| This study investigates the effect of bilingual lexicosemantic representation of Chinese EFL learners on lexical access in translation by means of an experimental design, as well as comparative and descriptive methods. It is an interdisciplinary attempt to connect translation studies and bilingualism studies.The research questions of this study are as follows:(1)What are the structures of bilingual lexicosemantic representation for Chinese EFL learners of different proficiency levels?(2)Whether is there any effect of different types of bilingual lexicosemantic representation of Chinese EFL learners on lexical access in the translation? Namely, what features do they show on the speed of word retrievals and error rates they have made in different word translation directions?This study randomly picks up 30 students whose scores for College English Test-Band Four are below 50, from two classes of first-year non-English majors and 30 third-year English-major postgraduates whose scores for Test for English Majors-Band Eight are above 70 as representing two groups of Chinese EFL learners of different proficiency levels, to participate in the picture-naming task and word translation task, in the hope of obtaining the first-hand data for the research.Quantitative and qualitative analyses of the data collected revealed major findings as follows:(1)Language proficiency is related to bilingual lexicosemantic representation. In other words, Chinese EFL learners of different English proficiency levels show different types of bilingual lexicosemantic representation. That is, learners with lower L2 proficiency level show word association representation, and those with higher L2 proficiency level show concept mediation representation in their bilingual memories.(2)Different types of bilingual lexicosemantic representation of Chinese learners are immune to difference in forward and backward translations.(3)The asymmetry exists in the translation process. In the light of the analysis of the data, learners of two different types of bilingual lexicosemantic representation respond in backward translation (L2-L1) much faster than in forward translation(L1-L2) and do more accurately in the former than in the latter, which validates the hierarchical model.The above findings are discussed from the perspectives of the degree of L2 proficiency of Chinese EFL learners and bilingual lexicosemantic representation. Finally, general implications for foreign language vocabulary teaching and bilingualism studies are examined, and recommendations for further research on the related issues are also proposed. |