Łukasz Grabowski

So far little attention has been paid to the corpus analysis of recurrent phraseologies found in Polish texts, in particular texts representing specialists registers of language use. Also, one may note the lack of corpus linguistic studies of lexical bundles (Biber et al. 1999) found in texts originally written in Polish. Conducted from a register perspective (Biber and Conrad 2009), this descriptive and exploratory study is intended as a first step towards a comprehensive corpus-driven description of the use and functions of the most frequent lexical bundles found in patient information leaflets (PILs), one of the most commonly used text types in the healthcare sector in Poland. The research material includes 100 PILs written originally in Polish, extracted from internet websites of ten pharmaceutical companies operating on the Polish market, compiled in a purpose-designed corpus of circa 197,000 words. Based largely on the methodology proposed by Biber, Conrad and Cortes (2003, 2004), Biber (2006), and Goźdź-Roszkowski (2011), which makes possible an analysis of the use and discourse functions of lexical bundles, the present study is primarily meant to provide methodological guidelines for future research on lexical bundles in Polish texts. This appears to be important since so far lexical bundles have been studied predominantly in texts originally written in English. The results of this preliminary research reveal salient links between the frequent occurrence of lexical bundles on the one hand, and situational and functional characteristics of the text variety under scrutiny on the other. 

This pilot study attempts to examine the potential of selected corpus linguistics and computational stylistics methods in the investigation of translation universals in translational literary Polish.More specifically, the study deals with T-universals (after Chesterman 2004), which are also referred to as intralingual translation universals (Grabowski 2011), with emphasis on core patterns of lexical use, as proposed by Laviosa (1998, 2002), and the leveling-out hypothesis, as proposed by Baker (1996). To that end, the custom-designed corpora,with approximately 500,000 tokens each, of contemporary translational and non-translational literary Polish were compiled. The results of the study reveal that on the whole translated texts are more varied lexically and have more repetitions and lower lexical variety among top-frequency words than non-translated Polish texts. On the other hand, the study shows that non-translational texts have higher lexical variety among bottom-frequency words, where usually one can find author-specific and creative vocabulary. The results of multivariate methods (Principal Components Analysis and Cluster Analysis) confirm the leveling-out hypothesis that translations are more alike as compared with native texts.