The germ of Tibetan phonology appeares first in the Sum cu pa composed by Thon mi sambhoṭa (7th century). At the end of the Sum cu pa, Thon mi states that the pronunciation of each syllable (/ka/, /ki/, /ku/, /ke/ etc.) should be learned in accordance with the method of pronouncing phonemes (nga ro, *svara) in the correct position of articulation. However, since Thon mi did not give a detailed explanation of Tibetan phonology in his Sum cu pa, Tibetan grammarians in later periods had to develop a theory of pronunciation of their language with the help of their knowledge of Sanskrit phonology. Among them, Si tu paṇ chen (1699/1700–1774) gives an explanation of Tibetan phonology in his Si tu’i ’grel chen on the basis of the Sanskrit phonology described in Candragomin’s Varṇasūtra. What emerges from Si tu’s explanation is the development of phonological theory that is characteristic to Tibetan but not found in Sanskrit. In addition, he reveals both the applicability of Sanskrit phonology to Tibetan and the limitations of its application. The purpose of this paper is to clarify the major characteristics of Tibetan phonology by examining Si tu’s explanation based on the Varṇasūtra and comparing it with the explanations given by other Tibetan grammarians.
