Corpus based Tatar Lexicography: Verbs in TatWordnet

The paper deals with the issue of presentation of Tatar verbs in a wordnet-like thesaurus relying on data of the Tatar National Corpus. In wordnet building we take into consideration such features of Turkic and Tatar verbs as presence of one-word and compound verbs, and a complicated system of grammatical voices when adding a voice affix modifies in many cases the verb meaning. Using data from the Tatar National Corpus we solve some key problems: improving word definition on corpus data in cases where the definitions given in vocabularies are incomplete; constructing new verbal synsets; constructing a hierarchical network of Tatar verbs.