Social context and the evolution of empathy
– by Jeremy Garrison
We are a social species. Whether in the context of mammoth hunts, food sharing, child-rearing, war, or climate change, cooperation has always been instrumental to our individual and collective survival and propagation. Both the cooperation problems we face and our capacity to solve them are affected by our environments. Our capacity to accumulate and pass down socially learned behavior (culture) has allowed us to adapt to a wide range of ecosystems and organize into a diverse array of societal configurations. These different environments shape the genetic and cultural evolution of preferences through selection and their behavioral expression through context. For example, two people (with different genetic and cultural inheritances) may behave differently in any given context, and any given person may behave differently in different contexts.
One important way environments differ across societies is the social context of economic interactions. Do you know who made your clothes, your house, your tools, and your food? If you are a hunter-gatherer, you probably do. If you live in a market society, you probably don’t. We might call this the degree of social embeddedness or the average anonymity of the economic environment. How does this aspect of human ecology influence cooperative behavior?
In economic experiments where interacting subjects choose whether to act cooperatively or selfishly, highly social contexts promote cooperation. Cooperation increases when subjects are identified, when the interaction is face-to-face, when subjects can communicate, and when subjects are closer in real-world social networks. Additionally, in questionnaires of hypothetical choices, people are willing to forego more money to help those who are perceived to be socially closer to themselves. This relationship between social proximity and cooperation is also described in the psychology literature on empathy.
Empathy is a cognitive, affective, and/or behavioral response of one individual to the state of another. Empathic responses increase with familiarity, similarity, past experience, and salience. So, the strongest empathic responses occur between close friends or family and the weakest between out-group strangers. Cognitive and affective empathy are often associated with cooperative behavior. How might empathy have evolved?
In general, the evolution of cooperation is a puzzle since the actor is separated from the benefits of their action. If individuals interact randomly and always cooperate or never cooperate, the payoff of the non-cooperators is inevitably larger than the payoff of the cooperators, who pay the cost of cooperation. For cooperation then to evolve, the benefits of cooperation must be sufficiently concentrated among cooperators. This can occur via assortment, where cooperators interact disproportionately with other cooperators, or via discrimination, where cooperators can tell who the other cooperators are and cooperate only with them. These mechanisms can produce innate cooperative behavior through genetic evolution as well as socially learned cooperative behavior through cultural evolution. And they can manifest in a variety of proximate motives such as social emotions and moral norms.
To analyze the evolution of empathy, I developed a theoretical model consisting of an evolutionary game that takes place on a network with two different layers. Game interactions occur on the economic layer, and behavior transmission and empathy occur on the social layer. Previous literature shows that static networks can promote cooperation because interactors may have obtained their strategy from one another. But separating the transmission and interaction networks undermines the evolution of cooperation by reducing this assortment. So, unconditional cooperation relies on highly social contexts to evolve. But what if cooperation is conditioned on social proximity?
I model empathy as the game theoretic strategy of cooperating contingent on proximity in the social network. To explore the effects of social context on the evolution of empathy, I vary the overlap of the economic and social layers and analyze the evolutionary success of empathy. I employ both numerical methods—where I run many simulations and count the number of times the cooperators win out—and analytical methods—where I use a technique called pair approximation to derive a rule for the conditions under which empathy will evolve.
The numerical and analytical results align: unconditional cooperation relies on highly social economic environments and large gains from cooperation to evolve, but empathy can evolve even in asocial contexts as long as the gains from cooperation are sufficiently large. Specifically, I find that empathy evolves when (1/s)*b > c , where s is the number of social neighbors each player has and b and c are the benefits and costs of cooperation, respectively. What is the intuition for this rule (which so closely resembles Hamilton’s famous rule)? The degree of assortment in the social layer is inversely proportional to the number of social neighbors (the probability I learned my behavior from friend A is one divided by the number of friends I might have learned from). This then determines what proportion of the benefits of cooperation falls on other cooperators. If it’s greater than the cost of cooperation, empathy evolves.
In summary, cooperation can evolve when its benefits are concentrated among cooperators. If strategies spread among familiars by social learning or genetic transmission, then familiars tend to behave similarly. So, decreasing overlap of social and economic networks lessens assortment, impeding the evolution of unconditional cooperation. Conditioning cooperation on familiarity—which I call empathy—is adaptive in this context. Since familiarity is a proxy for behavioral relatedness, empathy concentrates the benefits of cooperation among empathizers, allowing this behavior to persist and spread. However, as anonymity increases, empathy results in less cooperation. Consequently, anonymous and impersonal economies may undermine cooperation, either by selecting for self-interested preferences or by rendering empathy nonfunctional.
In simpler terms, you probably act similarly to the people you know since we learn from the people around us. Cooperative people are therefore more likely than selfish people to be surrounded by other cooperative people. In this context, cooperating only with the people you know results in cooperative people reaping the benefits of cooperation and selfish people missing out. But if you use this rule when you don’t know anyone you interact with, you end up acting the same as a selfish person.
