Paper short abstract:

In this presentation, we report on the construction and the utilisation of the corpus of Christian materials (Kirishitan shiryō). These are crucial materials reflecting colloquial late Middle Japanese, and this corpus will be released as part of the Corpus of Historical Japanese.

Paper long abstract:

As part of the Corpus of Historical Japanese, we are constructing a corpus of Christian materials (Kirishitan shiryō) using Feiqe Monogatari (1592) and Esopo no Fabulas (1593). In this presentation, we discuss characteristics of the Christian materials, and comment on the construction of this Christian materials corpus. Further, we introduce a study example of late Middle Japanese by using this corpus. Christian materials are the documents written by Catholic missionaries, mainly the Jesuits, from the 16th to 17th centuries AD. These are important materials for the study of late Middle Japanese. Feiqe is the digest text of Heike-monogatari. Esopo is the Japanese translation of Aesop's Fables. Both texts are written in the spoken language using the Roman alphabet in the Portuguese style; they reveal much information about the Japanese language, which we cannot know only by studying Japanese characters. In the construction of this corpus, for analysing the texts using UniDic (the dictionary for morphological analysis), we transliterate them to Japanese characters. In order to do this, we use previous research and reprints of the original texts as references. For converting the texts to Kanji, we use information from the representative orthography of the form of UniDic. In addition, we render the texts in phonetic spelling using information from the Roman alphabet in the originals. This gives us evidence about the pronunciation at that time, for example, the euphonic change, the voiceless or voiced consonant, and so on.

The corpus not only can function as an index, but also enables more advanced researches and statistical analysis. In the corpus of Christian materials written in the Roman alphabet, there is scope for quantitatively investigating the tendency of voiced consonants and euphonic changes, for example. There is also scope for studying politeness by combining morphology information and information about the speakers. We have already released the corpus of Kyōgen (the traditional Japanese theatrical works), and we plan to construct a corpus of Heike-monogatari. Therefore, there is scope for comparing and analysing multiple works reflecting colloquial late Middle Japanese.