京都 陽明文庫 所藏 「千字文音决」에 대하여

Abstract

This paper examines three aspects of Senjimon Onketsu owned by Yōmeibunko, Kyōto (hereafter abbreviated as Onketsu): notes on character shapes, notes on character sounds and Monzen-yomi through Kana-ten, Onketsu is a version of Senjimon, but unlike other common manuscript versions of Senjimon in Japan, it contains no annotation. Instead it contains numerous notes about character shapes. This fact suggests that the annotator was interested in shapes of Chinese characters. Most of notes on character shapes are written under each phrase (consisting of four Chinese characters). They have a mark indicating original references, which is written as small letters. These references include 115 instances of ‘智’ (referring to 智永), 104 instances of ‘歐’ (referring to  歐陽詢), 9 instances of ‘仝’,  2 instances of ‘立篆’ and 1 instance of ‘夢瑛篆’. Notes on character sounds use Kana-ten or Shōten (points located in the four corners of Chinese characters indicating four types of tone). 587 instances of them use Kanaten. 148 of them are different from the chracter sound information in Hancihai(漢辞海). 323 instances use Shōten, and 32 instances of them differ from the tone information in Hancihai. 236 characters contain both types of notes. In 8 of them, the two notes contradict each other. For example, in two instances, the Shōten indicates the entering tone, whereas the Kana-ten shows a non-entering tone sound. In two other instances, the Shōten indicates the departing tone, whereas the Kana-ten shows an enterning tone sound. In two other instances, the Shōten indicates a voiced obstruent sound, whereas the Kana-ten shows an initial [m] sound. The examination of Monzen-yomi concentrated on whether the Chinese character was read using the sound or meaning of the character. 39 phrases contain no notes about character sounds. These are considered to be read not using character sounds, judging from the same types of phrases preceding or following. In 10 phrases of them, the readings of A and B are separated. In 24 phrases, the reading order of A and B or C and D or AB and CD is reversed. The order of reading tends to be influenced by the type of constructions. Among 19 instances where A and B are read separately using the character meaning, two chatacters are not read using the character sound in 14 instances, i.e. 75%. This suggests that sound-based reading is avoided when A and B are separately read using the character meaning. We can guess that using Monzen-yomi in Senjimon which is for learners of Chinese characters, helped learners by providing information on both the sound and meaning of each Chinese character. Therefore, the annotator tried to provide both the sound and meaning of each Chinese character through Monzen-yomi as far as possible through the 250 phrases, but there were instances where sound-based reading is not possible or desirable because A and B are separated.