War Before Civilization

by Lawrence H Keeley

Until the end of the nineteenth century, civilized soldiers exhibited in battle an extraordinary preoccupation with protecting their own and seizing their enemy’s regimental colors, imperial eagles, and the like. Terrible dishonor was associated with losing these symbols (which were nevertheless carried in the front ranks, at the point of maximum exposure, during combat) to the enemy. When combat became close, especially fierce struggles developed around these standards as men fought to seize or retain them. Perhaps the clearest evidence of their purely symbolic nature was that two British officers were posthumously awarded Britain’s highest award for valor, the Victoria Cross, for flight in the face of the enemy, because they were attempting to save the colors of their regiment from capture by victorious Zulus.9 The men were staff officers of the British units that were defeated at Isandlwana in 1879, and they had fled several miles from the battlefield before being caught and killed. Behaviors that would therefore normally be regarded as cowardly and irresponsible in an officer—abandoning a command and fleeing from battle—were transmuted into acts of extreme courage because their purpose was to save a useless symbol. Compounding the irony of this incident, the Zulus showed no interest in these British colors and left them on the spot. The curious focus of civilized soldiers on capturing such gewgaws is surely no less stylized and impractical than the desire of Plains Indian warriors to count coups.

  As far as decorative regalia worn by warriors is concerned, only in the past fifty years have the field uniforms, aircraft, and ship colors of civilized societies become truly practical (that is, camouflaged). The last war fought by British soldiers in their famous red coats was the Zulu War of 1879. The French army entered World War I dressed in light blue, and initially some German troops in that war wore the preposterous pickelhalbe helmet. The “flying circuses” of the German air force in World War I are the most extreme modern example of impractical, assertive coloration, but the famous squadron art of American aircraft (and tanks) in World War II continued, in a more subdued fashion, this supposedly primitive fashion. Even as recently as the Persian Gulf War, American A-10 Warthog ground-attack aircraft were decorated with shark’s-mouth motifs on their noses.

  The practice of taunting the enemy before a battle has also not entirely disappeared from modern warfare, but the greater distance between contending front lines that modern weaponry imposes means that loudspeakers, leaflets, and radio broadcasts must be employed. Such devices are devoted to issuing appeals to surrender (usually in vain) and to propagating elaborate taunts or boasts prepared by psychological warfare and propaganda services. Where front lines lie close together and where linguistic knowledge is sufficient, the more concise and ethnographically familar form of taunting will occur. For example, during the War in the Pacific, Japanese soldiers tried to unnerve their adversaries at night by screaming taunts in broken English (in Burma, they used Hindi) that ranged from the banal “Marine, you die!” to such infuriating insults as “Joe DiMaggio, no good!” The retorts of Allied soldiers were mostly in their native tongues and scatological in nature.10 In terms of content and intent, there is little difference between a Tokyo Rose broadcast and a tribal prebattle harangue.

  If modern battles are thus not free of rituals and stylized or impractical behaviors, are they more deadly than their primitive counterparts? Many ethnographers vaguely note that primitive battles tend to be called off after a few casualties, but they seldom actually count the number of warriors engaged or lost. In Figure 4.1, some of the few available casualty figures for specific or “average” tribal battles have been compiled and compared with various high-casualty civilized battle (Marathon, Zama, Gettysburg, the first day of the Somme, and a battle lost by the Aztecs in Michoacán). The lowest proportion of total casualties (killed, wounded, and missing) is registered by those of the Union and Confederate forces at Gettysburg; the highest rates are those attending the Aztec invasion of Michoacán, the Carthaginians at Zama, and the Assiniboin raiders. In several instances involving formal primitive battles begun or terminated by agreement (the Mtetwa, Cahto-Yuki, and Mae Enga confrontations), the proportion of warriors killed in action is below 2 percent, a rate that tends to confirm the impressions of ethnographers and their informants. There are many descriptions of such battles with no fatal casualties, although the fighting would often then be renewed after brief intervals until fatalities did occur. However, the phenomenon of low casualties in arranged battles is not universal in prestate warfare. Sources among the Yokuts of central California insisted that half of the participants in one formal battle involving three “tribe-lets” were killed.11 Furthermore, the casualties in primitive encounter battles were often heavy. For example, as in the case of the Assiniboin raiders, when a war party of Plains Indians was caught and heavily outnumbered by its enemies, the smaller party was usually completely wiped out.12 The safest conclusion to draw from such a small and mixed sample is that no evidence consistently indicates that primitive battles are proportionately less lethal or less injurious than civilized ones.

  Figure 4.1 Casualties in various tribal, ancient, and modern battles (see Appendix, Table 4.1).

  The preceding figures, moreover, are for single sustained encounters of one to four days and do not account for the frequency of battle. For example, in the Cahto-Yuki case, battle was resumed twice with similar losses after ten-day truces. The Cahto fought six separate battles during that summer. By comparison, the armies of the Potomac and of Northern Virginia did not have another minor engagement after Gettysburg for three months, and the next full-scale battle (the Wilderness) occurred ten months later.13 The cumulative effect of frequent but low-casualty battles will be discussed in Chapter 6.

  When one of the contenting parties in prestate warfare was routed, the subsequent rampage by the victors through the losers’ territory often claimed the lives of many women and children as well as men.14 One Maring clan of 600 people in New Guinea lost 2 percent of its population in the rout that followed its loss of 3 percent of its people in the preceding battle. This total may not seem very severe, but to produce equivalent figures France (with a population of 42 million) would have had to lose over 1.2 million soldiers in its 1940 defeat and some 840,000 civilians in the immediate aftermath (or five times the total number of war-related French deaths during the whole war). Victorious Tahitian warriors killed so many people in a loser’s territory that an “intolerable stench” of decaying corpses “pervaded defeated districts for long periods after battle.” Similarly severe slaughters attended battlefield defeats among the chiefdoms of Fiji and Cauca Valley of Colombia. These examples illustrate the most important and universal rule of war: do not lose.

  In several ethnographic cases, formal battles with controlled casualties were restricted to fighting within a tribe or linguistic group. When the adversary was truly “foreign,” warfare was more relentless, ruthless, and uncontrolled.15 Thus the rules of war applied to only certain “related” adversaries, but unrestricted warfare, without rules and aimed at annihilation, was practiced against outsiders.

  RAIDS AND AMBUSHES

  The most common form of combat employed in primitive warfare but little used in formal civilized warfare has been small raids or ambushes. These have usually involved having a handful of men sneak into enemy territory to kill one or a few people on an encounter basis or by means of some more elaborate ambush. Women and children have commonly been killed in such raids.16 The Cahto-Yuki war mentioned earlier was started when some Yuki, angry over Cahto use of a disputed obsidian quarry and some plant-gathering territory, killed a gathering party of four Cahto girls. One common raiding technique (favored by groups as diverse as the Bering Straits Eskimo and the Mae Enga of New Guinea) consisted of quietly surrounding enemy houses just before dawn and killing the occupants by thrusting spears through the flimsy walls, shooting arrows through doorways and smoke holes, or firing as the victims emerged after the structure had been set afire. During hard winters, the Chilcotin of British Columbia would attack small isola
ted hamlets or family camps of other tribes, kill all the inhabitants, and live off their stored food. The East Cree of Quebec slaughtered any Inuit (Eskimo) families they encountered, taking only infants as captives. Neither age nor sex was any guarantee of protection during primitive raids.

  Because the victims were unprepared or unarmed and because raids were so frequent, a predictably high cumulative fatality rate resulted.17 One Yanomamo village was raided twenty-five times in just fifteen months, losing 5 percent of its population. In just one summer (1823), two Yellowknife raids killed eight Dogrib (four men and four women), representing 3 percent of the population of the two victimized Dogrib bands; similar raids had been endured for years. Even when formal battles occurred frequently, more deaths were inflicted by raids. Among the Dugum Dani, in fewer than six months, seven ritual battles killed only two men, but nine raids over the same period killed seven people. Figures cited in Chapter 2 indicate that nearly all western North American Indian groups were raided at least twice each year. A careful and open-eyed reading of ethnographies, early historical accounts, and recorded tribal traditions for some supposedly pacifistic Plateau tribes in British Columbia leaves no doubt that raiding and other forms of combat were both frequent and persistent in this area. The numbers killed as a result of these raids were sometimes extremely significant, as in the case of 400 Lilloet (approximately 10 percent of the tribal population) slain in the course of a week-long raid by a neighboring tribe.18 Many groups, such as the Yanomamo of Venezuela and Koaka of Guadalcanal, never resorted to formal battles at all. Raids and ambushes have been the most frequent and widely employed form of nonstate warfare because they are terribly effective at eliminating enemies with a minimum of risk.

  Raids characteristically kill only a few people at a time; they kill a higher proportion of women than do battles or even the routs that follow them; they kill individuals or small groups caught in isolated circumstances away from major population concentrations; and because the victims are outnumbered, surprised, and often unarmed, their wounds are often inflicted as they try to flee. Archaeologically, this pattern will thus be evidenced by four corresponding characteristics: burials of individual or small groups of homicide victims; women as a high proportion of the victims; burials sometimes located away from the major habitation zones (although raid victims were recovered and buried in the usual cemeteries); and evidence that most wounds, even on adult males, were inflicted from behind.19 Several isolated prehistoric burials in central Washington State fit this pattern precisely, and radiocarbon dates indicate that raiding went on in this region for over 1,500 years. Projectile points found embedded in these skeletons indicate that in some cases the killers were “foreigners.” Interestingly, the usual ethnographic descriptions of the tribes in this area—indeed, in the whole culture area of the Plateau—depict them as exceptionally peaceful. At a cemetery site in central Illinois dating to about A.D. 1300, 16 percent of the 264 individuals buried there met violent deaths and also fit the patterns expected for raid victims. Similar attritional violence is documented in prehistoric cemeteries in central British Columbia and in California, where burials of probable raid victims were accumulated over several hundred years. The homicide victims at the 13,000-year-old Gebel Sahaba cemetery in Egypt do not quite fit this small ambush-raid pattern: more victims were buried at one time; adult males’ wounds were commonly left frontal, indicating that they were wounded while fighting with their bows; and children were common among the victims. In this case, the attacks seem to have been on a larger scale—perhaps against small encampments rather than against isolated work parties. These burials accumulated over at least two generations. In each of the cases cited, the proportion of violent deaths is quite high. For example, the homicide rate of the prehistoric Illinois villagers would have been 1,400 times that of modern Britain or about 70 times that of the United States in 1980!20 There can be little doubt that the frequent, sustained, and deadly raids recorded for ethnographic tribal groups were also practiced in many prehistoric cases.

  MASSACRES

  A gradual scalar transition in primitive warfare leads from the small raid to massacres. The latter are larger surprise attacks whose purpose is to annihilate an enemy social unit. The simplest form involves surrounding or infiltrating an enemy village and, when a signal is given, attempting to kill everyone within reach.21 Such killing has usually been indiscriminate, although women and children evidently escape in the confusion more often than adult males. In one case of massacre in New Guinea, the victim group of 300 lost about 8 percent of its population. In a case from a different area, a tribal confederation of 1,000 people lost nearly 13 percent of its population in just the first hour of an attack by several other confederacies. Surprise attacks on California Pomo villages usually killed between 5 and 15 percent of their inhabitants. When the first Spanish explorers reached the coastal Barbareño Chumash of California, the latter had just had two of their villages surprised, burned, and completely annihilated by raiders from the interior, representing a minimum loss of 10 percent of their tribal population. After enduring years of raids by the Yellowknife tribe of northern Canada, several Dogrib bands combined to wipe out a Yellow-knife camp, killing four men, thirteen women, and seventeen children who accounted for 20 percent of the victims’ population. The Yellowknifes never recovered from this blow, and the descendants of the demoralized survivors were gradually absorbed by neighboring groups. The seldom-achieved goal of another subarctic tribe, the Kutchin, was to surround and annihilate an encampment of their traditional enemies, the Mackenzie Eskimo, leaving only one male alive. This male, called “The Survivor,” was spared only so he could spread word of the deed. The Upper Tanana or Nabesna of Alaska massacred most of one band (numbering perhaps 100 people) of Southern Tutchone. Similar slaughters have been recorded in South America, as in the case of a treacherous attack on guests at a Yanomamo feast in which 15 of 115 people were killed in a single day. The approximate average loss in these various instances was 10 percent. To put such massacre mortalities in perspective, this level of population loss would be equivalent to killing over 13 million Americans in 1941 or over 7 million Japanese in 1945 in a single air raid. The results of intertribal massacres could be devastating, especially to a social unit already decimated by battle and raids.

  Such explosive slaughters seem to have occurred infrequently.22 For the Dugum Dani of highland New Guinea, it is estimated that such massacres happened only once every ten or twenty years. Over a period of half a century, the Sambia of the same region fought six neighboring tribes in wars involving massacres. The Yellowknife tribe of northern Canada had been raiding the neighboring Dogribs for no more than twenty years when, as we have seen, the latter annihilated one of their camps. These few cases hardly suffice to support a generalization; but in a number of other ethnographies, such slaughters were recalled by older informants born a generation before colonial pacification, suggesting that massacres once a generation were not an unusual experience in many nonstate groups.

  Contrary to Brian Ferguson’s claim that such slaughters were a consequence of contact with modern European or other civilizations, archaeology yields evidence of prehistoric massacres more severe than any recounted in ethnography.23 For example, at Crow Creek in South Dakota, archaeologists found a mass grave containing the remains of more than 500 men, women, and children who had been slaughtered, scalped, and mutilated during an attack on their village a century and a half before Columbus’s arrival (ca. A.D. 1325). The attack seems to have occurred just when the village’s fortifications were being rebuilt. All the houses were burned, and most of the inhabitants were murdered. This death toll represented more than 60 percent of the village’s population, estimated from the number of houses to have been about 800. The survivors appear to have been primarily young women, as their skeletons are underrepresented among the bones; if so, they were probably taken away as captives. Certainly, the site was deserted for some time after the attack because the bodi
es evidendy remained exposed to scavenging animals for a few weeks before burial. In other words, this whole village was annihilated in a single attack and never reoccupied.

  A similar massacre occurred in the historic period (ca. 1785) at the fortified Larson site, where the dead had been similarly scalped, mutilated, and finally buried under the collapsed roofs and walls of their burned houses. This example clearly shows that except for introducing some new weapons (in particular, muskets and iron-headed arrows), contact with Western civilization caused no significant change in the tenor of warfare in this area. In other words, anthropologists are not justified in dismissing or discounting the ethnographic descriptions of Middle Missouri warfare since they apply equally well to the precontact period. Evidence of a similar slaughter and burning of a whole village, dating to the late thirteenth century, has been uncovered in southwestern Colorado at Sand Canyon Pueblo, where (as at the Larson site) the bodies of the victims were buried under the collapsed roofs of their burned houses.

  After surveying a large number of prehistoric burial populations in the eastern United States, archaeologist George Milner concluded that the pre-Columbian warfare of this whole region featured “repeated ambushes punctuated by devastating attacks at particularly opportune moments.”24 From North America at least, archaeological evidence reveals precisely the same pattern recorded ethnographically for tribal peoples the world over of frequent deadly raids and occasional horrific massacres. This was an indigenous and “native” pattern long before contact with Europeans complicated the situation. When the sailing ship released them from their own continent, Europeans brought many new ills and evils to the non-Western world, but neither war nor its worst features were among these novelties.

  Similar massacres are also documented for the prestate peoples of prehistoric western Europe (Chapter 2).25 At the time of the Talheim massacre 7,000 years ago, neither civilizations nor states had yet developed anywhere. At Roaix in France, 4,000 years ago, more than 100 people of both sexes and all ages were killed by bow-wielding adversaries and then hastily buried in a mass grave. When this French massacre occurred, the nearest civilization was 1,000 miles away in Minoan Crete. In both cases, the number of victims conforms closely to the average number of inhabitants estimated by archaeologists for the average Early Neolithic hamlet and the average Late Neolithic village—respectively, the most common size of settlement in each period. Before any possible contact with civilizations, the tribesmen of Neolithic Europe, like those of the prehistoric United States, were thus wiping out whole settlements.