In recent years, theoretical and empirical work done under the rubric of the cognitive science of religion (CSR) have led many to conclude that religion (or, at least, some aspects thereof) is "natural". By this, it is meant that human beings are predisposed to believe in supernatural agents, and that their beliefs about these agents are constrained in various ways. The details about how and why these predispositions and cognitive constraints developed and evolved are still largely unknown, though there is enough of a theoretical consensus in CSR for philosophers to have begun reflecting on the implications of CSR for religious belief. In particular, much philosophical work has been done on the implications of CSR for theism, on both sides of the debate. On one hand, CSR might contribute to defeating particular arguments for theism, or indeed theism altogether; on the other hand, CSR might provide support for specific theological views. In this paper, we argue that the CSR is largely irrelevant for classical theism, and in particular that the "naturalness hypothesis" is much less congenial to theism than some have previously argued.