Query Structuring for Automatic Selection of Analytical Future Tense in Slavic Languages

Abstract

Modern corpora tools allow language researchers to perform a wide range and type of analysis as well as automatize the selection of empiric material thanks to a detailed annotation of the texts. Since annotation is mainly word-oriented, some grammar categories which are expressed by means of analytical forms cannot be tagged as such in a word-oriented approach, which is why in many corpora their selection is hampered or produces mistaken results. To overcome this problem, we propose to use specific queries in Context Query Language (CQL), which can combine as many parameters and conditions as the researcher might need to indicate: lexemes with their characteristics or sequence of them. To illustrate the usage, possible results, and potentiality of this tool, we make an overview of the queries aimed at selecting analytical forms of future in 6 Slavic languages: Belorussian, Check, Polish, Slovak, Russian, and Ukrainian. CQL-queries should be adapted for each corpus by modifying the tagname, though the structure of the query remains identical. Some annotated corpora do not accept CQL-query, though they provide a specific user interface for selecting grammar categories. Nevertheless, CQL appears to be a more universal solution (though, more complicated to use) because it includes the possibility of modifying partially the parameters of the search, for example, selecting future forms in a certain voice or aspect.