Abstract:

More and more researchers have involved in how bilinguals represent two languages. They use cross-language priming paradigm and reach an agreement that the lexical of two languages stored separately but the semantic stored together. The stimulus onset asynchrony (SOA) of priming paradigm in previous research was so long that the subjects could response in conscious. It is difficult to avoid the strategy in process.In addition,the relation of bilinguals’ two languages modulates the priming effect.For example, if the prime and target in translation pairs share some phone or morphology,then the subjects will response more quickly. In order to shed light on these questions, masked priming paradigm was involved in the present study to investigate the effects of cross-language masked translation with Chinese-Japanese homographs and non-homographs. In experiment 1 and 2, participants were instructed to fulfill the lexical decision task and the results showed significant cross-language masked translation effects with both Chinese-Japanese homographs (Exp 1) and non-homographs (Exp 2). In Experiment 1, priming words always facilitated the lexical decision task no matter whether priming items were Chinese words or Japanese words, whereas in Experiment 2, Chinese priming items caused much facilitation. In Experiment 3 and 4, semantic classification task was chosen to check whether task type affected the results. Effects of cross–language masked translation were again found with both Chinese-Japanese homographs and non-homographs. Moreover, such effects were not influenced by the type of priming items (i.e. Chinese or Japanese). Since Chinese and Japanese are both non-alphabetic language, recent results firstly indicates that the script-dependent effect may be involved in the masked translation priming effect in this study. In general, priming effects are found when Chinese and Japanese are homographs, no matter whether the priming items are Chinese or Japanese; by contrast, when Chinese and Japanese are non-homographs, an asymmetrical priming effect occurs. Such an asymmetrical priming effect is influenced by the tasks. The consequence illustrates that, priming effect is more significant when Japanese are primed by Chinese than reverse with lexical decision task. However, in semantic classification task, there is symmetrical priming effect in both directions. The Sense modal prefers a representational symmetry between L1and L 2 at the semantic level when the task requirement is involved the shared meaning. Based on RHM and the Sense modal, a Chinese-Japanese bilingual mix model is proposed according to the morphology and task. That is, there is a strong connection between Chinese and Japanese when they are homographs, and the strengths of their connections to the concepts are similar, either. By contrast, Japanese and the concepts are more complicated when Chinese and Japanese are non-homographs: Strong connection occurs only between Chinese and the concepts, meanwhile week connection occurs between Japanese and Chinese, as well as between Japanese and the concepts.

Key words: masked translation effect, homographs, non-homographs