The Pre-activation Of Native Language Phonological Information During L2 Spoken Sentence Processing

Prediction,as a common phenomenon in language comprehension,refers to the pre-activation of language when language input unfolds.During L2 processing,bilinguals can activate not only the target language but also the native language.Recently,many researchers have found that native language phonological information can be automatically activated in L2 comprehension.However,most studies focus on phonological information activation in lexical recognition.Whether the phonological information can be predicted in constraining situations remains unexplored.The present study,using the eye-tracking visual world paradigm(VWP),investigates: 1)Can Mandarin-English bilinguals activate the native language phonological information in a highly constrained English sentence context? 2)Can Mandarin-English bilinguals predict both Chinese segmental and suprasegmental information(tone)in a highly constrained English L2 context? If the answer is yes,what characterizes the time course of these two types of phonological information during processing? Since many previous studies have proved the important role of the context,we constructed constraint English sentences in which the final target word is predictable(e.g.,Jack loves reading so he often goes to the library and borrows many books.).Four experimental conditions were set according to the phonological overlap type between the Mandarin name of critical pictures with the Mandarin translation of target words in sentences.Participants listened to auditory sentences with four pictures being presented on the screen,including one critical object and three distractor objects.The critical objects can be a target object corresponding to the target word(i.e.,book,书 /shū/),a segmental-tone object whose Mandarin name shares the identical phonological information with the Mandarin translation equivalents of the English target word(i.e.,comb,梳 /shū/),a segmental object that shares the segmental information with the Mandarin translation equivalents of the target word(i.e.,mouse,鼠/shǔ/)or a baseline object which is phonologically unrelated with the target word(i.e.,house,房/fáng/).Participants’ eye movements were recorded and their fixation on four condition objects was then analyzed.Results manifested that target objects got the most fixation among the four conditions in all items before the auditory target word onset.There was a difference on the fixation of critical objects in segmental-tone and segmental conditions under various constraint level contexts.According to the median-split of cloze probability of target words,the eye-tracking data were divided into high predictive items and low predictive items.In high predictive items,segmental-tone condition objects gained more fixation than baseline condition objects in the whole time window,while segmental objects attracted more fixation in the middle and later phase of the prediction process.In low predictive items,there was a delay in the time of fixation divergence.Participants fixated more on the segmental and segmental-tone objects compared to the baseline objects in the middle and later phase of prediction while no difference was found in the early phase.After the appearance of target words,participants looked at the segmental and segmental-tone objects more than the baseline objects.The findings showed that: 1)Mandarin-English bilinguals can predict upcoming target words during L2 spoken sentence comprehension,and they can achieve a crosslanguage activation and even prediction at phonological level in highly constrained contexts.2)Context plays a crucial role in segmental and supra-segmental information activation.Highly constrained context promotes the pre-activation of supra-segmental information.And the pre-activation of both supra-segmental and segmental information facilitates the prediction of target words.In conclusion,bilinguals can activate the semantic representation and get access to L1 lexical representation,including segmental and suprasegmental information.It offers new insights into bilingual studies.The processing difference of segmental and suprasegmental information indicates the importance of context in cross-language phonological activation.This study has provided direct evidence for the pre-activation of cross-language phonological processing,which is instructive for further study of bilingual processing.