In Middle Polish there were several comparative markers (niż, nad, od, jako and jak). Although replacing one marker with another had no consequences for the interpretation of an utterance, it seems unlikely that in any given language there should be fully interchangeable elements. This paper intends to answer the question what governed the choice of comparative marker. This problem has already been addressed both synchronically and diachronically (inter alia by, Greszczuk 1988, Kallas 1997a, and Bondaruk 1998); however, an examination of the data retrieved from a machine-readable corpus (the KorBa corpus) allows to reveal previously overlooked facts. The distribution of the markers is not random, though it is governed by tendencies rather than rules. We examine the role of such factors as part of speech of the feature compared (adjective vs adverb), part of speech of the comparans (noun vs pronoun), meaning of the compared feature, and text type.

The article is written in English.