Abstract: The language representation of bilinguals and monolinguals is an essential issue in cognitive psychology. Much research has examined the bilinguals by fulfilling picture naming tasks and picture classification tasks. Results showed that bilingualism affects picture naming but not picture classification. Two theories were put forward to explain this phenomenon. The theory of cross-language interference assumes that mental representations of the two languages (first and second languages) have the same semantic representation and different lexical representations. Competitions among different lexical representations were involved in the process of lexical selection, therefore bilinguals should have more difficulties than monolinguals in this task. Weaker links hypothesis suggests that bilinguals have the same lexical selection mechanisms as what monolinguals have. However, bilinguals have less chances to practice either the first or the second languages, whereas monolinguals have more chances to practice their native language, so that links between semantics and phonology in bilinguals’ lexical system are weaker. How do the Mandarin-English bilinguals and Cantonese-Mandarin bilinguals name pictures and categorize pictures? Is there any common points between the Mandarin-English bilinguals and Cantonese-Mandarin diglossia speakers and bilinguals of western languages?
Using the task of picture naming and picture classification, 4 experiments were conducted to examine the performances of Mandarin-English bilinguals and Cantonese-Mandarin diglossia speakers. Experiment 1 and 3 investigated the differences between Mandarin-English bilinguals and Mandarin monolinguals, Cantonese-Mandarin diglossia subjects and Mandarin monolinguals with picture naming and picture classification tasks, respectively. Experiment 2 and 4 examined the same issue of experiment 1 and 3 by repeated the same pictures 5 times. 40 Mandarin-English bilinguals, 82 Mandarin monolinguals and 42 Cantonese-Mandarin diglossia undergraduates took part in these experiments. All pictures came from Snodgrass and Vanderwart. In the naming task, participants were asked to name the pictures presented on the screen of the computer as quickly and correctly as possible. In the categorizing task, participants were asked to decide whether the item a picture represents belongs to a certain category by pressing F or J on the keyboards of the computer.
Results showed an asymmetric effect of bilingualism and diglossia on picture naming and picture classification. Bilinguals named pictures in their dominant language more slowly and with more errors than monolinguals did. In contrast, bilinguals, and Cantonese-Mandarin diglossia as well, named pictures as quickly as monolinguals did and classified them as quickly and as accurately as monolinguals did when the items were presented four or five times. The study suggests that bilingualism and diglossia affect picture naming but not affect picture classification. The results supported the theory of cross-language interference and the hypothesis of weaker links.

Key words: bilingualism, diglossia, picture naming, picture classification