汉字形音信息在韩国留学生判断句义中的作用

The Role of Orthographic and Phonological Information in Reading Chinese Sentences by Korean Chinese Learners

本研究旨在探讨以汉语为第二语言的学习者在提取汉字语义时是否受到形音信息的影响。实验以16名在华的韩国留学生作为被试,采用句尾歧义字实验范式,句尾字包括四种条件——符合句义的目标字、目标字的同音字、目标字的形近字和无关控制字,后三种类型的句尾字构成了歧义句。被试的任务是判断句义是否正确。实验结果发现,被试对同音干扰字的反应时明显短于对形近干扰字的反应时,对同音干扰字的判断正确率显著高于对形近干扰字和无关控制字的判断正确率,而形近干扰字和无关控制字无论在正确率还是反应时上均没有显著差异。这个结果表明,以汉语作为第二语言的韩国留学生在阅读汉语时,汉字的语音信息自动激活,并促进语义的提取。

The time courses of orthographic, phonological and semantic activation in silent reading of native speakers have been widely discussed through various approaches in the last two decades. However, the results across languages were quite controversial. For instance,most studies conducted in alphabetic languages（e, g. ,English and German）showed that phonology was an important mediator between orthography and semantics. But the findings from Chinese were three folded. Some evidence supported a phonological mediation model, some a direct access model and the other a dualroute model. Recently, researchers tried to address this issue by exploring the role of orthographic and phonological information in semantic processing of second language（L2）, and found interesting results. For example, early bilinguals showed a different activation pattern from late bilinguals when speaking L2. However, few studies have been done on speaking Chinese as a second language. In this study, we investigated the processes of accessing the meaning of Chinese words by foreign Chinese learners（ e. g., Koreans）. The nature of Korean writing system, Hangul, is alphabetic. It maps letters onto phonemes and is a shallow orthography. However, Hangul is nonlinear and its overall shape appears more similar to Chinese than to alphabetic orthographies. In other words, Hangul possesses the characteristics of both alphabetic and Chinese writing systems. Therefore, it raises the interesting question of how Korean Chinese learners process Chinese characters. Is it similar to the processing of native Chinese speakers, to the processing of an alphabetic language,or to the processing of Hangul characters （i. e. ,a unique way）？ This study could be significant for the understanding of the effect of first language （ L1 ） literacy experiences and the effect of differences between L1 and L2 on the learning processes of L2. In this study, the participants were 16 Korean students of medium to advanced Chinese proficiency level who were learning Chinese at Sun Yet-San University as the experiments were conducted. We adopted the Kutas and Hillyard＇s paradigm （1980）, in which congruent and incongruent Chinese sentences were used in the experiment. In reference to the final character, incongruent sentences were further divided into 3 types ： graphemically similar, phonologically similar or dissimilar sentences. Participants were asked to judge whether each sentence made sense or not. Reaction time and error rates were recorded through an IBM compatible PC with E-prime software. In analyzing reaction time, we excluded those trials over 3 standard deviations and only calculated the correct responses. Data were analyzed through Repeated-Measure ANOVA in which sentence type was regarded as the within-subjects factor. The main effect of sentence type for reaction time was significant, in which the reaction time of phonologically similar sentences was the shortest and that of graphemically similar sentences was the longest. The reaction time of phonologically similar sentences was significantly shorter than those of graphemically similar and dissimilar sentences, and no significant difference was found between the graphemically similar and dissimilar sentences. On the other hand, the main effect of sentence type for error rate was significant. The error rate of phonologically similar sentences was the lowest and that of graphemically similar sentences was the highest. There was a significant difference between phonologically similar and graphemically similar sentences but no difference was shown be- tween graphemically similar and dissimilar sentences. The results indicated that participants made more accurate and faster judgment on phonologically similar sentences, whereas graphemically similar characters didn＇t block participants from making responses. On the basis of these results, we can draw the conclusion that phonological information is activated automatically and has effects on semantic processing when Korean Chinese learners read Chinese sentences. The findings also suggest that the L1 background of L2 learners may affect the processing of the second language.