現代漢語寄生語缺結構與全域結構不對稱表現之句法分析及其對普遍語法之含蘊意義

Since Ross (1967), various researchers have proposed that parasitic gap (henceforth PG) and across-the-board (henceforth ATB) constructions are derived on a par (Williams 1990, Munn 2001, Nunes 2001, Hornstein & Nunes 2002, Bruening & Khalaf 2017, Levine et al. 2001, Levine & Hukari 2006, cf. Chaves 2012). This view is refuted based on asymmetries between these two constructions (Postal 1993, 1998, Nissenbaum 2000, Niinuma (2010). Counterexamples challenging the observed asymmetries, however, have been pointed out (Levine et al. 2001, Levine & Hukari 2006). In this research project, we show that PG and ATB constructions in Mandarin Chinese exhibit clear asymmetries. By Occam's razor and empirical evidence, we argue for a unified approach where PG and ATB constructions are derived by sideward movement (Nunes 2001). The findings include the following: First, PG and ATB constructions in Chinese should be analyzed under a proposal in which multiple gaps at issue are bound by the same antecedents, namely the so-called “shared-antecedent proposal” (cf. Nissenbaum 2000), given that both constructions exhibit variable binding reconstructions effects. Second, in Chinese PGs are subject to more restrictions than ATB gaps. Therefore, they are not entirely parallel. Third, the sideward movement approach as advanced by Hornstein and Nunes (2002) can be extended to capture the facts of PGs and ATB gaps in Chinese. Finally, there is a parametric difference between English and Chinese in the licensing of PGs: the sideward movement responsible for generating PGs is sensitive only to theta-marking in Chinese but is sensitive to both theta-marking and Case-addresses in English. The results of this study lend support for a parametric variation approach to cross-linguistic differences (Huang & Roberts 2017).