Zrozumiałość przekładu tekstów użytkowych: automatyczna analiza tłumaczeń ulotek leków na cukrzycę
Anna Bączkowska
Abstract
The article presents results of a corpus-assisted study of diabetes medicines Patient Information Leaflets written in English and their translations into Polish. Several computer programs were used (Sketch Engine, Jasnopis, Coh-Metrix, Oleander Studio) in order to find out which text aspects affect the readability of these leaflets. The study combined the methodology of classical readability tests, a corpus-driven lexical analysis as well as automated text analysis. The correlation between traditional tests and the lexicosyntactic analysis was checked. Moreover, lexical aspects of texts which have the strongest influence on text readability were determined. A parallel corpus was compiled in order to compare the results obtained for the corpus of Polish leaflets with those retrieved from the corpus of English leaflets. The frequency of select words was determined with the aid of the British National Corpus, Narodowy Korpus Jzyka Polskiego as well as the plTenTen15 corpus. It was noticed that both the English leaflets and their translations had a low readability level.