Revisiting hunter-gatherer war with Azar Gat

It’s been nearly a decade since I immersed myself in the popular and academic debates around prehistoric violence. It was a fascinating tangent to my research into prehistoric art and consciousness, and it resulted in the brief study War & the Noble Savage. I’ve caught wind of a few bits of related material since then, and generally the grounding provided by that research has enabled me to digest and react to them critically — and little has made me consider seriously rethinking my conclusions.

Of course, in such a complex and contested arena, some day something’s going to rear up and pointedly challenge your ideas. Recently being made aware of the work of the military historian (and Major in the Israeli army) Azar Gat — specifically his 2015 paper ‘Proving communal warfare among hunter-gatherers’1 — has been one of those moments. It hasn’t completely upended the position I reached in War & the Noble Savage — I rarely settle into viewpoints brittle enough to afford such revolutions. But let’s say that there’s some stuff to be taken on board. Let’s have a look.

First, let me recap a bit. My two key reference points in my research on this topic were Lawrence Keeley’s 1996 book War Before Civilization, and Raymond Kelly’s 2000 book Warless Societies and the Origin of War. (Mind those surnames!) In War Before Civilization, Keeley set out to demolish naive ‘Rousseauian’ ideas about civilisation, or the state, being the root of organised violence. He achieved his aim pretty well. However, what I gathered from my research was that the naive position Keeley set out to demolish was a bit of a straw man. Few if any serious anthropologists believe in an entirely peaceful ‘state of nature’. The real debate was more about the extent and nature of civilisation’s transformation of violence. Supposedly neo-Rousseauian anthropologist R. Brian Ferguson has even-handedly noted that ‘what has been called primitive or indigenous warfare was generally transformed, frequently intensified, and sometimes precipitated by Western contact.’2 The debate around those terms is complex — but notice, no denial of pre-civilised warfare. Keeley’s attack on Rousseauianism has more of the character of someone scrubbing away some romantic residues than someone destroying an edifice of naivety.

In any case, work by people such as Raymond Kelly quickly clarified the debate and moved it on. (Azar Gat paints this as the Rousseauians defensively moving the goalposts, or backtracking from one concession to another. This may have a grain of truth in it, though even then I would take this as evidence of the kind of good scientific practice that people like Keeley are so reluctant to ascribe to woolly social scientists — conceding to evidence as it becomes clear, and reformulating your position accordingly. Gat’s attack is largely rhetorical.) Kelly’s position — which Gat labels ‘quasi-Rousseauianism’ — is that the real breach between ‘warless’ and ‘warring’ societies is governed not by the appearance of the state, but by the much earlier appearance of close kin networks, and further by sedentism. Kelly maintains that ‘simple’ mobile hunter-gatherers — such as the !Kung in southern Africa or the Samay of the Malay peninsula — have relatively loose kin relationships between their nomadic bands. This leads to a reduction in ‘social substitutability’, where violence can escalate beyond specific disputes between individuals into a general feud between social groups — i.e., ‘war’.

Importantly, violence is never denied. And the homicide stats touted loudly by Steven Pinker and popular writers drawing on him, showing that supposedly ‘peaceful’ hunter-gatherers have murder rates as high or higher than modern urban hotspots are also generally accepted. However, they’re contextualised with a sensitivity to the relationship between statistics and social scale that is rarely, if ever, discussed in ‘anti-Rousseauian’ (or ‘neo-Hobbesian’) circles. This context can be summed up by this fact: if you apply the murder rate for modern Johannesburg to hunter-gatherers who live in small bands, this works out (in the latter) to one killing every fifteen to twenty years.3 It’s crucially important that we use per capita rates to compare things like this across societies. At the same time, we can’t be slavish to this important principle when such strong dissonances in comparison as this become evident. Kelly remarks:

Thus [despite the statistically high murder rate among mobile hunter-gatherers] the general tenor of daily social relations observed by the ethnographer can readily be a strongly positive one of friendship, camaraderie, and communal sharing that is very rarely disrupted by argument or physical fighting.4

Now, against this background, the significance of Gat’s work is that he moves beyond the somewhat redundant critique of naive Rousseauianism, and progresses to tackling Kelly’s claims about warlessness among mobile hunter-gatherers. His main target is Douglas P. Fry, an anthropologist who seems to have taken up the role of championing the sophisticated quasi-Rousseauian position pioneered by Kelly (I haven’t read his encyclopaedic volume, War, Peace & Human Nature). Gat says: ‘Fry argues that the vast majority of hunter-gatherer life is spent in peace, with violence, and especially deadly violence, erupting very rarely, usually as isolated incidents that last briefly and are separated by years.’5 (Such is what I took from Kelly.) ‘This is quite true,’ Gat continues, ‘as long as a few crucial points are added. First and foremost, these violent occasions are sufficient to accumulate into rates of killing that are, on average, far higher than in any state society.’ It’s hard to know what to make of this. ‘Far higher’ seems to be one of those debatable points of evidence selection. More interesting to me is that fact that while I concur with Kelly in seeing significance in ‘daily social relations’ being ‘strongly positive’, even though the murder rate is ‘high’, Gat sees significance in the murder rate being high, even though daily life can be positive. Beneath the legitimate points of debate here, there seems to be a rift between the valuation of concrete experience and of abstract data. It would be easy to be simplistic about this rift — from either side — so for now I’ll leave it hanging.

Gat, however, follows this privileging of abstract data by attacking at the level of concrete experience. ‘Moreover, death in general is a hugely significant occurrence in the life of people, even if it takes only a brief moment compared to an entire lifetime. Indeed, the imminent potential of violence among hunter-gatherers is a social fact that hangs over their lives and dominates them even when violence is not activated.’ This is the classic Hobbesian line: that the state of war isn’t necessarily characterised by fighting per se, but by the threat of fighting ‘when there is no assurance to the contrary’.6 Gat is dismissing the ethnographies referred to by Kelly, without citing other ethnographies to back his point up. Rhetorically, this dubiously relies on the prevalence of Hobbesian reflexes in civilisation — the image of pre-civilised life as nasty, brutish and short. It also ignores the immensely complex question of the effects of the ‘potential of violence’ in civilisation, i.e. ‘structural’ violence and its subtle effects on psychology and quality of life. Again, it’s not something easily resolved; it’s a vital point that is better left hanging than dealt with glibly.

While Gat fails to specifically back this point up, the core of his attack on the quasi-Rousseauian position demands attention. In short, he argues that Fry et al. have studiously ignored ethnographic evidence that radically challenges their view — a whole continent of evidence, in the form of aboriginal Australia. It’s true that I don’t recall Kelly including Australian data in his analysis. Apparently Fry looks at the data, but Gat claims he misrepresents it. Gat dismisses doubts about the reliability of early colonial reports, which seems rash. But even allowing for biased reporting, the reports he cites certainly paint a picture of violent raids and endemic inter-group conflict that seems to challenge Kelly’s image of hunter-gatherer life. Until I find a clear refutation of Gat’s take on Australia, I have to take this as a very large caveat to Kelly’s ideas.

There is a major objection to Gat’s point about Australia — which is all the more compelling since he never seems to address it. That is, there is little evidence that aboriginal Australian societies tended to conform to the model of mobile foraging which Kelly sees as ‘warless’. So the evidence from Australia isn’t conclusive at all in relation to Kelly’s argument. The most glaring problem here is that while Gat highlights Australia as a unique ‘hunter-gatherer continent’ free of agriculture and pastoralism until Europeans arrived, there are a growing number of voices arguing that various types of non-foraged food production were endemic before colonisation.7 Villages, fish-trapping, fire-stick farming — there are a number of elements which, together with the highly complex kinship systems found in Australia, make Gat’s use of this region as a counter to Kelly’s image of ‘simple hunter-gatherers’ problematic. Yet again, this isn’t a debate that’ll find easy resolution. But I always pay attention when someone putting their case forward seems to avoid addressing obvious objections.

Interestingly, as with Keeley and Pinker, if you pay attention to Gat you realise he’s actually not wholly fighting the corner that popular discourse places him in. He clearly states: ‘People habitually assume that if widespread deadly violence has always been with us, it must be a primary, irresistible drive that is nearly impossible to suppress. Many find in this reason enough to object to the idea that human fighting is primordial; others regard it as compelling evidence that war is inevitable. Both sides are wrong. Contrary to fashionable 1960s notions, traced back to Freud’s latter-day theorizing about a death drive or instinct, violence is not a primary drive that requires release, like hunger or sex.’8 Keeley and Pinker also concede that culture is more important than biology when it comes to organised violence.9 So to some extent — at least as it’s popularly framed in terms of what’s ‘natural’ — the debate is a red herring. Just as no one really believes in an entirely peaceful past, no one really believes that group conflict is ‘innate’. As with most debates, the magnetism of polarisation warps perspectives like iron filings, and work is needed to sift them into their original pattern: less ordered, but more edifying.

As with Pinker and Keeley, Gat contributes some useful points to the debate. But too often, at important junctures, he’s either disingenuous or a little blinkered. Overall, once you accept the debate’s polarised friction, and allow it, absorb it, the debate seems to be in good health. The main recent development for me comes from the related debate about the origins of social hierarchies. David Graeber and David Wengrow’s recent work on this indicates that inequalities may have ebbed and flowed seasonally during the Palaeolithic, as festive gatherings coalesced then dispersed. Sometimes these gatherings constellated hierarchies, and sometimes they dispersed them. The important point is that the search for the origins of inequality may be misplaced. Graeber rephrased the question: not, ‘When did inequality arise?’ but ‘When did inequality get stuck?’ Because of the implications of this for social structure, and because of the relationships between social structure and violent conflict, we have to admit that different scenarios probably came and went many times over during that long hunter-gatherer era. I’ve not seen Graeber and Wengrow’s research brought to bear on the question of prehistoric violence, but this perspective on social structure may open up fruitful crossovers between the opposing sides of the debate on war and violence.

In the end, while we’re discussing real people and living societies (and the contemporary ones stand to suffer or even be wiped out, depending on the image we cultivate of them), most important for us to bear in mind is that we’re always to an extent in a hall of mirrors here. The debate is, among other things, a way to reflect back to ourselves our own culture’s ideological conflicts — especially since the subject matter, the distant past, is so uncertain and Rorschach-like. I’ve found that in the experience of delving deeper, and dipping in and out of this debate, what I’ve learned about the process of research, and the dynamics of disagreement and bias, has been more important than my inevitably provisional conclusions.