The problem

In the last five chapters we have seen multiple examples of how structured interactions in a population of self-interested, boundedly rational agents tend to promote moral behavior. Although this does not hold universally for all structures and all games, it does seem that it holds often enough for it to be more than a mere coincidence. The central philosophical question, then, is what, if anything, does this imply for our understanding of morality and moral theories?

It is not immediately clear that the results discussed in chapters 3–7 affect our understanding of morality at all. Perhaps the best encapsulation of the general problem evolutionary explanations of morality face – and why they might not illuminate our understanding of morality – can be found in the following observation:

As moral agents, we care about the superstructure of concepts and principles that we use to describe and evaluate our behavior and the behavior of others.