The Skeptics Society & Skeptic magazine

Doubting Altruism
New research casts a skeptical eye
on the evolution of genuine altruism

by Kenneth W. Krause

As we soar into an inspiring new era of genomics, genetic manipulation, and, potentially, the directed evolution of our own species, naturalists urge us to remain partially grounded — to keep digging for long-buried evidence of key pre-historical developments. In so doing, however, the world’s leading anthropologists and primatologists have immersed themselves in a now-roiling debate over the origins of human morality in general and altruism in particular.

Some say that altruism — sometimes referred to as “other-regarding preferences” or “unsolicited prosociality” — is nothing more than a veneer, a cultural innovation humans alone have achieved in order to collectively restrain each individual’s natural proclivity to serve only herself, her close genetic relatives, and those who have demonstrated an adequate inclination to reciprocate to her eventual benefit. For these folks, no act can be characterized as wholly unselfish.

Others argue that altruism is more primitive than culture and, in fact, considerably more ancient than the human species itself. Other-regarding preferences, they say, are deeply innate, predating even the phylogenetic split that occurred six million years ago among the common ancestors of chimps and bonobos on the one hand and all species of hominid on the other. According to this camp’s credo, selflessness is as natural as appetite.

One line of experiments has confronted the issue directly, inquiring whether non-human primates will seize opportunities to assist others. In 2005, for example, UCLA anthropologist Joan Silk and others chose 18 chimpanzees (Pan troglodytes) as the subjects of two such experiments, conducted in Louisiana and Texas.¹ Chimps are among the primates most likely to exhibit unsolicited prosocial behavior, they reasoned, because in the wild they regularly hunt, patrol, and mate-guard cooperatively.

In each study, subject chimps were allowed to deliver food to other chimps, or “conspecifics,” at no cost to themselves. The test apparatuses provided each confined subject with two options — the 1/0 choice where it could acquire food only for itself, and the 1/1 choice where it could obtain food for both itself and its separately caged partner. As an essential control, acting chimps were given the same options with no partners present (Figure 1).

Figure 1. A test apparatus providing a subject with two options: acquire food only for itself (1/0), or obtain food for both itself and its separately caged partner (1/1).

Silk’s team predicted that if chimps are truly altruistic they should choose the 1/1 option more often than the 1/0 option when a conspecific is there to benefit. But that wasn’t the case. In Louisiana, not one of the seven subjects chose the 1/1 option significantly more often when partnered. In Texas, the remaining 11 actors went with both the 1/1 and the 1/0 option an average of only 48 percent of the time when another chimp was present.

“The absence of other-regarding preferences in chimpanzees,” the authors inferred, “may indicate that such preferences are a derived property of the human species tied to sophisticated capacities for cultural learning, theory of mind, perspective taking and moral judgment.” Nevertheless, Silk’s team remained open to the prospect that altruism might be detected among primates that, in some crucial ways, were even more cooperative than chimps. We will consider that possibility later.

Altruism’s Alter-Ego

A closely related line of experiments has tackled the same issue from a different direction, asking instead whether primates display a rudimentary sense of fairness in some form of “inequity aversion” (IA). If an animal reacts negatively to its own relative overcompensation, we say it has demonstrated some sensitivity to “advantageous inequity.” If it merely responds to a conspecific’s superior gain, on the other hand, the animal has shown aversion only to “disadvantageous inequity.”

The former inclination probably evolved after (and, morally speaking, is emphatically more advanced than) the latter because an animal sensitive to its own advantage can demonstrate not only an egocentric expectation of how it should be treated, but also a communal expectation of how all members of its species should be treated. In either case, if test subjects attempt to restore equity by sacrificing their own gains — even if only to simultaneously and unceremoniously deny superior gains to their luckier partners — according to many (but not all) researchers, they have nonetheless acted altruistically.

In 2003, Emory University primatologists Sarah Brosnan and Frans de Waal developed token exchange experiments where tufted capuchin monkeys (Cebus apella) were measured for their reactions to situations in which their partners received greater food rewards.² In the end, shortchanged subjects proved less likely to complete exchanges for identical tokens, and withdrew even more frequently when their partners received prizes for no tokens at all. These now-classic results have been widely interpreted as formidable evidence of disadvantageous IA in primates.

Two years later, Brosnan, de Waal, and Hillary Schiff released the outcomes of a similar study of adult chimpanzees.³ In order to distinguish the effects of social alignment, the team chose four animals that had lived continuously in pairs and 16 others that had been housed together at the Yerkes National Primate Research Center in Atlanta, Georgia for either 30 years or eight years prior to testing. As in the 2003 experiment, subjects were given tokens — in this case, rather useless and nondescript chunks of white PVC pipe — which they had been trained to return for either cucumber slices (the low-value reward) or grapes (the high-value reward).

During the inequity test, examiners initially allowed the partner chimps to exchange for a juicy, delicious grape — while eager subjects observed, of course — and then offered the subjects a relatively dry and no doubt disappointing cucumber slice. The examiners diligently recorded the subjects’ reactions, noting whether they had accepted or rejected their prizes. Brosnan discovered first that, when the tables were turned, subjects did not react negatively when given a superior reward and, thus, were likely not averse to advantageous inequity. Whether such a finding actually distinguished chimpanzees from humans in any meaningful way, the authors noted, was questionable.

Second, according to Brosnan, the results confirmed that disadvantageous IA was “present and robust” among chimpanzees, although to significantly different degrees depending on each subject’s social history. Chimps that had lived in pairs or in relatively novel groups reacted most intensely, while animals from older, more tightly-integrated groups appeared more accepting of inequity — all of which could be entirely consistent with human predilections to either “make waves” or “go with the flow,” depending primarily on their social milieu. Tolerance of inequity, Brosnan suggested, may be more a function of group size and intimacy than either moral choice or any isolated cognitive factor. So by the end of 2005, very little if anything had been truly settled. The experiments would continue and become ever more creative and exacting, but the already muddied anthropological waters would grow more cluttered and murkier still.

High Expectations

In 2006, three teams from the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany published studies that in one way or another challenged these landmark outcomes from Yerkes. Julian Brauer’s group tested for IA among chimps, bonobos, gorillas, and orangutans — 30 individual great apes in all — and produced a pattern of food rejection that was opposite to that reported by Brosnan.⁴ In other words, instead of snubbing more food after seeing their partners receive tastier treats, Brauer’s apes actually rejected less food.

All the same, the authors did not infer that apes were necessarily insensitive to unfairness. In fact, at one point they questioned whether food refusal was a fair test of IA to begin with. Inequity-wary apes, after all, might decide to accept lower-quality spoils simply in an effort to offset the higher-quality gifts bestowed upon their partners. Citing a then-recently published study questioning chimpanzee altruism, however, Brauer’s team finally betrayed a clear inclination to attribute their results to the so-called “food expectation hypothesis,” which asserts that the mere act of witnessing a conspecific’s receipt of superior food will create an anticipation of acquiring the same food for oneself. Such an expectation might explain why Brauer’s chimps begged more vigorously and why many of her apes generally remained at their testing stations much longer after having witnessed partner overcompensation.

The study Brauer cited had been conducted by a second German team led by Keith Jensen.⁵ The key problem with Brosnan’s examination, according to Jensen, was that subject chimps were never allowed to convey their mind-sets by actually correcting unequal outcomes. Silk’s group had devised a somewhat more effective experiment in this respect, the authors commended, but even they had failed to test for anything more than selfishness (the 1/0 option) or mutualism (the 1/1 option). In Jensen’s experiments, by comparison, 11 chimps participated in three separate studies collectively designed to reveal expressed tendencies toward altruism and spite as well.

In study one, each subject was allowed to pull one of two tables toward itself. The first table contained bananas accessible to both the subject and its partner; the second table held fruit accessible only to the subject. Either way, subjects received the same reward. But because Jensen’s chimps predominantly chose the mutually accessible table in both the test (partner present) and control (partner absent) conditions, the results were inconclusive as to selfishness and mutualism. Nevertheless, Jensen vied, this initial phase of the experiment did show that his chimps were not averse to disadvantageous inequity, at least with regard to relative effort expended.

Studies two and three tested for spite and altruism. In neither case could any subject receive a reward for pulling any table closer to itself or its partner. In experiment two, acting chimps could have conveyed an other-regarding preference by pulling tables accessible only to partners, or passive spite by doing nothing at all more frequently in the test condition than in the control condition. But they did neither.

In experiment three, the scheme was altered slightly such that, in order to deny food to their partners, subjects needed to actively draw those partners’ trays away. But, once again, the Leipzig chimps were as likely to do nothing in one condition as in the other, thus failing to demonstrate active spite as well. The authors noted, however, that two of their six chimps did express potential signs of altruism. But these animals also tended to beg or harass their partners following delivery of the fruit, thus raising the possibility that they intended to only benefit themselves.

Food for Thought

Felix Warneken and Michael Tomasello, also from the Planck Institute, decided to examine the issue through an entirely different type of experiment. If altruism exists among our primate cousins, they judged, it might more readily be elicited with something less critical to individual survival than food. So Warneken and Tomasello tested both human children (24 18-month-old infants) and three young chimpanzees (34, 54, and 54 months old) for their willingness to help human caretakers (quite familiar to the chimps) with some task absent of any possible expectation of reward.⁶

As predicted, the children assisted experimenters more often and in a greater variety of tasks than the chimps. Nonetheless, each of Warneken’s chimps reliably helped a reaching human obtain apparently desired objects. Although young humans clearly cooperate to degrees found in no other species, the authors concluded, “our nearest primate relatives show some skills and motivations in this direction as well.”

By the summer of 2007, Warneken had assembled another team and published the results of similar “instrumental helping” experiments calculated to address several important and yet unanswered questions — in particular, whether 36 semi-free ranging chimps would spontaneously help unfamiliar humans and genetically unrelated conspecifics in addition to their caretakers, and whether they would do so at some significant cost to themselves.⁷

Again, as expected, infant children helped more quickly. But the chimps performed just as reliably regardless of their partners’ familiarity or species, even when they had to expend a little extra effort to do so. “The roots of human altruism may go deeper than previously thought,” Warneken ultimately concluded, “reaching as far back as the last common ancestor of humans and chimpanzees.”

Later that year, however, Keith Jensen’s team cast a skeptical eye on Warneken’s conclusions in two well-focused examinations of potential IA among chimps. The first pair of experiments probed 11 animals’ capacities for spite — altruism’s evil twin, if you will.⁸ Jensen reasoned that Chimps might be characterized as altruistic, at least in a punitive sense, if they choose to act out against conspecifics due to an abstract sense of fairness. In the first study, caged subjects were allowed to pull ropes to collapse food-laden tables drawn away by humans toward different enclosures that either contained other chimps (test condition) or were empty (control condition). In the end, the chimps appeared to be indifferent. Actors were just as likely to collapse the platforms when they approached empty cages as when they neared hungry conspecifics.

In the second study, subjects were exposed to three conditions. In conditions one and two, much as before, human experimenters pulled the tables away from subject animals and toward partners or empty cages. In the last condition, however, it was the partner chimps that were allowed to drag the tables away from subjects. Between the first two conditions there was no real disparity, indicating again that subjects didn’t really care whether their partners benefited inequitably. Between the first two and the third conditions, on the other hand, subjects were significantly more likely to drop the tables when other chimps, as opposed to humans, began drawing them away. From these combined results, Jensen concluded that although chimps are certainly vengeful, “[s]pitefulness may thus be a peculiarly human phenomenon.”

Ultimatums (and more Expectations)

Hailing it as the “benchmark test for examining sensitivity to fairness and other-regarding preferences,” Jensen then unleashed his 11 subjects on a chimp-friendly version of the celebrated ultimatum game.⁹ Proposer animals were permitted to make one of two possible offers to their receiving partners, potentially retaining either 100, 80, 50, or 20 percent of the spoils for themselves in each trial. If the receiver accepted the offer, each party got what it wanted. But if the receiver rejected the offer — having noted what the proposer intended to keep for itself — neither animal received any reward.

Presumably out of some concern for fairness, humans proposers tend to make equitable offers of 40 to 50 percent or, as receivers, to reject offers of 20 percent or less, thus confounding the economic model of rational self-interest (so-called Homo economicus). This was not how the Leipzig chimps reacted, however. Proposers chose not to make fair offers and receivers opted to accept all nonzero offers without hesitation or perceptible sign of irritation. While the authors cautioned that these outcomes “may be in part be a reflection of the fact that active food sharing is rare among the species,” they were clearly inclined to attribute such behaviors to the chimps’ absent sense of justice.

Figure 2. A monkey returns a token to the experimenter while using her left paw to steady the human hand. Her partner looks on. In this “hidden-reward exchange” the monkey does not see the reward she is to receive before successful exchange. (Redrawn from an illustration done from a video still by Gwen Bragg and Frans de Waal.¹⁰)

In late 2007, Megan van Wolkenten, working with Brosnan and de Waal, finally published a narrowly tailored response to Brauer and others addressing the alleged preeminence of food expectation over IA.10 They used the now-familiar token exchange experiment — this time enhanced with an additional condition where food rewards were shown to subjects well prior to exchange — on 13 capuchin monkeys (Figure 2). But, contrary to the predictions of various expectation hypotheses, behavioral changes did not depend on either greed or frustration. Rates of refusal among subjects, in fact, increased not when higher-value grapes were merely visible, but only when they were actually bestowed upon partners.

Importantly, van Wolkenten’s subjects also made significantly fewer exchanges when forced to expend more effort for the same lower-value cucumbers received by partners. As the food value increased, however, effort became secondary, indicating that capuchins are willing to reprove inequity only when the cost of doing so is slight. This appreciable yet limited brand of IA, the authors proposed, “likely evolved in conjunction with cooperative enterprises,” and “may characterize a great variety of social animals.”

More Monkey Business

By the end of 2007, then, the combined body of research had established mixed results at best, especially with regard to the great apes. Recalling Joan Silk’s suggestion that true altruism might be discovered among primates even more social than chimpanzees, Swiss anthropologist Judith Burkhart’s team decided to test 26 common marmoset monkeys (Callithrix jacchus) in two studies — one for related, one for unrelated pairs — involving hungry partners and subject-operated food trays (Figure 3).¹¹

Figure 3. The donor has the choice between two trays representing the payoff distributions (0,1) (upper tray with a cricket in the left food bowl) or (0,0) (lower tray). If the donor pulls the tray with the (0,1) payoff distribution, it results in a payoff to the recipient but none to itself. ¹¹

The experimenters provided each actor with a 0/0 option and a 0/1 option only, thus eliminating all potential for subject rewards. Because marmosets are cooperative breeding New-World monkeys, Burkhart predicted that if any primate should display an unsolicited prosocial tendency capable of overcoming any penchant for envy, it would be this species, despite their theory of mind deficit and general cognitive shortcomings.

Burkhart was right. Kin or no kin, marmoset subjects — fully schooled with the test apparatus and, thus, aware of the experiments’ consequences — pulled the 0/1 tray more often when their partners were present in adjacent cages than in the control condition when their partners were absent. Remarkably, the disparity widened significantly when female “helpers” — which, despite this distinction, tend not to carry other monkeys’ infants in the wild — were eliminated from the analysis. Because humans and New-World monkeys are the only primates that behave as cooperative caretakers, Burkhart proposed, strong altruism may have evolved within such groups independently, and not necessarily among the ancestors common to chimps, bonobos, and humans.

More Food for Thought

The thick, swirling waters of controversy have spilled largely unabated into 2008. Working with Brosnan, Silk, and others, American evolutionary psychologist Jennifer Vonk published a detailed study of low-cost, conspecific-directed altruism among 18 chimpanzees at the University of Louisianna’s Cognitive Evolution Group laboratory.¹² In two separate experiments involving two different apparatuses and two distinct groups of chimps, actors were given the options to trigger rewards for themselves alone, for their partners alone, or for both themselves and their partners.

The team chose these three options in order to address important criticisms of previous experiments involving food. Because her chimps were allowed to act prosocially only after having fed themselves, Vonk argued, this method avoided the possibility that subject animals might be distracted from an otherwise spirited altruistic tendency by the potent and ever-present need to feed.

If chimps are really other-regarding, the authors reasoned, subjects should deliver rewards to partner enclosures at some point during the experiment, but more often in the partner-present test condition than in the partner-absent control condition. By contrast, if chimps are indifferent to the welfare of others, actors should minimize their personal costs by obtaining rewards only for themselves. Ultimately, the presence of awaiting partners in other enclosures had no significant effect on subjects in either experiment. At first, actors consistently released both rewards. But delivery rates to other cages always decreased as subjects learned that such efforts would not benefit them.

Notably, one of the 11 chimps tested in the second experiment did choose to act prosocially, but these results could not be replicated. “[W]hile chimpanzees’ behavior is consistent with standard evolutionary models based on kinship and reciprocity,” Vonk insisted, “human cooperation and prosociality may require an emerging class of evolutionary models, rooted in the coevolutionary interaction of genes and culture.”

Agreeing to Disagree

Despite these equivocal results, some scientists still see altruism as a considerably more ancient impulse, born of the intense parental and, thus, empathic instinct. Frans de Waal, as one prominent example, appears to be thoroughly convinced that some skeptics of primate altruism have their arguments backwards — at least in one crucial respect. “[E]mpathy evolved in animals as the main proximate mechanism for [individually] directed altruism,” he explained in a recent review, and it is empathy — not self-interest — that “causes altruism to be dispensed in accordance with predictions from kin selection and reciprocal altruism theory.”¹³ Although gene propagation and benefit exchange may be the evolutionary or ultimate cause of altruism, only a spontaneous emotional response to another being’s situation can possibly trigger or proximately cause an altruistic impetus.

In his latest study of non-cooperatively breeding monkeys, de Waal discovered that brown capuchins will predominantly choose the 1/1 mutual option over the 1/0 selfish option, depending on the subjects’ familiarity with their partners.¹⁴ Although his monkeys’ other-regarding tendencies clearly turned on social closeness, de Waal nevertheless concluded that because kinship was critical and because his subjects had no means of predicting return favors, only empathy could explain this study’s results.

When I asked him about the persisting debate, de Waal proposed that the scientific community has become polarized between evolutionary biologists on the one side and, on the other, a discrete group of economists and anthropologists that “has invested heavily in the idea of strong reciprocity,” which absolutely demands discontinuity between humans and all other animals. As for the results obtained by Silk and others, de Waal offered, experiments such as these involving repeated trials and frequent rewards are vulnerable to “side-biases” that can skew outcomes.

Sarah Brosnan, a former student of de Waal’s and now Assistant Professor of Psychology at Georgia State University, remains ambivalent. Her subjective though surely copious experience with both apes and monkeys informs her that at least some of these animals do seem altruistic. Even so, she told me, “there is not too much evidence for this outside some of Frans’ and Felix Warneken’s work.” But cooperation in all species, she emphasized, “is much more likely to be based on emotion and relationships than on cognitive calculations.”

Both de Waal and colleague Keith Jensen are doubtful that even chimpanzees possess the cognitive capacities requisite for delayed reciprocation. But for Jensen, the added conclusion that chimps must be altruists simply doesn’t follow. “De Waal’s use of the term ‘empathy’ is somewhat contentious,” he told me, “and the evidence he provides for empathy is [anecdotal and] not very robust.”¹⁵ More evidence is needed, he admitted, but, like Jennifer Vonk, his “working hypothesis” is that other-regarding preferences emerged at some point during human evolution only.

Even so, both Brosnan and Jensen conceded that the distinction between food exchange and instrumental helping is a potentially crucial one. Indeed, Jensen and Felix Warneken are now collaborating on a new project to determine whether food rewards might interfere with genuine other-regarding preferences. Although “food exchange is not a bad test for altruism,” Warneken reminded me, it explores “only one type of potentially altruistic behavior.” In the more sensitive context of instrumental tasks, he added, chimpanzees have repeatedly demonstrated solid helping tendencies.

When I asked Warneken about Vonk’s latest attempt to neutralize the nutritional imperative, he warned that Vonk’s chimps might not have fully understood how the apparatuses worked during that experiment’s altruism phase. “The pattern of results,” he argued, “still suggests that the subjects had a tendency to try to obtain the reward for themselves.” Plugging Jensen’s 2006 study as the most convincing presentation to date of limited prosociality among chimps, Warneken recommended that future researchers follow that team’s lead, at least with respect to designing an apparatus that animals might comprehend more intuitively.

Where to Go from Here

Everyone agrees that more work needs to be done, and that no research could be more germane to achieving a competent grasp of who we are as a species and where we might be headed. If altruism is in fact deeply innate to humanity’s collective being, we may have to rethink a number of things, including some of our most established political and economic assumptions.¹⁶ Jensen summed it up pretty well when I invited him to characterize his work’s significance:

This research is interesting to the question of what makes humans special, if, indeed, they are. Most research in the past has focused on “cold cognition” such as abstract reasoning, language and tool use. Social motivations and emotions — “hot cognition” — are just as important, and may even be central to the emergence of human ultrasociality. Holding a lens up to ourselves after focusing it on other species will help us see ourselves more clearly.

So it looks like we’ll be hearing a great deal more from these and other esteemed authorities during the coming years. Sadly, however, the indispensable subjects of these investigations seem to be living on borrowed time, the African great apes especially. If scientists can ever clear the dim, shadowy depths of altruistic origins, they’ll have to act quickly before our own dark natures drive our ancestral cousins into extinction.

References

1. Silk, J. B., Brosnan, S. F., Vonk, J., Henrich, J., Povinelli, D. J., Richardson, A. S., Lambeth, S. P., Mascaro, J. & Shapiro, S. J. 2005. “Chimpanzees Are Indifferent to the Welfare of Unrelated Group Members.“ Nature, 437, 1357–1359.
2. Brosnan, S. F. & de Waal, F. B. M. 2003. “Monkeys Reject Unequal Pay.” Nature, 425, 297–299.
3. Brosnan, S. F., Schiff, H. C. & de Waal, F. B. M. 2005. “Tolerance for Inequity May Increase With Social Closeness In Chimpanzees.” Proc. R. Soc. B, 272, 253–258.
4. Brauer, J., Call, J. & Tomasello, M. 2006. “Are Apes Really Inequity Averse?” Proc. R. Soc. B, 273, 3123–3128.
5. Jensen, K., Hare, B., Call, J. & Tomasello, M. 2006. “What’s in it for me? Self-regard Precludes Altruism and Spite In Chimpanzees.” Proc. R. Soc. B, 273, 1013–1021.
6. Warneken, F. & Tomasello, M. 2006. “Altruistic Helping In Human Infants and Young Chimpanzees.” Science, 311, 1301–1303.
7. Warneken, F., Hare, B., Melis, A. P., Hanus, D. & Tomasello, M. 2007. “ Spontaneous Altruism By Chimpanzees and Young Children.” PloS Biology, 5(7), e184.
8. Jensen, K., Call, J. & Tomasello, M. 2007. “Chimpanzees Are Vengeful But Not Spiteful.” Proc. Natl. Acad. Sci., USA, 104, 13046–13050.
9. Jensen, K., Call, J. & Tomasello, M. 2007. “Chimpanzees are Rational Maximizers In an Ultimatum Game.” Science, 318, 107–109.
10. van Wolkenten, M., Brosnan, S. F. & de Waal, F. B. M. 2007. “Inequity Responses of Monkeys Modified by Effort.” Proc. Natl. Acad. Sci., USA, 104, 18854–18859.
11. Burkhart, J. M., Fehr, E., Efferson, C. & van Schaik, C. P. 2007. “ Other-Regarding Preferences In a Non-Human Primate: Common Marmosets Provision Food Altruistically.” Proc. Natl. Acad. Sci., 104, 19762–19766.
12. Vonk, J., Brosnan, S. F., Silk, J. B., Henrich, J., Richardson, A., Lambeth, S., Schapiro, S. & Povinelli, D. J. 2008. “Chimpanzees Do Not Take Advantage of Very Low Cost Opportunities to Deliver Food to Unrelated Group Members.” Animal Behavior, 75, 1757–1770.
13. de Waal, F. B. M. 2008. “Putting the Altruism Back Into Altruism: The Evolution of Empathy.” Annu. Rev. Psychol., 59, 279–300.
14. de Waal, F. B. M., Leimgruber, K. & Greenberg, A. R. 2008. “Giving Is Self-rewarding for Monkeys.” Proc. Natl. Acad. Sci., USA. 105, 13685–13689.
15. See also, Silk, J. B. 2007. “Empathy, Sympathy, and Prosocial Preferences In Primates.’ In: The Oxford Handbook of Evolutionary Psychology. (Ed. by R. I. M. Dunbar & L. Barrett), pp. 115–126. Oxford: Oxford University Press (“ Current claims for the existence of empathy, sympathy, moral sentiments, and other-regarding preferences in other primates rest on an insecure empirical foundation.”).
16. See, e.g., Bowles, S. 2008. “Policies Designed for Self-interested Citizens May Undermine ‘the Moral Sentiments’: Evidence from Economic Experiments.” Science, 320, 1605–1609 (“Economists, psychologists, and others … are well on their way to constructing an economic psychology of the interplay of self-regarding and other-regarding motivation that may eventually enlighten mechanism design and public policy.”).