Examination of Syntactic Constraints and Modeling of Verbal Working Memory by Experimental Study of Sentence Comprehension

本研究的目的是通过日语句子理解的心理实验，考察句法限制在实时句子加工中的有效性，以及词语工作记忆限制与加工的相关性。第一个实验旨在考察日语句子(包括名词和动词之间有歧义的单词)在自进度非累积移动窗口阅读范式下的加工。研究结果表明，格标记和主位限制被逐渐用于建立主位关系，句子的语用似然程度影响歧义词的消解。在此基础上，我们指出了英语句法分析模型在日语应用中存在的类型学问题，并提出了一种新的适用于日语句子处理的句法分析模型。通过实验验证了该模型在多种类型的日语句子中的有效性，并讨论了日语阅读广度测验对词汇工作记忆的个体差异，分析了测验成绩与句子加工速度和准确率的相关性。实验结果表明，高分的读者比低分的读者更能准确地理解加工负荷大的重分析句子，但前者的理解时间要长于后者。在认知心理学的标准文献中，经常假设在执行速度和准确性之间进行权衡。然而，我们的结果表明，加工速度或准确性的优先顺序不是任意决定的，而是由言语工作记忆限制的个体差异决定的。

The objective of this study is to examine the validity of syntactic constraints in real-time sentence processing and the relevance of verbal working memory constraints to the processing by psychological experiments on Japanese sentence comprehension.Some experiments were carried out to examine the roles of case-markings and thematic constraint in Japanese sentence processing. The first series of experiments were intended to examine the processing of Japanese sentences including the words ambiguous between nouns and verbs by the self-paced non-cumulative moving-window reading paradigm. The results suggested that case-markings and thematic constraint were incrementally utilized to build thematic relations, and that the pragmatic plausibility of a sentence affected the disambiguation of the ambiguous words. On the basis of these results, we pointed out the typological problems of English parsing models in their application to Japanese, and we proposed a new model valid for Japanese sentence processing. We performed some experiments to suggest the validity of our model in many types of Japanese sentence.We also discussed the individual differences of verbal working memory estimated by Japanese Reading Span Test and analyzed the correlation between the scores in the test and the speed and accuracy in sentence processing. Our experimental results indicated that readers with high scores in the test tended to comprehend the reanalysis sentences with great processing load more accurately than those with low scores, but the former spent longer time for their comprehension than the latter. In the standard literature of cognitive psychology, trade-off is often assumed for performance speed and accuracy. Our results suggests, however, that the priority of processing speed or accuracy is determined not arbitrarily but by the individual difference of verbal working memory constraints.