Annotation methodology of supracorpora databases

The paper considers methodological principles of annotating linguistic units in parallel corpora using supracorpora databases. Supracorpora databases are a novel information resource in linguistics that allows researchers to save the results of linguistic analysis of corpus data in the form of annotations structured according to the research objectives. When dealing with parallel corpora, the annotation procedure consists of 4 basic stages: annotation objects lookup; definition of the linguistic unit and its context (both in original and translated texts); definition of the linguistic unit's attributes (both in original and translated texts); and combination of two linguistic units into a translation correspondence and definition of its attributes. The paper summarizes the previously described annotation techniques, examines functional potential of supracorpora databases, and concludes that it is possible to apply the developed methodology to a wide variety of research objects.