The focus of this paper is Aristotle's solution to the problem inherited from Socrates: How could a man fail to restrain himself when he believes that what he desires is wrong? In NE 7 Aristotle attempts to reconcile the Socratic denial of akrasia with the commonly held opinion that people act in ways they know to be bad, even when it is in their power to act otherwise. This project turns out to be largely successful, for what Aristotle shows us (...) is that if we distinguish between two ways of having knowledge (potentially and actually), the Socratic thesis can effectively account for a wide range of cases (collectively referred to here as drunk-akrasia) in which an agent acts contrary to his general knowledge of the Good, yet can still be said to know in the qualified sense that his actions are wrong. However, Book 7 also shows that the Socratic account of akrasia cannot take us any farther than drunk-akrasia, for unlike drunk-akrasia, genuine akrasia cannot be reduced to a failure of knowledge. This agent knows in the unqualified sense that his actions are wrong. The starting-point of my argument is that Aristotle's explanation of genuine akrasia requires a different solution than the one found in NE 7 which relies on the distinction between qualified and unqualified knowing: genuinely akratic behaviour is due to the absence of an internal conflict that a desire for the proper pleasures of temperance would create if he could experience them. (shrink)