I want to show that constrained maximization or any similar proposal cannot be coherently supported by expected utility theory. First, I point to an important implication of that theory. Second, I ...