R1. Introduction
Was life among our early human ancestors best characterized by harmonious intergroup relationships or was the threat of war ever present? I have argued that neither of these views is likely to capture the reality of intergroup social relationships during human evolution. Rather, I presented a framework in which individuals and concomitantly groups struggled to balance tensions between conflict and cooperation. It would have been individually advantageous for early humans to interact cooperatively with out-groups, especially once our reliance on cultural technologies increased. This is consistent with strong evidence of intergroup trade emerging around 300,000 years ago. But peace in humans is more than ad hoc or intermittent cooperation, it consists of stable harmonious interactions where conflicts are prevented and resolved.
While most individuals may be better off with peace, it only takes a small number of self-interested individuals to initiate a conflict that then involves the entire group. I argued that intergroup conflict would have been likely to occur at least periodically among our early human ancestors because some individuals may have benefited asymmetrically from lethal violence, such as through the theft of material goods, the taking of captives, and status benefits. Preventing individuals from initiating intergroup violence when they may benefit from it, and then resolving violence when it does occur requires the psychological capacity follow norms and the social mechanisms to enforce norms. I argued that enforcing norms that impose individual costs requires strong social institutions, including coercive mechanisms and leadership. Based on the available archaeological evidence, the social institutions required to prevent individuals from unilaterally using lethal violence against out-groups were unlikely to have been present before 100,000 years ago. Thus, cooperation with out-groups is very old, but peace as I have defined it, is likely to be much more recent, developing within the past 100 thousand years.
The 30 commentaries provided insightful and stimulating discussion that will surely advance the field. There was large agreement that human evolutionary history is not one solely of war and that cooperation, and possibly peace, predates agriculture extending much deeper into our history (>100 kya). In general, there was an appreciation that the evolutionary human sciences needs to move past the debate about whether war or peace characterizes our history and instead focus on the specific conditions that give rise to intergroup violence or cooperation and their adaptive consequences. I reply to the commentaries grouping them by theme. In section R2, I discuss commentaries that focus on the definition of peace; in section R3 I focus on the relationship between multilevel societies and peace; in section R4 I consider earlier origins for peace. Section R5 focuses on the distinction between internal and external war, while in section R6 I discuss commentaries that focus on how to study and enforce peace. In section R7, I consider the evolutionary psychology of peace; in section R8 I focus on extensions of the framework the target article develops. Finally, I conclude in section R9.
R2. Defining peace
Most scholarship on peace uses a negative conception of peace defining it as the absence of war. I argue that the lack of war does not capture what we mean when we use the term peace. An absence of war may be maintained through the threat of violence or avoidance, and lack the positive aspects of intergroup relationships that we generally infer with the term peace. The target article defines peace as a state where one can expect to harmoniously interact across social group boundaries and expect to do so in the future. Crucially, we expect that serious conflicts or violence is unusual and does not represent a group-held norm. In a state of peace, we can safely interact with, travel among, and even live with other groups all while maintaining membership in a distinct out-group. Once humans achieve peace, we do something extraordinary with it. We create interdependent systems where we depend on the resources or technologies that other groups provide for our survival. This intergroup interdependence is a key factor in enabling the success of our species.
Several commentators take issue with my definition of peace and by extension implications of my argument. Antony argues that a more useful definition of peace would include groups that have little interaction and interdependence with each other and use avoidance to prevent violence. Romano, Gross, & De Dreu (Romano et al.) argue that my restrictive definition of peace, in which stable positive-sum interactions are generally present, ignores situations in which interdependencies may be negative-sum, and in these cases it is much more difficult to reach peace. For them, a simpler and more useful definition of peace is simply the absence of conflict. This may or may not lead to cooperation, but is necessary for the establishment of cooperation. I agree with Romano et al. that a lack of conflict is likely required for cooperation (a point also made by Samuni, Wessling, & Surbeck [Samuni et al.] for bonobos). I recognize, as Antony points out, that groups sometimes avoid each other and in doing so lack conflict, which meets the criteria for a negative definition of peace. But peace is more than avoidance or even tolerance, and it is more than periodic cooperation scaled up. It is the ability to reliably interact with members of other groups with reasonable assurance of safety despite having membership in a distinct social group. In humans, this appears to require the ability to prevent and resolve conflicts. A state where two groups avoid each other, or lack overt conflict but do not cooperate appears substantively different than peace.
R3. Multilevel societies and other species with intergroup cooperation
Three commentaries (Grueter; Samuni et al.; Wilson) point toward the fact that some species, including humans, gelada monkeys, and hamadryas baboons, live in multilevel societies in which several social units coalesce creating higher-level forms of organization, such as several families forming a band or community, or members of a tribe belonging to different clans or territorial sections. The conditions that enable multilevel sociality in other species and our own may also facilitate peace, and a comparative perspective that includes them would be useful. Further, as Wilson points out, our prehuman ancestors may have lived in multilevel societies making the transition to cooperative relationships with out-groups easier. I agree that multilevel societies may provide important insights into the conditions which favor the evolution of cooperation between social groups. It is also plausible our prehuman ancestors lived in multilevel societies facilitating the development of peace but so little is known about the social organization of early Homo and earlier ancestors that inferences are tenuous. However, cooperation between subgroups of a multilevel society is less of a puzzle than peace, which occurs between socially distinct groups. This is because in a multilevel society members of different subgroups are still members of a single social group, just as a set of families in a community are all members of a community. This overlapping membership aligns incentives and provides more opportunities for monitoring and policing behavior than can occur between groups that otherwise have little contact. But I agree that consideration of multilevel societies can shed light on the factors by which peace emerges.
Several other commentators illustrate that other species may have aspects of peace or peace itself. Grüning & Krueger take issue with my claim that peace as I have defined it appears to be human-unique. Their reply is wary of anthropocentric arguments that take human behavior as unique and then seek to find the human characteristics enable it and I agree. Humans may not be alone in having peace, but peace on the positive definition is rare appearing in only a few other species. Grüning & Krueger also point toward evidence from chimpanzees showing social learning, and what they infer is norm compliance and enforcement. While I am sensitive to the myriad ways in which chimpanzees and many other species have complex social lives, including social learning, on my interpretation of the research from primate cognition and behavior, I do not see convincing evidence of norm compliance. Humans appear alone in being able to create and follow arbitrary and even self-harming norms.
Samuni et al. provide valuable insights into the behavior of wild bonobos including that while bonobos have intergroup aggression, the rates do not seem greater than within-group aggression. This is a fascinating insight suggesting reconsideration of what we consider as peaceful interactions and suggesting that bonobos have at least some aspects of peace, even on the restrictive definition I propose. They argue that unlike humans and chimpanzees, bonobos are able to avoid war through lacking high levels of proactive aggression that makes war likely in humans and chimpanzees. This, along with the commentary by Robinson, Rodrigues, & Barker (Robinson et al.) illustrate that there is more than one pathway to peace. Robinson et al. argue that some species of ants engage in long-term sharing of resources between nests. Because ants have an unusual reproductive structure, workers are unlikely to benefit from conflict except through benefits to their queen. This is in contrast to humans where individual participants in conflict can benefit reproductively. They also note that age-related costs and benefits are different between humans and ants. In humans, youth, especially young males, typically gain the most from participating in war, while ants lack similar age-related asymmetry. However, ants that have this form of intergroup cooperation typically are highly related to each other having been founded by a few individuals who then form interconnected nests. In this case, relatedness alongside the reproductive structure facilitates peace, while humans must rely on social institutions to create peace. These similarities and differences provide important insights into other pathways to peace.
These commentators, along with Majolo, all note that our knowledge of the behavior of other species is increasing so we should be wary of claims that attribute a unique behavior to humans. We need to engage in more detailed systematic cross-taxa studies to better understand whether other species have peace, and the pathway to the evolution of our own peace, and I agree.
R4. When did peace evolve?
R4.1. Peace within the past 100,000 years
I argue that there is evidence of human intergroup trade, and by extension, cooperation, beginning 300,000 years ago. However, as with other inferences based on the paleoarchaeological record, dates are unlikely to represent the earliest instances of a behavior due to sampling biases. It is plausible, then, that intergroup cooperation is older but that evidence for it has not yet been found.
However, I argue that peace is more than just cooperation – it refers to durable, harmonious relationships between groups. Societies in the middle Paleolithic were likely decentralized and egalitarian within age and sex. This social structure makes regulating the behavior of any group member difficult. At the same time, there were likely to have been some benefits from intergroup violence, such as obtaining valuable resources, or reproductive opportunities. At the same time, intergroup conflict could also arise simply from a misunderstanding. Given these conditions, it would have been exceedingly difficult to prevent unilateral aggression by motivated individuals and then to resolve it once it began. Thus, I argue that although cooperation is ancient, peace was not likely to occur until sufficiently strong social structures developed to regulate individual behavior and enforce group-based norms.
Many commentators disagree with the timing I propose and argue peace is much older. Antony for instance, argues that peace was a gradual process and emerging well before 100 kya because, he posits, our capacity to sanction within-group norm violations is much older. Fuentes, Kim, & Kissel argue that even at the birth of our species around 300 kya our species would have been able to engage in “peacefare.” Intriguingly, they argue that conflict well before the 100 kya date I give would have fueled the development of social structures that could promote peace. Wrangham seems to be in agreement with this timeline arguing that language was likely present at the birth of our species 300 kya and would have contributed to peacemaking between groups. Pisor, Smith, & Deminchuk (Pisor et al.) argue that intergroup cooperation, and possibly peace, would have been present possibly at the beginnings of our genus 2 million years ago due to marriage exogamy where individuals marry members of other groups. Such exchange, they argue, would have fueled intergroup cooperation and alliances. Montoya & Pinter make a more extreme claim appearing to argue that peace predates the Homo lineage when the brain size of prehuman ancestors began to increase due to navigating increasingly complex social relationships.
Hypothesizing about when peace emerged depends on the definition of peace one uses. If peace is periodically cooperative and generally nonaggressive interactions, then peace is likely to have emerged when individuals and groups would have benefited from interacting with out-groups. On this definition, peace probably well precedes the origin of modern humans, appearing at least since the birth of our species, and possibly earlier as Pisor et al., Montoya & Pinter, and others argue. If peace is stable, harmonious relationships without the threat of violence, then given the potential benefits one could possibly obtain from violence against out-groups (goods, reproductive opportunities, etc.), peace likely developed when we were able to create social institutions that can prevent and resolve such violence.
I am skeptical that an increase in brain size would have fueled peace as Montoya & Pinter argue because presumably it would also enabled conflict for the same reasons. I am also skeptical of making too much of the role of reproductive exogamy that Pisor et al. suggest because chimpanzees have exogamy but also high rates of lethal intergroup violence. However, when reproductive units (pair bonds, etc.) transformed into the cultural institution of marriage, then it seems more reasonable to think about reproduction being used to aid intergroup alliance formation (Walker, Hill, Flinn, & Ellsworth, Reference Walker, Hill, Flinn and Ellsworth2011). But marriage, on my interpretation, is likely to be a more recent development in our species' history. I agree with Antony that the emergence of peace was a gradual process and I did not intend my date of approximately 100,000 years to be taken as a bright line but instead to represent a range from approximately 150kya to 75kya or so. However, I disagree with Antony's assessment that peace would have been present significantly deeper than 100 kya because, as the target article shows, regulating norms for peace requires social institutions that were unlikely to have been present prior to these dates.
Overall, I am extremely sympathetic to the position of these commentators that intergroup cooperation and peacefare is much deeper than 100,000 years and likely predates our species. And if our prehuman ancestors lived in multilevel societies as other commentators have noted (Grueter; Wilson) we would have been well on our way to developing peace. Yet, I am still skeptical of whether there was full blown peace as defined in my paper earlier in our species' history for two reasons: (1) Intergroup cooperation would be expected to leave archaeological remains. Although humans and our prehuman ancestors have been using stone tools for several million years, the evidence for intergroup cooperation through stone tool trade is absent before 300,000 years. (2) Not only does intergroup contact leave archaeological remains, but it also fuels cumulative cultural evolution, and there is little evidence of this early in our species' history. On my interpretation prior to the past 100,000 years ago, the cultural took kit of humans was simple, not rich, diverse, and changing like one would expect with regular intergroup cooperation and peace. Further, the simple tool kit indicates to me that there are fewer benefits to be gained from regular intergroup interaction.
R4.2. The coevolution of material technology with peace
In the target article, I argued that between 615 and 500 thousand years ago, our prehuman ancestors became increasingly choosy about the lithic materials they worked with. Approximately 300,000 years ago, evidence of what I interpret as intergroup cooperation in the form of trade emerges in the form of long-distance lithic transport. From 615,000 years onward, and especially from 300,000 years I would expect increasingly cooperative intergroup interactions. However, it is only within the past hundred thousand years [150k–75k], that our species begins to produce distinctively human behaviors. During this period, innovation and wide variation in material and stone tool technology develops, alongside the other distinctive hallmarks of Homo sapiens, including art, music, status markers, and a complex and varied tool kit (Aubert et al., Reference Aubert, Lebe, Oktaviana, Tang, Burhan, Hamrullah and Brumm2019; Brumm et al., Reference Brumm, Oktaviana, Burhan, Hakim, Lebe, Zhao and Aubert2021; Henshilwood, d'Errico, Vanhaeren, van Niekerk, & Jacobs, Reference Henshilwood, d'Errico, Vanhaeren, van Niekerk and Jacobs2004; Mackay et al., Reference Mackay, Armitage, Niespolo, Sharp, Stahlschmidt, Blackwood and Steele2022; Marean et al., Reference Marean, Bar-Matthews, Bernatchez, Fisher, Goldberg, Herries and Williams2007; Režek, Dibble, McPherron, Braun, & Lin, Reference Režek, Dibble, McPherron, Braun and Lin2018), Taken together, these are often referred to as behavioral modernity and it is generally recognized to have emerged within the past 150 kya (though some argue for earlier origins; Kissel & Fuentes, Reference Kissel and Fuentes2018).
A leading explanation the origin of behavioral modernity is that demographic changes, including increases in population size and density, facilitated the development and maintenance of cultural complexity, which likely included increased complexity in social organization as well (Powell, Shennan, & Thomas, Reference Powell, Shennan and Thomas2009; Shennan, Reference Shennan2001; Stiner & Kuhn, Reference Stiner and Kuhn2006; Zilhão, Reference Zilhão2007). Prior to this, humans were thought to be biologically and cognitively adapted for complex cultural traits but the demographic processes to sustain them once developed were lacking. Once the requisite demographic shift occurred, complex cultural traits could be sustained. At the same time, the development of more complex and specialized material and social technologies would have fueled increased intergroup contact, which would have in turn fueled more cultural evolution.
While intergroup cooperation would have likely occurred prior to evolution of behavioral modernity, maintaining and developing it into stable harmonious relationships requires complex structures to prevent and resolve conflicts when individuals may benefit from intergroup conflict. Those social structures were unlikely to be present before the revolution that ushered in behavioral modernity. It seems a stretch that our species could build the complex social technologies for peace while lacking other behavioral modern traits. These include group-functional norms and the social structures to enforce them, alongside the flexible cultural technologies to resolve conflicts through peacemaking (sanctioning, rituals, wergild).
Instead, I think it more plausible that the factors that fueled one, fueled the other: When our material technology increased, so did our social technology. These both drove and facilitated intergroup contact, eventually resulting in the capacity to construct peace systems. It seems unlikely that a population of early humans who were incapable of flexibly responding to changing conditions with their material or social culture could build the flexible social systems required for peace. Prior to behavioral modernity, we may have had often cooperative relationships between groups where we sometimes exchange stone materials or mates through exogamy. But the transformation that enabled peace – the ability to regulate behavior with norms – only emerged with behaviorally modern humans. Based on the archaeological record, it is likely that most of those changes occurred in the past 100,000 [150kya to 75kya] years.
R5. Internal and external war
Hames; Pisor et al.; and Wrangham all note that I bypassed the discussion of internal and external war, and they argue that considering it may provide insights into the peace process. Internal war is war that occurs between groups that are members of the same society (such as feuding, civil war, etc.). External war is war between members of different societies, who typically have different languages or cultural practices. Hames; Zentall; and Wrangham all note that language may have had a crucial role in facilitating peacemaking among groups that speak intelligible languages (internal war). I agree that a shared language and cultural system appears conceptually to facilitate peacemaking.
I avoided the discussion of internal and external war for two reasons: A lack of space in the manuscript and I am unconvinced that it does much work in explaining war or peace. A shared language and culture may in theory facilitate peacemaking. But societies that share language and culture may be more likely to border each other, or to otherwise come into contact thus fueling conflict. It may be more difficult to create peace due to deep-seated grudges from a long history of conflict than between groups that have only had a few intermittent conflicts. These are empirical questions we do not know the answer too. But even when societies are distinct from each other, few human societies are as atomistic as those in the Andaman Islands, who are often used as a canonical example of external war. Instead, most societies, even ones that have external war with each other, often have friendships and some trade across group boundaries.
Pisor et al. note that in addition to not distinguishing internal and external war, I am also vague on the definition of a group (a point also made by Mathew & Zefferman), which creates ambiguity, especially with regard to pacification by state society which typically occurs through the threat of violence. They suggest resolving the apparent contradiction through distinguishing between internal peace and external pacification. This is a good point I had not considered. It also bears noting that the process of fusing and merging multiple groups who were previously at war occurs through violence and force either in the process of pacification by states, or through conquest.
R6. Paradigms for peace
R6.1. Peace as a social dilemma
Multiple commentators focus on my treatment of peace as a solution of a prisoner's dilemma, making useful suggestions about how to extend this work. Jeffries, Wright, & Lew-Levy (Jeffries et al.) see my argument as a starting point for modeling and ethnography to be used in tandem to better understand peace, and I heartily concur. They suggest a number of considerations for employing such a dual framework, including the use of iterated games, the revealing of in-group participant identities, and possibly using multiple games in parallel to capture varying aspects of the social dynamics. Their dual framework for exploring social phenomena through ethnography and well-chosen experiments and models could be, and perhaps should be, applied to multiple phenomena. Regrettably, much of contemporary evolutionary human sciences tends to overlook the crucial importance of ethnography. Ethnography is not only important to understand the phenomena but should also be a critical component of deciding which experimental paradigms are most likely to capture the relevant parameters of the social phenomena we seek to understand.
Böhm & Columbus argue that conceptual accounts such as the one I have provided need to be supplemented with economic games to better identify and study the mechanism of peace, and I fully agree. They note that some well-established games such as the intergroup parochial and universal cooperation game are appropriate to test the conditions of peace. Fischer, Avrashi, & Savranevski (Fischer et al.) correctly point out that the prisoner's dilemma I use as a conceptual foundation for understanding the problem of peace is the tiniest slice of potential ways to model social interactions. There are hundreds, if not thousands, of other ways to consider payoffs from various strategies that would provide insight into the evolution of war and peace, some of which may better capture the social dynamics. They suggest a different paradigm than the prisoner's dilemma, called subjective expected relative similarity which appears to cover the broad set of 2 × 2 games, as well as the preconditions I stipulate for peace. Their model integrates various kinds of information into a single-decision rule of whether to cooperate or defect and is arguably computational simpler. Whether their model, the PD, or some other game best captures the social dilemma of war and peace is an open question but one that demands more research.
As all these commentators note (Böhm & Columbus; Fischer et al.; Jeffries et al.; Romano et al.) a critical step in understanding the evolution of peace is beginning to use the well-developed experimental and analytic tools that have been used to study war to examine the conditions in which peace emerges. My hope is that this paper will stimulate new experimental and modeling work into better revealing the conditions in which peace emerges.
Several commentators took issue with my framing of peace as the outcome of a prisoner's dilemma where players choose between cooperation and defection (aggression) where peace is the solution. Montoya & Pinter argue “Groups avoid conflict – and even seek peace – when given the opportunity. Laboratory variations using ‘minimal’ groups (Tajfel et al., Reference Tajfel, Billig, Bundy and Flament1971) show that when group members allocate resources to the ingroup and outgroup, they divide money equally between them.” While I agree with a number of other points in their commentary, I strongly disagree with this assessment. A key contribution of my paper is that the framework whereby groups have agency such as in “groups avoiding conflict” is conceptually incompatible with small-scale decentralized groups. Rather than speaking of groups as having agency, the locus of action is in the individual and their actions then have consequences for the other members. Further, while the research by Tajfel et al. in Reference Tajfel, Billig, Bundy and Flament1971 may show that money between groups is allocated evenly, much recent experimental work demonstrates that the path to conflict between groups is easy, and group members often do not make even allocations between their own group members and other groups (Dogan, Glowacki, & Rusch, Reference Dogan, Glowacki and Rusch2018, Reference Dogan, Glowacki and Rusch2022).
Imada & Mifune make the important point that individuals and groups are free to do nothing – neither cooperate or defect – but simply avoid others. Without consideration of this third strategy of doing nothing, they argue I overestimate the likelihood of conflict. I focus on these two strategies because the strategy of doing nothing does not lead to peace or war and these are what I seek to understand. A strategy of “do nothing” leads to avoidance, which begs little explanation. Further, humans seldom avoid other groups, our psychology seems motivated to seek out and approach others, which is one reason we have been able to successfully colonize the globe. Further, experimental games to study conflict where the “do nothing” strategy may dominate typically involve making an allocation of a low-value currency between oneself or one's group members, and anonymous others of another group. While this may represent the state of the art for studying intergroup relationships, I'm skeptical about pushing it too far. Opportunities to interact with other groups versus avoiding them has a very different salience in the real world than in lab experiments where doing nothing involves not splitting or taking money with anonymous others. Humans are curious, often impulsive, and it will be exceedingly difficult to avoid visiting one's neighbors or strangers if one lacks information about them. The history of group relationships in our species often involves doing something, not nothing. So while I agree the strategy space does include doing nothing, I do not think it materially alters the conceptual framework I've developed about the difficult and requirements for peace.
The second important point Imada & Mifune make is that social structures can also instigate people into participating in conflict. I fully agree: My own research has shown how leaders and cultural institutions can solve the collective action problem in warfare (Glowacki et al., Reference Glowacki, Isakov, Wrangham, McDermott, Fowler and Christakis2016; Glowacki & McDermott, Reference Glowacki and McDermott2022; Glowacki & von Rueden, Reference Glowacki and von Rueden2015). In the target article I write, “The development of increased social complexity enables both peace and war; thus, tribes have a greater capacity for peace and more intense warfare than bands, chiefdoms more than tribes, states more than chiefdoms.” The point is that social institutions can be wielded to promote any behavior – for good or harm. As the institutions and their associated norms become more powerful, their efficacy at promoting war and peace also increase.
R6.2 Mechanisms enforcing peace
In the target article I argue that peace requires the ability to enforce potentially costly norms, including nonaggression, cooperation, and restitution. Based on a variety of ethnographic support I argue that the societies that appear to be able to do this successfully often do so with strong sanctions, potentially including physical punishment. While I am not explicit about it, the implication is that verbal sanctions, such as gossip, or exclusion are likely to often be insufficient for costly norm enforcement. Both Huang and Lie-Panis & André argue that I underestimate the importance of reputation in the evolution of peace. Lie-Panis & André elaborate on the framework I have outlined and provide new testable predictions. While they agree with the thesis in the target article that cultural technologies solve the cooperative dilemma between war and peace, they insightfully note that these technologies are also solutions to cooperative dilemmas. They hypothesize that they do so by leveraging reputation, which then solves the first-order cooperative dilemma. This is an intriguing proposition – and I agree reputation is a powerful mechanism in small communities. I am unconvinced, however, that reputation alone is enough to fill in the gap. I look forward to seeing the authors further develop this line of thinking.
Huang takes issue with the emphasis I put on strong sanctions, arguing that reputational sanctions are likely to be more effective at promoting peace. I agree with Huang that immaterial sanctions, including reputational damage and social rejection can be very effective at enforcing social norms, and in some cases perhaps more effective than physical sanctions. Further, reputational sanctions appear more common in hunter–gatherers than material or physical sanctions (Garfield et al., Reference Garfield, Ringen, Buckner, Medupe, Wrangham and Glowacki2023). My argument about the importance of stronger sanctions is based on ethnographic observation but is not intended to imply that reputation is not important, only that it often appears insufficient. During the course of my own research with east African pastoralists who regularly participate in intergroup raiding, discussions of raids would often involve consideration of the response of community members. However, what appeared to be of more concern was the withdrawal of material support, especially livestock. More research will hopefully resolve the relative importance of differing kinds of sanctions.
R6.3 Group structured cultural selection
Mathew & Zefferman's commentary focuses on areas I was unable to cover, especially why certain norms are transmitted. They note that the Turkana, like other multilevel societies, consist of multiple subgroups and norms regulating behavior between subgroups promote cooperation, while norms between different societies (such as Turkana and Samburu) are more likely to promote conflict. They make the point that I argue that peace is a challenge that requires norms to solve, but instead of developing norms for peace, groups could just drop war-promoting norms. I disagree, in part, with this assessment. In the target article, I make it clear that many societies do have norms that promote war. However, I do not mean to imply that norms are necessary for participation in small-scale decentralized warfare that primarily takes the place of ambush raids where attackers face low risk. Peace is more of a challenge than war because small-scale decentralized war does not necessarily require norms for participation, just as lethal raiding in chimpanzees appears to occur without norms. Participants may directly benefit from warfare without norms through capturing items of material value, being motivated by an evolved psychology for revenge, or by gaining reproductive opportunities through coercion, or through other pathways. Thus, small-scale decentralized war in the form of ambush raids in humans and chimpanzees may not be a significant collective action problem (Glowacki & Wrangham, Reference Glowacki and Wrangham2015; Massaro et al., Reference Massaro, Gilby, Desai, Weiss, Feldblum, Pusey and Wilson2022). As war becomes higher cost and on a larger scale, norms for participation are expected to become more important, as Mathew's previous work has shown (Mathew, Reference Mathew2017; Mathew & Boyd, Reference Mathew and Boyd2011).
Mathew & Zefferman note that I do not identify a mechanistic process for the patterning of peace and war. They argue that the patterning of peace and war can be explained by group-structured cultural selection. This is a plausible explanation for the spread of certain norms but does not explain where the norms for peace and war initially come from. Elsewhere I have argued that norms emerge from individuals who have overlapping self-interest and then enforce their interests on others (Singh, Glowacki, & Wrangham, Reference Singh, Glowacki and Wrangham2016, Reference Singh, Wrangham and Glowacki2017). Because groups are composed of heterogenous individuals with different interests, and the same individual may have competing interests, norms for peace and war may coexist in a group and eliminating a category of norms, such as those promoting war is difficult.
R7. The evolutionary psychology of peace
R7.1. Cooperation
McDermott; Baumeister & Bushman; and Montoya & Pinter all note the important role that within-group relationships had in the evolution of the capacity for peace. McDermott correctly points out that cooperation precedes intergroup cooperation and was likely a crucial aspect for the evolution of our species, both helping us compete against other groups, but also against nature. In part, our cooperative abilities may have emerged to help humans manage within-group conflict. Both McDermott and Baumeister & Bushman argue that leadership may be an important part of this. Cross-cultural work supports their hypothesis demonstrating that across small-scale societies leaders tend to have a primary role in conflict management (Garfield, Reference Garfield2021). McDermott makes the insightful observation that if intergroup war is mostly male coalitionary behavior, then the evolution of cooperative tendencies may have evolved differently for women. This is an underexplored area of research; however, among small-scale pastoralist societies, my own work shows that women, not just men, may have a critical role in resolving in-group conflicts (Garfield & Glowacki, Reference Garfield and Glowacki2023). There is convergence between the approach I develop in the target article and the commentary by Baumeister & Bushman who have argued that becoming human required the human mind to develop in a certain way so that it can both adopt and create culture. In the case of peace, this required capacities such as social identity, norm enforcement, and tolerance. They point out that the cognitive requirements that underlie peace, probably evolved due to pressures in other aspects of social life, such as in-group interactions, but then facilitated out-group cooperation.
R7.2 Cognition
Coolidge notes that radical evolutionary changes in behavior almost assuredly require neurobiological reorganization, which my account neglects. Coolidge proposes one such change, the expansion of the parietal lobe, which allowed our ancestors to better regulate their emotions and expanded our capacity for theory of mind. Both would have had made cooperative intergroup relationships more common. Coolidge argues that the function of the parietal lobes may be an exaptation, rather than the product of selection for that function. As the field of evolutionary neuroscience is still in its infancy, these questions remain unresolved. Riordan points out that peace requires other cognitive mechanisms to be in place, most importantly common knowledge and the ability to mentalize the states of others, and I agree that this is an important prerequisite for solving the challenges of maintaining harmonious relationships between human groups.
Sijilmassi, Safra, & Baumard (Sijilmassi et al.) appear to largely accept my argument and illustrate areas where there is articulation with work in social cognition. Crucially, they note that peace requires the ability of human agents to keep track of positive and negative social relationships, which they term an alliance detection system. Other factors underlying peace institutions, such as leadership and sanctioning, may also engage core aspects of social cognition. Indeed, the authors cite research showing that these sometimes function as cues that individuals use to infer cooperative networks. As Sijilmassi et al. note much more work remains to be done. When and why did these capacities evolve? Were they specific for intra- or intergroup relationships or coopted? To what extent are they shared with other species, or do other species achieve tolerant intergroup relationships while lacking them?
My account of peace requires that humans are able to follow and enforce group-based norms but understanding how and when humans evolved a norm psychology is one of the great challenges of human evolutionary studies. McCullough & Pietraszewski focus on the computational difficulties involved in achieving group-based norms. Specifically, they argue that doing so is achieved by simpler three-person interactions which are used to project group membership. Then the dynamics to reach a group-based decision, which is based on what the other group is projected to do are exceedingly complex, resulting from individuals with their evolved psychologies competing to promote their interests. This is consistent with the framing the target article provides and is one reason why norms that require individuals to act against their own self-interest are so hard to achieve.
R8. Extensions
Multiple commentators pointed out areas of the target article I neglected or that could be extended. Zentall argues that shared stories can bind groups together, or even transcend groups, enabling separate groups to find commonalities, though it could also be used to fuel conflict. Narrative is a powerful factor in shaping the behavior of individuals but it remains an open question of the extent to which it shapes intergroup relationships among hunter–gatherers. On my reading of the literature, it has only a small role at most. But further, while narrative may facilitate peace, it does not seem to be required for peace, and potentially can also be used to promote intergroup aggression. Kiper & Sosis note that I argue that cultural knowledge, including values and norms can enable peace, and imply that I argue that religion in early human societies promoted peace. However, in the target article I am conspicuously silent on religion precisely because we know so little about the timing of the first religions and their belief structure. Insofar as early religions resemble those found in more recent small-scale decentralized societies, religion probably did not feature prominently in regulating intergroup conflict. When and where it did, I agree with Kiper & Sosis that it may have facilitated both cooperation and conflict depending on the circumstance and that religious beliefs may developed in response to local conditions, including war and peace.
Both Hames and Zentall focus on the role that resources may have in promoting conflict. Zentall argued that the development of farming would have increased the need to negotiate boundaries between groups. He notes that this subsistence transition would have fueled changes in social organization, including expanded forms of alliances and hierarchies that may have facilitated both war and peace. I agree that there is a strong relationship between social organization and subsistence but I think the role of agriculture compared to hunting and gathering remains unclear. Some foragers are well-known for hierarchical and coercive forms of social organization including property ownership and even slavery. Hames notes that population density may be associated with higher rates of war among many small-scale societies, perhaps because of increased competition for food resources. Mobile hunter–gatherers typically have lower rates of war, while presumably having lower competition for food resources. While I agree with Hames that resource competition and food production are important factors in understanding war and peace, and a full explanation of war will connect group-level benefits such as potentially expanding territory, to individual motivations for war. If a group is successful in war and all members can expand into the neighboring territory, why would any individual participate in war, and not just free-ride on others? This is part of the central dilemma of the collective action problem in war.
Rusch takes the approach to framing the challenge of peace as one of policing. He points out that policing presents its own problems, and in decentralized groups with strong degrees of personal autonomy it may be especially difficult. Because one does not know if oneself will be the victim of retaliatory violence, sanctions for norm violations resemble third-party punishment. But third-party punishment appears rare in small-scale societies (Fitouchi & Singh, Reference Fitouchi and Singh2023; Singh & Garfield, Reference Singh and Garfield2022), exactly the type of punishment that would be needed to curtail a group of would-be raiders. Rusch notes that I do not address how early peacemakers, states, and colonizers are able to resolve the policing dilemma. He hypothesizes the emergence of something akin to a “peacelord,” or leaders who balance the urges of youth with the interests of the entire group. In societies without a strong governmental presence, Rusch notes that gangs and nonstate actors do solve the “policing” problem of peace with young men endogenously organizing enforcing their monopoly on violence, and creating a form of peace. These are tantalizing questions about which we know so little in small-scale societies.
Brown, Brown, Cavallino, Monterroza, Li, & Huang identify that aspects of my argument may be relevant to understanding how to create peace today among humans. They focus on how to get groups to increase their interdependence through promoting interpersonal relationships or avenues for cooperation. At the same time, many areas of shared interdependence exist that may be unrecognized and further recognition may promote peace. I largely agree with their conclusion that increased recognition of overlapping interdependence would be useful in promoting peace. Applied efforts to reduce war and other forms of intergroup violence would be advanced through consideration of both the motivation of individuals, but also through careful empirical work. Unfortunately, many peace efforts appear to proceed without these considerations.
R9. Conclusion
My goal in writing the target article was to move the debate past the question of war versus peace in human evolution and to shift the focus to the conditions which enable peace or war. In doing so, I have speculated on the nature of early human societies and when specific features such as intergroup cooperation, trade, and norm enforcement developed. It is my intention that this is a first statement of an ongoing discussion rather than a final statement. Thus, the account I offer is tentative, and will be surely updated as new evidence and new models emerge.
Several commentators make a strong case that polydomous ants, bonobos, and species living in multilevel societies have peace or important elements of it. Polydomous ants are typically highly related to their neighboring nests bonobos lack high rates of predatory aggression as well as potential benefits from raiding, and multilevel societies share social group membership. Thus, while humans may not be alone in having peace, we should be struck by the fact that this is still just a handful of species. Peace appears to be extremely rare, especially among species that regularly kill each other as humans do. This requires an explanation, and the target article provides one that I think still stands. At some point in our history, our species began to benefit from interacting with out-group members, possibly shifting from a strategy of avoidance, or tolerance to seeking out opportunities to interact. Our norm psychology alongside social institutions enabled us to create the conditions in which intergroup violence could be prevented thus allowing the development of durable harmonious intergroup interactions. We then developed cultural technologies allowing us to restore relationships after a conflict. Thus the pathway to human peace was long and not inevitable.
The last sentence of Mathew & Zefferman's commentary is a fitting statement for understanding human war and peace. They write “if primordial propensities for war or peace exist, they seem to be quite readily overwhelmed by local cultural norms.” While the human lineage may have contained conditions which favored lethal aggression between social groups, with the right norms and social institutions, whatever tendency there may be toward war can be suppressed allowing us to create peaceful harmonious societies.
Competing interest
None.
Target article
The evolution of peace
Related commentaries (30)
A game of raids: Expanding on a game theoretical approach utilising the prisoner's dilemma and ethnography in situ
A neurological foundation for peaceful negotiations
Capacities for peace, and war, are old and related to Homo construction of worlds and communities
Creating shared goals and experiences as a pathway to peace
Cultural technologies for peace may have shaped our social cognition
Economic games for the study of peace
Enhanced cooperation increases the capacity for conflict
Evolution, culture, and the possibility of peace
Experimental evidence suggests intergroup relations are, by default, neutral rather than aggressive
Group-structured cultural selection can explain both war and peace
How language and agriculture promote culture- and peace-promoting norms
Impediments to peace
Is peace a human phenomenon?
Language likely promoted peace before 100,000 ya
On peace and its logic
On the evolved psychological mechanisms that make peace and reconciliation between groups possible
Peace as prerequisite rather than consequence of cooperation
Peace in other primates
Peace is a form of cooperation, and so are the cultural technologies which make peace possible
Police for peace
Rethinking peace from a bonobo perspective
Social and economic interdependence as a basis for peaceful between-group relationships in nonhuman primates and humans
Social norms, mentalising, and common knowledge, in making peace and war
The evolution of (intergroup) peace hinges on how we define groups and peace
The evolution of peace (and war) is driven by an elementary social interaction mechanism
The importance of social rejection as reputational sanction in fostering peace
The intertwined nature of peace and war
The psychology of intergroup relations was grounded in intragroup processes
The role of religion in the evolution of peace
The roots of peace
Author response
Author's response: The challenge of peace