Most of the Westerners who did observe and describe traditional warfare have not been professional scholars. For instance, Sabine Kuegler, daughter of missionaries Klaus and Doris Kuegler, described in her popular book Child of the Jungle how, when she was six years old, a fight with bows and arrows erupted between the Tigre clan of the Fayu (among whom her family was living) and visitors from the Sefoidi clan, and how she saw arrows flying around her and wounded men being carried away in canoes. Similarly, the Spanish priest Juan Crespí, a member of the Gaspar de Portolá Expedition, which was the first overland European expedition to reach the Chumash Indians on the coast of southern California, in 1769–1770, wrote in detail about groups of Chumash shooting arrows at each other.
A problem associated with all of these accounts of traditional warfare by outside (usually European) observers, whether anthropologists or laypeople, is reminiscent of the Heisenberg Uncertainty Principle of physics: the observation itself perturbs the phenomenon observed. In anthropology this means that the mere presence of outsiders inevitably has large effects on previously “untouched” peoples. State governments routinely adopt a conscious policy of ending traditional warfare: for example, the first goal of 20th-century Australian patrol officers in the Territory of Papua and New Guinea, on entering a new area, was to stop warfare and cannibalism. Non-government outsiders may achieve that same result in different ways. For instance, Klaus Kuegler eventually had to insist that his host Fayu clan stop fighting around his house and go somewhere else to shoot each other, otherwise he and his family would have to leave for their own safety and peace of mind. The Fayu agreed, and gradually stopped fighting altogether.
Those are examples of Europeans intentionally ending or decreasing tribal fighting, but there are also claims of Europeans intentionally provoking tribal fighting. There are also many ways in which outsiders, through their activities or mere presence, may unintentionally increase or decrease fighting. Thus, whenever an outside visitor reports observations of traditional warfare (or lack of warfare), there is inevitable uncertainty about how much fighting there would have been if no outside observer had been present. I shall return to this question later in this chapter.
An alternative approach has been to scrutinize evidence of tribal fighting preserved in the archaeological record laid down before the arrival of outsiders. This approach carries the advantage of removing the influence of contemporary outside observers entirely. However, in analogy with the Heisenberg Uncertainty Principle, we gain that advantage at the cost of a disadvantage: increased uncertainty about the facts, because fighting was not observed directly nor was it described on the basis of reports of local eyewitnesses, but it instead had to be inferred from archaeological evidence, which is subject to various uncertainties. One undeniable type of archaeological evidence for fighting is piles of skeletons, thrown together without the usual hallmarks of intentional proper burial, with cut marks or breaks on bones recognizably made by weapons or tools. Such marks include bones with imbedded arrow points, bones with cut marks made by a sharp weapon such as an ax, skulls with long straight cut marks indicative of scalping, or skulls with the first two vertebrae attached as normally results from decapitation (e.g., for head-hunting). For instance, at Talheim in southwestern Germany, Joachim Wahl and Hans König studied 34 skeletons of what turned out to be identifiable as 18 adults (nine men, seven women, and two of uncertain sex) and 16 children. They had been heaped haphazardly around 5000 BC in a pit without the usual grave goods associated with respectful burial by relatives. Unhealed cut marks on the right rear surfaces of 18 skulls showed that those people had died of blows administered from behind by at least six different axes, evidently wielded by right-handed assailants. The victims were of all ages from young children to a man of about 60. Evidently, an entire group consisting of half a dozen families had been massacred simultaneously by a much larger group of attackers.
Other types of archaeological evidence for warfare include finds of weapons, armor and shields, and fortifications. While some weapons aren’t unequivocal signs of war, because spears and bows and arrows can be used to hunt animals as well as to kill people, battle axes and piles of large slingshot missiles do provide evidence of war, because they are used only or mainly against people, not against animals. Armor and shields are similarly employed only in war, not in hunting animals. Their use in war has been described ethnographically among many living traditional peoples, including New Guineans, Aboriginal Australians, and Inuit. Hence finds of similar armor and shields in archaeological sites are evidence of fighting in the past. Further archaeological signs of warfare are fortifications, such as walls, moats, defensible gates, and towers for launching missiles against enemy attempts to scale walls. For instance, when Europeans began to settle in New Zealand in the early 1800s, New Zealand’s indigenous Maori population had hill forts, called pa, used initially to fight each other and then eventually also to fight Europeans. About a thousand Maori pa are known, many of them excavated archaeologically and dated to many centuries before European arrival, but similar to the ones that Europeans saw in use. Hence there is no doubt that Maori were fighting each other long before European arrival.
Finally, other archaeological settlement sites are on hilltop, cliff-top, or cliff-face locations that make no sense except for defense against enemy assault. Familiar examples include Anasazi Indian settlements at Mesa Verde and elsewhere in the U.S. Southwest, on cliff ledges and overhangs accessible only by ladders. Their positions high above the valley floor meant that water and other supplies had to be carried hundreds of feet up to them. When Europeans arrived in the Southwest, Indians used such sites as retreats to hide or protect themselves against European attackers. It’s therefore assumed that cliff dwellings dated archaeologically to many centuries before European arrival were similarly used for defense against Indian attackers, especially as recourse to such sites increased with time as population density and evidence of violence were increasing. If all of this archaeological evidence weren’t enough, rock paintings dating back to the Upper Pleistocene show fighting between opposing groups, depict people being speared, and depict groups of people fighting each other with bows, arrows, shields, spears, and clubs. Sophisticated later but still pre-European art works in this tradition are the famous Maya wall paintings at Bonampak, from a society around AD 800, depicting battles and torture of prisoners in realistic gory detail.
Thus, we have three extensive bodies of information—from modern observers, from archaeologists, and from art historians—about traditional warfare, in small-scale societies of all sizes, ranging from small bands to large chiefdoms and early states.
Forms of traditional warfare
Warfare has assumed multiple forms, both in the past and today. Traditional warfare utilized all basic tactics that are now used by modern states and that were technologically possible for tribal societies. (Naturally, the means for aerial warfare were not available to tribes, and naval warfare with specialized warships is not documented until the emergence of state governments after 3000 BC.) One familiar and still-practised tactic is the pitched battle, in which large numbers of opposing combatants face off against each other and fight openly. This is the first tactic that comes to mind for us when we think of modern state warfare—famous examples including the Battles of Stalingrad, Gettysburg, and Waterloo. Except for scale and weapons, such battles would have been familiar to the Dani, whose battles developing spontaneously on June 7, August 2, and August 6, 1961, I described in Chapter 3.
The next familiar tactic is the raid, in which a group of warriors small enough to conceal itself, advancing under cover or at night, makes a surprise attack on enemy territory with the limited goal of killing some enemies or destroying enemy property and then retreating, but without the expectation of destroying the whole opposing army or permanently occupying enemy territory. This is perhaps the most widespread form of traditional warfare, documented in most traditional societies, such as the Nuer raids again
st the Dinka, or the Yanomamo raids against each other. I described Dani raids that occurred on May 10, May 26, May 29, June 8, June 15, July 5, and July 28, 1961. Examples of raids, by infantry and now also by ships and airplanes, abound in state warfare as well.
Related to raids, and also widespread in traditional warfare, are ambushes, another form of surprise attack in which the aggressors, instead of moving by stealth, hide themselves and remain in wait at a site to which unsuspecting enemies are likely to come. I described Dani ambushes that took place on April 27, May 10, June 4, June 10, July 12, and July 28, 1961. Ambushes remain equally popular in modern warfare, abetted by radar and code-breaking methods that facilitate detecting movements of enemy who are less likely to detect the ambushing party.
A traditional tactic without parallel in modern state warfare is the treacherous feast documented among the Yanomamo and in New Guinea: inviting neighbors to a feast, then surprising and killing them after they have laid down their weapons and focused their attention on eating and drinking. We moderns have to wonder why any Yanomamo group would let itself fall into that trap, having heard stories of previous such treachery. The explanation may be that honorable feasts are common, that accepting an invitation usually brings big advantages in terms of alliance-building and food-sharing, and that the hosts go to much effort to make their intentions appear friendly. The only modern example I can think of involving state governments is the massacre of the Boer commander Piet Retief and his whole party of a hundred men by the Zulu king Dingane on February 6, 1838, while the Boers were Dingane’s guests at a feast in his camp. This example may be considered the exception that proves the rule: the Zulus had been just one of hundreds of warring chiefdoms until unification and the foundation of the Zulu state a few decades previously.
Such blunt treachery has for the most part been abandoned under the rules of diplomacy which modern states now find it in their own self-interest to follow. Even Hitler and Japan issued formal declarations of war against the Soviet Union and the U.S., respectively, simultaneous with (but not before) their attacks on those countries. However, states do employ treachery against rebels whom they consider as not binding them to the usual rules of diplomacy between states. For instance, the French general Charles Leclerc had no qualms about inviting the Haitian independence leader Toussaint-Louverture to a parley on June 7, 1802, seizing him there, and shipping him to a French prison, where he died. Within modern states, treacherous killings are still carried out by urban gangs, drug cartels, and terrorist groups, which do not operate by the rules of state diplomacy.
Another form of traditional warfare without close modern parallels is the non-treacherous gathering that degenerates into fighting. Far commoner than a treacherous feast, this involves neighboring peoples meeting for a ceremony without any intention of fighting. But violence may nevertheless erupt because individuals who have unsettled grievances and who rarely encounter each other now find themselves face-to-face, can’t restrain themselves, and begin fighting, and relatives then join in on both sides. For example, an American friend of mine who was present at a rare gathering of several dozen Fayu people told me of the tension prevailing as men periodically burst out in mutual insults and explosions of anger, pounded the ground with their axes, and in one case rushed at each other with axes. The risk of such unplanned fighting breaking out at gatherings intended to be peaceful is high for traditional societies in which neighboring peoples meet rarely, revenge for grievances is left to the individual, and there is no leader or “government” able to monopolize force and restrain hotheads.
Escalation of spontaneous individual fighting into organized warfare of armies is rare in centralized state societies but does sometimes happen. One example is the so-called Soccer War of June–July 1969 between El Salvador and Honduras. At a time when tensions between the two countries were already high over economic disparities and immigrant squatters, their soccer teams met for three games in a qualifying round for the 1970 World Cup. Rival fans began fighting at the first game on June 8 in the Honduran capital (won 1–0 by Honduras), and the fans became even more violent at the second game on June 15 in the El Salvador capital (won 3–0 by El Salvador). When El Salvador won the decisive third game 3–2 in overtime on June 26 in Mexico City, the two countries broke diplomatic relations, and on July 14 the El Salvador army and air force began bombing and invading Honduras.
Mortality rates
How high is the mortality from traditional tribal warfare? How does it compare with the mortality from warfare between state governments?
Military historians routinely compile national casualty totals for each modern war: e.g., for Germany during World War II. That permits one to calculate national war-related mortality rates averaged over a century of a country’s history of alternating war and peace: e.g., for Germany over the whole of the 20th century. Such rates have also been calculated or estimated in dozens of studies of individual modern traditional societies. Four surveys—by Lawrence Keeley, by Samuel Bowles, by Steven Pinker, and by Richard Wrangham and Michael Wilson and Martin Muller—summarized such evaluations for between 23 and 32 traditional societies. Not surprisingly, there proves to be much variation between individual societies. The highest annual time-averaged war-related death tolls are 1% per year (i.e., 1 person killed per year per 100 members of the population) or higher for the Dani, Sudan’s Dinka, and two North American Indian groups, ranging down to 0.02% per year or less for Andaman Islanders and Malaysia’s Semang. Some of those differences are related to subsistence mode, with average rates for subsistence farmers being nearly 4 times those for hunter-gatherers in Wrangham, Wilson, and Muller’s analysis. An alternative measure of war’s impact is the percentage of total deaths that are related to warfare. That measure ranges from 56% for Ecuador’s Waorani Indians down to only 3%–7% for six traditional populations scattered around the globe.
For comparison with those measures of war-related mortality in traditional small-scale societies, Keeley extracted 10 values for societies with state government: one of them for 20th-century Sweden, which experienced no wars and hence zero war-related deaths, the other nine for states and time periods selected for notoriously horrible suffering in war. The highest-percentage long-term death tolls averaged over a century in modern times have been for 20th-century Germany and Russia, which reached 0.16% and 0.15% per year respectively (i.e., 16 or 15 people killed per year per 10,000 members of the population) due to the combined horrors of World Wars I and II. A lower value of 0.07% per year held for France in the century that included the Napoleonic Wars and the winter retreat of Napoleon’s army from Russia. Despite the deaths inflicted by the two atomic bombs on Hiroshima and Nagasaki, the fire-bombing and conventional bombings of most other large Japanese cities, and the deaths by gunfire and starvation and suicide and drowning of hundreds of thousands of Japanese soldiers overseas during World War II, plus the casualties from Japan’s invasion of China in the 1930s and the Russo-Japanese war of 1904–1905, Japan’s percentage war-related death toll averaged over the 20th century was much lower than Germany’s or Russia’s, “only” 0.03% per year. The highest long-term estimate for any state is 0.25% per year for the famously bloody Aztec Empire in the century leading up to its destruction by Spain.
Let’s now compare these war-related death rates (expressed again as percentages of the population dying per year of war-related causes, averaged over a long period of alternating war and peace) for traditional small-scale societies and for modern populous societies with state government. It turns out that the highest values for any modern states (20th-century Germany and Russia) are only one-third of the average values for traditional small-scale societies, and only one-sixth of Dani values. Average values for modern states are about one-tenth of average traditional values.
It may astonish you readers, as it initially astonished me, to learn that trench warfare, machine guns, napalm, atomic bombs, artillery, and submarine torpedoes produce time-averaged war-related death
tolls so much lower than those from spears, arrows, and clubs. The reasons become clear when one reflects on the differences between traditional and modern state warfare that we shall discuss in more detail below. First, state warfare is an intermittent exceptional condition, while tribal warfare is virtually continuous. During the 20th century Germany was at war for only 10 years (1914–1918 and 1939–1945), and its war deaths during the remaining 90 years were negligible, while the Dani were traditionally at war every month of every year. Second, casualties of state war are borne mainly just by male soldiers age 18 to 40 years; even within that age range, most state wars use only small professional armies, with the mass conscription of the two world wars being exceptional; and civilians were not at direct risk in large numbers until saturation aerial bombing was adopted in World War II. In contrast, in traditional societies everyone—men and women, prime-age adults and old adults, children and babies—is a target. Third, in state warfare soldiers who surrender or are captured are normally permitted to survive, whereas in traditional warfare all are routinely killed. Finally, traditional but not state wars are periodically punctuated by massacres in which much or all of the population on one side gets surrounded and exterminated, as in the Dani massacres of June 4, 1966, the late 1930s, 1952, June 1962, and September 1962. In contrast, victorious states nowadays routinely keep conquered populations alive in order to exploit them, rather than exterminating them.
Similarities and differences
In what respects is traditional warfare similar to state warfare, and in what respects is it different? Before answering this question, we should of course recognize that there isn’t a polar opposition between these two types of warfare, with no middle ground, but that warfare instead changes along a continuum from the smallest to the largest society. The larger the society, the larger the armed force that it can muster, hence the lower the possibility of concealing the force, the lower the potential for raids and ambushes by small concealed groups of a few men, and the greater the emphasis on open battles between large forces. The leadership becomes stronger, more centralized, and more hierarchical in larger societies: national armies have officers of various ranks, a war council, and a commander-in-chief, while small bands just have equal-ranked men fighters, and medium-sized groups (like the Gutelu Alliance among the Dani) have weak leaders directing by persuasion rather than by authority to give orders. Warfare in large centralized chiefdoms may approximate warfare in small states. Despite this continuity of societal size, it’s still useful to compare small and large societies in how they fight.