War Before Civilization
War Before Civilization: The Myth of the Peaceful Savage is a 1996 book by anthropologist Lawrence H. Keeley, published by Oxford University Press, that examines archaeological and ethnographic evidence to argue that warfare was widespread, frequent, and proportionally more lethal in prehistoric and non-state societies than commonly portrayed in academic and popular narratives.[1] Keeley contends that the notion of pre-civilized humans as peaceful "noble savages" stems from selective interpretation of data and ideological preferences rather than empirical findings, drawing on skeletal remains showing high rates of violent trauma, fortified settlements, and mass graves from Paleolithic through Neolithic periods.[2] The book quantifies prehistoric warfare's intensity, estimating casualty rates from combat at 10-60% in some tribal groups—far exceeding those in most historical state-level conflicts—and highlights tactics like ambushes, raids, and mutilation that rival or exceed modern atrocities in brutality per capita.[3] Keeley's analysis critiques the underreporting of violence in ethnographic accounts of hunter-gatherers and horticulturalists, attributing it to anthropologists' aversion to portraying "primitive" societies negatively, which he links to a Rousseauian bias favoring views of civilization as the corrupter of innate human harmony.[4] By comparing global datasets, he demonstrates that non-state warfare often lacked restraints like truces or mercy for combatants, leading to higher risks of annihilation for losing groups, and challenges the sequence that war emerged solely with agriculture or states, instead positing it as a near-universal feature of human social organization predating sedentary life.[5] The work's significance lies in its empirical rebuttal to pacifist interpretations of prehistory, influencing subsequent scholarship on human violence, including Steven Pinker's The Better Angels of Our Nature, though it has faced pushback from researchers emphasizing variability in warfare intensity across contexts.[6]Overview
Publication Details and Author Background
War Before Civilization: The Myth of the Peaceful Savage was first published in hardcover by Oxford University Press on February 22, 1996, with ISBN 978-0195091120.[7] A paperback edition followed on December 18, 1997, under ISBN 978-0195119121.[8] The book spans 272 pages and includes bibliographical references and an index.[9] Lawrence H. Keeley (1948–2017) was an American anthropologist and professor emeritus at the University of Illinois at Chicago, where he taught until his retirement in 2014.[6] Specializing in prehistoric archaeology and microwear analysis of stone tools, Keeley gained recognition for his empirical approaches to reconstructing ancient behaviors, including violence and warfare.[6] His work emphasized quantitative data from archaeological and ethnographic records to challenge prevailing assumptions about pre-state societies.[10] Keeley died on October 11, 2017, at age 69.[6]Core Thesis and Structure of the Book
In War Before Civilization: The Myth of the Peaceful Savage, published in 1996, Lawrence H. Keeley advances the central thesis that organized violence and warfare were endemic to prehistoric and non-state societies, often exceeding in frequency, lethality, and ferocity the conflicts of modern civilized nations. He directly refutes the influential anthropological orthodoxy, rooted in Enlightenment-era romanticism and perpetuated by 20th-century scholars like those influenced by Margaret Mead's Samoan fieldwork, which portrayed hunter-gatherers and early horticulturalists as inherently pacific, with war emerging only as a byproduct of state formation around 5,000 years ago. Keeley marshals empirical data to argue that this "pacified past" narrative stems from selective evidence interpretation, where ambiguous archaeological remains are deemed non-violent by default and ethnographic reports of tribal raids are dismissed as atypical or post-contact distortions. Instead, he posits that primitive warfare—characterized by ambushes, massacres, and trophy-taking—served adaptive functions like resource control and social cohesion, resulting in casualty rates (often 15-60% of adult male deaths from violence in studied groups) far surpassing the 1-2% typical in 20th-century industrialized wars.[1][11][12] Keeley's analysis underscores causal realism in violence: war arises from human predispositions toward aggression, amplified by small-scale societies' lack of centralized deterrence, rather than being a civilizational invention. He critiques institutional biases in academia, where interdisciplinary pressures favor narratives minimizing innate human conflict to align with progressive ideals, leading to underreporting of fortifications, skeletal trauma (e.g., arrow wounds in 20-40% of prehistoric remains from sites like Ofnet, Germany, circa 5000 BCE), and ethnographic parallels in groups like the Yanomami, where 30% of adult deaths trace to feuding. This thesis challenges deterministic models positing agriculture or population density as war's sole triggers, instead emphasizing ubiquitous intergroup competition in stateless contexts.[1][5][2] The book's structure proceeds logically from critique to evidence to synthesis. It opens with "The Pacified Past," dissecting the myth's origins in anthropology's aversion to ethnocentric judgments and its reliance on unrepresentative "peaceful" ethnographies. Subsequent chapters, such as those on Neolithic evidence (e.g., fortified villages like Jericho, circa 8000 BCE, with skull cults indicating ritual violence) and Paleolithic hypotheses, systematically review archaeological data, rejecting claims of a "pre-warfare" era by highlighting overlooked indicators like mass burials at Talheim, Germany (circa 5000 BCE), where 34 individuals show execution-style trauma. Ethnographic sections draw on over 40 non-state societies to quantify warfare's ubiquity, followed by comparative analyses revealing primitive conflicts' higher per capita destructiveness. The volume concludes with methodological reflections on evidence biases and philosophical inquiries into war's inevitability, advocating rigorous, data-driven reinterpretation over ideological priors.[13][9][5]Historical and Intellectual Context
Origins of the Noble Savage Myth
The concept of the noble savage, depicting pre-civilized humans as inherently virtuous, peaceful, and uncorrupted by societal institutions, emerged from ancient critiques of urban decadence rather than empirical observation of primitives. Roman historian Publius Cornelius Tacitus, in his ethnographic work Germania composed around 98 AD, portrayed Germanic tribes as embodying simplicity, loyalty, martial valor, and sexual restraint—qualities he contrasted sharply with the perceived moral decay of Roman imperial society. Tacitus' idealized depiction served a rhetorical purpose to shame his contemporaries, establishing an early template for viewing "barbarians" as morally superior foils to civilized excess.[14][15] Renaissance humanism refined this archetype through direct encounters with New World peoples. In his 1580 essay "Of Cannibals," Michel de Montaigne drew on reports of Tupinambá practices in Brazil to argue that these "savages" adhered more closely to natural justice and reason than Europeans, whose sophisticated laws masked hypocrisy and cruelty; he famously contended that cannibalism, while alien, paled against the barbarism of European judicial torture and warfare. Montaigne's relativism inverted ethnocentric hierarchies, suggesting primitives retained a purer form of humanity unmarred by artificial customs.[16] The specific phrase "noble savage" entered English usage in John Dryden's 1672 play The Conquest of Granada, where it evoked a primordial freedom predating oppressive laws: "I am as free as nature first made man, / Ere the base laws of servitude began, / When wild in woods the noble savage ran." This literary flourish romanticized untamed existence amid growing European awareness of colonial subjects. By the Enlightenment, Jean-Jacques Rousseau's 1755 Discourse on the Origin and Basis of Inequality Among Men amplified primitivist themes, hypothesizing a pre-social "natural man" driven by self-preservation and pity rather than vice, though Rousseau neither coined the term nor endorsed contemporary "savages" as paragons—his state of nature was an abstract construct critiquing inequality's corrupting effects.[17] Anthropologist Ter Ellingson, in his 2001 analysis, demonstrates that the enduring attribution of the myth to Rousseau as a wholesale endorsement of primitive nobility constitutes a 19th-century scholarly fabrication, primarily by figures like John Crawfurd, who weaponized it as a strawman to rationalize colonial hierarchies and dismiss egalitarian primitivism amid rising racial pseudoscience. This misconstruction persisted in academic discourse, embedding assumptions of pre-civilized harmony that overlooked ethnographic and archaeological evidence of pervasive tribal conflict.[18][19]Anthropological and Archaeological Biases Prior to Keeley
Prior to the 1996 publication of War Before Civilization, anthropological and archaeological scholarship largely dismissed or minimized the prevalence of organized violence in prehistoric societies, asserting that true warfare emerged only with the Neolithic Revolution and the formation of agricultural or state-level polities around 10,000 years ago. This consensus portrayed Paleolithic hunter-gatherers as engaging in little more than individualistic homicides, ritual executions, or rare feuds, rather than collective, lethal conflicts akin to war. For instance, influential mid-20th-century works like Harry Holbert Turney-High's Primitive War (1949) established a conceptual "military horizon," demarcating non-state societies below it as incapable of "real" warfare due to lacking formal organization, tactics, or objectives, thereby reclassifying intergroup raids and ambushes as sub-war phenomena.[1][20] Archaeological interpretations reinforced this view by systematically underinterpreting physical evidence of violence. Skeletal remains showing perimortem trauma—such as embedded projectile points, parry fractures, or decapitations—were frequently attributed to hunting mishaps, accidents, or isolated murders rather than interpersonal or intergroup aggression, even in cases of mass graves like the Crow Creek site in South Dakota (circa A.D. 1325, but analogous to earlier denials). Fortifications, palisades, and weapon caches from pre-agricultural contexts were explained away as defensive against animals or symbolic structures, not human threats. Keeley noted that surveys of major anthropology textbooks from the era revealed 13 out of 14 explicitly claiming warfare's absence or insignificance before settled societies, reflecting a disciplinary aversion to recognizing violence as endemic to small-scale groups.[1][4] These biases arose from methodological constraints, such as requiring "civilized" attributes like uniforms or declarations of war to qualify evidence as warfare, combined with ideological inclinations toward romanticizing non-state life as inherently pacific and egalitarian. Ethnographic data on contemporary hunter-gatherers, including high homicide rates in groups like the Yanomami (up to 30% of adult male deaths from violence in some villages) or Australian Aboriginals, were selectively downplayed in favor of peaceful exemplars like the !Kung San, whose low observed violence was atypical and context-dependent. This selective emphasis, often unmoored from comprehensive cross-cultural analysis, aligned with post-World War II academic trends prioritizing anti-militaristic narratives, potentially amplified by institutional preferences in anthropology for views that critiqued Western civilization's role in fostering aggression over empirical scrutiny of universal human propensities.[1][21]Key Arguments and Evidence
Archaeological Evidence of Prehistoric Warfare
Archaeological evidence for prehistoric warfare includes skeletal remains exhibiting trauma from weapons, mass graves indicative of group killings, fortified settlements suggesting defensive needs, and artifacts such as projectile points embedded in bones. These findings span from the Upper Paleolithic to the Neolithic, contradicting earlier scholarly tendencies to interpret such violence as rare or ritualistic rather than organized conflict. Rates of lethal trauma in some prehistoric populations reached 15-25% among adult males, exceeding those in many historical wars.[1] One of the earliest datasets comes from Jebel Sahaba (Site 117), a Late Pleistocene cemetery in Sudan dated to approximately 13,400-13,000 years ago, where over half of the 61 individuals show signs of interpersonal violence, including embedded lithic points from arrows or spears and perimortem fractures. Recent reanalysis of the skeletons reveals healed and unhealed injuries across multiple episodes, indicating sustained low-level conflict rather than a singular massacre, with trauma concentrated on extremities and torsos consistent with ambush or skirmish tactics.[22][23] In Neolithic Europe, the Talheim Death Pit in southwestern Germany, dated to around 5000 BCE, contains the remains of 34 individuals—mostly women, children, and elderly males—with clustered cranial blunt-force trauma suggesting systematic executions, possibly by a raiding party targeting non-combatants after killing adult males elsewhere. Isotopic and genetic analyses confirm the victims as local kin groups, ruling out intra-community violence and supporting intergroup aggression. Similar massacres occur at Schöneck-Kilianstädten (Germany, ~5000 BCE), where 26 skeletons display perimortem injuries from clubs and adzes, and Schletz (Austria), a fortified Linearbandkeramik site with evidence of burned structures and defensive ditches breached during assault.[24][25] The Crow Creek site in South Dakota, USA, dated to circa 1325 CE but representing late prehistoric Plains village societies, yielded a mass grave with over 450 commingled skeletons showing scalping, decapitation, and defensive wounds, with 90% of adults exhibiting trauma consistent with a coordinated village raid and slaughter. Bioarchaeological examination indicates attackers spared few, depositing mutilated bodies in a fortification ditch, reflecting escalated inter-village warfare amid resource competition.[26][27] Mesolithic evidence from Ofnet Cave in Bavaria, Germany (~7500 BCE), includes two clusters of 33 skulls—predominantly female and subadult—arranged in "nests" with perimortem cutmarks and fractures indicative of decapitation and trophy-taking post-mortem, patterns aligning with raid aftermaths rather than ritual sacrifice. Across Neolithic Europe, over 100 fortified enclosures with palisades, ditches, and gates appear from 5500-4000 BCE, such as those in the Danube valley, where burn layers and weapon caches suggest repeated sieges and counterattacks driven by territorial disputes.[28][29] These assemblages, when quantified, show violence accounting for 10-20% of mortality in many pre-state societies, with male-biased trauma patterns implying organized raiding for captives or resources, challenging interpretations that downplayed warfare due to incomplete preservation or ideological preferences for peaceful foraging models.[5]Ethnographic Data on Tribal Violence
Ethnographic studies of non-state societies, including hunter-gatherers and tribal groups, document elevated rates of lethal violence, often manifesting as chronic feuding, raids, and warfare that account for substantial portions of mortality. In the Yanomami of the Amazon, approximately 30% of adult male deaths result from violence, with 44% of males aged 25 or older having participated in at least one killing.[30] Nearly 70% of adults over 40 have lost a close genetic relative to such violence, underscoring the intergenerational cycle of revenge killings and raids.[30] Among the Waorani of the Ecuadorian Amazon, violence has caused about 60% of adult deaths across five generations, driven by persistent feuds and retaliatory killings.[31] The Grand Valley Dani of New Guinea exhibit homicide rates equivalent to 1,000 violent deaths per 100,000 people annually, reflecting frequent pig-raiding wars that escalate into lethal confrontations.[31] Similarly, the Kato (Cahto) of 1840s California recorded 1,450 violent deaths per 100,000 per year, indicative of intergroup conflicts over resources and territory.[31] In lowland South American tribal groups, compilations from 11 anthropological studies show violence accounting for roughly 30% of adult deaths, with males comprising about 70% of victims, often from ambushes and revenge expeditions.[32] The Gebusi of New Guinea report 35.2% of male and 29.3% of female deaths due to violence, including sorcery accusations leading to executions and homicides.[31] These patterns persist in unpacified or minimally contacted groups, where ethnographic fieldwork captures endemic raiding and homicide as normative responses to disputes over women, resources, or honor, contrasting sharply with the lower violence levels enforced by state monopolies on force.[31] While some researchers debate exact figures due to underreporting or definitional variances, the aggregate data affirm violence as a pervasive feature of tribal social organization.[31]Comparative Lethality and Frequency of Pre-Civilized vs. Civilized War
Keeley argues that the lethality of warfare, measured as the proportion of total deaths attributable to combat, was substantially higher in non-state societies than in civilized ones. Ethnographic data from tribal groups indicate that warfare caused an average of approximately 15% of all deaths, with rates varying widely but often exceeding 30% for adult males in particularly violent societies such as the Yanomami and Waorani.[31] [20] In contrast, even during the most destructive twentieth-century conflicts, such as World War II, war-related deaths constituted less than 1% of total deaths across combatant populations when normalized over lifetimes or generations, with annual combat mortality rates around 0.15%.[31] [33] Archaeological evidence from prehistoric sites reinforces this disparity, with violent death rates estimated at 10-20% or higher in many cases, compared to under 5% in ancient state societies like those in nineteenth-century France.[34] [31] Annual death rates from warfare in non-state societies were typically two to fifty times higher than those in state-level warfare of the twentieth century, reflecting tactics like ambushes and raids that inflicted near-total casualties on small groups rather than the partial losses in pitched battles of larger armies.[33] For instance, in the Dugum Dani of New Guinea, warfare accounted for about 35% of male deaths, while ethnographic surveys of groups like the Mae Enga show similar patterns.[31] Keeley attributes this elevated lethality to the absence of medical infrastructure, the targeting of non-combatants, and the lack of truces or quarter in tribal conflicts, contrasting with civilized warfare's logistical constraints and occasional restraints.[20] Regarding frequency, Keeley documents that non-state societies experienced warfare far more routinely than assumed, with ethnographic samples revealing 65% in continuous conflict, 87% fighting more than once per year, and only 13% enjoying a decade or more of peace.[20] Approximately 86% of surveyed non-state groups were either raiding or defending against raids at any given time, underscoring endemic violence driven by resource competition and revenge cycles.[33] In state societies, by comparison, full-scale wars occur less pervasively, often separated by years or decades of relative peace, though total casualties may accumulate higher in absolute terms due to scale; Keeley emphasizes that per capita exposure remains lower, as state monopolies on violence suppress internal raiding.[20] This pattern holds across 90-95% of known societies, where peaceful exceptions are typically isolated or transient refugees rather than normative.[35]| Society/Group | Estimated % of Male Deaths from Warfare | Source |
|---|---|---|
| Yanomami | 30% | Keeley (1996) via OWID[31] |
| Waorani | 60% | Keeley (1996) via OWID[31] |
| Jívaro | 60% | Ethnographic data cited by Keeley[20] |
| Grand Valley Dani | ~35% (males) | Keeley (1996)[31] |
| 20th-century state wars (avg.) | <1% of total deaths | Keeley (1996)[33] |
Methodological Foundations
Keeley's Data Selection and Analytical Framework
Keeley selected archaeological data by prioritizing physical indicators of collective violence, such as fortifications (e.g., ditches and palisades surrounding settlements), skeletal remains exhibiting trauma from weapons (e.g., embedded arrowheads, axe wounds, or spear injuries), mass burials suggestive of massacres, and evidence of settlement destruction like burn layers.[36] He applied strict criteria to distinguish warfare from individual homicide or accident, requiring contextual clustering of injuries (e.g., multiple skeletons with perimortem trauma in one site) and association with defensive structures, drawing from sites spanning the Mesolithic to Neolithic periods, such as Talheim, Germany (where 34 individuals showed axe-related wounds indicating a group execution around 5000 BCE), Gebel Sahaba, Sudan (over 40% of burials with projectile injuries circa 13,000 BCE), and Crow Creek, South Dakota (a massacre site with nearly 500 scalped and dismembered bodies from the 14th century CE).[36] This selection countered prior archaeological tendencies to underreport or reclassify violence evidence as non-warlike, emphasizing verifiable osteological and structural data over interpretive assumptions.[37] For ethnographic data, Keeley drew from cross-cultural compilations like the Ethnographic Atlas by George Murdock, focusing on non-state societies (bands, tribes, and chiefdoms) minimally influenced by external states to approximate pre-civilizational conditions.[36] He selected cases with detailed records of conflict participation, casualties, and social organization, excluding pacified or colonial-altered groups, and analyzed over 90 societies where warfare documentation was robust, such as the Yanomami of Venezuela (frequent raids causing 5% casualties in short periods) and the Dugum Dani of Papua New Guinea (endemic feuding with high male mortality).[36] Surveys indicated that 90% of such societies engaged in war, with peaceful exceptions (e.g., Semai of Malaysia) being rare and often geographically isolated.[36] This approach privileged quantitative ethnographic accounts from anthropologists like Chagnon and Heider, while acknowledging variability but rejecting selective sampling that overstated pacifism.[31] Keeley's analytical framework quantified warfare's impact through metrics of lethality, frequency, and mobilization scale, enabling direct comparisons between non-state and state societies. Lethality was assessed via the percentage of deaths attributable to violence (primarily warfare for males) or annualized death rates, revealing non-state societies often exceeding 15-60% lifetime violent mortality for adult males—far higher than 1-2% in 20th-century European states.[38] For instance, ethnographic data showed annual warfare death rates of 0.5% or more in tribal groups (e.g., 1.0% among the Dani, 1.45% among the Karo), compounding to 25% or higher over a 50-year lifespan, versus 0.07% in historical France.[36] Frequency was measured by time spent in conflict, with non-state groups at war 65-87% of years versus intermittent state engagements (e.g., Roman legions active once every 6-20 years).[36] Mobilization rates highlighted near-total societal involvement in primitive warfare, contrasting partial levies in states.[36] To address interpretive biases, Keeley incorporated meta-analysis of prior scholarship, critiquing post-World War II anthropology and archaeology for "pacifying the past" by dismissing violence evidence as atypical or ritualistic, often due to ideological preferences for Rousseauian narratives over empirical patterns.[37] He cross-validated archaeological and ethnographic findings against historical analogs, arguing that under-detection of low-visibility raids (common in non-state warfare) systematically underestimated prehistoric violence, and advocated conservative estimates to avoid overstatement.[36] This framework prioritized causal inference from material traces and longitudinal data, rejecting unsubstantiated assumptions of evolutionary pacifism in favor of observed regularities in human conflict.[31]| Metric | Non-State Societies (Examples) | State Societies (Examples) | Source |
|---|---|---|---|
| % Violent Deaths (Males) | 15-60% lifetime (e.g., Yanomami, prehistoric sites 7-40%) | 1-2% (20th-century Europe) | [38] [36] |
| Annual Death Rate | 0.5%+ (e.g., Dani 1.0%, Mae Enga up to 1.23%) | 0.07-0.16% (e.g., France, Germany) | [36] |
| Warfare Frequency | 65-87% of years | Intermittent (e.g., once per generation) | [36] |