It has been suggested that in order to aid students in fulfilling their potential and achieving optimum academic outcomes, both teachers and students themselves should hold beliefs that significantly overestimate student ability and go beyond the available evidence. In this article, I contend that teachers adopting such beliefs or instilling them in their students is neither desirable nor practicable. I instead propose a pair of alternative cognitive attitudes--hope and imaginings-- that can substitute for belief in producing self-fulfilling prophecy-type mechanisms, which I argue have often been misattributed to belief within the literature on this topic. These attitudes offer a plausible route to academic improvement that does not breach epistemic norms.