Nov 17, 1989 - organisation of talk about 'delicate' issues .... 3 P: em I've had other females I haven't had (0.5) a guy for. 4 ..... not saying (lets) assume you?
David Silverman and Anssi Perakyla AIDS counselling: the interactional organisation of talk about 'delicate' issues
Abstract Although the AIDS pandemic has generated considerable social science research, the focus has almost entirely been on epidemiology and on survey research studies of health knowledge and behaviour. In contrast, this paper offers an early report on ongoing work into HIV and AIDS counselling as it occurs in practice in a number of English clinics. An analysis of transcripts of such naturally occurring encounters reveals how professionals and clients organise their talk in relation to the 'delicate' issues to be discussed. Particular attention is paid to the phenomenon of delay in the production of 'delicate' items and to the social organisation of the description of such items. Drawing, in part, on methods from conversation analysis, it is shown how a detailed analysis of how talk is produced in naturally occurring settings can generatefindingsrelevant to professional practice.
Introduction
Although lots of people figure that expedence is a great thing . . . [experiences] are extraordinarily carefully regulated sorts of things. The occasions of entitlement to have them are carefully regulated and then the experience you are entitled to have on an occasion you are entitled to have one is further carefully regulated (Sacks 1984:428) Sex and death are culturally constituted as central human experiences. Yet, despite Sacks' observations, the descriptive apparatus through which we understand them has been more the topic of fiction writers than of scientific work (for recent exceptions see Foucault (1979) and Prior (1989)). Perhaps the Romantic impulse which threatens to engulf certain parts of the human sciences (see Silverman 1989, Strong and Dingwall 1989) makes us reluctant to treat them as simply language-games among others. Yet the legal requirement that HIV testing should be preceded by Sociology of Health & Illness Vol. 12 No. 3 1990 ISSN 0141-9889
294 David Silverman and Anssi Perakyla
counselling guarantees, tragically, that 'sex' and the prospect of death should become topicalised in large numbers of professional-client exchanges which, if the patient is seropositive, are likely to be lifelong. A social force, even prior to HIV, counselling now offers more people than ever the opportunity to 'learn what we are like, what our experience is, how things are with us' (Taylor 1986:78). It is now a commonplace that the HIV pandemic will not be resisted simply by giving people better information. For, as all studies of communication have shown, many other conditions have to be satisfied before knowledge leads to behaviour change (see Nelkin 1987, Aggleton 1989). In the context of HIV, we may assume that part of the difficulty resides in broaching the topic of safer sex with partners (for interviews on this topic with drug-users see McKeganey 1989). Yet we lack knowledge of how people talk about their sexual behaviour in naturally occurdng situations. In this context, Coxon's conclusion to a recent article on what he calls 'sexual research' is remarkably optimistic: The enormous growth in studies of sexual behaviour at an international and national level under the impact of Aids has . . . had the unanticipated and welcome result that scientific, intellectual and practical understanding of sexuality has now been immeasurably improved (Coxon 1989:24) Certainly, we have epidemiological work, together with the forthcoming British survey of people's sexual behaviour (thanks to a private charity). There is also Coxon's own pioneering study of sexual vocabularies, obtained by asking a cohort of gay men to keep sexual diaries. But where are the studies of such vocabularies as they are deployed in naturallyoccurring situations? It is difficult to argue that keeping a diary on one's sexual behaviour is a routine activity for many people. To the objection that such studies are impossible without prying or engaging in other improper behaviour (perhaps placing a tape-recorder under a bed), we can answer: why accept the essentialist assumption that sexuality only occurs between sexual partners? Increasingly, ethnographers are resisting the naive suggestion that phenomena like 'the family' or 'science' are only (or most 'authentically') constituted in single sites, such as households or laboratories (Silverman 1989a). Why should not the same be true about 'sexuality' or, indeed, 'death'?'
Methods
This paper is based on an analysis of over 100 tape-recorded extracts from counselling sessions with patients coming for an HIV test and patients diagnosed as seropositive. More than half of these sessions took place at the District AIDS Unit and Haemophilia Centre at the Royal Free
AIDS counselling 295 Hospital, London. The remainder occurred in three other hospital-based centres in England. All the matedal was obtained on the basis of informed consent on the understanding that the identity of patients would never be revealed. The studies from which this paper is drawn seek to combine a detailed analysis of the 'micro' particulars of forms of talk in counselling interviews with an examination of the organisational, cultural, political and economic constraints within which HIV Counsellors work. The latter more 'macro' concerns have been discussed, in a preliminary way, elsewhere (Chester 1987, Silverman 1990). In this paper, we draw mainly upon conversation analytic methods rather than conventional ethnography. An appendix gives the transcription symbols used. In the transcripts, C refers to counsellors and P to patients.
The sequential organisation of talk about sexuality in counselling
As a first observation, we note that patients typically pause before first mentioning terms relating to sexual intercourse or contraceptives. This seems to apply irrespective of gender. For instance: Extract A 1 C: 2 P: 3 C: 4 P: 5 C: 6 P: 7 C: 8 P:
um what are you actually doing? we're just using er condoms (0.3) [umm [for safe sex umm um we (0.5) intend to use spermicide umm er but we don't intend to use (0.3) the cap (Male P)
Extract B 1 C: right (1.0) what do you know about (0.5) catching AIDS or 2 how it can be how you can catch it [it's it er 3 P: [all I know is it is 4 transmitted em (1.0) through er sex (Female P) In the whole corpus of transcripts, we have been able to identify only two deviant cases where patients fail to mark their^r.9? use of such an item by a pause or other marker (both are men using the term 'condom'). Frequently, although not with such regularity, counsellors pause before delivering such terms. For example: Extract C 1 C: there are many couples who don't (1.0) have intercourse 2 every time they so called make love
296 David Silverman and Anssi Perakyla We will return later to possible cultural and sequential explanations of the presence or absence of such pauses. However, at this stage, it will be useful to introduce some definitions relevant to these phenomena. We will refer to pauses as part of a class of phenomena that we will call perturbations. Perturbations may mark items that members perceive to be delicate. Of course, they also do a range of other work. For instance, topic or speaker change is often associated with a pause. Where the only sequential or interactional explanation for a pause is the delicacy of the item that follows it, we will refer to it as a pre-delicate perturbation. Let us try to flesh out these concepts by further observations. Firstly, delicate items appear to extend beyond sexual intercourse or contraception. Notice how, in Extract B above, the counsellor pauses before saying 'catching AIDS'. Of course, you can't 'catch AIDS'. As we will see in Extract K below, the counsellor is mirroring the patient's term and her perturbations around it. Similar perturbations before the first use of this term is found with patients as well. For instance: Extract D (PF = patient's fiancee) 1 PF: we know that er HIV virus can be passed through (0.5) the 2 semen In Extract D, we get the usual pause before the production of the sexually delicate item ('semen'). More importantly, as in Extract B, we find a perturbation before the disease is mentioned ('er HIV virus'). Certain ways of naming sexual preference are also associated with delays or other perturbations. Consider the following extract from another Centre: Extract E (transcription simplified) (0.4) can I just ask you briefly (0.2) erm: one or two questions 3 C: before we start, hh have you ever had a test before? 4 no 5 P: no hhh have you ever injected drugs? 6 C: no 7 P: (2.0) have you ever had a homosexual relationship? 8 C: (0.5) no (0.5) and that's not really (0.5) (I mean) (0.2) put 9 P: me in a high risk group now [has it? 10 [no 11 C: We will return to this very rich extract later to take up the implications of the counsellor answering the question about risks (9-11) and using the term 'homosexual'. Staying, for the moment, with the issue of delays in the production of delicate items, we observe a clear contrast between the production of the first two question-answer sequences (lines 4-7) and the last (lines 8-9). Notice how the counsellor extends her inbreaths marking a change of topic before her first two questions to a full two seconds before
AIDS counselling 297 her last question. Again, the patient delivers his first two negative answers without any pause at all but pauses for half a second before producing his final 'no'. Both seem to be marking the issue of 'a homosexual relationship' as delicate. We might speculate that such delicacy may reside in cultural norms applied by a dominant culture when naming, in its own terms, a dispreferred sexual orientation. However, it would be wrong to assume that members are mere automatons responding to such external forces. Look, for instance, at the following exchange: Extract F 1 C: 2 3 P: 4
are there any other partners you've had in the past any any males any females? em I've had other females I haven't had (0.5) a guy for about two years.
The patient here has already revealed (turbulently) that she has a sexual relationship with another woman. Notice the counsellor's hesitation before she asks about male partners- but not about females. Although the patient hesitates before she replies, she too produces a pre-delicate turbulence before 'guy' but not before 'females'. It is possible that the patient is finely tuning heF turbulence to that of the counsellor (we will later call this phenomenon 'mirroring"). In any event, it is clear that we are dealing with a locally organised phenomenon. So, while there clearly may be cultural norms which identify delicate items, the identification and treatment of delicacy is clearly locally managed. We have stressed that turbulence is often present pdor to the first production of a delicate item. However, the rule appears to be: on repeating a delicate term, do not repeat the turbulence. For instance: Extract G When P/zr^f says 'sexual intercourse' it is preceded by 'er you know'; when he repeats it there is no further hesitation. When C first says 'intercourse' it is preceded by a 1.0 pause; when she repeats it four lines later there is no pause. Extract H P pauses before his first use of 'condoms' but his wife does not pause when she uses the term afterwards. We can only speculate about why the parties fail to repeat their predelicate work. Perhaps once you have exhibited the intersubjective issues involved in producing delicate items, further perturbations when they are repeated might indicate a subjective problem rather than a deference to cultural norms. Extract H is intriguing because it suggests that in Goffman's (1956)
298 David Silverman and Anssi Perakyla sense, patients and their partners (or families) may constitute a 'team' where each member is implicatedln the moves of the other. So, as here, where one team-member does the pre-delicate work before an item, there is no need for another member to repeat that work." We have already seen, in Extracts F and G, patients produce 'er's and 'um's before delicate items. It is clear, therefore, that the term 'perturbation' refers to more than mere pauses. We will identify a perturbation by the presence of any disturbed speech in a turn at talk, where the disturbance may be a pause, a hesitation, a change of intonation, a 'repair' or body movement. Extracts I and J below are both transcribed from videotapes and show the presence of body movements prior to a patient producing some delicate items: Extract I 1 P: 2 3 C: 4 P: 5 6 C: 7 P: 8 9 10 C: 11 P: 12 C: 1 2 3 4 5 6 7
we know that er HIV virus can be passed through (0.5) the semen mm ((hand movements begin)) and that's er you know sexual intercourse is a part [of marriage [mm and if you've not had it before marriage it will be something that will be ((end)) you know quite exciting and ((begin)) and new [and [mm we'll be breaking new ground heh heh so to speak ((end)) okay
Extract J C: is is there anything you're particularly worried about (0.5) that you think you might have been risking is some sort of way? P: ((rubbing face 2.0)) em (1.0) I dun heh that's a hard question em (4.0) it's hard to explain if you don't really know er ((small hand movement)) ((lowered voice)) what we do (1.0) er
'Repairs' occur when speakers correct their utterances. We find two examples of such repairs in Extract I line 4 and Extract J line 4. Once again, we would stress that we are searching for local explanations of the positioning of all these perturbations. This means rejecting reductionist explanations either in terms of 'culture' or 'identity' (for instance, any assumption that a particular body movement indicates a particular 'state of mind'). A consequence of these perturbations can sometimes be a prolonged delay in the production of a delicate item. Such an item may only emerge
AIDS counselling 299 after the counsellor makes several attempts to get the patient to specify her concerns: Extract K 1 C: 2 3 4 P: 5 6 7 8 P: 9 10 P: 11 12 13 C: 14
15 16 17 18 19 20 21 22 23 24
um (0.7) perhaps we could start by you telling me: (0.5) why yer wanted ter: (0.7) to be seen: (0.7) well erm (1.3) an ex girlfriend of mine: (.) I found out (0.5) about a week ago (now) (.) hh that erm she's a prostitute when I was with her: (1.0) an: (0.5) like (1.0) I dunno I was never worded about it before (0.7) but like (us:) knowing that (jus:) (started me) got me worded (see) ( ) right what in particular are you worried about^? (2.3)
P: P: C: P:
I don't know (0.7) that that's hard to say I jus:t(.) (eh) (0.9) of catching something really (1.0) I mean (1.0) I (.) whatever ( ) is there anything in particular: (0.3) that you're concerned about catching? (1.4) well I definitely don't wanna wanna get ca:: (.) catch AIDS
Extract L 1 C: what are the things you would like (.) to raise (.) today 2
(0.6)
3 PF: 4 5 PF:
er (0.3) um:: what (1.8) what the::
6
20 PF:
well (things) er::: on the sexual side
In Extract L, it takes 15 lines and two further questions for the patient's fiancee to produce the delicate item ('the sexual side'). Even then, it is immediately preceded by an extended hesitation. In Extract K, we find an even longer delay (18 lines of transcript) and two further questions, before the patient (turbulently) produces 'AIDS'. Note how she uses beforehand a series of glosses: 'it', 'that', 'something' and then only delivers 'catch AIDS' after a 1.4 pause (line 21) and two repairs."*
3(X) David Silverman and Anssi Perakyla
Delay of the delicate item in "dreaded issues" talk
A delay of the delivery of delicate objects also seems to be a characteristic feature when the participants are talking about serious things possibly happening to the patient in the future. Obviously one of these things if you are HIV-positive is becoming sedously ill and/or dying. Following Bor and Miller (1988) we will use the gloss 'dreaded issues' about such topics in counselling. Much of the talk about 'dreaded issues' in our corpus takes place within an Information Delivery footing, where the participants are stably aligned as speaker and recipient. The counsellor tells the patient about HIV-positive persons' possibilities of getting ill and dying and about typical psycho-social reactions in such circumstances. In the counsellor's talk, we observe recurrently the same features as in talking about sex: the delivery of the delicate items is delayed through the production of pre-delicate perturbations: Extract M 1 2 C: hhherm:(.)( 3 that w-w-we don't know what percentage of 4 people- (0.4) hhh probably more than fifty: 5 percent of people go alo- (.) go on to 6 develop: 7 P: unr 8 C: er::: at least some sort of symptoms and 9 (0.2) hhh ((with a slightly lowered 10 voice:)) some goon to develop 11 a:ids [ and: subsequently die from = 12 P: [um: 13 C: = that (0.2) hhh erm::
) you're right
In M, the delivery of 'develop a:ids' and 'die from that' is delayed through hesitations ('w-w-we'), repairs ('go alo- (.) go on to') and pauses ('symptoms and (0.2) hhhh some go on to develop a:ids'). More interesting are, however, the cases of 'dreaded issues' talk where counsellor and patient are aligned as Ouestioner and Answerer. In this Interview-footing we can examine how the participants locally orient to and make use of the publicly available ways of displaying the delicacy of the topic. In these cases there are basically two types of entrances into 'dreaded issues' talk. In one of them, the counsellor asks a closed question which already implies the delicate issue, and in the other, the counsellor makes an open ended question and the patient introduces the delicate issue as a response to that. Extracts N and O are instances of the first approach.
AIDS counselling 301 Extract N (BF = Patient's boy-fdend) 1 C: Have you just had the test (.) [ Gary 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
BF:
P: C: C:
P: C:
P:
I Yes I
have *Umh* (0.4) A:nd (1.8) from what you: kno:w of Ga:ry I mean(2.0) if it was to be positive (.) what do you think his main concern would be? (1.6) [ umh [ or how do you think- (0.4) how would you see him coping? (2.8) I (don't) think he coped_well: because we have discussed it ((continues))
Extract O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Dr:
P: C:
C:
C: C: C:
( . . . ) but it holds back the: (0.5) multiplication time dght (1.4) If (0.6) and we're just ta:lking very hypothetically (1.2) if you sh- (0.6) yo::ur (.) em (.) T-cells did drop an your immu:ne system (0.7) began not to work so well: an:d (1.5) you became unwell, (.) how do you see (2.0) Doreen as coping? (1.2)
P:
Erm
P:
(1.1) She'll r:- (0.4) she'll respond to the
The counsellors' ways of designing their questions here bear a strong resemblance to the way that utterances referdng to sex were designed in previous excerpts. Here also we can observe hesitations and pauses before
302 David Silverman and Anssi Perakyla the delicate items are delivered. Moreover, the patients, while initiating their answers, also follow the same pattern, through delaying the beginning of their answers. In Excerpt N, the first delay (line 12) is followed by the counsellor rephrasing her question (lines 14-15). A new gap emerges (line 16) before the patient initiates his answer (line 17). In Excerpt O, the counsellor's question is followed by a gap (line 16), an 'erm' by the patient (line 17), another gap (line 18), and an aborted initiation of the answer (line 19), and finally a restarted answer. Even though patients' answers do not lead to additional delicate material, their design can be seen as attending to the delicate nature of the topic initially displayed by the counsellor. Our suggestion is that the delay of initiating an anwer by the patient claims an agreement with the counsellor on the delicate nature of the topic, and a willingness to deal with it as such. An open-ended question and its consequences are displayed in Excerpt O: Extract Q• (W = Patient's wife)
1 C: 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
P: C: W: C: W: C: W: W: W:
W: C: W: C:
Can I just ask you what are your greatest conce:rns:: (.) Liz. [Liza [Liza: Ica:n'tgetit[( ) [ ((coughing)) Liza about- hhh (0.4) at this mo:ment in ti:me. (.) can you s:ay alou:d. (3.0) Erm:: (.) the uncertainty[:? [mmh: (1.5) obviously:?
(0.6)
r'
an::d (3.0) trying to get John to cope with it (0.2) an:d-(0.3) lead as normal a life as possible? (.) I'd (.) I don't see hhhh (1.0) I don't really see any f::easible r:ealistic alternaitive.= = mm:h (0.5) than (.) (both) to carry on:: (0.3) as (.) as no:r[mal. [mmh (1.6)
AIDS counselling 303 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
Baby: gJMU (0.7)
W: W: C: W: W: C: W:
an::d (1.6) what would happen to me:? mmh and the children (2.1) if he did devel[op [mmh something? (0.2)
C:
mmh
C:
(0.6) What's your greatest fe^ar about that?
As a response to the counsellor's enquiry about 'greatest concerns' the wife of a HIV-positive patient produces several issues. First, there is a general gloss 'uncertainty' (line 9), then considerations about helping her husband to lead as normal a life as possible (line 16-17) followed by an assessment of the situation (18-25). Only thereafter does she introduce the menacing future possibilities: 'what would happen to me:? ( . . . ) and the children ( . . . ) if he did develop ( . . . ) something'. This issue is then, after a short delay, topicalised by the counsellor for further development (line 42). Our argument is that both the wife and the counsellor here make use of the rule of delay. The wife doesn't produce the delicate concern immediately as a response to the counsellor's question. Neither cioes the counsellor topicalise the first offers, but waits instead until the cielicate item has emerged. The counsellor's policy of waiting for delicate issues to appear later is obvious, if we examine the possible slots in the wife's turn where she could have got the floor back. In her long utterance, there are several possible turn completion points (cf. Sacks, Schegloff and Jefferson 1974), where the counsellor - according to the commonsensical rules of turn-taking in conversation - could have initiated her utterance. She passes them all, either by producing a 'continuer' (cf. Schegloff 1981) (lines 10, 22, 26, 33) or by just allowing a gap to emerge (line 13) until the patient self-selects as the next speaker. In other words, the counsellor withholds her response until the wife finally produces the delicate item. Consonant with this, the wife produces delicate material after the counsellor has withheld an initial response. Our argument is that the delayed delivery of the wife's delicate concerns is not (merely) a result of psychological inhibitions related to talk about such things. On the contrary, we argue that both speakers orient to and make use of the socially and culturally prescribed etiquette of approaching
304 David Silverman and Anssi Perakyla a delicate issue. By producing less delicate material first, the wife leaves it up to the counsellor to decide whether she wants to treat this as a preliminary for the delicate issue, or as an issue per se. Through withholding her response, the counsellor chooses the former option - and as a consequence of that, the patient finally produces the delicate issue, which is in turn topicalised by the counsellor.
The descriptive organisation of 'delicate' talk
So far we have found striking resemblances in how talk about sex and 'dreaded issues' is constructed in these settings. The delicacy of these topics is finely attended to by the devices through which patients and counsellors alike delay their delivery of them. But why do we find such disturbed talk around these topics? One immediate temptation is to look for psychological explanations of this turbulence, appealing to notions which link hesitation to the experience of 'embarrassment'. However, as Sacks reminds us, we would then still need to examine the interactional organisation of that experience - an organisation which can apparently allow, even encourage, professionals to delay delicate items just as much as do their clients. Moreover, because that organisation is demonstrably local and sequential, purely cultural explanations of these turbulences are insufficient. Closer attention is required to the descriptive organisation of this talk. We might start by examining the implications of using and excluding certain items to describe people and their activities. Fortunately, we have a useful explanatory tool to hand: Sacks' (1972) account of the nature of everyday description. This can be introduced by referring to two extracts discussed above. In Extract E, we have so far looked at the turbulences associated with the item 'homosexual'. However, we have not yet asked why this item should produce turbulences. Following Sacks, we can make the following observations: 1 obviously, unlike a foreign film, in everyday life, people and their activities don't carry sub-titles 2 so any number of descriptive labels may adequately describe a person or an activity 3 however, each label or membership categorisation device carries a moral baggage e.g. the different implications in our society of asking someone if they've had a homosexual relationship versus if they have any girlfriends or boyfriends 4 moreover, using a label about an activity in which you are involved implies that you fit into the collection of people who engage in that activity - in this case, that you are a homosexual 5 note, further, that this particular label is resisted by those people to
AIDS counselling 305 whom it is applied: homosexual is now a term used by many medical professionals and the homophobic press but 'gay' is the preferred term of the community concerned. So, by using the term herself, before asking the patient to describe his own activities, the counsellor quite unintentionally may be creating difficulties for the patient whether or not he turns out to be gay. Precisely the same issues of description arise for the patient. We can see this in Extract E which continues: 8 9 10 11 12 13 14 15 16 17 18
C: P: C: P: C: P:
(2.0) have you ever had a homosexual relationship? (0.5) no (0.5) and that's not really (0.5) (I mean) (0.2) put me in a high risk group now [ has it? [ no no it doesn't umm er why did you want to have the test done? oh well (1.0) err I've just moved in with my current (0.3) gidfdend hmm hh and err (0.5) before we actually start (2.0) full sexual relationship we thought it was best if we both (1.0) screened.
The pause on lines 13-14 between 'current' and 'girlfriend' can be noted. Once again, 'girlfriend' is a membership categorisation device. The delicate issue here is that once you so categorise someone with whom you are involved, a label from the same categorisation can be applied to you yourself. Notice how the term 'girlfriend' is a neat half-way house between, say, partner or fiancee (which implies a permanent relationship) and lover (which may imply that you are promiscuous). Descriptive issues arise in Extract K above. Once again, in lines 4-6, we find a turbulent production of descriptive items: in this case 'ex-girlfriend' and 'prostitute'. Given that an item from these collections can legitimately now be used about the speaker (e.g. woman who has a girlfriend, woman who befriends prostitutes), the delicacy of the descriptive activity is revealed. One final example will underline the relevance of Sacks' account of description for an explanation of delicacy: Extract R 4 C: 5 6 P: 7 C: 8 P: 9 10 11
so what made you think that you need to have an HIV test just[( ) [er:::hh [( ) [just well I spent the last year (.) in A (.) and (1.0) had slept with a (1.0) with a lady down there (1.0) and although I had taken precautions it was only once (0.5) and I just think (0.5) it was likely ( ) heh (
306 David Silverman and Anssi Perakyla 12 13 C:
) when you say that you took precautions (.) what doyou
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
and were we talking about a girl that you actually knew (.) or (.) girl or prostitute or [what are we talking about? [er:: (2.5) well she hh (3.0) was a dancer at a nightclub so that sort of puts her in er (0.3) she's not (0.3) not really a [softly] prostitute but (0.5) they try (0.3) I think I realised later that tbey try and (0.3) pick up European (0.5) men hoping that (.) well some of them do end up marrying them and going (.) to Europe (0.5) so they're not (1.0) not completely prostitutes but you know heh (1.0) it means that tbey probably bave a little a few more partners tban you you normally would do if you er (1.0)
C:
P: P:
C:
Note bow the patient defines bis partner in relation to tbe counsellor's question at lines 4-5. As in the previous extract, note in lines 8-9 the patient's delay in producing a category for his partner: two 1.0 second pauses and a repetition of 'witb a'. Note tbe nice solution that the membership categorisation device 'lady' provides to problems of delicacy: it is even vaguer tban 'girlfriend' and so carries few problematic implications. Indeed, it may even create favourable implications: tbe partner of a lady is a gentleman (using Saeks' term: lady and gentleman is a standardised relational pair). However, in the second extract from this consultation (beginning line 25), the counsellor berself offers a category witb altogether different implications: line 34 'prostitute'. Note: 1 Tbe standardised relational pair of prostitute is client. 2 Tbe delicate implications of this are attended to by the counsellor in the long delay before the term is produced. Note at lines 33-34: tbe initial use of the term 'girl', then a micro-pause, tben the failure to put the delicate item after the 'or' (line 34) but the further micro-pause and the repetition of 'girl' before finally the counsellor produces the delicate item: 'prostitute' 3 Note further how in lines 37-39 the patient delays saying 'prostitute' via pauses, hesitations, and initially trying another membership categorisation device 'dancer at a nightclub'. Notice also at the end of line 38, the downgrade on the label and the dropped voice before the label is actually used. 4 Again, in lines 39-43 he attempts to downgrade the saliency of this label by introducing the activity of marriage (line 42). Marriage is what migbt
AIDS counselling 307 be termed a category-bound activity i.e. it produces a standardised relationship pair (SRP) of husband and wife. This SRP is incompatible witb the SRP of prostitute and client. Hence this supports tbe further downgrade of the latter collection at lines 42-43 complete with perturbations:'so they're not (l.O) not completely prostitutes'. This is followed by the patient's laugbter - offering tbe counsellor an option about bow seriously she is to take his account. 5 Notice how tbe counsellor eo-operates in this down-grading of the SRP prostitute and elient at lines 45-46. She does this by contrasting the activities of 'prostitutes' and himself by producing the equation: Prostitutes = many partners; You = fewer partners These observations show that tbe apparatus of description itself embeds members in delicate matters. Using Sacks' concepts, we have demonstrated the delicacy in using membership categorisation devices. Tbese locate tbose so described in standardized relational pairs and category-bound activities witb highly delicate implications. Moreover, tbe same process also works in reverse: by naming a particular activity ('sex' in B, 'make love' in C, 'sexual intercourse' in G and I, using a condom in A and H, and 'catching AIDS' in K), one implies a possibly delicate description of the kinds of parties who migbt engage in an activity of that kind. In a slightly modified form, tbe argument based on tbe concept of membership categorisation devices relates to tbe displays of delicacy in 'dreaded issues' talk. The core of the delicacy here lies not, however, in the relations implied by the categorisation, but rather in tbe activity of categorisation as such. 'Dreaded issues' talk trades off tbe hypothetical future possibilities, tbe serious things tbat may happen to tbe patient in the future. This implies a re-categorisation of the patient in the conversations: instead of maintaining the categorisation as 'well' or at least 'alive', in the 'dreaded issues' talk tbe patient becomes re-categorised as 'unwell' or even 'dead'. Looking at the matter in more detail, the patients in tbe previous excerpts were re-categorised as 'developing Aids and dying' in M, 'being positive' in N, 'becoming unwell' in O, and 'developing something' in O. This re-categorisation is a highly delicate matter, especially wben the patient's current physical status gives no overt warrant for it. We must stress tbat what we are seeing here should not be seen as difficulties in communication. Conversely, this represents elegant interactional work in which both parties subtly attend to the local consequences of the labels they use to describe people and, by implication, themselves. In so doing, tbey reveal tbat massive cultural apparatuses are nonetheless, inevitably and recurrently, a matter for local production and, indeed, play (for a similar observation, emerging from a different analytical approach, see Bourdieu (1977) on bow gift-exchange is a matter of local play).
308 David Silverman and Anssi Perakyla Professional dilemmas in managing delicacy
Elegant interactional work in relation to delicate items can nonetheless create dilemmas for professional practice. For instance, we remarked earlier that the counsellor in Extract E nicely marks the delicacy of asking the patient about 'a homosexual relationship'. However, given the baggage that this descriptive category carries (its associated SRPs and the categorybound activities it implies), we suggest that counsellors may need to think about the consequences of using such a term. We now turn to further consideration of the management of delicacy in the context of professional practice. In our extracts, we can find other examples of counsellors marking the delicacy of MCDs they have introduced. For instance. Extract R above offers, in common with Extract E, an example of a professional engaged in the turbulent production of a delicate item (in this case 'prostitute'). This does not exhaust the repertoire of tactics available for the local management of delicacy. For instance, in E, note how the counsellor asks the patient's permission to ask her subsequent questions. Schegloff (1980) has described this practice as a question projection. One function of such a projection is to mark a forthcoming delicate item. We have several other examples of question projection but we will include only the most elaborate one: Extract S 1 C: 2 3 4 5 6 7 P:
I have to ask you the couple of questions like the you know are a bit a bit cheeky if you like heh but I ask everybody because obviously when you're talking about HIV infection and AIDS people who use drugs people who are homosexual or who who've had homosexual contacts um are more at risk perhaps than other people hmm right
This is a nice example because it shows the working in unison of a question projection with considerable perturbations prior to the delivery of a delicate item ('cheeky') immediately followed by a laugh which shows the patient that the counsellor is aware of the incongruity of a counsellor asking 'cheeky' questions. As regards 'dreaded issues' talk, the delicacy can be managed by downplaying the relevance of the re-categorisation in relation to this particular patient: Extract T 1 C: hhh (.) we have to make sure you 2 understand (.) what happens when someone3 (0.3) gets a positive result hh[hh (.) rm==
AIDS counselling 309 4
P:
5 C: 6 7 8 9 10 11 12 13 14 15
C: P: C: P: C:
[ uhu? = not saying (lets) assume you? (0.8) that I think for one minute: (0.4) ( ) that would be the case uhu? (0.3) but people (0.2) you know have to be aware: (0.3) uhu? hhhhh when someone's getting a positive result (0.2) they ((continues))
Down-playing the relevance of the re-categorisations is in a way inherent in the 'information delivery' footing. In it, the counsellor does not talk about this particular patient but about patients in general. This may be one reason for the fact that information delivery is so widely applied in the talk about 'dreaded issues': it structurally minimises the delicacy of the re-categodsation of the patient. In T above, the counsellor simply explicates and formulates the structural neutrality of the information delivery footing that she is applying. But down-playing the relevance of the re-categorisations can be done also in an interview-footing. Extract U 1 2 3 4 5 6 7
C:
I mean as far as 1 understand if one-(0.7) I'm not saying this can happen but taking it to its very worst (.) if hhh you did begin to fi:nd it difficult to make decisions and all that, who would you want to make them for you? (0.2) Who would you want to help you?
In U, the counsellor designs her question emphasising the imaginary nature of the circumstances depicted. By doing this, she seems to be able to create a similar kind of distance between the re-categodsation and this particular patient that is structurally present in an Information Delivery footing. The participants are talking about the patient's possible dementia - but the relevancy of this re-categorisation is minimised. In talk about sexuality, there is still another option open to the professional. In Extracts E, R and S, the counsellors have used highly specific MCDs carrying delicate implications. Yet not all categories are so specific. For instance, 'couples' (Extract C) and 'partners' (F) constitute SRPs and suggest category-bound activities which are not delicate (the activity of a partner-partner pair is, after all, 'partnership'). The same applies to this question:
310 David Silverman and Anssi Perakyla
Extract V C: any girlfriends, boyfriends? Here, tbe potential delicacy, in tbe dominant culture of asking a man if he bas boyfriends, is balanced by setting up a less problematic category-bound activity ('friendship'). This unites botb categories and avoids the single, presumably stigmatising, MCD used in Extract E. Moreover, once tbe patient's answer has been produced, issues of delicacy seem to be minimised by feeding back to tbe patient ber own categories. We have already seen an example of this in Extract K where the counsellor follows an answer about 'catching something' with a question about 'catching anything'. However, tbis involves the description of an activity. Extract X below offers an example wbere the counsellor follows up tbe minimal delicacy of ber categorisations of people (line 1) witb a repetition of tbe term that the patient uses to describe bis brother ('straight'): Extract X 1 2 3 4 5 6 7
C: P: C: P:
C: 8 P:
Okay Boyfriend? Girlfriend? er:: (0.2) Neither. (0.2) Okay:, hbhh (0.7) wbich [d'you see [I suppose girlfrie:nd Yeah:= =A11 righ= = He's straight
Q y
10 C: 11
P:
He's straight. [Okay: [Mm hm
Conversely, in the latter part of Extract M, we saw the counsellor producing a considerably turbulent delivery of a delicate item ('prostitute') rather than feeding back the patient's own term ('lady') with a request for specification. These contrasting examples give us a useful basis to consider further the wider implications for professional practice. Space only allows us to address two such implications. Description and empathy 'Empathy' between professional and client is a central concern of counselling texts. Carl Rogers (1975:4) has offered a highly respected definition: (empathy) means entering the private perceptual world of tbe other and becoming thoroughly at home in it. It involves being sensitive, moment
AIDS counselling 311 to moment, to the changing felt meanings whichflowin the otber person ( . . . ) To be with another in tbis way means that, for the time being, you lay aside the views and values you hold for yourself in order to enter anotber's world without prejudice A technique held to embody this empathy is for counsellors to use their turns at talk to offer regular paraphrases of clients' utterances (NelsonJones 1988, Rogers 1975). Curiously, we have only one or two examples of the use of parapbrases in our transcripts. For instance, in line 23 below: Extract Y (continues Extract E) 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
C:
P: C: P:
C: P: C: P: C: P: C: P:
[no no it doesn't umm er why did you want to have the test done? ob well (1.0) err I've just moved in witb my current (0.3) girlfriend mm bm and err (0.5) before we actually start (2.0) full sexual relationship we thought it was best if we both (1.0) screened just to be just to be on the safe side it's not as if botb of us bave been (0.3) sleeping around for the last few years but just [to be [right it's just to put our minds at rest= =just to be safe it I mean it's unlikely that there again there is that cbance of right tbat chance exists have you been abroad a lot? no
Here, after tbe counsellor's paraphrase, the patient does indeed continue his turn at talk that looked as if it might have come to an end at line 22. This is in line with what we have observed about the possible functionality of the counsellor repeatitig the patient's terms. Happily, this is in line with recent theoretical work in linguistics. For instance, Tannen (1987) has noted the assumption that repetition in conversation is dysfunctional. However, this assumption depends upon an erroneous picture theory of language which holds that, since repetition does not appear to convey information, it is superfluous and an instance of bad communication. On the contrary, Tannen (1987:584) argues that repetition of another's words is functional because: Repeating the words, phrases or sentences of other speakers (a) accomplishes a conversation, (b) shows one's response to another's utterance, (c) shows acceptance of others' utterances and their
312 David Silverman and Anssi Perakyla participation, and (d) gives evidence of one's own participation. It provides a resource to keep talk going where talk itself is a show of involvement, of willingness to interact, to serve positive face (Tannen 1987:584) Although Tannen does not discuss the functions of repetition or mirroring of body language, this would be wholly consonant with her analysis. Indeed, this is taken up by some counselling psychologists (see Maurer and Tindall 1983). In ongoing work on videotapes of HIV counselling, we are examining how body movements accomplish various interactional tasks, including the display of 'delicacy'. This is consistent with Heath's trail-blazing work on the local and context-bound positioning of both patient and counsellor body-movement, particularly in relation to the organisation of 'embarrassment' (Heath 1988). All tbese considerations give us an interesting contrast with most accounts of empathy. They - exemplified by Rogers' definition above (see also Nelson Jones 1988; Chester 1988, Appendix 4) - imply a view of communication as a public process building a bridge between two private consciousnesses. This model has been neatly explicated by Nelson-Jones in an authoritative text-book (1982: 212-14). Within such a framework, empathy appears as the propensity of one mind to reach the lived meanings of the other one, via the medium of public expressions, verbal and nonverbal. Accordingly, research on empathy treats different verbal and non-verbal modes of action as indications of the underlying characteristics of empathic orientation between the two participants (see Hammer 1983, Ellickson 1983, Maurer and Tindall 1983, Barkham and Shapiro 1986). The problem of these studies appears to be the unacknowledged 'leap' from public to private and vice versa. All the modes of action examined in them (body and eye movements of tbe counsellor and their coordination with those of the client; the verbal response modes of the counsellor) take place in tbe public sphere; only empathy remains in the private. The public sphere is treated as an indication of the private. An important opening has been made by Barret-Lennard (1981), who wants to treat empathy as an interactional process, rather than as a quality of individuals or relations. But be also shares the above mentioned model of communication: the interaction is between two private selves, mediated through the public sphere. Contrary to both the textbooks and commonsense, our analysis suggests an approach to empathy less as the psychological propensity to attune to the private meanings of the patient, but more as the social ability to pick up behavioural and cultural cues present in what tbe patient is saying and doing. As a consequence, we methodologically 'bracket' all assumptions about the private minds represented by participants' public actions, but
AIDS counselling 313 concentrate on the logic of the public actions only. We too, as BarretLennard, see empathy as an interactional phenomenon. For us, however, the core of this process is not the interplay between two private selves, but the interplay of actions making use of publicly available apparatuses of description. It is evident, then, that in this perspective there are no a priori right or wrong ways of responding to the patient. Rather that what works has to be interactionally devised on each occasion - although there is some evidence that repetition or mirroring of the patient's linguistic and paralinguistic expressions may be functional in the sense already described. Advice-giving and counselling We have already noted that most discussion of sex and dreaded issues takes place in an information-delivery footing. Indeed, where patients talk about such matters in an interview-footing, counsellors often change footings or topics. For instance, referring back to Extract Y, although the counsellor's paraphrase of the patient's utterance (line 23) secures a further continuation of his story, she does not go on to question the patient further about his turbulent account. Rather than make use of the alignment of the patient to talking about his sexual practices, the counsellor here switches topic (line 28). Later, she will return to the delicate item of sexual behaviour but in the context of her own delivery of information. This seems related to the fact that most of the extracts we have looked at arise during pre-test counselling. Elsewhere, Silverman (1990) has remarked that such encounters are over-burdened by the number of topics that counsellors are expected to cover. One consequence of this is that they may be tempted to cut the patient's trouble-telling short in order to switch into information-delivery or, as in Extract Y, to ask a question about another topic. This is wholly consonant with an earlier observation that pre-test counselling and post-test counselling with the seronegative are weighted towards advice (Chester 1987:8). While the counselling texts recommend a systematic feeding-back of clients' statements and answers (presumably by further questions which repeat clients' terms), we have found that advice-giving is the focus of many of these consultations. Usually, this focus means that the counsellor spends most of her time in an information-delivery footing.** However, the 'dreaded issues' extracts showed how a menacing future that might include death can be addressed at pre-test counselling by inviting the patient to provide his own answers. A final extract will reveal that it is possible to handle the delicacy of a discussion about a patient's sexual practices within pre-test counselling while staying within an interview footing:
314 David Silverman and Anssi Perakyla Extract Z (BF = patient's boyfriend 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
P;
Well sexually I m more worried about him. Okay well let's think about- say you decided to have (0.4) a sexual relationship. (0.5) Uh:m (0.4) and you were the one making some decisions, =now what would be (0.7) the steps that you'd take (0.2) to make you feel less: (1.2) concerned about that. (0.6) P: Mm: Just that it had to be very safe (.) [er: [And C: what would be very [safe, let's talk [hhuhhhhbhhhhhuh= (): C: =What would you consider very safe. = John? BF Me:? C: Mm: = (BF): hhuh= BF: =(P) wearing a condom. (0.5) C: Would that be very safe. (0.2) BF: Ye:s. And no anal and no oral sex. C:
C:
Mm:
(P):
(1.0) Mm h[m
Notice how the counsellor constructs her initial question (lines 2-6) by repeating the patient's term at line 1. She then goes on to produce delicate talk by both the patient and his boyfriend simply by three simple requests to specify what they have just said. Elsewhere, however, most discussion of this topic occurs by the counsellor simply delivering information. Certainly, we find in tbese cases, pre-delicate turbulences by the counsellor but interactionally these are functionless since the patients concerned are only required to give minimal response-tokens to acknowledge the right of the counsellor to hold the floor over a number of turns-at-talk. In Extract Z, the important but delicate topic of safer sex is addressed by the patient and his partner without the counsellor having to resort to an informationdelivery footing. This implies that there is no conflict between an interview-footing and the coverage of topics in which the counsellor has specialist knowledge.'^
Conclusion
The excerpts above have shown counsellors and their clients managing talk about sex and a menacing future with the patient. Both the professionals
AIDS counselling 315 and their clients have drawn upon the same apparatus of description, which protrays these as potentially delicate issues. The experience per se of the counsellors and their clients has of course been beyond the reach of our gaze; following Sacks, what we have been able to reach and to analyse is the apparatus of description regulating talk about experiences. This is something shared by the counsellors and patients. This suggests a revision to the conceptions we may have about professional activity. The skills of the counsellors we have examined in these excerpts are not primarily based on owning a special (professional) body of knowledge. This type of conception of professionality is common (see e.g. Freidson 1970, Schutz and Luckman 1974). Instead, the skills we have been analysing depend upon an apparatus of description that is publicly available for anybody. The distinctive character of counselling arises in the systematic deployment of this apparatus in encouraging the patient to talk. Goldsmiths College London University Acknowledgements We would like to acknowledge the help of Christian Heath with specific data extracts and for encouraging us to develop the form of analysis developed here. Naturally, he is not responsible for the use we have made of his help. We owe a debt of gratitude to all the patients and staff who kindly gave us access to their conversations. In particular, we wish to thank Riva Miller and Eleanor Goldman (Royal Free Hospital Haemophilia Centre) and Robert Bor and Heather Salt (Royal Free Hospital AIDS Counselling Unit) whose counselling theory and practice has been a constant source of insight for us. Notes 1 For example. Prior's (1989) inventive ethnography shows how 'death' is differently constituted in the mortuary, the Coroner's Court, the official records' office, the funeral parlour and the announcements in the classified columns of the newspapers. In relation to another setting, Perakyla (1989) shows how 'death' is differently constituted in connection to the different aspects of hospital work. 2 However, we have to be cautious in our use of Goffman's concept. Goffman is concerned with how parties collaborate to present a front but: (1) we need not infer the presence of a social-psychological strategy to preserve a version of identity but rather local work to recognize that other parties are co-present; (2) in turn, this work is embedded in situated, local practices - it cannot be understood without reference to its place in a sequence of managed talk. 3 In both extracts, the way in which the participants initially respond to the C's questions points to a potentially delicate item. This item only emerges, in a stepwise fashion, when the C presses on with further enquiries. But the early turbulent answers suggest that the delicate item is there, from the start, some
316 David Silverman and Anssi Perakyla distance from its actual delivery as something implicit which all parties are orienting themselves towards. 4 Sometimes, however, it involves the C in giving a direct answer to a P's question. We have already seen this in Extract E (lines 9-11), where the C agrees with the P that his avoidance of a 'homosexual relationship' means that he is not in a highrisk group. Compare this answer (with its problematic prediction of the HIV-test result), to asking the P what he understands by 'a high-risk group'. 5 It may be argued that staying within an interview-footing works here because both clients are gay men who probably possess considerable knowledge of safer sex already; however, we have several examples of this approach being used where the Ps concerned may have a very different knowledge-base (e.g. heterosexual persons being pre-test counselled and seropositive haemophiliac teenagers). Here the P's own descriptive apparatus is nonetheless dredged up by the C continuously requesting specification of his answers. At that point, usually at the conclusion of the consultation, the C can use that apparatus to deliver specialist knowledge that has not otherwise been covered.
Appendix: The transcription symbols quite a [ while [yea
Left brackets indicate the point at which a current speaker's talk is overlapped by another's talk. W: that I'm aware of = Equal signs, one at the end of a line and one at C: = Yes. Would you the beginning, indicate no gap between the confirm that? two lines. (0.4) Yes (0.2) yeah Numbers in parentheses indicate elapsed time in silence in tenths of a second. to get (.) treatment A dot in parentheses indicates a tiny gap, probably no more than one-tenth of a second. Underscoring indicates some form of stress, via What's up? pitch and/or amplitude. Colons indicate prolongation of the O:kay? immediately prior sound. The length of the row of colons indicates the length of the prolongation. WORD Tve got ENOUGH TO Capitals, except at the beginnings of lines, WORRYABOUT indicate especially loud sounds relative to the surrounding talk. A row of h's indicates an inbreath or an hhhh I feel that (0.2) hhh outbreath. The length of the row of h's indicates the length of the in-or outbreath. Empty parentheses indicate the transcriber's ( ) future risks and ( ) inability to hear what was said. and life ( ) (word) Would you see (there) Parenthesised words are possible hearings. anything positive Double parentheses contain author's (( )) confirm that descriptions rather than transcriptions. ((continues))
AIDS counselling 317 References Aggleton, P. (1989) Evaluating health education about AIDS. In P. Aggleton, G. Hart and P. Davies (eds), AIDS: Social Representations. Social Practices, Lewes: Falmer Press. Barkham, M. and Shapiro, D.A. (1986) Counsellor Verbal Response Modes and Experienced Empathy. J. of Counselling Psych., 33, 3-10. Bor, R. and Miller, R. (1988) Addressing 'dreaded issues': a description of a unique counselling intervention with patients with AIDS/HIV. CounseUing Psychology Quarterly 1, 397-405. Bordieu, P. (1977) Outline of a Theory of Practice. Cambridge: Cambridge University Press. Chester, R. (1987) Advice, Support and Counselling for the HIV Positive (a report for DHSS). Institute for Health Studies, University of Hull. Coxon, A. (1989) Who's doing what and to whom . . . The Times Higher Education Supplement, 17.11.89. Ellickson, J.L. (1983) Representational Systems and Eye Movements in an Interview. J. of Counselling Psych., 30, 339-45. Foucault, M. (1979) The History of Sexuality, Volume I. London: Allen Lane. Freidson, E. (1970) Profession of Medicine. A Study of the Sociology of Applied Knowledge. New York: Harper and Row. Goffman, E. (1959) The Presentation of Self in Everyday Ufe. New York: Doubleday Anchor. Heath, C. (1988) Embarrassment and Interactional Organization. In P. Drew and A. Wootton (eds), Erving Goffman: Exploring the Interaction Order. Cambridge: Polity Press, 136-60. Maurer, R.A. and Tindall, J.H. (1983) Effects of postural congruence on client's perception of counsellor empathy. J. of Counselling Psych., 30, 158-63. McKeganey, N. (1989) Being positive: drug injectors' experiences of HIV infection. Social Paediatric and Obstetric Research Unit, University of Glasgow. Nelkin, D. (1987) AIDS and the social sciences: review of useful knowledge and research needs. Review of Infectious Diseases, 9, 947-60. Nelson-Jones, R. (1982) The Theory and Practice of Counselling Psychology. London: Cassell. Nelson-Jones, R. (1988) Practical Counselling and Helping Skills, London: Cassell. Perakyla, A. (1989) Appealing to the experience of the patient in the care of the dying. Sociology of Health and Illness, 11, 117-34. Perakyla, A. and Silverman, D. (forthcoming) Reinterpreting speech exchange systems: communication formats in AIDS counselling. Sociology. Rogers, C.R. (1975) Empathic: an unappreciated way of being. The Counselling Psychologists, 2-10. Sacks, H. (1984) On doing 'Being ordinary'. In J.M. Atkinson and J. Heritage (eds). Structures of Social Action, Cambridge: Cambridge University Press, 428. Sacks, H., Schegloff, E.A. and Jefferson, G. (1974) A simplest systematics for the organization of turn-taking for conversation. Language 50, 696-735. Schegloff, E.A. (1981) Discourse as an interactional achievement: some uses of 'uh huh' and other things that come between sentences. D. Tannen (ed.), Georgetown University Roundtable on Language and Linguistics, Washington DC: Georgetown University Press, 71-93.
318 David Silverman and Anssi Perakyla Schutz, A. and Luckmann, T. (1974) The Structures of the Lifeworld, London: Heinemann. Silverman, D. (1989) The impossible dreams of reformism and romanticism. In J. Gubrium and D. Silverman (eds). The PoUtics of Field Researeh: Sociology Beyond Enlightenment, London: Sage, 30-48. Silverman, D. (1989a) Six rules of qualitative research: a post-romantic argument. Symbolic Interaction 12, 25-40. Silverman, D. (1990) The social organization of HIV counselling. In P. Aggleton et al (eds), AIDS: Individual, Cultural and Policy Perspectives, Lewes: Falmer, 191-211. Strong, P.M. and Dingwall, R. (1989) Romantics and Stoics. In Gubrium and Silverman, op cit, 49-69. Tannen, D. (1987) Repetition in conversation: towards a poetics of talk. Language, 63, 574-605. Taylor, C. (1986) Foucault on freedom and truth. In D. Hoy (ed), Foueault: A Critical Reader, Oxford: Basil Blackwell, 69-102.
