Zero anaphora in Chinese spoken and written narrative monologues

This study investigates the discourse-pragmatic characteristics of zero anaphora (ZA) in Chinese spoken and written narrative monologues and the role modality plays in the use of ZA.;The written corpus consists of 8 fictional short stories by contemporary Chinese writers. The spoken corpus is comprised of 14 interview narrative monologues selected from two Studs-Terkel-style publications. A pragmatic-functional approach is adopted for the research. Based on the characteristics of ZA, this research focuses on how hierarchical structure, density of discourse entities and information packaging requirements interact with modality in motivating the use of ZA.;The study finds that ZA in Chinese are of two distinct types: accessibility-conditioned and negligibility-conditioned. ZA serve two discourse functions. They are used as a cohesive tie among clauses that are structurally close. They are also used as an information packaging device to highlight clausal focus and reduce linguistic redundancy.;The research confirms that the hierarchical structure of discourse influences the use of ZA. Zero pronouns are used in clauses that are closely related structurally even when there are intervening clauses in between. On the other hand, minor breaks in the inter-clausal coherence features such as shifts of themes, states, actions and goals prompt the use of overt pronouns despite the continuity of reference. Moreover, high density of minor, negligible discourse entities may prompt the use of ZA to avoid explicitly introducing entities into discourse and reduce linguistic redundancy.;The analysis shows that modality contributes to the higher frequency of pop-returns by ZA in the spoken narratives. Its influence manifests itself in the correlation between the need for after-thought explanations and comments and the time constraints on speech production. Modality also plays a role in the greater number of negligibility-conditioned ZA. The greater need for negligibility-conditioned ZA arises from a participatory pattern and the narrative structure that correlate with speech rather than writing. On the other hand, the use of ZA to indicate minor coherence breaks and rhetorical emphasis and contrast is not found sensitive to modality.