| Research Abstract |
The Indo-Aryan language is usually divided into Old Indo-Aryan, Middle Indo-Aryan (MIA) and Modern Indo-Aryan. Middle Indo-Aryan comprises various scriptures of different times and different localities, such as the Pali canon of early Buddhism, the Prakrit canon of the early Jainism and the Buddhist-Hybrid-Sanskrit canon, etc. Middle Indo-Aryan languages are very different from classical Sanskrit, although they are in some respects related to it, and have very complicated grammatical structures such as complete alteration of a particular word due to phonetic change, assimilation of conjunct consonants, etc. Unlike in the case of classical Sanskrit, most of the grammatical features of Middle Indo-Aryan have not yet been thoroughly investigated. In order to make a major breakthrough in the study of Middle Indo-Aryan, the systematic study of the canons is now required with respect to metrical analysis, grammatical structure, vocabulary and syntax. The metrical analysis of the canons is in
… More
dispensable to edit a critical edition. The compilation of word indexes of the canons is helpful for making a translation. A reverse index is necessary to investigate the grammatical structure of the sentences, and a pada index or a reverse pada index is useful to ensure a correct reading of the text. For these purposes much data needs to be processed. Fortunately, by using a computer to analyze texts, considerable advances in the systematic study of the texts could be made. In Particular, verses could be classified, metres could be collected, indexes could be compiled and the grammatical structure could be analysed. There is no doubt that this systematic study will be extremely valuable to all scholars engaged in the study of Indology and Buddhology.
The Buddhist-Hybrid-Sanskrit text, Mahavastu-avadana is one of the biographies of most important Buddha. We have accomplished the publication of the indexes in 2003 and 2006, the word and reverse word index to the Mahavastu for Vols. I and II, respectively. This text is very important in the study of the Buddhist-Hybrid-Sanskrit. We are now preparing the index to the third volume.
In the metre analysis so far, its metre name is identified by the pattern match with the basic metre scheme and the text data. The rate of the identification of the metre of the most important old canons runs from 70 to 80% on the average, and at most 20% level in the oldest one. The identification rate is able to be improved greatly by using a quite new technique different from the previous one, the neural network assisted by the discriminant analysis. We can obtain information on the metre that cannot be analyzed so far. A new technique for extracting a remarkably different metre from the standard Sanskrit is developed. Moreover, this technique can divide a half verse into two padas extremely efficiently and accurately. This research shows one application success in neural network.
We have constructed practicable database of the study in middle Indo-Aryan by collaboration with the linguistics and the information scientific researchers. Its kernel is three of composition, the font, the editor, and analysis tools (text data, meter analysis, and index production). A database concerned is discussed, including the discussion how to devise the analysis tools on new OS in each platform.
In collaboration with linguistic scholars and computer scientists, we have made the computer tools for the systematic analysis of the manuscripts in Middle Indo-Aryan (MIA) first on Macintosh PC, and subsequently extended it on Windows PC and on Linux PC. Recently computer environment on Mac OS and Windows OS has been changed into OSX and XP from old ones, OS 9.2 and Me etc., respectively. As a result, our several computer tools cannot work normally. Since our analysis system plays an important role in analysis of the manuscripts in MIA, we have rebuilt all these computer tools so that work well on every platforms, based on JAVA. We have accomplished the publication of the manual for computer tools of the compilation of the index, with the execution files on CD-ROM by Java. Less
|
