Volume 8, No. 2, Art. 3 – May 2007

Using the Technology of the Confessional as an Analytical Resource: Four Analytical Stances Towards Research Interviews in Discourse Analysis

Brendan K. O'Rourke & Martyn Pitt

Abstract: Among the various approaches that have developed from FOUCAULT's work is an Anglophone discourse analysis that has attempted to combine FOUCAULTian insights with the techniques of Conversation Analysis. An important current methodological issue in this discourse analytical approach is its theoretical preference for "naturally occurring" rather than research interview data. A FOUCAULTian perspective on the interview as a research instrument, questions the idea of "naturally-occurring discourse". The "technology of the confessional" operates, not only within research interviews, but permeates other interactions as well. Drawing on FOUCAULT does not dismiss the problems of the interview as research instrument rather it shows they cannot be escaped by simply switching to more "natural" interactions. Combining these insights with recent developments within discourse analysis can provide analytical resources for, rather than barriers to, the discourse analysis of research interviews. To aid such an approach, we develop a four-way categorisation of analytical stances towards the research interview in discourse analysis. A demonstration of how a research interview might be subjected to a discourse analysis using elements of this approach is then provided.

Key words: discourse analysis, FOUCAULT, technology of the confessional, naturally occurring data, research interviews

Table of Contents

1. Introduction

2. Constructing Research Interviews as Notably Unnaturally

2.1 The making of "naturally occurring" discourse

2.2 How "ordinary conversation" became the "natural" context for all interactions

2.3 Research interviews are researcher involved: Interview as truth-event

2.4 The technology of the confessional, the interview society and research interviews

3. Uses of the Research Interview in Discourse Analysis

3.1 Type I: Interview material analysed as interview discourse

3.2 Type S: Interview material analysed as the stimulated discourse of interest

3.3 Type R: Interview material analysed as reports of the discourse of interest

3.4 Type C: Interview material as a resource for constructing the discourse of interest

4. Some Features of the Interview as Research Instrument

4.1 Identity creation in the interview

4.2 Hierarchy in the research interview

4.3 The activity of the research interview

4.4 The research interview as an integrated whole

5. Exploration of a Researcher-Involved Interview

5.1 Type I analysis: Interview material analysed as interview discourse

5.2 Type S analysis: An illustration of the use of Type I analysis to support Type S analysis

6. Conclusions


Appendix 1: Jefferson Style Notation Used






1. Introduction

FOUCAULT's work has inspired and influenced various differing theoretical and empirical approaches to discourse analysis. One of these approaches in the Anglophone world is, rather imperialistically, referred (e.g. WOOFFIT, 2005) to as "Discourse Analysis" (DA). A central tendency in this approach can be traced back to the sociology of GILBERT and MULKAY (1984), through the social psychology of POTTER and WETHERELL (1987) to the work of authors like WOOD and KROGER (2000). Around this central tendency, and in tension with it, have been various approaches including the Critical Discourse Analysis of FAIRCLOUGH (1992), HARRÉ and GILLET's discursive psychology (1994) and BAXTER's (2003) Feminist post-structuralist discourse analysis. While accepting that the appellation DA in this sense is rather imperial we deploy it here because it captures a key feature of this Anglophone DA's approach: Its eclectic roots and on-going engagement with, indeed its absorbing embrace of, diverse approaches. The way DA has drawn on different sources has not been without its critics (e.g. SOYLAND & KENDALL, 1997). At times DA's understandings has been based on highly particular interpretations of the sources concerned. For example, the influence of FOUCAULT was initially traced in DA through a work by BURTON and CARLEN (1979) that GILBERT and MULKAY (1984, p.15) describe as seeing in official government publications "a language of class domination operating as an ideological legitimisation of the state". Nevertheless DA has attempted to develop its approach using FOUCAULT's work, drawing particularly heavily on a FOUCAULTian approach to the self (e.g. POTTER & WETHERELL, 1987, p.109). While trying to keep its synthesising nature by, for example, engaging with more critical approaches to understand context (WETHERELL, 1998), DA has also been keen to have "articulated a sustained empirical programme" (POTTER & WETHERELL, 1987, p.31). This orientation to detailed empirical study of contemporary interaction has meant DA has tended to draw heavily on the techniques of Conversation Analysis (CA). DA has neglected the historical aspect so central in FOUCAULT's approach (KENDALL & WICKHAM, 1999; HOOK, 2001). However, in one area where DA has been united, its suspicion of the research interview, historical argument has been atypically strong. We argue in this paper that DA's methodology concerning interviews can benefit from drawing more heavily on its FOUCAULTian roots. [1]

In the social sciences generally, the research interview is a commonly used and a theoretical celebrated method of research often seen as an inherently emancipatory research instrument (see BRINKMANN & KVALE, 2005 for a critical review). FOUCAULT (1978, p.60) warns against such romantic notions:

"one has to have an inverted image of power in order to believe that all these voices which have spoken so long in our civilization—repeating the formidable injunction to tell what one is and what one does, what one recollects and what one has forgotten, what one is thinking and what one thinks he is not thinking—are speaking to us of freedom." [2]

This critique is part of a more general FOUCAULTian critique of confession and its progeny, (including the interview) as made, for example, in FOUCAULT (1978, pp.57-73). FOUCAULT's warning against the emancipatory view is echoed in the more conversation analytical end of DA by an uncharacteristically historical but very influential argument within DA that condemned the interview as, " … an implicit endorsement of the contemporary culture, and [as a method that] does not offer a vantage point from which to question its taken-for-granted modes of reproduction" (ATKINSON & SILVERMAN, 1997, p.322). ATKINSON and SILVERMAN (1997) develop their critique of the interview society independently: the only reference to FOUCAULT in their argument is to the Birth of the Clinic where FOUCAULT (1973) is more concerned with the medical examination than confession. POTTER and HEPBURN (2005) consolidate the specific DA arguments against the current practice of the research interview in social science. Thus we can see that despite the diversity of DA it does indeed look at interview material in a way that is both unified and distinctive. These commonalities explain the unusual but unified DA suspicion of interviews. These commonalities also allow a meaningful discussion of DA and the research interview. [3]

Despite its suspicion of the research interview, DA work frequently uses interview data (examples include KNIGHTS & MORGAN, 1995; WIDDICOMBE & WOOFFITT, 1995; HARDY, PALMER & PHILLIPS, 2000; MCAULEY, DUBERLEY & COHEN, 2000; VAARA, 2002). So there is a tension within DA between methodological concern and frequent research practice. Using FOUCAULT's historical understanding of the confessional nature of the interview and an understanding of why interviews were seen as undesirable in DA's early history, we cast light on the use of the research interview. This combined with recent methodological developments (e.g. SPEER & HUTCHBY, 2003), together with an acceptance that the technology of the confessional is now all pervasive in society, lead us to urge a more pragmatic but more self-conscious use of the interview within DA. To achieve this we structure the rest of this article as follows: In the next section we consider what DA means by "unnaturalness" in its understanding of the interview, identify the component meanings of this "unnaturalness" and argue that the force of the term "unnatural" as pejorative in DA is somewhat anachronistic. Next we look at how, despite being labelled as "unnatural", the research interview is used in various forms of DA and we propose a four-way categorisation of the analytical stances towards the research interview in DA. We then take a deeper look at one way DA has looked at research interviews, namely, when DA has explicitly examined the research interview as topic (our "Type I"). The insights thus gained are used in the empirical section of the paper to show that Type I analysis of an interview is relevant to analysing an interview from a Type S analysis (a different type of discourse analytical stance in our categorisation). After this empirical demonstration, we draw out some implications of our arguments. [4]

2. Constructing Research Interviews as Notably Unnaturally

This section discuses objections DA has had to using research interviews. This involves initially looking at how the idea of "naturally-occurring" data became important in DA and in particular how CA provided a specific view of the natural interaction idealised as "ordinary conversation". Next we will look at how CA claimed that "ordinary conversation" provided a natural or "internally" constructed context independent from the analyst. As we move to look at the involvement of the researcher in the interview we see both the influence of CA on DA's caution about the interview and how FOUCAULT's conception of the truth-event can recast the issues in a more helpful way. As we conclude this section we see how FOUCAULT's "technology of the confessional" allows a more specific understanding of what is happening in research interviews, while revealing that the pervasiveness of the issues involved means that merely avoiding the research interview is rather futile. [5]

2.1 The making of "naturally occurring" discourse

As SPEER (2002) has illustrated "naturally occurring" data has been much prized in DA. However "naturally occurring" is a particularly unfortunate term since DA shows how many phenomena are artfully and therefore unnaturally constructed. FOUCAULT's work undermines many of the things we hold as most deeply naturally and urges us "to question over and over again what is postulated as self-evident, to dissipate what is familiar and accepted …" (FOUCAULT, 1988, p.265). For "naturally occurring" data, the FOUCAULTian approach points to the usefulness of history in questioning the term. [6]

Historically the term "naturally occurring" served to distinguish DA from both ordinary-language philosophy and mainstream linguistics. Though interested in "ordinary language", WITTGENSTEIN (1998, pp.89-90) and AUSTIN (1998, p.122) used hypothetical statements of their own construction. "Naturally occurring" signalled that DA (including empirical work drawing on FOUCAULT's archaeological investigation of possibilities e.g. KNIGHTS & MORGAN, 1991) was not just investigating what was possible to say but was interested in what was actually said. In linguistics the focus was on studying language rules and units (langue) which were systematic. In contrast parole (speech) or language use which was context specific was not considered to be systematic in any way (DE SAUSSURE, 1983, p.14; CHOMSKY, 1998, p.305). One of POTTER and WETHERELL's (1987) objections to traditional social psychology was that it restricted the natural variability in the language of people's accounts. When SACKS, SCHEGLOFF and others found themselves interested in "speech as it is actually used" they found that "[f]or a very long time, you could not get a 'license' to study ordinary interaction closely ..." (ČMEJRKOVÁ & PREVIGNANO, 2003, p.45). This licence to study naturally occurring language became a mark of both CA and DA. [7]

CA had at hand not only a "licence" but also a material technology to practice its study of "ordinary interaction". A technology of audio-recording was essential for the development of CA, as SCHEGLOFF admits "So, the fact of actually having the tape recorder as an available technology didn't determine anything. But it was almost certainly the case that without it we could never have had a field ..." (ČMEJRKOVÁ & PREVIGNANO, 2003, p.18). Another piece of technology that also played an important role in the development of CA was the telephone. As SACKS developed CA he found himself working a lot with the telephone

"… an object introduced into the world 75 years ago. And it's a technical thing which has a variety of aspects to it. It works only with voices … Now what happens is, like any other natural object, a culture secretes itself onto it in its well-shaped ways" (SACKS, 1995, p.548). [8]

Notwithstanding the assumption that a culture's "well-shaped ways" will not be too disturbed by this technology, the telephone was the median in which a lot of CA's "natural" conversations took place. Thus these "natural conversations" were worked on using an "only with voices" technology. [9]

The necessity of audio recording technology and the dominance of telephone "audio-only" conversation in the development of CA idealised the telephone conversation as the ordinary conversation. It became the standard of naturally occurring interaction. Anytime DA studied interactions it drew heavily on its CA influences so this ideal of ordinary conversation spread throughout DA generally. [10]

2.2 How "ordinary conversation" became the "natural" context for all interactions

The idea of how to deal with the context of any interaction is a key issue in DA. CA claims it provides an objective standpoint from which context can, at least initially, be handled. This solid point from which CA can get a context of an interaction, "… is not external, as in the classical imagery. It is internal to the object of analysis itself …" (SCHEGLOFF, 1997, p.185; italics in the original). While this has been a matter of some dispute (SCHEGLOFF, 1997; SCHEGLOFF, 1998; WETHERELL, 1998; BILLIG, 1999; SCHEGLOFF, 1999a, 1999b, 1999c), the potency of the ideal of ordinary conversation as the natural standard of interaction within DA has gained much from this CA claim. Thus in dealing with the disapproval of the interview as non-natural interaction it is useful to further examine CA's claims in this regard. We argue that that the solid point from which CA claims to analyse interaction is not completely "internal to the object of analysis", and necessarily imports (like all other approaches) prior theories about context. In the case of CA this prior theory depends on participants in an interaction orienting themselves to ideals of "ordinary conversation". [11]

CA's solid point is the idea that interactional participants orient to the context of generic conversational features: "… one can say that 'conversation' is talk-in-interaction produced by the participants' orientation to, and implementation of, the generic organizations for conversation …" (SCHEGLOFF, 1999a, p.412). It is worth quoting Harvey SACKS, the father of CA, at some length to get behind this idea of conversation as an orientation to a generic collection of invariant features that CA has identified:

"They [formal features of a conversation] are not features that vary with, e.g., the number of participants to a conversation. They hold for single conversations across whatever n participants there are for the conversation, and across changes in the personnel of a given conversation. They hold across various other compositional features of the parties. I don't know which features they don't hold across, but they seem to hold across things like: sex, occupation, political persuasion, etc. That is to say, it's not a peculiar feature of, e.g., male conversation or female conversation or female-male conversation that one party talks at a time and speaker change recurs. They hold also across types of conversations—argument, business talks, whatever else. They hold across the parts of a conversation—beginnings, middles, ends. They hold across topics. They seem to be, then, rather general, formal features of a conversation" (SACKS, 1995, p.34; italics in original). [12]

FOUCAULT's historical perspective suggests a feature that the formal rules of conversation might not hold across: the time across which an apparatus might change. That SACKS is disinclined to worry about such features is seen in his assumption that the world is largely unchanged by, for example, the introduction of the technology of the telephone: "Where what happens is that the object is made at home in the world that has whatever organization it already has" (SACKS, 1995, pp.548-549; our emphasis). FOUCAULT is wary of such unchanging givens:

"Moreover, this a priori does not elude historicity: it does not constitute, above events, and in an unmoving heaven, an atemporal structure; it is defined as the group of rules that characterise a discursive practice: but these rules are not imposed from the outside on the elements that they relate together; they are caught up in the very things that they connect; and if they are not modified with the least of them, they modify them, and are transformed with them ..." (FOUCAULT, 1972, p.127). [13]

Our argument is not that the norms of ordinary conversation do not form a relevant context from which many interactions can, and should, be usefully analysed. Such a position would ignore the genius of SACKS and SCHLEGLOFF in exposing the very relevant, but not only, context of ordinary conversation. Nonetheless, it seems that the insights yielded by CA provide a resource for the analysis of what is going on in interviews rather than an argument that interview data or any other data cannot or should not be studied because of its alleged unnaturalness. [14]

2.3 Research interviews are researcher involved: Interview as truth-event

Another dimension of the "naturally occurring" conception of discourse is the level of researcher involvement (POTTER & WETHERELL, 1995; SPEER, 2002). One strategy for dealing with researcher involvement is to aim for data that might pass POTTER's (2002, p.541) "dead social scientist" test: the interaction would have been the same even if the social scientist died before the interaction took place. The "dead social scientist" test is related to the idea of a discoverable universally available demonstrative truth that claims to have no relationship to the individual who produces it. Thus the "dead social scientist test" aims to make the researcher objective, however

"the supposedly universal subject of knowledge is really only an individual historically qualified according to certain modalities, and that discovery of truth is really a certain modality of truth; putting what is given as the truth of observation or demonstration back on the basis of rituals, of the qualifications of the knowing individual, of the truth-event system, is what I call the archaeology of knowledge" (FOUCAULT, 2006, p.238). [15]

While work has been produced near this "dead researcher" ideal (e.g. the clever use of consumer helpline data in TORODE, 1995), for much of DA however approaching the ideal of no researcher involvement has been problematic in practice, as even awareness of a recording device can have "unnatural" effects on interactants. SPEER and HUTCHBY not only confirm the presence of such effects but also provide a discursive analytical purchase on them:

"Instead of seeing the presence of a recording device as necessarily contaminating what would otherwise be a pristine occasion of real-world interaction … social scientists should investigate precisely what it is that participants are doing when they orient to being recorded …" (SPEER & HUTCHBY, 2003, p.317). [16]

Extrapolating the work of SPEER and HUTCHBY (2003) an alternative discourse analytical strategy can be discerned: to embrace the presence of the researcher in the interaction. The researcher can no longer be seen as entering into the interaction only in such a way as to reveal what was and would "naturally" be going on if the researcher were absent. Rather the researcher is inherently part of an "active interview" (HOLSTEIN & GUBRIUM, 1997, p.122). Such an approach can also adapt SPEER and HUTCHBY's (2003) analytical use of the recorder device as discursive resource by viewing the presence of the researcher as analytical resource. Clearly this "interactive social scientist" approach allows to DA deal with researcher-involved interviews. Even those who think that researcher-involved interviews are relied on too much, argue that it is important that any interviewer activity is considered in analysis (POTTER & HEPBURN, 2005). Here DA is taking account to some degree of FOUCAULTian concerns. [17]

A key issue, in preferring either the dead or interactive social scientist strategy, is how much one is prepared to depend on researchers' interpretations and constructions of their research. The dead social scientist strategy has advantages: The researcher's agenda is not injected into the interaction in hard to explicate ways. The resources of the "naturally occurring" actors whose "… own ability to artfully (and very helpfully) poke holes in each others' positions to reveal their constructed nature" (POTTER & WETHERELL, 1987, p.162) are not displaced or disabled by the presence of a researcher. The absence of the researcher from the studied interaction means analytical claims must be made without any apparent privileged insight into the interaction, free from any bias formed by non-transcribed impressions taken from the interaction. In contrast, with active researcher-involvement in the studied interaction, the analyst acts as interpreter twice: once as a participant in the interaction and secondly in the subsequent analysis of the interaction. The level of subjectivity of analysis might be said to have doubled. On the other hand, one might also say that the opportunity for the researcher to explore the discourse has also doubled. In such circumstances, SPEER and HUTCHBY (2003) along with HOLSTEIN and GUBRIUM (1997) give us an approach that can make that exploration rigorous. When the interviewer is also an analyst there is also a greater test, namely that the interpretation has to cohere with the non-textual impression formed of the interview by the interviewer-analyst. Of course this involves a level of trust in the researcher. This is a basic issue concerning the role of interpretation in DA. Some seem to wish to "delete the researcher" (to adapt POTTER & HEPBURN's 2005 phrase) from DA: In the dead social scientist strategy, any role for interpretation by the researcher that is not demonstrable in the transcript seems to be denied. We agree that arguing transparently from data is desirable. However there are many slips between cup and lip: analytical arguments are traced for the reader not from the interaction per se but from a researcher construction of it. Theoretically these constructions can never be neutral (ASHMORE & REED, 2000). For example, in practice the more detailed the construction of a transcript the more errors there are likely to be—and it is doubtful that such errors would be entirely unbiased. [18]

If one is prepared to recognise the interpretative and interactive involvement of the researcher, the interactive scientist strategy can yield rich insights from research interviews. These richer insights ought to involve a more explicit acknowledgement of, and critical reflection on the ineradicable role of the researcher in constructing research. DA provides good analytical resources available for such critical reflection on researcher-involved interviews. To use such resources we must acknowledge that an interview is in FOUCAULTian terms a truth-event, rather than an ever-present universally truth: "We could call this discontinuous truth the truth-thunderbolt, as opposed to the truth-sky that is universally present behind the clouds" (FOUCAULT, 2006, p.235). [19]

2.4 The technology of the confessional, the interview society and research interviews

Research interviews, because they are interviews, are said to risk uncritically reflecting "a wider cultural preoccupation with the interview and personal revelation as a technology of biographical construction" (ATKINSON & SILVERMAN, 1997, p.305). This "cultural preoccupation" seems to makes interviews particularly unnatural and so unsuitable for social scientific analysis in the "interview society". This implication seems to rely on the power of the interview society being peculiarly confined to the interview situation. Yet the kind of concerns that ATKINSON and SILVERMAN (1997) express regarding the interview society extend now to many other forms of interaction, as they themselves admit "Beyond the confines of the social sciences, the confessional voice is thoroughly characteristic of the interview society" (ATKINSON & SILVERMAN, 1997, p.314). So the personal revelation is not confined to the interview but pervades all our interactions: we have, "… become a singularly confessing society" (FOUCAULT, 1978, p.59). Our society is one where a multitude of interactional forms, contexts and technologies such as talk shows and web-cams have thrown off the "… bodice of the private-public divide" (VAN ZOONEN, 2001, p.672). A vast range of interaction types (see http://www.realitytvworld.com/ for an idea of the range) are the subject of the hyperreality of reality-television (BAKER & FREEMAN, 2003). GABLER (1998, p.234) notes that with video technology people, "… more than perform for the camera, people also began tailoring the major events of their lives to its demands, which were the demands of entertainment". So all of our interactions, not just interviews, are now subject to the technology of the confessional with "its obscure familiarity and its general baseness" (FOUCAULT, 1978, p.62). [20]

These effects, more general than the interview society, on all of our interactions can be treated as useful analytical resources. A generalisation of SPEER and HUTCHBY's (2003) approach would seem intellectually more sustainable than avoiding interviews and the increasing range of interactions "unnaturalised" by their elaboration in the media. [21]

All of the ways in which interviews have been considered "unnatural" have been concerned in some way with a supposed "corrupting" of the data. The argument in this paper is that all forms of interaction—including the great variety of interview forms—are "incorrigibly contaminated" by the fact of being the subject of research, by media constructions or by other public portrayals of similar interactions. Here one is reminded of the admonition that, "There is no point then in wanting to dismantle hierarchies, constraints, and prohibitions … as if the individual was something existing beneath all relationships of power, preexisting relationships of power and unduly weighed down by them" (FOUCAULT, 2006, p.56). As FOUCAULT contends it is with individuals, so it is with "natural" interactions. DA's strength lies in its sensitivity to the context of each interaction and the creativity of interactants, rather than as a mandate to search for a pure, uncontaminated form of interaction. [22]

3. Uses of the Research Interview in Discourse Analysis

Despite the prevailing theoretical mood within DA being against interviews, they continue to be used in DA. It is to an examination of these uses that we now turn. For this purpose it is useful to adopt the four-way categorisation (based on the analytical stance taken to the interview material) that we develop below. Each of these four analytical stances take as topic a discourse from the interview and so each can properly be termed an analysis of discourse. We find examples of each analytical stance in the discourse analytical literature. The four types of analytical stance do differ on the aspect of discourse that the instrument of the interview is used to produce. [23]

3.1 Type I: Interview material analysed as interview discourse

Talk within an interview constitutes discourse. This discourse can be analysed with the research focus on its nature as interview discourse (e.g. SLEMBROUCK, 2004, p.108). So "Type I" is where the interview interaction is viewed as the topic of interest in itself. This is because the researcher is interested specifically in what discourses do in interviews treated as interviews. Outside of DA such an interest in interviews as interviews is unusual (POTTER, 2002, p.541). However, within DA they are deservedly the first we chose to mention: Even the conversation analyst (who is wholly oriented to naturally-occurring discourse) could not resist an exposing of the mechanics of the technology of the confessional. This "Type I" treatment of interview data within DA will be the subject of greater examination in the next section. [24]

3.2 Type S: Interview material analysed as the stimulated discourse of interest

Interviews are usually about something and can be carried out in a variety of ways. Both the topic and the manner of an interview can be used to stimulate the production of discourse of interest to a particular researcher. Where this stimulated discourse is the focus of analysis of a research interview, we term this a "Type S" analysis of an interview. [25]

Interviews are occasions where some kinds of discourse of interest can be stimulated more easily than others. For example, DA researchers interested in the construction of identity might be well advised to choose interviews as their method of data generation. After all, the interview context seems to stimulate in particular the production of identity-building discourse (see the discussion of the nature of the research interview in the next section). This assumes the resources drawn on in interviews to produce identities have some commonality with the resources used on occasions where similar tasks are being performed (WIDDICOMBE & WOOFFITT, 1995, pp.73-74). Of course, there are some kinds of discourse that would be very hard to stimulate in an interview: Suppose the focus of the research was how issues of sexuality were constructed amongst young women. An interview between a male middle-aged interviewer and one such woman would be an interesting encounter. Material from this encounter could be usefully subjected to DA. However, it would be flawed material for a DA of how young women construct issues of sexuality, in the shared discourse of their own milieu. [26]

3.3 Type R: Interview material analysed as reports of the discourse of interest

A "Type R" analysis is where the interview material is treated as reports of other interactions: the topic of interest is the discourse of the interaction reported on. KITZINGER and FRITH (1999, p.296) make some observations on how women turn down requests from men for sex, from reports given subsequently by the women. MCGOWAN (2002), in her work on how the private domain is silenced within organisations, relied on reported discourse to explore how managerial discourses silenced managers' talk of their care-giving to elderly relatives. Both KITZINGER and FRITH (1999) and MCGOWAN (2002) provide insightful discourse analytical perspectives on the foci of their research, where the actual interactions were probably impossible to record. [27]

3.4 Type C: Interview material as a resource for constructing the discourse of interest

A "Type C" analytical stance is where interview material is used as a means to construct the "real" or "underlying" discourse of interest. Here the interview is not itself the topic of the DA, rather it is used as a resource from which to build the discourse to be studied. For example, RIGG's DA of managing in small business used "… a range of discursive practices, including logos, written signs, building layout, interactions within meetings, names of meetings and job roles, informal interactions, body language, conversations, as well as narrative accounts given to me in formal interviews" (RIGG, 2005, p.64). RIGG's readers are then presented with a highly-finished construction which sacrifices some textual transparency to the original interview data with greater insights based on the more nuanced interpretation of the researcher. Such work maximises the use of the researcher's ethnographic knowledge rather seeking to minimise or delete it. Even if there is representation of actual discourse, the aim of this type of research is to "… weave together direct field observation, interviews and systematic collection of textual data to … offer a collage—a mixture of production and reproduction that every reading and writing necessarily entails" (NG & DE COCK, 2002, p.27). So Type C uses interview discourse as a resource to construct a researcher-crafted discourse. It is this crafted discourse that then acts as the topic of the analysis. Despite the use of collected discourse as a resource from which to fashion the object of study these works claim to be engaged in DA. This claim is justified because the resulting crafted discourses constitute the topic of the analysis. Thus HARDY et al. (2000) show how the discourse typified by them as "localization" was used as a strategic resource in a non-governmental organisation operating in Palestine. [28]

Some might object that as we move from Type I to Type C we our getting further away from what should be considered DA. However we have already seen that there are works carrying the DA label that use interview material in ways that variously fall into all these four categories. There are works which do not fall within DA as we understand it: They apply conceptions of interview which are excluded from our four types of analytical stance. For example, MILLER and GLASSNER (1997, p.101) define their ideal interview as one in which the "… interviewees' subjective view as experienced by the interviewees themselves" is revealed. The issue here is that 'interviewees' subjective view as experienced by the interviewees themselves' seems to rely on a pre-constructed or even non-discursive concept of self. Yet the point of DA is to study how such subjectivities, relations and representations are constructed and used in discourse (for example, in the discourse of the interview interaction) rather than merely catch glimpses of underlying and unchanging entities like personality that existed prior to the discourse. [29]

4. Some Features of the Interview as Research Instrument

A key insight offered by DA into research interviews is they are not a transparent way into something beyond the interview but rather are incorrigibly and inherently interactional performances in themselves. Research interviews are about doing research interviews. Yet the problem remains as to what interviews do in detail and in particular. Interviews do not remain fixed and formulaic but are constructed interactively on each occasion. This means that DA researchers using interviews can usefully study the interactions produced as interview discourse (i.e. in a Type I way) whatever other interests they have in their data. It is possible to explore some features of interviews that might contribute to such analysis by drawing on the literature. To this task we now turn. [30]

4.1 Identity creation in the interview

In interviews the roles, subject positions, category membership or subjectivities of the participants are accomplished (RAPLEY & ANTAKI, 1998; LEE & ROTH, 2003). A lot of this work may be prompted by the interviewer for "... questioning is really a particular way of fixing the individual to the norm of his own identity … of pinning the individual to his social identity" (FOUCAULT, 2006, p.234). As the proponents of the active interview argue, "… there is always a working model of the subject lurking behind the persons assigned the roles of interviewer and respondent" (GUBRIUM & HOLSTEIN, 2003, p.30). Even innocent-looking terminology is likely to have some assumptions "lurking" within, for example, the grammar of the terms "interviewer" and "interviewee" imply that the former is implicated as more powerful and active (MISHLER, 1986). Other identity work is done by labels such as respondent, informant, client, subject and story-teller. This necessary labelling is only an example of the identity work that continues throughout an interview. [31]

4.2 Hierarchy in the research interview

Writing of the confessional, FOUCAULT (1978, p.62) pointed out that "… the agency of domination does not reside in the one who speaks (for it is he who is constrained), not in the one who knows and answers, but in the one who questions and is not supposed to know". So research interviews can "… go on—at the extreme—to subordinate that view [of the interviewee] to the interviewers' " (RAPLEY & ANTAKI, 1998, p.605): The interviewer comes to the interview with a greater knowledge of the interview's ultimate uses than the interviewee. In interviews particulars told by the interviewee are often generalised by the interviewer, thus positioning the interviewer as the one in control of the general categories. In the very arrangement of an interview the expectation is created that the interviewee has material of interest for the interviewer. Thus the interview can become a test of the interviewee's competence on the topic of interest with the interviewer positioned as the grader in that test. While these insights are useful we should acknowledge that the interview participants are unlikely to always remain dupes who passively enact their allotted roles in a predefined order determined by the interviewer. Careful analysis of how the context of the interview is used by participants is needed on each occasion. [32]

4.3 The activity of the research interview

In analysing the discourse of a particular information system, (e.g. DOOLIN, 2003; RÄISÄNEN & LINDE, 2004), or a newspaper's front page (e.g. KRESS & LEEUWEN, 1999), one sees the finished product of a particular discourse rather than its active construction. In researcher-involved interviews both the interviewee and interviewer are engaged in this active construction. In contrast to the views of the interviewee in experimental and information-gathering perspectives, the perspective of the "… active interview eschews the image of the vessel waiting to be tapped in favour of the notion that the subject's interpretative capabilities must be activated, stimulated and cultivated" (HOLSTEIN & GUBRIUM, 1997, p.122). There are preferred topics, roles and formats but these also develop and change in interaction. The data produced is not so much collected as negotiated (FONTANA & FREY, 2000). Both interviewer and interviewee are skilled and active and "seem to have routine knowledge of the rules of interviewing, with no need for instruction" (FONTANA, 2003, p.51). Thus we can use the "technology of the confessional" as a resource to enable and build interaction. Within this interaction, both interviewee and interviewer can adopt different positions, have (and negotiate) agendas as well as draw upon a range of discursive resources. [33]

4.4 The research interview as an integrated whole

An advantage of the interview is that it can be recognised by participants as an integrated whole, with time to balance one remark with another. It is a special occasion—an interaction within recognisable boundaries. This is not to deny that the interview itself has a wider context. Rather what we are arguing is that an interview can create for the interactants a recognisable element of context in itself. This means that as BRIGGS (1986, pp.102-103) has pointed out "the communicative structure of the entire interview affects the meaning of each utterance" (italics in original). This needs to be taken into account in its analysis, but it also means that the interview may allow the emergence of discourse that might be hard to capture in more naturally-occurring data. In one-to-one unstructured interviews, interviewees generally have considerable control over the context in which discourse is produced: their meanings are not so dependent on the unexpected turns of other conversational members as they might be in an interaction of three or more, or in more time-indefinite snatches of conversations. Discourse can thus emerge in a one-to-one interview that might emerge only rarely and in hard to achieve conditions "in the wild". [34]

The Type I use of the interview in DA was discussed in this section. We now have various insights into how participants can draw on the interview genre in interactions constructed as interviews: Interviews can be used for the construction of identity. Interviews are interactions where the genre itself may facilitate the emergence of hierarchy between interviewer and interviewee. Interviews are occasions of active interaction where the participants not only have many resources to draw upon in our interview society but also have the potential to do so quite creatively. Finally interviews can be used as occasions to portray more complexities than might be wise in situations where an interactant is less sure of the duration of the interaction. As we turn now to our empirical exploration of some of these ideas we will see that this creativity and activity of the interview is quite apparent so that a Type I analysis is useful for each particular interview. [35]

5. Exploration of a Researcher-Involved Interview

Below we present extracts from an interview conducted by the first author as part of a larger research project where the primary method of data collection was interviewing managers in small Irish firms. The extracts are the kind of material that would often be discarded in the analysis of interviews, yet are typical of the interviews conducted. We use the extracts to do some analysis of the kind we have labelled above as Type I and Type S. Some of the issues of identity creation, hierarchy and general "doing" of the interview can be seen in the interpretations and we also note how the kind of discourse stimulated (in particular enterprise discourse) can be seen in the doing of the interview. [36]

Paul (a pseudonym) is the owner-manager of a small firm. He has founded and sold a number of other businesses. Brendan is the researcher and full-time academic. Brendan and Paul had met each other previously but were not well acquainted. The e-mail sent to Paul asking for the interview referred to Brendan's interest in "the thinking of practicing managers" particularly in "the messy and challenging situations of small firms", and declared that the research project involved "unstructured interviews with directors of small to medium-sized enterprises". [37]

Turn numbers are given for ease of reference, and where turns from the interview itself have been omitted the timing of particular extracts within the overall interview are noted. The extracts include some set-up interaction (turns 1 to 9). A Jefferson-type transcription notation (ATKINSON & HERITAGE, 1984) is used and detailed in Appendix 1. Now that we have provided some information on the context of the interview we will proceed first to carry out, in terms of our categorisation of analytical stances, a Type I analysis of the interview before illustrating how insight from this Type I analysis can aid Type S analysis. [38]

5.1 Type I analysis: Interview material analysed as interview discourse

Extract 1 below shows the opening of the interaction. As we can see the conversation is initially orientated to Brendan's digital recording device. At the time this (an iRiver H340) was a relatively new kind of device which combined substantial storage (40 gigabytes) with the ability to record audio and other data in a small and stylishly designed palm-sized unit. The iRiver brand was a well-respected brand with greater functionality than the best selling alternative –the Apple iPod. During turns one to nine, there is much overlap between turns with a very light-hearted tone to the exchanges. When Brendan indicates that he is starting the interview proper by saying "'So yeah em, so how did you get here?' is my usual start," the tone changes with the speech being slower and more formal. Paul is ready with his story, immediately positioning himself as a man of experience in the sector concerned: "I am twenty years in [name of sector]" (Turn 11). In these exchanges we see that the interview is being constructed by the participants as an "occasion", compared to the more casual encounter of the initial turns.


((The section of the interview below is at the start of the 87 minute interview recording))


Brendan: Fine ya. That's great iRiver


Paul: >Oh this one!<


Brendan: 40 gig It is the cheap, the mean version of the iPod [ha ha


Paul: Ya] We've actually a review in for it for next one


Brendan: I must say I've found it very good. Very, very happy. It's got the external mike which the


Paul: ([ )


Brendan: ((Laughs))] for this kind of thing [anyway.


Paul: ( )]


Brendan: So yeah em, so how did you get here? is my usual start, so.


Paul: I'm twenty years in ((name of sector)).


Brendan: Okay.


Paul: And I fell into it by accident twenty years ago, when I was asked to do a survey on ((name of sector)). On behalf of the ((well known research company)).


Brendan: Oh the ((well known research company))


Paul: Yeah on ((street name)), yes. So I was in there for three months. So I was running around like an eejit ((Hiberno-English corruption of idiot)), asking ((producers in the sector)) all sorts of weird questions and then some poor unfortunate offered me a job.


Brendan: ((Laugh)). And who was that with? .


Paul: ((name of firm))


Brendan: Right () [very much sectoral


Paul: So I] went from there to, eh I worked for a couple of years then in Dublin. I went very quickly from selling ((sector products)) to actually ((producing)) the whole thing myself at the age of about twenty-two

Extract 1: An Extract from the Interview Showing Turns 1 to 19 [39]

That the interview is oriented to as an integrated whole is seen in the way one remark in this interview is best understood in the context of another. For example, in turns 13-15 (see Extract 1) Paul makes some self-deprecating remarks referring to himself as being like an idiot. These remarks taken out of context might portray Paul as a victim of happenstance but he has already told us that he has 20 years experience and that he worked for a well-regarded research company. Furthermore the self-deprecating remarks are followed immediately with turns 16-19 (Extract 1) where Paul tells of his rapid promotion within the firm. It is hard to imagine Paul risking the self-portrait of Turns 13-15 in an interaction that he didn't know was going to go on for sometime. The interview society means that the interviewee recognises the interview as an occasion of extended interaction—an integrated whole. Paul orients to the interview as an integrated whole—allowing him to make one remark in the context of another. This kind of orienting occurs frequently in the interviews conducted in the course of this research. [40]

Extract 2 begins with Turn 20 which is about 20 minutes into the 87 minute interview, the last turn of Extract 1 having ended at about 1 minute 18 seconds into the interview. In Turn 24, Paul assures Brendan that he—Paul—is engaged in "telling warts and all here" i.e. he has included in his account a full description including perhaps uglier bits of the story which, like warts, would be hidden in a purely flattering portrayal. Paul is performing his role as a co-operative interviewee. He certainly has made time for Brendan's interview and tells of mistakes he has made—his idiotic "running around" for example. To competently use the technology of the confessional the performance requires such a "warts and all" approach.


Paul:<I think that's something else, I think em I think that team development is (.) eh it's a an ongoing process (.) it's a (.) certainly for anybody running the company.> >Tell me more about the the objectives from all this < [( )



Brendan: Sure yeah] yeah, I realise I've kinda lured you in [to a conversation.



Paul: No]



Brendan: [about eh



Paul: I'm I'm] just telling warts and all here



Brendan: Sure, and that's that's really useful to me. Em, I kind of started off in being very interested in, I suppose how



Paul: ya, ya



Brendan:that we developed within ((name of third level institution))



Paul: ya



Brendan: Em and gradually that came to being well, let's find out what practitioners use at the moment.



Paul: [ya



Brendan: And] I was appalled in the literature,



There's very little [description really.



Paul: ya ya]



Brendan: Em, there's all sorts of time studies



Paul: [ya



Brendan: a]nd things like that and then there's kind of, what I'd call lab work, where you know we ask people like yourself to do case studies which seems [a little silly



Paul: ya]



Brendan: given [ that you are living your own case study.



Paul: ya (.) ya]



Brendan: Em so I came to the conclusion that a good way to study what was going on eh in terms of people’s em (.) thinking about their business, and I, I decided the word strategy



Paul: [Ya



Brendan: was ]a nice loose [ambiguous one



Paul: Well I can talk] I'll talk to you about strategy in a minute


Brendan: Sure ya


Paul: but go on go on.

Extract 2: An Extract from the Interview Showing Turns 20 to 45 [41]

We saw in Section 4.1 above that DA has pointed out how interviews are occasions for building the identity of the participants. What then is the identity being constructed in the account that Paul is producing for Brendan? The portrayal of turns 13 to 15 demonstrates that Paul is being very co-operative as an interviewee and that he is a competent and amusing interviewee. This is certainly one element of the identity Paul is building in this interaction. Below in Section 5.2, as we take a Type S analytical stance, we will see that awareness from this Type I analysis of the mechanisms by which Paul builds his "competent interviewee" identity, allows a clearer understanding of how Paul is simultaneously building an entrepreneurial identity. [42]

Turn 20 and the turns that follow it are typical of the interactions in this set of interviews which show that interviewees don't always accept the passive position into which being an interviewee may have cast them. In this interaction we can see that Paul takes over from the interviewer the role of generalising: At Turn 20, Paul declares "<I think that's something else, I think em I think that team development is (.) eh it's a an ongoing process (.) it's a (.) certainly for anybody running the company.>" Generally the interviewer is made more powerful by being the one who generalises the experience of the interviewee (RAPLEY & ANTAKI, 1998). Here, however, Paul does the generalising in his comment that "(.) it's a, certainly for anybody running the company". That the preferred manner of an interview is that it is interviewer-led is evidenced by Paul's hurried follow-up ">Tell me more about the the objectives from all this, ()<" (again Turn 20). This question not only signals an acceptance that something may be out of place in Paul's generalisation: The repetition of the word "the" could be read as a hesitation giving control of the interview back to Brendan, as also could the invitation to Brendan to explain what the interview is about. The interviewer—perhaps somewhat defensively (Turn 21) attempts to repair his position: "Sure yeah yeah, I realise I've kinda lured you in to a conversation". Brendan here asserts that he "realises" exactly what is going on and the word "lured" could imply that it was all expertly planned by him. Soon (Turn 27) Brendan is building on his institutional position "—that we developed within ((name of local higher education institute))", showing his expertise by being in such a knowledgeable position as to be "appalled in the literature" (Turn 31) and labelling bits in a familiar manner: "all sorts of time studies" (Turn 34) and "what I'd call lab work" (Turn 36). The interviewer stresses his power to decide and define "… I, I decided the word strategy" (end of Turn 40) and then leaps over an interruption from the interviewee to finish his clever characterisation of the word "strategy" as "a nice loose [ambiguous one" (Turn 42). This last quip perhaps positions Brendan as an intellectual who knows what he is about. Paul, the interviewee, then inserts (notice the overlap in talk between Turns 41 and 42, and in Turns 20 to 42 in general) an assertion of himself as both an expert practitioner and as the participant who controls the timing of events in the interview with " Well I can talk] I'll talk to you about strategy in a minute". Brendan signals his acceptance of this with "Sure ya"—which is implied as necessary since he is the interviewer. Paul reasserts his power by directing Brendan to continue his explanation: "but go on go on". It is hard to characterise Paul as passive in this interview and this is further illustrated in the next extract which follows immediately from Extract 2.


Brendan: em was simply to go out and ask them to see what kind of words they use in discussing that. … ((Paul’s mobile phone rings and he signals he's taking the call)) Sure ya no worries.


Paul: Just turn it off for a second, is that OK


Brendan: Em , Sure [ya


Paul: Hello ((on the phone))]


((Recording is halted for 2/3minutes phone call))


Brendan: Yes that's great ((laughs)) >So em as I was saying it was just in terms if I go out and actually talk to people I go out and actually talk to people about their [business<


Paul: mm]


Brendan: they might show me how they're thinking about it, in their in their talk. So that's why I'm (.)


Paul: Yep


Brendan: happy to follow your direction, if you like even though that might seem a little aimless at times.


Paul: If the (.) objective is about you know what is the modus operandi


Brendan: Mmmh


Paul: of entrepreneurs or business people


Brendan: Mmmh


Paul : and and what are the watch words they use


Brendan: Yeah


Paul :what are the touchstones they have how is it they [think


Brendan: Yeah]


Paul: how is it they sort of organise their mindset


Brendan: Mmmh


Paul: Em I have to say, I'm em (.) extremely well organized


Brendan: Okay


Paul: Extremely well organised and I put a lot of (.) eh, eh emphasis on planning


Brendan: Okay


Paul: Okay Eh here's something about planning, if you have a business and you don't have a plan, you will fail


Brendan: Right


Paul: I guarantee you (.) I guarantee ya (.) If you have a business (.)and I don't care how bad the situation is, I don't care, give me any business, right, it could be in the absolute shite. Right. If you have time to plan, there's a way out

Extract 3: An Extract from the Interview Showing Turns 46 to 72 [43]

Extract 3 begins as Brendan gets over his initial hesitation to expand on an explanation but Paul's mobile phone rings and interrupts him. Paul then instructs (Turn 47) Brendan—"Just turn it off for a second > is that OK<"—to stop the recording device. Notice that "is that OK" is said faster than the earlier bit of the talk. Note that there is no raising of intonation at the end of "is that OK": It seems that "is that OK" is more of an instruction than a question. Unlike for another mobile phone call later in the interview (not shown) Paul completes the phone conversation rather than excuses himself to call back later. In Turn 53 Brendan casts his rather long contribution (stretching back to before the telephone call) as an explanation with a "so that's why". Brendan has accepted that Paul is entitled to ask for and receive explanations. In this interview the interviewer is not the only one who can ask questions. The negotiation of a more active role for Paul is completed with Brendan's declaration in Turn 55 that he is happy to follow Paul's direction, though that a norm is being breached is acknowledged by his "even though that might seem a little aimless at times". With a more active role acknowledged as being in the interests of the interaction, Paul gets busy in constructing himself as an active and in-control "interviewee". This can be seen from turns 56 to 64 where there is a careful specification by Paul—with facilitation by Brendan—of what this bit of the interview is about. [44]

In a way this Type I analysis generates some interesting questions. Why is Paul (the interviewee) asserting himself so much? Why is Brendan engaging in such work to assert his interviewer position? Why is the interaction not running as we might expect a smooth interview to run? Thus Type I analysis generates some questions that provide an opportunity for Type S consideration. [45]

5.2 Type S analysis: An illustration of the use of Type I analysis to support Type S analysis

Type S analysis in our categorisation is where the analytical stance is focused on the discourse stimulated by the interests of the particular researcher. In this case an effort was being made to stimulate the talk of practicing managers and directors in small to medium-sized firms with a view to seeing how they drew on strategy discourse. As the project proceeded it was also clear that these interviews also stimulated enterprise discourse. [46]

Looking at Extract 1 from a Type S analytical stance we can see that Paul is not only performing as a good general interviewee (as our Type I analysis revealed) but he is also responding to Brendan's particular research interests. Brendan's interests have been revealed to Paul as being in "the thinking of practicing managers" particularly, in "the messy and challenging situations of small firms" (from the e-mail sent to Paul asking for the interview). Paul's account of being "twenty years in" the sector (Turn 11) and "of running around like an" idiot (Turn 15) but of also going "very quickly from selling ((sector products)) to actually ((producing)) the whole thing myself at the age of about twenty-two" (Turn 19), serves to build Paul's identity as one familiar with the messy situations of small firms but also as one who has the competence to running "the whole thing". Thus in the overall context of the interview turns 13-15 could be interpreted as acting in the manner of what rhetorics calls a procatalepsis or what American political commentators call a prebuttal: an anticipation of a counter-argument to Paul's overall tale of being in charge of his destiny. We can see that what Type I analysis revealed as a confessing of "warts and all" to meet the requirements of a good interview generally, Type S analysis shows that the specific content of those admissions and confessions mean that they also function to meet the particular interests of the research interview. [47]

Type I analysis of Extract 2 produced some interesting research challenges in the sense that the interaction here did not seem to function easily as an interview. Paul seemed to resist the passivity in which his role as interviewee seemed to cast him and Brendan too worked hard on his identity as the competent interviewer. What was happening here? Type S analysis provides a possible interpretation: We run into trouble at Turn 20 when Paul while seeking to perform his role as a good interviewee is also at work on his identity as somebody "running the company" (Turn 20). Unlike in Extract 1, the role of small business practitioner cannot simply be managed by the specificity of the confessions needed for managing the role of interviewee. Here the stimulated role of the strategist and entrepreneur clashes with the passivity required of a good interviewee. After all discourses of strategy and entrepreneurship tend to cast their subjects as quite active (KNIGHTS & MORGAN, 1991; CARR & BEAVER, 2004; GREY, 2004). Thus the puzzling activity revealed by Type I analysis can be interpreted in Type S as the effort to establish the stimulated identity of strategist revealed in the, "Well I can talk] I'll talk to you about strategy in a minute", of Turn 40. [48]

As we move into Extract 3 our Type I analysis concluded that there was a successful completion of the negotiation of a more active role in the interaction for Paul. Type S analysis suggests that this negotiation was needed to allow the stimulated discourse to emerge more starkly: Paul can thus talk authoritatively "of entrepreneurs or business people" (Turn 58). Paul can reveal that he—now as presumably an acknowledged member of this very active and in control group—is "extremely well organised" and puts a lot of emphasis on planning (Turn 68). His negotiation of a more active role in the interview means there is no clash now between being a good interviewee and expressing the power of a strategist who can be given "any business, right it could be in the absolute shite" and, given enough time to plan, will be able to find a way out. Type I analysis was productive in generating a research question the answer to which supports the interpretation of Type S analysis. [49]

At the beginning of the interview the stimulated discourse of business practitioner could be achieved in a way that was compatible with being a confessing and good interviewee. Paul merely used the particular content of the confessions to illustrate his experience as a business man in the messy real world. When the interview role gets in the way of the stimulated discourse Paul resists the passivity into which the interview genre casts him. Paul's assertions might thus be explained through Type S analysis as being more consistent with the active positions assigned to him by the discourses of strategy and enterprise, than the passive role assigned to him by the form of the interview evidenced by Type I analysis. Thus a close Type I analysis of an interview may throw up potential evidence of what else is at work in an interaction: Type I analysis helps Type S analysis. [50]

6. Conclusions

In this paper we have examined how DA—a particular approach developed from FOUCAULT's work—has come to regard the research interview with particular suspicion. The form of this suspicion we traced from another of DA's main influences—the work of Harvey SACKS: CA developed "ordinary conversation" as the ideal of "naturally occurring data". DA came to contrast this naturally occurring ideal with the unnaturalness of the research interview. Using FOUCAULTian insights we argued that this attempt, to find an instrument of research that would yield "… the primordial form of talk-in-interaction …" (SCHEGLOFF, 1999a, p.407) existing independent of any social scientist, is futile. The theoretical perspective of the researcher has power in creating knowledge and no instrument of research can avoid this: "It seems to me that power is 'always already there', that one is never 'outside' it, there are no 'margins' for those who break the system to gambol in" (FOUCAULT, 1980a, p.141, italics in original). Thus our interpretation of FOUCAULT does not mean that the research interview can be viewed naively as a transparent window on unsullied interaction. Rather FOUCAULT in examining the "technology of the confessional" gives us insights into how the instrument of the interview operates. [51]

To illustrate the various ways in which the interview is used as an instrument in creating knowledge within DA, we constructed a four-fold categorisation of analytical stances within this tradition: Type I analysed the interview in terms of its interactional form, Type S analysed the interview as an occasion where particular discourses were stimulated, Type R used reports within the interview of discourses heard outside the interview and Type C used the interview as a resource to help construct what the researcher thought to be the underlying discourse of the situation. We also used this categorisation of analytical stances to argue that knowledge of how the interview operates as interview discourse, produced by Type I analysis, can provide insight into the other three ways the interview can be analysed within DA to produce knowledge. Understanding the interactional form of the interview allows the analyst to understand more of what is going on in the interview, regardless of analytical stance. Type I analysis can also produce insight into how interviews, conducted for research or other purposes, operate as one of the inheritors of the "technology of the confessional", within contemporary society. [52]

Our understanding of FOUCAULT's technology of the confessional, including its ability to recast some of the arguments presented by DA's suspicion of research interviews, allowed us to expose some general features of the interview. A key feature of the interview is the way it acts as instrument of identity creation with the interviewer's questions and expectations demanding the creation of particular subjectivities. The interview situation is one whose history casts the interviewee as the one lower in the hierarchy with the interviewer setting the agenda. Research interviewers continue this tradition with power operating in many subtle ways despite efforts to make this instrument purely emancipatory. However there is room for creativity in the active nature of the interview. Furthermore the structure of the interview—its extended interaction, its negotiated ending—can facilitate the creativity of the interview provided its interpretation takes account of this integrated nature of the interaction. While these general features of the interview can be useful, to see how the interview actually operates in any particular situation demands an empirical examination of that particular event. [53]

In the empirical section of the paper we conducted an analysis of a research interview from two different analytical stances: what we have called Type I (interview material analysed as interview discourse) and Type S (interview material analysed as the stimulated discourse of interest) analysis of a research interview. Using a Type I analytical stance we have seen above how our interview was an active occasion for both interviewee and interviewer where matters of identity and hierarchy are worked on. In our interpretation, we argued that within the occasion of the interview, Paul constructed a "warts and all" identity of himself as someone very much in charge of his destiny. The "warts and all" aspect of the interview seemed to be a necessary part of its use as an instrument to create the appropriate identity. The price of being allocated a particular identity is to make a good confession—an honest exposition of errors. The individual is "… authenticated by the discourse of truth he was able or obliged to pronounce about himself" (FOUCAULT, 1978, p.58). Although not the focus of our analysis we could see too that our interview acted as an instrument that also constructed an identity for the interviewer. Brendan produced wordy explanations for the manner and functioning of the interview that reinforced his institutional identity as an academic. An interesting study might be made in this regard of the academy's present day equivalent of the penitential manuals (the methodological guides) that might detail a greater variety of positions like the ones that FOUCAULT exposed in his study of sexuality. [54]

Being able to draw on evidence from an interview—as an integrated whole—allows one to get a better understanding of the context in which a particular discourse is deployed. Here the problematisation, by Type I analysis, of the activity and the assertiveness in the interaction, gave extra weight to the Type S analysis's claim that the stimulated discourse was one which cast the interviewee in a particularly active role. The strategy-enterprise discourse may be the reason for this particular interviewee's assertiveness in a genre of interaction (the interview) that would typically position him as passive. Thus we can see that the interview as a research instrument can act as an analytical resource. This is analogous to SPEER and HUTCHBY's (2003) claim that orientation to the presence of recording devices can be treated as an analytical resource when analysing recorded conversation. Research which does this has the potential not only to aid all discourse analytical uses of interviews but also to aid a more rigorous use of the research interview more generally. [55]

The analyst has a lot of text from each essentially discreet interview from which to construct an interpretation of the interaction. Without reproducing the whole text of the interview, readers are reliant on our integrity as researchers to fairly represent this integrated whole. For example, readers need to trust that researchers will avoid at least the intra-interview version of what SILVERMAN (2001, p.34) terms anecdotalism. An interviewer-researcher has a further test of an interpretation: its coherence with the interviewer's holistic understanding of the interview. Much of this overall understanding will remain tacit. Analysis of interviews is therefore incorrigibly interpretative. The importance of interviewer's interpretations cannot and should not be deleted from the research. Always "… the revelation of confession had to be coupled with the decipherment of what was said" (FOUCAULT, 1978, p.66), but this remains true whether the confession takes the form of a research interview or any other interaction of the confessing society. Interpretation remains an inherent part of the production of truth from the interview as an instrument of research: "If one had to confess, this was ... because the work of producing the truth was obliged to pass through this relationship if it was to be scientifically validated" (FOUCAULT, 1978, p.66). In the discourse analysis of interviews (and indeed other interactions) it is important to be honestly reflexive about researcher interpretation of context, while not using it as an excuse for failing to make the trail from data to conclusions as transparent as possible. [56]

Deeper understanding of any discourse genre such as the research interview should not lead to its avoidance but rather its more judicious and informed use. As one of the earliest authors to critique the interview puts it: "I am not trying to persuade researchers to abandon interviewing altogether" (BRIGGS, 1986, p.4; italics in original). The interview should not be abandoned as some sort of unnatural abomination whose artificiality makes it too self-conscious and reflexive an interaction for the natives and thus an improper object of sociological inquiry. Extreme condemnations of interviews could even blind us to the more general lessons of the interview critique and the consequences of a confessing society for research methods, for no genre of interaction is untouched by the reflexivity of our age. The trick is not to find a purely natural interaction that can be analysed objectively but to understand how the discourse in the interaction is produced. As FOUCAULT (1980b p.118) puts it:

"Now I believe the problem does not consist in drawing the line between that in a discourse which falls under the category of scientificity or truth, and that which comes under some other category, but in seeing historically how effects of truth are produced within discourses which in themselves are neither true nor false." [57]

When using interview data we have rich sources of methodological reflection on the relevant issues. Both methodologically and in the life-world we, as researchers and researched, have the resources to know the interview more than in many forms of interaction. This article has sought to get to know the interview better—engaging with, rather than abandoning either the interview or its critique. Rather than seeking an unsullied natural interaction from which they can feel distanced and removed, researchers should work with our hard-won knowledge of the technology of the confessional when analysing interview data. The four-fold categorisation of discourse analytical stances we have proposed in this article allows DA researchers to do just that. [58]


The authors would like to thank the editors and reviewers of FQS for their very helpful and constructive comments during the redrafting process. Additionally members of the Discourse Analysis Group at Dublin Institute of Technology also made many helpful contributions on early drafts of the paper.

Appendix 1: Jefferson Style Notation Used1)




A brief but noticeable pause.


A stopping fall in tone


A continuing intonation


A rising inflection

(# )

A timed paused where # is the number of seconds


Enclosed speech was delivered more quickly than usual


Enclosed speech was delivered more slowly than usual

Deliberating excluded talk within a turn


Square brackets enclose overlapping speech


Shouted or increased volume speech

Underlined text

Speaker is stressing the underlined speech

( text )

Unclear speech in the transcript, enclosed speech is the transcriber's best guess

((text ))

Enclosed text is a report of non-verbal activity, deliberate replacement of part of speech or an inserted clarification


1) This transcription notation is based on the notation developed by Gail JEFFERSON as described in ATKINSON and HERITAGE (1984). <back>


Ashmore, Malcolm & Reed, Darren (2000). Innocence and nostalgia in conversation analysis: The dynamic relations of tape and transcript [45 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 1(3), Art. 3, http://www.qualitative-research.net/fqs-texte/3-00/3-00ashmorereed-e.htm [Date of Access: 12/30/2005].

Atkinson, J. Maxwell & Heritage, John (1984). Transcription notation. In J. Maxwell Atkinson & John Heritage (Eds.), Structures of social interaction (pp.ix-xvi). New York: Cambridge University Press.

Atkinson, Paul & Silverman, David (1997). Kundera's immortality: The interview society and the invention of the self. Qualitative Inquiry, 3(3), 304-25.

Austin, John Langshaw (1998). Performatives and constatives. In Maria Baghramian (Ed.), Modern philosophy of language (pp.110-123). London: J.M. Dent.

Baker, Fiona & Freeman, Olivia (2003). Big Brother: A postmodern consumption experience. Presented at the 17th Nordic Conference of Business Studies (NFF Scandinavian Academy of Management), 14th to the 16th of August 2003, University of Iceland, Reykjavik.

Baxter, Judith (2003). Positioning gender in discourse: A feminist methodology. New York: Palgrave MacMillan.

Billig, Michael (1999). Whose terms? Whose ordinariness? Rhetoric and ideology in conversation analysis. Discourse & Society, 10(4), 543-558.

Briggs, Charles L. (1986). Learning how to ask: A sociolinguistic appraisal of the role of the interview in social science research. Cambridge: Cambridge University Press.

Brinkmann, Svend & Kvale, Steinar (2005). Confronting the ethics of qualitative research. Journal of Constructivist Psychology, 18(2), 157-181.

Burton, Frank & Carlen, Pat (1979). Official discourses: On discourse analysis, government publications, ideology and the state. London: Routledge & Keegan Paul.

Carr, Patricia & Beaver, Graham (2002). The enterprise culture: Understanding a misunderstood concept. Journal of Strategic Change, 11(2), 105-113.

Chomsky, Noam (1998). Form and meaning in natural Languages. In Maria Baghramian (Ed.), Modern philosophy of language (pp.294-308). London: J.M. Dent.

Čmejrková, Svetla & Prevignano, Carlo L. (2003). Discussing conversation analysis: The work of Emmanuel A. Schegloff. In Carlo L. Prevignano & Paul J. Thilbault (Eds.), Discussing conversation analysis: the work of Emmanuel A. Schegloff (pp.11-55). Amsterdam: John Benjamins Publishing Company.

De Saussure, Ferdinand (1983). Course in general linguistics (translated by Roy Harris, edited by Charles Bally, Albert Sechehaye & Albert Riedlinger). London: Duckworth.

Doolin, Bill (2003). Narratives of change: Discourse, technology and organization. Organization, 10(4), 751-770.

Fairclough, Norman (1992). Discourse and text: Lingustic and intertextual analysis within discourse analysis. Discourse and Society, 3(2), 193-217.

Fontana, Andrea (2003). Postmodern trends in interviewing. In Jaber F. Gubrium & James A. Holstein (Eds.), Postmodern interviewing (pp.51-65). Thousand Oaks: Sage.

Fontana, Andrea & Frey, James H. (2000). The interview: From structured questions to negotiated text. In Norman Denzin & Yvonna Lincoln (Eds.), Handbook of qualitative research (pp.645-679). London: Sage.

Foucault, Michel (1972). The archaeology of knowledge (translated by Alan M. Sheridan Smith). London: Tavistock.

Foucault, Michel (1973). The birth of the clinic: An archaeology of medical perception (translated by Alan M. Sheridan). London, Routledge.

Foucault, Michel (1978) The history of sexuality: Volume 1, an introduction: The will to knowledge (translated by Robert Hurley). London: Penguin.

Foucault, Michel (1980a). Powers and strategies: Interview with the editorial collective of Les révoltes logiques (translated by Colin Gordon). In Colin Gordon (Ed.), Michel Foucault: Power/Knowledge: Selected interviews and other writings, 1972-1977 (pp.134-145). Brighton: Harvester.

Foucault, Michel (1980b). Truth and power: Interview with Alessandro Fontana and Pasquale Pasquino (translated by Colin Gordon). In Colin Gordon (Ed.), Michel Foucault: Power/Knowledge: Selected interviews and other writings, 1972-1977 (pp.109-133). Brighton: Harvester.

Foucault, Michel (1988). The concern for truth: An interview by Francis Ewald (translated by Alan Sheridan). In Lawrence D. Kritzman (Ed.), Michel Foucault: Politics, philosophy, culture: Interviews and other writings 1977-1984, (pp.255-267). London: Routledge.

Foucault, Michel (2006). Psychiatric power: Lectures at the Collège de France 1973-1974 (edited by Jacques Lagrange, translated by Graham Burcell). Basingstoke, Palgrave Macmillan.

Gabler, Neal (1998). Life the movie: How entertainment conquered reality. New York: Vintage Books.

Gilbert, Nigel G. & Mulkay, Michael (1984). Opening Pandora's box: A sociological analysis of scientists' discourse. Cambridge: Cambridge University Press.

Grey, Christopher (2004). Enterprise, management and politics. The International Journal of Entrepreneurship and Innovation, 5(1), 9-13.

Gubrium, Jaber F. & Holstein, James A. (2003). From the individual interview to the interview society. In Jaber F. Gubrium & James A. Holstein (Eds.), Postmodern interviewing (pp.21-50). Thousand Oaks: Sage.

Hardy, Cynthia; Palmer, Ian & Phillips, Nelson (2000). Discourse as a strategic resource. Human Relations, 53(9), 1227-48.

Harré, Rom, & Gillet, Grant (1994). The discursive mind. Thousand Oaks: Sage.

Holstein, James A. & Gubrium, Jaber F. (1997). Active interviewing. In David Silverman (Ed.), Qualitative research—theory, method and practice (pp.113-29) London: Sage.

Hook, Derek (2001). Discourse, knowledge, materiality, history: Foucault and discourse analysis. Theory and Psychology, 11(4), 521-547.

Kendall, Gavin & Wickham, Gary (1999). Using Foucault's methods. London: Sage.

Kitzinger, Celia & Frith, Hannah (1999). Just say no? The use of conversation analysis in developing a feminist perspective on sexual refusal. Discourse and Society, 10(3), 293-316.

Knights, David & Morgan, Glenn (1991). Corporate strategy, organizations, and subjectivity: A critique. Organization Studies, 12(291), 251-274.

Knights, David & Morgan, Glenn (1995). Strategy under the microscope: Strategic management and IT in financial services. Journal of Management Studies, 32(2), 191-214.

Kress, Gunther & Van Leeuwen, Theo (1999). Front pages: Analysis of newspaper layout. In Allan Bell & Peter Garret (Eds.), Approaches to media discourse (pp.186-218). Oxford: Blackwell.

Lee, Yew-Jin & Roth, Wolff-Michael (2003). Making a scientist: Discursive "doing" of identity and self-presentation during research interviews [37 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 5(1), Art. 12, http://www.qualitative-research.net/fqs-texte/1-04/1-04leeroth-e.htm [Date of Access: 09/13/ 2006].

McAuley, John, Duberley, Joanne & Cohen, Laurie (2000). The meaning professionals give to management.and strategy. Human Relations, 53(1), 87-116.

McGowan, Rosemary A. (2002). Organizational discourses: Sounds of silence. Phd Thesis, York University.

Miller, Jody & Glassner, Barry (1997). The "inside" and the "outside": finding realities in interviews. In David Silverman (Ed.), Qualitative research—theory, method and practice (pp.99-112). London: Sage.

Mishler, Elliot G. (1986). Research interviewing—context and narrative. Cambridge, MA: Harvard University Press.

Ng, Wilson, & De Cock, Christian (2002). Battle in the boardroom: A discursive perspective. Journal of Management Studies, 39(1), 23-49.

Potter, Jonathan (2002). Two kinds of natural. Discourse Studies, 4(4), 539-542.

Potter, Jonathan & Hepburn, Alexa (2005). Qualitative interviews in psychology: Problems and possibilities. Qualitative Research in Psychology, 2(4), 281-307.

Potter, Jonathan & Wetherell, Margaret (1987). Discourse and social psychology: Beyond attitudes and behaviour. London: Sage.

Potter, Jonathan & Wetherell, Margaret (1995). Natural order why social psychologists should study (a constructed version of) natural language, and why they have not done so. Journal of Language and Social Psychology, 14(1-2), 216-22.

Räisänen, Christine & Linde, Anneli (2004). Technologizing discourse to standardize projects in multi-project organizations: Hegemony by consensus. Organization, 11(1), 101-121.

Rapley, Mark & Antaki, Charles (1998). "What do you think about?" Generating views in an interview. Text, 18(4), 587-608.

Rigg, Clare (2005). "It"s in the way they talk": A discourse analysis of managing in two small businesses. International Journal of Entrepreneurial Behaviour & Research, 11(1), 58-75.

Sacks, Harvey (1995). Lectures on conversation Volume II. In Harvey Sacks, Lectures on conversation Volumes I & II (edited by Gail Jefferson). Oxford: Basil Blackwell.

Schegloff, Emanuel A. (1997). Whose text? Whose context? Discourse & Society, 8(2),165-187.

Schegloff, Emanuel A. (1998). Reply to Wetherell. Discourse & Society, 9(3), 413-416.

Schegloff, Emanuel A. (1999a). Discourse, pragmatics, conversation analysis. Discourse Studies, 1(4), 405-35.

Schegloff, Emanuel A. (1999b). "Schegloff's Text" as "Billig's Data": A critical reply. Discourse & Society, 10(4) 558-72.

Schegloff, Emanuel A. (1999c). Naivete vs sophistication or discipline vs self indulgence: A rejoinder to Billig. Discourse & Society, 10(4), 577-582.

Silverman, David (2001). Interpreting qualitative data—methods for analysing talk, text and interaction. (2nd ed.) London: Sage.

Slembrouck, Stef (2004). Reflexivity and the research interview. Critical Discourse Studies, 1(1), 91-112.

Soyland, Anders.John & Kendall, Gavin (1997). Abusing Foucault: Methodology, critique and subversion. History and Philosophy of Psychology Section Newsletter, 25, 9-17.

Speer, Susan A. (2002). "Natural" and "contrived" data: A sustainable distinction? Discourse Studies, 4(4), 511-525.

Speer, Susan A. & Hutchby, Ian (2003). From ethics to analytics: Aspects of participants' orientations to the presence and relevance of recording devices. Sociology, 37(2), 315-337.

Torode, Brian (1995). Negotiating "advice" in a call to a consumer help line. In Alan Firth (Ed.), The discourse of negotiation: Studies of language in the workplace (pp.345-372). Oxford: Pergamon.

Vaara, Eero (2002). On the discursive construction of success/failure in narratives of post-merger integration. Organization Studies, 23(2), 211-249.

Van Zoonen, Liesbet (2001). Desire and resistance: Big brother and the recognition of everyday life. Media, Culture & Society, 23(5), 669-77.

Wetherell, Margaret (1998). Positioning and interpretative repertoires: Conversation analysis and post-structuralism in dialogue. Discourse & Society, 9(3), 387-412.

Wood, Linda A & Kroger, Rolf (2000). Doing discourse analysis. London: Sage.

Wooffitt, Robin (2005). Conversation analysis and discourse analysis: A comparative and critical introduction. London: Sage.

Widdicombe, Sue & Wooffitt, Robin (1995).The language of youth subcultures: Social identity in action. New York: Harvester Wheatsheaf.

Wittgenstein, Ludwig (1998). Philosophical investigations. In Maria Baghramian (Ed.), Modern philosophy of language (pp.89-106). London: J.M. Dent.


Brendan K. O'ROURKE lectures in economics and strategy at the Faculty of Business, Dublin Institute of Technology, Ireland. He is also pursuing a PhD at Brunel University in the United Kingdom where his research is on discourses of small firm proprietors.


Brendan K. O'Rourke

Faculty of Business,
Dublin Institute of Technology
Aungier St., Dublin 2, Ireland

Tel.: +353-1-4027097
Fax: +353-1-4027198

E-mail: Brendan.ORourke@dit.ie
URL: http://www.BrendanKORourke.com/


Martyn PITT is Senior Lecturer in strategic management at Brunel University. His primary research interests are in the skills, knowledge and sense-making processes entailed in strategic management, especially in the context of organisational innovation and entrepreneurship. He has co-authored three books on strategy, innovation and small firms as well as being published in a variety of journals including Organization Studies and Strategic Change.


Martyn Pitt

School of Business and Management,
Brunel University
Uxbridge, Middlesex UB8 3PH, United Kingdom

Tel.: +44-1895-265296
Fax: +44-1895-203149

E-mail: Martyn.Pitt@brunel.ac.uk
URL: http://www.brunel.ac.uk/about/acad/bbs/bbsstaff/bm_staff/MartynPitt


O'Rourke, Brendan K. & Pitt, Martyn (2007). Using the Technology of the Confessional as an Analytical Resource: Four Analytical Stances Towards Research Interviews in Discourse Analysis [58 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 8(2), Art. 3, http://nbn-resolving.de/urn:nbn:de:0114-fqs070238.

Copyright (c) 2007 Brendan K. O'Rourke, Martyn Pitt

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.