藏-汉-英双语者字词识别中的语码切换及其代价

Language Switching and Switching Cost in Tibetan-Mandarin-English ’Visual Word Recognition

以母语为藏语、汉语熟练、英语不熟练的藏-汉-英双语者为被试,采用词汇判断的研究范式,在无切换、预期切换和无预期切换三种条件下,考察了藏-汉-英三种语言之间的语码切换及其代价。结果表明:⑴在三种条件下,被试对汉语词和藏语词的反应时和错误率的差异不显著,藏语词和汉语词的切换代价差异不显著。⑵在无切换条件下,对藏语词和英语词的反应时和错误率的差异不显著;在切换条件下,对英语词反应时长,错误率也高;英语词的切换代价显著大于藏语词的切换代价。⑶在无切换条件下,对汉语词和英语词的反应时和错误率的差异不显著;在切换条件下,对英语词反应时长,错误率也高;英语词的切换代价显著大于藏语词的切换代价。整个研究表明,藏-汉-英双语者在字词识别中的语码切换代价主要受语言的熟练程度影响。

Many studies have found language switching and switching costs in lexical decision tasks when bilingual participants switched languages for the recognition words. There were two main reasons for this, and the crux of the two reasons lay in whether the task-irrelevant language was deactivated when bilinguals were performing in only one of their languages. Until now, many studies have been conducted on the switching of languages in the case of alphabetic characters; however, few studies have been conducted using ideograms （such as Mandarin characters）. Meanwhile, Tibetian is one of the few languages whose characters have the same origin as Mandarin characters and have regular print-to-sound correspondences like alphabetic characters. What about the cognitive processing of such languages？ In this research, we chose English-Chinese-Tibetian multilinguals to perform the recognition word task. There were three experiments, and in each experiment, we chose two of the three languages. The aim of this research was to determine the essentials of language switching and switching costs in the case of ideogram-alphabetic and alphabetic-alphabetic language pairs. A 2 × 3 repeated-measures design was used. The independent variables were the languages used （two of the three languages in each experiment） and the nature of the task （no switching and anticipated/unanticipated switching）. All the materials were assessed by the homogeneity subjects, and a t-test revealed no remarkable difference. The criteria for the participants were as foUows： they needed to have lived inland for more than 8 years, needed to have been exposed to Chinese since a young age, and needed to have learned English for more than 5 years. The participants assessed their proficiency in the three languages, using a 7-point scale （1 ： extremely low proficiency; 7： extremely high proficiency）. The self-assessment scores for Tibetian/Chinese/English were 4.29, 3.97, and 2.64, respectively. A t-test revealed that there were no differences in the participants＇ proficiency for the Chinese-Tibetian pair; however, there were notable differences with re^trd to the other two language pairs. The reaction times for correct responses and error rates were analyzed by subject and item variance. The results revealed that in experiment 1, neither the main effects of language and nature of the task nor the interaction between them was significant for the Tibetian-Mandarin pair. However, in experiment 2, the main effect as well as the interaction was significant for the Tibetian-English pair, with similar results being obtained in experiment 3 for the Mandarin- English pair. The results indicate that for the Tibetian-Mandarin language pair, switching and switching costs were not significant; however, for the other two language pairs, there were remarkable differences with regard to switching and switching costs. In the cognition of words by the Tibetian-Mandarin-English multilinguals, there were switching costs in the switching among the three languages. The degree of switching costs was affected by the languages involved in the switching. The proficiency in the mastered language affected language switching and switching costs. Whether or not the task-irrelevant language was activated depended on the extent of language proficiency.