Anaphoric dependencies : reflexive binding and null arguments in child mandarin

Abstract

This thesis investigates children's acquisition of reflexive binding and null arguments in Mandarin Chinese, focusing on the interplay between linguistic theory and language acquisition. It provides data on Mandarin-speaking children's interpretation of ziji and their use of null subjects and null objects by experimenting on a large number of subjects, using a range of experimental techniques such as the truth value judgment task, the picture identification task, and the story-retelling task. It is found that children starting from 4 years of age were sensitive to the blocking effect of person feature conflicts between subjects. They were also sensitive to the contrastive discourse context that licenses a long-distance binding of ziji. Nevertheless, they differed from adults in their choices of the antecedents for ziji with respect to their sensitivity to the factors such as grammatical functions, thematic roles, and animacy. It finds that even up to the age of six they still differed from adults in being free from the subject orientation effect when choosing antecedents for ziji, confirming the experimental results of earlier studies (Chien 1992; Chien & Li 1998). The 3-year-old children's responses were different from those of the older children. They were insensitive to the blocking effect induced by person feature conflicts. They accepted the long-distance binding regardless of whether the subjects agreed in person features or not. They had difficulty in interpreting long-distance ziji used in a contrastive discourse context. They demonstrated non-adult-like sensitivity to grammatical functions, thematic roles and animacy in their choices of the antecedents for ziji. The present findings lend support to the hypothesis that children have early knowledge of locality, but have difficulty with the interface of syntactic properties with semantic and pragmatic factors. It is argued that children's sensitivity to the blocking effect reflects their knowledge of locality defined by the governing category whereas their choice of antecedents is regulated by their knowledge of prominence incorporated with syntactic, semantic and pragmatic factors. The results of this study suggest that children's mastery of locality, i.e., the governing category, as evidenced by 4 years of age, follows a biologically programmed maturational schedule (Borer & Wexler 1987; Wexler 1999) whereas their acquisition of prominence develops later as they need more time to learn how to co-ordinate syntactic properties with semantic and pragmatics factors. With respect to children's use of null arguments in narratives, the thesis has two findings. First, children's use of null subjects was constrained by the principle of topic continuity in discourse. They tended to use null subjects to refer back to antecedent subjects, and null objects to antecedent objects. Second, children were sensitive to the animacy effect on the use of overt and empty pronouns. They tended to use null objects to refer to inanimate entities rather than animate or human entities, and such preference strengthened with age.