关键词: 部件, 部件结合率, 汉字识别, ERPs

Abstract:

Prior research on sublexical processing in Chinese character recognition has almost consistently found that radicals were involved in lexical access. However, it is still controversial about the question of how radicals were accessed. For example, it remained unclear that whether sublexical processing relied on radical position information. Close examination to previous studies, we found material compositions might be an important factor that accounted for the different effects of radical positions in lexical access. For example, studies in which illegal characters were used as fillers showed strong positional effects. In contrast, studies in which illegal characters were not employed demonstrated the significant effects of position-general radicals instead of position-specific radicals. Following this argument, the present study was conducted with two purposes. First, by manipulating radical combinability, we aimed to investigate the roles of position-general and position-specific radicals in Chinese character recognition. Second, we examined the question of whether the existence of illegal characters would influence effects of position-specific radicals by changing the composition of materials. We recorded participants’ brain responses using event-related brain potentials. Two types of radical combinability were manipulated: the number of characters containing a specific radical at a particular position (i.e., position-specific radical combinability, SRC), and the number of characters containing a specific radical irrespective of position (i.e., position-general radical combinability, GRC). Orientation decision task was employed, in which a stimulus (a character or filler) was not presented at the center of the screen, but slightly to the right or the left. The participants were required to judge whether the target stimulus appeared on the right or the left, relative to the fixation at central of the screen. This procedure would not draw participants’ attention to orthographic information and make characters process implicitly. In addition, we manipulated the type of fillers across different experiments, with Experiment 1 employing illegal characters as fillers, whereas Experiment 2 using geometric figures as fillers. Results revealed that in the implicit task, effects of GRC appeared to be stable. More specifically, no matter which types of fillers were used (i.e., illegal characters or geometric figures), characters with high GRC evoked a larger right-brain lateralized N400 effect as compared with the items with low GRC. However, effects of SRC were influenced by material compositions. When using illegal characters as fillers, Characters with low SRC elicited a larger P200 than their high SRC counterparts. However, when replacing illegal characters with geometric figures, the P200 effect disappeared. These findings confirm the role of radical combinability in Chinese character recognition. Findings of the present study also suggest that the role of GRC might be relatively stable across different materials, while the effect of SRC might be enlarged by the existence of illegal characters. We argued that this might be due to the fact that illegal characters were generated by violating one or some of Chinese orthographic rules, which may sensitize participants to orthographic information. Radical position, as an important feature of Chinese orthography, might be over attended.

Key words: radical, radical combinability, Chinese character recognition, ERPs