Real-time Analysis of the comprehension process of sentences with empty subjects

研究我们如何“理解”一个给定的表达是一个基本问题。本文试图通过对句子加工的研究来阐明“理解”的过程是一个认知过程。本研究的实际对象是日语空主语句的理解过程。特别是，我们有兴趣在“实时”检测句子处理。我们进行了一个实验，使用计算合成的句子。任务是“是”或“否”的决定。此外，我们还考察了主语和宾语在句子中的位置互换的效果。在这两个订单中，我们观察到一个“主题偏好”的现象。这种影响随着时间的推移而消失。这一实验结果表明，句子加工机制的运作首先是“自上而下”的过程，然后是“自下而上”的过程。虽然有一些关于日语句法分析的研究工作，但这种语言学家和心理学家合作的工作并不多见。上述发现是联合研究的一些结果。今后，这种联合研究的新产品将越来越多。

It is one of the fundamental issues to investigate how we "understand" a given expression. We try to elucidate the process of "understanding" as a cognitive process through the study of sentence processing.First, we consider some processing models based on linguistic theory. The actual target of our investigation was the process of understanding sentences with empty subjects (or, missing subjects) in Japanese. Particularly, we are interested in detecting the sentence processing in "real-time". We conducted an experiment using computationally synthesized sentences. The task was Yes-No decision. Furthermore, we examined the effect of exchanging the order of the subject and object in the sentence. In both orders, we observed a "subject preference" phenomena. This effect disappears in the course of time. This experimental findings suggest that the sentence processing mechanism operates in the "top-down" procedure at first, and then in the "bottom-up" procedure. Thus, we obtained a new knowledge concerning the understanding of "empty" element in a sentence.Although there are some research works concerning parsing of Japanese, this kind of co-operative work by linguist and psychologist are rare. This findings mentioned above are some out-come of the joint research. In the future, new products of this kind of joint research will increase.