Twenty years ago Rwanda was torn apart by violence. The Hutu majority slaughtered their Tutsi neighbors, killing approximately 70% of the Tutsi minority in the space of only four months. Once the killing finally stopped, a difficult question arose: how to right these monstrous wrongs without creating a cycle of revenge and retribution?

Such a cycle would be the epitome of the ancient “eye for an eye” notion of justice, in which punishment is commensurate with the crime, an approach taken even today by most modern legal systems (including the United States). The aim is simple—when someone is wronged, the goal of punishment is to hurt the perpetrator. Decades of research have examined this preference for punishment, demonstrating that people have strong proclivities to punish transgressions. While punishment serves other functions like deterrence, there is no doubt that retribution is a central goal. In fact, we as individuals, punish even when it costs us to do so: one study found that if treated sufficiently unfairly, some people are willing to forgo up to three month’s salary to punish the perpetrator.

Although it seems that people like to punish, there’s a critical limitation to this research. Punishment is typically the only option made available for righting a wrong. It is either punish, or accept the transgression. It’s possible, however, that people prefer to restore justice without punishment, focusing instead on the needs of the victim. Because researchers typically don’t offer non-punitive options, we didn’t know—until now—how punishment stacked up against other ways of righting wrongs. A recent set of studies from our lab has found people may strongly prefer non-punitive options when restoring justice.

In our studies, instead of punishing, participants overwhelmingly (9 out of 10 times) preferred to compensate the victim (in this case, themselves) after being unfairly given a disproportionately small split of money. While that’s not particularly surprising (who doesn’t like money?), participants also felt that this was enough to right the wrong: they generally didn’t punish the transgressor—even when punishment was free and easy to do. This is straightforward evidence that sometimes, for some crimes, victims do not want to punish the perpetrator if there are other paths to justice. There is evidence for this in the real world as well: programs that prioritize victims’ needs and foster dialogue between victims and perpetrators show the highest rates of victim satisfaction and offender accountability.

Victims, however, typically don’t decide the fate of their perpetrators: judges and juries do. A core tenet of our own legal system is that victims are partial and thus should not be deciding the fate of their perpetrators. In contrast, third parties like judges and juries, are considered to be impartial and can more objectively and dispassionately mete out justice. Given these assumed asymmetries between how victims and third parties approach justice, we examined whether an individual’s perspective matters when deciding how to restore justice. Would third parties (juries) sanction social transgressions differently than those who had been personally affected (victims)?

To address this, participants in a series of follow-up studies were asked to make decisions on behalf of other people. Effectively, we asked participants to act like juries—doling out punishment to the perpetrator or compensation to the victim despite having no “skin in the game.” Unlike victims, third parties chose the most retributive option in our task, where the victim is compensated and the transgressor is simultaneously punished. Put simply, although participants rarely punished transgressors after being treated unfairly, when they saw someone else being wronged, they chose the harshest form of punishment, implementing the classic ‘eye for an eye’ form of retributive justice.

This finding sheds a new light on how people choose to rebalance the scales of justice. When we ourselves have been slighted, we appear to tend to our own needs rather than pursue punishment, but this changes when we make decisions on behalf of someone else: for bystanders or jurors, an eye-for-an-eye may be preferable. Our notion of justice seems to depend on where we stand. This leaves us with a challenge: there may be a gap between what we as victims want, and what third parties decide for us, calling into question our blind reliance on the putative impartiality of juries and judges.

Though more research is needed, this work offers a glimmer of hope for our overburdened and underfunded justice system. In fact, an emerging literature has begun to shift the emphasis away from punishment towards less punitive ways of restoring justice. A few years ago a group at Harvard demonstrated that having the opportunity to reward others in a cooperative game actually boosts cooperation more than punishment does alone. And a paper published earlier this year reveals that restricting people’s ability to restore justice to punishment alone only leads to an increased desire to punish. Essentially, the motivation to punish appears to be a function of the landscape of options it is presented with. Altogether, these studies suggest that punishment—while certainly desirable in some instances—should not always be considered the gold standard of justice restoration.

Although they are less frequently employed, reconciliatory non-punitive approaches towards restoring justice have proven to function well in the real world. Perhaps most surprisingly, these approaches can be successful for even the most egregious crimes against humanity. A national effort currently being undertaken in Rwanda is bringing together Hutus and Tutsis in order to reconcile one of the most horrific genocides in history. After being counseled over many months, a Hutu perpetrator asks for forgiveness from a Tutsi survivor of his crime. Wielding an impressive ability to forgive—even for murder—the Tutsi survivors become reconciled with their Hutu perpetrators. With an aim to heal ties between neighbors, this initiative stresses restoration and forgiveness, not punishment. In doing so, it has the potential to curb the hate and violence that fuels a cycle of revenge and retribution.