Testing a model of adult second language acquisition of Mandarin wh-questions.

摘要

The area of investigation of this study is the Second Language Acquisition (“SLA”) of Mandarin interrogatory pronouns (“wh-words”) by two groups of adult language learners whose first languages (“L”) are English and Japanese.; This study adopts a syntactic description of the wh-words of Mandarin, English and Japanese to develop an SLA Model for the two learner groups. From the SLA model a set of predictions is derived and tested. The syntactic model adopted as the basis for the SLA Model is the Unselective Binding Model of Mandarin wh-questions (Tsai 1994, 1999). This syntactic model assumes that Mandarin wh-words are variables bound by an operator.; The research design tests the prediction that the acquisition by the learner groups of non-interrogatory binding of wh-words by the overt universal operator dou or the covert operator in so-called Bare Conditionals will cause the learners to recognize that Mandarin wh-words are bindable as nominals, except for the adverbial how/manner and why/cause readings of Mandarin how (zenme yang) and why ( weile shenme), which are not bindable and for which there is no testable syntactic input. This change in the interlanguage settings is predicted to be reflected in the learners' proficiency in interpreting wh-questions embedded in strong syntactic islands, where movement would be ungrammatical because of island effects, but where binding of nominals is grammatical.; The major findings of this study are that the predictions derived from the SLA Model are largely fulfilled by the data from the English-speaking learner group, while the data from the Japanese-speaking group shows signs of transfer from the L1 grammar. This is indirect empirical support for the Unselective Binding Model as a syntactic description of Mandarin wh-words. In addition, the Bare Conditional operator seems to provide more useful input to the English-speaking group than the dou operator. This difference is not derivable from the syntactic model nor predicted by the SLA Model. The lack of acquisition of the adverbial readings of Mandarin how and why is reflected most clearly by a lack of statistical relations between the learners' proficiency in interpreting embedded wh-questions and their proficiency in non-interrogatory binding.