154 Qualitative Psychology (14) (expan 3 joy: 2 Les: 3 4 5 6 Joy: 7 8 Les: 9 10 n joy 12 Les 13 Les 14 3oy 15 16 17 Les: 18 19 ísion of (10)) [11011:2:15:4-51 =eh Well surely she's clever J mentally isn't s|hc [Oh ! don't know'bout jthat, ! mean u h 1 don'lhmk it's all thai difficult really (0.4) What. (0.5) If you've got- if you got the schooling an' the backTgľOu:nd ih-uh (.) ( )- (0.4) Oh[no(h)o perhaps that's what it is 1 don't know f( ) iNo[: : : , | [( ) j Oh well 1 don't jknow thougii I d- 5 should imagine she is clever her children')- clever aren't they, ,hhhh yih know l[mean] I MP:_:]"[: no they're not. Only: one is outstandingly clever wujv an:' the other- .hh an:'°Rebecca didn't get ť college,0 After Lesley's initial disagreement in lines 2-3, and subsequent elahcir. ation (lines 8-9), Joyce pursues her assessment of their friend's likely cleverness, stating as supporting evidence that her children are clever (line 15). Without tracing this in detail, it is reasonably clear that neither is entirely letting go of her position regarding their friend's cleverness, and that they have, in effect, 'upped the ante'. At this point, in line 17, Lesley further pursues and escalates the disagreement by very strongly contesting Joyce's claim that the children are smart: the extent to which she has escalated the strength of her disagreement is evident in its being strongly marked - lexically, through the outright negative tokens, and direct rejection of Joyce's statement; prosodically, through raised pitch and amplitude. So it is in this environment, in pursuing her disagreement, and doing so in a strongly marked form, that Lesley produces her rebuttal of Joyce's claim that her children are clever. Her rebuttal is designed to equal the strength of her (escalated) disagreement. Thus, the completeness and strength of her rejection ('NO::f: no they're not.') is matched by her claim that one of the children did not even get into college - while the (act that both children are at college (that is, university) would hardly be commensurate with or support her claim, against Joyce, that they are not clever. One further example: recall that in (1Í), line 3, Sandra claims initially never to have been to a disco. Looking at the sequence leading up to this claim, we see that she and Becky are talking about their friends/housemates going to a local club that night {[15], lines 1 -7). Conversation analysis 3 San: 'J Bee: 5 San: í) Bee: 7 San: 8 Bee: 9 ]() San: H Bee: 12 San: 5 3 Bee: M San: (15) (Expansion of (51) jDrew:Sl:I) ('Silks' is a local disco club) 1 liec: We were all talking about going out t- Silks lonight'cause 2 everyone's got the day off tomorrow? Are you- c/_ my house is all going t- Silks tonighl?= -Really Yea:h I'.jmma un Ges urn Qe- jiiet it's gormer be absolutely pa:cked though isn't it. Yeah and Ces has been radding my wardrobe. So: hhjh [.hhh Are you going. No::, j Why:: í don't know hhh hu hu .hhh 1 dunno it's not really me Mw:rh ( ) like it .hh I've never been to one yet, In response to her inquiry in lines 8-9, Sandra tells Becky that she is not going; and when Becky pursues this with an expression of evident disbelief (line II), Sandra explains that it's not really me - which she supports by adding that I've never been to one yet (line 14). Thus, her declining to go on the grounds that it's not really me is made more credible by her claiming never to have been to such a place; of course, this also detaches her reasons for not going this evening from anything which might relate to this particular occasion. In this way, her claim that she has never been to one yet is fitted to a sequence in which her friend has responded to her declining to go (in line 10) with disbelief: Sandra matches the strength of that disbelief with an account which seems incontrovertible. . What emerges, then, is that these strong, dramatic or perhaps cxagger- / a ted claims arise from, or are fitted to, the contingencies of the particular action sequence in which they are produced. They are constructed to 'work' in terms of the 'requirements' of the slots in which they are done. The 'weaker' versions to which they subsequently retreat would not have done the job in the slot in which they are produced. For instance, in the case with which we initially began, when she confirms that not a word lias been heard from her ex-husband, Nancy is simultaneously both agreeing with Emma, and complaining about him (simultaneously, because she is joining or collaborating with Emma's implied complaint about him, on her, Nancy's, behalf). The strong version riot a word works to agree/complain; had she began, in line 38 in extract (8), by reporting that his mother had heard from him, she would not have been agreeing/complaining ~ indeed, she might . have been heard as disaffiliating, or disassociating herself from Emma's (sympathetic) implied complaint. Thus, the weaker version which she comes to would not have achieved the responsive action which Emma's prior turn is built to get, that is, confirmation, agreement and collaboration in a 156 Qualitative Psychology 1 -*"* » -í complaint about her ex-husband. Hence, the subsequent weaker versions would not have accomplished, in a coherent fashion, the work of reporting, disagreeing, confirming/agreeing, complaining, giving an account etc., in the particular positions in which speakers construct those actions. Therefore, these initial (over-)strong versions are fitted to the slot in which speakers art-announcing, disagreeing, declining, etc.; here, we can see that the speakers are dealing, through these claims, with the exigencies which have arisen in the immediate (prior) sequential environments. Speakers produce versions which are fitted to those sequential moments- When the moment is past, so, too, is the 'requirement' for thai strong version: the speaker can put the record straight and retreat to a 'weaker' version {albeit in a manner which maintains an essential consistency with the initial 'false' claims). ':V-tf?^\-*: ':i .. ->t; *,', * -. J** '» E*a£. Conclusion 1 have tried here to give an account of the principal stages of analysis in CA research - the stages which are involved when developing research findings about the ways in which interaction, and particularly verbal interaction, are organized. Space has not allowed me to say much about the significance ot what we are looking for, or about the theoretical standpoint which this perspective adopts towards people's activities in talk - for instance, the reasons for considering talk as action rather than communication; and our identifying patterns associated with manifest behaviour, and not inner cognitive states and other states such as intentions, motives, personality, etc. However, i hope that showing how Nancy's apparent inconsistency in extract (8) is simply one in a recurrent pattern in which speakers initially produce over-strong versions, in order to fit the contingencies of the particular interactional sequences in which they are engaged, helps to illustrate that CA resists psychological accounts of behaviour which turn out to be general (rather than individual), systematic (and not particularistic) and interactional (rather than arising from the psyche of one of the participants, as though she were acting independently of the other and the interaction between them). What Nancy does here is what people - of whatever psychological dispositions or types - do generally, given these interactional circumstances. Nor have I been able to give any account of another very important aspect of CA research, which is to distributionatize phenomena - that is, to try to identify where in conversations certain devices or patterns lend to occur. We take the view that everything about conversation (and other forms of verbal interaction) is thoroughly organized, so that what people say, and how they say it, is not random or chance (see Chapter 8 in Heritage, 1984, for a beautifully clear account of this fundamental assumption o! CA). An example of research showing that a device in conversation is systematically distributed is Drew and Holt's account of the way in which idioms and figures of speech are used to terminate topics of conversations: Conversation analysis 157 hence, figures of speech are distributed in an organized fashion, occurring predominantly at points where speakers move from one topic to another (that is, topic transition) (Drew and Holt, 1998). J have, though, tried to describe how we cut in to looking at data, and start to make analytic observations about (verbal) conduct. Starting with a transcript of a recording of natural conversation, we begin by looking at the activities which participants may be managing through their talk; then, we examine in as much detail as possible how their talk is designed or constructed, in an effort to map the organized properties through which participants conduct their affairs in talk-in-interaction. And 1 have illustrated how we develop an analysis of such organized properties (patterns, devices and practices) by focusing on what at first sight appears to be an incidental curiosity in the extract with which 1 began - that is, Nancy's initially confirming that no one heard a word (from her ex-husband), but subsequently reporting that his mother had heard from him. This was to give you some feeling for how we can move from making observations about the details of talk, to developing an analysis of a conversational phenomenon or practice - in other words, arriving at findings about stable and systematic patterns in talk. If only it were as easy as this! I have to admit that the process or steps from beginning to notice things about the detail to be found in talk, to the end product of a publishable research finding, are not nearly as smooth as this account might have suggested. For one thing, there is the difficulty of knowing what kinds of details one might begin noticing, and what to say about them. The next hurdle is to decide whether what one is focusing on is actually a phenomenon (that is, a systematically organized pattern or practice). Afterwards, building a collection of the phenomenon can involve comparative questions, including what kinds of cases the phenomenon encompasses and what cases might be used for comparison and contrast -all of which are not dissimilar from the decisions which need to be made in experimental design. This is too brief an account of CA's methodological approach to do any justice to these complexities: the only way to find out more is to survey the research which has been published (see below), collect some naturally occurring interactional data relating to some topic in which you are interested, and have a go at doing it yourself. 158 Qualitative Psychology Further Reading Heritage, J. (1984) Gariinkel and Ethnomethodoiogy. Cambridge: Polity Press, ch. 8. This is perhaps the best concise overview of conversation analysis. Atkinson, j.M. and Heritage, J. (eds) (1984) Structures of Social Action; Studies in Conversation Analysis. Cambridge: Cambridge University Press. The editor's introduction gives an invaluable brief guide. Sacks, H. (1992) Lectures on Conversation, G. Jefferson (ed.), Vols. 1 and 2. Oxford: Blackweil. See Schegloff's introductions for a definitive account of Sacks's work and the development of the field, and the distinctiveness of CA's approach. Sacks's lectures themselves are an essential resource for anyone interested in CA's analytic approach. Ten Have, P. (1999) Doing Conversation Analysis: A Practical Guide. London: Sage. Hutchby, I. and Wooffitt, R. (1998) Conversation Analysis: Principles, Practices and Applications. Cambridge, Polity. These are useful texts on how CA works. Drew, P. and Heritage, J. (eds) (1992) Talk at Work. Cambridge: Cambridge University Press. CA is widely applied to forms of talk-in-interaction other than 'ordinary conversation' - for instance, to the study of interactions in such 'institutional' settings as courts, classrooms, medical consultations, news media, counselling and therapy. This book gives an overview and a collection of studies covering a wide range of settings. Chapters Discourse analysis Carla Willig in recent years, discourse analysis has gained popularity and acceptance as a qualitative research method in psychology. As an increasing number of lesearchers turn to the analysis of discourse, it is worth exploring what a discursive analysis can actually deliver and what kinds of research questions it can, and cannot, address. fiKi'O'rN In this chapter, 1 introduce two versions,_Qf-the..discourscanaiyti^..'-'^ :, -^,"1 metho^djscursivej)sychologyand Fůucauldictn disc<>uise_m.alysi$,J&.wJh.auKHy-' '■""""'"'' these Jw. a^ro^ i- constffictio.n...oi..5ociaLi^ ^ , . icsearch. questions. They also identify with different theoretical traditions^-";' i ' ; Burr (1995) and Parker (1997) provide detailed discussions of the distinction-ty^," \*i>^ between the two versions of discourse analysis. However, some discourse '{\\ anaäysľs do not welcome such a strong conceptual separation. For example,! ; ' jPgttexand..VVetherell-.(1995: 81) argue that the distinction between the two versions 'should not be painted too sharply' while Wetherell (1998) also advocates a synthesis of the two versions. This chapter introduces and describes the two approaches to discourse analysis and illustrates each with a worked example. The two versions of discourse analysis are applied to the same interview extract in order to highlight similarities and differences between them. The chapter concludes with a comparison between the two discursive methods. Psychology's Turn to Language Psychologists' turn to language was inspired by theories and research which had emerged within other disciplines over a period of time. From the 1950s onwards, philosophers, communication theorists, historians and sociologists became increasingly interested in language as a social performance. The assumption that language provides 3 set of unambiguous signs with which to label internal states and with which to describe external reality began to be challenged, instead, language was reconceptualized as productive; that is to 159 160 Qualitative Psychology say, language was seen to construct versions of social reality, and it was seen lo achieve social objectives. The focus of inquiry shifted from individuals and their intentions lo language and its productive potential. Wittgenstein's philosophy, Austin's speech-act theory and Foucauíťs historical studies of discursive practices are important examples of this shift. However, psychology remained relatively untouched by these intellectual developments throughout the 1950s and 1960s. Instead, it was concerned with the study of mental representations and the rules which control cognitive mediation of various types of input from the environment. In the 1970s, social psychologists began to challenge psychology's cognitivism (e.g., Gergen, 1973; 1989), and in the 1980s the 'turn to language' gained a serious foothold in psychology. Discursive Psychology This version of discourse analysis was introduced into socia! psychology with the publication of Potter and Wetherell's Discourse and Social Psychology: Beyond Attitudes and Behaviour in 1987. The label 'discursive psychology' was provided later by Edwards and Potter (1992). Potter and Wetherell's book presented a wide-ranging critique of cognitivism, followed by a detailed analysis of interview transcripts using a discourse analytic approach. Later publications developed the critique of psychology's preoccupation with cognition and its use as an all-purpose explanatory strategy which involved 'claiming for the cognitive processes of individuals the central role in shaping perception and action' (Edwards and Potter, 1992: 13). The critique of cognitivism argues that the cognitive approach is based upon a number of unfounded assumptions about the relationship between language and representation. These include: 1) that talk is a route to cognition, 2} thai cognitions are based on perception, 3) thai an objective perception of reality is theoretically possible, 4) that there are consensual objects of thought, and 5) that there are cognitive structures which are relatively enduring. Each of these assumptions can be challenged from a discursive psychology perspective. Talk Is a Reuiü to Coönition From a cognitive point of view, people's verba! expression of their beliefs and attitudes provides information about the cognitions which reside in their minds. In other words, talk is a route to cognition. As long as the researcher ensures that participants have no reason to lie, their words are Discourse analysis 161 taken to constitute true representations of their mental state (such as the ; beliefs they subscribe to or the attitudes they hold). Discourse analysts do ; not share this view of language. They argue that when people state a belief or : l''J express an opinion, they are taking part in a conversation which has a \ '/'.''" purpose and in which ail participants have a stake. In other words, in order ■ "ľ~,. to make sense of what people say, we need to take into account the social context within, which they speak. For example, when male participants are interviewed by a female researcher with the aim of identifying men's attitudes towards sharing housework, their responses may be best understood as a way of disclaiming undesirable social identities (as 'sexist slob', as dependent upon their female partners or as lazy). This is not to say that they are lying to the researcher about the amount of housework that they do; rather, it suggests that in their responses, participants orient towards a particular ; reading of the questions they are being asked (such as a challenge, a criti- ; cism or an opportunity to compiain), and that the accounts they provide [ need to be understood in relation to such a reading. As a result, we shouid not be surprised to find that p,epplť consistent across jpcial. contexts. Cognitions Are Based on Perception Ultimately, cognitivism has to assume that cognitions are based on perceptions. Cognitions are mental representations of real objects, events and processes which occur in the world. Even though cognitions are abstractions, and therefore often simplifications and distortions of such external events, they do constitute attempts to capture reality. Once established, cognitive schemata and representations facilitate perception and interpretation of novel experiences and observations. By contrast, discourse analysts argue that the s - »i. world can be 'read' in an unlimited number of ways, and that, far from ; A; giving rise to mentai representations, objects and events are, in fact, con- í a.-_, ' structed through language itself. As a result, .it; i s discourse_aj^cmiyersation : ""•"''' which should be the focus of study, because that is_where meanings_are created_.and negotiated. Objective Perception of Reality Is Theoretically Possible if cognitions are based on perceptions, as proposed by cognitivism, it follows that an objective perception of reality is theoretically possible. Errors and simplifications in representation are the result of the application of time-saving heuristics which introduce bias into cognition. Given the right circumstances, it should be possible to eliminate such biases from cognitive processes. Again, discourse analysts take issue with this assumption. If language constructs, rather than represents, socia! reality, it follows that there can be 162 Qualitative Psychology no objective perception of this reality. Instead, emphasis..«placed upon the waysin.wh.iciuodal. caie^piies.ai;e.cpnstriict.ed and with what conseq.uena'i: they are deployedin. conversation. There Are Co^sensua! Objeo's of Thought Attitudes describe how people feei about objects and events in the social world, whereas attribution theory is concerned with how people account for actions and events. In both cases, researchers assume that the social object ot event towards which participants have different attitudes and which participants attribute to different causes, is itself consensual. That is to say, even though people hold different attitudes and attributions in relation to something (for example, European Monetary Union, same-sex marriages or the break-up of the Soviet Union), that 'something' itself is not disputed. In other words, there are consensual objects of thought, in relation to which people form opinions. People agree on what it is they are talking about, hul they disagree about why it happened (attributions) and whether or not il is a good thing (attitudes). Discourse analysts do not accept that there are such consensual objects of thought. They argue that the sociai objects themselves are constructed through language and that one person's version of, say, 'the break-up of the Soviet Union' may be quite different from another person's. From this point of view, what have traditionally been referred to. as. 'attitudes'and .'attributions' are,.in fact,, aspects.of the discursive, construction „of thcobjecUtself. There Are Relatively Enduring Cognitive Sti'iicturos Finally, cognitivism is based upon the assumption that somewhere inside the human mind there arc cognitive structures which arc relatively enduring. People are said to hold views and have cognitive styles. Cognitive structures can change, but such change needs to be explained in terms of intervening variables such as persuasive messages or novel experiences. The assumption is that in the normal course of events, beliefs, attitudes, attributions and so forth remain stable and predictable from day to day. Discourse analysts' conceptualization of language as productive and performative is not compatible with such a view, instead, they argue that people's accounts, the views they express and the explanations they provide, depend upon the discursive context within which they are produced. Thus, what people say tells us something about what they are doing with their words (disclaiming, excusing, justifying, persuading, pleading, etc.) rather than about the cognitive structures these words represent. Discourse analysts' challenge to cognitivism shows that discourse analysis is not simply a research method. It is a critique of mainstream Discourse analysis 1 psychology, it provides an alternative way of conceptualizing language, and tl indicates a method of data analysis which can tell us something about the discursive construction of social reality. Discourse. anai.ys.is..is.m.ore,.th.an a methodology.because 'it involves, a. theoretical way. of. .understanding- the nature of discourse and the.nature .of-..psychological.phenomena'...(ISi!lig, i'W: 43). However, discursive psychology is stiii a psychology because it is concerned with psychological phenomena such as memory, attribution and , Identity, But, in line with its critique of cognitivism, discursive psychology, conceptualizes these phenomena as discursive actions rather than as cognitive/ processes. Psychological activities such as justification, rationalization, categorization, attribution, naming and blaming are understood as ways in whicn~~parfičipänts manage their interests. They are discursive practices which are "used by participants within particular contexts in order to achieve social and interpersonal objectives. As a result, psychological concepts such as prejudice, identity, memory or trust become something people do rather than something people have. The focus of analysis in.discursive._psychology is on..how .participants j use discursive resources and .with, what .efÍ£cts^.jn._.oth.er...words,..discursive :, psychologists pay attentionjo^.e...ň.c.fifiB^ corned with...the ways.in.which.speakers_manage issues, of.stake...andJiiier.est. They identify discursive strategies such as 'disclaiming' or 'footing' and explore trjeir function in a particular discursive context. For example, an ! interviewee may disclaim a racist social identity by saying '1 am not racist, but i think immigration controls should be strengthened' and then legitimate the statement by referring to a higher authority: '! agree with the Prime Minister's statement that the situation requires urgent action.' Other discursive devices used to manage interest and accountability include the use of metaphors and analogies, direct quotations, extreme case formulations, graphic descriptions, consensus formulations, stake inoculation and many more (see Edwards and Potter, 1992; Potter, 1996, for a detailed discussion of > such devices), ilox 8.1 summarizes discursive psychology's major concerns. How To Do Discursive Psychology Ideally, this type of analysis should be used to analyse naturally occurring text and talk. This is because the research questions addressed by discursive psychologists are concerned with how people manage accountability and stake in everyday life. For example, tape recordings of naturally occurring conversations in informal (for example, friends chatting on the telephone, families having meals together) and formal (for example, medical consultations, radio interviews) 'real-world' settings constitute suitable data for discursive analysis. However, both ethical and practical difficulties in 164 Quaíítative Psychology 8ox 8.1 Discursive psychology • emerged from ethnomelhodology and conversation analysis • is concerned with discourse practices » emphasizes the performative qualities of discourse • emphasizes the fluidity and variability of discourse • prioritizes action orientation and stake • asks, 'What are participants doing with their talk?' obtaining such naturally occurring data have led many discourse analysis to carry out seml-stru.ctoed_i^tcLrviews to generate data for analysts. In any case, discourse analysis works with texts, most of which are generated by transcribing tape recordings of some form of conversation (see Potter and Wetherell, 1987; O'Connell and Kowal, 1995, for guidance on transcription). ! It is important that the transcript contain at least some information about non-linguistic aspects of the conversation such as delay, hesitation or : emphasis. This is because the way in which something is said can affect its ; meaning. Discourse analysis may be described as a way of reading a text. This reading is informed by a conceptualization of language as performative. This means that the reader focuses upon the internal organization of the discourse in order to find out what the discourse is doing. It means moving beyond an understanding of its content and to trace its action orientation. Discourse analysis requires us to adopt an orientation to talk and text as social action, and it is this orientation which directs our analytic work. Although there is no universally agreed set of methodological procedures, the following guidelines for the analysis of discourse can help the analyst get started (see also Potter and Wetherell, 1987: 160-76; Billig, 1997: 54, for guidance). Reading First of all, the researcher needs to take the time to read the transcripts carefully. Although the researcher will continue to read and reread the transcripts throughout the process of coding and analysis, it is important that the transcripts are read, at least once, without any attempt at analysis. i This is because such a reading allows us to experience as a reader some of the ■: discursive effects of the text. For example, a text may come across as an Discourse analysis 165 apology even though the words '1 am sorry' arc not actually spoken. We may feel that a text 'makes it sound like' there is a war going on even though the topic of the transcribed speech was a forthcoming election. Reading a text / before analysing it allows us to become aware of what a text is doing. The ; purpose of analysis is to identify exactly how the text manages to accomplish tliis. Coding Reading and rereading of the transcripts is followed by the selection of material for analysis, or coding. Coding of the transcripts is done in the light of the research question. Ail relevant sections, at text are..highlighted, ■ copied and hied for analysis. A^ this _stage,JL.^ ; that all; material which is poKjntially_jeley.anrJ5Jncluded..,.This_rneans that ■ even instances ■which are .indirectly .or.p.nly vaguely related to_.the research . \ question ..should be .identified. Most importantly, use of certain key words is not required for selection of textual material. All implicit constructions (MacNaghten, 1993) must be included at this stage. The need for coding before analysis illustrates that we can never pro- ,■ duce a complete discourse analysis of a text. Our research question identifies \ a particular aspect of the discourse which we decide to explore in detail Coding helps us to select relevant sections of the texts which constitute our < data. There are always many aspects of the discourse which we will not ; analyse. This means that the same material can be analysed again, generating ; further insights. Analysis Discourse analysis proceeds on the basis of the researcher's interaction with the text. Potter and Wetherell {1987: 168) recommend that throughout the process of analysis the researcher asks, 'Why am I reading this passage in this •' way? What features [of the text] produce this reading?' Analysis of textual data is generated by paying close attention to the constructive and functional dimensions of discourse. In order to facilitate a systematic and sustained exploration of tiiese dimensions, context, variability and construction of discursive accounts need to be attended to. The researcher looks_at bow the jfi^...coj3&txuctsi.its,.pbjects and_ subjects,,,how such constructions vary across_dhcursiye contexts, and with w"hat_consequejices.,^iey__may. be deployed... In order to identify diverse constructions of subjects and objects in the text, we need to pay attention to terminology, stylistic and grammatical features,, and preferred metaphors and other figures of speech which may be used.in their construction. Potter and Wetherell (1987: 149) refer to such systems of terms as 'interpretative repertoires'. Different repertoires are used Qualitative Psychology to construct different versions of events, for example, a newspaper article may refer to young offenders as 'young icaraways', while defending lawyers may describe their clients as 'no-hope kids'. The former construction emphasizes the unconlroUability of young offenders and implies the need for stricter parenting and policing, while the latter draws attention to the unmet psychological and educational needs of young offenders and highlights the importance of social and economic deprivation. Different repertoires can be used by the same speaker in different discursive contexts in the pursuit of different social objectives. Part of the analysis of discourse is to identify the action orientation of accounts. In order to do this, the researcher needs to pay careful attention to the discursive contexts within which such accounts are produced and to trace their consequences for the participants in a conversation. This can he done satisfactorily only on the basis of an analysis of both the interviewer's and the interviewee's contribution to the conversation, ft is important to remember that discourse analysis requires us to examine language in context. I.nte.rj]rejaliye.xe^dQJTg-j ar^ used to construct alternative, and often contradictory, versions of events. Discourse analysts have identified conflicting repertoires within participants' talk about the same topic, for example, Potter and Wetherell (1995) found that their participants used two different repertoires in order to talk about Maori culture and its role in the lives of Maoris in New Zealand - 'culture-as-heritage' and 'culture-as-therapy'. Billig (1997) identifies two alternative, and contrasting, accounts of the meaning of history in participants' discussions of the British royal family: 'history as national decline' and 'history as national progress'. The presence of tensions and contradictions among the interpretative repertoires used by speakers demonstrates that the discursive resources which people draw on are inherently dilemmatic (see Billig et al., 1988; Billig, 1991). This is to say, they contain contrary themes which can be pitted against each other within rhetorical contexts. In order to understand why and how speakers are using a particular theme, we need to look to the rhetorical context within which they are deploying it. Again, the analytic focus is upon variability across contexts and the action orientation of talk. Writing Writing up discourse analytic research is not a process which is separate from the analysis of the texts. Both Potter and Wetherell (1987) and Billig (1997) draw attention to the fact that writing a report is itself a way of clarifying anaiysis. The attempt to produce a clear and coherent account of one's research in writing allows the researcher to identify inconsistencies and tensions which, in turn, may lead to new insights. Alternatively, the researcher may have to return to the data in order to address difficulties and problems raised in the process of writing. Discourse analysis 167 X Worked Example The extract in iiox 8.2 is taken from the transcript oí a sc m i-structured interview with a woman who had recently experienced the break-up of an intimate relationship, 'flic extract represents an exchange between the interviewer (1) and the participant (!i) which occurred about halfway through the hour-long interview. Discursive Psychology: a Reading Reading An initial reading of the first half of the extract (lines 1-26) leaves me feeling weary. The text appears to hear testimony to the speaker's considerable efforts in coming to a decision about how to end her relationship with hor partner. It invokes a decision not taken lightly. The speaker comes across as mature and responsible in her way of dealing with the task of breaking up. A first reading of the second half of the extract (lines 27-42) evokes a sense of finality. There appears to be no ambiguity in its message, and its conclusion (the end of the relationship) seems inevitable. The purpose of the j analysis is to understand how the text achieves these impressions. Coding The material for analysis was selected in the light of the research question, > which was concerned with the ways in which the participant accounted for the break-up of an intimate relationship. Both parts of the present extract ) (lines 1-26 and lines 27-42) represent occasions within the conversation which provided the participant with an opportunity to elaborate upon the circumstances surrounding the end of the relationship. This meant that they constituted suitable data for analysis within this context. Analysis Part 1 (lines 1-26). in response to the interviewer's question (1: 'did you talk to friends about it?', lines 1-2), the participant uses an extreme case formulation-Call the time')- In..this.way, her claim (to have discussed the situation with friends) is taken to its extreme in order to provide an effective warrant (Pomerantz, 1986) for her ultimate decision (to end the relationship). It is suggested that this decision is based on careful consideration informed by frequent discussions with friends. Through the use of list-like sentence constructions and the use of repetition ('How do I do it, how do 5 say it, what do 1 say', lines 5-7; and again 'How is he going to cope, what's 168 Qualitative Psychology Box 8.2 Extract from break-up interview 1 i: And when you made the decision urn when you were actually working 2 towards tinishing it did you talk to friends aboul it? 3 R: Oh of course 4 I: Yeah"..... 5 R: All thetime yeah it would always be a case of how do ä do it 6 1: Ail right 7 R: How do I say it what do I say 1 know I've got to do if how do 1 go about doing 8 it you know and and just sort of role-playing it through and and you know just 9 sort of just preparing myself to actuaiiy say to him I don't want to go out with 10 you anymore because it's so hard even though you know it's got to be done 11 It is just so hard because there's all these you know ties and emotional 12 baggage which is which you're carrying and you you you're worrying about 13 the other person and you're thinking you invested you know he's invested !4 maybe two years in me (5 I: Yes 16 R: by going out with me and suddenly I'm dumping him what if he doesn't find 17 anyone else to go out with 18 1: Oh right yes 19: R: You you start taking responsibility for them and for how they'll cope 20 afterwards you know maybe to the detriment to your own personal son of 21 well-being 22 I: Right 23 R: And it was a case of how is he going to cope what's going to happen to him 24 what if no one goes out with him what if this and what if that and it's all a 25 case of ifs anyway and you know as far as I was concerned i was I was 26 more concerned about him and how he would be [. . .] [and a utile later in the interview] 27 i: [. . .] if you sort of think about it as going on through time urn was there 28 anything that changed in the way you behaved towards each oltier or sex 29 life or anything like thai? Could you say you know something changed or 30 R: No it was the way I saw it was would I want to marry him was the sort of urn 31 you know foundation I would use 32 I: Right 33 R: because i thought OK we've Seen going out for two nearly two years if we 34 were going out for another two years would I want to marry him and the 35 answer was no 36 I: Right 37 R: And even though [. . .] I had no intentions of getting married say for another 38 you know four five whatever amount of years it was on ihat.basis I was 39 using the criteria of my wanting to continue going out with him 40 !: Right 41 R: because it was a case of where is this relationship going and as far as I was 42 concerned it had hit the the brick wall and it wasn't going any further Discourse analysis 169 going to happen to him, what if no one goes out with him, what if this and what if that', Sines 23-24), a commitment to thorough and careful consideration of ali eventualities is demonstrated. References to 'role-playing' (liřie"8)"ánd 'preparing myself (line 9) reinforce this impression by suggesting that such consideration includes the mental anticipation and practical rehearsal of possible scenarios. Use of terminology such as 'ties and emotional baggage . , . which you're carrying' (lines 11-12) and repeated references to it being 'so hard' (line 10 and line 11) invoke a sense of sustained effort and serve to counteract any impression of a decision taken lightly. Talk of 'investment' (line 13) and 'responsibility' (;ine 19) chime with a construction of breaking up as serious business. To summarize, part 1 of the extract uses language in such a way as to construct a version of decision-making which involves considerable effort and hard work. Such a construction of decision-making constitutes a warrant for the decision actually taken (that is, ending the relationship) because it removes any semblance of lightness or superficiality from the account. Part 2 (lines 27-42). The text accomplishes its sense of finality through its use of terminology and grammatical and stylistic features such as the use of metaphor, first, the use of the first person in assertions of the speaker's, perspective ('the way I saw if, line 30; 'as far as I was concerned', lines 41- ; 42) supports a singular and unambiguous point of view to which the speaker : has privileged access. The use of a question ('Would 1 want to marry him?' line 34) that requires a categorical answer (we cannot get 'a little biť married or choose to marry 'some of the time') also contributes to the finaiity of the extract; in the event, the 'answer was no' (lines 34-35), and this leaves no room for doubt or negotiation. References to the 'foundation' (line 31) and the 'basis' (line 38) of her decision to terminate the reiationship invoke a bottom line beyond which considerations cannot be made. This serves as a warrant for the finality of the decision. Finally, and most dramatically, the use of the metaphor in the last sentence (line 42) provides a visual image of the inevitability of the end of the relationship: 'it had hit the brick wall and it wasn't going any further'. By invoking the image of an object hitting a physical barrier, the speaker underlines the finality of her decision. There is no room for second thoughts or reappraisals because it is simply too late: the reiationship has 'hit the brick wall' and it cannot continue. To summarize, part 2 of the extract uses language in such a way as to ' construct a version of the participant's decision that is characterized by inevitability and finality. Such a construction of the decision constitutes a warrant for the decision taken (that is, to end the relationship) because it does not allow for the possibility of an alternative outcome. From a discursive psychology perspective, both parts of the extract t serve as a warrant for the participant's decision to terminate her relationship !; with her partner. However, two different constructions of the decision are :1 170 Qualitative Psychology _ produced in the same interview (thai is, as involving effort and hard work. í and as final and inevitable, respectively) which demonstrates some of the variability that characterizes discourse. A look at preceding sections of text (not reproduced here) can throw further light on the variable deployment of discursive constructions of decision-making within the interview. The portion of text which constructs the decision as the product of considerable effort on the part of the participant is produced in response to a question about the involvement of friends in the decision-making process (lines 20-22, I: 'And when you made the decision urn when you were actually working towards finishing it did you talk to friends about it?'). This question, in turn, is preceded by an account of how the participant's friends had 'taken a dislike' to her ex-partner and how they had 'talked about him with disdain'. As a result, the participant pointed out, 'everyone was glad when !'e..and its role in the constitution of social and psychological, life. From ajjoucauldian. point of view, discourses faci!itaíe..and.!imií,. enable.and constrain, what can be said, by whom, where and wh.en_.lPa.rker, .1992}.....F.oucauldia.n discourse anaiysts.focus upon the availability.of..discui3ive.resources.withina.culture -something like a discu.rsiv.e...econom.y--..and.its_.impiJc.a.tions_fo.r...those,who. live within it. Here, discourses may be defined as 'sets of statements that construct objects and an array of subject positions' (Parker, 1994: 245). These constructions, in turn, make available certain ways of seeing the world, and certain ways of being in the world. Discourses.offer subject.f>o$itions\~: 'i'..^. which, when taken, up,, have.im pli cations, for..subjectivity. and. experience. Forj..,,.. ■/. . ,. example, within a biomedical discourse, those who experience ill health;1 occupy the subject position of 'the patient', which locates them as the passive recipient of expert care within a trajectory of cure. The concept of positioning has received increasing attention in recent years (Harre and van Langenhove, 1999). Foucauldian.discourse^^^^ of, dis-;' CQ.urse.jn wider social.processes of..lcgi.tim.ati.on._and.power. Since discourses ..;..;■: make available ways of seeing and ways of being, they are strongly impli- v cated in the exercise of power. Dominant discourses privilege those versions of social reality which legitimate existing power relations and social structures. Some discourses are so entrenched that it is very difficult to see how we may challenge them. They have become 'common sense'. At the same time, if is in the nature of language that alternative constructions are always possible and that counter-discourses can, and do, emerge. Fojicauldiari. Oi'.Yu-::'';;.. discourse analysts.also take a historical perspective and. explore, the., ways J a '.-'VPa; .;;.• which.discourses have changed over time, and how this may have,shaped'; "'"" >"'! historical subjectivities (see aiso Rose, 1999). Finally, the Foucauldianj •/< \\<)ý version of discourse analysis aiso pays attention to the relationship between! ,-r-,...., discourses and institutions. Here, discourses are not conceptualized simplyi " 'J* as""ways of speaking or writing. Rather, discourses are bound up with! 172 Qualitative Psychology Box 8.3 Foucauldian discourse analysis was inspired by Foucault and post-structuralism is concerned wifh discursive resources explores the role of discourse in the constitution of subjectivity and selfhood explores the reiationship between discourse and power links discourse with institutions and social practices asks, 'How does discourse construct subjects and objects?' y\ -..m institutional practices - that is, with ways of organizing, regulating and administering sociai life. Thus, while discourses legitimate and reinforce existing social and institutional structures, these structures, in turn, also support and validate the discourses. For instance, being positioned as 'the patient' within a biomedical discourse means that one's body becomes an object of legitimate interest to doctors and nurses, that it may be exposed, touched and invaded in the process of treatment which forms part of the practice of medicíne and its institutions (see also Parker, 1992: 17). _expJorej;hejü.mphca.t»^ Stage 1: Discursive Constructions The first stage of analysis is concerned with the ways in which discursive / objects are constructed. Which discursive object we focus on depends on our research question. For example, if we are interested in how people talk about 'love' and with what consequences, our discursive object would be 'love'. The first stage of analysis involves the identification of,the different ways in which the discursive object is constructed in the text. It is important that we do not simply look for key words. Both implicit and explicit references need to be included. Our search for constructions of the discursive object is guided by shared meaning rather than lexical comparability. The fact that a text does not contain a direct reference to the discursive object can tell us a lot about the way in which the object is constructed. For example, someone may talk about a relative's terminal illness without directly naming it. Here, references to 'it', 'this awful thing' or 'the condition' construct the discursive object (that is, terminal illness) as something unspeakable and perhaps \ also unknowable. Stage 2: Discourses Having identified all sections of text which contribute to the construction of \ the discursive object, we focus on the differences between constructions. ! What appears to be one and the same discursive object can be constructed in very different ways. The second stage of analysis aims to locate the various discursive constructions of the object within wider discourses. For example, '* within the context of an interview about her experience of her husband's 174 Qualitative Psychology prostate cancer, a woman may draw on a biomedical discourse when she ialks about the process of diagnosis and treatment, a psychological discourse when she explains why she thinks her husband developed the illness in the first place, and a romantic discourse when she describes how she and her husband find the strength to fight the illness together. Thus, the husband's illness is constructed as a biochemical disease process, as the somatic manifestation of psychological traits, and as the enemy in a battle between good ■(the loving couple) and evil (separation through death) within the same text. Strige 3: Action Orientation The third stage of analysis involves a closer examination of the discursive contexts within which the different constructions of the object arc being deployed. What is gainedj'rom constructing theobject in! this.particular.way at.thísp_articuÍar.Roinl.wÁthin.the..text?,_WÍiat is its. function...a ndVhow.do.es j t relate to. other.constructions.produced in the surrounding text?.These questions are concerned with what discursive psychology refers to as the action orientation of talk and text. To return to our example of a wife talking about her husband's cancer, it may be that her use of biomedical discourse allows her to attribute responsibility for diagnosis and treatment to medical professionals and to emphasise that her husband is being taken good care of. Her use of romantic discourse may have been produced in response to a question about her own role in her husband's recovery after surgery and may have served to emphasize that she is, in fact, contributing significantly to his recovery. Finally, psychological discourse may have been used to account for her husband's cancer in order to disclaim responsibility for sharing in a carcinogenic lifestyle (for example, T told him to slow down and take better care of himself but he wouldn't listen'). A focus on action orientation allows us to gain a clearer understanding of what the various constructions of the discursive object are capable of achieving within the , text. Stage 4: Positioning;; Having identified the various constructions of the discursive object within the text, and having located them within wider discourses, we now take a closer look at the subject positions which they offer. Asubject position within a discourse identifies 'a location for persons within the structure of rights and duties for those who use that repertoire' (Davies and Harré, 1999: 35). In other wordšZ^ make .available, positions, within, networks .ol.rn.eaning which.speakers._can. take up (as well as place, others, with in). -Subject positions are. different from I roles in that they offer discursive.locations.from which to speak and act Discourse analysis 175 rather than prescribing a particular.pari to be acted out. In addition, roles can be played without subjective identification, whereas taking up a subject position has direct implications for subjectivity (see stage 6 below). Stsfse 5: Practice This stage is concerned with the relationship between discourse and practice, it requires a systematic exploration of the ways in which discursive constructions and the subject positions contained within them open up and/or close down opportunities for action. By constructing particular .versions, of the world,, and...by....posito.ning-.^ discourses limit what can..be..sai.d..and..done..I-'urthermore, non-verbal practices can, and do, form part of discourses. For example, the practice of unprotected sex can be bound up with a marital discourse which constructs marriage and 'its equivalent, the 'long-term relationship', as incompatible with the use of condoms (Willig, 1995). Thus, certain practices become legitimate forms of behaviour from within particular discourses. Such practices, in turn, reproduce the discourses which legitimate them in the first place, in this way, speaking and doing support one another in the construction of subjects and objects. Stage 5 of the analysis of discourse maps the possibilities for action contained within the discursive constructions identified in the text. Stsae 6: Subjectivity The final stage in the analysis exj).l„o.res...th.e.rela.ti.onsh.ip...between..discourse and subjectivity. Discourses make ayaHabje certain„waysof seeing the world and..cer.tam...w3ys of being..in...the..world.__T.hey construct social.as well.as psychological realities. Discursive positioning plays an important role in this process. As Davies and Harré (1999: 35) put it: Once having taken up a particular position as one's own, a person inevitably sees the world from the vantage point of that position and in terms of the particular images, metaphors, storylines and concepts which are made relevant within the particular discursive practice in which they are positioned. This stage in the analysis traces the consequences of taking up various .subject positions for the participants' subjective experience. Having asked; /questions about what can be said and done from within different discourses. (Stage 5), we are now concerned with what can be felt, thought and experienced from within various subject positions. ';