The Tocharian gender system. A diachronic study

Tocharian is the name given to two closely related Indo-European languages, Tocharian A and Tocharian B, known from manuscripts discovered in the Tarim basin (present-day Xīnjiāng Uyghur Autonomous Region, China), dating from the 5th to 10th centuries CE. Despite its late attestation, Tocharian has proved to be archaic, particularly in some sections of the morphology. However, the exact relationship of Tocharian with the other Indo-European branches remains an unresolved issue. The problem is that a strong impact of language contact and internal drift has resulted in an intricate combination of archaisms and innovations that are often difficult to be disentangle. Examining the category of gender, this thesis contributes to the investigation of archaisms and innovations in Tocharian nominal morphology. It aims at providing a comprehensive treatment of the Tocharian gender system, describing how it historically derived from the Indo-European proto-language and why it typologically deviates from most of the other Indo-European languages. Next to the masculine and the feminine, Tocharian has a third category, which is named genus alternans. Nouns pertaining to this category combine agreement traits of the masculine and the feminine, taking masculine agreement in the singular and feminine agreement in the plural. Using a cross-linguistic approach, the synchronic analysis could show that Tocharian has a three-gender system, with the genus alternans being a gender value of its own. The largest part of the thesis deals with the diachronic investigation of a large number of nominal and adjectival classes whose endings and forms are relevant to the historical analysis of the Tocharian gender system. These classes are analysed from the point of view of their derivation and inflection, in order to clarify for each of them the origin and the development. The evolution of Tocharian gender is a long-standing problem, which has recently become even more pressing because of the scholarly debate on the chronology and the relative order of the split of the Indo-European proto-language into the various branches. It is sometimes assumed that Tocharian has split off second after the Anatolian branch and that the gender system has retained evidence of this early split. In order to test this claim, the origin of the Tocharian gender system in nominal, adjectival, and pronominal morphology has been discussed. It has been argued that, despite the many peculiarities of its gender system, Tocharian does not require the reconstruction of a more archaic gender marking system than the other non-Anatolian Indo-European languages. The peculiarities of Tocharian have been caused by internal developments that took place within the evolution of the language, which have often blurred the boundary between inherited archaisms and internal innovations.