The Evolution of War and Lethal Violence

Over the years, I have written a long series of posts on whether evolutionary science can adjudicate the debate between Thomas Hobbes and Jean-Jacques Rousseau over whether our earliest human ancestors were naturally violent (as Hobbes argued) or naturally peaceful (as Rousseau argued).  Many social scientists have been vehement in taking one side or the other in this debate.  But I have argued that John Locke took a third position that is closest to the truth--that our foraging ancestors lived in a state of peace that tended to become a state of war.  Hobbes is partly right. Rousseau is mostly wrong. And Locke is mostly right.

Now a new study published in Nature has stirred up this debate again.  This research shows how human lethal violence against members of our own species is part of our evolutionary history as mammals. 

Jose Maria Gomez and his colleagues have compiled and analyzed the sources of mortality from a comprehensive sample of more than 4 million deaths from 1,024 mammalian species drawn from 137 mammalian families, which is 80% of the total number of mammalian families.  The data for humans was from over 600 studies, ranging from Paleolithic samples (50,000 to 12,000 years ago) to anthropological studies of the last few centuries. (J. M. Gomez, M. Verdu, A. Gonzalez-Megias, & M. Mendez, "The Phylogenetic Roots of Human Lethal Violence," Nature 538 [13 October, 2016]: 233-37). 

They applied comparative statistical techniques to the phylogeny (the family tree) of mammalian violence to reconstruct the rates of lethal violence, defined as killing by members of the same species.  They calculated that the rate of lethal violence at the phylogenetic origin of mammals was about 0.30%, which is approximately 1 in 300 deaths.  Rates of lethal violence increased across the family tree leading to primates--2.3% for the common ancestor of primates and tree shrews, declining slightly to 1.8% for the ancestor of the great apes (gorillas, orangutans, chimpanzees, and bonobos), and then increasing to 2% at the origin of the human species, which is about six times higher than the rate at the origin of mammals.  Increases in lethal violence were correlated with increasing group living and territoriality.  Apparently, lethal violence increases when individuals are living close together and competing for territorial resources.

Gomez and his colleagues compared the phylogenetically inferred levels of human lethal violence with the levels observed in archaeological and anthropological records.  They found that the phylogenetic prediction of about 2% for prehistoric foragers was about the same as what emerged from their calculations from the archaeological records.  But they also found that levels of lethal violence rose far above 2% in historic times when human beings began to live in chiefdoms and states.  Levels of deaths by lethal violence did not begin to decline until about 500 years ago.  Finally, in the last 100 years, these levels fell to below 2% for the first time in human history.

In his commentary in Nature, Mark Pagel claims that this Gomez study supports the Hobbesian position that human beings are innately violent (Pagel, "Lethal Violence Deep in the Human Lineage," Nature 538 [13 October, 2016]: 180-81).  In an article in The Guardian, Steven Pinker is quoted as claiming that this study does support the Hobbesian position and Pinker's argument in The Better Angels of Our Nature about how modern societies have brought a dramatic reduction in violence from the high levels in prehistoric foraging societies.

On the contrary, John Horgan and Brian Ferguson have claimed that the Gomez study actually supports the Rousseauean position that war is not innate for human beings but largely a product of the cultural environments that arose after human beings left the ancient foraging life.

There are a number of points in dispute here.

The first point is that in measuring how often humans and other mammals kill members of their own species, the Gomez study makes no distinction between individual killing and killing in war.  So they include all forms of human-on-human killing, which includes individual homicides and infanticide.  Rousseaueans like Brian Ferguson and Douglas Fry agree that prehistoric foragers often killed one another, but almost all of this homicidal killing arose in personal disputes or feuding, and so, they argue, this was not warfare.  But notice what this means: the Rousseaueans are conceding that our foraging ancestors were not utterly peaceful, as Rousseau claimed, because they did sometimes kill one another, although this was not killing in war between groups.  One can find high rates of individual killing among ancient foragers, but without war.

The second point is that the Gomez study's estimate of 2% of deaths among ancient foragers being due to lethal violence is much lower than the estimates by Pinker and others of 15-25% of deaths due to warfare.  While Pagel claims that the Gomez study shows "hunter-gatherer societies as being engaged in constant battles" (181), the Gomez study actually concludes that the 2% rate "contrasts with some previous observations," such as that of Pinker (235).

The third point is that the Gomez study shows that rates of lethal killing increase (from 2% to 9%) with the move from ancient foraging bands to chiefdoms.  These high rates do not drop below 2% until the last 100 years.  This looks like the pattern suggested by Douglas Fry--like the letter n--low violence in the ancient foraging past, then increasing violence in more complex societies, followed by declining violence in contemporary states.  Pinker might argue that this conforms to the Hobbesian pacification thesis, because the only form of social order that reduces violence is the modern centralized state.  Ferguson and Fry seem to concede this.  This shows that both the Hobbesians and the Rousseaueans can agree that violence is not genetically determined, because the expression of the natural capacity for violence depends on the social environment.  As Azar Gat has said, "war is innate, but optional."

The fourth point is that the rates of lethal violence in prehistoric bands and tribes is much lower than for bands and tribes studied by anthropologists over the past few centuries.  This suggests that foraging societies become more violent after having contact with colonial societies, and that such high levels of violence cannot be found in the archaeological record for ancient prehistoric foragers.

The final point is that the two living ape species most closely related to humans--chimps and bonobos--show conflicting evidence as to whether the primate phylogeny of humans favors an innate propensity to violence.  Chimpanzees show male coalitional raiding leading to killing that looks like primitive warfare similar to what people like Richard Wrangham see among human foragers.  But there is no clear evidence for even a single case of conspecific killing among bonobos.  Pagel writes: "Even the usually peaceful bonobo Pan paniscus can sometimes display violent behavior" (181).  But he does not cite any evidence for bonobo lethal violence.

In the Supplementary Material for their article (found online at the Nature website), Gomez and his colleagues provide the data for mammalian lethal violence.  For chimpanzees, they say there is documentation for 734 deaths and for 4.49% of these being deaths from conspecific killing.  For bonobos, it's 145 deaths and 0.68% conspecific deaths.  For the bonobo data, they cite five articles.  Apparently, they are relying on one "suspected" case of conspecific killing for bonobos.  But when I contacted Frances White, the leading observer of bonobos in the wild, and asked her about this, she said that this one case is very dubious, and that there is no clear case of conspecific killing among bonobos.  She said that there were at least three cases of bonobo cannibalism, but there was no evidence that these were cases of killing rather than eating already dead carcasses.

It's easy to understand why the Rousseaueans love the bonobos--there're the hippie apes who make love not war.

On the other hand, it's also easy to understand why the Hobbesians love the chimps and not the bonobos, because the chimps are closer to Hobbesian expectations for a evolutionarily close human relative.  Frances White once observed:

"As we have found out more about how bonobos behave in the wild, they have declined in favor as a model for our ancestor . . . This makes me wonder how much our use of models is influenced by how we would like our ancestor to have behaved--clearly the bonobo has fallen from grace because it shows what, for many, is behavior that is socially unacceptable for a close relative of ours."

Some of my other posts on the evolution of war and violence are here, here, here, here, here, here, here, here, here, here, here, here, here, here, here, and here.