Abstrakt
The paper deals with the problem of decoding the meaning of some Czech predicates in Polish. The analysis of vague and ambiguous lexemes shows that some concepts cannot be denoted easily in the target language.
As a result, it is difficult to find an appropriate equivalent: every choice is accompanied by a loss or simplification of the translated meaning. A field of inliterality opens.
We analyse examples excerpted from InterCorp, a multilingual parallel corpus, and pairs of lexemes generated from the corpus using an automatic method of word-to-word alignment. Using the corpus data, we try to discover to what extent the factors relevant for the choice of an equivalent are determined also by a syntactico-semantic structure of the given segment.
In this paper we focus on the Czech verb toužit 'desire' (approximate translation).