The Semantics of Words Comprising Phraseological Structures – the Observations of Polish Corpus of Wrocław University of Technology Editor

Abstrakt

The aim of the article is to present several problems encountered by the editor compiling the Corpus of Wrocław University of Technology. The problems are connected with the application of the method of disambiguation of meanings of words in text (WSD – Word Sense Disambiguation) during the work on material of phraseological character. Any researcher using this method confronts the problem of the lack of semantic independence of words entering into the composition of fixed word combinations. The necessity to develop a unique procedure of conduct for this type of words, inclined the author of the article to experiment in the isolation of independent meanings of words co-forming phraseologisms and to make an attempt at their typologization. The isolation of such meanings has proved possible, although the accuracy of their description is very unstable: in some cases the motivation of semantic changes—which the meaning of a word undergoes under the influence of “being petrified” in a set phrase—is transparent, in other cases, however, we can speak of a hypothetical source of changes rather than a definite modification of meaning. It is worth mentioning here about the rigour imposed by the adopted WSD marking method: a set of all meanings which a given word form may assume in the analysed text must be entered into the database. That is why the problem of providing the definition of words forming phraseologisms is vital from the point of view of a linguist dealing with machine processing of language.