EN
Semantic contextualism is recently one of the most discussed epistemological theories. So far, the main part of discussion concerns contextualist solution of the skeptical problem. Nowadays it has become more clear that this theory needs strong and independent justification from a linguistic and language-philosophical point of view. In this paper The author outlines several treads concerning linguistic basis for contextualism. In part one there are presented some fundamental contextualism thesis and an example proposed by S. Cohen which has to support it. Next the following question is considered: which semantic feature is responsible for context dependence of knowledge ascription sentences? Is it indexicality, vagueness, ambiguity, ellipticity or unspecificity? Debate sketched in part three is concentrated on an analogy between linguistic behavior of knowledge ascription sentences and other context-dependent expressions, especially indexicals and gradable adjectives. The last part of this paper contains a new argument for contextualism proposed by K. De Rose. In addition such argument points out that contextualism does not confuse a truth conditions of knowledge ascribing sentences with their condition of warranted assertability.