Social preferences such as altruism, reciprocity, intrinsic motivation and a desire to uphold ethical norms are essential to good government, often facilitating socially desirable allocations that would be unattainable by incentives that appeal solely to self-interest. But experimental and other evidence indicates that conventional economic incentives and social preferences may be either complements or substitutes, explicit incentives crowding in or crowding out social preferences. We investigate the design of optimal incentives to contribute to a public good under these conditions. We identify cases in which a sophisticated planner cognizant of these non-additive effects would make either more or less use of explicit incentives, by comparison to a naive planner who assumes they are absent.