| Summary: | 碩士 === 國立臺灣師範大學 === 華語文教學系 === 105 === The sentence processing study in the framework of Competition model reveals that the cue strengths of animacy are larger than word order in Mandarin Chinese, while Good-enough theory denies the effect of animacy, upholding that the shallow structure of "Proto-agent-verb-proto-patient" and plausibility are the strategies that humans use to comprehend the sentence. On the other hand, Shallow structure hypothesis argues that regardless of their proficiency and native language, the second language learner relies on lexical-semantic information during processing of the target language. Their syntactic representation is not as detailed as the native speakers. But Competition model predicts that L2er’s proficiency can be more and more native-like as the increase of proficiency. The current study examines the processing strategy of active sentence, ba-construction, and passive sentences in different plausibility and reversibility conditions by Mandarin native speakers and French L2 learners of Chinese. The results show that Mandarin native speakers utilize the shallow structure of "Proto-agent-verb-proto-patient" to comprehend the sentence. However, the difference of power hierarchy between the two arguments (including contrast of animacy) is found to have a similar effect with plausibility. The difference of power hierarchy and the plausibility can both accelerate the comprehension of difficult structures and interact with the sentence structure. Moreover, the native speakers appear to show a deep processing. These results are not predicted by Good-enough theory. L2ers’ results show that both intermediate and advanced learners use the word order strategy of "Proto-agent-verb-proto-patient" to comprehend Mandarin sentences. But, the lexical-semantic information has greater influence on intermediate learners than advanced learners. Advanced learners can use the word order strategy more consistently. Nonetheless, none of the L2ers' results show an interaction between reversibility and structure and between plausibility and structure like native speakers. Furthurmore, L2 learners not only cannot distinguish the difference of power hierarchy between two arguments, they cannot process differently two actives but structurally different sentence: SVO and ba-construction like native speakers, either. Therefore, the results of L2 learners partly fit the predictions of Shallow structure hypothesis: L2ers’ syntactic representation is not as deep as native speakers. Nevertheless, as the increase of proficiency, learners’ L2 processing strategies are more and more native-like because of using word order strategy and their animacy contrast effect. L2ers’formation of syntactic strategies can be affected by input frequencies and animacy contrast embedded in the sentences in question, which also fits the prediction of Competition Model.
|
|---|
