The Long-term Cross-language Repetition Priming of Less Proficient Cantonese-mandarin Bilinguals: An ERP Study

Objective:Using event-related potentials (ERP) technology to explore the language representationof less proficient Cantonese-Mandarin bilinguals, and reveal the difference of proficientlanguage (Mandarin) and less proficient language (Cantonese) in the ERP.Methods:The participants are twenty less proficient Cantonese-Mandarin bilinguals fromUniversity.the Mandarin is their proficient language. Cantonese is their less proficientlanguage.Each experiment using the same research paradigm(cross-language long repetitionpriming paradigm) in Zeelenberg et al (2003),. Learning phase presents60Cantonese word(convenience, represented by A), test phase showing120Mandarin words, of which60Mandarin translation of words and A has translational relationship (expressed by B1), another60Mandarin words and A no translation relationship (with B2represents). Cantonese andMandarin words are visually presented words, and the ERP system will recond the behaviordata and EEG ingredients of participants. Study is divided into four small experiments,including:1. The study of concept representation in less proficient Cantonese-Mandarinbilinguals.2. The study of lexical representation in less proficient Cantonese-Mandarinbilinguals.3, The study of memory representation model in less proficient Cantonese-Mandarin bilinguals.4, The study of difference between proficient language and lessproficient Cantonese-Mandarin bilinguals..Using SPSS13.0statistical software analyse the behavioral data and ERP N400component. Respectively. For descriptive analysis, measurement data were described asmean±standard deviation, and count data were described as number of cases or thepercentage. Paired t-test was used to analyse behavioral data and ERP N400averagevolatility, the significance level alpha=0.05, P <0.05means statistically significant.Results:1.The study of concept representation in less proficient Cantonese-Mandarin bilinguals Behavior results: the average reaction time under the conditions of the Mandarin word B1(542.17ms) significantly high faster than that under the conditions of the Mandarinword B2(654.34ms),P <0.05),the difference was statistically significant.ERP results: the N400average volatility under the conditions of the B1and B2word aresignificant difference (P <0.05).2.The study of lexical representation in less proficient Cantonese-Mandarin bilingualsBehavior results: the average reaction time under the conditions of the Mandarin word B1(521.65ms) compared with the average reaction time under the conditions B2(536.82ms) is not very different (P>0.05), the difference was not statisticallysignificant.ERP results: the N400average volatility under the conditions of the B1and B2word was nosignificant difference (P>0.05).3. The study of memory representation model in less proficient Cantonese-Mandarinbilinguals.Behavior results: the average reaction time under the conditions of the Mandarin word B1(532.45ms) was significantly faster than the average reaction B2conditions(646.74ms), P <0.05, the difference was statistically significant.ERP results: the N400average volatility under the conditions of the B1and B2wordsignificant difference (P <0.05).4. The study of difference between proficient language and less proficient Cantonese-Mandarin bilinguals.ERP results: the proficient language Mandarin word B2N400amplitude higher than the lessproficient language Cantonese word A N400amplitude (P <0.05), the difference wasstatistically significant.Conclusion:1. The findings of long-term cross-language repetition priming paradigm in lessproficient Cantonese-Mandarin bilinguals support the point of “conceptual representationshared and lexical representation separation”.2. The findings of Cantonese semantic access in less proficient Cantonese-Mandarin bilinguals support the vocabulary connection model.3. The N400wave,evoked by the proficient language in less proficient Cantonese-Mandarin bilinguals may be more negative deflection.