Why can’t Napoleon Chagnon prove anything?
By Stephen Corry
Napoleon Chagnon’s latest book, ‘Noble Savages’, is a synopsis of his work with the Yanomami Indians of Amazonia, and is intended for non-specialists. This anthropologist is key to the recent revival of what Corry calls the ‘Brutal Savage’ myth. Chagnon contends that he holds the scientific truth, saying that his critics, especially those who have worked with the same Indians, downplay their violence. Corry shows how Chagnon makes unsupported claims, quotes from dubious sources, misrepresents his own data, and contradicts himself. Corry points out how close Chagnon was to United States officials, and how his promulgation of the ‘Brutal Savage’ underscores that of fundamentalist missionaries. Corry argues that Chagnon’s characterizations are unscientific and dangerous.
Jared Diamond and Steven Pinker’s sermon – that ‘warfare’ is ‘chronic’ in most tribal societies, as well as prehistoric ones, and that it diminished with the arrival of the state – relies in good part on Napoleon Chagnon and his ideas about the Yanomami. He is the most controversial anthropologist in America, and had he spent his life doing something other than promoting his studies of this Amazon tribe (which he calls ‘Yanomamö’),1 it’s difficult to imagine that Diamond or Pinker would have nearly as much traction with their ‘Brutal Savage’ myth.
In that sense, Chagnon’s new popular book, sarcastically entitled ‘Noble Savages: My Life Among Two Dangerous Tribes – the Yanomamö and the Anthropologists’,2 can be taken as one part of a trilogy together with Diamond and Pinker. Although it hit the bookstores last – and is clearly intended as a retrospective score-settling with his many critics – Chagnon’s revelations are primal to the renaissance of the Brutal Savage, and so should be looked at in detail.
Before doing so, let’s get some minor points out of the way. The least important is that he confuses the two organizations, Survival International and Cultural Survival,3 though it certainly shows sloppy fact-checking.
The more surprising aspect about Chagnon is his lifestyle. He requires prodigious quantities of stuff, including ‘trade goods’ far in excess of what most anthropologists would consider necessary, or could afford.4 These are primarily large numbers of steel axes, machetes and cooking pots, used to pay Yanomami to give him the information he seeks. He hardly treads lightly; for example, when traveling in his motorized dugout, he also loads his metal boat, rather like a large motorhome carrying a small car. He needs Yanomami ‘bearers’ (my word, not his), partly to transport all these things, but also to build his houses, and carry out his bidding. He orders them around with no hint that he is a guest in their territory. On the contrary, he constantly presses them to his will.5
He doesn’t make it easy for them, or himself. He has to carry, or have carried, his preserved foods,6 even to remote areas. As well as a camera, he needs a Polaroid with its bulky and expensive film; as well as one shotgun, he needs two. Even if they wanted to emulate him, few anthropologists would have the resources.
Another surprise is that for someone who has long promoted himself as an Indiana Jones figure, he often appears out of his depth, even floundering. He’s candid about this: he fails to keep insects out of his peanut butter and fungus out of his loin cloth, gets mad when the Indians play tricks on him, and upsets just about everyone.7 The Yanomami have reason enough to be dismayed, not least by his data gathering which plays on their enmities.8 To cap it all, Chagnon nearly shoots a young Indian boy, blames his gun, and reacts by being ‘badly rattled’ – he isn’t referring to the child!9
His book won’t lead many to empathize with the Indians’ humanity, or with Chagnon’s: neither are much in evidence. All this is relevant because Chagnon has long cast himself as the lead actor in his fieldwork, supposedly unfairly maligned, misunderstood, unwanted. In his latest book, the reader learns much about the anthropologist’s endless problems, but less about the Indians – apart from just how nasty many of them are.
Of course, none of this affects whether or not his theories might be right. Though even if you think they are, no one can deny that Chagnon makes extrapolations into transparently unsupported surmises.
For example, he doesn’t explain how he could come up with two sweeping generalizations on his first day of fieldwork in 1964. Arriving after a fight has ended, he asks his evangelical missionary friend to teach him his very first – perhaps prophetic – Yanomami words, ‘Don’t do that. Your hands are dirty.’10 He then immediately decides that ‘native warfare… was a chronic threat’ (his emphasis) and that ‘most Yanomamö arguments and fights started over women.’ He spends the rest of his book – and life, for that matter – trying to substantiate this blitzkrieg insight, drawing it out much further, and concluding it ‘seems’ that is how all tribal societies were, until they lost their ‘pristine’ status. Apart from his kinship studies of some Yanomami communities, how does he support such a weighty pronouncement about humankind’s history?
Let’s start by going back to 1968, when Chagnon claimed unequivocally (in his first book) that the Yanomami practiced female infanticide,11 and so had to fight for women because there just weren’t enough to go around. He provided no evidence for it, which isn’t surprising, because it wasn’t true: like most societies, the Yanomami do, very occasionally, kill babies,12 but they don’t especially single out girls. Although Chagnon repeated his claim about female infanticide in the second edition of Yanomamö (1977), he dropped it completely six years later.13 Students relying on Chagnon up to 1983 would wrongly believe the Yanomami practiced female infanticide. Those studying the same book in later editions wouldn’t encounter ‘female infanticide’ at all.14
His more generalized claim nowadays is also slippery – sometimes it’s a probability; other times it’s more definite – but it seems to be twofold. The first part is supposed to be unarguable: ‘the archeological record reveals abundant evidence that fighting and warfare were common prior to the origin of the political state and, in much of the Americas, prior to the coming of Europeans.’ As far as ‘fighting’ is concerned, so far, so good: ‘common’ is vague enough, so it’s impossible to argue with that, even though ‘the archeological record’ cannot possibly reveal its frequency.
Whether or not there was ‘warfare’ as such, is another question. Some think that didn’t start until after the invention of the state, but that hangs on what you mean by ‘war’.15 In any event, everyone knows that fighting and war were certainly extremely common after the Europeans turned up.
Chagnon then goes on to his evolutionary key, ‘Females appear to have been prized booty in those cases where large numbers of skeletons – victims of massacres – have been found together.’ In a nutshell, the killer gets the girl.
He gives just one example: Crow Creek in the Great Plains, where nearly five hundred such victims were buried in about 1325. Chagnon says there are gaps in the body count: many children and teenage girls are missing. He decides it’s ‘most likely’ they’d been captured, and ‘presumably’ the girls had become ‘extra mates for their captors’. He doesn’t explain why he labors such qualifiers – ‘appear to’, ‘most likely’, ‘presumably’ – when his thesis hinges on this being the preponderant case throughout history. Indeed, it’s the only reason he mentions it at all in his chapter entitled, ‘Conflicts over Women’.
He then, curiously, recites three confessions: ‘we don’t know directly how common fighting over women… was in the past’ (his emphasis); sites like Crow Creek are ‘rare’ (in fact, it’s unique);16 and ‘ethnographic accounts are often silent about fights over women even if they take place while the anthropologist is there.’ His omniscience about what other anthropologists conceal is, to say the least, surprising, or perhaps he simply means, ‘most don’t mention fights, but don’t let that spoil the theory.’
Moving on from what he claims is direct evidence, he turns to indirect accounts about fighting over women, and advances just two examples: writings of Spanish conquistadors; and convict William Buckley, who escaped to the Australian interior in 1803, and subsequently recounted his stories about the Aboriginals. The resulting 1852 book (also cited by Pinker)17 was an effort to make some money at the end of Buckley’s life. It includes claims which are clearly fabricated, or at least mistaken.18
That’s not very convincing, so let’s return briefly to the ‘direct’ evidence, the apparent absence of dead children and teenage girls at Crow Creek. Firstly, it’s by no means certain: both age and gender are difficult to determine from the remains, as archaeologists have stressed. Secondly, if there is any imbalance it might have been reflected in the living population as well: they were not in great shape, perhaps resulting from a lack of food due to climate change. Anyway, even if we give Chagnon the benefit of the doubt and do assume a lack of girls amongst the skeletal remains, that might still be explained in different ways. They may have fled or been sent away when attack was feared; perhaps they were spared and allowed to leave, or maybe captured and kept as slaves, or integrated into the attacking group, but not as ‘extra mates’. Perhaps they were killed after all, but their bodies not buried, or just not yet found. Who knows? Not me, not Chagnon, not anyone.19
Chagnon’s assertion that ‘females were prize booty’ is just his guess. He might be right, but if there’s any evidence that this was common, he doesn’t tell us what it is. In fact, if there’s a single shred of ‘archeological evidence that earlier people fought over women’ (the chapter subheading), Chagnon doesn’t reveal it: there’s none at Crow Creek.
There is a – literal – world of difference between saying that people kill each other and one of the things men fight about is women – both banal and obvious – and advancing a ‘scientific’ claim that men fought ‘chronically’, that the ‘primary source’ of conflicts was women, and that this was a key in the evolution of the state, and so the world into which we’re now born.
According to Chagnon, killers have more women, and more children – and grandchildren, and so on – than non-killers, and so have a genetic advantage. Genetic selection favors killers because (at least, Yanomami) society rewards them with enhanced prestige. That’s supposedly where we all came from.
Chagnon has never seen a raid: he’s going on what he’s told (sometimes, by missionaries).20 His conclusions are based on his studies of 380 Yanomami men, of whom 137 say they’ve killed someone (according to Chagnon).21 That’s a summary of ‘twenty-five years of findings on Yanomamö warfare’, and seems to be a total amassed over decades. The data was originally published in 1988 in the journal Science,22 where he cites 282 violent deaths ‘during the past 50 to 60 years’. In brief, Chagnon spends a quarter century looking for ‘warfare’ and going to where he thinks it most common, and comes up with a total of 137 Yanomami supposed ‘killers’.
To what degree are they the norm? Chagnon writes, ‘Approximately 45 percent of all the living adult males in my study were unokais,23 that is, had participated in the killing of at least one person. That is an extraordinarily high percentage’. (It’s clear from the Science article, by the way, that several must have ‘killed’ the same victim.)
What Chagnon doesn’t mention in his book, incidentally, is that many Yanomami, including some respected leaders, avoid raiding or fighting, and that this is a position accepted by their kin.
Before going on, it’s relevant to understand typical raids because they’re rather different to the impression conveyed in the book, where ‘war’ is chronic, obsessive, frequent and bloody. Twenty-five years ago, Chagnon described Yanomami attacks to specialist academic readers, ‘The number of victims per raid is usually small – one or two individuals… they usually kill the first man they encounter… as many raiders as possible… shoot the victim… and hastily retreat.’24 It’s obviously nasty stuff, and no one wants to be on the receiving end, but it’s not a matter of massacres.
Let’s take another look at the percentage given in the quotation above. The study comprises 380 men, fifty-four of whom apparently say they’ve killed two or more people, with another eighty-three having ‘participated in’ the killing of one person. As I’ve said, Chagnon concludes that ‘killers’ number ‘approximately 45 percent of all the living adult males’. He’s wrong: the actual number according to his own data is thirty-six percent. He’s inflated that by one-quarter.25
You can arrive at a figure of forty-four percent (not forty-five) only by excluding those aged twenty to twenty-four, though these men are included in the book’s tables (their ages, incidentally, are just guesses, as Chagnon says). Cutting out those in their younger twenties bolsters the conclusion Chagnon seeks, but it’s a clear massaging of the numbers. The total exclusion of all those under twenty is also a relevant failure. Some younger men would undoubtedly join in Yanomami raiding, just as teenagers and pre-teens fight in industrialized wars.26 Had Chagnon included them, it could only further weaken his conclusions.
Let’s accept his numbers at face value anyway, but just rephrase his analysis: most Yanomami don’t kill;27 and most of those who do claim to have ‘killed’, had only ever done so once.
The total sample which led Chagnon to his theory about violence – the paean for those promulgating the Brutal Savage myth worldwide – wouldn’t fill a couple of subway cars. Moreover, nearly all Yanomami ‘serial killers’, those who say they’ve killed ten or more times, were from a single village which had a reputation for unusually excessive violence. Eight of the eleven major ‘killers’ were from there,28 though these facts are not mentioned in Chagnon’s latest book.
Extracting this one exceptional settlement from the data would presumably move the averages significantly, but we’re not given enough information to do this.
So what do his total numbers really show? The Science article (but not the book) says there were 282 violent deaths over a fifty to sixty year period, in villages with a (1987) population of 1,394, and some others nearby.29 That’s a maximum of 5.1 violent deaths per year, less than 0.4% of the population. It’s a large figure, though less than in recent European wars. In WWII, for example, Soviets had proportionally six times more deaths30 than Yanomami, who are, remember, supposed to live with chronic war all the time.
It’s nearly twenty-five years since Brian Ferguson pointed out that Chagnon’s data does not demonstrate his thesis for another reason: it omits counting any children of men who are dead.31 For example, if Indians who had killed someone then died having had, say, only one child (or none), then that would change Chagnon’s averages too, and still further dilute his conclusions. This is so glaring an omission that anthropologist Daniel Lende couldn’t understand how Chagnon’s Science paper passed peer review.32 The fact that it has been criticized by scientists for its bad science, however, is conveniently ignored by those who recite it in support of their beliefs. Chagnon pretends that it’s his critics who are ‘anti-science’ – it’s simply not true.33
Chagnon retorted to Ferguson that he had collected all the information about the children of dead killers too, and would publish it, but as far as I know he hasn’t.34
There’s an even deeper concern: have all Chagnon’s supposed ‘killers’ really killed anyone at all? Marta Miklikowska and Douglas Fry have pointed out a problem with his defining the Yanomami word ‘unokais’ as ‘warriors who had killed someone’. In fact, the term includes raiders who shoot arrows without necessarily even hitting a live target (for example, into a corpse), as well as those who put fatal ‘spells’ on enemies, ‘killing’ them from a distance, shamanically rather than physically.35 In other words, in Western eyes not all ‘unokais’ are killers.
Miklikowska and Fry don’t stop there: they cite studies of other tribal peoples, both those with a propensity for violence and others with none, which come up with entirely different results to Chagnon’s, sometimes the opposite. Killers in other societies have fewer children because their lives are likely to be cut short by revenge attacks.36 In a stroke, this proves that Chagnon’s data cannot be extrapolated to social evolution in general. Miklikowska and Fry also point out that, unusually in the Yanomami case, Chagnon’s supposed ‘killers’ are on average about ten years older than the non-killers: so are likely to have had more children anyway.
Gabriele Herzog-Schröder has highlighted another big problem with Chagnon’s definition: exactly the same word, ‘unokai’, is also used for a man who accompanies his future bride during the ritual which embraces her passage from childhood to adult.37
When scrutinized, Chagnon’s vision looks less like meticulous Darwinian observation, and rather more like a Biblical fall from grace, with women as the source of all strife. Perhaps we shouldn’t be surprised that Chagnon thinks, ‘most Yanomamö men are trying to copulate with available women most of the time’!38
Chagnon claims unequivocally –without explaining how his data supports it – that ‘most [Yanomamö] disputes, fights, and wars can be ultimately traced back to conflicts… over alleged or actual infidelity by wives’. It’s worth noting his insertion of ‘ultimately’ and ‘alleged’, and then looking at his other statements, which don’t say exactly the same thing. He writes: ‘if I had to specify the single most frequent cause of lethal [Yanomamö] conflicts, it would be revenge for a previous killing’; ‘the Yanomamö have frequent fights over women but it would be inaccurate and misleading to say that they “go to war” over women’, (contradicted, by the way, in his 1977 edition, where he unequivocally states, ‘A few wars… are started with the intention of abducting women’);39 and, ‘the wars are generally the result of cumulative grievances of many kinds.’ Also, consider: ‘The motive the Yanomamö give for lethal raids almost always has to do with revenge for the death of some person. As emphasized elsewhere, the previous killing is often a result of some fight over women’ (my emphasis). (In his 1977 book, he also gives another reason: ‘the possession of the gun caused wars where none previously existed.’)40
Although it’s key in his theory, he doesn’t seem to have made up his mind about the extent women actually play in these disputes. Do ‘most’ originate with women, or is it just ‘often’? Are any wars fought over women, or not? If an Indian claims ‘women’ are the cause of fights, could he in fact mean ‘kin relations’ in a wider sense than sex or marriage partners? Could it be that occasionally men voice this, when pressed to give a reason for belligerence? After all, hooligans desperate to start a brawl often blame a victim for ogling a girlfriend. This raises a key point about whether you can justifiably examine societies in the same way you can other species: comparing what people say they do, to the direct observation of what animals do do.41
This is hardly semantics: Chagnon is advancing a theory about the evolution of social life; he’s disagreeing with most other anthropologists; he’s trying to shake our view of ourselves, to persuade us that society is a result of men’s chronic belligerence – we succeed because we kill people. He might believe it’s so, but if this is science then the onus is on him to present (at least some!) convincing evidence.
You don’t have to spend years in Amazonia to realize that some men squabble over woman (and vice-versa), and in spite of Chagnon’s view of his critics – that we supposedly romanticize tribes – I don’t know anyone who would claim otherwise. The disagreement is not that such violence exists, it’s about the degree to which it’s ‘chronic’, whether there’s any evidence that most raids originate with women, and whether the tendency to be a killer is really naturally selected for or not.
Chagnon doesn’t bring many other authors to his defense, but when he does, it’s worth noting who they are. For example, he cites the evangelical, Gordon Larson, who studied the Dani in West Papua. Larson collected the ‘stated causes’ of 179 disputes recorded over thirty years. Chagnon presents his conclusion like this: ‘The most frequent cause of these disputes was women, some 73 of the 179 cases (41 percent).’ That’s certainly a large proportion, but as always we can express the same data in a quite different way: the majority of stated causes of disputes (nearly sixty percent) were not about women. The preface of Larson’s dissertation says, ‘fighting breaks out primarily as the result of population build-up’; it doesn’t mention women at all.42
Returning to the Yanomami, Chagnon admits that other anthropologists who have worked there accuse him, ‘of inventing or exaggerating Yanomamö violence’. He cites in his defense unspecified missionaries, and state employees, but quotes from just one source, which he thinks ‘fascinating’. It describes a horrible beating of a teenage girl, and a ghastly attack by a man on his wife. Both originate (or do they?) with Mark Ritchie, a commodities trader and evangelical, who ‘befriended a group of… missionaries’ and visited their base in the 1980s.
Chagnon doesn’t tell us this, but there is no indication that Ritchie witnessed any of the incidents recounted in his 1996 book.43 All the stories really come from evangelical missionary Gary Dawson. Ritchie doesn’t make this clear. On the contrary, he writes, ‘To represent [the Indians’] story authentically [sic!], I have told it through the eyes of Jungleman, one of their most charismatic leaders. Readers may be troubled… by this shaman’s ability to get inside everyone’s head… I found myself… asking, “How did you know that, if you weren’t there and [the eyewitnesses] didn’t tell you?” He always answered the same, “I just knew,” an answer that presents no confusion to rainforest peoples.’44
It certainly confuses me. These stories, repeated by Chagnon, were told to Ritchie, who got them from Dawson. So was missionary Dawson a first-hand witness for what’s recounted in the book? He’s spent most of his life with the Indians and will have seen much, but it seems that originally most stories were told to him by his Indian converts (who are often, incidentally, the most zealous at denouncing their former beliefs). They are ‘represented’ (presumably for literary purposes) by someone called ‘Jungleman’ who ‘just knew’ about them! It’s clear from reading Ritchie’s book that Jungleman knew an uncanny amount, extending to what people experienced at the moment of their deaths.45 Such omniscience rivals Chagnon’s on first meeting the Yanomami. (Like Jungleman, Dawson and Ritchie, Chagnon also gives graphic and precisely detailed accounts of raids and killings he never saw.)46