[Presented at G. Stanley Hall Symposium 2001, Williams College, “Gender and Aggression”]


The Importance of War in the Gendered Socialization of Aggression


Joshua S. Goldstein

Professor of International Relations (on leave)

American University, Washington DC



413-256-6363 voice/fax





The Importance of War in the Gendered Socialization of Aggression

Joshua S. Goldstein




I. The Need to Be Prepared for War

The Nearly Universal Potential for War

Gendered War Roles

Traumatic Effects of War

Biological and Cultural Aspects of Gendered War


II. Gendered Socialization of Aggression

Toughening Boys

Exposure to Violence

Childhood Gender Segregation



[Author’s note:] Parts of this chapter are excerpted and adapted with permission from Joshua S. Goldstein, War and Gender: How Gender Shapes the War System and Vice Versa (Cambridge University Press, 2001), which also contains further scholarly references. For more resources on this topic, see www.warandgender.com.



I. The Need to Be Prepared for War


            War casts a long shadow over many aspects of gender and aggression. Societies’ need to prepare for potential war affects how they socialize boys and girls differently concerning aggression. In turn, in a feedback relationship of mutual causality, the gender differences in how societies socialize children for aggression affect the strongly dimorphous gender roles found in war.


            Virtually all human societies need to be prepared to fight wars. With the exception of a handful of cases – all of them small isolated communities – all societies have war in their repertoire of social behaviors. (I define war broadly, as lethal intergroup violence.) Because war universally relies on a gender-based division of labor, this need to prepare for war affects gender roles extensively across society and through the life span.


The Nearly Universal Potential for War

            Virtually every society must successfully address the problem of being able to fight wars. This is because no known societies have found stable long-term peace. They all go to war at least occasionally, or they did before being “pacified” by European colonial authorities. (The very small number of exceptions are explained by physical isolation that makes war impractical; see Goldstein 2001: 22-34.)


            Furthermore, simple societies are at least as war prone as modern industrialized ones. Modern non-pacified preindustrial societies are not generally peaceful. Ember and Ember (1994; 1997) find over half of a sample of 90 societies were in a constant state of war or readiness for war, and half of the remaining societies fought every year during a particular season. In only eight societies did wars occur less frequently than once in 10 years on average. Of 31 gathering-hunting societies surveyed in another study, 20 typically had warfare more than once every two years, and only three had “no or rare warfare” (Ember 1978: 444). Nonstate societies have as much warfare as states do. Relatively peaceful societies can become warlike and vice versa, as the !Kung did (Goldstein 2001: 28-29).


            The twentieth century was one of total war, and the twenty-first starts with the world’s cities living under threat of attack by weapons of mass destruction. The need to be prepared for war therefore is still immediate.


Gendered War Roles

            Societies have universally met the challenge of war preparation in a gendered manner. War is among the most consistently gendered of human activities. Every known society assigns war roles differentially by gender, with men as the primary fighters (and usually the only ones). Exceptions are numerous and informative, but these exceptions together amount to fewer than 1 percent of all warriors in history (Goldstein 2001: 10-22).


            Of about 23 million soldiers in uniformed standing armies, about 97 percent are male (somewhat over 500,000 are women). In only six of the world’s nearly 200 states do women make up more than 5 percent of the armed forces. And most of these women in military forces worldwide occupy traditional women’s roles such as typists and nurses. Designated combat forces in the world’s state armies today include several million soldiers (the exact number depending on definitions of combat), of whom 99 percent are male. These disparities persist despite women’s having reached historically high levels of social and political power globally, and despite the world’s predominant military force’s carrying out the largest-scale military gender integration in history, with 200,000 women comprising one-sixth of U.S. forces (Goldstein 2001: 10-11).


            Today’s 97% male military worldwide may be the all-time low for this variable through history – a variable that has shown amazing consistency and robustness against technological, military, and politico-diplomatic evolution through history. When war shaped the rise of states and civilizations after the neolithic revolution, it was already a male domain. The importance of horses in historical warfare did not alter the gender division despite the fact that women ride as well as men (only equestrian events are gender-integrated in the Olympics). The introduction of firearms, and later the mechanization of war, radically changed the importance of physical strength in war, but still barely affected the gender division.


            Nor do simple societies offer counterexamples. No empirically corroborated cases are known of Amazon societies in which all (or even a majority) of fighters were female (Goldstein 2001: 11-19). Some archaeological evidence suggests that early-iron-age nomadic women of the Eurasian steppes rode horses, may have used weapons, and may even have had some political influence, though probably not dominance. But excavated graves yielded war-related artifacts for about 90 percent of men and only 15–20 percent of women (Davis-Kimball 1997: 47). Little evidence exists for purported Amazon societies in ancient Greece or South America.

            Among contemporary preindustrial societies, both the very war-prone and the relatively peaceful ones share a gender division in war with men as the primary (and usually exclusive) fighters. For example, although gender relations on Vanatinai island (where war is rare) are radically more egalitarian than those among the war-centered Sambia, one commonality is war fighting – a male occupation. In many present-day gathering-hunting and agrarian societies, special gender taboos apply to weapons, and special practices focus on men’s roles as warriors. Sometimes war and hunting are the only two spheres of social life to exclude women.


Traumatic Effects of War

            One might think that war fighting is a male occupation because men like to fight. But this argument is untenable. All evidence indicates that killing does not come naturally to men. Combat is a horrific experience marked by confusion, noise, terror, and atrocity in addition to any physical injury. Societies historically have worked hard to get men to fight – drafting them, disciplining them (e.g., shooting deserters), sometimes drugging them, and sometimes abruptly breaking family and community ties and replacing them with military bonding. After a war, many cultures honor veterans and confer special status or rewards on them. In some societies, war participation and war leadership open opportunities for political leadership. By contrast, men who do not fight may be shunned as cowards. All these inducements to participate in combat show the difficulty of getting men to fight (Goldschmidt 1989: 16–17, 22–23; Goldstein 2001: 253).


            Many war survivors suffer lasting psychological effects, including Post-Traumatic Stress Disorder (PTSD). The experience of battle is inherently traumatic – scenes of violent horror, the killing of best friends before one’s eyes, blood, gore, physical pain (Bourke 1999). Judith Herman (1992) argues that in war trauma, as in rape trauma, isolation is central. Civilian society’s common lack of interest in hearing about war traumas, along with survivors’ own denial, make healing difficult. PTSD has gone by various names in different wars, notably “shell shock” in World War I. Men and women are about equally susceptible to PTSD. Women war veterans are as prone as men to PTSD, but more men than women are exposed to combat trauma (Goldstein 2001: 259-63).


            Thus, every society must prepare for war specifically by motivating the participation of combatants despite the high individual costs of that participation both physically and emotionally. Every society focuses this effort on males.


Biological and Cultural Aspects of Gendered War

            The commonality of gendered war roles cross-culturally suggests a biological basis. I argue that biology leans, so to speak, in one direction but is insufficient to account for gendered war roles. If there were no relevant biological gender differences, many more cultures would have used women as combatants. However, if biological gender differences alone accounted for male participation in war, cultures would not have to work so persistently to get men to fight. Rather, I believe, relatively modest average gender differences are reinforced and amplified by cultures’ socialization practices in the course of societies’ effort to prepare for war.


            Men’s biology favors them for fighting in several ways. On average, men are larger and stronger than women – though men’s and women’s bell-curve distributions overlap. Men’s spatial abilities, modestly higher than women’s on average, might also slightly help men in war. And, owing to androgen exposure in utero, boys have a tendency towards rough-and-tumble play. Size and strength differences presumably would have mattered most early in human history before such inventions as horse-riding, firearms, and airplanes made sheer strength less decisive in war. However, gendered war roles have persisted despite these technological changes, so size differences do not explain the gender divide in combat participation.


            The main biological gender differences relevant to war all derive from effects of testosterone and other “male” and “female” sex hormones (both of which both men and women have, though in different amounts). These effects occur first in utero – resulting in sex organ differentiation and relatively subtle differences in brain functioning – and later around puberty and thereafter when androgens make men bigger and stronger (among other effects). Dabbs and Dabbs (2000) argue that inherited tendencies for high or low testosterone levels drive many aspects of individual behavior. However the reverse direction of causality – from social behavior to testosterone levels – seems stronger.


            Men’s fluctuating short-term testosterone levels respond to competitive situations, such as a tennis or wrestling match, a chess game, or a competitive task in a psychology laboratory (Mazur and Booth 1998; Mazur and Lamb 1980; Archer 1991, 17–18; Goldstein 2001: 153-57). Levels rise in preparation for the competition, and then go up afterwards in winners, and down in losers. This effect does not depend on direct aggression. It applies to any changes in an individual’s perceived status in a social hierarchy. Winning or losing a physical fight often has that effect, but so do other competitions. The testosterone high of competitive victory has been measured in males participating in a ceremony to receive their M.D. degrees, and even in sports fans when their team wins (Bernhardt et al. 1998).


            The effect on testosterone levels depends on subjective judgments about triumph or defeat, and is strongest when a victory is decisive and results from an individual’s own efforts (Geen 1998: 321). Among eight men participating in a New York chess tournament over eight weeks, testosterone levels rose about 10 percent on average in winners of games where the chess ratings of the players were close (the players expected to have to fight hard to win). However, testosterone actually decreased after winning a game where the ratings showed ahead of time that a win would be easy (down about 10 percent, as with losers) (Mazur, Booth, and Dabbs 1992). In 17 young male first-offenders in a shock-incarceration (“boot camp”) program, testosterone levels dropped dramatically in the first month, but less so in six men who started out with a bad attitude and may have refused to feel defeated (Thompson, Dabbs, and Frady 1990).


            Outcomes of aggressive interactions also affect testosterone levels among animals. When male rodents fight over status and territory, the winner of the fight produces more testosterone and the loser produces less. In rhesus monkeys, researchers studied whether levels of testosterone, prior to the formation of a group from unfamiliar males, would predict the eventual status hierarchy that emerged in that group. They did not. But once that hierarchy was established, the testosterone levels in the top monkey rose dramatically, as much as tenfold. After fighting, defeated males’ testosterone levels dropped to 10–15 percent of the prior level (Monaghan and Glickman 1992: 281–82). Thus, testosterone levels appear to reflect changes in status – i.e., winning and losing. Similarly, in experiments where male monkeys displayed aggression but did not win or lose an encounter, their testosterone levels were unaffected.


            In adult men who are not criminals, some researchers find no correlation of testosterone levels with aggression unconnected to sex (Monaghan and Glickman 1992: 283; Baron and Richardson 1994: 255). Others find a small but positive correlation (Archer 1991: 14). Aggression does not rise and fall with the daily and seasonal fluctuations in men’s testosterone levels. Thus, testosterone does not appear to cause aggression, though it may intensify it.


            The complexity and bidirectional causality that marks androgens and aggression suggests that men are not biologically driven to fight wars (or more generally to engage in aggression and violence), though not everyone would agree with this conclusion (see Wrangham and Peterson 1996). Some scholars have argued that men have evolved for violence because such an adaptation confers evolutionary advantage (on the related issue of homicide in evolutionary perspective, see Buss’s chapter in this volume). These issues remain controversial. Generally, in species where males are adapted for fighting, there is a large size difference between males and females, but humans seem to have moved the opposite direction with a relatively small gender gap in size (Hall 1985: 134–39).


            Thus, biological aspects of gendered war roles are important, yet still indeterminate. Men do not follow programming that sends them merrily off to war. They have biological adaptations for war-fighting, yet still resist individually paying the cost for the “collective good” of war fighting (be it for defense or aggrandizement of the community). Cultures channel and amplify gendered biological potentials as part of preparation for war. My main point about biology is that what the war system requires – that societies come up with capable combatants – biology does not fully provide.


II. Gendered Socialization of Aggression

            Because biology does not solve the dilemma of getting individuals to participate in fighting wars, cultures devote considerable effort to solving it. Some of these efforts involve gender, while many do not. (Soldiers may believe in a cause, or strongly identify with a country or group they are fighting to defend, for instance.)


            Among the gender-defined approaches found in a wide range of cultures, conceptions of masculinity and femininity often promote ideals adapted for war. Masculine men are brave, tough, and ready to fight to protect their women and communities. Most importantly, they are able to suppress emotions – especially fear and grief – and thus stay functional amid the horrors of battle. Feminine women are skilled nurturers who can nurse men emotionally after traumatic war experiences, among other qualities.


            Cultures socialize children to prepare them for gender roles appropriate to a society prepared for war (a war in which the genders will have starkly different roles). Socialization involves motivating children to comply with norms of behavior appropriate in a given culture. Means of socialization include teaching, reinforcement, and modeling, by parents and others. Societies socialize children differentially by gender to prepare them for adult gender roles, including war.


            Socialization is a powerful force shaping behaviors, even in adulthood. Socialization for aggression succeeds in motivating combat participation where biology alone does not do so. Indeed, desensitization and training can turn ordinary people into monsters who commit atrocities, as with Japanese troops during the 1937 “rape of Nanking.” One Japanese commander said of his new troops’ shock at seeing hardened veterans torturing civilians to death, “All new recruits are like this, but soon they will be doing the same things themselves” (Chang 1996: 57–59).


            Socialization of children for aggression, plays a very important role in facilitating a society’s ability to fight wars, and in turn that socialization occurs under the influence of the society’s war experiences. Causality between war and the socialization of aggression is hard to pin down, and in my view runs in both directions. One study of a balanced sample of 186 largely preindustrial societies, “strongly suggest[s] that more homicide and assault is a consequence of socialization for aggression which in turn is a consequence of more war” (Ember and Ember 1997). This makes war the driving variable.


            But it would be illogical to think socialization for aggression does not in turn affect the next generation’s propensity for war. After all, harshness toward children contributes to their later manifestation of aggression and violence. Growing up in a violent context increases the likelihood of violence later in life. Ross’s (1993: 98-102; 1990: 55-56, 60) statistical analysis of ninety “small scale, preindustrial societies” finds that war frequency correlates with harsh or affectionate child socialization practices, among other variables.


            Warmth and affection towards children may facilitate attachment and reduce fears of object loss, reducing hostility in social relationships later in life. Most societies with infrequent war are marked by permissive child-rearing practices. In a 1978 study of “peaceful” societies, Ashley Montagu found parental affection to be the key. Among seven small-scale societies that are all low on internal conflict and aggression, “great affection is frequently directed toward the child... Overt expression of aggression is discouraged, but not through physical punishment. Finally, these societies lack models of highly aggressive persons ...” (Ross 1990: 55–56).


            In addition to the circular causality of war and socialization for aggression, the causality regarding gender in these two phenomena is indeterminate. Do cultures socialize boys and girls differently regarding aggression because only men will have to fight wars later? Or do only men fight wars because they were socialized for aggression differently as children? Again, I think, probably both are true.


            Differential socialization of aggression in boys and girls is hard to verify empirically. Maccoby and Jacklin found parents equally permissive of aggression with preschool girls as with boys. A study of 700 mothers of English 4-year-olds found that mothers reported encouraging daughters to fight back as often as sons. A variety of studies have found parents “on average, equally permissive (or nonpermissive) toward aggression in sons and in daughters.” A substantial amount of school aggression derives from a “relatively small subset of children,” mainly boys, who consistently engage in fighting or bullying. These are not typical or ideal-type boys, and gender researchers perhaps focus on these extreme cases too much. The “aggressive” boys tend to come from families in which coercive behavior is high among various family members (Maccoby 1998: 130–34).


            Yet aggression itself has shown very consistent gender differences (Coie and Dodge 1998: 789–92). Maccoby and Jacklin (1974: 349–53) find a “fairly well established” gender difference in aggression, “observed in all cultures in which the relevant behavior has been observed. Boys are more aggressive both physically and verbally,” and have more play-fighting and aggressive fantasies. The gender difference in aggression starts “as early as social play begins – at age 2 or 2½. Although the aggressiveness of both sexes declines with age, boys and men remain more aggressive through the college years” (p.352). Males’ primary victims from an early age are other males. Whiting and Whiting’s (1975: 167) study of six widely differing cultures worldwide found that overall, “[g]irls are more nurturant and boys more aggressive ...”


            Several recent studies of aggression find significant gender differences in physical aggression, but fewer gender differences – even a slight female edge at certain ages – in verbal or social aggression, such as excluding someone from a group (which can devastate the victim). Some researchers find that “relational” aggression occurs significantly more often among girls than boys, both in preschool and primary years (Crick, Casas and Ku 1999; Crick and Grotpeter 1995). Girls and boys appear to follow similar developmental paths, in which aggression decreases from childhood to adulthood, and direct physical aggression gives way to indirect aggression based on social relationships, such as ostracism.


            During elementary school, children’s aggression decreases and becomes more limited to a few individuals. The previous “nonsocial, instrumental nature of aggression” gives way increasingly to “person-oriented and hostile” aggression elicited by “threats and derogations to one’s ego” (Coie and Dodge 1998: 791). However, boys lag behind girls at each developmental stage, especially in adolescence when boys’ aggression remains as direct as ever (and more injurious) whereas girls’ aggression has become mainly indirect. Aggression decreases with age for girl-girl and mixed-gender interactions, but remains high through early adolescence for boy-boy aggression (Cairns and Cairns 1994, 56–67; Maccoby 1998: 40–41, 57–58).


            Despite these average gender differences, gender differences in aggression are not categorical but a matter of degree. Many women commit violence against men (see Huesmann’s chapter in this volume). Meta-analyses support the characterization of gender differences in aggression as relatively modest. Eagly and Steffen (1986: 325) trace gender differences in adult aggression to different beliefs about the consequences of aggression, using meta-analysis to combine results of fifty previous quantitative studies seeking to explain aggressive behavior. (The method of meta-analysis combines a number of past studies’ results in a statistically appropriate manner.) “In general, this meta-analysis shows that men are more aggressive than women and that this sex difference is more pronounced for physical than psychological aggression.” However, the sizes of gender effects vary and are often quite small. Women’s aggression is reduced when the expected effect would harm the victim, endanger themselves, or provoke intense guilt or anxiety. Similarly, Hyde’s (1986: 51, 63) meta-analysis of 143 studies also finds significant gender differences for “all types of aggression, all methods of measurement, and all designs, although effect sizes were generally small to moderate ...” (the distribution curves for males and females overlap a lot).


            Similarly, many animal species show gender differences in aggression that are relative, not categorical. Even though females can be as aggressive as males, notably when their young are threatened, nonetheless behavior patterns regarding violence often differ for male and females of a given species. Primatologist Meredith Small (1993: 27–28, 57) finds overlap in the distributions of aggressiveness in male and female primate populations. Males are more aggressive and females more social on average, yet females can be aggressive when called for and males care about social interactions.


            Overall, then, researchers find modest but persistent gender differences in aggression. I think these are both an effect and a cause of men’s roles as potential and actual warriors. Gender differences in aggression are in the same direction as, but less pronounced than, gender differences in war roles. This suggests that gendered aggression is not a sufficient explanation of gendered war roles.


Toughening Boys

            The kind of person needed to fight wars is tough, both emotionally and physically. Basic training of new recruits in military forces – which clearly entails socialization for qualities adaptive to war-fighting – essentially is a process of toughening them up. Actual induction into the military often serves as a final stage of toughening for young men, but for boys this toughening starts much earlier in life.


            Rites of passage into manhood vary cross-culturally but often include common elements relevant to war preparation. Gilmore (1990: 11-20) argues that a broad sweep of cultures reflect the central theme that men are made, not born. Men must take actions, undergo ordeals, or pass tests in order to become men. In rites of passage, only select men can achieve “manhood,” and it must be won individually. Rituals typically inflict pain on adolescent males and force them not to cry out, on pain of lifelong shame if they fail. In some especially war-prone societies, men have had to kill an enemy to be considered a man, or to marry. In others, near-universal male conscription marks a passage to manhood. These various passages, based on passing harsh tests bravely, adapt males for war (Goldstein 2001: 264-67).


            Still earlier in their development, boys face harsh teasing and ridicule if they act like “sissies” (i.e., girls – the root of sissy being “sister”) by not being tough enough. Treatment of girlish boys and of boyish girls is not symmetrical. There is a greater urgency to keeping boys in their gender roles.


            “[T]omboy girls seldom receive criticism or rejection from other girls” for associating with boys. Thus “girls have considerable freedom to choose how much contact they will have with boys” (Maccoby 1998: 66). Cultural tomboy themes contain a “mixed message” – abnormality combined with freedom, neither a “full-fledged insult” nor a compliment (Thorne 1993: 112–17, 127–30). (Children themselves seldom use the term “tomboy.”)


            Boys, however, “have less freedom of choice.” (Maccoby 1998: 66). By contrast with “tomboy,” the label “sissy” (applied mostly to gender-deviant boys) has “relentlessly negative connotations” (Thorne 1993: 115–17). Applied to boys in U.S. elementary schools, the term also conveys “gender and sexual deviance.” “In short, a ‘sissy” is a failed male” – synonymous with “girl.” Being called a sissy, or its equivalents, constitutes a “stinging insult,” and men years later find such memories “painful,” in contrast to women who “reminisce quite positively” about being called a tomboy (ibid). Thorne (1993: 168–69) concludes that “the specter of ‘sissies’ ... helps sustain hegemonic masculinity and the structuring of gender as opposition and inequality.”


            It does more than that, I would add. The intense shame and teasing that faces boys who show weakness, cry, or play with girls is functional for the war system. Without these socialization pressures, many boys might follow their own individual instincts (such as crying) or self-interests (such as sometimes playing with girls) and in time of war the society might not have enough tough, hardened, aggressive males available. Boys who act weak in middle childhood, those who fail rites of passage in adolescence, and those who shirk war service or desert in battle, all have the same character of males dropping out from the prescribed gender path, culminating in war service. The steady pressure to toughen up throughout boys’ development, and the intense social pressure to conform to these expectations, reflect the urgency of society’s dilemma of inducing war participation by young men.


Exposure to Violence

            An important aspect of the toughening of boys is to expose them to violence in order to desensitize them. This ultimately lessens the shock of experiencing the violence of battle (and has other effects across society).


            Children traumatized by being in a war zone will likely become the next generation of warriors. Political violence may be “more stressful for children than other forms of violence,” because it threatens social identity by attacking the child’s group (Cairns 1996: 5–8). With the terrorist attacks of September 2001, parts of the United States became a war zone for children, in a way it had not been before.


            In addition to the real thing, U.S. children receive massive doses of media violence which have well-documented effects on aggressive behavior (see Donnerstein chapter in this volume). “[C]onsiderable evidence” shows that “watching such violence is associated with increased aggression in the viewer,” both verbal and physical (Geen 1998: 335). For boys, the effect is robust in various experimental designs, including a 22-year longitudinal study following children from ages 8 to 30 (p. 336). Several studies find that U.S. children watch at least several hours of TV daily, the amount increasing with age (Carlsson-Paige and Levin 1987, 14). The popular children’s TV show, Transformers, contained an attempted murder every 30 seconds on average (p.13). Television portrayals are “more stereotypic in roles and behaviors than the world children see around them everyday” (Ruble and Martin 1998: 982). These stereotypes include those concerning gender and aggression.


            In recent years, questions about effects of media violence on children have also focused on violent video games. Unlike TV and movie violence, these games are interactive. In “shooter” games the screen shows what the player would see as he blasts away at realistic people (creating realistic wounds, such as severed heads). Grossman (1995: 312–16) finds these games identical to the methods used in military training in which soldiers fire at “man-shaped silhouettes that pop up for brief periods ...” This method creates “an automatic, conditioned response ... to the appropriate stimulus...”

            Carlsson-Paige and Levin describe a “dilemma” facing peace-loving teachers whose students engage in war play: Banning such play may be inappropriate to the developmental needs of the children engaged in it, but allowing it may encourage later real-world violence and legitimize war. Parents complain they cannot “turn on the television, visit other children, or go to a supermarket, toy store, or playground with their children without encountering some reminder of war and weapons play” (Carlsson-Paige and Levin 1987: 9). School peer groups often seem to be the source of children’s interests in war play, and “while some girls are attracted to war play, it is most often boys who show a compelling interest” (p.21).


            Most teachers and parents surveyed by Carlsson-Paige and Levin (1987: 43–51) used one of four approaches to the war-play dilemma: ban it, let it be, allow it with set limits, or actively facilitate it. Banning was the most popular approach with the teachers, while parents either used a laissez-faire approach or imposed limits (e.g., not in the house). Carlsson-Paige and Levin (p.66) argue, however, that the facilitation approach is best for both parents and teachers. It lets adults engage the child in dialogue, share the adult’s personal feelings, and model alternatives, while decreasing the salience of TV-based scripts. Among other ideas, they suggest helping girls become involved in war games in non-feminized roles.


            Carlsson-Paige and Levin (1987: 21) argue that part of the appeal of war characters to boys is that they are “clearly defined male models with which to identify.” War figures may also offer boys “alternative models to their nurturing mothers” as they master separation. They “offer boys the concrete, powerful models that they are seeking, as well as opportunities to express the possible anger and frustration they are feeling” (p.22). Boys use war play to “mask ... feelings of helplessness and insecurity. ... The boys who are most passionately involved in war play are often those who are most insecure and in need of support in learning to express and act out a range of feelings” (p.67).


            Many writers have treated present-day boys’ competitive sports as training for war. In the 19th-century United States, some observers thought sports could promote some of the “manly traits” developed in war, without the violence, while others thought sports and militarism complemented each other in this regard (Rotundo 1993: 239–44). Cross-cultural data show a correlation between frequency of war and importance of combative sports in the culture. Montagu (1976: 277) argues that sports do not serve as a release valve or alternative channel for an instinctual aggressive drive, but rather “combative sports ... represent the embodiments of the same theme” as war. But Eibl-Eibesfeldt (1979: 236) concludes that both are true: “Competitive games can in fact divert aggression, but at the same time, they train the aggressive system.”


Childhood Gender Segregation

            Childhood gender segregation also socializes children for war. All-boy groups in middle childhood develop the social interaction patterns used later in armies.


            Gender segregation in children’s peer interaction “appears to be virtually universal in Western and non-Western societies ... although the extent varies ...” (Ruble and Martin 1998: 961). Even young monkeys and apes “show gender segregation in play” (Maccoby 1998: 99). For humans, segregation tends to be more pronounced in cultures with high male dominance (Hartup 1983, 109; Whiting and Edwards 1988).


            Maccoby (1998: 5, 27–29) details the ways that children grow up in largely separate gendered peer cultures – in boys’ and girls’ group activities – with different styles and norms of behavior. Gender segregation is found across different cultures, and is “fairly resistant to change” since intervention effects are temporary. These findings support the possibility that the genders segregate naturally. (Maccoby sees segregation resulting from biological, socialization, and cognitive components.) The key biological aspect is young males’ greater propensity for rough-and-tumble play, which has been shown to result from prenatal exposure to androgens. Since boys play rougher than girls, distinct play styles emerge, and crossing into the opposite gender group is problematical (Maccoby 1998: 103–111, 293, 161, 170, 175).


            Gender segregation varies considerably across settings and cultures, through time, and along the developmental lifespan. Sociologist Barrie Thorne (1993: 31–40) is more focused than Maccoby on U.S. elementary schools, more attuned to variations in patterns that Maccoby treats as categorical, and more optimistic about potentials for change.


            Even if nearly absolute gender segregation is limited to certain settings and ages – such as playgrounds during middle childhood – those settings still provide contexts to develop the social skills needed for all-male armies. Gender segregation need only operate in some settings, not in all of a child’s life, to be functional for war preparation.


            Play themes revolve around aggression more for boys than girls, and boys often assume “the role of a heroic character” and act out fantasies with themes of “danger and righteous combat.” Appropriate props or costumes will be used but in their absence “children improvise.” A 4–8 year-old boy “playing alone will also enact heroic or warlike themes by himself,” which implies that the script does not depend on the all-boy peer group setting, although it may have been learned or practiced there. Girls’ pretend play centers on “cooperative role-taking” around “domestic or school themes.” Girls frequently enact family scenes, taking on either the mother or father (the boys seldom play either role). These scenarios revolve around “preparation and serving of food” and, in mother-child roles, other nurturing activities such as feeding baby, putting on Band-Aids, rocking to sleep, or soothing a hurt child (Maccoby 1998: 41–42).


            Characteristic boys’ play styles and themes are often tied directly to the boys’ future roles in wartime (play-fighting, dominance, heroic themes, and specific war scripts). If “boys’ culture” is seen as functional in socializing males for adult roles, it surely does so most efficiently with regard to war roles, and somewhat less directly with regard to the work roles – such as heavy agriculture or herding – that Maccoby (1998: 270-78) emphasizes, for example. Maccoby (1998: 167) enumerates boys’ preferred play materials (guns, swords, monsters, dinosaurs) and themes (danger, struggle) – a list that seems influenced by the potential future role of boys as warriors.


            The extent of childhood gender segregation is impressive. Boys and girls show a substantial preference for same-gender playmates, increasing through early childhood, solidifying around ages 5–8, and peaking around ages 8–11 (Maccoby 1998: 169, 18–19, 27–29). Already at preschool age, a “substantial degree” of gender segregation has been found in “many different cultural settings” as well as showing “remarkably consistent results” across a variety of psychology studies (Maccoby 1998: 21). In ten “small societies” worldwide, children choose same-gender playmates (excluding sibling play) about two-thirds of the time by ages 3–6 and three-quarters by ages 6–10 – similar proportions to those found in middle-class American children (Whiting and Edwards 1988).


            Gender differentiation occurs primarily in settings that bring together multiple children of similar age, especially with minimal adult structuring of interactions – such as a school playground at recess – and it first develops at the preschool age when parallel play gives way to extended sequences of interactions between children. Children code each other by gender, and react accordingly, whether or not they have met before (Maccoby 1998, 20–24, 27).


            By middle childhood (elementary school years), “patterns of gender segregation become more firmly entrenched ... but only in certain settings” (Maccoby 1998: 23-24). In non-Western, nonindustrialized societies, gender segregation is reinforced by the assignment of girls and boys to different tasks, such as herding and child-care, which place them in same-gender groupings. In Western contexts, such as U.S. elementary schools, adults more often create gender-integrated social structures (e.g., in classroom seating), so same-gender preferences emerge mostly in contexts not structured by adults (e.g., corridors, playgrounds).


            Play in mixed-gender dyads and groups does occur in middle childhood, however. In one study of first-graders, about one-quarter (boys) to one-third (girls) of “affiliative activity” was directed to opposite-gender peers. Maccoby reports the ratio of same-gender with opposite-gender play (age 6½) as 11 to 1, but this omits mixed-gender groups (those with at least two boys and two girls), which make up about one-quarter of social play. In a 1993 study, 4th- and 6th-graders during lunch and recess spent 64 percent of time with same-gender groups, 27 percent with mixed-gender groups, and 1 percent with one or more members of the opposite gender and no other members of their own gender. Nonetheless, even mixed-gender interactions often reinforce gender boundaries, as when boys and girls play on opposite teams, or chase each other.


            Researchers have paid less attention to mixed-gender groups than same-gender ones, and interactions have been studied almost entirely in dyads and occasionally triads (Thorne 1993, 49; Maccoby 1998: 95.). Yet by middle childhood most of the cross-gender contact occurs in larger groups, not dyads.


            Gender segregation is strongest in one developmental phase, middle childhood, and even then only in certain settings. Although peers account for 30 percent of social interactions in middle childhood, compared with 10 percent in two-year-olds, this still leaves 70 percent of interactions that are with siblings, parents, and adults – contexts where gender segregation is far less salient (Rubin, Bukowski, and Parker 1998: 638). Of the three “major sites” of contemporary childhood – families, neighborhoods, and schools – schools have by far the least mixing of girls and boys. Since 90 percent of research on children occurs in schools (owing to ease of access), the psychology literature may overemphasize gender segregation (Thorne 1993: 29, 46).


            Overall, then, childhood gender segregation is important for war, but is less absolute than the gender division regarding combatants. This suggests that gender segregation alone does not explain the gendering of war roles.


            Gender segregation appears to be reinforced at three levels. First, it is self-reinforcing, both by individual preference and by peer socialization. Second, gender rules are enforced by peers, primarily by teasing. Third, adults promote gender segregation (including by promoting divergent gender norms, which reinforce segregation), especially fathers with sons.


            Segregation is enforced by boys more than girls. After age 4, “boys appear to play a more active role in establishing and maintaining the separation of the sexes” (Maccoby 1998: 29). Boys seem more intent on excluding girls from groups than vice versa. Girls are more interested in interacting with boys and pursuing masculine activities (notwithstanding many girls’ wariness of boys’ style) than boys are in girls and feminine activities. Boys’ groups, more than girls’ groups, “exclude and ignore” the other gender and “vigilantly monitor the boundaries.” Boys’ groups “achieve more autonomy” and probably more group cohesion despite higher conflict within the group (Maccoby 1998: 289).


            Asymmetries skew gender segregation towards male dominance. Boys control as much as ten times the space in school playgrounds as girls do, and more frequently invade and disrupt the remaining girls’ spaces. Nonetheless, children’s power dynamics are “extremely complex” (Thorne 1993: 82–84, 39).


            Parents’ treat sons and daughters differently. This does not have to do with loving one more, nor with giving warmth, responsiveness, or restrictiveness differently. Nor do parents promote independence more in young boys and attachment in girls. Parents are somewhat rougher with boys, but not dramatically so. The major difference, rather, is in promoting specific gender-appropriate behavior, rather narrowly defined, and discouraging cross-gender behavior. Four routes of parent-to-child gender socialization in early childhood are offering gendered toys, supporting gender-appropriate play themes, playing more roughly with boys than girls, and talking less about feelings with boys (Maccoby 1998, 124–27).


            Maccoby’s longitudinal study found “few consistent differences” in how parents treated girls and boys up to age two. By age four, however, “[b]oth parents offered masculine toys more to boys, feminine ones more to girls, and ... [chose different] play themes that were acted out jointly by the parent-child pair ...” (Maccoby 1998: 124–25). Providing gender-linked toys may stimulate certain themes and scripts which in turn trigger other gendered patterns of social interaction. Even without parental direction, however, preschool children’s toy choices mostly follow gender lines.


            Parents both talk less about feelings with boys, and “actively suppress emotional displays” in them. This may in part derive from boys’ greater proneness to “impulsive emotional outbursts” (as girls more quickly develop self-regulation), and their lower compliance to their mothers’ demands (found across cultures and age groups). However, parents also find it “especially important for boys not to display the kind of weakness or vulnerability or ‘babyishness’ that is implied by crying ...” (Maccoby 1998: 139).


            Parents also interpret emotions differently based on gender. In lab experiments, adults gave different assessments of videotaped infants and toddlers depending if they thought the child was male or female. A toddler named Chris was rated as “stronger, more assertive, and more aggressive” by adults watching the videotape who thought Chris was a boy than adults told that Chris was a girl. An infant’s videotaped reaction to a jack-in-the-box was more often rated as “anger” when the infant was labeled male and “fear” when labeled female (Cummings et al. 1986: 185; Coie and Dodge 1998: 790).


            Fathers in contemporary U.S. culture enforce gender roles on young children more than mothers do, and on boys more than on girls. In one study where toddlers started playing in a room with gender-typed toys and then their parents joined them, mothers intervened to discourage cross-gender play for boys and girls equally, whereas fathers disapproved five times as often of boys’ cross-gender play than that of girls (Campbell 1993, 20–21). Snow, Jacklin, and Maccoby (1983: 229) report that fathers interacted differently with their 12-month-old boys than their girls. In their fathers’ presence, girls played with dolls about twice as often as boys, while boys touched tempting objects twice as often as girls (these were the two most significant gender differences). In their child’s presence, fathers gave dolls two or three times as often to girls as boys, and gave prohibitions 2–3 times as often to boys as girls (again the two most significant items).


            A comparison of 39 studies found a “clear” trend showing that fathers differentiate between sons and daughters more than mothers do, often by being “stricter with boys.” Fathers “maintain a stance of male dominance toward their sons, as an older boy might do,” thereby inducting them into the male status hierarchy, despite fathers’ sometimes contradictory role as nurturing parent. Fathers’ greater role in socializing daughters for gender, if any, is less clear (Maccoby 1998, 142–43).


            Fathers’ own “masculinity” influences boys’ development in ways that are unclear at best. Overall, the most important variable seems to be fathers’ active involvement rather than any particular high rating on some scale of “masculinity.” (Archer and Lloyd 1985: 221–22).


            These various active efforts by parents to socialize children into gender roles reinforce the children’s tendency towards self-segregation, a tendency that in turn results from divergent play styles arising from differences in rough-and-tumble play. The reinforcement of childhood segregation plays a functional role, along with toughening up and desensitizing boys, in preparing a society to field an army if necessary or desirable.




            The ubiquity of war creates various social dilemmas, of which a central problem is motivating individuals to sacrifice by becoming combatants. The war system, historically and still today, is extremely rigid in its gender asymmetry regarding such participation. In this chapter, I have tried to show that several aspects of the gendered socialization of aggression plausibly play a functional role in turning boys into male combatants – something that does not come naturally. In each area, gender divisions were somewhat more nuanced and less stark than the gender divide in war roles. This suggests that war may play a more primal role in the connection of gender and aggression than most past psychology research gives it.


            On the whole, I believe, individual boys would be better off without being socialized for combat participation. But their societies might not be. Marvin Harris (1974: 85–87) argues that although women could subdue their sons, they cannot control their enemies’ sons: “As soon as males ... bear the burden of intergroup conflict, women have no choice but to rear large numbers of fierce males of their own.” If the threat of potential war is indeed a central cause of the gendered socialization of aggression, then efforts to reduce war and manage conflict could affect developmental psychology in interesting ways.


            The connections among gender, aggression, and war are complex. I have argued that causality runs both ways, with war as both cause and effect of gender patterns in aggression. Above all, war is an important influence – even in “peacetime” – that should not be omitted from psychologists’ models, theories, and experiments.




Archer, John. 1991.” The Influence of Testosterone on Human Aggression.” British Journal of Psychology 82: 1–28.


Archer, John and Barbara Lloyd. [1982] 1985. Sex and Gender. Revised North American Edition. Cambridge University Press.


Baron, Robert A. and Deborah R. Richardson. 1994. Human Aggression. 2nd Edition. New York: Plenum.


Bernhardt, P.C., J.M. Dabbs, J.A. Fielden, and C.D. Lutter. 1998. “Testosterone Changes During Vicarious Experiences of Winning and Losing among Fans at Sporting Events. Physiology and Behavior 65, 1 (August): 59–62.


Bourke, Joanna. 1999. An Intimate History of Killing: Face-to-Face Killing in Twentieth-Century Warfare. New York: Basic.


Cairns, Ed. 1996. Children and Political Violence. Cambridge, MA: Blackwell.


Cairns Robert B. and Beverly D. Cairns. 1994. Lifelines and Risks: Pathways of Youth in Our Time. Cambridge University Press.


Campbell, Anne. 1993. Men, Women, and Aggression. New York: Basic.


Carlsson-Paige, Nancy and Diane E. Levin. 1987. The War Play Dilemma: Balancing Needs and Values in the Early Childhood Classroom. New York: Teachers College Press.


Chang, Iris. 1997. The Rape of Nanking: The Forgotten Holocaust of World War II. New York: Basic.


Coie, John D. and Kenneth A. Dodge. 1998. “Aggression and Antisocial Behavior.” In William Damon and Nancy Eisenberg, eds. 1998. Handbook of Child Psychology. 5th Edition. Volume 3: Social, Emotional, and Personality Development. New York: Wiley: 779–862.


Crick, Nicki R., Juan F. Casas, and Hyon-Chin Ku. 1999. “Relational and Physical Forms of Peer Victimization in Preschool.” Developmental Psychology 35, 2 (March): 376–85.


Crick, Nicki R. and Jennifer K. Grotpeter. 1995. “Relational Aggression, Gender, and Social-Psychological Adjustment.” Child Development 66, 3 (June): 710–22.


Cummings, E. Mark, Barbara Hollenbeck, Ronald Iannotti, Marian Radke-Yarrow and Carolyn Zahn-Waxler. 1986. “Early Organization of Altruism and Aggression: Developmental Patterns and Individual Differences.” In Carolyn Zahn-Waxler, E. Mark Cummings, and Ronald Iannotti, eds., Altruism and Aggression: Biological and Social Origins. Cambridge University Press: 165–188.

Dabbs, James McBride and Mary Godwin Dabbs. 2000. Heroes, Rogues, and Lovers: Testosterone and Behavior. New York: McGraw Hill.


Davis-Kimball, Jeannine. 1997. “Warrior Women of the Eurasian Steppes.” Archaeology 50, 1: 44–48.


Eagly, Alice H. and Valerie J. Steffen. 1986. “Gender and Aggressive Behavior: A Meta-Analytic Review of the Social Psychological Literature.” Psychological Bulletin 100, 3: 309–30.


Eibl-Eibesfeldt, Irenäus. 1979. The Biology of Peace and War: Men, Animals, and Aggression. New York: Viking.


Ember, Carol R. 1978. “Myths About Hunter-Gatherers.” Ethnology 17, 4: 439–48.


Ember, Carol R. and Melvin Ember. 1997. “Violence in the Ethnographic Record: Results of Cross-Cultural Research on War and Aggression.” In Debra L. Martin and David W. Frayer, eds., Troubled Times: Violence and Warfare in the Past. Langhorne, PA: Gordon & Breach: 1–20.


Ember, Melvin and Carol R. Ember. 1994. “Cross-Cultural Studies of War and Peace: Recent Achievements and Future Possibilities.” In S.P. Reyna and R.E. Downs, eds., Studying War: Anthropological Perspectives. Langhorne, PA: Gordon & Breach: 185–208.


Geen, Russell G. 1998. “Aggression and Antisocial Behavior.” In Daniel T. Gilbert, Susan T. Fiske, and Gardner Lindzey, eds., The Handbook of Social Psychology. 4th Edition. Volume I1. Boston: McGraw-Hill: 317–56.


Gilmore, David D. 1990. Manhood in the Making: Cultural Concepts of Masculinity. New Haven, CT: Yale University Press.


Goldschmidt, Walter. 1989. “Inducement to Military Participation in Tribal Societies.” In Paul R. Turner and David Pitt, eds., The Anthropology of War and Peace: Perspectives on the Nuclear Age. Granby, MA: Bergin & Garvey: 15–31.


Goldstein, Joshua S. 2001. War and Gender: How Gender Shapes the War System and Vice Versa. Cambridge: Cambridge University Press.


Grossman, Dave. 1995. On Killing: The Psychological Cost of Learning to Kill in War and Society. Boston: Little Brown.


Hall, Roberta L. 1985. “The Question of Size.” In Roberta L. Hall, ed. Male-Female Differences: A Bio-Cultural Perspective. New York: Praeger.


Hartup, Willard W. 1983. “Peer Relations.” In Paul H. Mussen, ed., Handbook of Child Psychology, 4th Edition. Volume IV: Socialization, Personality, and Social Development. E. Mavis Hetherington, volume editor. New York: Wiley: 103–96.


Harris, Marvin. 1974. Cows, Pigs, Wars and Witches: The Riddles of Culture. New York: Random House.


Herman, Judith Lewis. 1992. Trauma and Recovery. New York: Basic.


Hyde, Janet Shibley. 1986. “Gender Differences in Aggression.” In Janet Shibley Hyde and Marcia C. Linn, eds. The Psychology of Gender: Advances Through Meta-Analysis. Baltimore: Johns Hopkins University Press : 51–66.


Maccoby, Eleanor E. 1998. The Two Sexes: Growing Up Apart, Coming Together. Cambridge, MA: Belknap/Harvard University Press.


Maccoby, Eleanor Emmons and Carol Nagy Jacklin. 1974. The Psychology of Sex Differences. Stanford University Press.


Mazur, Allan and Alan Booth. 1998. “Testosterone and Dominance in Men.” Behavioral and Brain Sciences 21, 3 (June): 353–63.


Mazur, Allan and T.A. Lamb. 1980. “Testosterone, Status, and Mood in Human Males.” Hormones and Behavior 41: 236–46.


Mazur, Allan, Alan Booth and James M. Dabbs, Jr. 1992. “Testosterone and Chess Competition.” Social Psychology Quarterly 55, 1: 70–77.

Monaghan, Edward P. and Stephen E. Glickman. 1992. “Hormones and Aggressive Behavior.” In Jill B. Becker, S. Marc Breedlove, and David Crews, eds. 1992. Behavioral Endocrinology. Cambridge, MA: MIT Press: 261–86.


Montagu, Ashley. 1976. The Nature of Human Aggression. Oxford University Press.


Ross, Marc Howard. 1990. “Children and War in Different Cultures.” In Francesca M. Cancian and James William Gibson, eds. Making War/Making Peace: The Social Foundations of Violent Conflict. Belmont, CA: Wadsworth: 51–63.


Ross, Marc Howard. 1993. The Culture of Conflict: Interpretations and Interests in Comparative Perspective. New Haven, CT: Yale University Press.


Rotundo, E. Anthony. 1993. American Manhood: Transformations in Masculinity from the Revolution to the Modern Era. New York: Basic.


Rubin, Kenneth H., William Bukowski, and Jeffrey G. Parker. 1998. “Peer Interactions, Relationships, and Groups.” In William Damon and Nancy Eisenberg, eds. 1998. Handbook of Child Psychology. 5th Edition. Volume 3: Social, Emotional, and Personality Development. New York: Wiley: 619–700.


Ruble, Diane N. and Carol Lynn Martin. 1998. “Gender Development.” In William Damon and Nancy Eisenberg, eds. 1998. Handbook of Child Psychology. 5th Edition. Volume 3: Social, Emotional, and Personality Development. New York: Wiley: 993–1016.


Thompson, Wendy M., James M. Dabbs, Jr. and Robert L. Frady. 1990. “Changes in Saliva Testosterone Levels During a 90-Day Shock Incarceration Program.” Criminal Justice and Behavior 17, 2: 246–52.


Thorne, Barrie. 1993. Gender Play: Girls and Boys in School. New Brunswick, NJ: Rutgers University Press.


Whiting, Beatrice B. and C.P. Edwards. 1988. Children of Different Worlds: The Formation of Social Behavior. Cambridge, MA: Harvard University Press.


Whiting, Beatrice B. and John W. M. Whiting [in Collaboration with Richard Longabaugh]. 1975. Children of Six Cultures: A Psycho-Cultural Analysis. Cambridge: Harvard University Press.


Wrangham, Richard and Dale Peterson. 1996. Demonic Males: Apes and the Origins of Human Violence. New York: Houghton Mifflin.