Late assignment of syntax theory : evidence from Chinese and English
The attraction of the well-structured arguments of the mental syntactic processing device (parser) in Chomsky’s theory has led to an overemphasis on syntactic processing to the exclusion of semantic and other processing in the initial sentence processing stage (Frazier & Clifton, 1996; Gibson & Hickok, 1993; Pickering & van Gompel, 2006). The current thesis joins some others (Green & Mitchell, 2006; MacDonald et al., 1994; Townsend & Bever, 2001, etc.), investigating the timecourse of the information processing of sentences. The first interest centres on ambiguous sentence resolution. Crosslinguistic studies have shown different resolutions in processing the relative clause (RC) attachment as in “the servant of the actress who was on the balcony” (Cuetos & Mitchell, 1988). Three studies confirmed that there is an NP-low preference in Chinese; however, this effect was delayed in comparison to its English counterparts. The NP-low preference can be explained by syntax-first, syntax parallel, and syntax later theories. However, the delay effect questions the traditional syntax-first theories. This leads to the second investigation of direct comparison of the timecourse of syntactic and semantic processing using anomalous materials in English and Chinese. Two experiments have confirmed that the syntactic anomaly is recognised later than semantic anomaly in both languages. The empirical investigation in the current thesis used various methodologies, including self-paced reading, a questionnaire, and eye-tracking studies, where the design of materials strictly followed linguistic principles. All the results support the late assignment of syntax theory (LAST) (Townsend & Bever, 2001). In fact, LAST can explain most of the evidence for syntax-first and syntax-parallel theories, and it is in line with the latest development of the linguistic UG theories (the Minimalist Programme).