The processing of homographic morphemes in Chinese: an ERP study

The processing of homographic morphemes during Chinese word recognition was investigated by recording event-related potentials (ERPs) in a masked priming lexical decision experiment. There were four conditions. In the morpheme condition, the homographic morphemes in primes and targets had the same interpretations (e.g. "-public garden/park" and "-public people/the public"). In the homograph condition, they had different interpretations (e.g. "-male chicken/cock"). Semantic-sharing (e.g. "-lawn") and unrelated (e.g. "-olfaction") conditions were also included. Compared to the unrelated condition, the morpheme and homograph primes produced a comparable P200. In contrast, N400 priming was identified only in the morpheme condition. Semantic sharing produced relatively weaker effects. These findings indicate that homographic morphemes are activated during word recognition even though in principle, whole-word processing is more efficient. The results are discussed with reference to morpho-orthographic and morpho-semantic processing.

通过在掩蔽启动词汇判断实验中记录事件相关电位（ERPs），对汉语词汇识别过程中同形语素的加工进行了研究。实验设置了四种条件。在语素条件下，启动词和目标词中的同形语素具有相同释义（例如“公园”中的“公”和“公众”中的“公”）。在同形异义词条件下，它们具有不同释义（例如“公鸡”中的“公”）。还包括语义相关（例如“草坪”）和语义无关（例如“嗅觉”）条件。与语义无关条件相比，语素启动词和同形异义词启动词产生了相似的P200。相比之下，仅在语素条件下发现了N400启动效应。语义相关产生的效应相对较弱。这些发现表明，在词汇识别过程中同形语素会被激活，尽管从原则上讲，整词加工效率更高。研究结果结合正字法语素和语义语素加工进行了讨论。