2.1. The structure of decentralized war

Understanding how peace is achieved in small-scale decentralized societies requires first understanding how and why individuals participate in war in these same types of groups. Small-scale decentralized societies have a fundamentally different pattern of conflict than state societies with militaries (Wright, Reference Wright1942). Counterintuitively, the individual costs of participation in war appear to be relatively low and the potential marginal benefits significant. Small-scale warfare is acephalous and decentralized, occurring without formal leadership or chains of command, mechanisms to compel participation, and mechanisms to restrain conflict. Membership is typically ad hoc, composed of available people who want to participate, and leadership is informal, situational, and noncoercive. Unlike militaries which can involve years of compelled participation, small-scale warfare lasts for the duration of the event – hours to days – after which the participant returns to their ordinary life. Raiding parties often form without consent or even the knowledge of the larger social group, coordinated by one or two people who convince others to join them.Footnote ¹ Unlike warfare in state societies, war in small-scale societies does “not seem to be carried out with any global strategy in mind” (Tornay, Reference Tornay1979, p. 114).

The most common pattern of war is the raid, primarily composed of young men. Raids are usually undertaken to fulfill the proximate goals of the raiders themselves which may include revenge, capturing loot, or gaining status. Raiding parties use strategic timing and ambush to attack one or two victims at very low risk to themselves, usually while the victims collect water, do daily activities, or exit their village in the morning (Gat, Reference Gat1999). The victims may be members of another ethnolinguistic community or members of the same ethnolinguistic community, but of a different lineage or clan (as in feuding). Because the primary tactic in small-scale war is surprise, raiders can choose to attack when the odds heavily favor their success. As a result, attackers on raiding parties face an extremely low risk of being killed or injured during an attack (Beckerman et al., Reference Beckerman, Erickson, Yost, Regalado, Jaramillo, Sparks and Long2009; Chagnon, Reference Chagnon1988; Glowacki et al., Reference Glowacki, Isakov, Wrangham, McDermott, Fowler and Christakis2016; Mathew & Boyd, Reference Mathew and Boyd2011; Wrangham & Glowacki, Reference Wrangham and Glowacki2012). A similar pattern is found in chimpanzees, who also form raiding parties that attack members of other groups when they have a significant imbalance of power (approximately eight attackers to one victim) with little evidence of chimpanzee attackers being seriously injured or killed (Wilson & Wrangham, Reference Wilson and Wrangham2003; Wilson et al., Reference Wilson, Boesch, Fruth, Furuichi, Gilby, Hashimoto and Koops2014). When there are casualties among human attackers, it is usually because they are detected and ambushed while traveling to the site of their intended raid but such accounts are rare (Wrangham & Glowacki, Reference Wrangham and Glowacki2012). Despite the low risk to attackers, members of raiding parties still must overcome fear and confrontational tension (Collins, Reference Collins2009; Mathew & Boyd, Reference Mathew and Boyd2011; Roscoe, Reference Roscoe2007). “This fear is curious because there is no memory of any Wao raider being killed, or even seriously injured, by the Waorani he attacked” (Beckerman et al., Reference Beckerman, Erickson, Yost, Regalado, Jaramillo, Sparks and Long2009, p. SI: 1). While the risks to attackers on raids are low, the overall mortality rates from intergroup violence can be high, though the severity is primarily driven by victims of raiding parties rather injuries to attackers.

Thus far we have described the most common pattern of small-scale warfare that has close parallels to intergroup conflict in chimpanzees. As societies increase in sociopolitical complexity, they often adopt more structured forms of intergroup violence, such as battles (Dye, Reference Dye2009, Reference Dye2013; Glowacki, Wilson, & Wrangham, Reference Glowacki, Wilson and Wrangham2020), which cangreatly increase the mortality rate of attackers as well as the chances of the defenders being successful (Dreu & Gross, Reference Dreu and Gross2019). Structured organized conflict such as high-risk battles presents a different set of strategic dynamics that may better approximate the conditions under which states wage war than the pattern commonly found in decentralized societies (Buckner & Glowacki, Reference Buckner and Glowacki2019).

2.2. The individual benefits to attackers

Attackers in small-scale warfare often benefit personally from their participation through private incentives. Status is almost universally accorded to warriors, thus war often provides an important arena for men in the same society to compete with each other for status (Gat, Reference Gat2009; Glowacki & Wrangham, Reference Glowacki and Wrangham2013; Wright, Reference Wright1942). Across societies, even among mobile hunter–gatherers, warriors frequently take material plunder, including captives or goods (though mobile foragers appear to do so to a much lesser extent than other types of subsistence) (Cameron, Reference Cameron2011; Gat, Reference Gat1999, Reference Gat2000). Captives can be used as reproductive partners, for labor as slaves, or to expand one's kin networks through adoption. In the few cases where the individual benefits of warfare have been quantified, they appear to improve the reproductive opportunities of warriors (Chagnon, Reference Chagnon1988; Dunbar, Reference Dunbar1991; Fleisher & Holloway, Reference Fleisher and Holloway2004; Glowacki & Wrangham, Reference Glowacki and Wrangham2015; Hames, Reference Hames2020; Macfarlan, Walker, Flinn, & Chagnon, Reference Macfarlan, Walker, Flinn and Chagnon2014, Reference Macfarlan, Erickson, Yost, Regalado, Jaramillo and Beckerman2018). The specific mechanisms are likely to vary between societies ranging from increased access to bridewealth, opportunities to make alliances with people who may provide reproductive partners, increased desirability as a potential partner, or other cultural mechanisms (though see Beckerman et al., Reference Beckerman, Erickson, Yost, Regalado, Jaramillo, Sparks and Long2009, for a potential counterexample).

Even in instances where intergroup violence is not socially endorsed, attackers still often receive social benefits from their peers. The ethnography of small-scale societies is replete with examples in which intergroup violence is subject to general reprobation or even punished, but a smaller subset of society may laud warfare, providing the attackers with status among their peers. In the absence of material or social incentives, war can provide endogenous motivations through “offer[ing] excitement not found in the village” (Westermark, Reference Westermark1984, p. 116). “Old informants speak about the pleasurable excitement in preparing for and setting out on a… raid…. [which] might even have been welcomed as a break to long, tedious hours of work…” (Dozier, Reference Dozier1967, p. 78). Thus, even if society at large does not accord warriors with prestige, and war is unlikely to result in captured loot, warriors may still be endogenously motivated to participate in raids.

2.3. The collective costs and benefits of war

Despite the common assumption that warfare in human groups is driven by competition for natural resources, there is mixed evidence of a relationship between competition for resources and the intensity, frequency, or scale of war in small-scale societies (Adano, Dietz, Witsenburg, & Zaal, Reference Adano, Dietz, Witsenburg and Zaal2012; Scheffran, Brzoska, Kominek, Link, & Schilling, Reference Scheffran, Brzoska, Kominek, Link and Schilling2012). Many ethnographers argue that there is no relationship, as warfare commonly occurs in regions with abundant resources including territory. In many cases, successful groups may not acquire the territory of the defeated groups. Moreover, any territory acquired through war would be a collective benefit available to both warriors and nonwarriors, exacerbating the collective action problem of intergroup violence.

While individual warriors may benefit from participating in war, there are two major collective costs from warfare borne by all members of the attackers' group: The risk of being killed or injured in a revenge attack and decreased access to resources through reduced opportunities for intergroup contact and the creation of unused buffer zones. The desire for revenge is a major proximate cause of war in small-scale societies and often results in the deaths of more people than the initial offense (Boehm, Reference Boehm2012a; Walker & Bailey, Reference Walker and Bailey2013). After an attack, the most likely response from the attacked group is to launch an attack of their own against the offender's group, thus leading to tit-for-tat raiding. Because the specific identity of individual attackers is usually unknown, any member of the offender's groups will suffice as a target. As a result, the original attackers are usually at no or little more at risk of being a victim of revenge than any other group member. The risk of retaliation then falls on all group members, regardless of their participation in the initial intergroup conflict.Footnote ²

In addition to the risk of being killed in revenge, wars impose collective costs by reducing opportunities for trade, the exchange of information, and access to potential reproductive partners both within and between groups. While cooperation frequently continues across group boundaries during intergroup conflict, it is often reduced or severely curtailed as people avoid interacting with members of groups that are hostile to them. War also has the often-devastating effect of producing large unused border or buffer areas that people avoid (Evans-Pritchard, Reference Evans-Pritchard1957; Glowacki & Gonc, Reference Glowacki and Gonc2013; Turton, Reference Turton, Fukui and Turton1979). People may also flee areas at high risk of conflict even if those regions are resource abundant, losing access to valuable resources.Footnote ³ For subsistence populations, these large unused border zones can mean the devastating loss of access to productive game land, grazing areas, and water sources.

2.4. The cooperative dilemma of war and peace

I have shown that participation in small-scale war is low risk to attackers because of the strategic use of ambush and assymeteries in the number of attackers and defenders. At the same time, attackers are likely to receive important material and social benefits, especially status. The costs of war, however, are primarily borne by all members of the attacker's group, including the risk of retaliation, the creation of unused buffer zones, and the loss of opportunities that come from intergroup contact. As a result, a dynamic exists in which it may be individually beneficial to initiate intergroup violence because of the possiblity of receiving private benefits, but simultaneously costly for other members of the group.

The insight that war may be hard to avoid even when peace is the most beneficial strategy for a group as a whole has been long recognized (Schelling, Reference Schelling1980). In fact, efforts to make one's own group more secure may ultimately increase the likelihood of conflict. This is because other groups are likely to respond in kind, particularly when they have incomplete information (known as the security dilemma) (Blattman, Reference Blattman2022; Levy, Reference Levy1998). The dynamic between war and peace is commonly modeled as a prisoner's dilemma where any individual member may be better off defecting (initiating aggression against out-groups), but the entire group would be better off with peace (cooperating) (Cohen & Insko, Reference Cohen and Insko2008; Coombs & Avrunin, Reference Coombs and Avrunin1988; Rusch, Reference Rusch2013; Snyder, Reference Snyder1971; van der Dennen, Reference van der Dennen2014). Depending on the dynamics of the conflict, other cooperative dilemmas may better match the specific context, including games of chicken or the stag hunt, or attacker–defender games (Dreu & Gross, Reference Dreu and Gross2019; Dreu et al., Reference Dreu, Gross, Méder, Giffin, Prochazkova, Krikeb and Columbus2016; Rusch, Reference Rusch2022; Schelling, Reference Schelling1980). Regardless of which cooperative dilemma is the best match for the specific group dynamics, the difficulty of limiting the payoffs of aggression by individuals is one of the most formidable barriers to the emergence of peace in small-scale societies.

Preventing conflict is difficult because a single act of aggression by one group member can be enough to trigger conflict (Fig. 1), as other members of the attacked group seek revenge. Thus peace requires coordinating the interests of all group members for nonaggression making sustained peaceful relationships difficult to achieve, especially once a conflict has started. “A fundamental reason for the perpetuation of cycles of raiding… was that a unilateral decision to cease fighting was impractical… so long as neighboring villages continued to be willing to fight” (Ploeg, Reference Ploeg, Rodman and Cooper1979, p. 143). It also means that even one individual acting unilaterally can determine the nature of intergroup relationships. As Clastres notes (Reference Clastres2010, p. 193), “The power to decide on… war and peace… no longer belong[s] to society as such, but… to the … warriors, which would place its private interests before the collective interest of society… The warrior would involve society in a cycle of wars it wanted nothing to do with.”

Figure 1. Peace as a prisoner's dilemma. Intergroup conflict can be studied as an iterated prisoner's dilemma. The key challenge to peace is developing payoff systems that favor cooperation by member of both groups that are resilient against real or perceived defection.

The payoffs from aggression are not symmetric across a population because individuals vary in how much they are likely to benefit from their participation. Young men, in particular, are especially prone to status-seeking behaviors, including acts of aggression, exacerbating the conditions for war (Yair & Miodownik, Reference Yair and Miodownik2016). While women in small-scale societies rarely participate in violence themselves, they often have an important role in encouraging men toward violence through teasing or ridiculing men who abstain from violence.

Thus, achieving peace requires solving an iterated cooperative problem like the prisoner's dilemma that each member of a group plays repeatedly in encounters with any member of another group. This dynamic is further exacerbated by the fact that war does not necessarily have to originate with unprovoked aggression but can instead arise from routine conflicts between individuals. Conflicts are an inevitable feature of social life no matter how pacific the cultural values. Any conflict has the potential to escalate, resulting in violence and triggering retaliation. Furthermore, peaceful exchanges or interactions may inadvertently result in the injury or death of a group member; an accidental death or injury may be interpreted as an act of aggression leading to retaliation and initiating a cycle of tit-for-tat war. Therefore, the conditions that give rise to peace must not only coordinate the interests of individuals toward cooperation but must also be tolerant and resilient against instances of real or perceived defection.

2.5. Relevance to centralized (state) warfare

My analysis focuses on intergroup violence in small-scale decentralized societies because these kinds of society best resemble our understanding of ancestral human groups. This analysis is both relevant to and diverges from warfare in centralized societies. In centralized societies such as states, or chiefdoms such as many Plains Indians, intergroup violence is typically directed through an organizational structure including chiefs, officers, or militaries. This organizational structure solves the coordination problems inherent in warfare by incentivizing and organizing combatants, preventing defection from cowardice and desertion (often through severe sanctions), and mitigating the risk of unprovoked aggression by group members. The organizational structure can also incorporate a global view of the group and use violence to achieve the goals of the group. Because of the centralization through which war is waged by states to advance the strategic aims of the group, the appropriate level of analysis is the group itself, not the individuals who compose the group (Schelling, Reference Schelling1980). Thus, Blattman (Reference Blattman2022, p. 17) writes about war in state societies, “Wars are long struggles…. Big groups are deliberative and strategic.”

This quotation highlights the fundamental difference between small-scale decentralized war and centralized war that underlies the game theoretical logic of war and peace: Whether the most appropriate level of analysis is the individual or the group. Small-scale war typically occurs through a raid that lacks any overall strategic objectives. Instead of raids being directed toward advancing the strategic objectives of the group, they are initiated to satisfy the often-short-term aims of the individual attackers, especially revenge and status. Although I focus on small-scale societies, similar dynamics are often found in decentralized urban violence (Buford, Reference Buford2001; Mays, Reference Mays1997; Shakur, Reference Shakur2007). Thus, the most appropriate level of analysis for the conditions of war in decentralized small-scale societies is the individual. It is the individual, notthe society that decides to initiate war.

Despite the differences between state and decentralized war, there are important similarities in the logic of war and peace. For both decentralized and centralized societies, peace is often more beneficial than war for both the group as a whole and the individuals within the group. Because of this, individuals often seek to maintain peace and prevent conflict. Many of the primary drivers of war are the same between decentralized and centralized societies (Blattman, Reference Blattman2022; Schelling, Reference Schelling1980): Individual actors who are able to initiate conflict without feedback from the group, such as group of young men who decide attack their neighbors in the case of a small-scale society or an authoritarian leader in control of the military (Putin) (Kleinfeld, Reference Kleinfeld2019); incentives for war that cannot be shared with the other group or are intangible, such as revenge or status (Levy, Reference Levy1998); and finally commitment problems. Groups cannot necessarily trust that their adversaries will honor their commitments toward peace, and to assume that the other side has cooperative nonaggressive intentions may leave them open for attack (Powell, Reference Powell2006; Walter, Reference Walter2009).

3.1. Capacity for tolerant interactions

Peace requires the psychological capacity for tolerant, nonaggressive interactions that cross group boundaries. While humans clearly have this capacity, many social species lack this ability. Chimpanzees, for example, rarely have tolerant intercommunity interactions; instead they usually avoid each other and when an imbalance of power exists, the larger group often aggresses the smaller group (Wilson & Wrangham, Reference Wilson and Wrangham2003). While bonobos do have intergroup aggression, they also have tolerant and cooperative intergroup relationships that can involve copulation and occasional food sharing. The fact that bonobos have intergroup tolerance suggests that the capacity for tolerance between groups may have developed early in the hominid lineage or even predate it. Once a capacity for tolerance was in place, social conditions such as the expansion of kinship networks (Chapais, Reference Chapais2009) or sanctions against overly aggressive individuals (Boehm, Reference Boehm2012b; Wrangham, Reference Wrangham2019) may have further increased our ability to tolerate strangers. Regardless of when a human capacity of tolerance emerged, intergroup cooperation requires the ability to tolerate strange individuals, something our chimpanzee cousins are incapable of. Thus, identifying when and how this ability arose will provide insight into the first crucial step necessary for peaceful intergroup relationships.

3.2. Payoff structure favors cooperation

Peace requires the psychological ability to tolerate strangers but tolerance itself is not sufficient for peace. Peace also requires the motivation to interact with members of other groups (unlike most group-living species, in which groups generally avoid each other). Positive intergroup interactions will be favored when individuals of both parties can benefit from their interactions, such as by accessing resources that would otherwise be unavailable (Pisor & Gurven, Reference Pisor and Gurven2016, Reference Pisor and Gurven2018). In nonhuman social animals, the potential benefits from intergroup interactions include opportunities to interact with potential reproductive partners, infer information about groups for future transfers, or learn about the relative size and strength of neighboring groups (Pisor & Surbeck, Reference Pisor and Surbeck2019). These potential benefits would apply to early humans. However, as early humans developed a more specialized subsistence niche, especially one that depends on complementarity (extra-household food sharing) and cultural technologies (spears, traps, tracking), the potential benefits would have expanded leading to increased incentives for intergroup cooperation.

The creation of interdependencies would have greatly amplified the potential payoffs for intercommunity cooperation. A common form of interdependency among subsistence societies is one in which groups that depend on unpredictable and variable resources allow others to access resources in their territory in times of need, such as water, game lands, or grazing (Cronk & Aktipis, Reference Cronk and Aktipis2021; Glowacki, Reference Glowacki2020; Kelly, Reference Kelly and Fry2013a, Reference Kelly2013b; Pisor & Jones, Reference Pisor and Jones2021). A potentially more important form of interdependence would have developed when groups began to rely on nonlocal resources or goods that other groups had access to and that could be procured through trade or social relationships (Schulz, Reference Schulz2022; Smith et al., Reference Smith, Pisor, Aron, Bernard, Fimbo, Kimesera and Borgerhoff Mulder2022). In small-scale societies, these include material goods, such as tools, stones for toolmaking, and ochre, as well as cultural knowledge including religious, ceremonial, or ritual information.

If intergroup conflict disrupts access to goods or other benefits from other groups, group members have a strong incentive to avoid conflict. This occurred in the Solomon Islands, for example, where “it must have required extraordinary self-control… to withstand the tantalizing temptation of having a go at each other. The remarkable thing is that peace of any duration obtained. What probably occurred was that each side badly wanted what the other had to offer; these considerations overrode appetites for bloodletting for more or less extensive periods of truce” (Oliver, Reference Oliver1955, p. 296).

3.3. Norms promote intergroup interactions

The capacity for tolerance and the possibility of benefiting from interactions with out-groups creates the conditions for intergroup cooperation of the type seen in bonobos, but these alone are insufficient for peace. When severe or lethal violence is a possibility, as in chimpanzees and many human groups, individuals are more likely to avoid interactions or even engage in preemptive aggression. Thus, peace also requires the ability to have reasonable expectations about whether interactions with out-groups are likely to be neutral, aggressive, or positive (avoiding neutral and aggressive interactions and seeking out positive interactions). This depends on the ability to predict both the behavior of one's own group members and the behavior of members of other groups. But how do we reasonably anticipate the behavior of our group members and members of other groups? We do so by adhering to and enforcing norms regulating the behavior of our group members with the knowledge that the other group does the same.

3.4. Mechanisms to resolve conflicts

Resolving conflicts is the most serious challenge to the development and maintenance of peace in small-scale societies. Conflicts often spread beyond the original parties to include the larger social group creating a cycle of tit-for-tat violence making resolution even more challenging (Garfield, Reference Garfield2021). Even when individuals who have been aggrieved do not wish to seek revenge, the social pressures to do so may be enormous. There also exists the possibility that unintentional harm caused by out-group members will be misinterpreted as having aggressive intent, triggering intergroup conflict (Fig. 2).

Figure 2. Examples of peacemaking rituals: (A) Andaman Islands: Peacemaking involves a ritualized dance between hostile groups where aggressive feelings are displayed culminating in an exchange of weapons (Radcliffe-Brown, Reference Radcliffe-Brown1922). (B) Enga: Distribution of compensation after a death, approximately 100 pigs were slaughtered and money distributed (courtesy of Polly Wiessner). (C) Peace agreements with Arbore and other groups in southwest Ethiopia involve symbolically blunting spears and (D) then breaking and burying the broken spears (Streker & Pankhurst, Reference Streker and Pankhurst2004).

3.4.1. Restitution and signaling cooperative intent

War [can be] triggered by an individual, [but] peace can only be re-established communally. (Girke, Reference Girke, Bruchhaus and Sommer2008, p. 202)

The key challenge after intergroup conflict is to prevent members of the aggrieved group from taking revenge. This often requires restitution to the aggrieved party for the harm they have suffered (see Table 1). This may involve in-kind exchanges, such as replacing stolen livestock with other livestock or the utilization of different currencies, such as providing the aggrieved group with a person from the offender's group (usually a young woman). Because blame is often ascribed to the group rather than the individual, restitution frequently comes from members of the perpetrator's group, rather than from the perpetrators themselves.

Table 1. Common conflict resolution mechanisms

Not only does the offending group have to offer restitution, but the aggrieved group must accept it as satisfactory. This negotiation provides another arena for conflict between groups as they determine an adequate level of restitution that satisfies both groups. For example, among the Kalinga, “kindreds [of the victim] are rarely satisfied with simply being paid off, and often retaliate by a counter-killing” (Dozier, Reference Dozier1967, p. 93). Reaching satisfactory compensation can be difficult, especially when tensions between groups are high.

At the same time, the offending group needs to signal cooperative intent, for example, that future interactions are likely to be positive and that the offender's actions do not represent a new norm on the part of the offender's group (Roscoe, Reference Roscoe and Fry2013). The need to signal cooperative intent is why peacemaking after a violent conflict often requires that the offending group execute one of their own group members. For example, among the Curripaco “lineage members decided to execute ritually their kinsman who had killed, rather than provoke a spate of tit-for-tat revenge killings” (Valentine, Reference Valentine, Valentine and Beckerman2008, p. 36). While among the Erbore of southwest Ethiopia, one elder reported “We brought about peace by allowing two Erbores…to be killed by our enemies. I, myself, have handed over one of our sons to be killed” (Sullivan, Reference Sullivan2008, p. 16). Drastic actions such as the execution of the offender can signal to the aggrieved group that future interactions are likely to be positive.

Because restoring or creating peace requires the community to reaffirm norms of cooperation and nonaggression toward the out-group, peacemaking often involves many people from both groups meeting to discuss the conflict and its resolution, often engaging in symbolic ceremonies indicating resolution (Table 1). This will commonly involve eating and drinking together, as well as rituals that symbolize that the conflict has been resolved and neither party desires revenge. Groups may break or bury items related to conflict such as spears or weapons, believing that peace may hold as long as these items remain buried (Strecker, Reference Strecker, Elwert, Feuchtwang and Neubert1999). Symbolic gifts may be given between members of the opposing groups that indicate a desire for peace (Bacdayan, Reference Bacdayan1969). Such traditions also exist in centralized societies, including states, with militaries often indicating surrender by turning over ceremonial swords.

3.5. Third-party mediators and leadership

We have seen that restoring relationships after a conflict requires the ability to sanction peace violators, the coordination of compensation between groups, and the ability to signal cooperative intent. These are difficult conditions to satisfy especially in the context of an ongoing conflict. Two factors can greatly increase the likelihood of peace: Leadership and third-party mediators. Despite their potential efficacy, small-scale decentralized societies often lack strong leadership and third-party institutions due to their egalitarian nature.

Leadership facilitates peace because individuals who wield asymmetric power can prevent war or establish peace using their influence over others (such leaders can also use their influence to motivate warfare) (Garfield, Syme, & Hagen, Reference Garfield, Syme and Hagen2020). As a result, peace efforts in small-scale societies are frequently led by prominent individuals who motivate in-group members to maintain peace, sanction offenders, and negotiate with out-group members (Fry, Reference Fry2007; Fry et al., Reference Fry, Souillac, Liebovitch, Coleman, Agan, Nicholson-Cox and Strauss2021; Glowacki & Gonc, Reference Glowacki and Gonc2013). Some societies institutionalized the role of peacemaker into a position such as a peace chief or peace leader (Bacdayan, Reference Bacdayan1969; Goldschmidt, Reference Goldschmidt, Sponsel and Gregor1994; Moore, Reference Moore1990), who “appeared at the scene of battle… and attempted to induce disputants to come to amicable agreement” (Goldschmidt, Reference Goldschmidt1951, p. 326). However, these kinds of formal peace leaders occur more frequently in societies with significant social stratification such as the Kalinga and Cheyenne. The absence of prominent leadership who can negotiate for peace is a key impediment to the development of peace in decentralized societies.

Third parties have an important role in restoring relationships after conflict in small-scale societies, whether within or between groups (Fitouchi & Singh, Reference Fitouchi and Singh2023; Hoebel, Reference Hoebel2009). Third-party mediators may be customary leaders or institutions, such as groups of elders or other bodies of prominent individuals, while in contemporary contexts they are more likley to consist of government representatives or nongovernmental organizations (NGOs) (Box 1). They often facilitate the negotiations about compensation and restitution such that they are acceptable to both parties, rarely relying on punishment for restoring relationships (Fitouchi & Singh, Reference Fitouchi and Singh2023; Singh & Garfield, Reference Singh and Garfield2022; Wiessner, Reference Wiessner2020). The absence of strong third parties to facilitate conflict resolution can be a serious impediment to peace. For example, among Wanggular of Melanesia “De-escalation was difficult…. There was no intermediary party… who could assist the two hostile parties to agree on the size and content of the payment…. Thus it seemed almost impossible for Wanggularm to settle quarrels” (Ploeg, Reference Ploeg, Rodman and Cooper1979, pp. 170–171).

6.1. Intergroup cooperation in the late Middle Pleistocene

Dramatic changes in early human behavior began around 300,000 years ago. Some of the earliest reliable evidence of regular long-distance transport of stone materials appears between 295,000 and 320,000 years ago, with raw stone materials being transported more than 50 km in straight line distance (walking distance would have been much greater), exceeding the typical home range of many recent hunter–gatherers (Brooks et al., Reference Brooks, Yellen, Potts, Behrensmeyer, Deino, Leslie and Clark2018). Similarly, at the Sibilo School Road Site in Kenya, there is strong evidence for long-distance transport of stone materials dating back to more than 200,000 years ago from sources located 25, 144, and 166 km away. Surprisingly, most of the transported obsidian is from the farthest source at 166 km away, not the closest source at 25 km away (Blegen, Reference Blegen2017). The distance these materials were transported is far greater than the estimated home ranges of forager bands and is more consistent with the exchange networks for modern hunter–gatherers, which could involve scores of people across hundreds of miles (Ambrose, Reference Ambrose, Liritzis and Stevenson2012; Bird, Bird, Codding, & Zeanah, Reference Bird, Bird, Codding and Zeanah2019; Yellen & Harpending, Reference Yellen and Harpending1972). This kind of resource movement suggests “intensive, perhaps even obligate intergroup exchange rather than down-the-line-exchange” such as the exchanges that characterize the Kula cycle (Ambrose, Reference Ambrose, Liritzis and Stevenson2012, p. 65). Around the same time, the use of ochre was increasing, and by 300,000 years ago it was in regular use in some regions, with much of it also being transported long distances, at a minimum of 38 km but potentially up to 170 km away (Watts, Chazan, & Wilkins, Reference Watts, Chazan and Wilkins2016).

Increases in intergroup exchange around 300,000 years ago are paralleled by skeletal changes in the human lineage toward increasing gracility. Skeletal and cranial gracility is often used as a proxy for reduced reactive aggression (Chirchir, Reference Chirchir2021; Wrangham, Reference Wrangham2019). Reduced reactive aggression allows for increased out-group tolerance, enabling affiliation with strangers. The earliest evidence for gracility among human ancestors comes from archaic Homo sapiens around 320,000 years ago (Wrangham, Reference Wrangham2019), around the same time as the emergence of long-distance stone transport, suggesting that humans around this period were becoming less reactively aggressive while simultaneously increasingly relying on intergroup trade.

The development of long-distance transportation networks, increased selectiveness of stone tool materials, bodily adornment with ochre, and reduced reactive aggression all around 300,000 years ago or earlier suggest strongly that the early human social environment was changing dramatically. These changes would have both enabled and promoted positive intergroup interactions, leading groups of early humans to seek out interaction with other groups they could possibly benefit from (Wilson & Glowacki, Reference Wilson, Glowacki, Muller, Wrangham and Pilbeam2017). The payoffs from cooperation are significant enough that beginning around 300,000 years ago, the ability to identify cooperative possibilities across intergroup boundaries would potentially have been a selective force favoring increased prosociality (Hames, Reference Hames2019; Wilson, Reference Wilson and Fry2013). Thus, by 300,000 years ago at the latest, humans would have been capable of intergroup tolerance, relationships across group boundaries would have at least been periodically cooperative, and these relationships would have provided access to valuable resources including stone for making tools and ochre (Pisor & Ross, Reference Pisor and Ross2021).Footnote ⁵

Peace, however, requires more than periodic cooperative intergroup exchange. It requires social structures to develop and enforce group-based norms, and prevent and resolve conflicts. Direct and circumstantial evidence in support of these prior to the last 100,000 years ago is lacking. Given what we can reasonably infer about group size and social complexity this deep in the Pleistocene, they were highly unlikely to be present. Societies at this time were likely to be small and unstratified, with few means to regulate and enforce norms against intergroup aggression and with little evidence of the types of specialization that would promote intergroup interdependence. Without these social structures in place to regulate intergroup interactions, the increased frequency of intergroup interactions during this period (300 to 100 kya) increases the likelihood that some intergroup disputes would result in violence. Without the ability to prevent and resolve conflicts, it would have been extremely difficult to turn periodic cooperative intergroup interactions into the stable harmonious relationships that characterize peace.

6.2. The potential for peace in the Late Pleistocene

Our more recent evolutionary history provides strong evidence that humans were developing material and social technologies that would have made peace more likely within the past 100,000 years. The development of new lithic techniques and specialized hunting, as well as the regular exchange of stone, shell, and ochre all during the last 100 kya (Foley & Lahr, Reference Foley and Lahr2003; Mcbrearty & Brooks, Reference Mcbrearty and Brooks2000) created the conditions for high levels of interdependence, which is a crucial means of incentivizing intergroup cooperation and preventing conflict. Between 75 and 100 kya there appears to have been a large increase in the development of complex material technologies, status symbols such as shell beads, and symbolic behaviors (Bouzouggar et al., Reference Bouzouggar, Barton, Vanhaeren, d'Errico, Collcutt, Higham and Stambouli2007; Roberts & Stewart, Reference Roberts and Stewart2018; Shipton et al., Reference Shipton, Roberts, Archer, Armitage, Bita, Blinkhorn and Boivin2018). Access to the materials and knowledge of how to produce these items would have increased the incentives for intergroup cooperation to obtain these materials and the cultural knowledge of their manufacture and meaning. The development of decorative and status items indicates that group identity and social structures were becoming important, which enables the capacity for group-enforced norms and informal leadership, both of which would have facilitated the emergence of peace.

Rather than intergroup relationships being mostly local, evidence of extremely widespread trade emerges beginning 50,000 years ago when humans in East Africa began creating beads from ostrich eggshells (Miller & Wang, Reference Miller and Wang2022). Not only were ostrich eggshell beads traded, but also a comprehensive study mapping the spread of bead patterns across eastern and southern Africa found that beads were exchanged over an area of 3,000 kilometers connecting both eastern and southern Africa (Fig. 4) lasting from 50–30,000 kya (Miller & Wang, Reference Miller and Wang2022). Even after this pan-African trade broke down, regional trade within eastern and southern Africa over vast distances persisted until present times. Wide social networks like the ostrich eggshell trade are consistent with ethnographically recent hunter–gatherers who also were embedded in extensive exchange networks spanning hundreds of miles (Bird et al., Reference Bird, Bird, Codding and Zeanah2019; Boyd & Richerson, Reference Boyd and Richerson2022) (Fig. 4).

Figure 4. Long-distance trade and networks. (A) Long-distance trade networks of ostrich eggshell beads connected eastern and southern Africa from 50 to 30 kya. Reconstructed from Figure 4c in Miller and Wang (Reference Miller and Wang2022). (B) Hunter–gatherer social organization in Western Australia where individuals are embedded in multiple levels of networks that span wide regions, including numerous language groups facilitating trade and the sharing of ritual knowledge. Courtesy of Douglas Bird.

While we cannot confidently date the beginnings of peace, circumstantially, societies would have been able to create peace when they developed social structures that promoted high levels of interdependence, group-based norms, and socially integrative mechanisms to prevent and resolve conflicts. This likely began by 100,000 years ago, when evidence of large-scale trade, cooperation, and increasing sociopolitical complexity emerges (Boyd & Richerson, Reference Boyd and Richerson2022; Miller & Wang, Reference Miller and Wang2022; Singh & Glowacki, Reference Singh and Glowacki2022), though regular intergroup cooperation likely dates back to at least several hundred thousand years ago. Once the positive benefits created through peace appeared, they would have created more selective pressure for the tolerance of strangers, affiliation across group boundaries, against reactive aggression, and cultural selection for the institutions and norms to promote conflict resolution.

In addition to intergroup cooperation, lethal intergroup conflict would have occurred at least intermittently during this period. This is supported by the fact that many recent hunter–gatherer and other small-scale groups have at least occasional warfare (Ember, Reference Ember1978; Fry & Söderberg, Reference Fry and Söderberg2013; Otterbein, Reference Otterbein1989; Wrangham & Glowacki, Reference Wrangham and Glowacki2012; Wright, Reference Wright1942). Boehm (Reference Boehm and Fry2013), for example, found that nearly half of Late-Pleistocene Appropriate foraging groups in a sample of 100 societies had lethal intergroup conflict, though he argues this is an underestimate due to inadequate ethnographic accounts.

The frequency and importance of war during the Late Pleistocene is uncertain but it would have had the potential to become intense. The presence of status items during the Late Pleistocene suggests the presence of cultural incentive systems for individuals who distinguished themselves. Cross-culturally among small-scale societies, war is the primary pathway to status for individual men, and status after age is the most important predictor of reproductive success (Hill, Reference Hill1984; von Rueden & Jaeggi, Reference von Rueden and Jaeggi2016). In the few recent small-scale societies where it has been studied, participation in small-scale intergroup war appears to be associated with success in reproductive competition. Based on this, we would expect that in addition to intergroup cooperation, lethal intergroup conflict would have been present during this period and likely had fitness consequences for aggressors..

Thus, it is reasonable to expect that when Pleistocene societies developed social structures similar to more recent small-scale groups, such as status hierarchies and social incentive systems, intergroup coalitionary aggression as well as intergroup cooperation may have been a selective factor in our species' evolution. Insofar as humans during this period resemble more recent small-scale societies, we would expect that intergroup cooperation would continue alongside intergroup conflict and that groups may have simultaneously had peace with one or more groups while also having conflict with other groups.

The timeline I have developed here is tentative and will likely be updated as new evidence emerges. I argue that by 300,000 years ago and until approximately 100,000 kya, early H. sapiens had intergroup cooperation, including trade, that was likely to have been an important part of their livelihoods. However, without evidence for cultural and social complexity, we cannot infer that the conditions for high levels of interdependence or the social structures to prohibit violence or resolve conflicts existed during this period. Thus, while intergroup cooperation occurred and may have been a selective force for increased prosociality during this period, it was likely accompanied by at least intermittent intergroup conflict. Intergroup conflicts would have been opportunistic, occasional, and low intensity, with one or two victims, as opposed to the intense tit-for-tat raids seen among many contemporary small-scale societies. Beginning sometime between 100 and 80 kya, or slightly earlier, humans developed the social structures and cultural technologies to facilitate high levels of interdependence, creating greater benefits to cooperation, and to regulate conflict through norms that prohibit aggression and can be enforced through sanctions. These social structures would have created the conditions for societies to achieve peace, but also increased the potential severity of conflict through creating group-based identities, norms that reward aggression, and enabled the organization of individuals for violence. Thus, from 100,000 years ago or so until the rise of hierarchical centralized societies, intergroup relationships likely consisted of both war and peace just as the more recent ethnographic record reflects.