Analiza fleksyjna tekstów historycznych i zmienność fleksji polskiej z perspektywy danych korpusowych

The subject matter of this paper is Chronofleks, a computer system (http://chronofleks.nlp.ipipan.waw.pl/) modelling Polish inflection based on a corpus material. The system visualises changes of inflectional paradigms of individual lexemes over time and enables examination of the variability of the frequency of inflected form groups distinguished based on various criteria. Feeding Chronofleks with corpus data required development of IT tools to ensure an inflectional processing sequence of texts analogous to the ones used for modern language; they comprise a transcriber, a morphological analyser, and a tagger. The work was performed on data from three historical periods (1601–1772, 1830–1918, and modern ones) elaborated in independent projects. Therefore, finding a common manner of describing data from the individual periods was a significant element of the work.