A Study Of The Cognitive Mechanism Of Chinese Zero Anaphora

This research comprehensively explores the cognitive mechanism of Chinese zero anaphora by combining theoretic study with corpus analysis and cogntive experiments.It is proposed that two major factors account for the cognitive mechanism of Chinese zero anaphora, namely, antecedent saliency factor and structural factor. The antecedent saliency factor is represented in three dimensions:grammatical prominence, semantic prominence, and information prominence. The grammatical prominence of an entity is mainly determined by its grammatical role. On the ground of previous researches, a modified grammatical prominence hierarchy is formulated as follows:"{Subject, Subject Possessor}>Object(Preposed Object)> Oblique". The semantic prominence of an entity is determined by both its animacy and the agentivity of its verb predicate. A verb agentivity hierarchy is formulated as follows: "action verb>{ psychological verb, resultative verb}>linking verb". The higher the agentivity of a verb, the higher the semantic prominence of its subject, the lower the semantic prominence of its object (in active sentences). A notion of "information prominence" is proposed to represent the superior information status of new entities to old entities. An information prominence hierarchy is proposed as follows:"a. indefinite NP > definite NP; b. indefinite NP without modifiers> indefinite NP with modifiers; c. indefinite NP with non-specific number classifier> indefinite NP with specific number classifier "。 Verbs with the semantic function of conveying new information, such as linking verbs, psychological verbs, and resultative verbs, are often followed by entities of high information prominence. The grammatical prominence, semantic prominence and information prominence work together to contribute to the salient status of some entities, which as a result, can potentially be referred to as the antecedents of zero anaphors.The structural factor is represented in two dimensions:distance and inter-clausal correlation. In the discussion of distance, a notion of "cognitive distance" is introduced and the following hypotheses about linear distance and cognitive distance are proposed: ① The accessibility of potential antecedents to the zero anaphor is firstly determined by the cognitive distance between the antecedent and the zero anaphor; ② in case several potential antecedents have the same cognitive distance from the zero anaphor, the accessibility of the antecedents is then determined by the linear distance between the antecedent and the zero anaphor; ③The antecedent is inaccessible to the zero anaphor if the mental path from the antecedent to the target of the zero anaphor traverses more than 2 dominions. On the basis of these hypotheses, an accessibility hierarchy for antecdents of zero anaphors in embedded complex sentences is established. The "local subject> matrix subject> local object" hierarchy is a modification of the previously established hierarchies.In the discussion of inter-clausal correlation, the two inter-clausal correlation types, namely, topic chain correlation and information chain correlation, are associated with corresponding zero anaphora patterns, and their respective linguistic markers are identified. In a topic chain compound sentence, the antecedent of the zero anaphor is usually the subject of the previous clause. And in an information chain compound sentence, the antecedent of the zero anaphor is usually the object of the previous clause. The linguistic markers for topic chain compound sentence mainly include:①correlatives indicating coordination, coherence, progression, alternation; ② syntactic parallelism. The linguistic markers for information chain compound sentence mainly include:① information prominence marker in nucleus constituent; ② static state marker in satellite constituent. The distance factor and inter-clausal correlation factor determine whether a salient entity will finally be picked as the antecedent of the zero anaphor.4 experimental series (including 7 experiments and 8 data analysis tasks) are conducted to verify the influence of the above-mentioned factors on Chinese zero anaphora. The results support the hypotheses. Experiment Serie A, which tests the influence of grammatical prominence on the processing of Chinese zero anaphora, includes two experiments:①Position of Antecedent Strategy (PAS) experiment; ②active sentence and passive sentence experiment. Experiment Serie B, which tests the influence of semantic prominence on the processing of Chinese zero anaphora, include two experiments:①NP animacy experiment; ②verb semantics experiment. Experiment Serie C, which tests the influence of information prominence on the processing of Chinese zero anaphora, includes two experiments: ①indefinite NP v.s. definite NP experiment; ②number classifier specificality experiment. In Experiment Serie D, the influence of distance on Chinese zero anaphora is tested in embedded complex sentences. The data generated from Experiment Serie B is also used to verify the influence of the inter-clausal correlation factor. All experiments are conducted in a paper-pen questionnaire form and in one of the two paradigms of forced choice preference task or sentence completion task. The SPSS software is employed for data analysis and frequency statistics.On the strength of the exploration of the cognitive factors for Chinese zero anaphora, five cognitive principles for Chinese zero anaphora are established, and the generation and interpretation processes of Chinese zero anaphora are outlined in terms of dynamic diagrams.