Abstract
This paper considers the relationship between evidence and self-fulfilling beliefs—beliefs whose propositional contents will be true just in case—and because—an agent believes them. Following Grice (1971), many philosophers hold that believing such propositions would involve an impermissible form of bootstrapping. This paper argues that such objections get their force from a popular but problematic function-model of theoretical deliberation, and that attending to the case of self-fulfilling belief can help us see why such a model is mistaken. The paper shows that on a revised model of theoretical deliberation our evidence will problematically underdetermine any appropriate doxastic attitude: when belief in a proposition is self-fulfilling, our evidence is insufficient to support belief, disbelief, or even suspended judgment toward that proposition.