Other-Repair In Semi-Institutional Discourse

This dissertation is a systematic study of other-repair in Chinese TV talk shows (CTVTS) from the perspective of conversation analysis (CA). For this purpose, both quantitative and qualitative approaches are carried out to analyze the 277 extracts of conversation containing other-repairs collected from our data sources, aiming to uncover the relationship between the frequent occurrence of other-repair in CTVTS and the semi-institutional nature of this type of discourse. The major types of analyses that have been made in the present study and the significant findings resulting from these analyses are as follows:Firstly, two basic trajectories of other-repair and the three types of environment of its occurrence have been closely scrutinized and systematically classified. Findings from our analysis of the data suggest that other-initiated other-repairs outnumber self-initiated other-repairs, which is consistent with the results of the studies on repair in Chinese conversation, as well as with the strong skewing found in the English empirical studies. As the present study is more interested in the repair outcome, a different practice is adopted in our analysis, which treats the refusal or rejection of the reparans as indications of"disagreement"or"disaffiliation". We find that the three types of environment in which other-repair arises enjoy nearly equal distribution in our data, among which the"X, Y, Y"format is further divided into four variations reflecting structural flexibility in other-repair. Humor used in other-repair, especially jokes, is for the first time recognized as the fourth independent type of environment, based on the belief that the use of this rhetorical device helps relieve face threat and create entertaining effect in the production of other-repair.Secondly, according to the nature of the trouble source and the linguistic levels at which other-repairs occur, all extracts in the data are formally and functionally classified and statistically analyzed. Findings suggest that the causes of other-repair are mostly attributed to speaking inaccuracy/error, difficulty in lexical understanding and word search gap. As to the linguistic level, other-repair occurs most frequently at the lexical level, especially in the form of providing synonymous alternatives. The less frequent are the syntactic repairs, which either rephrase the problematic sentence or eliminate its ambiguity. This shows that other-repair is largely a process of meaning negotiation to clarify the semantic meaning and achieve the expressive appropriacy.Thirdly, a systematic study of initiation techniques, placements and repair strategies has been carried out. No explicit initiation markers, question/request and sound stretch are three most frequently distributed initiation techniques with different degrees of strength in their capacity of locating the trouble source. As a commonly recognized initiation placement, next turn repair initiations (NTRIs) in our data testify what has been assumed as the most preferable positioning although there exist a small number of extracts which involve third turn initiations and delayed NTRIs. The repair strategies under which other-repair operates turn out to work in concert with the formal and functional classifications in that completion, modification and explanation are most frequently used to deal with the three major causes for other-repair.Fourthly, statistical analyses are carried out to identify the characteristics of other-repair in CTVTS. Two claims about repair in ordinary conversation are reviewed, namely, the relevance of repair to syntax and the preference for self-repair as a universal structure in ordinary conversation. We find that a larger amount of lexical repairs in our data are related to the meaning of the problematic lexical items. The syntactic structure of the problematic sentence, however, is only the superficial representation providing the context for proffering appropriate lexical meaning. Moreover, the large amount of other-initiated other-repairs found in our data indicates that the preference for self-repair as a universal structure can not account for the use of repairs in CTVTS. It also shows that discourse type may affect the preference structure of repair organization. As for turn management, other-repairs in our data follow similar initiation placement patterns with those in English conversation. Multiples are also found in our data, which lends support to their cross-linguistic universality. Last but not the least, our statistical analyses reveal one feature of TV talk shows as semi-institutional discourse, that is, the host-controlled speech event leaves large room for casual conversation. It is this skewing that makes the present study on other-repair better fitted in the analytic framework of ordinary conversation.