Analytical grammar forms extraction as a new challenge for corpora (Case of conditional mood in Polish and Ukrainian)

Abstract

A particular challenge for modern textual corpora is the tagging of analytical grammar categories. The com-ponents of these categories may be separated in certain contexts by other words or may even be inverted. A particular interest regarding the selection of analytical grammatical forms is centred around the conditional mood in some Slavic languages, as expressed by means of two words: a past verb form and the particle by/б/би/бы, which is why in most modern corpora, this category lacks a specific tag for these compound forms. The case of Polish is particularly complicated because the particle by may either be merged with the parti-ciple or used separately; furthermore, its separated form may contain a personal verb ending. Specific que-ries subject to experiment on Polish and Ukrainian corpora allow selecting the analytical forms in question.