This study investigates the occurrence/nonoccurrence, distribution, and interpretation of zero anaphora (ZA) in Chinese conversational discourse. It explores three major questions: (1) what factors account for the occurrence of ZA, and the nonoccurrence of ZA in potential positions; (2) whether there are differences in occurrence and nonoccurrence by genre and style; and (3) what linguistic research approaches are most appropriate for the analysis of ZA in Chinese. The database consists of complete dialogues from the two-hour Chinese film, Li, Shuang-shuang, and an uninterrupted one-hour recording of spontaneous conversations among three adult Chinese speakers. The analysis employs a taxonomy adapted from Halliday and Hasan (1976), and Givon's (1989) discourse measures, and involves investigation of linguistic (syntactic and semantic) and extralinguistic (social and cultural) variables, and linear versus hierarchical discourse structure. Difference levels between the scripted and spontaneous conversations were assessed using independent t-tests and chi-square tests. The study revealed the following major findings. First, knowledge sources, linguistic and/or extralinguistic, that were shared by the conversational participants formed the basis for the occurrence of ZA; in the absence of adequate shared knowledge, ZA could potentially be used to attract attention by inviting clarification questions from addressees. Second, overt referents in positions of potential ZA had the function of enhancing referent saliency by specific means including rhetorical, interactional, and stylistic devices. Third, there were great differences between conversational data from this study and narrative discourse from prior research in the proportion of syntactically shifted ZA and in referential distance. Fourth, referent continuity was not subject to linear factors in hierarchical discourse structure. There was a statistically significant difference between the planned and unplanned conversations in: (1) clause type and discourse context in terms of event boundary, role, and person; and (2) referential distance, potential interference, and referential persistence. The overall results of the study revealed the highly variable nature of the ZA phenomenon in Chinese, which suggests the need for multiple perspectives in the analysis of ZA in typologically similar languages, and questions the adequacy of the Government and Binding Theory account of Empty Categories as a universal explanation for ZA.
展开▼
