Vengeance, retribution meted out by an individual or social group against someone who has harmed them, has long fascinated social scientists. From blood feuds in the Balkan mountains to honor killings in the US South to rido in the Philippines, ethnographers have thoroughly studied these extreme forms of retaliation and the environments in which they arise.

Revenge as a collective pattern of behavior is most commonly associated with particular types of social organization, such as the tribe or clan (with segmentary lineage systems, discussed by Sara in this post, being one type of a tribal society). It has also been tied to specific modes of subsistence, such as pastoralism, in which property is mobile and easy to expropriate or destroy. Psychologists Nisbett and Cohen traced the origins of the “culture of honor” of the US South to Scots-Irish herders. Pauline Grosjean provided systematic evidence for the connection between settlers of Scots-Irish origin and homicide rates in that region.

Correlated with both tribalism and the predominance of herding is the third commonly identified predictor of revenge: the absence of law enforcement and weak or absent formal institutions. Cultures of honor flourish in areas where the state has not penetrated. In a cross-cultural study of 192 societies, psychologists Karen Ericksen and Heather Horton found a strong correlation between the prevalence of blood revenge and the absence of institutionalized justice systems. In theoretical work using agent-based models, Nowak et al. (2016) showed that honor cultures are more likely to emerge when enforcement is ineffective or unreliable.

The reason why that is the case is not difficult to see. In the absence of external enforcement, punishment of wrongdoers can help deter future infractions and maintain social order. This is true regardless of whether revenge is exacted by the individual or the social group. In Robert Ellickson’s classic study of the norms governing dispute resolution among cattle ranchers in California’s rural Shasta county, individual cattlemen would inflict costly punishment on trespassers, for instance by driving trespassing cattle to a place inconvenient for the owner. In tribal societies, revenge is prescribed at the level of the kin group. The person who takes revenge need not be the individual slighted, and the victim of retribution need not be the wrongdoer themselves, but either side can be represented by a member of their kin group. When the revenge is not commensurate to the crime, counter-revenge may ensue; in societies where the unit of organization is not the individual but the group this can quickly escalate into vendettas and large-scale violence, which works as an additional mechanism of deterrence. Evidence supports this function of revenge. Sociologist Roger Gould studied court records in Corsica, a society known for its honor culture, and found acts of revenge to be rare. Their occurrence was consistent with revenge playing a strategic role in social order maintenance.

Global prevalence of negative reciprocity, partly capturing willingness to take revenge. Darker blue shade indicates higher values. From Falk et al. (2018).

Central in this view of revenge is the role of reputation as a mechanism for sustaining cooperation. Individuals or groups who punish those who slight them build a reputation for being tough and not to be antagonized. This deters future transgressors. Without institutional enforcement, it seems reasonable that reputation would take center stage as a solution to human cooperation problems and that a culture of honor would develop, as both strategic behavior and an internalized norm of honor in the long-run.

Yet this is not the only way in which reputation can facilitate cooperation in the absence of external enforcement. “Image scoring”, building a good reputation as someone who helps others, who “plays nice” and who can be trusted in individual interactions, can also theoretically support cooperative behavior in a group. Cooperators are then seen as reliable transaction partners and good community members, and “cheaters” are avoided or excluded. Consistent with such a view of reputation, Leung and Cohen (2011) extend the typology of cultures beyond that of honor to “cultures of face”, in which honor is derived not by avenging misconduct, but by behaving appropriately within the system in which one is embedded. A third cultural type, the culture of dignity, does not respond to offenses by revenge, but by exercising “covert avoidance, quietly cutting off relations with the offender without any confrontation” (Campbell and Manning, 2014).

In joint work with Alain Schläpfer, we show in a game-theoretic setup how both punishment-based and image scoring-based reputation can support social cooperation. People interact in situations that have the structure of a prisoner’s dilemma game – both sides would be better off cooperating, but individual incentives make defection the optimal choice for each player. We assume a large group, such that pairs rarely interact multiple times, but past actions follow people in the form of reputation that spreads in the group through gossip. We give people the option to retaliate against cheating, but also to avoid engaging in a transaction with a disreputable individual in the first place.

In such a setup, a culture of honor is one type of equilibrium that can emerge. In a society of revenge, people behave as if according to the following rule: punish cheaters, cheat on the “suckers”. Those who don’t punish are considered weak and can be exploited. Tough guys on the other hand induce cooperation through the threat of retaliation.

Another type of cooperative equilibrium is more “cooperative” in nature. The social prescription there is to cooperate unless met with someone known to cheat. In that case, avoidance, and not aggression, is the prescribed action. To ensure others will want to interact with them in the future, people have an incentive to play nice.

Alain and I usually think of the two equilibria in terms of our respective origin cultures. Greeks might go a long way to establish themselves as people who are not taken advantage of, even if they themselves may sometimes violate the rules (which is not necessarily frowned upon, if the violator is not caught). Swiss people instead tend to be more rule-following – and more attuned to potential rule violators as people best avoided (sometimes those individuals to be avoided could be Greek, or other immigrants, and one of our model’s applications is cross-cultural conflict). Neither of the two cultures is at the extremes of the continuum, but I hope you get the idea.

Retaliatory behavior in soccer. We examine all players in Europe’s big 5 soccer league between 2017 and 2020 and compute the likelihood that a player who was fouled will retaliate with another foul against the person who fouled them in the same game. Greek players tend to retaliate frequently. Swiss players less so – but many major Swiss players are second-generation immigrants from Balkan countries.

The insight of a setup like this – which is not entirely new, but merely a synthesis of two strands of literature on reputation and cooperation that have developed in parallel – is that it adds a new dimension to the relation of vengeance and weakly institutionalized environments. First, it shows that cooperation can be sustained without retaliation in the absence of enforcement, making weak institutions one of many requirements for cultures of honor to arise. Second, and most importantly, it shows how a culture of honor can be a driver rather than a consequence of weak institutions.

In our setup, we show this by first establishing that the introduction of external enforcement is a better complement of reputational systems reliant on image scoring, than it is of honor cultures. In the latter, any enforcement punishes both the victim and the avenger, crowding out the informal means society used to maintain order. In the short-run, enforcement may even prove counterproductive and increase rates of defection. Cooperation-avoidance cultures instead are compatible with external enforcement. Punishment of defectors only increases the incentives of “good” players to play cooperatively.

We then consider what happens when societies can choose their own institutions. Because of the way formal justice interacts with informal reputation mechanisms, cultures of retaliation are likely to introduce external enforcement under a narrower set of situations. A culture of honor ends up hindering the development of institutionalized order.

Revenge was once prevalent across most human societies. Ericksen and Horton find kin group vengeance to occur in 90% of the societies they study. Over time, the practice of vengeance has become increasingly rare and there is no doubt that the penetration of state power has had a lot to do with this development (see evidence for this in Pauline Grosjean’s piece cited above, or Jensen and Ramey (2020) on state capacity and dueling in the US). But beyond this broad pattern, there is considerable variation in the prevalence of retaliation as an element of culture, from its most extreme forms like blood feuds to its more subtle versions whereby people ensure their actions will establish them as tough actors not to be meddled with. One reason for this variation may be the variation in the ways in which societies have historically solved their problems in the absence of institutional enforcement. Pre-existing cultural equilibria then hinder the adoption of the rules that would eradicate them, contributing to their persistence.


  • Vicky Fouka

    Vicky Fouka is an assistant professor of Political Science at Stanford University. Her research is at the intersection of history, political economy and political behavior, with a focus on group identity, intergroup relations and immigrant integration. Her work is published or forthcoming in the American Political Science Review, Review of Economic Studies, the Journal of Politics and Economic Journal. She received a PhD in Economics from Pompeu Fabra University in 2015.

Leave a Reply