Evolution of human cooperation

Human beings cooperate on a scale and in ways that have no parallel in the rest of the animal kingdom. Ant colonies and termite mounds represent sophisticated cooperation, but their members are close genetic relatives, and the cooperation is rigidly encoded in caste-specific behaviours. Among non-human primates, cooperation is limited to small groups of individuals linked by kinship or reciprocal exchange. Humans, by contrast, cooperate with vast numbers of unrelated strangers: they fight wars in armies of tens of thousands, build cities of millions, maintain international trade networks spanning continents, and sustain institutions — from religions to legal systems to scientific communities — that coordinate behaviour across generations.^{4, 5} Understanding how and why this extraordinary capacity for cooperation evolved is one of the central problems in the study of human evolution.

The puzzle of cooperation among unrelated individuals

The evolution of cooperation poses a theoretical challenge because natural selection typically favours individuals who maximise their own reproductive fitness. Cooperation is costly to the cooperator and beneficial to others, creating an incentive for free-riders who enjoy the benefits of others' cooperation without paying the costs. William Hamilton's theory of kin selection explains cooperation among genetic relatives: an individual can increase the transmission of its genes by helping relatives who share copies of those genes, a principle formalised as Hamilton's rule (the cost to the altruist must be less than the benefit to the recipient multiplied by their coefficient of genetic relatedness).⁸ Robert Trivers' theory of reciprocal altruism extends cooperation to non-relatives: in repeated interactions, individuals can benefit from helping others provided the favour is returned, as demonstrated by the success of tit-for-tat strategies in Robert Axelrod's computer tournaments of the iterated prisoner's dilemma.^{8, 9}

But neither kin selection nor reciprocal altruism can account for the scale of human cooperation. Humans routinely cooperate with strangers they will never meet again, contribute to public goods from which they could free-ride, and incur costs to punish norm violators even when punishment provides no personal benefit. These behaviours are not easily explained by genetic relatedness (the cooperators are often unrelated) or by reciprocity (the interactions are often one-shot or anonymous).^{6, 10} Evolutionary psychologists have proposed that domain-specific cognitive adaptations, such as cheater-detection mechanisms and coalitional psychology, evolved to manage the computational demands of navigating complex cooperative environments.¹⁸

Shared intentionality: the cognitive foundation

Michael Tomasello has argued that the cognitive capacity underlying uniquely human cooperation is shared intentionality — the ability to participate with others in collaborative activities with shared goals and coordinated roles. Human children, beginning around nine months of age, engage in triadic interactions in which they share attention to objects with adults, pointing and following gaze in ways that great ape infants do not. By fourteen months, children participate in cooperative activities with shared goals, understanding both their own role and their partner's role in joint actions. This capacity for joint intentionality, Tomasello argues, is the foundation on which all distinctively human forms of cooperation are built.^{2, 3}

In A Natural History of Human Morality (2016) and Becoming Human (2019), Tomasello proposed a two-stage evolutionary model. In the first stage, driven by the ecological demands of collaborative foraging (particularly cooperative hunting and gathering), early humans evolved the capacity for joint intentionality: two individuals collaborating toward a shared goal with mutual awareness of each other's roles and expectations. This stage produced a "second-personal morality" of fairness and mutual respect between partners in joint activities. In the second stage, driven by intergroup competition and the demands of large-scale social coordination, humans evolved the capacity for collective intentionality: the ability to create and enforce group-wide conventions, norms, and institutions. This stage produced "group-minded morality" — loyalty to one's group, conformity to shared norms, and the capacity for cultural life.^{1, 17}

Cultural group selection

Peter Richerson, Robert Boyd, and Joseph Henrich have developed the most influential evolutionary framework for explaining large-scale human cooperation: cultural group selection. Their theory rests on two key insights. First, humans possess uniquely powerful capacities for cultural learning — particularly imitation, teaching, and conformist transmission (the tendency to adopt behaviours that are common in one's group). These capacities allow cultural traits, including cooperative norms, to spread and persist within groups far more rapidly than genetic traits.^{4, 5}

Second, cultural variation between groups creates the conditions for group selection to operate. Groups with more cooperative norms — norms that promote sharing, coordination, and collective defence — tend to outcompete groups with less cooperative norms in intergroup competition, whether through warfare, resource competition, or differential migration. As more cooperative groups expand and less cooperative groups shrink or are absorbed, cooperative norms spread through the population. Crucially, this process does not require genetic differences between groups: the variation on which selection acts is cultural, transmitted through learning rather than genes.^{4, 7}

Henrich has argued that this process created a gene-culture coevolutionary feedback loop. As cultural norms favouring cooperation became prevalent, they altered the selective environment faced by individuals. Individuals who were psychologically disposed to learn, follow, and enforce cooperative norms were more successful within cooperative groups, selecting for genetic traits that supported norm-following and prosocial behaviour. Over many generations, this feedback loop produced a species that is innately inclined toward conformity, norm enforcement, and cooperation with in-group members — a process Henrich calls "self-domestication."^{5, 16}

Punishment and norm enforcement

A critical mechanism sustaining human cooperation is altruistic punishment — the willingness of individuals to bear personal costs to punish those who violate cooperative norms, even when the punisher receives no direct benefit. Boyd, Gintis, Bowles, and Richerson demonstrated in theoretical models that punishment can stabilise cooperation in large groups where reciprocity alone would fail, because the threat of punishment raises the cost of free-riding above the cost of cooperating.^{6, 7}

Experimental evidence from cross-cultural studies supports the importance of punishment. Henrich and colleagues conducted economic experiments (such as the ultimatum game, the dictator game, and the public goods game) across fifteen diverse small-scale societies on five continents. They found that the willingness to punish unfair behaviour varied substantially across societies but was present in every society studied, and that societies with greater market integration and participation in world religions displayed stronger prosocial and punishment behaviours, suggesting that cultural institutions amplify the human disposition to enforce norms.^{12, 15}

Martin Nowak and Karl Sigmund have emphasised the role of indirect reciprocity — cooperation sustained by reputation rather than direct exchange. In groups where individuals' reputations are tracked through gossip and observation, cooperators gain reputational benefits that make others more willing to cooperate with them, while defectors suffer reputational costs. Language, Nowak and Sigmund argue, may have evolved in part to facilitate the gossip and reputation tracking that sustain indirect reciprocity.¹⁴

Archaeological evidence for cooperative behaviour

The archaeological record provides evidence for increasingly sophisticated cooperation throughout hominin evolution. The oldest known stone tools, dating to approximately 3.3 million years ago, imply coordinated action and possibly social learning, as the skills required to produce effective flakes from stone cores are difficult to acquire without instruction or imitation.⁵ By the Middle Pleistocene, approximately 400,000 years ago, evidence from sites such as Qesem Cave in Israel suggests cooperative large-game hunting and systematic meat-sharing, activities that require coordination among multiple hunters and some system for dividing the spoils equitably.¹³

The control of fire, well established by at least 400,000 years ago and possibly considerably earlier, represents a form of cooperation in itself: maintaining a fire requires collective vigilance, fuel gathering, and shared use of the hearth for cooking, warmth, and light. Richard Wrangham has argued that cooking, by softening food and increasing caloric returns, altered the selective environment in ways that favoured longer periods of social interaction around the hearth, deepening social bonds and enabling more complex forms of cooperation.⁵

By the Upper Palaeolithic (approximately 50,000–12,000 years ago), the archaeological record reveals long-distance exchange networks spanning hundreds of kilometres, the production of symbolic artefacts such as cave art and personal ornaments that signal group identity, and the construction of dwellings and storage facilities that imply collective planning and labour. These developments suggest that the full apparatus of cultural group selection — group identity, symbolic marking, norm enforcement, and institutional coordination — was in place by this period, enabling the cooperative feats that would culminate in the agricultural revolution and the rise of urban civilisations.^{4, 5}

Parochial altruism and the dark side of cooperation

The evolution of cooperation is inseparable from the evolution of intergroup conflict. Jung-Kyoo Choi and Samuel Bowles proposed the concept of parochial altruism — the combination of in-group cooperation and out-group hostility — and demonstrated in evolutionary models that these two traits coevolve: groups in which members are willing to sacrifice for their group and fight against other groups outcompete groups lacking these dispositions.¹¹ This finding suggests that human cooperativeness evolved not as a universal benevolence but as a group-bounded trait: a disposition to cooperate intensely with those inside the group boundary and to compete, sometimes violently, with those outside it.^{4, 11}

The expansion of cooperative circles beyond small bands to encompass nations, religions, and eventually the entire species is, on this account, a cultural achievement rather than a biological inevitability. Henrich has argued that large-scale institutions — markets, legal systems, universal religions, and international organisations — represent cultural technologies that extend the scope of cooperation beyond the bounds of parochial altruism, enabling cooperation among millions of unrelated strangers by creating shared identities, enforceable contracts, and reputational systems that substitute for the face-to-face monitoring of small groups.^{5, 15, 16} The story of human cooperation is thus not only a story of evolved psychology but of cultural innovation: the invention of institutions that harness, extend, and redirect the cooperative instincts that evolved in small-scale societies to operate in a world of cities, states, and global networks.^{4, 5}