Context-dependent behavior can make cooperation flourish
A generous and caring person at home can be ruthless at work, striving to make the most sales or move up a chain of business management. Along the same lines, an egocentric neighbor can be a model of altruism on Twitter.
This is a widespread characteristic of human society: people can adopt different behaviors depending on the social context in which they find themselves. Yet according to a new study by Penn biologists published today in Scientists progressthat context-dependent behavior tends to favor the diffusion of cooperative behavior throughout a society.
Using models rooted in game theory, the researchers show that cooperation is particularly favored when there is room for “spillovers” between domains. In other words, a worker can observe how his colleague behaves with his friends when deciding how to interact with this person and others in the workplace.
“We’ve studied small and large groups,” says Joshua Plotkin, a professor in Penn’s Department of Biology and lead author of the new paper, “and we find that the simple idea of conditioning behavior to social context, while allowing imitation of behaviors in different contexts – this alone facilitates cooperation in all areas simultaneously.”
This work, and a related study in Nature Human behaviorsuggests that the greater the number of domains of social life, the higher the likelihood that cooperative interactions will eventually dominate.
“It shows that the structure of interactions in different aspects of our social lives can galvanize each other – to the benefit of mutual cooperation,” Plotkin said.
Since Charles Darwin, scientists have wondered about the enigma of cooperation. It is clear that cooperation is essential to human society, but from an evolutionary perspective, it is difficult to explain why people would give up something to help others. Plotkin and his colleagues modeled and explored this question from many angles, considering how the structure of social interaction networks, the presence of memory and reputation, and the capacity for empathy, among other characteristics, influence the likelihood that cooperation flourishes in a group.
In these new papers, led by first author Qi Su, a Simons postdoctoral fellow at Penn, the researchers introduced another nuance to their analysis of strategic interactions: a multilevel society, in which actors interact in distinct social domains. In their model, the actors can choose different strategies to adopt in the different areas – perhaps being selfish in one and cooperative in the other.
“For example, I interact with co-workers at work, I interact with my family and friends outside of work, I interact with people online and people offline,” says Plotkin. “Each of these areas may have an internal structure – I may be closer to some people at work than others – but the strategies I employ in my interactions at work may differ from the interactions in my personal life.”
In Nature Human behavior, analysis by Plotkin and his team showed that when these interactions take place in a pattern where actors in a given context can only imitate the strategies of other players in that context, cooperation can thrive in a domain, but selfish strategies prevail in another. Overall, however, the likelihood of cooperation dominating in a given domain increases as the number of social domains increases.
“Let’s say there’s a cooperator in layer one, but the same person selfishly takes advantage of others in layer two and receives a lot of excessive rewards from their behavior in layer two,” Plotkin explains. “Overall, he appears to be a successful individual, and so layer one individuals might tend to copy his behavior in layer one – so the cooperation then propagates into layer one, following of interactions occurring in layer 2. Sometimes this dynamic facilitates cooperation in one domain at the expense of cooperation in another, depending on the structure of the network in each domain.
The domains tended to act synergistically when Plotkin and his colleagues added additional functionality, explored in the Scientists progress paper. In the model presented here, actors could observe what strategies others were using in the other layers, allowing them to copy strategies from one sphere of interaction to use in a different sphere.
“Here the results are more striking,” says Plotkin, “because multiple domains with overflow tend to facilitate cooperation in all domains simultaneously, even though cooperation would never propagate in a single domain.”
For example, in a population of six individuals interacting in a single domain, approximately half of all possible interaction networks promote the evolution of malicious behavior – individuals paying to harm others. By taking the same networks of six individuals and separating the pairwise interactions into four different social contexts, on the other hand, cooperation is then favored to evolve for all network structures.
“A similar thing happens in much larger groups,” Plotkin says. “When you have multiple domains, cooperation will tend to predominate, even if the benefit-cost ratio is low.”
Su adds, “Context-dependent behavior can help understand why human societies are often very cooperative, even though they are densely connected.”
In addition to their modeling, the researchers examined empirical evidence from real-world interaction networks, which supported the idea that multilevel social domains would lead to greater cooperation through “coupling” – when the strategy that an individual employs in one domain influences the strategy employed in another.
Most models of cooperation assume two-way interactions – one person chooses where to act altruistically or not in relation to another, and vice versa. In a third related article published last month in the Proceedings of the National Academy of SciencesPlotkin and his colleagues considered one-way interactions that are common in human society: pure altruism, when one individual helps another without the possibility of receiving anything directly in return.
By including one-way interactions in their model, removing the possibility of direct reciprocity between pairs, the researchers found that cooperation was still favored in most scenarios, a “really shocking” result, says Plotkin. What drives this, he notes, is a kind of “third-party reciprocity,” where trios of actors form, each giving in one direction, but all benefiting from it. Thus, the dominant trend is towards the cooperation of the actors.
The researchers note that many real-world interactions are one-way: In the pandemic, volunteer first responders put themselves at risk to help others; supervisors have the ability to support subordinates who cannot be reciprocated; you can follow another person on Twitter without that person following them back. In future work, the Penn researchers hope to explore how to possibly intervene in a directed network to promote cooperation.
“Careful moderation of directed interactions can help facilitate more prosocial behavior,” says Plotkin. “I’m thinking of the algorithms by which Twitter suggests new people to follow, for example. These networks could be put together in a way that fosters cooperation.”
Joshua Plotkin is the Walter H. and Leonore C. Annenberg Professor of Natural Sciences at the School of Arts & Sciences at the University of Pennsylvania.
co-authors of Plotkin on Scientists progress paper were Qi Su and Alex McAvoy of Penn’s Center for Mathematical Biology. Plotkin, Su and Benjamin Allen of Emmanuel College wrote the Proceedings of the National Academy of Sciences paper. Penn’s Su, McAvoy, and Yoichiro Mori co-wrote the Nature Human behavior paper.
Support for these studies came from the National Science Foundation (grants 1907583 and 2042144), the Simons Foundation, the David and Lucile Packard Foundation, and the John Templeton Foundation.