Human Nature

, Volume 29, Issue 3, pp 219–244 | Cite as

Coalitional Play Fighting and the Evolution of Coalitional Intergroup Aggression

  • Michelle Scalise SugiyamaEmail author
  • Marcela Mendoza
  • Frances White
  • Lawrence Sugiyama


Dyadic play fighting occurs in many species, but only humans are known to engage in coalitional play fighting. Dyadic play fighting is hypothesized to build motor skills involved in actual dyadic fighting; thus, coalitional play fighting may build skills involved in actual coalitional fighting, operationalized as forager lethal raiding. If human psychology includes a motivational component that encourages engagement in this type of play, evidence of this play in forager societies is necessary to determine that it is not an artifact of agricultural or industrial conditions. We examine whether coalitional play fighting appears in the hunter-gatherer record and includes motor skills used in lethal raiding. Using the ethnographic record, we generated a list of motor patterns regularly used in forager warfare. Then, using Murdock’s Ethnographic Atlas, we identified 100 culture clusters containing forager societies and searched the ethnographic records of these societies for descriptions of coalitional play fighting, operationalized as contact games played in teams. Resulting games were coded for the presence of eight motor patterns regularly used in forager lethal raiding. Although play does not tend to be systematically documented in the hunter-gatherer literature, sufficiently detailed descriptions of coalitional play were found for 46 of the 100 culture clusters: all 46 exhibited coalitional play using at least one of the predicted motor patterns; 39 exhibited coalitional play using four or more of the eight predicted motor patterns. These results provide evidence that coalitional play fighting (a) occurs across a diverse range of hunter-gatherer cultures and habitats, (b) regularly recruits motor patterns used in lethal raiding, and (c) is not an artifact of agricultural or industrial life. This is a first step in a new line of research on whether human male psychology includes motivations to engage in play that develops the deployment of coordinated coalitional action involving key motor patterns used in lethal raiding.

Key Words

Coalitional play fighting Team sports Lethal raiding Warfare Play Hunter-gatherers 

Play as Adaptation Assembly

Play behavior in humans and other animals is widely regarded as an adaptation, the function of which is to develop, rehearse, and/or refine skills critical to the organism’s survival and/or reproduction later in life (e.g., Dolhinow and Bishop 1970; Fagen 1974, 1981; Lancaster 1971; Symons 1974; Van Lawick-Goodall 1967). For example, chase play may build the stamina and speed demanded by sustained, rapid locomotion, which is integral to avoiding predation (Boulton and Smith 1992; Symons 1978). It can also develop, rehearse, and test different interception and avoidance tactics at speed against an opponent, all under relatively safe conditions. The motor training hypothesis is supported by research showing that, in early development, two important effects of exercise on motor performance in mammals are modification of cerebellar synaptogenesis and modification of skeletal muscle fiber type differentiation (Byers and Walker 1995). Even spectating appears to confer training benefits: mental rehearsal, visualization, and/or watching skilled performances can enhance physical performance (e.g., Calvo-Merino et al. 2004; Cumming and Ramsey 2008; Cumming and Williams 2012; Moran et al. 2012). Similar mental training, visualization, and computer simulations appear to be useful in combat training as well (Kaufman 2004; Oskarsson et al. 2010).

Play thus appears to provide an alternate means of acquiring and testing skill/knowledge sets when there are constraints on doing so through direct experience (Bock and Johnson 2004). The most obvious of these constraints is risk of injury or death. It is much safer to practice predator evasion with a playmate, for example, than with a predator. Another constraint is opportunity: the necessary learning experience might not be encountered by the time it is needed, or it might not be encountered with sufficient frequency to allow adequate development of the skill or knowledge set(s) in question. Play circumvents this problem by enabling organisms to simulate critical features of the real-world experience under less-costly learning conditions (Tooby and Cosmides 2001). In so doing, play also provides a means of skill assessment and display, enabling individuals to gage and advertise their proficiency relative to that of their playmates (e.g., Apostolou 2015).

In humans, much play behavior can be productively understood in terms of the prolonged juvenile period, which is believed to have evolved to support acquisition of the extensive skill and knowledge sets demanded by the human ecological niche, and development of the biological structures that scaffold this task (e.g., Bjorklund and Blasi 2005; Bock 2005; Kaplan et al. 2007; Walker et al. 2002; cf. Bird and Bliege Bird 2005). Development is a broad term encompassing the growth, organization, and/or calibration of anatomical, physiological, and cognitive structures. According to this view, many adaptations are expected to have two modes, a functional mode and an organizational mode. In the functional mode, the adaptation performs its evolved function; in the organizational mode, the adaptation is assembled (Tooby and Cosmides 2001). During assembly, the adaptation acquires information inputs requisite to performance of its evolved function.

Some information relevant to assembly cannot be stored in the genome and must be acquired via the organism’s environment, in which case the organizational mode must be designed to acquire needed information inputs through experience. To this end, the organizational mode is expected to have a motivational component that guides the organism to interact with its environment in ways that advance development of the adaptation. These motivational mechanisms generate affective states that reward the organism with pleasure or excitement when it interacts with relevant aspects of the environment in ways (e.g., wrestling with a buddy, throwing rocks at targets) that provide the requisite information (Tooby and Cosmides 2001). In other words, these motivational mechanisms encourage the organism to play, to play in particular ways, and to enjoy that play. Much of what we perceive as play behavior may be the product of adaptations operating in their organizational mode.

One developmental task that likely faced our ancestors was the acquisition of skills related to coalitional intergroup aggression (CIA), or warfare, defined as an interaction in which members of one social group cooperate to pursue, capture, assault, kill or otherwise harm at least one member of another social group (Manson and Wrangham 1991). Based on the weaponry and tactics of historically documented small-scale societies, early warfare is believed to have consisted largely of lethal raiding, although pitched and running battles may also have occurred (Gat 1999). Raiding characteristically involves a group of allied males who collectively attempt to invade an outgroup territory, seek out vulnerable outgroup members, assess the probability of success, conduct a quick surprise attack, and then return to their home territory without being drawn into a battle (Wrangham 1999). Prior to the acquisition of firearms, lethal raiding among foragers would have been characterized by the use of short-range weapons (e.g., spears, clubs, slings, knives) and hand-to-hand combat.

Multiple lines of evidence suggest that warfare, largely in the form of raiding, may be an ancient feature of H. sapiens life. The archaeological record indicates that foraging peoples engaged in warfare well before contact with non-foraging peoples (e.g., Lambert 2002; Maschner and Reedy-Maschner 1998; Moss and Erlandson 1992), and thus that the emergence of warfare was not contingent on the emergence of agriculture and state societies. Further evidence comes from massacre sites such as Ofnet (~7500 bp; Frayer 1997) and Nataruk (9500–10,500 bp; Lahr et al. 2016), as well as Mesolithic cave art in Spain depicting a pitched battle (LeBlanc and Register 2003) and Australian rock paintings dating to 10,000 bp depicting humans fighting each other with spears, clubs, knives, and shields (Taçon and Chippendale 1994). Although it is impossible to determine ancient frequencies, a systematic survey found that 64% of foraging societies in the ethnographic record had combat between communities or larger groups at least every two years (Ember 1978), a likely underestimate since pacification by external powers was not taken into account (Ember and Ember 1997). Our nearest living relatives, chimpanzees, also engage in lethal raiding (Wrangham 1999). Like humans, chimpanzees live in territorial groups characterized by a fission-fusion pattern of social organization, which enables the development of intergroup hostilities and imbalances of power, the conditions hypothesized to be necessary and sufficient for warfare to evolve (Wrangham 1999). Although chimpanzees are not straightforward models for early hominids, chimpanzee raiding suggests that human warfare could date back to when our ancestors began engaging in central place foraging, the earliest evidence for which dates to approximately 2 mya (Isaac 1978). Hominids living around this time had average brain volumes that far exceeded those of modern chimpanzees (630 cc for Homo habilis and 1000 cc for Homo erectus, compared with 350–400 cc for chimpanzees). Although brain volume does not provide evidence for specific computational abilities per se, it suggests that early hominids had the processing capacity requisite to the computational demands of lethal raiding.

These converging lines of evidence have led some researchers to hypothesize that male H. sapiens evolved adaptations dedicated to engaging in CIA (McDonald et al. 2012; Tooby and Cosmides 1988, 2010; Wrangham 1999). A hypothesis that a given trait is an adaptation is a hypothesis that the trait per se was shaped by natural selection. Only natural selection (or its products) produces functional design, so demonstrating evidence of functional design constitutes the necessary and sufficient evidence that a trait is an adaptation. This requires a probability assessment of the degree to which the hypothesized trait is an efficient, reliable solution to an adaptive problem faced by ancestral members of the species (e.g., Darwin 1859; Dawkins 1986; Symons 1987, 1992; Thornhill 1997; Tooby and Cosmides 1989; Williams 1966). In other words, one must demonstrate that the hypothesized trait is so well suited to solving a given adaptive problem that the probability of its having arisen by chance alone, or as the nonfunctional by-product of another adaptation, is so small that we can reject these explanations.

With regard to warfare, evidence of adaptation would consist of psychological mechanisms improbably well suited to solving problems inherent to engaging in CIA. If such mechanisms are present in male H. sapiens, it follows that they require assembly. Among other tasks, this would include development and rehearsal of physical and cognitive skills instrumental to engaging in coalitional lethal raiding. At minimum, this would include (a) the use of offensive and defensive body movements (e.g., striking, kicking, blocking, dodging) in (b) the deployment of coordinated coalitional action against (c) an opposing coalition, while (d) monitoring, anticipating, and responding to the actions of both coalitions, and (e) continually reassessing comparative fighting formidability as both sides lose combatants (owing to defection, capture, injury, death, etc.). As part of the assembly phase, then, we would expect CIA psychology to have an emotional component motivating individuals to engage in activities that provide opportunities to develop these abilities. We posit that this emotional component is manifest as motivation to engage in play behavior that simulates critical features of raiding.

One such candidate is motivation to engage in play fighting, behavior in which the action patterns used in actual fighting, such as hitting and kicking, are modified in ways that reduce the likelihood of injury (Symons 1978). Dyadic (one-on-one) play fighting occurs in many species, including humans, and is hypothesized to be an adaptation that builds skills involved in actual dyadic fighting (Boulton and Smith 1992). Dyadic play fighting in rhesus monkeys has been shown to recruit key motor patterns used in dyadic agonistic interactions (Symons 1978) and may perform a similar function in humans (Boulton and Smith 1992). However, because dyadic play fighting involves interactions between individuals rather than coalitions, it does not simulate the key features of coalitional fighting outlined above. On this point, a striking feature of human play is that it includes coalitional play, and this coalitional play includes play fighting. Team contact sports are a case in point: these play activities combine the agonistic and coalitional components of CIA and thus may simulate features of raiding that cannot be simulated by dyadic play fighting alone.

We refer to the phenomenon of play fighting in teams as coalitional play fighting (CPF), defined as play activity in which one coalition uses coordinated action and nonlethal physical force to attain, and prevent an opposing coalition from attaining, a predetermined physical objective (i.e., “goal”). Team contact sports are illustrative: these games typically involve coordinated group action aimed at advancing or propelling a ball (or similar object) into a pre-specified zone, while thwarting an opposing coalition’s efforts to do the same. Because the ball is advanced with the body, these games enable participants to rehearse body positioning and more fine-grained movements involved in blocking, intercepting, striking, kicking, and/or throwing at moving human targets. At the same time, the chance of injury is reduced by directing the thrust of the action at the ball rather than the body per se. In many games, the ball is advanced with a modified club or stick, which may rehearse skills related to landing, parrying, and dodging blows delivered with a weapon. This is similar to the use of implements in formal martial training—such as the wooden bokken and bamboo sword (fukuro shinai) used in Kenjutsu training—to practice distance, reach, timing, and technique with less risk than that entailed in using an actual sword.

In sum, we hypothesize that human CPF is a manifestation of adaptations for CIA operating in their organizational mode. Specifically, we hypothesize that human males are motivated to engage in CPF because it provides information inputs relevant to building and rehearsing key skills involved in lethal raiding: (a) the use of offensive and defensive body positioning and movements (e.g., striking, kicking, blocking, dodging) in (b) the deployment of coordinated coalitional action against (c) an opposing coalition, while (d) anticipating, monitoring, and strategically responding to the actions of both coalitions, and (e) continually reassessing comparative fighting formidability as both sides tire and/or lose combatants. If our hypothesis is correct, one expectation is that we will find evidence of motivation to engage in CPF, and a correspondence between movement patterns used in CPF and those used in actual coalitional fighting. It is important to note that we are not arguing that the predicted movement patterns evolved specifically for warfare. Many of the motor patterns examined herein likely evolved in the context of predator evasion, hunting, and/or dyadic fighting and thus may antedate the emergence of lethal raiding. The adaptation we hypothesize is psychological: it lies in (a) an emotional system that motivates and organizes engagement in (b) behaviors that recruit specific motor patterns (c) in the context of coalitional play involving the use of coordinated action to attain, and prevent an opposing coalition from attaining, a predetermined physical goal. Our hypothesis does not require that a given game recruit all of the predicted motor patterns, because human cultures typically exhibit more than one form of CPF: collectively, we expect a society’s CPF games to rehearse many of the key motor patterns associated with lethal raiding.

Tests of predictions derived from the CPF hypothesis might include investigating whether motivation to engage in CPF is evidenced in boys and young men cross-culturally, and whether this behavior develops spontaneously (e.g., extemporaneous emergence of coalitions in snowball fights or similar activities). One might investigate the comparative popularity of sports that engage many versus few motor patterns involved in human lethal raiding, or whether CPF enhances combat unit effectiveness. However, these lines of evidence could be interpreted as the product of modern conditions and/or cultural transmission (e.g., media exposure, video games, the military-industrial complex). Although historically documented foragers are not isolated facsimiles of ancestral foragers, they share a dependence on foraging for all or a sizable part of subsistence, carried out in comparatively small, natural-fertility populations with low population densities and no reliable access to telecommunication, motorized labor-saving devices, or Western medicine (Lee and DeVore 1968; Marlowe 2005). Thus, as a first step toward testing the hypothesis that human males have motivational mechanisms dedicated to CPF, we examine whether CPF is evident in the forager ethnographic record, and whether these games engage key motor patterns involved in lethal raiding.

Design of Coalitional Play Fighting

The play-as-practice hypothesis has been challenged on the grounds that there are fundamental physical differences between play fighting and actual fighting (e.g., Loizos 1966; Stamps 1995; Sutton-Smith 1995). For example, Biben (1998) notes that serious squirrel monkey fighting is characterized by the infliction of bites, but this behavior is absent from play fighting: “we never saw biting, inhibited or otherwise, or wounding in play, making play look like a less-than-perfect model for at least this important aspect of fighting” (1998:173). This point overlooks a key criterion of play behavior: it not aimed at inflicting serious injury (Fagen 1977). On the contrary, play is characterized by the practice of self-handicapping, whereby bigger, stronger, faster, and/or more experienced individuals voluntarily inhibit or redirect the force of their actions to make the contest more even (Boulton and Smith 1992). As a result, “consequences of playful activity differ from consequences of serious performance of the same motor patterns” (Fagen 1977:395). In accordance with this logic, we would not expect the actions used in play fighting to exactly mirror those used in actual fighting; rather, we would expect these actions to be modified in ways that reduce the likelihood of lethal or disabling injury. For example, in CPF, we would expect blows to be aimed at a proxy (e.g., a ball) and/or less vulnerable parts of the body in order to reduce the chance of injury. By the same logic, we might expect to find games in which a projectile is driven along the ground because strikes aimed at ground level reduce the chances of injury to the head or internal organs, while providing safe practice for striking a downed opponent in actual fighting. In short, arguing that play fighting is not practice for actual fighting because participants modify or inhibit actions likely to injure their opponent is akin to arguing that Karate, Krav Maga, Pankration, Silat, Jiu Jitsu, and Laamb wrestling are not practice for fighting because biting, eye gouging, throat striking, and fish hooking are not carried out in sparring.

Although skills developed in dyadic play fighting can be deployed in coalitional fighting, dyadic play fighting does not rehearse, and thus cannot develop, all of the skills demanded by CIA. Coalitional fighting imposes attentional, coordination, strategic, and evaluative demands that dyadic fighting does not. Chief of these is “orchestrating one’s behavior so that it meshes simultaneously with that of several others” in order to thwart the orchestrated actions of an opposing coalition (Tooby and Cosmides 1988:3). Coalitional play that requires passing, receiving, and interception (e.g., lacrosse, rugby, soccer, hockey) is more likely to engage these capacities than dyadic play fighting because it simulates the complex coordinated action and the main obstacle involved in raiding—namely, resistance by a comparably armed human coalition. This observation is echoed by sports commentators, who note that in modern team contact sports, individual performance is not sufficient to produce wins because coordinated action is critical to success: “With few exceptions, it doesn’t matter which team has the best player, the final result hinges on the entire team. . . . No matter your talent level in team sport, you must rely on your teammates. . . . Without a group of players performing simultaneously complex motions pointed toward the same goal, the team performance will falter” ( Dyadic play fighting does not require an individual to track, anticipate, assist, or impede the goals and aggressive actions of multiple human agents in two opposed groups comprising individuals with different skills and attributes, and it is thus unlikely to develop those skills specific to cooperative fighting or assessment of group-fighting formidability and commitment. On this point, although the Roman legions practiced individual and dyadic weapons training, as do all professional armies, they also incessantly drilled coordinated tactical maneuvers, upon which their fighting success was dependent (Webster 1998).

Like dyadic play fighting, play hunting is unlikely to develop all the skills critical to engaging in CIA. For example, although “games of skill” (Roberts and Sutton-Smith 1962) that involve shooting at targets are common play activities among hunter-gatherer children (e.g., hoop-and-pole; Culin 1907), these games do not involve coordinated action against an opposing coalition. On this point, there are hints in the ethnographic record that achieving hunting and warfare prowess later in life (Garfield et al. 2016; MacDonald 2007) required differential training in youth (Mendoza 1985). Among the Tahltan, for example, boys actively pursued their preferences: “most lads desiring to be good hunters prayed to become such and practiced the real hunting of game,” while “lads who desired to be good warriors prayed to be such and practiced mimic warfare” (Teit 1956:137–38). Thus, although hunting play provides experience in the use of weapons, and may develop strength and stamina as stalking and warfare games do (MacFarlan and Macfarlan 1958), it does not fully simulate the tasks involved in CIA. One such task is anticipating the moves of human opponents; the vast majority of prey species are incapable of mounting the strategic, coordinated group maneuvers exhibited by humans (e.g., fortification, encirclement, treachery). This point is underscored in an eighteenth-century description of a Reindeer Chukchi game played by adult males using real weapons. Participants draw the bow but “do not release the arrows from the hand, and then they take a spear, making the same try as would be necessary in a battle. . . . The primary difference from battle lies in the fact that the bow is not shot and the spear not thrust” (Nefëdkin 2014:29). The observer adds that, through this game, “the reaction of the warrior to the action of an opponent was perfected. The sequence of exercises is purely combat” (Nefëdkin 2014:29).

Chase games, too, are common among foragers but tend to exhibit a many-vs.-one or one-vs.-many rather than a coalition-vs.-coalition structure. For example, the Netsilingmiut played a game in which one person was a “caribou” and the others were “wolves”: the first “wolf” to touch the “caribou’s” bare skin became the next “caribou” (Rasmussen 1931). Kiowa children played a game in which one person pretended to be a bear and chased the other players, who were pretending to pick berries near the bear’s den (Parsons 1929). Thus, chase games appear to simulate pursuit by two different types of animal threats: social carnivores and solitary predators. A third threat potentially simulated by these games is charges by large game animals. For example, the Netsilingmiut played a game in which one player, pretending to be a musk oxen, chased the others and tried to “horn” (gore) them (Rasmussen 1931). Such games do not mimic the coalitional structure of CIA.

Moreover, their explicit references to animals suggest that many chase games simulate—and rehearse responses to—the offensive, defensive, and evasive behaviors of nonhuman rather than human agents. This is an important distinction because animal fighting tactics and weaponry are considerably different from those deployed by humans. On this point, it may be significant that in some chase games, players imitate key behavioral and/or physical traits of the animal they are pretending to be, as in the musk oxen example above. Another example is a Netsilingmiut game in which one player pretending to be a bear crawls on all fours while chasing the others (Rasmussen 1931). Such simulations may encourage participants to develop offensive, defensive, and/or evasive strategies that are appropriate to the animal in question. In any case, although they provide experience with pursuit and evasion, chase games do not simulate other important features of coalitional human attack. With their highly developed ability to innovate (Tooby and DeVore 1987; Barrett et al. 2007), humans are better able to improvise tactics on the fly and tailor them to immediate constraints and opportunities. Their capacities for innovation, coordinated action, and division of labor make human groups much less predictable in aggressive contexts than animal targets and pose problems that require dedicated mechanisms for their solution (Duntley 2005; Tooby and Cosmides 1988). These problems are not simulated by play hunting or dyadic play fighting, but they are simulated by CPF (e.g., the use of strategic plays in football, hockey, or lacrosse). The converse is also true: CPF is not well suited to simulating predator evasion or hunting. Group hunting may appear to be an exception in that, like CIA, it involves the use of coordinated aggressive action. However, group hunting typically involves a small party of humans targeting a single nonhuman agent, and while game drives target multiple agents, they typically exploit nonhuman animal herd behavior. CPF does not reflect these aspects of group hunting.

We therefore posit that CPF provides a means by which, in an open-field context, individuals can learn, practice, and refine the coordinated deployment of fighting skills with those of coalition members while thwarting the fighting maneuvers of an opposing coalition. Periodic participation in such games during childhood, adolescence, and early to middle adulthood provides individuals with opportunities to viscerally assess the aggressive formidability and commitment of their own and—when played with neighboring groups—other coalitions as their composition and skills change through time. Thus, we further hypothesize that team play fighting provides practice at making such assessments with accuracy and rapidity. (The display and assessment components of intergroup CPF will be discussed in a future publication.)


To test the prediction that motivation to engage in CPF occurs across diverse hunter-gatherer populations, and regularly recruits key motor patterns used in forager lethal raiding, we first searched the ethnographic record for detailed qualitative descriptions of forager warfare from all continents for which such information was available (Asia, Australia, North America, South America). We used these qualitative descriptions to deduce motor patterns requisite to the deployment of, and avoidance of injury by, the weaponry and tactics referenced therein. We then surveyed a sample of forager societies for the presence of team contact games and coded these games for the presence of the motor patterns deployed in forager CIA.


To assemble the study sample, we searched Murdock’s (1967) Ethnographic Atlas for forager societies, operationalized as a combined score of 7 or more for the categories of dependence on gathering, hunting, and/or fishing (as measured in Table A, Column 7 of the Atlas). This resulted in the inclusion of some forager-horticulturalist and forager-pastoralist societies but was deemed necessary because a stricter criterion of 100% dependence on gathering, hunting, and/or fishing would have greatly reduced the sample size. The search yielded 235 culture groups distributed across 100 culture clusters in five of Murdock’s six geographical regions (there are no forager groups in the Circum-Mediterranean region by our criteria). Mounted foragers were included in the sample because the acquisition of horses did not substantially alter the nature of forager warfare, in that it continued to take the form of lethal raiding. Moreover, some groups continued to raid on foot even though they possessed horses (Goodwin 1971). Additionally, references in oral tradition to attacks by Oroks and Chukchi mounted on reindeer (Batchelor 1926; Bogoras 1918) indicate that mounted attacks may have occurred in some forager groups before the acquisition of the horse. Ethnographically documented forager societies are not uniformly distributed across continents; as a result, the representation of these cultures in the Ethnographic Atlas is biased by geographical region, with 47 culture clusters from North America, 23 from South America, 12 from East Eurasia, 11 from the Insular Pacific, and 7 from Africa, and this bias carries over into our study sample.

Motor Patterns in Warfare

With the aim of elucidating design, we examined detailed descriptions of forager lethal raiding to generate a list of motor patterns that we predicted would be present in CPF. These motor patterns were inferred in large part from the weaponry used. An example is seen in Burch’s account of Iñupiat warfare: “At a distance bows and arrows were obviously the weapons of choice. At close quarters . . . thrusting spears might have been more useful . . . while in hand-to-hand combat knives and clubs might have been more effective than any other weapon” (2005:85). Similarly, among the Murngin, “The wooden or stone-headed spear, the spear-thrower and the club, as well as the stone knife, are used in these [combat] engagements. The spear is the chief weapon. . . . The stone axe . . . also serves as a weapon” (Warner 1931:458). Slings and bolas were also used in warfare (e.g., Haenke 1943). For example, in the eleventh century, a Norse expedition led by Thorfinn Karlsefni was attacked by “Skrellings” on the east coast of North America. According to the sagas, “their major weapons were large ball-shaped objects, resembling the stomachs of sheep, which were flung from poles” (Oswalt 1999:12). It is unclear whether the “Skrellings” were Inuit or Algonkians, but both used sling weapons. The Algonkian type was made by “sewing a large stone inside a fresh skin that tightened about the stone as it dried” and was “attached to a pole and hurled as a weapon” (Oswalt 1999:12). The Inuit type is described in an account of a Danish expedition attacked in 1605 while anchored off northwestern Greenland: the party of thirty men in kayaks “paddled ashore to find stones that they hurled at the vessel with slings” (Oswalt 1999:43). Slings were also used in combat by peoples of the Gran Chaco region (Hutchinson 1865:326). Accounts also reference dodging, as seen in a description of Ona warfare: “While they fought, the Ona did not stand still for a moment, but jumped continuously to the right and to the left to avoid presenting himself as an easy target to the enemy” (De Agostini 1956).

With regard to motor patterns, the use of slings indicates throwing and implies dodging, and the use of spears and clubs indicates throwing, dodging, striking, and parrying. Spears and clubs are used both as hand weapons and projectiles. For example, Maguire reports that Dorobo use the club “indiscriminately as a throwing or hitting weapon” and “are excellent shots with it” (1928:256). Spear-throwers, too, may be deployed as weapons, as seen among the Murngin: “No shield is found here. The Murngin depend on the spear-thrower to ward off spears, and also on their well-developed agility to avoid being hit” (Warner 1931:458). Similar defensive tactics (dodging, parrying, catching) are documented among the Chukchi: “When he is shot at, he avoids the arrows by springing to one side, or parries them all with the butt-end of his spear, or simply catches them between the fingers” (Bogoras 1904–1909:646; for arrow-catching, see also Abreu de Galindo 1767:33). In hand-to-hand combat, grappling comes into play. Among the Ona, for example, “the men shoot first from ambush, then in the open, and finally close empty-handed, the object being to break the opponent’s back or neck by wrestling” (Barclay 1904:73, our emphasis). Although grappling includes several motor patterns (e.g., blocking, tackling, pinning, clinching, choking), in most cases it was impossible to tease out different movements from the descriptions of lethal raiding or play; accordingly, these movements were collapsed into one category. Although not explicitly referenced in descriptions of lethal raiding, kicking was inferred from the common use of sweeping in fighting, the use of strikes with the legs/knees/feet in many martial arts, and the widespread use of kicking in assaults (a Google search for “assault kicking” returned ~3,360,000 hits in .5 seconds). For example, 20% of head injuries in UK accident and emergency patients were the result of assault (Kay and Teasdale 2001), with the head a “target of choice” for kicking and trampling in homicidal attacks (Henn and Lignitz 2004). Running was inferred from the fact that lethal raiding is characterized by brief, surprise attack followed by immediate, rapid retreat to the aggressor’s home territory (Wrangham 1999), and because raiding often involves pursuit and evasion. Among the indigenous nations of Alaska, for example, “unless the battle had gone on for a very long time, they [the vanquished party] were usually hotly pursued by the victors” (Burch 2005:108).

Our survey of forager lethal raiding yielded eight motor pattern categories (ESM Table 1): catching, dodging, grappling, kicking, parrying, running, striking, and throwing. These motor patterns formed the basis of the ethogram used for coding (reproduced in the ESM). The inclusion of catching may appear to be at odds with the claim that CPF serves as practice for CIA: other than scattered references to arrow-catching, this motor pattern is rare in descriptions of lethal raiding. However, catching may provide practice for dodging, in that both movements require calculating the trajectory of a projectile and quickly positioning the body in relation to it. In terms of position, the chief difference between catching and dodging is whether the body is placed in line with the trajectory or immediately outside of it. Thus, catching games offer a less risky means of developing skills related to dodging than dodging games do, which is what we would expect if play actions are modified to reduce the likelihood of serious injury (Fagen 1977; Burghardt 2004).

Forager CPF

We then searched the ethnographic record for descriptions of team contact games played by forager societies in each of the 100 culture clusters from our sample, using materials available through the University of Oregon Knight Library (including WorldCat). We searched for descriptions documented at a time when traditional ways of life were still practiced or practiced within living memory. The survey included sources in English, Spanish, Portuguese, Italian, French, and German. When ethnographic sources included mention of team games, the chapter or section was saved as a pdf for later coding.

Owing to the paucity of rich, detailed descriptions of forager games in the ethnographic record, statistical sampling of societies was not feasible. Such an approach would have resulted in the exclusion of societies for which such descriptions were available and/or the inclusion of societies for which they were not. A further problem was that, when play information was lacking for a given culture, it was impossible to determine whether this was because of the behavior being absent or simply not being recorded by the ethnographer.

The study sample was analyzed by culture cluster rather than individual cultures because, in some cases, ethnographers reported on play behavior for a general region (e.g., “Aborigines of Victoria,” “Bering Strait Eskimo”) rather than a specific culture group. In most cases, information about team games was found for only one culture within a given culture cluster. However, when we found sufficiently detailed play descriptions for two or more cultures within a cluster, all were copied for coding, on the logic that such information was too rare to be overlooked. When this occurred, it was often because the informant or ethnographer mentioned that a given game was played with or by neighboring groups. For example, in his description of hockey as played in the Gran Chaco region, Nordenskiöld notes that he “saw it among the Choroti, Asluslay and Mataco” (1910:430–32).

Because the play behavior examined in this study is hypothesized to be generated by adaptations dedicated to male coalitional aggression, we limited our survey to male team play, defined as play in which two or more male players engaged another group of two or more male players. This eliminated “team” sports such as wrestling matches, in which only one player from each side competes at a time and which is thus more akin to dyadic play. In the handful of descriptions where the sex of the players was not specified, we assumed male participation owing to a pronounced tendency on the part of ethnographers to identify games played by women only. For example, Culin notes that “the game of double-ball throughout the eastern United States and among the Plains tribes is played exclusively by women, and is commonly known as the woman’s ball game” (1907:647). This pattern of calling attention to “women’s games” indicated a default assumption on the part of the (usually male) investigator that, unless otherwise indicated, male activity was being reported. We included mixed-sex team games in our sample because they were rare and because our prediction that males will engage in these games is neutral with respect to females.

Motor Pattern Coding

We coded descriptions of team games for motor patterns rather than game types because physical play is composed of motor patterns and its hypothesized adaptive function is to train motor and associated cognitive skills. As Burghardt notes, “Repetition of patterns of movement is found in all play and games in human and nonhuman animals” (2004:234). We therefore expect to find design at the level of the movement patterns used in the context of CPF. In other words, we would expect selection to produce motivations to engage in specific motor activities (e.g., throwing or dodging projectiles) in a competitive coalitional context, not for engaging in specific games per se. We took as our model Symons’s (1978) study, which breaks down rhesus play fighting into a sequence of distinct motor patterns. Symons argues that rhesus play fighting exhibits design, in that “each monkey attempts simultaneously to bite without being bitten. Thus a playfighting monkey succeeds in achieving its goal to the extent that its partner fails” (Symons 1978:87). This same design structure is inherent in team play fighting, but instead of two individual opponents there are two sets of opponents. Like dyadic play fighting, then, CPF is a structured activity: each team uses coordinated action and nonlethal physical force to attain, and prevent an opposing team from attaining, a predetermined physical objective.

Since our survey focused on motor patterns used across team game types, all CPF games documented within a given culture cluster were collapsed into a single category. The game descriptions collected for each cluster were then surveyed for the presence of the predicted motor patterns. For example, Hoffman’s (1890:134) description of Eastern Ojibwa lacrosse evinced the presence of the following motor patterns:
  • Catching: “The ball is tossed into the air in the center of the field. As soon as it descends it is caught with the ball stick by one of the players”

  • Running: “he immediately sets out at full speed towards the opposite goal”

  • Throwing: “If too closely pursued, or if intercepted by an opponent, he throws the ball in the direction of one of his own side, who takes up the race”

  • Striking: The usual method of depriving a player of the ball is to strike the handle of the ball stick so as to dislodge the ball; but this is frequently a difficult matter on account of a peculiar horizontal motion of the ball stick maintained by the runner. Frequently the ball carrier is disabled by being struck across the arm or leg, thus compelling his retirement”

For games identified by their Western name only (e.g., hockey, rounders, handball), we followed Deaner and Smith (2013:8) and coded them as their modern equivalents, reasoning that the investigator considered the label a sufficient description. In such cases, we assumed that motor patterns used in the indigenous game were the same as those used in the Western version, and coded the game accordingly (see the ESM). For example, the observation that the Omaha played “shinny in the fall” (Culin 1907:781) was coded for the same motor patterns involved in Western shinny. An exception to this rule was “football,” which can refer to soccer, American football, or rugby. When it was impossible to determine which form of the game the investigator was referencing, “football” was coded only for running. Variants of hockey were coded for both striking and parrying, on the grounds that the player in possession of the projectile must block strikes made by other players trying to steal it (i.e., stick checking). This is seen in a description of Sioux shinny: “the club may be used in any manner to make a play [coded as striking], or to prevent an opponent from making a play [coded as parrying]” (Walker 1905:285). Because hockey allows body, shoulder, and hip checking, we also coded hockey and its variants for grappling. Games that do not typically involve grappling in modern societies were coded for the presence of this motor pattern only if it was explicitly mentioned, as in a description of Micmac soccer: “a player may catch his opponent by the neck and thus hold him back until he can obtain the ball himself” (Hager 1895:35–36). Because many games prohibit players from using the hands to restrain or take down opponents, much of the action coded as grappling consisted of checking—driving the shoulder, arm, elbow, or hip into one’s opponent.

Interestingly, one of the team games we encountered in the ethnographic record was mock warfare. Because this activity is an obvious instance of CPF, we included it in our analysis. These games typically involve the throwing and dodging of projectiles, such as rocks and sticks, and are similar to dodgeball in that action is expressly directed at the body. In North Western Australia, for example, players “pick sides and throw blunted miniature spears at each other. They can either dodge these spears or turn them off with a small stick held in the left hand, and they seem to enjoy this sport immensely” (Withnell 1901:29). Similarly, Bambuti children “frequently have pitched battles, which open with mutual taunts and end with cudgeling and mud-slinging” (Schebesta 1933:60). Many of these games involve the hurling of projectiles using sticks and thus underscore the motor pattern parallels between lethal raiding and stick-based throwing games such as lacrosse. For example, in the “mud-and-willow” game of the Santee Sioux, “a lump of soft clay was stuck on the end of a limber and springy willow wand and thrown as boys throw apples from sticks, with considerable force” (Eastman 1902:55–56). The reference to apple-throwing suggests that mock warfare games may be more widespread across human societies than indicated by our study, which focused exclusively on foraging peoples. For example, in a survey of indigenous South American games and amusements, Cooper (1949:511) reports that “mock battles of various kinds” were played in the Andean, Eastern Bolivian, and Eastern Brazilian regions, and snowball fights and similar activities are common among children in the United States.

All activities coded as play met Burghardt’s (2004:234) five criteria for identifying play behavior: (1) performance of the behavior does not aid immediate survival in the form or context in which it is expressed; (2) the behavior is voluntary, intentional, and/or pleasurable; (3) compared with “serious” performance of ethotypic behavior, play behavior is incomplete, exaggerated, awkward, and/or involves special signals; (4) it occurs repeatedly in a similar form during at least a portion of the animal’s ontogeny; and (5) it is initiated during leisure—when the animals are adequately fed, healthy, and free from stress (e.g., predator threat, social instability, inclement climatic conditions, feeding or mating competition).

The first author and an undergraduate assistant blind to the research hypothesis independently coded ethnographic materials for presence of predicted motor patterns. Motor patterns were coded for presence rather than frequency: if a motor pattern was present in multiple games within a given culture cluster, it was coded only once. Cohen’s κ calculated using SAS v. 9.49 shows agreement significantly above chance between the two independent raters’ judgements overall (κ =.825, p<.0001) and for all eight motor patterns with: almost perfect agreement between raters for strike (κ =.934, p<.0001) and parry (κ =.826, p<.0001); substantial agreement for catch (κ =.784, p<.0001), kick (κ =.771, p<.0001), throw (κ =.758, p<.0001), grapple (κ =.75, p<.0001), and dodge (κ =.727, p<.0001); and moderate agreement for run (κ =.427, p=.0031) (Landis and Koch 1977). There was also substantial agreement for the one type of game coded independently, mock warfare (κ =.718, p<.0001). Running may have shown less agreement than other categories because it was often implied rather than explicitly stated in descriptions of team contact games.


Despite the fact that play (or its absence) was not commonly or extensively documented by early ethnographers, team game information was found for 46 of the 100 culture clusters, across all five geographic regions in the sample (Africa, East Eurasia, Insular Pacific, North America, and South America; ESM Table 1). Interestingly, mock warfare was found in 39% of culture clusters and boys’ mock warfare in 26%, the latter suggesting that motivation to engage in CPF emerges in childhood. As seen in ESM Table 2, games using sticks to hit objects (and sometimes people) were the most common game type (76%), followed by games involving kicking (41.2%) and team handball (32.6%).

All eight of the predicted motor patterns were present in multiple culture clusters, although some were more widespread than others (ESM Table 1). As noted above, our prediction is not that every type of coalitional fighting game engages all eight predicted motor patterns. Rather, we predicted that, cross-culturally, forager CPF games regularly engage different subsets of these eight motor patterns. As seen in ESM Table 1, the majority of culture clusters for which information regarding team games was available evince participation in team games that, collectively, recruit several of the predicted motor patterns. ESM Table 2 shows that most forager cultures engage in more than one type of CPF game, with multiple CPF games present in 58.7% of culture clusters.

Running was the most common of the predicted motor patterns, being present in 39 (85%) of the culture clusters. It is also arguably the least diagnostic because running ability is critical in other aspects of forager life, such as predator evasion and hunting. Thus, the motivation to engage in running-based play activity might be driven by multiple fitness advantages. However, this explanation does not satisfactorily account for running play involving two opposing coalitions. Chase play more closely simulates the dynamics of predator evasion than CPF does, and the degree to which forager hunting depends on running is unclear. Endurance running is hypothesized to be a derived capability of our genus that may have been used in the context of hunting or scavenging (Carrier 1984). However, endurance running is not common in modern forager hunting (Bramble and Lieberman 2004:351): although persistence hunting (running game to exhaustion) has been documented in some hunter-gatherer groups, it does not appear to be a widespread or frequently used tactic (Liebenberg 2006). Forager hunting techniques rely more on walking and stalking than running down prey (e.g., Raichlen et al. 2016). Even game drives do not tend to involve sustained, high-speed running (e.g., McClellan 1987:119–22; Turnbull 1983; cf. Anell 1969). Thus, although “it is reasonable to hypothesize that Homo evolved to travel long distances by both walking and running” (Bramble and Lieberman 2004:351), the demands of hunting may not adequately account for human running abilities. In any case, running, often in a coalitional context, is an inherent part of lethal raiding, which typically involves a sudden strike followed by immediate retreat. Iñupiat warfare is illustrative: “the overwhelming majority of armed confrontations were raids on single settlements, after which the surviving invaders headed for home as fast as they could go” (Burch 2005:69). Fleeing enemies were often pursued; in such circumstances, stamina as well as speed was critical, as evinced by Helena Valero’s account of a Yanoáma (Yąnomamö) attack in which she and her fellow villagers were pursued by enemy warriors for nearly twenty-four hours (Biocca 1970:31–37). Running also appears to have been critical to mustering allies or securing refuge in a timely fashion: “when raiders were discovered to be on the way, residents had to travel several miles to the nearest village to escape or get help” (Burch 2005:69).

Games involving striking were present in 37 (80%) culture clusters. Since early hunting may have involved the use of clubs, the motivation to engage in striking play may be driven by more than one set of selective forces. But hunting doesn’t explain play in which two coalitions compete to strike an object: although foragers often hunt in groups, they don’t compete with other teams when doing so, nor do prey animals defend themselves in teams. Games involving parrying were present in 35 (76%) of the culture clusters. Parrying occurs in most games that involve striking, with the exception of rounders. Throwing games that use sticks also involve parrying in the course of blocking another player’s throw, as seen in a description of Chippewa lacrosse: “when one is on the point of hurling it [the ball] to a great distance, an antagonist overtakes him, and by a sudden stroke dashes down the ball” (Carver 1796). Parrying provides some of the strongest evidence that CPF subserves the development of skills related to lethal raiding rather than hunting: only humans parry with handheld implements. Indeed, parry fractures are one of the clear osteological signatures of interpersonal violence, including warfare (Lambert 2007). Games involving grappling were present in 33 (72%) of the culture clusters. As with parrying, the presence of grappling is compelling evidence that CPF develops skills related to lethal raiding rather than hunting. As the use of projectile weapons attests, hunter-gatherers prefer to kill animals from a distance because doing so reduces the chances of injury. Hence, grappling does not normally occur in the course of hunting. In contrast, once a person’s supply of projectile weapons is spent, lethal raiding comes down to hand-to-hand combat using clubs, knives, and/or the body, at which point effective grappling skills are critical.

Games involving throwing were present in 31 (67%) of the culture clusters. Like striking play, throwing play may also be driven by adaptations related to hunting. However, throwing play in which two teams compete to drive a projectile into enemy territory better simulates the maneuvers and calculations demanded by lethal raiding rather than those required by group hunting. Catching was somewhat less common than throwing, being present in 24 (52%) of the culture clusters. This may be explained by the fact that some throwing games in the sample involved dodging rather than catching. For example, the Netsilingmiut played a game called “Pretending to Kill Another” in which players throw stones at each other and try to dodge them (Rasmussen 1931). Games involving kicking were present in 18 (39%) of the culture clusters. Like grappling, kicking is instrumental in hand-to-hand combat, and more diagnostic of fighting than hunting: hunters don’t kick animals to death, or take them down by sweeping. Games involving dodging were present in 16 (35%) of the culture clusters.

Limitations of This Study

This study is limited by the sparse nature of the early ethnographic record on forager coalitional games. Geographically, evidence of CPF is heavily biased toward North America, but this reflects the Ethographic Atlas forager sample as a whole. Whether or not CPF was present in culture clusters for which data are lacking remains unknown. Also unknown is the degree to which lack of data is due to the behavior not being present or simply not having been recorded by early ethnographers. Further, except for boys’ mock warfare, almost all reports are for formally recognized games rather than spontaneously emerging coalitional contact play. Our finding that 46% of the culture clusters in our sample show evidence of CPF is thus likely an underestimate of the pervasiveness of this behavior. On this point, rough and tumble play is expected to be ubiquitous in humans, yet according to the Atlas’s own coding, only 66 of the 100 clusters included in our survey show evidence of games of physical skill of any type. Significantly, we found evidence of CPF in 43 (74%) of those 66 clusters, and evidence of CPF in four clusters for which the presence of games of physical skill was coded in the Atlas as unknown.

If CPF is the product of adaptations for CIA operating in their organizational mode, we might expect to see CPF in other animals that exhibit CIA (e.g., chimpanzees, wolves, hyenas, dolphins), yet a survey of the play literature yielded no evidence of CPF in these species. There are a number of possible explanations for this, the first being that absence of evidence is not evidence of absence: CPF may occur despite not having been observed by investigators. Given the extensive behavioral observations that have been made of these species, however, this seems unlikely, although it may be that CPF has not been recognized as such. Secondly, the expectation that CPF will occur in all species that exhibit CIA assumes that the benefit of engaging in CPF will outweigh the costs across species, which may not be the case. Indeed, the cost/benefit ratio of engaging in CPF is likely to vary considerably across human groups. Besides the fitness costs and benefits of CIA itself, other variables (e.g., life history, group size, demographic structure, diet, habitat, range size) are expected to affect the costs and benefits of engaging in CPF. For example, population structure might not allow, or might reduce opportunities for, coalitional play: a group might lack sufficient individuals of the same age/sex class to support play in teams.

Another possibility is that, in some species, limits imposed by diet and feeding behavior may make an energetically taxing activity such as CPF more costly than any fitness benefits derived from it. Compared with the human diet, for example, the chimpanzee diet is much less nutritionally dense and requires a higher volume of food intake to supply adequate energy and nutrients (Leonard et al. 2007). Moreover, in humans, post-weaning parental provisioning greatly increases juvenile survival rates relative to nonhuman primates and social carnivores: in chimpanzees, only about one in three juveniles survives adolescence, and in lions and wolves, survival rates drop to one in five; in contrast, human foragers successfully rear one out of two offspring to adulthood (Lancaster and Lancaster 1983). Similarly, juvenile survival rates are higher—nearly double—in primate groups regularly provisioned by humans than in wild, self-feeding groups (Lancaster and Lancaster 1983:39). In addition, the extensive use of skilled extraction techniques by humans provides reliable access to large resource packages that are pooled across the foraging group, adding further energy savings (Kaplan et al. 2007). Notably, forager intergroup CPF often occurs during festivals or trade fairs—times when work demands are low and resources are abundant (e.g., Wells Jr and Kelly 1890:25; Rand and Webster 1894:200; Stern 1934:55).

Diet plays a highly consequential role in the expression of play behavior because of its effects on the organism’s energy budget. Play is defined, in part, as behavior initiated during leisure—when the animal is adequately fed, healthy, and free from stressors such as predation threat (Burghart 2004:234). In other words, play occurs when the animal’s time and energy are not being allocated to more pressing tasks. The enabling effect of leisure vis-à-vis play is illustrated in a study that found extensive tool use in a range of captive species (e.g., birds, ungulates, monkeys) not known to use tools in the wild (Beck 1980). Beck attributes this novel behavior to leisure: because the animals were not burdened with tasks such as searching for food or fending off predators, their time and energy could be invested in exploration and manipulation of their environment. In short, a life history approach suggests that ecological constraints on an animal’s energy budget are highly likely to influence the nature and frequency of play and should be taken into consideration in cross-species comparisons.

On this point, synchronous motor activities in adult bottlenose dolphins may be instructive. Males in this species are known to form first-order and second-order alliances: the former (dyads and triads) cooperate to guard and herd individual females, whereas the latter (coalitions comprising two or more first-order alliances) cooperate in conflicts with other alliances over females (Connor et al. 2010). Interestingly, synchronous surfacing, aerial leaping, and underwater turning have been documented in male dyads and (less commonly) triads, both between alliance members and between members of different alliances (Connor et al. 2006). These synchronized activities have been observed when alliances are alone, when they are affiliating with another alliance, and when female consorts are present in one or both alliances (Connor et al. 2006). Connor and his colleagues posit that this behavior serves a signaling function within or between male alliances (Connor et al. 1992). However, it might also be explained as play behavior whose function is to develop the coordinated action utilized in the guarding, herding, and stealing of females, with signaling as a related functional component. Although it is unclear whether this behavior is CPF, synchronized movement is critical to coordinating action, and thus a critical component of CPF. Connor’s findings suggest that, with regard to CPF, synchronized motor activities performed during leisure among allied individuals may be a productive line of inquiry in species that exhibit CIA.

Besides the adaptation assembly hypothesis, a number of other explanations for the occurrence of CPF in humans are possible. For example, it could be argued that CPF games function as alternatives to CIA. It may well be that some CPF play is initiated in this spirit, either consciously or unconsciously, but this hypothesis is challenged by the playing of mock warfare games by young boys, who do not customarily engage in warfare (ESM Table 1). Furthermore, if CPF were a substitute for intergroup aggression, we would expect to find a negative correlation between frequency of warfare and participation in contact sports. However, research shows the opposite: “where we find warlike behavior we typically find combative sports and where war is relatively rare combative sports tend to be absent” (Sipes 1973:71; see also Chick et al. 1997). Indeed, Chick et al. (1997) found that frequency of external warfare correlates strongly with the presence of sham combat games. These results are what we would expect if CPF is preparation for warfare.

A purely costly signaling explanation for CPF is problematic as well. We don’t deny that CPF recruits some motor patterns used in hunting, predator evasion, and dyadic fighting, or that it provides an opportunity for individuals to display qualities related to these endeavors. But if the point is to display individual qualities, why play in teams? Individual performance is likely to be obscured by the presence of other players on the field. Racing, target practice, and wrestling are all common amusements in forager societies (Culin 1907), provide ample opportunity to display hunting and dyadic fighting ability, and offer an unobstructed view of each performer. Another possibility is that CPF is costly signaling of physical and cognitive abilities related to CIA. However, this is only a partial explanation: it presumes that these abilities are valuable but overlooks the source of that value—namely, their instrumentality to success in warfare. A more parsimonious explanation is that motivation to engage in CPF improves abilities important for CIA and, in so doing, provides a context in which these abilities can be effectively signaled.

Cultural transmission is also insufficient to account for CPF. Although particular forms of this behavior are often transmitted through social learning, a content-free social learning hypothesis for CPF takes us back to the question of motivation and appeal: why—despite differences in language, cultural practices, and habitat—have so many societies invented or adopted this behavior? Such widespread imitation of these types of behaviors is not an explanation but a phenomenon to be explained. On this point, from the limited ethnographic information available, it was impossible to determine whether the games we examined were independently invented or acquired through cultural borrowing. This is unfortunate because, if motivation to engage in CPF is part of the CIA assembly process, we would expect CPF to emerge spontaneously among groups of human males. The presence of informal mock warfare play among young boys in diverse forager cultures (ESM Table 1) hints that this may indeed be the case—especially when it occurs in famously peaceful societies such as the Bambuti (Schebesta 1933:60). Additionally, although team contact play could take any number of forms (e.g., team boulder carrying, team nose-touching), cross-culturally this type of play converges on a particular constellation of features: the use of offensive and defensive body positioning and movements in the deployment of coordinated coalitional action against an opposing coalition. Finally, regardless of how they emerge, these play behaviors must be sufficiently appealing to be maintained from generation to generation. Thus, the presence of voluntary and enthusiastic CPF across diverse cultures attests to a powerful human motivation to engage in this activity.

Directions for Future Research

Although we have focused here on motor patterns deployed in coalitional contact games, other aspects of CPF warrant investigation as well. First, further cross-cultural research on the spontaneous emergence of coalitional play in childhood is needed. Another consideration is age distribution: if CPF builds, rehearses, and calibrates motor systems and cognitive circuitry involved in lethal raiding, we would expect participation in CPF to begin in childhood and persist into the prime of life in order to provide practice against opponents who have reached peak adult strength and stamina. Although play is widely hypothesized to serve a developmental function, and is commonly observed in juveniles (e.g., Dolhinow 1999; Fagen 1981), it also occurs in adulthood (Cordoni 2009; Logan and Longino 2013; Pellis and Iwaniuk 1999). This suggests that play might serve other functions besides development and/or that some developmental processes might continue into adulthood. In the case of CPF, adult participation is expected because of the need to keep skills sharp and the fluidity of coalition membership, which varies for several reasons. Participation in a given raid is typically optional, and men may opt out of attacks directed at villages in which they have close social ties (e.g., Burch 2005; Chagnon 1997; Goodwin 1971). Coalition composition also varies as a result of attrition when members move away, are injured, die, or become too old to fight. Conversely, new coalition members are periodically gained as juvenile males reach adulthood and new males marry into the group. Thus, individuals are regularly presented with a different constellation of persons with whom they must orchestrate their actions, and a different constellation of persons against whom they must orchestrate their actions. Unfortunately, numerical age is rarely mentioned in ethnographic descriptions of team games, making it impossible to determine age-specific rates of play. However, occasional references to the age categories of participants tend to support the expected age distribution. For example: Twana “shinny was played by older boys and young men” (Elmendorf and Kroeber 1992:234); Assiniboin shinny was “usually played by men of 20 to 30 years of age” (Culin 1907:636–37); Makah shinny was “played, as a rule, only by young men” (Dorsey 1901:70); and Chiricahua shinny was played by “men thirty or forty years” of age (Opler 1941:445–46). CPF is common in boys (Mendoza 2016), and the participation of old men appears to be rare (Nordenskiöld 1910:431).

A question raised by but beyond the scope of this study is the degree to which forager girls and women engage in CPF. If CIA in humans is male-specific, and if CPF functions to build and calibrate some of its components, we would expect such games to be played primarily by males or to be played between teams of the same sex, for the simple reason that competing against females would not give males an accurate sense of their fighting formidability. Symons makes a similar point with regard to dyadic fighting: “A male rhesus monkey who, by chance, had only female playmates during his formative years could conceivably develop a dysfunctional, exaggerated estimate of his fighting skills” (1978:195). However, since women were often captured, raped, or killed in raids (Scalise Sugiyama 2014), they too may have reaped fitness benefits from engaging in CPF. Team contact play might have increased female proficiency at thwarting aggression, which could be mustered for self-defense in actual attacks. On this point, we found that shinny and double-ball were popular women’s games in many forager groups, which suggests that the female mind might also have been shaped by CIA, albeit in somewhat different ways (Scalise Sugiyama 2014).

Although difficult to quantify, the strenuousness and roughness of CPF are further aspects of this behavior that simulate the conditions of lethal raiding. The arduousness of play can be inferred from the length of playing fields. For example: the Omaha and Ponca shinny field was “300 to 400 yards in length” (Dorsey 1884:336); Menomini lacrosse goals were “erected about one-third of a mile apart” (Hoffman 1896:128); Chippewa lacrosse goals were “about 600 yards apart” (Culin 1907:566); Yokuts lacrosse goals were “about 1,200 yards” apart (Culin 1907:596); and Chinook lacrosse goals were “about a mile apart” (Culin 1907:573). The arduousness of play can also be inferred from game duration. For example, in Ojibwa lacrosse, “a single inning may be continued for an hour or more” (Hoffman 1890:134), and Pomo shinny games might last “a half day or more” (Powers 1877:193). These extremes of time and distance were a keen test of stamina, and observers often commented on how exhausting play was. For example, Powers notes that Nisenan double-ball “goals are several hundred yards apart . . . and the players often race up and down the champaign . . . until they are dead blown and perspiring like top-sawyers” (1877:193). In lethal raiding, stamina would have been integral to retreat since mounting a raid often involved several days’ travel, and it was not uncommon for attackers to be pursued as they made their way home (Burch 2005:87; see also Ingstad 1987:347; Opler 1938:382; Wilbert 1975:74).

Another aspect of CPF that simulates the conditions of lethal raiding is roughness. In the game of mungan-mungan, for example, “the stick is often thrown through the air while men . . . are tripped and tackled, pushed and shoved” (Craig 2002:245). In Micmac lacrosse, “regard for neither life nor limb was allowed to stand in the way of possible success” (Brown 1889:45-46), and Ojibwa lacrosse players “fall upon and tread upon each other, and in the struggle some get rather rough treatment” (Copway 1860:4–5). Similarly, the Copper Inuit played a version of handball in which “nothing is disallowed: players may fight for the ball, trip one another, and so on” (Rasmussen and Calvert 1932:269–70). Not surprisingly, injuries—including the occasional broken bone—sometimes resulted from this vigorous play (e.g., Culin 1907:566; Guinnard 1861:93–94; Hoffman 1890:134; Métraux 1943:207–8; Powers 1877:193). For example, Tututni shinny players exhibited “great activity and strength, whacking away at each other’s shins . . . with a refreshing disregard for bruises” (Chase 1869:433), and in Menomini lacrosse, “frequently the ball carrier is disabled by being struck across the arm or leg, thus compelling his retirement” (Culin 1907:573). The roughness of these games may appear to contradict the claim that play is aimed at avoiding injury. However, observers rarely report that injuries were deliberately inflicted, permanently disabling, or lethal, and in at least some groups games were refereed (e.g., Gilij 1992; Hager 1895; Nordenskiöld 1910). Moreover, injury—or the prospect of it—can be beneficial: non-debilitating injury may develop pain tolerance, and the desire to avoid injury is a powerful incentive to rapidly hone one’s defensive skills. As Dolhinow and Bishop argue, “if learning which animals are stronger involves some pain, the young primate may learn the rules rapidly” (1970:170).


In this study, we have identified and described a previously unrecognized type of play, coalitional play fighting, and its possible function. We have shown that this play type occurs in more recent hunter-gatherer populations and thus may have occurred in earlier H. sapiens. Despite the fact that data are sparse, we have also shown that CPF is present across diverse hunter-gatherer societies on five continents, suggesting that it may be a statistical universal (Brown 1991) in humans. Finally, we have applied a new theoretical framework that conceptualizes play as part of the process of adaptation assembly and approaches design in terms of experiences available to the organism that can provide it with information inputs requisite to constructing and calibrating the adaptation in question.

As Symons (1978:6) observes, addressing the question of design vis-à-vis play requires granular description of the behavior in question. Although it has been argued that the widespread human affinity for watching and participating in sports is rooted in adaptations that evolved “in the context of male-male physical competition and primitive hunting and warfare” (Lombardo 2012:5), this claim has not been tested by comparing the motor patterns, body positioning, coordinated action, and strategies used in sports to those used in forager lethal raiding and/or hunting. To the best of our knowledge, this claim has only been tested by looking for sex differences in sports participation, spectatorship, and postgame affiliation in modern societies (e.g., Apostolou 2015; Apostolou and Zacharia 2015; Benenson and Wrangham 2016; Deaner and Smith 2013; Deaner et al. 2015). Although valuable, this approach does not illuminate design features beyond sex-specificity, and it overlooks the degree to which forager women engaged in CPF. We believe that identifying the motor patterns used in lethal raiding and demonstrating that these motor patterns are regularly recruited in competitive coalitional contact games across diverse forager cultures and habitats is an important step toward delineating design. If CPF builds, rehearses, and calibrates motor skills critical to engaging in lethal raiding, we would expect the former to exhibit the motor patterns demanded by the latter, which is precisely what we find. Alternative hypotheses must account for the recruitment of these particular motor patterns in the context of two male coalitions using coordinated action and physical force to prevent each other from penetrating each other’s “territory.” Given that team contact games presently enjoy worldwide popularity, the finding of CPF in 46 forager culture clusters across five continents—despite the fact that few ethnographers have given detailed attention to play in forager populations—suggests the presence of a robust phenomenon across human societies.

Supplementary material

12110_2018_9319_MOESM1_ESM.pdf (218 kb)
ESM 1 (PDF 218 kb)


  1. Abreu de Galindo, Juan de (1767). Historia de la conquista de las siete islas de Gran Canaria. The history of the discovery and conquest of the Canary Islands: translated from a Spanish manuscript, lately found in the island of Palma. With an enquiry into the origin of the ancient inhabitants. To which is added, a description of the Canary Islands, including the modern history of the inhabitants by Capt. George Glas, 2 vols. London: A. Pope and J. Swift.Google Scholar
  2. Anell, B. (1969). Running down and driving of game in North America. Studia Ethnographica Upsaliensia 30. Uppsala: Inst. för Allm. och Jämförande Etnografi.Google Scholar
  3. Apostolou, M. (2015). The athlete and the spectator inside the man: A cross-cultural investigation of the evolutionary origins of athletic behavior. Cross-Cultural Research, 49(2), 151–173.CrossRefGoogle Scholar
  4. Apostolou, M., & Zacharia, M. (2015). The evolution of sports: Exploring parental interest in watching sports. Evolutionary Psychological Science, 1(3), 155–162.CrossRefGoogle Scholar
  5. Barclay, W. S. (1904). The Land of Magellanes, with some account of the Ona and other Indians. The Geographical Journal, 23(1), 62–79.CrossRefGoogle Scholar
  6. Barrett, H. C., Cosmides, L., & Tooby, J. (2007). The hominid entry into the cognitive niche. In S. W. Gangestad & J. A. Simpson (Eds.), Evolution of mind: Fundamental questions and controversies (pp. 241–248). New York: Guilford.Google Scholar
  7. Batchelor, J. (1926). Ainu life and lore. Tokyo: Kyobunkwan.Google Scholar
  8. Beck, B. B. (1980). Animal tool behavior: The use and manufacture of tools by animals. New York: Garland STPM Pub.Google Scholar
  9. Benenson, J. F., & Wrangham, R. W. (2016). Cross-cultural sex differences in post-conflict affiliation following sports matches. Current Biology, 26(16), 2208–2212.CrossRefGoogle Scholar
  10. Biben, M. (1998). Squirrel monkey playfighting: Making the case for a cognitive training function for play. In M. Bekoff & J. A. Byers (Eds.), Animal play: Evolutionary, comparative, and ecological perspectives (pp. 161–182). Cambridge: Cambridge University Press.Google Scholar
  11. Biocca, E. (1970). Yanoáma: The narrative of a white girl kidnapped by Amazonian Indians. New York: Dutton.Google Scholar
  12. Bird, D. & Bliege Bird, R. (2005). Martu children’s hunting strategies in the Western Desert, Australia. In B. Hewlett & M. Lamb (Eds.), Hunter-gatherer childhoods (pp. 29–46). Piscataway, NJ: Aldine Transaction.Google Scholar
  13. Bjorklund, D. F., & Blasi, C. H. (2005). Evolutionary developmental psychology. In D. Buss (Ed.), The handbook of evolutionary psychology (pp. 828–850). Hoboken: Wiley.Google Scholar
  14. Bock, J. (2005). What makes a competent adult forager? In B. Hewlett & M. Lamb (Eds.), Hunter-gatherer childhoods (pp. 109–128). Piscataway, NJ: Aldine Transaction.Google Scholar
  15. Bock, J., & Johnson, S. (2004). Subsistence ecology and play among the Okavango Delta people of Botswana. Human Nature, 15, 63–81.CrossRefGoogle Scholar
  16. Bogoras, V. (1904–1909). The Chukchee. The Jessup North Pacific Expedition. Memoir of the American Museum of Natural History VII. New York: G. E. Stechert.Google Scholar
  17. Bogoras, W. (1918). Tales of the Yukaghir, Lamut, and Russianized natives of eastern Siberia. Anthropological Papers of the American Museum of Natural History XX, 1–148. New York: American Museum Press.Google Scholar
  18. Boulton, M. & Smith, P. (1992). The social nature of play fighting and play chasing: Mechanisms and strategies underlying cooperation and compromise. In J. Barkow, L. Cosmides & J. Tooby (Eds.), The adapted mind (pp. 429–444). New York: Oxford University Press.Google Scholar
  19. Bramble, D. M., & Lieberman, D. E. (2004). Endurance running and the evolution of Homo. Nature, 432(7015), 345–352.CrossRefGoogle Scholar
  20. Brown, W.W. (1889). Some indoor and outdoor games of the Wabanaki Indians. Transactions of the Royal Society of Canada 6(2):35–55.Google Scholar
  21. Brown, D. E. (1991). Human universals. New York: McGraw-Hill.Google Scholar
  22. Burch, E. (2005). Alliance and conflict: The World System of the Iñupiaq Eskimos. Lincoln. University of Nebraska Press.Google Scholar
  23. Burghardt, G. M. (2004) Play: How evolution can explain the most mysterious. In A. Moya & E. Font (Eds.), Evolution: From molecules to ecosystems (pp. 231–246). Oxford & New York: Oxford University Press.Google Scholar
  24. Byers, J., & Walker, C. (1995). Refining the motor training hypothesis for the evolution of play. The American Naturalist, 146(1), 25–40.CrossRefGoogle Scholar
  25. Calvo-Merino, B., Glaser, D. E., Grèzes, J., Passingham, R. E., & Haggard, P. (2004). Action observation and acquired motor skills: An FMRI study with expert dancers. Cerebral Cortex, 15(8), 1243–1249.CrossRefGoogle Scholar
  26. Carrier, D. R. (1984). The energetic paradox of human running and hominid evolution [and comments and reply]. Current Anthropology, 25(4), 483–495.CrossRefGoogle Scholar
  27. Carver, J. (1796). Travels through the interior parts of North America, in the years 1766, 1767, and 1768, etc., with maps. Philadelphia: Key and Simpson.Google Scholar
  28. Chagnon, N. (1997). Yąnomamö: The fierce people (fifth ed.). Fort Worth: Harcourt Brace College.Google Scholar
  29. Chase, A. W. (1869). Siletz, or “Lo” reconstructed. The Overland Monthly, 2, 424–441.Google Scholar
  30. Chick, G., Loy, J. W., & Miracle, A. W. (1997). Combative sport and warfare: A reappraisal of the spillover and catharsis hypotheses. Cross-Cultural Research, 31(3), 249–267.CrossRefGoogle Scholar
  31. Connor, R. C., Smolker, R. A., & Richards, A. F. (1992). Dolphin alliances and coalitions. In A. Harcourt and F. B. M. de Waal (Eds.), Coalitions and alliances in humans and other animals (pp. 415-443). Oxford and New York: Oxford University Press.Google Scholar
  32. Connor, R. C., Smolker, R., & Bejder, L. (2006). Synchrony, social behaviour and alliance affiliation in Indian Ocean bottlenose dolphins, Tursiops aduncus. Animal Behaviour, 72(6), 1371–1378.CrossRefGoogle Scholar
  33. Connor, R. C., Watson-Capps, J. J., Sherwin, W. B., & Krützen, M. (2010). A new level of complexity in the male alliance networks of Indian Ocean bottlenose dolphins (Tursiops sp.). Biology Letters, rsbl20100852.Google Scholar
  34. Cooper, J. M. (1949). Games and gambling. Bureau of American Ethnography Bulletin, 5(143), 503–524.Google Scholar
  35. Copway, G. (1860). The Ojibways. Boston: Albert Colby.Google Scholar
  36. Cordoni, G. (2009). Social play in captive wolves (Canis lupus): Not only an immature affair. Behaviour, 146(10), 1363–1385.CrossRefGoogle Scholar
  37. Craig, S. (2002). Sports and games of the ancients. Westport, CT: Greenwood Press.Google Scholar
  38. Culin, S. (1907) Games of the North American Indians. 24th Annual Report of the Bureau of American Ethnology. Washington, DC: Smithsonian Institution.Google Scholar
  39. Cumming, J., & Ramsey, R. (2008). Imagery interventions in sport. In S. Mellalieu & S. Hanton (Eds.), Advances in applied sport psychology: A review (pp. 5–36). London: Routledge.Google Scholar
  40. Cumming, J., & Williams, S. E. (2012). The role of imagery in performance. In S. Murphy (Ed.), Handbook of Sport and Performance Psychology (pp. 213–232). New York: Oxford University Press.Google Scholar
  41. Darwin, C. (1859). On the origin of the species by natural selection. London: J. Murray.Google Scholar
  42. Dawkins, R. (1986). The blind watchmaker: Why the evidence of evolution reveals a universe without design. New York: W.W. Norton.Google Scholar
  43. De Agostini, Alberto M. (1956) 30 años en Tierra del Fuego. Buenos Aires: Ediciones Peuser.Google Scholar
  44. Deaner, R. O., & Smith, B. A. (2013). Sex differences in sports across 50 societies. Cross-Cultural Research, 47(3), 268–309.CrossRefGoogle Scholar
  45. Deaner, R. O., Geary, D. C., Puts, D. A., et al. (2015). A sex difference in the predisposition for physical competition: Males play sports much more than females even in the contemporary U.S. PLoS One, 7(11), e49168.CrossRefGoogle Scholar
  46. Dolhinow, P. (1999). Play: A critical process in the developmental system. In P. Dolhinow & A. Fuentes (Eds.), The nonhuman primates (pp. 231–236). Mountain View, CA: Mayfield.Google Scholar
  47. Dolhinow, P. J., & Bishop, N. (1970). The development of motor skills and social relationships among primates through play. Minnesota Symposia on Child Psychology, 4, 141–198.Google Scholar
  48. Dorsey, J. O. (1884). Omaha sociology. Third Annual Report of the Bureau of American Ethnology, part 2, pp. 205–370. Washington DC.Google Scholar
  49. Dorsey, G. A. (1901). Games of the Makah Indians of Neah Bay. The American Antiquarian, 23, 69–73.Google Scholar
  50. Duntley, J. D. (2005). Adaptations to dangers from humans. In Buss, D. (Ed.), The handbook of evolutionary psychology (pp. 224–249). Hoboken: John Wiley & Sons.Google Scholar
  51. Eastman, C. A. (1902). Indian boyhood. New York: Dover.Google Scholar
  52. Elmendorf, W. W., & Kroeber, A. L. (1992). The structure of Twana culture with comparative notes on the structure of Yurok culture: Pre-white tribal lifeways on Washington’s Hood Canal. Pullman: Washington State. University Press.Google Scholar
  53. Ember, C. R. (1978). Myths about hunter-gatherers. Ethnology, 17(4), 439–448.CrossRefGoogle Scholar
  54. Ember, C. R., & Ember, M. (1997). Violence in the ethnographic record: Results of cross-cultural research on war and aggression. In D. Martin & D. W. Frayer (Eds.), Troubled times: Violence and warfare in the past (pp. 1–20). Amsterdam: Gordon and Breach.Google Scholar
  55. Fagen, R. (1974). Selective and evolutionary aspects of animal play. The American Naturalist, 108(964), 850–858.CrossRefGoogle Scholar
  56. Fagen, R. M. (1977). Selection for optimal age-dependent schedules of play behavior. The American Naturalist, 111(979), 395–414.CrossRefGoogle Scholar
  57. Fagen, R. (1981). Animal play behavior. London: Oxford University Press.Google Scholar
  58. Frayer, D. W. (1997). Ofnet: Evidence for a Mesolithic massacre. In D. Martin & D. W. Frayer (Eds.), Troubled times: Violence and warfare in the past (pp. 181–216). Amsterdam: Gordon and Breach.Google Scholar
  59. Garfield, Z. H., Garfield, M. J., & Hewlett, B. S. (2016). A cross-cultural analysis of hunter-gatherer social learning. In H. Terashima & B. Hewlett (Eds.), Social learning and innovation in contemporary hunter-gatherers (pp. 19–34). Tokyo: Springer.CrossRefGoogle Scholar
  60. Gat, A. (1999). The pattern of fighting in simple, small-scale, pre-state societies. Journal of Anthropological Research, 55(4), 563–583.CrossRefGoogle Scholar
  61. Gilij, F. S. (1992). Ensayo de Historia Americana. Translator Antonio Tovar. 3 vols. Caracas, Venezuela: Editorial Arte. (Originally published in 1731).Google Scholar
  62. Goodwin, G. (1971). Western Apache raiding and warfare, In K. Basso (Ed.). Tucson: University of Arizona Press.Google Scholar
  63. Guinnard, A. (1861). Tres años de esclavitud entre los Patagones (Relato de Mi Cautiverio). Doral, FL: Stockcero.Google Scholar
  64. Haenke, T. (1943). Viaje por el Virreinato del Río de la Plata. Buenos Aires: Emecé Editores. (Originally published in 1794).Google Scholar
  65. Hager, S. (1895). Micmac customs and traditions. American Anthropologist, 8(1), 31–42.CrossRefGoogle Scholar
  66. Henn, V., & Lignitz, E. (2004). Kicking and trampling to death. In M. Tsokos (Ed.), Forensic pathology reviews (pp. 31–50). Totowa, NJ: Humana Press.CrossRefGoogle Scholar
  67. Hoffman, W. J. (1890). Remarks on Ojibwa ball play. American Anthropologist, 3(2), 133–136.CrossRefGoogle Scholar
  68. Hoffman, W. J. (1896). The Menomini Indians. Fourteenth Annual Report of the Bulletin of American Ethnology. Washington, DC: US Government Printing Office.Google Scholar
  69. Hutchinson, T. J. (1865). On the Chaco and other Indians of South America. Transactions of the Ethnological Society of London, 3, 321–334.CrossRefGoogle Scholar
  70. Ingstad, H. (1987). Nunamiut stories. K. Bergsland (Trans. and Ed.). Barrow: North Slope Borough Commission on Iñupiat History, Language and Culture.Google Scholar
  71. Isaac, G. (1978). The food-sharing behavior of protohuman hominids. Scientific American, 238(4), 90–109.CrossRefGoogle Scholar
  72. Kaplan, H., Gangestad, S., Gurven, M., Lancaster, J., Mueller, T., & Robson, A. (2007). The evolution of diet, brain and life history among primates and humans. In W. Roebroeks (Ed.), Guts and brains: An integrative approach to the hominin record (pp. 47–90). Leiden: Leiden University Press.Google Scholar
  73. Kaufman, S. F. (2004). Musashi’s Book of Five Rings. Rutland, VT: Tuttle.Google Scholar
  74. Kay, A., & Teasdale, G. (2001). Head injury in the United Kingdom. World Journal of Surgery, 25(9), 1210–1220.CrossRefGoogle Scholar
  75. Lahr, M. M., Rivera, F., Power, R. K., Mounier, A., Copsey, B., Crivellaro, F., et al. (2016). Inter-group violence among early Holocene hunter-gatherers of West Turkana, Kenya. Nature, 529(7586), 394–398.CrossRefGoogle Scholar
  76. Lambert, P. M. (2002). The archaeology of war: A North American perspective. Journal of Archaeological Research, 10(3), 207–241.Google Scholar
  77. Lambert, P. M. (2007). The osteological evidence for indigenous warfare in North America. In R. Chacon & R. G. Mendoza (Eds.), North American indigenous warfare and ritual violence (pp. 202–221). Tucson: University of Arizona Press.Google Scholar
  78. Lancaster, J. B. (1971). Play-mothering: The relations between juvenile females and young infants among free-ranging vervet monkeys (Cevcopithecus aethiops). Folia Primatologica, 15(3-4), 161–182.Google Scholar
  79. Lancaster, J. B., & Lancaster, C. S. (1983). Parental investment: The hominid adaptation. In D. Ortner (Ed.), How humans adapt: A biocultural odyssey (pp. 33–56). Washington, DC: Smithsonian Institution.Google Scholar
  80. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174.CrossRefGoogle Scholar
  81. LeBlanc, S. A., & Register, K. E. (2003). Constant battles: Why we fight. New York: St. Martin’s Press.Google Scholar
  82. Lee, R. B., & DeVore, I. (Eds.). (1968). Man the hunter. Chicago: Aldine.Google Scholar
  83. Leonard, W. R., Robertson, M. L., & Snodgrass, J. J. (2007). Energetics and the evolution of brain size in early Homo. In W. Roebroeks (Ed.), Guts and brains: An integrative approach to the hominin record (pp. 29–46). Leiden: Leiden University Press.Google Scholar
  84. Liebenberg, L. (2006). Persistence hunting by modern hunter-gatherers. Current Anthropology, 47(6), 1017–1026.CrossRefGoogle Scholar
  85. Logan, C. J., & Longino, J. T. (2013). Adult male coatis play with a band of juveniles. Brazilian Journal of Biology, 73(2), 353–355.CrossRefGoogle Scholar
  86. Loizos, C. (1966). Play in mammals. In P. A. Jewell & Caroline Loizos (Eds.), Play, exploration, and territory in mammals (pp. 1–9). New York: Academic Press.Google Scholar
  87. Lombardo, M. P. (2012). On the evolution of sport. Evolutionary Psychology, 10(1), 147470491201000101.CrossRefGoogle Scholar
  88. MacDonald, K. (2007). Cross-cultural comparison of learning in human hunting. Human Nature, 18(4), 386–402.CrossRefGoogle Scholar
  89. Macfarlan, A., & Macfarlan, P. (1958). Book of American Indian games. New York: Association Press.Google Scholar
  90. Maguire, R. A. J. (1928). “Il-torōbo”: Part II. Journal of the Royal African Society, 27(107), 249–268.Google Scholar
  91. Manson, J. H., & Wrangham, R. W. (1991). Intergroup aggression in chimpanzees and humans [and comments and replies]. Current Anthropology, 32(4), 369–390.CrossRefGoogle Scholar
  92. Marlowe, F. (2005). Hunter-gatherers and human evolution. Evolutionary Anthropology, 14, 54–67.CrossRefGoogle Scholar
  93. Maschner, H. D., & Reedy-Maschner, K. L. (1998). Raid, retreat, defend (repeat): The archaeology and ethnohistory of warfare on the North Pacific Rim. Journal of Anthropological Archaeology, 17(1), 19–51.Google Scholar
  94. McClellan, C. (1987). Part of the land, part of the water: A history of the Yukon Indians. Vancouver, BC: Douglas and McIntyre.Google Scholar
  95. McDonald, M. M., Navarrete, C. D., & Van Vugt, M. (2012). Evolution and the psychology of intergroup conflict: The male warrior hypothesis. Philosophical Transactions of the Royal Society B, 367(1589), 670–679.CrossRefGoogle Scholar
  96. Mendoza, M. (1985). Hombres-de-guerra y hombres-de-paz: integración de los impulsos agresivos y pacíficos entre los Tobas del Pilcomayo Medio. Publicaciones del CEFyL, Departamento de Antropología, Universidad de Buenos Aires.Google Scholar
  97. Mendoza, M. (2016). Juegos de combate entre varones de grupos etnográficos cazadores-recolectores. Resistencia, Argentina: Universidad Nacional del Nordeste.Google Scholar
  98. Métraux, A. (1943). Suicide among the Matako of the Argentine Gran Chaco. América Indígena, 3, 199–210.Google Scholar
  99. Moran, A., Guillot, A., MacIntyre, T., & Collet, C. (2012). Re-imagining motor imagery: Building bridges between cognitive neuroscience and sport psychology. British Journal of Psychology, 103(2), 224–247.CrossRefGoogle Scholar
  100. Moss, M. L., & Erlandson, J. M. (1992). Forts, refuge rocks, and defensive sites: The antiquity of warfare along the north Pacific Coast of North America. Arctic Anthropology, 29(2), 73–90.Google Scholar
  101. Murdock, G. (1967). Ethnographic atlas. Pittsburgh: University of Pittsburgh Press.Google Scholar
  102. Nefëdkin, Alexander K. (2014). Warfare of the Chukchi (mid 17 th to early 20 th century). Translated by Richard L. Bland. Anchorage: US Department of Interior, National Park Service, Shared Beringian Heritage Program.Google Scholar
  103. Nordenskiöld, E. (1910). Spiele und Spielsachen im Gran Chaco und in Nordamerika. Zeitschrift für Ethnologie, 42(3/4), 427–433.Google Scholar
  104. Opler, M. (1938). Myths and tales of the Jicarilla Apache Indians. New York: American Folklore Society.Google Scholar
  105. Opler, M. E. (1941). An Apache life-way: The economic, social, and religious institutions of the Chiricahua Indians. Lincoln: University of Nebraska Press.Google Scholar
  106. Oskarsson, P. A., Nählinder, S., & Svensson, E. (2010). A meta study of transfer of training. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 54(28), 2422–2426. Los Angeles: SAGE Publications.Google Scholar
  107. Oswalt, W. (1999). Eskimos and explorers, second ed. Lincoln: University of Nebraska Press.Google Scholar
  108. Parsons, E. W. C. (Ed.). (1929). Kiowa tales. Memoirs of the American Folklore Society 22.Google Scholar
  109. Pellis, S. M., & Iwaniuk, A. N. (1999). The problem of adult play fighting: A comparative analysis of play and courtship in primates. Ethology, 105(9), 783–806.CrossRefGoogle Scholar
  110. Powers, S. (1877). Tribes of California. US Geological Survey Contributions to American Ethnology 3. Washington, DC.Google Scholar
  111. Raichlen, D., et al. (2016). Physical activity patterns and biomarkers of cardiovascular disease risk in hunter-gatherers. American Journal of Human Biology.
  112. Rand, S. T., & Webster, H. L. (1894). Legends of the Micmacs. New York and London: Longmans, Green.Google Scholar
  113. Rasmussen, K. (1931). The Netsilik Eskimos: Social life and spiritual culture. Report of the 5th Thule Expedition 1921–24. Copenhagen: Gyldendalske Boghandel, Nordisk Forlag.Google Scholar
  114. Rasmussen, K., & Calvert, W. E. (1932). Intellectual culture of the Copper Eskimos. Copenhagen: Gyldendalske Boghandel.Google Scholar
  115. Roberts, J. M., & Sutton-Smith, B. (1962). Child training and game involvement. Ethnology, 1(2), 166–185.CrossRefGoogle Scholar
  116. Scalise Sugiyama, M. (2014). Fitness costs of warfare for women. Human Nature, 25(4), 476–495.CrossRefGoogle Scholar
  117. Schebesta, P. (1933). Among Congo Pygmies. London: Hutchinson.Google Scholar
  118. Sipes, R. G. (1973). War, sports and aggression: An empirical test of two rival theories. American Anthropologist, 75(1), 64–86.CrossRefGoogle Scholar
  119. Stamps, J. (1995). Motor learning and the value of familiar space. The American Naturalist, 146(1), 41–58.CrossRefGoogle Scholar
  120. Stern, B. J. (1934). The Lummi Indians of northwest Washington. New York: Columbia University Press.Google Scholar
  121. Sutton-Smith, B. (1995). Conclusion: The persuasive rhetorics of play. In Anthony D. Pellegrini (Ed.), The future of play theory: A multidisciplinary inquiry into the contributions of Brian Sutton-Smith (pp. 275–295). Albany: SUNY Press. Google Scholar
  122. Symons, D. (1974). Aggressive play and communication in rhesus monkeys (Macaca mulatta). American Zoologist, 317–322.Google Scholar
  123. Symons, D. (1978). Play and aggression: A study of rhesus monkeys. New York: Columbia University Press.Google Scholar
  124. Symons, D. (1987). If we're all Darwinians, what's the fuss about? In C. Crawford, M. Smith, & D. Krebs (Eds.), Sociobiology and psychology: Ideas, issues, and applications (pp. 121–146). Hillsdale, NJ: L Erlbaum Associates.Google Scholar
  125. Symons, D. (1992). On the use and misuse of Darwinism in the study of human behavior. In J. Barkow, L. Cosmides, & J. Tooby (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 137–159). New York: Oxford University Press.Google Scholar
  126. Taçon, P., & Chippendale, C. (1994). Australia’s ancient warriors: Changing depictions of fighting in the rock art of Arnhem Land, N.T. Cambridge Archaeological Journal, 4, 211–248.CrossRefGoogle Scholar
  127. Teit, J. A. (1956). Field notes on the Tahltan and Kaska Indians: 1912-1915, edited by J. H. MacNeish. Anthropologica, 3, 39–171.Google Scholar
  128. Thornhill, R. (1997). The concept of an evolved adaptation. In Bock, G., Cardew, G. (Eds.), Characterizing human psychological adaptations (pp. 4-22). Ciba Foundation Symposium 208. Chichester, UK, and New York: Wiley.Google Scholar
  129. Tooby, J., & Cosmides, L. (1988). The evolution of war and its cognitive foundations. Institute for Evolutionary Studies Technical Report 88–81.Google Scholar
  130. Tooby, J., & Cosmides, L. (1989). Evolutionary psychology and the generation of culture, part I. Theoretical considerations. Evolution and Human Behavior, 10(1), 29–49.Google Scholar
  131. Tooby, J., & Cosmides, L. (2001). Does beauty build adapted minds? Toward an evolutionary theory of aesthetics, fiction, and the arts. SubStance, 94(95), 6–27.Google Scholar
  132. Tooby, J., & Cosmides, L. (2010). Groups in mind: The coalitional roots of warfare and morality. In H. Høgh-Olesen (Ed.), Human morality and sociality: Evolutionary and comparative perspectives (pp. 191–234). New York: Palgrave Macmillan.Google Scholar
  133. Tooby, J. and DeVore, I. (1987). The reconstruction of hominid behavioral evolution through strategic modeling. In W. Kinzey (Ed.), The evolution of human behavior: Primate models (pp. 183–237). Albany: SUNY Press.Google Scholar
  134. Turnbull, C. M. (1983). The Mbuti Pygmies: Change and adaptation. NewYork: Holt Rinehart and Winston.Google Scholar
  135. Van Lawick-Goodall, J. (1967). My friends the wild chimpanzees. Washington, DC: National Geographic Society.Google Scholar
  136. Walker, J. R. (1905). Sioux games I. Journal of American Folklore, 18(71), 277–290.CrossRefGoogle Scholar
  137. Walker, R., Hill, K., Kaplan, H., & McMillan, G. (2002). Age dependency of strength, skill, and hunting ability among the Ache of Paraguay. Journal of Human Evolution, 42, 639–657.CrossRefGoogle Scholar
  138. Warner, W. L. (1931). Murngin warfare. Oceania, 1(4), 457–494.CrossRefGoogle Scholar
  139. Webster, G. (1998). The Roman Imperial Army of the first and second centuries AD. Norman: University of Oklahoma Press.Google Scholar
  140. Wells, R. Jr., & John W. Kelly. (1890). English-Eskimo and Eskimo-English vocabularies (compiled by Wells), preceded by ethnographical memoranda concerning the Arctic Eskimos in Alaska and Siberia (by Kelly). Bureau of Education Circular 2. Washington, DC: Government Printing Office.Google Scholar
  141. Wilbert, J. (Ed.). (1975). Folk literature of the Selknam Indians. Los Angeles: UCLA Latin American Center Publications.Google Scholar
  142. Williams, G. C. (1966). Adaptation and natural selection: A critique of some current evolutionary thought. Princeton: Princeton University Press.Google Scholar
  143. Withnell, J. G. (1901). The customs and traditions of the Aboriginal natives of North Western Australia. Roebourne, Australia: Hugh B. Geyer, Printer.Google Scholar
  144. Wrangham, R. W. (1999). Evolution of coalitionary killing. Yearbook of Physical Anthropology, 42, 1–39.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Anthropology DepartmentUniversity of OregonEugeneUSA

Personalised recommendations