中文摘要：

以中、英文的多义词为材料，采用跨语言隐蔽启动的实验范式，考察在无语境条件下熟练中一英双语者对多义词的早期识别。结果表明：（1）熟练中一英双语者表现出显著的隐蔽翻译启动效应，当目标词是主要翻译对等词时，启动效应更加显著。（2）翻译对等词的性质、多义词的义项联系程度以及是否是词类歧义词都影响熟练中一英双语者对中、英多义词的早期加工。（3）在无语境条件下，词类信息在意义通达的早期就已经激活，并且同语义信息之间存在着交互作用。根据研究结果，扩展和修正了分布式概念表征模型。

英文摘要：

Polysemy is an important linguistic phenomenon, and it appears widely in every language. The main issue of polysemy research includes whether there is semantic ambiguity advantage and its＇ influencing factors. Studies of polysemy in bilingual research are few in number and are contradictory about polysemy recognition. This study, based on out-of-context condition, discussed the proficient Chinese-English bilinguals＇ recognition process and relative factors using masked priming paradigm. Two experiments were involved： The influence of the dominant and non-dominant translation equivalence, syntactic category information and relatedness of translation equivalence were investigated in Experiment 1 with word decision task and English ambiguous words. The experimental procedure was classical masked priming procedure. Subjects were told to decide whether the word was true or not. All the participants were tested individually on computers. Reaction times for correct responses and error rates were analyzed by subject and item variance. Results showed a significant masked priming effect and semantic relatedness effect in English polysemous word recognition. Furthermore, when words were presented with their dominant translation, the recognition process was faster than when words were presented with their non -dominant translation. Experiment 2 was exactly the same with experimentl, except for the material. The aim of experiment 2 was to discuss how the above-mentioned factors influenced the Chinese ambiguous words. Surprisingly, the results were also similar with experimentl despite the huge difference between Chinese and English ambiguous words. That is to say, We had observed the effect of dominance of translations and semantic relatedness of multiple translations. Responses were faster when the two translations of a word were related in meaning than when the two translations of a word were unrelated in meaning. It is well known that the distributed conceptual feature model has been proposed for bilingual memory.