David Sloan Wilson argues that religion has evolved in human beings as a group adaptation. Part of Wilson’s argument relies on an analysis of a randomized sample of religions that he selects from The Encyclopedia of Religion. One significant methodological problem with this strategy is that Wilson offloads the work of defining the boundaries of each religious tradition to the encyclopedia he uses and allows the category ‘religion’ to do the conceptual heavy lifting in his argument. An examination of the way this category is used by Wilson will demonstrate that an insufficient attentiveness to the use of the word ‘religion’ makes Wilson’s argument circular and invalid. Wilson’s argument would be strengthened by rejecting any causal role for the category ‘religion’ and examining specific practices, rituals, and other acts of identification for an adaptive advantage irrespective of their association with the category ‘religion’.