Why Become a Farmer?



Join 36.9K other subscribers

The previous blog discussed Göbekli Tepe, which achieved a surprisingly high level of social complexity before the adoption of agriculture. In the language of philosophy of science, Göbekli Tepe is an anomaly for the reigning paradigm in theoretical archaeology, which posits that the adoption of agriculture was the pre-condition for, or even the cause of, the rise of complex societies. As such anomalies accumulate (and we shall see later that Göbekli Tepe is not an isolated case), eventually the paradigm collapses, and is replaced by another (at least, so argued Thomas Kuhn in The Structure of Scientific Revolutions).

So what kind of theory may replace the standard ‘bottom-up’ theory, when, or if, it collapses under the accumulated weight of anomalies? In the previous blog I found the alternative explanation, described in the National Geographic, unsatisfactory.

Actually, the bottom-up theory also has a huge logical gap. It assumes that when foragers discovered how to cultivate plants, they gladly switched to this more productive way of making a living.


Comic by David Steinlicht

But this makes sense only from the point of view of a person who came from a culture with long-established farming practices (such as academic anthropologists). Real foragers, on the other hand, are very resistant to abandoning their ways of life, and they have good reasons for such reluctance.

If you think about it, why humans switched from hunting and gathering to agriculture is a big, big puzzle. This has been most effectively argued by Sam Bowles (most recently in a PNAS paper co-authored with Jung-Kyoo Choi).

The first problem is that farming involves a lot of back-breaking labor. Early agriculture was not particularly productive, and it required many more hours of work, compared to foraging.

Second, hunter-gatherer societies share the food; hoarding marks you as an anti-social deviant. What this means is that you could put all that work into growing plants (clearing the field, planting, weeding), but others would think nothing of harvesting the plants when they ripen. Or you could get to the point of harvesting and storing the crop, but then everybody else in your community would expect you to share it.

Third, agriculture has its dark side – it’s negative effect on human health. Evidence is overwhelming that after switching to agriculture human heights decreased, and that is a very reliable indicator of a decline in biological well-being. People fell sick more often because of higher population density and because pathogens jumped from domesticated animals to humans. The quality of nutrition declined, as is abundantly documented in ancient bones and teeth.


A skull fragment from a teen buried at Rome’s Casal Bertone necropolis. The teen ate a millet-heavy diet in childhood but switched to wheat in the years before death. Pores in the bone of the eye socket known as cribra orbitalia suggest the teen was anemic. CREDIT: Kristina Killgrove Source

But the switch to agriculture did occur, and farming did spread, so there had to be a compelling reason why. Let me sketch a possible explanation, which fits the various data pretty well (or so it seems to me). Long-time readers of this blog will not be surprised to discover that the explanation focuses on warfare, or more properly cultural multi-level selection.

Consider a landscape inhabited by many local groups, each with a territory that they use for hunting and gathering. These groups (typically around 50 individuals) are in turn grouped into larger ‘tribes’ – ethno-linguistic communities (typical size = 1-3 thousand individuals) who cooperate in warfare against other ethno-linguistic communities.

Now suppose that something happens that leads to more intense warfare. In historical times, the most likely trigger has been the invention or spread of a new military technology — bows and arrows, cavalry, gunpowder. (We have a paper in review modeling this process, and it predicts the rise of large-scale states and empires in history remarkably well.)

But the most likely trigger at the dawn of agriculture was climate change. When climate became dryer and cooler during the period, known as the Younger Dryas, the productivity of plant communities declined, which caused a decrease in the carrying capacity for humans who depended on these resources (game animals also decreased because their food base shrank).

Lower overall resources led to a spike in conflict between ‘tribes’ (ethno-linguistic groups) because each group attempted to expand its territory to compensate for lower carrying capacity. A possible example of this dynamic is the intensification of warfare among the Anasazi after climatic conditions in the American Southwest deteriorated.

Most likely humans who lived in the Fertile Crescent had already known about techniques needed to intensify plant production, but for reasons we have discussed, did not deploy them. The new conditions of widespread warfare, however, imposed an intense selection regime for larger group size, because the best way to ensure tribal survival was to have more warriors. Growing their own food enabled human groups to raise more warriors and concentrate them within larger war bands. Such groups then expanded at the expense of groups that didn’t have agriculture. So we have a typical process of evolution by cultural group selection.

Why was the cultural group selection necessary? Because you cannot switch to farming when everybody else in your group is foraging. The whole group needs to shift to farming together and to acquire a new set of cultural norms, most notably, private property rights. Bowles and Choi in their paper model how this dual switch can occur.

However, more is needed. Switching to farming makes evolutionary sense only if it leads to a larger tribe size, which is key for surviving under conditions of intense intertribal warfare. But it is not easy to keep a large group of people internally cohesive. You need a new type of social glue.

In his recent articles, including one on the Social Evolution Forum, Harvey Whitehouse argues that large-scale human societies can build up cohesion by inventing and conducting regular symbolic, or ‘doctrinal’ rituals that bring together large numbers of people. Everything we know about Göbekli Tepe suggests that it was used precisely for such rituals, and that it served a very large ritual community. It took many hundreds of people to build the monument, so there had to be a large community numbering in many thousands, since somebody had to provide the food. And it brought together population from a large area.

The cultural group selection logic also explains why agriculture was adopted despite its huge health costs. Groups of poorly nourished and perhaps even chronically sick, but numerous farmers exterminated healthy and tall foragers because of the group size advantage. So individual fitness (both in the evolutionary sense, and in the common sense of physical fitness) declined, but the evolutionary fitness of the group increased, and that is what drove the whole process.

So here’s the logic of my explanation. Rampant warfare, resulting from climate change, leads to intense selection for larger society size. In order to make this transition, a number of seemingly disparate, but actually synergistic cultural traits need to coevolve. One bundle of cultural traits, which is needed, is what makes agriculture possible – not only knowledge of how to cultivate plants and herd livestock, but also new social norms such as property rights. Another set of cultural traits, which actually had to appear first, was what glued together large groups. This is why monuments, used for ritualistic purposes by large groups of people, according to this theory, can (in fact are expected) to appear before the transition to agriculture.


Notes on the Margin: I will be away from the Internet during the next week; when I come back I will continue with this series. Next is a discussion of Poverty Point.

Notify of
Most Voted
Newest Oldest
Inline Feedbacks
View all comments

“This is why monuments, used for ritualistic purposes by large groups of people, according to this theory, can (in fact are expected) to appear before the transition to agriculture.”

Makes sense i think except i don’t believe you need climate change to hit a tipping point in the level of conflict just a region where the resources available to HGs were particularly abundant thus leading to relatively high population densities. It could even be positive climate change in the sense that it *increased* resources available in certain regions leading to an increase in population and a subsequent increase in conflict and a search for a way of defusing it.

A lot of ancient priesthoods were the judges.

Peter Turchin

That’s right, a favorable climate change could increase the population density, and when optimal conditions are over and the carrying capacity declines, this could serve as a trigger for a spike of warfare.

T. Greer

Great essay.Your summary statement in the concluding paragraph is well said, and I agree that the group of cultural traits you identify are prerequisites for functioning complex societies and agriculture.

But I am still left with a few questions.

1) There still seems to be a “then a miracle occurs” moment in the progression you set forth. You arrange things thus:

1) Climate change leads to 2) intense competition between tribes which favors 3) larger tribal groups which need 4) a strong sense of social cohesion to work effectively which leads to 4) the capacity to adopt agriculture.

The problem with the sequence is that it does not explain why agriculture was adopted. It explains the necessary prerequisites to agriculture (social cohesion, complex society), and it explains why agriculture was such a useful adaptation (larger population, and by extension, larger fighting force). But we are missing the mechanism that brought this transition about.

Did hunter gatherer tribesmen realize that adopting agricultural production would raise their population over the next few generations, thus allowing them to establish martial dominance? If we asked one of these people in the midst of the transition why they were growing plants in the ground, is this what they would say?

Probably not. It is a good explanation for why agriculture succeeded after it began; it is less useful for explaining why individual communities actually made the shift.

I would submit that sedentarism plays an important role here. In North America we see large, complex hunter-gatherer societies in places of unusual resource abundance. The pacific Northwest is a great example – it has a higher biomass per square foot than anywhere else on the continent. Tribal groups in the Pacific northwest did not need to cover large areas or move around often to find game or forage. They could survive by staying put.

This leads to a few important things: increased population, the chance to build and use buildings/ritual centers, and perhaps most importantly, the chance to raise animals or domesticate plants in one place. It was probably very accidental to start out with – the seeds of what the H-G folks ate started growing next to their camps (perhaps in their waste areas?), and the H-G folks were still around when it was time for harvest. (Similar process with animals – only dogs could be domesticated before settled society arose).

2) Is there any evidence that Younger Dryas climate change affected plant communities adversely in the regions where agriculture developed? (Coming out of the ice age boosted plant productivity in the Pacific Northwest – but then again, they never had agriculture. As suggested above, they survived just fine off of the abundant forage the new landscape provided).

Peter Turchin

How a cultural innovation arises, and why it spreads are to separate questions. I have focused on the second, but I agree that the first is an interesting and important one. As others below suggest sedentism seems implicated, and it looks like a very plausible mechanism to me.

Nikolai Rozov

Peter, why don’t you use the circumscription theory by Bob Carneiro for explanation of political evolution? This model contains such powerful conceptual components as: 1) phisical circumscription (nowhere to escape) 2) social circumscription (all neibouring places are already occupied), 3) resource gradient (any other place is worse), 4) population pressure (following 1-3), 5) growth of shiefdoms’ size because of warfare and alliances, 6) class relations between winners and loosers and coercion for labour – especially for hard labour on fields. All these concepts (factors) do not contadict your group evolutionary idea but make it more rich and clear.

Peter Turchin

I respect Bob’s thinking and contributions a lot. His influence on posing the right questions, and his emphasis on, and defense of warfare as the primary driver have been extremelky productive. However, I disagree with that part of his theory that emphasises circumscription. This is a big topic and needs to be addressed in a separate blog. For now, I will just note that his theory is still a variant of the ‘bottom-up’ scenario.


I think that the main problem with both Peter’s and Sam’s scenarios is that the transition was not from standard mobile egalitarian hunters straight to farmers. Rather, mobile egalitarian hunter-gatherers with high percentages of meat in their diet first evolved into semi-sedentary plant intensive gatherer-hunters. The Natufian and similar Fertile Crescent gather-hunters are one example. The Jamon tradition in Japan was a very long-duration example of gatherer-hunters. Californians had had such a tradition for a few thousand years before Europeans arrived. These are already very labor intensive systems. Some proto-agricultural techniques are often present. For example Californians cultivated tobacco and Great Basin people diverted water onto grasses and other herbaceous wild plants from which they then harvested a richer crop. As far as we know, these systems are restricted to the Allerod-Bolling warm period at the very end of the Pleistocene and to the Holocene. People certainly used plant resources in the Pleistocene, but Bob Bettinger and I don’t see any good evidence that the labor and plant intensive systems existed during the last glacial. The signature artifacts of these systems are abundant heavy stone milling equipment which should be archaeologically highly visible.

The idea that the Younger Dryas millennial cold snap triggered agriculture in the Fertile Crescent is perhaps the standard hypothesis, championed by Ofer Bar-Yosef, the Dean of Fertile Crescent archaeology. Bob and I are skeptical on two grounds. First, Younger Dryas people seem to have substantially moved out of plant intensive subsistence back into more mobile hunting and gathering. That is,away from plant intensive techniques not toward more intensive versions of the Natufian and related systems. Second, most of the other primary developments of agriculture seem to have been later than the Fertile Crescent one. The Younger Dryas push hypothesis can’t be a general explanation.

The gatherer-hunter economies are pre-adapted to move into agriculture but few did. Mark Blumler argues that the pre-adapted economic system has to meet up with a pre-adapted plant species or two for cultivation to take off, specifically large seeded annual plants (or suitable tubers in some cases). California’s wild flora lacks an analog of wheat, barley, or teosinte (wild maize). The most important tubers were thumbnail sized. In California and Japan, one of the staples was acorns. They are plenty big, but not good candidates for domestication. In both California and Japan, agriculture was brought in by invaders, not adopted by natives by cultural diffusion. The speculation is that, at the margin, growing maize or rice did not look attractive to gatherer-hunters who had a system of acorns and whatever down pat. Switching to cultivation would required a lot of costly experimentation to perfect and even then might be no better or worse in terms of labor per unit food energy at prevailing population densities. On the other hand, if your target wild crop was a large seeded grass or forb, skills at handling it would already be in place and at the margin steps like deliberate sowing might have positive returns. You’ve got less of an innovation hump to get over and can proceed to agriculture one small step at a time.

The role of warfare, if any, might be fairly subtle. Gatherer-hunter California was relatively peaceful at the time of contact due, according to Bob, to specific institutions that damped down feuding and inter-tribal warfare. Polly Wiessner’s work in violence among the horticultural Enga of the New Guinea Highlands points to cycles of violence and peacemaking similar to what Peter has studied in agrarian and modern societies. I like to think that the social organizational innovations are disproportionately in the peacemaking phase of the cycle rather than in the warmaking one.


Re: ice age milling equipment.
A recent report says that Chinese archaeologists have unearthed grinding stones 23,000 years old from the Yellow River area.
So, they were grinding seeds ten millennia before agriculture, during the ice age.

To my mind, it implies that there needed to be a few millennia of eating ground starches to gradually pre-adapt people to the switch-over to agriculture and surviving primarily on this strange diet. (Not saying this is a sufficient explanation though.)


Grinding equipment goes back well into the late Pleistocene, In addition to possibly grinding seeds, such equipment was used to grind ochre for pigment and possibly for other uses like polishing wood. The existence of this technology is definitely a preadaptation to agriculture. According to Bob, before agriculture grinding kit is relatively rare and often small. He says the commonest non-specialized way to deal with plant seeds was to parch them–essentially make them into crude popcorn. From this you only get a fraction of the calories you get from labor intensive milling, but it is quick and easy to do. Once people start grinding 1500 kcals of seeds per capita per day, grindstones become abundant and large. He claims it is easy to recognize the advent of large scale labor intensive seed grinding. His experience is in California and in China specifically looking for the Chinese Neolithic (advent of farming). I think he represents a pretty is pretty firm consensus that Chinese agriculture is nearly as early or as early as in the Near East, but in the Holocene.

Peter Turchin

Thanks, Pete and Martin, for your thoughtful comments. I need to digest these points before I can offer a considered response! Two things I want to say, however. First, I agree that sedentism was probably an important stage in the transition to agriculture. But I don’t think that it affects the hypothesis that I offered in the blog. Sedentism together with gathering and storing resources could help with the evolution of property norms, but it does not explain why monumental architecture and large-scale rituals evolved. Second, paleolithic humans consumed a lot of carbohydrate, and not only from seeds. They also dug up, using sharpened sticks, a lot of roots and tubers, which could be ground up into flour and store for long times.


I follow Bob Bettinger on these things. He has thought a lot about the issues because the sequence in the American West changes so dramatically over the Holocene. Early Holocene people in the West were highly mobile specialists on big game. In this he thinks they resemble late Pleistocene hunters. They made relatively little use of plant resources.

Now, humans are omnivores. They will seldom pass up nice nuts, sweet berries, and other easy to harvest tasty plant resources. They would probably go after a wide variety of plants in emergencies. Bob says everybody knows how to parch seeds, what tubers are nutritious and what poisonous. Maybe how to detoxify some of the poisonous ones. But the emphasis was on moving frequently to pursue hunting opportunities to the exclusion of focusing on accumulating big stores of plant resources that are hard work to process and can’t be moved great distances.

The switch to intensive use of plant resources he thinks requires a socioeconomic revolution. Plant collection and plant processing is women’s work and to specialize on plants requires that women get a much larger say in how the economy is run. Semi-sedentism prevents men from pursuing the most advantageous hunting opportunities as does even the highly mobile plant specialization of the desert Great Basin. It required prioritizing women’s pinon nut groves over men’s mountain sheep hunting opportunities. In the Great Basin this switch was very late, only a few hundred years before Europeans arrived so the archaeological data are very good.

James Waddington

As usual, a challenging hypothesis, very thought provoking.
There is something in the air. In The Guardian on Saturday there was a review of Daniel Dennet’s Intuition Pumps and other Tools for Thinking, in which Steven Rose generally slagged off Dennet, and then himself got a bit of a going over in the comments. A CJ Hacket wrote, for instance, “You can’t talk favourably of group selection and expect to be taken seriously.”
CJ Hacket has a point. Quite a few philosophers, sociologists and anthropologists invoke the notion of human culture being an evolutionary process, while in fact ignoring evolution and even having a ball with teleological accounts which run directly counter to anything Darwin might have had in mind.
A case in point. Eminent men (not just Steven Rose, and I think they’re all men) have recently talked about evolution taking place at the level of the group. Maybe I’m an old-fashioned literalist, but to me evolution involves re-iteration, variation and selection. All three of them. The only selection process that might be exerted on a group is by that group’s environment, which will include other groups (ethno-lingusitic in Peter’s example). A group in this sense has a certain amount in common with a species in biological evolution. You might say that competition between species is in some sense evolutionary, in the competition between hunter and prey perhaps. But it would be a very partial assertion, because at the same time you would have to acknolwledge that evolution only takes place at the level of a genetic, epigenetic, somatic and behavioural envelope; the individual. It is individuals who are selected or not by their environment. No other hypothesis is scientifically current.
In Peter’s article, he is seeking to explain at least two things. The first is the evolution of agriculture itself. As Peter Richerson sugests with his usual gentle cogency, this must have been a very lengthy process, going back as far as you like, but at least 100 ka, and immensely complex, and also in an evolutionary interdependence with the rest of the behaviour of the human organism, fighting, painting and so on. Peter (Turchin) is therefore very sensible to take this evolution as a given, and present agriculture as a fully functioning system which was available for adoption around 11 ka ago.
The main theme of Peter’s article, as I understand it, was why adopt a system, agriculture, when it was inferior in terms of economy and health to hunting and gathering?
Which is perhaps less of a question if one takes the evolution of culture seriously. Because then conditions in the human environment, which includes all other human beings and all the rest of the universe, including earth, flora, fauna, weather, the heavens, must have acted as selective forces on the thousand or ten thousand or a million evolutionary strands, behavioural and material, that ended up as what we used to call the Birth of Agriculture. Which probably didn’t seem to be at all like the birth of agriculture at the time, but merely what people did.
I sometimes feel that we might understand this better if we’d spent a year or two doing subsistence farming as a matter of life and death.

Peter Turchin

Hi James – I don’t know who CJ Hacket is, but he is sadly behind on one of the most important recent developments in social sciences – the impending triumph of cultural multilevel selection (if I may be allowed a little hyperbole here). I wrote about it here:


I take cultural evolution very seriously. So to explain the adoption of agriculture you need to present a coherent evolutionary scenario of how this cultural innovation arose and why it spread. For reasons detailed in the last two blogs, I don’t feel that the standard bottom-up theory, or the religion-first theory pass the logical and empirical tests.So we need a better alternative.


There is a bit of a missing piece in the hypothesis, and that’s why the adoption of agriculture led to higher group size/population density. This might seem like a trivial query, and certainly the evidence seems to support the fact that it did, but a careful reading of the hypothesis indicates that there is more going on here.

To boil down the story to simple postulates, it states that the adoption of agriculture had two effects. One, it led to higher population size/density. Two, it resulted in worse health and lower nutrition. But these two phenomenon are actually linked together. Health, and nutrition are both tied to resources per person (and the mix of these – but that’s essentially a measurement/index issue) and so is population growth. So the question becomes how did population growth increase but health/resource per capita deteriorate?

In a simple Malthusian framework, the only way this can happen is if there is an increase in exogenous (i.e. non resource related) fertility. And that would be the missing piece here; if the story is correct and the stylized facts taken at face value then agricultural societies for some reason favored higher fertility rates than non-agricultural ones. This is also a testable prediction. At the same time it opens up some puzzles of its own.


I think the main effect is due to lowering the trophic level at which people feed. People who depend substantially on animals for calories are going to be fairly scarce on the ground. If you switch to a mainly plant based diet you can increase the human population’s caloric intake substantially. The ecologists’ rule of thumb is that 90% of available energy is dissipated as it passes from one trophic level to the next. People, however are not very efficient plant eaters. Our gut is too short to efficiently extract calories from plant matter, among other problems. Farmers grow plants with calorie dense parts like tubers and seeds. Then we grind and cook them so that even our short guts can get most of the nutrition out. We also evolve terracing, irrigation, fertilization and other techniques to up the productivity of our agroecosystems.


But which effect, the increase in population, or the decline in health/nutrition/height? You need an explanation which fits both (or a different interpretation of the data).


Both. Plants provide a lot more calories per unit land, hence human population densities can increase. At the same time, they present all sorts of nutritional challenges for a short-gut omnivore not faced by people on a high meat diet. Low and poor quality protein. Toxins. Vitamin and mineral deficiencies. Starchy diets promote tooth wear (from grindstone grit) and tooth decay.

Also, sedentary life is easier than the constant mobility of hunters. Active hunters are close to being pro athletes and gatherers have to carry heavy burdens long distances. Even old folks have to be up for frequent long marches. If you are a mobile hunter-gatherer, you are either pretty healthy or dead. Farmers can linger for years in a cozy hut before dying. Semi-sedentary gatherer-hunters are in between.

The last time I had a look at the primary literature, it wasn’t clear that all early farmers had poor health. Some plant resources are better than others and some vegetarian diets are not at all bad.The corn-beans-squash diet of Amerindians is pretty good. Amazonian horticulturalists consume a lot of fish. Many populations had access to fish and game in reasonable quantities. As usual, sweeping generalizations about a very diverse and ever-changing set of agrarian economies are bound to oversimplify.

Peter Turchin

Agreed. A few comments.

“Farmers can linger for years in a cozy hut before dying.” – so, a better system for retaining and transmitting cultural knowledge?

“Some plant resources are better than others” – precisely, see my posts on breadfruit and the traditional Polynesian diet.


Ah. You are working in what economists call the “2×2 model”. Two goods (“offspring” and “health”) and two factors of production (proteins and calories). The tricky part is that both of the goods should be increasing in both factors. I had to think about it a bit, write down a few equations and “translate it” into economics, but yes. Even if both health and offspring increase with both calories and proteins, but the “production” of health is relatively intensive in proteins while the “production” of offspring is relatively intensive in calories (which is the plausible case), then you do get the effect of lower health and higher population.

Here’s a little graph I made of the effect of adopting agriculture, drawn in terms of “iso-lines” (along which either population growth or health) are constant. The fact that health is more intensive in proteins is reflected by the fact that the iso-health line has flatter slope than the iso-population growth line:

(hope that works – advance apologies if it doesn’t)

I drew it so that adoption of agriculture only increases calories per person without immediately affecting proteins. This results in higher health on impact but lower health once population has adjusted. Of course one can draw it so that calories increase but proteins decrease with adoption in which case you may get worse health right away.

One could actually tell a simple story where people like to have better health and more children (an iso-preference line, which balances these two, could be added and it would have a slope in between the iso-health and iso-pop growth line). If you (and your children) start out really healthy to begin with you’d be willing to trade off some of that health for a greater quantity of (less-healthy) children. That would be sort of an ironic twist on the economic “child quantity-quality” model that is often used to explain why after the Industrial Revolution people reduced their fertility (although that story is usually in terms of quantity and education per child)


All that sounds reasonable. You could also add (1) that the curve of productivity of farmers is probably higher that that of hunter gatherers when plotted against health status, (2) that higher population densities of farmers expose them to more infectious diseases that manifest skeletal abnormalities (TB for example), and (3) farmers were probably sicker when they died than hunter-gatherers, biasing the skeletal sample relative to the averaging living person. (2) is related to one of the reasons that farmers will tend to out-compete hunter-gatherers. In contact with farmers, they will tend to contract their diseases to which they are not adapted. This was devastating to American natives in contact with Europeans and Africans. (3) is related to another reason. It is said that San females tended to marry Herero herders because life was easier in their camps. At least in one well studied population in Southern Oregon, virtually all contemporary survivors of the native population were the offspring of Indian women married to White men because of (2) and (3).


Pete a better reason for ‘San’ women to reproduce with Herero is that mortality of their offspring is cut in half. Their children get milk and other foods, have greater body weight, etc. Don’t know about life being easier…..

BTW I wish ‘San’ would go away: in the Central Kalahari it is the equivalent of our famous ‘N-word’.


Henry, Ju/hoansi, sorry about the San usage. Your point about the health effects on children of joining a Herero camp undermines the over-generalization that that hunter-gatherers have better nutrition than farmers–OK pastoralists in this case.


Right. Sahlins made much of their leisure, sleeping all the time. People also sleep all the time during famines when they are starving. I have nothing positive to say about being a hunter-gatherer except that the freedom is great when there is plenty of food.

Peter Turchin

Radek, I see what you are doing, and it makes sense. However, note that instead of protein vs. calories phase space you should be using paleolithic foods (healthy, limited in productivity per unit of area) vs. neolithic foods (unhealthy, but very productive per area). Neolithic foods such as cereals and legumes supply adequate amounts of proteins; the main problem is the additional stuff – anti-nutrients and toxins. Furthermore, different kinds of neolithic foods have very different effects on health. In particular, ‘Mediterranean’ diet is the worst (funny as it may sound), Polynesian is the best, and Southern Chinese (rice-based) in the middle. More on this in my blogs on this topic (use the search term ‘diet’ to locate them).


Easy life due to agriculture is not the only possible explanation for native women marrying colonialist men. There is a huge difference in social status. Also, note that mass invasion by agriculturalists destroys the game populations and makes hunting/gathering way of life difficult or impossible.

What would be the correct generic name for non-Negroid peoples of Southern Africa? “Bushmen” is also derogatory, as far as I know.


You are absolutely right (in my parochial experience) that cattle ruin the ecology of foraging: many large mammals are displaced and the hoof traffic destroys many of the root foods. Not so sure about “status”: it must depend on the groups. Among northern Bushmen the offspring are not proper members of the dominant invading group but are looked down on as half-breeds. But they do have body fat because they get food.

“Bushman” sounds not quite right to us English speakers but the origin is a Dutch word that means something like bandit. No one has a problem with it except Peace Corps volunteers and the like. There was an organization several decades ago called “First People of Southern Africa” that said the correct term for themselves was Bushmen.

There are (in my limited experience) three linguistic groups of Bushmen. The northern folk are famous in the literature in Anthropology as the Dobe !Kung. Their word for themselves is simply “humans”, “Zu”, where the Z is pronounced like the j in the French “Je”. The u has a high tone. No one seems to know where the word “!Kung” comes from but I have heard people way north, on the Angolan border, refer to themselves with a word something like that.

The central group is the most numerous and their names mostly have the “-kwe” suffix. These include the Nama Hottentots of Namibia, they extend across Botswana into Zimbabwe and more perhaps. This “-kwe” is the same as “Khoi” of Khoisan. These are the people who would not wish to be called “San”.

There is a third, southern group, extending into South Africa. I have never met any of them and know nothing at all about them.

I do apologise for sounding like a PC speech policeman. On the other hand there is no good reason to insult people when it can be avoided.

Peter Turchin

Henry, thanks for laying this out. To summarize: avoid ‘San’, ‘!Kung’ is of unknown provenance, and ‘Bushmen’ seems like the best neutral term. I completely agree that there is no need to insult people when it can be avoided.

Peter Turchin

In the Malthusian situation health deteriorates because individuals don’t get enough resources per capita, for example, energy. In the scenario I described negative health effects struck immediately, before the populations that adopted agriculture reached the Malthusian ceiling. This happened because of antinutirents and other poisons in grass seeds to which humans were not adapted. So the problem was not lack of resources (e.g., energy), but their nature. In terms of abundance of energy, switching to agriculture allows an order of magnitude increase in the carrying capacity (productiveity per unit of land). The general ecological rule is that you increase carrying capacity and population will approach it by whatever mechanism (lower death rates, higher birth rates).

In the scenario that I propose, increased carrying capacity has several important effect on the ability to wage war successfully. First, and most obviously, it allows to raise more warriors in the same territory. Secondly, and especially initially, it allows to concentrate the population – you only need to cultivate fields in the immediate vicinity of the village, rather than spreading out over large foraging territory. Population concentration allows faster response to an outside threat. Spread-out foragers are more easily picked off, one by one.

Henry Harpending

Aggregation into larger and larger polities is a logical outcome of increasing local raiding and warfare but it depends on topography does it not? In order to organize and control a polity I, as emperor or priest, need to be able to project power and control over a large area. I have always thought that the appearance of ‘civilizations’ in the fertile crescent and the Nile valley had to do with the openness of the terrain and the ease of movement across it.

Another possible outcome seems to be fragmentation into more or less endogamous warring clans as in Highland PNG or Afghanistan.

A caveat about this site is that nearby sites (with 50 or 100 km) that are roughly contemporaneous do have early domesticates present. It is not so clear, to me at least, what “pre-agricultural” means and implies here.

Peter Turchin

The effect of topography is very important. We have a paper, currently in review, that shows that you need to include the effect of topgraphy (which we proxy by elevation) to capture the rise and spread of large-scale societies in ancient and medieval Afroeurasia.

On the caveat: apparently, no domesticated plants or animals were consumed at G-Tepe, and since there was a lot of feasting, the arch. record is pretty good. But the past is changeable… at least our understanding of it.

James Waddington

A mea culpa. I am probably not unique in using this blog as a launch point for the bees in my own particular bonnet; but I do do it, while neglecting such non-academic trivia as that yesterday morning I was recounting Peter Turchin’s Why Become a Farmer? to a sixteen year old grandchild when my daughter arrived downstairs and said to him, lightly but with conviction, “I’ve come to save you from your grandfather”. To which he replied, “No, it’s incredibly interesting”. It may be significant that he plans to do both maths and history at A level (university entrance exam). What I’m trying to illustrate is what is obvious from the depth and width of the discussion, that this is a valuable blog, the ramifications of which go way beyond what appears in the comments. (Oh, and Peter, when he’s finished his present exams, I’ll give him How to Become a Cliodynamicist to read. Beyond it’s immediate focus, it’s an great model for how to choose what degree to do.)

Peter Turchin

James, thanks for this! The ultimate test of a scientific theory is being able to explain it in such a way that our children and grandchildren understand it and are fascinated by it. It is heartening to see that it we are reaching the 16-year olds (surely, your skill in translating it to non-tech language played a role). The golden standard is to do it in terms understandable to a 4-year old.

James Waddington

Right, I’ll have to wait for great grandchildren (even now legally and biologically possible, though one hopes in practice not yet). The other thing, selection at the level of the group, still leaves me puzzled.
Yes, I read about the Strüngmann conference. Humph! Nicaea, crossed my mind In fact it was that that started me thinking that there is a tendency among “developmental psychologist, primatologists, anthropologists” and others to find it much more comfortable to take certain evolved forms of culture, such as religion or agriculture, as givens, because then they fit more easily into (very productive and fascinating) hypotheses; whereas I would submit that, contrary to what you say above, “How a cultural innovation arises, and why it spreads are two separate questions”; how cultural innovation arises and how it spreads are as interdependent as how species evolution arises and how species increase or decrease. Sure, you can look at the success of a social species, such as African Hunting Dogs, and see why one group may be more successful than the other, but this does not suggest that that group is itself evolving, or part of the evolutionary process. All it suggests is that Hunting Dog proto-culture has evolved to lead them to live in a proximity with each other that in turn leads them to behave in a particular way. This “group” will then become an aspect of the Hunting Dog phenotype and, as in the case of humans, the phenotype is a significant component of the environment in which evolution, at the level of the individual dog, is taking place.
But this does not suggest that evolution is taking place at the level of the group, as if the group is some sort of envelope analogous to the organisms of which of which it is composed. I don’t know that anything has happened to displace the view popularised by Dawkins that variation, reiteration and selection can only take place at the smallest locus of irreducible difference (which can be quite dispersed, certainly more so than at the naked “gene”.) There is a lacuna here, and I’m not saying the more Darwinian view is correct, but the lacuna, I feel has to be filled.
For a suggestion as to how group behaviour evolves (at the level of the individual) I can only recommend Brother Jero’s no doubt unhinged attack on the very notion of group selection.

Peter Turchin

I am back from my field trip to Tampa (lots of material for future blogs – ‘Cooperation from Bottom Up’, but first to finish with this theme). Great discussion here. I think we have a SEF Special Feature in the making – the topic is very important, and the old paradigm is groaning under the weight of accumulated anomalies. If anybody has ideas about how to organize a discussion and who to invite to participate, please comment here, or send me an e-mail.

Also any suggestions about references proposing explicit alternatives to the bottom-up theory would be greatly appreciated. Here’s one that Jason Collins suggested (via Twitter):


Henry Harpending

Peter what is the “bottom up” theory? Not something, I trust, from English boarding schools.

If the first farmers were partly or mostly pastoralists then foragers likely treated them as a new and valuable resource since, barring disastrous herd losses, there is a great surplus of food. There may have only rarely been actual territorial conflict and violence.

Peter Turchin

See my preceding blog:


Starting with “The standard archaeological model (which is so standard that it is rarely formulated in explicit terms) explains it this way”


“Second, hunter-gatherer societies share the food; hoarding marks you as an anti-social deviant. What this means is that you could put all that work into growing plants (clearing the field, planting, weeding), but others would think nothing of harvesting the plants when they ripen. Or you could get to the point of harvesting and storing the crop, but then everybody else in your community would expect you to share it.”

I’ve heard this often, but I wonder. Why is it that it’s natural/customary among foragers to share a gazelle that a hunter may have spent days getting, and not share the corn from a field that another person in the tribe planted? I mean, if the ethos of your tribe is sharing, then the person who is a gifted hunter shares the gazelles, and the person who is gifted with plant tending shares the corn, no?

Peter Richerson


The usual explanation is that many hunter-gather groups are essentially little insurance pools. Medium and large size game are very risky for an individual man to hunt. Even good hunters are successful only every few weeks. Then, if they kill a large animal, there is more meat than can easily be eaten by one family. The answer is to assemble camps with 5-10 good hunters. This team can insure a comparatively steady flow of meat to each family if they share. The gathered resources that women have less variation in return and are not generally widely shared. Small game caught by men, women and kids is also so widely shared.


Peter, yes, thank you for reminding of that explanation. What is gathered is usually not shared. But consider the scenario where a tribal person decides to intensify. Let us consider a tribal person who’s seen neighbors growing field corn, and decides to give it a try. Now growing anything is also a risky proposition, what with weather and pests and so on. At harvest time, suddenly there is glut. Why not insure for the considerable risks of growing stuff by sharing it?

There are stories of the Montagnais throwing a big feast even in fall or midwinter and eating through their stores of food (some of it gifted). They shared whatever they had, meat or otherwise… why would the ethos change just because one person grows something? And it doesn’t… the anthropologists recount such occurrences where the field gets “pilfered” and say, well, that is totally demotivating, people would only cultivate if they get to keep it all, but if the whole tribe just helps themselves, that’s failure for the incipient farmer. I am asking, why not consider it success instead? Bill feeds the tribe with gazelles, Bob feeds the tribe with a field of corn.

I am not sure if I am getting my meaning across right.


In other words, if it’s not demotivating to share a gazelle, why is it demotivating to share a harvest?

Henry Harpending

The gazelle, or even the giraffe, is only good for a few days. The harvest has to last the farmer for months and he has to save seed else no harvest the next year. Moochers are relatively cost free at the gazelle party, devastating at the corn feast.


The storage qualities of grain also make it handy to sell or trade.


It also makes it easy to steal. These properties are not shared with root crops, which are hard to steal and hard to sell or trade. The kinds of social elaborations we associate with ‘agriculture’ seem mostly to accompany cereal agriculture while root crops are more likely to leave a region stuck at the level of swiddening, or such is my impression.


Yes, root crops contain much more water than grain so they are heavy and bulky per unit food value. They are correspondingly unattractive to trade or loot. In Highland New Guinea the medium of trade was/is pigs fed on sweet potatoes. The complex societies in the Altiplano around Lake Titicaca were highly dependent on potatoes though they had access to quinoa locally and maize traded up from lower elevations. They also made some varieties of potatoes into a dried product, chuno. Root crop growers can work around the limitations of their dominant crop, but they often seem not to.


Thank you for all your responses. I am still not quite sure what I am getting at myself, let me try again. I want to avoid, just for the sake of the argument, the “grain can be stored and sold” part. I am trying to point out that both hunters and farmers need insurance because both are highly uncertain activities dependent on luck and other factors out of the person’s control. That’s why the Chacoans spread the risks over a large area. That’s why we have crop insurance today.

So we have this tribe of predominantly foragers, and one of them decides to grow a crop. The harvest need not last the farmer for months. He is still part of the sharing economy of the foragers around him. Yes, he has to save the planting seed, but that small quantity can be safely hidden. So looking at it from this angle, why would moochers be any more costly during the harvest feast than during the gazelle feasts during the rest of the year?


I don’t understand why you don’t understand. I suppose the main assumption is that people tend not to want to share what they produce with their hard work if it doesn’t spoil over time. That is true of grains. That is not true of gazelle meat.

In any case, agriculture likely didn’t start out with one guy deciding to plant. As noted above, hunter-gathers may know the basics of agriculture, but tend not to want to give it up unless they are forced to. Such as war, famine, and desperate people willing to follow some messiah (or authoritarian) figure who says that the way to salvation is to start being farmers.

Peter Turchin

Gazelle meat can be dried or salted or smoked to preserve.


Exactly. A variety of foods can be preserved.

As for one guy deciding to plant — I was referring to the tribal group where actually one guy did decide to plant corn. And it got “pilfered” (shared) just before harvest. Anthropologists give this as an example of why farming cannot take off among the forager values. But I am saying, in a sharing economy, why not farm and share it? Hunting is hard work, and uncertain, and so is cultivation.


Except, I don’t think that is how agriculture happened in history. I’d like to know what the evidence shows, but while hunter-gatherers knew some basic planting skills, full-scale agriculture probably happened through coercion. Like Sparta and its slaves. Some bright tribal leader got the idea that instead of slaughtering the neighboring tribe after defeating and capturing them, he’d make them in to unarmed slaving peasants instead.

Peter Turchin

Gazelle meat can be dried or salted or smoked to preserve.

Peter Turchin

Pete and Henry, I don’t see that grains and root crops are so different. When raiders strike, they don’t take grain or potatoes, but take more valuable things. Which is usually livestock and humans (as slaves). The additional advantage is that this kind of loot is self-propelling.

Henry Harpending

Peter you are right, maybe, but it would be better if we knew more about bronze-age warfare. If my folks were really hungry I might take the wheat and leave the girls.

Another consideration is that it is easier to destroy grain. A few years ago I got mildly chewed out by the caretaker at Ain Dara in northern Syria. As he showed me a large basket of burnt and charred wheat he complained that “you Romans”, i.e.me, burned all their wheat. It is a Kurdish region, for all I know he might have been talking about Crusaders. They sure do nourish grudges in that part of the world.



My understanding is that pastoralists were/are typically quite dependent on grain grown by settled farmers. The trade in crafts was also important. They sometimes controlled irrigable land farmed by dependent farmers, sometimes traded with independent farming societies, and sometimes acquired grain in raids. Ibn Khaldun remarks on the dependence of nomads on the settled agriculturaists for basic necessities and Khazanov makes a big point of it in his Nomads and the Outside World. The Plains horse hunters both raided and traded for supplies of grain and manufactured items.

I visited the Atlas mountains a few years ago. The river flowing out of the Atlas toward the Sahara forms a long linear oasis. The guide said that it formed part of one of the trans Sahara caravan routes and that the camel nomads using the route normally traded for grain, dates, fodder, and other supplies as they passed through. The traditional towns were compact adobe fortresses presumably build to encourage the nomads to trade rather than raid.

Peter Turchin

Right, see my reply to Henry.

BTW, see you in Bavaria in mid-September?

Peter Turchin

Hi, Henry. It’s true that we don’t know as much about bronze-age warfare, so I am basing my comments on more recent history. My models are Crimean Tatars and the Comanches. Sure they would grab any grain that was lying around, but mostly they cared about slaves and livestock. Then they sold it to the middlemen. The Tatars at Kaffa, the Comanches at Santa Fe. After that they could purchase all the grain they needed, as well as other products of agrarian civilization they wanted (fabrics, some luxury/prestige goods, etc).

If you are smart about it, you don’t wait to go raiding until you are weak with hunger. You do it strategically and you get high value, easy to move loot, which you can later exchange for bulky things like grain.


Hope, you don’t mind that I’m digging up (pun indended) a two year-old post, but I found your line of thought rather interesting (a shame, I only found your posts yet, but thoroughly read them with great interest) and would like to add a few remarks, if you don’t mind.

Why people of the Neolithic shifted to this peculiar and radically new way of living is indeed the crucial question and it’s its chronological (as much as geographical) position which makes Göbekli Tepe such an interesting site to ponder the social implications and (probable) economic necessities behind this incisive step. It’s the mode of construction of these monumental enclosures which could offer an important clue here. The necessary workforce clearly exceeds the number of individuals of a single hunter-gatherer band, thus we tend to see these building projects as collaborative events of several of such groups – already gathering at special places for certain occasions (exchange, communication, etc. pp.). A useful instrument to attract workforce for such projects is known from more recent ethnographic contexts with the prospect of a lavish feast. The remarkable fact here is that organisation and realization of feasts and work alike surely would have asked for some coordinating instance, hinting at a social complexity not suspected among these early Neolithic groups as of yet. The constant supply of food to fuel these feasts is the crucial basis of this scenario and may have caused some stress to an economy based on hunting and gathering. Exploration for and experimentation with new ressources therefor might have been an acceptable consequence of the ever-growing demand.

A paper in which we discussed this scenario some time ago should be available online here: http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=9530672&fulltextType=RA&fileId=S0003598X00047840

  1. Home
  2. /
  3. Cliodynamica
  4. /
  5. Regular Posts
  6. /
  7. Why Become a Farmer?

© Peter Turchin 2023 All rights reserved

Privacy Policy