关键词: 双言者, 语码切换, 切换代价, 认知控制

Abstract:

Recent studies have been focused on the relationship between bilingual cognitive control in language switching and general cognitive control in task-switching. So far, however, evidences for whether the two mechanisms are same or not are still controversial. Among which, two major arguments have been achieved: (1) bilingual cognitive control works out as a set of processes which are fully subsidiary to the general cognitive control mechanism; (2) the bilingual cognitive control in language-switching is compulsory independent from general cognitive control in task-switching. To investigate the bilingual cognitive control mechanism of highly proficient Cantonese-Mandarin speakers and its relationship with the general cognitive control in task-switching, a dual-task switching paradigm was conducted in two experiments, including letters-digital classification task and picture naming task in simple and mixed contexts. Besides, phonological similarity of Cantonese and Mandarin were specifically controlled in the two experiments to ascertain sources of the switch cost in bilingual performance. 24 undergraduates who are highly proficient Cantonese-Mandarin speakers participated in the two experiments. They were asked to classify letters or digits in Cantonese in letters-digital classification task and name the picture in Mandarin or Cantonese according to which half of the screen the stimuli would appear. Specifically, pictures in experiment 1 shared high phonological similarity between Cantonese and Mandarin; while pictures in experiment 2 shared low phonological similarity. All the participants were tested individually on computers during the whole experiment. Reaction time and error rates were analyzed by subject and item variance. In Experiment 1, the data showed that the reaction time was shorter and the error rates were lower in non-switch condition in both tasks. Participants showed a symmetric code switching cost in picture naming task and symmetric task switching cost in letters-digital classification task, but no significant correlations was found between switch costs in linguistic and non-linguistic switching tasks. In Experiment 2, the data revealed a typical language switching cost as well as same pattern of the switch costs in the linguistic and non-linguistic switching. Besides, no significantly correlation was found in switch costs of the two tasks. Under different phonological conditions, high proficient Mandarin-Cantonese bilingual showed a symmetric code switching cost in picture naming task, with shorter reaction time in high phonological similarity than low phonological similarity. This indicated that switching process might be modulated by the phonological similarity, however code-switching cost is not affected. No matter what phonological conditions were, they showed symmetric task switching cost, which suggested that the pattern of task switch cost might be similar or same to that of code switch cost. The data of the two experiments suggested that the patterns of the switch costs in the linguistic and non-linguistic switching tasks appeared to be similar or same, though switch costs of the two tasks were not significantly correlated. Based on the qualitative and quantitative results, the study indicated that language cognitive control mechanisms of high proficient Mandarin-Cantonese bilingual are closely related, yet not completely subsidiary to the general cognitive control mechanism. Meanwhile, as evidences revealed that phonological similarity did not exert any significant effect on the language switching cost, the study supposed that the switching cost might not come from the word-recognition system itself.

Key words: bilingual, language switching, switch cost, cognitive control