Volume 2, No. 3, Art. 11 – September 2001

Reconciling Theory with Method: From Conversation Analysis and Critical Discourse Analysis to Positioning Analysis

Neill Korobov

Abstract: Not only is it often challenging to wade through the many different discourse analytic approaches to studying talk-in-interaction, but it is also often challenging to understand how certain methods adequately capture the complexity of the theories that lie behind them. What is needed are methods that are analytically sophisticated enough to empirically demonstrate the complexity of the theories that make fashionable and relevant the analysis in the first place. To illustrate this quandary, I will trade on some of the recent tensions between two of the most popular approaches—Critical Discourse analysis (CDA) and Conversation Analysis (CA). More specifically, attention is given to recent methodological attempts to synthesize a middle-ground position between CDA and CA. The focus of my overall argument will be that Positioning Analysis offers a viable analytic way to reconcile the discrepant methodological orientations while trading on the shared theoretical convictions of both CDA and CA.

Key words: Anti-positioning, positioning analysis, conversation analysis, critical discourse analysis, theory, method

Table of Contents

1. Introduction

2. Framing the Debate—CA and CDA

3. Navigating between CA and CDA—EDLEY and WETHERELL

4. Positioning Analysis

5. Final Thoughts





1. Introduction

Reconciling the theoretical insights that ground different qualitative methodologies with the actual analytic methods that supposedly follow from such theories is a crucial undertaking that is far too often left obscure in qualitative social research. For instance, it is highly fashionable these days to embrace some version of a "social constructionist" mantra—the idea that cultural/social/historical discourses play (to some degree) a constituting role in the semiotic or discursive establishment of our "realities". And further, as we practice these discourses in our everyday conversations, we in turn are perpetually re-constituting these discourses by expanding, challenging, rejecting, or re-inventing them. Many theorists craft this general, yet paradoxical sentiment vis-à-vis different theoretical movements. But, I will argue, far too few explicitly tie these theoretical insights to concrete qualitative methods that can analytically make sense of the way this theoretical paradox actually gets played out in the interactive domain. Far too few explicitly delimit the ways in which their methods embody their theories. More to the point, there are far too few explicit analytic methods available that effectively reconcile the theoretical tension that we both constitute and are constituted by the social and cultural discourses in / by / through which we speak. [1]

One of the more recent and heated theoretical and methodological arenas where this type of insight is debated is between proponents of Critical Discourse Analysis (CDA) and Conversation Analysis (CA). Part of my goal for this paper is to lay out this debate in an effort to illustrate the way similar theoretical orientations can diverge into very different methodological orientations. My more central goal is to suggest an analytic way to reconcile the methodological discrepancies between CDA and CA while retaining their shared theoretical convictions. As such, my argument is that Positioning Analysis (BAMBERG, 1997a, 1997b, 1997c, 1999a, 1999b, 2000a, 2000b, 2000c) occupies a salubrious "middle-ground" for collapsing some of the thorny tensions between proponents of Conversation analysis (CA) and critical discourse analysis (CDA). My argument is structured in three parts. First, I will situate the relevance of Positioning Analysis within the broader methodological field by making relevant exactly what these "thorny tensions" are between CA and CDA, and why I believe these tensions are highly germane as indexes of important methodological questions facing current social scientific research. To broach this debate, I will trade on the recent exchanges between Michael BILLIG (1996, 1999), Emanuel SCHEGLOFF (1997, 1999), Margaret WETHERELL (1998), Nigel EDLEY & Margaret WETHERELL (1997, 1999), and Norman FAIRCLOUGH (1993, 2001). Second, I will discuss EDLEY and WETHERELL's (1997, 1999) recent attempts to synthesize or move beyond the dichotomies of CA and CDA, but will suggest that despite their efforts their work remains too couched in CDA. Finally, I will present Positioning Analysis as a useful compromise. I will focus specifically on BAMBERG'S (see above citations) development of Positioning Analysis. [2]

2. Framing the Debate—CA and CDA

CA and CDA often begin with many of the same theoretical assumptions. For instance, both are discursive approaches to the social order and to the study of talk in interaction. Each posits that identity is an active, discursive and/or semiotic accomplishment maintained and transformed within interactions. Each maintains that identities are organized out of the social order (vis-à-vis procedures or skills called "ethno-folk-methods"), are actively mobilized within the ongoing details of talk and communication, are sequentially organized, and are thus the product of joint social action. In addition, they are united in general in their assumption that we construct and are constructed by societal and historical discourses. The differences arise as each orientation methodologically conceptualizes and pursues these insights differently. The differences have to do with the fact that each has a different methodological way of invoking context. Each has a different degree of willingness or criteria for invoking those contexts (or broader discourses) in the interpretation of social action. In addition, each orientation shares different convictions about the possibility for "studying participants orientations" or "studying participants in their own terms". While CA proponents embrace this dictum without apology, CDA proponents hear it as a vestige of some incipient form of naïve realism. And these tensions, to be unintentionally curt, are only the tip of the iceberg. [3]

As noted above, I will explore the details of these differences as they are articulated in the recent exchanges between SCHEGLOFF (1997, 1999), BILLIG (1996, 1999), and WETHERELL (1998). I will also trade on FAIRCLOUGH'S (1993, 2001) and EDLEY'S (2001) development of CDA. To begin, proponents of CA (particularly SCHEGLOFF) reject what has been called a "bucket theory of context" in which some preestablished social framework is viewed as containing the participants actions (in other words, that the frameworks carry around inherent meanings). Instead, CA argues for a more dynamic approach in which context is treated both as the project and product of the participants own actions and therefore as locally produced and transformed at any moment (SCHEGLOFF, 1999). Note, there is a slight difference between saying that the context both produces and is produced by the participants actions (CDA premise) and treating context as the project and product of the participants own actions (CA premise). The latter way of framing it removes the idea of the context (or societal discourses) as having some autonomous or preestablished volition in directing the participants actions. CA maintains that utterances and the social / historical actions they embody (or index) are doubly contextual in the sense that they are context shaped and context renewing (HERITAGE, 1984). But this view of "context" is more local than is "context" in CDA terminology. In CA, utterances and actions are context shaped in that their meanings and subsequent relation to the ongoing sequence of actions depends on what has come before. And further, utterances are context renewing because every current utterance will form the immediate context for the next action. This form of being "context shaped" is a much softer and more local version than the more radical or broad socio-political notion of "context shaped/produced" among some CDA proponents. [4]

Another point of contention concerns CA's dictum to understand participants "in their own terms" or "own orientations". The general idea for CA is that analysis should illuminate the interpretive mechanisms and understandings that are relevant for the participants and the local practices that make possible those understandings. Analysis should focus on the linguistic resources used in the production of such local knowledge, paying attention to when and how it is made salient. When invoking the relevance of gender or social categories in their interpretations, analysts must show how these categories are oriented to by the participants and how they are procedurally consequential in the interaction. Analysis should NOT simply invoke categories out of some pre-established theoretical or academic. For instance, invoking the concept of "power" without showing how "power" is actually being done within the interaction is to be avoided. CA proponents—while not denying the theoretical insight that power is parasitic on all social interaction—do not believe that this insight necessarily sanctions one to see power at work in every interaction. CA believes that sometimes power differences are not mobilized as the basis for every interaction. They believe in a certain degree of rigor and systematic-ness—a conviction to stick with what is actually said and to ground observations in the details of interactions more so than in the theoretical apparatuses of this or that analyst. [5]

One of BILLIG'S (1999) main problems with CA is that he believes it is based on an unexamined, or naïve epistemology and methodology. In short, BILLIG (1999) believes it is both impossible and misleading to ever study participants "in their own terms" and that an analysis that does not incorporate (or attempts to bracket or avoid) the broader backdrop of social and cultural discourses entirely misses the point of doing social analysis in the first place. BILLIG believes that CA can never and will never be able to study participants "in their own terms" because such an endeavor is one of the many scurrilous of "realist tales". He believes that the whole notion of getting at the unique and actual terms in which the participants speak rests on the supposedly "out-dated" idea that the "facts" can actually "speak for themselves". In other words, CA seems (at least to BILLIG) to be saying that it is possible and desirable to uncover what the participants are actually saying (in pure form) without polluting their actual words with our interpretive mechanisms. BILLIG finds this laughably naïve. [6]

WETHERELL (1998) seems to agree. While she lauds CA for its attention to small pieces of conversational detail, she cautions against making this the goal for analysis. She finds this impractical and restricted. It is impractical because there are always an infinite amount of "fine-details" to look at—making the admonition to simply "look at" the actual interaction an oversimplified feat. How one conceptualizes "detail", WETHERELL argues, is an ideological point that should be built into the analysis. Secondly, even if one could grasp the actual "fine-detail" of the interaction, such an approach is far too restrictive. It is restrictive because SCHEGLOFF's sense of the participants' orientation is far too narrow. For starters, the notion of a participants' orientation is unclear. When does it cease being the analyst importing their own preoccupations and begin being what the participants are actually orienting themselves to? WETHERELL (1998) asks: Isn't it always the analyst who selects certain aspects of the conversation to highlight, thus participating in the construction of what becomes "relevant"? And finally (in consonance with BILLIG), by restricting the analytic gaze to the fragment, doesn't the analyst marginalize the broader social and cultural argumentative textures of which the fragment is a part? [7]

WETHERELL's (1998) basic challenge to CA is that by focusing too narrowly on the interactive moment, we run the risk of forgetting that the positions drawn up in that moment are one of many variations or options for reflecting the larger patterns or threads of intelligibility that make possible that very interaction and all of the possible detailed exigencies that are then available for analysis in the first place. Her feeling is that SCHEGLOFF is performing his own act of colonization by proposing an overly narrow understanding of what it is that participants are orienting themselves. She believes that an adequate analysis must not only look at the conversational details of talk-in-sequence, but must also trace these detailed linguistic formulations through the larger argumentative threads that are displayed in the participants orientations. We must interrogate, she says, the taken for granted discursive backcloth that organizes and makes possible participant orientations. This, she argues, should be one large part of the larger genealogical analysis of socio-political issues. This is what she means when she claims that we need a scholarly analysis and not simply a technical one. Without it, the analysis runs the risk of being an irrelevant or moot point. [8]

In defense of CA, SCHEGLOFF (1997, 1999) has several responses to challenges like these. For starters, SCHEGLOFF (1997, 1999) argues that CA's insistence on looking at the participants "in their own terms" is not the same thing as looking for the pure form (free from interpretation) of what the participants are "actually" (in some "truly" real way) saying. SCHEGLOFF believes that BILLIG is criticizing a straw-man, and is perhaps overly pre-occupied with imbuing a false sense of ontology to phrases like "actual" or "own terms". When SCHEGLOFF and other CA proponents advocate these convictions they are not suggesting an interpretation-free, getting-at-the-real-thing form of analysis. SCHEGLOFF is simply stressing that all interpretations must be grounded FIRST in the actual talk and practices of the participants. This is not an endorsement for naive objectivity, nor is it a recipe for all together obviating the larger socio-political backdrops at play. [9]

Analysis of the participants "own terms" is a methodological admonition to ground interpretation within the ongoing sequence of talk-in-interaction, and NOT grounding interpretation FIRST in this or that theoretical or political orientation and SECOND in however one can find such predilections embodied in the data. Inevitably (and SCHEGLOFF knows this), even our most fundamental observations and attempts to be rigorous or systematic are historically shaped by our theoretical and political orientations. SCHEGLOFF realizes that of course everything we do (even rigor of CA) is historically connected to our theoretical and political interests. SCHEGLOFF (1999) does not deny this, but he does not take this "warming fact" to be the starting point for analysis. He argues against any methodological approach that side-steps a rigorous and patient investigation into the actual utterances and exchanges in interactions for a self-consciously political projection into the data of ones own interests. [10]

SCHEGLOFF (1999) believes that far too many working in CDA side-step an explicit grounding of their analysis in linguistic detail. Instead, they far too quickly invoke something like "power differences" or "hegemony" when it is not always so clear how the participants are linguistically indexing (or orienting themselves to) something like "power"—or for that matter, what "power" even is. SCHEGLOFF asks time and time again of CDA what standards are being used to ground interpretation. One answer BILLIG (1999) gives over and over again is "communities of relevance" as the standard (for instance, other academics in various fields who have thought hard about such issues). SCHEGLOFF (1999) feels that the "community of relevance" ought to be composed of the actual participants under investigation. It is what they demonstrably orient themselves to (as best we can establish it) that determines what counts as an observable or as standing evidence. [11]

In any case, whatever community of relevance is chosen, SCHEGLOFF argues the analyst must show that the observation being advanced and the analytic line being taken is resonant with the actual orientations of the people who matter most—the ones who actually perform the social practices of which we speak. They are the ones who are engaged in the conduct and whose understandings of the relevances the actual ensuing trajectory of the interaction was built. SCHEGLOFF positions himself as one who opts to actually allow the participants to act as the community of relevance, while BILLIG is seen to argue that it is our academic communities that act as our interpretive guides. SCHEGLOFF comes across looking fair, yet naïve, while BILLIG comes across looking imperialistic, yet sophisticated. One of SCHEGLOFF's (1999) big concerns for people doing CDA is that their critical analyses end up not "binding" to the data, and risk ending up merely ideological. One of BILLIG's (1999) concerns is that CA's approach to detail is an overly-focused analysis of detail at the expense of the broader and (perhaps more important) social and political issues that need attention. [12]

After wading through the different convictions of each orientation, it is tempting to simply conclude that each has different analytic agendas and starting points—that each orientation, in short, is operating at a different level of analysis. It is tempting to think that CA is simply designed to reveal how things like pronomial self-repair strategies are accomplished during question-and-answer exchanges while CDA is designed (as FAIRCLOUGH 2001 suggests) to uncover something like the ideological workings of hegemonic language practices. It is my belief that both of these stereotypes are misplaced oversimplifications. The debate isn't that CA is completely myopic to the larger socio-political contexts or that CDA is all together ignorant to detailed linguistic patterns and micro-discursive constructions. The debate is really about when and how things like "context" and "participant orientation" are brought into the analytic discussion, and how they ground claims-making. BILLIG (1999) argues that analysts should not have to wait until "power" or "abuse" are actually brought up or attended to before the analyst can invoke them. Invoking them need not be an imperialistic move, but rather an informed and cautionary attempt to fill-out the social and cultural forces which have come to make possible the encounter in the first place. BILLIG (1999) basically argues that CA should become more ideological in its fine-grained efforts and less neutral. Because BILLIG believes that a non-ideological analysis is impossible, he wants to argue that CA should aim less for pure empiricism and more for an open and reflexive ideological presentation of its assumptions and motives. [13]

SCHEGLOFF (1999) demurs with this sterile characterization of CA as overly-empirical, and argues that when we are analyzing strips or moments of discourse in which power is featured, it is far from obvious that a CA approach would be inappropriate, misplaced, or politically insensitive. Rather, SCHEGLOFF (1999) maintains, by examining in close detail the ways that instances of conversational interactions escalate (in cases of abuse, for instance)—even when the abuse is not yet magnified in the particular interaction—we can see the micro-genesis of such powerful and devastating episodes. And it is only through understanding how exactly such complex episodes are orchestrated that understanding or intervention is possible. Far from being an inappropriate or insensitive way to approach instances of abuse or inequality, SCHEGLOFF believes that CA can be one of the most helpful. As such, by using the tools of CA we can gather detailed records of how extant social categories (like abuse or inequality) are accomplished interactively. That is, we can see what units and resources are locally used by participants to build the broader social categories. As SCHEGLOFF (1999, p.562) notes:

"If interaction is produced within a matrix of turns organized into sequences, etc., and if it is from these that motives and intentions are inferred, identities made relevant, stance embodied and interpreted, etc., how else—when confronted by the record of singular episodes—are we to understand their genesis and course, how else to try to understand what unwilling participants can do to try to manage that course to safer outcomes, how else to try to understand how others might intervene to detoxify those settings?" [14]

SCHEGLOFF believes strongly that those committed to understanding how inequality and oppression operate interactively should harness CA's tools as a resource for their work rather than complain about them as ideological distractions. For those in CDA who are convinced that social interactions are the marketplace for oppression and power-relations, they should undertake to demonstrate that rather than simply assume it and then find it. And they should realize that CA's tools of analysis in no way preclude the empirical demonstration of such ideological convictions. SCHEGLOFF (1999) adds that people in CDA should be more about the business developing and deploying these linguistic and analytic skills on actual discursive materials and less with belaboring them with ideological character assassination. SCHEGLOFF notes that for many in CDA the great risk is they get drawn further and further into criticizing theory and methods and less into actually doing good empirical work—that is, their critical theory becomes the work they choose to do. [15]

3. Navigating between CA and CDA—EDLEY and WETHERELL

By now it is hopefully obvious how there are some points of overlap between CDA and CA and many points of contention. By emphasizing important analytical insights, each orientation can potentially enliven discourse analysis in general, particularly if ways of working between CA and CDA can be imagined and enacted. One such attempt comes with the recent work of EDLEY and WETHERELL (1997, 1999). Like BILLIG, WETHERELL (1998) and EDLEY and WETHERELL (1997, 1999) have been engaged with CA advocates (particularly SCHEGLOFF) about these exact issues. To begin, EDLEY and WETHERELL (1997) argue that it is simply time to move beyond the dichotomies of CA and CDA. They imagine "forms of discursive psychology which draw more eclectically on both styles of work and which study the ways in which people are simultaneously the master and slave of discourse" (EDLEY & WETHERELL, 1997, p.206). Yet while attempting to synthesize these two orientations, it is not so clear exactly how the points of contention between CA and CDA are being reconciled. My feeling is that it is simply the general "simultaneous master and slave of discourse" slogan that they are out to synthesize. If so, it is difficult to see how it specifically addresses the wide ranging and specific tensions between CA and CDA. [16]

EDLEY and WETHERELL's synthesis of CA and CDA does not (to the delight of SCHEGLOFF) see broader social and cultural discourses as constituting the subject. But they do take it that the participant is positioned by these discourses. What determines how and what a participant does in a particular interaction is their orientations to (or accountability within) those discourses and the emergent conversational activities. In this way, what a participant does is only partly the consequence of the discourses that co-inhabit his or her settings. To find a way to analyze the both/and of being positioned and simultaneously actively positioning-back (so to say), EDLEY and WETHERELL (1997) introduce the concepts of "interpretive repertoires", "ideological dilemmas", and "subject positions". It is really with the interplay of these concepts that EDLEY and WETHERELL attempt to find a methodological synthesis between CA and CDA. As such, my central criticism with their attempted synthesis has to do with their lack of grounding of these concepts within CA. [17]

EDLEY and WETHERELL's form of critical discourse analysis takes the form of scholarly-informed paraphrases of the interplay between interpretive repertoires, ideological dilemmas, and subject positions. While they do analyze transcripts and language constructions, their main agenda is in delimiting the interplay of these three ideologically-loaded concepts for particular socio-political issues—such as the fragmentation and contradictory nature of our shared cultural conceptions of masculinity, fatherhood, or gender relations (see FAIRCLOUGH, 2001 as an example of a similar starting point). Approaching analysis in this way, according to EDLEY (2001), requires a violation of the CA inspired maxim to focus attention to the participants orientations within ongoing sequences of interaction. Instead, EDLEY (2001) believes the participants talk should be understood and analyzed as embodying certain interpretive repertoires and as the attempt to manage the dilemmatic nature of often conflicting lived ideologies. This conviction lays the groundwork for EDLEY and WETHERELL's strong inclination towards a more critical (and less CA-oriented) form of analysis. [18]

With "interpretive repertoires" (or "argumentative threads"), EDLEY and WETHERELL are referring to the culturally familiar and habitual lines of argument comprised from recognizable themes, common places, and tropes. EDLEY (2001, p.202) refers to them as "repositories of meaning"—giving them a rather "out there", discourse-independent quality. As one example that EDLEY and WETHERELL (1997) discuss, the idea of "scoring" in the context of sexual-talk indexes the notion of a sexual conquest for some males. As such, it is an interpretive repertoire that invokes certain culturally familiar narratives of competition, conquest, popularity, and an ethic that legitimates sexuality and equates it with permissiveness, promiscuity, frivolity, success, not to mention the idea of women as "commodities". Interpretive repertoires are not simply the pre-figured cultural resources used to make sense and convey meaning, but they are also the "building blocks" of conversation (EDLEY, 2001). They place an essential, constituting role. For EDLEY and WETHERELL's analysis, the goal is to find interpretive repertoires at work, how they are being utilized, where the boundaries lie between them, etc. (see EDLEY, 2001). According to EDLEY (2001), what determines whether or not one has located an interpretive repertoire has to do with how skilled one is as a scholarly analyst—that is, how well one can see ideological patterns within data. Again, while this is a rather fashionable way to operate within qualitative research, it is nonetheless decidedly critical and not conversation-analytic. [19]

The same is true for their use of "ideological dilemmas" and "subject positions". Interpretive repertoires makes possible different "subject positions"—which are psychologically-laden locations that people take-up or inhabit as they intentionally or unintentionally paint a picture of "who they are", or how they want to be seen in the particular conversation. Subject positions are locations made possible as subjects are interpellated within certain discourses. As such, subject positions are ideological effects (EDLEY, 2001). In addition, they comprise discourses and can work to change cultural and historical meanings over time. To connect interpretive repertoires with subject positions, EDLEY and WETHERELL trade on BILLIG's (1987) idea of "ideological dilemmas". Ideological dilemmas are indexed when there is tension between interpretive repertoires and the management of certain subject positions. The back-and-forth tension between the moves and versions that are made available vis-à-vis interpretive repertoires and the active subject positioning that indexes these interpretive repertoires is ideological dilemma management—and it is the site for EDLEY and WETHERELL's critical discourse analysis. Their efforts, then, involve analyzing the "ideological dilemmas"—dilemmas demonstrated as participants try out, manage, resist, or affirm different versions of culturally available argumentative threads. The distinctive site for their analysis is the tracing of the lived ideological tensions between the use of competing interpretive repertoires. In other words, ideological dilemmas drive conversations (EDLEY 2001, p.207). The participants talk is the "battleground" where opposing ideological realities are played out (EDLEY 2001, p.209). Analysis, then—in a markedly critical fashion—concerns a socio-political interpretation of various ideological fields. Even the specification of "subject positions" is marked less through the active and sequential use of linguistic constructions and more through culturally familiar descriptive tropes—like "macho", "heroic", "rebellious", etc. (again, a marked departure from a CA approach). [20]

While I find it an interesting and noteworthy endeavor to study the use of interpretive repertoires and ideological dilemma management, I do not believe EDLEY and WETHERELL's approach steps very far beyond the shadow of CDA. In addition, I agree with WETHERELL (1998) when she argues that the measure of whether or not analysis is done well is not simply our level of rigor or systematicity, but our skill as historical and cultural commentators who are able to say useful and interesting things about ideological contexts, structures, and the possibility for change. Nonetheless, I do not believe they have adequately found a way to synthesize the orientations of CA and CDA under this insight. What we are left with is indeed a scholarly and interdisciplinary analysis, but one that is underdeveloped in demonstrating the fine-mechanics of how participants are actually—linguistically and sequentially (and not simply ideologically)—drawing up subject positions or indexing (from the ground up) patterns of lived ideology. With my discussion of Positioning analysis, I will argue that what is missing is 1) an immanentist account of discursive meaning-making (see DAVIES & HARRÉ 1990; HARRÉ & VAN LANGENHOVE 1999, 1992) that posits that subject positions are immanent within (and not transcendent to) discourses and 2) an account of indexicality—that is, a micro-discursive way of demonstrating how the interactive use of language forms index (or draw-up into a kind of communicative space) versions or perspectives that in turn index certain subject positions, or social acts / social identities. This process has a decidedly "bottom-up" emphasis, while not neglecting the broader ideological discourses that are linguistically resisted, maintained, and re-worked. [21]

4. Positioning Analysis

One of the central reasons why BAMBERG's Positioning Analysis avoids the overly top-down trappings of CDA or the overly myopic technicalities of CA is because it derives from neither orientation. Rather, the development of BAMBERG's Positioning analysis really begins with his concept of narrative positioning—a concept that results from a critical functionalist extension of what the concept narrative means, particularly as it involves the tension between 1) narrating as a way to simply refer to a world of past events (ordered in time) and 2) narrating as a way to establish perspective or point-of-view (BAMBERG 1997a, 1997b). BAMBERG (1997a, 1999a) traces the first of these two possible orientations to LABOV and WALETCKY's (1967/1997) original narrative framework, and notes that within this framework narratives of personal experience were seen as representations of events that once happened, and which now have meaning to the narrator. Dissatisfied with this highly referential orientation to narrative, BAMBERG (1996, 1997a) began to argue for a more functionalist, performance-based approach to narrative. This orientation culminated and extended his work on perspective, agency, and event construal (1994, 1996). In fact, much of his work on how perspective and point-of-view are linguistically established (from the bottom-up) forms the backbone of his development and application of positioning analysis. [22]

Rather than seeing the establishment of perspective as that which is the result of the preverbal construction of cognitive decisions, BAMBERG wants to argue that perspective-taking involves the active and discursive management of a vantage point (or viewpoint—or "event view") that is manufactured linguistically through the marking of agency. BAMBERG (1994) points out that certain agency constellations (linguistically marked) are inextricably connected to the discursive purposes at hand, and that the agency relationships established are central in the construction of a perspective such that the events, character alignments, storylines, and the use of certain cultural repertoires appear as linguistic products. Further, vantage-point (BAMBERG, 1994) is the term he reserves to refer to that which binds together action, space, and time—thus unifying the regulatory nature of the interactive system. In this way, perspective and vantage-point do not enter the construal process independent of the discursive purposes to which they are relevant. Rather, the choice of perspective and event construction (the establishment of topic, loci of control, event view, degrees of agency, etc.) are functions of the discursive purposes (attributing blame, saving face, etc.) that are relevant for the narrator. In this way, the narrator is always positioning him/her-self not simply referentially with regard to the figure-ground textual construction of some state-of-affairs, nor as the conduit for pre-established repertoires, but also performatively vis-à-vis an actual or imagined audience for some type of discursive effect. The referential establishment of character, events, norms, or cultural repertoires are a means to an end—that is, of establishing a self that is positioned within a moral order. This moral order is indexed linguistically vis-à-vis the degree of agency marked in the speaker's perspective. [23]

Rather than beginning by looking at what (content) is being talked about, or which interpretive repertoires are being utilized, this performative orientation focuses more on how the narrative is linguistically performed and what this performance means in terms of establishing the narrator's perspective. With this orientation comes a distinct view of the function of language, one that further differentiates BAMBERG from EDLEY and WETHERELL. According to BAMBERG (2000a), language has traditionally come to be defined in two ways—1) as knowledge, where one first gains knowledge of certain language forms or rules that one then puts to use for certain communicative activities, and as 2) as practice, where the simple participation and use of cultural repertoires somehow affords one the correct knowledge and use of language forms and linguistic rules. For BAMBERG (2000a), both views of language are problematic. If language is a kind of knowledge (or competence!), then there is the problematic dualism between some hypothetical realm where this social competence or knowledge of language rules/forms is purportedly stored and the interactive social realm of practices where they are put into use. Further, if language is practice, implying that knowledge of language forms/rules is derived from one's experience in social situations, then there is the problem of accounting for how participation within certain socio-political discursive repertoires (or practices) can somehow afford one either the correct knowledge of language forms/rules or the ability to adopt or revise a perspective. In other words, the linguistic establishment of a position / perspective is not simply an ideological effect (as it is in EDLEY & WETHERELL's view of "subject positions"). [24]

In contrast to approaching language in terms of either of these views, BAMBERG (2000a) stresses that to understand language is to understand language use, meaning that language is inextricably situated within human practices. It is here Positioning Analysis embraces DAVIES and HARRE's (1990) argument for an immanentist conception of human meaning production. The general idea is that interpretive repertoires, rules, or norms do not pre-exist actual language production. Instead, interpretive repertoires are immanent within actual conversations, and are indexed into communicative reality by the use of certain linguistic devices that have been used historically in actual past conversations. This notion does not suggest (as noted above) that language is simply the medium for conveying cultural repertoires. Nor does it suggest that language is simply the site for ideological dilemma management. Rather, language is as language is used (i.e., language-in-use rather than language as a transcendent system, or tool, or site). Again, this is not the same as saying that language is derived from or the site for the management of something independent of it—i.e., interpretive repertoires. It is also counter to the view that language is a way of reflecting social competence or knowledge, although it is true that language can be treated as a kind of knowledge system that can be reflected upon (BAMBERG, 2000a). The key point here is that the systematic tools, interpretive genres, and normative discourses used to discuss language use are themselves a form of language use, immanent within actual conversations and not independent realms that are subsequently manipulated in social interactions. This immanentist orientation to language-as-use is a cornerstone for weaving a middle ground position between CA and CDA. [25]

Circumscribed to an immanentist view of language as use, subjectivity and the notion of "identity" becomes something that we do, an active, interactional accomplishment (BAMBERG, 2000a). Because language is not fundamentally seen as something that people possess, subjectivity or identity is thus not viewed as an attribute that is possessed differently by different people. Rather, identities are discursive, meaning that identity is immanent and made relevant in the ongoing, fine-detailed patterns of "talk". As talk, identity is done through the use of things like turn taking, topical shifts, contrasts, repairs, lexical and pronoun choice, formulaic expressions, language varieties, intonational patterns, figures of speech, and so on. Like the view of language being presented here, identity does not refer to an overarching structure of the individual who synthesizes and organizes various conversations and social practices. Nor does it paint a picture of the "subject" as an ideological effect, or as simply the site for the study of socio-political repertoires, thus obviating an analysis of the linguistic construction of agency and volition. Rather, identity is seen as the local, or "ethno"-ways in which talk is used in interactive contexts to evince the local display of perspective, or the positioning of self vis-à-vis the other, and vice-versa. [26]

In arguing this way, BAMBERG is clearly trading on an "ethnomethodological" or "CA" stance for language and identity. The import this has for the development of positioning is the onus to create a method that is sensitive to "hearing" the participants as much as possible. Rather than being a naïve endorsement of realism, this simply implies situating positioning at the micro-level of the turn-by-turn interaction, fastening our gaze on the way the linguistic forms are utilized to do certain things and not others. In studying language production this way, focus is paid to the active subjective positioning and construction process of participants using discourses in situ. From a conversation analytic frame, the gaze is on the micro-patterned, turn-by-turn active accomplishment of various linguistic forms in particular contexts. From an ethnomethodological frame, the gaze is on the kinds of subjective stances accomplished through talk, such as conveying blame, taking responsibility, criticizing, and making denials. This level of discursive analysis emphasizes people's linguistic activities, highlighting the remarkable subtleness and sophistication of ordinary people's talk and its designed features. [27]

In line with ethnomethodology, interactive talk does not arrive pre-packaged and pre-ordained according to the interpretive repertoires that constitute our social categories, but is actively constituted and re-constituted in interaction. The focus here is on the person actively interacting and making sense of the given social and historical conversations of which s/he is a part. Thus, in line with CDA, one does not deny the existence of interpretive repertoires or ideology, but instead views identity formation as resulting less from the imposition of interpretive repertoires than in the active indexing of (or "wrestling with") the linguistic constructions that have conventionally been linked to certain cultural repertoires. In a way that parallels EDLEY and WETHERELL's general aim, talk is not simply about events and occurrences (in the sense of a "conduit" transmitting the larger cultural discourses) but is also a constitutive part of those broader discourses. However, this insight is more fruitfully worked out as BAMBERG appropriates OCHS' (1996) and GUMPERZ's (1996) discussion of the indexical property of language. Because language is not simply referential, but also indexical, the content or linguistic construction of ones talk can "point to" different possible versions. [28]

It is with the discussion of indexicality that Positioning Analysis allows the interests of both CA and CDA to coalesce. In these terms, subject positions do not change as they ideologically manage certain interpretive repertoires, but because the local ordering activities themselves produce linguistic forms which index versions—or arrays of discursive possibilities. Because speakers possess many different ways of talking about the world, a rhetorical topic need not be discussed in just one way. Here, the variation that is produced is the result of linguistic forms (or construction types) indexing different types of stances or performances. As BAMBERG (2000b, p.11) points out:

"Speakers' choice of particular construction types (intonation, lexical or grammatical) indicate claims to particular memberships as well as stances toward particular categories. However, rather than assuming that these choices are indicative of symbolic knowledge systems which in turn are implemented with the discursive effect "in mind", I would like to suggest that they are residues of language practices that have been practiced elsewhere and now are locally instantiated to claim positions vis-à-vis a particular category." [29]

What BAMBERG is doing here is specifying a mechanism (linguistic variation a la construction types) that explains how subject positions are engaged in identity projects. BAMBERG's (2000b) point is that the variability of possible linguistic constructions is the resource that subjects use in practice (not simply the adoption of interpretive repertoires), and in so doing, ideological norms or repertoires are indexed into the here-and-now (into a kind of "communicative reality") as subjects position themselves within and against certain interpretive repertoires. [30]

This indexical view of language ought not be overlooked, for it is really with it that BAMBERG's approach to positioning (and his view of language in general) is unique. In stressing indexicality, BAMBERG (2000c) seems to diverge from the assumption that subject positions ought to be studied by isolating the interpretive repertoires that make them possible. BAMBERG (2000c) stresses that these ideational, cultural, cognitive, meta-discursive (or whatever one wants to call them) repertoires do not direct the practice of text and context construction as EDLEY and WETHERELL seem to suggest. By denying this assumption, BAMBERG is not denying the CDA conviction that language is historically situated or that people have "histories", nor is he advocating the scurrilous view that "everything is language" or that everything is constructed anew during each interaction. Rather, he is taking note that neither analysts nor participants have direct access to their backgrounds, cultural discourses, rules, norms, or intersubjective repertoires in such a way as to draw from them in text/context construction. And if and when they do, it is because they are indexed within the course of language use. As a result, not only do we position ourselves as a function of being language users, but we are simultaneously positioned as we use language forms that have been used in prior conversations for antecedent purposes (BAMBERG, 2000c). [31]

By arguing this way, BAMBERG is carefully rearranging the way we (as analysts) would go about understanding how subject positions are drawn up. What is crucial to note is that it is a methodological re-description, not an ontological negation. What is being suspended is simply the assumption (not the existence of) that things like shared backgrounds or interpretive repertoires have the kind of galvanizing force that many have posited. For BAMBERG, Positioning analysis begins at the linguistic level (CA), but it does not necessarily end there (CDA). While BAMBERG agrees that "not everything is language", he plays with the idea that one's notion of "something beyond language" (read: "interpretive repertoire", "mind", etc) is itself a notion made possible vis-à-vis language (2000c). Again, as language forms index communicative functions, they simultaneously do not index other possible figure/ground or text/context relationships. Thus, within the indexical act there is communicated the other possibilities that have not been selected. It is as a result of this, BAMBERG (2000c) suggests, that UZGIRIS's (2000) notion that "there seems to be more to our thoughts and existence than can be expressed in language" rings true. [32]

Through an empirical focus on socio-linguistic variability for understanding how subject positions are indexed by the use of linguistic forms—which in turn index interpretive repertoires—BAMBERG is able to use positioning analysis to connect CA and CDA. By trading on the multifunctionality of construction types, positioning analysis affords a multi-level analysis (between CA and CDA) of how language forms convey what the talk is about and how it is structured (CA), with an ideological and rhetorical discussion of how it works (conveying blame, taking responsibility, satirizing, demeaning, etc.) to establish various subject positions within a moral order (BAMBERG, 1997a, 1997b, 1997c, 1999a, 1999b, 2000a, 2000b). Said differently, positioning analysis analyzes the different linguistic forms used to position oneself within different topics, during different interactive situations, and for the management of certain ideological tensions in the overall establishment of "who I am" or "who I am becoming". [33]

BAMBERG (1997a) specifies three distinct, but interrelated levels of positioning. In closing, I will highlight these levels for the purpose of underscoring exactly how both a CA and CDA orientation are methodologically accounted for as one analytically moves from level one positioning through level three. Each level is intricately connected and interdependent. An actual analysis would proceed cumulatively from level one through level three. They are as follows (summarized from 1997a):

  • Level 1How the conversational units (i.e, characters, events, topics, verb structure, etc) or general conversational structure are positioned in relation to one another within the reported events. In line with a CA orientation, the general concern is with analyzing how the contents or units of conversational organization are sequentially situated across turns. In other words, which linguistic devices and sequential arrangements are being used and in what order? For instance, at this first level of positioning attention is given to how characters are constructed ("he", "she", "Mike", "they", "us", etc.) and how they are set within the ongoing series of unfolding events or descriptions. With this level of linguistic analysis, characters are linguistically marked in different ways. For instance, as 1) the agent who is in control of the situation or others, or 2) the victim or passive recipient who is at the mercy of outside forces or who benefits from luck, fate, etc. In addition, attention is paid to how the opening and closing of turns are structured, what are the linguistic structural preferences or "unit types" and how are they organized, and how is turn-taking or distribution patterns made salient?

  • Level 2How the speaker both is positioned by and positions him/herself to the actual or imagined audience. This level concerns how the content and structure of the talk are actually interactive effects. In other words, how conversational units are distributed and managed within (and as an effect of) certain distinctively interactive or discursive modalities—i.e., what are the conversational units doing interactively in institutionally-saturated settings like interviews or focus group interactions or in more quasi-natural / free-interactive settings? With this level of positioning, a CDA focus begins to emerge. The structure and content of conversation is analyzed as a means to an end—one that is concerned with situating conversational structure within certain distinctive audience-driven interpretive modalities. For instance, given the way the characters and topics are structured at level 1 positioning, the analytic focus at level 2 concerns the indexical establishment of certain subject positions and social acts (attributing blame, making a denial, giving advice, making excuses) that are ideologically meaningful as indexes of the particular interactive moment? Focus is given to the particular discourse modes being employed and of the ensuing effects for the interaction.

  • Level 3How do the narrators position themselves in answering the specific and general question of "who am I?" and "how do I want to be understood". This level of positioning is a culmination of the previous two levels. Because the linguistic devices used do more than simply specify the structure or content of what is being said, but additionally "point to" broader identity projects at work, what emerges at this level of positioning is a more distinctively CDA focus on the question of "who am I?", or "what kind of person do I want to be seen as?" The analysis at this level addresses how various interpretive repertoires are indexed into the interaction. Attention is also given to the various ideological tensions produced as various subject positions are managed, resisted, or reworked. The different positions or identity claims described at this level are not meant to hold across contexts. Because they are highly interpretive as scholarly-exercises, they are debatable given different academic predilections. Thus, an important feature of positioning at this level is that the broader interpretive repertoires invoked during claims-making are used as interpretive guides, and not as causal mechanisms used in an overly simplistic or generalizable sense. Positioning at this level is thus revisable, tentative, and of limited range. Nevertheless, it allows for a distinctive CDA orientation. [34]

5. Final Thoughts

My hope with this paper is to have made a tendentious argument concerning the discrepancies between CA and CDA, to have presented a noteworthy attempt (ELDEY & WETHERELL) to draw eclectically from both orientations, and a case for why and actually how Positioning analysis can be a useful method for synthesizing the orientations of CA and CDA. In closing, I will not attempt to draw overly general similarities or differences between the different methods or thinkers explored thus far. Hopefully by now, these similarities and differences have been clearly articulated. I do want to underscore that it is crucial for qualitative researchers to adequately create methods that are sophisticated enough to capture analytic complexity and not simply theoretical complexity. More specifically, it is crucial for one's analysis to analytically illustrate what the theoretical idea of a "discourse" looks like operationally, how discourses are linguistically indexed, how they are interactionally managed, and how subjectivity/agency is not simply an ideological effect but also a set of micro-managed semiotic/linguistic constructions. In addition, it is one thing to debate whether or not the function of discourse is denotative, ideational, referential, or performative, and another thing to deploy analytic tools to reflect such predilections. It is one thing to theoretically argue that one's level of analysis is the person, or the linguistic utterance, or the interaction, or the broader interpretive repertoires, or the ideological dilemma, and another thing to specify how the analytic focus is situated at that particular level. It is not sufficient to simply point out that there are a myriad of choices and subsequent orientations. It is key to make connections (theory to method) within one's orientation and (if possible) between different orientations (for instance, with Positioning analysis as an attempt to connect CDA and CA). [35]

As a theoretical tool, it is my contention that "positioning analysis" is straightforwardly an epistemological answer to the question of how meaning-making is done. It shrugs off the thorny meta-discursive or ontological question/problems by re-appropriating things formerly thought of as entities (like "interpretive repertoires", "cognition", or "mind") to activities or forms of participation that are always immanent within linguistic practices, not transcendent or independent to them. And most important, through its dual analytic focus on an immanent and indexical account of meaning-making, it is able to illustrate (with data) the theoretical dictum that participants are simultaneously constituting and constituted. Positioning analysis presents an argument for (rather than denying or bracketing, as CA sometimes does) broader cultural discourses as the product of the indexical function of language forms that are both interactively maintained and which have historical and cultural histories of usage. Yet it does not, in line with SCHEGLOFF (1997, 1999), a priori endorse interpretive repertoires as the starting point or focus of analysis. Positioning analysis is a double-edged tool for making sense of the interactive moment. Its focus conceptualizes the interactive moment in a way that is not simply myopic to the on-line encounter only. It is also focused on providing a fruitful way of appropriating the sense of that which is beyond the discursive moment. It thus concretely embodies the insight that the subject is both constituted and constituting. It seeks to ground such explanations in a way that provides a tangible methodological [linguistic] starting point rather than simply scholarly-informed interpretive predilections vis-à-vis this or that interpretive lens. As such, "positioning" is an analytic tool that can be applied at many different levels—from the micro-linguistic level of construction types, to the interactive level of rhetorical strategizing, and to the broader, ideological level of identity construction. It stands, I will argue, as a bold and cross-disciplinary initiative for connecting progressive theorizing with analytic methods (like CA and CDA) that, if reconciled in appropriate ways, can emerge as sound and pedagogically commensurate. [36]


Bamberg, Michael (1994). Actions, events, scenes, plots, and the drama. Language and the construction of part-whole relationships. Language Sciences, 16(1), 39-79.

Bamberg, Michael (1996). Perspective and agency in the construal of narrative events. In Andy Stringfellow, Dalia Cahana-Amitay, Elizabeth Hughes, & Andrea Zukowski (Eds.), Proceedings of the 20th Annual Boston Conference on Language Development, Vol. 1 (pp.30-39). Somerville, MA: Cascadilla Press.

Bamberg, Michael (1997a). Positioning between structure and performance. Journal of Narrative and Life History, 7, 335-342.

Bamberg, Michael (1997b). A constructivist approach to narrative development. In Michael Bamberg (Ed.), Narrative development—six approaches (pp.89-132). Mahwah, NJ: Lawrence Erlbaum.

Bamberg, Michael (1997c). Language, concepts, and emotions. The role of language in the construction of emotions. Language Sciences, 19, 309-340.

Bamberg, Michael (1999a). Is there anything behind discourse? Narrative and the local accomplishment of identities. In Wolfgang Maiers, Betty Bayer, Barbara Duarte Esgalhando, Rene Jorna & Ernst Schraube (Eds.), Challenges to theoretical psychology (pp.220-227). North York, Ontario: Captus University Publications.

Bamberg, Michael (1999b). Talk and what talk is about: The development of form and content in discourse practices. Human Development, 42, 247-252.

Bamberg, Michael (2000a). Language and Communication: what develops?—Determining therole of language practices for a theory of development. In Nancy Budwig, Ina Uzgiris & Jim Wertsch (Eds.), Communication: An arena of development (pp.55-77). Stamford, CT: Ablex/JAJ.

Bamberg, Michael (2000b, grant proposal submitted). Transformations from innocence to responsibility: Identity construction in boys between 10 and 15 years of age. Grant application, submitted to Spencer Foundation, 1999-2000.

Bamberg, Michael (2000c). Critical personalism, language development. Theory & Psychology, 10(6), 749-767.

Billig, Michael (1987). Arguing and thinking: A rhetorical approach to social psychology. Cambridge: Cambridge University Press.

Billig, Michael (1996). Arguing and thinking. A rhetorical view of social psychology (2nd edn). Cambridge, MA: Cambridge University Press.

Billig, Michael (1999). Whose terms? Whose ordinariness? Rhetoric and Ideology in Conversation Analyis. Discourse and Society, 10(4), 543-558.

Davies, Browyn & Harré, Rom (1990). Positioning: the discursive production of selves. Journal for the Theory of Social Behavior, 20(1), 43-63.

Edley, Nigel (2001). Analysing masculinity: Interpretive repertoires, ideological dilemmas and subject positions. In Margaret Wetherell, Stephanie Taylor & Simeon Yates (Eds.), Discourse as data. A guide for anlaysis (pp.189-228). The Open University: Sage.

Edley, Nigel & Wetherell, Margaret (1997). Jockeying for position: The construction of masculine identities. Discourse and Society, 8(2), 203-217.

Edley, Nigel & Wetherell, Margaret (1999). Imagined futures: young men's talk about fatherhood and domestic life. British Journal of Social Psychology, 38, 181-194.

Fairclough, Norman (1993). Critical discourse analysis and the marketization of public discourse. Discourse and Society, 4, 133-168.

Fairclough, Norman (2001). The discourse of new labour: Critical discourse analysis. In Margaret Wetherell, Stephanie Taylor & Simeon Yates (Eds.), Discourse as data. A guide for analysis (pp.229-266). The Open University: Sage.

Gumperz, John (1996). The linguistic and cultural relativity of conversational inference. In John Gumperz & Stephen Levinson (Eds.), Rethinking linguistic relativity (pp.374-406). Cambridge, MA: Cambridge University Press.

Harré, Rom & van Langenhove, Luk (1992). Varieties of positioning. Journal for the Theory of Social Behavior, 20, 393-407.

Harré, Rom & van Langenhove, Luk (1999). Positioning theory. Oxford, UK: Blackwell Publishers.

Heritage, John (1984). Garfinkel and Ethnomethodology. Oxford: Basil Blackwell.

Labov, William & Waletcky, Joshua (1967/1997). Narrative Analysis. Oral versions of personal experience. In June Helm (Ed.), Essays on the verbal and visual arts (pp.12-44). Seattle, WA: University of Washington Press. [Reprinted in Journal of Narrative and Life History, 7, 1997, 3-38]

Ochs, Eleanor (1996). Linguistic resources for socializing humanity. In John Gumperz & Stephen Levinson (Eds.), Rethinking linguistic relativity (pp.407-437). Cambridge, MA: Cambridge University Press.

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

Schegloff, Emanuel A. (1999). Schegloff's texts' as Billig's data: A critical reply. Discourse and Society, 10(4), 558-572.

Uzgiris, Ina C. (2000). Words don't tell all: Some thoughts on early communication development. In Nancy Budwig, Ina Uzgiris, & Jim Wertsch (Eds.), Communication: An arena for development (pp.131-141). Stamford, CT: Ablex.

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


Neill KOROBOV is a doctoral student in the Developmental Psychology program at Clark University. His dissertation work is a sociolinguistic exploration into the production of masculinity among adolescent males, with an emphasis on using Positioning analysis (and a distinctively CDA and CA blend of technique and perspective) in developmentally analyzing the socialization functions of "teasing" and "gossiping" interactions. Broadly, his interests involve philosophical / critical approaches to knowledge construction, hermeneutics, theory, and social scientific methods.


Neill Korobov

Clark University, Psychology Department
Worcester, MA, 01610

E-mail: Nkorobov@clarku.edu


Korobov, Neill (2001). Reconciling Theory with Method: From Conversation Analysis and Critical Discourse Analysis to Positioning Analysis [36 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 2(3), Art. 11, http://nbn-resolving.de/urn:nbn:de:0114-fqs0103119.

Revised 6/2008

Copyright (c) 2001 Neill Korobov

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