Corpus-based Analysis of Czech Units Expressing Mental States and Their Polish Equivalents Identification of Meaning and Establishing Polish Equivalents Referring to Different Theories
Abstract
The analysis is focused on Czech polysemousverbs expressing mental states. We test if different linguistic theories can lead to the closest equivalents of these verbs in Polish.
The analysis proper is preceded by automatic extraction of pairs of equivalents from the parallel corpus InterCorp. These pairs constitute a kind of bilingual dictionary.
The research includes automatic excerption of chosen verbs (with aligned segments). The segments are analysed manually.
We check how the key verb was translated and what kinds of collocations and arguments it has. Subsequently we applymethods specific for each linguistic theory.
The analysis is complementedby collocation profiles, obtained by Word Sketchesthat proved to be anessentialtool. The analysis showsthe most effective theories but also shortcomings ofthe proposed algorithmto spur its optimization and development..