Payoff-based learning best explains the rate of decline in cooperation across 237 public-goods games

The effect of intergroup competition outcome on ingroup cooperation: insights from the male warrior hypothesis

Introduction

The Male Warrior Hypothesis (MWH) proposes that sex-specific selective pressures have promoted male cooperation with the ingroup members to outcompete rival groups. However, intergroup conflicts do not occur in isolation and the outcomes of previous competitions may influence group cooperativeness. Since this phenomenon is not well understood, we aimed to shed light on the effect of previous competition outcome on later cooperative behavior under intergroup conflicts. Based on the MWH, we hypothesized that repeated contests between groups could enhance ingroup cooperation, regardless of the outcome of the previous contest because status is at risk, but when competition is not present, participants would move to the symmetric equilibria.

Methods

To test this hypothesis, we recruited 246 individuals organized in groups of 6 and measured cooperation using a threshold public good game over two rounds, manipulating the outcome in the first round to create groups of winners and losers.

Results

Our results show that intergroup conflict scenarios promoted cooperation in both victory and defeat conditions, whereas, in the control scenario only losers increased their cooperation.

Discussion

We argue that winners under the presence of an external threat may enhance in-group cooperation in order to assure their status; whereas, losers may be attempting to regain it.

Resolving selfish and spiteful interdependent conflict

Interdependence occurs when individuals have a stake in the success or failure of others, such that the outcomes experienced by one individual also generate costs or benefits for others. Discussion on this topic has typically focused on positive interdependence (where gains for one individual result in gains for another) and on the consequences for cooperation. However, interdependence can also be negative (where gains for one individual result in losses for another), which can spark conflict. In this article, we explain when negative interdependence is likely to arise and, crucially, the role played by (mis)perception in shaping an individual's understanding of their interdependent relationships. We argue that, owing to the difficulty in accurately perceiving interdependence with others, individuals might often be mistaken about the stake they hold in each other's outcomes, which can spark needless, resolvable forms of conflict. We then discuss when and how reducing misperceptions can help to resolve such conflicts. We argue that a key mechanism for resolving interdependent conflict, along with better sources of exogenous information, is to reduce reliance on heuristics such as stereotypes when assessing the nature of our interdependent relationships.

Confusion cannot explain cooperative behavior in public goods games

Some scholars find that behavioral variation in the public goods game is explained by variations in participants' understanding of how to maximize payoff and that confusion leads to cooperation. Their findings lead them to question the common assumption in behavioral economics experiments that choices reflect motivations. We conduct two experiments, in which we minimize confusion by providing participants with increased training. We also introduce a question that specifically assesses participants' understanding of payoff maximization choices. Our experimental results show that the distribution of behavior types is significantly different when participants play with computers versus humans. A significant increase in contributions is also observed when participants play with humans compared to when they play with computers. Moreover, social norms may be the main motive for contributions when playing with computers. Our findings suggest that social preferences, rather than confusion, play a crucial role in determining contributions in public goods games when playing with humans. We therefore argue that the assumption in behavioral economics experiments that choices reveal motivations is indeed valid.

Perspectives on adaptive dynamical systems

Adaptivity is a dynamical feature that is omnipresent in nature, socio-economics, and technology. For example, adaptive couplings appear in various real-world systems, such as the power grid, social, and neural networks, and they form the backbone of closed-loop control strategies and machine learning algorithms. In this article, we provide an interdisciplinary perspective on adaptive systems. We reflect on the notion and terminology of adaptivity in different disciplines and discuss which role adaptivity plays for various fields. We highlight common open challenges and give perspectives on future research directions, looking to inspire interdisciplinary approaches.

Random migration with tie retention promotes cooperation in the prisoner's dilemma game

Migration has the potential to induce outbreaks of cooperation, yet little is known about random migration. Does random migration really inhibit cooperation as often as previously thought? Besides, prior literature has often ignored the stickiness of social ties when designing migration protocols and assumed that players always immediately disconnect from their ex-neighbors once they migrate. However, this is not always true. Here, we propose a model where players can still retain some bonds with their ex-partners after they move from one place to another. The results show that maintaining a certain number of social ties, regardless of prosocial, exploitative, or punitive, can nevertheless facilitate cooperation even if migration occurs in a totally random fashion. Notably, it reflects that tie retention can help random migration, previously thought to be harmful to cooperation, restore the ability to spark bursts of cooperation. The maximum number of retained ex-neighbors plays an important role in facilitating cooperation. We analyze the impact of social diversity in terms of the maximum number of retained ex-neighbors and migration probability, and find that the former enhances cooperation while the latter often engenders an optimal dependence between cooperation and migration. Our results instantiate a scenario in which random migration yields the outbreak of cooperation and highlight the importance of social stickiness.

The restart effect in social dilemmas shows humans are self-interested not altruistic

Do economic games show evidence of altruistic or self-interested motivations in humans? A huge body of empirical work has found contrasting results. While many participants routinely make costly decisions that benefit strangers, consistent with the hypothesis that humans exhibit a biologically novel form of altruism (or "prosociality"), many participants also typically learn to pay fewer costs with experience, consistent with self-interested individuals adapting to an unfamiliar environment. Key to resolving this debate is explaining the famous "restart effect," a puzzling enigma whereby failing cooperation in public goods games can be briefly rescued by a surprise restart. Here we replicate this canonical result, often taken as evidence of uniquely human altruism, and show that it 1) disappears when cooperation is invisible, meaning individuals can no longer affect the behavior of their groupmates, consistent with strategically motivated, self-interested, cooperation; and 2) still occurs even when individuals are knowingly grouped with computer players programmed to replicate human decisions, consistent with confusion. These results show that the restart effect can be explained by a mixture of self-interest and irrational beliefs about the game's payoffs, and not altruism. Consequently, our results suggest that public goods games have often been measuring self-interested but confused behaviors and reject the idea that conventional theories of evolution cannot explain the results of economic games.

Ancestral social environments plus nonlinear benefits can explain cooperation in human societies

Human cooperation (paying a cost to benefit others) is puzzling from a Darwinian perspective, particularly in groups with strangers who cannot repay nor are family members. The beneficial effects of cooperation typically increase nonlinearly with the number of cooperators, e.g., increasing returns when cooperation is low and diminishing returns when cooperation is high. Such nonlinearity can allow cooperation between strangers to persist evolutionarily if a large enough proportion of the population are already cooperators. However, if a lone cooperator faces a conflict between the group's and its own interests (a social dilemma), that raises the question of how cooperation arose in the first place. We use a mathematically tractable evolutionary model to formalise a chronological narrative that has previously only been investigated verbally: given that ancient humans interacted mostly with family members (genetic homophily), cooperation evolved first by kin selection, and then persisted in situations with nonlinear benefits as homophily declined or even if interactions with strangers became the norm. The model also predicts the coexistence of cooperators and defectors observed in the human population (polymorphism), and may explain why cooperators in behavioural experiments prefer to condition their contribution on the contributions of others (conditional cooperation in public goods games).

The Black Box as a Control for Payoff-Based Learning in Economic Games

The black box method was developed as an "asocial control" to allow for payoff-based learning while eliminating social responses in repeated public goods games. Players are told they must decide how many virtual coins they want to input into a virtual black box that will provide uncertain returns. However, in truth, they are playing with each other in a repeated social game. By "black boxing" the game's social aspects and payoff structure, the method creates a population of self-interested but ignorant or confused individuals that must learn the game's payoffs. This low-information environment, stripped of social concerns, provides an alternative, empirically derived null hypothesis for testing social behaviours, as opposed to the theoretical predictions of rational self-interested agents (Homo economicus). However, a potential problem is that participants can unwittingly affect the learning of other participants. Here, we test a solution to this problem in a range of public goods games by making participants interact, unknowingly, with simulated players ("computerised black box"). We find no significant differences in rates of learning between the original and the computerised black box, therefore either method can be used to investigate learning in games. These results, along with the fact that simulated agents can be programmed to behave in different ways, mean that the computerised black box has great potential for complementing studies of how individuals and groups learn under different environments in social dilemmas.

Self-interested learning is more important than fair-minded conditional cooperation in public-goods games

Why does human cooperation often unravel in economic experiments despite a promising start? Previous studies have interpreted the decline as the reaction of disappointed altruists retaliating in response to non-altruists (Conditional Cooperators hypothesis). This interpretation has been considered evidence of a uniquely human form of cooperation, motivated by an altruistic concern for equality ('fairness') and requiring special evolutionary explanations. However, experiments have typically shown individuals not only information about the decisions of their groupmates (social information) but also information about their own payoffs. Showing both confounds explanations based on conditional cooperation with explanations based on confused individuals learning how to better play the game (Confused Learners hypothesis). Here we experimentally decouple these two forms of information, and thus these two hypotheses, in a repeated public-goods game. Analysing 616 Swiss university participants, we find that payoff information leads to a greater decline, supporting the Confused Learners hypothesis. In contrast, social information has a small or negligible effect, contradicting the Conditional Cooperators hypothesis. We also find widespread evidence of both confusion and selfish motives, suggesting that human cooperation is maybe not so unique after all.

Evolutionary instability of selfish learning in repeated games

Across many domains of interaction, both natural and artificial, individuals use past experience to shape future behaviors. The results of such learning processes depend on what individuals wish to maximize. A natural objective is one's own success. However, when two such "selfish" learners interact with each other, the outcome can be detrimental to both, especially when there are conflicts of interest. Here, we explore how a learner can align incentives with a selfish opponent. Moreover, we consider the dynamics that arise when learning rules themselves are subject to evolutionary pressure. By combining extensive simulations and analytical techniques, we demonstrate that selfish learning is unstable in most classical two-player repeated games. If evolution operates on the level of long-run payoffs, selection instead favors learning rules that incorporate social (other-regarding) preferences. To further corroborate these results, we analyze data from a repeated prisoner's dilemma experiment. We find that selfish learning is insufficient to explain human behavior when there is a trade-off between payoff maximization and fairness.

Network Characteristic Control of Social Dilemmas in a Public Good Game: Numerical Simulation of Agent-Based Nonlinear Dynamics

This paper proposes a possible mechanism for obtaining sizeable behavioral structures by simulating a network–agent dynamic on an evolutionary public good game with available social .learning. The model considers a population with a fixed number of players. In each round, the chosen players may contribute part of their value to a common pool. Then, each player may imitate the strategy of another player based on relative payoffs (whoever has the lower payoff adopts the strategy of the other player) and change his or her strategy using different exploratory variables. Relative payoffs are subject to incentives, including participation costs, but may also be subject to mutation, whose rate is sensitized by the network characteristics (social ties). The process discussed in this report is interesting and relevant across a broad range of disciplines that use game theory, including cultural evolutionary dynamics.

Religious celibacy brings inclusive fitness benefits

The influence of inclusive fitness interests on the evolution of human institutions remains unclear. Religious celibacy constitutes an especially puzzling institution, often deemed maladaptive. Here, we present sociodemographic data from an agropastoralist Buddhist population in western China, where parents sometimes sent a son to the monastery. We find that men with a monk brother father more children, and grandparents with a monk son have more grandchildren, suggesting that the practice is adaptive. We develop a model of celibacy to elucidate the inclusive fitness costs and benefits associated with this behaviour. We show that a minority of sons being celibate can be favoured if this increases their brothers' reproductive success, but only if the decision is under parental, rather than individual, control. These conditions apply to monks in our study site. Inclusive fitness considerations appear to play a key role in shaping parental preferences to adopt this cultural practice.

Four levers of reciprocity across human societies: concepts, analysis and predictions

This paper surveys five human societal types - mobile foragers, horticulturalists, pre-state agriculturalists, state-based agriculturalists and liberal democracies - from the perspective of three core social problems faced by interacting individuals: coordination problems, social dilemmas and contest problems. We characterise the occurrence of these problems in the different societal types and enquire into the main force keeping societies together given the prevalence of these. To address this, we consider the social problems in light of the theory of repeated games, and delineate the role of intertemporal incentives in sustaining cooperative behaviour through the reciprocity principle. We analyse the population, economic and political structural features of the five societal types, and show that intertemporal incentives have been adapted to the changes in scope and scale of the core social problems as societies have grown in size. In all societies, reciprocity mechanisms appear to solve the social problems by enabling lifetime direct benefits to individuals for cooperation. Our analysis leads us to predict that as societies increase in complexity, they need more of the following four features to enable the scalability and adaptability of the reciprocity principle: nested grouping, decentralised enforcement and local information, centralised enforcement and coercive power, and formal rules.

Bounded rational agents playing a public goods game

An agent-based model for human behavior in the well-known public goods game (PGG) is developed making use of bounded rationality, but without invoking mechanisms of learning. The underlying Markov decision process is driven by a path integral formulation of reward maximization. The parameters of the model can be related to human preferences accessible to measurement. Fitting simulated game trajectories to available experimental data, we demonstrate that our agents are capable of modeling human behavior in PGG quite well, including aspects of cooperation emerging from the game. We find that only two fitting parameters are relevant to account for the variations in playing behavior observed in 16 cities from all over the world. We thereby find that learning is not a necessary ingredient to account for empirical data.

A preference to learn from successful rather than common behaviours in human social dilemmas

Human cooperation is often claimed to be special and requiring explanations based on gene-culture coevolution favouring a desire to copy common social behaviours. If this is true, then individuals should be motivated to both observe and copy common social behaviours. Previous economic experiments, using the public goods game, have suggested individuals' desire to sacrifice for the common good and to copy common social behaviours. However, previous experiments have often not shown examples of success. Here we test, on 489 participants, whether individuals are more motivated to learn about, and more likely to copy, either common or successful behaviours. Using the same social dilemma and standard instructions, we find that individuals were primarily motivated to learn from successful rather than common behaviours. Consequently, social learning disfavoured costly cooperation, even when individuals could observe a stable, pro-social level of cooperation. Our results call into question explanations for human cooperation based on cultural evolution and/or a desire to conform with common social behaviours. Instead, our results indicate that participants were motivated by personal gain, but initially confused, despite receiving standard instructions. When individuals could learn from success, they learned to cooperate less, suggesting that human cooperation is maybe not so special after all.

Sharing Money With Humans Versus Computers: On the Role of Honesty-Humility and (Non-)Social Preferences

Does giving behavior in economic games reflect true prosocial preferences or is it due to confusion? Research showing that trait Honesty-Humility accounts for giving behavior suggests the former, whereas research showing that participants give money to a computer might suggest the latter. In three preregistered, well-powered studies, we examined the relation of Honesty-Humility with behavior in the Dictator Game (Study 1, N = 468) and Public Goods Game (Studies 2 and 3, each N = 313), while participants interacted either with humans (“social game”) or with a computer (“non-social game”). We found that (a) decisions in the non-social game predicted decisions in the social game, supporting the confusion hypothesis; (b) the effect of Honesty-Humility differed within and between games; and (b) participants who gave money to the computer reported acting as if they were playing with humans. Overall, the studies suggest that both prosocial preferences and confusion underlie giving behavior.

Social Learning Strategies and Cooperative Behaviour: Evidence of Payoff Bias, but Not Prestige or Conformity, in a Social Dilemma Game

Human cooperation, occurring without reciprocation and between unrelated individuals in large populations, represents an evolutionary puzzle. One potential explanation is that cooperative behaviour may be transmitted between individuals via social learning. Using an online social dilemma experiment, we find evidence that participants’ contributions were more consistent with payoff-biased transmission than prestige-biased transmission or conformity. We also found some evidence for lower cooperation (i) when exposed to social information about peer cooperation levels than without such information, and (ii) in the prisoners’ dilemma game compared to the snowdrift game. A simulation model established that the observed cooperation was more likely to be caused by participants’ general propensity to cooperate than by the effect of social learning strategies employed within the experiment, but that this cooperative propensity could be reduced through selection. Overall, our results support previous experimental evidence indicating the role of payoff-biased transmission in explaining cooperative behaviour, but we find that this effect was small and was overwhelmed by participants’ general propensity for cooperation.

Decoupling cooperation and punishment in humans shows that punishment is not an altruistic trait

Economic experiments have suggested that cooperative humans will altruistically match local levels of cooperation (conditional cooperation) and pay to punish non-cooperators (altruistic punishment). Evolutionary models have suggested that if altruists punish non-altruists this could favour the evolution of costly helping behaviours (cooperation) among strangers. An often-key requirement is that helping behaviours and punishing behaviours form one single conjoined trait (strong reciprocity). Previous economics experiments have provided support for the hypothesis that punishment and cooperation form one conjoined, altruistically motivated, trait. However, such a conjoined trait may be evolutionarily unstable, and previous experiments have confounded a fear of being punished with being surrounded by cooperators, two factors that could favour cooperation. Here, we experimentally decouple the fear of punishment from a cooperative environment and allow cooperation and punishment behaviour to freely separate (420 participants). We show, that if a minority of individuals is made immune to punishment, they (i) learn to stop cooperating on average despite being surrounded by high levels of cooperation, contradicting the idea of conditional cooperation and (ii) often continue to punish, 'hypocritically', showing that cooperation and punishment do not form one, altruistically motivated, linked trait.