Skip to main content

War Scholar Critiques New Study of Roots of Violence

Study contradicts claim that humans prior to civilization were “plagued by war.”

A new study undercuts the claim of many modern scholars, notably Steven Pinker, that rates of violence were higher before civilization than at any time since. The painting above by William Sadler shows the Battle of Waterloo, which took place between French and British forces in 1815. 

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


Is human violence a product primarily of nature or nurture? Yesterday Nature published two articles on this extremely important question: “The Phylogenetic Roots of Human Lethal Violence,” by Jose Maria Gomez of the University of Granada and three colleagues; and a commentary, “Lethal Violence Deep in the Human Lineage,” by Mark Pagel. I commented on the articles yesterday in “Was Civilization the Cure for Primordial Human Violence?” Many other journalists and scholars have weighed in as well. I asked Rutgers anthropologist Brian Ferguson, an authority on warfare who has previously been a guest on this blog, for his take on the Nature papers. He sent me the following.—John Horgan

“The Phylogenetic Roots of Human Lethal Violence” is an important and highly ambitious study, representing a heroic amount of work.  It combines a comprehensive survey of mammalian conspecific killing with an empirical examination of rates of killing among human populations.  Although I am sure their findings will be taken as support for those who believe war has gone back through the human lineage all the way to our Last Common Ancestor with chimpanzees, the results starkly contradict that idea.  What is demonstrated is the dramatically increasing likelihood of violent death, both with social evolution of complexity; and, among their categories of  bands and tribes, after historical contact with expanding colonialism.  For these reasons their findings, rather than the spin to come, are a very important contribution to the debate over the antiquity of violence and war.     

The main finding is that their phylogenetic study produced an estimated death rate of 1.8% attributable to conspecific violence for the ancestor of all great apes, and 2% for ancestral humans.  2% also is their calculation of human-on-human killing in prehistoric bands and tribes.  This is much higher than for other clades, with .3% calculated for all mammals.  This, they argue, supports a deep phylogenetic tendency for lethal violence, which is reflected in human killing.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


This quick comment is written without having seen their supplemental materials, which would seem to offer a very valuable compilation of archaeological evidence on violent death.  I was surprised, for instance, that no increase in killings is registered in human remains as one moves from the Old World Paleolithic to the Bronze Age.  Several students of war identify the Mesolithic as when war began, and it is clearly evident that war increased over the course of the Neolithic and into the Bronze Age.  However, that conclusion is bolstered by other evidence, such as weapons and settlement characteristics.  Gomez et al. use just skeletal material. Nor are they dealing with war deaths exclusively.  They include all forms of human-on-human killing, including execution of incorrigibles and individual homicides (which are fairly common in many situations), and infanticide (which is prevalent among some peoples, usually within a domestic context).  This response relies exclusively on the findings as published in Nature, supplemented by the “News and Views” comment by Mark Pagel, one of the peer reviewers for the submitted study.

Simply put, “Phylogenetic Roots” shows that humans are more likely to kill other humans than many other mammals are to kill members of their own species, such as bats, whales, and lagomorphs.  This pan-mammalian cladistic comparison is a great achievement, and may become a touchstone for comparison of non-human animals.  But for showing something new about humans themselves, there is no revelation.  We have always known that people are capable of killing other people.  On the other hand, I would be quite surprised to hear of one bunny killing another.

However, drawing a straight line between the cladistic ancestral and Paleolithic dots runs into a rather severe bump: the fossils of ancient Hominins.  There is continuing debate over whether Ardipithecus ramidus is directly ancestral to modern humans, but there is no doubt that it is descended from the common ancestors of all great apes, and at around 4.4 million years old, is much closer to them than other well described species.  In “Reexamining Human Origins in the Light of Ardipithecus ramidus” (pg. 74) Owen Lovejoy writes: “Comparisons of the Ar. ramidus dentition with those of all other higher primates indicate that the species retained virtually no anatomical correlates of male-to-male conflict.  Consistent with a diminished role of such agonism, the body size of Ar. ramidus males was only slightly larger than that of females.”  In Russell Tuttle’s magisterial Apes and Human Evolution, after careful review of all evidence bearing on violence, he concludes (pg. 593) “there is no paleontological support for the notion that our stem ancestors regularly engaged in intragroup and intergroup killing, infanticide, cannibalism, or female bashing.”  Then there is the obvious point that bonobos have not been recorded as killing any conspecifics, as long emphasized by Frans de Waal.

Setting that aside, the 2% through-line would not in itself indicate an inherited genetic proclivity, rather than a species capacity, for killing.  An innate tendency to kill, and an ability to kill, are very different things, the latter more plastic in response to environmental influences.  “Phylogenetic Roots” finds that across clades, the two socioecological factors of sociality and territoriality are significantly correlated with rates of killing.  Pagel comments on this.  “The increases in lethal violence coincide with species having increasing amounts of group living and territoriality.  Group living places individuals routinely in close contact, and territoriality means that groups might potentially compete over resources.”  But after noting other possible environmental influences, he adds, “species that live in particular kinds of environments over long periods of time tend to adapt to those environments genetically.”  Those points raise issues not addressed in the Nature article.  As it is, the study does not provide direct evidence of  genetically determined predispositions to killing.

In my view, the importance of this study is not in confirming that humans as a species, unlike most other animals, have killing in their behavioral repertoire.  Point granted.  Rather, this weighs in strongly and perhaps surprisingly for long-standing and currently heated debates about the antiquity of war, collective killing of members of one group by another.  The key findings are apparent in their Figures 3c and 3d.

First, their finding of about 2% total killings in Old World Paleolithic and New World Archaic populations is drastically different from the commonly cited figure of 15-25% of deaths being violent. (See Ferguson “Pinker’s List: Exaggerating Prehistoric War Mortality”).  Oddly, Pagel says the study is consistent with portrayals by Steven LeBlanc and Lawrence Keeley of “hunter-gatherer societies as being engaged in ‘constant battle”-- which predicts how this article will be publicized in broader media.  Yet Gomez et al. themselves note the opposite: “This result contrasts with some previous observations,” citing Keeley and Pinker as examples.

On the same issue, neo-Hobbesians commonly dismiss the absence of signs of deadly violence in very early skeletal remains with the old phrase, “absence of evidence is not evidence of absence.”  (Maybe the projectiles killed by penetrating soft issues, and were removed before interment).    Gomez et al. consider such a possibility.  Their conclusion: “no underestimation was detected for the periods in which both skeletal remains and statistical yearbooks are available.”

Second, their rates of killing increase dramatically over archaeological time periods, jumping upward in the Old World Iron Age, and the New World Formative periods.  This is consistent with previous arguments against war being common throughout the archaeological record, although I put the increase much earlier than the Iron Age (See Ferguson, “Archaeology, Cultural Anthropology, and the Origins and Intensifications of War”).

Third, their rates of killing increase dramatically with developing social complexity and political hierarchy.  Prehistoric bands and tribes come in around that 2% marker, but the chiefdom bar looks like 8 or 9%.  This finding is consistent with a wide spectrum of anthropological writing, that social complexity leads to more war.  Their measured fall-off in killings with historic and contemporary states is not a surprise, since war in pre-state societies involves combat participation for virtually all fighting-age men, whereas states use specialized armies representing a much small fraction of the population.  And as Max Weber told us and Gomez et al. reaffirm, state governments claim and enforce “a monopolization of the legitimate use of violence.”

Fourth, their rates of killing indicate that deadly violence in prehistoric bands and tribes is dwarfed by kill-rates in “contemporary” or “present day” bands and tribes.  They suggest that this may be due to higher population density, “or because they have contacted colonial societies where warfare or interpersonal violence is frequent.”  This is precisely the point argued by “Tribal Zone” theory (Ferguson and Whitehead 1992): expanding colonial systems generate high levels of killings such as are not found in earlier archaeological remains.

It must be clarified that what is often misleadingly called the “Rousseauian” position–we do not argue for an idyllic “state of nature”–asserts a late onset of war, not of homicide. Comparatively high rates of individual killing in the absence of war were shown in Richard Lee’s classic The !Kung San in 1979,  theoretically discussed by Eric Wolf in his 1987 “Cycles of Violence,” cross-culturally elucidated that same year by Bruce Knauft’s “Reconsidering Violence in Simple Human Societies,” and meticulously documented in 2013 by Douglas Fry and Patrick Soderberg’s “Lethal Aggression in Mobile Forager Bands and Implications for the Origins of War.”  That humans have killed other humans for a very long time is not in doubt, except in neo-Hobbesian’s distorted portrayals of their scholarly adversaries.  But murder is not the same as the collective social process of war.

In his commentary, Pagel acknowledges “Still, the Rousseau camp might have a corner to fight,” in that Gomez et al. note changes in kill-rates happen “too quickly to be attributable to genetic changes;” and comparison of a “pre-societal ‘state of nature’” to “organized political societies that have a warrior class” demonstrates social influences–“society can also modify our innate tendencies.”

The findings in Gomez et al. do not leave “the Rousseau camp” with only a corner.  They support much of what that broad position has claimed for years.  Biology shows us what we always knew--we kill each other.  That is a characteristic of our species, and is not for many others, so it can be legitimately called a part of our evolutionary heritage.  Their cladistic analysis is very impressive, quantifying and clarifying contrasts, and I look forward to using their compiled archaeological statistics on perimortem trauma.  But the real news in this article is that their numbers contradict the notion that human beings have always been plagued by war.  If I were one of those who claim that war has always been with us, I would demand a recount. --Brian Ferguson

Further Reading:

Chimp Violence Fails to Support Deep-Roots Theory of War.

Was Civilization the Cure for Primordial Human Violence?

Anthropologist Brian Ferguson Challenges Claim that Chimp Violence is Adaptive

Anthropologist Finds Flaw in Claim that Chimp Raids Are “Adaptive

Chimp-Violence Researchers Respond to Criticism on Cross-Check

Has a Bogus Theory of War Kept Obama from Being a Peace President?

10,000-Year-Old Massacre Does Not Bolster Claim That War Is Innate.

Quitting the hominid fight club: The evidence is flimsy for innate chimpanzee–let alone human–warfare.

New Study of Foragers Undermines Claim That War Has Deep Evolutionary Roots.

New Study of Prehistoric Skeletons Undermines Claim that War Has Deep Evolutionary Roots.

Survey of Earliest Human Settlements Undermines Claim That War Has Deep Evolutionary Roots.

Margaret Mead’s War Theory Kicks Butt of Neo-Darwinian and Malthusian Models.

RIP Military Historian John Keegan, Who Saw War As Product of Culture Rather than Biology.

War Is Our Most Urgent Problem; Let's Solve It.