From Palmer, 2000: 669. ‘C’ is the interviewer, ‘R’ is the person

Cognitions in action The next three extracts show speakers using a variety of words and phrases which either report or invoke the relevance of cognitive processes or mental states. The first comes from routine conversational interaction.

6.1 From JS:II: 219–20

Ben: Lissena pigeons Ellen: Coo-coo:::coo::: Bill: Quail, I think Ben notices the sound made by pigeons. But at the same time that Ellen begins to mimic the sound of the birds, Bill corrects Ben by pointing out that they are quail, not pigeons. This correction is accompanied by a report of Ben’s state of mind concerning the birds: he thinks they are quail. The second illustration is taken from McGuiniss’ 1983 account of the inves- tigation of a notorious murder in 1970. Captain Jeffrey MacDonald, a respected army doctor, claimed that drug-crazed hippies broke into his apartment, knocked him unconscious and brutally murdered his family. The police, how- ever, suspected that MacDonald was the murderer. McGuiniss was able to inter- view the main people involved in the investigation, including MacDonald himself. The following passage comes from a taped interview with MacDonald in which he is describing how, during a mealtime in the Officers’ Mess, he first heard that the police had named him as the prime suspect in their investigation.

6.2 From McGuiniss, 1983: 168

I was standing in line getting food, and I had just gotten through the cash register area and was beginning to sit down, when they had a news bulletin that Captain Jeffrey MacDonald, the Green Beret officer from Fort Bragg who six weeks earlier had claimed that his wife and children were brutally beaten and stabbed by four hippies, was himself named chief suspect. And I remember the truly – I don’t mean to use clichés, but I don’t know how else to explain it – the room was spinning again. MacDonald is able to recall in some detail this traumatic moment. He has clear and detailed recall of his movements at the time, and recollection of the news broadcast itself. Clearly then, here is evidence for the operation of memory, and a good one at that: the detail of MacDonald’s description implies the successful operation of processes by which information or perceptions at the time are stored as memories and than later retrieved. Finally, the last instance comes from an interview with two young women about personal style, appearance and membership of youth subcultures. 114 CONVERSATION ANALYSIS AND DISCOURSE ANALYSIS 关关

6.3 From Widdicombe and Wooffitt, 1995: 96–7

1 I: can you tell me something about your style and the way 2 you look, 3 0.7 4 I: how would you descri:be yourselves 5 0.7 6 R1: hh 7 0.7 8 R1: I dunno I hate those sorts of quest ions uhm 9 R2: yeah horrible 10 isn’t it Consider the first respondent’s R1 reply to the interviewer’s question: what kinds of inner psychological conditions are being reported here? First, she refers to her knowledge: ‘I dunno’ indicates that the speaker either does not know how to describe herself, or is at least uncertain as to how to go about that task. Second, her claim, ‘I hate those sorts of questions’, displays a firm stance towards a particular object. We can see, then, that in the space of a few words, the speaker has drawn upon and described her state of knowledge and has expressed an attitude. We might be tempted to assume that cognitive terms are used because they correspond to or represent inner mental states: we say ‘I dunno’ because our brains do not contain the relevant information; we provide detailed reports of past experiences because the information is stored in our memory, and so on. This is a common assumption in traditional psychology. It may not be so crudely expressed: psychologists may work on the assump- tion that in principle language can reveal the workings of the mind while tak- ing account of the various ways in which those processes might be obscured or distorted. But it is not hard to find instances in which, broadly, the way language is used is taken to reveal the causal influence of mental states, or the basic properties of cognitive entities. For example, in social representa- tion research, it is assumed that talk can be examined to investigate the expression and use of underlying cognitive representations for example, Jodelet, 1991. And in research on various kinds of autobiographical memory, the contents of reports of past experience are taken as more or less accurate indicators of what information is actually stored in the head; the goal for the psychologist is to explain how it got there, and how it was sub- sequently accessed for example, Brown and Kulik, 1977. Thus traditional cognitivist psychology regards language as a window on, or expression of, the workings of cognitive procedures. Discourse analysis, however, challenges this version of the relationship between discourse and cognition. It examines talk and texts to show how des- criptions and reports have been constructed to perform interactional or inter- personal functions. It studies what people do in language. The discourse analytic focus on the way that language is used to do things has implications DISCURSIVE PSYCHOLOGY 115 关 for the study of cognition. First, it means that we cannot rely on people’s accounts to reveal the inner properties of the mind as the organisation of dis- cursive acts might be informed by the social actions for which they have been designed. But more radically, it suggests a new research programme which takes as its subject the ways in which references to and descriptions of mental states, and a cognitivist vocabulary, are used to perform social actions. For example, in discourse analysis, the attribution of causal relationships is analysed as a discursive activity, designed and performed with respect to the interpersonal context for which it is produced. A speaker may be attributing a relationship between two events not because of the operation of some cognitive processes by which causal relationships are identified, but to attend to inferential concerns such as managing the imputation of blame, warranting the factual basis of a claim, and so on. There was a concern to explore the implications of the action orientation of language for traditional psychological concerns early on in discourse analytic research. The term ‘discursive psychology’ was adopted to focus on this aspect of discourse studies. Discourse analysis is particularly concerned with examining discourse for how cognitive issues are dealt with. … It is a further development of these issues here which leads us to move beyond talking of discourse analysis and to describe the enterprise as ‘discursive psychology’. Edwards and Potter, 1992: 29 Discursive psychology, then, is focused on the ways in which cognitive notions can be treated analytically as situated practices which address interactional and inferential concerns in everyday circumstances. Discursive psychologists ask: What does a ‘memory’ do in some interaction? How is a version of the past constructed to sustain some action? Or: what is an ‘attitude’ used to do? How is an evaluation built to assign blame to a minority group, say, or how is an evaluation used to persuade a reluctant adolescent to eat tuna pasta? Potter, 2000: 35; original italics So what do they do? How do they work? We will begin to sketch some of the main analytic issues by considering the three kinds of cognitive states illus- trated in extracts 6.1 to 6.3: reports of ‘thinking’, memory; and in particular, memory formulations which indicate good recall, and ‘I dunno’-type reports of a lack of knowledge. Thinking What happens when we characterise a claim or position as something we ‘think’ is the case? Consider the following example. 116 CONVERSATION ANALYSIS AND DISCOURSE ANALYSIS

6.4 From Atkinson and Drew, 1979: 58

1 B: Uh if you’d care to come over and 2 visit a little while this morning 3 I’ll give you a cup of coffee 4 A: hehh Well that’s awfully sweet of you, 5 I don’t think I can make it this morning 6 hh uhm I’m running an ad in the paper and and uh I 7 have to stay near the phone. B invites A round for coffee. In declining this offer, A says ‘I don’t think I can make it this morning’. This is the refusal component of a dispreferred response to an invitation or offer Heritage, 1984a: 266. As such it handles delicate issues concerning the face of the person making the inviting. A blunt refusal might well appear brusque and insensitive. To say ‘I don’t think X’, however, modulates the strength of the refusal, in that it portrays the uncertain or conditional basis of the action. Now consider the exchange in extract 6.1. Ben: Lissena pigeons Ellen: Coo-coo:::coo::: Bill: Quail, I think Again, sensitive matters are at hand, as Bill is correcting Ben. Overt repair of this kind is a delicate interactional activity because it may be taken as a slight, a ‘put down’ or deliberate rudeness. Repair work, then, has potential implica- tions for the coordination of the interpersonal relations of the relevant parties Jefferson, 1987. ‘I think’ formulations address this potential sensitivity because they modify the force of the correction Schegloff et al, 1977. ‘I think’ formulations of knowledge claims are rooted in social activities. They allow speakers to manage sensitive interpersonal matters in delicate and subtle ways. Of course, this is not to say that this is the only kind of work they do: a withering scepticism can be achieved by following a report of someone’s stated plans or intentions with ‘I don’t think so’. The point is that the design of utterances is informed by the discursive actions and contexts for which they are produced. It is a mistake to assume that reference to a speaker’s ‘thoughts’, ‘thinking’ and so on simply expresses the operations of inner mental states. Memoryremembering For discursive psychology, the focus is on remembering and forgetting, and these are treated as social actions embodied in everyday social practices. Verbalised recollections are treated as ‘pragmatically variable versions that are constructed with regard to particular communicative circumstances’ Middleton and Edwards, 1990: 11. For example, Edwards and Potter’s study of John Dean’s DISCURSIVE PSYCHOLOGY 117 关关 testimony to the senate committee investigating the ‘Watergate’ scandal explores the contextualised and pragmatic work embedded in memory formulations Edwards and Potter, 1992: 30–53. Their analysis was prompted by an earlier study by Ulric Neisser, a cogni- tive psychologist who had been trying to broaden the scope of psychological research on memory to take account of ecologically valid data, such as mem- ories of real events for example, Neisser, 1982; Neisser and Harsch, 1992; Neisser and Winograd, 1988. The publication of the transcripts of the senate hearings allowed Neisser to investigate the extent to which Dean’s recall, noted at the time for its extensiveness and apparent detail, was accurate. He claimed to have identified three kinds of memory functioning in Dean’s testi- mony: verbatim, gist and repisodic memory. Neisser argued that repisodic memory works to preserve the key themes of an event or discussion while allowing for errors. In short, it allowed Dean to be telling the truth while ostensibly getting things wrong Neisser, 1981. While welcoming Neisser’s attempt to move the study of memory beyond the laboratory, Edwards and Potter argued that his analysis was problematic because it confused accounting practices with memory processes. Aspects of the testimony which Neisser had seen as evidence of good recall, bad recall or a certain type of recall verbatim, gist, repisodic were ways of managing Dean’s accountability in a courtroom setting; they were methods for dealing with actual or implied attributions of guilt, and so on. So, for example, producing the gist of a prior event or conversation allows the speaker to perform delicate descrip- tive operations: they can preserve, delete or transform aspects of the prior talk to attend to current interactional purposes Heritage and Watson, 1979. Moreover, verbatim recall of prior conversation is a rhetorical strategy through which the speaker is able to manage the evidential value of their claims Goffman, 1981; Holt, 1996; Wooffitt, 1992. Thus what Neisser interpreted as expressions of the workings of memory were discursive practices oriented to inferential tasks generated in the context of official hearings to identify responsibility and blame for illegal activities see also Molotch and Boden, 1985. Dean was noted for his good memory: he drew attention to it at various points in the hearings, and on occasions was able to give detailed accounts of past events. Edwards and Potter 1992 select some illustrative examples from Neisser’s 1981 paper, which serve to show that Dean presented himself as … someone with a virtually direct perceptual access to the original events: ‘you know the way there are two chairs at the side of the President’s desk … on the left hand chair Mr. Haldeman was sitting’ [Neisser] 1981: 11; ‘I can very vividly recall the way he sort of rolled his chair back from the desk and leaned over to Mr. Haldeman and said “A million dollars is no problem” ’. 1981: 18, cited in Edwards and Potter, 1992: 42 However, Edwards and Potter argue that we must not take Dean’s pronounce- ments on his own memory abilities as literal or neutral statements as they are 118 CONVERSATION ANALYSIS AND DISCOURSE ANALYSIS