The author confronts the common-sense realism (understood as the thesis, that in linguistic utterances we speak about the world as it is in itself, independently on our language and conceptual apparatus) with the linguistic constructivism (claiming that what we speak about is ipso facto structured by language). He argues that realism is compatible with the plurality of possible ways of speaking about the world, that it provides a good basis for the pragmatic account of language and passes the reflexivity test (the realist claims about language are applicable to these claims themselves). On the contrary, constructivism fails in the last two respects. The confrontation of realism and constructivism on the epistemological level leads to analogical results. In his comments to Wright's and Dummett's explications of the realist-antirealist controversy, the author points out that the common-sense realism does not include committment to the non-epistemic notion of truth and to the bivalence principle.