Cannibals and Kings

by Marvin Harris

  Hence when spears began to be thrown, they were thrown by individuals for reasons based on individual disputes.

  Since the old men did most of the spear throwing, marksmanship tended to be highly inaccurate.

  Not infrequently the person hit was some innocent noncombattant or one of the screaming old women who weaved through the fighting men, yelling obscenities at everybody, and whose reflexes for dodging spears were not as fast as those of the men.… As soon as somebody was wounded, even a seemingly irrelevant crone, fighting stopped immediately until the implications of this new incident could be assessed by both sides.

  I do not mean to liken hunter-collector warfare to slapstick comedy. W. Lloyd Warner reported high rates of fatalities for at least one other northern Australian hunter-collector group called the Murngin. According to Warner, 28 percent of adult male Murngin deaths were caused by wounds inflicted on the battlefield. Bear in mind that when a whole band contains only ten adult males, one death per battle every ten years is all it takes to rack up this kind of body count.

  Warfare after the development of agriculture probably became more frequent and more deadly. Certainly the scale of combat increased. Permanent houses, food-processing equipment, and crops growing in the fields sharpened the sense of territorial identity. Villages tended to remain enemies across the generations, repeatedly attacking and plundering, seeking to rout each other from their territories. Among the village-dwelling Dani of West Irian, New Guinea, warfare has a regulated “nothing fight” phase, similar to that of the Tiwi, in which there are few casualties. But the Dani also launch all-out sneak attacks that result in the destruction and rout of whole villages and the deaths of several hundred people at a time. Karl Heider estimates that 29 percent of Dani men die as a result of injuries sustained during raids and ambushes. Among the Yanomamo village horticulturalists along the Brazil-Venezuela border, raids and ambushes account for 33 percent of adult male deaths from all causes. Since the Yanomamo are an important test case, I’ve set aside a whole chapter for them following this one.

  The reason some anthropologists deny the reality of high levels of combat among band and village peoples is that the populations involved are so small and spread out as to make even one or two intergroup killings seem utterly irrational and wasteful. The Murngin and the Yanomamo, for example, have population densities of less than one person per square mile. But even groups with such low densities are subject to reproductive pressure. There is considerable evidence indicating that the balance between people and resources does in fact lie behind band and village warfare and that the origin of this scourge stems from the inability of preindustrial peoples to develop a less costly or more benign means of achieving low population densities and low rates of population growth.

  Before I discuss this evidence, let me review some alternative explanations and show why I think none of these is adequate to the task. The major alternatives include war as solidarity, war as play, war as human nature, and war as politics.

  War as solidarity. According to this theory, war is the price that is paid for building up group togetherness. Having external enemies creates a sense of group identity and enhances esprit de corps. The group that fights together stays together.

  I must admit that aspects of this explanation are compatible with one based on reproductive pressure. If a group is undergoing stress caused by intensification, declining efficiencies, and increased abortions and infanticides, the deflection of aggressive behavior onto neighboring bands or villages would certainly be preferable to letting it fester within the community. I do not doubt that deflecting aggressive behavior onto foreigners can act as a “safety valve.” What this approach fails to explain, however, is why the safety valve has to be so deadly. Wouldn’t verbal abuse, mock combat, or competitive sports be less costly ways of achieving solidarity? The claim that mutual slaughter is “functional” cannot be based on some vague abstract advantage of togetherness. It must be shown how and why such deadly recourse is necessary to prevent an even more deadly consequence—how, in other words, the benefits of war outweigh its costs. No one ever has shown or will be able to show that the consequences of less solidarity would be worse than deaths in combat.

  War as play. Some anthropologists have tried to balance the material costs and benefits of warfare by representing it as a pleasureable, competitive team sport. If people actually enjoy risking their lives in combat, then war can be materially wasteful but psychologically valuable and the puzzle disappears. I agree that people, especially men, are frequently brought up to believe that warfare is a zestful or ennobling activity and that one should enjoy stalking and killing other human beings. Many of the mounted Indians of the Great Plains—the Sioux, the Crow, the Cheyenne—kept a tally of their acts of bravado in war. A man’s reputation lay in counting coups. They gave the most points not to the warrior with the highest body count, but to the one who took the most risks. The greatest feat of all was to sneak in and out of an enemy camp without being detected. But indoctrination for military bravado among band and village peoples was not always successful. The Crow and other Indians of the Great Plains took care of their pacifists by letting them put on women’s clothing and making them serve as attendants to the warriors. And even the bravest of warriors, as among the Yanomamo, have to be emotionally prepared for the fray by performing rituals and taking drugs. If people can be taught to value war and to enjoy stalking and killing other human beings, one must also grant that they can be taught to hate and fear war and to be revolted by the spectacle of human beings trying to kill each other. Both kinds of teaching and learning actually do take place. So if warlike values cause wars, the crucial problem becomes that of specifying the conditions under which people are taught to value war rather than to abhor it. And this the war as play theory cannot do.

  War as human nature. A perennially favorite way for anthropologists to avoid the problem of specifying the conditions under which war will be regarded as a valuable or an abhorrent activity is to endow human nature with an urge to kill. War occurs because human beings, especially males, have a “killer instinct.” We kill because such behavior has been proved successful from the standpoint of natural selection in the struggle for existence. But war as human nature runs into difficulties as soon as one observes that killing is not universally admired and that intensity and frequency of warfare are highly variable. I fail to see how anyone can doubt that these variations are caused by cultural rather than genetic differences, since sharp reversals from extremely warlike to peaceful behavior may occur in one or two generations without any genetic changes whatsoever. The Pueblo Indians in the Southwest of the United States, for example, are known to contemporary observers as peaceful, religious, unaggressive, cooperative peoples. Yet not so long ago they were known to the Spanish governor of New Spain as the Indians who tried to kill every white settler they could get their hands on and who burned every church in New Mexico, together with as many priests as they could lock inside and tie to the altars. One need merely recall the astonishing flip-flop in post-World War II Japanese attitudes toward militarism, or the sudden emergence of the Israeli survivors of Nazi persecution as the leaders of a highly militarized society, to grasp the central weakness of the war as human nature argument.

  Obviously it is part of human nature to be able to become aggressive and to wage war. But how and when we become aggressive is controlled by our cultures rather than by our genes. To explain the origin of warfare one must be able to explain why aggressive responses take the specific form of organized intergroup combat. As Ashley Montagu has warned us, even in infrahuman species killing is not the goal of aggression. There are no drives or instincts or predispositions in human beings to kill other human beings on the battlefield, although under certain conditions they can easily be taught to do so.

  War as politics. Another recurrent explanation of war holds that armed conflict is the logical outcome of an attempt of one group to protect or increase its political, social
, and economic welfare at the expense of another group. War occurs because it leads to the expropriation of territory and resources, the capture of slaves or booty, and the collection of tribute and taxes—“To the victor belong the spoils.” The negative consequences for the vanquished can simply be written off as a miscalculation—“the fortunes of war.”

  This explanation makes perfectly good sense in relation to the wars of history, which are primarily conflicts between sovereign states. Such wars clearly involve the attempt on the part of one state to raise its standard of living at the expense of others (although the underlying economic interests may be covered up by religious and political themes). The form of political organization which we call the state came into existence precisely because it was able to carry out wars of territorial conquest and economic plunder.

  But band and village warfare lacks this dimension. Band and village societies do not conquer territories or subjugate their enemies. Lacking the bureaucratic, military, and legal apparatus of statehood, victorious bands or villages cannot reap benefits in the form of annual taxes or tribute. And given the absence of large amounts of stored foods or other valuables, the “spoils” of war are not very impressive. Taking prisoners and making slaves of them is impractical for a society that cannot intensify its system of production without depleting its resource base and that lacks the organizational capacity to exploit a hostile, underfed labor force. For all of these reasons, the victors in pre-state wars often returned home carrying a few scalps or heads as trophies or with no spoils at all—except the right to boast about how manly they were in combat. In other words, political expansion cannot explain warfare among band and village societies because most such societies do not engage in political expansion. Their entire mode of existence is dominated by the need not to expand in order to preserve the favorable ratio of people to resources. Hence we must look to the contributions of warfare to the conservation of favorable ecological and demographic relationships in order to understand why it is practiced by band and village societies.

  The first such contribution is the dispersal of populations over wider territories. While bands and villages do not conquer each other’s lands the way states do, they nonetheless destroy settlements and rout each other from portions of the habitat that they would otherwise jointly exploit. Raids, routs, and the destruction of settlements tend to increase the average distance between settlements and thereby lower the overall regional density of population.

  One of the most important benefits of this dispersion—a benefit shared by both victor and vanquished—is the creation of “no man’s lands” in areas normally providing game animals, fish, wild fruits, firewood, and other resources. Because the threat of ambush renders them too dangerous for such purposes, these “no man’s lands” play an important role in the overall ecosystem as preserves of plant and animal species that might otherwise be permanently depleted by human activity. Recent ecological studies show that in order to protect endangered species—especially large animals that breed slowly—very extensive refuge areas are needed.

  The dispersal of populations and the creation of ecologically vital “no man’s lands” are very considerable benefits which derive from intergroup hostilities among band and village peoples despite the costs of combat. With one proviso: having dispersed the enemy camps and settlements, the victors cannot allow the population of their camps and settlements to increase to the point where game and other resources are threatened by their own population growth and intensification effort. Warfare under pre-state conditions cannot satisfy this proviso—at least not through the direct effect of combat deaths. The problem is that the combatants are almost always males, which means that most of the battle fatalities are men. Warfare causes only 3 percent of adult female deaths among the Dani and 7 percent among the Yanomamo. Moreover, war-making band and village societies are almost always polygynous, that is, the husband services several wives. Thus there is no possibility that warfare alone can depress the rate at which a band or village—especially if it is victorious—grows and depletes its environment. Male combat deaths, like geronticide, can produce short-run relief from population pressure, but they cannot influence overall trends as long as a few polygynous male survivors continue to service all the noncombatant females. The biological reality is that most males are reproductively superfluous. As Joseph Birdsell has put it, the fertility of a group is determined by the number of its adult women, rather than by its adult men. “Undoubtedly, one able-bodied male could keep ten women continuously pregnant.” This is obviously a conservative statement, since at ten pregnancies per woman the male in question would have only a maximum of 100 children while many Arab sheiks and Eastern potentates seem to encounter no great difficulty in siring well over 500 children.

  But let us follow Birdsell’s logic, which is unassailable even though it is based on the hypothetical example of one man and only ten women:

  This would produce the same number of births as if the group consisted of ten men and ten women. But if we can imagine a local group consisting of ten men and only one woman, the birth rate would necessarily be ten percent of the former example. The number of women determines the rate of fertility.

  As I will show, warfare does drastically affect the number of women and thus does have a powerful effect on the human crop. But the manner in which it achieves this has hitherto not been understood.

  Before I explain how warfare limits the rate at which settlements grow, I want to emphasize one point. The twin demographic effects that warfare produces among band and village societies are not characteristic of state-level military complexes. For the moment, I will address myself only to the origin of pre-state warfare. Among state-level societies warfare may disperse populations, but it seldom depresses their rate of growth. Each of the major wars of this century—World Wars I and II, Korea and Vietnam—has failed to reduce the long-term rate of growth of the combatant populations. While it is true that during World War I the deficit between the projected and actual population for Russia reached 5 million, it took only ten years for this to be overcome. Even short-time population growth may be unaffected. All through the decade of the Vietnam War, the population of Vietnam grew at the phenomenal rate of 3 percent per year. That warfare does not automatically depress the rate of population growth should be obvious from European history. Scarcely a decade went by during the past three centuries without large-scale combat, and yet the European population soared from 103 million in 1650 to 594 million in 1950. One might more readily conclude that European wars—and wars of states in general—have been part of a system for stimulating rapid population growth.

  What no one seems to have realized is that, unlike state societies, bands and villages were exceptional in their use of warfare to achieve very low rates of population growth. They achieved this not primarily through male combat deaths—which, as we have just seen, are always easily compensated for by calling upon the remarkable reproductive reserves of the human female—but by another means that was intimately conjoined with and dependent upon the practice of warfare yet was not part of the actual fighting. I refer to female infanticide. Warfare in band and village societies made the practice of infanticide sex-specific. It encouraged the rearing of sons, whose masculinity was glorified in preparation for combat, and the devaluation of daughters, who did not fight. This in turn led to the limitation of female children by neglect, abuse, and outright killing.

  Studies recently carried out by William Divale show that among band and village societies practicing warfare when they were first censused the number of males aged fourteen or under greatly exceeded the number of females in the same age bracket. Divale found that the ratio of boys to girls was 128:100, whereas the ratio of adult men to women was 101:100. Since the expected world-wide sex ratio at birth is 105 males to 100 females, the discrepancy between 105 and 128 is a measure of preferential treatment of male children and the drop to 101:100 is probably a measure of the rate of adult male combat deaths. Thi
s interpretation was strengthened when Divale compared the sex ratios of groups that had practiced warfare at progressively more remote periods in the past with those that were actively practicing warfare when they were first censused.

  For populations that were censused five to twenty-five years after warfare had been stopped, usually by colonial authorities, the average sex ratios were 113 boys and 113 adult men per 100 girls and 100 adult women. (The increase in the adult sex ratio from 101:100 when war was present to 113:100 when war had been stopped is probably the result of the survival of males who previously would have been killed in war.) Among populations that were censused more than twenty-five years after warfare, the sex ratio of persons fifteen years and younger was still lower—106:100, approximating the world norm of 105:100, at birth.

  These shifts are even more dramatic when the reported frequency of any kind of infanticide, male or female, and the presence of warfare are taken into consideration. Among populations who were still practicing warfare at the time of census and who according to ethnographers’ reports were commonly or occasionally practicing some kind of infanticide, the average sex ratio among the young was 133 boys to 100 girls. Yet among adults it declined to 96 men to 100 women. For populations in which warfare had been stopped twenty-five or more years prior to the census and in which infanticide was reported as not common or not practiced, the sex ratio among the young was 104 boys to 100 girls and 92 men to 100 women.

  I am not suggesting that war caused female infanticide or that the practice of female infanticide caused war. Rather, I propose that without reproductive pressure neither warfare nor female infanticide would have become widespread and that the conjunction of the two represents a savage but uniquely effective solution to the Malthusian dilemma.