The aim of this paper is to describe a morphosyntactic tagset used to mark up a large-scale corpus of Polish being developed at the Institute of Computer Science, Polish Academy of Sciences.The key point of this tagset is the notion offlaceme - a set of wordforms which inflect in a uniform way. Lexemes are seen as sets of flexemes. The paper discusses lemmentation and lemmatization of Polish text, grammatical categories and lexemic and flexemic classes which were distinguished in this tagset.