中文摘要：

拼音文字的研究发现，在词汇通达过程中，语音和词形之间存在着相互作用的共振关系。与高频词比，低频词能够更有效地引发词形效应。通过2个实验，探讨了在汉字高频同音字通达中语音激活对词形激活的反馈作用。结果表明，汉字高频同音字的通达过程受语音激活的反馈影响。在汉字加工的早期，共享典型部件的汉字之间存在着词形竞争。在高频汉字的通达中，存在着拼写几率效应，但对高频同音字而言，拼写几率效应的促进作用比语音激活的反馈作用要弱。研究结果支持词汇识别的动力系统原则和词汇激活的共振模型。

英文摘要：

In the last few decades, three different models, including phonological mediation model, direct access model, and dual-rout cascaded model, were trying to explain how visual words are processed. As we all know, one of the most important differences among them is whether rime plays an important role in orthography. Till 1994, Van Orden and Goldinger suggested a more complicated relationship between spelling and phonology in processing English words, and a resonance model was proposed. More and more researchers have used this resonance model to explain not only visual words processing, but also auditory words processing. However, it has not been really convinced in all recent studies, and evidence from ideographs, such as Chinese, was still missing. Chinese is a very special language whose pronunciations and orthographies are not one-to-one correspondence. Most of the pronunciations have more than two writing forms （homophones）, while two similar forms are pronounced totally differently sometimes. Relationships between pronunciations and orthography in Chinese characters might be even more complicated than those in English. Whether would the resonance model be also fit for Chinese？ To answer this question, two lexical decision experiments were employed to investigate how rime affects the activation of orthography in Chinese. In Experiment 1, a cross model lexical decision task was used, and results showed longer response latency with words having more than two orthography candidates （homophones）. However, no significant differences were found between homophones with dominant spelling and those with subdominant spelling. It was not consistent with the results reported by Ziegler, Ferrand and Montant （2004）. In Ziegler et al.＇s research （2004）, facilitation in processing words with dominant spelling in English was found. The probably reason for the missing of spelling probability effect in the recent research is that dominant and subdominant spelling components of homophone words were activated by