Pragmatics and Discourse | A Resource Book for Students | Joan Cutting

Book

Pragmatics and Discourse

ABSTRACT

Pragmatics and Discourse, 2^nd edition:

has been revised and reorganised to place more emphasis on pragmatics
covers the core areas of the subject: context and co-text, Speech Act Theory, Conversation Analysis, Exchange Structure, Interactional Sociolinguistics, the Cooperative Principle, Politeness Theory and extends to more applied areas: Corpus Linguistics & Communities of Practice, and Intercultural Pragmatics, Interlanguage Pragmatics & language learning
draws on a wealth of texts: from Bend it Like Beckham and The Motorcycle Diaries to political speeches, newspaper extracts and blogs.
provides classic readings from the key names in the discipline, from Sperber and Wilson to Fairclough, Wodak and Gumperz
is accompanied by a supporting website

Key features of the new edition include: two new strands on Corpora & Communities and Culture & Language Learning; the merging of two strands on Context and Co-text; new material from speaker-based cognitive linguistics; updated references; and fresh examples and exercises.

Written by an experienced teacher and author, this accessible textbook is an essential resource for all students of English language and linguistics.

TABLE OF CONTENTS

chapter |1 pages

Common knowledge, then – that is, ‘what everyone knows’ – is necessarily something that is culture-loaded and varies from group to group. Much of what everyone knows is also either scientiﬁcally unwarranted or very superﬁcial. For example, there are numerous stereotypes in this kind of knowledge – ideas we have about the ‘typical’ behaviour and characteristics of people or objects. But that should not surprise us, because, after all, that is essentially what norms them-selves are in one sense – abstractions based on certain kinds of experiences which apparently typify some kind of general behaviour. Many people go through life holding the view that common knowledge and stereotypes characterize a sort of truth about the world; others are somewhat more critical and conscious of the complexities that lie behind such a simple belief. What we must not assume, how-ever, is that common knowledge is always false and stereotyping is always bad; social harmony is possible only if there are things we can agree on, and there are measures of agreement. What may be important is how ﬁxed are the measures any society uses, not the existence of the measures themselves. In periods of rapid social change old norms and stereotypes come under attack at a time when new ones are not available, so it is not surprising that confusion results. Linguistic behaviour at such times tends to reﬂect the disorder. Some strive to preserve the old ways, as conservative factions in Greece did in the 1960s to reimpose a ‘high’ variety of Greek. Others want to create a new set of condi-tions, for example, to rid a language of a tu–vous distinction in address forms, as did both the French and Russian revolutionaries (but eliminating the vous form in one case and the tu form in the other). Eventually new norms emerge, new appear-ances, new conventions, and new ways of using language to express these new norms with all the advantages, and disadvantages, of the old, offering as they do a way of constructing a certain kind of reality as well as providing blinkers which make other realities somewhat inaccessible to view. One consequence of all this is that we must set limits on the amount of trust we place in others and in our view of the world. Similarly, in conversation we should not trust absolutely: that is too severe a demand to make both of our-selves and of others. Those who give their absolute trust to others are almost cer-tain at one time or another to be disappointed. But we must also be aware that distrust cannot be the norm either, for a climate in which everyone distrusts every-one else would prohibit entirely all hope of mutually beneﬁcial social contact. Therefore, we must err at all times on the side of trust. Unfortunately, those who would deceive us know that too, and, having conﬁdence in their ability to exploit this basic social need, proceed to do so, often with impunity. For any particular conversation it is also possible to show that there are differences between the parties in the speciﬁc things that they know in contrast to the kinds of background knowledge that they share. No two people have identical backgrounds, so in any conversation the participants will have different kinds of knowledge about almost any topic that is likely to be mentioned. If only two people, Fred and Sally, are involved, there will be certain matters known to both, some because ‘everybody knows such things’ and others because both Fred and Sally happen to know them. Then there will be matters known to only one

chapter |1 pages

of the speakers, so that Fred will know something that Sally does not know, or Sally something that Fred does not know. In addition, there will be partly known information: Fred or Sally, or both, may partly know something or know parts of something, but not necessarily the same parts. And Fred or Sally, or both again, may believe that the other knows something that the other actually does not know. As we can see, there are numerous possible permutations in who knows what, who believes who knows what, and so on. Again, there are predictable conse-quences: conversation can proceed only on the basis that the participants share a set of beliefs, that is, certain things must be known to all parties; others may be known; some will have to be explained; questions may be asked for clariﬁcation; difﬁculties will be negotiated or cleared up somehow; people will be understand-ing and tolerant; and the various processes that are involved will be conducted decently. If only one participant in a conversation refuses to subscribe to these beliefs and to conduct himself or herself accordingly, the others will become irri-tated, confused, or frustrated, and may well abandon any attempt to continue what they have begun. Since most participants in a conversation usually do share a certain amount of background knowledge about ‘proper’ behaviour and the ‘right’ way to do things, much of what they say can be understood if we, too, are familiar with the know-ledge they share. Their references to places, times, and events, and their accounts and descriptions are related to what they know and what they believe the others know. A participant in a conversation must believe that he or she has access to the same set of reference points that all the other participants have access to; all he or she needs do in conversing is use those points for orientation, and listeners will comprehend. And such a belief is largely justiﬁed. What is hardly ever necessary in a conversation is to begin at the very beginning of anything and to treat everyone and everything as unique and somehow without antecedents. In a trivial sense every occasion is unique, but procedures exist which minimize novelty and maximize normality – accepted ways of asking and giving directions, rules for regulating who speaks to whom and about what, and basic principles for conducting yourself, for example, with complete strangers. A conversation between familiars offers a very special mix of knowledge. There are matters in it which the parties know but are reluctant to refer to directly, although they may allude to them if necessary. There are matters which are not in the conversation by reason of the fact that they are deliberately avoided – their absence is conspicuous. And then there are the actual topics of the conversation. However, these topics are not introduced logically, as it were, but rather in a variety of ways according to the needs of the individuals and of the occasion, with each participant willing to let a topic emerge as seems natural at the time in the expectation that its various bits and pieces will hold together. In general conversation with others it is ordinary, everyday, ‘commonsense’ knowledge that we assume they share with us. In certain circumstances, as between professionals, we can also assume a sharing of specialized knowledge. We must always take great care when we refer to items outside these shared areas. We cannot rely on others knowing what we know. They may not even share the same

chapter |4 pages

assumptions about what it means to ‘know’ something. A physicist’s knowledge of matter is different from a lay person’s, and an actor’s view of character is unlikely to be the same as that of a psychiatrist. Explanations may well become necessary, and they may not be easily provided. Brieﬁng is one kind of explain-ing behaviour in such circumstances. But a recurrent difﬁculty is knowing just how much to say on a particular occasion and then judging how successful we have been in saying it. This is particularly crucial if we then proceed to treat this ‘new’ information we supply henceforth as part of our listeners’ everyday knowledge. It may not be easily incorporated into existing knowledge, as anyone who has ever taught well knows, for it is one thing to teach something and quite another to learn it.

chapter |1 pages

process to swamp them. Although our ﬁndings do not support the distinction between macrostructure and microstructure, at least as it is usually deﬁned, the implica-tions of the existence of central sentences are that writers who see what they have to say in terms of a series of interconnected macropropositions are more likely to succeed in producing highly valued writing than those who make it up as they go along. One thing seems certain: the traditional advice to avoid repetition needs to be couched with special care if it is not to interfere actively with the development of mature writing skills. The advice grew out of two quite reasonable worries. First, when an inexperienced writer does not know what else to say, they some-times resort to restating what they have already said. Nothing in this book should have shaken the reader’s conviction that this is an unsatisfactory practice; the existence of patterns of lexis in text is not to be interpreted as an incitement to padding. Second, especially among less experienced writers, limitations of vocabulary and ignorance of the means whereby one can repeat in a language may lead a learner to juxtapose the same lexical item clumsily in adjacent sentences. Again, it has been noted in earlier chapters that the tendency for adjacent sentences to bond is not great; the reason is that, in English, care is usually taken to avoid the clumsy juxtaposition just referred to. So, here too, the advice as traditionally given still stands. But it cannot rest there. Reasonable as the worries concerning repetition may be, the advice to avoid repetition may be harmful unless it is immediately sup-plemented by something more. To begin with, if a learner is to avoid clumsiness, he or she must be taught how to avoid it. One of the most important ways is by means of complex repetition. So, in the ﬁrst sentence of the previous paragraph, I used the lexical item clumsily; in the following sentence it has become clumsy while in the third sentence of this paragraph it appears as clumsiness. Similarly, juxtapose becomes juxtaposition, and repeat becomes repetition. There is nothing contrived about these examples; my practice is that of most writers need-ing to repeat without making the repetition obtrusive. Stotsky (1983) comments that ‘an increase in the use of morphologically complex words [i.e. complex rep-etition], rather than repetition of a simple word or the use of a cumbersome para-phrase, may be an important index of growth.’ If we need to protect our learners against this aspect of avoiding repetition, still more must we protect them against misuse of the counsel to avoid padding. Learners should not be encouraged to say the same thing over and over again, but they should be advised to make connections between what they are currently saying and what they said before. There should, in non-narrative text, be some relationship between sentences at a distance from each other. What this means for learners is that they need to take time out of grappling with the difﬁculties of composing the sentence they are currently working on to consider its relationship with what they have already written. This may impose on the writer an additional burden but it also relieves him or her of at least some of the task of lexical selec-tion. Indeed, knowledge that it is legitimate to reuse in different combinations

chapter |1 pages

lexical items already brought into play, may actually serve to lighten some aspects of the writing task. The advice a learner needs will vary from person to person and from group to group. For some the main need may be to avoid going around in circles; such learners will need reminding that new information always accompanies repetition in mature writing. For others the need may be to prevent the text drifting; such learners will need telling that repetition usually provides the grounding for new information. In either case, materials may need developing that give the student practice in bonding back to earlier sentences. The well-tried strategy of supply-ing the learners with an incomplete text and asking them to complete it may be adaptable to this purpose; indeed, there is no reason why more than one text organizing principle at a time might not be practised in this way. All of this is rather obvious, but it is easy for second language learners to lose sight of the applicability of what they already know in their ﬁrst language when faced with the twin tasks of selecting appropriate lexis and avoiding ungram-maticality. As so often is the case with language learning, encouraging the learners to transfer a skill from their ﬁrst language is half the battle.

chapter |3 pages

Imprecise references, pronominalizations, and a lack of feed-back are thus also characteristics of these texts. Typically they are produced unconstrained by a need for self-justiﬁcation. Consequently, they cannot really be considered stories, as the latter are normally conceived. The relevance of the story is never really explained, background knowledge and hints as to orientation are absent, no frame is avail-able within which to embed the news item. As a result there is almost no possi-bility of ‘updating’ it. This situation was conﬁrmed by Larsen, who analysed the intelligibility of Danish news spots (Larsen 1983) (see Wodak/Lutz 1986: 202ff ). This suggests that it is often impossible to integrate new information into already available knowledge as long as the present form of providing news items pertains. Little or no acquisition of new knowledge takes place. As Larsen writes:

chapter |1 pages

warning; or, on a different plane, referring to people or things, presupposing the existence of people or things or the truth of propositions, and implicating mean-ings which are not overtly expressed. The idea of uttering as acting is an impor-tant one, and it is also central to CLS in the form of the claim, that discourse is social practice. The main weakness of pragmatics from a critical point of view is its individ-ualism: ‘action’ is thought of atomistically as emanating wholly from the individ-ual, and is often conceptualized in terms of the ‘strategies’ adopted by the individual speaker to achieve her ‘goals’ or ‘intentions’. This understates the extent to which people are caught up in, constrained by, and indeed derive their individual iden-tities from social conventions, and gives the implausible impression that conven-tionalized ways of speaking or writing are ‘reinvented’ on each occasion of their use by the speaker generating a suitable strategy for her particular goals. And it correspondingly overstates the extent to which people manipulate language for strate-gic purposes. Of course, people do act strategically in certain circumstances and use conventions rather than simply following them; but in other circumstances they do simply follow them, and what one needs is a theory of social action – social practice – which accounts for both the determining effect of conventions and the strategic creativity of individual speakers, without reducing practice to one or the other. The individuals postulated in pragmatics, moreover, are generally assumed to be involved in cooperative interactions whose ground rules they have equal con-trol over, and to which they are able to contribute equally. Cooperative interac-tion between equals is elevated into a prototype for social interaction in general, rather than being seen as a form of interaction whose occurrence is limited and socially constrained. The result is an idealized and Utopian image of verbal inter-action which is in stark contrast with the image offered by CLS of a sociolinguistic order moulded in social struggles and riven with inequalities of power. Pragmatics often appears to describe discourse as it might be in a better world, rather than discourse as it is. Pragmatics is also limited in having been mainly developed with reference to single invented utterances rather than real extended discourse, and central notions like ‘speech act’ have turned out to be problematic when people try to use them to analyse real discourse. Finally, Anglo-American pragmatics bears the scars of the way in which it has developed in relation to ‘linguistics proper’. While it has provided a space for investigating the interdependence of language and social con-text which was not available before its inception, it is a strictly constrained space, for pragmatics tends to be seen as an additional ‘level’ of language study which ﬁlls in gaps left by the more ‘core’ levels of grammar and semantics. Social con-text is acknowledged but kept in its place, which does it less than justice.

chapter |1 pages

of comprehension involved, as well as of processes of production, has been under-taken by cognitive psychologists, and workers in artiﬁcial intelligence concerned with the computer simulation of production and comprehension. From the per-spective of CLS, the most important result of work on comprehension is the stress which has been placed upon its active nature: you do not simply ‘decode’ an utter-ance, you arrive at an interpretation through an active process of matching fea-tures of the utterance at various levels with representations you have stored in your long-term memory. These representations are prototypes for a very diverse collection of things – the shapes of words, the grammatical forms of sentences, the typical structure of a narrative, the properties of types of object and person, the expected sequence of events in a particular situation type, and so forth. Some of these are linguistic, and some of them are not. Anticipating later discussion, let us refer to these prototypes collectively as ‘members’ resources’, or MR for short. The main point is that comprehension is the outcome of interactions between the utterance being interpreted, and MR. Not surprisingly, cognitive pyschology and artiﬁcial intelligence have given little attention to the social origins or signiﬁcance of MR. I shall argue later that attention to the processes of production and comprehension is essential to an under-standing of the interrelations of language, power and ideology, and that this is so because MR are socially determined and ideologically shaped, though their ‘common sense’ and automatic character typically disguises that fact. Routine and unselfconscious resort to MR in the ordinary business of discourse is, I shall sug-gest, a powerful mechanism for sustaining the relations of power which ultimately underlie them.

chapter |2 pages

described, so the question becomes why business executives and other power-holders systematically avoid too much overt marking of their power.

chapter |1 pages

The student reports that the interview that followed was stiff and quite unsatis-factory. Being black himself, he knew that he had ‘blown it’ by failing to recognize the signiﬁcance of the husband’s speech style in this particular case. The style is that of a formulaic opening gambit used to ‘check out’ strangers, to see whether or not they can come up with the appropriate formulaic reply. Intent on follow-ing the instructions he had received in his methodological training and doing well in what he saw as a formal interview, the interviewer failed to notice the hus-band’s stylistic cues. Reﬂecting on the incident, he himself states that, in order to show that he was on the husband’s wave-length, he should have replied with a typically black response like ‘Yea, I’ma git some info’ (I’m going to get some information) to prove his familiarity with and his ability to understand local verbal etiquette and values. Instead, his Standard English reply was taken by the husband as an indication that the interviewer was not one of them and, perhaps, not to be trusted. The opener ‘So y’re gonna check out ma ol lady’ is similar to the ‘Ahma git me a gig’ discussed elsewhere. Both are formulaic phrases identiﬁable through co-occurrent selections of phonological, prosodic, morphological and lexical options. Linguists have come to recognize that, as Fillmore (1976) puts it, ‘an enormous amount of natural language is formulaic, automatic and rehearsed, rather than pro-positional, creative or freely generated.’ But it must be emphasized that although such formulas have some of the characteristics of common idioms like kick the bucket and spill the beans, their meaning cannot be adequately described by lexical glosses. They occur as part of routinized interactive exchanges, such as Goffman describes as ‘replies and responses’ (1981). Their use signals both expectations about what is to be accomplished and about the form that replies must take. They are similar in function to code switching strategies. Like the latter they are learned by interacting with others in institutionally deﬁned networks of relationships. Where these relationships are ethnically speciﬁc they are often regarded as markers of ethnic background. But, as our example shows, their use in actual encounters is ultimately determined by activity speciﬁc pre-suppositions so that failure to react is not in itself a clear sign of ethnic identity. Basically, these formulaic phrases reﬂect indirect conversational strategies that make conditions favorable to estab-lishing personal contact and negotiating shared interpretations. . . .

chapter |1 pages

Although it might seem at ﬁrst glance that contextualization cues are surface phe-nomena, their systematic analysis can lay the foundation for research strategies to gain insights into otherwise inaccessible symbolic processes of interpretation. On the practical level, the study of conversational inference may lead to an explanation for the endemic and increasingly serious communication problems that affect private and public affairs in our society. We can begin to see why individ-uals who speak English well and have no difﬁculty in producing grammatical English sentences may nevertheless differ signiﬁcantly in what they perceive as meaning-ful discourse cues. Accordingly, their assumptions about what information is to be conveyed, how it is to be ordered and put into words and their ability to ﬁll in the unverbalized information they need to make sense of what transpires may also vary. This may lead to misunderstandings that go unnoticed in the course of an interaction, but can be revealed and studied empirically through conversational analysis. The main purpose of earlier chapters was to illustrate the nature of the cues and the inferential mechanisms involved. To that end, the discussion largely relied on examples of brief encounters. Miscommunications occurring in such brief encoun-ters are annoying and their communicative effect may be serious. But the social import of the phenomena in question and their bases in participants’ cultural back-ground is most clearly revealed through case studies of longer events. The fol-lowing two chapters present in depth analyses of two such events. To begin with, let me give one more brief example to illustrate the scope of the analysis and the subconscious nature of the interpretive processes involved. In a staff cafeteria at a major British airport, newly hired Indian and Pakistani women were perceived as surly and uncooperative by their supervisor as well as by the cargo handlers whom they served. Observation revealed that while rela-tively few words were exchanged, the intonation and manner in which these words were pronounced were interpreted negatively. For example, when a cargo handler who had chosen meat was asked whether he wanted gravy, a British assistant would say ‘Gravy?’ using rising intonation. The Indian assistants, on the other hand, would say the word using falling intonation: ‘Gravy.’ We taped relevant sequences, includ-ing interchanges like these, and asked the employees to paraphrase what was meant in each case. At ﬁrst the Indian workers saw no difference. However, the English teacher and the cafeteria supervisor could point out that ‘Gravy,’ said with a falling intonation, is likely to be interpreted as ‘This is gravy,’ i.e. not interpreted as an offer but rather as a statement, which in the context seems redundant and con-sequently rude. When the Indian women heard this, they began to understand the reactions they had been getting all along which had until then seemed incompre-hensible. They then spontaneously recalled intonation patterns which had seemed strange to them when spoken by native English speakers. At the same time, super-visors learned that the Indian women’s falling intonation was their normal way of asking questions in that situation, and that no rudeness or indifference was intended. After several discussion/teaching sessions of this sort, both the teacher and the cafeteria supervisor reported a distinct improvement in the attitude of the Indian workers both to their work and to their customers. It seemed that the

chapter |3 pages

Indian workers had long sensed they had been misunderstood but, having no way of talking about this in objective terms, they had felt they were being discrimi-nated against. We had not taught the cafeteria workers to speak appropriate English; rather, by discussing the results of our analysis in mixed sessions and focusing on context bound interpretive preferences rather than on attitudes and stereotypes, we have suggested a strategy for self-diagnosis of communication difﬁculties. In short, they regained conﬁdence in their own innate ability to learn. The ﬁrst of the longer case studies examines excerpts from an interview– counselling session recorded in an industrial suburb in London. The participants are both educated speakers of English; one is a Pakistani teacher of mathematics, who although born in South Asia went to secondary school and university in England. The other is a staff member of a center funded by the Department of Employment to deal with interethnic communication problems in British industry. The teacher has been unable to secure permanent employment and having been told that he lacks communication skills for high school teaching, he has been referred to the center. While both participants agree on the general deﬁnition of the event as an interview–counselling session, their expectations of what is to be accomplished, and especially about what needs to be said, differ radically. Such differences in expectation are of course not unusual even where conversationalists have similar cultural backgrounds. Conversations often begin with an introductory phase where common themes are negotiated and differences in expectation adjusted. What is unusual about this situation is that participants, in spite of repeated attempts at adjustment over a period of more than an hour, utterly fail to achieve such nego-tiation. Our analysis concentrates on the reasons for this failure and shows how it is based on differences in linguistic and socio-cultural knowledge.

chapter |1 pages

from its independently observed or inferred effects. This pattern of inference is generally not available to an audience trying to recognise a communicator’s infor-mative intention. As we have seen, the informative effects of communication are normally achieved, if at all, via recognition of the informative intention. Hence, it seems, the audience cannot ﬁrst observe or infer these effects, and then use them to infer the informative intention. However, the problem is not that it is hard to come up with hypotheses about what the communicator might have intended to convey: it is that too many hypo-theses are possible. Even a linguistic utterance is generally full of semantic ambi-guities and referential ambivalences, and is open to a wide range of ﬁgurative interpretations. For non-coded behaviour there is, by deﬁnition, no predetermined range of information it might be used to communicate. The problem, then, is to choose the right hypothesis from an indeﬁnite range of possible hypotheses. How can this be done? First, it is easy enough to infer that a certain piece of behaviour is communicative. Communicative behaviour has at least one characteristic effect which is achieved before the communicator’s informative intention is recognised: it overtly claims the audience’s attention. Grice’s fundamental idea in his William James Lectures is that once a certain piece of behaviour is identiﬁed as communicative, it is reasonable to assume that the communicator is trying to meet certain general standards. From knowledge of these general standards, observation of the communicator’s behaviour, and the con-text, it should be possible to infer the communicator’s speciﬁc informative inten-tion. Grice, talking only of verbal communication, argues,

chapter |1 pages

Maxim of relation Be relevant. Maxims of manner Supermaxim: Be perspicuous. 1 Avoid obscurity of expression. 2 Avoid ambiguity. 3 Be brief (avoid unnecessary prolixity). 4 Be orderly. This account of the general standards governing verbal communication makes it possible to explain how the utterance of a sentence, which provides only an incomplete and ambiguous representation of a thought, can nevertheless express a complete and unambiguous thought. Of the various thoughts which the sentence uttered could be taken to represent, the hearer can eliminate any that are incom-patible with the assumption that the speaker is obeying the co-operative principle and maxims. If only one thought is left, then the hearer can infer that it is this thought that the speaker is trying to communicate. Thus, to communicate efﬁci-ently, all the speaker has to do is utter a sentence only one interpretation of which is compatible with the assumption that she is obeying the co-operative principle and maxims. Recall, for instance, our example (16)–(18):

chapter |1 pages

In just the same way, if Peter is aware of (35), he could infer conclusion (36):

chapter |1 pages

Grice’s view of implicature raises even more basic questions. What is the ratio-nale behind the co-operative principle and maxims? Are there just the nine max-ims Grice mentioned, or might others be needed, as he suggested himself ? It might be tempting to add a maxim every time a regularity has to be accounted for. However, this would be entirely ad hoc. What criteria, then, do individual maxims have to meet? Could the number of maxims be not expanded but reduced? How are the maxims to be used in inference? Grice himself seems to think that the hearer uses the assumption that the speaker has observed the maxims as a premise in inference. Others have tried to reinterpret the maxims as ‘conver-sational postulates’ (Gordon and Lakoff 1975), or even as code-like rules which take semantic representations of sentences and descriptions of context as input, and yield pragmatic representations of utterances as output (Gazdar 1979). The ﬂavour of such proposals can be seen from the following remarks:

chapter |1 pages

more explicit and systematic than the intuitive reconstructions supplied by un-sophisticated speakers, the analyses of implicature which have been proposed by pragmatists have shared with these intuitive reconstructions the defect of being almost entirely ex post facto. Given that an utterance in context was found to carry particular implicatures, what both the hearer and the pragmatic theorist can do, the latter in a slightly more sophisticated way, is to show how in very intuitive terms there was an argu-ment based on the context, the utterance and general expectations about the behaviour of speakers, that would justify the particular interpretation chosen. What they fail to show is that on the same basis, an equally convincing justiﬁcation could not have been given for some other interpretation that was not in fact chosen. There may be a whole variety of interpretations that would meet whatever standards of truthfulness, informativeness, relevance and clarity have been proposed or envis-aged so far. The theory needs improving at a fundamental level before it can be fruitfully applied to particular cases. In his William James Lectures, Grice put forward an idea of fundamental import-ance: that the very act of communicating creates expectations which it then exploits. Grice himself ﬁrst applied this idea and its elaboration in terms of the maxims to a rather limited problem of linguistic philosophy: do logical connectives (‘and’, ‘or’, ‘if . . . then’) have the same meaning in natural languages as they do in logic? He argued that the richer meaning these connectives seem to have in natural languages can be explained in terms not of word meaning but of implicature. He then suggested that this approach could have wider applications: that the task of linguistic semantics could be considerably simpliﬁed by treating a large array of problems in terms of implicatures. And indeed, the study of implicature along Gricean lines has become a major concern of pragmatics. We believe that the basic idea of Grice’s William James Lectures has even wider implications: it offers a way of developing the analysis of inferential communication, suggested by Grice him-self in ‘Meaning’ (1957), into an explanatory model. To achieve this, however, we must leave aside the various elaborations of Grice’s original hunches and the sophisticated, though empirically rather empty debates they have given rise to. What is needed is an attempt to rethink, in psychologically realistic terms, such basic questions as: What form of shared information is available to humans? How is shared information exploited in communication? What is relevance and how is it achieved? What role does the search for relevance play in communication? It is to these questions that we now turn.

chapter |1 pages

conceptually vague notion of ‘shared information’. We will discuss in what sense humans share information, and to what extent they share information about the information they share. All humans live in the same physical world. We are all engaged in a lifetime’s enterprise of deriving information from this common environment and construct-ing the best possible mental representation of it. We do not all construct the same representation, because of differences in our narrower physical environments on the one hand, and in our cognitive abilities on the other. Perceptual abilities vary in effectiveness from one individual to another. Inferential abilities also vary, and not just in effectiveness. People speak different languages, they have mastered dif-ferent concepts; as a result, they can construct different representations and make different inferences. They have different memories, too, different theories that they bring to bear on their experience in different ways. Hence, even if they all shared the same narrow physical environment, what we propose to call their cognitive environments would still differ. To introduce the notion of a cognitive environment, let us consider a parallel case. One human cognitive ability is sight. With respect to sight, each individual is in a visual environment which can be characterised as the set of all phenomena visible to him. What is visible to him is a function both of his physical environ-ment and of his visual abilities. In studying communication, we are interested in conceptual cognitive abilities. We want to suggest that what visible phenomena are for visual cognition, mani-fest facts are for conceptual cognition. Let us deﬁne:

chapter |1 pages

in a cognitive environment if the environment provides sufﬁcient evidence for its adoption, and as we all know, mistaken assumptions are sometimes very well evidenced. Anything that can be seen at all is visible, but some things are much more visible than others. Similarly, we have deﬁned ‘manifest’ so that any assumption that an individual is capable of constructing and accepting as true or probably true is manifest to him. We also want to say that manifest assumptions which are more likely to be entertained are more manifest. Which assumptions are more manifest to an individual during a given period or at a given moment is again a function of his physical environment on the one hand and his cognitive abilities on the other. Human cognitive organisation makes certain types of phenomena (i.e. perceptible objects or events) particularly salient. For instance, the noise of an explosion or a doorbell ringing is highly salient, a background buzz or a ticking clock much less so. When a phenomenon is noticed, some assumptions about it are standardly more accessible than others. In an environment where the doorbell has just rung, it will normally be strongly manifest that there is someone at the door, less strongly so that whoever is at the door is tall enough to reach the bell, and less strongly still that the bell has not been stolen. The most strongly manifest assumption of all is the assumption that the doorbell has just rung, the evidence for which is both salient and conclusive. We will have more to say, in chapter 3, about the factors which make some assumptions more manifest than others in a given situ-ation. For the moment it is the fact rather than the explanation that matters. Our notion of what is manifest to an individual is clearly weaker than the notion of what is actually known or assumed. A fact can be manifest without being known; all the individual’s actual assumptions are manifest to him, but many more assumptions which he has not actually made are manifest to him too. This is so however weakly the terms ‘knowledge’ and ‘assumption’ are construed. In a strong sense, to know some fact involves having a mental representation of it. In a weaker sense, to say that an individual knows some fact is not necessarily to imply that he has ever entertained a mental representation of it. For instance, before read-ing this sentence you all knew, in that weak sense, that Noam Chomsky never had breakfast with Julius Caesar, although until now the thought of it had never crossed your mind. It is generally accepted that people have not only the know-ledge that they actually entertain, but also the knowledge that they are capable of deducing from the knowledge that they entertain. However, something can be manifest without being known, even in this virtual way, if only because some-thing can be manifest and false, whereas nothing can be known and false. Can something be manifest without being actually assumed? The answer must again be yes. Assumptions are unlike knowledge in that they need not be true. As with knowledge, people can be said to assume, in a weak sense, what they are capable of deducing from what they assume. However, people do not assume, in any sense, what they are merely capable of inferring non-demonstratively – that is, by some creative process of hypothesis formation and conﬁrmation – from what they assume. Although it presumably followed non-demonstratively from what you

chapter |2 pages

knew and assumed before you read this sentence that Ronald Reagan and Noam Chomsky never played billiards together, this was not, until now, an assumption of yours: it was only an assumption that was manifest to you. Moreover, some-thing can be manifest merely by being perceptible, and without being inferable at all from previously held knowledge and assumptions. A car is audibly passing in the street. You have not yet paid any attention to it, so you have no knowledge of it, no assumptions about it, even in the weakest sense of ‘knowledge’ and ‘assumption’. But the fact that a car is passing in the street is manifest to you. We will now show that because ‘manifest’ is weaker than ‘known’ or ‘assumed’, a notion of mutual manifestness can be developed which does not suffer from the same psychological implausibility as ‘mutual knowledge’ or ‘mutual assumptions’. To the extent that two organisms have the same visual abilities and the same physical environment, the same phenomena are visible to them and they can be said to share a visual environment. Since visual abilities and physical environments are never exactly identical, organisms never share their total visual environments. Moreover, two organisms which share a visual environment need not actually see the same phenomena; they are merely capable of doing so. Similarly, the same facts and assumptions may be manifest in the cognitive environments of two different people. In that case, these cognitive environments intersect, and their intersection is a cognitive environment that these two people share. The total shared cognitive environment of two people is the intersection of their two total cognitive environments: i.e. the set of all facts that are manifest to them both. Clearly, if people share cognitive environments, it is because they share physical environments and have similar cognitive abilities. Since physical envir-onments are never strictly identical, and since cognitive abilities are affected by previously memorised information and thus differ in many respects from one per-son to another, people never share their total cognitive environments. Moreover, to say that two people share a cognitive environment does not imply that they make the same assumptions: merely that they are capable of doing so. One thing that can be manifest in a given cognitive environment is a charac-terisation of the people who have access to it. For instance, every Freemason has access to a number of secret assumptions which include the assumption that all Freemasons have access to these same secret assumptions. In other words, all Freemasons share a cognitive environment which contains the assumption that all Freemasons share this environment. To take another example, Peter and Mary are talking to each other in the same room: they share a cognitive environment which consists of all the facts made manifest to them by their presence in this room. One of these facts is the fact that they share this environment. Any shared cognitive environment in which it is manifest which people share it is what we will call a mutual cognitive environment. In a mutual cognitive environ-ment, for every manifest assumption, the fact that it is manifest to the people who share this environment is itself manifest. In other words, in a mutual cognitive environment, every manifest assumption is what we will call mutually manifest. Consider, for example, a cognitive environment E shared by Peter and Mary, in which (41) and (42) are manifest:

chapter |1 pages

knowledge and adopting the weaker notion of mutual manifestness, we deprive ourselves of a certain type of explanation in the study of communication. Communication requires some degree of co-ordination between communica-tor and audience on the choice of a code and a context. The notion of mutual knowledge is used to explain how this co-ordination can be achieved: given enough mutual knowledge, communicator and audience can make symmetrical choices of code and context. A realistic notion of mutual manifestness, on the other hand, is not strong enough to explain such symmetrical co-ordination. However, before concluding that mutual manifestness is too weak after all, ask yourself what are the grounds for assuming that responsibility for co-ordination is equally shared between communicator and audience, and that both must worry, symmetrically, about what the other is thinking. Asymmetrical co-ordination is often easier to achieve, and communication is an asymmetrical process anyhow. Consider what would happen in ballroom dancing if the responsibility for choosing steps was left equally to both partners (and how little help the mutual-knowledge framework would be for solving the resulting co-ordination problems in real time). Co-ordination problems are avoided, or considerably reduced, in dancing, by leaving the responsibility to one partner who leads, while the other has merely to follow. We assume that the same goes for communication. It is left to the communicator to make correct assumptions about the codes and contex-tual information that the audience will have accessible and be likely to use in the comprehension process. The responsibility for avoiding misunderstandings also lies with the speaker, so that all the hearer has to do is go ahead and use whatever code and contextual information come most easily to hand. Suppose Mary and Peter are looking at a landscape where she has noticed a distant church. She says to him,

chapter |1 pages

he need never have made himself before she spoke. What she expects, rightly, is that her utterance will act as a prompt, making him recall parts of the book that he had previously forgotten, and construct the assumptions needed to understand the allusion. In both these examples Mary makes assumptions about what assumptions are, or will be, manifest to Peter. Peter trusts that the assumptions he spontaneously makes about the church and about Sense and Sensibility, which help him understand Mary’s utterances, are those she expected him to make. To communicate success-fully, Mary had to have some knowledge of Peter’s cognitive environment. As a result of their successful communication, their mutual cognitive environment is enlarged. Note that symmetrical co-ordination and mutual knowledge do not enter into the picture at all. The most fundamental reason for adopting the mutual-knowledge framework, as for adopting the code model, is the desire to show how successful communi-cation can be guaranteed, how there is some failsafe algorithm by which the hearer can reconstruct the speaker’s exact meaning. Within this framework the fact that communication often fails is explained in one of two ways: either the code mech-anism has been imperfectly implemented, or there has been some disruption due to ‘noise’. A noiseless, well-implemented code mechanism should guarantee per-fect communication. In rejecting the mutual-knowledge framework, we abandon the possibility of using a failsafe algorithm as a model of human communication. But since it is obvious that the communication process takes place at a risk, why assume that it is governed by a failsafe procedure? Moreover, if there is one conclusion to be drawn from work on artiﬁcial intelligence, it is that most cognitive processes are so complex that they must be modelled in terms of heuristics rather than failsafe algorithms. We assume, then, that communication is governed by a less-than-perfect heuristic. On this approach, failures in communication are to be expected: what is mysterious and requires explanation is not failure but success. As we have seen, the notion of mutual manifestness is not strong enough to salvage the code theory of communication. But then, this was never one of our aims. Instead of taking the code theory for granted and concluding that mutual knowledge must therefore exist, we prefer to look at what kind of assumptions people are actually in a position to make about each other’s assumptions, and then see what this implies for an account of communication. Sometimes, we have direct evidence about other people’s assumptions: for instance, when they tell us what they assume. More generally, because we mani-festly share cognitive environments with other people, we have direct evidence about what is manifest to them. When a cognitive environment we share with other people is mutual, we have evidence about what is mutually manifest to all of us. Note that this evidence can never be conclusive: the boundaries of cogni-tive environments cannot be precisely determined, if only because the threshold between very weakly manifest assumptions and inaccessible ones is unmarked. From assumptions about what is manifest to other people, and in particular about what is strongly manifest to them, we are in a position to derive further,

chapter |1 pages

though necessarily weaker, assumptions about what assumptions they are actually making. From assumptions about what is mutually manifest to all of us, we are in a position to derive further, and weaker, assumptions about the assumptions they attribute to us. And essentially, this is it. Human beings somehow manage to communicate in situations where a great deal can be assumed about what is manifest to others, a lot can be assumed about what is mutually manifest to them-selves and others, but nothing can be assumed to be truly mutually known or assumed. The situations which establish a mutual cognitive environment are essentially those that have been treated as establishing mutual knowledge. We have argued that assumptions of mutual knowledge are never truly warranted. Examples (49) and (50) are anecdotal evidence that they are unnecessary. The detour via mutual knowledge is superﬂuous: mutual cognitive environments directly provide all the information needed for communication and comprehension. The notions of cognitive environment and of manifestness, mutual or other-wise, are psychologically realistic, but by themselves shed little light on what goes on in human minds. A cognitive environment is merely a set of assumptions which the individual is capable of mentally representing and accepting as true. The ques-tion then is: which of these assumptions will the individual actually make? This question is of interest not only to the psychologist, but also to every ordinary communicator. We will argue that when you communicate, your intention is to alter the cognitive environment of your addressees; but of course you expect their actual thought processes to be affected as a result. In the next section we will argue that human cognition is relevance-oriented, and that as a result, someone who knows an individual’s cognitive environment can infer which assumptions he is actually likely to entertain.

chapter |1 pages

possible expenditure of whatever resource (time, money, energy . . . ) it takes. Efﬁciency with respect to relative goals is a matter of striking a balance between degree of achievement and expenditure. In the special case where the expenditure is ﬁxed – say all the time available is going to be spent anyhow – efﬁciency con-sists in achieving the goal to the highest possible degree. Most discussions of information processing, whether in experimental psycho-logy or in artiﬁcial intelligence, have been concerned with the realisation of abso-lute goals. ‘Problem solving’ has become the paradigm of information processing. The problems considered have a ﬁxed solution; the goal of the information-processing device is to ﬁnd this solution; efﬁciency consists in ﬁnding it at the minimal cost. However, not all cognitive tasks ﬁt this description; many tasks con-sist not in reaching an absolute goal, but in improving on an existing state of affairs. Hence, cognitive efﬁciency may have to be characterised differently for dif-ferent devices. Simpler information-processing devices, whether natural, such as a frog, or artiﬁcial, such as an electronic alarm system, process only very speciﬁc informa-tion: for example, metabolic changes and ﬂy movements for frogs, noises and other vibrations for alarm systems. Their information-processing activity consists in mon-itoring changes in the values of a few variables. They could be informally described as engaged in answering a few set questions: ‘Is there a ﬂy-like object within reach?’, ‘Is there a large body moving in the room?’ More complex information-processing devices, by contrast, can deﬁne and monitor new variables or formu-late and answer new questions. For the simpler devices, efﬁciency consists in answering their set questions at the minimal processing cost. Efﬁciency cannot be so easily deﬁned for more com-plex devices such as human beings. For such devices, efﬁcient information pro-cessing may involve formulating and trying to answer new questions despite the extra processing costs incurred. Formulating and answering speciﬁc questions must then be seen as subservient to a more general and abstract goal. It is in relation to this general goal that the efﬁciency of complex information-processing devices must be characterised. On the general goal of human cognition, we have nothing better to offer than rather trivial speculative remarks. However, these remarks have important and non-trivial consequences. It seems that human cognition is aimed at improving the individual’s knowledge of the world. This means adding more information, infor-mation that is more accurate, more easily retrievable, and more developed in areas of greater concern to the individual. Information processing is a permanent life-long task. An individual’s overall resources for information processing are, if not quite ﬁxed, at least not very ﬂexible. Thus, long-term cognitive efﬁciency consists in improving one’s knowledge of the world as much as possible given the avail-able resources. What, then, is short-term cognitive efﬁciency – efﬁciency, say, in the way your mind spends the next few seconds or milliseconds? This is a more concrete question, and one that is harder to answer. At every moment, many different cog-nitive tasks could be performed, and this for two reasons: ﬁrst, human sensory

chapter |1 pages

abilities monitor much more information than central conceptual abilities can pro-cess; and second, central abilities always have plenty of unﬁnished business. The key problem for efﬁcient short-term information processing is thus to achieve an optimal allocation of central processing resources. Resources have to be allocated to the processing of information which is likely to bring about the greatest con-tribution to the mind’s general cognitive goals at the smallest processing cost. Some information is old: it is already present in the individual’s representa-tion of the world. Unless it is needed for the performance of a particular cogni-tive task, and is easier to access from the environment than from memory, such information is not worth processing at all. Other information is not only new but entirely unconnected with anything in the individual’s representation of the world. It can only be added to this representation as isolated bits and pieces, and this usually means too much processing cost for too little beneﬁt. Still other informa-tion is new but connected with old information. When these interconnected new and old items of information are used together as premises in an inference pro-cess, further new information can be derived: information which could not have been inferred without this combination of old and new premises. When the pro-cessing of new information gives rise to such a multiplication effect, we call it relevant. The greater the multiplication effect, the greater the relevance. Consider an example. Mary and Peter are sitting on a park bench. He leans back, which alters her view. By leaning back, he modiﬁes her cognitive environ-ment; he reveals to her certain phenomena, which she may look at or not, and describe to herself in different ways. Why should she pay attention to one phe-nomenon rather than another, or describe it to herself in one way rather than another? In other words, why should she mentally process any of the assumptions which have become manifest or more manifest to her as a result of the change in her environment? Our answer is that she should process those assumptions that are most relevant to her at the time. Imagine, for instance, that as a result of Peter’s leaning back she can see, among other things, three people: an ice-cream vendor who she had noticed before when she sat down on the bench, an ordinary stroller who she has never seen before, and her acquaintance William, who is coming towards them and is a dread-ful bore. Many assumptions about each of these characters are more or less man-ifest to her. She may already have considered the implications of the presence of the ice-cream vendor when she ﬁrst noticed him; if so, it would be a waste of processing resources to pay further attention to him now. The presence of the unknown stroller is new information to her, but little or nothing follows from it; so there again, what she can perceive and infer about him is not likely to be of much relevance to her. By contrast, from the fact that William is coming her way, she can draw many conclusions from which many more conclusions will fol-low. This, then, is the one truly relevant change in her cognitive environment; this is the particular phenomenon she should pay attention to. She should do so, that is, if she is aiming at cognitive efﬁciency. Our claim is that all human beings automatically aim at the most efﬁcient information processing possible. This is so whether they are conscious of it or not;

chapter |1 pages

in fact, the very diverse and shifting conscious interests of individuals result from the pursuit of this permanent aim in changing conditions. In other words, an indi-vidual’s particular cognitive goal at a given moment is always an instance of a more general goal: maximising the relevance of the information processed. We will show that this is a crucial factor in human interaction. Among the facts made manifest to Mary by Peter’s behaviour is the very fact that he has behaved in a certain way. Suppose now that she pays attention to this behaviour, and comes to the conclusion that it must have been deliberate: per-haps he is leaning back more rigidly than if he were merely trying to ﬁnd a more comfortable position. She might then ask herself why he is doing it. There may be many possible answers; suppose that the most plausible one she can ﬁnd is that he is leaning back in order to attract her attention to some particular phenomenon. Then Peter’s behaviour has made it manifest to Mary that he intends to make some particular assumptions manifest to her. We will call such behaviour – behaviour which makes manifest an intention to make something manifest – ostensive behav-iour or simply ostension. Showing someone something is a case of ostension. So too, we will argue, is human intentional communication. The existence of ostension is beyond doubt. What is puzzling is how it works. Any perceptible behaviour makes manifest indeﬁnitely many assumptions. How is the audience of an act of ostension to discover which of them have been inten-tionally made manifest? For instance, how is Mary to discover which of the phe-nomena which have become manifest to her as a result of Peter’s behaviour are the ones he intended her to pay attention to? Information processing involves effort; it will only be undertaken in the expec-tation of some reward. There is thus no point in drawing someone’s attention to a phenomenon unless it will seem relevant enough to him to be worth his attention. By requesting Mary’s attention, Peter suggests that he has reason to think that by paying attention, she will gain some relevant information. He may, of course, be mistaken, or trying to distract her attention from relevant informa-tion elsewhere, as the maker of an assertion may be mistaken or lying; but just as an assertion comes with a tacit guarantee of truth, so ostension comes with a tacit guarantee of relevance. This guarantee of relevance makes it possible for Mary to infer which of the newly manifest assumptions have been intentionally made manifest. Here is how the inference process might go. First, Mary notices Peter’s behaviour and assumes that it is ostensive: i.e. that it is intended to attract her attention to some phe-nomenon. If she has enough conﬁdence in his guarantee of relevance, she will infer that some of the information which his behaviour has made manifest to her is indeed relevant to her. She then pays attention to the area that has become visible to her as a result of his leaning back, and discovers the ice-cream vendor, the stroller, this dreadful William, and so on. Assumptions about William are the only newly manifest assumptions relevant enough to be worth her attention. From this, she can infer that Peter’s intention was precisely to draw her attention to William’s arrival. Any other assumption about his ostensive behaviour is inconsis-tent with her conﬁdence in the guarantee of relevance it carries.

chapter |1 pages

Mary has become aware not only that there is someone coming who she wants to avoid, but also that Peter intended her to become aware of it, and that he is aware of it too. On the basis of his observable behaviour, she has discovered some of his thoughts. Ostensive behaviour provides evidence of one’s thoughts. It succeeds in doing so because it implies a guarantee of relevance. It implies such a guarantee because humans automatically turn their attention to what seems most relevant to them. The main thesis of section is that an act of ostension carries a guarantee of relevance, and that this fact – which we will call the principle of relevance – makes manifest the intention behind the ostension. We believe that it is this principle of rele-vance that is needed to make the inferential model of communication explanatory.

chapter |1 pages

think that the clouds are going to get worse and turn to rain. Such an assump-tion is of a very standard sort and would probably be the ﬁrst to come to mind. The old man can thus be reasonably conﬁdent that, prompted by his behaviour, she will have no difﬁculty in deciding that this is what he believes. If it were not manifest to the old man that it was going to rain, it would be hard to explain his behaviour at all. The girl thus has reason to think that in drawing her atten-tion to the clouds, he intended to make manifest to her that he believed it was going to rain. As a result of this act of ostension, she now has some information that was not available to her before: that he thinks it is going to rain, and hence that there is a genuine risk of rain. In this example, the state of affairs that the old man drew the girl’s attention to had been partly manifest to her, and partly not. The presence of the clouds and the fact that clouds may always turn to rain had been manifest and merely became more so. However, until that moment she had regarded the fact that the weather was beautiful as strong evidence that it would not rain. The risk of rain in that particular situation was not manifest to her at all. In other words, the clouds were already evidence of oncoming rain, but evidence that was much too weak. The old man made that evidence much stronger by pointing it out; as his intentions became manifest, the assumption that it would rain became manifest too. Sometimes, all the evidence displayed in an act of ostension bears directly on the agent’s intentions. In these cases, only by discovering the agent’s intentions can the audience also discover, indirectly, the basic information that the agent intended to make manifest. The relation between the evidence produced and the basic information conveyed is arbitrary. The same piece of evidence can be used, on different occasions, to make manifest different assumptions, even mutually in-consistent assumptions, as long as it makes manifest the intention behind the ostension. Here is an example. Two prisoners, from different tribes with no common language, are put in a quarry to work back to back breaking rocks. Suddenly, prisoner A starts putting some distinct rhythm into the sound of his hammer – one–two–three, one–two, one–two–three, one–two – a rhythm that is both arbi-trary and noticeable enough to attract the attention of prisoner B. This arbitrary pattern in the way the rocks are being broken has no direct relevance for B. However, there are grounds for thinking that it has been intentionally produced, and B might ask himself what A’s intentions were in producing it. One plausible assumption is that this is a piece of ostensive behaviour: that is, that A intended B to notice the pattern. This would in turn make manifest A’s desire to interact with B, which in the circumstances would be relevant enough. Here is a more substantial example. Prisoners A and B are at work in their quarry, each with a guard at his shoulder, when suddenly the attention of the guards is distracted. Both prisoners realise that they have a good chance of escaping, but only if they can co-ordinate their attack and overpower their guards simultaneously. Here, it is clear what information would be relevant: each wants to know when the other will start the attack. Prisoner A suddenly whistles, the prisoners overpower their guards and escape. Again, there is no need for a pre-

chapter |1 pages

existing code correlating a whistle with the information that now is the moment to attack. The information is obvious enough: it is the only information that A could conceivably have intended to make manifest in the circumstances. Could not the repetition of such a situation lead to the development of a code? Imagine that the two prisoners, caught again, ﬁnd themselves in the same predicament: again a whistle, again an escape, and again they are caught. The next time, prisoner B, who has not realised that both guards are distracted, hears pris-oner A whistle: this time, fortunately, B does not have to infer what the whistle is intended to make manifest: he knows. The whistle has become a signal associ-ated by an underlying code to the message ‘Let us overpower our guards now!’ Inferential theorists might be tempted to see language as a whole as having developed in this way: to see conventional meanings as growing out of natural inferences. This is reminiscent of the story of how Rockefeller became a million-aire. One day, when he was young and very poor, Rockefeller found a one-cent coin in the street. He bought an apple, polished it, sold it for two cents, bought two apples, polished them, sold them for four cents . . . After one month he bought a cart, after two years he was about to buy a grocery store, when he inherited the fortune of his millionaire uncle. We will never know how far hominid efforts at conventionalising inference might have gone towards establishing a full-ﬂedged human language. The fact is that the development of human languages was made possible by a specialised biological endowment. Whatever the origin of the language or code employed, a piece of coded behaviour may be used ostensively – that is, to provide two layers of information: a basic layer of information, which may be about anything at all, and a second layer con-sisting of the information that the ﬁrst layer of information has been intentionally made manifest. When a coded signal, or any other arbitrary piece of behaviour, is used ostensively, the evidence displayed bears directly on the individual’s intention, and only indirectly on the basic layer of information that she intends to make manifest. We are now, of course, dealing with standard cases of Gricean communication. Is there a dividing line between instances of ostension which one would be more inclined to describe as ‘showing something’, and clear cases of communica-tion where the communicator unquestionably ‘means something’? One of Grice’s main concerns was to draw such a line: to distinguish what he called ‘natural meaning’ – smoke meaning ﬁre, clouds meaning rain, and so on – from ‘non-natural meaning’: the word ‘ﬁre’ meaning ﬁre, Peter’s utterance meaning that it will rain, and so on. Essential to this distinction was the third type of communi-cator’s intention Grice mentioned in his analysis: a true communicator intends the recognition of his informative intention to function as at least part of the audi-ence’s reason for fulﬁlling that intention. In other words, the ﬁrst, basic, layer of information must not be entirely recoverable without reference to the second. What we have tried to show so far in this section is that there are not two distinct and well-deﬁned classes, but a continuum of cases of ostension ranging from ‘showing’, where strong direct evidence for the basic layer of information is provided, to ‘saying that’, where all the evidence is indirect. Even in our very

chapter |1 pages

ﬁrst case of Peter leaning back ostensively to let Mary see William approaching, it is arguable that some of the basic information is made manifest indirectly, through Peter’s intention being made manifest. Someone who engages in any kind of osten-sive behaviour intentionally draws some attention to himself and intentionally makes manifest a few assumptions about himself: for instance, that he is aware of the basic information involved, and that he is trying to be relevant. Peter’s ostension might make it manifest not just that William is approaching, but also that Peter expects Mary to be concerned, and that he is concerned too. Would we want to say, though, that Peter ‘meant something’ by his behav-iour? Like most English speakers, we would be reluctant to do so; but this is irrelevant to our pursuit, which is not to analyse ordinary language usage, but to describe and explain forms of human communication. Our argument at this stage is this: either inferential communication consists in providing evidence for what the communicator means, in the sense of ‘meaning’ which Grice calls ‘non-natural meaning’, and in that case inferential communication is not a well-deﬁned class of phenomena at all; or else showing something should be considered a form of inferential communication, on a par with meaning something by a certain behaviour, and inferential communication and ostension should be equated. There are two questions involved here. One is substantive: which domains of facts are to be described and explained together? Our answer is that ostension is such a domain, and that inferential communication narrowly understood (i.e. under-stood as excluding cases of ostension where talk of ‘meaning’ would be awkward) is not. The second question is terminological (and hence not worth much argu-ment): can the term ‘communication’ be legitimately applied to all cases of osten-sion? Our answer is yes, and from now on we will treat ostensive communication, inferential communication, and ostensive–inferential communication as the same thing. Inferential communication and ostension are one and the same process, but seen from two different points of view: that of the communicator who is involved in ostension and that of the audience who is involved in inference. Ostensive–inferential communication consists in making manifest to an audi-ence one’s intention to make manifest a basic layer of information. It can there-fore be described in terms of an informative and a communicative intention. In the next two sections, we want to reanalyse the notions of informative and com-municative intention in terms of manifestness and mutual manifestness, and to sketch in some of the empirical implications of this reformulation.

chapter |1 pages

linguistic utterance. The verbal communication of an explicit meaning is then taken as the model of communication in general. This is true of semiotic approaches, which are not only generalisations of a linguistic model, but are also based on the assumption that to communicate is always, in Saussure’s terms, to transmit a ‘signiﬁed’ by use of a ‘signiﬁer’. It is true of inferential approaches, which regard all com-municative acts as ‘utterances’ in an extended sense, used to convey an ‘utterer’s meaning’. We believe that the kind of explicit communication that can be achieved by the use of language is not a typical but a limiting case. Treating linguistic com-munication as the model of communication in general has led to theoretical distortions and misperceptions of the data. The effects of most forms of human communication, including some of the effects of verbal communication, are far too vague to be properly analysed along these lines. Moreover, there is not a dichotomy but a continuum of cases, from vaguer to more precise effects. Let us ﬁrst illustrate this point with two examples of non-verbal communica-tion. Mary comes home; Peter opens the door. Mary stops at the door and sniffs ostensively; Peter follows suit and notices that there is a smell of gas. This fact is highly relevant, and in the absence of contextual counterevidence or any obvi-ous alternative candidate, Peter will assume that Mary intended to make it man-ifest to him that there was a smell of gas. Here, at least part of what is communicated could be reasonably well paraphrased by saying that there is a smell of gas; and it could be argued that this is what Mary means. She could indeed have achieved essentially the same result by speaking rather than snifﬁng ostensively. Contrast this with the following case. Mary and Peter are newly arrived at the seaside. She opens the window overlooking the sea and sniffs appreciatively and ostensively. When Peter follows suit, there is no one particular good thing that comes to his attention: the air smells fresh, fresher than it did in town, it reminds him of their previous holidays, he can smell the sea, seaweed, ozone, ﬁsh; all sorts of pleasant things come to mind, and while, because her sniff was appreciative, he is reasonably safe in assuming that she must have intended him to notice at least some of them, he is unlikely to be able to pin her intentions down any further. Is there any reason to assume that her intentions were more speciﬁc? Is there a plausible answer, in the form of an explicit linguistic para-phrase, to the question, what does she mean? Could she have achieved the same communicative effect by speaking? Clearly not. Examples like the one of Mary smelling gas, where it is reasonable to impute a meaning to the communicator, are the only ones normally considered in dis-cussions of communication; examples like the one of Mary at the seaside – clearly communicating, but what? – are generally ignored. Yet these examples do not belong to distinct classes of phenomena, and it is easy enough to imagine inter-mediate cases: say, a guest snifﬁng appreciatively and ostensively when the stew is brought to the table, and so on. The distortions and misperceptions introduced by the explicit communication model are also found in the study of verbal communication itself. Some essential aspects of implicit verbal communication are overlooked. Pragmatists assume that

chapter |1 pages

what is communicated by an utterance is a speaker’s meaning, which in the case of an assertion is a set of assumptions. One of these assumptions is explicitly expressed; the others (if any) are implicitly conveyed, or implicated. The only dif-ference between the explicit content of an utterance and its implicatures is sup-posed to be that the explicit content is decoded, while the implicatures are inferred. Now we all know, as speakers and hearers, that what is implicitly conveyed by an utterance is generally much vaguer than what is explicitly expressed, and that when the implicit import of an utterance is explicitly spelled out, it tends to be distorted by the elimination of this often intentional vagueness. The distortion is even greater in the case of metaphor and other ﬁgures of speech, whose poetic effects are generally destroyed by being explicitly spelled out. In an effort to minimise the distortion, pragmatists have tended to focus on examples such as (32), where the implicit import is fairly precise, and to ignore equally ordinary cases of implicit vagueness such as (51):

chapter |2 pages

At ﬁrst sight, it might look as if semioticians had a more comprehensive view. They have an a priori account of how any kind of representation, propositional or not, might be conveyed: namely, by means of a code. However, studies by semioticians of what they call ‘connotation’, i.e. the vaguer aspect of what is com-municated, are highly programmatic and do not offer the beginnings of a psycho-logically adequate account of the type of mental representation involved. The semiotic approach is more comprehensive only by being more superﬁcial. The only people who have been quite consistently concerned with the vaguer aspects of communication are the Romantics, from the Schlegel brothers and Coleridge to I. A. Richards, and their many acknowledged or unacknowledged followers, including many semioticians such as Roman Jakobson in some of his writings, Victor Turner, or Roland Barthes. However, they have all dealt with vagueness in vague terms, with metaphors in metaphorical terms, and used the term ‘meaning’ so broadly that it becomes quite meaningless. We see it as a major challenge for any account of human communication to give a precise description and explanation of its vaguer effects. Distinguishing mean-ing from communication, accepting that something can be communicated without being strictly speaking meant by the communicator or the communicator’s behaviour, is a ﬁrst essential step – a step away from the traditional approach to communication and most modern approaches. Once this step is taken, we believe that the framework we propose, unlike the others we have discussed, can rise to this challenge. Accounts of communication either are not psychological at all, and avoid all talk of thoughts, intentions, etc., or else they assume that a communicator’s inten-tion is to induce certain speciﬁc thoughts in an audience. We want to suggest that the communicator’s informative intention is better described as an intention to modify directly not the thoughts but the cognitive environment of the audience. The actual cognitive effects of a modiﬁcation of the cognitive environment are only partly predictable. Communicators – like human agents in general – form intentions over whose fulﬁlment they have some control: they can have some con-trollable effect on their audience’s cognitive environment, much less on their audi-ence’s actual thoughts, and they form their intentions accordingly. We therefore propose to reformulate the notion of an informative intention along the following lines. A communicator produces a stimulus intending thereby

chapter |1 pages

to increase simultaneously the manifestness of a wide range of assumptions, so that her intention concerning each of these assumptions is weakly manifest, then each of them is weakly communicated. An example would be snifﬁng ecstatically and osten-sively at the fresh seaside air. There is, of course, a continuum of cases in between. In the case of strong communication, the communicator can have fairly precise expectations about some of the thoughts that the audience will actually entertain. With weaker forms of communication, the communicator can merely expect to steer the thoughts of the audience in a certain direction. Often, in human interaction, weak communication is found sufﬁcient or even preferable to the stronger forms. Non-verbal communication tends to be relatively weak. One of the advantages of verbal communication is that it gives rise to the strongest possible form of com-munication; it enables the hearer to pin down the speaker’s intentions about the explicit content of her utterance to a single, strongly manifest candidate, with no alternative worth considering at all. On the other hand, what is implicit in ver-bal communication is generally weakly communicated: the hearer can often fulﬁl part of the speaker’s informative intention by forming any of several roughly similar but not identical assumptions. Because all communication has been seen as strong communication, descriptions of non-verbal communication have been marred by spurious attributions of ‘meaning’; in the case of verbal communication, the difference between explicit content and implicit import has been seen as a differ-ence not in what gets communicated but merely in the means by which it is com-municated, and the vagueness of implicatures and non-literal forms of expression has been idealised away. Our account of informative intentions in terms of man-ifestness of assumptions corrects these distortions without introducing either ad hoc machinery or vagueness of description.

chapter |1 pages

This takes care of the types of example which Strawson and Schiffer used to show that, in order to communicate, it is not quite enough to inform an audience of one’s informative intention. For instance, in the example in an earlier section, Mary leaves the pieces of her broken hair-drier lying around, intending thereby to inform Peter that she would like him to mend it. She wants this informative intention to be manifest to Peter, but at the same time, she does not want it to be ‘overt’. In our terms, she does not want her informative intention to be mutu-ally manifest. Intuitively, what she does is not quite communicate. Our redeﬁnition of a communicative intention accounts for this intuition. What difference does it make whether an informative intention is merely manifest to the audience or mutually manifest to audience and communicator? Should this really be a criterion for distinguishing communication from other forms of information transmission? Is it more than a technicality designed to take care of implausible borderline cases dreamed up by philosophers? Our answer is that there is indeed an essential difference. Consider ﬁrst a more general question: why should someone who has an in-formative intention bother to make it known to her audience that she has this intention? In other words, what are the reasons for engaging in ostensive com-munication? Grice discussed only one of these reasons: sometimes, making one’s informative intention known is the best way, or the only way, of fulﬁlling it. We have shown that people sometimes engage in ostensive communication even though the informative intention could be fulﬁlled without being made manifest: for exam-ple, by providing direct evidence for the information to be conveyed. However, even in these cases, ostension helps focus the attention of the audience on the rel-evant information, and thus contributes to the fulﬁlment of the informative inten-tion. This is still the Gricean reason for engaging in communication, just slightly extended in scope. However, we want to argue that there is another major reason for engaging in ostensive communication, apart from helping to fulﬁl an informative intention. Mere informing alters the cognitive environment of the audience. Communication alters the mutual cognitive environment of the audience and communicator. Mutual manifestness may be of little cognitive importance, but it is of crucial social impor-tance. A change in the mutual cognitive environment of two people is a change in their possibilities of interaction (and, in particular, in their possibilities of fur-ther communication). Recall, for instance, the case of Peter leaning back to let Mary see William coming their way. If, as a result of his behaviour, it becomes mutually manifest to them that William is coming, that they are in danger of being bored by his conversation, and so on, then they are in a position to act efﬁciently: i.e. promptly. All Mary may have to do is say, ‘Let’s go!’; she can feel conﬁdent that Peter will understand her reasons, and, if he shares them, will be ready to act without ques-tion or delay. In the case of the broken hair-drier, if Mary had made mutually manifest her wish that Peter would mend it, one of two things would have happened. Either he would have mended it, thus granting her wish and possibly putting her in his

chapter |1 pages

debt; or he would have failed to mend it, which would have amounted to a refusal or rejection. Mary avoids putting herself in his debt or meeting with a refusal by avoiding any modiﬁcation of their mutual cognitive environment. If Peter mends the hair-drier, he is being kind on his own initiative, and she does not owe him anything. If Peter decides not to mend the hair-drier, he might reason as follows: she doesn’t know I know she intended to inform me of her wish, so if I ignore it, she will attribute this to her failure to inform me; she may ﬁnd me stupid, but not unkind. As for Mary, she may have intentionally left this line of reason-ing open to Peter. If he does not mend her hair-drier, she will ﬁnd him unkind, but not hostile. His failure to grant her wish will not be in the nature of a rebuff. They will stand in exactly the same social relationship to each other as before. This shows how ostensive communication may have social implications that other forms of information transmission do not. By making her informative intention mutually manifest, the communicator creates the following situation: it becomes mutually manifest that the fulﬁlment of her informative intention is, so to speak, in the hands of the audience. If the assumptions that she intends to make manifest to the audience become manifest, then she is successful; if the audience refuses to accept these assumptions as true or probably true, then she has failed in her informative intention. Suppose – we will soon see how this may happen – that the audience’s behaviour makes it mutu-ally manifest that the informative intention is fulﬁlled. Then the set of assump-tions I that the communicator intended to make manifest to the audience becomes, at least apparently, mutually manifest. We say ‘at least apparently’ because, if the communicator is not sincere and some of the assumptions in I are not manifest to her, then by our deﬁnition of mutual manifestness, these assumptions cannot be mutually manifest to her and others. A communicator is normally interested in knowing whether or not she has succeeded in fulﬁlling her informative intention, and this interest is mutually manifest to her and her audience. In face-to-face communication, the audience is generally expected to respond to this interest in fairly conventional ways. Often, for instance, the audience is expected to communicate its refusal to accept the information communicated, or else it becomes mutually manifest that the com-municator’s informative intention is fulﬁlled. Where communication is non-reciprocal, there are various possible situations to be taken into account. The communicator may be in a position of such author-ity over her audience that the success of her informative intention is mutually manifest in advance. Journalists, professors, religious or political leaders assume, alas often on good grounds, that what they communicate automatically becomes mutually manifest. When the communicator lacks that kind of authority, but still wants to establish a mutual cognitive environment with her audience, all she has to do is adapt her informative intentions to her credibility. For instance, in writ-ing this book we merely intend to make mutually manifest that we have devel-oped certain hypotheses and have done so on certain grounds. That is, we take it as mutually manifest that you will accept our authority on what we actually think. The mutual cognitive environment thus created is enough for us to go on

chapter |3 pages

to communicate further thoughts which we would otherwise have been unable to communicate. (Of course we would also like to convince you, but we hope to do this by the force of our arguments, and not by making you recognise our infor-mative intentions.) We began this section by asking how human beings communicate with one another. Our answer is that they use two quite different modes of communica-tion: coded communication and ostensive–inferential communication. However, the two modes of communication are used in fundamentally different ways. Whereas ostensive–inferential communication can be used on its own, and sometimes is, coded communication is only used as a means of strengthening ostensive–inferential communication. This is how language is used in verbal communication, as we will argue later. Ostensive–inferential communication can be deﬁned as follows:

chapter |1 pages

The American tendency to associate indirectness with female style is not culturally universal. The above description of typical Japanese style operates for men as well as women. My own research (Tannen 1981, 1984, 1986) suggests that Americans of some cultural and geographic backgrounds, female as well as male, are more likely than others to use relatively direct rather than indirect styles. In an early study I compared Greeks and Americans with regard to their tendency to interpret a question as an indirect means of making a request. I found that whereas American women were more likely to take an indirect interpretation of a sample conversation, Greek men were as likely as Greek women, and more likely than American men or women, to take an indirect interpretation. Greek men, of course, are not less powerful vis-à-vis women than American men. Perhaps most striking is the ﬁnding of Keenan (1974) that in a Malagasy-speaking village on the island of Madagascar, women are seen as direct and men as indirect. But this in no way implies that the women are more powerful than men in this society. Quite the contrary, Malagasy men are socially dominant, and their indirect style is more highly valued. Keenan found that women were widely believed to debase the language with their artless directness, whereas men’s elab-orate indirectness was widely admired. Indirectness, then, is not in itself a strategy of subordination. Rather, it can be used either by the powerful or the powerless. The interpretation of a given utterance, and the likely response to it, depends on the setting, on individuals’ status and their relationship to each other, and also on the linguistic conventions that are ritualized in the cultural context.

chapter |1 pages

that both Syrians and Americans are more likely to either accept or mitigate the force of the compliment than to reject it. Both groups employed similar response types (e.g. agreeing utterances, compliment returns, and deﬂecting or qualifying comments); however, they also differed in their responses. US recipients were much more likely than the Syrians to use appreciation tokens and a preferred Syrian response, acceptance + formula, does not appear in the US data at all. Recently, in a conversation with an American who had taught EFL in Damascus for two years, one of the researchers mentioned that she was investigating the strategies Syrians use in responding to compliments. The teacher looked surprised and asked, ‘What’s there to study? Syrians just say Shukran (“thank you”). When I’m complimented in Arabic, that’s what I say – Shukran.’ This teacher was apply-ing a rule from his L1 speech community to an L2 speech community. The rule he was transferring is one that American parents teach their children and one that is taught in etiquette books: ‘When you are complimented, the only response nec-essary is “Thank you” ’ (Johnson 1979: 43). Compliment responses in Syrian Arabic, as shall become clear later, are much more complex than saying Shukran when praised. In this paper, we report on a study of Syrian Arabic speakers’ and American English speakers’ verbal responses to compliments. The purpose of the study is to better understand the strategies used by Syrians and Americans in responding to compliments, to discover similarities and differences between the two groups, and to relate the ﬁndings to second language acquisition and second language teaching.

chapter |1 pages

In the context of language learning, one cause of pragmalinguistic failure is pragmalinguistic transfer, the use of L1 speech act strategies or formulas when interacting with members of an L2 speech community (Leech 1983). This trans-fer has been addressed in a number of speech act/event studies (e.g. Blum-Kulka 1982, 1983; Olshtain 1983; Olshtain and Cohen 1983; Edmonson, House, Kasper, and Stemmer 1984; Thomas 1984; Eisenstein and Bodman 1986; Garcia 1989; Wolfson 1989a; Beebe, Takahashi, and Uliss-Weltz 1990; Takahashi and Beebe 1993). In the anecdote at the beginning of this paper, the American, in respond-ing to Arabic compliments by transferring an appropriate response from his L1 to an L2, believes that he is politely accepting the compliment. However, if the native Arabic speaker interprets the illocutionary force of the utterance differently (e.g. interprets the response as impolite and inappropriate) pragmatic failure has occurred. It is, however, difﬁcult, at times, to determine whether the pragmatic failure results from L1 transfer or from other factors. Hurley (1992), for example, notes that pragmatic failure may also result from developmental and proﬁciency factors or from L2 learners overgeneralizing the use of an L2 form to inappropriate set-tings. Stated differently, it is sometimes difﬁcult to know why language learners experience certain kinds of pragmatic failure. In order to understand the reasons behind pragmatic failure, it is helpful, and perhaps even necessary, to conduct cross-cultural research to investigate students’ L1 strategies (Wolfson 1989a). Speech act and speech event studies have been criticized as being ethnocentric in that most have investigated variations of English (Blum-Kulka, House, and Kasper 1989). Rose (1994) further points out that, in particular, little work has been done in non-Western contexts. The present study is valuable, in part, because it was conducted in Arabic as well as English. Compliment responses were selected for cross-cultural study for two reasons. First, although a body of knowledge exists on the speech act of complimenting (Wolfson 1981, 1983; Manes 1983; Knapp, Hopper, and Bell 1984; Barnlund and Araki 1985; Holmes and Brown 1987; Nelson, El Bakary, and Al-Batal 1993), less research has been conducted on responses to compliments. For non-native English speak-ing (NNES) students, knowing how to compliment is important, but it is equally important to know how to respond to a compliment. In fact, it could be argued that for NNES students in the United States, appropriately responding to compli-ments is more important than complimenting because of the frequency with which Americans compliment (Wolfson 1983; Holmes and Brown 1987; Herbert 1988). In other words, ESL students may receive more compliments than they initiate. A second reason is that, although a few studies have been conducted on compli-ment responses in English-speaking countries (Pomerantz 1978; Herbert 1988; Herbert and Straight 1989), few, if any, cross-cultural studies have investigated compli-ment responses in an Arabic-speaking country. For the purpose of this study, a compliment response is deﬁned as a verbal acknowledgement that the recipient of the compliment heard and reacted to the compliment. Compliment/compliment response interactions have been referred to

chapter |1 pages

as adjacency pairs (Schegloff and Sacks 1973), action chain events (Pomerantz 1978), interchanges (Herbert 1988), and sequences (Wolfson 1989b). For ease of refer-ence, Speaker will refer to the person issuing the compliment and Speaker to

chapter |1 pages

Herbert (1988) provided such an analysis in a study comparing the compli-ment/compliment response interchanges from American university students to South African university students. In analyzing his data, he grouped the responses as (a) Agreeing, (b) Nonagreeing, or (c) Requesting interpretation. Overall, nearly 66 per cent of the American compliment responses were broadly classiﬁed as Agree-ments, 31 per cent as Nonagreements, and 3 per cent as Request Interpretations. Of those Agreements (66 per cent), 7 per cent were categorized as Comment Acceptances and 29 per cent as Appreciation Tokens. In contrast, 88 per cent of

chapter |1 pages

ments tended to be adjectival (e.g. ‘You look great’). A major difference between Egyptian and American compliments was that Egyptian compliments tended to be longer and contained more comparatives and metaphors than the US compliments (e.g. shaklak a ariis innaharda [‘You look like a bridegroom today’] ). Both Egyptians and Americans complimented the attributes of physical appear-ance, personality traits, and skills/work. Because these are attributes complimented in both Arabic-speaking and English-speaking countries (see Holmes and Brown 1987; Holmes 1988 for studies on New Zealand compliments) these were the qualities complimented in this study.

chapter |1 pages

female. One was attending college part-time and was 29 years of age, single, and a dental technician. The other was 25, single, a translator and secretary, and an English literature graduate from Damascus University. The other two interview-ers were male. One studied English literature at the University, managed his fam-ily farm property, was 27 and was single. The fourth also studied English literature at the University and was 22. All four were middle class. The Syrian compliment/compliment responses were not audiotaped. The Syrian interviewers reported that tape recorders were likely to make the interviewees feel uncomfortable; that, in general, Syrians are not familiar with the practice of conducting sociological or sociolinguistic studies about themselves; and that the tape recording would be culturally inappropriate. The Syrian interviewers praised 32 recipients, 20 males and 12 females, on physical appearance, on personality traits, or on a skill or job; listened to the responses; responded in turn; and after the interaction was completed, wrote down what was said. In some cases, the interviewers felt uncomfortable complimenting a person of a different gender or a person that was older. In these cases, they observed others giving and respond-ing to compliments and wrote down what was said. These observations resulted in an additional 20 compliment/compliment response sequences. In 7 cases, males were complimented, and in 13 cases, females were complimented. These proce-dures resulted in naturalistic data and yielded 52 Syrian compliment/compliment responses from 52 recipients, 27 males and 25 females. To insure the accuracy of the transcriptions, the Syrian interviewers were trained by one of the researchers. The trainer instructed them (l) to write down the exact words used in the complement/compliment response interaction, and (2) to do so as soon as possible after the interaction took place. In addition, the trainer gave each interviewer note cards and instructed them to write each interaction on a separate card. The trainer met with the interviewers at least once a week. At these meetings, the interviewers reported on their progress and the trainer again emphasized the importance of recording the interactions verbatim. To native speakers of English, recalling compliment responses word-for-word may seem difﬁcult, but the task is less difﬁcult for native speakers of Arabic. Many of the Syrian utterances consist of set formulas. The Syrian interviewers would remember the responses because they exist as formulaic chunks of discourse. The potential for varying the formulas is minimal. For the non-formulaic responses, it is possible that an interviewer might have made a minor change in the wording. However, if such a change occurred, the wording of the compliment response would still be an appropriate Syrian response to the situation. The Arabic compliments/compliment responses were translated into Eng-lish, but the primary analysis was based on the Arabic transcripts, not the English translations.

chapter |2 pages

This examination suggested classiﬁcation schemes similar to existing schemes (e.g. Pomerantz 1978; Herbert 1988; Herbert and Straight 1989). In the end, the classiﬁcation scheme that most appropriately ﬁtted the data was similar to, but still different from, earlier classiﬁcations. It consisted of three broad categories (i.e. acceptances, mitigations, and rejections) and subcategories. The speciﬁc subcate-gories are provided in the Results and Discussion section of this article in Tables D5.1 and D5.2. Following guide-lines set forth by Krippendorf (1980) and Holsti (1969), the categories were exhaustive (i.e. all data were represented in one of the categories) and mutually exclusive (i.e. a response could belong to only one category). After the classiﬁcation scheme was developed, one of the researchers and a graduate research assistant coded each of the English compliment responses as belong-ing to one of the categories. The Arabic compliment responses were coded by two of the researchers; one of whom is a native Arabic speaker. The coders worked independently and coded all of the compliment responses. Intercoder reliability was determined by comparing both coders’ scores. Intercoder reliability was 92 per cent for the American data and 88 per cent for the Arabic. Next, the coders reviewed the coding guide-lines and the items on which there was disagreement. They recoded until they came to a consensus; thus, in the end, agreement on all compliment responses was achieved. This section presents the analysis of the American and Syrian compliment response types.

chapter |2 pages

In none of her examples did a person agree with a compliment about him or herself. c. Compliment Return. A Compliment Return consisted of two parts – (a) a stated or implied acceptance of the force of the compliment, and (b) praise for the orig-inal sender.

chapter |3 pages

The infrequency of this response type in the United States is consistent with the work of other researchers (Herbert 1988; Chen 1993). By using this response type, Speaker clearly and directly disagrees with the judgment of Speaker, thus

chapter |3 pages

were coded as Agreement + Formula if they included a particular utterance or saying that is commonly used in Arabic when responding to a particular kind of compliment. These expressions are automatic and often ritualistic. They fulﬁll a particular social function and should not be interpreted primarily at the semantic level. As far as we know, this response type does not appear in any other lan-guage group studied. One common ritualistic compliment response was m’addam (‘[It is] presented [to you]’). With this response, Speaker offered the object of the compliment to

chapter |1 pages

Data for this study was obtained from one strata of the larger population of Syria and the US. The Syrian compliment responses were uttered by middle class people from an urban area (i.e. Damascus) and most of the American compliment responses were given by Caucasian university graduate students. One cannot assume that these ﬁndings generalize to other groups within Syria or the US or to other Arabic-speaking or English-speaking countries. Further research is needed to know how generalizable these ﬁndings are. In order for students to become communicatively competent in a second language, they need both grammatical and pragmatic competence (Thomas 1983). However, achieving pragmatic competence may, at times, be complicated due to pragmatic transfer – using the rules governing speech events from one’s L1 speech com-munity when interacting with members of an L2 speech community. Pragmatic transfer can lead to pragmatic failure, to not understanding the illocutionary force of an utterance, to not understanding what is meant by what is said (Thomas 1983). Such situations can result in cross-cultural misunderstandings and communi-cation breakdowns. Cross-cultural studies such as this one contribute to our know-ledge of appropriate compliment/compliment response competence in Syrian Arabic and American English and also to our understanding of pragmatic transfer as a possible cause for pragmatic failure. The results of this study suggest similarities and differences in Syrian Arabic and American English compliment responses. Similarities include the overall manner of responding – both Syrians and Americans are much more likely to either accept or mitigate the force of the compliment than to reject it outright. In addition, members of both groups use some similar response types (e.g. Agreeing Utterances, Compliment Returns, Deﬂecting or Qualifying Comments, and Reas-surance or Repetition Requests). Finally, males and females in both groups employ most of the response types. An exception is Agreeing Utterances; Syrian females did not use this response. Students of English and Arabic can use these similari-ties between Arabic and English compliment responses to their advantage by learn-ing the responses that are similar in both languages. As Kasper and Blum-Kulka (1993) point out, behaviors that are consistent across L1 and L2 usually result in communicative success. However, Hurley (1992) warns that the similarity of an L2 form to a form in the learner’s L1 can also be a pragmalinguistic problem. The danger is that the L2 learner may overgeneralize the form to inappropriate settings. Although the two groups share similarities in compliment responses, they also differ in important ways. In responding to compliments, US recipients are much more likely than Syrians to use Appreciation Tokens (e.g. thanks). The infrequency of this response in the Arabic data suggests that the utterance Shukran (‘thank you’) by itself is not usually a sufﬁcient response to an Arabic compliment and needs to be supplemented by additional words. By itself, it may sound ﬂat and awkward because it appears to signal the end of the conversation. As illustrated

chapter |1 pages

at the beginning of this article, American students of Arabic may respond to a compliment given by a native-speaker of Arabic by saying Shukran. If the intent of the American, drawing from his or her L1 strategies, is to respond in an appro-priately polite manner and if the native Arabic speaker interprets the force of the utterance differently (e.g. that Speaker wants to end the conversation), pragma-

chapter 1986|11 pages

175). English-speaking students of Arabic, on the other hand, may have difﬁculty with the number of words in Arabic compliment/compliment response sequences, particularly with the formulaic expressions expected in response to cer-tain compliments. If the length of the sequence results in their feeling phony and insincere, they may fall back on their L1 strategies. In this case, pragmatic fail-ure may result not from too many words, but from too few. (Revised version received October 1995)