The purpose of this paper is to inquire into an ingenious defense of preference utilitarianism put forward by R. M. Hare in Moral Thinking: its Levels, Method, and Point. He insists that we can derive preference utilitarianism from logical properties of moral words and several other plausible premises. I admit that his argument is forceful, but there are some weak points in its logical construction. First, he posits as one of his premises that to know another's experience is to represent his preference. However, his argument for this premise is not sufficient. Secondly, even if we accept this premise, there remains a slight gap in his derivation which allows us to escape from preference utilitarianism.