An investigation of the computational algorithm for dependency formation in human sentence processing and its learning processes

点击翻译按钮获取中文摘要

This study aimed at investigating what kind of algorithm the human sentence processor is endowed with in order to predict the absence of a certain element. A particular type of Japanese wh-interrogative sentences, namely "aggressively non-D-linked wh-interrogative sentences", played a crucial role in the study to investigate the properties of the human sentence processor, with particular relation to the working memory system. Also, detailed procedures were examined that are involved in processing English multiple wh-interrogative sentences. It was argued that backward search is crucially involved, and it was suggested that Japanese in situ wh-phrases are processed somewhat differently.