyperbole (Cutler 1974; Kreuz and Roberts 1995). More recent research, however, has shown people do make use of specific acoustic information to infer ironic intent (e.g., Bryant and Fox Tree 2002). A number of specific acoustic parameters have been identified as characterizing the ITOV, and these include particular rates of articulation, pitch, and amplitude. However, there are discrepancies among the recent research findings. For example, Rockwell (2000) recruited six male and six female speakers to record non-sarcastic, natural sarcastic and posed sarcastic utterances. These speakers were chosen for their expressive voices. Results showed that listeners were unable to differentiate between spontaneous sarcasm and nonsarcasm, but they were able to tell when the utterances were posed sarcastic as opposed to nonsarcastic. Rockwell claimed that the sarcastic utterances could be characterized by having slower tempo, greater intensity, and a lower pitch. In contrast, Bryant and Fox Tree (2005) found amplitude variability to be the only significant factor in their analysis of the ironic and nonironic speech of radio talk show hosts. Anolli, Ciceri, and Infantino (2002) also attempted to determine acoustic parameters associated with ITOV. In this case, however, the researchers wanted to differentiate between kind irony and sarcastic irony. Kind irony was defined as implying admiration through negative words (praise by blame), whereas sarcastic irony was defined as an intention to harm someone through words of praise (blame by praise). Another variable the researchers thought might affect how one expresses irony is how heavily the utterance depends on the context to be understood. Low context utterances should not depend as much on context to be understood and in turn should require less in terms of vocal cues to be understood as ironic. However, high context utterances should depend greatly on context to be understood and in turn require more vocal cues from the speaker. In four different experiments, target utterances were varied based on low context (LC) versus high context (HC) utterances. All utterances in the first experiment were LC utterances. Overall, results from the first study revealed that there was an increase in pitch for ironic speech compared to normal speech.

Social aspects of verbal irony use   339

Also, sarcastic irony differed significantly from kind irony on this parameter, with the former being characterized as having higher pitch. Both ironic types showed a higher amplitude than normal speech. The second experiment compared only HC utterances. Results from this study showed that only sarcastic irony had higher pitch, but both ironic types were spoken with higher amplitude than normal speech. Finally, both ironic types were characterized as having a slow rate of articulation and more pauses. Exaggeration can also refer to prosody rather than just hyperbole. Haiman (1998) noted that ironic and literal speech often differs only in stress or exaggeration. For example, the phrase “Don’t you just love . . .” could be spoken in either a literal tone or an ironic tone. If the stress were put on the word “just,” for example, the sentence would be perceived as more ironic (“Don’t you JUST love people who use turn signals!”). On the other hand, if the speaker applies stress to the word “love,” the sentence would be perceived as more literal (“Don’t you just LOVE the sunset!”). 9. Determining ironic intent How accurate are listeners in inferring ironic intent? Given the plethora of cues that speakers can make use of, one might suppose that accuracy must be rather high. However, we can all think of cases in which our conversational partners either missed our ironic intentions, or mistakenly believed that a literal statement was intended ironically (see Gibbs and Colston 2001; Gibbs, O’Brien, and Doolittle 1995). This suggests that the process of communicating irony is far from foolproof. How can we explain these communicative failures? 9.1. Principle of inferability As part of a heuristics-based approach to verbal irony (Section 4.4), Kreuz (1996) proposed a principle of inferability: speakers will employ ironic statements if they believe it is likely that they will be understood. If inferability is low, however, then the speaker should play it safe, and not use irony. This approach is appealing because it suggests that determining ironic intent should be relatively accurate, although as with any heuristic, success is not guaranteed. A number of studies support this idea (see Sections 7.2 and 7.3), although at least one has not. Hancock (2004) found that his participants used verbal irony in a computer-mediated situation in which inferability should have been low (see Section 7.4). Additional research may be needed to determine under which

340   Roger J. Kreuz and Gina M. Caucci circumstances people employ a heuristics-based approach, and under which circumstances they do not. 10. Conclusions This review of the literature suggests that verbal irony is a complex pragmatic phenomenon, and one in which social, affective, situational, and linguistic factors all play a role. The study of verbal irony tells us a great deal about language in general, and future research on this topic seems certain to increase our understanding of one of humanity’s most important behaviors. Note Support for the preparation of this chapter was provided by a Center of Excellence grant from the State of Tennessee to the Department of Psychology at the University of Memphis. Correspondence concerning this chapter may be addressed to Roger J. Kreuz, Department of Psychology, The University of Memphis, Memphis, TN 38152 USA ([email protected]).

References Anolli, L., R. Ciceri, and M.G. Infantino From “blame by praise” to “praise by blame”: Analysis of vocal pat2002 terns in ironic communication. International Journal of Psychology 37: 266–276. Attardo, S. 2000 Irony markers and functions: Towards a goal-oriented theory of irony and its processing. International Journal of Language and Communication 12: 3–20. Attardo, S., J. Eisterhold, J. Hay, and I. Poggi 2003 Multimodal markers of irony and sarcasm. Humor 16: 243–260. Barbe, K. 1995 Irony in Context. Amsterdam: John Benjamins Publishing. Booth, W.C. 1974 A Rhetoric of Irony. Chicago: University of Chicago Press. Brown, P., and S.C. Levinson 1987 Politeness: Some Universals in Language Use. Cambridge: Cambridge University Press. Bryant, G.A., and J.E. Fox Tree Recognizing verbal irony in spontaneous speech. Metaphor and Symbol 2002 17: 99–117.

Social aspects of verbal irony use   341 Bryant, G.A., and J.E. Fox Tree 2005 Is there an ironic tone of voice? Language and Speech 48: 257–277. Clark, H.H. 1996 Using Language. Cambridge: Cambridge University Press. Clark, H.H., and R.J. Gerrig 1984 On the pretense theory of irony. Journal of Experimental Psychology: General 113: 121–126. Colston, H.L. 1997a “I’ve never seen anything like it”: Overstatement, understatement, and irony. Metaphor and Symbol 12: 43–58. 1997b Salting a wound or sugaring a pill: The pragmatic functions of ironic criticism. Discourse Processes 23: 35–45. Colston, H.L., and S.Y. Lee Gender differences in verbal irony use. Metaphor and Symbol 19: 289– 2004 306. Colston, H.L., and J. O’Brien Contrast of kind versus contrast of magnitude: The pragmatic accom2000 plishments of irony and hyperbole. Discourse Processes 30: 179–199. Cutler, A. 1974 On saying what you mean without meaning what you say. Papers from the Tenth Regional Meeting of the Chicago Linguistics Society, 117– 127. Chicago: Department of Linguistics, University of Chicago. Dews, S., J. Kaplan, and E. Winner Why not say it directly? The social functions of irony. Discourse Proc1995 esses 19: 347–367. Dews, S., and E. Winner Muting the meaning: A social function of irony. Metaphor and Symbol 1995 10: 3–19. 1997 Attributing meaning to deliberately false remarks: The case of irony. In The Problem of Meaning: Behavioral and Cognitive Perspectives, C. Mandell and A. McCabe (eds.), 377–414. Amsterdam: Elsevier. Dress, M.L., R.J. Kreuz, K.E. Link, and G.M. Caucci Regional variation in the use of sarcasm. Journal of Language and So2008 cial Psychology 27: 71–85. Du Bois, J.W., W.L., Chafe, C. Meyer, and S.A. Thompson Santa Barbara corpus of spoken American English, Part 1. Philadelphia: 2000 Linguistic Data Consortium. Eisenberg, A. 1994, April E-mail and the new epistolary age. Scientific American 270: 128. Fainsilber, L., and A. Ortony Metaphorical uses of language in the expression of emotion. Metaphor 1987 and Symbolic Activity 2: 239–250. Fowler, H.W. 1965 A Dictionary of Modern English Usage, 2d ed. Oxford: Oxford University Press.

342   Roger J. Kreuz and Gina M. Caucci Gibbs, R.W., Jr. 1994 The Poetics of Mind: Figurative Thought; Language, and Understanding. Cambridge: Cambridge University Press. 2000 Irony in talk among friends. Metaphor and Symbol 15: 5–27. 2003 Nonliteral speech acts in text and discourse. In Handbook of Discourse Processes, A.C. Graesser, M.A. Gernsbacher, and S.R. Goldman (eds.), 357–393. Mahwah, NJ: Lawrence Erlbaum Associates. Gibbs, R.W., Jr., and H.L. Colston The risks and rewards of ironic communication. In Say not to Say: New 2001 Perspectives on Miscommunication, L. Anolli, R. Ciceri, and G. Riva (eds.). Amsterdam: IOS Press. Gibbs, R.W., Jr., and C.D. Izett Irony as persuasive communication. In Figurative Language Compre2005 hension: Social and Cultural Influences, H.L. Colston and A.N. Katz (eds.), 131–151. Mahwah, NJ: Lawrence Erlbaum Associates. Gibbs, R.W., Jr., and J.E. O’Brien 1991 Psychological aspects of irony understanding. Journal of Pragmatics 16: 523–530. Gibbs, R.W., Jr., J.E. O’Brien, and S. Doolittle Inferring meanings that are not intended: Speaker’s intentions and irony 1995 comprehension. Discourse Processes 20: 187–203. Giora, R. 2003 On Our Mind: Salience, Context, and Figurative Language. Oxford: Oxford: University Press. Grice, H.P. 1975 Logic and conversation. In Syntax and Semantics: Vol. 3. Speech Acts, P. Cole and J. Morgan (eds.), 41–58. New York: Academic Press. 1978 Further notes on logic and conversation. In Syntax and Semantics: Vol. 9. Speech Acts, P. Cole and J. Morgan (eds.), 113–127. New York: Academic Press. Haiman, J. 1998 Talk is Cheap: Sarcasm, Alienation, and the Evolution of Language. Oxford: Oxford University Press. Hancock, J.T. 2004 Verbal irony use in face-to-face and computer-mediated conversations. Journal of Language and Social Psychology 23: 447–463. Ivanko, S.L., P.M. Pexman, and K.M. Olineck 2004 How sarcastic are you? Individual differences and verbal irony. Journal of Language and Social Psychology 23: 244–271. Jorgensen, J. 1996 The functions of sarcastic irony in speech. Journal of Pragmatics 26: 613–634. Katz, A.N. 2005 Discourse and sociocultural factors in understanding nonliteral language. In Figurative Language Comprehension: Social and Cultural

Social aspects of verbal irony use   343 Influences, H.L. Colston and A.N. Katz (eds.), 183–207. Mahwah, NJ: Lawrence Erlbaum Associates. Katz, A.N., and P.M. Pexman Interpreting figurative statements: Speaker occupation can change met1997 aphor to irony. Metaphor and Symbol 12: 19–41. Katz, A.N., I. Piasecka, and M. Toplak 2001 Comprehending the sarcastic comments of males and females. Abstracts of the Psychonomic Society 6: 77. Kreuz, R.J. 1996 The use of verbal irony: Cues and constraints. In Metaphor: Implications and Applications, J.S. Mio and A.N. Katz (eds.), 23–38. Mahwah, NJ: Lawrence Erlbaum Associates. 2000 The production and processing of verbal irony. Metaphor and Symbol 15: 99–107. Kreuz, R.J., and A. Ashley 2006 Nonliteral language, persuasion, and memory. In Language and Memory: Aspects of Knowledge Representation, H. Pishwa (ed.), 425–443. Berlin: Mouton de Gruyter. Kreuz, R.J., and G.M. Caucci Lexical influences on the perception of sarcasm. Proceedings of the 2007 Workshop on Computational Approaches to Figurative Language 1–4, 1777–1806. New Brunswick, NJ: Association for Computational Linguistics. Kreuz, R.J., and S. Glucksberg How to be sarcastic: The echoic reminder theory of verbal irony. Jour1989 nal of Experimental Psychology: General 118: 374–386. Kreuz, R.J., M.A. Kassler, L. Coppenrath, and B. McLain Allen 1999 Tag question and common ground effects in the perception of verbal irony. Journal of Pragmatics 31: 1685–1700. Kreuz, R.J., and K.E. Link 2002 Asymmetries in the use of verbal irony. Journal of Language and Social Psychology 21: 127–143. Kreuz, R.J., D.L. Long, and M.B. Church On being ironic: Pragmatic and mnemonic implications. Metaphor and 1991 Symbolic Activity 6: 149–162. Kreuz, R.J., and R.M. Roberts 1993 On satire and parody: The importance of being ironic. Metaphor and Symbolic Activity 10: 21–31. Kreuz, R.J., and R.M. Roberts 1995 Two cues for verbal irony: Hyperbole and the ironic tone of voice. Metaphor and Symbolic Activity 10: 21–31. Kumon-Nakamura, S., S. Glucksberg, and M. Brown How about another piece of pie: The allusional pretense theory of dis1995 course irony. Journal of Experimental Psychology: General 124: 3–21.

344   Roger J. Kreuz and Gina M. Caucci Lee, C.J., and A.N. Katz 1998 The differential role of ridicule in sarcasm and irony. Metaphor and Symbol 13: 1–15. Leggitt, J.S., and R.W. Gibbs, Jr. 2000 Emotional reactions to verbal irony. Discourse Processes 29: 1–24. Link, K.E., and R.J. Kreuz 2005 Do men and women differ in their use of nonliteral language when they talk about emotions? In Figurative Language Comprehension: Social and Cultural Influences, H.L. Colston and A.N. Katz (eds.), 153–180. Mahwah, NJ: Lawrence Erlbaum Associates. Long, D.L., and A.C. Graesser Wit and humor in discourse processing. Discourse Processes 11: 35–60. 1988 Lucariello, J. 1994 Situational irony: A concept of events gone awry. Journal of Experimental Psychology: General 123: 129–145. Matlin, M.W., and D.J. Stang 1978 The Pollyanna Principle: Selectivity in Language, Memory, and Thought. Cambridge, MA: Schenkman. Matthews, J.K., J.T. Hancock, and P.J. Dunham 2006 The roles of politeness and humor in the asymmetry of affect in verbal irony. Discourse Processes 41: 3–24. Mish, F.C. 1993 Merriam-Webster’s Collegiate Dictionary. 10th ed. Springfield, MA: Merriam-Webster, Inc. Myers Roy, A. 1981 The functions of irony in discourse. Text 1: 407–423. Nunberg, G. 2001 The Way We Talk Now: Commentaries on Language and Culture. Boston: Houghton Mifflin. Ortony, A. 1975 Why metaphors are necessary and not just nice. Educational Theory 25: 45–53. Pexman, P.M. 2005 Social factors in the interpretation of verbal irony: The roles of speaker and listener characteristics. In Figurative Language Comprehension: Social and Cultural Influences, H.L. Colston and A.N. Katz (eds.), 209– 232. Mahwah, NJ: Lawrence Erlbaum Associates. Pexman, P.M., and K.M. Olineck Does sarcasm always sting? Investigating the impact of ironic insults 2002a and ironic compliments. Discourse Processes 33: 199–217. 2002b Understanding irony: How do stereotypes cue speaker intent? Journal of Language and Social Psychology 21: 245–274. Preminger, A., and T.V.F. Brogan (eds.) The Princeton Encyclopedia of Poetry and Poetics. New York: MJF 1993 Books.

Social aspects of verbal irony use   345 Roberts, R.M., and R.J. Kreuz Why do people use figurative language? Psychological Science 5: 159– 1994 163. Rockwell, P. 2000 Lower, slower, louder: Vocal cues of sarcasm. Journal of Psycholinguistic Research 29: 483–495. Roncero, C., J.M. Kennedy, and R. Smith Similes on the Internet have explanations. Psychonomic Bulletin and 2006 Review 13: 74–77. Slugoski, B., and W. Turnbull Cruel to be kind and kind to be cruel: Sarcasm, banter, and social rela1988 tions. Journal of Language and Social Psychology 7: 101–121. Tannen, D. 1984 Conversational Style: Analyzing Talk Among Friends. Norwood, NJ: Ablex. Toplak, M., and A.N. Katz 2000 On the uses of sarcastic irony. Journal of Pragmatics 32: 1467–1488. Trumble, W.R., and A. Stevenson 2003 Shorter Oxford English Dictionary on Historical Principles. 5th ed. Oxford: Oxford University Press. Utsumi, A. 2000 Verbal irony as implicit display of ironic environment: Distinguishing ironic utterances from nonirony. Journal of Pragmatics 32: 177–180. Wasserman, P., and M.F. Schober 2006 Variability in judgments of spoken irony. Abstracts of the Psychonomic Society 11: 43. Wiener, P.P. 1973 Dictionary of the History of Ideas, Vol 2. New York: Charles Scribner.

Section III Social cognitive functions of single structures

Chapter 13 Attribution theories wired into linguistic categories* Klaus Fiedler and Peter Freytag

1. Introduction Language affords a rich and versatile toolbox for doing all kinds of things. We use language, first of all, to share our thoughts and intentions with others. Apart from this intentionally cooperative function, language tools can also serve an instrumentally manipulative function. For instance, when language is used to deter or to intimidate others, to deceive or to distract, to ingratiate or to negotiate, the true function of linguistic stimuli is often to conceal the information that the speaker has in mind – and to prevent the recipient from inferring the underlying intention. In addition to the cooperative and manipulative functions of language, however, there are still other, unintended consequences of verbal interaction. Some of the most consequential and creative functions of language arise as unintended, emergent by-products of language use. Examples of such emergent effects would be amusement arising from communication errors, problem solutions resulting from free verbal associations, irritation in interaction partners originating in communicators’ lack of verbal or non-verbal control, or the automatic activation of goals through verbal priming. The present article is devoted to the exciting topic of language and attribution. We believe that studying this topic provides us with insights and vivid examples of all functions of language – cooperative, manipulative, and unin* The present research was supported by various grants from the Deutsche For­ schungsgemeinschaft. Correspondence should be addressed to Klaus Fiedler or to Peter Freytag, Department of Psychology, University of Heidelberg, Hauptstrasse 47–51, 69117 Heidelberg, Germany. Fax: + 49 6221 547745. Requests via electronic mail may be sent to [email protected] or to [email protected].

350   Klaus Fiedler and Peter Freytag tended haphazard functions. In the next section, we first provide an overview of what the psychology of attribution is about, and we outline two major theories of attribution. Then, in another section, we introduce a taxonomy of linguistic categories and we point out that the two major theories of attribution are wired into the semantic and pragmatic meaning of two word classes, adjectives and interpretive action verbs. The remainder of this chapter will then describe studies that illustrate the manner in which these two word classes trigger different attribution processes, some of which may be intended by the speaker, some of which may be experienced consciously by the receiver, but many of which are neither intended nor consciously recognized. The final discussion will address practical implications of the close link between language and attribution. 2. The psychology of attribution The theoretical and practical questions that we come to mind when we refer to “attribution” are clearly conveyed in the title of the seminal book by Jones et al. (1972) that constitutes the beginning of a huge research program: Attribution: Perceiving the causes of behavior. Thus, the psychology of attribution is concerned with the processes through which people arrive at an explanation of behavioral episodes or events – or, put differently, with the question of how people answer why-questions. These questions can take many different forms: What is the cause of a student’s success, of a crime, of an accident, or of a historical event, such as the emergence of globalization? Why do consumers prefer one brand to another? Who started the fight? No doubt, the answers to such why-questions are at the heart of many social, educational, economical, political, and health-related problems. Parents’ attribution of children’s behavior determines reward and punishment. Jurors’ attributions of a defendant’s behavior guide sentencing decisions. Partner attributions are crucial for the satisfaction and the longevity of partnerships. Students’ achievement motivation depends on whether they attribute their academic success to ability or to good luck, and whether they attribute their failure to inability or to a lack of effort. Similarly, the recovery of in-patients is partly determined by their compliance with medical treatments, and compliance in turn depends on in-patients’ self-efficacy beliefs. Indeed, a major lesson from psychological research on stress, health, achievement, and life-satisfaction is that the subjective attribution of states is more important for people’s wellbeing than the objective states themselves.

Attribution and linguistic categories   351

2.1. Dimensions of attribution Attribution processes proceed along distinct dimensions. By far the most prominent dimension involves the distinction between internal factors within individuals and external conditions in the situation (Jones and McGillis 1976). For example, in a close relationship, the occurrence of a conflict may be attributed to the partner’s internal disposition or personality (“because of your laziness”) or to external circumstances in the situation (“because of the permanent stress on the job”). The term “locus of attribution” usually refers to this fundamental distinction. The generalized tendency, especially in Western cultures, to explain behavior in terms of internal dispositions and to neglect external constraints, is called the fundamental attribution error (Nisbett and Ross 1980). This overall bias is moderated by perspective. The actor-observer bias (Jones and Nisbett 1972) reflects the well-established finding that actors are less susceptible to the fundamental attribution error than are observers. That is, actors are more likely to consider situational factors than observers, whose perspective hardly reveals the situational constraints imposed on the actor’s behavior. The actor-observer bias in turn is moderated by the self-serving bias (Kunda 1990) in that the tendency to consider others, but not oneself, as a cause is more pronounced for explanations of undesirable as opposed to desirable behaviors. Attributional conflicts along the dimension of internal versus external causation are thus often a matter of perspective and desirability. The second attributional dimension to be emphasized here is controllability (Reisenzein 1986). To keep within the example, partnership problems may be attributed to controllable factors, which may be internal (“because you did not try hard enough”) or external (“because your new job is too distressing”). Or problems may be ascribed to uncontrollable factors, which may also be internal (“because of your allergy”) or external (“because society has changed”). A third dimension that matters a lot is stability across time and task situations (Weiner 1985). For instance, attributing a conflict to the partner’s restricted intelligence implies a stable and global trait that will hardly ever change, and that generalizes across many occasions. In contrast, blaming the partner for his or her use of a sensitive phrase, or failure to listen carefully, would imply a local and transient cause that is unlikely to preserve or to generalize to other occasions. All three dimensions can greatly influence the consequences of attribution. In the partnership context, attribution styles are highly predictive of satisfaction and psychological health. Attributions of negative behavior to internal, controllable, general causes are diagnostic of dissatisfaction, break-up, and divorce. Attributions to external, controllable and more specific factors are typically

352   Klaus Fiedler and Peter Freytag observed in more satisfying relationships (Fincham 1985). Self-attributions of stress or failure to internal, stable, and uncontrollable causes are a major determinant in the etiology of depression (Peterson and Seligman 1984). In other contexts, such as academic achievement or sports, the self-attribution of success to global, internal and uncontrollable factors, such as talent or genius, may be the secret underlying strong achievement motivation. In any case, whether behaviors or events are attributed to internal or external causes, to controllable or uncontrollable factors, and to general or specific conditions, is of utmost importance for health, well-being and mutual understanding. 3. Linguistic categories and their attributional implications 3.1. The Linguistic Category Model Language styles and language use have a marked influence on everyday attributions exactly because the semantic and pragmatic meaning of different word classes varies systematically along the major attributional dimensions depicted in the previous section. The Linguistic Category Model (LCM; Semin and Fiedler 1988) distinguishes between four classes of linguistic terms that can be used as predicates in sentences that describe interpersonal behaviors. The defining and characteristic features of these four word classes – descriptive action verbs (DAVs), interpretive action verbs (IAVs), state verbs (SVs), and adjectives (ADJs) – are summarized in Table 1, along with some examples. All four word classes can be clearly defined linguistically. Adjectives and verbs are easy to distinguish anyway. SVs differ from action verbs (IAVs, DAVs) in terms of several criteria (e.g., no progressive from in English, anomalous imperative, “how long” rather than “how often” etc.). And IAVs differ from DAVs in that the former entail interpretations and evaluations and refer to classes of behaviors rather than to concrete behaviors or events. More important to the present analysis, however, the four lexical categories also differ systematically in terms of their semantic and pragmatic implications for the sentence subject and object, respectively. Semantic judgment studies using random samples of verbs and adjectives from the entire lexicon demonstrate, for instance, that with increasing abstractness of language – from DAVs to IAVs to SVs to ADJs – the informativeness of the sentence concerning the sentence subject increases, whereas the dependence of the meaning on the specific situation and the sentence object decreases. Moreover, from DAVs to ADJs, the temporal stability of the behavior described increases, and the

Attribution and linguistic categories   353

sentences become more debatable and less clearly verifiable, because the sentence meaning becomes increasingly dependent on subjective interpretation. At the same time, with increasing abstractness, the degree of control exerted by the subject over his or her behavior decreases, with ADJs and SVs denoting less controllable attributes than IAVs and DAVs. Figure 1 summarizes these major differences. It should be noted that these semantic and pragmatic differences between linguistic categories, which generalize across the entire lexicon, could be easily mapped onto the aforementioned attributional dimensions. First, informativeness concerning the sentence subject and independence from object and situation are obvious aspects of meaning that imply internal attributions to the subject of the sentence, as opposed to external attributions to the situation. Second, the generality dimension is inherent in the temporal stability and in the overall abstractness and globality of linguistic terms. Finally, the controllability of actions varies with word class, too, reaching its maximum for IAVs (cf. Figure 1). 3.2. Mapping theories of attribution onto linguistic categories Psychologically and rhetorically, there are two different ways or strategies for attributing causality or responsibility to internal factors within the sentence subject rather than to external factors originating in the situation or in the object of behavior. Intriguingly, these two modes of attribution represent the two most prominent attribution theories in social psychology, and they are wired into the semantics and pragmatics of two distinct word classes, ADJs and IAVs. 3.2.1. ADJs in Kelley’s covariation model Kelley’s (1967, 1973) covariation model predicts that behavior is attributed to internal causes within the subject when the behavior (a) is more pronounced in this particular person than in other persons (low consensus), (b) generalizes across other object persons (low distinctiveness), and (c) is stable over time (high consistency). Exactly this pattern – a behavior that covaries with subjects but that does not covary with either objects or time – is implicated by the meaning of ADJs. Adjectives load high on subject informativeness, thus setting the subject apart from other subjects (low consensus). However, the meaning of ADJs also implies that the behavior generalizes across objects (low distinctiveness) and across time (high consistency). Thus, in the context of an interper-

354   Klaus Fiedler and Peter Freytag sonal conflict, the use of an ADJ such as “aggressive” or “peaceful” conveys an internal attribution to the subject of the sentence based on these predicates. A subject person called aggressive or peaceful is implicitly said to be the cause or origin of the behavior under consideration, because ADJs imply an attribute distinguishing the subject from most other people, and covarying neither with object nor with time. 3.2.2. IAVs in Jones and Davis’ correspondent inference theory The other major attribution theory is based on Jones and Davis’ (1965) correspondent inference theory (see also Jones and McGillis 1976). According to this model, a correspondent inference from an observed behavior to an analogous trait or disposition within the subject person should be drawn when the following conditions are met: a clear-cut, informative effect pattern that apparently reflects an intended action rather than a random or unintended outcome, together with evidence that the actor has the necessary ability and knowledge to perform the action voluntary, and to understand what he or she is doing. Given these boundary conditions, correspondent inferences (from an observed behavior to an internal person disposition) will be even stronger when the behavior is unexpected or unusual for the social category to which the actor belongs. Translated into attributional terms, this means that attributions to internal dispositions arise when the observed behavior is (a) informative concerning the actor, (b) controllable, and (c) perceived to be intentional. Although the antecedent conditions of internal trait attributions specified by the two models converge in some respects, they diverge in other important ways. Although both models imply that behaviors are explained by internal factors when consensus is low (i.e., when behaviors are unexpected and informative about the actor), they differ in terms distinctiveness and consistency. According to the covariation model, distinctiveness has to be low (i.e., the behavior generalizes across many target objects) and consistency has to be high (i.e., the behavior generalizes across time and occasions) for internal attributions. In contrast, according to correspondent inference theory, an internal attribution is supported by high distinctiveness (i.e., the effect pattern is distinct) and low consistency (i.e., the behavior deviates from category-based expectancies). Apparently, then, the two models reflect two different processes leading to internal attributions. These qualitatively different processes correspond nicely to the semantic implications conveyed by adjectives and interpretive action verbs, respectively. Using ADJs to describe the partner’s behavior implies dispositions that set the partner apart from other people (low consensus) and that

Attribution and linguistic categories   355

generalize across time and object persons (high consistency, low distinctiveness). Thus, the semantic meaning of ADJs implies internal attributions to the sentence subject in accordance with Kelley’s model. Using IAVs to describe the partner’s behavior also implies an attribution to internal causes within the partner. However, the semantic implications of these linguistic tools are different from the implications of ADJs. Although IAVs are also informative about the sentence subject (i.e., relatively low consensus), these behaviors are distinct with respect to time (relatively low consistency, not extended in time) and object (high distinctiveness, specific object person). Together, the semantic features of IAVs imply a high degree of controllability and intentionality, the crucial condition for an internal attribution via Jones and Davis’ correspondent inference process. Note that ADJs do not imply a high degree of control and intentionality; many ADJs do not even allow for an imperative (e.g., be extravert, be brave). Behaviors expressed by ADJs appear to originate in the subject because ADJs imply something specific to the subject but something that generalizes over objects and occasions (time) – an informative covariation pattern à la Kelley. In contrast, behaviors expressed by IAVs appear to originate in the subject because the verbs imply intentionality and internal control, while being sufficiently unusual to be informative about the subject. Against this background, it may be no coincidence that the experimental tasks used in studies testing the covariation model differ markedly from those used in studies testing correspondent inference theory. Correspondent inferences are typically derived from a detailed description of a single behavioral episode, providing participants with numerous cues that allow for inferences regarding a target person’s intention, knowledge, and ability, whereas person attributions based on covariation information are typically derived from a series of brief statements, providing information about multiple subjects and objects and allowing for inferences regarding the presence of global dispositions in a target person. At this point, a comment is in place regarding the remaining two linguistic categories, SVs and DAVs, both of which do not imply subject causation, but for different reasons. The semantic features of SVs regularly imply object causation, opposite to the way in which IAVs imply subject causation. Given a subject–SV–object sentence frame (such as “She admires her partner”; “S abhors O”); the impression arises that the behavior is due to something inherent to the object (i.e., “the partner is admirable”; “O is outrageous”). These semantic inferences are similarly regular as the subject-causation inferences for IAVs; they generalize across the vast majority of verbs in the lexicon. Whereas IAV sentences (“S hurts O”) imply a cause in S and an emotional state in O, SV

356   Klaus Fiedler and Peter Freytag sentences (“S hates O”) imply an emotional state in S and a cause in O. The last verb category, DAVs (such as ‘to call’, ‘to tickle’, ‘to touch’), also implies subject control and intentional behavior, but its meaning is too shallow, context-specific, and uninformative about subject and object person to elicit strong dispositional inferences. 3.2.3. Summary In summary, then, two distinct attribution processes that both lead to behavioral explanations in terms of internal causes within the sentence subject are wired into the semantics of two word classes, ADJs and IAVs. Apart from the different inference schemes underlying these two linguistic tools, they also differ in terms of their pragmatic consequences. Whereas ADJs indicate global and stable causes in the subject, which need not be under intentional control but allow for strong predictions of behaviors that generalize across targets and occasions, IAVs indicate controlled, intentional reasons motivating deliberate, goal-driven behaviors directed at distinct targets. Other word classes that can also be used to describe social behaviors either support external attributions to the sentence object (SVs), or they do not have any strong attributional implications (DAVs). 4. Evidence for the impact of linguistic stimuli on attributions The remainder of this chapter is devoted to a cursory review of existing empirical evidence on language and attribution. The purpose here is to demonstrate that the attributional implications of different word classes that have been derived in the preceding section on purely linguistic grounds can indeed have a profound impact on attributional judgments and their psychological consequences. We shall first summarize the findings from several semantic judgment studies using simple sentence frames (of the form S-predicate-O) as stimuli and straightforward attribution judgments as dependent measures. We then turn to the impact of linguistic stimuli on prominent attribution biases, such as the fundamental attribution bias, the actor-observer bias, and the self-serving or group-serving bias. As we will see, several of these phenomena can be easily linked to either the cooperative or the manipulative function of communication. The final subsection will deal with the subtle but influential role played by language in the constructive cognitive processes underlying attribution. As these processes often evade conscious experience, they shall serve here as an illustration of the unintended function of communication.

Attribution and linguistic categories   357

4.1. Predicate causality A number of studies have assessed the attributional consequences of different verb classes through explicit semantic judgments. The notion of implicit verb causality, which is as old as the seminal papers by Kanouse and Abelson (1967) and Garvey and Caramazza (1974), focuses on the opposite implications of IAVs and SVs. Common to all studies on implicit verb causality (for an overview, see Rudolph and Försterling 1997) is that a sample of verbs from both categories is drawn from the lexicon, and embedded in stimulus sentences involving IAVs (e.g., “Peter rescues Paul”; “S insults O”) or SV predicates (e.g., “Peter likes Paul”; “S abhors O”). The general finding is that over 80 % of the IAV sentences induce internal attributions to S, whereas a similarly high proportion of SV sentences suggests a causal origin in O. Convergent evidence for this highly regular finding has been obtained with a variety of different judgments tasks. Some studies have relied on direct judgment tasks like “Is the sentence due to something about S or something about O?” or “Who caused the behavior expressed in the sentence, S or O?”. Other studies have used a sentence completion task with an ambiguous pronoun, like “Peter abhors Paul because he ….” When respondents’ completion of the sentence implies that the personal pronoun he is interpreted as referring to the object person, Paul (e.g., “because his behavior is so primitive”), the implicit cause is located in O. If he is disambiguated as referring to the subject, Peter (e.g., “because his aspiration level for accepting people is very high”), the implicit locus of causation lies in S. Fiedler and Semin (1988) presented S-Name verb O-Name sentence frames and asked language users to make inferences about plausible antecedent and consequent events: “What might have plausibly occurred before?” and “What might plausibly happen next?” Inferred antecedents of IAV sentences tended to be S behaviors, whereas inferred antecedents of SV sentences tended to be O behaviors. In contrast, inferred consequences of IAV sentences tended to be located in O, whereas consequences of SV sentences were more likely S behaviors. Last but not least, McKoon, Greene, and Ratcliff (1993) have used a priming technique in text comprehension studies, showing that participants who have just read an IAV sentence (such as “Peter insults Paul in the presence of other people …”) will later on be faster to verify a trait of Peter rather than Paul’s. In contrast, reading SV sentences can be shown to facilitate inferences about the object (Paul) rather than about the subject (Peter). Note that, with respect to the aforementioned attribution models, the contrast between IAVs and SVs is reflective of Jones and Davis’ (1965) correspondent inference theory. The semantic meaning of most IAVs entails a controlled, intentional behavior of S, causing an emotional or hedonic consequence in O,

358   Klaus Fiedler and Peter Freytag the target of S’s goal-directed action. In contrast, most SVs imply an emotional state or reaction in S, which is not subject to S’ control or intention but constitutes a reaction to some stimulation originating in O. In a similar vein, other studies have included ADJs and DAVs, along with IAVs and SVs, using judgment tasks like “Who is more likely to be the cause of the behavior expressed in the sentence?”; “Does the sentence express a stable attribute of S?”, or “Is the behavior expressed in the sentence specific to the situation in which it occurred?” (Semin and Fiedler 1992). Factor analyses of these and similar judgment items result in two dimensions, which can be interpreted as an abstractness dimension and a locus of causation dimension, respectively. The factor scores of the different linguistic categories (i.e., their coordinates on these two dimensions) reveal a clear-cut pattern. ADJs and DAVs represent opposite ends of the abstractness dimension (reflecting different consensus and consistency values within covariation model). However, IAVs and SVs differ on the locus of causation dimension, with IAVs implying S causation and SVs implying O causation (in accordance with correspondent inference theory), while taking on intermediate positions on the abstractness dimension. 4.2. Predicate diagnosticity While the studies reviewed so far are based on direct judgments of attribution, a few other studies have tackled relevant boundary conditions, such as domainspecific attribution thresholds. As Gidron, Koehler and Tversky (1993) have pointed out, different trait terms (and also nouns) are characterized by different implicit quantifiers, or “scope” parameters. For instance, being “honest” means behaving honestly all the time, whereas one or two dishonest behaviors are sufficient to infer the attribute “dishonest”. Similarly, the implicit quantifier of being a murderer is having committed one murder, whereas the implicit quantifier of being a pacifist is to be always peaceful. In the domain of morality, negative traits (“dishonest”; “rude”; “arrogant”) tend to have clearly lower implicit quantifiers than positive traits (“honest”; “gentle”; “decent”). One obvious implication of this general rule is that much more evidence is required to support positive trait inferences than to support negative trait inferences or, stated differently, negative behavior is attributionally more diagnostic than positive behavior. Thus, by using negative linguistic terms rather than their positive antonyms, language users can facilitate attributions of dispositional traits. Supplementary evidence comes from Hampson, John, and Goldberg (1986), showing that trait ADJs differ in the breadth of behaviors to which they refer, and by Rothbart and Park (1986), showing that ADJ terms not only

Attribution and linguistic categories   359

vary in terms of the amount of evidence needed for confirmation, but also in terms of ease of disconfirmation. Whereas less behavioral evidence is required to confirm that somebody is dishonest than to confirm the he or she is honest, more evidence is required to disconfirm that somebody is dishonest rather than honest. It is interesting to note that the implicit quantifiers, or diagnosticities, of positive and negative terms depend on the behavioral domain. As already mentioned, in the morality domain negative terms are more diagnostic, due to a lower scope or quantifier, than positive terms. In contrast, in the domain of skills and abilities, positive behaviors or terms are more diagnostic than negative terms. An athlete who once jumped over two meters high must possess the corresponding ability; this single event cannot be due to chance, regardless of how often the athlete failed even on easier trials. Thus, language users’ toolbox offers even more ways of influencing or manipulating attributions. In addition to framing the same behavior in positive or negative terms, they are often free to describe the same behavior by morality-related terms (“fair” vs. “unfair”; “deceive”, “praise”) or by ability-related terms (“clever” vs. “naïve”; “outperform”, “give up”). Apparently, the implicit quantifiers built into the semantics of linguistic terms afford a rich repertoire of linguistic tools for influencing the attribution of traits, intentions, dispositions, and stereotypes – although, or exactly because, they typically evade language users’ attention and conscious control. 4.3. Attribution biases Our major source of empirical evidence for the impact of linguistic stimuli on everyday reasoning comes from research on the attribution biases introduced above. As we shall see, these biases can be derived from the semantic properties of linguistic terms, and they can therefore be induced linguistically. As to the fundamental attribution bias, for instance, there is a general tendency for communication to become more and more abstract and interpretive over time. Fiedler, Semin, and Bolten (1989) employed a serial reproduction method to study the way in which person descriptions (e.g., of a male porter or a female nurse) change when communicated from one person to another in a serial communication chain. Descriptions in this communication game became more and more abstract, moving from the concrete verb levels to ADJs, which imply stable dispositions in the persons described, detached from concrete situational contexts. Similar findings have been found in more recent studies on the serial reproduction of both well-established gender stereotypes (Kashima 2000) and

360   Klaus Fiedler and Peter Freytag novel stereotypes about fictitious groups (Lyons and Kashima 2003). These authors found that the proportion of stereotype-inconsistent information included in a reproduction decreases more readily than the proportion of stereotypeconsistent information. Moreover, a recent study by Karasawa (2005; see also Suga and Karasawa 2006) has shown that the two major tendencies in serial communication studies (i.e., more abstract descriptions on the one hand, and more stereotype-consistent descriptions on the other) coexist in the same data. The reasons for the general tendency toward highly abstract, interpretive language levels may be diverse. Although it may partly reflect memory loss, it also clearly reflects people’s adherence to Grice’s (1975) maxims of quantity and quality, respectively, which oblige speakers to refrain from telling too many subsidiary details, and to pass on what they belief to be true. In a related fashion, an inspection of the semantic differences between linguistic categories can also help resolve seeming contradictions between different attribution biases. The actor-observer bias, for instance, implies more internal attributions to others, whereas the so-called egocentric bias implies more internal attributions to oneself. As clarified in a study by Fiedler, Semin, and Koppetsch (1991), the former bias reflects a tendency to apply more ADJs to others (i.e., more dispositional attributions in accordance with the covariation model), whereas the latter bias reflects a tendency to apply IAVs to oneself rather than to others (i.e., more intentional action in accordance with Jones and Davis’ correspondent inference theory). Interestingly, the insights gained here from a linguistic analysis of self-related as opposed to other-related language nicely converge with a retrospective analysis of the procedures used in prior research. Whereas the experimental tasks employed in studies on the actor-observer bias typically call for trait judgments at the ADJ level, classical studies on the egocentric bias (Ross and Sicoly 1979) involve judgments of behavioral intentions at the level of IAVs. Another prominent example of the role of linguistic stimuli can be found in the domain of self-serving biases. Remember that the actor-observer bias is moderated by the desirability of the behavior under consideration in that actors tend to provide external (internal) attributions when explaining their own negative (positive) behavior. An analogous phenomenon at intergroup level is commonly called group-serving bias. Self-serving biases and group-serving biases have been the focus of countless studies on the linguistic-intergroup bias (Maass et al. 1989; for a review, see Maass 1999). The linguistic-intergroup bias (LIB) consists in the tendency to describe negative behaviors of outgroup members in more abstract terms than negative behaviors of ingroup members – and vice versa, to describe positive behaviors of ingroup members more abstractly than positive behaviors of outgroup members. It is customary

Attribution and linguistic categories   361

in these studies to measure abstractness of language use on a scale that results from counting 1, 2, 3, or 4 scores for DAVs, IAVs, SVs, and ADJ predicates, respectively. Given this scoring rule, which gives the highest score to ADJs, the lowest score to DAVs and the second lowest to IAVs, it is clear that the LIB reflects attributions of the Kelley-type. Language use facilitates attributions of global negative dispositions to the outgroup and global positive attributions to the ingroup. Originally, the LIB was interpreted as a self-serving language style that strengthens positive ingroup stereotypes and negative outgroup stereotypes. More recent research, though, found that the LIB may be another by-product of cooperative communication, reflecting communicators’ expectancies regarding the behavior of ingroup and outgroup members (Wigboldus, Semin, and Spears 2000). That is, the reason why language users vary the level of abstractness when talking about the ingroup’s as opposed to the outgroup’s assets and deficits may not be selfish or egocentric. The bias may rather reflect a prosocial motive, namely, to express abstractly what can be expected or what is stereotype-consistent, and to express in concrete, contextualized ways only what could not be expected from members of a given target group. To the degree that positive aspects of the ingroup and negative aspects of the outgroup are expectancy-consistent, the LIB may be nothing but a reflection of speakers’ attempts to communicate efficiently (i.e., to tune messages length to receivers’ prior knowledge) as required by Grice’s (1975) maxim of quantity in cooperative communication. Ironically, then, the adherence to genuinely cooperative language rules may contribute to the maintenance of stereotypes. Interestingly, the idea that linguistic abstraction may be used to convey information about the expectedness of a target person’s behavior implies that communicators can actively vary linguistic abstraction so as to meet certain communication goals. For instance, a speaker might deliberately describe a certain behavior in abstract terms in order to create the impression that the behavior is quite characteristic of a target person. Put differently, variation in linguistic abstraction may be considered a communicative skill that may be used in the service of the manipulative function of language, too. Following these considerations, Douglas and Sutton (2003) had participants describe positive and negative behavior of a target person in a way that would affect the behavior’s perceived favorability. Descriptions turned out to be more abstract for positive (negative) behavior when instructions called for favorable (unfavorable) descriptions. Moreover, these biased patterns of language use obtained independently of participants’ privately held beliefs about the target person, thereby lending further support to the notion of the manipulative potential of linguistic stimuli.

362   Klaus Fiedler and Peter Freytag 4.4. Priming attributions As noted at the outset, language may be used not only to cooperate and to get along with others, or to manipulate others instrumentally. It may also have incidental, unintended, emergent effects that do not reflect any communication goals or planned speech acts, but simply uncontrolled side effects of using words. Such incidental, non-instrumental language influences – which typically evade the communication partners’ awareness – may be understood as verbal priming effects on social attributions. The notion of priming means that the very presentation of a stimulus cue, or prime, is sufficient to activate memory contents and to solicit reflex-like particular behaviors or reactions, as distinguished from planned action. For example, exposure to an aggressionrelated word or phrase might prime an aggressive behavior, even when the priming remains subliminal or unconscious. Likewise, the interpretation of the same behavior depends on whether either the trait clumsy or the trait aggressive has been primed immediately before. Priming effects on person judgments have been shown to be stronger, and more congruent with the prime, when abstract words (trait words) rather than specific words (verbs) were used as primes (Stapel, Koomen, and van der Pligt 1996). Abstract concepts (“professor”) will more likely support inferences of high intelligence in other persons than specific concepts (“Albert Einstein”), which may even produce contrast effects, because specific primes are less likely to be applicable to judgment targets than abstract, more inclusive primes (cf. Schwarz and Bless 1992). This asymmetry also holds for the word classes of the linguistic category model. Fiedler et al. (1989) asked participants to give two successive ratings of the same well-known target person on two sets of 12 attributes. Attributes were all of the same constant valence (only positive vs. only negative), but abstractness either increased (e.g., an SV preceding an ADJ) or decreased from the first to the second rating (e.g., an ADJ preceding an IAV). Across all transitions of word classes, priming effects from abstract to concrete terms were strongest. That is, the second ratings were more extreme when the first ratings involved more abstract attributes. Verbal priming effects may not only influence the outcomes of social judgments but may also affect the language users’ cognitive style and the mental operations. As nicely illustrated in a recent investigation by Stapel and Semin (2007), the very priming of abstract or concrete words (i.e., adjectives vs. action verbs) can trigger a global versus more local perceptual focus, respectively. Apparently, then, the same power of linguistic tools that can be utilized for intentional, goal-driven communication may be also apparent, with a similar

Attribution and linguistic categories   363

level of intensity, in the unwanted, incidental side effects and priming effects solicited by linguistic stimuli. 5. Concluding remarks The psychology of attribution is intimately linked to the psychology of language. When we share information with others, the very manner in which we describe our environment does not only reflect the mental representations we hold, but also the communicative goals we pursue, and the state of mind we are in. Some of these links are well understood, such as the close correspondence between internal attributions expressed by ADJs and covariation information, or between internal attributions expressed by IAVs and information about controllability and intention. Empirical evidence regarding these links does no longer draw exclusively on data obtained in the laboratory. For instance, the closing speeches of prosecutors have been shown to emphasize the defendant’s control over his behavior by an increase in the utilization of IAVs, whereas the closing speeches of attorneys seek to diminish the defendant’s accountability by an increase in the utilization of ADJs and SVs, thereby stressing the involuntary, unintentional nature of the defendant’s behavior (Schmid and Fiedler 1998). Similar findings in the domains of health and romantic relationships have already been discussed in the introductory section. The common denominator of most findings reviewed so far says that the same behaviors or events can be represented at different levels of abstractness (Semin and Fiedler 1988; Trope and Liberman 2003; Vallacher and Wegner 1987), and the way in which this latitude or ambiguity of possible representations is resolved depends to a large degree on the linguistic terms used for communication. However, the role of language is not confined to selecting an appropriate representation of the original events. Language may also promote distorted representations through the deliberate use of predicates that mislead the recipient of a message, and it may even create illusory representations through constructive memory processes. We believe that the latter class of effects, which often evade communication partners’ conscious control, still deserves our attention. A final example shall illustrate this particularly marked influence of language on social judgment. In a study by Fiedler et al. (1996), participants were first presented with a videotaped group discussion. Afterwards they were asked to answer a list of questions about various behaviors of protagonists, of which most behaviors had actually not occurred. Depending on the experimental condition, these questions included positive or negative action verbs (implying internal causation) or positive or negative state verbs (implying external causation). Re-

364   Klaus Fiedler and Peter Freytag gardless of whether the questions referred to observed or unobserved behavior, the predicates used for questioning affected judgments of the protagonists. Specifically, merely considering behaviors expressed in actions words, protagonists were attributed more traits that matched the meaning of the verbs. Conversely, merely considering state verbs resulted in more external attributions. These constructive influences on social judgments have apparent implications for the instrumental and manipulative use of language. Just like the experimenter in the above study, rhetorical questions may be used to mislead eyewitnesses in the courtroom, or voters scanning an interview of political candidates in the newspaper, or clients retrieving memories of alleged abuse during childhood in psychotherapy. Presumably, indirect measures based on response latencies, signal detection data, or parameter estimates in multinomial modeling, will play a prominent role during the next stage of empirical investigation on the interplay of language and social cognition. The degree to which these measures converge with direct judgments of perceived causal impact may help determine the range of phenomena that can be accounted for by contemporary theories of causal attribution.

References Douglas, K.M., and R.M. Sutton 2003 Effects of communication goals and expectancies on language abstraction. Journal of Personality and Social Psychology 84: 682–696. Fiedler, K., T. Armbruster, S. Nickel, E. Walther, and J. Asbeck 1996 Constructive biases in social judgment: Experiments on the self-verification of question contents. Journal of Personality and Social Psychology 71: 861–873. Fiedler, K.T., and G.R. Semin On the causal information conveyed by different interpersonal verbs: 1988 The role of implicit sentence context. Social Cognition 6: 21–39. Fiedler, K., G.R. Semin, and S. Bolten Language use and reification of social information: Top-down and 1989 bottom-up processing in person cognition. European Journal of Social Psychology 19: 271–295. Fiedler, K., G.R. Semin, and C. Koppetsch 1991 Language use and attributional biases in close personal relationships. Personality and Social Psychology Bulletin 17: 147–156. Fincham, F.D. 1985 Attribution processes in distressed and nondistressed couples: 2. Responsibility for marital problems. Journal of Abnormal Psychology 94: 183–190.

Attribution and linguistic categories   365 Garvey, C., and A. Caramazza 1974 Implicit causality in verbs. Linguistic Inquiry 5: 459–464. Gidron, D., D.J. Koehler, and A. Tversky 1993 Implicit quantification of personality traits. Personality and Social Psychology Bulletin 19: 594–604. Grice, H.P. 1975 Logic and conversation. Syntax and Semantics 3: Speech Acts, P. Cole and J.L. Morgan (eds.), 41–58. New York: Academic Press. Hampson, S.E., O.P. John, and L.R. Goldberg Category breadth and hierarchical structure in personality: Studies of 1986 asymmetries in judgments of trait implications. Journal of Personality and Social Psychology 51: 37–54. Jones, E.E., and K.E. Davis 1965 From acts to dispositions: The attribution process in social perception. In Advances of Experimental Social Psychology, Vol. 2, L. Berkowitz (ed.), 220–266. New York: Academic Press. Jones, E.E., D.E. Kanouse, H.H. Kelley, R.E. Nisbett, S. Valins, and B. Weiner (eds.) 1992 Attribution: Perceiving the Causes of Behavior. Morristown, NJ: General Learning Press. Jones, E.E., and D. McGillis Correspondent inferences and the attribution cube: A comparative re1976 appraisal. In New Directions in Attribution Research, Vol. 1, John H. Harvey, Robert F. Kidd, William John Ickes (eds.), 389–429. Hillsdale, NJ: Lawrence Erlbaum. Jones, E.E., and R.E. Nisbett The actor and the observer: Divergent perceptions of the causes of be1972 havior. In Attribution: Perceiving the Causes of Behaviour, E.E. Jones, D.E. Kanouse, H.H. Kelley, R.E. Nisbett, S. Valins, and B. Weiner (eds.), 79–94. Morristown, NJ: General Learning Press. Kanouse, D.E., and R.P. Abelson 1967 Language variables affecting the persuasiveness of simple communications. Journal of Personality and Social Psychology 7: 158–163. Karasawa, M. 2005 The linguistic intergroup bias in the context of message transmission. Paper presented at the 14th General Meeting of the European Association of Experimental Social Psychology, held July 19–23, 2005, at the University of Würzburg, Germany. Kashima, Y. 2000 Maintaining cultural stereotypes in the serial reproduction of narratives. Personality and Social Psychology Bulletin 25: 594–604. Kelley, H.H. 1967 Attribution theory in social psychology. Nebraska Symposium on Motivation 15: 192–238. 1973 The processes of causal attribution. American Psychologist 28: 107– 128.

366   Klaus Fiedler and Peter Freytag Kunda, Z. 1990 The case for motivated reasoning. Psychological Bulletin 108: 480–498. Lyons, A., and Y. Kashima 2003 How are stereotypes maintained through communication? The influence of stereotype sharedness? Journal of Personality and Social Psychology 85: 989–1005. Maass, A. 1999 Linguistic intergroup bias: Stereotype perpetuation through language. In Advances in Experimental Social Psychology, Vol. 31, M.P. Zanna (ed.), 79–121. San Diego, CA: Academic Press. Maass, A., D. Salvi, , L. Arcuri, and G.R. Semin Language use in intergroup contexts: The linguistic intergroup bias. 1989 Journal of Personality and Social Psychology 57: 981–993. McKoon, G., S.B. Greene, and R. Ratcliff Discourse models, pronoun resolution, and the implicit causality of 1993 verbs. Journal of Experimental Psychology: Learning, Memory and Cognition 19: 593–607. Nisbett, R.E., and L. Ross Human Inference: Strategies and Shortcomings of Social Judgment. 1980 Englewood Cliffs, NJ: Prentice-Hall. Peterson, C., and M.E.P. Seligman 1984 Causal explanations as a risk factor for depression: Theory and evidence. Psychological Review 91: 347–374. Reisenzein, R. 1986 A structural equation analysis of Weiner’s attribution-affect model of helping behavior. Journal of Personality and Social Psychology 50: 1123–1133. Ross, M., and F. Sicoly 1979 Egocentric biases in availability and attribution. Journal of Personality and Social Psychology 37: 322–336. Rothbart, M., and B. Park On the confirmability and disconfirmability of trait concepts. Journal of 1986 Personality and Social Psychology 50: 131–142. Rudolph, U., and F. Försterling The psychological causality implicit in verbs: A review. Psychological 1997 Bulletin 121: 192–218. Schmid, J., and K. Fiedler The backbone of closing speeches: The impact of prosecution versus 1998 defense language on judicial attributions. Journal of Applied Social Psychology 28: 1140–1172. Schwarz, N., and H. Bless Constructing reality and its alternatives: An inclusion/exclusion model 1992 of assimilation and contrast effects in social judgment. In The Construction of Social Judgment, L.L. Martin and A. Tesser (eds.), 217–245. Hillsdale, NJ: Lawrence Erlbaum Associates.

Attribution and linguistic categories   367 Semin, G.R., and K. Fiedler The cognitive functions of linguistic categories in describing persons: 1988 Social cognition and language. Journal of Personality and Social Psychology 54: 558–568. 1992 The inferential properties of interpersonal verbs. In Language, Interaction and Social Cognition, G.R. Semin and K. Fiedler (eds), 58–78. Thousand Oaks: Sage Publications. Stapel, D.A., W. Koomen, and J. van der Pligt The referents of trait inferences: The impact of trait concepts versus 1996 actor-trait links on subsequent judgments. Journal of Personality and Social Psychology 70: 437–450. Stapel, D.A., and G.R. Semin The magic spell of language: Linguistic categories and their perceptual 2007 consequences. Journal of Personality and Social Psychology 93: 23–33. Suga, S., and M. Karasawa 2006 Effects of social stereotypes on language use in the description of person dispositions. Japanese Journal of Social Psychology 22: 180–188. Trope, Y., and N. Liberman 2003 Temporal construal. Psychological Review 110: 403–421. Vallacher, R.R., and D.M. Wegner 1987 What do people think they’re doing? Action identification and human behavior. Psychological Review 94: 3–15. Weiner, B. 1985 An attributional theory of achievement motivation and emotion. Psychological Review 92: 548–573. Wigboldus, D.H.J., G.R. Semin, and R. Spears 2000 How do we communicate stereotypes? Linguistic bases and inferential consequences. Journal of Personality and Social Psychology 78: 5–18.

368   Klaus Fiedler and Peter Freytag Table 1. Defining features of the four word classes of the LCM.

Examples DAV call Descriptive Action Verbs kick comb touch IAV deceive Interpretive Action Verbs hurt help save SV admire State Verbs hate like envy ADJ honest Adjectives aggressive nice clever

Defining Features Singular behavioral episode Concrete situation reference Invariant physical attribute Context-dependent meaning Singular behavioral episode Concrete situation reference Interpretation und valence Context-independent meaning Affective or cognitive state No situation reference Interpretation und valence Abstract of specific action Enduring disposition No situation reference Highly interpretive Abstract of action and person

Attribution and linguistic categories   369 DAV

IAV

SV

ADJ

Controllability

Debatability

Verifiability

Context Dependence

Temporal Stability

Subject Informativeness

0

Figure 1.

10

20

30

40

Major differences between the four word classes identified by the LCM regarding the key dimensions underlying attribution processes.

Chapter 14 Tuned to hidden messages: Exploring recurrent word combinations in English Rainer Schulze

1. Introduction Philosophers and linguists alike have, for some time, been attracted to the idea of tackling the implicitness issue in language, an issue traditionally being approached either from a pragmatic and/or semantic angle. Hidden messages (or ‘pictorially’ speaking: the fine print which often escapes the attention of the language user), the topic of this paper, seem to lend themselves quite readily to a pragmatic analysis based on the distinction between what is being said and what is being intended, as it has become popularised in the notions of conversational implicature, invited inference, invocature or allusion (Lennon 2004; Robinson 2006). Even if both the pragmatic and the semantic angle vis-à-vis the ‘unsaid’ or implied have provided valuable insights, neither of them has been capable of explaining the true nature and specific function of language use, and of recurrent word combinations in particular, in relation to the epistemological possibilities provided by the lexicon and linguistic constructions in general. Yet if we take the principle of the word-concept relationship seriously, which suggests “…that the meaning of a word consists of word-specific properties plus the properties of the associated concept” (Cruse 2000: 129), the specific function of language in use, whether it be from a linguistic, social, and/or more generally cognitive point of view, is obviously something worth investigating. One reason why these aspects have attracted limited attention among scholars from different scholarly strands so far is perhaps that most of them have viewed linguistic issues from a purely linguistic point of view (cf. semantic or lexical field theory, denotational semantics, etc.), from a more societal point of view (cf. sociolinguistics) or from a stance with an explicit perceptual orientation (cf. cognitive linguistics). The renewed interest in the ‘performative’

372   Rainer Schulze aspect of language is partly linked with the work of Bless, Fiedler and Strack (2004), Moskovici (1981, 1984), or Tomasello (2002), partly linked with the writings of Halliday (1994), Hoey (2005, 2009), Sinclair (1991) or Stubbs (1994, 2001) and partly linked with publications by Biber, Conrad and Reppen (1998), Kennedy (1998), McEnery and Wilson (2001), McEnery, Xiao and Tono (2006) or Tognini-Bonelli (2001). What all these different, yet related frameworks have in common is the fact that they “…cover more than perception, categorization, and figurative language” (Pishwa 2006: 308). Although the present study shares the concern for a social-cognitive interpretation of language in performance and a conceptual basis for recurrent word combinations with the above-mentioned authors, its goals are much more modest. Far from proposing a comprehensive view on recurrent word combinations, this paper takes up the challenge of predominantly modular approaches to language in which a clear conceptual distinction between the language user, the lexicon, the grammar, etc. is advocated, merely resulting in a coarse-grained analysis of language in use. The paper thus seeks to capture the specific social and cognitive function that authentic language, and more specifically, recurrent word combinations seem to fulfil. The starting-point includes the notions of knowledge and memory, intertextuality, experience as textually mediated knowledge structure and communicative collaboration (Antos 1997; Assmann 1997; Berger and Luckmann 1966; Bless, Fiedler and Strack 2004; Fairclough 1999; Goffman 1974; Moscovici 1981, 1984; Schmidt 1996; Varela 1988); the background is provided by the cognitive view of language as it has been advocated by Lakoff and Langacker (Ungerer and Schmid 2006) and by the social, psychological and linguistic assumptions centring around the use of lexical items and recurrent word combinations (Goldberg 1995, 2006; Hoey 2005, 2009; Hunston and Francis 2000; Sinclair 1991; Stubbs 1994, 2001), as popularised by the lexical priming approach, the pattern grammar approach and the construction grammar approach. 2. Setting the scene: Knowledge, memory, and texts as the media of social cognition Research on the most essential components of social cognition has long centred on questions related to the encoding, storage and retrieval of social information, the structure and (linguistic) representation of social knowledge and the concomitant processes employed by the language user when forming judgements and making decisions (Bless, Fiedler and Strack 2004). Given its

Exploring recurrent word combinations in English   373

emphasis on the structure and (linguistic) representation of social knowledge, it comes as no surprise that social cognition has strong links with social psychology, cognitive linguistics and the social-cognitive foundations for language acquisition and development, partially following Emil Durkheim’s notion of représentations collectives. Crucial to the notion of social cognition is the notion of knowledge which, in a more simplified way, is assumed to be a dynamic concept, including aspects of accumulation, anchoring, erasure, objectification, organisation, reconfiguration, reflection, replacement, revision, testing, usage and communication (for the notion of social representation, see Moscovici (1981); in this survey article, Moscovici spells out the general characteristics of different sets of concepts, statements and explanations originating in daily life in the course of inter-individual communications). According to this view, ‘knowledge’ can be seen as a permanently operating, largely self-organising process that is preferably fed by (personal and communal) experience, previously gained knowledge (thus enabling recursivity) and/or probably some portion of innate knowledge. If we accept that knowledge is something that tends to settle in and at the bottom of languages, then the proper and adequate description and analysis of languages should lead us to a better understanding of the impact of information, skills, and understanding that we have gained through learning or experience through the linguistic representation of our concepts, the fundamental units of thinking. It is probably not a coincidence that ‘knowledge’ can be viewed from two different, but mutually dependent perspectives, i.e. ‘knowledge’ as an intraindividual, cognitive-psychological phenomenon and ‘knowledge’ as a supraindividual phenomenon marked by societal, cultural and historical aspects (Berger and Luckmann 1966; Schmidt 1996; Varela 1988). The strong association between linguistic, and semantic knowledge in particular, and nonlinguistic or encyclopedic knowledge (including factual and expert knowledge) has already been shown by adherents of prototype or stereotype theory (Taylor 1995: 59 ff.). Here, it is experiential knowledge in particular that speakers, in the context of everyday interpretations, subconsciously combine with linguistic forms. Or to put it differently: intuitive, experiential knowledge is not a separate, language-independent ‘module’ that is derived via complex inferential processes; language does not affect cognition – rather, it is one form that cognition can take. The tenet that we acquire a language in much the same way as we acquire everything else can be translated into a claim concerning the continuous interaction between the cognitive system and general cognitive abilities, as has been captured in Figure 1 (after Dirven and Verspoor 2004: ix):

374   Rainer Schulze cognitive system: -perception -emotions -empathy/sympathy -categorization -abstraction processes -reasoning

cognitive abilities

(interact) (influence) -language

Figure 1. The interaction of cognitive abilities

The centrality of knowledge and memory in and through language has recently come to be acknowledged in a number of studies (e.g. Pishwa 2006 and numerous papers therein). Jan Assmann (1997), a widely acclaimed expert in cultural theory, claims that the human memory is social in nature: the human being acquires memory in the course of his/her socialisation, and it is within this process of socialisation that the gradual adoption of values and norms associated with social knowledge can be observed. That is, acts of remembering and forgetting are mediated through social experiences. For Assmann, ‘collective memory’ functions as an umbrella term, covering both ‘communicative memory’ (i.e. a type of memory that is experienced and mediated by contemporary witnesses) and ‘cultural memory’ (i.e. a type of memory which is typically represented in the written medium and which guarantees experiential continuity) as extreme poles of a continuum with transitional and overlapping areas. Seen from that point of view, ‘cultural knowledge’, as one of the possible outcomes of cultural memory, turns out to be the result of a very special encounter with texts in all their possible shapes, making intertextuality or conversational invocature (Robinson 2006: 209–225) one of its driving forces. Taking the “…textually mediated nature of contemporary social life…” (Fairclough 1999: 71) for granted, it becomes obvious that the mere focus of a linguist on isolated words in authentic language should turn out to be a futile enterprise, or, as Malinowski (1935: 11) already stated: “…isolated words are in fact only linguistic figments, the products of an advanced linguistic analysis.”

Exploring recurrent word combinations in English   375

This observation can be seen as a clear plea for a more meticulous analysis of recurrent word combinations in texts, since these can be viewed as the ‘true’ carriers of collectively shared knowledge. We may conclude then that in literate societies, texts (in the broadest sense of the word) function as the media for the constitution and transmission of knowledge and as an essential component of everyday communication processes. Collective consciousness or conscience and cultural memory become manifest in writing via a number of processes which strengthen the idea of intertextuality or invocature (Robinson 2006): as soon as knowledge becomes manifest through ‘textualisation’, it becomes available as a collectively shared stock of wisdom; it changes from an intra-psychological phenomenon towards a more supra-individual one and thus becomes subject to acts of communal ratification, i.e. putative cultural memory is constantly being checked for its social fit. In the same vein, both individual-cognitive and supra-individual, social aspects of knowledge are subsumed under the category of social cognition (Antos 1997: 46; Feilke 1996: 55–57). Knowledge of that type is accepted by the speech community only if it is adaptable to a wide range of similar experiences in that community, provided that it is consonant with the community’s predominant system of thought, patterns of interpretation and patterns of evaluation. The notion of adaptability or social fit, as it is understood in this context, tries to capture the interaction of previously produced acts of communication and their follow-ups, and from a more socio-cultural and constructivist point of view, this notion also helps to account for the structural pairing of cognition and communication (Schmidt 1996). Within this general perspective, communication and the ratification of knowledge is a process which is largely determined by the mutual production and reception of texts and their intertextual linkage; and this view is tempting since large amounts of texts, stored in digitally re-mastered electronic corpora, have been made and are collectively available as, so to speak, ‘frozen’ material equivalent to commonly shared cultural knowledge. It is vitally important to note that it is only the inspection of large bodies of naturally occurring text (i.e. corpora) that promises to provide insights into the multi-faceted nature of general, accumulated and cultural knowledge. Consequently, it would be embarrassingly naïve to assume that few or single texts as self-contained database packages are semantically sufficient to uncover the speaker’s or writer’s messages and intentions; on the contrary, the single text is largely underspecified, i.e. its meaning emerges incrementally only following the recipient’s interpretive skills in activating and constituting his or her encyclopedic knowledge which goes beyond the verbalised material in the source text. Given, however, that vast textual databases allow some systematic and structured analysis, it becomes evident that the semantic and pragmatic ‘underspecificity’ of a single,

376   Rainer Schulze isolated text can be ‘remedied’ through the exploration of large text corpora. Thus, knowledge, ‘unsaid’ and only tacitly presupposed in single texts, can be made ‘said’, explicit and therefore recoverable in larger corpora; in larger corpora, the multiplicity of texts makes tacit knowledge and assumptions verbal, explicit and real (the closest approximation to our conception of the terms ‘explicit vs. ‘implicit’ or ‘said vs. ‘unsaid’ that we know of is that of Moscovici (1984: 4) who emphasises the ‘visible’ vs. the ‘invisible’ distinction as a result of some pre-established fragmentation of reality). Having accepted that single texts contain both explicit and many more implicit traces of communal knowledge, it seems reasonable to look for methodological tools that are able to focus on aspects of knowledge, being translated into linguistic forms by a particular speech community and taken from large corpora by the analyst. This view necessarily calls for an integrated approach towards knowledge and language, in this contribution simultaneously fed by cognitivists and sociologists in an attempt to get a grasp of the cognitive and societal construction of reality. In general, cognitivists start out from the assumption that both utterance and sentence meaning become manifest through the contextualisation of concepts, i.e. by being embedded in particular syntactic environments and thus by being related to verbal structures. The constitution of meaning and related interpretive processes are thus made available via the systematic study and analysis of the linguistic environments of particular lexical items or linguistic expressions. This viewpoint should be accompanied by some cautioning remarks: the elicitation of reliable and valid data is possible only if ad-hoc combinations or occasional formations are simultaneously excluded from the analysis. In order to explore and exploit syntactic environments as empirical bases for sound semantic and pragmatic analyses, it is utmost necessary to capture and record non-arbitrary, essential linguistic environments of lexical items and linguistic expressions only. Talking about linguistic environments means that we are now concerned with socially ‘loaded’ linguistic patterns that can be seen as the outcomes of processes of communication and cognition. More general in scope and closer to our linguistic concerns is the suggestion proposed by Erving Goffman in his still highly valued Frame Analysis (1974). Although some of the concepts in the book are ill-defined, Goffman’s work is still remarkable in that it examines how the situations of interaction (situations of physical co-presence) determine the social character and shape the social self that is presented to others. I.e., his assumption that knowledge frames can be seen as viable frameworks (!) for processes of concept-formation and interpretation functions as a useful theoretical and methodological basis for more large-scale explorations into the specific nature of language-in-use.

Exploring recurrent word combinations in English   377

His suggestion also enables us to view cognition as a social phenomenon, since frames in the sense of patterned knowledge do not really pertain to individual knowledge, but to knowledge in the sense of commonly shared and collaboratively negotiated knowledge. As it emerges from this presentation, it seems to be quite reasonable to describe and empirically confirm interpretive processes on the basis of large corpora via the notion of frame. This approach offers a number of advantages. First, this approach enables us to recover traces of past and ‘textually mediated’ communicative processes that reside in vast amounts of texts. In this way, social reference frames can be reconstructed via a number of contextualisation patterns against which concepts and expressions from different periods of time can be formed and interpreted. Second, the introduction of the frame-model can be seen as a useful methodological tool: ideas in a speech community about socially relevant processes, states or values, simultaneously linked to abstract lexical domains, can be made available to linguistic analyses. The analysis of vast amounts of data does not only demonstrate how conceptual changes are brought about, it also shows how linguistic, discursive and textual manifestations of conceptual and communal knowledge contribute to these alterations. Given the grounding of concepts in linguistic, communicative and textually-mediated communication, it is not surprising that concepts as fundamental units of knowledge cannot any longer be viewed as context-free meaning-carriers (‘kontextfreie Sinnträger’ (Knobloch 1992: 11)). 3. Materials and methods The present study offers a usage-based account of selected multi-word units, as exemplified in (1) (2) (3) (4)

With hindsight he may have acted differently. (BNC) Yet what looks obvious in hindsight was not at all obvious before. (BNC) The discussion presented has the benefit of hindsight. (BNC) With the wisdom of hindsight and the benefit of several years’ experience, (BNC)

The analysis, which builds on an investigation of the British National Corpus (BNC), is part of a larger project that aims at an exhaustive description of the pattern or construction as a symbolic unit, carried out within the frameworks of cognitive linguistics, construction grammar, pattern grammar and lexical

378   Rainer Schulze priming. The reason for choosing these frameworks as explanatory models has been the fact that all these approaches promise to live up to their basic manifesto of representing usage-based linguistics, thus allowing wide-ranging applications of empirical methods (e.g. corpus investigations, statistical surveys of language usage, etc.) (Geeraerts 2006). All in all, the present investigation offers an attempt to bridge the gap between theory and implementation, presenting an analysis of structural patterns (i.e. with n, in n or n of n in Hunston’s and Francis’ terminology (2000: 45–48)) which is based on the careful analysis of language data obtained from actual usage as represented in the BNC (via the VIEW interface, http://corpus.byu. edu/bnc/). There is no denying the fact that the search for and isolation of usage patterns and the concomitant contextual factors give rise to a major methodological challenge: for one thing, the search for typical patterns in a language cannot be relegated to individual intuition or anecdotal evidence, since native speakers tend to ‘unravel’ unusual patterns instead of the much more frequent ones (Biber, Conrad and Reppen 1998: 3); for another, generalisations based on the language production of very few informants are far from being trustworthy and reliable, thus a vast number of samples produced by very many different informants has to be made available in order to avoid any time-consuming and laborious manual work on the part of the analyst. Looking at both sides of the coin, Biber and his colleagues summarise the ultimate aims of corpus linguistics as follows: Studies of language can be divided into two main areas: studies of structure and studies of use. […] There are two central research goals in such analyses of use: (1) the assessing the extent to which a pattern is found, and (2) analyzing the contextual factors that influence variability. […] Association patterns represent quantitative relations, measuring the extent to which features and variants are associated with contextual factors. However, functional (qualitative) interpretation is also an essential step in any corpus-based analysis, […] (Biber, Conrad and Reppen 1998: 1–4)

4. A short portrait of semi-preconstructed phrases As it emerges from the previous discussion, the analysis of recurrent word combinations or multi-word units as meaningful syntactic arrangements should take precedence over the analysis of ad-hoc combinations or occasional formations: on the assumption that the meanings of utterances become manifest through the contextualisation of concepts (i.e. embedded in syntactic environments and thus related to linguistic structures), the systematic analysis of

Exploring recurrent word combinations in English   379

linguistic environments should make available processes of meaning constitution and interpretation. Granted that there are syntactic environments as an empirical fund or supply for semantic analyses, it is vitally important that data are captured from essential linguistic environments rather than more arbitrary ones. It is a characteristic feature of multi-word units or recurrent word combinations that they “…appear frequently in each other’s company” (Hoey 2005: 2), and that they display both “pervasive and subversive” (2005: 3) traits, properties that make these ‘petrified’ structures the ideal candidates for our study. Most of these aspects have been investigated by the late John Sinclair, who in the early 1990s set out to explore the relevance of the notion of context in the identification of the meaning of an utterance. Sinclair’s view of context is predominantly linguistic (as opposed to Searle’s notion of context as sets of background assumptions, Malinowski’s ‘context of culture’ or Langacker’s notion of domain as some coherent knowledge structure or mental phenomenon). He elaborates two principles which account for how language actually works and which explain the way in which meaning (unexpectedly and unpredictably?) emerges out of larger chunks of text: the open choice principle and the idiom principle (1991: 109–115). It is the idiom principle in particular that explains the phraseological tendency of a language in general: “The principle of idiom is that a language user has available to him or her a large number of semipreconstructed phrases that constitute single choices, even though they might appear to be analysable into segments” (Sinclair 1991: 110). Sinclair claims that the occurrence of such preconstructed chunks may be due to the recurrence of similar situations in human affairs or may illustrate the language user’s natural tendency to communicate as efficiently as possible. ‘Preconstructed chunk’ in this context is used as an umbrella term for notions such as ‘collocation’, ‘fixed expression’, ‘phrasal pattern’ or ‘idiom’ and can be related to more recent terms such as ‘pattern’ (Hunston and Francis 2000) or ‘construction’ (Goldberg 1995, 2006). Neglecting finer nuances between the pattern grammar and the construction grammar approach, we may view ‘patterns’ as exhibiting meanings that are distinct from those of the lexical items used in these predominantly grammatical patterns; similarly, we apply the term ‘construction’ in its more general sense to preconstructed chunks as a symbolic pairing of form and meaning at any level of abstraction, thus allowing for a number of distinct combinations of meaningful elements, including complex morphological or syntactic patterns in all shapes and sizes. To round off this picture, these analytical tools have to be supplemented by insights from another corpus-driven approach to grammar, i.e. the lexical priming approach. This approach, both linguistic, pragmatic and mental

380   Rainer Schulze from its very inception, is firmly associated with Michael Hoey’s work resulting in a new theory of the lexicon. Having established that existing theories of language cannot provide us with convincing explanations of fluency, with systematic and comprehensive models of variation, with thorough accounts of polysemy or preconstructed chunks in language, he advocates a number of lexical priming claims “[…] whenever we encounter a word, syllable or combination of words, we note subconsciously the words it occurs with (its collocations), the meanings with which it is associated (its semantic associations), the grammatical patterns it is associated with (its colligations), and the interactive function it contributes to serving (its pragmatic associations) (Hoey 2009: 34). These claims have spawned a number of publications, including those of Bublitz (1998), Louw (1993), Römer (2005), Schulze (2006) or Stubbs (2001). Another set of assumptions describes the tendency for certain words to be related to a particular genre and/or style and/or social situation they are used in, a claim that cannot be validated through simple concordance lines, but a claim that can be corroborated by annotated corpora only. Finally, lexical priming claims can also be related to the textual dimension of language. More specifically and in addition to the co-textual and contextual features mentioned so far, Hoey posits that “[…] whenever we encounter a word (or syllable or combination of words), we also note subconsciously the positions in a text that it occurs in (its textual colligations), the cohesion it favours or avoids (its textual collocation) and the textual relations it contributes to forming (its semantic associations)” (Hoey 2009: 35). Having worked out these finer distinctions, we will, in the following, attend to issues of the actual uses of preconstructed phrases as represented in the BNC. Preconstructed phrases will be seen as templates which speakers and writers use for the production and comprehension of conventionalised linguistic expressions. As we aim at the usage-based description and analysis of hindsight-occurrences, frequencies of usage of the various hindsight types and tokens are of primary concern to us; as already shown in the discussion on the social and cognitive nature of language, the frequency of usage of a particular experiential unit or linguistic form, as evidenced by large databases, is assumed to both reflect and affect the representation of units in the mind of speakers (Kemmer and Barlow 2000). More specifically, the frequency of units or forms as recorded in language corpora is taken to provide insights into the degree of entrenchment (or automatization) of a particular pattern in speakers and therefore permits judgements about the status and function of such patterns or constructions in a language. In a nutshell: since preconstructed phrases, patterns or constructions do not represent haphazard manifestations of conceptual information, but frequently

Exploring recurrent word combinations in English   381

occurring linguistic structures, they help to gain easy access to the analysis of processes of meaning construction and interpretation; thus, recurrent word combinations can be seen as reliable indicators of meaning ascription in the narrow sense and of concept-bound knowledge and relations to superordinate knowledge schemata in the broadest sense of the word (Römer and Schulze 2008; Römer and Schulze 2009). 5. A usage-based description and analysis of hindsight-occurrences The corpus material was studied with the aid of the VIEW interface (http:// corpus.byu.edu/bnc/) whose primary aim is to run concordance searches, to produce frequency information and to provide us with a useful window on the collocational and/or colligational patterns of the language under investigation. The initial input for the analysis was gathered by extracting two-word (with hindsight or in hindsight), three-word (wisdom of hindsight), four-word and five-word clusters ((with) the benefit of hindsight). The BNC was exhaustively searched for examples of hindsight, yielding 422 tokens or occurrences. However, raw frequencies of hindsight (itself being a hidden ‘space’ and ‘vision’ term) do not take into account functional differences amongst them. Therefore, it was crucial to establish which contexts were actually part of the relevant sets. Thus, it emerges that hindsight is predominantly primed to occur in written registers (with exceptions proving the rules…). Having extracted differently sized clusters or preconstructed phrases, the following quantitative picture emerges: –– –– –– ––

with hindsight (with) the benefit of hindsight in hindsight (with the) wisdom of hindsight

appears appears appears appears

211 times 45 times 38 times 15 times

This is also to say that seemingly ‘non-patterned’ occurrences of hindsight, as in (5)

The desire can be surmised, without recourse to hindsight, in some of what he wrote, but is far from (BNC)

were excluded from the analysis. Let us now turn to the examination of the prevalent preconstructed phrases and meaning associations speakers generally attach to these patterns or con-

382   Rainer Schulze structions. The complex word hindsight has a relatively common collocation with with, which therefore seems to have a psychological as well as a statistical reality for most native speakers of English. The cluster of words thus formed, with hindsight, then typically has for speakers a semantic association with processes of understanding and evaluation of a situation, decision, judgement or event only after it has happened or developed: (6) (7)

volume still to be erupted. In this case, with hindsight, this suggestion turns out to be true (BNC) whales swim but they don’t fly. It is with hindsight that we decide whether to judge the success of (BNC)

It does not come as a surprise that the evaluation of a prior decision or judgement can be confirmatory (as in 8) or non-confirmatory (as in 9) in nature: (8) (9)

commissions, reports, surveys, and designs. With hindsight, his dismissal in 1878 was a blessing (BNC) Both the barrister, and his court, accept with hindsight that his intervention was perhaps inapt or (BNC)

Clearly, with hindsight expresses a revaluation of a prior decision or judgement, and not always and necessarily on the basis of conclusive evidence; more specifically, the decision and/or situation being commented on sometimes turn out to be events in progress which, with the aid of new information, enable the observer to put a different complexion on the matter. It is this possible change in perspective that is reflected in numerous collocates following with hindsight: clear, clearly, obvious, easy to see, become clear, viewpoint, look, apparent, etc. In other words, with hindsight plus its favourite collocates establish and reflect a very special knowledge structure at the conceptual level that centres around the frame of vision which, in turn, is put into language in the following way: (10) put forward eighty or so estimates for different schemes, with hindsight it’s clearly more prudent an (BNC) (11) because the question of reasonableness is bound to be viewed with hindsight, is a valid criticism (BNC) (12) what looks like progress towards some distant goal seems, with hindsight, to have been achieved (BNC) (13) powers realized that war was on the horizon, but with hindsight it is possible to see that the battle (BNC)

Exploring recurrent word combinations in English   383

Combinations of with hindsight with its typical right-hand collocates has for most users a pragmatic association with ‘improved understanding’. The combinations of all these primings in turn frequently colligate for most speakers with impersonal constructions, including modal verbs. As will be quickly seen from the instances below, these are far from being aberrant: (14) Brahé to conclude that the Copernican theory was false. With hindsight, it can be appreciated (BNC) (15) not flung from the surface of a spinning earth. With hindsight, this can be attributed to the inadequacy (BNC) (16) 3. Opposition to international fascism: collective security. With hindsight, there is a tendency to (BNC) (17) and for a long time I hid it from myself. With hindsight, it’s clear that I was too single-minded about (BNC) (18) equally recognized this as the heart of the dispute. With hindsight, one can read from their response (BNC) Thus, in the examples we have been considering so far, with hindsight predominantly occurs in clause-initial position, followed by an impersonal construction hiding any human agency, emotion or volition. This observation can be corroborated by the fact that with hindsight frequently functions as a hedging device that helps to weaken the speaker’s or writer’s commitment to some aspect of the proposition. If with hindsight occurs anywhere else, it can be found in second position in dependent clauses, preceded by co-ordinating or subordinating conjunctions such as and, but, although or that, as in (19) and it was greeted with general disbelief. But with hindsight, it ought to have been obvious (BNC) (20) was that he had not been compromised, although, with hindsight, it seems possible that Syrian (BNC) (21) brochures? No it didn’t require approval no and with hindsight, although, don’t think I’m being over (BNC) (22) who died in the Newton collision, he conceded that with hindsight the new layout at Newton (BNC) leaving all its ‘hidden’ properties stable. Only fairly few instances (if at all) show with hindsight in clause-final position: (23) statistically improbable in a direction that is specified not with hindsight.–; may seem (BNC)

384   Rainer Schulze (24) But Mont Blanc as we know it is defined with hindsight. Any one of a very large number of ways of (BNC) (25) or that opens the safe, is nothing to do with hindsight. It is specified in advance. The lock-manufacturer (BNC) The most obvious observation that can be made here is that the putatively strong bond between the preposition and hindsight is weakened in favour of another collocation including the preceding verb and with. In the examples we have already been considering, with hindsight is typically found at the beginning of a main clause or subordinate clause (although not always in initial position), thus displaying a clear case of textual colligational priming. We are also primed to expect the pattern with hindsight not to be repeated cohesively (i.e. a textual collocational priming), since the pattern is equivalent to a value judgement expressing the speaker’s or writer’s stance. The use of with hindsight, however, is sometimes pragmatically and/or semantically anticipated by a co-ordinating conjunction or adverb as in (26) is. Yes. But we’re looking at it with hindsight, we can’t just say that because of this this that and (BNC) (27) brought into the hospital purely for observation. Tragically, with hindsight, when Allitt came on duty (BNC) (28) the invasion as the end of an era; only with hindsight would that become clear. Instead, despite (BNC) Finally, we are primed to expect the whole expression to be accompanied by a statement of positive or negative revaluation, and with hindsight is thus highly significant with regard to semantic prosody. On closer inspection, we realize that the two-word cluster cannot be assumed to exclusively evoke either negative or positive prosodies; with hindsight typically revaluates some prior decision or event, in the meantime being enriched by additional information. What seems to be crucial here is that the cluster seems to license gradable revaluations that enable speakers and/or writers to abstain from viewing issues in black and white: (29) go cautiously back to take another look. I realize with hindsight that the dread word was preceded (BNC) (30) what looks like progress towards some distant goal seems, with hindsight, to have been achieved, (BNC) Once we have acknowledged that with hindsight constitutes a conventional unit pairing form and meaning and that the overall meaning of the two-

Exploring recurrent word combinations in English   385

word cluster cannot be exhaustively predicted on the basis of the relevant sub-parts, we must ask ourselves what the analysis of a multi-word cluster such as (with) the benefit of hindsight might contribute to the uncovering of hidden meanings. To start with syntactic issues: In a data set of 45 instances, 34 occurrences can be found with initial with, two with the initial preposition given (to the exclusion of with), eight instances of the noun phrase as a subject complement and only one instance with the noun phrase in subject position, as in (31) and fighting a war in Indochina, but the benefit of hindsight indicates now that in other ways it was (BNC) With the benefit of hindsight is less frequently available in clause-initial position; and if it occurs in this slot, the fully-fledged cluster displays basic properties of a stance adverbial: (32) the pressing problems of modern life. With the benefit of hindsight it can be seen that the 1980 (BNC) (33) threat of substitution and competitive rivalry. With the benefit of hindsight it is possible at a later stage (BNC) (34) were not those of a revolutionary. With the benefit of hindsight, historians have been able to find (BNC) It can be safely assumed that the examples provided are meant to make the distinction between the with hindsight and the with the benefit of hindsight combination as clear as possible: whereas the former cluster evokes a meaning that is related to revaluations of prior events or decisions, the latter type activates a reading of a revaluation of an earlier evaluation; this is to say that an earlier evaluation of events or decisions to which the current evaluation in the text refers has already been made. These two distinct readings do not really come as a surprise, thus echoing one of Hoey’s claims that “…synonymous word sequences […] are primed similarly but distribute themselves differently across the lexical, semantic and grammatical terrain” (2005: 79). Both constructions collocate with similar linguistic material, but one of them is far more strongly primed than the other for such habitual co-occurrences. With the benefit of hindsight is a reflection of the prior evaluation of the decision or event, facilitated by the fact that the outcome of the process is obvious and unambiguous. But this is only half the story. What is equally noticeable is that the decision or event referred to or invoked in the concordance line was either wrong or fatal, reflected in the following material:

386   Rainer Schulze (35) of whom would not unjustly, though with the benefit of hindsight, be condemned later as Fascist or (BNC) (36) begun to believe something which now, with the benefit of hindsight, appears utterly incredible: that (BNC) (37) it is easy to argue now, with the benefit of hindsight, that Dennis was doomed anyway, but it didn’t (BNC) (38) can be shown to be wrong, with the benefit of hindsight, every creditor can blame the auditors, is (BNC) In this sense, the construction itself represents a re-revaluation (or meta-evaluation) and a major critique of the benefit of later knowledge (or self-imposed self-criticism). Since the process the decision was meant to affect was completed in the past, collocation partners of with the benefit of hindsight clearly belong to the frame of history, i.e. knowledge verified by history. This is also to say that with the benefit of hindsight is less frequently primed to occur with lexical items related to vision, another indication of the fact that the process at issue has been completed and the newly adopted perspective is not really in need of any further elaboration. Typical collocation partners here include now, know, modern, smart-aleck, etc. and clearly show time-based knowledge differences: (39) falls. Now you’ve had the, benefit of hindsight which would you rather go with, the structured (BNC) (40) contained, with what we now know with the benefit of hindsight about Liechtenstein, which was (BNC) (41) But let’s not get too smart-aleck with the benefit of hindsight. Circulation before the re-launch as (BNC) As far as the evaluative or attitudinal meaning (i.e. the semantic prosody) invoked by the multi-word pattern is concerned, critique pertaining to the first evaluation or event and the positive and protective function of the expression in relation to the current decision-maker are immediately obvious. An important point to make here is that the benefit of hindsight, a four-word cluster, can be slotted into typical NP-positions in a sentence or utterance, leaving us with constructions (except for 31.) such as (42) not in a position to do. It has the benefit of hindsight. But it would also appear to know its own place (BNC) (43) and theory. Thirdly, there is of course the benefit of hindsight, which always makes the critic’s job (BNC)

Exploring recurrent word combinations in English   387

(44) Rubik’s cube. The discussion presented has the benefit of hindsight: although each idea appears (BNC) (45) goes to press, readers will probably have the benefit of hindsight.) The end of sight? That possibility (BNC) (46) after a key event, and usually enjoy the benefit of hindsight. Some sources are both primary and secondary (BNC) (47) falls. Now you’ve had the, the benefit of hindsight which would you rather go with, the structured (BNC) Here, the benefit of hindsight fills the obligatory complement slot, so that all these instances are usually found at the end of independent clauses (a textual colligational priming); and we are also primed to expect the four-word cluster not to be repeated cohesively (a textual collocational priming). As mentioned before, we are primed to expect the whole expression to be accompanied by a statement of positive evaluation, i.e. primings for textual semantic association that account for almost all the instances of the data set. Apart from the findings presented so far, the analysis of in hindsight yields another interesting result: although in hindsight also displays an expressive meaning insofar as it is indexical of the speaker or writer and expresses a personal evaluation, the two-word cluster can be credited with a property that was listed neither for with hindsight (the ‘vision’ frame) or (with) the benefit of hindsight (the ‘history’ frame): it is the ‘regret’ frame in particular which is neatly documented in 38 instances from the BNC. Where does this property come from? A widely accepted, basically cognitive answer is that we typically rely on our ‘knowledge of the world’. Indisputable as it is, this explanation is not really satisfactory because the notion of world knowledge is very general indeed. Looking at some selected in hindsight instances, we will find that the attribute is fairly visible in the following concordance lines: (48) was wanting to get things done in a hurry. In hindsight he wishes he had been more patient, particularly (BNC) (49) the cinema chains which, as Heston said, was in hindsight a death blow which merely served to hasten (BNC) (50) was a nice earner all the time and now, in hindsight, I wish we hadn’t frittered the money away so easily (BNC) (51) the lifting of the ban, wouldn’t he? In hindsight, it might have been better for him to have taken up (BNC) What we find in these instances is a direct revaluation of a decision or judgement that is now assumed to be wrong, linguistically accompanied by elements

388   Rainer Schulze of regret or sadness: the speaker or writer simply feels sorry about someone or something in the past. The right-hand collocation partners of the node expression include verbs such as wish or expressions such as it might have been better, I wish we hadn’t, I wish we had, X would have helped, etc. and express (at least in cases in which the speaker or writer is or was not involved in the judgement) empathy with the people involved or possibly a more intimate relationship between the speaker or writer and the prior decision-maker. In hindsight clearly exhibits negative semantic prosody. As far as the syntactic preferences of this cluster are concerned, in hindsight clearly functions as a clause element, with an equal distribution of clause-initial and clause-final positions. The final pattern under review shows 15 occurrences in the BNC only, (with the) wisdom of hindsight showing 10 instances. The ‘fixedness’ of the construction is infrequently weakened by premodifiers of the first noun in the pattern such as mature or painful: (52) are its (avoidable) weaknesses. With the wisdom of hindsight, it is easy to see now that the Stag’s (BNC) (53) narrowest of samples; nevertheless we know with the wisdom of hindsight that the trends in question (BNC) (54) motions of caring she saw now with the painful wisdom of hindsight that, despite all his assurances (BNC) (55) not yet have been created. If the mature wisdom of hindsight takes too long coming, the record will (BNC) As the examples show, there seems to be a tendency for speakers and writers to refer with this pattern to decisions or judgements only; compared with in hindsight, the element of uncertainty is absent. In most cases, it is plausible to assume that the current evaluator expresses his or her own experience with the putatively wrong decision or judgement in the past and wisdom is obviously gained by some painful experiences only. The instances also show that the wisdom sub-part of the expression is occasionally used in an ironic or even sarcastic way by projecting his or her own painful experiences on someone else who is unaffected by the decision or judgement. With typical collocates such as mature, pain, experience or benefit, the superordinate frame for this pattern might be difficult to grasp and possibly oscillates between the frame of pain and the frame of regret; a similar observation obtains when we look into the evaluative meaning of the construction: we get a negative prosody with regard to the prior decision or judgement, but possibly also a positive one related to the increase of wisdom. In sum, these rare instances

Exploring recurrent word combinations in English   389

unambiguously reveal that hidden meanings require a substantial amount of data for detection. 6. Summary and conclusion To conclude this paper, here is a summary of the main issues that have been addressed: speakers and writers use language for a number of broader social purposes such as greeting others, asking questions, requesting favours or informing others. In general, they make use of conventionalised and preconstructed expressions for the most frequently occurring of these functions. Additionally, speakers and writers use different patterns or constructions depending on who the addressee is, what he or she knows and expects and what subject is being negotiated (cf. the notions of social fit and adaptability). In theories of language such as cognitive linguistics, lexical priming, pattern grammar or construction grammar, different kinds of constructions or patterns, firmly entrenched as (non-) linguistic knowledge in the speaker’s or writer’s mind, such as with hindsight, with the benefit of hindsight, in hindsight or with the wisdom of hindsight constitute different linguistic units that pair a linguistic form and meaning; these in turn are specified with regard to their semantic and syntactic structure as well as their pragmatic applicability. This is to say that speakers and writers have internalised a particular perspective or stance that a particular construction embodies, and the communicative situation in which it is used. Conventionalised, firmly entrenched and preconstructed expressions exhibit meanings that are not derivable from their sub-parts, but derivable from the context in which they are used. Thus, the semantic and pragmatic contribution of with hindsight, in hindsight, with the benefit of hindsight or with the wisdom of hindsight is adjusted as a consequence of the context in which they are embedded, yielding different frames such as ‘vision’, ‘history’, ‘regret’ or even ‘miscellaneous’ In transferring the notion of context to large bodies of machine-readable texts, we are able to extract (even hidden) meanings from concordance lines which themselves form part of large data sets which in turn can be assumed to be the ‘true’ carriers of collectively shared knowledge. Sequential patterns as the basic units of language representation emerge from the collaboration of the speakers’ and writers’ memories of all the sentences, utterances and clauses in their entire history of language use and from the frequency-based abstraction of regularities within them, giving rise to the detection of hidden meanings on the part of the analyst. From a broader perspective, it turns out that the notion of knowledge in this paper has been fed by two different traditions: the cognitive-psychological type

390   Rainer Schulze and the societal, cultural and historical type. Knowledge seen from that point of view is firmly associated with cognition and communication in its broadest sense, with language in general and texts in particular representing salient ‘connecting pieces’ between these two. Different types of knowledge can be made available through linguistic contextualisations, and knowledge made available from a linguistic point of view implies knowledge that in turn is commonly shared by the members of the speech community. It thus seems plausible to assume that empirical access to semantic and pragmatic knowledge can be made possible through the analysis of recurrent word combinations taken from large data sets. The corpus-based analysis of linguistic co-occurrences, as shown in this study, does not focus on arbitrary or occasional contextualisations, but on frequency-based language use and essential semantic relationships. In this sense, recurrent word combinations refer to patterns of contextualisation and interpretation and thus function as important clues about sentence or utterance meanings, about concept-bound knowledge and about relations to superordinate knowledge-schemata or frames.

References Antos, Gerd 1997 Texte als Konstitutionsformen von Wissen. Thesen zu einer evolutions­ theoretischen Begründung der Textlinguistik. In Die Zukunft der Text­ linguistik. Traditionen, Transformationen, Trends, Gerd Antos and Heike Tietz (eds.), 43–63. Tübingen: Max Niemeyer Verlag. Assmann, Jan 1997 Das kulturelle Gedächtnis: Schrift, Erinnerung und politische Identität in frühen Hochkulturen. München: Beck. Berger, Peter L., and Thomas Luckmann 1966 The Social Construction of Reality: A Treatise in the Sociology of Knowledge. Garden City, NY: Anchor Books. Bless, Herbert, Klaus Fiedler, and Fritz Strack Social Cognition: How Individuals Construct Social Reality. Hove/ 2004 New York: Psychology Press. Biber, Douglas, Susan Conrad, and Randi Reppen Corpus Linguistics: Investigating Language Structure and Use. Cam1998 bridge: Cambridge University Press. Bublitz, Wolfram “I entirely dot dot dot”: Copying semantic features in collocations with 1998 up-scaling intensifiers. In Making Meaningful Choices in English: On Dimensions, Perspectives, Methodology and Evidence, Rainer Schulze (ed.), 11–32. Tübingen: Gunter Narr Verlag.

Exploring recurrent word combinations in English   391 Cruse, David Alan Meaning in Language: An Introduction to Semantics and Pragmatics. 2000 Oxford: Oxford University Press. Dirven, René, and Marjolijn Verspoor (eds.) Cognitive Exploration of Language and Linguistics. 2d ed. Amsterdam/ 2004 Philadelphia: John Benjamins. Fairclough, Norman 1999 Global capitalism and critical awareness of language. Language and Awareness 8: 71–83. Feilke, Helmuth 1996 Sprache als soziale Gestalt. Frankfurt am Main: Suhrkamp. Geeraerts, Dirk Methodology in cognitive linguistics. In Cognitive Linguistics: Cur2006 rent Applications and Future Perspectives, Gitte Kristiansen, Michel Achard, René Dirven and Francisco J. Ruiz de Mendoza Ibánez (eds.), 21–49. Berlin/New York: Mouton de Gruyter. Goffman, Erving Frame Analysis. New York: Harper and Row. 1974 Goldberg, Adele E. 1995 Constructions: A Construction Grammar Approach to Argument Structure. Chicago/London: The University of Chicago Press. 2006 Constructions at Work. Oxford: Oxford University Press. Halliday, Michael Alexander Kirkwood 1994 An Introduction to Functional Grammar. 2d ed. London: Arnold. Hoey, Michael 2005 Lexical Priming: A New Theory of Words and Language. London/New York: Routledge. 2009 Corpus-driven approaches to grammar: The search for common ground. In Exploring the Lexis-Grammar Interface, Ute Römer and Rainer Schulze (eds.), 33–47. Amsterdam/Philadelphia: John Benjamins. Hunston, Susan, and Gill Francis 2000 Pattern Grammar. A Corpus-Driven Approach to the Lexical Grammar of English. Amsterdam/Philadelphia: John Benjamins. Kemmer, Suzanne, and Michael Barlow 2000 Introduction: A usage-based conception of language. In Usage-based Models of Language, Michael Barlow and Suzanne Kemmer (eds.), vii– xxviii. Stanford, CA: CSLI Publications. Kennedy, Graeme 1998 An Introduction to Corpus Linguistics. London: Longman. Knobloch, Clemens Überlegungen zur Theorie der Begriffsgeschichte aus sprach- und kom1992 munikationswissenschaftlicher Sicht. Archiv für Begriffsgeschichte 1992: 7–24. Lennon, Paul 2004 Allusions in the Press. An Applied Linguistic Study. Berlin/New York: Mouton de Gruyter.

392   Rainer Schulze Louw, Bill 1993

Irony in the text or insincerity in the writer: The diagnostic potential of semantic prosodies. In Text and Technology, Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds.), 157–176. Amsterdam/Philadelphia: John Benjamins. Malinowski, Bronislaw Coral Gardens and Their Magic, Vol. 2: The Language of Magic and 1935 Gardening. London: Allen & Unwin. McEnery, Tony, and Andrew Wilson 2001 Corpus Linguistics. 2d ed. Edinburgh: Edinburgh University Press. McEnery, Tony, Richard Xiao and Yukio Tono 2005 Corpus-based Language Studies: An Advanced Resource Book. London/New York: Routledge. Moscovici, Serge On social representations. In: Social Cognition. Perspectives on Every1981 day Understanding, Joseph P. Forgas (ed.), 181–209. London, etc.: Academic Press. 1984 The phenomenon of social representations. In Social Representations, Robert M. Farr and Serge Moscovici (eds.), 3–69. Cambridge: Cambridge University Press/Paris: Editions de la Maison des Sciences de l’Homme. Pishwa, Hanna Expression of uncertain goals in communication: The case of multi2006 functional try. In Language and Memory: Aspects of Knowledge Representation, Hanna Pishwa (ed.), 269–311. Berlin/New York: Mouton de Gruyter. Robinson, Douglas Introducing Performative Pragmatics. London/New York: Routledge. 2006 Römer, Ute 2005 Progressives, Patterns, Pedagogy: A Corpus-Driven Approach to English Progressive Forms, Functions, Contexts and Didactics. Amsterdam/Philadelphia: John Benjamins. Römer, Ute, and Rainer Schulze (eds.) 2008 Special issue: Patterns, Meaningful Units and Specialized Discourses. In International Journal of Corpus Linguistics 13:3. Römer, Ute, and Rainer Schulze (eds.) 2009 Exploring the Lexis-Grammar Interface. Amsterdam/Philadelphia: John Benjamins. Schmidt, Siegfried J. 1996 Kognitive Autonomie und soziale Orientierung. Konstruktivistische Bemerkungen zum Zusammenhang von Kognition, Kommunikation, Kultur und Medien. Frankfurt/M.: Suhrkamp. Schulze, Rainer 2006 Licence to mate: Zum Paarungsverhalten eines ausgewählten Verbs im Deutschen und Englischen. Ein Entwurf. In Italienisch und Deutsch als

Exploring recurrent word combinations in English   393 Wissenschaftssprachen. Bestandsaufnahmen, Analysen, Perspektiven/ Italiano e tedesco come lingue della comunicazione scientifica. Ricognizioni, analisi e prospettive, Emilia Calaresu, Christina Guardiano and Klaus Hölker (eds.), 177–193. Münster: LIT-Verlag. Sinclair, John M. 1991 Corpus, Concordance, Collocation. Oxford: Oxford University Press. Stubbs, Michael Corpus evidence for norms of lexical collocation. In Principle and Prac1994 tice in Applied Linguistics, Guy Cook and Barbara Seidlhofer (eds.), 245–256. Oxford: Oxford University Press. 2001 Words and Phrases: Corpus Studies of Lexical Semantics. Oxford: Blackwell. Taylor, John R. 1995 Linguistic Categorization: Prototypes in Linguistic Theory. 2d ed. Oxford: Clarendon Press. Tognini-Bonelli, Elena 2001 Corpus Linguistics at Work. Amsterdam/Philadelphia: John Benjamins. Tomasello, Michael 2002 The key is social cognition. In Language in Mind: Advances in the Study of Language and Thought, Dedre Gentner and Susan Goldin-Meadow (eds.), 47–57. Cambridge, MA/ London: The MIT Press, A Bradford Book. Ungerer, Friedrich, and Hans-Jörg Schmid 2006 An Introduction to Cognitive Linguistics. 2d ed. London: Longman. Varela, Francisco J. 1988 Connaître: Les sciences cognitives, tendences et perspectives. Paris: Editions du Seuil.

Chapter 15 Emotion talk and emotional talk: Cognitive and discursive perspectives Monika Bednarek

1. Introduction The aim of this chapter is to examine emotion talk and emotional talk both from a cognitive and from a discursive perspective. To illustrate what I mean by emotion talk and emotional talk let us look at a passage from Lionel Shriver’s novel We Need to Talk about Kevin: (1)

Franklin, I was absolutely terrified of having a child. Before I got pregnant, my visions of child rearing – reading stories about cabooses with smiley faces at bedtime, feeding glop into slack mouths – all seemed like pictures of someone else. I dreaded confrontation with what could prove a closed, stony nature, my own selfishness and lack of generosity, the thick, tarry powers of my own resentment. However intrigued by a “turn of the page,” I was mortified by the prospect of becoming hopelessly trapped in someone else’s story. And I believe that this terror is precisely what must have snagged me, the way a ledge will tempt one to jump off. (Shriver 2006: 37, bold face mine, italics in original)

In this passage the first-person narrator records her essentially negative emotional responses to the prospect of child rearing, which is very significant in the context of a novel that deals with how someone (the narrator’s son) can become a murderer. On the one hand, she does this by using a number of words that directly denote emotions (marked by bold face in the passage above), such as terrified, dreaded, resentment, intrigued, mortified, and terror. Such words can be described as part of emotion talk. On the other hand, the narrator uses phrasings that more indirectly signal negative affective meaning, for instance cabooses with smiley faces at bedtime, feeding glop into slack mouths, a closed, stony

396   Monika Bednarek nature, my own selfishness and lack of generosity, the thick, tarry powers of, hopelessly trapped, etc. I shall use the term emotional talk as a shorthand for such expressions (Bednarek 2008). In other words, emotion talk makes use of expressions that directly and explicitly name a particular emotional response (e.g. fear), whereas emotional talk uses expressions that can be more indirectly related to some kind of emotional experience, which need not be clearly identifiable (Section 3 below lists different types of emotional talk). For instance, we can infer some kind of negative emotional experience from feeding glop into slack mouths but a particular emotion is not explicitly identified (e.g. as dislike, distaste, disgust, hate). The use of both emotion talk and emotional talk is best considered as a discursive strategy rather than representing the speaker’s ‘real’ internal affective state (compare also Galasiński 2004: 6). The two phenomena will be discussed to some extent in this chapter, drawing on research from cognitive science (including cognitive psychology) as well as from functional linguistics and discursive psychology (Edwards 1999). Thus, this paper aims to make a contribution to research that examines socio-cognitive dimensions of language, with emotions being considered as both cognitive and social phenomena. The interest is in the relation between language and emotion, how language use both reflects and construes emotion schemata, and in the role that emotion talk plays in cause-effect relations and justification. It is thus in a sense about how people construe social reality (emotion). The chapter starts by looking at emotion schemata and the relation between language and emotion/affect, before outlining a functional theory of affective meaning (Appraisal), and suggesting a new classification that is based on schema theory. It then adopts a more discursive perspective in analyzing the role of affect in justification in casual conversation as evidenced in data from the 100-million word British National Corpus (BNC).1 2. Emotions and emotion schemata Since this chapter deals with references to emotional experience, it may be useful first to look at theories of emotion, and at theories concerning schemata or belief/knowledge structures about emotional experience. Indeed, although there are manifold definitions of emotions (Jahr 2000: 7; see Mees 2006; Oatley, Keltner and Jenkins 2006 for overviews), emotions are 1 In the examples from the BNC, the following codes are occasionally used: PS000 (‘speaker cannot be identified’) and PS001 (‘group of unidentified speakers’). More information on the data is provided in the Appendix.

Cognitive and discursive perspectives   397

often described as involving various components such as an eliciting condition or an antecedent event, a cognitive evaluation or appraisal, a physiological response and an action readiness/an action. In this definition, “emotions are complex physiological-affective-cognitive responses to the physical and sociocultural environment” (Schrauf and Sanchez 2004: 267). It seems that people’s folk knowledge includes an awareness of these various components of emotional experience (Kövecses 2000: 130) and that their emotion schemata comprise such aspects.2 These schemata are presumably based on people’s actual emotional experience (e.g. increased body heat when angry), and observing it in others (perception) as well as exposure to discourse on emotions and other socializing processes. Following Kövecses we can say that emotion schemata are both motivated by human physiology and produced by the socio-cultural environment (Kövecses 2000: 14). Such structured folk knowledge is important for interpreting emotional experience and its representation as well as for adequate interpersonal interaction. Aspects that may be involved in emotion schemata are: –– the kind of cognitive evaluation associated with an emotion;3 –– prototypical and potential antecedent events (eliciting conditions) that cause an emotion; –– psycho-physiological expressions that are prototypically and potentially caused by an emotion; –– prototypical and potential expressions of action readiness4 –– prototypical and potential subsequent actions caused by an emotion. In other words, it is hypothesized that people are (at least subconsciously) aware that emotional experience involves evaluations and eliciting conditions, that emotions are accompanied by bodily symptoms (e.g. increased heart rate), and that they might cause an Emoter to be prepared for and, consequently, to 2 Schema theory suggests that mental knowledge is organized in structures or chunks (schemata) which capture the typical features of the world (see Bednarek 2005a, 2006b). 3 See, e.g., Ortony, Glore and Collins (1988) or Ellsworth and Scherer (2003) on cognitive appraisal processes. The hypothesis is that “for events to prompt emotions, they must be evaluated, or appraised, in relation to the individual’s goals” (Oatley, Keltner and Jenkins 2006: 167). 4 Action readiness items are for example ‘I wanted to oppose, to assault, hurt, or insult’ (for antagonism), ‘I felt inhibited, paralyzed, or frozen’ (for inhibition), ‘I wanted to do something, but I did not know what’ (for helplessness) (Oatley, Keltner and Jenkins 2006: 132; more examples in Fridja et al. 1995: 129).

398   Monika Bednarek perform a certain action. For example, most people would agree that an emotion like sadness typically involves some of the following (plus other) aspects: –– –– –– ––

Cause: death, loss Feelings: helpless, tired Expression: drooping posture, crying Actions: negative talk (Oatley, Keltner and Jenkins 2006: 184)

A large body of research from cognitive linguistics as well as cognitive and cognitive-social psychology, and other research on emotion (e.g. Shaver et al. 1987; Johnson-Laird and Oatley 1989: 92–93; Russell 1991; White 1990; Gottman et al. 1996: 251; Kövecses 2000; Mees 2006: 7; Oatley, Keltner and Jenkins 2006: 184) appears to support this claim. For instance, it has been shown that speakers are aware of typical behaviour and situations that are associated with emotions (Ortony, Glore and Collins 1988: 3; Wowk 1989). Whether or not emotion schemata mirror actual emotional experience is another matter, as is the question of how and whether the various components of emotional experience are temporally related. In any case, emotion schemata are at least partly influenced by culture, and it should therefore be emphasized that the comments in this chapter only relate to English. 3. Emotion talk and emotion schemata Now, where does language come into play? The relation between language and emotion is a very complex one and cannot be explored in depth here. Instead, I shall concentrate only on the above-introduced difference between emotion talk and emotional talk (for an extended discussion see Bednarek 2008: 10–12), and the relation of emotion talk to emotion schemata. As said, emotional talk concerns expressions that work conventionally to signal some kind of emotional experience (without this necessarily being experienced by the speaker). Such expressions can be verbal or non-verbal. Examples that are listed in the relevant literature include intonation, mental process verbs, grading (intensifiers, comparison, quantifiers, mood, modality, negation), repetition, interpersonal metaphor, figurativeness, punctuation, interjections, affective derivation (diminutives/ augmentatives), inversion, exclamation, syntactic markedness, pronoun use, emphatic particles, intensifiers, expletives, vagueness, affective connotations, evaluative adjectives, and many more. Paralinguistic devices are facial expressions, vocal cues, gestures, body posture, body movement, and

Cognitive and discursive perspectives   399

physiological cues. (This means that talk is used in an extended sense here.) The term emotion talk, on the other hand, is used to refer to lexical items that denote emotional experience for example love, hate, joy, envy, sad, mad, enjoy, dislike etc (as well as fixed expressions such as He had a broken heart). Emotion talk can refer both to transient, momentary emotional states and to more permanent emotional dispositions: (2)

Mark always experienced the same feelings of contentment when entering the home straight. [emotional disposition] Today he felt particularly pleased to be back. [momentary emotional state] (BNC, AC2 21–22, bold face mine)

The main focus of this section is on emotion talk, but emotional talk will be relevant further below. Let us start the discussion of emotion talk and its relation to emotion schemata by looking at an example: (3)

He tried to hide his wild consuming fear, but I saw it. In the early hours of the morning between two and half-past-two, I jolt awake from his rushed breathing. Rolls this way, that side of the bed. He’s pale. Ice cold in a sweltering night – sopping wet with sweat. (from Martin and Rose 2003: 10, bold face mine)

From a cognitive perspective, we can regard fear in (3) as a cue or trigger which activates the relevant (folk) emotion schema (Belleli 1995: 493; Bednarek 2005a), “a shorthand … to refer to the various events and processes which comprise the phenomenon of emotion” (Ekman 1997: 3). This means that we would say that both the first-person narrator and the reader can draw on this schema to make sense of the references to rushed breathing, rolling in bed, paleness, low body temperature, and sweat as components of a possible schema of fear. Their schemata can help to establish coherence. The particular aspects of this schema are cued by the context in that it is made clear that the symptoms and behaviour are in fact caused by fear (He tried to hide his wild consuming fear, but I saw it. In the early hours of the morning between two and half-past-two, I jolt awake from …). There is a difference then between potential schemata (schemata in the mind) and actualized schemata – emotion schemata that are construed by discourse (by the word in context; this includes larger phrasal units, see below). In other words, we can look at emotion talk from two perspectives: the cognitive and the discursive. Whereas the cognitive perspective looks at “discourse as a pathway to individuals’ inner life, whether it be cognitive processes, motivations or some other mental stuff” (Edwards and Potter 1992: 127), the dis-

400   Monika Bednarek cursive perspective regards “psychological issues as constructed and deployed in the discourse itself” (Edwards and Potter 1992: 127). Consider a second example: (4) PS000 George had no affection in him, I found that out. Annette Really? PS000 No affection. Annette Now you surprise me cos I would have thought cos you always went round holding hands and looked so happy. PS000 Not with George wouldn’t have hold your hand. Wouldn’t have touched you! Annette Oh? PS000 No! No way! The most affection you got from George was a pat on head! Annette Oh crumbs ! PS000 Oh no, you see he used to say about Tracy PS000 people that held hands. He used to be quite Annette Yeah! PS000 about people. Tracy Oh! I see PS000 Yeah! Yeah! Annette Yeah. PS000 See it’s funny, he used to think all people just, he used to say, oh he’s doing that for show. Annette Mm. PS000 He’d never had affection so he couldn’t give it. Annette No. Annette No. PS000 You know, but Annette No. PS000 I used to make allowances for the way he was brought up. Annette Yeah. PS000 He was brought up with his grandma and that. Well I Annette Yeah. PS000 thought he’s never had affection. Tracy No. People of that age, it’s like David’s grandma they don’t know what affection means. Annette No!

Cognitive and discursive perspectives   401

PS000 No! Annette They say we’re daft for showing our affection. PS000 Yeah. But you should do. I only just realized that. Well it’s taken me a long time to even be able to give affection. Annette Yeah. (BNC KB9 1281–1324 italics and bold face mine) One of the functions of emotion talk is to reveal speakers’ attitudes towards emotion – what has been called meta-emotion philosophy (Gottman et al. 1996: 243) or emotion ideology (Turner and Stets 2005: 36). In example (4) PS000, Annette and Tracy discuss and negotiate their conception of affection (with respect to George’s behaviour): Affection for them is something that you have in you, that you give (to other people), that you get (have) from other people and that you show (mostly by holding hands but also by touching each other). They also explicitly comment on other people’s deviant, and, in their view, deficient emotion schemata. In so doing, they set themselves up against people who don’t know what affection means, who say we’re daft for showing our affection, and who think that holding hands is for show only. In contrast, they communally negotiate and construe the showing and giving of affection as important (you should do), even if it may be difficult (it’s taken me a long time to even be able to give affection). The passage is thus an interesting illustration of the discursive construal of emotion schemata, and of the sub-cultural variability of such schemata. At the same time, it is reasonable to assume that e.g. the potential connection between holding hands and feelings of affection is one that is part of a generalized emotion schema that would be shared by many people. That is, even someone who disapproves of public displays of emotion is aware of the schema that connects affection and holding hands, though they may evaluate the latter negatively. However, explicit and detailed schema construal, as exemplified by example (4), does not always take place. Thus, although the following passage also makes reference to affection as something to be given and to receive, no references are made as to how this is to be done. The author can draw on the assumption that readers will share a generalized schema for affection that includes knowledge about how it can be shown to animals (e.g. via patting, stroking, talking etc.). (5)

A horse’s natural desire for affection can be extended and enlarged. A horse needs to give affection and to receive affection from its mother or offspring, and its close companions. Thus it can learn to need affection from people too. This can be valuable to us. The desire to gain

402   Monika Bednarek affection and approval from its owner can be a prime motivator for the horse to try and do what we ask of it.5 (BNC ADF 440–445, bold face mine) Similarly, in example (1) above, there are no detailed discursive schematic construals for each of the emotions referred to via emotion talk (terrified, dreaded, resentment, intrigued, mortified, terror). Rather, the emotional talk construes a general negative affective attitude on the part of the first-person narrator, which, however, is in line with schematic knowledge that negative evaluations are components of negative emotions (such as fear/resentment). This negative affective attitude is furthermore contextually associated with these emotions through emotion talk. In other words, given the context, we read the negative evaluations as relating to the narrator’s negative emotions of fear and resentment towards child rearing. Generally speaking, speakers/writers can draw on aspects of assumed common schemata until challenged by the hearer/reader. Differences and mismatches of schemata therefore result in explanations, contradictions and questions (Tannen and Wallat 1993). Research in cognitive science has shown that inconsistent information can be ignored, sub-categorized, give rise to new schemata, or result in a modification of an existing schema. Many words in English refer explicitly to parts of emotion schemata (antecedent event, evaluative process, physiological state, action, situational circumstances, etc. (Heelas 1986; Johnson-Laird and Oatley 1989: Bellelli 1995; Fiehler 2002), triggering inferences about the emotional response involved, often guided by the context. Thus, while speechless may be glossed in the dictionary as ‘not able to speak, especially because you are extremely angry or surprised’ (Oxford Advanced Learner’s Dictionary), the context will make clear which 5 The schema that is at stake here is one that relates to affection as shown to horses rather than humans, and involves cultural background assumptions (for instance, that it is morally right to condition horses to behave according to our wishes). This becomes easily apparent if we substitute ‘horse’ vocabulary in (5) with ‘child’ vocabulary, with the resulting text only imaginable in the context of brain washing ‘manuals’ or similar documents:

A [child’s] natural desire for affection can be extended and enlarged. A [child] needs to give affection and to receive affection from its mother […], and its close companions. Thus it can learn to need affection from [other] people too. This can be valuable to us. The desire to gain affection and approval from its [parents] can be a prime motivator for the [child] to try and do what we ask of it. So far, so good. (BNC ADF 440–445, bold face mine)

Cognitive and discursive perspectives   403

emotional experience is involved (if any). Here are some examples from the BNC for the phrase speechless with + emotion (bold face mine): … which left Jenny almost speechless with delight. (B0B 796) … Chairman of the Action Committee, was almost speechless with anger. (CAR 1675) (III) She was almost speechless with anger now. (JY3 934) (IV) … the artists before whom he became speechless with respect. (FB9 1514) (V) All morning everyone had been speechless with depression about “Sonnet”. (FBM 2658) (VI) … at the famous Hurlingham Club which left her speechless with wonder (CA0 2399) (VII) … of flanking attack which always left Matthew speechless with admiration. (GV8 2936) (VIII) … which has left me speechless with rage. (CH1 6719) (IX) … I was speechless with a mixture of fear and embarrassment. (AC6 1365) (X) … the same route they had just walked, I was speechless with indignation. (AS3 1241) (XI) Naturally, he was speechless with rage and he seemed to go a funny colour (J56 297) (XII) For a time the eagles were speechless with surprise at what had happened (FP3 1022) (XIII) … and hurried down the path, and Katherine, speechless with rage (EVG 1390) (XIV) … governor could only stand and watch it fall, speechless with anger. (FU8 725) (XV) She went scarlet, speechless with rage as she glared at him in the sunlit (JYD 2120) (XVI) Speechless with rage, Mait hurled the card in the Doctor’s (FSR 1984) (XVII) Speechless with weary anger, she cut herself two crooked slices (H8N 2285)

(I) (II)

There are other ways of contextually associating speechless with emotions (e.g. an adjective + noun structure: But the others could only stare at him in speechless amazement, CAB2901), but considering only the pattern speechless with, this is most frequently associated with an emotion of anger (anger, indignation, rage), followed by surprise and admiration. However, speechlessness can

404   Monika Bednarek potentially also be associated with delight, fear, embarrassment and depression (Figure 1):

delight

anger, indignation, rage

admiration, respect

fear & embarrassment speechless with

(+)

surprise, wonder

depression (-)

Figure 1. Emotions associated with speechless with

Consequently, while speechlessness may potentially be a part of schemata for various emotions, the context can tell us which specific emotion is at stake. Where there are no contextual cues (e.g. in an artificial experimental situation), the degree of inferencing that has to be done may depend on the strength of association between action, behaviour etc. and a particular emotion. Potentially, there are (at least) two different ways in which the strength of an association can be tested by researchers: –– discourse frequency –– psychological salience Firstly, we can consider corpus evidence and systematically investigate the association of speechlessness in discourse with emotions, and the frequency with which certain emotions are associated with it. For example, just examining the phrase speechless with, the strongest connection would be with an anger-like emotion. Using Hoey’s (2005) terminology, the phrase speechless with is “primed” for speakers to occur with these emotion terms, i.e. “is associated in the mind of a language user” (Hoey 2005: 24) with the semantic set of anger emotions. (For a comparison of Hoey’s concept of priming and the psychological concept of priming see Hoey 2005: 7–8.) Secondly, we can present informants with a list of emotion related behaviour such as that of speechlessness and see which emotions they elicit for them most frequently and consistently across time and across informants. For example, speechlessness might perhaps be more frequently and consistently associated with surprise rather than anger. (This is a hypothesis, as it was beyond the scope of this chapter to conduct any such studies). It would be

Cognitive and discursive perspectives   405

interesting to compare the results of the two kinds of studies in terms of what they tell us about language and cognition (especially if the results differ). A discussion of these issues is well beyond the scope of this paper, as it would take us deep into different notions about language, meaning, culture and cognition. And problems about the relation between these “are clearly very difficult – perhaps impossible – otherwise they would already have been solved, especially as they have been discussed by scholars such as Popper, Searle and Halliday, not to mention the whole of the cognitive sciences” (Stubbs 2006: 28). A final note on emotion terms as ‘pathway to individuals’ inner life’. The usage of emotion talk and emotional talk does not reliably tell us anything about speakers’ current emotions, e.g. whether or not someone who says I’m happy really is happy (whatever that may mean). Werth, among many others, also mentions “the practical difficulty – if not impossibility – of verifying mental states. … One partial answer is the Gricean one: if you say you are delighted, then we should tend to believe that you are cooperatively reporting a genuine state of mind. However, this is not a semantic solution of course, but a pragmatic one” (Werth 1998: 413). Such a purely Gricean approach also disregards other principles that may be at work, such as principles of politeness (interacting with factors such as sex/gender, ethnic group and power status of speakers and hearers) or cultural norms that are particularly relevant where references to emotions are concerned (Downes 2000: 108). In any case, the genuineness of emotional experience, though interesting from a psychological viewpoint, is perhaps not the most important issue in linguistic analysis. Instead, we take up the notion that emotion/ emotional talk does not necessarily stand for or represent the speaker’s or others’ ‘real’ internal affective state but that it rather represents what Galasiński (2004: 6) calls a discursive practice, and that it is strategically related to notions such as self-presentation (Caffi and Janney 1994: 328–329) or constitutes conventionalized, habituated discourse. To recapitulate, both speakers and readers draw on schematic knowledge to construe and interpret discourse, but they also establish context-dependent, dynamic and discursive schemata, which may be communally negotiated. Further, this construal has a social significance, and accomplishes social actions (Edwards and Potter 1992: 2–3). Edwards (1999) speaks of “a set of rhetorical affordances, in which different parts or potentials of meaning, even contrasting ones for the same word, may be worked up and deployed, on and for occasions.” (Edwards 1999: 281–282). Both perspectives are hence needed when examining language and emotion: the cognitive and the discursive. Research on emotion schemata can provide underlying explanatory power and a grid for classifying affective language,

406   Monika Bednarek whereas adopting the discursive perspective can show us important social functions of emotion talk. Both perspectives are now explored using some ideas from Appraisal theory (as developed in linguistics – not to be confused with psychological theories of appraisal that are discussed in emotion research). 4. Appraisal theory Appraisal theory has been developed in systemic functional linguistics to investigate interpersonal meaning (e.g., Martin and White 2005). It sub-divides interpersonal Appraisal resources into three main systems: Engagement, Graduation, and Attitude. Engagement has to do with voicing, modality, evidentiality and similar concerns; Graduation concerns grading (e.g. intensifying/ amplifying or downtoning); Attitude is about positive and negative evaluations. Attitude itself can be further sub-divided into Affect (positive/negative emotions), Appreciation (positive/negative aesthetic evaluations of phenomena, such as texts or processes), and Judgement (positive/negative ethic evaluations of people and their behaviour). These three Attitude systems have further, more detailed sub-classifications (In/security, Dis/satisfaction, Un/happiness, Dis/ inclination; Impact, Quality, Balance, Complexity, Valuation; Normality, Capacity, Tenacity, Veracity, Propriety) which, however, are not relevant in this chapter (Graduation and Engagement are not discussed, either). Table 1 gives an overview of Attitude lexis according to Appraisal theory.6 Attitude: positive/negative evaluation Affect: reference to emotional experience Appreciation: aesthetic evaluation of phenomena/products Judgement: ethic evaluation of people/behaviour Table 1. Attitude lexis Appraisal systems

Attitude

Example lexis (from Martin and White 2005)

Affect

miss, fearful, love, sad, trusting, depressed, uneasy, astonished, pleased, sick of, bored with…

Appreciation

captivating, boring, good, nasty, harmonious, irregular, elegant, arcane, profound, shallow…

Judgement

fashionable, dated, robust, weak, brave, cowardly, reliable, disloyal…

6 The classification of attitudinal lexis is context-dependent, and this is not to be considered as a ‘dictionary’ of attitudinal lexis, cf. Martin and White (2005: 52).

Cognitive and discursive perspectives   407

Starting with references to emotional experience (Affect), Appraisal researchers use a number of factors to classify these. Martin and Rose (2003) list three different ways of making reference to emotions: 1) writers employ emotion labels ( fear), 2) they use expressions describing “behaviour that also directly expresses emotion” (Martin and Rose 2003: 26) or 3) they mention “unusual behaviour which we read as an indirect sign of emotion” (Martin and Rose 2003: 27) (be very quiet). The last represents what Martin (2002) has called “behaviour which indexes emotion” (Martin 2002: 203), and the reader needs to spend some cognitive effort in retrieving the emotion: from this unusual behaviour we know something is wrong but we can’t be quite so sure about the exact emotion being expressed; we need to use a bit of psychology perhaps. Read in context, however, we do know what Helena’s on about, because these symptoms are surrounded by explicit references to emotions which tell us what the strange behaviour means. (Martin and Rose 2003: 27)

In Martin and White (2005: 47), the difference is between 1) behavioural surge (weep, smile) and 2) mental process/state (dislike, sad). A further distinction that is relevant to the discussion in this chapter is the difference that is made between inscribing and evoking (sometimes also called invoking) Attitude (Martin and White 2005; Bednarek 2006b). This distinction is currently still being elaborated on in Appraisal research (e.g. Hood and Martin 2007). According to Hood and Martin (2007), inscribed Appraisal refers to lexis that is explicitly evaluative, whereas evoked Appraisal covers a range of more indirect options for evaluation: provoked, flagged and afforded Appraisal. They elaborate: “The term [provoke] refers to implicit ATTITUDE which is evoked through lexical metaphors. Choosing [flag] means that we deploy some kind of GRADUATION to alert readers to the feelings at risk. The [afford] option makes room for the ways in which ideational meanings alone imply evaluation” (Hood and Martin 2007: 746). Although these mechanisms for indirect realization are supposed to apply to all sub-systems of Attitude (Affect, Appreciation and Judgement) in Appraisal theory, I shall limit the discussion to Affect. Another point worth making is that expletives, interjections and swearing are not included in Attitude (Hood and Martin 2007) though Martin and White have suggested to look at these phenomena “as outbursts of evaluation which are underspecified as far as type of attitude is concerned” (Martin and White 2005: 69, original emphasis). Finally, it remains to be mentioned that whoever experiences the emotion is classified as the Emoter, and what evokes the emotion as the Trigger, e.g.: The boy [Emoter] liked the present [Trigger].

Figure 2 . Resources for authorial and non-authorial affect

408 Monika Bednarek

Cognitive and discursive perspectives   409

5. Affect: A cognitive perspective Appraisal is a social-functional theory, which, to my knowledge, has not yet been related to cognitive research on emotion schemata. In this section I shall explore how the cognitive perspective (the concept of emotion schemata) can give us an alternative grid for a classification of Appraisal (here Attitude) and related resources. I do this by relating the introduced aspects of Appraisal to the components of emotion schemata mentioned in Section 2 above. It is useful to do this with respect to different Emoters. On the one hand, we can look at how speakers/writers can convey affective meanings to listeners/readers, showing that their own emotional experience is at state (authorial affect (White 1998): Emoter = Speaker). On the other hand, speakers/writers can describe emotional experience by Emoters other than themselves (non-authorial affect (White 1998): Emoter ≠ Speaker). Figure 2 incorporates both of these possibilities and also includes options for reporting speech and thought, with the main focus on affective resources identified in Appraisal theory, in order to keep the complexity manageable. (Note that this chapter is not at all concerned with the evoking of emotions in readers/hearers, and that this issue is therefore not represented in Figure 2.) In Figure 2 (which is not a system network as used in systemic functional linguistics but rather a simple taxonomy), the resources shown above the dotted line are the resources that are available with authorial affect, whereas the resources shown below the dotted line are the resources that are available with non-authorial affect. The main difference is that when the Emoter is not the Self, emotional talk devices are only available in reported speech/thought for showing the Emoter’s emotions. Compare Figure 3.

Emoter = Self (authorial affect): Emoter ≠ Self (non-authorial affect):

Fuck! I said/thought: “Fuck!” S/he said/thought: “Fuck!”

Figure 3. Authorial and non-authorial affect

The curved lines leading from I said/thought: “…” (above the dotted line) and from S/he said/thought: “…” (below the dotted line) to the left-hand side of Figure 2 show that the resources above the dotted line can be used as the content of one’s own and others’ reported speech and thought to show an Emoter’s experience. For example: I was afraid vs. I thought: “I’m afraid” and S/he was

410   Monika Bednarek afraid vs. S/he thought: “I’m afraid” or S/he thought that she was afraid (and so on, for diverse options for reporting speech and thought see Leech and Short 1981; Halliday and Matthiessen 2004: 441–482; Bednarek and Bublitz 2006). Although the linguistic examples in Figure 2 (listed on the right-hand side) are invented, many similar authentic examples from different types of discourse can be found in relevant emotion research, for example: My heart sank; He had a broken heart; his voice broke; he became very much more taciturn, more difficult to talk to, more tense, more withdrawn; he seemed to have aged and lost weight (from Bednarek 2005b); we fenced them in like sheep; all we children being herded up, like a mob of cattle; we smashed their way of life; you touch my kids and you fight me (from Martin and White 2005: 62–67); I could kill you; I could strangle you; I could kiss you (heard on TV); my biggest sister got into a car accident so she died; I moved to Worcester and I couldn’t see my neighbors and their dogs; my Mommy hit me she hit me in the eye (from Bamberg 1997a). In Figure 2 the resources that have been introduced above (emotion talk and emotional talk, inscribed and evoked Appraisal, and non-gradable Involvement (Martin and White 2005: 33) resources such as swearing) are treated as affective linguistic resources. In other words, affect is taken as primary. This means that: –– Appreciation and Judgement are re-interpreted simply as references to aesthetic and moral-ethic evaluative meanings, and treated as a way of indicating affect.7 Appreciating and judging lexis has to do with social standards that are esteemed in particular cultures and societies with respect to aesthetic and moral value. –– Expletives and interjections are seen to express Affect rather than Appreciation/Judgement; they are ‘underspecified’ only as far as Affect is concerned. –– Affording, provoking and flagging are also considered primarily with reference to Affect. That is, utterances such as we smashed their way of life are regarded as showing the writer’s negative affective stance towards the 7 Rather than classifying as Judgement appraisals of people and their behaviour, and as Appreciation appraisals of entities, phenomena and processes, all and only those appraisals that make reference to ethic-moral standards are considered as Judgement and all and only those appraisals that make reference to aesthetic standards are considered as Appreciation. Note that in Appraisal theory Affect also has a special role, in that Judgement and Appreciation are regarded as ‘reworked’ feelings, with affect “at the heart of institutionalised feelings” (Martin and White 2005: 45).

Cognitive and discursive perspectives   411

proposition (the ideational content) and described as flagging Affect rather than Judgement. –– Where we draw the line between inscribed and evoked/invoked Affect (as well as between emotion talk and emotional talk) is in fact not clear. For instance, fixed figurative expressions can be considered as inscribing Affect (and thereby part of emotion talk) since they are conventionalized to a very high degree. On the other hand, they involve metaphor and metonymy (Kövecses 2000) and could thus also be classified as provoked rather than inscribed Affect. There is also no clear dividing line between behavioural Affect and the description of behaviour (see also Martin and Rose 2003: 27). It is important to emphasize that this classification is clearly not a systemicfunctional linguistic one. Rather, it is a classification of how emotional experience can be represented linguistically, according to components of emotion schemata. It suggests that the different linguistic resources work conventionally to realize affective meanings because speakers are aware of the various components of such schemata (and vice versa, exposure to such discourse contributes to the construal of such schemata, even if not exclusively). Even though Figure 2 above lists linguistic examples in an isolated, contextindependent form, no assumption is made that affective meanings occur in isolation in discourse. Rather, as Martin and Rose have mentioned (see above), explicit references to emotions that occur in the context of more indirect references can tell us what the latter mean. Some examples (bold face mine): (6)

As he explained when I met him, he became so angry when he read what they had to say about the Princess that he wanted to ring the editors and complain, but realized that, rather than spending every day on the telephone, a simpler solution was to stop reading them. (BNC A7H 1421)

(7)

With her penetrating instinct she did not like him, and was so angry with me for, as she said, “wasting myself upon such rubbish”, that in the end she turned me out of my room … (BNC AC6 951)

(8)

I was so angry with him, I called him a jerk in the street, and worse than that, and I told my mum about it. (BNC ADG 429)

In example (6) a mental state term (angry) frames both the references to cause (when-clause) and desired behaviour (complain) in the context of being angry.

412   Monika Bednarek In (7), the matter is slightly more complex, though again a mental state term (angry) frames references for cause (angry for doing x …) and behaviour (she turned me out of my room). The reported direct speech (“wasting myself upon such rubbish”) while referring to cause simultaneously involves emotional talk (wasting, rubbish) that can be read as a sign of the reported speaker’s anger. In (8), the linguistic behaviour that is reported (I called him a jerk in the street, I told my mum about it), and that involves emotional talk (jerk) is again directly framed as caused by anger through the use of a mental state term (angry). In fact, the phrase Emoter BE so ‘emotion’ (when/for etc) that can be said to discursively construe a complex emotion schema that links a particular emotion to particular causes of emotion and particular caused behaviour. The existence and acceptability of such phrases also clearly shows the culturally accepted notion of ‘double causality’ with respect to emotion: for English speakers, at least, emotions are caused by events and are the cause of behaviour. This is part of their emotion schema (Kövecses 2000: 64, 85). This kind of research takes us into frame semantics (http://framenet.icsi.berkeley.edu; see Bednarek 2008 for a discussion). For an extended example, let us turn to a news report from the BNC. In this example, literal and figural affect terms are enclosed in square brackets ([]), references to Emoter’s behaviour is in bold face, and references to causes of Emotions are in italics: (9) THE [HATE] TRAP Homes can’t be sold Couples can’t split Families [live in hell] THOUSANDS of couples who [hate] each other are trapped together [in a living hell] because of the slump. The [tormented] partners have had to put off their divorce actions as they can’t sell their homes. Now they are stuck with each other, locked in violence and [misery] with no end in sight, lawyers revealed yesterday. MEN [explode in frustration] and batter their wives. WOMEN strike back with anything that comes to hand – including lknives and rolling pins. SOME have nervous breakdowns and suffer stress and high blood pressure. (BNC CH2, from The Daily Mirror)

Cognitive and discursive perspectives   413

This extract illustrates well the prosodic nature of affect noted e.g. by Martin and Rose (2003), and that different types of affect “often work together” (Martin and Rose 2003: 28) to create textual effects. This particular text starts with references to emotions and causes of emotions in the headlines and lead. It continues with this portrayal in its description of past events (The tormented partners have had to put off their divorce actions as they can’t sell their homes). As we move on to the present (Now) we encounter the first description of behaviour caused by emotion (alongside references to causes and emotions), which is then continued in the following three sentences, with the final two sentences only referring to behaviour caused by emotions. The patterning is as follows: E–C–C–E–E–E–C–E–C–C–B–E–E–B–B–B (E = referring to emotions, C = referring to causes of emotions, B = referring to behaviour caused by emotions) It seems as if the first reference to emotion (E) – The hate trap – functions as an emotional ‘flash’, which tells the reader that the text s/he is about to read is interesting in terms of emotional significance. This reference to emotional experience is then repeatedly encountered again in the text: families live in hell, couples who hate each other, a living hell, tormented, misery, frustration, nervous breakdowns etc. Both hate and trap are picked up again in the body of the text: trap is referred back to with modality (Homes can’t be sold, Couples can’t split, can’t sell their homes) as well as expressions that belong to the lexical field of trap (trapped together, stuck with each other, locked). This textual patterning confirms White’s finding that news reports typically have two phases: “an opening nucleus containing the text’s core informational and interpersonal meanings; a subsequent development stage which acts not to introduce new meanings but to qualify, elaborate, explain and appraise the meanings already presented in the opening ‘nucleus’.” (White 1997: 111). 6. Affect: A discursive perspective We have already begun the move from a cognitive to a discursive perspective in the sections above by showing the co-textual patterning of different affective meanings, and the discursive construal of affective schemata. In this section we will complete the move from the cognitive to the discursive with a focus on the role of affect in discourse, specifically in justification. Rosenberg has noted that there is “too little exemplification of how actual conversations proceed” (Rosenberg 1990: 170; similar observations are made by Anderson

414   Monika Bednarek and Leaper 1998: 419; Fiehler 2002: 80) in emotion research, and as late as in 2004, Galasińksi has pointed out that there “are relatively few studies of discourse strategies employed by people accounting for, explaining, or simply telling stories about their emotions” (Galasińksi 2004: 2–3). In fact, two functions of affect in narrative have caught the attention of many linguists in the past: this is the use of affect terms by speakers to describe the emotive reaction of characters in a narrative or by hearers to offer an evaluative comment on the narrative. The classical reference is Labov (1972). Cortazzi and Jin (2000: 108) give an overview of the application of Labov’s model (see also Bamberg 1997b and Bednarek 2008 for further references). Other functions of affect in conversation have also tentatively been identified (Bednarek 2008), though it is true that much more linguistic research in this area needs to be done. In psychology, too, not much research has focussed on discourse (Edwards and Potter 1992), although lately a few studies have started to address this issue under the name of Discursive Psychology, defined as the discourse analysis of psychological topics (Edwards 1999: 271). Such analyses show how flexible emotion discourse is (Edwards 1999: 277) and are interested in discursive construals: “Rather than focusing on standardized scenarios and their cognitive representations, discourse analysis focuses on how specific stories are constructed on and for their occasions, including the ways in which links between emotions and scenarios can be discursively worked up and made relevant” (Edwards 1999: 278). Taking the cue both from linguistic and psychological studies of what emotion talk ‘does’, the following analyses concentrate on the role of affect in cause-effect relations and justification in casual conversation. The data for these analyses consist of seven conversations (22,613 words) from the British National Corpus (BNC) spoken demographic part. When analyzing this corpus it was noted that affect terms are often used to motivate actions, justify them, or give explanations for them as well as in other cause-effect relations. In linguistic research, this particular discourse function of affect has been somewhat neglected so far in contrast to other functions of affect. Similarly, there is not much research on the discursive construction of reasoning in psychology, as Edwards and Potter state: “The psychology of […] everyday causal reasoning [.] has scant regard for the way versions of events are actively put together to bolster particular causal stories and undermine others.” (Edwards and Potter 1992: 1). In fact, even a small corpus shows that different types of causal connections are made with affect terms. A broad distinction can be made between the use of affect as part of justifications for behaviour and subjective judgments, in contrast to affect as something that is itself justified in discourse. Starting with

Cognitive and discursive perspectives   415

the former, let us look at some examples (utterances with affect in bold face, parts of justifying structure in italics): (10) GORDON AUDREY GORDON AUDREY GORDON

[67] She said we always sleep together [laugh] [68] sister and me. [69] Yeah. [70] She said, why do you do that? [71] Says, well … if I wake up in the night and my sister’s not there I’ll think someone’s taken her or she’s run away, I’ll be frightened. [72] And the same applies the other way so we always stick together. (BNC, KBC 67–72) What is at stake here is a justification of the children sleeping together rather than by themselves. This is justified by reference to the (hypothetical) emotion of fear (I’ll be frightened – And the same applies the other way), which in itself is causally related through an if-clause. Interestingly, the girl’s fear is not simply construed as caused by waking up in the middle of the night and being by herself, it is that this will in turn cause her to think that her sister has gone forever (someone’s taken her or she’s run away). The causal construal (which is reported by Gordon) thus is as follows: wake up + sister not there  induces mental state of belief that sister gone  induces fear That being frightened is something to be avoided (i.e. evaluated negatively by the child and presumably by the culture as a whole) becomes clear from the fact that the avoidance of fear causes action which has to be explained (sleeping together/sticking together), and which is hence in some sense exceptional – compare the comments above on schema deviations and their linguistic indicators. Another example where affect occurs in the context of trying to explain behaviour is (11): (11) CLARENCE

[1521] [laughing] And Sambo flying around yeah [] [1522] … No he just crawled under the door … sort of glare, […] glare at me and then scootled, shooted past, you know. [laugh]

416   Monika Bednarek NINA CLARENCE NINA CLARENCE NINA CLARENCE

[1523] Why? [1524] Shot I should say. [1525] [laughing] Shooted past [] . [1526] Shooted past. [1527] He shooted past as well as shot past. [1528] [laugh] Mm. [1529] I wonder why? [1530] I don’t know because he he’s never really been frightfully keen on me, that cat. (BNC, KBP 1521–1530)

Here it is simply a cat’s (Sambo) unusual behaviour of crawling under the door, glaring and shooting past Clarence that is in need of explanation. I read utterance 1530 as meaning ‘I don’t know, but maybe he acted that way because he’s never really been frightfully keen on me’, i.e. as the because relating back to utterance 1522 rather than being structurally related to the first clause of 1530 (I don’t know). The explanation offered by Clarence is one that makes potentially sense because the cat’s actions (e.g. glaring) can schematically be associated with an emotional disposition of dislike. For a different example consider (12): (12) AUDREY GORDON AUDREY GORDON AUDREY GORDON AUDREY GORDON AUDREY GORDON AUDREY

[191] D’ya know those were reduced from thirty one pounds to fourteen. [192] Were they? [193] [laughing] Yes [] . [194] Yeah! [195] Good buy that then. [196] So I thought, well i yo in a, in a presentation box [197] Yeah! [198] and everything. [199] It was a good [200] Er, I just think yo er, you know, you couldn’t be bit for that. [201] So while I was there [202] Couldn’t for that could you? [203] Well yo … you know when I looked round Gordon there was nothing, nothing less than that really. [204] And I thought well, it’s for an engagement present, just sort of two

Cognitive and discursive perspectives   417

GORDON [205] Yes, it’s ideal. AUDREY [206] it’s er ideal, and they like wine … so GORDON [207] Yeah. AUDREY [208] I thought it would be er … ideal. (BNC, KBC 191–208) This extract is all about Gordon and Audrey evaluating (and justifying the buy of) an engagement present, and the emotion talk is only part of a whole list of causes why this present and the act of buying it should be evaluated positively (good, buy, good, ideal) (implying the givers’ emotional satisfaction). While some of the causes relate to more material considerations (reduced price, there was nothing less expensive) or to standards that a present should fulfil (in a presentation box and everything) from the givers’ (Gordon and Audrey) point of view, the explicit emotion talk (they like wine) makes reference to the donors’ emotional satisfaction. This is one reason why such justifications are interesting; they contribute to speakers’ construal of themselves as a particular identity, personality or persona. For instance, it is possible to contrast this extract with one in which speakers only make reference to their material considerations or only to the donors’ emotional satisfaction. A special sub-class of affect, terms of desire or Dis/inclination (Martin and White 2005: 48), can also be part of a chain of reasoning. Terms of desire such as want/would like to are clearly important with reference to goals, which are vital for the regulation of human life and well-being, as social cognitive psychologists have shown. Crucially, “human activity is planful and is directed toward desired ends and the avoidance of unwanted ends.” (Barone, Maddux and Snyder 1997: 247). This means that what contributes to desired ends is judged positively: (13) BARRY

[1953] So it’s a … it’s a nice place cos it’s like … it … they do … a great spread of … I mean you can go for like a nice meal there, but you can like also … eat, you know … dips and […] ALAN [1954] Yeah. BARRY [1955] hamburger ALAN [1956] Yeah. BARRY [1957] or whatever, a big steak or whatever you want and the ALAN [1958] And how do you cook it? (BNC, KBD 1952–1968) Similar to example (12) above, emotion talk (want) in (13) is only part of a chain of reasoning given to justify a subjective judgment (it’s a nice place). On

418   Monika Bednarek the one hand, we can find explicit positive emotional talk involving subjective lexis (a great spread, a nice meal), on the other hand, this is complemented by a more indirect evaluation in terms of the fulfilment of the Emoter’s desire (you can eat whatever you want). Using the notion of schemata, we can say that features such as good and plentiful food (great spread, nice meal, big steak) and a wide choice (and you can also eat dips and hamburger or whatever, a big steak or whatever you want) are discursively construed as part of what makes up a good place to eat (part of the speaker’s schema for ‘good place to eat’), on account of the causal structuring (cos). At the same time, the discourse makes sense, is coherent, because these are all features that would in general be expected of such a place (part of a shared schema for ‘good place to eat’). A related example, which does not involve justification but also shows the importance of desire terms for evaluation, is (14): (14) CLARENCE NINA CLARENCE

[1233] It’s not too bad finding [cough] where you want to go [1234] Mm. [1235] when it comes to looking for x-ray or path lab or whatever it is … but finding your way out. [1236] … Very seldom they have continuing … er continuity [cough] of markers saying exit, you know, or this way out. NINA [1237] Wouldn’t it, it’s stupid that (BNC, KBP 1233–1237) Again, human desires and goals turn out to be important reference points in evaluation. What is positive about the hospital is that it allows visitors to achieve their goals in finding their way around (It’s not too bad finding where you want to go); what is negative about it is that the infrequent sign posting makes it more difficult for visitors to achieve their goals in finding their way out. This is also retrospectively and explicitly framed by Nina in terms of negative Judgement (stupid as indicating negative capacity). On prospective and retrospective Appraisal see Lemke (1998) and Hood (2006). Importantly, affect terms can also be used to soften a demand by framing it in terms of emotional experience. They can therefore occur in the context of deontic modality (modulation), as in (15) below: (15) PS000

[7331] his leg, so erm, he [Trevor from the pub]’s been moved to Royal hospital, I went to see him yesterday and he didn’t stop talking about you, bless him [laugh]

Cognitive and discursive perspectives   419

[23 utterances] [7355] So, er he put me onto it, oh he said you’ve got to go down, he’s missing you all, he’s, and he’d love a bit of [sic] (BNC, KD8 7331–7355)

PS000

Trevor’s demand (reported by PS000) that ‘you’ must come visit him in the hospital is framed in terms of Trevor’s positive affect towards the addressees of his demand. This can be explained with recourse to two ‘classic’ politeness theories: In Brown and Levinson’s (1994) terminology, we can interpret Trevor’s demand (you’ve got to go down) as a face threatening act, which threatens his addressees’ negative face (their desire to be free from imposition). Using affect softens this face threatening act by help of a positive politeness strategy, which claims in-group membership, and creates solidarity and emotional bonding between Trevor and his addressees, fulfilling their need to be liked and cared about. In Leech’s (1991) terminology, it is also possible to say that the directness (you’ve got to) of the impositive (demand) is justified by making reference to the social closeness between Trevor and his addressees (since Leech’s Tact Maxim proposes that the greater the distance between speakers, the greater the need for indirectness, and vice versa). Using Martin’s (1992: 533) metaphor of affect as a stereo system, Trevor ‘turns on’ the affect between him and his addressees in order to mitigate his demand. Example (16) below also involves deontic modality, but in this example affect is used to help justify a subjective judgment of obligation (realized by deontic modality/modulation) that impacts on the speaker only: (16) AUDREY

[310] I saw a pair of shoes, if my grey shoes don’t go with this dress, I’ve seen a pair of shoes that I like. [311] You know. [312] Well I have to, I, I’ll have to have Elaine with me about the hat because er … I prefer somebody else with me (BNC, KBC 310–312)

What Audrey feels she has to justify in example (16) is her statement of having to have Elaine with her in order to be able to choose a hat. In Halliday and Matthiessen’s (2004) terms, have to is an implicit and subjective modulation of a high degree of force. This means that it leaves implicit the source of the obligation but makes a subjective rather than objective statement (on notions of subjectivity in a cognitive-pragmatic framework, which are not incompatible with the systemic functional linguistic framework, see Nutys 2001). This

420   Monika Bednarek is perhaps why such a statement demands some sort of explanation, and why Audrey struggles to provide this, as evidenced by the hesitation phenomenon (er) and the pause (…). In fact, Audrey does not really give any kind of ‘rational’ justification, but rather makes her statement even more subjective, marking it as private and personal affect, referring to a general emotional disposition (I prefer). The utterance thus moves from subjective obligation to subjective emotion, which is perhaps less easily challengeable or demands no rational explanation according to the prevalent Western conceptualization of emotion as opposed to rationality (Bednarek 2008).8 However, emotional experience sometimes does seem to be followed up by causal explanation, as becomes evidenced in examples (17) to (20) below where affect is justified. (17) BARRY

[1979] I remember once, you know, getting and cutting … a rabbit and things, by the time you’d actually cut the thi damn thing [1980] You’ve gone off it. ALAN BARRY [1981] you’d gone off it cos you’d been … heading, and gutting it, and [laughing] whatever else [] ! [1982] And you know, you’re sat there and you think [1983] Yeah. ALAN BARRY [1984] I don’t [laughing] really wanna eat this, you know []! [1985] It’s alright if it comes to you in a ca in a, in a, as a, as a meal ALAN [1986] Comes up ready-made. BARRY [1987] Yeah. [1988] But er, when you actually see what you’re doing to it, you know. (BNC, KBD 1979–1988)

8 Examples where the ‘obligation’ meaning of have to seems to shade into a ‘desire’ meaning are common, e.g.: He [Anton Assaad] fell in love with the 1959 Arne Vodder sideboard on a buying trip to Denmark last year. “I saw it in the back of a warehouse in the countryside and I had to have it,” he says. (Sydney Morning Herald Good Weekend magazine, 28 April 2007). A definition for one meaning of have to as ‘used to say that you want something to happen’ could thus be added to the already existing definitions in dictionaries such as the Oxford Advanced Learner’s Dictionary (7th ed): ‘used to show that you must do sth’, ‘used to give advice or recommend sth’, ‘used to say that sth must be true or must happen’, ‘used to suggest that an annoying event happens in order to annoy you, or that sb does sth in order to annoy you’.

Cognitive and discursive perspectives   421

In (17) we are dealing with what might be called generalized affect, with you as Emoter referring “to people in general, including the speaker/writer” (Biber et al. 1999: 330). Affective meanings in this extract include emotion talk and emotional talk that imply negativity (the damn thing, you’ve gone off it, I don’t really wanna eat this) and emotional talk that signals positivity (alright). Discursively, two contrasting (compare e.g. the use of but) schemata for eating animals are negotiated by Alan and Barry, with positive and negative affective meanings communally attached to animal-food preparation and processing: getting and cutting heading and gutting it you actually see what you’re doing to it NEGATIVE (gone off it, don’t really wanna eat this)

it comes to you as a meal ready-made POSITIVE (it’s alright, Yeah)

In (18) below, it is non-volition (refusal) that has to be explained: (18) NINA

[1271] when Carolyn was working there they had this intruder. CLARENCE [1272] Well they’re often having intruders, […] go on … NINA [1273] I can’t remember all the details. [1274] There was quite a to-do about it in the paper. CLARENCE [1275] Mm. … NINA [1276] I, I know Carolyn said that they … in the end CLARENCE [1277] Yeah. NINA [1278] they refused to work on their own CLARENCE [1279] Oh that’s right I remember. [1280] Yes yes. [1281] Yeah. NINA [1282] because they were so way out. [1283] There was no way CLARENCE [1284] Mm. NINA [1285] that they could have summoned help or anything else. (BNC, KBP 1271–1285) It has been suggested that refused acting – instances where actors are described as “unwilling to do something (and consequently do not do it)” (Bednarek 2006a: 175) – can potentially trigger negative evaluation, even though this does

422   Monika Bednarek not always happen (Bednarek 2006a: 178). In (18) above, no negative evaluation is triggered because the refusal to act is justified by reference to an emergency situation (they had this intruder; they were so way out; there was no way that they could have summoned help). Other examples involve a justification of non-authorial affect, i.e. where the speaker is not the Emoter.9 In Western culture, emotions are typically construed as private states (Bednarek 2008), and have a special status in that it is not easy to challenge or attribute them to other people. It is thus probable that projecting emotional experience onto others (what Fiehler (1990: 132) has called “projektive Erlebensthematisierung” [‘a thematization of projected experience’] frequently has to be mitigated or justified in some way. Let us look at an example from the corpus: (19) NONE [2105] But she wouldn’t like tonight because BARRY [2106] No. NONE [2107] she’s not into, what I call is it heavy metal? [2108] Or summat like BARRY [2109] Yeah. NONE [2110] that, heavy rock BARRY [2111] Yeah. NONE [2112] or something? BARRY [2113] Heavy rock. NONE [2114] Well, erm, she’s not into that but BARRY [2115] Yeah. NONE [2116] I said we’d bring her. [2117] She’ll like it I think, she can dance. (BNC, KBD 2105–2113) In (19) a projection of a momentary (non-authorial) emotional state (she wouldn’t like tonight) is justified in terms of a general emotional disposition (she’s not into) that the speaker is assumed to have insight into. Further below, this is qualified (She’ll like it), and this time the momentary emotional state is 9 Strictly speaking, example (18) above also involves non-authorial affect. However, there is no reason for Nina to provide reasons for attributing a mental state to someone other than herself, since, on the one hand, it has explicitly been reported by Carolyn (Carolyn said that they … they refused to work on their own), and, on the other hand, refuse to denotes a linguistic action that only implies non-volition, rather than denoting non-volition itself.

Cognitive and discursive perspectives   423

justified in terms of a general capacity (she can dance). Perhaps this is part of a general patterning of reasoning where the particular (the instance) is explained by reference to the general (the system), such as making sense of an Emoter’s transitory emotional experience (‘instance’) in terms of their emotional disposition (‘system’) or making sense of a person’s behaviour in terms of their character. As a final example from the corpus, consider (20): (20) MERIELLE

[7384] I’ve got, I’ve got to go and have a word with Rob about something. MARTINE [7385] You never liked to do the jobs I say because they’re always pooey jobs MERIELLE [7386] What, what is it? MARTINE [7387] [laugh] Calculating a percentage of heavy goods [laugh] and MARTINE [7388] no, I might try and get that done MERIELLE [7389] that’s a pooey job (BNC, KD8 7384–7389)

In (20) affect is directly and explicitly attributed by a speaker to her hearer, a potentially challenging attribution in terms of interpersonal dynamics. The causal conjunction (because) does not relate to the speaker’s reasons for attributing an emotional experience to someone other than herself (as in (19) above), but gives reasons for the Emoter’s emotions. In other words, the challenging nature of the attribution (You never liked to do the jobs I say) is diminished by Martine’s saying that there are good reasons for Merielle to feel this way (they’re always pooey jobs). Again, this makes sense in terms of our schemata for emotions which associate negative evaluations with negative emotions. That the exchange is construed by the participants as non-challenging and non-threatening becomes apparent by laughter and “involvement” (Martin and White 2005: 35) resources (such as the non-standard pooey) that work to create solidarity and bonding. By way of a summary, Table 2 gives an overview of the role of affect in justification in casual conversation. This research would need to be complemented by large-scale studies into affect and justification. Areas of interest include the difference between authorial and non-authorial affect where affect is part of the justification; further roles that affect plays in reasoning and the kind of utterances that are justified (whether affect is part of the justification or whether affect is justified).

424   Monika Bednarek Table 2: Affect in justification Examples Role of affect in justification Affect as part 10, 11 Affect justifyof ing non-standjustification ard behaviour

What is justi- Justification fied we always sleep if I wake up in the together night and my sister’s not there I’ll think someone’s taken her or she’s run away, I’ll be frightened. he just crawled he’s never really been under the door frightfully keen on me … sort of glare, … glare at me and then scootled, shooted past 12, 13, 16 Affect bolster- it’s ideal they like wine ing subjective it’s a nice place you can eat whatever judgments you want I’ll have to have I prefer somebody Elaine with me else with me about the hat 15 Affect justify- you’ve got to go he’s missing you all ing/ softening down demand

Justifying affect authorial

17 18

non-authorial

19

20

Justifying you’d gone off it generalized affect Justifying non- they refused to volition work on their own

you’d been … heading, and gutting it

Justifying speaker’s assumption of Emoter’s emotion Justifying Emoter’s emotion

she wouldn’t like tonight

she’s not into, what I call is it heavy metal

You never liked to do the jobs I say

they’re always pooey jobs

they had this intruder; they were so way out; there was no way that they could have summoned help

Cognitive and discursive perspectives   425

Table 2 clearly shows that affect has an important role to play in cause-effect relations or justification. This supports a hypothesis made by Painter regarding mother-child talk that “apparently impersonal areas such as causal relations and generalizations arise initially from the impetus to share ‘attitude’” (Painter 2003: 183, italics in original). It also confirms claims made by the philosopher Errol Bedford in the 1950s that affect terms “explain [behaviour] by giving the reason for an action, in the sense of giving justification for it.” (Bedford 1956/57: 303) and backs up findings in discursive psychology on emotion discourse and rational accountability in relationship counselling sessions (Edwards 1999). 7. Concluding comments One of the objectives of studies in social cognition is to examine “how humans understand themselves and others” (Piswha, call for papers). In this paper I have shown that affect references are crucial in casual conversation for construing social reality, with a particular focus on the role of affect in justification. Affect allows speakers to provide reasoning for unusual behaviour, bolster justifications and soften demands, providing understanding for social actions. On the other hand, affect itself sometimes has to be justified. To some extent this paper also tried to combine cognitive and discursive perspectives into emotion talk and emotional talk. It is only recently that linguists have started to discuss connections between corpus and other functional approaches and cognitive research (Butler 2004; Gonzálvez-García and Butler 2006; Stubbs 2006). More research in this area is desirable if the aim of linguistic research is to account for language both as a cognitive and as a social phenomenon. Appendix This appendix provides further information on the files used in this paper (from the BNC), such as sample type, title, speakers etc. File AC2: End sample containing about 37622 words from Man at the Sharp End. Kilby, M, The Book Guild Ltd, Lewes, East Sussex (1991). File ADF: Beginning sample containing about 36847 words from Understanding Horses. Langley, Garda, David, Charles Publishers plc, Newton Abbot, Devon (1989). File A7H:

426   Monika Bednarek Beginning sample containing about 38700 words from Charles and Diana. Junor, Penny, Headline Book Publishing plc, London (1991). File AC6: End sample containing about 41199 words from A Poet Could Not But Be Gay. Kirkup, James, Peter Owen Publishers, London (1991). File ADG: Whole text sample containing about 33828 words from How do I Look?. Dawson, Jill, Virago Press Ltd, London (1990). File CH2: Unknown sample containing about 187088 words from The Daily Mirror. Mirror Group Newspapers, London (1992-08/1992-10). File KB9: 44 conversations recorded by ‘Annette’ (PS1CX) in 1992 Speakers: ‘Annette’, 44, administrative assistant, Lancashire, C1, female ‘Teresa’, 19, stable hand, Lancashire, C1, female ‘David’, 26, engineer, Lancashire, male ‘Tracy’, 24, housewife, Lancashire, female ‘Donald’, 71, retired, Northern England, male ‘Pat’, 48+, Lancashire, female File KBC: 14 conversations recorded by ‘Audrey’ (PS1A9) between 2 and 9 April 1992 Speakers ‘Audrey’, 61, housewife, Lancashire, AB, female ‘Gordon’, 61, teacher, Lancashire, AB, male ‘Margaret’, 45, nurse, Lancashire, female ‘Joan’, 50+, clerk, Central Northern England, female ‘Kevin’, 29, computer engineer, Northern England, male ‘Carl’, 31, pharmacist, Northern England, AB, male ‘None’ ‘Elaine’, 28, housewife, Northern England, female ‘Iris’, 60, housewife, Lancashire, female File KBP: 15 conversations recorded by ‘Clarence’ (PS065) between 13 and 19 March 1992 Speakers: ‘Clarence’, 65, retired, Lancashire, DE, male ‘Nina’, 67, retired, Lancashire, DE, female ‘Nev’, 72, retired, North-east Midlands, male ‘Lil’, 70, retired, Lancashire, female File KBD: 24 conversations recorded by ‘Barry’ (PS03W) between 1 and 6 February 1992 Speakers:

Cognitive and discursive perspectives   427 ‘Barry’, 41, entertainments consultant, Central Northern England, AB, male ‘Terri’, 35, bar staff, Home Counties, female ‘Hugh’, 30, bar staff, Irish, male ‘Alan’, 38, security, Lancashire, male ‘None’ ‘Mark’, 30, dj, London, male ‘Ken’, 30, security, Lancashire, male ‘None’, 35, housewife, Lancashire, female ‘Sergio’, 9, student (state primary), Lancashire, AB, male File KD8: 31 conversations recorded by ‘Martine’ (PS0LK) between 12 and 20 March Speakers: ‘Martine’, 25, senior technician, Welsh, C1, female ‘Mike’, 28, construction worker, Welsh, C1, male ‘Merielle’, 55, housewife, Welsh, female ‘None’, 45, pub landlord, Home Counties, male ‘Harold’, 58, engineer, Welsh, male ‘Nora’, 76, housewife, Welsh, female ‘Will’, 45, civil engineer, Merseyside, male ‘Michael’, 40, technical director, Home Counties, male ‘Jim’, 27, technician, Home Counties

References Anderson, Kristin J., and Campbell Leaper 1998 Emotion talk between same- and mixed-gender friends. Form and function. Journal of Language and Social Psychology 17: 419–448. Bamberg, Michael 1997a Language, concepts, and emotions: The role of language in the construction of emotions. http://www.massey.ac.nz/~alock/virtual/ bamberg.htm (last accessed 16 August 2006). Bamberg, Michael (ed.) Oral Versions of Personal Experience: Three Decades of Narrative 1997b Analysis. Special issue of Journal of Narrative and Life History 7. Barone, David, James Maddux and C.R. Snyder 1997 Social Cognitive Psychology: History and Current Domains. New York: Plenum Press. Bedford, Errol 1956/57 Emotions and statements about them. Proceedings of the Aristotelian Society LVII: 77–89. Bednarek, Monika Frames revisited: The coherence-inducing function of frames. Journal 2005a of Pragmatics 37: 685–705.

428   Monika Bednarek Construing the world: Conceptual metaphors and event construals in news stories. Metaphorik.de 09/2005. http://www.metaphorik.de/09/ bednarek.pdf (last accessed 15 January 2007). 2006a Evaluation in Media Discourse. Analysis of a Newspaper Corpus. London/New York: Continuum (Research in Corpus and Discourse). 2006b Subjectivity and cognition: Inscribing, evoking and provoking opinion. In Language and Memory: Aspects of Knowledge Representations, Hanna Pishwa (ed.), 187–221. Berlin: Mouton de Gruyter. 2008 Emotion Talk across Corpora. Houndmills: Palgrave Macmillan. Bednarek, Monika, and Wolfram Bublitz 2006 Reported speech: Pragmatic aspects. In Encyclopedia of Language and Linguistics, Keith Brown (ed.), 550–553. 2d ed. Oxford: Elsevier. Bellelli, G. 1995 Knowing and labelling emotions. In Everyday Conceptions of Emotion: An Introduction to the Psychology, Anthropology and Linguistics of Emotion, James A. Russell, Jose Miguel Fernández-Dols, Anthony S. Manstead, and Jane C. Wellenkamp (eds.), 491–504. Dordrecht: Kluwer Academic Publishers. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward Finegan Longman Grammar of Spoken and Written English. London: Longman. 1999 Brown, Penelope, and Levinson, Stephen Politeness: Some Universals in Language Usage. Cambridge: Cam1994 bridge University Press. Butler, Christopher Corpus studies and functional linguistic theories. Functions of Lan2004 guage 11: 147–186. Caffi, Claudia, and Richard W. Janney 1994 Toward a pragmatics of emotive communication. Journal of Pragmatics 22: 325–373. Cortazzi, Martin, and Lixian Jin 2000 Evaluating evaluation in narrative. In Evaluation in Text: Authorial Stance and the Construction of Discourse, Susan Hunston and Geoff Thompson (eds.), 102–120. Oxford: Oxford University Press. Downes, William 2000 The language of felt experience: Emotional, evaluative and intuitive. Language and Literature 9: 99–121. Edwards, Derek Emotion discourse. Culture & Psychology 5 (3): 271–291. 1999 Edwards, Derek and Jonathan Potter Discursive Psychology. London: Sage. 1992 Ekman, Paul Should we call it expression or communication? http://www.paulekman.com/pdfs/should_we_call_it_exp.pdf (last accessed 14 July 2006). [Published in Innovations in Social Science Research 10 (4): 333–344.] 2005b

Cognitive and discursive perspectives   429 Ellsworth, Phoebe C., and Klaus Scherer Appraisal processes in emotion. In Handbook of Affective Sciences, 2003 Richard J. Davidson, Klaus R. Scherer, and H. Hill Goldsmith (eds.), 572–595. Oxford: Oxford University Press. Fiehler, Reinhard 1990 Kommunikation und Emotion. Berlin: Mouton de Gruyter. 2002 How to do emotions with words: Emotionality in conversation. In The Verbal Communication of Emotions: Interdisciplinary Perspectives, Susan R. Fussell (ed.), 79–106. Mahwah N.J.: L. Erlbaum. Frijda, Nico H., Suprapti Markam, Kaori Sato, and Reinout Wiers 1995 Emotions and emotion words. In Everyday Conceptions of Emotion. An Introduction to the Psychology, Anthropology and Linguistics of Emotion, James A. Russell, Jose Miguel Fernández-Dols, Anthony S. Manstead, and Jane C. Wellenkamp (eds.), 121–143. Dordrecht etc: Kluwer Academic Publishers. Galasiński, Darius 2004 Men and the Language of Emotions. Houndmills: Palgrave Macmillan. Gonzálvez-García, Francisco, and Christopher S. Butler Mapping functional-cognitive space. Annual Review of Cognitive Lin2006 guistics 4: 39–96. Gottman, John, Lynn Fainsilber Katz, and Carole Hooven 1996 Parental meta-emotion philosophy and the emotional life of families: Theoretical models and preliminary data. Journal of Family Psychology 10 (3): 243–268. Halliday, M.A.K., and Christian M.I.M. Matthiessen 2004 An Introduction to Functional Grammar. 3d ed. London: Arnold. Heelas, Paul Emotion talk across cultures. In The Social Construction of Emotions, 1986 Rom Harré, (ed.), 234–266. Oxford: Basil Blackwell. Hoey, Michael 2005 Lexical Priming: A New Theory of Words and Language. New York: Routledge. Hood, Susan The persuasive power of prosodies: Radiating values in academic writ2006 ing. Journal of English for Academic Purposes 5: 37–49. Hood, Susan, and Jim R. Martin Invoking attitude: The play of graduation in appraising discourse. In 2007 Continuing Discourse on Language, Ruqaiya Hasan, Christian M.I.M. Matthiessen, and Jonathan Webster (eds.), 739–764. London: Equinox. Jahr, Silke 2000 Emotionen und Emotionsstrukturen in Sachtexten. Berlin/New York: Walter de Gruyter. Johnson-Laird, Philip N., and Keith Oatley The language of emotions: An analysis of a semantic field. Cognition 1989 and Emotion 3: 81–123.

430   Monika Bednarek Kövecses, Zoltan Metaphor and Emotion: Language, Culture, and Body in Human Feel2000 ing. Cambridge: Cambridge University Press. Labov, William Language in the Inner City. Philadelphia: University of Pennsylvania 1972 Press. Leech, Geoffrey 1991 Principles of Pragmatics. London: Longman. Leech, Geoffrey, and Mick Short 1981 Style in Fiction. London: Longman Lemke, Jay L. 1998 Resources for attitudinal meaning: Evaluative orientations in text semantics. Functions of Language 5 (1): 33–56. Martin, Jim R. 1992 English Text: System and Structure. Amsterdam/Philadelphia: John Benjamins. 2002 Blessed are the peacemakers: Reconciliation and evaluation. In Research and Practice in Professional Discourse, Chris Candlin (ed.), 187–227. Hong Kong: City University of Hong Kong Press. Martin, Jim R., and David Rose 2003 Working with Discourse: Meaning beyond the Clause. London: Continuum. Martin, Jim R., and Peter R.R. White The Language of Evaluation: Appraisal in English. London/New York: 2005 Palgrave/Macmillan. Mees, Ulrich Zum Forschungsstand der Emotionspsychologie – eine Skizze. http:// 2006 www.psychologie.unioldenburg.de/mub/Zum%20Forschungsstand%20 der%20Emotions-psychologie.pdf (last accessed 13 July 2006). [Published in Emotionen und Sozialtheorie. Disziplinäre Ansätze, R. Schützreichel (ed.), 104–124. Frankfurt: Campus.] Nuyts, Jan 2001 Subjectivity as an evidential dimension in epistemic modal expressions. Journal of Pragmatics 33 (3): 383–400. Oatley, Keith, Dacher Keltner, and Jennifer Jenkins 2006 Understanding Emotions. 2d ed. Oxford: Blackwell. Ortony, Andrew, Gerald L. Clore, and Allan Collins The Cognitive Structure of Emotions. Cambridge: Cambridge Univer1988 sity Press. Oxford Advanced Learner’s Dictionary. 7th ed. Oxford: Oxford University Press. (used on CD-ROM: Oxford Advanced Learner’s Compass) Painter, Claire 2003 Developing attitude: An ontogenetic perspective on appraisal. Text 23 (2): 183–209.

Cognitive and discursive perspectives   431 Rosenberg, Daniel 1990 Language in the discourse of the emotions. In Language and the Politics of Emotion, Catherine Lutz and Lila Abu-Lughod (eds.), 162–185. Cambridge: CUP. Russell, James In defense of a prototype approach to emotion concepts. Journal of Per1991 sonality and Social Psychology 60: 37–47. Schrauf, Robert W. and Julia Sanchez The preponderance of negative emotion words across generations and 2004 across cultures. Journal of Multilingual and Multicultural Development 25 (2–3): 266–284. Shaver, Phillip, Judith Schwartz, Donald Kirson, and Cary O’Connor 1987 Emotion knowledge: Further exploration of a prototype approach. Journal of Personality and Social Psychology 52: 1061–1086. Shriver, Lionel We Need to Talk About Kevin. Melbourne: The Text Publishing Com2006 pany. Stubbs, Michael Corpus analysis: the state of the art and three types of unanswered ques2006 tions. In System and Corpus: Exploring Connections, Geoff Thompson and Susan Hunston (eds.), 15–36. London: Equinox. Tannen, Deborah, and Cynthia Wallat 1993 Interactive frames and knowledge schemas in interaction: Examples from a medical examination/interview. In Framing in Discourse, Deborah Tannen (ed.), 57–76. Oxford: Oxford University Press. Turner, Jonathan, and Jan E. Stets The Sociology of Emotions. Cambridge: Cambridge University Press. 2005 Werth, Paul On the semantics and pragmatics of emotion verb complementation. In 1998 Speaking of Emotions: Conceptualisation and Expression, Angeliki Athanasiadou and Elzbieta Tabakowska (eds.), 409–440. Berlin: Mouton de Gruyter. White, Geoffrey 1990 Moral discourse and the rhetoric of emotions. In Language and the Politics of Emotion, Catherine Lutz and Lila Abu-Lughod (eds.), 46–68. Cambridge: Cambridge University Press. White, Peter R.R. 1997 Death, disruption and the moral order: The narrative impulse in massmedia ‘hard news’ reporting. In Genres and Institutions: Social Processes in the Workplace and School, Fran Christie and Jim Martin (eds.), 101–133. London: Cassell. 1998 Telling media tales: The news story as rhetoric. Unpublished PhD dissertation, University of Sydney, Sydney. Wowk, Maria T. 1989 Emotion talk. In Text and Talk as Social Practice, Brian Torode (ed.), 51–71. Dordrecht/Providence, RI: Foris Publications.

Chapter 16 From motion to emotion to interpersonal function: The case of fear predicates Eliza Kitis

All that emotion is, in many circumstances, is a particular form of communication. (Parkinson 1995: 170)

1. Introduction* The connection between feeling and thought, emotion1 and cognition, has been noticed and widely discussed in the literature. In philosophy this connection seems to have been well established for a very long time. According to Aristotle’s Poetics (19 ‘Thought’), under ‘thought’ fall all the effects that have to be deliberately and consciously achieved through the use of speech. Elements of this endeavor are (1) proof and refutation and (2) the stimulation of feelings such as pity, fear, anger, and the like. On the other hand, Hume (1962: 77) writes: It seems a proposition which will not admit of much dispute that all our ideas are nothing but copies of our impressions, or, in other words, that it is impossible for us to think of anything which we have not antecedently felt, either by our external or internal senses.

Hume almost puts his finger on the imprint of the scars but misses the deeper source of thinking, that of concrete perception and expression. But he does trace ideas, that is thinking, to impressions, that is feeling. In this study, I will provide linguistic evidence that points to the priority of external behavioural * I thank Maria Theodoropoulou and Alexandros Tantos for useful comments, as well as Karin Boklund for checking the translations from OE and ME. 1 For a discussion on the distinction between emotion and feeling, see Wierzbicka (1999).

434   Eliza Kitis aspects of presumed emotion that lead through their perception and interpretation to the attribution of cognitive aspects of emotion. Such behavioural aspects are metonymically connected to assumed emotional states, which in their turn are metonymically used in social situations at an interpersonal level functioning as mitigation devices, amongst other uses, or, more generally, serving a variety of social functions at a presumed expressive rather than propositional level of communication. I will demonstrate this route of semantic evolution by examining two emotion verbs in two languages, English and Greek. I will focus on the Greek psych-verb fovame and its translational equivalents fear/be afraid.2 I will claim firstly that, not only mental verbs in the broad sense, but also the special class of psych-verbs can be shown to contribute to the class of speech act verbs. Secondly, additionally to Vendler’s (1972) claim that there is close affinity between mental and, in particular, thought verbs and speech act verbs, I will show that the category of psychological verbs, or the domain of emotions, has been a source for the development of propositional attitude and speech act verbs, and interpersonal meanings. Moreover, contrary to Vendler’s claim that there is a two way leakage between speech act verbs and mental verbs, I will show, by tracing back the diachronic evolution of both fovame and fear/be afraid, that the leakage goes in one direction only, from emotions to propositional attitude and speech acts, and not the other way round or both ways. At a crosslinguistic level, this study is intended to add further evidence to Traugott’s (1989) thesis that this directionality, from emotions to performativity and interpersonal meaning, is consistent with a universal tendency in language to move from propositional meanings via expressivity to interpersonal ones, since emotions are represented in propositional form while performativity involves propositional attitude and interpersonal meanings. It will also emerge from an examination of the origins of both fear verbs that their source domain is spatial rather than emotional. In both the cases of fovame and fear/be afraid there has been a shift from concrete spatial source domains to the more abstract domain of emotions. Our diachronic itinerary is expected to partially explain the syntactic patterns into which fovame enters. Therefore, let us first concentrate on the Modern Greek (MG) predicate fovame. 2 Let it be noted that I regard as translational counterparts those terms which occur more frequently or readily as their translational equivalents and which appear first in dictionaries: fear = fovos, tromos, etc.; afraid = fovismenos, etc., be afraid = fovame (Penguin-Hellenews English-Greek Dictionary, 1975); fovos = fear, fright, dread, anxiety; fovume = be afraid, be fearful, be/stand in fear of, etc. (Stavropoulos, 1988, Oxford Greek-English Learner’s Dictionary, OUP). There are two morphological versions of the predicate: fovame and fovume, the latter being closer to its Ancient Greek equivalent form.

The case of ‘fear’ predicates   435

2. The case of fovame The following examples illustrate the main constructions of the predicate fovame in MG. (1) Fovate ala apokalipti. is afraid-3SG but reveal-3SG ‘He is afraid/scared but reveals.’ (newspaper title) (2) Pjios fovate ton Lenin? Who is afraid-3SG the-ACC-SG Lenin? ‘Who’ s afraid of Lenin?’ (newspaper title) (3) Fovame neo ktipima. Am afraid/fear-1SG new blow ‘I fear a new blow.’ (newspaper title) (4) Fenonte na mi fovunte tus paparatsi seem-3PL to not be afraid-3PL the-ACC-PL paparazzi ‘They seem not to be afraid of the paparazzi.’ (5) Fovame na su po perisotera. Am afraid-1SG to you tell more ‘I’m afraid to tell you more.’ (6) Foviθike apo ton sismo. Was afraid-3SG of/from the-ACC earthquake. ‘He got frightened/scared by/of/at the earthquake.’ pu o filos tu piastike. (7) Foviθike Was afraid-3SG that the friend his was arrested-3SG ‘He was afraid/frightened when his friend was arrested.’ (8) Fovame oti δen θa pliroθume. Am afraid-1SG that not will(partcl.) be paid-1PL ‘I’m afraid that we won’t be paid.’ (9) Fovame oti/pos δen mporo na apokalipso Am afraid-1SG that not can to reveal-1SG to onoma. the-ACC name ‘I’m afraid that I can’t reveal the name.’ (10) Fovame min/mipos apotiho ksana. Fear-1SG that(lest) fail-1SG again. ‘I’m afraid/fear that I might fail again.’ (11) Fovame mu ine aδinato na ti Am afraid-1SG me-GEN is-3SG impossible to it δehto. accept-1SG

436   Eliza Kitis ‘I’m afraid I can’t accept it (an application).’ (12) Mu .δehto, ine aδinato na ti Me-GEN is-3SG impossible to it accept-1SG, poli fovame. very much am afraid. ‘I can’t accept it, I’ m afraid.’ (13) A: θa ’rθis sto parti? will(partcl.) come-2SG to the party? ‘Will you come to the party?’ B: Fovame (pos) ohi. am afraid-1SG (that) no. ‘I’m afraid no.’ As can be seen the syntactic pattern of the predicate fovame is quite wide ranging. Fovame as an one-place predicate admits its experiencer in subject position in the structure NP V, as in I Maria fovate ‘Maria is afraid’ or in (1). As a two-place predicate it can take an internal argument in the accusative NP V NP, as in (2), (3) and (4), or in a prepositional phrase (PP) expressing cause, as in (6). Moreover, it admits an oti ‘that’-complement, as in (8) and (9), or a min/ mipos ‘that/lest’-one as in (10), as well as a factive pu ‘that’-complement, as in (7). It also admits a na ‘to’-clause, as in (5) (cf. Kakouriotis and Kitis 1999; Theodoropoulou 2003). Vendler (1972) wanted to prove that the semantic organization of verbs of “speech and thought” is very similar. In what follows, I will demonstrate that his claim can be extended by showing that the semantic organization of verbs of “speech and emotions” is very similar, too. 2.1. The evolution of fovame First, I will try to substantiate the claim that there is only one direction in Vendler’s leakage, that from the source domain of space and motion to the target domain of emotions. Tracing back the history of fovame or fovume, which is a middle formation, we see that it originates from the Ancient Greek (AG) transitive predicate (it checks objective case) foveo [phobeô]3 which means ‘put to flight’ as in the following examples from Homer’s Iliad: 3 phobeô; aor. (e)phobêsa; mid. pres. part. phobeumenos; fut. phobêsomai; pass. aor. 3 pl. (e)phobêthen; perf. part. pephobêmenos; plup. 3 pl. pephobêato: act., ‘put to flight’, tina (=one)-3 sl.; mid. and pass., ‘flee, be put to flight’, hupo tinos or hupo tini

The case of ‘fear’ predicates   437

(14) tous d’ allous Danaous ephobêse Kroniôn. (Il. 11, 2.36) ‘the rest of the Danaans hath the son of Cronos scattered in flight’ (15) hos te kai alkimon andra phobei (Il. 16, 2.36) ‘for he [Zeus] driveth even a valiant man in rout’ In both (14) and (15) the predicate selects a nominal in object accusative: tous …allous Danaous, alkimon andra. The Passive and Middle phebomai [phobeomai] in Homer denoted a bodily activity and was always used in the sense of ‘to be put to flight’, ‘flee in terror’: (16) mega men kakon ai ke phebômai plêthun tarbêsas (Il. 11, 2.71) ‘[then what is to befall me?] Great evil were it if I flee seized with fear of the throng’ (17) tophra d’ Achaioi taphrôi kai skolopessin eniplêxantes oruktêi entha kai entha phebonto (Il. 15, 2.48) ‘meanwhile the Achaeans were flinging themselves into the digged trench and against the palisade, fleeing this and that way’ (18) opisô de pulas lipe, bê de phobêtheis: (Il. 22, 1.93) ‘but left the gates behind him, and fled in fear’ From this meaning of motion away from one point (due to imminent danger or compulsion) there develops the meaning of the emotion of fear as we know it now, as in (19), in which the motion (fleeing) is denoted separately (ôichonto) (but see also (18)): (19) phobêthentes hoi xenoi ôichonto (Aeschines 1.43) ‘having been frightened, the strangers fled’ And in Thucidides we encounter the transitive verb meaning ‘terrify’ checking an object accusative (autous) as in (20), and (Alkibiadên) as in (21), which is also interestingly followed by a telic complement clause (mê . . epagagôntai). (20) su de, Klearida, husteron, hotan eme horais êdê proskeimenon kai kata to eikos phobounta autous, tous meta seautou tous t’ Amphipolitas. (Th. 2.79) (=by one); tina. Transliteration key for AG (different from that for MG, the first sign is in AG, the second in transliteration, if a third, it is for pronunciation): β/b=v. φ/ ph=f, ω/ô =o (long), η/ê=i (long), υ=u, θ/th=th (as in ‘thing’), γ=g, δ=d, x=x, ψ=ps, c=ch, as in Perseus. Breathings are not reflected. Underlining in examples reflects exact lexical correspondences. Bold indicates corresponding fear predicates.

438   Eliza Kitis ‘and do you, Clearidas, afterwards, when you see me already upon them, and, as is likely, terrifying them, take with you the Amphipolitans’ (21) Alkibiadên ephoboun, mê kai,… tauta legôsin, epagagôntai to plêthos. (Th. 5.45) ‘[they] frightened Alcibiades lest, if they were to say these, there would be a crowd’ Middle morphology is apparently motivated by dethematizing the agent, a process common to both reflexive and middle formations, as noticed by Abraham (1995). The poetic use of the passive form of phobeô, phebomai [phobeomai] is used only in early texts with the passive meaning of ‘being put to flight’. While in the Iliad the transitive predicate with its motion or spatial meaning of ‘put to flight’ is very common, soon this form gives way to its middle formation as the agent gets dethematized and the semantic focus shifts on emotion aspects of meaning, that is, as the experiencer or affected becomes the focus of the discourse, as in the following example from Herodotus. The focus is on the barbarians as the experiencers and the theme of the discourse, but one can still detect the spatial meaning of the past participle pephobêmenoi (signifying that they were pushed into this small space) on top of the emotion meaning of fear. (22) hoi de barbaroi … aluktazon te hoia en oligôi chôrôi pephobêmenoi (Herod. 9.70.1) ‘but the barbarians … in distress in such a small space were pushed terrified’ So the middle formation of fovame seems to be predicted by Iwata (1995). As observed in there, “psych-verbs behave like ordinary verbs in several respects when the notions of external argument and direct internal argument are crucially involved. First, middle formation is generally taken to suppress the external argument and externalize the direct internal argument” (96). So we see that the meaning of fovame originates from the meaning of its transitive form denoting the spatial source domain of motion and it only later shifts to the domain of emotions: terrify, alarm, as in the following middle examples from Thucydides: (23) pantôn te ephobounto malista tous Lakedaimonious, hoti echontas ti ischuron autous enomizon ouketi sphisin epikêrukeuesthai (Th. 2.68) ‘out of all, however, they feared the Lacedaemonians most, as they must, it was thought [by the Athenians], feel themselves on strong ground not to send them any more envoys’

The case of ‘fear’ predicates   439

(24) phoboumenos de mê hoi pempomenoi … ou ta onta apangellôsin (Th. 2.10) ‘Fearing, however, that the messengers … might not report the truth’ It is interesting to note the transitive form of the verb in (20) blending both meanings denoting motion and emotion, while the middle formations in both (23) and (24) present a distinct evolution from emotional aspects of meaning to cognitive ones. The (quasi-) complementizers in both cases (23: hoti echontas, etc. -in fact a relative-causal clause-; 24: mê… apangellôsin) attest to this meaning development. Both structures are reflected in MG, the former by examples such as (2), (3) and (4) selecting object nominals that can also be followed by oti ‘that’-clauses or causal ones, the latter by examples such as (10). In later texts this semantic evolution of the predicate is stabilized. In (25) the predicate is constructed with an infinitival complement (phanai) marking a move to cognitive meanings signifying emotion (in some cases) but often doxastic views: (25) egô men gar phoboumai sophistas phanai. (Plato, Sophist 3.52) ‘for I fear/hesitate to declare (them) sophists’ In (25) the emotion meaning of phoboumai gives way to cognitive doxastic meanings and the verb functions as a mental or cognitive verb, rather than just an emotion one, signifying attitude towards the proposition of its complement clause. This structure is reflected in MG examples like (5). However, even though in MG this predicate never denotes motion or bodily movement, vestiges of this motion meaning in AG use are to be found in some of its current constructions. The middle predicate phoboumai in AG can check a prepositional phrase (PP) apo tinos ‘to be afraid of one’, just as in MG. According to L&S this co-occurrence is probably a Hebraism. In MG apo is encountered primarily (if not only) with the past perfective of fovume, fovithika, but Tzartzanos mentions expressions such as fovate apo ta skilia ‘s/he’s afraid of dogs’, and Thoedoropoulou (2003: 268) notes features such as [+specificity] and [–animacy] in this respect. Whatever the case may be, MG fovume in its past perfective form can be followed by the preposition apo ‘by/from’ signifying the cause of fear as in (6), repeated here: (6)

Foviθike apo ton sismo. Was scared-3SG of/from the-ACC earthquake. ‘He got frightened/scared by/of/at the earthquake.’

440   Eliza Kitis Followed by apo-PP, MG fovume always signifies emotion only, and, as this construction occurs overwhelmingly in the past perfective (foviθika, perfectives typically signify events rather than states), we can mark it as a structure reminiscent of the initial motion or spatial meaning of the AG (Homeric) predicate. Moreover, the past perfective signifies characteristics, [+specificity], [+factivity], consonant with those of action predicates such as the AG phobeô/phoboumai denoting bodily movement, while tracing the etymology and semantic evolution of the preposition apo ‘from’ will witness a similar route to that of phoboumai, i.e. its evolution from concrete spatial domains, signifying movement from a point, to abstract domains of causality (cf. Fraser 1987; Kitis 2009). Its etymology as a verb signifying action (motion) also seems to be implicated in the use of the imperative form or subjunctive forms in directive speech acts in later texts: (26) ton phobon autôn mê phobêthête (New Testament, Ep.Pet. 3.14) ‘do not keep away from/because of their fear’ (27) Phobou tous Danaous kai dôra pherontes (moto) ‘keep away from the Danaans even if they are bearing presents’ It is interesting to note that in (26) the predicate checks its cognate noun phobon in accusative form in object nominal position. The action meaning of the verb rather than its emotion one is implicated in imperative forms of the predicate, as you cannot ‘impere’ emotions but you can direct actions. In this case, the verb can admit a person accusative signifying person and can mean ‘stand in awe of’, ‘dread’: (28) phobou tous anô theous (Plato) ‘dread/stand in awe of the gods above’ Indeed, the New Testament is bristling with injunctions featuring this predicate in the sense of ‘revere’ and ‘stand in awe’, not to be found in its MG version. As AG phoboumai is de-semanticized as an action verb and re-semanticized as an emotion one, it soon comes to admit (quasi) complement clauses with an ho ti ‘that’ complementizer. In fact, these clauses are relative-causal, further elaborating the cause of fear always articulated in an object accusative nominal in internal thematic position. This development marks the predicate’s acquired potential to act as a cognitive verb, not so much expressing feelings or emotion, but as expressing (unwelcome) views. This further parallel development is exemplified in (29); in other instances, it directly admits a mê[n] ‘lest’-clause, without an internal thematic nominal, as exemplified in (24) or (30). This struc-

The case of ‘fear’ predicates   441

ture is very common in both early and later AG texts, as well as in the New Testament: (29) phoboumai tode, ho ti…(Th. 7.67). ‘I fear this, that…’ (30) kai hoi Athênaioi, … phoboumenoi mê sphisi dicha gignomenois rhaion machôntai, (Th. 2.10). ‘and the Athenians, fearing lest they get divided and so fighting at a disadvantage,’ Goodwin (1889) notes that complements of verbs of fearing, selecting mê or hopôs mê followed by the present subjunctive, refer to a future object of fear, and “may also denote what may hereafter prove to be an object of fear.” (II. [17]§ 92). This structure is reflected in MG examples such as (10) and, even though the verb expresses the emotion of fear, its function is not very dissimilar to that of cognitive verbs. What must be noted is that structures of the predicate admitting directly an hoti complement (as in MG example [9]) is a much later development as this structure does not occur in the New Testament either. In later texts, there is the occasional occurrence but then hoti introduces a causal clause, often after a comma, as in (31) from Xenophon: (31) ho de Armenios… kai to megiston, ephobeito, hoti ophthêsesthai emelle ta basileia oikodomein archomenos (Xenophon, Cyropaedia, 3.1.1) ‘but the Armenian [king] …was afraid most of all because (that) he saw that he was sure to be seen in the act of beginning to build his palace.’ But (31) does not feature the verb in first person singular, a structure promoting the doxastic (propositional attitude) aspect of meaning. Such structures were not found in AG texts, including Lycias’ and Isocrates’ texts (expected to have more rhetorical structures). As will be claimed further below, it is this structure that in MG has evolved to function as a ‘regret’ speech act verb at an interpersonal level. 3. The case of fear In this section, we will briefly examine the translational equivalent predicates in English, fear and afraid. Extending our diachronic approach to them, we

442   Eliza Kitis shall find that the basic configuration of their semantics owes a great deal to their etymology and the evolutionary aspects of their development. Let us first turn to the verb fear. As a noun, O(ld) S(axon) fâr meant ‘ambush’, ‘stratagem’, and ‘danger’. The OS verb fârôn meant ‘to lie in wait’, and in O(ld) H(igh) G(erman) fârên meant ‘to plot against’ (OED). Fear, just like the AG phobeo, originally is an active verb with an object as patient. The OE verb færan meant ‘to terrify’, ‘to take by surprise’. It has a transitive use admitting an object in internal thematic position. (32) is from the very early text of Middle English Ormulum: (32) He wĭle himm færenn. (Ormin 675, c 1200) ‘He will frighten him (drive him away)’ (33) His huntes to chace he commaunde,/ Here Bugles boldely for to blowe,/ To fere the beestis in þat launde. (The Sowdone of Babylone 57–59, c 1400). ‘He commanded his hunters to chase, and blow their horns strong and far to frighten (drive) away the beasts in that land’. But also much later (functioning as an interjection in this context): (34) “Where’s Miss Kitty… or gone to see somebody’s child with the measles, devil fear her! She has plenty on her hands to do anywhere but at home…” (Father’s angry talk about his daughter, Miss Kitty) (Charles J. Lever Lord. Kilgobbin xviii, 1872). Just like AG phobeo, that means ‘to put to flight’, fear also meant ‘to drive away by fear’, ‘to frighten away’, especially birds or animals (OED): (35) Eddres to sleyn and foules oute to fere is. (Palladius on Husbondrie 1.147, c 1420). ‘it is to destroy adders and to put to flight fowl’ (36) O Thou good ihesu, … fight strongly for me, & fere away the euyll bestes, that is to say my lecherous concupysseus, that I am moued & tempted by… (Atkynson tr. De Imitatione III.xxvii, 1504) ‘Oh, Good Jesus, … fight strongly for me and drive away the evil beasts, that is to say my lecherous carnal desire, that I’m moved and tempted by…’ (37) A scar-crow …to feare the Birds of prey. (Shakesp. Measure for M. II.i.2, 1602) (OED). ‘A scare-crow to frighten away the birds of prey’

The case of ‘fear’ predicates   443

It also meant ‘to deter from a course of conduct or action’ or to ‘to drive by fear’ (OED): (38) Eueriche busshope…sholde…Feden hem [hus peple]… and fere hem fro synne. (Langland P.Pl.C. xviii, 1393) ‘Every bishop should nourish them [his people] and stop them from sinning (keep them away from sin)’ (39) And it should somwhat touche them to be sene by werynes of pryson to feare him to it. (Foxe Acts & Monuments 788a, 1563). ‘And they should be concerned if they were to be seen that wariness of prison would drive him (the bishop) to it’ (to give an answer for fear of imprisonment) From these two motion and locative meanings, signifying bodily movement (‘to put to flight’, ‘to frighten’, ‘to drive away by fear’) evolved the meaning of emotion of both these verbs. It is interesting to note that, while in Greek a morphological Middle develops from phobeô, phoboumai, in English we witness the form fear developing into a reflexive form, meaning ‘to feel fear’, ‘to regard with fear’, ‘to be afraid’: (Both agentive fear and phobeô eventually become obsolete). The construction develops from the expletive structure witnessed in (40) to reflexive structures (I fear me) such as in (41) (42), expressing propositional attitude alongside the emotion of fear: (40) I, now symple and moost rude…It fereth me sore for to endyte;/ But at auenture I wyll now wryte (Stephen Hawes The Example of Vertu 8–14, 1503) ‘I, now humble and most ignorant … I very much fear to compose; But I will now venture to write …’ (41) I feared me always that it wolde be so. (Palsgr. 547/2, 1530). ‘I was afraid that it would be so’ (42) I fear me …some… earthly love mingles with his friendship (R.A. Vaughan Mystics I.167, 1856) (OED). ‘I fear that some earthly (coarse, material) love interferes with his friendship’ If we accept Meyer-Lee’s (2007: 180) commentary on this stanza of the lines in (40), as framing the poet’s self-derogation in seeking to position himself as a poet in the context of his predecessors, then it is likely that ‘It fereth me sore’ expresses propositional attitude rather than any emotion of fear. Furthermore, we notice the structure’s development to first person singular

444   Eliza Kitis present uses (I fear me) as in (42) – compare with (41) – a structure used to register propositional attitude and perform speech acts. Subsequent uses as a parenthetical, and indeed in medial position, is an unsurprising development, as in (43): (43) A flash, I fear me, that will strike my blossom dead. (Tennyson Lancelot & Elaine 966, 1859) (OED). What emerges from the examination of both these verbs (phobeô, fear) is that they have followed parallel routes in their evolutionary semantics and morphosyntactic patterns. It appears, however, that whereas in the synchronic use fear seems to be more dynamic, projecting into the future (with afraid taking over in more stative aspects of the emotion of fear -generic statements are expressed with afraid), the Greek fovame is both stative and dynamic according to its context and its aspectual properties. (cf. Kakouriotis and Kitis 1999; Tissari 2007). Afraid is the past participle of affray, meaning ‘alarmed from a previous state of peace’, affraien ‘to attack’, ‘invade a country’. (From Old French esfreer, ‘to disturb’, from Vulgar Latin exfredare, ‘to break the peace’, from ex-, ‘out’, ‘away’, AHD), while later it acquires the meaning of fear as in (45): (44) þe Kyng was alle affraied. (R. Brunne Chron. 16, 1330) (OED). ‘the king was wholly alarmed’ (45) Moses couered his face, for he was afrayed to loke vpon God. (Coverdale Ex. iii. 6, 1535) (OED). As it was used frequently in its participial form, this predicate soon acquired an independent status (16th c) (OED). Therefore, its stative fear meaning is accountable on the grounds of both its morphology (being a past participle) and its original meaning (locative). This predicate soon came to be used to express more subjective discourse-meanings. Tissari’s findings attest to a ‘I suspect’, ‘I regret to say’ use of the predicate in 1501, earlier than the analogous one recorded in OED. In the next section, we will see how subjective meanings of the emotion are due to processes of metonymization. These processes are also at work in pushing fear predicates to function further in the social domain of interpersonal transaction. We will see how initially observation or evidence of external behaviour is used to infer subjective meanings which ultimately are capitalized upon to perform specific speech acts.

The case of ‘fear’ predicates   445

4. Processes of metonymization Metonymy is a pervasive linguistic, conceptual and processual phenomenon (Gibbs 1999), enforced by instincts of economizing on resources, but it also marks the whole representational system of a language, since even writing is metonymic originating from drawing. An eye metonymically represents omniscient and omnipresent God, the drawing of the eye itself metonymizes God and all the qualities attributed to God, the concrete element of the physical domain standing for, or invoking, the abstract non-physical qualities. As Foucault (2002/1966: 122) writes: And it is by following the nervure laid down by these figures [synecdoche, metonymy, catachresis] that those languages paralleled with a symbolic form of writing will be able to evolve. They become endowed, little by little, with poetic powers; their primary nominations become the starting-points for long metaphors; these metaphors become progressively more complicated, and are soon so far from their points of origin that it is difficult to recall them. This is how superstitions arise whereby people believe that the sun is a crocodile, or that God is a great eye keeping watch on the world; it is also how esoteric forms of knowledge arise among those (the priests) who pass on the metaphors to their successors from generation to generation; and it is how allegorical discourse (so frequent in the most ancient literatures) comes into being, as well as the illusion that knowledge consists in understanding resemblances.

It is reasonable to assume that in the case at issue the fear meaning, as a later development, was inferred and arrived at on the basis of the external demonstrable behaviour, especially as inner psychological states are abstract and inaccessible. We are not concerned here with the debated issue of whether emotions are caused by their symptoms or vice versa. (cf. Damasio 2003), but we can assume that the verb denoting the external behaviour was metonymically used later to also attribute the presumed psychological state to the experiencer until it eventually came to denote it. The behaviour of fleeing away from something (cf. Kövecses 1990; Wierzbicka 1990, 1999 on the specification of fear), or being disturbed or alarmed from a previous peaceful state is endemic in corresponding psychological states or vice versa. We need not claim that there is a cause effect relationship, but we can assume that reference to the initial bodily action also at some later stage came to implicate the existence of the psychological state, probably initially conversationally and later this implicature was conventionalized (Grice 1989). In other words, reference to the concrete behaviour can serve as a bridge to the emotional state, and this bridge is what we call metonymy. Kövecses (2002: 148) explains the function of metonymy quite succinctly:

446   Eliza Kitis The main function of metonymy seems to be to provide mental, cognitive access to a target entity that is less readily or easily available; typically, a more concrete or salient vehicle entity is used to give or gain access to a more abstract or less salient target entity within the same domain.

Moreover, the metonymic relation in the case of the fear verbs examined here is that of contiguity as the bodily action is ordinarily considered a symptom of the caused (and causing) psychological state. So, as regards the derivation of the semantic organization of the predicates fovume, fear and afraid, we see the following schema being in full force: Source Domain (concrete, spatial) motion

Target Domain >

(abstract) emotion

Figure 1. The semantic evolution of fear predicates from motion to emotion

5. The evolution of interpersonal meaning It must be clear by now, that both fear verbs examined here have followed similar trajectories from the meaning of bodily activity (from motion), which is obsolete in both cases, to the meaning of emotion. However, as we have already seen, this trajectory has spanned further into the discursive space of social interaction, and the semantics of these predicates has extended further: these predicates (fovume, fear, be afraid) have been shown to shift from more semantically-based space to more pragmatic functions determined by discursive needs. Both the Greek fovume and its translational counterparts exhibit a high degree of variation in position within the clause. Already in its reflexive form fear occurs interclausally as a parenthetical as in (43), and later, shedding the reflexive form, as below: (46) The account …will hardly, I fear, render my letters very interesting. (F.A. Kemble Resid. in Georgia 16, 1863) (OED). However, this function seems to have been very largely taken over by the predicate be afraid for English, as it has become current since the 16th century. (Cf. Kakouriotis and Kitis 1999; Tissari 2007) On the other hand, as AG phoboumai acquires the meaning of fear-emotion and its constructional schema mutates into a middle formation, as we have seen, it admits an internal argument in the accusative often followed by an hoti relative-causal clause elaborating on the object of fear, as in (23). It is this

The case of ‘fear’ predicates   447

structure that later is assumed to give rise to the predicate’s selecting directly an oti ‘that’-clause as in (31)4, which occurs at an early stage. Both fovume and fear/be afraid drop their complementizers at some stage and gradually acquire potential for free movement within the formerly complement clause, which is naturally elevated to the status of the main clause as in (12) and (43). Thus, both fovume and fear/be afraid, having already undergone a process of desemanticization as we have seen, can be used as modalizers (11), highlighting the speaker’s attitude towards what s/he is saying as well as to couch negative predictive speech acts, as in (9) and (10). In this respect they seem to have followed a similar route as that of cognitive verbs such as think. It is evident that there is a very small step from this state to the state of performativity. If a condition for speech acting is a certain state of mind, we can assume, by extending this condition, that a certain psychological state of mind is a prerequisite for a speech act verb, which eventually may acquire performative status. And this constitutes a kind of metonymy (cf. Kakouriotis and Kitis 1999).5 Naturally, the emotion is perfunctorily assumed rather than existing, and socially motivated. We may recall Austin’s (1962) comments in this respect: There are numerous cases in human life where the feeling of a certain ‘emotion’ (save the word!) or ‘wish’ or the adoption of an attitude is conventionally considered an appropriate or fitting response or reaction to a certain state of affairs, including the performance by someone of a certain act, cases where such a response is natural (or we should like to think so!). In such cases it is, of course, possible and usual actually to feel the emotion or wish in question; and since our emotions or wishes are not readily detectable by others, it is common to wish to inform others that we have them. Understandably, though for slightly different and perhaps less estimable reasons in different cases, it becomes de rigueur to ‘express’ these feelings if we have them, and further even to express them when they are felt fitting, regardless of whether we really feel anything at all which we are reporting. (78–79).

Indeed, the name for the emotional condition might evolve into the name for the speech act; and in the case of both the Greek fovume and the English be afraid, they are both used to perform negatively viewed predictive speech acts as we have seen. The fear verb can act as a sine qua non mitigating device of 4 This is not absolutely correct, as, the verb ephobeito in fact checks to megiston as a direct internal object even if the latter functions as a quantifying adverb: ‘he feared a great deal’. I have in fact not found an instance of the verb directly selecting an hoti ‘that’ complement in my AG corpus. This construction appears to be a later development. 5 For a similar argumentation in respect of indirect speech acts, see Panther and Thornburg (1998), and Thornburg and Panther (1997).

448   Eliza Kitis the illocutionary force, as shown in the following news item: “Upon hearing that his son suffered from Hahira syndrome, a rare neurological disorder which would cause him ‘very serious difficulties’, David Cameron asked a pediatrician: ‘Does that mean he’s going to have trouble doing his maths, or does that mean he’s never going to be able to walk and talk?’ The doctor replied: ‘I’m afraid it means he probably won’t walk or talk.’” (The Mail, 11.03.07) Further, in (11), fovume functions as a performative as its speaker uses the predicate, not so much to mitigate the force of the rejection (a socially motivated tendency) but to ‘seal’ the speech act as a rejection, etc., depending on the context. Indeed, often a fear construction (fovume or be afraid) “as a performative, … expresses speaker’s appropriation of authority as an actor attempting to match world to word.” (Traugott and Dasher 2002: 209). The following text from a cartoon, underlines the point: A: Let’s get down to it, Mr. Duke. I’ll pay you $10,000 a week to represent Berzerkistan! B: I’m afraid we’d need $50,000, Excellency. A: I wasn’t negotiating. And don’t push it-I have a very long reach! (from Doonesbury, Garry Trudeau, The Guardian 18.10.07) ‘I’m afraid’ in B’s utterance turns it into a clear negotiatory move couching a rejection, hence A’s reply.

Therefore, fear verbs are prime examples of what Traugott (1989) has shown to be a well attested process in language, i.e. cases of shift from propositional (‘to flee’, ‘to unsettle’) to expressive meanings (emotions) to interpersonal ones (propositional attitude verbs, speech act hedges, speech acts, modalizers and parentheticals), as shown in the diagram below: Semantic Meaning Conceptual/Propositional Objective

Pragmatic Meaning Procedural Function Subjective

propositional

expressive

>

textual >

>

interpersonal

Figure 2. The evolution of fear predicates from semantic to pragmatic domains (adapted from Traugott 1989)

While spatial and emotion meanings are propositional and conceptual, performatives, parentheticals and modalizers have a procedural function. As Traugott and Dasher (2002: 209) write, “[r]ecruiting a verb designating a certain kind of locution (itself ultimately derived from an originally spatial lexeme) to performative use involves recruiting it from the domain of content semantics to

The case of ‘fear’ predicates   449

function also as a procedural indexing the kind of discourse being engaged in. Recruiting this verb …to the class of epistemic parentheticals involves recruiting it to a primarily procedural class.” As has often been shown in the literature, this process engages both the semantic and the pragmatic domains as it is a discourse-based evolutionary process, but as meaning evolution proceeds from the former domain to the latter, we can witness a transition from more objective domains to more subjective ones (Kitis 2006; Pishwa 2006a; Traugott 2003; Traugott and Dasher 2002; Verhagen 2005).6 However, it must be stressed that, while the fear predicates examined here have developed a concurrent procedural pragmatic function, in both languages they also retain their semantic content of emotion and are the main predicates for expressing the corresponding emotions. 6. Conclusion In conclusion, in this study we have witnessed a course of parallel shifts in the meanings of the fear verbs examined here. The signification of action was metonymically used to indicate, or as the name of, the emotion, gradually giving up its place in the semantic domain of meaning. So, we can safely assume that in both cases there has been a shift of meaning from the concrete source domain of motion and locality to the target domain of emotions. In the case of these verbs the source domain is currently inactive but its traces are felt in current uses of the verbs discussed. We have also witnessed a further shift from the domain of emotions to the domain of propositional attitude meaning; in this case, the domain of emotions has been used as a source domain. Both fovume and fear (originally, and afraid later) joined the course of propositional attitude verbs very early in their evolution. Having dropped the need for the complementizer at some stage, they 6 For a not very dissimilar account of these processes, see Tissari (2007). She bases her account of the subjectification of the propositional meaning of fear verbs on Fauconnier and Turner’s (2000) notion of compression, and on politeness principles. She also offers a comprehensive corpus-based account of the semantic evolution of the fear verbs. However, my account differs from hers in viewing the direction of the process: while she thinks that “[f]ear in the chain stands first for a suggestion of danger, then for evidence of danger, then for knowledge of danger, and lastly for the hearer/reader’s reaction”, this study adopts a rather phenomenological approach that guides interpretations to proceed from outer demonstrable behaviour to inner inaccessible emotional states. These states are then in their turn capitalized upon at a perfunctory semiological level in performing certain social speech acts.

450   Eliza Kitis gradually exhibited potential for free movement within the clause.7 They are currently used (alongside their propositional uses) also as parentheticals promoting their embedded clauses to the status of the main. We acknowledge that it is often extremely difficult to tell whether we have a performative use of a verb; and we also acknowledge that both afraid and fovume are currently used to introduce (rather than perform) speech acts. However, they both have the meaning of ‘regret’ and seem to be well on their way towards becoming full-fledged performatives as some of their uses indicate. Moreover, if regret is a performative verb, so be it for ‘regret’ uses of the predicates. It also appears that we need to take a more constructional view of speech acts and performativity that would take on board contextual features contributing to interpretations (Cf. Stefanowitsch 2003). Evidence for their performativity acquiring status is also afforded by the fact that an external negative operator with scope over these predicates is not licensed by such uses of the predicates either in Greek or English: *I’m not afraid I can’t help you. Neither can such uses be reported as emotion uses: *He was afraid he couldn’t help her. Moreover, we have provided evidence that supports Traugott’s (1989) hypothesis that propositional meanings give rise to interpersonal ones and not vice versa. This evidence does not support Vendler’s claims that there is a bidirectional leakage between thought and speech, and consequently between mental verbs and speech act verbs. “This shows once more”, as Traugott and Dasher (1987: 571) would say, “that there are powerful regularities in semantic change of a far more specific sort than the ‘extension of meaning, metaphoric shift, metonymic shift, amelioration’ or ‘pejoration’ we hear so much about in earlier treatments of semantic change.” Moreover, we believe that we have provided evidence that extends Traugott and Dasher’s (1987) claim that “as far as lexicalization of metalinguistic repertoires is concerned, [not only] ‘having in mind’ seems to be more fundamental than ‘asserting that’” (571), but also ‘having the emotion’ has been proven to be very essential, too. Our evidence supports Traugott and Dasher’s (1987) claim that thought (and we would add ‘emotion’ too) and speech “are not the same thing as the metacognitive and metalinguistic terms that lexicalize them” (572). Besides, fear is a native term, and this constitutes further evidence for the postulated anteriority and priority of the emotion over its articulation as a performative or parenthetical verb. The relation, indeed, is not mutually constitutive. 7 That does not mean, of course, that the that-complementizer uses are not synchronic uses in both languages.

The case of ‘fear’ predicates   451

In short, the discussion has demonstrated that the semantic evolution of the fear verbs examined in this study evidences a configuration of cognitive meanings that is put in the service of social demands. Indeed, sociality is cognitively configured. But then established social norms shape and reinforce cognitive structures to an extent that they reach the status of being regarded as collective cognition, probably of a different caliber. This type of social cognition is often blaringly missing in autistic conversational behaviour, and more particularly in Asperger’s syndrome (extensive personal knowledge), which is characterized by social impairment in the relevant literature. While neurotypical individuals can handle the interpersonal procedural function of fear predicates efficiently in their conversational transactions, Asperger patients can only use them as emotion predicates (semantic meaning) but not in their interpersonal pragmatic function as mitigating devices, etc. This evidence may indicate that semantic bleaching of expressions attributing to them a procedural function may implicate procedural strategies of social cognition stored separately from semantic and episodic memory (cf. Pishwa 2006b). 7. Implications for psychological accounts of emotions Appreciation of linguistic aspects of the evolution of emotions from experiential domains to the interpersonal domain of communication will further corroborate psychological accounts that view emotions as social constructs played out in the real world, i.e., as having communicative function, rather than merely as internal states (Parkinson 1995). In particular, an account of the development of fear predicates in the terms analyzed here might further inform interpersonal accounts of emotion vis-à-vis intrapsychic appraisal theories of emotions in psychology. According to the latter, emotion needs to be first encoded as a private meaning before it can enter the interpersonal world, as a type of translation of inner states. Moreover, my account of the history of the two predicates points to a dynamic ‘on-line’ construction of the emotion of fear rather than to a static, preformed cognitive schema, or internal script that is waiting to be activated by some external or other eventuality (Wierzbicka 1990, 1994; Kitis 2009, for criticism). On the other hand, interpersonal accounts of emotions in the field of psychology need to take into account findings of linguistic research in the field, which may balance their claim that emotion is primarily interpersonal.

452   Eliza Kitis References Abraham, Werner Diathesis: The middle, particularly in West-Germanic. In Discourse, 1995 Grammar and Typology: Papers in Honour of John W.M. Verhaar, Werner Abraham, Talmy Givón, and Sandra Thompson (eds.), 3–47. Amsterdam/New York: Benjamins. AHD 2000 The American Heritage Dictionary of the English Language. 4th ed. Aristotle Poetics, Transl. with an introd. and notes by G.F. Else, The University of 1970 Michigan: Ann Arbor Paperbacks. Austin, John 1962 How to Do Things with Words. Oxford: Clarendon Press. Damasio, Antonio Looking for Spinoza Joy, Sorrow, and the Feeling Brain. New York: 2003 Harcourt. Fauconnier, Gilles, and Mark Turner 2000 Compression and global insight. Cognitive Linguistics. 11 (3–4): 283–304. Foucault, Michel Reprint. The Order of Things. London and New York: Routledge, 1966. 2002 Fraser, Thomas The establishment of ‘by’ to denote agency in English passive construc1987 tions. In Papers from the 7th international conference on Historical Linguistics, Anna Giacalone Ramat, Onofrio Carruba, and Giuliano Bernini (eds.), 239–49. Amsterdam/Philadelphia: Benjamins. Giacalone Ramat, Anna, Onofrio Carruba, and Giuliano Bernini (eds.) Papers from the 7th International Conference on Historical Linguistics. 1987 Amsterdam/Philadelphia: Benjamins. Gibbs, Raymond W. 1999 ‘Speaking and thinking with metonymy’. In Metonymy in Language and Thought, K-U. Panther and G. Radden (eds.), 62–76. Amsterdam/Philadelphia: John Benjamins. Goodwin William Syntax of the Moods and Tenses of the Greek Verb. London: MacMillan. 1889 Grice, Paul H. 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Homer 1948[1898] The Iliad. Samuel Butler (ed.). ebooks@adelaide, 2006. Hume, David Reprint. On Human Nature and the Understanding, Anthony Flew (ed.). 1962 New York, London: Collier Books, 1748. Iwata, Seizi 1995 The distinctive character of psych-verbs as causatives. Linguistic Analysis 25: 95–120.

The case of ‘fear’ predicates   453 Kakouriotis, Athanasios, and Eliza Kitis 1999 The case of ‘fovame’ and other psychological verbs. In Greek Linguistics. Proceedings of the 3rd International Conference on Greek Linguistics, Amalia Mozer (ed.), 131–140. Athens: Ellinika Grammata. Kitis, Eliza 2006 Causality and subjectivity: The causal connectives of Modern Greek. In Language and Memory: Aspects of knowledge Represenation, Hanna Pishwa (ed.), 223–267. 2009 Emotions as discursive constructs: The case of the psych-verb ‘fear’. In Studies in Cognitive Corpus Linguistics, Katarzyna Dziwirek and Barbara Lewandowska-Tomaszczyk (eds.), 147–172. Frankfurt: Peter Lang. Kövecses, Zoltán 1990 Emotion Concepts. New York: Springer. 2002 Metaphor: A Practical Introduction. Oxford/New York: Oxford University Press. (L&S), Liddell Henry George, and Robert Scott 1948 A Greek-English Lexicon. Oxford: Clarendon Press. Meyer-Lee, Robert J. 2007 Poets and Power from Chaucer to Wyatt. Cambridge: Cambridge University Press. (OED) Oxford English Dictionary. Oxford: Oxford University Press. Panther, Klaus-Uwe, and Linda Thornburg 1998 A cognitive approach to inferencing in conversation. Journal of Pragmatics 30: 755–769. Parkinson, Brian Ideas and Realities of Emotion. London and New York: Routledge. 1995 Pishwa, Hanna 2006a Expression of uncertain goals in communication: The case of multifunctional ‘try’. In Language and Memory: Aspects of Knowledge Representation, Hanna Pishwa (ed.), 269–311. Berlin/ New York: Mouton de Gruyter. 2006b Memory and language: Introduction. In Language and Memory: Aspects of Knowledge Representation, Hanna Pishwa (ed.), 1–34. Berlin/ New York: Mouton de Gruyter. Pishwa, Hanna (ed.) Language and Memory: Aspects of Knowledge Representation. Berlin/ 2006c New York: Mouton de Gruyter. Stefanowitsch, Anatol 2003 A constructional-based approach to indirect speech acts. In Metonymy and Pragmatic Inferencing, Klaus-Uwe Panther and Linda Thornburg (eds.), 105–126. Amsterdam/Philadelphia: Benjamins. Theodoropoulou, Maria Στα Γλωσσικ Μoνoπτια τoυ Φboυ. Ψυcισμς και Γλσσα 2003 [On the Linguistic Path to Fear. Psyche and Language]. Athens: Nisos.

454   Eliza Kitis Thornburg, Linda, and Klaus-Uwe Panther Speech act metonymies. In Discourse and Perspective in Cognitive 1997 Linguistics, Wolf-Andreas Liebert, Gisela Redeker, and Linda Waugh (eds.), 205–219. Amsterdam/Philadelphia: Benjamins. Tissari, Heli 2007 Compressing emotion to politeness: On I fear and I’m afraid. In Change in Meaning and the Meaning of Change: Studies in Semantics and Grammar from Old to Present-Day English, (Mémoires de la Société Néophilologique de Helsinki LXXII) M. Matti Rissanen, Marianna Hintikka, Leena Kahlas-Tarkka and Roderick McConchie (eds.), 57–90. Helsinki: Société Néophilologique. Traugott, Elizabeth Closs 1989 On the rise of epistemic meaning in English: An example of subjectification in semantic change. Language 65 (1): 31–55. 2003 From subjectification to intersubjectification. In Motive for Language Change, Raymond Hickey (ed.), 124–139. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs, and Richard Dasher 1987 On the historical relation between mental and speech act verbs in English and Japanese. In Papers from the 7th International Conference on Historical Linguistics, Anna Giacalone Ramat, Onofrio Carruba, and Giuliano Bernini (eds.), 561–573. Amsterdam/Philadelphia: Benjamins. 2002 Regularity in Semantic Change. Cambridge: Cambridge University Press. Tzartzanos, Achilleas 1946 Νεoελληνικ Σνταxις B [Modern Greek Syntax]. B. Thessaloniki: Kiriakidi. Vendler, Zeno Res Cogitans: An Essay in Rational Psychology. Ithaca and London: 1972 Cornel University Press. Verhagen, Arie 2005 Constructions of Intersubjectivity. Discourse, Syntax, and Cognition. Oxford/New York: Oxford University Press. Wierzbicka, Anna The semantics of emotions: fear and its relatives in English. Australian 1990 Journal of Linguistics 10 (2): 359–375. 1994 Cognitive domains and the structure of the lexicon: The case of the emotions. In Mapping the Mind. Domain Specificity in Cognition and Culture, Lawrence A. Hirschfeld and Susan A. Gelman (eds.), 431–452. Cambridge: Cambridge University Press. 1999 Emotions Across Languages and Cultures: Diversity and Universals. Cambridge: Cambridge University Press.

Chapter 17 Metaphor in mental representations of space, time and society: The cognitive linguistic approach Paul Chilton

1. Introduction In contrast with Chomsky’s theory of language, cognitive linguistics (CL) is based on the hypothesis that human language is a specialised ability related to other cognitive abilities of human beings. For Chomsky, language is an autonomous module which interfaces with, but does not share the attributes of other cognitive systems. CL is a theory about the human mind as well as about human language. This places it in contrast with M.A.K. Halliday’s approach to language, which emphasises the social situation of language and treats linguistic expressions as the product of systems of choices between meanings which it does not link to cognitive abilities. CL does not ignore social meanings, as will be seen, but it assumes that social meanings exist in people’s minds, that is, in their brains. There is no universal agreement as to how exactly language is linked to non-linguistic cognition: CL is a research programme. However, certain plausible ideas on the matter have emerged over several decades of CL research. In the influential work of Langacker (1987, 1991), and also Talmy (2000, 2001), for instance, the various visual systems of the human brain are clearly crucial. These include the scanning of scenes and the distinction between foreground-background and alternating viewpoints on scenes. CL also hypothesis, on the basis of linguistic evidence, that other bodily systems enter into linguistic meanings. Linguists (Lakoff 1987, 1993), philosophers (Johnson 1987) and psychologists (Mandler 2004) agree that human minds use ‘image schemas’ that arise from the interaction between the human body and its physical environment. Such image schemas include representations of motion (path schema, moving from point a to point b), bounded space

456   Paul Chilton (container schema, the experience of being inside something), vertical orientation (up-down schema, related to human upright stance), grasping schemas (unconscious control of manual actions). Meaning, as expressed and communicated in language, can be explained in terms of these kinds of mental structures, which are inherently meaningful for humans, and arise because humans have a certain kind of anatomy and physiology in a certain kind for physical environment. For these reasons, some cognitive linguists stress that meaning is ‘embodied’.

up

front

over, above in front of

left side

self behind

under, beneath, below

right side

back down Figure 1. Spatial prepositions related to bodily orientation

2. Language and space As can be seen from the above outline, we would expect that spatial representation in general plays an important role in cognition and language. It is certainly true that all languages have forms that code spatial meanings of various kinds, though of course languages vary in the way they do this. The most obvious coding of spatial concepts in languages is in prepositions. Although all human beings are subject to the same laws of space and time, there is a certain amount

The cognitive linguistic approach   457

of variation in the way the prepositions in different languages denote the possible spatial relations. Here we shall focus on the English prepositions. Most of these denote spatial relationships, and some of those that in modern English denote only temporal relations are derived form spatial meanings (before, after). The variety of spatial prepositions is considerable. One way of grouping them is in terms of their complexity. Some encode more of the three physical dimensions than others. The preposition at seems to encode a point, i.e. no dimensions. The prepositions from, to and toward encode a one-dimensional path or direction specified by the complement NP, not any particular direction such as upward, downward or sideways. The preposition on, on the other hand, can refer simply to a plane, i.e. two-dimensional surface, while in is prototypically associated with a three-dimensional volume. Through and across imply that their complement NPs are, or can be construed as, having a particular geometrical shape. Some prepositions, for example, by have a basic spatial meaning that denotes merely proximity, but also many derived meanings. The best example of a polysemous preposition of this type is of, which serves to indicate a range of associations between two entities. These brief remarks do not do justice to the complex meanings and clusters of derived meanings associated with the English prepositions. Lakoff (1987) and others have shown, for instance, that the preposition over expresses a range of related spatial schemas. Other linguists (Herskovits 1986; Vandeloise 1999) stress that prepositions express not just spatial relations but also functional properties (e.g. on implies support). A specific case of polysemy is metaphor and it is on the metaphorical extensions of one particular set of English prepositions that we are going to concentrate – the set of related prepositions that are defined in relation to the structure of the human body. As shown schematically in Figure 1, we can think of the self as located at the geometrical origin of three intersecting planes: the back-to-front plane, which is oriented because of the fact that the main organs of perception lie in the direction of movement, the vertical orientation given by and oriented by the gravitation signal, and the lateral plane. We can also see that some of these spatial prepositions are linked with image schemas for certain kinds of movement – forward motion, falling and perhaps swaying from side to side. There is much more to say about the semantics of these basic prepositions, but we are going to move forward a step and consider the metaphorical meanings that are built upon them. Cognitive Linguistics hypothesises that spatial representation is pre-linguistic and provides basis for meanings coded in linguistic expressions. By metaphor, spatial representations are used for more abstract concepts. To explain this idea further we next have to outline the cognitive theory of metaphor.

458   Paul Chilton 3. Metaphor: The theory Cognitive Linguistics hypothesises that spatial representation is pre-linguistic and provides basis for meanings coded in linguistic expressions. Metaphor is crucially involved in the extension of basic physical spatial meanings into non-physical, non-spatial ones and has been extensively, though not conclusively researched by cognitive linguists and some psychologists (amongst others, cf. Lakoff 1987; Lakoff and Johnson 1980, 1999; Gibbs 1994; Boroditsky 2000). The term metaphor comes from the Greek morphemes meta + phor (across +carry), which are equivalent to the Latin trans + fer. The idea of metaphor has been discussed in western philosophy and the arts of rhetoric at least since Aristotle (384–322 BC), whose definitions are the best known and most influential. Aristotle thought of metaphor as the use of a word to refer to something to which it would not normally refer. This could be done, he felt, when there was an inherent similarity in the objects referred to. A later manual of rhetoric that was widely influential in Europe up to the Renaissance put it as follows: Metaphor occurs when a word applying to one thing is transferred to another, because the similarity seems to justify the transference. (Rhetorica ad Herennium (c. 86 BC)

Until relatively recently metaphor was thought of in the west as an ornament of literary style or as a persuasive device in speech making. Some philosophers, for example Thomas Hobbes (1588–1679), were deeply suspicious of metaphor and believed that metaphor led the mind astray. Cognitive Linguistics makes a number of different claims about metaphor, although the central idea of transfer remains. First, metaphor is a conceptual operation, something that happens in the non-linguistic part of the mind; it is not simply the use of a word in an unusual way. Second, it follows from the first point that metaphor is confined to the literary or oratorical uses of language, although it might be particularly important in those uses. What this means is that many everyday lexical items and grammatical constructions have an active metaphorical component; others are polysemes, the origins of whose meanings can be explained in terms of historical semantic changes that are themselves metaphorical. Third, metaphors are not isolated occurrences, but are systematic and productive. This means that certain kinds of concept recur in metaphors and it means that when new metaphors appear they are usually related to existing ones. This can happen because of ‘metaphorical entailment’, the counterpart of logical entailment. In metaphorical entailment, we have a

The cognitive linguistic approach   459

mental operation which has roughly the following structure: if x can be metaphorically viewed as y, then y has the properties of x. The extent to which this generalisation is actually true is one of the contended issues in current research on metaphor (cf. Croft and Cruse 2004). The CL explanation of metaphor also includes two further important parts. The first part is the notion of mapping from a source domain to a target domain. ‘Mapping’ here means roughly ‘correspondence’. These mappings are usually summarised in the following way: life is a journey, where journey is the source domain and life is the target domain. In actual metaphorical expressions (e.g. ‘young Tom was just starting out in life’, ‘Henry was led astray in his youth’) that derive from this underlying mapping one highly schematic concept (‘life’) is viewed in terms of another concept that is understood in a more detailed and concrete way. The second part concerns the source domain. What exactly is the conceptual source of metaphorical mapping? There is a general claim that source domains are more basic and intuitively understood by the human mind, target domains more schematic, less well understood and more abstract. Although research is still proceeding, many cognitive linguists think that many source domains are primarily spatial and are present in some form in the neural networks of the human brain, either at birth or in early child development. Because it is impossible at the present stage of research in neuroscience to make exact claims, CL and psychologists have operated on the more general level of ‘image schemas’, mentioned above. Neuroscientists have of course established the existence of a certain number of cognitive systems, especially for vision. For example, there are visual systems for grasping, for object recognition, for locomotion. CL metaphor theory (and also the cognitive grammar of Langacker) postulates ‘image schemas’ such as path. In theory it should be possible to link such schemas with the neural pathways discovered by neuroscientists, but this has not yet been done. In any case, it should be noted that the ‘image schemas’ are complex constructs that would involve more than one neural system: for example, the path schema could involve not only vision but also vestibular, sensory-motor and somatosensory systems. It seems unlikely that source domains are entirely accounted for by such innate or universal bodily systems, however. In addition, the study of metaphors shows that many source domains must be culture-based. For example, it is well known that common source domains include the family, warfare, journeys and dwelling places. It is possible of course that these source concepts are in some sense basic for human beings, but it is also clear that the structure of the family, the structure of houses, the conduct of war, etc. varies from culture to culture and from time to time

460   Paul Chilton 4. Metaphor: The case of time One of the most frequently mentioned examples in the CL literature is the mapping from spatial representation to the abstract domain of time (cf. Lakoff and Johnson 1980; Evans 2004; Radden 2004; Yu 1998). In English expressions relating to time, points in time or periods of time are expressed by lexical items that are systematically metaphorical. First, times are expressed as entities; second they are represented in relation to a path schema. For example: (1) (2)

The election date is approaching. We are getting closer to the election date.

In (1) a time point is represented as an object ‘moving closer’ to us; in (2), the same time point is represented as an object towards which we are ‘moving’. We can imagine variations on this theme: ‘the elections are coming’, ‘we are fast approaching the day when…’, ‘the examinations are slowly creeping up on us…’. In English and other languages there is thus a productive metaphor system that has two complementary aspects. The two aspects of conceptualising time, illustrated in (1) and (2), simply reflect the two ways the human conceptual and perceptual apparatus views moving objects, and specifically the parallax effect. Despite this alternation, however, it should be noticed that future time is always ‘in front’, or ‘ahead’ of us, so that in general we always ‘face the future’, in English, but also in many unrelated languages. English also uses a range of spatial prepositions to locate points in time (note that ‘point’ is itself primarily a spatial concept). For example, several prepositions are used in (3): (3)

The meeting is at ten o’clock in the morning on Thursday.

Here CL does not take the view that the view that at, in and on are simply arbitrary, decided by rule. Rather, the hypothesis is that such constructions are ‘motivated’. A tentative explanation in this case would go roughly as follows. The preposition at treats both the landmark location and the object being located in a highly schematic way, as if they were points: e.g. Mary is at the counter. This preposition is also used for ‘location’ in a sense that is not entirely physical: Mary is at university. And it occurs in a much more abstract sense in (3), where both the meeting and the time ‘ten o’clock’ are viewed as points. The preposition in, in its physical spatial meaning, has meaning because it is linked to the image schema of container: the larger period of time ‘contains’ the smaller unit, the point ‘ten o’clock’. As for ‘on Thursday’, the preposition on, its meaning includes the concept of one plane in contact with, and perhaps

The cognitive linguistic approach   461

supported by another plane. The supporting plane typically has a greater surface area, which two-dimensionally ‘contains’ the located plane: e.g. the book is lying on the table. Analogously, ‘the morning’ is located on the metaphorical plane of Thursday, occupying an ‘area’ of its ‘surface’. Of course, this way of linguistically encoding the temporal representation in (3) is probably peculiar to English, nor are these the only peculiarities of English expressions. Other European languages do not work in this way, nor do many other languages of the world. It is therefore interesting to ask whether or to what degree the spatial mapping from space to time is universal. This means asking, for example, if all three spatial dimensions are used, and if not whether all languages use the same one(s) or not. We have seen that in English time expressions use the path image schema, which is essentially a one-dimensional line, and we have seen that the spatial prepositions two-dimensions for the purpose of locating time points and periods. It makes sense to ask, as Radden (2004) does, how the three directions oriented on the human body as the origin (see Figure 1) come into play in expressions of time. For English speakers it seems natural to think of time as a line running primarily from back to front. It is this line that is presupposed in examples (1) and (2). There are some English expressions that apparently evoke the vertical dimension: (4)

The exams are coming up in a week.

However, this does not seem to be a productive metaphor, if one assumes that it is the up-down schema that is involved: (5)

*The exams went down last week.

Rather, the explanation of (4) would seem to be related to the visual experience of moving through space, just as the parallax effect explains the alternation of (1) and (2). In (4) we probably have the cognitive and linguistic effect of the experience in which distant objects appear to get taller, or rise higher from ground level, as we or they approach, and in fact project a larger image on the retina.1 1 Radden (2004) citing Yu (1998) argues that English does actually use a genuine vertical schema when talking about genealogical time: stories are passed down form one generation to the next, from descendant to descendant, have lasted down to the present… In the register of genealogy we speak of ascendants, though more ordinarily of ancestors, which is not spatial. These examples appear to restricted and not productive. Radden 2004 also claims that a vertical schema metaphor is

462   Paul Chilton The vertical axis seems therefore not to be utilised in the English linguistic system; nor is it used, it would appear, in any non-linguistic way of conceptualising and reasoning about time. Equally, English does not use the lateral axis: there are no expressions that represent time as going from left to right or vice versa.2 However, this does not mean that some languages may utilise conceptual sources different from the ones used in English or by English speakers. In Chinese there exist certain expressions for time that do in fact use the vertical axis, as well as expressions that use the back-front axis (Yu 1998; Radden 2004). For example: Chinese has shàngyuè (up + month), equivalent to English ‘last month’, and xiàyuè (down + month), equivalent to English ‘next month’. Chinese also exploits the three axes of Figure 1 to a much greater degree than English. Chinese expressions equivalent to the English phrase ‘present time’ include: shǒubiān (hand + beside), dāngqián (just at + front), yǎnqián, mùqián (eye + front), yǎnxià (eye + below). In other words, the Chinese expressions are using as a source domain the lateral axis (hands), the back-front axis and orientation (front) and the vertical axis (underfoot). The important point here is that these expressions all draw on another feature of spatial representation, namely, distance. Each of the expressions just listed refer to the spatial area close to the body, or within reach of the hands; this is the area known to cognitive scientists and psychologists as peripersonal space, the space around the person. Incidentally, the etymology of the English word ‘present’ indicates something very similar: the word present the first part of which pre- derives from the Latin prae, a preposition meaning ‘in front of’. This meaning is of course lost for modern speakers, but etymologies are often evidence of a direction in semantic change that is consistent with synchronic evidence. Although modern English does not use this type of expression, it does use the distance scale: points in time can be ‘close’, ‘near’, ‘far off’, ‘remote’, etc. Chinese, then, uses a different spatial axis for its source domain. There is also the question of direction, which way the axis that is chosen points. In certain Chinese expressions time flows from top to bottom. In other languages, time can flow on the same axis as for English, but in the opposite direction. While future time is always ‘in front’ for English, for Aymara (northern Chile) present in This year went down in family history. The phrase go down does not seem to me to denote time in any sense; rather it seems to be related to expressions such as write down, note down, nail down, put down to, etc., which do not have time as target but the fixing or stabilising of some thing. 2 True, people may imagine a time-line as running from left to right, especially if the written language in their culture is organised horizontally left to right, but there are no linguistic expressions for time in English that are based on such an image.

The cognitive linguistic approach   463

(the future is represented as behind the person (Núñez and Sweetser 2006). In Aymara the word for ‘past’ is nayra, which is roughly equivalent to ‘eye’, ‘sight’ or ‘front’, and the word for future is qhipa equivalent of ‘behind’ or ‘the back’. The Aymara word for ‘tomorrow’ is qhipüru, composed of qhipa and uru (the word for ‘day’), literally meaning something like ‘the day behind one’s back’. Another characteristic of the Aymara time metaphors is that they are static: time points do not metaphorically move in any direction. Because there is consistency over a range of linguistic expressions, we can generalise and say that Aymara has an underlying metaphor for time which uses the front-back axis of the human body but fixes the ‘direction’ of time such that the person is facing the past and the future is behind his or her back. An explanation is that what is in front of you can be seen, that is known, which the case is for the past. The Aymara language has evidential morphemes that focus attention on what is seen as epistemically more reliable: this may explain why forward-facing vision is important for facing the past as something known. The above is only a small sample of the data, so it is dangerous to generalise. Very tentatively, however, the following can be proposed. For the purpose of mapping spatial source domains to the time target domain languages draw variously on spatial cognition in the following ways. –– Languages draw on different body-related axes, probably only on the frontback axis and the up-down axis (not the left-right axis) –– Languages can orientate the axes in different ways –– Time points can be viewed in relation to the self or in relation to other times –– Time is viewed in terms of spatial orientation relative to self –– Time is viewed in terms of spatial distance relative to self –– Time or the self may or may not be viewed as moving. Finally, it should be noted that the existence of metaphorical mappings from space to time does not, logically or empirically, mean that the human mind has no independent cognition of time. Its role is not to make the conceptualisation of time possible, but to make it possible to reason about time and to communicate about time (cf. Evans 2004). 5. Metaphor: Social structures and relations Despite pertinent remarks about metaphors of power, control and influence in Lakoff and Johnson (1980), there is to my knowledge no systematic study of the spatial metaphors for concepts of society and social interaction. One would

464   Paul Chilton expect these concepts to vary not only cross-linguistically but also according to historical period, social groups and social ideologies. However, the building blocks of culturally variable concepts of society could, in principle, come from the same source domain, with different languages and cultures using them in different ways. The question that I am opening in this final section is: In which ways is the spatial source domain mapped on to the social structures and relations? As in the case of time, there are probably language-independent cognitions of the social. These may be innate or developed in childhood, or some combination of both processes. Some psychologists and ethologists infer from observation that there is an innate capacity for social cognition. Such cognition seems to include the following: the ability to understand intentions of others (Leslie 1991), the ability to cooperate, to deceive and manipulate (Byrne and Whiten 1988), and its counterpart, the ability to detect deception (Sperber 2000). But in order to communicate about societies (a process that is itself constitutive of human societies), some form of linguistic representation would be essential and it may be that a major part of such representation relies on metaphorical mappings from spatial cognition. Here I am not focussing on specific discourses in specific societies. I will merely suggest some ways in which spatial metaphors play a role in the ways language enables conceptualisation and communication about social structures and social relations. Let us examine each of the body-based, oriented axes outlined in Figure 1. The best known case is what Lakoff and Johnson (1980) call the power is up metaphor. Metaphorical expressions show that the metaphor uses both direction and distance on the vertical axis, but there are some subtle conceptual consequences in the choice of certain prepositions, specifically the choice between over and above: (6) (7)

Albert is over Bill, Bill is under Albert Albert is above Bill, Bill is below Albert

Intuitively there are differences between (6) and (7). The meanings of the preposition over constitute a radial category (Lakoff 1987), many of which include contact between landmark and trajector, complete coverage and movement. This is not the case for above, whose meanings do not include contact (cf. Tyler and Evans 2003: 111 ff.), as can be seen from the meaning possibilities of (8) and (9), where in (8)

The flag was over the table

The cognitive linguistic approach   465

the flag could be spread in contact with the surface of the table, while in (9)

The flag was above the table

this meaning is not available. This difference explains why (6) appears to imply power and control of Albert over Bill, while (7) seems to imply only higher status, not necessarily also power – though of course in practice those with higher status may in fact have power over those with lower status. This much is apparent from expressions containing the term power, such as: (10) Albert has power over (*above) Bill, Bill is under (*below) Bill’s power. English has a number of expressions that denote change of social status in terms of movement up or down the vertical scale: one may be a ‘social climber’, and ‘go up in the world’, for example, or one may ‘drop down’ the hierarchy, ending up ‘at the bottom’ of society. It is of course the case that the up-down spatial axis is also productively mapped onto rather diverse non-social target domains, such as mood (‘her spirits rose’, etc.), processes of fragmentation (‘tear up the paper’, etc.), quantity (‘the rising numbers of immigrants’, etc.) and states of finality (‘finish up the food’, ‘be fed up’, etc.), which we cannot discuss in detail here. However, if a given society has any sort of status recognition it is likely that this will be encoded in up-down terms. The motivation is clearly derived from physical experience: control and power are indeed, because of gravitation and also because of visual advantage, greater if one occupies a relatively high position. Some of the target domains that are not specifically to do with social structure do none the less cohere with the status is up metaphor. A particularly interesting example is the good is up metaphor in its deontic sense ‘morally right is up’. In English this is a productive mapping which includes examples such as ‘high moral standards’, ‘he is above such low behaviour’, ‘standards are slipping’, ‘falling standards’, ‘it is beneath him to do that’, ‘he is of low moral character’, ‘lofty ideals’, and so forth. We should note here once again the choice between over and above. The latter, as noted above, implies lack of contact, that is, distance. Hence we have (cf. Tyler and Evans 2003: 119): (11) He is above (*over) such low behaviour. It remains an open question whether other languages use the vertical axis as a source domain for concepts of moral values. It is also an open question whether

466   Paul Chilton there is any significant cognitive consequences for the parallelism between the status is up metaphor and the good is up metaphor. Turning to the front-back axis, we find here also a rich collection of target domains that relate to social relationships of various kinds. Perhaps the most fundamental of these is social precedence. Dominance hierarchies are correlated with ‘pecking order’, i.e. access to food, on the horizontal forward-facing axis. Such social ordering is reflected in many social rituals, such as processions and passing through doorways, which are not directly related to food but are certainly related to (perceived) status. So the concept of order and priority, which can be highly abstract, has both a social and an embodied source. The following sentences have a variety of possible meanings that deserve further research, but for present purposes we can note the social meanings: (12) (13) (14) (15)

The President comes ahead of the minister. The minister is behind the President. John is ahead of Mary (for promotion) Mary is behind John (for promotion).

These meanings are derived by metaphorical mapping from a source in which people are viewed as positioned one in front of another, facing the same direction, along a line. In (12) and (13) the meaning is both literal (for example, it describes a procession of dignitaries) and metaphorical (it describes the relative status of the dignitaries), while (14) and (15) has only the metaphorical meaning. However, it seems that front-back axis metaphors do not express a stable position in a hierarchy, but intrinsically evoke a dynamic process. (14) and (15) do not seem to entail that John will remain ‘ahead’ of Mary or Mary ‘behind’ John. The implication of movement is what would be expected, given the salience of forward-facing motion in human embodied experience. This particular mapping of front-back onto precedence in various types of social activity is closely related to another social concept, namely leadership. The difference is that instead of prepositions such as ahead of, in front of and behind, we have verbs which semantically imply both spatial location and spatial action, as in the examples below: (16) John Smith is leading the people toward prosperity. (17) The people follow him. Such examples rely on viewing a spatial relation in the horizontal plane between one and many individuals. However, the metaphorically induced con-

The cognitive linguistic approach   467

cept is a social (or political) concept, not a physically spatial one. In (16) we can see that the source image schema is not in fact simply a relationship between one individual and a set of individuals but involves the components of the path schema. The path schema is a mental structure made of up a trajectory, its source and its goal or end-point. Sentence (16) illustrates the goal element of the path schema; it is easy to invent an example that also includes the source of the ‘movement’ involved in ‘leading’ and ‘following’. As in the case of social metaphors derived metaphorically from the vertical axis, those derived metaphorically from the front-back axis have a naturalseeming relationship with an axiological scale, a scale of values expressing what is right or desirable. Thus, one finds examples such as backward nation, a nation that is lagging behind, as opposed to one that is forging ahead, making great strides forward. In general to be ‘backward’ is negatively valued while to be ‘go-ahead’ is good. There is thus a metaphor of the form forward motion is good, which appears to be productive in English and other languages. This metaphor is also coherent with the values that are associated with the leaderfollower schema and the first-is-best schema, discussed above. Finally, let us consider the lateral axis, which runs from left to right or right to left. The English prepositional phrase at the side of can take as complement any NP that can be viewed as having sides, and also viewed as having a back-to front orientation. However, the form at NP’s side can apply only to humans: (18) John stood at Mary’s side (19) *John stood at the car’s side. A possible explanation is that the possessive construction used in (19) implies (perhaps iconically) greater proximity and intrinsic possession: the ‘sides’, ‘fronts’ and ‘backs’ of cars are projected from human body parts that are intrinsically oriented onto an inanimate object.3 It is perhaps for this reason that (18) can imply social support, as one can see if to (18) one adds, for example, ‘throughout her troubles’. It is also the case that one can say (20) John stood by Mary

3 This is only an outline of an explanation. With regard to the possibly iconic difference between the possessive NP’s NP construction and the NP of NP construction, the suggestion is that the auditory or visual proximity of the NPs is related to the conceptual proximity of what the NPs denote: consider the difference between the teacher’s car and the car of the teacher.

468   Paul Chilton and mean, depending on context, that John was physically at the side of Mary or that John gave Mary social support, maybe while she was under some form of criticism or ‘attack’. The preposition by has in fact several subtly related meanings that cluster together in a radial category. The most basic spatial meaning seems to require the located object to be in some spatial regions surrounding the landmark object. However, there is also a connection with the lateral axis. Even when an object manifestly does not have ‘sides’, for example a tree, it is still possible to say (21) There is a rock by the tree. This will tend to be understood as ‘at the side of the tree’. This is because there is a default assumption that speakers are facing the object they are referring to, and because facing an object projects an axis from speaker to object, with parallel lines to the edges of the tree: this leads the viewer to attribute ‘sides’ to the tree. It has other meanings, for example the agency meaning (as in ‘the book by John’), an explanation for which would take too much space here. For present purposes, the point is that many meanings of the English preposition by are closely connected to laterality, and this fact explains how (20) can have a social meaning. There is more, however, to the explanation of this metaphor. Again, the cognitive hypothesis concerning the experiential basis of metaphorical meanings is relevant: physical closeness of human beings provides physical support and protection. We have considered the three bodily axes as separate source domains for metaphorical mappings onto social target domains. However, perhaps the most productive spatial source domain in this context is not related to the dimensional axes but to spatial distance and direction (the elements of vectors). There are two types of closely related mapping from distance and direction. The target domain of the first includes several important types of social relationships; including kinship, social solidarity, and alliance in conflict contexts. The following are examples: (22) She is a close relative of Bill’s (23) She is very close to the section leader (24) The UK is close to the US. Clearly in none of these cases is close interpreted in a spatial sense, although the motivation for the metaphorical mapping onto non-spatial targets comes from spatial experiences that are socially significant. Different degrees of distance can be metaphorically mapped onto social relationships and compared,

The cognitive linguistic approach   469

and metaphorical movement along the distance scale also permits corresponding conceptualisation. Thus, for example, someone can be more or less distant from oneself, a relative can be remote, and one can distance oneself from former friends or allies. That directionality is involved is evident form the way deictic verbs come and go are used, as shown in Figure 2. close to

neutral middle ground

come towards our position

distance oneself

self here

distant from go too far the limit other there

Figure 2. Social relationships in terms of distance and direction

Particularly significant is the opposition of self and other on a distance scale. The agreement of some other party with oneself is represented in terms of a movement of the other from one location to a close location. English expressions such as ‘they just went too far: their action was totally unacceptable’ and ‘that was the limit: their action was totally unacceptable’ are also metaphorical mappings from the basic spatial concepts of scalar distance, where one end of the scale is located at the self and the remote end (the other) is distant. In generalising over the spatial metaphors for social relationships, we might tentatively suggest the following hypotheses for further investigation: –– –– –– –– –– ––

Languages use all three bodily axes as source domains for social concepts The vertical and front-back axes may be a favoured source Languages may differ in the way they use the three axes Spatial distance relative to the speaker’s location is an important source Spatial direction is also important but possibly less used. Spatial concepts that are metaphorically mapped onto social concepts also involve physical relations between humans that have social significance.

6. Conclusion Much work remains to be done to substantiate the claims of cognitive linguistics. However, at least for English, there is now a lot of linguistic evidence that suggest that spatial concepts are important for the formation of the meanings – that is, concepts – that are associated with particular lexical items. A number

470   Paul Chilton of theoretical issues remain under discussion. For example, it may be argued that metaphor is not required in order to conceptualise time or society but only arises as a means of creating the symbolic means for communicating about them. Another issue is the direction of metaphorical mappings: are they always from the more physical to the more abstract? Moreover, the notion of image schema, in the sense used by Johnson, Langacker and Lakoff was initially a hypothesis based on rational speculation that is now only gradually being given substance in terms of the neurological systems of the brain. Finally, it is English that has been most intensively researched and there is now a need to carry out research in other languages that enable us to establish whether there are universal tendencies in metaphorical mappings across diverse languages. References Boroditsky, L. 2000 Metaphoric structuring: Understanding time through spatial metaphors. Cognition 75: 1–28. Byrne, R., and A. Whiten (eds.) Machiavellian Intelligence: Social Expertise and the Evolution of Intel1988 lect in Monkeys, Apes and Humans. Oxford: Clarendon Press. Croft, W., and D.A. Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. Evans, V. 2004 The Structure of Time: Language, Meaning and Temporal Cognition. Amsterdam: John Benjamins. Gibbs, R. 1994 The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Herskovits, A. 1986 Language and Spatial Cognition: An Interdisciplinary Study of Prepositions in English. Cambridge: Cambridge University Press. Johnson, Mark The Body in the Mind. Chicago: Chicago University Press. 1987 Lakoff, G. 1987 Women, Fire and Dangerous Things. Chicago: Chicago University Press. 1993 The Contemporary Theory of Metaphor. In Metaphor and Thought, A. Ortony (ed.), 202–251. Cambridge: Cambridge University Press. 1996 Moral Politics: What Conservatives Know that Liberals Don’t. Chicago: Chicago University Press, Chicago. Lakoff, G., and Johnson, M. 1980 Metaphors We Live By. Chicago: Chicago University Press

The cognitive linguistic approach   471 1999 Philosophy in the Flesh. New York: Basic Books. Langacker, R.W. 1987 Foundations of Cognitive Grammar, Vol. 1. Stanford: Stanford University Press. 1991 Foundations of Cognitive Grammar, Vol. 2. Stanford: Stanford University Press. Leslie, A. 1991 The theory of mind impairment in autism: Evidence for a modular mechanism of development. In Natural Theories of Mind: Evolution and Development and Simulation of Everyday Mindreading, A. Whiten (ed.), 63–78. Oxford: Blackwell. Mandler, J.M. 2004 The Foundations of Mind. Oxford: Oxford University Press. Núñez, R., and E. Sweetser With the future behind them: Convergent evidence from Aymara lan2006 guage and gesture in the croslinguistic comparison of spatial construals of time. Cognitive Science 30: 401–450. Radden, G. 2004 The metaphor time as space across languages, http://www.spz.tu-darmstadt.de/projekt_ejournal/jg-08-2-3/beitrag/ Radden1.htm [accessed 26.12.2004]. Sperber, D. 2000 Metarepresentation in an evolutionary perspective. In Metarepresentations: A Multidisciplinary Perspective, D. Sperber (ed.), 117–138. Oxford: Oxford University Press. Tyler, A., and V. Evans 2003 The Semantics of English Prepositions. Cambridge: Cambridge University Press. Talmy, L. 2000 Toward a Cognitive Semantic. Volume I: Concept Structuring Systems. Cambridge, MA.: MIT Press. 2001 Toward a Cognitive Semantic. Volume II: Typology and Process in Concept Structuring. Cambridge, MA: MIT Press. Vandeloise, C. 1999 Spatial Prepositions: A Case Study from French. Chicago: University of Chicago Press. 1986. Yu, N. 1998 The Contemporary Theory of Metaphor: A Perspective from Chinese. Amsterdam: John Benjamins.

Subject index

abstract  4–5, 7, 10, 12, 40, 43, 66, 86, 153, 210, 269, 278–279, 282, 294, 305, 316, 359–362, 377, 434, 440, 445–446, 457, 459–460, 466, 470 abstraction  33, 361, 364, 374 abstractness  4, 10, 18, 352–353, 358, 361–363 actor-observer  190, 200 –– bias  351, 356, 360 adjective  9, 11, 14, 17–18, 134–136, 138, 140, 144, 146, 148, 157, 174, 274, 280, 336–337, 350, 352–354, 362, 368, 398, 403 adverb  9, 17, 135, 336–337, 384, 447 adverbial  137, 208, 385 affect  8–10, 18–19, 334, 344, 396, 406–415, 417–425 affection  13, 35, 205, 401–402 affective  19, 171, 269–270, 275–276, 281, 283, 325, 329, 340, 368, 395–397, 398, 405, 409–411, 413, 421 appraisal  396–397, 406–407, 409–410, 418, 428 –– theory  19, 406–407, 409, 451 attention  4, 6, 50, 68, 83–84, 134, 196, 293 attitude  1, 8, 156, 183, 188, 192, 195, 218–219, 282, 326, 336, 401–402, 406–407, 409, 425, 434, 439, 441, 443–444, 447–449 attribution  5–6, 18–19, 183, 189-.190, 195, 273–274, 280, 282–283, 349–365, 369, 423, 434 –– error  189, 190, 351 –– theory  18, 349 attributional  18, 20, 351–358, 364 belief  5, 6, 12, 16, 30, 129, 190, 195–196, 269, 272, 283, 326, 335, 350, 361, 396

body  277, 292, 294, 397, 398, 455, 457, 461–464, 467 categorization  3, 16, 167, 169–171, 179, 206–207, 212, 215, 217–218, 372, 374 –– social  203, 210–211 category  4, 14, 18, 86–87, 104, 109, 166, 168–169, 212–213, 230, 255, 270, 352, 254, 362, 375, 434, 464, 468 –– social  1, 4, 204, 209, 270, 354 cognition –– collective 13, 15, 18, 163–164, 166, 175, 451 –– distributed  5, 164, 289–290, 292–293, 304, 307, 315 –– extended  17, 289, 293, 302, 313, 315–318 –– social  2–3, 7–8, 12–13, 15–16, 25, 34, 47, 49, 52, 72, 127–129, 163, 166, 173, 183, 185, 204–207, 224, 231, 247, 256, 267–269, 290, 293, 298, 318–319, 372–373, 375, 425, 451, 464 cognitive –– linguistics  1, 14, 47–48, 52–53, 79–80, 82, 128–129, 166–168, 292, 371, 373, 377, 389, 455, 457–459 –– model  16, 204–210, 212–217, 223–224, 226, 268, 270–271, 273, 277, 282–284 collaboration  6, 18, 48, 304, 308, 312, 372, 389 common ground  47, 215, 219, 221, 224, 335 concept –– social  4–5, 8, 14, 205, 466, 469 conceptualization  13–16, 66, 69, 79, 86, 88, 114–115, 168, 171–174, 176, 204, 206–207, 209, 211–213, 215, 226–227, 230–232, 294

474   Subject index –– cultural  166–170, 172, 174–176 –– social  209, 211–212, 215, 221, 227, 230–231 conversation analysis  2, 48–49, 206, 243–244 cooperative principle  1, 12, 184–185, 191, 195 corpus linguistics  378 discourse analysis  47, 52, 79, 82–83, 414 –– critical  (CDA)  17, 283 discursive  16, 48–49, 71, 203–204, 208, 214–215, 217, 219, 224–225, 227, 231, 233, 267, 269–271, 274, 277, 282, 284–285, 377, 395–397, 399, 401–402, 405–406, 412–414, 418, 421, 425, 446 embodied  47, 54, 57, 69, 168, 213, 292, 302, 319, 456, 466 embodiment  51, 54, 69–70, 167 embody  283, 302 emotion  2, 6, 8–11, 18–19, 167, 171, 276, 297, 329, 331, 333–334, 374, 383, 395–399, 401–407, 409–415, 417, 420–425, 448–451 emotional  40, 112, 128, 166, 171, 193, 212, 270, 276, 289, 298, 331, 333, 336, 355–358, 395–399, 402–403, 405–407, 409–412, 416, 418–423, 425, 434, 447, 449 ethnomethodology  48–49, 241, 243 evaluation  17, 219, 227, 242, 247–249, 252–254, 274, 329, 334, 336, 376, 382, 385–387, 397, 402, 406–407, 418, 421–323 experiental  18, 167, 242, 373–374, 380, 451, 468 expressive  8, 11, 18–19, 298, 338, 387, 434, 448 figurative language  2, 172, 185, 298, 320 frame  7, 9, 18, 60, 82, 86, 128–129, 131, 159, 184, 212, 217, 268, 355–357, 376–377, 382, 387–390, 411–412

gestural  40, 45, 57, 72 gesture  14, 44, 49, 52, 54, 56–58, 63–64, 66, 69, 71, 398 grammar  15, 168, 170, 295, 372, 377, 379 –– cognitive  9, 459 –– construction  372, 377, 379, 389 grammatical  15, 34, 36, 171, 273, 379–380, 385, 458 humor  6, 16, 227, 296, 298, 317, 326, 331–333 humorous  206, 329, 331, 333–334, 336 image schema  82, 455, 457, 459–461, 467, 470 implicature  11–12, 15, 173, 183–185, 191, 194, 196–198, 200, 208, 217–218, 371, 445 inference  3, 185, 187, 194, 196–197, 212, 221, 354–358, 360, 362, 365, 371, 402 inferencing  194, 197, 404 inferential  191, 206 inferability  6, 17, 339 interpersonal  11–12, 15–16, 18–19, 87, 128–130, 144, 183, 197, 204, 241, 268, 306, 313, 318, 335, 352, 364, 367, 397–398, 406, 413, 423, 433–434, 441, 444, 446, 448, 450–451 intersubjective  8, 64, 66, 68, 72, 237, 240–241, 257 intrapersonal  86–88, 94, 195–196 irony  9, 11, 15, 17, 54, 205, 229–230, 295, 291, 301, 323, 325–340 knowledge structure  5, 15, 210, 372, 379, 382, 396 lexical  17, 26, 34, 64, 134, 143–144, 146, 157, 168, 175, 208, 229–230, 336–337, 343, 352, 371–372, 376–377, 379–380, 385–386, 389, 399, 407, 413, 437, 450, 458, 460, 469 lexicon  167, 175, 352–353, 355, 357, 371–372, 380

Subject index   475 manipulate  188, 196, 204, 206, 231, 295, 302, 316, 359, 362, 464 manipulation  207 manipulative  349, 356, 364 maxim  15–16, 184.185, 188, 191–193, 195, 197, 326, 331, 360–361, 419 memory  3–4, 6, 13, 128, 168, 172, 175–176, 183, 192, 196, 205, 239, 292, 360, 362–363, 372, 374–375 –– collective  5, 7, 10, 15, 163, 168–169, 374 –– semantic  12, 211 –– working  110, 297–298 mental model  4, 210, 302 mental space  4, 16, 82, 204, 209–212, 223–224, 226, 229–231 metaphor  6, 10, 15, 17, 19, 33, 68–69, 81–82, 148, 166–167, 171, 273, 276–277, 283, 292, 295–296, 298, 300, 309, 311, 327, 329, 331–332, 334, 398, 407, 411, 419, 427, 445, 450, 457–461, 463–468, 470 metaphoric/al  8, 18–19, 268–269, 273–276, 282–283, 292, 457–461, 466–470 metonymic  9–10, 19, 270, 273–274, 276, 283, 434, 445–446, 449–450 metonymy  17, 19, 22, 82, 269, 273, 283, 411, 445–447 modality  17, 56–58, 273, 275–276, 281–283, 398, 406, 413, 418–419

pragmatic  50, 52, 172–173, 183, 181, 195, 203, 340–341, 350, 352–353, 356, 371, 375–376, 379–380, 383, 389–390, 405, 419, 446, 448–449, 451 pragmatics  1, 11–12, 47, 172, 183, 187–190, 195–197, 210, 353 priming  119, 349, 357, 362–263, 372, 378–380, 383–384, 387, 389, 404 procedural  12, 50, 129, 246, 448–449, 451 pronoun  9, 15, 17, 169–171, 273, 275, 314, 357, 366, 398 prototype  7, 330, 373 prototypical  88, 397, 457 psychological  14–15, 18, 54, 60–61, 68, 79, 82, 105–106, 238, 325, 342, 350–351, 353, 356, 372–373, 375, 382, 389, 400, 404–405, 414, 434, 445–447, 451, 453 psychology  3, 38, 42, 53, 58, 79, 84–86, 92, 105, 109–110, 117, 163, 191, 240, 350, 363, 407, 414, 451 –– clinical 128–129 –– cognitive  3, 7, 14, 20, 47–48, 50, 68–69, 128, 237, 239, 396 –– cultural  82–83, 107 –– discoursive  48–49, 396, 414, 425 –– social  1, 5, 353, 373, 398

nonliteral  2, 183, 185, 326–333, 336 nonverbal  7, 8, 14, 59, 195, 208, 349, 398

salience  4, 88, 155, 404, 466 salient  113, 152, 218, 221, 223, 270, 282, 390, 446 sarcasm  325–326, 328, 333–338 schema –– cultural  166, 168, 171, 173–175 –– interpersonal  128–130, 144 –– relational 129, 158 script  7, 227, 270, 283, 451 self  4–5, 8–9, 14–17, 19, 63, 70, 83–84, 110, 117, 128–129, 135, 153, 173–175, 208, 211, 230, 267–273, 275–276, 279–284, 297, 299, 316, 350, 352, 360,

paralinguistic  8, 17, 335–337, 398 perception  3–6, 8, 62, 69, 92, 111, 113, 117, 120, 131, 156, 163, 166, 183, 195, 200, 219, 221, 238–239, 242, 251, 254–256, 261, 263, 282, 296–299, 304, 318, 372, 374, 397, 433–434, 457 persuasion  188, 198, 201 persuasive  187, 276, 278, 332, 458 polysemy  380, 457

representativeness  133, 185–186, 188, 194, 200

476   Subject index 364, 376, 386, 405, 409, 443, 456–457, 463, 469 –– image  5, 204, 218, 230, 274, 320 –– -organising  15, 165, 176, 373 –– schema  5, 8, 15, 173 –– -serving bias  5, 284, 351, 356, 360–361 semantic  12, 16, 19, 50, 52, 56, 124, 128, 148, 208, 211, 213–215, 217, 223–227, 276, 350, 352–357, 359–360, 371, 373, 375–376, 379–380, 382, 384–390, 404–405, 429–430, 434, 436, 438–440, 448–451, 458, 462, 466 –– frame  82, 412, 457 semantics  9, 209–210, 353, 442, 444, 446 social –– category, see category –– cognition,  see cognition –– cognitive,  see cognitive –– concept,  see concept –– knowledge  7, 12, 18, 213, 372–374 –– psychology, see psychology –– role  128–129, 205, 212, 220, 224, 226–227, 283 –– schema, see schema socio-cognitive  204–205, 207, 217–218, 229, 231, 267–271, 273–275, 277, 280, 282–284, 396 spatial  8, 19, 206, 434, 438, 440, 446, 448, 456–469 speech act  31–32, 47, 52, 127, 172, 206, 208–209, 215, 218, 225, 362, 434, 440, 444, 447–450 –– verbs  19, 434, 441, 447, 450

stereotype  1, 186, 193, 195, 211, 318, 344, 359–361, 365–367, 373 strategy  227, 230, 272, 279, 329–330, 396, 419 subject (grammatical)  9, 11, 18, 352–358, 369, 385, 436 subjective  19, 83, 106, 112, 193, 208, 350, 352, 414, 417–420, 424, 444, 448–449 syntactic  9, 18, 54, 73, 168, 376, 378–379, 385, 388–389, 398, 434, 436, 444 syntax  14, 44, 52 temporal  86, 197, 273, 352–353, 369, 398, 457, 461 tense  8, 17, 158, 273, 276, 281–282, 410 thought  14, 26–28, 32–33, 35–39, 43, 51, 64, 69–70, 79, 83, 97, 135, 145, 148, 155, 165–167, 171, 173, 175–176, 191, 195, 199, 239–240, 256, 258, 263, 291, 349, 375, 409–410, 433– 434, 436, 450 trait  3, 17, 41, 43, 173, 175, 205, 273, 283, 334, 351, 354, 357–360, 362, 364–367, 379 verb  8–9, 17–19, 21, 35, 128, 170–172, 208, 276, 280–281, 350, 352, 354–357, 359, 362–364, 368, 383–384, 388, 398, 434, 436–451, 466, 469 verbal  8–9, 11, 15, 17, 21, 49, 51, 56, 59, 106, 195–196, 205, 227, 238–239, 245, 275–276, 325–337, 339–343, 349, 362, 376, 398