This paper is to conduct a quantitative analysis of the styles, periods and themes of the texts in Japanese language textbooks used at jingdu (intensive reading) course at Chinese universities, whose resemblances with Japanese kokugo textbooks have been pointed out in previous studies. As a result, it is found out that, with respect to the styles ‘critical articles', ‘essays' and ‘novels' cover the majority of the texts, followed by ‘classic texts' and ‘tanka and haiku'- of these, more than 70% are to be the texts also found in Japanese high school kokugo textbooks. As for the periods when the texts were written, it became clear that texts published from the 1970's to the 1990's represent more than 50% of the total; moreover, it is also discovered that the older the text, the higher the probability that it is also found in kokugo textbooks. As far as the themes are concerned, ‘literature' covers 62.7% of the total. Therefore, it became evident that Japanese language textbooks used at Chinese universities have a dule imbalance with respect to the themes (‘literature') and the styles (‘critical articles', ‘essays', ‘novels') of the texts. Based on the above, in order to ensure the selection of a wide variety of texts, it is difficult to solve this problem if one does not also take the resemblance of themes into consideration, even if the matter of the resemblance in style which has been pointed out by previous researches is addressed to widen the variety. It is necessary to first grasp the dual imbalance and then to make a selection of texts that takes the balance of themes into consideration.

View full abstract

This paper introduces a 3-gram list for automatic experimentation with Japanese pop song writing. The songs are derived from a randomgeneration of third-order word approximation based on C.E. Shannon's information theory. The words in the list come from 73 Japanese popular songs written by Matsutoya Yumi. Using this list, it is possible to make Japanese popular songs automatically.

View full abstract