This study explores talk, and what talking may amount to, in terms of social organization. It was carried out in a Swedish Youth Detention Home specializing in assessments. The analyses draw on conversational data from multiparty conferences, so-called network meetings, involving the assessment of ten girls/young women. Apart from the young person herself and her family, the detention home staff, social workers and sometimes other persons also took part in the meetings. One chapter also analyzes ten research interviews conducted with the girls/ymmg women. The epistemological framework of the thesis rests on several discourse-oriented approaches to the analysis of talk, notably on the interrelated research programs of ethnomethodology, conversation analysis and discursive psychology, all of which advocate a practice-based program of social inquiry. A central research problem deals with how institutionality may bear on and infonn identity descriptions and the interactional work through which different identities are collaboratively constructed. In tenns of institutionality, the present study highlights some formalizing devices, which orient the interaction towards particular organizational goals. In addition, a systematic production of iriformality was observed to facilitate the organization's work, and a locally relevantpreference stmcture is introduced to account for these findings.

Important parts of the meetings proceeded according to set agendas and, for someone who was not directly addressed, considerable effort was required to take a turn at talk. Here laughter was found to be a particularly useful tool for the structuring of interaction, as it provided both lay and professional participants with opportunities to participate meaningfully in the flow of talk without actually expressing much through words. The analyses focus on the fine-grained and delicateaspects that comprised social interaction at the juncture of distinct institutional projects, highlighting the production of contrastive versions, which cast the young person as either accountable or not accountable for past events. In this sort of competition or politics ofrepresentation, participants could, for example, be seen to "do nonnality" as a situated accomplislunent of talk. In sum, the thesis highlights various devices used to manage institutional categorizations and identities, invoking different notions of accountability and different notions of normality/deviance. In Swedish discussions concerning directions for social work, the virtue of involving members of the clients' social network in the assessment and treatment work has beenstressed. Yet, the present fmdings show the ways in which such meetings are highly complex. The participants need to orient to various subtle contextualization cues. In this complex undertaking, the monitoring work of a chair is crucial, especially when the girl/young woman under assessment is present at the meeting, listening to different versions of ascribed problems and identities that concern her past and future.