Introduction

The Tocharian manuscripts were discovered on the Silk Road in the Tarim Basin of Kashgar valley located in the Xinjiang region of China. The documents which date from the 6th till the 8th century represent two dialects: TA in the north-east (Turfan, Karashar) and TB in the south-west (mainly in Kucha and elsewhere). Tocharian A was used for religious purposes at that time, whereas Tocharian B was the language of current and widespread usage. The Tocharian texts consist mainly of translations of Buddhist religious literature. The Tocharian language became a tool for the dissemination of Buddhist and Manichean religions in China and among the Turkic tribes, testimony of which are the translations of religious texts from Tocharian into the Old Uighur language. The two branches of Tocharian represent two, very different (and supposedly mutually unintelligible) dialects which are derived from one common proto-language. The extent of the differences makes us infer that they must have separated from their parent language in a relatively remote past and that they were shaped independently of each other, but on a shared substrate.

Despite the common origin of Tocharian and the remaining Indo-European languages, the differences between them are too great to be explained by internal development. We meet further difficulties in that most of the phenomena in Tocharian have been attributed to which foreign influence occur simultaneously in several languages of northern and central Asia. In this situation a typological confrontation of separate phenomena, within the languages in which they occur, seems to be most appropriate here. Thus, we can list the following features:

Phonology

The lack of voiced consonants

As it is generally known, the correlation of sonority does not occur in Finno-Ugric languages (Łytkin 1974: 117–128), Samoyed (Janhunen 1998: 462–463) and probably it was not found at the initial stage of the development of the Turkic languages. According to A. Róna-Tas (1998: 71) “Proto-Turkic has a fortis (strong) vs. lenis (weak) system of obstruents, though the actual phonetic features are not known. The strong member of each opposition was unvoiced, and maybe aspirated, whereas the other member was weak and most probably also unvoiced.” There is a similar situation in certain Siberian languages and in Sino-Tibetan languages: Mandarin Chinese: p/ph, t/th, k/kh, etc., whereas in Lhasa Tibetan we have (non)aspirated voiceless and prenasalised voiced stops.