On The Time-course Of L1 Activation In English Compounds Processing For Chinese-English Bilinguals

L1 activation has long been a hot topic in the field of psycholinguistics.So far,research on L1 activation is limited to simple words and only very few touches on compounds.It is proved that compounds are processed quite differently from simple word.There are four models for compounds processing,sublexical model,supralexical model,dual route model and multiple route model.Though there are many disputes over models for compounds processing,it is clear that compounds processing is different from simple words processing.So it is doubtful whether current theories of L1 activation which are established on simple words can be applied to explain L1 activation in compounds.In this research,we intend to find out(1)whether Chinese-English bilinguals activate constituents of the compound in L1;(2)the way of L1 activation;(3)the time-course of L1 activation in English compounds processing;(4)similarities and differences of processing patterns between non-proficient bilinguals and proficient bilinguals.Six experiments with a lexical decision task were conducted for 61 non-proficient and 71 proficient Chinese-English bilinguals by using within-language repetition priming paradigm.Participants were instructed to judge four groups of compound pairs: group A whose initial constituent was the same with the same L1 translation(+O+M+S+T;e.g.,backdoor-backyard.back is translated into “ ? ” in Chinese);Group B whose initial constituent was the same morpheme but with different Chinese translations(+O+M+S-T;e.g.,seashell-seafood.sea is translated into “?” in Chinese in the first compound while “?” in the second);Group C whose final constituent was the same with the same L1 translation(+O+M+S+T;e.g.,eyeball-football.Ball is translated into “?” in Chinese);Group D whose final constituent was the same with different L1 translations(+O+M+S-T;e.g.,farmland-motherland.Land is translated into “?” in the first compound while “?” in the second).Findings revealed that,for non-proficient bilinguals,there were no significantpriming and interactive priming effects for all 4 groups of compounds at 150 ms.This meant that no L1 activation was observed at 150 ms across all groups.At 300 ms,Group A and Group C revealed a positive priming while Group B and Group D showed a reverse negative priming,which indicated that translational effect facilitated word recognition when constituents shared the same translation,otherwise,inhibited word recognition.However,the positive priming effect for Group C failed to reach significant level.This meant that L1 activation of the final constituent was weak at300 compared to that of the initial constituent.At 700 ms,each group showed a positive priming effect.Group A and Group C had a larger priming effect than Group B and Group D and priming effects of Group B and Group D turned to be positive at700 ms,which suggested that semantic priming was stronger than translational priming,compared to priming effects at 300 ms.Besides,Group A had a similar priming effect at both 300 ms and 700 ms while a sharp increase of priming effect was observed for Group C from 300 ms to 700 ms.It indicated that the final constituent was processed more at 700 ms and thus,its L1 activation was stronger than that of the initial constituent.For proficient bilinguals,a significant priming effect across all groups was observed at 150 ms,which meant that proficient bilinguals could decompose the compound quickly and the recognition of constituents could facilitate the retrieval of the compound.And at 300 ms,a similar significant priming effect was found for Group A and C,and a marginally significant priming effect was found for Group B and D,but there was no interactive effect.This meant that semantic effect was stronger than translational effect and both constituents were processed equally.At700 ms,a significant priming effect was only found for Group A and C,not Group B and D,suggesting that translational effect became stronger.A deeper analysis showed that the priming effect for Group A at 700 ms was similar to that at 300 ms while for Group C there was a larger priming effect than at 300 ms.Meanwhile,the priming effect for Group D was smaller than for Group B at 700 ms.All these meant that L1 for final constituent was activated more at 700 ms.All in all,from six experiments,we observed that non-proficientChinese-English bilinguals activated concepts of the constituents via their L1 lexical forms,whereas proficient Chinese-English bilinguals first activated concepts of the constituents and then activated their L1 lexical forms through concepts.This finding supports RHM.In addition,L1 activation occurred in sub-lexical level and its magnitude was modulated by L2 proficiency and time-course.For non-proficient bilinguals,the L1 sub-lexical activation of the initial constituent occurred earlier and was stronger than that of the final constituent on middle stage of processing,while on late stage of processing,the second constituent was processed more and its L1sub-lexical activation was stronger than that of the initial constituent.On the other hand,for proficient bilinguals,the initial constituent was activated slightly earlier than the final constituent but L1 sub-lexical activation occurred equally in both constituents on middle stage.But on late stage,L1 sub-lexical activation was stronger in final constituent than in initial constituent.Our findings supplement RHM and support multiple route model.