I usually find that this argument has a very narrow definition of incentives. By definition all behaviour is maintained by reinforcement in the learning history, whether ontogenic or phylogenic. Multilevel selection is the most robust understanding of life we have, the opposite of a just so story.