One of the great challenges for successful consilience in the social sciences is knowing when and where methodological individualism – exemplified by characterizing subpersonal psychological mechanisms (i.e., cognitive adaptations) – starts and stops. Even if you knew every synapse, fiber, and Rube Goldberg convolution of functional specialization housed within individual skulls, you still would not understand the complex dynamics that play out when individual minds start interacting.
Complicating matters, a lot of computational machinery is precisely for social dynamics, baked into our minds courtesy of natural selection. Because social life emerges from interacting minds possessing adaptive machinery for life in groups, it is possible to overestimate the role of the evolved psychology while simultaneously underestimating the complexity of that evolved psychology.
In the target article, Glowacki points to group norms (e.g., punishing individuals who raid out-groups) and group decisions (e.g., paying restitution to victimized out-groups) as factors that foster peace. But noting such factors is different from understanding how they arise. To do that, we contend, requires (i) unpacking the complexity of the underlying psychological mechanisms involved, including describing what is being represented and acted upon within individual minds, and (ii) unpacking the complexity of the social dynamics created by these individual minds in interaction with one another. In turn, we suggest that “group” decisions and norms are not themselves psychological mechanisms, but rather strange and fragile outcomes, arising from individual psychologies having already vied and interacted with one another.
We also seek to avoid analyses that may fall prey to “instinct blindness.” The psychological mechanisms that parse the world into social groups operate so effortlessly that we might fail to recognize that they exist at all (Cosmides & Tooby, Reference Cosmides and Tooby1994). Even something as seemingly simple as perceiving individuals as members of in-groups and out-groups poses formidable computational challenges. These are in part solved by mechanisms that represent simpler three-person interactions – including (a) who you are willing to harm after you have harmed me, (b) who is willing to harm you after you have harmed me, (c) who is willing to harm me after you have harmed me, and (d) who I am willing to harm after you have harmed me – all of which are then used to infer (or better still, project) group membership onto individual agents (Pietraszewski, Reference Pietraszewski2022).
Acknowledging the complex dynamics underwritten by these and other representational systems may help us to better understand why restitution encourages peace after intergroup raids. In dyadic interaction, there is no mystery: The offender pays a cost now to advertise that they will be paid back these short-term costs by restoring a productive relationship and avoiding (potentially much steeper) retaliatory costs (McCullough, Kurzban, & Tabak, Reference McCullough, Kurzban and Tabak2013). But this reasoning does not scale up to the intergroup level because not all group members have the same incentives. For instance, suppose I did not participate in the raid: Why should I be willing to concede that the raid was “our fault?” And if I am in the victimized group, why should I expect that restitution from those who did not participate in the raid is even relevant?
The reason, we suggest, has to do with concerns about the future. To the extent that raiders in pursuit of spoils kill randomly, and that raids motivated by revenge against specific people can easily draw in the targets' family members and allies, each person in group V (“V” for victim) is right to assume that they are as likely to be killed in a future raid as anyone else. They are therefore better off representing that the attack that just killed “some people I sometimes fish with” was actually an attack against all of us; our “group.” And this is why the narrative becomes “Group A attacked us” (“A” for attacker) rather than “some random people who live over there attacked a bunch of people over here.”
We can also consider things from the perspective of individuals in group A: If they want to predict the future, they will need to represent how group V is framing the raid. Because members of group V are unlikely to know which members of group A actually conducted the raid, they will hold all of the people in group A culpable (for the reasons above). Consequently, group A's members can safely infer that members of group V will view the members of group A as interchangeable.
Even so, some in group A have an incentive to resist this framing. Raiders might be inclined to accept the group framing in order to dilute their individual culpability, while nonraiders might be inclined to reject it. Likewise, those who would be called upon to make restitution (in the form of money, weapons, cows, daughters, or sons) are more likely to resist the group framing than those who lack anything to make restitution with. The individuals within these subgroups can come into conflict as they deliberate about whether to provide restitution: Their costs and benefits differ. So, they in turn must be able to represent additional group or alliance dynamics within their group as they try to convince others about what they should do. Eventually, the subgroups of individuals within group A who resist the “We raided you” framing must be either convinced to change their minds or coerced into going with the flow by those who support it.
The point, then, is that a norm to maintain positive ties or make restitution, or a group “decision” to do so, is a social dynamic above and beyond the individual psychologies; it is already the outcome of individual psychologies vying for their particular interests.
One final note: We (along with the individuals who make up our hypothetical groups A and V) have indulged the idea that groups A and V objectively exist. But they do not. Groups do not exist in the same way that rocks exist; they are “useful delusions” that help agents coordinate action in the service of cooperation and conflict (Tooby & Cosmides, Reference Tooby, Cosmides and Høgh-Olesen2010, p. 206). Group identities are promissory notes, built atop hopes of future cooperation and fears of future conflict.
You have
Access
One of the great challenges for successful consilience in the social sciences is knowing when and where methodological individualism – exemplified by characterizing subpersonal psychological mechanisms (i.e., cognitive adaptations) – starts and stops. Even if you knew every synapse, fiber, and Rube Goldberg convolution of functional specialization housed within individual skulls, you still would not understand the complex dynamics that play out when individual minds start interacting.
Complicating matters, a lot of computational machinery is precisely for social dynamics, baked into our minds courtesy of natural selection. Because social life emerges from interacting minds possessing adaptive machinery for life in groups, it is possible to overestimate the role of the evolved psychology while simultaneously underestimating the complexity of that evolved psychology.
In the target article, Glowacki points to group norms (e.g., punishing individuals who raid out-groups) and group decisions (e.g., paying restitution to victimized out-groups) as factors that foster peace. But noting such factors is different from understanding how they arise. To do that, we contend, requires (i) unpacking the complexity of the underlying psychological mechanisms involved, including describing what is being represented and acted upon within individual minds, and (ii) unpacking the complexity of the social dynamics created by these individual minds in interaction with one another. In turn, we suggest that “group” decisions and norms are not themselves psychological mechanisms, but rather strange and fragile outcomes, arising from individual psychologies having already vied and interacted with one another.
We also seek to avoid analyses that may fall prey to “instinct blindness.” The psychological mechanisms that parse the world into social groups operate so effortlessly that we might fail to recognize that they exist at all (Cosmides & Tooby, Reference Cosmides and Tooby1994). Even something as seemingly simple as perceiving individuals as members of in-groups and out-groups poses formidable computational challenges. These are in part solved by mechanisms that represent simpler three-person interactions – including (a) who you are willing to harm after you have harmed me, (b) who is willing to harm you after you have harmed me, (c) who is willing to harm me after you have harmed me, and (d) who I am willing to harm after you have harmed me – all of which are then used to infer (or better still, project) group membership onto individual agents (Pietraszewski, Reference Pietraszewski2022).
Acknowledging the complex dynamics underwritten by these and other representational systems may help us to better understand why restitution encourages peace after intergroup raids. In dyadic interaction, there is no mystery: The offender pays a cost now to advertise that they will be paid back these short-term costs by restoring a productive relationship and avoiding (potentially much steeper) retaliatory costs (McCullough, Kurzban, & Tabak, Reference McCullough, Kurzban and Tabak2013). But this reasoning does not scale up to the intergroup level because not all group members have the same incentives. For instance, suppose I did not participate in the raid: Why should I be willing to concede that the raid was “our fault?” And if I am in the victimized group, why should I expect that restitution from those who did not participate in the raid is even relevant?
The reason, we suggest, has to do with concerns about the future. To the extent that raiders in pursuit of spoils kill randomly, and that raids motivated by revenge against specific people can easily draw in the targets' family members and allies, each person in group V (“V” for victim) is right to assume that they are as likely to be killed in a future raid as anyone else. They are therefore better off representing that the attack that just killed “some people I sometimes fish with” was actually an attack against all of us; our “group.” And this is why the narrative becomes “Group A attacked us” (“A” for attacker) rather than “some random people who live over there attacked a bunch of people over here.”
We can also consider things from the perspective of individuals in group A: If they want to predict the future, they will need to represent how group V is framing the raid. Because members of group V are unlikely to know which members of group A actually conducted the raid, they will hold all of the people in group A culpable (for the reasons above). Consequently, group A's members can safely infer that members of group V will view the members of group A as interchangeable.
Even so, some in group A have an incentive to resist this framing. Raiders might be inclined to accept the group framing in order to dilute their individual culpability, while nonraiders might be inclined to reject it. Likewise, those who would be called upon to make restitution (in the form of money, weapons, cows, daughters, or sons) are more likely to resist the group framing than those who lack anything to make restitution with. The individuals within these subgroups can come into conflict as they deliberate about whether to provide restitution: Their costs and benefits differ. So, they in turn must be able to represent additional group or alliance dynamics within their group as they try to convince others about what they should do. Eventually, the subgroups of individuals within group A who resist the “We raided you” framing must be either convinced to change their minds or coerced into going with the flow by those who support it.
The point, then, is that a norm to maintain positive ties or make restitution, or a group “decision” to do so, is a social dynamic above and beyond the individual psychologies; it is already the outcome of individual psychologies vying for their particular interests.
One final note: We (along with the individuals who make up our hypothetical groups A and V) have indulged the idea that groups A and V objectively exist. But they do not. Groups do not exist in the same way that rocks exist; they are “useful delusions” that help agents coordinate action in the service of cooperation and conflict (Tooby & Cosmides, Reference Tooby, Cosmides and Høgh-Olesen2010, p. 206). Group identities are promissory notes, built atop hopes of future cooperation and fears of future conflict.
Financial support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Competing interest
None.