Enriching CA through MCA? Stokoe’s MCA keys

Introduction

Fifty or so years of Conversation Analysis (CA) has given us insights into issues such as how turns are allocated between speakers in an orderly manner, how turns are the medium through which actions are constructed, how these actions follow one another in coherently structured sequences, how non-lexical vocalizations and visually apprehended activities are part of the warp and weft of talk, and how breakdowns of intersubjectivity are quickly and efficiently dealt with. Where CA has had relatively little to say is about how sociocultural knowledge and cultural understandings can be accounted for in talk-in-interaction. In Stokoe’s terms (this issue), this is about how members’ practices describe and show understanding of the world and the commonsense and routine workings of society, or as Hester and Eglin (1997: 3) put it, the ‘presumed common-sense knowledge of social structures’. These workings may be of social phenomena such as gender/sex, race or other social identities, but also how the little pieces of sociocultural knowledge become part of the local and broader structural organization of talk. It is in this respect that a joining of forces between CA and Membership Categorization Analysis (MCA) may enrich the project of describing and explicating the phenomenon of talk- in-interaction. Whilst, as Raymond and Heritage (2006) argue, finding the ‘mechanisms’ for identifying ‘identities’ has remained elusive, Stokoe has done us the service of proposing a set of tools with which we can at least explore such questions. As a contribution towards this goal, her article is to be very much welcomed.

Categories, and the tools for analysing categories summarized in Stokoe’s keys, surely hold promise for a systematic way of opening these doors. From a CA practitioner’s perspective, this would require that the analytic and methodological strengths of CA not be compromised. Analytic claims need to be rooted firmly in the interactional data. What the work of Stokoe and others (e.g. Butler and Fitzgerald, 2010; Fitzgerald et al., 2009; Stokoe, 2010) suggests is that categories, whilst not omnipresent on the surface of talk, do leave their traces across sequences of talk, through category-tied predicates or categorization maxims for instance. Indeed, these scholars show that it is possible to do a systematic sequential analysis using membership categorization. It may thus well be possible to explicate how categorization is linked to the sequential structures of turn-taking, action sequences, repair or turn-construction features. The aim of this article, in the service of the challenge set down by Stokoe, is to explore how her MCA keys might be used to enrich and extend a CA analysis of a stretch of talk. For this purpose, I use a sequence of talk from my teaching for which a CA analysis has previously been developed.

CA and MCA approaches to an analysis of a fragment of conversation

The first step in this section is to present a CA analysis of an extract from a conversation conducted in a restaurant in Sydney, Australia. The conversation was audio recorded, which means that some potentially important information is lacking. There were five participants, four of whom contributed in this extract. Denise and Colin are an Anglo-Australian couple. Edina, from Germany, and Roberta, from Brazil, are on an exchange visit to an Australian university. One of them is being hosted by Colin and Denise, who are the parents of the fifth participant, Annabel, who does not talk here. The topic of conversation is barbecues. The focus of the analysis will be on lines 17 to 36.

At the beginning of this extract, Colin is talking about how often his family barbecues in summer and the kind of food they barbecue, with Denise adding that he is the one who does the barbecuing. The first task is to provide a sketch of a CA analysis, using some of the CA keys that Stokoe refers to in her article, namely turn-taking rules and transition relevance places, sequence organization, actions and action formation, and turn design.

In lines 17–19, Denise asks whether men do the barbecuing in Brazil and Germany. Of four potential recipients of the question, common sense tells us this one selects two, Roberta and Edina, as the participants know that Roberta is from Brazil and Edina is from Germany. The action-based preference for a question is an answer, and furthermore, Denise’s question is designed for a positive response (Raymond, 2003). Edina’s ‘Yeah’ response satisfies these design principles on both counts, as it is both an answer and it reflects the positive polarity within the question. Beyond this, the answer is type-conforming, in the sense that it answers the Yes–No interrogative with a ‘Yes’. However, the second response to the question, from Roberta, is delayed by about 0.5 seconds (line 23). Roberta also produces a ‘Yeah’ answer, which is preferred on all counts, except that of delay. This delay is long enough for Colin to self-select with a follow-up question to Denise’s, even though only one of the two selected speakers has answered.

Colin’s question in lines 22 and 24–25, like Denise’s, selects both Roberta and Edina: his ‘they’ links back to ‘men in Brazil and Germany’. It also passes over three (or even four) points of possible completion, after ‘virtually’, ‘barbecue’, ‘here’ and possibly the ‘do’ in line 22, each of which, however, is marked by non-final intonation and a straight-through articulation without a break in the flow of talk. This time Roberta alone answers (she may have been selected by gaze). Her response: ‘It’s a man thing’, is constructed to be an answer, and as a positive answer it conforms to the polarity of Colin’s question, and is thus preferred both in action-based and design-based terms – but it is not type-conforming, in that it contains no ‘Yes’ or ‘No’. Note that her answer to Denise’s first question was both quiet and overlapped by Colin’s question, and was thus vulnerable to not being heard. Indeed her answer (‘It’s a man thing’) seems more fitted to Denise’s question (‘Is it the men . . .’), than to Colin’s (‘Is that the only cooking they do . . .’). This may reflect its non-type-conforming design. Roberta’s response might be construed as a way of appearing to answer Colin by its placement as a next turn within the transition space, while actually answering Denise. She is not saying whether the only cooking men do is barbecuing, as in Australia. Roberta’s turn subsequently is acknowledged both by the participant who asked the first question, Denise, and by the second questioner, Colin.

Denise’s question in lines 32–34 is a third question in this sequence of sequences (Schegloff, 2007). It is a reduced, phrasal question, dependent on the previous two questions, and can be seen as a grammatically fitting increment both to her first question and to Colin’s. Again it could be seen as selecting both Roberta and Edina, except what is most contingently relevant is Roberta’s response that follows Colin’s question, which is about men in Brazil, not men in Germany. And it is Roberta who responds, this time satisfying all three design preferences of action, design and type-conformity.

Thus from line 17 to 36, in terms of sequence structure, there are three Question–Answer adjacency pairs (with minimal post-expansions to the middle adjacency pair: Denise’s and Colin’s Yeahs in lines 28 and 30 respectively), forming a sequence of sequences that are intricately tied together. There are features here related to turn-taking: who selects whom to speak, the timing of responses, overlapping talk; sequence organization, with three questions in a sequence closely related topically, each receiving second pair part answers, and each answer reflecting aspects of quite complex preference features. More, of course, could be added to a CA analysis. What more can we gain by adding a layer of MCA analysis? Does it enrich our understanding of what is going on?

Denise’s first question evokes two membership category devices (MCDs): gender and nationality. The former evokes one of the most basic and natural MCDs, men and women. The latter contrasts two members of the 200-plus nations of the world, Brazil and Germany, and compares these with Australia (‘same as here’). It could be argued, though, that merely invoking MCDs does not add to ‘commonsense’ understandings. As we saw in the CA analysis, it is not necessary to appeal to MCDs to establish that Denise’s original question selected the two participants from the countries mentioned. That these categories remained relevant can be explained by tying devices: the ‘they’ in line 22 tying to ‘men in Brazil and Germany’, and ‘with a beer at the same time as they’re [i.e. men] cooking’ in line 32 tying grammatically and through the pronoun ‘they’ to the two prior questions. On the other hand, these two MCDs remain ‘active’ throughout (and indeed beyond) the extract analysed above. Is it possible to examine the maintenance of these devices in a systematic, even structural way? Appealing to the notion of membership category may alert us to other aspects of its organization, and it is here that Stokoe’s keys may help in this endeavour.

Through juxtaposition, Denise in her first question brings these two category devices to bear on each other, and thereby creates a category-bound activity, ‘men in Brazil and Germany cooking on barbecues’. According to the economy rule, the mention of ‘men’ suffices for it to count as a member of the MCD ‘gender’. There is no mention of the other member of this device, ‘women’. A question this raises (and one I cannot answer) is how we know as analysts, without appealing to common sense, that ‘men’ are not being contrasted, for example, with ‘boys’? If we were to stick to CA methodological principles, would we not need the participants to bring into the talk the other member of the invoked MCD if we were to claim an empirical basis for this assumption? Sacks (1974) himself cautioned that most of us come to hear the mommy in the ‘The Baby Cried’ story as the baby’s mommy, but such observations were ‘not proposed as sociological findings, but rather do they pose some of the problems which social science will have to resolve’ (Sacks, 1974: 330), despite his later claim that his observations reflect ‘the fine power of a culture’ (Sacks, 1974: 332). Clearly he is invoking a different domain of the workings of talk from the sequential analyses developed by CA. So how do we move towards sociological – or interactional – findings?

The concept of category-bound activities might help further, and there is one in the extract above: Denise’s ‘men do the barbecue cooking’, extended by Colin to ‘the only cooking men do is barbecue cooking’. Common sense and cultural knowledge tell us that women are bound (in some way) to ‘general’ cooking. ‘Men’ in association with ‘cooking’ can be seen as ‘inference rich’. Women do most of the cooking, men barbecue, which might lead us to infer again that a predicate tied to the category ‘men’ and the activity ‘cooking’ is that ‘men who cook, barbecue’. A corollary of Colin’s question is the that ‘men (virtually) do not cook in the kitchen’, which in turn leads us to a ‘category-activity puzzle’: men don’t cook except when they barbecue, and he recognizes that this may not be a human ‘law’, but may be culturally bound, and thus possibly different in Brazil and Germany (wrong at least for Brazil, as it turns out). We might infer that Colin’s question thus achieves its askability by the fact of it being a ‘category-activity puzzle’ extending Denise’s question. This may also be seen as activities tied to ‘standard relational pairs’ of ‘men–women’, and some of the typical activities we associate with them. The category ‘men’ is maintained across questions, as are the categories ‘Brazil’ and ‘Germany’; Colin does not, as he might have, ask a question about what cooking women do – and why does it seem so unlikely, at least to me, that he would ask here what cooking boys do? Denise’s third question in the series asking if they cook ‘with a beer’ raises another, though highly implicit, categorial association – between ‘men’ and ‘beer’.

There is a consistency here: men barbecuing has been maintained across sequences, and the rich inferences relating to the category ‘men’ are maintained, extended to them drinking beer as they barbecue, and set in cross-cultural and cross-national categories. Denise has mentioned ‘men’, and in the absence of subsequent references to ‘men’ being repaired or challenged, it is at least likely that participants have followed the economy rule that describing a category member once suffices for it to count as a member of the MCD/population of ‘gender’ or ‘sex’. One might also claim that Roberta at least has followed the ‘hearer’s maxim’, that is, that the reference to ‘men’ is one of ‘duplicatively organized category’ of ‘men’ and ‘women’ (gender), and has heard ‘men’ in that category.

A final observation is that Roberta and Denise used very similarly designed turns to assess the categories of men – Roberta’s ‘It’s a man thing’ in line 27 – and of national cultures – Denise’s ‘that’s Australian thing’ in line 37, and Roberta’s ‘That’s a Brazilian thing too’ in lines 40–41. Categories are highly salient here, as well as their associated activities and predicates.

Concluding remarks

A question posed at the beginning of the MCA analysis above asked what it adds to commonsense understandings. It seems that simply identifying categories added nothing much. ‘Men’, ‘Brazilians’ and ‘Germans’ are indeed commonsense categories. However, when the richer array of MCA tools and keys are appealed to and employed in analysis, a more nuanced, less commonsense – and an enriched – analysis of this passage of talk emerged. Notions such as category-bound activities, inference rich categories, category-tied predicates, standardized relational pairs, category-activity puzzles, the economy and consistency rules, and the hearer’s maxim usefully added a layer of understanding, rooted in a close engagement with the data, of ‘the commonsense and routine workings of society’ (Stokoe, this issue), and ‘common-sense knowledge of social structures’ (Hester and Eglin, 1997: 3). We saw how categories played out across sequences, how they were re-invoked in subsequent sequences, and how the MCA keys provided a richer description.

Schegloff (2007: 475) cautioned:

. . . if we want to characterize the parties to some interaction with some category terms, we need in principle to show that the parties were oriented to that categorization device in producing and understanding – moment-by-moment – the conduct that composed its progressive realization. In doing so, we will need to be alert to the ways in which the parties make accessible to one another these orientations, because that is the most serious and compelling evidence of their indigenous-to-the-interaction status.

It seems that there are signs in the work of Stokoe and other MCA scholars that it may be possible to comply with Schegloff’s requirements, though it also needs to be said that it seems there is much to be done to discover how, for instance, inferences about categories work, or how activities are bound by parties in a conversation to categories. However, we have seen in the sequence analysed here how the participants orient to categories and activities in their little exploration of cultural similarities and differences, and the MCA keys provide us with the concepts and language to discuss these cultural features beyond what CA alone is able to do, and beyond mere ‘commonsense’ analysis.