Epistemic status: I find the descriptive claims in this piece quite well-supported and the prescriptive claims considerably less so.

I.

We have for most of recent history been living under a boot stomping on our faces. More recently the laces may have been loosened and the boot had changed size, even shifted just enough that we forgot it’s there. This boot is civilisation. We have come up with all sorts of ingenious apologetics for this boot. With out it stomping in out face, Hobbes had argued, we return to the “state of nature”, an actual past from which we had only narrowly escaped. One of the most terribly and embarrasing enduring consequences for this 17th-century English story for defending 17th-century English boots (Hobbes, who had tutored the young Charles II as a child, was writing in the aftermath of the English Civil War and thought that without a father-figure on the throne, we will tear at each other’s throats like starved wolves) has been the caricature of hunter-gatherers as inherently violent and stupid and miserable, “the savage people in many places of America,” in Hobbes’s phrase, who “live at this day in that brutish manner.”

Hobbes in fact was simply recycling something much older I want to call the Neolithic prejudice, and most of the next section is going to be about how it works and where it came from and how thoroughly it has misled us about our own past.

II. The Neolithic prejudice

Put crudely, it is thus: agriculturalists are real people and foragers are not.

From the moment that some human groups started tilling soil and others didn’t, the tillers began constructing their own identity in opposition to the wanderers beyond the fence. This is not a claim I can defend in the finest of details with respect to the very first farmers, who were small-scale, often relatively egalitarian and left no written record. Nevertheless, the early literate agrarian states, say, Mesopotamia, Egypt, China, contains this prejudice most clearly.

Identities, as a general matter, are constitued through negation. They are constitutively exclusionary, we may say. Who you are depends, ontologically, on not being some other group, and the boundary between groups gets drawn through categorical oppositions that eventually are reified over centuries. Often the oppositions maps on to ecological and economic realities, to sedentary vs. nomadic, hill vs. valley, coastal vs. desert, serf vs. free Cossack, farmer vs. forager. James Scott has pointed out that many of the group categories we now think of as ethnic names began life as flat descriptions of where people lived or how they made a living, “those who live in the hills,” “those who live by the river” and so on, and only later grew into ethnonyms.

Often the first version is an exonym, a name applied by outsiders, and usually a slur. The San Bushmen of southern Africa get their name from a Khoikhoi word meaning roughly “those without cattle,” i.e., losers. The Eskimo are stuck with an Algonquian term that’s something like “eaters of raw meat.” Sometimes the slur gets reappropriated for pragmatic reasons (the legal and NGO categories use “San” and “Bushmen,” so people use them too) but the etymology is what it is. Hunter-gatherers, who as a class had limited access to writing or imperial bureaucracies, mostly didn’t get to write back. So what we have, in the textual record, is a more or less unbroken millennia-long monologue in which settled people describe foraging people as savages.

Therefore: the Mesopotamians called the highland Gutians and Lullu animals. The Epic of Gilgamesh gives us Enkidu, the wild man from the steppe, who lives among animals and drinks at watering holes and doesn’t know what clothes are, until a temple prostitute “civilises” him into the superior ways of Uruk. The Western Zhou and subsequent Chinese dynasties lumped their non-sedentary neighbours into the Four Barbarians i.e., peripheral peoples who didn’t follow the rites and spoke unintelligible languages and were, by definition, lesser.

The Neolithic prejudice, in other words, is what it feels like from the inside to be a settled person looking at a non-settled one from one frame of reference. The forager’s mobility is therefore rootlessness. The forager’s leisure is laziness. The forager’s lack of accumulation is poverty. The forager’s adaptability is ignorance. And finally the Neolithic farmer, who is in fact working much harder than his hunter-gatherer neighbour, can only justify his own toil by reframing the hunter-gatherer’s easier life as a deficiency. He is not lazy; the forager is. He is not impoverished; the forager is. He is not unfree; the forager is.

This is an extraordinary feat of psychological projection and the engine of about four thousand years of subsequent historical bad behaviour.

III. The Original Affluent Society

In 1968, Marshall Sahlins delivered a paper at a conference in Chicago called ‘Man the Hunter’ which became in one form or the other the most widely-cited essay in twentieth-century economic anthropology. The paper was called The Original Affluent Society and claimed that hunter-gatherer societies, judged by the standards relevant to actual human well-being, were considerably better off than agricultural and industrial societies have been since.

The claim was backed up by a small but solid set of ethnographic studies (Lee on the Dobe !Kung, McCarthy and McArthur on the Arnhem Land Aborigines, several others) that had measured, in the careful way that mid-century ethnographers measured things, how many hours per day people in surviving hunter-gatherer groups actually spent on subsistence work, how many calories they consumed, how varied their diets were, and how much leisure they had. It turned out that the !Kung worked, on average, around two and a half hours a day on food acquisition, and ate about 2,140 calories a day, which exceeded the calculated minimum for their body weight and activity level. The Arnhem Land groups worked roughly four hours a day, with the women breaking up the work with extended pauses in the shade. “Foragers”, Sahlins quipped, “keep bankers’ hours”, because their production was driven by immediate needs rather than an endless pursuit of surplus. The leisure that came about from this short workday was spent on things (visiting, talking, ritual, music, sleeping) that most contemporary readers, when they think about what they would actually like to be doing with their time, would recognise as the things they themselves wish they had more of.

Sahlins wanted to figure out why this had been so consistently misread, why is the standard narrative that hunter-gatherer socities lived miserable lives and that the story of humanity is one of cumulative progress. He suggested that Western observers who had described hunter-gatherer life had brought to their observations the assumptions of a culture in which affluence is taken to be the accumulation of material goods, in which scarcity is the foundational economic fact, and in which leisure refers to the absence of productive work rather than as a positive state. From this specific cultural/historically conditioned frame of reference, hunter-gatherers naturally are impoverished (since they had few possessions, they did not accumulate, they did not work much, they did not produce surplus) and the impoverishment was assumed to be involuntary, a function of their inability to do better. Sahlins asked, what if what if these were people who had organised their cultural lives around having few wants, easily satisfied, and the appearance of poverty to outside observers was a function of those observers being unable to recognise affluence in any form other than the one they were used to?

This is, as far as I can tell, broadly correct. Not in every detail (Robert Kelly’s The Foraging Spectrum (1995) is the standard work qualifying the Sahlins image og hunter-gatherers, in which he points out that the !Kung numbers are from a relatively favourable ecological setting and that hunter-gatherers in less favourable settings worked considerably more, that hunger was not absent everywhere it was claimed to be absent, and that the cross-cultural variability among foraging peoples is much greater than the iconic Sahlins examples suggested). But on the whole, hunter-gatherers in reasonably favourable ecological conditions worked far less than agriculturalists, ate better, had more leisure, were healthier, and lived under social arrangements considerably less hierarchical than anything that came after.

III. The Paleopathological Record

Sahlins’ case is contingent, in part, on ethnographic data from contemporary or recently-contacted foraging peoples. This is invariably open to the objection that contemporary foragers are not representative of pre-agricultural humans, that they have been displaced into marginal lands, have had centuries of contact with state societies, and may have adapted in ways that obscure what life was actually like before agriculture. This objection however is contested by an entirely separate body of evidence: the bones.

The transition from foraging to agriculture is one of the most thoroughly studied transitions in paleopathology. The methodology is to take skeletons from foraging populations immediately before the agricultural transition in a given region, take skeletons from agricultural populations from the same region in the centuries immediately after, and compare them on the standard markers of population health. The markers include adult stature (a function of childhood nutrition), enamel hypoplasias (markers of childhood nutritional or disease stress), porotic hyperostosis and cribra orbitalia (markers of iron deficiency anemia), evidence of infectious disease, dental health, evidence of chronic skeletal stress, and so on. The comparisons are not always perfect (preservation conditions vary, samples are not always comparable, the timing of agricultural adoption differs from region to region) but the field has been doing this work for fifty years and the results are remarkably consistent across regions and across research teams.

And the results are that agriculturalists, on essentially every standard marker of population health, did worse than the foragers they replaced. They were shorter (in the relevant European and Mediterranean comparisons, the drop is on the order of four to six inches, which is large). They had more dental caries. They had more enamel hypoplasias, indicating more childhood stress. They had more anemia. They had more infectious disease, both because of increased population density and because of zoonotic transmission from domesticated animals. They had more skeletal markers of chronic physical stress. The pattern is generalisable across the Old World and the New, across the Neolithic and the later transitions to agriculture in places such as North America, and across more than two dozen large comparative studies. These include the first major synthesis in Cohen and Armelagos’s 1984 Paleopathology at the Origins of Agriculture, and most recently, the 2011 Origins of Agriculture and Human Health edited by Pinhasi and Stock, which is a methodologically more sophisticated update. The data largely speaking has not substantially changed..

This is, on balance, a strange thing to know. The transition from foraging to agriculture is, in the standard story we are taught, the beginning of human progress, the moment at which our species got serious, started building things, and began the long climb toward the conditions that produced us, the readers of this essay, with our smartphones and our central heating. But the actual evidence is that the people who made the transition were, by any standard available to us, considerably worse off than the people they replaced, and that the worsening continued for thousands of years before any compensating improvements began to show up in the skeletal record. In other words, the story of the agricultural transition as the beginning of progress is not consistent with the evidence of what happened to the people involved. It is a retrospective justification reconstructed from the vantage of the eventual winners, of a development that, at the time and for a very long time afterwards, looked and felt like a disaster.

I want to flag here that I am not making the strong primitivist claim that everything about agriculture was a mistake. I am making a better-supported claim that the transition was bad for the people who made it, and remained bad for their descendants for thousands of years. Whether the very long-term consequences of agriculture are net positive or net negative is orthognal to the issue at hand, but one I will return to.

IV. Why Agriculture Won

The obvious question, given the previous section, is that if agriculture was so much worse, why did it spread? There must be some sense in which it was better than foraging, or it would not have replaced it. A kind of evolution by natural selection.

The answer is that better for the people involved and more likely to spread are quite distinct things. Agriculture spread for a set of reasons that are not very flattering to it.

First, agriculture supports much higher population densities than foraging. A patch of land that can support one forager can support fifty or a hundred farmers. This is not because the farmers are better off (they are worse off) but because they extract more calories per unit of land even when the calories are of poorer quality and the work required to extract them is considerably more onerous. So the per-capita welfare goes down while the total carrying capacity of the land goes up.

Second, agricultural populations grow faster than foraging populations. Foraging women space their children at intervals of around four years, in part because nomadic life makes it difficult to carry more than one infant at a time. Settled women can space their children at intervals of around two years, and they do. Over a few generations, this difference compounds enormously. An agricultural community that began with a hundred people becomes a thousand in twelve generations; a foraging community that began with a hundred people becomes two hundred in the same time.

Third, the higher densities of agricultural populations allow them to dominate the foraging populations they come in contact with. A small group of healthy foragers cannot stand against a much larger group of less healthy farmers, and so over time the farmers expand and the foragers either adopt agriculture themselves (under pressure rather than by preference) or are pushed into the marginal lands that the farmers do not want.

So the spread of agriculture is a story of higher-density-but-less-pleasant living arrangements outcompeting lower-density-but-more-pleasant living arrangements through the simple mechanism of producing more bodies per acre. In technical terms, agriculture is a highly aggressive form of niche construction where human-engineered ecosystems artificially expand their carrying capacity and effectively force mobile foragers out of their optimal ecological niches through competitive exclusion and habitat fragmentation. This phenomena became particualrly acute with the advent of State systems. James Scott, in Against the Grain (2017), argues persuasively that early states actively coerced their populations into staying put and continuing to farm, because farmers were a much more legible and taxable population than foragers; the spread of agriculture is, on his account, partly a story of state-imposed labour discipline and not one of voluntary uptake. (Whether one buys the full Scott account or not, the available evidence, including extensive evidence of foraging populations actively resisting agricultural neighbours, often successfully, over long periods, is clearly inconsistent with the standard story in which agriculture spread because everyone could see it was better.)

The mechanism by which agriculture won is a mechanism that aselects for competitive expansion and not for welfare. What came from it is a world in which the dominant mode of human life is the one that out-bred and out-coerced its alternatives not the one that made human beings happiest or healthiest or most fulfilled. There is no reason in principle to expect such a world to be a particularly good one for the human beings living in it, and on the available evidence, it has not been.

The same pattern shows up at several other major transitions. The shift from peasant agriculture to industrial wage labour, for example, was for most of the people who made it a sharp decline in welfare. The hours got longer (peasants worked roughly 150 days a year on average; early factory workers worked 300 days a year, twelve hours a day). The conditions got worse. The diets got worse. The communities got worse. Life expectancy in early industrial cities was substantially below life expectancy in the rural areas the workers had come from, and remained so for decades. The transition occured anyway because industrial economies could support higher populations and could outcompete agricultural economies militarily and economically. It was not until well into the twentieth century, in the parts of the world that had industrialised first, that material conditions for the median person caught up with where they had been before industrialisation, and even then the catching-up was uneven and contested.

I am not making the strong claim that everything is getting worse (central heating and modern medicine seem to be indivdually good) and I am not making the strong claim that there is no progress, because some things have gotten better in ways that are unambiguously valuable. I am making the specific claim that the mechanism by which large-scale human social transitions tend to happen is a mechanism that does not select for the welfare of the people involved and that the result of running this mechanism for ten thousand years is a world in which the people living in it are, on a number of measures that matter, worse off than they would have been if the mechanism had not run.

V. The Standard Objections

I want to now spend a section on the standard objections to this view.

Objection 1: But infant mortality. Pre-modern populations had very high infant mortality rates, and the comparison between modern life and pre-modern life is therefore wildly unfavourable to the pre-modern side once you factor in the children who did not survive to grow up. A foraging society in which a third of children die before age five is not, on any reasonable accounting, a society in which life is good.

Infant and child mortality rates among pre-agricultural foragers are indeed higher than ours, but were not as catastrophic as the rates among agricultural and early industrial populations, which is the comparison that counts for the did agriculture make things better? question. The high infant mortality numbers everyone has heard of (a third of children dying before five, life expectancy at birth in the thirties) are mostly numbers from agricultural and early industrial populations and not from foragers. Foragers had better infant mortality than agriculturalists. The best current estimates, from Gurven and Kaplan’s 2007 review suggest pre-modern forager life expectancy at birth was around 30-37, with a substantial fraction of children dying young, but with adult life expectancy conditional on surviving childhood being something like 70, which is not catastrophically different from contemporary numbers.

Modern medicine has dramatically reduced infant and child mortality, and this is an unambiguous good that the pre-modern world did not have. I agree. This is one of the things I meant when I said I am not making the strong claim that everything is getting worse. Infant mortality is an unambiguous improvement and it is a large one. What the data does not support is the inference from infant mortality has improved to therefore pre-agricultural life was bad on every dimension and modern life is better on every dimension. The infant mortality improvement is true and it is recent (mostly post-1900) and it does not redeem the long pre-modern period during which infant mortality was, if anything, worse than among the foragers.

Objection 2: But violence. Foraging societies had higher rates of violent death than modern societies, by a substantial margin. Steven Pinker’s Better Angels of Our Nature (2011) makes the case [deceptively, I believe] at length, that pre-state societies had violent death rates of around 15% of all deaths, modern Western societies have rates of well under 1%. The reduction in violence is one of the great achievements of state civilisation and it is being elided by the picture I have just painted.

The data Pinker draws on is partly true but the case is considerably weaker than the popular reception of Better Angels would suggest. The 15% figure is an average that conceals enormous cross-cultural variability - many ethnographically-attested forager societies had violent death rates well below 5%, and the high figures come disproportionately from a small number of cases (Yanomami, some New Guinea highland groups) that are not representative of the foraging mode of life. Brian Ferguson’s careful re-examination of the archaeological data (Pinker’s List, 2013) suggests that the deep paleolithic record shows much lower rates of inter-personal violence than the Pinker thesis claims, and that the high rates appear in the archaeological record only with the emergence of agriculture, sedentism, and the social conditions that produce defended settlements and stored surplus worth fighting over. So yes, contemporary state societies have lower violence than pre-state societies, but the lowest-violence societies in the historical record are the deep-paleolithic foragers, not the moderns; the violence increased with agriculture and sedentism, peaked in the early state period, declined again through the long modern period, and is now lower than at most points in the historical record but probably not lower than at the deep-paleolithic baseline.

Objection 3: But culture. The argument I have just made measures human welfare on a fairly limited set of axes (calories, height, leisure, mortality) and ignores the dimensions of human life on which civilisation has produced enormous gains, such as literature, science, philosophy, art, the cumulative cultural achievements that distinguish a literate post-Enlightenment society from a paleolithic band. A defender of civilisation might concede every empirical point I have made and respond: yes, but the achievements of human culture are worth the costs, and any accounting that does not include them is an accounting that fails to take seriously what it is to be a human being. I am the kind of person who reads books and listens to music and finds these things genuinely valuable so I take this objection seriously.

However, I take what this objection is actually communicating is thus: but I would not want to live the way the Hadza live, because the things I value most are not available there, and therefore the civilisational arrangement that makes those things available is justified. The problem with the objection in other words is that the I who is doing the valuing in I would not want to live without these things is not a stance-independent observer of the relative merits of civilisational and pre-civilisational life. It is a person whose matrix of valuing (the categories under which experiences become salient, the things that count as fulfilment, the practices that constitute a life worth living, the capacity to be the kind of person for whom reading novels in the afternoon is a coherent description of well-being) has been produced by the civilisation that is under examination. The desires that civilisation does not satisfy in pre-civilisational arrangements are desires that civilisation itself instilled. That is, they are not pre-existing human needs that civilisation arose to meet, and are rather products of the cultural conditions civilisation produced, and they could not have been desires in any robust sense before the conditions that made them intelligible were instantiated. The Hadza do not lack these desires and suffer their absence, they have a different set of desires, also culturally produced, which their arrangements satisfy adequately well, and on the available evidence they appear to be at least as content with their lives as we are with ours, and probably considerably more so.

Affluence, on the Sahlins definition, is the ratio of desires to means. There are therefore two ways to be affluent: produce more, or want less. Industrial civilisation has bet everything on the first strategy and has produced an enormous amount but it has also been continuously producing new wants at a rate that exceeds the rate at which the production keeps up with them, and with the consequence that contemporary populations are, by their own self-report, no more satisfied than populations a generation or two ago and considerably less satisfied than the foraging baseline. The Hadza have bet on the second strategy and have ended up with a ratio of desires to means that comes out, on most days, in their favour. To then say but I would not be content with the Hadza means is to say I have been successfully conditioned into wanting more than the Hadza want, which is true and which is a description of the problem not a defence of the system that produced it.

The civilisationalist at this point may retort: fine, the desires are produced, but I am the person who has them, and the question of what I should value cannot be answered by appealing to what some other version of me, raised under different conditions, would have valued instead. The desires I actually have are the desires that count for me. This is correct as far as it goes. However, it forces her to bite the bullet that one cannot use but I value these things as an argument for the system that produced the valuing, because the valuing is, on this version, simply a brute fact about a particular person rather than a normative assessment of the system. I like ice cream and I am glad ice cream exists is true and does not constitute an argument that the entire industrial dairy complex is justified. I value novels and I am glad civilisation exists is pretty structurally similar.

So, yes, civilisation has produced things I value, and yes, my valuing of them is itself a product of the civilisation that produced them, and yes, I am stuck with both facts and have to live as the person I am rather than as some hypothetical version of myself who would have wanted different things. This is fine. It is also not a defence of civilisation in the sense the objection requires. It is a description of the inheritance one happens to have received, with no particular normative implications about whether the inheritance was, on balance, worth what was paid for it.

VI. The Anti-Civ Position

If you have followed me this far, you can see roughly where the argument is going. There is a position called anti-civilisation, or primitivism, which takes the descriptive claims I have just made and draws from them the conclusion that civilisation is a mistake, that we should dismantle it, and that the project of human flourishing requires a return to something closer to the foraging mode of life. The position has a small but very ardent and serious literature (John Zerzan, Derrick Jensen, Kevin Tucker, the late John Moore) and it has a slightly larger fellow-traveller literature in writers such as James Scott and David Graeber, who do not endorse the prescriptive part of primitivism but whose descriptive work largely supports the primitivist diagnosis.

I find the descriptive case strong. I find the prescriptive case weak for reasons I will give in a moment. But I want to insist on something before I get to the prescriptive disagreement, which is that the descriptive case is strong whether or not you accept the prescription. One can hold that civilisation has made human beings worse off in important ways, and that the standard story we tell about progress is largely ideological, without thereby being committed to the view that we should burn down the cities and return to the savannah. These are different claims and they should be evaluated separately. Most of the energy spent by mainstream commentators dismissing primitivism is spent on the prescriptive claim, with the descriptive claim being assumed to fall along with it. This is sloppy. The descriptive claim is not refuted by the impracticality of the prescription.

I have my own reservations about the prescriptive claims.

First, is that I do not think we can return. There are eight billion of us. The carrying capacity of the planet under foraging conditions is on the order of five to ten million people. The transition from where we are to where the primitivists want us to be involves the death of a number of people that is difficult to write about plainly - between 99.9% and 99.95% of the current population, depending on how you do the arithmetic. I do not see any pathway from here to there that does not involve either the most catastrophic mass dying in human history or some kind of voluntary mass suicide of the technological infrastructure that is keeping eight billion people alive. The first is monstrous. The second is not going to happen. So the primitivist prescription is either a counsel of mass death or a counsel of impossibility, and neither is a basis for politics.

The second reason is that pre-agricultural life was not uniformly egalitarian and fulfilling. The Hadza are wonderful in many ways but they are also a people in whom violence between men over women is not unknown, in whom the elderly and the disabled are sometimes left behind when the band moves, in whom the lives of women are constrained in ways that contemporary feminists would not endorse. The deep-paleolithic record shows similar patterns where they were better than what came after but not amazing by contemporary standards. So I do not endorse the prescription of the call to returning to exactly those conditions.

The third reason is that I value some of the things civilisation has produced, including some of the things I am using to write this essay, and I am not willing to commit to a position that requires me to deny that I value them. Modern medicine and antibiotics (though arguably a large reason we require them is a product of civilisation). The novel. Recorded music. The ability to read what other people who lived a very long time ago thought about questions that I am also trying to think about. These are not nothing. The strong primitivist position requires me to either pretend they are nothing or to say they are not worth the cost of producing them and I am not prepared to do either.

So I am not really a primitivist. I find the descriptive case the primitivists make to be largely correct, and I find their prescriptive case to be largely wrong, and I think it is possible to hold both of these views consistently.

VII. What I Actually Think, A Forward-looking Anti-Civ

Here is what I have come to think after several years of working through this material.

The strong primitivist position is wrong but the strong civilisationist position (the standard story of progress) is also wrong, and is wronger. The strong primitivist position is wrong because we cannot return and most of us would not want to even if we could. The strong civilisationist position is wrong (more wrong!) because the evidence does not support it, because most of the gains it celebrates are very recent and partial, and because the ideology of progress conceals a great deeal in our collective sense of what kind of world we live in and what kind of trajectory we are on.

What the descriptive case holds is something like the following. Human social arrangements have, over the long sweep of history, tended to evolve in directions that are not selected for human welfare, and the mechanism by which this happens is a mechanism that runs roughly continuously and that is operating right now. The things we are currently doing to ourselves and to one another and to the planet are the contemporary expression of this mechanism, in which arrangements that produce more output, more legibility, and more competitive position outcompete arrangements that produce more human flourishing. We are not at the end of this process. We are probably somewhere in the middle of it and the next iteration is likely to make many of the things we currently take for granted to look like, in retrospect, the false plenitude of the early agricultural surplus.

The position I personally occupy is closer to the anti-civ position only in the sense that I take the anti-civ analysis to also be a very potent structural claim about the present and the near future. The claim is that civilisation (in the specific sense of the large-scale grain-cultivating arrangement that began about ten thousand years ago and that all subsequent state societies have been variations on) is structurally self-terminating. The kind of arrangement that it is cannot exist without producing ecological damage at a rate that exceeds the rate at which the underlying ecosystems can regenerate. Things typically considered ‘externalities’ to civilisation, such as topsoil, fresh water, atmospheric carbon, biodiversity, the integrity of the nitrogen and phosphorus cycles, are all in fact the substrate it consumes merely in order to operate. And we are running out of all of them. We have run out of new landscape to expand into and there is no further frontier to move into when the current one fails. Add to this that the rate of depletion is, by every measure available to us, considerably higher than the rate any previous civilisation managed. Soil loss in the contemporary American Midwest is on the order of an inch per decade against soil formation rates of an inch per several centuries. Atmospheric carbon is at concentrations not seen in the past four million years. Biodiversity loss is in the range that paleontologists call mass extinction events, of which there have been five in the previous five hundred million years and which we are now in the early stages of the sixth. This is made worse when we take into account the fact that technological infrastructure we have developed to extract resources is itself dependent on the resources being extracted, such that the system becomes increasingly fragile as the extraction proceeds. We cannot run the carbon-extraction infrastructure without carbon, the rare-earth extraction infrastructure without rare earths, the food production system without the topsoil and the phosphate and the fresh water that are running out. The system has, in short, no exit. It can continue along its current trajectory and run out or it can be deliberately scaled down, and there is no third option in which it continues at scale on different inputs because the inputs it depends on are the only ones available in the quantities it needs.

The anti-civ position, construed in this way, seems to be a kind of call for the immediate recognition and amelioration of our common trajectory, which is to say, to the end of civilisation. Civilisation in its current form is self-terminating. The only question is whether the termination happens through deliberate scale-down or through systemic collapse. On the available evidence the second is considerably more likely than the first because the institutions that would have to lead the scale-down are themselves dependent on the system continuing at its current scale and have no incentive to start the work. In a sense, anti-civ is therefore the most rational position available; it is not a position about what one would prefer if one could choose freely between modes of human life. It is rather a position about what is going to happen, on roughly geological time scales which are now in fact human time scales because the rate has accelerated, regardless of what anyone prefers.

I think this position is true and also the position that, if more people held it, would produce slightly better outcomes than the alternatives. I do not think most people will hold it because the ideology of progress is very effective.

If the standard story of progress were true, there would be a much stronger presumption in favour of further scaling, further centralisation, further technological dependence. Because most people would benefit from each new development, the burden of proof would be on those resisting it. If the standard story of progress is false (if the actual record is one in which large-scale social-technical changes typically make most people worse off for long periods, with the eventual benefits being concentrated and partial) then the burden of proof should run the other way. Each proposed scaling-up is more likely than not to be welfare-reducing for most of the people affected, which begs the difficult questions about who is going to bear the costs, who is going to capture the benefits and whether there are smaller and less-disruptive alternatives that could deliver the actual welfare gains without the full cost.

I would rather be wrong in this direction than wrong in either of the other two, and I would rather have the conversation that this position makes possible than the conversations the other two foreclose.

That is roughly where I am.