Are self-deceivers irrational? I shall argue that whether they are or not depends upon the kind of rationality being considered. One kind is epistemic or cognitive rationality.

While there is widespread agreement that the epistemic rationality of a belief is compatible with its falsity – the person who adds 67, 81, 37 and gets 175, having forgotten to carry the 1, can be epistemically rational although he believes something false – there is disagreement about what epistemic rationality is. Is epistemically rational belief, for example, belief that is the outcome of a reliable belief-forming process, or is it belief that is well founded on reasons? While I concentrate in this discussion on the latter way of understanding epistemic rationality, and argue that on that way of understanding it the self-deceiver is not epistemically rational in believing that which he deceives himself into believing, nevertheless, given that the self-deceiver is, as I have argued, biased, the self-deceiver would not be epistemically rational in his belief on the reliabilist's conception either. The process of self-deception does not yield beliefs most of which are true.

As I shall understand epistemic rationality, if one's belief state lacks epistemic rationality, and that belief state is not an immediate perceptual belief, an immediate experiential belief, or a belief in a necessary truth, then one's belief state is epistemically irrational. Every actual belief token (barring the exceptions mentioned) is either epistemically rational or epistemically irrational. In showing, therefore, that self-deceivers are not epistemically rational in believing that p, I will be showing that self-deceivers are epistemically irrational in believing that p.