英文摘要 |
In this study, we explore how different components like syntactic knowledge and discourse context may interact with each other in sentence acceptability judgment performance. We examined the acceptability judgment patterns on Chinese topic constructions in different contexts and we also looked at how different types of topic NPs (definite topic NPs vs. bare topic NPs) may affect the judgment patterns. Our findings show that island-obeying topic sentences were consistently judged better than their island-violating counterparts, whether presented in a null context or in a relevant context. This contrast suggests that discourse context and syntactic knowledge do not interact in sentence acceptability judgment performance, and favors a categorical approach to grammar. In addition, our findings also show that topic sentences with a definite topic NP were consistently rated better than those with a bare topic NP, and that canonical sentences which served as the baseline were also consistently judged to be better than grammatical topic sentences. These patterns suggest that acceptability judgment is continuous with a wider range of variations due to the ambiguity, frequency, and processing complexity involved in the given sentences. Overall, our study not only suggests that the traditional acceptability judgment task can elicit good data about grammaticality, but it also recognizes that extra-grammatical factors can affect acceptability judgment performance and should therefore be taken into consideration to ensure the quality of the collected data. |