Polysemous words in and out of the context

The paper discusses various techniques of discovering and describing lexical ambiguity. This is one of the top issues in computational linguistics. A variety of techniques are used for word sense disambiguation, but all of them are based on context. Yet, studying how word senses work without context and what patterns of polysemous words could be found in speakers’ minds also seems an interesting and important issue. The main approaches to WSD with or without context (in narrow and broad sense, including the situational context) are evaluated. The importance of corpora in discovering word senses is substantiated. New experimental data are presented, which allow defining subsets of senses for polysemous words for different speakers and rating the senses in the dictionary. Finally, the paper proposes to distinguish between absolute and relative polysemy and to search for ways of their adequate lexicographic description.