Abstract: The issue whether or not semantic processing can be controlled remains controversial. In recent years, the researchers mainly ask participants to complete the shallow-level processing tasks or unconsciously present the prime word to explore whether or not these processes and manipulations can control semantic processing. There are significant inconsistencies in these studies. For example, semantic priming effect has been found reduced or disappeared when attention is allocated to low-level features of the words. Whereas, some other studies has suggested that semantic processing can occur in low-level processing tasks and no-consciousness situations. However, some researchers pointed out that semantic priming effect in these previous studies might be overwritten by other processes so that it was not found. The compact and simple structure of Chinese characters other than alphabetic scripts used by most of the previous studies might be able to avoid this problem. Moreover, few studies have focused on whether or not the semantic processing of Chinese characters can be controlled. This present study manipulated different levels of control to explore whether the semantic processing of the Chinese characters can be controlled by them. Experiment 1 investigated whether semantics processes might be influenced by shallow level of processing and unconscious manipulation (the low-level control); and Experiment 2 presented the part in advance to investigate whether or not the enhanced level of control could prevent the processing of semantics. In Experiment 1 and 2, half of the participants were asked to indicate whether or not the part was a component of the prime word; and the other half were instructed to indicate whether the prime stimulus was an animate or inanimate word. Unconsciousness was achieved by presenting the prime words for a brief exposure (20 ms), then following masking stimuli. Negative priming effect was observed in part-search task irrespective of the masked or unmasked condition in Experiment 1. A reliable positive priming effect was found in the unmasked condition, and a significant negative priming effect was observed in the masked condition regardless of part-search or categorization task in Experiment 2. The present study demonstrated that the semantic processing of Chinese characters was independent of the strength of control. The occurrence of semantic negative priming indicated that the control mechanism could play a role on the semantic activation occurred to allow participants to fulfill the task-goals, instead of stopping the activation itself. The current study showed that semantic processing of Chinese characters can not be controlled.

Key words: semantic processing, part search, priming effects, control mechanism

摘要： 字词语义的加工是否受其它加工过程或操纵手段的控制一直是自动化研究的热点，以往针对英文字词的研究结果不尽一致。本研究以在该类研究中比英文单词更具优越性的汉字为材料，采用部件搜索任务，通过操纵控制水平的强弱探讨汉字的语义加工是否受控制这一问题。实验一考察汉字的语义是否受浅水平加工和无意识操纵的控制（低控制水平）；实验二探讨增强了的控制水平是否能够控制语义的加工。结果在实验一中发现了语义负启动效应；实验二出现了正启动和负启动效应。该结果表明，控制机制并不能阻止汉字语义加工本身的发生，仅能对已经产生激活的语义进行调控，提示汉字的语义加工是不可控的。

关键词: 语义加工 部件搜索 启动效应 控制机制