Project 10

Cultural Norms and Cooperation

Norms, third-party enforcement, and the evolution of human cooperation
People using public transit in a large city

Principal Investigators

ASU Collaborators

External Collaborators

Project 10 focuses on theoretical research that explores how and why extensive cooperation in humans is a predictable outcome of the evolution of our unique “moral machinery.”

This project will help explain how singular features of human behavior, such as mutual aid, trade, and division of labor are linked to more general mechanisms sustaining uniquely complex forms of cooperation in human societies on larger scales.

Detailed Project Description

Humans cooperate much more extensively than do other vertebrates. We cooperate both at the small-scale of kin and close associates and, uniquely, in large groups comprising unrelated strangers. This pattern of increased cooperation at every scale is not satisfactorily explained by existing theories. It is generally accepted that large-scale cooperation is not feasible in other vertebrate societies because they lack language and culture. However, the mechanisms that are thought to maintain small-scale cooperation—kin selection and direct reciprocity—should work equally well in many other vertebrates. Yet, most vertebrates do not seize opportunities for significant gains from pairwise exchange, even with kin or close associates. For example, division of labor and delayed exchange of valuable commodities is virtually absent in nonhuman vertebrates; sick or injured individuals are on their own. In contrast, trade is omnipresent in human societies, and the sick and injured are fed and cared for. While a number of researchers have sought to explain the increased levels of small-scale cooperation in humans, there is no consensus.

Here we propose a program of theoretical work that could potentially answer this question. Our hypothesis is that our evolved moral machinery is crucial for sustaining both large-scale cooperation and small-scale cooperation in humans. This is a novel idea, and there is compelling reason to think it is a promising line of enquiry. Empirically, small-scale cooperation in contemporary human societies is very commonly regulated by community-enforced norms—a surprising pattern if small-scale cooperation rests on nepotism and pairwise reciprocity. Interactions between mates, between parents and offspring, sharing partners, and friends are observed and judged by others to be right or wrong, and their behavior reinforced through positive or negative reactions by those around them. We presented this idea to a large group of researchers in the field at an Ernst Strungmann conference, and it was extremely well received. However, the existing theoretical work does not show clearly why moral machinery would make it easier for small-scale cooperation to evolve.

We will conduct theoretical work aimed at explaining how the moral machinery allows for the evolution of a higher degree of small-scale cooperation than can evolve through the action of kin selection and reciprocity. In particular, we will undertake four theoretical projects that investigate a different avenue by which third-party judgment and norms may increase the range of conditions under which small-scale cooperation can take place.

1. Norms and third-party judgment may resolve problems resulting from perception errors. Errors play a crucial role in the evolution of reciprocity. Errors occur when an individual’s behavior deviates from that specified by his or her inherited strategy (for instance, an individual might defect when it should cooperate). Theorists distinguish between “execution errors,” in which all actors know that the error occurred, and “perception errors,” in which there is disagreement about whether an error has occurred. Execution errors create little problem for the evolution of cooperation. However, perception errors cause serious problems for the evolution of reciprocity. For direct reciprocity, the strategy labeled “Pavlov” is evolutionarily stable, but cannot easily increase when rare. For indirect reciprocity, no one has devised a robust strategy for dealing with perception errors.

We believe that it is possible that adjudication by third parties can solve this problem because it aligns the beliefs of interactants. In pilot work, we have studied a model of the evolution of pairwise direct reciprocity in which the behavior of reciprocating partners is observed by a third party, the “adjudicator.” When a perception error occurs, the individual disadvantaged by the error announces that he/she thinks an error has occurred and the adjudicator then makes a public judgment about whether he/she thinks an error actually occurred. We have shown that a reciprocating strategy that judges standing based on the adjudicators decision is evolutionarily stable, and can increase when rare even when the adjudicator is only moderately fair or accurate.

We will extend this model so that it applies to indirect reciprocity. We will also model the behavior of the adjudicator. In foraging bands there is much discussion about the behavior of others. We will model this by assuming that interdependence causes all individuals in the band to have an interest in reducing conflict, but each band member has a small number of relatives who they favor. Adjudication arises from discussion and consensus among members.

2. Norms and third-party judgment may resolve information asymmetries that can inhibit the evolution of cooperation. In many situations, needy individuals must communicate their state of need in order to receive help, and this fact restricts the conditions under which cooperation can evolve, even among kin. For example, the benefit from aiding a sick or injured individual may often be much greater than the cost of providing such need. However, the party that would benefit from a transfer is better informed about its state of need than possible donors, and must signal when he/she is in need. This creates the possibility of false signals. For example, it may be difficult for a mother to determine whether an older offspring is really ill, or just malingering because any signal used by a sick offspring to transmit information about its need can also be sent by a healthy offspring who just wants special treatment.

Existing modeling work suggests that this problem is difficult to resolve when pairs of individuals interact. This is because the cost necessary to ensure that signals are honest is often higher than the benefits of the transfer and so the basins of attraction of the signaling equilibria are often very small.

Shared social norms, and judgment from the community may help resolve this problem. Existing work indicates that repeated interactions allow honest low cost signals in pairwise interactions. However, models in which liars are punished require that their deception be exposed. Community monitoring could greatly enhance the efficiency of this process. For example, without community monitoring, a juvenile who signals to his mother that he is sick only needs to provide the right cues in her presence. He can go out and play when she is not looking. With community monitoring the cost of malingering will be much greater—he can’t play when anyone is looking. We will investigate this conjecture by creating models of pairwise signaling in which signals are observed by third parties, who then condition their own help to an individual on whether that individual has been detected lying to others.

3. Norms can help identify how to get the benefits of cooperation in the local ecology. It is possible that different patterns of cooperation are favored in different environments and culturally transmitted norms are needed to adapt to local conditions. Most models do not explain how actors identify when and what to exchange with whom in order to reap the benefits from the exchange. However, the ability to adapt to different environments is limited to what can be acquired through individual learning and experiments indicate that strategies supporting contingent cooperation are difficult to learn. We will model this situation by embedding models of indirect reciprocity in a spatially variable environment in which widely available cues allow cooperation, but the utilization of different cues in different localities allows more accurate assessment of the behavior of others. We will then investigate how the relative rates of adaptation and migration between localities affect the levels of cooperation favored by natural selection.

4. Norms and third-party judgment can eliminate inefficiencies caused by asymmetric interests among pairs of individuals. Norms can enforce pairwise exchange that may be beneficial only at a scale larger than that of the dyad. For instance, parents have a greater interest in cooperation between their children than do the children themselves. For example, Hamilton’s rule predicts that sisters will help each other when the benefits exceed twice the costs. However, parents want their offspring to help each other whenever the benefit exceeds the cost. Such conflicts of interest lead to a tug of war between parents and offspring, where the outcome depends on the mutations that are available to both sides. However, a shared norm that siblings should help each other could work better because everyone is better off living in a group that has a higher level of cooperation between-siblings. We will investigate whether parent enforcement of helping among siblings can increase level of sibling-sibling cooperation and whether this behavior can increase in frequency when rare and is evolutionarily stable.

Our hypothesis suggests that the psychology that governs norms and the sanctioning of norm violators may have arisen to support small-scale cooperation early in the evolution of the genus Homo, and explains the high levels cooperation within families that is needed to sustain human life-history attributes, such as prolonged maturation (which can be read from the fossil record through studies described in Project 8). This may explain how, as the rate of cultural evolution accelerated during the Pleistocene, cultural-group selection could have led to norm enforcement at the larger scales of cultural groups.