Situating stigma: Accounting for deviancy, difference and categorial relations

Abstract This article returns to Goffman's early formulations of ‘stigma’ in outlining a critique of contemporary social scientific uses and abuses of the concept. We argue that whilst Goffman's discussion of stigma is not without its troubles, it has mostly been approached in a manner that treats the concept outside of an appreciation of stigma as a phenomenon of interaction order. More specifically, we discuss and demonstrate how stigma serves an analytic gloss for social relations observable in social settings and in accounts of difference, deviance and degradation. We analyse both social scientific and lay uses of the stigma concept in relation to care‐experienced young children and self‐harm to demonstrate the shared categorisational practices and logics that are often obscured through theoretical treatments of stigma. The recommendation is, then, that an attention to ‘stigma’ in care settings must begin with the conditions in and from which stigma might come to feature as a sense‐making device for all parties.


| INTRODUCTION
The concept of stigma has, since the seminal writings of Goffman, 1 endured a varied career (see Muller 2 ). It is a career recently undergoing a renaissance of sorts which finds stigma enroled as an explanatory concept for an array of social relations and experiences.
It is, however, increasingly applied without the kind of conceptual accuracy that Goffman 3 stressed 'students of society' should strive for. Critiques, readings and revisions of Goffman's work in relation to stigma are routinely produced without regard for the project that Goffman was pursuing; the delineation of a situational sociology that took seriously the interaction order, in its own right. 4 Anthony Giddens 5 once wrote that he was setting out to rescue Goffman from his fans. In this article, we are not quite setting out to rescue Goffman from his critics, but we do intend to make serious use of Goffman's formulations of stigma as a matter of social relations. In that sense, we contribute to arguments for understanding stigma as a thoroughly interactional and situational matter. 6 In contrast to dominant contemporary treatments, stigma ought to be treated as belonging to interactions and settings, not a discrete 'thing' that is attached to individuals. Stigma is a product of interaction or, to put things another way, an accomplishment. 7 In doing so, we do not analyse clinical interactions themselves, but, instead, describe some of the work that the stigma concept is put to-as an explanatory device and theoretical gloss in formal social science analyses and by lay members -in making sense of difficult, shaming or degrading experiences.
In attending to how stigma is talked about and 'used' as an explanatory device, we pay particular attention here to membership categorization practices. [8][9][10][11] Membership categorization practices are, among other things, the methods though which members of a given scene come to categorize-and thus make sense of-fellow members. Members' selections of relevant categories are thus foundational for social organization. As such, it seems to us that such practices must be fundamental to the interactional production of stigma. The import of the approach is that, rather than assuming that certain categories are relevant and 'bound' to stigma at all points, we argue for an attention to the detail of interaction and people's actual orientations to categories which, in turn, reveals the grounds upon which stigma is produced.
In what follows, we describe some of the 'categorial logics' present in both social scientific and lay uses of the stigma concept. Our first set of examples are drawn from social science publications. Here, we show something of the heavy lifting that the stigma concept is made to do. Our second set of examples are drawn from interviews with carers of young people, and their accounts of those young people's treatment, and exclusion, by others. Here, we show how the stigma concept is deployed to make sense of the troubles of those with experience of being in care. In drawing from both formal social scientific and lay uses of the concept, we demonstrate a family resemblance in the work the concept is put to; work that is grounded in the same categorial logics.
The pay-off of our discussion is recognizing that troubles of shame, degradation, exclusion and so on, deserve careful attention which, perhaps ironically, can be obscured by using stigma as a catchall gloss. We begin with a different order of trouble, and a discussion of the career of the concept of stigma in mainstream social scientific usage.

| Stigma trouble: recovering and topicalising Goffman's formulations
Over the course of its career-a career increasingly independent of Erving Goffman's original formulations-the concept of 'stigma' has become widely diffused. The term itself, of course, predates Goffman's discussion. It is a term 'out there' in society which, through its very familiarity, causes issues of clarity and consistency when applied in a social scientific context. It is a 'natural language' concept par excellence. In his use, Goffman 1 makes a careful delineation between the common usage of the term and the work he puts the concept to in his own writing, emphasizing that 'it should be seen that a language of relationships, not attributes, is really needed' other words, we need to understand that any given attribute is not necessarily stigmatic in own right but becomes so in and through situated social relations.
Despite the apparent clarity of Goffman's opening remarks, the career of the stigma concept has been shaped by various forms of sociological amnesia, misunderstanding, and misapplication. We point readers to Müller's 2 overview of treatments of the stigma concept 'after Goffman' and find ourselves very much aligned with his assessments; not least his insistence that one really needs to read the whole of Stigma to understand, properly, what Goffman is up to in those first few, possibly misleading, chapters. Goffman's entire sociology does not proceed from the point of individual social actors on the one hand and societal structures on the other. 4 Contemporary writings on stigma critical of Goffman (e.g., Tyler 12 ) trade in the same 'logic of exteriority' 13 as Giddens' earlier critique. In other words, in attempting to 'repair' a perceived gap between Goffman's social actor and the historical and structural orders that he was apparently 'blind' to, much of this work reproduces a false relation between 'structure' and 'agents'; a relation not recognized either in Goffman's work or interactionist sociology more generally. 14 Outside of such a treatment, 'stigma' becomes a catch-all concept for relations between 'normal' and 'deviant' categories. As various critiques of contemporary usages have noted, it is enroled to gloss relations and practices of shame, degradation, 'self-stigma' and other interactional phenomena worthy of careful treatment. 2,6,15 Conceptual clarity is not quite our concern here. We are, instead, concerned with how stigma is invoked as an explanation without consideration of the production of stigma in actual situations. Stigma is treated by social scientists as a thing of the world and, at the same time, a conceptual device for finding order in the world (in the manner critiqued by Garfinkel 7 and the ethnomethodologists). It can be found operating as an elevator concept, 16 adding a conceptual lift to mundane observations. 'Stigma' is similarly used as a placeholder concept, marking a conceptual, analytic space without the obligation of filling said space with necessary detail. Stigma is also, we suggest, a bidet concept. This likely needs further explanation. Years ago, Paul heard someone ask what the function is of a nonexecutive director in a company. The reply to the question was 'It's a bit like a bidet. It adds a touch of class, but nobody knows how to use it'.
All of these (metaphorical) usages have something in common in allowing an author to do several things. A primary use is an affiliation of their work to a very broad range of sources; it offers professional membership in linking the work to some canonical texts to which readers can knowingly nod. Another is the deployment of 'stigma' in providing for a 'black-boxing' of key analytic elements of their argument, in the very course of their objectivation. 17 The practices which produce the context in and through which an account of 'stigma' is meaningfully produced are obscured by the very terms which serve to describe it. Such is glosses' work.
For reasons of brevity, we can only summarize general trends in recent stigma literature. They can be organized through the following observations: 1) Stigma is treated as a 'thing', attached to and sometimes even equivalent to a person. This means that the emphasis on specific social settings is lost 18 2) Stigma is approached as a directly describable subjective 'experience'.
3) The recognition that categories are situated and always occasioned is ignored in an insistence that attribute X is a stigmatizing predicate of social category Y. SMITH ET AL. | 891 4) Stigma is given agency, and is seen to 'play a role', or to be 'promoting a view'. This position in relation to the analytic category 'deviant' reminds us that Goffman was attempting to shift the attention from the study of 'deviants' to an understanding of situated rule-breaking which, again, is grounded in an understanding of the rules, demands and obligations that hold in each social setting. 19

| Beyond elevators, placeholders, and bidets
Admittedly, Goffman's own treatment of 'stigma' is not perfectly consistent. Nor is it unproblematic. It should, however, be remembered that attributes glossed as discrediting or discreditable can only In working out the properly situational approach outlined by Goffman, we think that some of the insights of Harvey Sacks,20 and particularly as they were developed into what became 'membership categorization analysis' 8,9,11,21,22 are useful here.
Drawing, albeit sparingly here, from this tradition, a first point is that-contra to most social science treatments-people do not straightforwardly 'belong to' this or that category, but, in a far more radical sense, the availability of a relevant category (and category device) organizes and accomplishes a person as a member of category, within a specific social setting and context. A second point is that any member can be categorized by an almost endless inventory of categories. That this is the case appears to be trivially obvious, and yet the question of how person-description categories are produced and selected and made relevant in any case is a deeply analytic one for members and sociologists alike. As noted by Coulter, 23 the analytic task is to demonstrate how a category becomes 'operationally relevant' in some scene, rather than rely upon its apparent perceptual assignability. A third is an alternate treatment of culture (which can be said to 'produce' the norms, values and rules which in turn create 'normals', 'deviants' and 'stigma'), as culture-in-action. 8 A situational treatment of stigma turns on a proper treatment of how categorizations are actually organized in any given setting. As suggested above, the stigma concept provides for the glossing of this categorial work routinely engaged in by members and social scientific analysts alike.
When it is said, for example, that 'he might have got some help, if only there wasn't so much stigma around talking about mental health', we can readily hear that there is a relation between the category 'male', the activity 'seeking help for mental health', and the assumed perception of that category of person engaging in that activity, in the eyes of some community (presumably 'other men', in this case). Thus, practical reasoning regarding social relations is necessarily categorial in character. Goffman,[1][2][3][4][5][6][7][8][9][10][11]   'people' (which is hearable as non-members of either category). These interactions are accounted for as the breaching of some rule, or the inability to meet certain demands and expectations, for which the category 'mentally ill' might have some mitigating or explanatory effect. This kind of categorial trouble is not necessarily equivalent to stigma, and yet is made to stand in for it in the methodologically ironic social scientific 'correction' of mundane practical reasoning.
This, again, establishes the 'normal/stigmatized' two-set category pair 20 which is present in the account and the analysis by the authors.
Stigma as a concept is deployed ahead of the talk, to make a formal analytic matter of the relatively mundane, selected, remarks of the participant which, in our reading at least, have more to do with the speaker's reported troubles in making relevant the category 'mentally ill' in unspecified interactions with others.
To repeat the main point of the article, if we are to take social relations describable as 'stigmatic' seriously, then we must recognize that rolling out stigma as an elevator, place-holder, or bidet concept obscures the actual lived detail that produces both the affective experience of a 'stigmatic' encounter, as well as related practices of degradation, discrimination and devaluation. The stigma concept, deployed as a catch-all, simplified construct, can obscure the lived detail of the complex social relations that shape care-experienced young people's encounters.
In the same paper, 24(p9) another participant produces an account grounded in this same categorial reasoning, which closely matches the sociological use of the concept: One youth explained her understanding of stigma, and described how people with mental health are divided and labelled: The overall picture is stigma can really separate you from the rest of the population. In terms of you're a crazy person, and these are normal people. And like, it's not something that, that type of stigma is a little more discrete. It's kind of, nobody will say that to you, outright, it's just how you are treated, and there's no faith in you. (Participant 5) Here, we can note that the stigma concept can be deployed in the construction of contrasts that invite interactants to find relevant differences, and to make consequential moral evaluations. Contrastive devices are characteristic of everyday and professional talk.
Smith 25 for instance analyses a description of the behaviour of another ('K') that invites the hearer to find particular instances as evidence for eccentricity or even mental illness. Contrastive talk is also used by professionals to make implicit, or even explicit, comparisons between their own practice and the work of others (found to be wanting in some way). Atkinson 26 documents such contrastive talk on the part of American haematologists, who distinguished their own diagnoses with the apparent shortcomings of physicians 'elsewhere'. Stigma has been explicitly addressed in this vein by discursive psychologists. 27 They argue, building like us on Garfinkel and Goffman, that the analytic resources of discursive analysis need to be brought to bear on 'stigma-in-practice'. Attention to the discursive work of invoking and attributing 'stigma' may be some distance removed from Goffman's original conception of spoiled identity, but attention to how stigma is used as a discursive resource remains faithful to the general spirit of the sociological and social-psychological origins. It is an approach that still resists the dilution of stigma as an analytic concept.
In the follow section, we point to an alternative approach to analysis of 'stigma talk' and highlight the shared modes of (categorial) reasoning employed in social scientific and lay accounts.

| Talking stigma: members' constructions and uses of a concept
In further pursuing the categorial logics that underpin the use of the stigma concept, the following data excerpts are taken from a study of foster and residential carers' experiences of preventing and managing self-harm among care-experienced young people, conducted between 2015 and 2018. 28,29 The excerpts below are drawn from a focus group in which in which the mental health and wellbeing of young people with experience of being in care is discussed. Rather than treating such accounts as providing direct and measurable access to stigma as an object, we aim to demonstrate how 'stigma' is presented as a resource for mundane and practical reasoning relating to difference and deviance and, in this way, how 'stigma' is intersubjectively understood and 'objectified' within the talk, as an accounting device providing for generalized cultural expectations. Here, the interviewer works with the two-set category pair 'deviant/normal' to put the idea on the table that 'normal' behaviours (i.e., actions that are 'category free' in that they could be done by anybody), are recategorized as 'self-harm' in and through the availability of the category 'looked after'. This suggestion is treated, with apparent caution (the 'umm' [l. 17]), with a revision which, again, first produces a relationship between the category and an attribute (of being naughty) that is 'out there' as thing that has conceivably being said, before moving on to invoke stigma as a product of that relationship with consequences of (again) a partition established between the 'normal' and the 'deviant'.
Whilst we, of course, recognize that what we have been working with here is abstracted interview, we do suggest that the materials provide for an insight in the categorially-based reasoning that underpins both lay and professional sociologies of stigma. This is, in the absence of access to clinical interactions themselves (described across the rest of this special issue), one way in which to avoid a key critique of the mainstream treatment of deviance and difference 30 ; that, on close inspection, 'labelling accounts' reveal an overwhelming preoccupation with generalized descriptions of cultures of understanding, often elicited from interviews in the field, allegedly brought to bear upon interactional scenes.

| CONCLUSION
In this article we have argued for an attention to the work that the stigma concept does in social scientific and lay discourse. We do so in order to demonstrate that 'stigma' is not a 'thing' in itself, but arises in social interaction and can be used as a device for making sense of problematic social relations. In this way, we have approached We then moved on to consider the uses of 'stigma' as a concept in both professional social scientific and lay sociological reasoning. In both analyses, we aimed to at least point to how 'stigma' is routinely deployed as a gloss with which members readily make sense of a whole range of matters of membership, belonging, exclusions, identity, shame, motivation, action and in action and so on. In what could only be an illustrative analysis, we have also pointed towards how this practical reasoning is categorially organized. We have not had room to fully elaborate this here but suggest that a necessary critical engagement with Goffman's treatment of stigma can be found in the ethnomethodological treatment of categorization practices.
What we have hinted at, but not at all developed is how the discussion of the 'stigmatized' and 'normals', and case-specific variants thereof, can be shown to be routinely embedded in both social scientific and lay analyses, accomplished through the recurrent construction of a two-set category class. As Harvey Sacks observed 20 ; 'Two-set classes would seem to have certain kinds of attractions… they're tremendously easy to compare… you can apparently make an observation of comparative lack much more easily than otherwise'. Sacks goes on to note that all sorts of social arrangements, from conflicts and revolutions, to games, are organized in terms of two-set classes, the 'haves' and the 'have nots', and speculates that formulating social relations as a two-set class appears to be 'a method of doing things' that might be significant as a device that forms 'a basic mechanism of social control'. Whether and how that is the case requires further examination. Either way, the orientation to 'stigma' outlined here, finds the 'machinery' which SMITH ET AL. | 895 occasions (discussions of) stigma, located not in social structure or exterior historical process but, rather, as existing in the mundane, diffuse and endogenous categorization practices employed in production of members' accounts of the organization of any given scene. That is a step towards working out the 'language of relations', which, in turn, does not rely on the analytic use of 'stigma' as a comforting and convenient conceptual gloss. In terms of the degree to which a social scientific analysis might contribute to improving encounters in clinical settings, this step may go a long way toward recovering what is submerged in standing treatments of 'stigma'.