EN
The aim of the article is to determine what role the liar sentence plays in our language. On the one hand, it seems to be well formed formula, and on the other, it does not seem to have any clear sense. At the beginning of the article I point what form an adequate solution of the liar paradox should take. In my opinion it could not consist in giving rules which do not allow to build such a sentence. The paradox remains unsolved until there is such a language in which it could be expressed. In the first part of the text I try to explain why Tarski’s solution is not satisfactory. If the semantical definition of truth is correct, the liar sentence could not lead to a contradiction because formulas which are not well formed could not be premises of any inference. From that follows that the so called liar paradox does not arise and that leads to the conclusion: ‘the reconstruction’ of the liar propounded by Tarski could not be correct. In the second part I present an approach to the liar which appeals to Frege’s and Wittgenstein’s conceptions of language. The conclusion of my consideration is that the liar sentence is nonsense, which means it is not given any sense – either its logical form is determined but we do not fix any definite meaning to some parts of it, or an attempt to determine its logical form in the standard way leads to regress ad infinitum.