部件表征在合体字识别中的作用——来自ERP的证据

文采用延迟反应的同一字判断任务,记录了大学生被试由部件数不同的两种残缺汉字诱发的事件相关电位(ERP)。结果发现部件少(笔画删除)比部件多(部件删除)的残缺字产生更小的P200,更大的N400,表明简单部件在合体字识别中起到一种中间层单元的作用;P200效应符合关于部件少比部件多的残缺字激活更多的部件家族成员的假设,而更多家族成员的激活使残缺字的整字内部表征受到的侧抑制更多,体现为更大的N400。双效应支持多层交互激活模型及其侧抑制机制。

The Intermediate Role of Radicals in Chinese Characters Recognition: ERP Evidence for the MIA Model

Words whose processing are primed by a word that share a sub-lexical unite such as syllable and radical are responded to faster than words whose processing are not primed, which reflects an intermediate, activation role of the priming sub-lexical units. On the other hand, words with a high-frequency syllable, elicit higher N400 of event-related brain potentials (ERPs), reflecting greater lateral inhibition associated with larger neighborhood. To explore the controversial intermediate roles of sub-lexical units, the present study examined whether the number of simple radicals in the fragments of Chinese characters could modulate ERPs. The present study recorded ERPs elicited by radical- and stroke-deleted(RD and SD) fragments of Chinese characters, while the participants performed a delayed character matching task on the fragment stimuli and the probe characters that followed a stimulus. The RD fragments were created by deleting a simple radical from a root character that contained in average 4.66 simple radicals. TheSD fragments, which contain at least one radical, were created by deleting the same number of strokes as deleting the corresponding radical, and deleting at random except that the deleted strokes belonged to as many different radicals of the root character as possible; the SD fragments nevertheless contain at least one radical. A radical-deleted and a stroke-deleted fragments were created from one root character only which was drawn in Kai Ti, and in total two types of 200 fragments each were generated with 200 root characters. Two complementary stimulus sets, one for each participant, were formed based on the two types of fragments. Half of one type was combined with fragments of the other type created from the remaining root characters. Therefore, the probability of every character being presented under each condition was balanced across all participants. The main results showed that the SD fragments elicited a smaller P200 component and a larger N400-like component than did the RD fragments. These results demonstrated an intermediate role of simple radicals in the recognition of complex Chinese characters, given equal numbers of strokes in the two types of fragments. Further, the SD fragments contained fewer intact radicals, presumably having a larger number of radical neighbors again due to fewer radicals than do the RD fragments.Thus, as expected a higher N400 was elicited by the SD fragments as compared with the RD fragments. A lower P200 usually reflects a larger number of neighbors activated by the stimuli, and so it could reflect a larger number of neighbors associated with the SD fregraments. The N400 effect is interpreted in a multilayer interactive-activation model as reflecting relatively stronger competition from the neighbors of the SD characters . Radicals should be embedded into the multilayer model as sub-lexical units. The P200 effect is interpreted as reflecting the number of neighbors of the stimuli.