USE OF UNIVERSAL DEPENDENCIES FOR GRAMMATICAL ANALYSIS OF THE MULTILINGUAL TEXT (ON THE EXAMPLE OF PREDICATIVE)

Abstract

The paper combines the initiative of universal dependencies (UD), with the aim to promote developing cross-linguistically consistent annotation scheme of grammatical analysis. The purpose of this initiative is to simplify cross-language research methods, develop a unified interlanguage linguistic typology, and thus lay the foundations for the use of automated multilingual systems and a universal cross-language text parser. The paper points out the main problems of grammatical analysis pertinent to a multilingual text, advantages of using common language features, and purposes of the universal dependencies project. The major principles of universal dependencies are discussed based on the example of morphology and syntax features of Russian, English and German. In the second part we illustrate how to conduct a corpus research on predicatives using UD. The article proposes a technique of automatic identification of predicatives, examines their frequency distribution in the Russian UD corpus, and the semantic categorization of the most often used predicatives.