Researchers have utilized mathematical modeling to discover a method of preserving cooperation among individuals without needing intricate norms or institutions.
The concept of indirect reciprocity suggests that individuals who build a positive reputation through helping others are likely to receive rewards from outside parties. However, for sustained cooperation to exist, there needs to be a shared understanding of what reputations mean. Traditional theoretical models focus on reputations as either good or bad and are based on limited information. In today’s society, particularly with the rise of social media, there is an abundance of behavioral data available.
Joshua B. Plotkin from the University of Pennsylvania and Corina Tarnita from Princeton University are biology professors leading collaborative theoretical research on cooperation. Sebastián Michel-Mata, a doctoral student in Tarnita’s lab, proposed examining how individuals are evaluated in environments rich with information.
“Current indirect reciprocity theory implies that reputations are only effective in a few societies that have sophisticated judgment norms and public systems to enforce consensus,” states Michel-Mata. However, as an anthropologist, he argues that such societies are exceptions, prompting him to consider the straightforward notion that reputations reflect numerous actions.
“Most previous models have assumed that a single action defines a person’s reputation, but I believe there’s more complexity in how we judge people. We frequently assess multiple actions to determine if they are largely positive or negative,” remarks Mari Kawakatsu, a postdoc in Plotkin’s lab.
Through their mathematical modeling, the research team concluded that evaluating multiple actions and allowing for some forgiveness of negative actions is an effective way to judge behavior, a strategy they term “look twice, forgive once.” Their research is published in Nature.
This research builds upon earlier work by Plotkin regarding indirect reciprocity, where he, along with Kawakatsu and fellow researcher Taylor A. Kessinger, examined the amount of gossip necessary to achieve enough consensus for sustained cooperation.
Plotkin emphasizes that “even when individuals within a society have differing judgment norms, ‘look twice, forgive once’ fosters enough consensus to encourage cooperation.” He further notes that this strategy upholds cooperation without the need for gossip or public institutions, supporting Michel-Mata’s initial hypothesis that such institutions aren’t essential for cooperation based on reputation. This insight provides a viable alternative in scenarios where trust in public institutions is waning, thus hindering cooperation.
Kessinger mentions that, akin to their study on gossip, this research utilizes a one-time donation game model, which resembles a simplified prisoner’s dilemma. In this setup, each player decides whether to help a partner, frequently updating their perceptions of each other’s reputations by observing how they interact with other players, determining if their partner cooperates or defects with others. Players also periodically revise their strategies.
The notion of indirect reciprocity doesn’t imply that kindness is reciprocated directly; rather, it’s about being kind to someone because they were kind to another whom you respect. Kessinger states, “In this research, if you witness two interactions from an individual and at least one is something you deem positive, you will cooperate with that player; otherwise, you won’t.”
Kawakatsu expressed that the co-authors were all surprised that the “look twice, forgive once” strategy couldn’t be overruled by other strategies, such as consistently cooperating or defecting, considering more than two actions from another player, or varying the amount of forgiveness for negative actions. Tarnita added that, interestingly, looking at more than two actions did not provide a significant advantage. “Information turned out to be a double-edged sword; even with ample information available, individuals typically didn’t evolve to utilize all of it,” she comments.
Michel-Mata highlights that the overall simplicity and durability of their outcomes suggest that this behavioral strategy may have deep roots in human society. The authors see an opportunity for anthropologists and behavioral scientists to build upon their findings.
The labs of Plotkin and Tarnita are continuing their collaborative efforts by investigating how individuals interact across multiple contexts, like in professional and personal environments. Kessinger states, “This addresses a variety of modern social issues, particularly where personal misconduct becomes publicly recognized.”
