A Cooperative Species Herbert Gintis Santa Fe Institute

A Cooperative Species* Herbert Gintis Santa Fe Institute Central European University of Siena *Material based on Samuel Bowles and Herbert Gintis, A Cooperative Species: Human Reciprocity and its Evolution (Princeton University Press, 2011)

Princeton University Press 2011

Two Propositions on the Nature of Human Cooperation I will advance two propositions. First, people cooperate not only for self-interested reasons, but also because they are concerned about the well being of others, try to uphold social norms, and value behaving ethically for its own sake. People punish those who free-ride on the cooperative behavior of others even when they cannot expect to gain anything from it other than the pleasure of hurting someone who has done wrong.

Two Propositions on the Nature of Human Cooperation Second, we have these moral sentiments, because our ancestors lived in environments in which groups of individuals who are predisposed to cooperate and uphold ethical norms tended to survive and expand relative to other groups, thereby proliferating these pro-social motivations, both culturally and genetically.

Mutualism and Altruism Cooperation may be mutualistic; i. e. , an activity that confers net benefits both on the actor and on others. Cooperation may also impose net costs upon the individual, in which case we say that it is altruistic. By contrast to mutualistic cooperation, altruistic cooperation would not be undertaken by an individual whose motives were entirely selfregarding and thus did not take account of the effects of his actions on others or the norms of the group.

Mutualism and Altruism The evolution of cooperation that is mutualistic or involving only kin altruism---sacrifice on behalf of close relatives---are easily explained by natural selection because such behaviors help proliferate one’s genes irrespective of whose body they are in including the genes that induced the helping behavior. Cooperation could also have evolved because of reciprocal altruism, a form of mutualism in which I help you today in the expectation that you will help me at a future date when I am in need.

Mutualism and Altruism Kin altruism, mutualism, and reciprocal altruism (which is really long-run mutualism) are popular among biologists and economists alike and explain many forms of human cooperation, particularly those occurring among close kin or in very small groups (two or three individuals).

Mutualism and Altruism These models fail to explain two facts about human cooperation: that it takes place in large groups of non-kin, and it occurs in interactions where there is no repetition, and where there is complete anonymity (i. e. , there are no reputational gains from behaving prosocially) Only a model in which humans have social preferences that motivate them to participate and to punish the transgressors of social norms can explain human cooperation in large groups of non- kin.

The Evolution of Social Preferences Early modern humans inhabited the large, mammal-rich African savannah and other environments in which cooperation in acquiring and sharing food yielded substantial benefits at relatively low cost. The slow human life-history with prolonged periods of dependency of the young also made the cooperation of non-kin in child rearing beneficial.

The Evolution of Social Preferences As a result, groups that fostered cooperation in hunting, child-rearing, punishing, prevailing against hostile neighbors, and sharing truthfully transmitted information had significant advantages over other groups. Prosociality in general and altruism in particular were successful for three reasons.

The Evolution of Social Preferences First, human groups devised ways to protect their altruistic members from exploitation by the selfinterested. Prominent among these is the shunning, ostracism, and even execution of those who violate cooperative norms. Prominent also are the egalitarian practices that limit hierarchy and inequality, including the sharing of food and information, within the group: There is no “big man” in a hunter-gatherer group.

The Evolution of Social Preferences Second, humans adopted prolonged and elaborate systems of socialization that lead individuals to internalize the norms that induce cooperation, so that contributing to common projects and punishing defectors became objectives in their own right rather than constraints on behavior. Allied with norm internalization are the social emotions, such as shame, guilt, pride, and empathy that are at best rudimentary in almost all other creatures.

The Evolution of Social Preferences Third, between-group competition for resources and survival was a decisive force in human evolutionary dynamics. Groups with many cooperative members tended to dominate less cooperative groups, thereby gaining both through genetic and cultural selection.

The Evolution of Social Preferences Throughout human history, group extinction, costly group dispersal, and ostracism have been powerful mechanisms supporting the evolution of human cooperation. The extraordinarily high evolutionary stakes of intergroup competition and the contribution of altruistic cooperators to success in these contests meant that sacrifice on behalf of others, extending beyond the immediate family and even to virtual strangers in one’s group, could proliferate.

The Science of Morality The relationship between self-regard and moral sentiments has been illuminated in recent years by behavioral game theory. The experiments confirmed that self-interest is indeed a powerful motive, but also that other motives are no less important. Even when substantial sums of money are at stake, most experimental subjects are fair-minded, generous towards those similarly inclined, and nasty towards those who violate these pro-social precepts. Let us review some of this evidence.

Honesty Why would a human be unconditionally honest, as opposed to being honest only where it pays off and ruthlessly dishonest when one can get away with it? Numerous experiments indicate that most human subjects are willing to sacrifice material reward to maintain a virtuous character, even in one-shot encounters with perfect anonymity.

Honesty in the Laboratory Gneezy (2005) used a two-player game played once, under completely anonymity. Call the two players Bob and Alice. Bob is shown two options. Option A pays him $5 and pays Alice $6. Option B reverses the payoffs. Alice can’t see the options, but she must choose one based on Bob’s advice (which she is free to ignore). Bob then chooses one of two sentences to tell Alice: “Option A will pay you more than Option B, ” or “Option B will pay you more than Option A. ” Alice then chooses either Option A or Option B, and the money is paid.

Honesty in the Laboratory Before making her choice, Alice is informed that one option is better for Bob, and the other is better for her. Gneezy (2005) found that 83% of Bobs told the truth. Moreover, 78% of Alices believed their Bob, even though they knew that Bob had an incentive to lie (provided Bob believed that Alices would likely follow his advice---which he did indeed believe).

Honesty in the Laboratory People make trade-offs among values and between values and material rewards. Morality is not categorical. If the options were A: (5, 500) and B: (500, 5), many more Bobs might lie. In Gneezy’s experiment with options A: (5, 15) and B: (6, 5), more Bobs told the truth (lying is very costly to Alice). In Gneezy’s experiment with options A: (5, 15) and B: (15, 5), more Bobs lied (the cost of honestly was very high).

Honesty in the Laboratory Many such experiments show that people value virtue for its own sake. People are generally not selfish, even when they can behave selfishly with complete impunity. Of course, some fraction of people are completely selfish (usually about 20% in behavioral game theory experiments). If society does not find a way to punish the free-riders, cooperation will very often unravel, as we shall see.

Parochial Altruism – conferring benefits on others at a cost to oneself and parochialism --favoring ethnic, racial or other insiders over outsiders – are commonly observed human behaviors that are well documented in experiments. In fact, they are closely interrelated in the evolution of Homo sapiens.

Parochial Altruism Parochialism and altruism are both puzzling from an evolutionary perspective because they reduce the individual’s personal payoffs by comparison to others who behave purely selfishly.

Parochial Altruism We consider an explanation: parochial altruistic behavior contributes to success in between-group competition. However, neither altruism nor parochialism alone is likely to evolve.

Parochial Altruism While within-group altruism and parochialism could not have evolved singly, they could have co-evolved, each providing the exceptional conditions allowing for the evolutionary success of the other. This is why other-regarding preferences are often conditional on group membership, and may involve negative as well as positive sentiments toward the wellbeing of others.

The Big Questions Was war sufficiently common and lethal enough to allow the proliferation of an altruistic trait? Can we explain the co-evolution of such an altruistic trait along with hostility towards ‘outsiders’?

Is There a Hamilton’s Rule for Groups? The probability that in a single generation a group engages in a contest for survival (κ) The degree of genetic diversity between groups that come into conflict (r) The contribution of altruistic behavior to the probability that a group will win a conflict (b) The fitness cost of the altruistic behavior (c) From all of the above, we can derive a “Hamilton’s Rule” for groups (approximate): 2κbr>c

Ethnographic (black dots) and archaeological (black squares) evidence on mortality and (white dots) genetic differentiation

Table 1 Fraction (δ) deaths due to warfare (or other violence) among hunter-gatherers: archaeological evidence Site, source Approx date (yrs before present) Citation (complete citations in (20)) δ British Columbia 5500 BP-334 Cybulski (1994) 0. 23 Nubia (site 117) 14 -12000 (23) Wendoff (1968) 0. 46 Nubia (near site 117) 14 -12000 Wendoff (1968) 0. 03 Ukraine (Vasylivka) 11000 Vencl (1991) 0. 18 Ukraine (Volos’ke) Epipalaeolithic Danilenko (1955) 0. 17 S. California 5500 -628 Lambert (1997) 0. 04 Central California 3500 -508 Moratto (1984) 0. 05 Sweden (Skateholm I) 6300 Price (1985) 0. 07 Central California (SCL 674) 2415 -1773 Andrushko et al (2005) 0. 08 India (Uttar Pradesh) 3140 -2854 Clark (1977) 0. 27 Central California (SCI-038) 2240 -238 Jurmain (2001) 0. 04 Central California (Ala-329) 1500 – 238 Jurmain (2001) 0. 04 Niger (Gobero) 16, 000– 8200 Sereno et al. (2008) 0. 00 Algeria (Calumnata) 8300 -7300 Chamla et al (1970) 0. 04 France (Ile Teviec) 6600 Newall et al (1979) 0. 12 Denmark (Bogebakken) 6300 -5800 Newall et al (1979) 0. 12

Table 2. Fraction of adult deaths due to warfare (δ) among hunter gatherers: ethnographic evidence Population, region Dates Livelihood and society Citation (complete citations in (20)) δ Ache, Eastern Paraguay pre-contact (1970) foragers Hill and Hurtado (1996) 0. 37 Hiwi, Venezuela. Colombia pre- contact (1960) foragers Hill et al. (2007) 0. 17 Murngin NE Australia: 1910 -1930 forager including maritime Warner (1931) 0. 21 Ayoreo Bolivia. Paraguay 1920 -1979 seasonal foragerhorticulturalists Bugos (1985) 0. 15 Tiwi N. Australia 1893 -1903 sedentary hunter gatherers Pilling (1968) 0. 10 Modoc N. California 1934 sources re ‘aboriginal times’ sedentary hunter-gatherer Ray (1963) 0. 13 Anbara N. Australia 1940 -1960 Recently settled nomadic maritime foragers Hiatt (1965) 0. 04 Yuki N. California before 1850 sedentary hunter-gatherer Kroeber (1953) 0. 27 Kato N. California before 1850 sedentary hunter-gatherer Kroeber (1953) 0. 01 Yurok N. California 1830 -1840 sedentary hunter-gatherer (storage) Kroeber (1953) 0. 04

Table 3. Critical cost ( c*) of an altruistic behavior for given estimates of genetic differentiation (F) and mortality in intergroup hostilities (δ) among Australian foragers. Murngin δ = 0. 207 Tiwi δ = 0. 100 Anbara δ = 0. 045 Aboriginal Australian groups F = 0. 042 0. 136 (0. 073) 0. 066 (0. 035) 0. 030 (0. 016) Kaiadilt-Lardiil F = 0. 081 0. 209 (0. 146) 0. 101 (0. 704) 0. 046 (0. 032) If c = 0. 03 in the absence or group conflict, the behavior would go from 90% to 10% of a population in just 150 generations Inter-demic genetic differentiation is measured by the FDT. Mortality in intergroup hostility (δ) is from Table 2. b = 2. Group size is 26 which could be a coalition of 3 groups of the size considered to be typical of foragers during the late Pleistocene Entries in parentheses are large group size t statistics. 2 κ b r = c*

Consider a population in which Parochial Tolerant individuals may be either Altruistic or Not and either Altruist PA TA Tolerant or Parochial towards Not NP NT other groups. A’s contribute to the fitness of other group members at a cost to themselves Only the Parochial Altruists fight wars. P’s induce hostilities and forgo the benefits of peaceful interactions with other groups enjoyed by the T’s

Four behavioral types; two selection processes (multi-level selection) Within group selection: a) N’s payoffs exceed A’s b) in absence of war, T’s payoffs exceed P’s A N Inter-groups interactions: (a) hostile conflict or (b) trade, insurance, exploiting buffer zones P PA NP T TA NT Between group selection: Group with fewer PA’s lose conflicts, a fraction of their members are replaced by a draw the from the winner group Within group selection favors N over A and in the absence of hostile inter-group conflict favors T over N; between group conflict favors PA over other behaviors.

The model parameter values (per generation, where relevant) • • • # of groups = 20 Group effective size =26 (i. e. census size about 70 = 3 bands) Mutation =0. 005 Two loci, two alleles at each locus, full recombination Between group island (random) migration =0. 25 Benefits and costs: b=0. 02, c=0. 01, baseline fitness=1 Benefit from peaceful interaction: g=0. 001 Between group interactions per generation: k=1 Fitness loss of losing group (non fighters) = 2. 5 c Fighters’ mortality in warfare = 0. 14

An empirically estimated distribution The height of the bars gives the fraction of a very long period in which we observe the indicated pair of population-level frequencies of altruists and parochials in the population.

Conclusion Under conditions approximating those experienced by our Late Pleistocene ancestors, groups of parochial altruists could emerge, and such groups would frequently engage in and win hostile conflicts with other groups. The model indicates that neither parochialism nor altruism is viable singly but that warfare, altruism and parochialism could have evolved jointly. Cooperation against others may be the distinctive capacity of our ancestors that explains their success in the great exodus from Africa around 60 thousand years ago.

Conclusion There is no evidence of a genetic basis for war, altruism or parochialism in humans. What we show is that if genes supporting these behaviors were to exist they could have evolved. Parochial altruism may be our (genetic and cultural) legacy, but it need not be our fate. As an especially cultural species, we may be able to invent, redirect, and overcome, outsider hostility.