I. ‘Strong’ self‐deception

Even if many instances of reflexive, and even of interpersonal, deception do not involve knowledge or belief of the deceiver to the contrary of the belief he fosters, it is conceivable that some instances could. This is obscured in Stanley Paluch's treatment of self‐deception by the dubious contention that one couldn't be self‐deceived if one could affirm that one knew (was aware) that P and believed not‐P, and that one couldn't be described as knowing P and believing not‐P unless one could affirm this (Inquiry, Vol. 10, 1967). The former claim would actually render the affirmation absurd, which it is not; and if it is not, the latter claim is harmless. Whatever can be said of ‘self‐deception’ involving deviant uses of ‘know’, the question remains how ‘X knows P and believes not‐P’ could be true given a standard use. The standards of rationality permit one to sustain rival beliefs so long as one does not reflect on all the facts alleged in one of the beliefs. Self‐deceit relies on withholding attention not from an unwanted belief as such, as Demos suggests, but from its detailed contents.