句子完成与时间缓冲对信息整合的启动作用

探讨句子完成和时间缓冲因素在阅读过程启动背景信息整合的作用。包括3个实验，被试是华南师范大学198名一年级本科生。采用移动窗口技术，要求被试阅读14篇一致性版本与不一致版本的文章，然后对不同条件下目标句阅读时间或探测反应时进行统计分析。实验1重复Guzman和Klin的研究，并进一步探讨单纯的句子完成因素是否能即时启动整合。结果表明，在单纯的句子完成条件下，一致版本和不一致版本的目标句阅读时间和探测反应时有显著差异，说明目标句阅读时已经即时启动了背景信息的整合。实验2探讨Guman和Klin关于缓冲可以即时启动整合的研究是否混合了句子完成因素的作用，结果表明，当在目标句后面增加连接词时，会产生句子完成的暗示，从而可以即时启动整合，Guzman和Klin关于缓冲可以启动整合的结果是值得商榷的。实验3从Guzman和Klin的研究设计中分离出句子完成因素，探讨单纯的时间缓冲是否可以即时启动整合，结果表明，单纯的缓冲不能即时启动整合。研究的结果初步证明，句子完成是启动信息整合的充分必要条件，而单纯的时间缓冲不能起到这种即时启动作用。

ROLE OF SENTENCE BOUNDARIES AND BUFFER TIME IN PRIMING THE INTEGRATION OF BACKGROUND INFORMATION

This research was designed to explore the roleof sentence boundariesand buffer time in priming the integration of background information during reading process. Three experiments were administered. Ss were 198 university studentsThe on-line window display technique was used and the Ss were asked to read 14 discourses which included the consistent version and inconsistent version. The time of reading the target sentences and responding to the probe words in different conditions were analyzed. In Experiment 1,the research made by Guzman & Klinin 2000 was replicated under the condition with Chinese materials. And it was further explored whether only the sentence completion could prime the on-line information integration. The results showed thatthe difference of the times both in reading the target line and in responding the probe approached significant level between the consistent condition and inconsistent condition under the condition of only sentence completion. It was indicatedthat background information integration was primed immediately during the course of reading the target line.In Experiment 2, it was examined whether the results which indicated the buffer time could prime the information integration on line in the experiment by Guzman & Klinin 2000 mixed the effect of sentence completion. The results showed that adding the conjunction to the end of target sentence could become the cue of sentence completion and immediately prime the integration. It was indicated that the conclusion got by Guzman & Klinwas doubtful. In Experiment 3, it was further explored whether only the buffer time wassufficient to prime the background information integration on line without the sentence completion cue. The result showed that only the buffer time could not prime the on-line information integration. It can be concluded according to this research that the factor of sentence completion can prime the on-line background information integration during reading course and the buffer time can not do so.