Abstract
In response to an article by Waring, Creider, Tarpey and Black (2012), the author argues that the nature of the analytic aims of a research project determines whether or not participants’ reported goals and motives are relevant and useful. Her position is that for traditional conversation analytic studies aimed at explicating culturally shared methods for producing conversational actions and for interpreting interactional behavior (i.e. talk, facial expressions, hand movements, etc.), participants’ reports of their goals and motives are irrelevant. She differentiates between explanations based on participants’ reports of their goals and motives and conversation analytic explanations consisting of preferences or principles to which members of a culture orient while interacting.
Waring, Creider, Tarpey and Black (2012) address the question of whether conversation analysts gain any analytic purchase by eliciting reports from the participants of an interaction. The authors’ answer is that participants’ reports can enhance CA analyses in a number of ways, including confirming, specifying, correcting, and disambiguating CA claims. However, the authors do not adequately address the programmatic aims of CA, nor the distinctions between different kinds of explanations. The ways in which these issues are clarified bear on one’s position as to whether or not different types of participants’ reports can enhance CA analytic claims.
Conversation analysts have been engaged in a variety of types of research projects, sometimes singly, sometimes in collaboration with other conversation analysts, and sometimes in collaboration with researchers having other interests and methods. I would consider some research projects as mainstream CA while others as informed by CA methods. Throughout this article, I use the following conception of CA’s aims: to analyze practices for accomplishing conversational actions and for interpreting behavior in interaction, that is, talk, facial expressions, gestures, etc. An analysis of a practice includes describing the procedures for performing an action and the sense-making work associated with it. Ideally, it also includes an explication of how the practice in question differs sequentially and relationally from other ways of performing the action. Furthermore, an analysis should serve as a resource that others could employ for accomplishing the action and for interpreting behaviors as instances of that action.
I chose to contribute to this colloquy for two reasons. First, I want to highlight the importance of keeping one’s analytic aims in focus while making decisions about the value of different kinds of data sources and data. Even with a given data source such as a stimulated recall session, the matters reported on by participants may range from descriptions of the setting, interpretations of interactional moments, character attributions of the participants, participants’ motives, agendas, concerns, interests, and so on. These different types of reports may have more or less value, depending on the project’s analytic aims.
The second reason I chose to contribute this article is that I want to highlight the distinctions between two different kinds of explanations and discuss whether and how each type is relevant for CA analyses. Waring et al. claim that motives, goals, and agendas as reported by participants may be useful in CA as explanations of the participants’ discursive choices. They imply that these types of explanations have some similarity with the types of explanations offered within mainstream CA studies. I take issue with that position and elaborate my views in the Discussion section.
To illustrate the different ways in which participants’ reports enhance CA analyses, Waring et al. provide four examples. For each illustration, they present a CA analysis, provide some points from the interactants’ reports, and discuss the ways the informants’ reports are analytically useful. To illuminate my perspective, I discuss three of the illustrations discussed by Waring et al., using the same format they used. I offer CA analytic claims followed by a summary of key aspects of the participants’ reports, and then I discuss why I think the reported goals and motives do not contribute to the analytic claims. Following that, I discuss the difference between explanations that are based on the reported goals, motives, and explanations of the particular participants and those based on preferences or principles that can be shown to be oriented to by ‘whomsoever’ or ‘anyone’, that is, by persons who share the culture.
In the first illustration, called ‘Writing Center Data’, Waring et al. targeted a turn by the tutor as their focal object of interest: ‘D’y ever read it out loud to yourself?’ With this turn, the tutor started to make a suggestion. More specifically, the tutor’s query served as a preliminary to explicitly making a suggestion. The inquiry was in the form of a candidate answer query in which the candidate incorporated within the query, reading one’s writing out loud, named a legitimate, valued activity (Pomerantz, 1988). The preliminary was formed by asking the recipient if he already was engaged in an activity, that is, the activity that would be suggested. An affirmative response would obviate the need to explicitly make the suggestion, whereas a negative response would provide an opportunity for the tutor to advocate the merits of the suggested activity. The use of the preliminary functioned to allow the speaker to avert giving an unneeded suggestion and/or to minimize the possibility of a rejection. That type of query is recognizable to participants as leading to a suggestion or as implicitly making a suggestion (Pudlinski, 1998).
The tutor provided the following report: she was concerned about the student’s lack of confidence, and she encouraged him to read his writing out loud to himself in order to instill in him the confidence to edit his papers without the tutor’s assistance. With respect to the query with which she started the suggestion sequence (‘D’y ever read it out loud to yourself’), the tutor reported: ‘[it] was designed to check that hunch and to seque into instilling in him the confidence to edit papers without her assistance’ (p. 482). For Waring et al., the interview provided CA analysts with access to the tutor’s concerns and goals, which were at best only broadly inferable from the analysis of the interaction.
Our CA analysis identified a method of launching a suggestion sequence, that is, asking the recipient if he or she already engages in the activity to be suggested. I don’t see how knowing the tutor’s reported goal of increasing the student’s self-confidence in correcting his own writing errors enhances the analysis of the preliminary to the suggestion. While reading one’s own drafts out loud is a well-known strategy for identifying errors, our understanding of the sequential and relational implications of the use of the preliminary query is not illuminated by knowing the tutor’s reported goals or motives.
In the second illustration, called ‘Toddler Tutoring Data’, an educator attempted to gain a child’s compliance to sort the chips by color instead of by size. Waring et al.’s focal interest was on the following turn at talk by the educator: ‘We could collect them in same sizes but let’s collect them- we could collect them by ↑colors ↑too.=We could make pi:les of each color.’ The practices that the tutor used to seek compliance involved maximizing commonalities and affiliation between the educator and the child. The educator first validated the child’s proposal for sorting by size by replicating, nearly word for word, the child’s proposal. She then joined her own proposal to the now validated child’s proposal, putting her own forth as simply another possibility. She concluded the turn by proposing the possibility of the activity ‘making piles of each color’ that would follow their collecting the chips of the same color. The educator reported that her agenda or goal was to have the child categorize the chips by color so as to prepare for a math game that she was planning to have the child play. She also reported having the relational concerns of not wanting to sound authoritative or to disenfranchise the child.
The goal that the educator reported, that is, that the child practice a skill required for a future game, seems irrelevant to an analysis of the practices she used to seek compliance. Whatever her actual and/or reported reasons might have been for wanting the child to categorize by color, they would not give us greater understanding of how the compliance-seeking strategies function sequentially and relationally. The concerns that the educator expressed about not wanting to sound authoritative and not disenfranchising the child seem compatible with the methods that she used to seek compliance. Seeking compliance by establishing commonalities differs from seeking compliance by offering tangible incentives, threatening punishment, etc., and each set of practices carries its own relational implications. The educator’s report of her concerns lets us know that in her case, the practices she employed seem to fit well with her reported concerns. However, it is not clear what that adds to our analysis of the practices that we didn’t already know.
The tutor’s compliance-seeking strategy of maximizing affinities and affiliation was unsuccessful; the child did not comply with the educator’s proposal but continued sorting by size. However, it seems that the educator’s next attempt at compliance seeking involved a somewhat different strategy that had different relational implications. The educator reached for the child’s pile and said, ‘Make your p↓i::le, (0.2) take your pile ’n make l- make- see how high a pile of red ones, okay?’. With that turn, the tutor adopted the strategy of challenging the child to accomplish an achievement. Proposing a challenge for the child sets up a different relationship between the tutor and child compared with the relational implications of the tutor’s earlier strategy of building commonalities with the child.
In the third illustration, called ‘Dinner Data’, a participant reported one experience with a female student and then another experience with another female student that her boyfriend, who was a teaching assistant, had in his classes. Waring et al.’s focal interest concerns the recipient’s stance as contrary to the stance of the teller. In various ways, the teller displayed an amused stance toward both the first and second event that she reported, whereas the recipient’s stance toward both reported events was that the female student’s conduct was inappropriate.
Both the story recipient and the story teller provided reports. The story recipient initially repeated and confirmed the stance he displayed during the original telling that the conduct of the girl in the story was ‘not cool’. However, once it was explained to him that the girl in the second story was a different girl than the one in the first story, he agreed that the second girl’s behavior could be seen as innocent. The story teller reported that she had a long-standing friendship with the recipient and saw his responses to the reported incidents as characteristic of his personality and of the role he typically enacted with her. She reported that the recipient was very opinionated about these matters, took on the role of friendly counselor, and evaluated the lives and relationships of his friends.
Waring et al. suggest that the story recipient’s report serves as a corrective to what was broadly inferable from the CA analysis. They argue that the cause of the misalignment between the story teller and the story recipient appeared to be based on their different standards for what constitutes inappropriate behavior. However, with the discovery of the recipient’s misunderstanding, the authors propose that the participants actually might not have been as misaligned as was displayed in their interaction. Waring et al. say that finding out that the story recipient partially misunderstood the second story serves as a corrective to our inferences regarding why he responded as he did.
I disagree that inferences regarding why a participant responded as he or she did are part of CA analyses. Analytically, it is irrelevant whether the story recipient disapproved of the girls’ conduct because of his standards for moral conduct or because of misunderstanding the second story. The datum consists of interaction in which the story teller’s and story recipient’s stances were non-aligned. Explicating why this story recipient responded as he did is a far different aim than explicating the practices for displaying a non-aligned stance with the story teller’s stance and holding firm on the stance in the face of opposition. As a conversation analyst, I would explicate those practices rather than explicate the reasons that the story recipient held fast to his position.
Discussion
Waring et al.’s article raises the question as to whether a CA analysis should include explanations of the participants’ discursive choices and, if so, on what bases. The authors consider the goals, agendas, and motives as reported by the interactants to be useful in explaining those participants’ discursive choices. We need to go further in thinking about various kinds of explanations and how they fit with different research programs.
Waring et al. suggest that while CA analyses primarily are aimed at documenting conduct, explanations of participants’ discursive choices are inferable from the analyses. Before taking up the issue of explanations, I want to address the authors’ portrayal of CA’s primary program of research. The characterization that CA generally ‘documents conduct’ is misleading at best. Assuming that the authors meant that conversation analysts primarily demonstrate that particular patterns of action exist, that does not constitute CA analysis. Using their terminology, I would say that CA analyses primarily document how actions are performed and interpreted. The primary aim of CA is not to document patterns of conduct, but rather to analyze the resources that participants draw on, resources that have as their product observable patterns of conduct.
Returning to the issue of explanations, Waring et al. claim that both CA analyses and participants’ reports provide explanations (i.e. answer ‘why’ a participant performs an action in a particular way). For the authors, one difference between explanations derived from participants’ reports and those derived from CA analyses is that participants explicitly report their motives, whereas CA analyses are written in ways that allow readers to infer the motives of the participants. The example they use to illustrate that participants’ motives are implied in CA analyses is drawn from my paper on compliment responses (Pomerantz, 1978).
In sum, both CA and talk-extrinsic data address the what (i.e., what is done) and the why (i.e., why X is done). While the CA analysis primarily documents conduct – the what, we can often infer why from that conduct. For example, by showing the different ways compliment responses are done in American English, Pomerantz (1978) infers the competing preference for both agreement and avoiding self-praise. In other words, these preferences explain why the compliment [responses] are done in their particular ways (e.g., praise downgrade or referent shift). (p. 488)
Waring et al. seem to be implying that the individuals who produce praise downgrades and referent shifts have two conscious motives that conflict with one another: they want to agree and also they want to avoid self-praise. Feeling the conflict, they find solutions which take the form of praise downgrades or referent shifts. The explanation of the discursive choices, in this scenario, brings in conscious inclinations, interests, and/or motives.
In the aforementioned article, I discussed the constraints that operate in the environment just after receiving a compliment. I suggested that responding to a compliment with a referent shift (e.g. ‘
In more recent work, I refer to preferences as oriented to principles (Pomerantz and Heritage, 2012). ‘The core idea of preference is that participants follow principles, often implicit, when they act and react in a variety of interactional situations’ (p. 210). Conversation analytic research has indeed proposed a number of principles to which participants orient when they interact, principles which bear on their discursive choices. A few examples of such principles include ‘A speaker has a right to complete a turn’, or ‘Upon being greeted, a return greeting is appropriate’, or ‘Explicit self-praise should be avoided, if possible’. These principles often are implicit; interactants are not necessarily conscious of them and generally would not be able to articulate them. They are background, taken-for-granted, or assumed, and only become salient and talked about when something goes awry and/or makes them noticeable.
In conclusion, I would like to emphasize three points. First, it is important to be clear about the aims of one’s research projects, regardless of whether the project is mainstream CA or informed by CA. Without being clear about the kind of news one wants to deliver, a researcher risks producing an unfocused analysis or analytic bits and pieces that do not cohere. Second, depending on the aims of the study, there may be good reasons to seek participants’ reports, perspectives, and versions of events in addition to capturing interactional data. However, in these circumstances it is crucial to distinguish between the different types of reports being offered and to remember that reports are interactional productions shaped by the context in which they are produced rather than mere representations of the participants’ interests, goals, etc. Finally, conversation analysts need to continue to consider whether they are explaining the conduct of particular individuals or of ‘an anyone’ in the culture and, in relation to that, the kind of data needed to substantiate that kind of explanation. As a conversation analyst, I do not see myself explaining why particular individuals make particular discursive choices. Instead, I would rather uncover shared taken-for-granted principles that help us to understand (and hence explain) the ways participants perform actions and interpret particular configurations of behavior as instances of those actions. And for those projects, I draw on interactional data.