Submitted by
for the degree of Master of Philosophy
at The
University of Hong Kong
in October 1998
| This thesis is concerned with two major issues
in Conversation Analysis: turn-taking and overlapping talk. For the first
issue, the validity of the turn-taking system proposed by Sacks, Schegloff,
and Jefferson (SSJ) (1974) is examined against a body of original conversational
data. In connection with the second phenomenon, questions are asked about
the extent to which overlaps occur in the data, and what interactional
tasks are relevant when overlaps occur. The research is based on 161 minutes
of audio-recordings of conversations among friends in Hong Kong. The language
of the conversations is Cantonese.
The study of turn-taking in the data has confirmed the validity of SSJ's system for Cantonese conversation. It is found that a turn in Cantonese conversation consists of one or more turn-constructional units (TCUs) which can be a word, a phrase, a clause or a sentence. Turn transition tends to take place at the completion point of a TCU. Turns are usually allocated by a set of three rules: current speaker selects next, second speaker self-selects and current speaker continues. Overlaps are frequently noticed in the data. They are found regularly at four positions in the course of an on-going TCU. Overlaps may take place slightly (not more than two syllables) before the unit's completion, exactly at the point of completion, slightly after this point, and in the middle of the unit. The first three positions are around a place where turn transition tends to occur. The occurrences of overlaps of these positions are accountable by the turn-allocational rule-set, while onsets of overlapping of the fourth position are largely associated with the syntactic structure of the overlapped utterance. A number of interactional tasks are relevant in overlapping. From the perspective of current speaker, overlapping may take place as a result of his continuation after a TCU's completion. This extension of the turn may serve to hold the speaking right, to make self-repair, or to minimise the transition gap from one turn to another. From the point of view of next speaker, he may start speaking before the completion of the on-going utterance to express agreement, to assist the development of the utterance, or to show understanding. He may also cut in to disagree with the first speaker, to compete for the speaking right, or to initiate repair. Overlaps also come about as a product of topic management. In three-party conversations, overlaps may involve two recipients’ simultaneous responses to the prior turn. |