Introduction

According to Sacks, Schegloff and Jefferson (1974), smooth turntaking in conversation is based on participants' recognition of certain stretches of talk as ‘turn-constructional units’, the completeness of which occasions the possibility of turn-transition. The turnallocation component of the turn-taking system, assigning turns according to certain ordered options to another or the same speaker, thus depends crucially on the ‘visible’ production of such turn-constructional units. It is these units that determine turn-transition places.

Sacks, Schegloff and Jefferson remain somewhat vague – as does subsequent conversation analytic research – about the structural bases according to which turn-constructional units are recognized. They seem to conceive of them basically in syntactic terms (as ‘sentences’ or smaller syntactically independent structures). The notion of syntactic closure is left up to linguists to investigate. At the same time, the role of prosody (intonation) is mentioned in determining turn-constructional units. From research on gaze, it is additionally known that turn-yielding is regularly indicated by speaker-gaze at the recipient as a possible (intended) next speaker. It also seems obvious that semantico-pragmatic aspects of completeness enter into the recognition of turn-constructional units as well. For a non-speaking participant in a conversation to know where speakership may change, i.e. when it may be ‘his (or her) turn’, it is therefore necessary to monitor on-going speech production together with its accompanying non-verbal activities in a very comprehensive manner, taking into account not only syntax, but also, minimally, prosody, gaze and the content of the utterance against the background of what is being talked about.