Neither evolutionary theory nor behavioral evidence is consistent with Rachlin's view of altruism as a learned, domain-general learned habit displayed because of its intrinsic value. But human beings can be psychologically motivated by altruism while still reaping a genetic benefit from their altruistic actions.

Rachlin uses the word “choice” 80 times, whereas “emotion” does not appear. In contrast, “Empathy: Its ultimate and proximate bases” by Preston and de Waal, uses the word “emotion” 139 times and “choice” once. This commentary compares these ways of approaching empathy and altruism, relating Rachlin's approach to Gilligan's Morality of Justice and Preston and de Waal's to the Morality of Caring.

There are three concerns regarding Rachlin's altruism model. First, proximal causal mechanisms such as those identified by cognitive neuroscientists and behavioral neuropharmacologists are not emphasized. Second, there is a lack of clear testable hypotheses. And third, extreme forms of altruism are emphasized rather than common forms. We focus on an overarching theme – proximal mechanisms of individual differences in altruism.

Using a simple learning agent, we show that learning self-control in the primrose path experiment does parallel learning cooperation in the prisoner's dilemma. But Rachlin's claim that “there is no essential difference between self-control and altruism” is too strong. Only iterated prisoner's dilemmas played against reciprocators are reduced to self-control problems. There is more to cooperation than self-control and even altruism in a strong sense.

We agree with Rachlin's argument that altruism is best understood as a case of self-control, and that a behavioral analysis is appropriate. However, the appeal to teleological behaviorism and the value of behavioral patterns may be unnecessary. Instead, we argue that altruism can generally be explained with traditional behavioral principles such as negative reinforcement, conditioned reinforcement, and rule-governed behavior.

If altruism requires self-control, people must consider altruistic acts as costly, the benefits of which will only be recouped in the future. By contrast, I shall present evidence that altruism is dictated by emotions: Altruists secure an immediate payoff from performing altruistic acts, so no element of self-control is present, and no future reward is required or expected.

The primrose path and prisoner's dilemma paradigms may require cognitive (executive) control: The active maintenance of context representations in lateral prefrontal cortex to provide top-down support for specific behaviors in the face of short delays or stronger response tendencies. This perspective suggests further tests of whether altruism is a type of self-control, including brain imaging, induced affect, and dual-task studies.

Altruism can be understood in terms of traditional principles of reinforcement if an outcome that is beneficial to another person reinforces the behavior of the actor who produces it. This account depends on a generalization of reinforcement across persons and might be more amenable to experimental investigation than the one proposed by Rachlin.

Any behavior belongs to innumerable overlapping types. Any adequate theory of emergence and retention of behavior, whether psychological or biological, must give us not only a general mechanism – reinforcement or selection, for example – but a reason why that mechanism applies to a particular behavior in terms of one of its types rather than others. Why is it as this type that the behavior is reinforced or selected?

Altruism is traditionally encouraged by promoting a goal, for example, going to heaven. In contrast, Rachlin argues that altruistic behavior can be sufficiently reinforced by the abstract intrinsic reward that comes from maintaining an unbroken pattern of altruistic behavior. In my experience, there are very few people for whom this is true. For fellow atheists and anti-theists, I suggest an alternative.

The concept of “intrinsic reinforcement” stretches the use of “reinforcement” beyond where it is valuable. The concept of the “self-system,” though fuzzy at the edges, can cover experience as well as the behaviour of altruistic acts.

Rachlin's thought-provoking analysis could be strengthened by greater openness to evolutionary interpretation and the use of the directed attention concept as a component of self-control. His contribution to the understanding of prosocial behavior would also benefit from abandoning the traditional (and excessively restrictive) definition of altruism.

Rachlin basically marshals three reasons behind his unconventional claim that altruism is a subcategory of self-control and that, hence, the prisoner's dilemma is the appropriate metaphor of altruism. I do not find any of the three reasons convincing. Therefore, the prisoner's dilemma metaphor is unsuitable for explaining altruism.

Biological, cognitive, and learning explanations of altruism, selfishness, and self-control can be integrated in terms of adaptive strategies. The key to understanding why humans and other animals sometimes resist temptation and sacrifice their immediate interests for the sake of others lies in mapping the design of the evolved mental mechanisms that give rise to the decisions in question.

The mother sacrificing herself while rescuing someone else's child is a red herring. Neither behaviorism nor cognitivism can explain it. Unlike behaviorism, however, the cognitive process of projection can explain cooperation in one-shot social dilemmas.

Rachlin shows that experiments about social cooperation may fruitfully be grouped with experiments on self-control, and that this suggests interesting possibilities for practical behavioral controls. The concepts of selfishness and altruism, however, that inform his theorizing about these experiments, do not serve to provide understanding of the behavior that commonly is referred to, derogatorily, as selfish.

Altruism by definition involves the self's evaluation of costs and benefits of an act of the self, which must include cost to the self and benefits to the other. Reinforcement value to the self of such acts is greater than the costs to the self. Without consideration of a self-system of evaluation, there is little meaning to altruistic acts.

Few disagree that altruism exists. The frequency and source of altruistic behavior remain mysterious, however.

Rachlin adds to the already long list of proposals for reducing what might be seen as social motivation to some roundabout form of self-interest. But his argument exhibits the usual limitations, and prompts questions about what drives this apparently unending quest.

The act-pattern model of altruism is primarily a brand-equity model, which holds that being altruistic can be traded for social benefits. This is a variant of the “selfish” altruism that Rachlin decries, with altruism being dictated by cold calculations. Moreover, personal and social “self-control” may not be as similar as Rachlin suggests – although we have good (biological) reasons to sacrifice the interests of our current selves in favour of our future selves, we have no such reason to sacrifice ourselves for our neighbours. When we do sacrifice ourselves – giving in to true altruism – we will be repaid with extinction.