ACKNOWLEDGMENTS
Earlier incarnations of this article benefited from comments from John Heritage, Celia Kitzinger, and Geoff Raymond; all weaknesses are my own.
Notes
1 Conduct-in-interaction is short for talk-and-other-conduct-in-interaction. CA's definition of conduct includes vocal and nonvocal behavior, the use of artifacts, and any other feature of context that participants can be shown to orient to (vocally or nonvocally), including identities, roles, and relationships.
2There isa distinction between a practice and a practice of action, the former being used to build the latter. For example, multiple practices of turn design, lexical choice, intonation, and sequential position frequently get orchestrated, in context-sensitive ways, to achieve single practices of action. Participants are concerned with the product of this orchestration in terms of action, whereas analysts are concerned in addition with the practices used to build actions.
3Lest we become overly pessimistic about the reliability of single cases, re-consider the first observation made about Extract 1. From prior research (CitationLabov & Fanshel, 1977), we know that the HV's “you were a nurse at thuh churchill.” (line 3) implements a practice called a B-event statement. Although space limitations prevent a full analysis, the husband's and wife's responses (lines 4–5) stand as data-internal evidence for this practice.
4There is only audio data for Extract 1. It is possible to “dream up” (plausible) reasons why the husband responds first, such as that his wife is occupied with their baby (note, though, that this reason is contraindicated by the fact that the wife responds first at line 8 and, given the observations made about Extract 3, does not account for why the husband responds immediately at line 11). However, such reasons are not empirically demonstrable, and if analysts choose to include audio-only data in their core collection, Extract 1 would be an unexplainable disconfirmatory case.