Action and representation – A comment on Batel and Castro ‘Re‐opening the dialogue between the theory of social representations and discursive psychology’

Abstract

Batel and Castro propose a reconciliation of social representation theory and discursive psychology. This comment highlights the continuing relevance of long‐standing critiques of social representation theory from discursive psychologists as well as their central focus on both how representations are built to appear factual and the role of representations in practices. It suggests that the analytic approaches proposed by Batel and Castro (e.g., focus groups and thematic analysis) are not sufficient to the analytic task. The proposed ‘Pragmatic Discourse Analysis’ falls short on its central task of identifying pragmatics. The virtues of working with naturalistic data using methods that attend to the action orientation of talk and text are pressed.

Background

One of the virtues of Social Representations Theory (SRT) over a range of other social psychology traditions was that it highlighted the significance of representations for action and opened up a less individualistic, more group‐based, approach to key topics such as racism, social change, and collective behaviour. Stimulated by the challenge of understanding representations of UK crowd disturbances in the early 1980s, I hoped SRT would help make sense of the way different representations were built in newspaper reports and individual accounts to assign blame to crowd members, say, or less commonly, the police, and how these might be linked to broader collective processes. SRT ought to have been a resource that supported analysis and provided theoretical context, and yet working with actual materials highlighted its problems over its virtues.

SRT's focus on representation and action is one reason discursive psychologists took the time to critically, but sympathetically, interrogate SRT research, exploring the coherence of its notion(s) of representation, its assumptions about action, and its methodological practices. This work highlighted a range of problems that clustered around the sense in which representations could be shared and how they are used (Potter & Litton, 1985) and the difficulty in coherently applying the notion of a social representation to actual materials – for example media reports of crowd disturbances (Litton & Potter, 1985). Further difficulties with the theoretical and methodological coherence of the notion of representation and its (then) core elements anchoring and objectification were identified (McKinlay & Potter, 1987; Potter & Wetherell, 1987). A series of basic concepts over which there is seemingly unbridgeable disagreement between SRT and discursive psychology (DP) were highlighted by Potter and Edwards (1999): action, representation, communication, cognition, construction, epistemology, and method.

Batel and Castro (2018) are to be applauded for continuing to explore the possibility for making links and building areas of shared assumption. Although their claims and arguments are interesting and creative, let me lay out why I do not believe they succeed and why the early DP critiques indicated above still have traction. First, some comments on the way they define disciplines. It is not beneficial to the progress of social science to blend disparate traditions, concepts, and theories as that risks confusion about methodological and theoretical issues. Theoretical and methodological developments are central to progress; it can come from creative blending of traditions as Batel and Castro propose, or from further precision and refinement which is what I suggest will be more productive here. With that in mind, I will be working from the distinctive tradition of DP recently explicated in Potter (2010, 2012) and overviewed in Wiggins (2017) rather than the broad, mutually inconsistent, spread of perspectives that Batel and Castro gloss as DP. Burr, Burman, Parker, Derrida, Deleuze, Foucault, Fairclough, Walkerdine, Laclau, and Willig are interesting figures but work with a disparate range of epistemological, methodological, and cognitive assumptions; I do not believe it is helpful to suggest they are part of DP. Few of these figures would self‐identify in this way.

Discourse analysis means doing analysis

In DP, representations have been treated as discursive objects embedded in practices – for example descriptions, versions, accounts (Potter, 1996). Analysis has focused on the procedures through which descriptions are built (to appear) as solid and factual. That is, the building of something as a representation is a major focus of DP – and this involved careful attention to lexical items, tropes, grammatical structures themselves embedded in unfolding action. This precision falls through the net of thematic analysis. Discursive psychology has focused on how these (built as) factual descriptions are used to form actions (expressing gratitude, condemning a social group, issuing a threat, and so on). The point about this kind of analysis is that the precise way the representation is assembled (its grammar, lexical items, prosody, real‐time unfolding) is precisely tuned to its role in an activity. Fundamental to DP is the focus on talk as a medium for action (rather than a channel for communicating ideas between one mind and another). This connection of the precise form of representation to the specific action it builds continues to be missing from SRT work. This is partly because of the ambiguous ontological status of representations in SRT (are they mental or embedded in settings? are they linguistic or pictorial? do they lie behind and yet govern action or are they a product of that action?). It is partly because of SRT's continued adherence to perceptual‐cognitivist assumptions – and the mismatch between perceptual‐cognitivist theory and the textual materials and thematic coding that is typical of its analytic practice. And it is partly because the standard methods used in SRT are not only not adequate to that task but actually introduce more confusions over what the representation is, where it is, and what its role is.

Let me illustrate this set of claims with the best‐case example provided by Batel and Castro (2018). They go a familiar social science route and use groups of people recruited to take part in a moderator‐guided conversation on a particular topic. The talk generated is subject to a rudimentary orthographic transcription, and this is searched for ‘themes’ following the suggestions of Braun and Clarke (2006). For example, one theme is ‘ascertaining the essence of rural landscapes/the British countryside versus the essence of energy infrastructures’ of which an example is ‘the atmosphere [of the countryside] …. It is spiritual dare I say, you know, and it's not going to get destroyed by these …’ (ms. p. 14). They assert, but do not demonstrate, that themes such as these are connected to ‘pragmatic functions’ identified through what they call Pragmatic Discourse Analysis. This kind of analysis does not pass the basic tests offered in Antaki, Billig, Edwards, and Potter (2003) that researchers claiming to do discourse analysis do indeed need to do discourse analysis. In their case, there is under analysis through summary (Antaki et al.'s problem 1) and circular identification of mental constructs (in this case representations, Antaki, et al.'s problem 4). It is not clear how Batel and Castro's claims of the compatibility of SRT with DP can be achieved if even basic discourse analysis is not being conducted. Moreover, one of the elusive attractions of SRT was the idea that it would capture something shared across a social group and therefore provide a basis for group behaviour. The ability to show sharing of representations was always a challenge because although the theory specifies sharing representations, the method involves conversational encounters such as interviews, focus groups, and questionnaires. Thematic analysis collects together topics or issues into a single theme, but this seems a much weaker form of collective sharing than Moscovici ((1984) originally proposed.

The relationship between representation and action has been central to DP. The title of Potter (1996) was chosen to highlight precisely that. At the heart of this project was a move from considering cognitive and psychological entities (memories, scripts, heuristics, schemas etc.) to focusing on how descriptions are assembled to be parts of actions. This analytic field has implications for the heart of SRT. It is encouraging that Batel and Castro (2018) wish to continue the dialogue. In my view, they need to build on this dialogue by considering more carefully the concrete, textual and conversational, nature of representations, and understanding them as embedded in actions, sequences, and everyday and institutional practices.

Naturalistic data

Discursive psychology has encouraged searching questions about the adequate and rigorous use of interviews in research (Potter & Hepburn, 2005, 2012). These questions are applicable to the examples that Batel and Castro (2018) offer. The interactional production of the focus group conversation is mostly wiped out by their transcript that excludes prosody, timing and stress, and, crucially, all but hides the moderator's involvement in its production.

The sharp point here is that for social psychologists who take social interaction as a significant part of social life to systematically obscure its role in the production of data is both a theoretical irony and a methodological problem. This irony is not unique to Batel and Castro, of course. Moreover, as is common with many interviews, and focus group studies, the consequential interaction that takes place at the point of recruitment is not captured or reported. Thus, we are not told what categories the participants are recruited as members of nor how the research is characterized to them. The recruitment, the schedule, the questions, and the interviewer uptake flood the research interaction with social science assumptions, categories, and research agendas. This flooding is largely invisible in the fragments that Batel and Castro present from their material and not discoverable in the tables listing Pragmatic Discourse Analysis tools. This invisibility is a product of the common representational practices in qualitative research, which obscure the role of the researcher.

Potter and Hepburn (2012) also highlight the insensitivity of analysis to the footing on which participants are speaking – is the speaker quoted in Batel and Castro's study speaking as a member of the rural population, offering her or his own distinct view, formulating the general countryside view, or some other footing? These are fundamental concerns for discourse analysis because they are critical for how interactants understand one another. Finally, interviews and focus groups (as well as numerous other techniques used by social psychologists) largely presuppose a cognitivist perspective. Individuals are recruited to discover their views, representations, attitudes, and so on. The organization of the interaction, the cooperativeness of the participants, and the selective representational practices of the social psychologists reproduce that presupposition. At the same time, representational practices that wipe out the interactions through which research findings are generated and analytic practices that turn specific claims, accounts, avowals, and other action into generic themes further contribute to cognitivist interpretations. Contrast the loose way in which materials are ‘prepared’ for thematic analysis with the precision required with the conversation analytic discursive psychological tradition (Hepburn & Bolden, 2017).

The discursive psychological tradition relevant here has worked for at least three decades primarily with naturalistic data (Potter & Shaw, 2018). Representations do not need to be elicited through structured, semi‐structured, or unstructured researcher interactions but are central and essential parts of everyday and institutional practices. Ethically secure materials can be collected from the most challenging of situations when the researcher is able to build trust and provide safeguards. Considering representations in their live context is a powerful way of linking them to practices, which is one of the long‐term aims of SRT. In my view, the challenge for SRT researchers is to take a very interesting and ambitious body of thinking and to show that it can make sense of records of actual human life. This might be media reports of crowd events or formulations used in public hearings. The risk of such materials, however, is that the preponderance of long individual monologues or texts can strip off the interactional uptake and response and therefore provide a little restriction on analytic speculation. But considering more interactional settings such as relationship counselling conversations, medical consultations, or family meal conversations would either show the genuine theoretical power of SRT, or more effectively expose its weaknesses. Working with interactional settings would inevitably require a consideration of how representations (accounts, descriptions, versions) are built, turn by unfolding turn in real time, and how their embodied or voiced nature becomes interactionally live, through gaze or prosody say. And this may require further consideration of different modalities as screens, displays, mobile devices, and documents are interactionally relevant. These are not issues that have been addressed in SRT, and that absence is one of the reasons why the traditions of conversation analysis and ethnomethodology have played a more central part in the development of DP.

Potter and Shaw (2018) document the fragmentary attempts by psychologists to build research using naturalistic materials. It seems that social psychologists found that they could not sustain their metatheoretical allegiances while at the same time working with records of real life. Rather than chose to the riskier path of metatheoretical innovation, and of learning how to engage with the world as it happens, with its prosody, institutional specificity, and intricate unfolding in real time, they chose to stay with their familiar methods and theories and fended off the social world. SRT had the promise to move in a different direction, and it could still move in this direction, and yet Batel and Castro (2018) do not follow this path.