In Thought Experiments and Utilitarianism, which is a comment on my Utilitarianism by Way of Preference Change?, Krzysztof Saja delineates three alternative ways in which one might interpret Richard Hare's famous thought experiments involving role reversals. He suggests that each of them would underwrite Hare's claim that moral deliberation transforms an interpersonal conflict of preferences into an intrapersonal one, which obtains between the deliberator's own preferences. In this reply to Saja (Role Reversals), I discuss the three proposals in turn and argue that the first two do not solve the problem. The third one, which was already considered in my original paper, is based on Zeno Vendler's suggestion that role reversals in moral thought-experiments are merely different 'takes' on one and the same real situation, as seen from different subjective perspectives. I argue that, in order to succeed in transforming interpersonal preference conflicts into intrapersonal ones, this proposal requires equating empathy, i.e. putting oneself in someone else's shoes, with sympathy, which Hare would not be prepared to do.