汉语动宾结构惯用语加工的基本单元：来自词切分的证据

采用词切分范式研究汉语动宾结构惯用语的理解机制, 探讨在惯用语理解中加工的基本单元, 揭示汉语动宾结构惯用语的表征方式。研究发现, 词切分范式对语义分解性不同的惯用语有不同的影响。在空格切分的形式下, 语素切分和非语素切分抑制了对语义高分解的动宾结构惯用语的通达, 却未影响对语义低分解的动宾结构惯用语的理解; 语素切分、非语素切分和空格切分均增加了对语义高分解的惯用语的错误率。在阴影切分的形式下, 非语素切分抑制了对语义高分解的动宾结构惯用语的加工, 同样未影响对语义低分解的惯用语的理解。整个研究表明, 在汉语动宾结构惯用语的表征中, 既存在着语素单元, 也存在着整语单元, 但整语是汉语动宾结构惯用语加工的优势单元。

Dominant Units of Idioms Processing: Evidence from Word Segmentation

Idiomatic language is special in the way of understanding its meaning. The entire meaning of an idiomatic phrase is commonly different from a simple combination of the meaning of its constituents, because the entire meaning of an idiomatic phrase arises from but means more than the literal meaning. Idioms were traditionally viewed as separate lexical units; later on, however, Gibbs and Nayak (1989) proved that individual word also actively contributed to the processing of the whole idioms. Models of idioms comprehension differ in their predictions concerning compositionality: some claimed that idiomatic meaning was the result of compositional analysis initiated at the earliest stage of comprehension, whereas others claimed that compositional analysis occurred only at late stage of comprehension process, subsequent to direct retrieval. Since there is space between words in written English, could there be the possibility that the morpheme segmentation affects the processing of idioms? Existed evidence showed that morpheme segmentation played a role in reading Chinese sentences and passages; we are interested in investigating what role does morpheme segmentation play in understanding Chinese idioms. The morpheme segmentation research paradigm was used in the present research to explore the dominant units within idioms comprehension processing. The results showed that: morpheme segmentation produced various influences on idioms with different semantic decomposabilities. Within the space segmentation format, morpheme segmentation as well as non-morpheme segmentation inhibited the access to adverb-verb structure idioms with high semantic decomposability, but the inhibition was not found in those idioms with low semantic decomposability. We also observed the decrease of reaction accuracy on idioms with high semantic decomposability, which was possibly reduced from segmentation. Within the shadow segmentation format, non-morpheme segmentation inhibited the access to adverb-verb structure idioms with high semantic decomposability, but not to adverb-verb structure idioms with low semantic decomposability. The present research showed that morpheme and whole idiom co-existed in understanding Chinese adverb-verb idioms, but whole idiom manifested large priority in this process.