In this essay I will outline Donald Davidson's account of two forms of irrationality, akrasia and self-deception, and relate this account to ethical action and belief.1 His view of irrationality is generally a Freudian one, to the effect that agents can act akratically or believe self-deceptively only because they compartmentalize mental contents. Davidson says that agents must compartmentalize both offending particular mental contents, and governing second order principles. Irrationality is engendered when particular contents and offending principles are excluded from reasoning without justification, and so agents are free to reason so as to act akratically or to hold unwarranted beliefs, for the offending contents and governing principles are no longer considered. But Davidson also hints that his account of akrasia and self-deception might show certain normative and meta-ethical theories to be irrational, insofar as they too engender irrationality. I explore these hints, showing both that Davidson is correct about irrationality, and that his view does show some theories to be irrational as well. Kantian deontology and moral realism will engender irrationality when agents follow them. I believe this to be no great loss to ethics generally, but hopefully serves to aid our understanding of how ethical action and belief actually happen. But if I am correct here, it should be said that Davidson himself would probably not agree with me. This is because he extends his of to ethical intentions and beliefs, saying that even here everything depends finding agreement.2 He even says that when it comes to moral intentions and beliefs, he is on the side of objectivity.3 But I think that Davidson betrays a failing here. He should not try to apply his principle of charity to moral intents and beliefs, for there is little reason to think that they are subject to it. Davidson claims that akrasia is what occurs when agents intentionally do what they think is not the best thing to do, all things considered. Agents weigh the evidence and judge that a certain action is the best thing to do. They judge that they should intentionally do that particular action. Then, against this judgment, such agents do not perform that action, but rather they intentionally do something else. This other action, moreover, akratic agents concede is not best to do, but is rather undesirable. For example, imagine that night I am lying in bed with my wife. I judge this the best thing to do all things considered. I then get up, and intentionally walk all around the house in a vigilant fashion, and finally return to bed. But my view, this was not the best thing to do. I did not even want to do it, and yet I did. Davidson says the reason that akrasia is irrational is not because it flouts any convention, but because it offends against what he calls the of continence. This is just that one should not intentionally perform an action when judges the basis of what deems to be all the available considerations that an alternative and accessible course of action would be better.4 Irrationality is engendered here because I have gone against my own second order principle of continence, and because I am violating this principle without justification. Davidson says that such irrationality allows reasoning to cause action without rationalizing it.5 The reasoning is still intentional, but is not the product of any good arguments. Therefore, in my example, I should have stayed in bed with my wife, but somehow I violated my principle of continence, and got up for my evening rounds. The intentional reason why I did this caused but did not rationalize my action. Davidson's account of self-deception is similar to but not the same as his account of akrasia, for while akrasia concerns intention formation and actions, self-deception concerns belief formation and the resultant beliefs. Self-deception is the state that an agent enters when, against his own estimation of the evidence, he knowingly forms a belief that contradicts another of his beliefs. …