| This study approaches the interpretation of zero anaphora, the anaphoric form which is widely used in natural Chinese discourse by sub-classifying 1) degree of topicality and accessibility of the related referents in terms of the different syntactic positions they adopt, and 2) degree of accessibility of the referents in relation to the semantic structural hierarchy of the clauses in which they occur, to facilitate the anaphoric resolution process for zero anaphora in Y. L. Xu's model. Xu's anaphora interpretation model was established on the basis of an interaction between the topicality of the antecedents and the accessibility expressed by the anaphors. Following the general practice in computational processing of natural languages, it operated by matching the topics in the Topic Stack with the anaphors encountered in the on-going process of text comprehension, which was cross-checked by semantic and pragmatic compatibility principles. With a data-based analysis of the syntactic positions and structural hierarchy of the use of 2132 zero anaphors in our data of Chinese text of 58,689 words, it is that the degree of topicality of the antecedents and the degree of accessibility expressed by the anaphors can be sub-classified in terms of the syntactic positions the antecedents and anaphors adopt. Adapting Keenan & Comrie's accessibility hierarchy of syntactic positions to this study, and with an analysis of the different syntactic positions of NPs as potential antecedents in relation to the frequency of the use of zero anaphors, this study proposes a syntactic hierarchy to indicate the degrees of topicality of these NPs as follows:T > S > SM > O > OM > OBLThis hierarchy indicates that the NPs in the positions on the left side of the hierarchy are more accessible and more likely to be referred to with zero anaphors. Second, with an analysis of the syntactic positions of both the antecedents and zero anaphors, it is further found that their positions are related and another hierarchy of the positional relations of zero anaphora is proposed:T-T/S/O > S-T/S > SM-SM > O-O > SM-S > S-SM > O-T/S > S-O> OM-S > OBL-O > SM-OBoth hierarchies can be applied to zero anaphora resolution. Based on these hierarchies, we can interpret an overwhelming majority of the zeros in the data of this study. With an application of the analytic method of structural hierarchy used in Fox and Chen to studying the cases in which zeros occur in syntactic positions different from those of their antecedents, it is found that the majority of the shifts occur in the discourse structure where hierarchical change is observed. It is found that in conjoining structures and those adjoining structures in which the adjunct precedes the nucleus, regular reference patterns can be maintained while in the adjoining structures in which the adjunct follows the nucleus, positional shift may occur. Close studies tell us that it is mainly the adjoining predicate of elaboration that causes change to the syntactic positions of zero anaphors. Finally, an analysis of the relation between structural hierarchy and long distance zero co-reference is conducted to ascertain that the right clause/sentence can be tracked for the identification of antecedents. There are two discourse structures in which long distance reference can take place: the structure in which there is intervening lower-level clause/sentence and the structure in which two levels of parallel clauses are involved. It is also found that in long distance zero co-reference the co-referential NPs usually occur in parallel positions. Incorporating the findings in this study into Xu's model, this study proposes the following principles for zero anaphora interpretation in Chinese text:1) ZA Interpretation Principle Based on Syntactic Positions2) ZA Interpretation Principle Based on Structural HierarchyApplying the first principle, we can correctly predict the antecedents of 1903 zeros, which constitute 89.3% of all the zeros in our data. With the second principle, we ca...
|
PDF Full Text Request