To appear in te Molder, H. & Potter, J. (Eds) (2005). Conversation and Cognition. Cambridge: Cambridge University Press.

11

Discursive psychology, mental states and descriptions

Derek Edwards and Jonathan Potter

Introduction Our aim in this chapter is to show how discursive psychology (DP) deals with psychological states and characteristics. We do this in several ways: by dening what DP is, by demonstrating it analytically, and by discussing various criticisms and misunderstandings of it. As for dening it, DP works in three closely related ways: 1. Respecication and critique. Standard psychological topics are respecied as discourse practices. Topics recognized in mainstream psychology such as memory, causal attribution, script knowledge, and so on, are re-worked in terms of discourse practices. We study how people ordinarily, as part of everyday activities, report and explain actions and events, how they characterize the actors in those events, and how they manage various implications generated in the act of reporting. DP often generates a critical stance on cognitive psychology. For example, cognitive theory and measurement of attitudes is criticized and replaced by the study of argumentative and evaluative practices in discourse (Billig, 1987; Potter, 1998a; Potter and Wetherell, 1987; Wiggins, 2002; Wiggins and Potter, 2003). Similarly, cognitive methods and theory on causal attribution are critically opposed by analyses of how people manage accountability in everyday talk (Antaki, 1994; Edwards and Potter, 1992a, 1993). 2. The psychological thesaurus. DP explores the situated, occasioned, rhetorical uses of the rich common sense psychological lexicon or thesaurus: terms such as angry, jealous, know, believe, feel, want, and so on. For example, expressions such as I dont know, or your angry stage are examined for the local contrasts and interactional work for which they are used (e.g., Edwards, 1995; Potter, 1998b). By grounding such studies in empirical materials, we are able to explore the ways in which concepts such as know or angry are used interactionally and rhetorically, with regard to specic, locally relevant alternative descriptions. We develop some examples in this chapter.
241

242

Conversation and cognition

3. Managing psychological implications. DP examines discourse for how psychological themes are handled and managed, without necessarily being overtly labelled. We explore how agency, intent, doubt, belief, prejudice, commitment, and so on, are built, made available, or countered indirectly, through descriptions of actions, events, objects, persons and settings. In fact, this has been a key feature of the kinds of respecication done in DP type 1, where attributions of intent and blame are shown to be handled, not by overt descriptions of intent or motive, but through what look like (or are produced as) straightforward event descriptions. This is the basis of DPs explorations of fact and accountability (Edwards and Potter, 1992b; Potter and Edwards, 2001), where we show how factual descriptions are used to implicate a range of psychological states and attributions, and vice versa. Again, this kind of psychologically implicative use of factual descriptions is closely tied, in actual talk and its analysis, to participants uses of the psychological thesaurus. This is a good point at which to address a recurrent misunderstanding about DP. It is sometimes assumed that DP is concerned only with overt talk about mental states, what we have just listed as DP type 2. For example Coulter summarizes Edwards (1997) to the effect that in DP the mental is thus to be construed solely in terms of what people say about it (1999: 166). We will return to Coulters characterizations of DP later in this chapter. Similarly, Drew (this volume) identies DP with what he calls the attributional approach, which is the study of the rhetorical uses of terms describing mental states (p. 00). But exploring uses of the psychological thesaurus is only part of DP. DP also explores how mental states feature as talks business, rather in the way that CA deals with the relevance of institutional settings and social structures. Psychological categories are analyzed as matters being handled, managed, produced, made relevant (etc.) in the talk, rather than as something sitting outside of the talk, for analysts to use in explaining it (Edwards, 1997; Potter, 1998b; cf. Drew and Heritage, 1992). One further point of clarication: when we use the term psychological, we do so under the auspices of DP. We use it provisionally, commonsensically, bracketed off for respecication, or else as DP re-denes it, but not as any kind of commitment to an inner life of the mind, nor to individualism, nor to whatever else academic psychology might assume or propose. Yet the main thrust of DP is not to close down psychology departments, but to counter and invert what mainstream psychology has done with discourse, which is to treat it as the expression of thoughts, intentions and cognitive structures. The inversion offered by DP is to start with discourse itself, and to see how all of those presumptively prior and

Discursive psychology and mental states

243

independent notions of mind, intention, motive, etc., are topicalized, categorized and, in various less direct ways, handled and managed within discourse itself. So, DP focuses on person and event descriptions in talk and text. It examines how factual descriptions are assembled, how they are built as solidly grounded or undermined as false, and how they handle the rational accountability (or otherwise) of actors and speakers. We focus particularly on what we (provisionally) call mind and reality on how people deploy common sense notions of an external reality as a kind of setting for, and evidential domain for inferences about, a range of mental states and personal characteristics. We also emphasize rhetorical organization, how descriptions and their inferences routinely (and not only in adversarial contexts) attend to possible or actual counter versions. Descriptions are constructive of their objects. This is not to say that talk brings things into the world, but rather, that descriptions are categorizations, distinctions, contrasts; there are always relevant alternatives available. This permits descriptions to be performative; they offer one construction rather than another, produced in sequential and rhetorical contexts, where the specics matter for the actions being done. Again, this is a suitable juncture for some clarication. The sense of construction that we are using here, is what we have called epistemic rather than ontological (see, for example, Edwards, 1997: 478). In his review of various sorts of discursive psychology, including the latter volume, Coulter glosses DP as a thesis which proposes that the human mind and its various properties are generated in and through discourse: in essence, the mind is revealed in and through analyseable features of the things that people say and do through their talk (1999: 163). In fact, the DP that we promote here, and that Coulter reviewed, is programmatically opposed to any of that. The term epistemic construction is designed to distinguish DP from constructivist developmental-psychological theories such as Piagets, Vygotskys and Bruners in which actual minds (Bruner, 1986) are produced and shaped through language and action. Similarly, Coulters further gloss on DP, that minds are revealed in discourse, is clearly repudiated in DP; indeed that repudiation is virtually the essence of it. It is probably the most salient characteristic of DP, at least as we practice it, that it rejects the cognitivist assumption that minds are revealed or expressed in what people say. To use a vernacular expression, it is where we came in! Coulter has blended together a range of different and sometimes opposed approaches that have adopted the term (social) construction, rather as he did, in his major work The Social Construction of Mind (Coulter, 1979).

244

Conversation and cognition

While opposing the cognitivist assumption that talk is driven by the workings of an inner life of the mind, that is not the end of the matter. The status of it as a poor general theory of language and mind does not prevent people from making use of it as a way of talking. This is not merely a matter of people making false theoretical assertions such as I think in my head (Coulter, 1999: 166), but of talking as if that were the case, as part of talks everyday practices. There are practical, common sense uses of such a notion, that a persons words may be produced or taken on occasions, to be expressions of a private and prior realm of mental life. This can be a practical basis for talking and doing things with words, whose investigation requires no commitment to mentalism on the analysts part. Descriptions, inferences and not knowing In our three-way typology of DP, we distinguished between investigating uses of the psychological thesaurus, and studying how event descriptions may be produced for their psychological implications. But even the overt labelling of psychological states needs to be analyzed within the descriptive, narrative and turn-by-turn sequences where they occur, where inferences about intentions, thoughts, feelings, motives (and so on) are made available and countered in how events are told. Extract (1)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 J:

[DE-JF:C2:S1:p.9]

And uh:: (1.0) Connie had a short skirt on I donknow. (1.0) And I knew this- (0.6) uh ah- maybe I had met him. (1.0) Ye:h. (.) I musta met Da:ve before. (0.8) ). But Id heard he was a bit of a la:d ( He didnt care: (1.0) who he (0.2) chatted up or (.) who was in Ireland (.) yknow those were (unavailable) to chat up with. (1.0) So Connie stood up (0.8) pulled her skirt right up her side (0.6) and she was looking straight at Da: ve (.) > like that < (0.6) and then turned and looked at me (1.2) and then she said w- (.) turned and then (.) back to Dave and said (.) by the way that wasnt for you.

Discursive psychology and mental states

245

Let us start with how descriptions of things, actions and events can provide for psychological inferences. This is a small, specic illustration of a very general feature of everyday discourse. Extract (1) is part of an extended sequence from an hour-long counselling session, in which a husband and wife (Connie and Jimmy) are recounting their problems with each other. Connie has described Jimmy as an endemically jealous person, prone to recurrent and unreasonable ts of jealousy.1 The descriptive detail in line 1, Connie had a short skirt on, depicts her (and her subsequently narrated actions build on this) as dressed in a sexually relevant or provocative way. It is part of how Jimmy builds his wifes character and motives as irtatious (and therefore making his jealousy reasonable), along with directing her behaviour on this particular occasion at a reputed bit of a lad named Dave (line 9), in which Dave is positioned as a recipient of Connies actions and appearance. The descriptive details work, not only through what they include, but through what they might well have mentioned but do not. Of all that she wore, it is only the short skirt that is mentioned (here or elsewhere in the session), and of all the skirts describable characteristics, only its brevity is remarked on. The sequential positioning of the description is also important. It occurs immediately before (and thus hearably relevant to) the introduction of Dave. We can hear Connies motivation and character being set up here, as irtatious, sexy and targeting Dave (explicitly, in lines 1516). The specic words that we are using, sexy and irtatious, are ours and not denitive (though they are used later by the counsellor as formulations of what Jimmy was saying), but the point is that something like that is implicitly conveyed. It is done not by overtly calling his wife sexually motivated and irtatious, but by making those kinds of characteristics inferentially available, as categories that a hearer of these descriptions can infer for themselves. Indeed, they are categories that Connie herself orients to in a similarly indirect manner (in a subsequent turn not presented here), by providing alternative, contrastive descriptions of the length of her skirt, what she did with it, and at whom she was looking. So Connies character, motives and intentions (psychological matters) are built and countered by descriptions of witnessable things her skirt and her actions, the proximity of those things to the descriptions of Dave, and their place in the narrative sequence. Consider now the use of an overt psychological term, I donknow, in line 2. It would be a mistake to hear this as simply an assertion of ignorance or uncertainty, or even as an assertion at all (it is said parenthetically, with no explicit object). What it does, like the rest of the sequence in which it occurs, is attend to Jimmys own character, as a

246

Conversation and cognition

purportedly jealous and suspicious husband who may be prone (in this case) to some kind of obsessive monitoring of the details of his wifes clothing and behaviour. The use here, and just here, of I donknow counters that. It implies that he wasnt paying particular attention, and does not have a lot hanging on it. In fact, this kind of interpersonal use of I donknow or I dunno (used in this parenthetical, framing kind of way, rather than as a bald answer to a factual question), recurs across a range of discourse materials as a way of handling, or playing down, the speakers stake or interest in the content of a description (Potter, 1998b; see also Beach and Metzger, 1997). Clearly, any such uses of I dont know are not independent of what know signies as a concept2 or its dictionary denition. However, what Jimmy does with it is not well catered for by this conceptual understanding. Indeed, his subsequent narrative conveys his close attention to the details of Connies skirt, its length, and what she did with it. Expressions such as I dont know can do interactional work of this kind, attending to potential common sense inferences and rhetorical alternatives at stake in the interaction; such as, in this case, that he is a suspicious, over-jealous husband prone to persecuting his wife over small, exaggerated things that are more in his head than in the world, more a reection of his own preoccupations than of his wifes actual clothing, actions or character. Clearly, the way that discourse handles psychological categories is not just a matter of using overt psychological labels. What makes these kinds of analytic observations more than merely ad hoc comments on particular stretches of talk, is that they are recurrently applicable to a wide range of materials. Motives and intentions (vernacularly understood) are built inferentially out of descriptions of actions and events; they are built to contrast with alternatives; they attend to matters local to the interactional context in which they occur; and they attend reexively to the speakers stake or investment in producing those descriptions. There is also a range of devices through which this kind of thing is done, including situated uses of I dont know, and various linguistic ways of establishing a persons recurrent actions as implying motive and character (see Edwards, 1994, 1997 on script formulations). Knowing, telling and wanting Wittgenstein (1958) argued that what we are calling mental state avowals (i.e., descriptions of ones own thoughts and feelings) do not and could not obtain their meaning from referring to privately experienced mental states. Bilmes (1992) has suggested that folk may nevertheless actually, and reasonably, consider themselves to be doing that, or may talk as if they were doing that. Coulters (1992) basic response is that Bilmes and

Discursive psychology and mental states

247

the lay public are wrong about it, at least as a theoretical explanation, or conceptual analysis, of what they are doing. Yet we have suggested that the concept of private mental states, and of the ability to provide experiential reports about them (which is clearly a meaningful idea, if only to allow Coulter and ourselves to call it cognitivism and oppose it!), may be a rhetorically useful way of talking. We can ask how, when, and in the performance of what kinds of actions, do people talk as if they were in possession of a privately available mental life, which their words may either truthfully or falsely express? A routine account for not telling something is not knowing it. Indeed we can invert that, and say that handling the practical accountability of not telling something, where telling is called for, is a routine environment for formulations of not knowing. So not knowing is analyzable for what it does, not just for what it says about states of mind. However, in extract (2) W must rely on a different kind of account for not telling, given that it is his own recent actions that he is talking about. Here he knows but is not saying. W is a witness being questioned by a police ofcer (P1) about the theft from a shop, and the subsequent selling on, of some cigarettes: a theft to which W has confessed. Extract (2)
1 P1: 2 3 W: 4 5 W: 6 7 W: 8 W: 9 10 11 P1:

[DE:West:p. 10]
So: (.) where were they sold to (1.2) In a pub in Pen. (2.4) Im ngonna tell ythe name a the pub. (1.8) Bcause: I jus don want the person in there (0.2) involved (.) the- manager in- involved in any a this, (0.4) Right,

The notion that W might normatively be expected to tell what he relevantly knows, at least under questioning in this setting, concerning the disposal of stolen goods, is oriented to by the fact that W accounts for it (lines 79). Note how the initial response in a pub in Pen (line 3) answers the prior question, but in a manner that would not allow the police ofcer to specically identify the place. Ws statement, that he is not going to provide specic identifying information (line 5), and the account that follows it (lines 79), are provided only after delays in lines 4 and 6 that W may be taking as signs that his answers are insufcient. The delay in line 10, prior to P1s receipt of the account, is notably shorter.3

248

Conversation and cognition

Clearly in this case, knowing and telling are handled under recognizable and recognized (line 11) normative auspices. The content of Ws account for not telling includes another common sense psychological avowal, not wanting (line 7). Again, what W may or may not want is produced as part of a normative order, recognized by P1 in line 11, not to get someone else into trouble. Avowals of this kind, of knowing and wanting, are produced not simply as reports on mental states, as if for interests sake, but as claims that are recognizable and somewhat conrmable within a conventional framework. There is a kind of intersubjectivity at work, in which thoughts and wishes are made plausible in context, with regard to what a person might be expected to think, do or feel under the circumstances. Of course it works both ways, and this is important in Ws general uses of it. Given the availability of circumstantial, normative warrants for psychological avowals, those psychological avowals can work reexively, to establish Ws version of circumstances (the events he is describing) as precisely what happened. W makes regular use of mental state avowals and their normative reportability, in building factual versions of the events that those mental states are about. This is done while attending to his own accountability in, and through, reporting some things and not other things that he is being questioned about. Extract (3) is an example parallel to extract (2). Extract (3)
1 P1: 2 3 W: 4 5 6 P1:

[DE:West:p. 22]

Who are you owin the money to? (2.6) I don wanna h (0.3) give you his name because a the(h) the e(h)nd a the day hes a dealer? (0.3) Yeh

Again, W accounts for not providing the requested information, an account also based on not wanting to get another person into trouble, and again P1 acknowledges it (line 6).4 There is an additional relevance beyond this short extract, in that owing money is Ws repeated account for stealing the cigarettes and selling them on; he claims to have needed the money to pay a debt for some cannabis, and was being threatened by the dealer if he did not pay up. So Ws failure to substantiate that story is less incidental than it might appear in the extract. As seen in extract (2), refusal (not inability) to comply with a request is accounted for in terms of a normatively overriding, indeed somewhat laudable, desire or motive to protect another individual. The fact that the individual concerned is a drug dealer probably heightens rather than softens the obligation, given the kind of trouble they would be in with the police, and perhaps the

Discursive psychology and mental states

249

trouble that W might be in with the dealer. Ws laughter (the (h) signs) and smiling delivery (bracketed by the signs) orient to that problem, as perhaps both obvious and ironic, but in any case recognizably to P1. W and P1 are using, as part of practical accountability, a common sense notion that what we know is tellable, but sometimes may and should remain a private matter, within a eld of normatively recognizable motives and circumstances. Further uses of wanting, hoping and liking A persons wishes, motives, likes or wants can play a crucial part in establishing or undermining contested versions of events as factual. In DP we have called this stake management (Edwards and Potter, 1992a; Potter, 1996). Claiming that a person has some kind of stake, interest, wish or motive in favour of their particular version of events, can be a way of undermining that version. As a corollary, people routinely work at dening preferred versions as disinterested, or even as contrary to their hopes and interests (stake inoculation). More broadly, we are dealing with the common sense play-off between reality and mind, within forms of practical reasoning and accountability. Avowals of hoping and wanting, and other psychological states, work in concert with, or in contrast to, factual descriptions of circumstances and events, particularly in environments of scepticism or dispute. Extract (4)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 W:

[DE:West:p.16]
(...) I didn push the woman or nothin, (0.5) I really did not do that.= Im not that type of person yknow what I me:an. .h Fair enough I stole the ciggies, (0.9) I wouldn hurt an old lady. (0.5) [Right.] [No. ] (.) Not a chance.= =What Im saying is that this (0.5) I hope theres a Ceet- a camera at the shop= =There i[s. [security [camera ] [There were] cameras in the shop, but Ill tell y now that the cameras (0.7) werent (1.1) functionin, (.) at that particular time. Bit coincidental innit?

P1: W: W: P1: W: P1: W: P1:

W:

250

Conversation and cognition

The factual dispute is displayed in the opening lines. W is denying a version previously put to him by P1, citing the old lady shopkeepers testimony, that he came across the shop counter and pushed her, causing an injury, turning the admitted theft into a more serious crime (robbery). W emphatically denies this, using the extreme expressions (Pomerantz, 1986) or nothin, really, and not a chance (lines 1, 2 & 10). He ties his denial to a claim about himself as a kind of person (lines 25), using the practical logic of scripts and dispositions (Edwards, 1995, 1997), building the case that he did not do it on this particular occasion because he would not do it on any occasion. Again, we have psychological formulations working alongside contentious event descriptions, building one version and undermining another. We can now focus on something for which these opening lines provide a relevant setting: the mental state avowal I hope (line 12), and its object. One would not generally expect a person to hope to see the evidence of their crime on videotape. It is not merely that the video evidence is potentially decisive. In fact W may already know that there is no such evidence; it becomes clear earlier in the interview that W has been briefed by his solicitor on the content of the shopkeepers testimony. It is the hoping that works rhetorically here. It signals, in the absence of the tape recording, the truth of the object of that hope, the factual content of what any such tape would reveal, which of course corresponds to Ws contested version of events. Ws I hope is a mental state avowal at work in a rhetorical environment of stake and interest (cf. Edwards and Potter, 1992a), where Ws contended version of events risks being heard as functional, self-serving, inaccurate, motivated. What W hopes for (the existence of a video recording) counters an alternative motivational story, alive in the interrogation by P1, that he is lying in order to downgrade the seriousness of his crime. So mental state avowals can work to reify their objects the states of affairs, as described, to which they pertain. In extract (5) a second police ofcer, P2, is asking about the possible involvement of an accomplice named Alan. Extract (5)
1 P2: 2 3 W: 4 5 W: 6 7 W: 8 9

[DE:West:p.24]

When did Alan come into the shop with you then. (1.8) I didn wan im in the shop. (1.5) Didn wan im in the shop. (0.2) Jus told im to walk down I said Ill be back now Im goin to get ciggies. (1.8)

Discursive psychology and mental states

251

It has been a major and repeated feature of the interview up to this point, that W consistently denies the shopkeepers version, repeatedly put to him by P1 or P2, that he committed the theft with an accomplice, Alan. W has insisted that Alan never came into the shop and took no part, but rather continued on down the road and waited for him. So P2s question in line 1 is hearable against that background of repeated proposal and denial. Ws response is not simply to deny it again, but to offer a mental state description that provides a credible basis for that denial he did not want Alan in the shop. W goes on soon after to explain this as a feature of his routine modus operandi when stealing cigarettes from shops, that it just complicates things: >The enda the day< its easier for me to get away on me own than havin tlook out for somebody else as well. But by reporting on his state of mind with regard to the matter, W provides a credible basis for his contested (and otherwise uncorroborated) version of events. Those events are again the intentional object of the not wanting i.e., having him in the shop (lines 3 & 5). As with extract (4), Ws mental state avowals work closely with his event descriptions, establishing the factual plausibility of his version of events within a normative order of mind and reality of how one would think, what one would want or hope for or expect, given the truth of events as he tells them. Later in the same interview (see extract (6)) W is again referring to Alans (lack of) involvement in the theft from the shop. Extract (6)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 W:

[DE: West: p. 37]

this is truly, (0.5) right, (0.3) from the bottm me he art, (0.5) he was not with me, (0.6) >know waramean<, (0.3) tell the truth I don even like the lad, (0.8) hes implicated me (.) in somein. (0.5) Right? 0.4) In a burglary (0.6) that Ive had no (.) no part of? (1.4) I dont really like im, hes more of an acquaintance >yknow waramean.< (0.4) But (.) he (.) did not have nothin to do with this (0.3) If you don like im youve been (.) youve put im up in your at haven you.

W: W:

W:

W:

P2:

252

Conversation and cognition

Not liking Alan (lines 4, 12), whom W has been consistently defending as not involved in the crime, serves as a stake inoculation (Potter, 1996). Note various details of Ws talk: the term even, in I don even like the lad (line 4) signals the rhetorical, contrastive nature of this avowal, working against what might be assumed, which is that W has a motive to lie, in order to protect a friend. Again, the category acquaintance is introduced in the same contrastive way; Alan is more of an acquaintance (lines 1213). The distancing between W and Alan is further brought off by referring to him impersonally as the lad (line 4). So the whole construction, I don even like the lad, works together, through all its detailed components (including contrastive stress on like), to inoculate W from any motive he may be imputed with, to lie on Alans behalf. Further, W actually claims a contrary motive to implicate Alan, as a piece of quid pro quo for Alans having implicated W in another crime of which W was innocent (lines 410). So again, as in extracts (4) and (5), we have a mental state avowal (not liking) serving to reinforce the factuality of a particular, disputed version of events. The avowals in extract (6), occurring in an environment of contestation and scepticism, are accompanied by emphatic claims to sincerity: truly, from the bottom of my heart, tell the truth, really. These expressions exploit one of the features of mental state avowals, which is their common sense (if philosophically disputed) reference to private mental states, that are the speakers privilege to know and report. W is claiming to speak sincerely from within, indeed from the bottom of his heart. One useful thing about talking in this manner, is that it can counter evidence to the contrary, though of course its rhetorical uses against such evidence do not guarantee success. That is what we see in extract (6), where P2 counters Ws avowals to dislike this mere acquaintance, on the grounds that W has acted in a manner associated with friends, putting him up at his at. In membership category terms, such actions are associated with friends rather than disliked acquaintances. We are not pursuing here the consequences of this dispute between W and P2. Rather, the important things are the conceptual, discursive resources through which the dispute is conducted. Descriptions of witnessable events in the world (what happened in the shop, Ws putting Alan up for the night) are offered in support of, or against, mental state avowals. Similarly, and as a corollary, claiming a mental state or psychological disposition (liking, hoping, wanting, etc.) can be used to build or refute a particular version of external events. Within these common sense descriptive practices, mind and world are invoked in concert or in contrast with each other. Whereas one can justiably argue that all words, including mental state avowals, can be meaningful only via their public

Discursive psychology and mental states

253

uses and ratications, it remains part of such public uses, that people talk as if there is a world of inner experience that can be reliably reported on by the experiencer and set against evidence to the contrary. Some situated uses of thinking A cognitive conception of thinking is that it refers to mental cogitation, ideas occurring in the mind, a process of ratiocination, information processing, or mental problem solving, whether conscious or unconscious. Of course the term can be used that way in everyday talk, to propose that some train of ideas is passing through our minds (whatever that might, in turn, amount to). Billig (1987, 1999) suggests that thought is intrinsically more social than psychologists have generally taken it to be, that it is better characterized as argumentative and ideological in nature. Also, the earliest denitions of think in the Oxford English Dictionary (2002) specify it as deriving from old English and Anglo-Saxon, with an initial sense of to seem, appear. This provides for a function in the discourse of mind and world that we are discussing in this chapter: a contrast between how things appear, or are assumed to be, and how they turn out. So we are able to say I thought so! when saying that assumptions are conrmed, or else At rst I thought . . . when saying that things turned out otherwise (see Sacks, 1992: 787ff; and Wooftt, 1991, for useful observations on that). Extract (7) is rich in uses of the verb think, including echoes of this ancient sense of proposing alignment or misalignment with events in the world. Extract (7)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 L: M: L: L:

[Holt:1:1: p.7]

M: L: M: M:

.hh Oh m:Mum, I thought she wz gonsendim mone:y. She told me she was. (0.3) Oh::::. (1.4) .hh I think Ill aftuh teller in the letter not tuh send any mo::re. (1.4) ( ) that wd be a good i:dea (.) Yes. You dont think shell take umbrage, (.) No::, (.) N-o:? (0.2)

254 18 19 20 21 22 23 24 M:

Conversation and cognition Uh: no: if you jussay thet uh (0.5) uh if yuh (0.3) tha- thank her very muchn say thet (0.2) but Gordon doesnt wear that sortv thi:ng= =mn::Ye:s. (0.4) I think thatd be a good idea. Ye:s. I think so:

L: M: L:

In extract (7) Lesley (L) and her mother (M) are talking about the prospect of Ann (named elsewhere) sending a gift for Gordon, Ls son. It is a gift that Ann has sent on previous occasions (line 8), some white Tshirts with a company logo on them, items that Gordon reportedly (line 20) never wears. The thing of interest is the ways in which notions of thinking are occasioned and accountable in this public way, rather than merely reports of a private life of the mind. One feature of their accountability is projected by past or present tense, thought versus think. As Sacks (1992: 7878) noted, both tense and intonation are important to the factual status of whatever was thought. I thought (line 1), together with the emphasis and falling pitch on money, signal that what was thought turned out not to be the case; money was not sent. Using past tense, saying what one thought can be a way of setting up what actually happened as something normal, routine, deviant, unusual, indicative of the actors and current speaker, and so on. Further, in the event that such thoughts turned out to be wrong, it is an accountable matter why the mistake was made. Accounts may take the form of how unusual the events were, and/or how reasonable the presumptions. In this case, a reasonable basis is provided by M (line 3): it is what Ann said she would do. On the other hand, present tense uses of think (lines 7, 12, 23, 24) are used here to project actions and events that are as yet unknown, or have not happened. These are: the proposal to tell Ann not to send any more t-shirts (lines 78), and the consequence that Ann might take umbrage (line 12). Ls telling M what she is thinking is a way of sounding M out (occasioning a second assessment from her: cf. Pomerantz, 1984) on the delicate matter of rejecting a gift and anticipating Anns possible reaction. As with making false predictions (I thought X but Y happened), rejecting a gift is an accountable matter. The accounts provided here are: (1) that it has to be done (line 7), where have to routinely contrasts with what anyone might freely choose to do (cf. Drew, 1984, on inability accounts in invitation refusals); (2) that it is a good idea (lines 9, 22, 23); and (3) that Ann will anyway not take umbrage (lines 1214) if it is properly done, with some normative restitution consisting of a thank you and an account (lines 1820). So the various mental state terms in extract (7) are made subject to public ratication, as being reasonably

Discursive psychology and mental states

255

grounded or appropriate. Note Ms various interventions throughout the extract, corroborating Ls thoughts and assessments, and in lines 1820 providing an account, when prompted by L, of how and why Anns umbrage might be avoided. Extract (7) displays a rich interplay of mental states, normative actions and factual descriptions. Although we have provided no systematic analysis of these uses of think and thought here, for which a range of cases would be required, the analysis illustrates DPs central concern with how psychological avowals, attributions and implications work in conjunction with factual descriptions and normative accountability (Edwards and Potter, 1992a). False ideas, ones that do not correspond with a given reality, along with non-normative actions, are treated as requiring accounts. Relationships between psychological states and the nature of the world to which those states pertain, feature in discourse as part of the normative order of social life and conversational interaction, as central elements in what talk does. Discussion We can distinguish three theoretical versions (there may be others) of what is going on when someone says how they feel, what they think, want or believe. 1. Utterances of that kind might be taken as reports from within, expressions of mental states. This is what Harris (1981) has called the telementation myth, that thoughts and feelings occur in our minds, and (then) we express them in our words. It is the starting point of a wide range of philosophy, psychology, clinical assessment, interviewing and general research methodology. Within this perspective, talk need not be a straightforward reection of mind. People can lie, be wrong about themselves, strategic, or inarticulate. So caution is always required when nding mind behind words, and sometimes special procedures are deployed to reduce unreliability. But according to this approach, the expression and communication of thoughts is basically what talk does. 2. A contrasting approach argues that descriptions of mental states are best analyzed in the manner recommended by linguistic philosophers such as Ryle and Wittgenstein. Here, the notion that words such as think or feel express or refer to mental states is rejected in favour of an analysis of how those words are properly used, in publicly ratiable ways. Anyone, whether philosopher or lay person, who claims to be using such words to refer to an essentially private mental experience is making a conceptual error. This approach often takes the form of

256

Conversation and cognition

analyzing assertions or propositions, often single sentences, that people might be imagined to make, such as I feel unhappy or I believe in ghosts. Empirical study of recordings of conversation plays a minor role, if any. 3. While acknowledging the arguments in extract (2), a third approach recognizes that there is nevertheless some substance to the idea of referring to private mental states, though not as the analysts favoured theory of language and mind. The very idea of reference to mental states is at least conceivable, if only for it to be attributed by their critics to cognitivists and misguided lay persons. In this third approach, the status of reference to internal mental states is not something to be refuted, even though it is conceptually refutable, but rather, studied as a practice within public forms of life. People may sometimes talk as if, or on the proposed and oriented-to basis, that their words are expressing inner thoughts and feelings. It is a basis and orientation found in clinical psychology for example, if we approach clinical attributions as studiable practices rather than rival theories. But it is found also in everyday talk, where again it is available for study as a social practice. It is also something that can be countered, not only by philosophical argument, but as part of everyday practices. People may point to inconsistencies, or to evidence to the contrary, or to disagreements, or to the speakers proposed strategic aims, and thus challenge any claim that they are merely reporting what they think. DP rejects position (1), in the main aligns theoretically with position (2), but mostly develops position (3), by approaching mentalistic claims as performative rather than merely erroneous (cf. the invocation of indifference in Lynch and Bogen, this volume). The notion that mental states are private and reportable emerges as part of, rather than merely refuted by, their public sense and ratication. There are, therefore, two different kinds of relationship between DP and cognitive psychology. One relationship is that of opposition and critique, in which DP is a rival way of understanding what is going on when people talk, or remember, or attribute causes, or express their attitudes and feelings, or whatever; it is a rival approach to discourse (Edwards, 1997 is an example of this style of DP). The second relationship is that of studying cognitive and social cognitive psychology as a form of life like any other, studying the set of scientic concepts and practices through which the nature of the cognitive is produced (examples of such work include: Antaki et al., 2000; Edwards and Potter, 1992b; Puchta and Potter, 2002). Neither approach entails any kind of endorsement of cognitivism. Critical reactions to DP range from defences of traditional cognitive psychology (e.g., Conway, 1992), to the kind of politically grounded

Discursive psychology and mental states

257

critique (e.g., Parker and Burman, 1993) that is commonly also directed at CA. Rather than engaging with those alternatives here, we focus instead on a discussion by Coulter (1999) that deals critically with DP more directly in its own terms and has more direct ramications for the issues raised in this volume. In distinguishing his own praxiological approach from DP, Coulter characterizes DP in ways that align it with position (1) above. This characterization is wrong. However, rather than dwell on that here (see Potter and Edwards, 2003 for an elaborate rebuttal), our concern is with the different practices of DP and Coulters approach, as well as the different materials that are the focus of study. We shall explore that a little bit here, given its central relevance to DP as a project and given the signicance of Coulters own contribution to the development of a non-cognitivist approach to mind, which we have acknowledged elsewhere, and do again here. We hope that our discussion will clarify some quite subtle issues. Coulter mistakenly interprets Edwards (1997) as studying the mental . . . solely in terms of what people say about it (1999: 166). This is in turn conated with members lay . . . theorizing (ibid.). We have already noted that DP is not solely concerned with what people say about the mental; but nor does the latter amount to lay theorizing, and this is a signicant point of contrast worth highlighting. Coulter goes on (ibid.: 1667) to discuss the status of someone asserting5 I think in my head. Even when DP is concerned with overt uses of psychological concepts (DP type 2), it is not concerned with members disembedded quasitheoretical assertions. In fact, it is not clear how, when, and doing what, people might make such assertions. Our reaction as analysts of discourse to such an example is to ask what anyone might be doing by uttering it somewhere. Rather than being concerned that such a person has a mistaken understanding of themselves, or of the concepts that they use, we would investigate what kinds of work such utterances might do, how and when they may occur, what they are used in contrast with, indeed what makes them locally intelligible things to say. This contrasts with Coulters use of such an item as a kind of hypothetical interview response, a proposition to refute. We do not have to hand a case of somebody asserting I think in my head, but extract (8) includes a situated use of a similar expression (lines 67). It comes from a counselling session with a husband and wife, Jeff and Mary. Extract (8)
1 2 3 J: M:

[DE-JF:C1:S1:p.5]

And I TRI:ed, befor yo- I dunno a couple o ) and all] of a= days it might work [( [ ((coughs)) ]

258 4 5 6 7 8 J:

Conversation and cognition sudden Id- (0.9) youknow Id (.) Id go to wor:k or: Id be on my ow:n, (0.4) somewhere an:d (0.4) all these thoughts would ll my hea:d and the anger would build up< and next >time I see Mary Id have a go at her you see,

Jeff is responding to a complaint by Mary, that he is spoiling her efforts to get closer to him, following an extramarital affair she had, because of his continuing bad feelings about it.6 Jeff tells how he does make genuine efforts (lines 12) but they are thwarted by a kind of process he undergoes, that he cannot help, that happens to him against his best efforts, all of a sudden (lines 24). His expression all these thoughts would ll my head serves as a kind of observational report on things happening within him, where a notion of passively and reluctantly undergoing mental experiences is something other than performing actions for which he might be blamed. Note how it is nicely situated in contrast to his concerted efforts (I TRI:ed, line 1). Jeff is using a familiar, common sense mind-as-container metaphor (cf. Lakoff, 1987). What it does here is enable him to say and feel bad things about Mary, to avow and express anger and resentment, and not be very receptive to her overtures, without those being actions that he culpably does. So reporting on in the head experiences is a culturally recognizable practice in talk, that has its uses. Of course, analyzing it in this way, as an intelligible practice of accountability, implies no ontological endorsement on our part of what Jeff claims about events in his head, any more than if we were analyzing the uses of Azande oracles (Evans-Pritchard, 1937). Other issues with Coulter are discussed in Potter and Edwards (2003). In this chapter have used his critique as an opportunity to clarify the discursive psychological position on mental states, and to correct some misleading claims before they become too established. More generally, in this chapter we have tried to sketch out the contours of an approach to psychology, mind, cognition, psychological states and so on which is both non-cognitivist and empirical. We have illustrated this approach with a set of examples. They point the way to the kind of analysis that would appear in full DP studies, but should be treated as no more than illustrations so far. Up to now research in discursive psychology has developed three themes: (1) respecication of psychological topics; (2) studies of the practical use of the psychological thesaurus; (3) studies of psychological orientations. Cognitive psychology and cognitive science are huge elds employing hundreds of researchers, working in close relation to applied disciplines such as human computer interaction, ergonomics and cognitive neuroscience. The implications of a

Discursive psychology and mental states

259

respecication of cognition along discursive psychological lines would be immense.

1. Connie described Jimmy a couple of minutes earlier as extremely jealous. Extremely jealous person. Has always been, from the day we met . . . it was totally ridiculous the way he goes on, through this problem that he has. Extract (1) is part of Jimmys extended response to that and other things. See also Edwards (1995, 1997). 2. Our scare quotes round concept are to distance us from the common cognitive interpretation of concepts, which treats them, roughly speaking, as mental meaning packets associated with words. In a sense our enterprise will, and should, generate major problems for a traditional understanding of what a concept is. 3. Notably shorter is reected in its being well inside Jeffersons (1989) normatively standard maximum silence of one second, whereas the pauses in lines 4 and 6 are well beyond that. 4. The minimal nature of that acknowledgement, as in extract (2), serves as a receipt of what W is saying that stops short of what might, for a police ofcer, be a problematic endorsement of it. 5. The use of this speech act category asserting is itself interesting, and rather typical of philosophical analysis, where people are imagined to assert things, uttering denitive theoretical statements, and so on. Ironically, it is the very stuff of cognitivism to deal with objects of that kind, which readily lend themselves to a notion of people saying what they think; it is a way in which psychological, philosophical and linguistic treatments of language often get off on the wrong foot right from the start. 6. Marys complaint included: I feel every time I try to get closer to him hes hurting me, and then it makes me distant again.