Representation Of Mental Lexicon In Chinese-English Late Proficient Bilinguals

The study of bilingual mental lexicon has now become a major concern in second language acquisition and psycholinguistic research, focusing on bilingual semantic representation and lexical selection mechanism. Two major theoretical viewpoints are dominant in bilingual semantic representation research: separate hypothesis and shared hypothesis. The former proposes that lexical knowledge of bilinguals may be represented in two language-specific memory systems, one for each of the bilingual's languages. The latter holds that bilinguals have a common, language independent, conceptual representation for words in their two languages. The debate on whether bilingual lexical representation is structured in separate memory systems or in a common conceptual store has been continuing because of conflicting empirical evidences. On the issue of bilingual lexical selection, language-specific mechanism only considers the activation-levels of words in the intended language, while language non-specific mechanism (Inhibitory Control Model) assumes that the speaker would consider all activated lexical nodes for selection, irrespective of the language to which they belong.Traditional experiment methods get their results by comparing the multitudes of subjects' reaction time, which are measured at discrete points in time, making it difficult to interpret how specific processes contribute to an effect. In contrast, Event Related Potentials measure on-going electrical brain activity and multidimensional data such as latency, amplitude, polarity, and scalp distribution of a component. Thus, this method is more precise and sensitive to the nature of cognitive processing and neural mechanisms involved. The study attempts to use the new neuroimaging technique—ERP to investigate whether late proficient Chinese-English bilinguals have two mental lexicon semantic systems or share a common one. What is the mechanism of bilingual lexical selection?The results indicate: 1) From time window 130 ms—230 ms, we haven't detected any significant differences in subjects'performances for Chinese naming and English naming tasks in the left extrastriate cortex and inferior temporal areas, which are related to early visual image processing. 2) In the prefrontal cortex area which is related to language processing, no statistic difference of peak latency is found in the ERP waveforms and there are similar activations in this area. 3) From time window 242 ms—380 ms (the lexical selection period), there is a more negative amplitude and more intensive activations are observed for Chinese naming task, reflecting that naming in Chinese costs more energy of the brain than naming in English.Findings of this study lead us to three conclusions. First, for late proficient bilinguals, there is no difference in the early visual image processing. Second, for late proficient Chinese–English learners, the conceptual representation of two languages is a common one, which supports the shared hypothesis. Third, our results are consistent with the language-specific mechanism perspective. The question still under discussion is whether the results may be attributed to learners'language proficiency or their age of acquisition.