Morality Games 311 Psychologists and economists sometimes take this “norm” as given, without asking where it comes from, and a naive reading of Trivers would lead one to think that we should be sensitive to the magnitude of the initial favor and whether it is manipulative. However, according to the above model, reciprocity is the Nash equilibrium, even if the favors are not evenly matched or manipulative, since, in equilibrium, we are neither sensitive to such quantitative distinctions nor to whether the initial reci- procity was manipulative, unless these facts are commonly known. Self-Image Concerns. People sometimes play mental tricks in order to appear to themselves as pro-social. For example, in an experiment, subjects will voluntarily take on a boring task to save another subject from doing it, but if given the option of pri- vately flipping a coin to determine who gets the task, they often flip—and flip, and flip again—until the “coin” assigns the task to the other subject (Batson, Kobrynowicz, Dinnerstein, Kampf, & Wilson, 1997). Why would we be able to fool ourselves and not, say, recognize that we are gaming the coin flip? Why do we care what we think of ourselves at all? Are there any constraints on how we will deceive ourselves? Such self-image considerations can be explained by noting that our self-image can act as a simple proxy, albeit an imperfect one, for what others think of us, and also that we are more convincing to others when we believe something ourselves (Kurzban, 2012; Trivers, 2011). This explanation suggests that the ways we deceive ourselves correspond to quirks described throughout this section—for example, we will absolve ourselves of remaining strategically ignorant even when it is easy not to, or be con- vinced that we have done good by voting, even if we cannot swing an election. Framing Effects. Whether we contribute is highly dependent on the details of the experiment, such as the choice set (List, 2007) and the labels for the d