People cooperate in public goods games even when an individual’s utility maximizing strategy is to defect. A form of non-institutionalized punishment called altruistic punishment—or strong reciprocity—may explain this cooperative behavior. I consider laboratory experiments of public goods games that provide evidence of altruistic punishment and proximate explanations for that behavior. I also present theories of the evolution of altruistic punishment via group-selection, multi-level selection, and gene and culture co-evolution. Furthermore, I consider criticisms of both laboratory results and evolutionary theories that suggest weaknesses in the current research on altruistic punishment. In sum, we will likely never have a definitive explanation of the origins and evolution of human cooperation. I conclude, however, that altruistic punishment may form an integral part of that trajectory.