However, some East African tribes did recognize surrender because the practice of ransoming prisoners with cattle was common. Among the Meru herdsmen of Kenya, a warrior wishing to surrender lifted his spear above his head and shouted “Take cattle!” But if his opponents had deaths to avenge, acceptance of his capitulation was by no means a foregone conclusion. The custom of capturing adult males and incorporating them into a tribe was extremely rare anywhere.8 The Shawnee and Fox tribes of the United States (and very occasionally a few other tribes in the Northeast) spared only those male captives who had survived the hardships and tortures inflicted during their journey to their captors’ village and who were immediately claimed by families who needed replacements for war casualties. A few South American petty chiefdoms saved some captive young men and married them to the daughters of their captors, in order to incorporate them as a despised “servile” class. The Nuer of Sudan adopted boys captured from their enemies (the Dinka), and women of marriageable age and girls were incorporated less formally. On the other hand, old women and babies captured in Nuer raids were clubbed to death and their bodies burned with the Dinka huts. Dinka adult males were simply killed.9 Groups that used or sold war captives as slaves usually preferred to subjugate women and children and therefore immediately dispatched all adult male captives.10 In general, the primitive warrior had only three means of surviving combat: an arranged truce, victory, or (in defeat) fleetness of foot.
In some primitive societies, women were spared injury or capture by enemy warriors.11 Even in societies where women were often slain in the small raids and rare wild slaughters attending massacres and routs, they could enjoy remarkable immunity from harm on formal battlefields. In Kapauku battles in New Guinea, for instance, married women wandered freely about the battlefield collecting arrows to resupply their men, “as if they were harvesting potatoes or cucumbers,” and even acting as scouts or lookouts. Unmarried Kapauku girls had to be more circumspect because if caught by the enemy, they might be raped. When Tuareg tribesmen of the Sahara were defeated near their encampments, they bolted, leaving their women and children in the hands of their enemies. This behavior reflected the Tuaregs’ expectation, given their own customs, that women and children were inviolable in warfare.12 Such chivalrous behavior toward women and children was, however, not the norm among non-state groups.
The capture of women was one of the spoils of victory—and occasionally one of the primary aims of warfare—for many tribal warriors. In many societies, if the men lost a fight, the women were subject to capture and forced incorporation into the captor’s society. Most Indian tribes in western North America at least occasionally conducted raids to capture women.13 The social position of captive women varied widely among cultures, from abject slaves to concubines to secondary wives to full spouses. In a few cases, female captives could be ransomed or, of course, escape.14 In situations where ransom or escape were not possible, the treatment of captive young women amounted to rape, whether actual violence was used against them to enforce cohabitation with their captors or was only implicit in their situation.
Sometimes the number of captive women held by a group represented a considerable proportion of its female population. According to their traditions, the Island Caribs of the Lesser Antilles had conquered these islands a century before Columbus by exterminating the resident Arawak men and taking the women for wives. After a few generations, there developed a peculiar linguistic situation in which the women and children spoke Arawak to one another, but the men spoke a corrupt form of Carib among themselves and to the women.15 Although the loss of even a large percentage of males will have no direct influence on a group’s demographic fortunes (whatever the effect on its military viability may be), the loss or gain of fertile women can mean the difference between population decline and growth.16
Female captives were also very valuable economically. In many societies, women’s labor provided the greater part of staple food. In California, acorns (a dietary staple) were gathered and processed by women. On the Pacific Northwest Coast, women performed the time-consuming but essential work of cleaning and drying salmon, which could be caught only during brief annual runs but were a staple food year-round. Since salmon were not difficult to locate or to catch, one man could supply several women with full-time work. Throughout Melanesia, gardening and pig rearing were female specialties. The widespread practice of polygamy indicates that many societies found that having several adult women in a household was not burdensome but was usually an economic boon. It became even more of an advantage if the additional women could be acquired without the costs of a bride-price or interfering in-laws.
Of course, many tribal societies took no prisoners and retained no captives of any sex or age.17 The Chemehuevi of the Southwest and several tribes in California spared no one. Perhaps the harshest treatment of captives was meted out in Polynesia. The Tahitians are described as leaving enemy children pinned to their mothers with spears or “pierced through the head and strung on cords.” The Maoris sometimes disabled captive women so that they could not escape, permitting the warriors to rape, kill, and eat them when it was more convenient to do so. Even in societies where captives were taken, once general killing started it could be difficult to stop. For example, in an Asmat head-hunting raid in New Guinea, anyone interested in saving a woman or child as a captive (something rarely done) experienced considerable difficulty in preventing his overexcited comrades from dispatching his chosen prisoners.18
In general, nonstate groups preserved the lives of captives only when some material benefit would accrue; this approach generally limited the persons spared to women and children. States, by contrast, often have a strong material interest in preserving the lives of defeated enemies—even adult males—because they can become tax- and tribute-paying subjects, serfs, or slaves. The life-preserving rituals of formal surrender and widespread official distaste for killing noncombatants are expressions of this interest. Economically, the state is usually best served by the submission of its enemies, not by genocide. The atrocities that do occur in civilized warfare usually happen when commanders lose control of their soldiers, whose primary motive may be the primitive one of avenging combat losses or previous real or fictive enemy atrocities. And slaughters of noncombatants can occur as a matter of policy, when the policymakers themselves are consumed by ethnic hatred or when they make a calculated attempt to use state terrorism to cow a conquered populace.
The reaction of the German government during the Herero-Nama uprising in Southwest Africa in 1904 is an example of the self-interested mercy of states and of the conditions under which it fails. The local military governor, General von Trotha, issued an extermination order against the Hereros. The imperial chancellor and the German colonial office successfully demanded that this order be countermanded by the Kaiser: it was inhumane, was bad for public relations, and (perhaps most important) would “undermine the potential for development” by eliminating native labor. The governor, his troops, and the German colonists paid little heed to the Kaiser’s order, however. When the fighting ended several years later, only one-half of the Nama and one-sixth of the Herero had survived.19 Precisely this weakness of state control over frontier “militias” made massacres of native peoples more common by such agents than by the “regular” forces of the state. Indeed, the most notorious massacres of North American Indians, such as those at Sand Creek and Camp Grant, and the only actual genocides (that is, complete extinction of a tribe primarily by homicide) during the European conquest were all inflicted by local militias.20 In many respects, these frontier struggles played out as tribal wars in which one tribe happened to be composed of European colonists. In general, though, the prospects for the defeated were slightly brighter (if still dim) if they were vanquished by a state than by a nonstate society.
Only the “rules of war,” cultural expectations, and tribal or national loyalties make it possible to distinguish between legitimate warfare and atrocities. Is there any beha
vioral difference between Caesar’s extermination of the Bituriges at Bourges, the slaughter of Minnesota settlers by the Sioux in 1862, the massacres by the U.S. Army at Wounded Knee and My Lai, the Allied air strikes at Dresden and Hiroshima, the massacres committed by Japanese soldiers in Nanking and Manila, and the similar accomplishments of primitives described earlier, except the body counts and the assignment of our sympathies with the perpetrators or the victims? Apologists for such massacres always claim that their perpetrators were “provoked.” But war always seems full to overflowing with provocations. At any rate, the treatment of captives and prisoners by nonstate groups has usually and literally been atrocious.
WAR DEATHS
Citizens of modern states tend to believe that everything they do is more efficient and effective than the corresponding efforts of primitives or ancients. Given the neo-Rousseauian tenor of the present day, this expectation about modern civilization finds ready acceptance in relation to distasteful or harmful behavior. Therefore, it comes as a shock to discover that the proportion of war casualties in primitive societies almost always exceeds that suffered by even the most bellicose or war-torn modern states.
Actual casualty figures from primitive warfare are scarce, and only in the past few decades have ethnographers attempted to collect such information. Figure 6.1 compares these casualty rates with those of the most war-torn modern states. Following the practice of several ethnographers, to facilitate comparison, these war death rates are expressed as annual percentages of mean population. Another measure of the deadliness of warfare is the proportion of all deaths caused by war; these figures are given in Figure 6.2. By either measure, primitive warfare was much deadlier than its modern counterpart.21 The death rates shown for civilized states overestimate the deadliness of combat, since most war deaths were caused by disease until very recently. For example, two-thirds of the deaths suffered by the Union armed forces during the Civil War were due to disease.22 Such disease casualties are included in the war death rates for civilized states but not in those of primitive groups. Moreover, many civilized war deaths were the result of accidents involving horses, vehicles, and weapons. For example, approximately 20 percent of British deaths in the Crimean War and 14 percent in the Boer War were accidental.23 The deaths recorded for the primitive groups were all the direct result of wounds suffered in combat and inflicted by the enemy. Were such noncombat deaths deleted from the civilized rates given in Figure 6.1, the terrible deadliness of primitive compared with modern combat would be even more one-sided.
Figure 6.1 War fatality rates (percentage of population killed per annum) for various prestate and civilized societies (see Appendix, Table 6.1).
But what of civilian deaths from disease or starvation resulting from the disruptions and dislocations caused by war? Again such deaths are included in the civilized rates but not in those of the primitive groups. These are difficult to calculate for modern states, and no figures are available for any primitive society. However, they were probably just as common in primitive warfare as in the civilized variety. For example, the ethnographer of the Mae Enga of New Guinea describes the wartime consequences of the “sudden and forced movements of women and children, the elderly and the ill, over difficult terrain in bleak and often wet weather”:
Figure 6.2 Percentage of male (upper) and all (lower) deaths caused by warfare in various societies (see Appendix, Table 6.2).
We simply do not know how many infants and old people succumb to pneumonia in these flights, how many refugees are drowned when trying to cross boulder-strewn torrents, how many already sick and weak people the because food supplies are interrupted. These less obvious costs of war, I believe, accumulate significantly through time and … have played their part in effecting a relatively low rate of population growth in the recent past.24
One may also quickly dispose of the argument that these high casualty rates only reflect contact between tribal peoples and Westerners by citing the very similar proportions of violent deaths documented in several prehistoric populations (Figure 6.2).25 My own first excavation training was on a prehistoric village site on the San Francisco Bay in California. Thousand-year-old skeletons with obsidian arrowheads embedded in the bones, missing heads, and other signs of violent death were so common that our excavation crew referred to burials as “bad sights.” As a matter of fact, one distinctive characteristic of this period in central Californian prehistory is that about 5 percent of all human skeletons contain embedded arrowheads—which, of course, represent only the most obvious evidence of death in warfare. The actual percentage of violent deaths must have been much higher. Indeed, several of these prehistoric cases seriously underestimate the number of violent deaths because only individuals with projectiles in their bones are counted as war deaths. Judging from the Gebel Sahaba cemetery, where only 25 percent of the skeletons that show signs of arrow wounds have the points so embedded, the real proportion of war deaths in the California and Scandinavian cases in Figure 6.2 probably ranged from about 7 percent to as much as 40 percent of all deaths. Contrary to arguments that tribal violence increased after contact with Europeans, the percentage of burials in coastal British Columbia bearing evidence of violent traumas was actually lower after European contact (13 percent from 1774 to 1874) than the very high levels (20 to 32 percent) evidenced in the prehistoric periods.26 It is clear from these archaeological examples that the casualty rates recorded by ethnographers are neither improbable nor exceptional. Tribal peoples needed no instructions or inducements from Europeans to make real war.
But how can such high losses be reconciled with the low casualty rates generally observed in primitive battles, where action is often broken off when both sides have suffered a few dead? Part of the answer lies in the higher sortie rate of primitive warriors. As was noted earlier, warfare occurs much more frequently in most primitive societies than in civilized ones. Thus a relatively low loss rate per war, battle, or raid can cumulate very rapidly to catastrophic levels. Suppose that a tribe with 100 warriors breaks off fighting or arranges a truce in a battle after the loss of just 5 percent dead or mortally wounded. If such battles occurred about four times a year, the cumulative loss in just five years would be 64 percent, leaving only about 36 warriors alive to defend the group. Given a high frequency of warfare, likely losses due to small raids and ambushes, and other sources of losses to warrior strength from accidents or disease, no small group could afford to accept losses in battle exceeding 2 percent. Even that loss rate per fight, if battles take place four times a year, would reduce the group’s fighting strength by a third in just five years. When debilitating wounds that do not result in death are also taken into account, it becomes clear why warriors from small societies were so prone to end battles after just a few casualties. At issue was not just their personal survival but literally that of their group.
Restricting the number and severity of casualties can be done only in the context of formal battles. Small raids and ambushes, which are more frequent than battles and more indiscriminately deadly, are less subject to control. Larger-scale surprise attacks, not uncommon in primitive wars, can cause extremely high levels of casualties. The uncontrollable violence and frequency of such raids and massacres thus make primitive warfare deadlier than modern wars.
The usual primitive practices of not recognizing surrender and of dispatching all male captives also contribute to the lethality of such warfare. Despite the great difficulties inherent in successfully surrendering in the heat of battle, this life-saving option is often resorted to by civilized soldiers; as we have seen, however, it is unavailable to his primitive counterparts. Wounded soldiers who fall into the hands of civilized foes may receive very poor care or may be killed to prevent them from becoming a burden to their captors, but a similar misfortune for a primitive warrior meant certain death.
As for women, even when the ritual conventions of civilized war are not observed by modern states, their female mortalities do not exceed those usually infli
cted over time by tribal warfare. The Allied “strategic” bombing of Germany killed more women than men; but compared with Germany’s total male losses, these deaths still represented a ratio of only one female to every sixteen to twenty males.27 Of course, the Nazi death squads and death camps killed so indiscriminately that the sex ratio of Germany’s victims must be much closer to unity.28 The corresponding ratios for the prestate groups range from about 1:1 to 1:15 (with a median of 1:7)—in every case a higher proportion of female deaths than that caused by the Allies and in a few cases little better than the proportion inflicted by the Nazis.29
The cumulative effects of all these forms of violence can decimate a small clan or tribe.30 One small New Guinea community began a war with twenty-two married men. After just four and a half months of fighting, six men (27 percent) had been killed, eight men had moved away to safety, and the group had been forced to merge with another unit in order to survive. In a war between two Papuan village confederacies (each with populations of 600 to 700 people) that lasted for more than a year, over 250 people were killed, and one side was left with almost no adult males. At the beginning of the nineteenth century, the Blackfoot tribe of the northern Plains was 50 percent deficient in adult males—presumably because of combat deaths since the deficiency disappeared rapidly after intertribal warfare ended later in the century. These percentages equal or exceed the decimations suffered by any modern state in its wars.31
Even complete annihilation of enemy social units has not been unknown in primitive wars. Instances of tribes or subtribes being driven to extinction by persistent tribal warfare have been recorded from several areas of the world.32 Such genocides were sometimes accomplished by a single surprise massacre, on other occasions by longer-term attrition from repeated raids, or by a combination of both. The case of the Woriau Maring of New Guinea illustrates one method by which such annihilations were accomplished, and it also indicates why such occurrences tend to be rare. As was mentioned in Chapter 4, a favorite raid tactic in highland New Guinea consisted of stealthily surrounding the men’s houses of an enemy, settting them afire, and killing all those who emerged. Usually, one Maring clan had insufficient manpower to attack all of an enemy’s men’s houses simultaneously and had to retreat in the face of counterattacks from the unattacked houses after killing a few men. In the Woriau case, two enemy clans allied themselves for the attack and were able to cover every house, annihilating the Woriau’s manpower in a single day. The defenseless survivors then dispersed and ceased to exist as a collective group. Indeed, social extinction in tribal societies seems not to have entailed the killing of every person in the victimized group; rather, after a significant portion of the group (including most of its adult men) was killed, the surviving remnants were incorporated into the societies of the victors or into friendly groups with whom they sought refuge. Thus a social or linguistic entity was destroyed, if not necessarily the whole of the biological population that composed it. These may be social versions of “the death of a thousand cuts,” but they are extinctions just the same.
War Before Civilization Page 13