POST-EDITING IN MACHINE TRANSLATION OF POLISH TEXTS INTO UKRAINIAN

Abstract

The article is devoted to the study of the peculiarities of machine translation of Polish texts into Ukrainian and their post-editing. The advantages and disadvantages of machine translation are identified, and the ways and principles of post-editing Polish technical texts are outlined. It is noted that machine translation is quite accurate, but not perfect. Machine translation allows you to fully or partially automate speech and text recognition, translate individual words and create translation databases. However, an automated translation system does not guarantee quality, but is only an effective tool that needs to be used correctly. The article identifies and categorizes the main errors and inaccuracies in machine translation of Polish technical texts. It has been determined that the main shortcomings of machine translation are violation of phonetic norms; incorrect selection of unified and other terms; replacement of a term with non-terminological lexemes; use of lexemes that do not correspond to the lexical norms of the Ukrainian language (incorrect selection of synonyms, paronyms; use of surzhik); incorrect grammatical form of the translated terms; inconsistency of linguistic units by persons, verb tense (active/passive), numbers, cases; incorrect construction of syntactic structures; erroneous deletion of lexemes; redundancy of lexical units in translation. The author offers recommendations for post-editing Polish machine translation: carefully check cases of literal translation, check the use of terms, unify them and create terminological translation dictionaries; edit grammatical, lexical and syntactic errors. It is noted that for a full-fledged translation, post-editing of the text is necessary, taking into account the identified typology of machine translation errors.