汉语同音异形词意义识别中的抑制过程

使用被试自定步速逐词阅读的移动窗口技术，考察了视觉呈现的均衡型和偏向型两类汉语同音异形词意义识别中的抑制过程。实验要求被试阅读由三个分句组成的句子。重复条件下，句子第一个分句中的一个词与第三个分句中的一个词同音异形。同不重复条件(用控制词替换重复条件下第一个分句中的同音异形词)相比，重复条件下，被试对第三个分句中均衡型同音异形词之后的第一个区段阅读时间要长，即出现重复性效应。这种效应在偏向型同音异形词中并未出现。这一结果支持VHMR模型的抑制控制模块假设。简单的抑制假设或衰退假设均无法解释上述结果。此外，上述重复性效应仅出现在阅读速度较快的被试身上。这说明正是对无关信息的高效能的抑制，使得被试表现出较快的阅读速度。

INHIBITORY PROCESSES IN THE RECOGNITION OF HOMOPHONE MEANINGS IN CHINESE

A participant self paced word by word moving window paradigm was used to investigate the fate of inappropriate meanings of Chinese disyllable homophones once their appropriate meanings were selected and enhanced by sentence context. Both balanced and biased homophones were included in the present study. The differences on relative frequencies of alternative meanings are smaller for the balanced homophones than for the biased ones. Forty participants were asked to read a discourse, which was composed of three sentences and coherent in meanings. In repeated condition, a word in the first sentence was homophonic to a word in the third sentence such that a homophone was presented twice. In non repeated condition, participants read almost the same discourses as those used in repeated condition with only one exception: The homophone on its first presentation in repeated condition was replaced with a non homophone control word in non repeated condition. The repeatedness effects were defined by longer reading time of the homophone on its second presentation and/or its continuous regions in repeated condition than that in non repeated condition. The repeatedness effects were found only for balanced homophones: The reading time for the region just after the homophone on its second presentation was longer in repeated condition than in the non repeated condition, suggesting that the inappropriate meanings of balanced homophones were inhibited after the appropriate ones had been selected. Neither simple inhibition nor passive decay hypotheses can explain these findings. However, the visual homophone meaning recognition (VHMR) model, which proposed both conflict monitoring and inhibitory control modules, can explain these findings well. In addition, the repeatedness effects were observed only in the participants with faster reading speed, suggesting that it was the efficient inhibition of irrelevant information during reading that promoted the reading speed.