Counterintuitive concepts have been identified as major aspects of religious belief, and have been used to explain the retention and transmission of such beliefs. To resolve some inconsistencies in the literature concerning counterintuitiveness, we conducted three experiments to study the effect of context on recall. Five types of items were used: intuitive, minimally counterintuitive, maximally counterintuitive, minimally counterintuitive with contradictory context, and intuitive with contradictory context. Items were presented with context or without context and participants were asked to recall them. Maximally counterintuitive concepts were found to have the poorest recall in both immediate and delayed recall conditions and regardless of the presence or absence of context. No significant differences were found in the recall rates of minimally counterintuitive concepts and intuitive concepts, although delayed recall affected minimally counterintuitive concepts less than intuitive concepts, suggesting the possibility of differential "fitness." Presence of contradictory context was found to be able to change minimally counterintuitive items into the functional equivalents of intuitive items (and vice versa). When relevant context was present, minimally counterintuitive concepts were recalled significantly better than intuitive concepts, which is consistent with the findings of Barrett & Nyhof (2001). For items presented as lists, intuitive items were recalled better, consistent with the findings of Norenzayan & Atran (2004b). Thus, context was the key element affecting recall and the discrepancy among prior studies (and the much earlier studies of Bartlett, 1932) was resolved. The results imply that no "item-centered" explanation of the formation and transmission of religious concepts can be adequate in itself. Instead, the nature of the surrounding context must be included in any such account.