Abstract: The access of the non-literal meaning reflects the abstract cognitive language process, which is influenced by various internal and external factors. Chengyu is a special kind of Chinese idioms with their fixed forms and meanings, many of which are compositional. While their non-literal (usu. figurative) meanings are more likely to be comprehended as the salient meaning, the literal meaning of these compositional idioms can also be accessed in some special contexts. Studies have shown that the biased meanings of contexts always influence the access of the dynamic meaning of ambiguous phrases. But few studies have found concerning the influence of context location on idioms’ access. This study was to explore how contexts at different locations influence the agent’s access to the meaning of Chengyu. It involved three contributory questions: 1) As ambiguous phrases, how can the meaning of compositional Chengyu be accessed? 2) Can the access to the meaning of Chengyu be influenced by post-context as well as pre-context? 3) If it is true, what is the difference between the function of pre- and post- contexts on the access to the meaning of Chengyu? This study adopted a self-paced reading test. It was designed in the form of 2 (location of the context: pre-context vs. post-context) × 2(biased meaning of the context: salient/figurative meaning vs. non-salient/literal meaning). Subjects in the test were all college students with Chinese as mother tongue. Preliminary tests were conducted to ensure all the target Chengyu items in the test are compositional, familiar to the subjects and with the figurative meaning as their salient meanings. Chengyu items in the self-paced test were showed first without context and then again with biased pre-context or post-contexts. Subjects were required to read these Chengyu items and identify their first accessed meaning or sentences with Chengyu and identify their dynamic meanings in that particular context. With SPSS 23.0, this study analyzed the reaction times that the subjects used when they pressed the keyboard as they continued their phrase or sentence reading. The statistics showed that semantic access of Chengyu included at least two levels: the activation of multiple potential lexical meanings and their match with the biased meaning of sentence context. In addition, context location had significant influence on semantic access of Chengyu’s figurative (salient) meaning. To be specific, pre-context location might preset the access of Chengyu’s figurative (salient) meaning. It was found that if the context-biased meaning was in accordance with the figurative (salient) meaning, the pre-context might speed up the access of figurative (salient) meaning. If not, it might inhibit the access of figurative meaning, which meant it would take the agent longer time until the literal meaning was accessed. In addition, post-context was found to function in checking the meaning of Chengyu reversely and then rectifying the under-identifying meaning in two ways:1) If the first accessed figurative meaning was suitable to the post-context, then the access process would succeed and the literal meaning would be suspended meanwhile; 2) if not, the post-context would prohibit it and the process would not finish until its literal meaning was accessed. Such conclusion was in accordance to the dynamic self-organization model.

Key words: pre-context, post-context, semantic access of Chengyu, speed up, check and rectification

摘要： 通过自定步速阅读实验研究发现，语境前后位置对成语义识别有显著性影响，且影响方式和程度存在差异，符合多义词重组通达模型。根据动态自组织模式，句子中的成语义识别至少包括多种潜在词义启动和与语境倾向义匹配两个过程。前置语境对成语比喻义通达具有预设作用，加速或抑制成语比喻义通达；而后置语境则在句子语义整合过程中对成语比喻义起到返查纠偏的作用。

关键词: 前置语境, 后置语境, 汉语成语义识别, 加速, 返查纠偏