![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
Garbage in, garbage out
The unmitigated pedantry blog recently posted about the "long peace" - the much reduced frequency of warfare, first in Europe, then worldwide, that started in 1815 with the end of Napoleon and has lasted (with a few interruptions) until now. The pedantry blog spoke approvingly of Azar Gat's work in providing an explanation of the Long Peace, and disapprovingly of Steven Pinker's bestseller "The Better Angels of Our Nature." I knew of the Pinker book but had never read it, and had never heard of Gat. I was intrigued by the explanation proffered in the blog, and decided to read Gat and Pinker. Which I have now done, and... Gat's argument about the long peace is interesting, but it's embedded in a steaming pile of authoritarian, fash-apologist garbage. Pinker's book makes almost exactly the same overarching fash-apologist argument, but with extra helpings of racism. What follows is my response to both author's discussion of human nature and war in prehistory. Their arguments about the "long peace" will have to wait for another post. This got long.
Gat's magnum opus is "War in Human Civilization," which I did not look at, because he's published a much shorter book, "The Causes of War and the Spread of Peace" that summarizes the longer one's thesis and then looks specifically at the Long Peace. Gat begins by summarizing his grand theory - that the propensity for war is genetic in humans, that pre-agricultural foragers actually warred on and murdered each other more often than their farming successors, that anarchy is inherently violent and that we are all better off with cops and soldiers and coercive governments around to keep us in check with the threat of retaliation if we step out of line. Sure they might be brutal thugs, but the alternative is worse. He is a modern day Hobbesian - quite literally, he refers to Hobbes's Leviathan multiple times and with approval.
I have zero patience for biological determinists and sociobiologists and Gat's argument that humans have a natural innate propensity to violence brings nothing new to the table. All the usual garbage arguments are there - cherry picking evidence, highlighting animal species that are violent towards their own kind and ignoring or glossing over those that aren't. He mentions the research that found mobs of common chimpanzees will sometimes gang up on and murder individual members of rival troops. He admits hurriedly that bonobo chimpanzees do not do this, and then declares that "the common chimpanzees, with their dominant aggressive male coalitions, resemble the known patterns of aboriginal human social life far more than the bonobos, who are dominated by female alliances." Notice the sudden reversal of goal posts. From looking at animals for clues to what human nature might be, he turns to dismissing animals that are not violent because they aren't like what he says humans are like.
The research that he doesn't cite is that bonobo chimps are more similar to humans along every metric - physiological, genetic, and behavioural - than common chimps. He also doesn't mention at all that common chimps are the *only* great ape that murders its own kind - bonobos do not, and gorillas and orangutangs have been seen to do it only a few times each under unusual circumstances. He also he tries to make the assassinations that chimps perform sound like human raiding, when they are not (the males of one chimp troop will gang up on and kill a single isolated male from a neighbouring troop, then repeat until the neighbouring troop has no males left, at which point they befriend the surviving females and incorporate them into their own troop. Terroristically descending on the homes of a neighbouring group to indiscriminately kill, rape, steal, and burn, this is not).
Steven Pinker is, like Gat, a Hobbesian. Like Gat, he cherry picks his evidence, omitting or slanting and distorting anything that doesn't support the conclusion he wants to reach. Like Gat, Pinker approves of Hobbes and thinks that the murderous cruelty of state governments is all that saves us from a "Warre Of Every One Against Every One." And like Gat, Pinker subscribes to sociobiological bullshit (although he calls it "evolutionary psychology," potato, potahto), and he goes for an even more disingenuous way of dismissing bonobos than Gat. He says that they're an outlier amongst the great apes, because they're neotenous (retaining juvenile traits into adulthood) and regular chimps/other apes are not. Which is complete bullshit because human beings are the most neotenous species of all. The fact that bonobo chimps don't kill each other and live in peace doesn't support the conclusion that these guys want, so they bullshit around it.
Moving on from animals, Gat draws a distinction between the farmers of the past 12,000-ish years and the ice age foragers who preceded them (This is artificially dichotomizing a diverse and complex spectrum of lifestyles in between full time farmers and full time nomadic foragers. He is doing this for a reason which will become clear) He says that there are just far too few ice age human fossils to make any conclusions about how violent pre-agricultural foragers were one way or another (false), and thus, the best evidence of our (alleged) innate tendency towards violence lies in anthropology, with records of violence among foraging peoples who survived long enough for Europeans to encounter and write about them. He admits that most foragers who survived to be written about were people living precariously on the unfarmable peripheries of the fertile lands claimed by farmers. We only have reports of how violent they tended to be after living in contact with farmers. So, he turns to Australia, a land of (he claims) only foragers who had remained utterly isolated from modern societies until colonizers arrived in the 1700's. This is one reason for the dichotomy he set up - pre-contact Australia was a diverse continent with everything from highly nomadic, low-population density foragers to farming-adjacent people who, without domesticating any crops, had still managed to create a food system that enabled them to live semi-sedentary lives with high population densities (see my post from last year). But for his argument, he needs the continent to all be one homogenous mass of foragers. He claims "aboriginal Australia is the closest to a pure, uncontaminated laboratory of hunter-gatherer communities on a continental scale that we are ever going to get. I came across the Australian dream laboratory in my search for evidence, and found... numerous volumes of field research carried out by early explorers and anthropologists among the Aboriginal tribes during the nineteenth and early twentieth centuries."
Setting aside the question of whether Australians were actually foragers, and setting aside the racist fetishizing of Aborigines as living fossils, lab specimens who somehow avoided any change over 50,000 years and can be used to reliably infer what life was like for other pre-agricultural peoples world wide... Gat relies on evidence from colonists who wrote about inter-aboriginal conflicts. But. BUT! Many of those early colonizers were racist as fuck, inclined to exaggerate Aboriginal people's "savagery" and violence levels. And every single one, without exception, was writing about people who were suffering an unprecedented cataclysm.
The invaders of Australia brought their diseases with them. Diseases that they hardly noticed amongst themselves except for the much lamented but taken for granted fact that half their children died young. Everyone back in England caught nearly all of these diseases in childhood. At least half died. Girls that survived to grow up passed antibodies on to their children in breast milk, helping bring the overall death rate for a host of deadly pandemic diseases - smallpox, typhus, diphtheria, tuberculosis, etc - down to "only" 50% or so.
The Australians had no pre-existing exposure to any of these diseases, no antibodies to pass to their children. Fifteen months after the First Fleet arrived at Botany Bay, an invader with smallpox passed the disease on to an Australian. The invaders only noticed the pandemic when half the Aborigines living near Botany bay died. They weren't able to see the disease spread beyond that corner of Australia in a continent-wide pandemic, but we know it did. And it was just the first imported virus to do so, the first of many pandemics that swept through Australia.
So: every early written account talking about Aborigines is a record of people who were dealing with an unprecedented calamity of mass death from terrifying new diseases disrupting their communities. And disease was only one of several calamities - there was also the ongoing invasion of murderous foreigners who saw them as subhuman and who brutally murdered anyone who refused to surrender their land to them. There was a massive cultural disruption from the introduction of extremely desirable foreign goods (new drugs, metal tools, cloth, etc), which were only available from those murderous invaders. And yet more disruption from new and foreign concepts (money and work) which had to be navigated to obtain those imported goods.
If we are going to look at the frequency of violence amongst Australians, as reported by European invaders, we have to acknowledge that our records are of a people suffering from lethal pandemics causing mass death, from brutal, genocidal ethnic cleansing, and from attendant breakdown of cultural norms. Naturally the accounts that Gat chooses to quote depict Australians as violent and warlike, engaging in brutal massacres of each other. Given how he dealt with apes, Gat is almost certainly cherry picking here. Some pre-contact Australian cultures clearly had warfare, with weapons designed for use against humans and shields, but I've found critiques of Pinker saying not all Australians engaged in the kind of murderous warfare that he and Gat are so fond of citing). But completely aside from cherry picking: the amount of violence going on even amongst the most warlike Australians must have been greatly increased once their social structure started being torn apart by an existential omnicalamity.
Gat points to two works of paleopathology (looking at causes of death and injury on bones dug up by archaeologists) to buttress his argument that pre-contact Australians had a lot of wars and killed each other a lot. One is a conference paper (completely unobtainable online, as usual for conference proceedings), which he claims found that 57% "of 366 adult crania from all over Australia reveal human-inflicted injuries." Wow, that sounds like a lot of people beating each other over the head. His other reference is a book, "Paleopathology of Aboriginal Australians" by Stephen Webb, available for e-loan from Archive.org. Gat says Webb's numbers are lower and more varied, "but the range is still high and very similar to that we have seen elsewhere in the world. Moreover, unlike in most other places, the Australian figures can be compared with the ethnographic evidence of Aboriginal killing rates. They reinforce the conclusion that these were very high."
But they don't. Web says "An added complication of analysing the placement and origin of cranial trauma in Aboriginal people comes from the quite common practice of striking the head with stones and other objects, as a sign of grief during mourning" (202). What he doesn't say, anywhere, is what fraction of the cranial injuries he's looking at were fatal vs non-fatal, but his discussion of mourning rituals and some photo captions make clear that not all the dented skulls he's tabulating resulted in death. Also, his samples show either the same or more frequent injuries to women's skulls than to men's (depending on the region). Which again doesn't seem to point to manly warfare as the origin for most of the injuries (I have no time to talk about Gat's sexism, but he definitely thinks war is all-male and that women are merely prizes to be won by the winners). But Gat plows on, undeterred. The bones reinforce his conclusions, but two "adjustments" are necessary: "injuries to the crania were mostly suffered in non-lethal... disputes, including a particularly high percentage... among the Aboriginal women; killings in inter-tribal night raids are largely unrecorded in the skeletal evidence because the spearing of those taken by surprise and unable to defend themselves mostly resulted in fatal injuries to soft tissues." So the bones support his thesis, even though they don't actually record the kinds of injuries he's interested in looking at. What exactly was the point of bringing up the bones, then? This is more of that goalpost swapping again. Evidence B is brought in to support conclusions drawn from evidence A, but when it doesn't actually do that, then evidence A is pointed to as support for conclusions about the equivocal evidence B. The only stable point in the shell game is the conclusion, which was decided ahead of time.
The fundamental problem with paleopathology is that the bones that have been found and dug up aren't necessarily representative of what was typical. You'll get wildly different rates of violent death from a church graveyard than from the Douaumont Ossuary at Verdun. Steven Pinker spends a lot more time dwelling on bones than Gat. He has a table that tabulates 21 prehistoric sites ordered by "percentage of deaths in warfare," ranging from 60% down to zero, with an average of 15%. My bullshit detector went off when I saw the 60% figure - how the hell does a community even exist with that level of war deaths? And how the hell can we know that these deaths are indeed "in warfare" as opposed to murder? It's not like the archaeologists are going to find little plaques on prehistoric graves telling us that this one was stabbed by their brother but this one was killed by the neighbouring tribe.
Fortunately Pinker's bestselling book has attracted a number of debunkers. One such, Brian Ferguson, delved deeply into the references for Pinker's table of prehistoric war deaths. "Pinker's List: exaggerating prehistoric war mortality" appears in the anthology "War, Peace, and Human Nature" edited by Douglas Fry, which I was also able to borrow from Archive org. The 60% casualty rate that leads the table is from the 14th century CE Crow Creek massacre in South Dakota - 100% of the remains died from violence, and the 60% figure is the ratio of remains found to a guess of the overall population of the village. The next line on the table, claiming 40% war deaths, represents one of two graveyards at Jebel Sahaba, one on each side of the Nile River (the site is now under Lake Nasser). The other graveyard is also in the table - Pinker was quite liberal with combining some datasets, but he splits this one. One one side of the Nile, out of 59 skeletons, 24 had sharp stone bits associated with them, and the original writeup of the find classed them as killed by violence. In the graveyard on the other side of the river, dated to the same era, the same writeup counted one violent death (out of 39 skeletons). Splitting the two sites lets Pinker give the top of the table (which is ordered by claimed percentage of war deaths) a more dramatic start.
Jebel Sahaba is often cited as the earliest proof of ancient warfare known. Ferguson notes multiple problems with this characterization. There are problems with the dating (the stone technology matches one used from 13000-5000 BCE. The original investigators placed the site at the beginning of that huge range, but on flimsy evidence). The high number of violent deaths there is based on the number of skeletons that had worked stone bits inside or very close to the bodies. Besides ordinary stone tools used to assign the date to the site, there are these bits, which are not arrowheads or spear points, but rather small, unretouched flakes. The original writeup of the site admitted that "in a normal assemblage all of these would be classified as debitage or debris, and none would be considered tools." It's only because they were found in the skeletons that the author went on to say that these bits were weapon tips and the bodies found with them were victims of violence. But some of the bits were found inside skulls that had no breakage - they must have washed in there along with dirt, through the foramen magnum. And then there's the excessive number of weapon tips - 110 stone bits were found associated with skeletons (four per body) plus another 73 in the dirt above the skeletons (Reading the rest of Ferguson's article, it's clear that most of the time, you find at most one, sometimes two weapon tips per body). It all adds up to suggest that the alleged weapon tips are in fact detritus from toolmaking that happened to be mixed in with the dirt of the graveyard, and the number of violent deaths needs to be drastically reduced. Ferguson cites a paleopathologist who looked at the more complete skeletons from the site and decided that there was evidence for only 4 violent deaths out of 41. So the warfare here was more in the heads of the investigating archaeologists than in the past.
The 4th entry in the table is another example of disingenuous massaging of the data: labelled "British Columbia, 30 sites, 3500 BCE- 1674 CE," it covers a vast time period and homogenizes three different epochs. Ferguson breaks it down: in the Early Pacific period, you have "8 out of 12 adult males show signs of some sort of violence at Namu, not necessarily lethal; but at Blue Jackets Creek series on the Queen Charlotte Islands," there's almost no signs of violent injury. "Signs of war multiply as populations grow" in the next era (1800 BCE - 500 CE), but they concentrate around Prince Rupert Harbor, which had a highly militarized culture. Farther south, you have defensive fortifications designed to repulse raids from the north, but no indications of a war based culture. And finally, in the last era, starting around 500 CE, everyone is warring on everyone else.
Elsewhere in North America, you see a similar trajectory - war is absent, then begins sporadically and gradually spreads. Ferguson: "Evidence of war in the Eastern Woodlands dates to several thousand years before it appears in the American Southwest. War in the Southwest is one of the best studied of all areas, but it is temporally and geographically complicated, interrupted by long periods of peace. The northern Great Plains has some of the most extreme evidence of mass killings from anywhere in the prehistoric world [ie, the Crow Creek massacre]. Yet in the southern Plains, prior to 500 AD, of 173 skeletal remains, only one shows signs of violent death." The Eastern Woodlands eventually became extremely warlike from around 500 CE up until the European invasion, but that was preceded by centuries of peace with little to no signs of warfare back to 1000 BCE. And before that, wars appear to have been infrequent and sporadic, with just three sites "noted as having multiple [violent] deaths" for the entire middle-late Archaic period (3000-1000 BCE). Pinker naturally picks one of those three for his table.
The rest of Pinker's table is riddled with similar problems - duplicate entries (Pinker combined two sources to get his table but failed to remove redundancies), entries where the claimed rate is based on just one violent death (Ferguson notes one death is just a murder, you need multiple deaths happening together in time and space to call it a war), and so on. Pinker's alleged evidence for high rates of violent death and war amongst prehistoric people exists only in his cherry picked table.
So what does archaeology actually tell us about prehistoric warfare? Ferguson's answer is in another essay in the same collection, "The Prehistory of War and Peace in Europe and the Near East." In ice age Europe, a tally of 103 ice age skeletons found one death by violence (a child with a stone point embedded in their spine), and one violent assault that did not end in death (a stone point embedded in someone's hip, but the bone had healed). No warfare. Then, from 8000-5000 BCE, we have a LOT more bones but still very little violent death. One (incomplete) tabulation of 1,107 burials from this period counted a maximum of 61 deaths by violence if an extreme outlier is included (5.5%), but only 27 (2.4%) excluding that outlier.
(Content note: disturbing prehistoric multiple murder.
feel free to skip the next paragraph.)
The outlier is Ofnet, Bavaria, where someone had a collection of 34 human skulls in a pit, arranged like a basket of eggs. Fourteen of the skulls were fractured from blows to the back of the head at the time of death. We have no way of knowing how the rest died, over what time frame the skulls were collected, or whether those killed were community members (sacrifices? victims of a serial murderer?) or outsiders (war trophies?).
In all, the era has a few small and one big cluster of violent deaths, but none of the other telltale signs of war like fortifications or artefacts designed primarily as weapons.
Weapons and fortifications only start to show up in the Neolithic (defined by the arrival in a particular place of a cluster of technologies like pottery, farming and herding). Ferguson sums up: "The earliest centuries of farming exhibit, in some places, some signs of individual violence, but no evidence that persuasively establishes the existence of war [except maybe England]... In all other regions, after 500–1000 years or more without it, clear evidence of war appears in skeletons and settlements, in some places but not others.... [more and more signs of war accumulate over time] But there is no simple line of increase... [2, maybe 3 massacres in 5000 BCE Germany were] far worse than what followed, and other areas had ups and downs in active war." But by the end of the Neolithic, with the arrival of metal tools, all of Europe was having quite a lot of warfare, and a culture of war, with a military aristocracy and art that mythologizes war, celebrates war gods, etc, was well established.
Here's a specific example of how localized and spotty warfare was at the beginning of the Neolithic: a style of pottery called LBK originated in Hungary around 5700 and spread west. By 5000, the style could be found all across Northern Europe south of the Baltic from parts of Ukraine to parts of France. LBK settlements at the western end (Germany) had defensive fortifications, but those in the east (Hungary) did not. Skeletal trauma (fatal and non-fatal combined) for all LBK finds is 20%. At the German end, it's 32%. That includes those two definite plus one maybe massacres in Germany around 5000. Exclude those three sites, and skeletal trauma among all LBK finds drops from 20% to just over 6%.
Moving on to the middle east, Ferguson traces war and its absence in three areas: the upper reaches of the Tigris river (the northeast corner of the fertile crescent, now straddling the Iran-Turkey border), Anatolia (the northwest corner, now south-central Turkey), and the Levantine corridor (the western limb of the crescent, from the Dead Sea along the Jordan to Galilee). These areas invented agriculture, so all the technological and cultural transitions happened earlier here than in Europe.
On the upper Tigris, there are a few signs of violence at Qermez Dere and Nemrik 9, sometime between 9750 and 8750. These 2 settlements, slightly separated in time and space, were both situated in natural defensive positions (on a steep sided hilltop, or between ravines). At one, archaeologists found an inordinate number of arrowheads, many with broken points, as if they'd been fired at a wall. At the other, they found three bodies with violent trauma, two of them killed by arrows, but the arrowheads were not made in the local style, so the killers had come from elsewhere. These finds could be the first, signs of war in the area, but you don't get anything more clearly indicative until sometime in the 6000's, with Tell Maghzaliyah (in northern Iraq), where, several centuries after the town was built, fortifications were added. This is the earliest fortified settlement in Mesopotamia and possibly the entire middle east (the much discussed wall of Jerico is earlier, but was built to defend against floodwaters, not attacks). From there, signs of war spread. In Anatolia, the city of Catal Huyuk (7000-ish) had no defences and its extensive art murals do not depict war, but by 6200, war had clearly arrived in the region, with fortified towns, stockpiles of sling bullets, and settlements being burnt with unburied bodies in the houses, then abandoned or rebuilt by a different group.
However, in the Levant, along the Jordan river, there's no signs of violence. All the way down to 3300-ish, people there lived without fortifications, without weapons other than ceremonial maces that were not sturdy enough to use in combat, and without violence other than the occasional murder. Only when warlike outsiders (the kingdom of Egypt) came knocking on their doors did cities in the region build fortifications. Only when they had to start paying Egypt tribute did they start to war on each other to amass the required goods. Ferguson concludes with a call for archaeologists to stop dismissing the absence of evidence of violence, to stop looking for war and refusing to conclude anything when they don't find it, and to instead start paying attention to and looking for signs of peace.
In short: war has a starting point. It's not an innate, natural part of human behaviour, it's something that was invented. It didn't show up with the transition from foragers to farmers (at least in some parts of Europe, Ferguson says there's little evidence for conflict between the two). Ferguson lays out a few "preconditions" that make war more likely to appear: relatively dense populations, a degree of social hierarchy, and social divisions on the one hand (enough people that you can afford to lose some in battle, someone with the power to order attacks, and a "them" to attack), and on the other, one or more economic factors that add up to "something worth taking" - things like one side having a scarce, concentrated high value resource (an obsidian deposit, say), or else enjoying a monopoly in the trade of high value items (due to geographic bottlenecks). The last ingredient would be some kind of survival stress - droughts or shifts in climate leading to crop failures or long term loss of fertility, creating precarious situations for at least one side. He also identifies one factor leading to long term peace and avoidance of war: shared inter-communal shrines or temples that regularly brought multiple far-flung communities together characterize both the Levant and pre-warfare Anatolia. He suggests they may have helped prevent an us/them split from developing.
Returning to Gat: after presenting his cherry picked evidence from Australia, he turns to North America and the Plains Indians, another group that he claims had high levels of violence. "The ethnographic record from the sparse populations of simple hunter-gatherers on the Great Plains offers the closest analogy to early populations in North America, and elsewhere." And there's a footnote in which he dismisses the claims of the scholars he is using that North Americans had higher levels of violence amongst denser communities of farmers than amongst sparse communities of foragers. "However, this is a theoretical presupposition rather than an empirical finding .... so what we have is a (false) theory guiding the empirical investigation." When other people have theories he disagrees with, that's doing your research backwards. But when he does it, it's science. Pot, kettle.
Gat's book is quite short and remains laser-focused on war. Pinker's is much more expansive, and he has a long section talking about "pre-state" and medieval rates of murder as well. Murder rates, he says, mirror death rates from war: more and more the further back in time you go. Medieval people murdered each other far more often than modern people do, and non-state societies have the highest murder rates of all. This is all bullshit, of course. One of several debunkings of the section on medieval murder rates I found was a blog entry here: https://bedejournal.blogspot.com/2011/11/steven-pinkers-medieval-murder-rates.html
Let's start with his claimed murder rate for medieval Oxford and London, two of the labelled points (most are not labelled) in a scatter plot in chapter 3 showing murder rates from medieval Europe through to the modern day. The blog linked above finds three main methodological problems with Pinker's approach. The underlying data for the medieval dots on the plot are archives of death certificates and coroners' inquests. Right off the bat, the data that survives is sparse and patchy. Norwich, England, for instance, has medieval coroner's records for 1263-68, and no other years. Now it would be very simple to simply count the percentage of violent vs natural deaths, which would provide data similar to what Pinker used counting violent deaths in graveyards in the previous chapter. But instead he combines that data with population estimates to get murders year per 100,000 people. Except censuses did not exist then; all we can do is guesstimate total population, and Pinker's sources use guesses which are probably too low, artificially raising the murder rate. Next, before around 150 years ago, injuries were vastly more likely to become infected and end in death than after doctors began using sterile procedures and antibiotics. One estimate is that 3/4 of injuries that resulted in death (and a finding of murder) then would be survivable (mere assaults) today. Finally there's an apples and oranges problem: modern death certificates provide contextual information that tell us whether a death is accidental (not caused by another person), manslaughter (caused by another person but not deliberately), or murder. In medieval inquest documents, that information is sometimes there and sometimes not. For Norwich again, out of 36 cases, 14 are accidental or not a death. In 5, there's context that tells us the person responsible did not mean to kill. Which leaves 17 deaths which were either murder or manslaughter, but we cannot tell which. Because the numbers are so small, just a handful of manslaughters would halve the calculated murder rate. In all, Pinker's chart of murder going down from the middle ages to the modern era has multiple fundamental problems, and it doesn't support the conclusions he's making about premodern Europe being murder central.
As for murder rates among what Pinker calls non-state societies: he has a bar chart in chapter 2 contrasting murder in modern countries and cities with 4 "non state" groups. Other than the Inuit, he seems to have picked groups with a reputation for being peaceful and nonviolent. All are 20th century aboriginal peoples, with data coming from modern anthropologists. So all the groups in the table are dealing with the effects of colonialist policies disrupting their way of life. Like Gat's histories of post-contact Australians, rates of violence are going to be higher than they would be without colonialist meddling. Plus, of course, given how he engineered his table of war deaths, we know Pinker is cherry picking his data points. Now, since I went and did the work (don't ask how long it took), here are each of the 4 "non state" data points for Pinker's bar graph of murder rates, with the sources he used, and the sources those sources used, going back to the original research. Archive org and my library card enabled me to dig up all but one of the original papers/books underlying the claims in the table.
1 Inuit: 100/100k. Pinker cites Gat's War in Human Civilization. Gat cites Donald Symons's 1979 book "The evolution of human sexuality" (Big surprise, he's a sociobiologist peddling bullshit about innate sex differences). Apologies for offensive language: in the cited passage, Symons says, in a section headed "Violence among preliterate peoples" that "Eskimos killed each other at the rate of about one person per thousand per year" and cites a conference paper, N Graburn's 1968 presentation "Inuriat: the killings." (at the Symposium on Primitive Law at the Annual Meetings of the American Anthropological Association, Seattle), and a "personal communication" from Graburn as well. The conference paper is, as usual, not online anywhere, even as a citation.
The amount of racist condescension here is egregious - these are most likely 20th century Canadian or Alaskan Inuit. Thanks to universal education at genocidal residential schools, they're literate. And no, they are not "non state" people and haven't been for quite some time. And their murder rate, regardless of how high it may actually be, is almost entirely attributable to social and psychological trauma inflicted by those residential schools.
2-3. !Kung "before state control" (40/100k) and "after state control." (30/100k). Pinker cites Gat's war in human civilization. Gat writes "The rate for the !Kung Bushmen of the Kalahari, the famous ‘harmless people’, was 0.29 per 1,000 per year, and had been 0.42 before the coming of firm state authority." And he cites Richard Lee's 1979 "The !Kung San: men, women, and work in a foraging society." (protip: do not try searching archive org with the exclamation point in !Kung, you'll get an error).
On checking Lee's book, I discovered Gat engaging in a blatant lie. Lee was there in the late 60's, gathering data until 1969. He was told of 22 killings that happened between 1920 and 1955, all well before he arrived. Zero had happened in the 15 years since 1955. Lee counted this as 22 deaths over 50 years among about 1500 people, for a rate of 29/100k. Gat cries, ah-ha, but all those took place over a 35 year period, and recalculates the rate as 42. Then he uses Lee's figure for the period since "firm state authority," which means he's counting the same number of killings twice. The correct figure since "firm state authority" is zero.
If we look at Lee's actual data, there were 2 killings in the 20's and 4 in the 50's. All the rest happened in the 30's and 40's, so the bulk of the deaths are even more compressed in time. A handful of the killings involve spears, the rest used poisoned arrows (these are all standard hunting kit: the !Kung don't have weapons meant to be used on other humans). Lee says that the death rate is much higher than you would expect because of the poison. Without medical facilities, the poison is fatal 50% of the time even if the wound is survivable. Most of the deaths were part of what Lee calls "feuds," where A kills B, then B's friends/family go after A and try to kill him. They shoot at A, who, with his friends, shoot back. One or more of A's friends may die, and one or more of B's friends may die before A finally goes down; sometimes a bystander also gets killed by a stray arrow. Sometimes it takes multiple attempts by B's friends to finally bring down A. Going through Lee's table of deaths, I counted 4 incidents which accounted for 14 of the deaths: two were simple tit for tat (2 deaths each), and two (one each in the 30's and 40's) were complex affairs with 4 and 6 deaths respectively, including one bystander death in each. Subtracting the bystander deaths as accidental and the killing of killers as (by !Kung norms) justice, Lee's data includes 16 murders spread across 12 incidents - 2 in the 20s, 4 in the 30s, 3 each in the 40s and 50s, for a rather lower murder rate than Gat claims. Without reading far more of Lee's book than I have time for, I can't say what stresses - of colonial rule, drought, or otherwise - contributed to the number of murders, nor what factors caused murder to utterly cease after the early 50's.
As a postscript, while searching Gat's book for !Kung, I came across another use of Lee's data (part of a section treating women as objects which men kill each other over). Citing an article by Lee, he talks about the number of quarrels, then says "in his study area in the period 1963–9, there were 22 cases of homicide; 19 of the victims were males, as were all of the 25 killers." This is obviously the same data vis-a-vis the number of killings, now misattributed to Lee's brief study period in the 60's in order to make things sound extra violent (I checked the cited article. Lee could have made a better distinction in the relevant paragraph between the fights he witnessed in the 60s' and the murders he was told about happening long ago, but given that Gat has read the relevant chapter in Lee's book, where it's all made very clear, I'll call this another lie rather than a mistake).
4. Semai, 30/100k. The citation here is a 1987 article in Current Anthropology v28, no4 (accessed via my library card), "Reconsidering Violence in Simple Human Societies: Homicide among the Gebusi of New Guinea" by Bruce M. Knauft. Knauft talks about the Semai and a few other groups briefly before getting to his own research among the Gebusi. In the relevant passage about the Semai, he cites an article and a book by Robert Dentan, an anthropologist who lived among the Semai, quoting Dentan's article saying "at least two murders have been committed between 1955 and 1977, and there is gossip about a couple of others." Knauft adds that this produces a high "homicide rate, since Dentan's study population totaled only about 300," and cites the book.
I checked both the book and the article by Dentan. The article, "Notes on childhood in a nonviolent context: The Semai case (Malaysia)," appeared in a 1978 anthology edited by Ashley Montagu, "Learning non-aggression." The relevant part of the article tries to explain what Dentan means when he characterizes the Semai as culturally nonviolent. He explains how they define unacceptable antisocial violence much more broadly than Westerners are wont to do. He then clarifies that his previous book about them has given some readers the idea that the Semai are "preternaturally gentle" - this is not so, he says. "Six months in a Semai settlement will see at least three or four serious quarrels in which voices are raised and threats of physical violence are at least alleged, if not actually made. Robarchek (1977b) records quarrels so bitter that only a full meeting of the whole settlement could calm people down. At least two murders have been committed between 1955 and 1977, and there is gossip about a couple of others."
Turning to the book, "The Semai: a nonviolent people of Malaya" (1968). The cited page (where Dentan talks about the fieldwork he and his spouse did) says they lived in two settlements, one in the east, another in the west. Statements about East Semai are based on "data collected in the first settlement and in one across the river which we often visited. The combined population of these two settlements was about one hundred at any given time." Statements about the West Semai are based on "the second settlement, which included about two hundred people, and on data from two other settlements to which we made extended visits." So 300 is obviously the wrong number because the Dentans were also gathering data from people in those two other Western settlements.
However. I screwed up and at first missed seeing the population figures on that page, instead fixating on an earlier mention of government census data. Thinking Knauft made a mistake in his citation, I searched the whole PDF for "census" and found this passage: "Since a census of the Semai was first taken in 1956, not one instance of murder, attempted murder, or maiming has come to the attention of either government or hospital authorities" (p58). So: at the time he was doing his fieldwork amongst 5 small villages with a population somewhere over 300, word of those two murders had not yet reached him. The passage in the later article must be talking not about 2 killings amongst a few tiny villages, but amongst all of the Semai (my search for "census" also turned up the number of Semai in the entire country: 12,700. Which works out to an *extremely* low murder rate). Knauft screwed up, big time, and his error in this one paper back in the 80's becomes a citable "fact" which Pinker can mine for his graph.
And again, it's a huge distortion of the facts to label the Semai as "nonstate" people. Under British Colonial rule, during a communist uprising in the 50s, the communists hid out in the forested hills which were the Semai's home. The government "first tried to relocate the aborigines in camps outside the rain forest. The death rate in these camps was so high that some Semai still regard their relocation as the first step in a campaign to exterminate" them (p3). The government later relented and relocated them in the forest, but near forts which could keep an eye on them to make sure they didn't help the communists. Faced with demands from the government downstream of them and the rebels upstream, Semai bands collaborated on a 6 point policy of deceit: "(1) Bands upstream and thus near the Communists were to aver that they supported the Communists. (2) Bands down river in contact with the government forces were to "support” the government." People in between were to play dumb and not give out information. "In the event of a Communist victory, the 'pro-Communist' bands were to cover up for all the bands downstream from them, claiming that all had been pro-Communist," and vice-versa (80-81). They were definitely living under a state, with all the disadvantages that entailed.
Having grossly distorted indigenous and premodern murder rates, Pinker goes on to explain the drop in rates over time as due to a "civilizing process." And yes, that means he thinks that places and people whose rates have not fallen as far as Europe or white America's have not enjoyed as much civilizing. This is the racist garbage I spoke of at the start.
Two last points about all this. First: Pinker beats on how murder can only be studied in terms of its statistical rate. But: people are not mathematicians. Societies mostly frown upon murder and have norms and rules to keep unacceptable violence in check (remembering that some murders are allowed - for self defence in America, for "honor" or revenge elsewhere). But no one in the community is going to regard murder as a problem as long as it *feels* rare enough. Today we depend on governments and news media to tell us how rare murder is (and the media, at least, make it seem far more common than it is). But for people without access to telecommunications, it was word of mouth, which can only travel so far. In a society with a murder rate of 100/100k, if each community is only 50 or so people, then any given community will have one murder every 20 years - too rare to seem like a problem that needs addressing. Once either the community size grows or the interconnectedness of communities increases to the point that people feel murder is becoming too common, then they take action to reduce violence. Even if murder was indeed as common as Pinker claims in some of these communities, because of their small size, a high rate wouldn't have felt high. And even if murder rates have indeed gone down over time, as he claims, the credit for that change does not belong to the state (modern studies show that more cops and more punishment of has little to no impact on crime rates), or to a mythical "civilizing process," it belongs to the communities that saw an excess of violence and made changes in their cultural norms to curb that violence.
Second: Gat says that he doesn't think it's useful to separate out murder and feuds from warfare. This is nonsense. War is indiscriminate. It's also directed against outsiders, and finally killing enemies is sanctioned and rewarded. Murder and feuds are directed against members of the community. They are personal, targeted, and (generally) forbidden. Dentan writes about Semai who were recruited into the military during the communist insurgency. On their own, they flee from anyone who tries to war against them. "When the British raised troops among the Semai, mainly in the west.... Many [recruits] did not realize that soldiers kill people." Asked to talk about his experiences, one veteran said "We killed, killed, killed. The Malays would stop and go through people’s pockets and take their watches and money. We did not think of watches or money. We thought only of killing. Wah, truly we were drunk with blood." But that had no effect on them once they returned to civilian life: "Back in Semai society they seem as gentle and afraid of violence as anyone else. To them their one burst of violence appears to be as remote as something that happened to someone else, in another country." If Semai can learn to set aside the killing they have done and go back to being peaceful, maybe we all can eventually learn to do the same.
Gat's magnum opus is "War in Human Civilization," which I did not look at, because he's published a much shorter book, "The Causes of War and the Spread of Peace" that summarizes the longer one's thesis and then looks specifically at the Long Peace. Gat begins by summarizing his grand theory - that the propensity for war is genetic in humans, that pre-agricultural foragers actually warred on and murdered each other more often than their farming successors, that anarchy is inherently violent and that we are all better off with cops and soldiers and coercive governments around to keep us in check with the threat of retaliation if we step out of line. Sure they might be brutal thugs, but the alternative is worse. He is a modern day Hobbesian - quite literally, he refers to Hobbes's Leviathan multiple times and with approval.
I have zero patience for biological determinists and sociobiologists and Gat's argument that humans have a natural innate propensity to violence brings nothing new to the table. All the usual garbage arguments are there - cherry picking evidence, highlighting animal species that are violent towards their own kind and ignoring or glossing over those that aren't. He mentions the research that found mobs of common chimpanzees will sometimes gang up on and murder individual members of rival troops. He admits hurriedly that bonobo chimpanzees do not do this, and then declares that "the common chimpanzees, with their dominant aggressive male coalitions, resemble the known patterns of aboriginal human social life far more than the bonobos, who are dominated by female alliances." Notice the sudden reversal of goal posts. From looking at animals for clues to what human nature might be, he turns to dismissing animals that are not violent because they aren't like what he says humans are like.
The research that he doesn't cite is that bonobo chimps are more similar to humans along every metric - physiological, genetic, and behavioural - than common chimps. He also doesn't mention at all that common chimps are the *only* great ape that murders its own kind - bonobos do not, and gorillas and orangutangs have been seen to do it only a few times each under unusual circumstances. He also he tries to make the assassinations that chimps perform sound like human raiding, when they are not (the males of one chimp troop will gang up on and kill a single isolated male from a neighbouring troop, then repeat until the neighbouring troop has no males left, at which point they befriend the surviving females and incorporate them into their own troop. Terroristically descending on the homes of a neighbouring group to indiscriminately kill, rape, steal, and burn, this is not).
Steven Pinker is, like Gat, a Hobbesian. Like Gat, he cherry picks his evidence, omitting or slanting and distorting anything that doesn't support the conclusion he wants to reach. Like Gat, Pinker approves of Hobbes and thinks that the murderous cruelty of state governments is all that saves us from a "Warre Of Every One Against Every One." And like Gat, Pinker subscribes to sociobiological bullshit (although he calls it "evolutionary psychology," potato, potahto), and he goes for an even more disingenuous way of dismissing bonobos than Gat. He says that they're an outlier amongst the great apes, because they're neotenous (retaining juvenile traits into adulthood) and regular chimps/other apes are not. Which is complete bullshit because human beings are the most neotenous species of all. The fact that bonobo chimps don't kill each other and live in peace doesn't support the conclusion that these guys want, so they bullshit around it.
Moving on from animals, Gat draws a distinction between the farmers of the past 12,000-ish years and the ice age foragers who preceded them (This is artificially dichotomizing a diverse and complex spectrum of lifestyles in between full time farmers and full time nomadic foragers. He is doing this for a reason which will become clear) He says that there are just far too few ice age human fossils to make any conclusions about how violent pre-agricultural foragers were one way or another (false), and thus, the best evidence of our (alleged) innate tendency towards violence lies in anthropology, with records of violence among foraging peoples who survived long enough for Europeans to encounter and write about them. He admits that most foragers who survived to be written about were people living precariously on the unfarmable peripheries of the fertile lands claimed by farmers. We only have reports of how violent they tended to be after living in contact with farmers. So, he turns to Australia, a land of (he claims) only foragers who had remained utterly isolated from modern societies until colonizers arrived in the 1700's. This is one reason for the dichotomy he set up - pre-contact Australia was a diverse continent with everything from highly nomadic, low-population density foragers to farming-adjacent people who, without domesticating any crops, had still managed to create a food system that enabled them to live semi-sedentary lives with high population densities (see my post from last year). But for his argument, he needs the continent to all be one homogenous mass of foragers. He claims "aboriginal Australia is the closest to a pure, uncontaminated laboratory of hunter-gatherer communities on a continental scale that we are ever going to get. I came across the Australian dream laboratory in my search for evidence, and found... numerous volumes of field research carried out by early explorers and anthropologists among the Aboriginal tribes during the nineteenth and early twentieth centuries."
Setting aside the question of whether Australians were actually foragers, and setting aside the racist fetishizing of Aborigines as living fossils, lab specimens who somehow avoided any change over 50,000 years and can be used to reliably infer what life was like for other pre-agricultural peoples world wide... Gat relies on evidence from colonists who wrote about inter-aboriginal conflicts. But. BUT! Many of those early colonizers were racist as fuck, inclined to exaggerate Aboriginal people's "savagery" and violence levels. And every single one, without exception, was writing about people who were suffering an unprecedented cataclysm.
The invaders of Australia brought their diseases with them. Diseases that they hardly noticed amongst themselves except for the much lamented but taken for granted fact that half their children died young. Everyone back in England caught nearly all of these diseases in childhood. At least half died. Girls that survived to grow up passed antibodies on to their children in breast milk, helping bring the overall death rate for a host of deadly pandemic diseases - smallpox, typhus, diphtheria, tuberculosis, etc - down to "only" 50% or so.
The Australians had no pre-existing exposure to any of these diseases, no antibodies to pass to their children. Fifteen months after the First Fleet arrived at Botany Bay, an invader with smallpox passed the disease on to an Australian. The invaders only noticed the pandemic when half the Aborigines living near Botany bay died. They weren't able to see the disease spread beyond that corner of Australia in a continent-wide pandemic, but we know it did. And it was just the first imported virus to do so, the first of many pandemics that swept through Australia.
So: every early written account talking about Aborigines is a record of people who were dealing with an unprecedented calamity of mass death from terrifying new diseases disrupting their communities. And disease was only one of several calamities - there was also the ongoing invasion of murderous foreigners who saw them as subhuman and who brutally murdered anyone who refused to surrender their land to them. There was a massive cultural disruption from the introduction of extremely desirable foreign goods (new drugs, metal tools, cloth, etc), which were only available from those murderous invaders. And yet more disruption from new and foreign concepts (money and work) which had to be navigated to obtain those imported goods.
If we are going to look at the frequency of violence amongst Australians, as reported by European invaders, we have to acknowledge that our records are of a people suffering from lethal pandemics causing mass death, from brutal, genocidal ethnic cleansing, and from attendant breakdown of cultural norms. Naturally the accounts that Gat chooses to quote depict Australians as violent and warlike, engaging in brutal massacres of each other. Given how he dealt with apes, Gat is almost certainly cherry picking here. Some pre-contact Australian cultures clearly had warfare, with weapons designed for use against humans and shields, but I've found critiques of Pinker saying not all Australians engaged in the kind of murderous warfare that he and Gat are so fond of citing). But completely aside from cherry picking: the amount of violence going on even amongst the most warlike Australians must have been greatly increased once their social structure started being torn apart by an existential omnicalamity.
Gat points to two works of paleopathology (looking at causes of death and injury on bones dug up by archaeologists) to buttress his argument that pre-contact Australians had a lot of wars and killed each other a lot. One is a conference paper (completely unobtainable online, as usual for conference proceedings), which he claims found that 57% "of 366 adult crania from all over Australia reveal human-inflicted injuries." Wow, that sounds like a lot of people beating each other over the head. His other reference is a book, "Paleopathology of Aboriginal Australians" by Stephen Webb, available for e-loan from Archive.org. Gat says Webb's numbers are lower and more varied, "but the range is still high and very similar to that we have seen elsewhere in the world. Moreover, unlike in most other places, the Australian figures can be compared with the ethnographic evidence of Aboriginal killing rates. They reinforce the conclusion that these were very high."
But they don't. Web says "An added complication of analysing the placement and origin of cranial trauma in Aboriginal people comes from the quite common practice of striking the head with stones and other objects, as a sign of grief during mourning" (202). What he doesn't say, anywhere, is what fraction of the cranial injuries he's looking at were fatal vs non-fatal, but his discussion of mourning rituals and some photo captions make clear that not all the dented skulls he's tabulating resulted in death. Also, his samples show either the same or more frequent injuries to women's skulls than to men's (depending on the region). Which again doesn't seem to point to manly warfare as the origin for most of the injuries (I have no time to talk about Gat's sexism, but he definitely thinks war is all-male and that women are merely prizes to be won by the winners). But Gat plows on, undeterred. The bones reinforce his conclusions, but two "adjustments" are necessary: "injuries to the crania were mostly suffered in non-lethal... disputes, including a particularly high percentage... among the Aboriginal women; killings in inter-tribal night raids are largely unrecorded in the skeletal evidence because the spearing of those taken by surprise and unable to defend themselves mostly resulted in fatal injuries to soft tissues." So the bones support his thesis, even though they don't actually record the kinds of injuries he's interested in looking at. What exactly was the point of bringing up the bones, then? This is more of that goalpost swapping again. Evidence B is brought in to support conclusions drawn from evidence A, but when it doesn't actually do that, then evidence A is pointed to as support for conclusions about the equivocal evidence B. The only stable point in the shell game is the conclusion, which was decided ahead of time.
The fundamental problem with paleopathology is that the bones that have been found and dug up aren't necessarily representative of what was typical. You'll get wildly different rates of violent death from a church graveyard than from the Douaumont Ossuary at Verdun. Steven Pinker spends a lot more time dwelling on bones than Gat. He has a table that tabulates 21 prehistoric sites ordered by "percentage of deaths in warfare," ranging from 60% down to zero, with an average of 15%. My bullshit detector went off when I saw the 60% figure - how the hell does a community even exist with that level of war deaths? And how the hell can we know that these deaths are indeed "in warfare" as opposed to murder? It's not like the archaeologists are going to find little plaques on prehistoric graves telling us that this one was stabbed by their brother but this one was killed by the neighbouring tribe.
Fortunately Pinker's bestselling book has attracted a number of debunkers. One such, Brian Ferguson, delved deeply into the references for Pinker's table of prehistoric war deaths. "Pinker's List: exaggerating prehistoric war mortality" appears in the anthology "War, Peace, and Human Nature" edited by Douglas Fry, which I was also able to borrow from Archive org. The 60% casualty rate that leads the table is from the 14th century CE Crow Creek massacre in South Dakota - 100% of the remains died from violence, and the 60% figure is the ratio of remains found to a guess of the overall population of the village. The next line on the table, claiming 40% war deaths, represents one of two graveyards at Jebel Sahaba, one on each side of the Nile River (the site is now under Lake Nasser). The other graveyard is also in the table - Pinker was quite liberal with combining some datasets, but he splits this one. One one side of the Nile, out of 59 skeletons, 24 had sharp stone bits associated with them, and the original writeup of the find classed them as killed by violence. In the graveyard on the other side of the river, dated to the same era, the same writeup counted one violent death (out of 39 skeletons). Splitting the two sites lets Pinker give the top of the table (which is ordered by claimed percentage of war deaths) a more dramatic start.
Jebel Sahaba is often cited as the earliest proof of ancient warfare known. Ferguson notes multiple problems with this characterization. There are problems with the dating (the stone technology matches one used from 13000-5000 BCE. The original investigators placed the site at the beginning of that huge range, but on flimsy evidence). The high number of violent deaths there is based on the number of skeletons that had worked stone bits inside or very close to the bodies. Besides ordinary stone tools used to assign the date to the site, there are these bits, which are not arrowheads or spear points, but rather small, unretouched flakes. The original writeup of the site admitted that "in a normal assemblage all of these would be classified as debitage or debris, and none would be considered tools." It's only because they were found in the skeletons that the author went on to say that these bits were weapon tips and the bodies found with them were victims of violence. But some of the bits were found inside skulls that had no breakage - they must have washed in there along with dirt, through the foramen magnum. And then there's the excessive number of weapon tips - 110 stone bits were found associated with skeletons (four per body) plus another 73 in the dirt above the skeletons (Reading the rest of Ferguson's article, it's clear that most of the time, you find at most one, sometimes two weapon tips per body). It all adds up to suggest that the alleged weapon tips are in fact detritus from toolmaking that happened to be mixed in with the dirt of the graveyard, and the number of violent deaths needs to be drastically reduced. Ferguson cites a paleopathologist who looked at the more complete skeletons from the site and decided that there was evidence for only 4 violent deaths out of 41. So the warfare here was more in the heads of the investigating archaeologists than in the past.
The 4th entry in the table is another example of disingenuous massaging of the data: labelled "British Columbia, 30 sites, 3500 BCE- 1674 CE," it covers a vast time period and homogenizes three different epochs. Ferguson breaks it down: in the Early Pacific period, you have "8 out of 12 adult males show signs of some sort of violence at Namu, not necessarily lethal; but at Blue Jackets Creek series on the Queen Charlotte Islands," there's almost no signs of violent injury. "Signs of war multiply as populations grow" in the next era (1800 BCE - 500 CE), but they concentrate around Prince Rupert Harbor, which had a highly militarized culture. Farther south, you have defensive fortifications designed to repulse raids from the north, but no indications of a war based culture. And finally, in the last era, starting around 500 CE, everyone is warring on everyone else.
Elsewhere in North America, you see a similar trajectory - war is absent, then begins sporadically and gradually spreads. Ferguson: "Evidence of war in the Eastern Woodlands dates to several thousand years before it appears in the American Southwest. War in the Southwest is one of the best studied of all areas, but it is temporally and geographically complicated, interrupted by long periods of peace. The northern Great Plains has some of the most extreme evidence of mass killings from anywhere in the prehistoric world [ie, the Crow Creek massacre]. Yet in the southern Plains, prior to 500 AD, of 173 skeletal remains, only one shows signs of violent death." The Eastern Woodlands eventually became extremely warlike from around 500 CE up until the European invasion, but that was preceded by centuries of peace with little to no signs of warfare back to 1000 BCE. And before that, wars appear to have been infrequent and sporadic, with just three sites "noted as having multiple [violent] deaths" for the entire middle-late Archaic period (3000-1000 BCE). Pinker naturally picks one of those three for his table.
The rest of Pinker's table is riddled with similar problems - duplicate entries (Pinker combined two sources to get his table but failed to remove redundancies), entries where the claimed rate is based on just one violent death (Ferguson notes one death is just a murder, you need multiple deaths happening together in time and space to call it a war), and so on. Pinker's alleged evidence for high rates of violent death and war amongst prehistoric people exists only in his cherry picked table.
So what does archaeology actually tell us about prehistoric warfare? Ferguson's answer is in another essay in the same collection, "The Prehistory of War and Peace in Europe and the Near East." In ice age Europe, a tally of 103 ice age skeletons found one death by violence (a child with a stone point embedded in their spine), and one violent assault that did not end in death (a stone point embedded in someone's hip, but the bone had healed). No warfare. Then, from 8000-5000 BCE, we have a LOT more bones but still very little violent death. One (incomplete) tabulation of 1,107 burials from this period counted a maximum of 61 deaths by violence if an extreme outlier is included (5.5%), but only 27 (2.4%) excluding that outlier.
(Content note: disturbing prehistoric multiple murder.
feel free to skip the next paragraph.)
The outlier is Ofnet, Bavaria, where someone had a collection of 34 human skulls in a pit, arranged like a basket of eggs. Fourteen of the skulls were fractured from blows to the back of the head at the time of death. We have no way of knowing how the rest died, over what time frame the skulls were collected, or whether those killed were community members (sacrifices? victims of a serial murderer?) or outsiders (war trophies?).
In all, the era has a few small and one big cluster of violent deaths, but none of the other telltale signs of war like fortifications or artefacts designed primarily as weapons.
Weapons and fortifications only start to show up in the Neolithic (defined by the arrival in a particular place of a cluster of technologies like pottery, farming and herding). Ferguson sums up: "The earliest centuries of farming exhibit, in some places, some signs of individual violence, but no evidence that persuasively establishes the existence of war [except maybe England]... In all other regions, after 500–1000 years or more without it, clear evidence of war appears in skeletons and settlements, in some places but not others.... [more and more signs of war accumulate over time] But there is no simple line of increase... [2, maybe 3 massacres in 5000 BCE Germany were] far worse than what followed, and other areas had ups and downs in active war." But by the end of the Neolithic, with the arrival of metal tools, all of Europe was having quite a lot of warfare, and a culture of war, with a military aristocracy and art that mythologizes war, celebrates war gods, etc, was well established.
Here's a specific example of how localized and spotty warfare was at the beginning of the Neolithic: a style of pottery called LBK originated in Hungary around 5700 and spread west. By 5000, the style could be found all across Northern Europe south of the Baltic from parts of Ukraine to parts of France. LBK settlements at the western end (Germany) had defensive fortifications, but those in the east (Hungary) did not. Skeletal trauma (fatal and non-fatal combined) for all LBK finds is 20%. At the German end, it's 32%. That includes those two definite plus one maybe massacres in Germany around 5000. Exclude those three sites, and skeletal trauma among all LBK finds drops from 20% to just over 6%.
Moving on to the middle east, Ferguson traces war and its absence in three areas: the upper reaches of the Tigris river (the northeast corner of the fertile crescent, now straddling the Iran-Turkey border), Anatolia (the northwest corner, now south-central Turkey), and the Levantine corridor (the western limb of the crescent, from the Dead Sea along the Jordan to Galilee). These areas invented agriculture, so all the technological and cultural transitions happened earlier here than in Europe.
On the upper Tigris, there are a few signs of violence at Qermez Dere and Nemrik 9, sometime between 9750 and 8750. These 2 settlements, slightly separated in time and space, were both situated in natural defensive positions (on a steep sided hilltop, or between ravines). At one, archaeologists found an inordinate number of arrowheads, many with broken points, as if they'd been fired at a wall. At the other, they found three bodies with violent trauma, two of them killed by arrows, but the arrowheads were not made in the local style, so the killers had come from elsewhere. These finds could be the first, signs of war in the area, but you don't get anything more clearly indicative until sometime in the 6000's, with Tell Maghzaliyah (in northern Iraq), where, several centuries after the town was built, fortifications were added. This is the earliest fortified settlement in Mesopotamia and possibly the entire middle east (the much discussed wall of Jerico is earlier, but was built to defend against floodwaters, not attacks). From there, signs of war spread. In Anatolia, the city of Catal Huyuk (7000-ish) had no defences and its extensive art murals do not depict war, but by 6200, war had clearly arrived in the region, with fortified towns, stockpiles of sling bullets, and settlements being burnt with unburied bodies in the houses, then abandoned or rebuilt by a different group.
However, in the Levant, along the Jordan river, there's no signs of violence. All the way down to 3300-ish, people there lived without fortifications, without weapons other than ceremonial maces that were not sturdy enough to use in combat, and without violence other than the occasional murder. Only when warlike outsiders (the kingdom of Egypt) came knocking on their doors did cities in the region build fortifications. Only when they had to start paying Egypt tribute did they start to war on each other to amass the required goods. Ferguson concludes with a call for archaeologists to stop dismissing the absence of evidence of violence, to stop looking for war and refusing to conclude anything when they don't find it, and to instead start paying attention to and looking for signs of peace.
In short: war has a starting point. It's not an innate, natural part of human behaviour, it's something that was invented. It didn't show up with the transition from foragers to farmers (at least in some parts of Europe, Ferguson says there's little evidence for conflict between the two). Ferguson lays out a few "preconditions" that make war more likely to appear: relatively dense populations, a degree of social hierarchy, and social divisions on the one hand (enough people that you can afford to lose some in battle, someone with the power to order attacks, and a "them" to attack), and on the other, one or more economic factors that add up to "something worth taking" - things like one side having a scarce, concentrated high value resource (an obsidian deposit, say), or else enjoying a monopoly in the trade of high value items (due to geographic bottlenecks). The last ingredient would be some kind of survival stress - droughts or shifts in climate leading to crop failures or long term loss of fertility, creating precarious situations for at least one side. He also identifies one factor leading to long term peace and avoidance of war: shared inter-communal shrines or temples that regularly brought multiple far-flung communities together characterize both the Levant and pre-warfare Anatolia. He suggests they may have helped prevent an us/them split from developing.
Returning to Gat: after presenting his cherry picked evidence from Australia, he turns to North America and the Plains Indians, another group that he claims had high levels of violence. "The ethnographic record from the sparse populations of simple hunter-gatherers on the Great Plains offers the closest analogy to early populations in North America, and elsewhere." And there's a footnote in which he dismisses the claims of the scholars he is using that North Americans had higher levels of violence amongst denser communities of farmers than amongst sparse communities of foragers. "However, this is a theoretical presupposition rather than an empirical finding .... so what we have is a (false) theory guiding the empirical investigation." When other people have theories he disagrees with, that's doing your research backwards. But when he does it, it's science. Pot, kettle.
Gat's book is quite short and remains laser-focused on war. Pinker's is much more expansive, and he has a long section talking about "pre-state" and medieval rates of murder as well. Murder rates, he says, mirror death rates from war: more and more the further back in time you go. Medieval people murdered each other far more often than modern people do, and non-state societies have the highest murder rates of all. This is all bullshit, of course. One of several debunkings of the section on medieval murder rates I found was a blog entry here: https://bedejournal.blogspot.com/2011/11/steven-pinkers-medieval-murder-rates.html
Let's start with his claimed murder rate for medieval Oxford and London, two of the labelled points (most are not labelled) in a scatter plot in chapter 3 showing murder rates from medieval Europe through to the modern day. The blog linked above finds three main methodological problems with Pinker's approach. The underlying data for the medieval dots on the plot are archives of death certificates and coroners' inquests. Right off the bat, the data that survives is sparse and patchy. Norwich, England, for instance, has medieval coroner's records for 1263-68, and no other years. Now it would be very simple to simply count the percentage of violent vs natural deaths, which would provide data similar to what Pinker used counting violent deaths in graveyards in the previous chapter. But instead he combines that data with population estimates to get murders year per 100,000 people. Except censuses did not exist then; all we can do is guesstimate total population, and Pinker's sources use guesses which are probably too low, artificially raising the murder rate. Next, before around 150 years ago, injuries were vastly more likely to become infected and end in death than after doctors began using sterile procedures and antibiotics. One estimate is that 3/4 of injuries that resulted in death (and a finding of murder) then would be survivable (mere assaults) today. Finally there's an apples and oranges problem: modern death certificates provide contextual information that tell us whether a death is accidental (not caused by another person), manslaughter (caused by another person but not deliberately), or murder. In medieval inquest documents, that information is sometimes there and sometimes not. For Norwich again, out of 36 cases, 14 are accidental or not a death. In 5, there's context that tells us the person responsible did not mean to kill. Which leaves 17 deaths which were either murder or manslaughter, but we cannot tell which. Because the numbers are so small, just a handful of manslaughters would halve the calculated murder rate. In all, Pinker's chart of murder going down from the middle ages to the modern era has multiple fundamental problems, and it doesn't support the conclusions he's making about premodern Europe being murder central.
As for murder rates among what Pinker calls non-state societies: he has a bar chart in chapter 2 contrasting murder in modern countries and cities with 4 "non state" groups. Other than the Inuit, he seems to have picked groups with a reputation for being peaceful and nonviolent. All are 20th century aboriginal peoples, with data coming from modern anthropologists. So all the groups in the table are dealing with the effects of colonialist policies disrupting their way of life. Like Gat's histories of post-contact Australians, rates of violence are going to be higher than they would be without colonialist meddling. Plus, of course, given how he engineered his table of war deaths, we know Pinker is cherry picking his data points. Now, since I went and did the work (don't ask how long it took), here are each of the 4 "non state" data points for Pinker's bar graph of murder rates, with the sources he used, and the sources those sources used, going back to the original research. Archive org and my library card enabled me to dig up all but one of the original papers/books underlying the claims in the table.
1 Inuit: 100/100k. Pinker cites Gat's War in Human Civilization. Gat cites Donald Symons's 1979 book "The evolution of human sexuality" (Big surprise, he's a sociobiologist peddling bullshit about innate sex differences). Apologies for offensive language: in the cited passage, Symons says, in a section headed "Violence among preliterate peoples" that "Eskimos killed each other at the rate of about one person per thousand per year" and cites a conference paper, N Graburn's 1968 presentation "Inuriat: the killings." (at the Symposium on Primitive Law at the Annual Meetings of the American Anthropological Association, Seattle), and a "personal communication" from Graburn as well. The conference paper is, as usual, not online anywhere, even as a citation.
The amount of racist condescension here is egregious - these are most likely 20th century Canadian or Alaskan Inuit. Thanks to universal education at genocidal residential schools, they're literate. And no, they are not "non state" people and haven't been for quite some time. And their murder rate, regardless of how high it may actually be, is almost entirely attributable to social and psychological trauma inflicted by those residential schools.
2-3. !Kung "before state control" (40/100k) and "after state control." (30/100k). Pinker cites Gat's war in human civilization. Gat writes "The rate for the !Kung Bushmen of the Kalahari, the famous ‘harmless people’, was 0.29 per 1,000 per year, and had been 0.42 before the coming of firm state authority." And he cites Richard Lee's 1979 "The !Kung San: men, women, and work in a foraging society." (protip: do not try searching archive org with the exclamation point in !Kung, you'll get an error).
On checking Lee's book, I discovered Gat engaging in a blatant lie. Lee was there in the late 60's, gathering data until 1969. He was told of 22 killings that happened between 1920 and 1955, all well before he arrived. Zero had happened in the 15 years since 1955. Lee counted this as 22 deaths over 50 years among about 1500 people, for a rate of 29/100k. Gat cries, ah-ha, but all those took place over a 35 year period, and recalculates the rate as 42. Then he uses Lee's figure for the period since "firm state authority," which means he's counting the same number of killings twice. The correct figure since "firm state authority" is zero.
If we look at Lee's actual data, there were 2 killings in the 20's and 4 in the 50's. All the rest happened in the 30's and 40's, so the bulk of the deaths are even more compressed in time. A handful of the killings involve spears, the rest used poisoned arrows (these are all standard hunting kit: the !Kung don't have weapons meant to be used on other humans). Lee says that the death rate is much higher than you would expect because of the poison. Without medical facilities, the poison is fatal 50% of the time even if the wound is survivable. Most of the deaths were part of what Lee calls "feuds," where A kills B, then B's friends/family go after A and try to kill him. They shoot at A, who, with his friends, shoot back. One or more of A's friends may die, and one or more of B's friends may die before A finally goes down; sometimes a bystander also gets killed by a stray arrow. Sometimes it takes multiple attempts by B's friends to finally bring down A. Going through Lee's table of deaths, I counted 4 incidents which accounted for 14 of the deaths: two were simple tit for tat (2 deaths each), and two (one each in the 30's and 40's) were complex affairs with 4 and 6 deaths respectively, including one bystander death in each. Subtracting the bystander deaths as accidental and the killing of killers as (by !Kung norms) justice, Lee's data includes 16 murders spread across 12 incidents - 2 in the 20s, 4 in the 30s, 3 each in the 40s and 50s, for a rather lower murder rate than Gat claims. Without reading far more of Lee's book than I have time for, I can't say what stresses - of colonial rule, drought, or otherwise - contributed to the number of murders, nor what factors caused murder to utterly cease after the early 50's.
As a postscript, while searching Gat's book for !Kung, I came across another use of Lee's data (part of a section treating women as objects which men kill each other over). Citing an article by Lee, he talks about the number of quarrels, then says "in his study area in the period 1963–9, there were 22 cases of homicide; 19 of the victims were males, as were all of the 25 killers." This is obviously the same data vis-a-vis the number of killings, now misattributed to Lee's brief study period in the 60's in order to make things sound extra violent (I checked the cited article. Lee could have made a better distinction in the relevant paragraph between the fights he witnessed in the 60s' and the murders he was told about happening long ago, but given that Gat has read the relevant chapter in Lee's book, where it's all made very clear, I'll call this another lie rather than a mistake).
4. Semai, 30/100k. The citation here is a 1987 article in Current Anthropology v28, no4 (accessed via my library card), "Reconsidering Violence in Simple Human Societies: Homicide among the Gebusi of New Guinea" by Bruce M. Knauft. Knauft talks about the Semai and a few other groups briefly before getting to his own research among the Gebusi. In the relevant passage about the Semai, he cites an article and a book by Robert Dentan, an anthropologist who lived among the Semai, quoting Dentan's article saying "at least two murders have been committed between 1955 and 1977, and there is gossip about a couple of others." Knauft adds that this produces a high "homicide rate, since Dentan's study population totaled only about 300," and cites the book.
I checked both the book and the article by Dentan. The article, "Notes on childhood in a nonviolent context: The Semai case (Malaysia)," appeared in a 1978 anthology edited by Ashley Montagu, "Learning non-aggression." The relevant part of the article tries to explain what Dentan means when he characterizes the Semai as culturally nonviolent. He explains how they define unacceptable antisocial violence much more broadly than Westerners are wont to do. He then clarifies that his previous book about them has given some readers the idea that the Semai are "preternaturally gentle" - this is not so, he says. "Six months in a Semai settlement will see at least three or four serious quarrels in which voices are raised and threats of physical violence are at least alleged, if not actually made. Robarchek (1977b) records quarrels so bitter that only a full meeting of the whole settlement could calm people down. At least two murders have been committed between 1955 and 1977, and there is gossip about a couple of others."
Turning to the book, "The Semai: a nonviolent people of Malaya" (1968). The cited page (where Dentan talks about the fieldwork he and his spouse did) says they lived in two settlements, one in the east, another in the west. Statements about East Semai are based on "data collected in the first settlement and in one across the river which we often visited. The combined population of these two settlements was about one hundred at any given time." Statements about the West Semai are based on "the second settlement, which included about two hundred people, and on data from two other settlements to which we made extended visits." So 300 is obviously the wrong number because the Dentans were also gathering data from people in those two other Western settlements.
However. I screwed up and at first missed seeing the population figures on that page, instead fixating on an earlier mention of government census data. Thinking Knauft made a mistake in his citation, I searched the whole PDF for "census" and found this passage: "Since a census of the Semai was first taken in 1956, not one instance of murder, attempted murder, or maiming has come to the attention of either government or hospital authorities" (p58). So: at the time he was doing his fieldwork amongst 5 small villages with a population somewhere over 300, word of those two murders had not yet reached him. The passage in the later article must be talking not about 2 killings amongst a few tiny villages, but amongst all of the Semai (my search for "census" also turned up the number of Semai in the entire country: 12,700. Which works out to an *extremely* low murder rate). Knauft screwed up, big time, and his error in this one paper back in the 80's becomes a citable "fact" which Pinker can mine for his graph.
And again, it's a huge distortion of the facts to label the Semai as "nonstate" people. Under British Colonial rule, during a communist uprising in the 50s, the communists hid out in the forested hills which were the Semai's home. The government "first tried to relocate the aborigines in camps outside the rain forest. The death rate in these camps was so high that some Semai still regard their relocation as the first step in a campaign to exterminate" them (p3). The government later relented and relocated them in the forest, but near forts which could keep an eye on them to make sure they didn't help the communists. Faced with demands from the government downstream of them and the rebels upstream, Semai bands collaborated on a 6 point policy of deceit: "(1) Bands upstream and thus near the Communists were to aver that they supported the Communists. (2) Bands down river in contact with the government forces were to "support” the government." People in between were to play dumb and not give out information. "In the event of a Communist victory, the 'pro-Communist' bands were to cover up for all the bands downstream from them, claiming that all had been pro-Communist," and vice-versa (80-81). They were definitely living under a state, with all the disadvantages that entailed.
Having grossly distorted indigenous and premodern murder rates, Pinker goes on to explain the drop in rates over time as due to a "civilizing process." And yes, that means he thinks that places and people whose rates have not fallen as far as Europe or white America's have not enjoyed as much civilizing. This is the racist garbage I spoke of at the start.
Two last points about all this. First: Pinker beats on how murder can only be studied in terms of its statistical rate. But: people are not mathematicians. Societies mostly frown upon murder and have norms and rules to keep unacceptable violence in check (remembering that some murders are allowed - for self defence in America, for "honor" or revenge elsewhere). But no one in the community is going to regard murder as a problem as long as it *feels* rare enough. Today we depend on governments and news media to tell us how rare murder is (and the media, at least, make it seem far more common than it is). But for people without access to telecommunications, it was word of mouth, which can only travel so far. In a society with a murder rate of 100/100k, if each community is only 50 or so people, then any given community will have one murder every 20 years - too rare to seem like a problem that needs addressing. Once either the community size grows or the interconnectedness of communities increases to the point that people feel murder is becoming too common, then they take action to reduce violence. Even if murder was indeed as common as Pinker claims in some of these communities, because of their small size, a high rate wouldn't have felt high. And even if murder rates have indeed gone down over time, as he claims, the credit for that change does not belong to the state (modern studies show that more cops and more punishment of has little to no impact on crime rates), or to a mythical "civilizing process," it belongs to the communities that saw an excess of violence and made changes in their cultural norms to curb that violence.
Second: Gat says that he doesn't think it's useful to separate out murder and feuds from warfare. This is nonsense. War is indiscriminate. It's also directed against outsiders, and finally killing enemies is sanctioned and rewarded. Murder and feuds are directed against members of the community. They are personal, targeted, and (generally) forbidden. Dentan writes about Semai who were recruited into the military during the communist insurgency. On their own, they flee from anyone who tries to war against them. "When the British raised troops among the Semai, mainly in the west.... Many [recruits] did not realize that soldiers kill people." Asked to talk about his experiences, one veteran said "We killed, killed, killed. The Malays would stop and go through people’s pockets and take their watches and money. We did not think of watches or money. We thought only of killing. Wah, truly we were drunk with blood." But that had no effect on them once they returned to civilian life: "Back in Semai society they seem as gentle and afraid of violence as anyone else. To them their one burst of violence appears to be as remote as something that happened to someone else, in another country." If Semai can learn to set aside the killing they have done and go back to being peaceful, maybe we all can eventually learn to do the same.
no subject
no subject