Electronic Textualization of a Japanese Rendition of Kanbun Using Aozora Bunko Annotations, and Visualization of the Workload of Reading the Analects of Confucius

Abstract

The reading of classical Chinese texts in Japanese involves adding kunten marks to classical Chinese texts, then rearranging the word order to read them in Japanese.
The Aozora Bunko annotations are simple markup, but there is a kunten notation, and many viewing software programs adopt it. We converted the readings of the Analects, consisting of 499 chapters in 20 volumes, into electronic text using the Aozora Bunko annotations. In order to see the characteristics of the reading, we introduced the number of steps in length including the return reading, the number of loops in the number of times of return reading, and the CR notation indicating the structure of the return reading. For short sentences up to 5 characters, we exhausted all permutations that correspond to reordering the word order and looked at the structure of the return reading. The results showed that the number of permutations, sentences made with parentheses and inverted signs, and sentences that can be represented by actual return marks, in that order, were less, indicating the characteristics of actual return marks. We analyzed the readings of the Analects with a unit of sentences divided by punctuation marks. There were 3,820 sentences, with a median length of 4 characters. The median of number of steps was 6, and the majority of sentences had more than one loop, indicating the workload of the reading. On the other hand, there were only about 60 main return reading structures, and the more complex structures were less common.