This blog post examines how selfish choices and altruistic behavior evolve under certain conditions through the Prisoner’s Dilemma, and why cooperation becomes a rational strategy in repeated interactions.
The question of whether selfish or altruistic behavior is more advantageous for survival has long been a subject of academic and philosophical debate. The concept of the “Prisoner’s Dilemma” provides one clue to answering this question. The “Prisoner’s Dilemma” describes a situation where two accomplices, despite the best possible outcome of reduced sentences due to insufficient evidence if they cooperate to conceal the crime, succumb to an investigator’s temptation to reduce their sentence if they reveal the other’s crime. Consequently, they end up incriminating each other and receiving heavier sentences. Here, let us set aside the moral question of whether the two prisoners should cooperate. From the prisoners’ perspective, choosing to reduce each other’s sentences is an altruistic act. However, in this experiment, prisoners tend to assume their counterpart will make a selfish choice, leading them to ultimately choose selfish behavior themselves. This mutual distrust results in greater losses for both individuals.
If the two prisoners were never to meet again after serving their sentences, betraying each other would likely pose no significant problem. However, if their relationship persists and the same situation recurs—that is, if the prisoner’s dilemma repeats—choosing to cooperate rather than betray each other yields greater long-term benefits. In other words, people tend to behave more cooperatively toward each other when relationships within a group are enduring rather than temporary. Ultimately, whether selfish or altruistic behavior is more advantageous depends on the persistence of the relationship.
The hypothesis explaining this characteristic is the “Repetition-Reciprocity Hypothesis.” This hypothesis starts from the perspective that if there is absolutely no chance of encountering a particular counterpart again, there is no reason to extend favors to them, as there is no risk of suffering losses or retaliation later. However, in real-world societies, interacting individuals are not randomly assembled; people generally encounter the same limited group repeatedly. In such situations, if encounters persist and the premise holds that when I show favor, the other person will reciprocate with favor, altruistic behavior naturally emerges.
In Korean rural society, the practice of “mutual aid labor exchange” has existed since ancient times. This is a form of labor exchange based on mutual support: one goes to another’s home to help with farm work or personal tasks, and later receives help from others when needed. While this may appear superficially like a labor exchange without compensation, it was actually not a waste of labor but an institution ensuring mutual benefit in the long term. Because relationships among members in rural communities are highly persistent rather than temporary, mutual aid serves as a prime example illustrating the repetition-reciprocity hypothesis.
Another example illustrating that altruistic behavior—that is, mutually cooperative actions—is more advantageous when group relationships are close and enduring is Rousseau’s “deer hunting game.” For hunters to successfully hunt a deer, each must faithfully guard their assigned path. However, if a rabbit passes by during this process, the hunter who sees it may be tempted to abandon their assigned position and chase the rabbit. In this situation, hunters must choose one of two strategies: either remain fully committed to the shared goal of hunting the deer, or chase the rabbit for immediate gain.
If the deer hunt succeeds, each participant gains 10. Assuming the rabbit hunt yields 5, it is rational for me to guard my assigned position if I believe the other party will faithfully pursue the deer hunt. Conversely, if the other party is likely to chase the rabbit, chasing the rabbit myself becomes the more advantageous choice. The problem is that guarding my assigned path to the end offers no guarantee that the deer hunt will succeed. If I cooperate in the deer hunt, the possible outcomes are 10 or 0. If the other party cooperates, success is achieved, but if they defect, I gain nothing. In contrast, abandoning the deer hunt to chase rabbits yields a fixed gain of 5 regardless of the other party’s choice. This makes chasing rabbits a relatively “less risky” option.
Nevertheless, because the value of the deer is greater than that of the rabbit, the more this game is repeated and the longer the relationship between the hunters persists—that is, the longer the interactions continue—the greater the likelihood that they will cooperate. Thus, the Repeated-Interaction Hypothesis explains how altruistic behavior can become a rational strategy within a sustained relationship, rather than a one-off choice.
This hypothesis applies to everyday situations as well. For instance, at a regular restaurant, the owner strives to provide quality food anticipating the customer’s continued visits, and in return, the customer repeatedly patronizes the establishment. The repeat-reciprocity hypothesis operates in this process, where altruistic behavior ultimately creates a structure of mutual benefit.
Ultimately, the Repetition-Reciprocity Hypothesis suggests that our altruistic actions are not driven solely by pure sacrifice. Acts of kindness toward others carry the nature of a carefully calculated transaction; in the long term, such altruistic behavior becomes a choice that ultimately benefits oneself. We have examined the definition of the Repetition-Reciprocity Hypothesis and reviewed several cases that align with it. To summarize, the longer-lasting and more enduring the relationships within a group, the more choosing altruistic behavior over selfish behavior benefits not only the individual but the entire group. This hypothesis helps predict what actions people will choose when faced with recurring dilemma situations and provides crucial clues for understanding the motivations behind various behaviors in our society.
However, this hypothesis also has clear limitations. We do not necessarily extend kindness only to those we are likely to meet again. We often act altruistically even toward people who are not close acquaintances and whom we expect to never see again. For example, one of the author’s acquaintances once handed several thousand won to someone claiming they couldn’t return home due to lack of bus fare, without even questioning the truth of their statement. Furthermore, most people have likely donated money to charity drives for the underprivileged. These donations benefit people we have never seen or met. The fact that altruistic behavior occurs even in situations without repeated interaction clearly indicates that the repetition-reciprocity hypothesis alone cannot fully explain the motivation behind human altruism.
