Submission to Discourse Analysis Online

Discourse Analysis Means Analysing Discourse: Some Comments On Antaki, Billig, Edwards And Potter 'Discourse Analysis Means Doing Analysis: A Critique Of Six Analytic Shortcomings'

Erica Burman

Discourse Unit/Women's Studies Research Centre
Department of Psychology and Speech Pathology
Manchester Metropolitan University
Gaskell Campus
Hathersage Road
Manchester M13 0JA

Abstract: In this paper I discuss the 'six analytic shortcomings' of discourse analytic work identified by Antaki et al. as concerned with contextual and part-whole relations. I then move on to offer three more addressing questions of location: under-analysis through uncontested readings, under-analysis through decontextualisation and underanalysis through not having a question. I suggest that, while Antaki et al. have usefully highlighted some prevalent limitations on current, especially introductory, work put forward as discourse analysis, their analysis benefits from some further elaboration in order to acknowledge and refer to the wider spectrum of discursive approaches.

Keywords: template, DAOL submissions

1. Introduction

Antaki et al.'s paper offers welcome illustrations of weaknesses that I recognise only too well within some contemporary (including student) purported discourse analyses. My comments here arise from a general sense of sympathy and agreement with their arguments, and are offered in the spirit of supplementing rather than supplanting them. I will therefore be drawing on, and then elaborating on, the framework provided by their paper to offer three more examples of forms of analysis that fail to fulfil claims to discursive analyses (or what Antaki et al. calls 'non-analyses').

Antaki et al. declare two reasons for highlighting these problems: firstly, to 'help those who approach DA enthusiastically, but in an environment where there is less support than there would be for more traditional methods of analysis, and so less opportunity to test and refine methods among sympathetic colleagues' (p.15); and second, 'to scotch the sort of errors that give comfort to the traditionally-minded who accuse DA of "anything goes"' (ibid.). While (as I will discuss later) there are many responses to the 'anything goes' argument (see also Burman, 1990, Burman, 1991, Burman, 1992), my previous attention to the 'errors' or shortcomings of novice engagements with discursive and qualitative research has focused on how these highlight with particular clarity ideological as well as conceptual and methodological features of the discipline, and contests within the discipline, that researchers new to the arena are labouring to join (Burman, 1996, Burman, 1997, Burman, 1998).

Like Antaki et al., my comments here are similarly offered to lend support to, as well as to strengthen the profile of, the wide range of analytic research that currently qualifies itself as discursive. As will become clear, my discussion diverges from that of Antaki et al. in suggesting, firstly, that they do not go far enough in identifying limitations of currently circulating forms of discourse analyses; secondly, that, paradoxically, the restrictions on the critical account they offer potentially work to proscribe forms of discourse analytic work that I would want to see supported; and thirdly and finally, that their argument bolsters a limited notion of discourse (as transcribed interview text) that unduly forecloses the political as well as analytic project of discursive research. I will start by revisiting the six problems identified by Antaki et al., and then move on to highlight three more that I would want to add. I should reiterate that I am fully in agreement with their desire to emphasise 'the analytic basis to discursive studies' (p.3). Indeed equivalent concerns motivate my account here.

2. Contextual Questions

The first problem they discuss is where a summary or descriptive account of the transcript seems to be offered as a substitute for an analysis. Antaki et al. suggest that this approach 'will lose information and add none' (p.8), including being 'likely to lose the detail and discursive subtlety of the original' (ibid.). Whilst in full agreement with the specific examples of this they provide (of how the summary strips away important cues provided by the ordering of the account and paralinguistic features included within the transcription conventions), two further points might be added.
Firstly such 'summarising' might also inadvertently work to decontextualise the analyst's own account. Not only would this be to return its claims to those more familiar parameters of positivist and objectivist psychological research. It would also thereby be masking the political position adopted by the commentator in her or his framing of the account. I will return to this point shortly in relation to the second 'non-analytic' strategy Antaki et al. discuss. On their first point, though, there is a further issue, which is: where is the analysis of the analytic framework? Any analysis (discursive or otherwise) has to be undertaken in relation to a declared set of theoretical presuppositions as well as specific questions generated in relation to these, which provide the basis on which the analysis can be evaluated. Any 'summary' proffering decontextualised truth claims simply bolsters common sense or, worse still, expert (for example, 'psychological') knowledge. While Antaki et al.'s transcript, and discussion of transcript conventions, offers some useful examples and techniques, they too assume an analytic framework that is not declared.
The second form of 'non-analysis' characterised by Antaki et al. is where the analyst's opinions or political commitments substitute for the analysis. This is said to 'lead to the sort of simplification that is the antithesis of analysis …produc[ing] a flattening of the discursive complexity, as the analyst selects quotations for the rhetorical effect of appealing to the readers as co-sympathises or co-scolders' (p.10). This recognisable weakness is particularly worthwhile to point out. Here, though, I would want to add two further points, the first of which is that it can also be a problem to presume that one can avoid 'taking sides' or engaging in strategies of 'enlistment' (p.8): both through the form of the analytic framework that is adopted, as well as through the form of language in which the analysis is presented. Whilst clearly (now claiming my own 'expert' status!), as I have indicated elsewhere (Burman, 1991, Burman, 1997 and Burman, 1998) reflexive analysis does not substitute for analysis, nevertheless the 'solidarity/hostility' or 'sympathy/scolding' dichotomies noted by Antaki et al. (on p.9) are inevitably to some extent present within the analytic account - albeit that, as feminist analyses have indicated, our malestream academic trainings have perhaps schooled us into failing to recognise lack of commitment as a subjective position (c.f. Henwood et al., 1998; Hollway 1989). Objectivity is not the absence of subjectivity but a particular form of it. Put simply, there is no way of avoiding adopting some kind of position. The question therefore is rather which, and on what grounds this is evaluated.
Secondly, just as we may show an inappropriate 'solidarity' or 'sympathy' (or indeed presuming the transparency of the account by claiming access to 'feelings', 'beliefs' or 'views') by (in Antaki et al.'s examples on p.9) discussing how a speaker 'realises' or 'appreciates', sometimes the language of 'stating', 'claiming' or 'goes on to speak about' which are less 'value-laden' descriptions that we are familiar with in discursive work - as labelling speech acts rather than intentional states - also produces 'rhetorical effects'. But these 'effects' are of an ironizing character that in some ways - especially to those new to discourse analysis and so importing their everyday language practices - appears to devalue the speaker's account because it implies that something else is being manifested through it. Now this of course illustrates some of the humanist objections to discursive work that in my view are usually misplaced and arise from a mis-conceptualisation of the purpose of discourse work - which is not to focus on individuals but rather the cultural frameworks of meaning that they reproduce. However, my point here is that we need to take seriously how the tools of our own discursive practice inevitably speak of their own assumptions, and failure to attend to these can lead us back towards precisely the kinds of decontextualised and objectivist claims to knowledge that discourse work in psychology was formulated to critique (cf. Burman and Parker, 1993; Burman et al, 1996; Parker, 2002).
Finally, while Antaki et al. take pains to point out their recognition of a diversity of positions among discursive researchers as to the desirability of taking sides, my arguments here would invite a further qualification that 'enlistment' is not in itself a problem, and that 'taking sides' is not the same as under-analysis. Whilst all analyses require the kind of detailed examination indicated by Antaki et al., it could further be argued that under-analysis occurs when the analysis substitutes detailed examination of the text for the adoption of a theorised position.

3. Questions Of Part-Whole Relations

Antaki et al.'s fourth point concerns under-analysis though over-quotation or isolated quotation, both of which are familiar within weak examples of discourse analysis. Yet there is a danger that we may be closing off the potential range of interesting and relevant discursive analyses if we offer without qualification technical norms about the 'ratio of analyst's comments to data extracts' (p.10). What about approaches to discourse analysis that include a performative or poetic character, in which the analysis is explicitly selectively crafted towards a specific audience for a specific intervention? An example of this would include the research recently completed by Burns (2001) of performing back to a collective audience of her activist participants the edited (and thereby analysed) versions of their own accounts. This functioned not only as a form of poetic presentation but also of organisational analysis, that also fulfilled commitments of making the academic research accountable to the researched. Much feminist and action research draws on equivalent (if possibly less dramatic) forms, and discourse work surely has a key contribution here in making interpretations explicit and accountable and thereby making use of its intervention.
Further, the 'profiling' process that Antaki et al. identify as underlying several of the under-analyses they discuss, including where quotations are 'pieced together' (p.11) from different respondents, is said to potentially 'impede analysis by removing utterances from their discursive context' (ibid.). Here we need to pause to consider what conceptual framework defines the 'discursive context'. Opinions between discourse analysts on this matter vary significantly. But perhaps we could agree that such a context should be explicitly identified.
Antaki et al.'s following discussion of '[t]he circular discovery of a) discourses and b) mental constructs' (p.11) helpfully illuminates some familiar limitations of under-analysed discourse work, and my concurrence with their points is already indicated in my comments above. However, it might be useful to include a further 'analytic extra' (p.11) element to provide a way out of the 'circularity' of warranting an interpretation via simply re-describing what is said in the text. This would be to elaborate the analysis or categories to relate to structures outside the detail of the text, for example via analysis of institutional practices and systemic patternings. So, far from merely being designated a theme (p.8), 'gender inequality in marriage' might be drawn upon as one of various institutional analyses under interrogation throughout the interview transcript. Clearly discursive analysts vary in their claims as to what lies inside and outside the text (see e.g. Parker, 1998, for a review). Notwithstanding this, I would suggest that proper analysis should not only 'consult the relevant previous research on all these conversational moves and apply the accumulated insights to the present data' (Antaki et al., p.14), but that this also includes consulting theoretical analyses of a historical and cultural kind that inform how such conversational moves come to be possible, and how they function. How are we able to recognise the theme of 'gender equality in marriage', for example, except by drawing on common sense and other contested cultural resources in circulation around us as embodied and historically and culturally located analysts.
As regards the charges of 'mentalism', perhaps it is helpful to remind ourselves that some such varieties of mentalism will always inevitably creep back into our accounts. Antaki et al. themselves impute mentalism when they write of discourses being 'drawn upon to deal with specific features of the conversational interaction' (p12) and that the 'respondent knows that they are expected to … in order to avoid appearing dogmatic and to gain recognition' (p.14). Indeed perhaps the issue is not only to struggle to avoid mentalistic categories of analysis (since they return us to the traditional individualist psychology we have tried to escape) but rather to attend to the kinds of mentalism being imputed. The framework implicitly put forward by Antaki et al. imputes a mentalism of a dramaturgical, but nevertheless voluntarist, kind that retains the dichotomy between individual agency and determinism. Unless supplemented by other (less voluntarist) accounts of the construction of subjectivity, this offers little alternative to more traditional psychological approaches.
The sixth under-analysis discussed by Antaki et al. arises through 'false survey'. There clearly are dangers of generalisation from the specificity of one discursive context to others (reminiscent of many discussions about the relevance and reformulation of the criterion of 'representativeness' as a criterion for evaluating qualitative research). Here it would be useful to add that there are equal dangers in failing to situate the text within the cultural-historical conditions that gave rise to it. For this, one needs an analytic framework that permits the conceptualisation of how the talk (if such is the text) has arisen in terms of broader institutional practices. Here two further points might be made about Antaki et al.'s account. Firstly, that it seems somewhat paradoxical to have been offered access to an MP3 recording of the interaction, yet to have been provided with no information of the circumstances in which the interview was conducted, nor the research question (or other purpose) that gave rise to it. Here I do not mean to be unfair - for I think it is perfectly acceptable that the material is offered as a medium from which to illustrate under- or non-analysis, without actually providing an analysis. Rather my point is that the generosity of contextual framing we are offered is nevertheless still partial. To this end I would suggest that contextual information regarding the historical moment, cultural setting, institutional position of the interview, and gender (as well as other structural positions) of participants would be more informative with respect to generating and evaluating an analysis than hearing the recording (cf. Burman et al., 1996). Perhaps we are drawn into the metaphysics of presence of in seeming able to 'replay' the interview? But surely without this other information aspects of the contextual specificity of the 'original' moment are obscured, while what is added to it by the process of iteration is equally left out of the analysis.
Secondly, and perhaps this is an aside, but I was surprised by the attribution (on pages 14 and 16) of 'unconscious' motivations to account for the implicit or artefactual character of these analytic difficulties. Surely, just as with its psychoanalytic application, such attribution implies an individual origin and motivation, rather than an institutional analysis accounting for how such difficulties so readily arise (of the role of methodological technologies in psychology, for example).
Antaki et al.'s final example of 'under-analysis' is characterised as 'spotting'. As they note: 'research does not and should not, consist principally of feature-spotting, just as analysing the history and functions of the railway system cannot be accomplished through train-spotting' (p.15). The point that 'feature-spotting' may well indicate some acquaintance with particular literatures and techniques of analysis, but does not in itself constitute the analysis, is particularly valuable. Yet even here the critique itself takes the technology of vision (and its metaphorical resources) for granted. This threatens to naturalise - or take for granted - which 'features' there are to 'spot', and thereby leaving out of the investigation the very conceptual frameworks (optics? Surveillance? Etc…).that specify such items and render them 'visible'.
They conclude this point by arguing that 'good analysis always moves convincingly back and forth between the general and the specific'. There can be no disagreement with this, yet a key word here that invites a further discussion and disputes over analysis is 'convincing'. What is 'convincing' to whom, and why? 'Conviction' here would seem to stand for a whole set of discussions around criteria used to justify analysis that would be a key task for the analyst to address. Further, what is meant by 'the general' here? This leads me to put forward some further issues.

4. Three Further Points: Questions Of Location

Having addressed myself to Antaki et al.'s six key points, in part by extending them, I move briefly now to add three to them. A seventh would be under-analysis through uncontested readings. Rather than formulating a monovocal account, good discursive analyses acknowledge the multiple and contested character of the interplay of discourses by showing how different discursive representations are built to interact with and ward off others. They have their own integrity of categories (in the sense that subject positions are elaborated in complementary ways) but these are frequently contested (if only to refute and reassert them) within any text. Highlighting such contests provides some internal reliability regarding the function and effects of any analysis, and the better discourse analyses that I have more recently encountered have formulated, or else demonstrated consideration of, several alternative ways of conceptualising the discursive domain under analysis.
An eighth proposition would be under-analysis through decontextualisation, whether this decontextualisation is on the part of the analyst in their account, or of their representation of the text. This point connects with my comments about lack of engagement, by which I mean the need to situate the text and one's analysis of it socially, historically, culturally and politically. While it is clearly logically impossible (as well as probably undesirable) to claim to identify the 'whole context' of the text, the issue is to provide both a rationale for it as a meaningful text to analyse (how it has come about, why it is important, who has which kinds of stake in it, and why and how), and to indicate the stance from which the analysis is conducted. Doing this helps to ward off the incipient objectivism that dogs so much psychological research, and (as I discuss in Burman, 1992, Burman, 1998), it also wards off the corresponding alternative (but equally mistaken) position of subjectivism. To use Antaki et al.'s phrase 'insightful and technically sophisticated work' (p. 16) can also, and I would argue inevitably, includes being politically committed and institutionally located. The alternative to 'anything goes' is surely not 'nothing goes'.
My final - but perhaps most important - addition is perhaps one that was assumed or taken for granted by Antaki et al., but I think is still worth making explicit: under-analysis through not having a question. The most uninteresting and weak examples of discourse work that I have encountered principally founder through the failure to specify why this analysis is being done, and is worth doing. Just as under-analysis, as Antaki et al. so well highlight, arises through presumption and premature selection of quotations and correspondingly unwarranted interpretation, it can also arise through an insufficiency of critical selectivity on the part of the analyst. In this respect the production of the data is actually necessary to the analysis, and we might expect innovative forms of data collection to always operate already analytically.

5. In conclusion….

Antaki et al. have provided six useful and illustrations of unnecessary limitations within current examples of discourse analysis. I have here commented on and elaborated upon these six to propose a further three (and doubtless the list could continue). Perhaps the list of 'shortcomings' needs to get longer, or even to become less linear and instead wind itself into a spiral, or double helix, or preferably some more open-ended kind of structure - so that evaluating particular advantages and limitations of each analytic strategy or point depends not only where you are, but also when...
Antaki et al. seem to have made efforts to formulate their account to connect with and be relevant to the broad range of discursive work currently in circulation in and around psychology and the social sciences. Perhaps in trying to be inclusive, however, and notwithstanding the acknowledged diversity of their own positions, they have inadvertently become prescriptive by omitting from their analysis attention to the specificity of their own framework for discourse analysis. My comments here are thus intended to ward off a potentially restrictive reading of the valuable points they have made that could work to close off forms of analysis that I would want to admit as relevant contenders for discourse work. Hence while Antaki et al. refer themselves primarily (though they make clear they do not intend this to be exclusionary) to forms of discourse analysis focusing on written text, (more specifically connected talk, and even more specifically transcription of interviews), their account could give the impression that there are actually prescribed features to (in their terms) be 'spotted'.
Moreover the model of discourse analysis they presume treats this as a form of reflection rather than action, in a way that potentially restricts the space of analysis to that of academic production of text rather than to other practices of socially-oriented accountability. Here I am thinking of the 'practical deconstruction' (cf. Parker et al., 1995) style of discourse analysis that occurs within mental health and other forms of political activism concerned with contesting the definitions and practices of contemporary social policies, as well as feminist work, such as that of Burns' (Burns, 2001) mentioned earlier.
Not only, then, does 'doing discourse analysis mean doing analysis', but discourse analysis means analysing discourse. One has to have a theory of discourse (or text or transcript) as well as of analysis to do discourse analysis - and this also includes having an analysis of the technologies of one's own analysis.


This paper has benefited in particular from comments by Paul Duckett, Angel Gordo-Lopez and Ken McLaughlin on an earlier draft.


Antaki, C., Billig, M., Edwards, D. and Potter (2002) 'Discourse analysis means doing analysis: a critique of six analytical shortcomings', Discourse Analysis Online

Burman, E. (1990) 'Differing with deconstruction: a feminist approach', in I. Parker and J. Shotter (eds) Deconstructing Social Psychology. London; Routledge [cited]

Burman, E. (1991) 'What discourse is not', Philosophical Psychology, 4, 3: 325-342. [cited] [cited]

Burman, E. (1992) 'Feminism and discourse in developmental psychology: Power, subjectivity and interpretation', Feminism & Psychology, 2, 1: 45-60. [cited] [cited]

Burman, E. (1996) 'The crisis in modern social psychology and how to find it', South African Journal of Psychology, 26, 3: 135-142. [cited]

Burman, E. (1997) 'Minding the gap: positivism, psychology and the politics of qualitative research', Journal of Social Issues, 53, 4: 785-803. [cited] [cited]

Burman, E. (1998) 'Disciplinary apprentices: "qualitative methods" in student psychological research', International Journal of Social Research Methodology, 1, 1: 25-45. [cited] [cited] [cited]

Burman, E., Alldred, P., Allwood, R., Goldberg, B., Gordo-Lopez, A., Heenan, C., Marks, D., Marshall, J., Taylor, K., Ullah, R. and Warner, S. (1996) Psychology Discourse Practice: from regulation to resistance. London: Taylor & Francis. [cited] [cited]

Burman, E. and Parker, I. (eds) (1993) Discourse Analytic Research: repertoires and readings of text in action. London: Routledge. [cited]

Burns, D. (2001) The Single Parent Action Network UK: an organisational analysis of 'grassroots', 'multiracial', 'participatory' practices, Unpublished doctoral thesis, Manchester Metropolitan University. [cited] [cited]

Henwood, K., Griffin, C. and Phoenix, A. (eds) (1998) Standpoints and Differences: essays in the practice of feminist psychology. London: Sage. [cited]

Hollway, W. (1989) Subjectivity and Method in Psychology: gender, meaning and science. London: Sage. [cited]

Parker, I. (2002) Critical Discursive Psychology. London; Palgrave. [cited]

Parker, I. (ed) (1998) Social Constructionism, Discourse and Realism. London: Sage. [cited]

Parker, I., Georgaca, E., Harper, D., McLaughlin, T. and Stowell-Smith, M. (1995) Deconstructing Psychopathology, London; Sage. [cited]