Conversations and Conversation Analysis

Ca's Starting Point: Naturally Occurring, Recorded Interactions

Starting with the data is especially apt for CA because access to specific kinds of data sparked its development. Harvey Sacks, who – together with Emanuel Schegloff and Gail [Page 328]Jefferson – founded CA in the early 1960s, was a Fellow at the Center for the Scientific Study of Suicide in Los Angeles. Consequently, he had access to recorded calls to the Suicide Prevention Center. Today, when mobile phones act as video recorders and we routinely hear that our calls ‘may be recorded for training purposes’, it is difficult to appreciate how significant those recordings were. Then, as Schegloff describes in his introduction to Sacks's (1995: xvii) Lectures on Conversation, Sacks was in the habit of jotting down fragments of talk in a notebook. This reflected his search for a way to study precisely how ‘ordinary activities get done’. When he came across the suicide prevention recordings, he realized they were a tremendous resource because:

Two features of this dataset were profoundly significant for CA's development (see Kitzinger and Toerien, 2009):

• 1.The data were naturally occurring, allowing Sacks to study the phenomenon of interest first hand. The contrast is with data generated for research purposes, such as interviews with people who have called a suicide helpline. These would reveal a great deal about callers' accounts of the service, but not the actual practices occurring within the calls themselves. This methodological shift remains innovative. Even today, interviews and focus groups tend to be the method of choice for qualitative research. Yet researcher-generated data pose some well-recognized difficulties, including the recognition that people, unavoidably, put a ‘slant’ on events when reporting them. This makes it difficult to theorize the relationship between talk about a phenomenon and the phenomenon itself (see Kelle, Chapter 38, this volume). CA avoids this gap by recording the latter (Kitzinger, 2004). This is often taken to mean that CA cannot be used to study interview or focus data. It can. However, conversation analysts treat the data in a distinctive way (see Wilkinson, 2006). The fundamental shift is from thinking of talk as a proxy for, or conduit to, some other event or experience (described in the talk), to thinking about talk as the object of study in its own right.
• 2.The data were recorded, allowing Sacks to study the precise details of the interaction. No matter how carefully you take field notes, it is impossible to write down every detail of a conversation in real time. Coding events during observations fails to solve this problem since, again, the details are irretrievably lost. Equally, no matter when you interview someone, they could never remember precisely what was said. This loss of detail matters because there is substantial evidence that the details matter to interactants. For example, in primary care consultations, Heritage et al. (2007) have shown that one word can make a difference. They compared two question forms used to elicit patients' additional concerns (other than their main reason for the visit): ‘is there ANYthing else’ versus ‘is there SOMEthing else you want to address today?’ In response to ‘anything’, patients were no more likely to express their additional concerns (as recorded in a pre-con-sultation questionnaire) than if the question was not asked at all. In contrast, the use of ‘something’ eliminated 78% of unvoiced concerns compared with no question (odds ratio = 0.154, p = 0.001). The only difference was the change from ‘any-’ to ‘something’.

Similarly, in ordinary conversation, even a tiny pause can be consequential. For example, Levinson (1983) has shown how the silence at line 2 in Extract 1 (shown by the dot inside parentheses; see the Appendix below) is treated by R as indicating a likely rejection of his proposal (at line 1). The evidence lies in line 3: instead of waiting for C to respond, R produces another bit of talk, which conveys an expectation that C may have some difficulty with the proposal. On the basis of only a ‘micro-pause’ (line 2), R has shifted from proposing a plan (‘coming here on the way’) to offering C a candidate reason for rejecting that plan (that C will not have ‘enough time’). C endorses this at line 4.

...