How Transcription may Influence Analysis: The Case of Silences

Silence as a human interactional phenomenon has rarely been studied in the scientific community. Notable exceptions are the works included in Tannen and Saville-Troike (1985),which study silence from a sociolinguistic and ethnographic perspective; and the seminalwork by Sacks, Schegloff and Jefferson (1974), where silence is a phenomenon occurring between turns at talk, and may thus signal a problematic interaction if the silence is deemed too long. Thus, on the one side, silence is a universal feature of human conversation, on theother, it has different meanings in different cultures.The present paper aims to show that the way analysts transcribe silence influences the ensuing analysis of the data. Generally silences are transcribed either with dots between parentheses, distinguishing only between a short or a long pause; or by milliseconds in parentheses. Neither of the two notations says much about what that silence is doing in that particular moment and in that particular interactional environment. In particular, when dealing with long silences – 8 to 9 seconds – the general assumption is that silence may signal a trouble-source in the preceding turn in the conversation.Drawing from a corpus of real videorecorded data of institutional interactions (lawyer-client counseling) it will be shown that, at least in these sorts of interactions, participants are engaged in other courses of action which are simultaneous to the ‘main’ one, that is the verbal interaction. Long silences are ‘occupied’ by the manipulation of objects and documents that are at the center of the attention focus (Goodwin, 2007), even though not accessible to everybody (as, for example, in Mondada, 2006a, 2006b, 2006c, 2007): one document is accessible to the lawyer only, who can turn pages, write on them and read aloud or to himself,or scan through it; the client on the other hand, is engaged in retrieving a text message, which will be taken as one of the evidences for applying for divorce. The transcription process is,thus, an account of the theory subsumed (Ochs, 1979): transcripts will take into account all the action, taking place during the interaction, which is significant to the participants.Drawing from seminal works by Goodwin (1980, 1981) and following developments by Mondada (2006a, 2006b, 2006c, 2007), a transcription notation has been devised which accounts for the notation of concomitant courses of action which are either simultaneous to talk or to silence.