摘要
本文从语言加工视角出发,通过两个在线阅读理解实验和两个离线问卷测试,以汉语母语者和母语为韩语的高水平汉语学习者为研究对象,以汉语反身代词和先行词的回指关系为目标结构,操控先行词的句法位置和有生性语义信息,探讨汉语反身代词的在线加工机制。研究发现,母语语言处理器优先提取近距离先行词,同时对有生性语义信息敏感,后者能够调节句法信息起作用的时间窗口,支持句法优先加工的内容可及性记忆机制假说。和母语有本质不同,二语语言处理器不能有效提取句法信息,也不能提取语义信息,对语篇显著的先行词表现出短暂的倾向性,整体上支持浅层加工假说。
Using two online self-paced reading comprehension tests and two offline questionnaire tests, we examined whether L1 native Chinese and L2 Korean second language learners at a high proficiency level interpreted Mandarin reflexive ziji with the same processing mechanism. Results suggested that the L1 parser first retrieved the local antecedent within the sub-clause and was also sensitive to the animacy feature of the antecedent, the latter modulating the time window as for when syntactic cue could be applied. The results supported the syntax-constrained content addressable memory mechanism. The L2 results suggested that L2 presented a different processing mechanism from L1 in the sense that the L2 parser failed to retrieve either the syntax information or the animacy information. Transient advantage was found in retrieving the discourse-prominent, long-distant antecedent in the matrix clause. In general, L2 results were collaborative with the Shallow Structure Hypothesis.
出处
《语言教学与研究》
CSSCI
北大核心
2017年第1期25-35,共11页
Language Teaching and Linguistic Studies
基金
上海交通大学文科科研创新计划(项目号13QN08)
上海交通大学2016年教改项目“混合式教学法在《英语语言学概论》中的研究与应用”的资助