In this article the author reflects on the ongoing debate between the intentional and non-intentional accounts of self-deception. On the intentional interpretation self-deception is brought about by an intentional act of the belief holder. Since the self-deceiver sets out to replace one belief with another, he must hold contradictory beliefs at the same time, which makes him irrational. On the non-intentional account the person is subjected to psychical motives and urges that do not create a contradiction but rather a tension or a psychological conflict. Non-intentional interpretations lead conveniently to a neurological account of the phenomenon in question. On this account self-deception is generated by uncontrolled, unconscious, and non-conceptual mechanisms which have evolved to deceive the animal or human subjects for their own sake or for the sake of the species. In this paper the author makes a point in favor of the intentional interpretation, but he tries to avoid the conclusion that the self-deceiver is utterly irrational. The problem is philosophically important because we need to judge someone's actions (and react to them) on the basis of the beliefs the persons actually hold, even if they do not admit they do. Besides, seeing the nature of self-deceptive acts helps us to get a better notion of responsibility. In the last part on his paper he distinguishes between causal responsibility and rational responsibility. The former is always there, but the latter is a cognitive and moral achievement. I argue that the intentional, non-naturalistic approach to self-deception provides a better basis for accepting and refining our rational responsibility. His argument against the naturalistic account goes as follows: 1) on this account a person is said to be unintentionally (unconsciously) suppressing certain beliefs. 2) But the suppression cannot disable logic. The person will still be capable of inferring the suppressed belief from pieces of evidence and thus retrieving it. 3) This seems to make the person irrational. 4) But if the person were thinking of herself as irrational, the whole effort of self-deceiving would be futile. The person must at least think of herself as rational being in order to deceive herself. 5) And not only think - a self-deceiving person must ascribe to herself sufficient grounds to think of herself as rational, otherwise the very thought would be irrational. The author calls this ascription a rationalization and he claims that it is the means of self-deception. 6) In order to see how the rationalization works he resorts to the logical account of belief revision (as he takes self-deception to be a kind of belief revision). His claim is that in order to deceive oneself by incorporating a new belief contradicting the one a person already holds and still to avoid the contradiction, a person has to change the intensional part of meanings in her idiolect.