Abstract: Embodied cognition theory assumes that while sentence comprehension, readers would activate the perceptual mental representation of what they read, such as object shape, orientation etc. Some of previous studies that including implicit object color information in language comprehension support the hypothesis of embodied cognition theory, however others do not. The current research intended to investigate mental simulation of implied object color during sentence reading, which the prototypicality of object color is considered. The current experiment is 2×2 two factors within-subjects design. One of the independent variables is the object color implied in the sentence (typical color or atypical color of the object); the other is sentence-picture matching condition (match or mismatch). The dependent variables are the reaction times and the brain electrical activities (N400 as index) when participants judge pictures after reading sentences. During the experiment, participants first read sentences implying a typical or atypical color of one object and then watched the pictures. They were asked to judge whether the object in the picture was mentioned in previous sentences. All of subjects’ reaction times were analyzed using 2×2 repeated measures ANOVAs. The result showed that reaction times of the pictures after sentences implying typical color of objects were faster than those after sentences implying atypical color. Reaction times of pictures in matched condition were faster than those in mismatched condition. After all, the interaction effect between two factors was not significant. Besides behavioral results, subjects’ ERPs data were analyzed using 2×5×5 repeated measures ANOVAs separately in both cases of typical color sentence and atypical color sentence. Analyzing factors were sentence-picture matching condition (match or mismatch), brain posteriority (frontal, fronto-central, central, centro-parietal, parietal), and hemisphere laterality (left temporal, left, midline, right, right temporal). The results showed that, after reading sentences implying typical color, the peaks of N400 elicited in the mismatch condition were more negative compared to the match condition. On the contrary, after reading sentences implying atypical color, the peaks of N400 elicited between the match condition and mismatch condition were similar. Both behavioral results and ERP findings together suggested that, readers do activate perceptual mental representation of objects and their properties during sentence comprehension, even characters of objects were not directly mentioned. More importantly, the prototypicality of object color implied in the sentence has influence on the specific activation of the object color representation. When reading sentences implying objects’ atypical colors, mental representation of both typical and atypical color were activated, therefore had different impacts on the sentence comprehension and other related cognitive processing. The current experiment is 2×2 two factors within-subjects design. One of the independent variables is the object color implied in the sentence (typical color or atypical color of the object); the other is sentence-picture matching condition (match or mismatch). The dependent variables are the reaction times and the brain electrical activities (N400 as index) when participants judge pictures after reading sentences. During the experiment, participants first read sentences implying a typical or atypical color of one object and then watched the pictures. They were asked to judge whether the object in the picture was mentioned in the previous sentences. All of subjects’ reaction times were analyzed using 2×2 repeated measures ANOVAs, with the object color implied in the sentence (typical color or atypical color) and sentence-picture matching condition (match or mismatch) as factors. The result showed the main effect of sentence version was significant, reaction times of the pictures after sentences implying typical color of objects were faster than those after sentences implying atypical color; the main effect of sentence-picture matching condition was significant, reaction times of pictures in matched condition were faster than those in mismatched condition; after all, the interaction effect between two factors was not significant. Besides behavioral results, All of subjects’ ERPs data were analyzed using 2×5×5 repeated measures ANOVAs separately in both cases of typical color sentence and atypical color sentence. Analyzing factors were sentence-picture matching condition (match or mismatch), brain posteriority (frontal, fronto-central, central, centro-parietal, parietal), and hemisphere laterality (left temporal, left, midline, right, right temporal). The results showed that after reading sentences implying typical color the main effect of sentence-picture matching condition was significant at the 300-500ms interval. The peaks of N400 elicited in the mismatch condition were more negative compared to the match condition. On the contrary, after reading sentences implying atypical color the main effect of sentence-picture matching condition was not significant at the 300-500ms interval. The peaks of N400 elicited between the match condition and mismatch condition were similar. Both behavioral results and ERP findings together suggested that readers do activate perceptual mental representation of objects and their properties during sentence comprehension, even characters of objects were not directly mentioned. More importantly, the prototypicality of object color implied in the sentence has influence on the specific activation of the object color representation. When reading sentences implying objects’ typical colors, only mental representation of typical color of the objects were activated, however when reading sentences implying objects’ atypical colors, mental representation of both typical color and atypical color of the objects were activated, therefore had different impact on the sentence comprehension and other related cognitive processing.

Key words: Embodied cognition, Mental simulation, Chinese sentence comprehension, Color prototypicality, Event-related Potentials

摘要： 当前研究采用事件相关电位（ERPs）技术，同时沿用经典的句图匹配范式，考察句子理解中物体典型颜色与非典型颜色的心理加工过程在脑电活动上的反映。实验中被试先阅读句子再判断句子后呈现的图片物体是否在句子中出现过，句子中隐含的物体颜色或是关键物体的典型颜色或是非典型颜色。实验结果发现，典型颜色句子版本下句图不匹配条件比句图匹配条件引发了更大的N400效应；而非典型颜色句子版本下两者N400差异不显著。研究结果表明，人们在汉语句子理解过程中会实时对隐含的物体颜色信息进行心理模拟。并且，句子隐含物体颜色的典型性是造成匹配易化或不匹配易化的关键因素之一。

关键词: 具身认知, 心理模拟, 汉语句子理解, 颜色典型性, ERPs