To say something meaningful about the evolutionary psychology of evil, it is also necessary to say something about the evolutionary psychology of good, or altruism (Duntley & Buss, 2004). 1 Recent research on applied human moral psychology has documented a bundle of evolved cognitive modules and attendant heuristics governing our intuitions (Bloom, 2013; Cushman & Greene, 2012; Cosmides & Tooby, 2006; Haidt, 2013; Warneken & Tomasello, 2013). In a nutshell, we tend to think agents and acts are moral or immoral based on gut feeling. Yet we are highly adept at rationalizing these whims post hoc, making them seem principled and consistent. In his synthesis of recent moral psychology research, philosopher and neuroscientist Joshua Greene (2014) lays out our best current understanding of the evolved human righteousness: “Morality is a set of psychological adaptations that allow otherwise selfish individuals to reap the benefits of cooperation” (p. 185). This definition neatly captures the evolutionary logic of kin selection and reciprocal altruism, the central dimensions along which prosocial cooperation can unfold (Axelrod & Hamilton, 1981).2 The problem these mechanisms evolved to negotiate is that competition, not cooperation, is the default mode of interaction between agents coexisting in ecologies of finite resources (Dawkins, 2006a, 2006b; Williams, 1966). Kin selection and reciprocal altruism enable individuals to reconfigure the incentive structure of Darwinian first-order competition by exploiting two gametheoretic facts. In the case of kin selection, related organisms’ likelihood of sharing altruistic genes increases in proportion to the organisms’ degree of relatedness (Krebs, 1987). This allows such altruistic genes to proliferate among kin, when otherwise they would only be to the self-sacrificing agent’s detriment. In the case of reciprocal altruism, if organisms are faced with options for iterated rounds of positive-sum interaction, and provided this interaction has no preset terminus (Pinker, 2011, pp. 645–646), it pays to sign the social contract and cooperate according to a Tit for Tat algorithm.3 Indiscriminate altruism is a losing strategy that falls prey to exploiters. From the perspective of individual organisms, of course, the very best scenario is to derive fitness benefits from interactions without reciprocating. This tension can select for selfish exploitation even in a context of cooperating altruists. The fundamental payoff matrix of cooperation and defection undergirds the rich human moral tapestry, including feelings of gratitude, honor, guilt, and righteous indignation (Greene, 2014). It is also the first part of the evolutionary story about why we designate certain acts, individuals, and groups as evil, as I will show. At its core, then, human intuitive morality is about within-group cooperation. Kin selection and reciprocal altruism enable selfish genes coding for unselfish behaviors to spread through the population. The flipside is that cheaters, or ‘free-riders,’ may take advantage of altruists cooperating by default. As psychologist Steven Pinker puts it, “a person’s own level of virtue is a tradeoff between the esteem that comes from cultivating a reputation as a cooperator and the ill-gotten gains of stealthy cheating” (2011, p. 591). In a cognitive species such as ♥♥♥♥ sapiens, this means that cooperating agents must gauge the prospective contributions of potential associates. Evolutionary psychologists Leda Cosmide and John Tooby (Tooby & Cosmide, 2008, 2010) systematized this evolutionary problem in the concept of the Welfare Tradeoff Ratio (WTR). If organism A is willing to trade off a relatively high measure of its own welfare against a relatively low measure of organism B’s welfare, then organism A exhibits a high WTR in that particular relation. As members of a social species, humans should preferentially court high-WTR conspecifics, which is why they are programmed to assess the prospective fitness returns of current and potential liaisons. But the argument cuts both ways. Others are only motivated to engage us insofar as we willingly incur costs to promote the welfare of in-group members. Failing this, we effectively advertise a limited association value to them, thereby motivating them to avoid or even punish us. And humans have numerous indirect ways of figuring out on whom to bet. In fact, evolutionary thinkers argue that gossip evolved precisely to assess the association value of ingroup members through communal coregistration of offending behaviors (Nowak & Sigmund, 1998). This mechanism of indirect reciprocity is supported by the fact that gossip is primarily about transgressions of social and moral mores (Dunbar, 2004). Cheating, exploitation, and malice, from this perspective, are expressions of a socially unacceptable WTR configuration that negates the prosocial ethos of the group, threatening to unravel it in the process. Depending on a variety of factors, such as the association value of the offender, a restorative, avoidant, or retributive effort may upregulate offenders’ WTR, or at least its extrinsic expression.4 This is supported by a sizable literature demonstrating that costly prosocial punishing occurs spontaneously in Public Goods games 5 (Greene, 2014, pp. 58–59). Prosocial punishers pay a sum of their own money to reduce the payoff of free-riders. The upshot is that contributions to the common pot begin to ratchet upward. Moderate retributive sanctions thus seem to carry a recalibrational rationale at the group level (Boehm, 2014). But, as Tooby and Cosmides point out, all of this is mediated by emotions, or, in the language of the neuroscientist Antonio Damasio (2006), ‘somatic markers’ that lubricate the neural pathways of specific behaviors. And indeed prosocial punishers in Public Goods games report that they are driven by the emotional satisfaction of setting things straight (Haidt, 2013, p. 209). The point is that none of us is consciously computing or recalibrating the WTRs of our associates; we just feel good or bad based on the outcome of a social exchange. These emotions in effect constitute a data format that promotes context-specific action, often as an answer to the quintessential adaptive questions for any member of a social species: approach or avoid? Deter or reinforce? Tooby and Cosmides’ conceptual framework makes it possible to describe human intuitions about evil in principled terms. We feel that someone is evil when that person grossly violates our marginal expectations of their WTR in relation to ourselves, our friends, or some other real or symbolic entity held in high regard. The category is of course fuzzy; there is no set boundary between ‘disagreeable,’ ‘nasty,’ and ‘evil.’With this in mind, folk psychological evil emerges as an antisocial agentic designation that licenses extreme retributive action. You may avoid or chastise the disagreeable person, but you may ostracize or even kill the evil person, who, by definition, is beyond the recalibrational reach of social sanction. The punishment regime thus tracks a functional gradient that is entirely consistent with its Darwinian origin (Boehm, 1999; Keltner, Haidt, & Shiota, 2006). Notice that this skeleton of evil has a fractal character. Because the human intuitive psychology can conceptualize groups as if they were people with singular goals and inclinations, the WTR concept maps uniformly onto groups, subgroups, and individuals alike. It is similarly applicable to superstitions such as spirits, monsters, and other supernatural agents because they satisfy the input conditions of our neural agency detection hardware (Boyer, 2002; Norenzayan, 2010). Such phantasms may even subvert the scale entirely by trading their own welfare against their victims’ suffering, making them, and any human ‘monsters’ like them, truly antisocial and evil. The WTR concept has expl