Pitfalls of shared innovations: Genealogical classification of the Tsezic languages and the controversial position of Hinukh (North Caucasus)

This paper deals with the genealogical structure of the Tsezic language group with a special focus on the position of the Hinukh language within the group, which has remained controversial in previous research. Most specialists have treated Hinukh as the closest relative of the Tsez (Dido) language, and some previously proposed formal classifications based on lexical data appear to confirm this. However, the new and advanced lexicostatistical classification used in this paper suggests that Hinukh and Tsez do not in fact form a distinct clade. We examine Hinukh and Tsez historical phonology as well as morphological and syntactic features, upon which the traditional expert classifications have rested, and find that all exclusive traits seemingly shared by Hinukh and Tsez are either illusory or due to contact-driven parallel development or may simply represent Proto-Tsezic retentions. In other words, there is no phonological and grammatical evidence for a hypothetical Tsez-Hinukh protolanguage. On the contrary, there are some exclusive grammatical features of the Tsez and Khwarshi languages that may be regarded as innovations of a Tsez-Khwarshi protolanguage, conforming with a distinct Tsez-Khwarshi node in the new lexicostatistical classification of Tsezic. Therefore the traditional criterion of shared innovations supports the validity of the modern version of the lexicostatistical method employed in this paper.