Abstract In cooperative breeding systems, some individuals help to raise offspring that are not their own. While early explanations for such altruistic behaviour were predominantly based on kin selection, recent evidence suggests that direct benefits may be important in the maintenance of cooperation. To date, however, discussions of cooperative breeding have made little reference to more general theories of cooperation between unrelated individuals (while these theories rarely address cooperative breeding). Here, we attempt to integrate the two fields. We identify four key questions that can be used to categorise different mechanisms for the maintenance of cooperative behaviour: (1) whether or not individuals invest in others; (2) whether or not this initial investment elicits a return investment by the beneficiary; (3) whether the interaction is direct, i.e. between two partners, or indirect (involving third parties) and (4) whether only actions that increase the fitness of the partner or also fitness reducing actions (punishment) are involved in the interaction. Asking these questions with regards to concepts in the literature on cooperative breeding, we found that (a) it is often straightforward to relate these concepts to general mechanisms of cooperation, but that (b) a single term (such as ‘pay-to-stay’, ‘group augmentation’ or ‘prestige’) may sometimes subsume two or more distinct mechanisms, and that (c) at least some mechanisms that are thought to be important in cooperative breeding systems have remained largely unexplored in the theoretical literature on the evolution of cooperation. Future theoretical models should incorporate asymmetries in power and pay off structure caused for instance by dominance hierarchies or partner choice, and the use of N-player games. The key challenges for both theoreticians and empiricists will be to integrate the hitherto disparate fields and to disentangle the parallel effects of kin and non-kin based mechanisms of cooperation.