A conversation analytic approach to interaction: methodological reflections on disruptions of the common-sense world.

Certain schools of phenomenological psychiatry conceive of as a pathology of common-sense. Ethnomethodological enquiry, with its roots in Schutzian social phenomenology, takes as its domain, topic, and substance of study the ongoing achievement of a common-sense world between social members. Yet, dialogue between psychiatry and ethnomethodological approaches is thin. In this article, we discuss a conversation analytic approach to interaction which has generated and utilized a model of a five-world manifold to frame analyses of talk-in-interaction. 'Worlds' are conceived, after Schutz, as finite domains of meaning, and the model operates as a breach of natural attitude assumptions to examine mechanisms of the constitution of the one-world-in-common of common-sense. It is suggested that certain aspects of talk might receive account in terms of a loss of integration between these five domains of meaning. Conversation Analytic methods were applied to transcripts of audio recordings of psychiatric interviews but encountered hurdles that motivated the broadening of methodological scope. Such hurdles included a weakening of the next turn proof procedure, implicit reification of the construct, and problems of translation presented by the analyst's normative membership encountering non-normative life-worlds of experience. Strategic responses to these hurdles included exploring linkages between phenomenological psychiatry and ethnomethodological approaches, as well as an engagement of ethnomethodological self-reflection and conceptual clarification of the construct in line with Garfinkel's unique adequacy requirement. The manifold model is glossed, and interaction between two of its worlds - a world of concrete, situational immediacies and another of abstract organizations - is explored in more detail via analysis of conversational data. It is suggested that the five-world model, along with further micro-analysis of talk-in-interaction, might have implications in psychiatry for topics such as autism, double bookkeeping, concretism, theories of disturbed indexicality, and insight attribution. We conclude that the consideration of atypical interaction obliges the interaction analyst to take account of their own implicit normative world-frames and that the use of domain-specific top-down models in conjunction with the inductive approach of Conversation Analysis may extend the reach of CA to facilitate productive dialogue with other disciplines.