中文摘要：

本文运用ERPs技术，通过语义判断任务，考察汉语单音节单位的加工差异。实验发现，汉语中非自由语素的判断准确率明显低于词和无意义的字，无意义的字诱发的N400成分的波幅明显小于词和非自由语素，非自由语素诱发的LPC成分的波幅显著小于词和无意义的字，词诱发的LPC波幅也明显小于无意义的字。结果表明，无意义的字、非自由语素和词具有不同的加工机制，由于缺乏意义，无意义的字在实验中加工的难度最小，非自由语素的加工难度明显大于词。最后得出结论，实验结果支持N400是语义加工的敏感指标，LPC成分的波幅和加工难度负相关；词更应该是大脑词库中的基本语言单位；词在大脑中存在状态比较稳定，非自由语素可能以不稳定的后备信息的概念存储或附着在大脑词库中；词和语素的划分以及语素的下位分类都是具有心理现实性的。

英文摘要：

The paper probes into the different ERPs components of Chinese monosyllabic units based on the semantic judgement iask. The experiment illustrates that the right responses to the bound Chinese morphemes are obviously lower than those to words and meaningless characters. The N400 amplitude of meaningless characters is smaller than those of bound morphemes and words. The LPC amplitude of bound Chinese morphemes is smaller than those of meaningless characters and words and the LPC amplitude of words is smaller than those of meaningless characters. It suggests that there will be different processing mechanisms among the three language units. Oweing to the lack of meaning, meaningless characters are the easiest to process. Bound morphemes are the most difficult to process. In the end, the paper concludes that N400 and the amplitude of LPC is negatively correlated with is the sensitive indicator of semantic processing the difficulty of processing, that the basic units in the mental lexicon are rather words than others, and that words remain steady in the mind while bound Chinese morphemes are not. Thus there might be psychological reality for the subclassification of Chinese words and morphemes.