2008 年 26 巻 2 号 p. 129-139
This paper examines the relevance of the constraints of working memory to Japanese syntactic reanalyses. A phrase-by-phrase, self-paced reading experiment was performed with two kinds of sentences which were assumed to involve reanalyses as experimental sentences. The comprehension accuracy and the reading time were analyzed in relation to the constraints of working memory estimated by the Japanese Reading Span Test. Our results indicated that a reader with a high score in the test, comprehended the sentences with costly reanalyses more accurately, and spent a longer time than a reader with a low score. This suggests that the efficiency generally assumed for readers with high (Japanese) Reading Span Scores does not necessarily imply rapidity in reanalyses. Some implications of our results for models of working memory, and the significance of the limitations of working memory, are discussed.