Typologia tekstów oparta na miarach kwantytatywnych: studium korpusowe o zróżnicowaniu polszczyzny

The article seeks to answer the question whether it is possible to establish a typology of Polish texts based exclusively on their grammatical features. An additional aim was to find whether the typology adopted in the National Corpus of Polish (NCP), based on purely extra-linguistic criteria, groups together texts that are grammatically similar. The study was conducted on a corpus of 1190 texts randomly chosen from the NCP. For each text the frequency of some 60 grammatical features was counted, such as the number words belonging to a part of speech, occurring in a particular case, person or tense etc. With these data Bootstrap Consensus Network analysis as well as multidimensional scaling was conducted. The results show that most members of a text type cluster together showing similarity one to another. Moreover, the typology of texts adopted in the NCP gains additional support.