Conversation Analysis

Conversation analysis (CA) has become the established label for a quite specific approach to the analysis of interaction that emerged during the 1960s in the work [Page 130]of Harvey Sacks and his coworkers Emanuel Schegloff and Gail Jefferson. Its basic interest was sociological—understanding social order. As such, it was inspired by the sociological perspective of Erving Goffman and the ethnomethodology as developed at the time by Harold Garfinkel. Conversation, or talk-in-interaction as it later came to be called, was chosen as its field of application because the creation and maintenance of social order could be studied in detail by inspecting recordings of actual interactions. As it developed, the method of CA was refined and attracted a still growing number of practitioners, not only from sociology but also from linguistics, anthropology, communication studies, and psychology. Its original impetus and way of working have, however, remained essentially the same. What follows is a sketch of the general properties of CA as a unique approach in the human sciences and then an indication of some of its applications.

To explain what CA is, one can characterize the typical CA research process in terms of a sequence of phases. CA research is essentially a data-driven endeavor, so it starts with the collection of data. Researchers in CA work on audio- or videorecordings of interactions that are “naturally occurring,” meaning that they are not arranged or provoked by the researcher as in experiments or interviews. For CA as such—“pure CA”—there are in principle, and often in practice, no further requirements or limitations, although for specialized forms of “applied CA” it makes sense to collect recordings of specific types of situations.

These recordings are then carefully transcribed using a set of conventions developed by Gail Jefferson. Apart from the words-as-spoken, these conventions allow the researcher to highlight a range of “production detail” concerning timing, intonation, and the like that have been proven to be important for the organization of the interaction.

Listening to the recording and reading the transcript, the analyst tries to understand what the interactants are doing “organizationally” when they speak as they do. They may, for instance, be requesting information, offering to tell a story, or changing the topic. Such understandings will be based, first, on the researcher's membership knowledge as, one might say, a “cultural colleague” of the speakers. Second, however, the analyst will check the sequential context and especially the uptake of the utterances in question in subsequent talk immediately following (e.g., by granting a request) or later in the conversation.

However, understanding the actions is not the purpose of the research but rather a necessary requirement for the next step, which is to formulate the procedures used to accomplish the actions-as-understood. CA's interest is organizational and procedural. The ultimate object of CA research is what Schegloff called the “procedural infrastructure of interaction” and, in particular, the practices of talking in conversation. This means that conversational practices are analyzed not in terms of individual properties or institutional expectations but rather as situated accomplishments.

It is often recommended that the researcher approach data with an open mind, that is, without pre- formulated interests, questions, or hypotheses (except the general organizational and procedural orientation). The idea is that inspecting the data in this way will raise an interest in the researcher's mind that can be used as a starting point for a more systematic exploration of an emerging analytic theme. The researcher searches the available data for instances that seem to be similar to the “candidate phenomenon” that inspired the first formulation of the theme as well as data that seem to point in a different direction—the so-called deviant case analysis. It may also be useful to collect new data to expand the analysis. In short, the researcher builds a collection of relevant cases in search of patterns that help to elucidate some procedural issues.

...