中文摘要：

采用句-图匹配范式考察汉语双重否定的加工方式以及句子形式与语义对句子加工的影响,实验1和实验2分别以祈使句和反问句为对象进行探讨。结果发现：（1）双重否定是整体加工的;（2）在理解初期（250ms）,否定反问句、双重否定祈使句与肯定祈使句的加工模式一致,都直接表征事件的实际状态;（3）当句子形式与语义不一致时,读者更多地受到句子语义影响。研究表明,在双重否定加工机制的理论解释中,命题表征理论和经验模拟理论都有一定的合理性,也都存在一定片面性。

英文摘要：

Psychologists have proposed some convincing theories about the processing of negation, among which propositional theory and experiential-simulations view are especially well known. Double negative sentence is a special negative sentence. It expresses an affi rmative meaning with a negative form. Using sentence-picture-verification paradigm this research aimed to investigate the early processing mechanism of double negative words. This research offered new experimental evidence for propositional theory and experiential-simulations view. Experiment 1 used imperative sentences to examine the early processing mechanism of double negative words. The experiment used three factors mixed design. Sentence presentation is a between-subjects factor and other two variables, including sentence types and sentence-fi gure matching relationships, are within-subjects factors. This experiment demanded the participants to judge whether the pictures are consistent with the sentences. The main results were as follows： a main effect of sentence presentation emerged. Participants reacted more quickly when the sentences appeared as a whole than part by part. This didn＇t conform to propositional theory. In addition, an interaction between sentence types and matching relationship was observed. Affi rmative sentences and double negative sentences reached the match effect. This didn＇t conform to two-step-simulation hypothesis.Experiment 2 used rhetorical questions to observe the infl uence made by the discord of sentence structure and semantic meaning upon the sentence processing. The procedure and design are the same as Experiment 1. However, the experimental materials were changed by rhetorical questions. The result showed that in the early state of understanding, negative rhetorical questions had match effect, while affi rmative rhetorical questions and double negative rhetorical questions didn＇t. It implied that the processing of negative rhetorical questions is the same as affi rmative imperative sentences and double negative