The Rise and Demise of Complex Societies. The Effect of Climate. Science



Join 36.7K other subscribers

Having your research written about in Science or Nature, or getting your article published in one of these journals (and PNAS) is one of the highlights of a scientist’s career. Competition is pretty fierce, and the editors of these prestigious journals tend to behave in a pretty high-handed manner. Although we complain about it among ourselves, we play the game, because potential rewards are worth it. Much of it is pure status competition, but publishing a paper in Science or Nature practically guarantees that it will have high impact, and not just on the scientific community.

I don’t have hard numbers, but my impression is that if you want to publish a paper on the evolution of complex societies in Science, your best bet is to first get trained as a climatologist. Somehow the combination of civilizational rise or collapse (a deservedly newsworthy topic) and ‘hard science,’ represented by climatology, proves irresistible to Science editors.

One recent example is a News Focus article “Roots of Empire: A climate history project in Mongolia is charting the unexpected conditions that may have propelled the rise of Genghis Khan” which was published in the September 28 issue of Science. Among other things the article claims that “the relationship between climate and the decline or collapse of civilizations is well established.” But its main focus is the claim that the rise of the Mongolian Empire under Chinggis (a more accurate spelling of his name) was due to a period of unusually good climate between 1211 and 1230.

The Mongol army (source)

Although I am very much in favor of reconstructing past climate and analyzing its effect on historical dynamics (more on this in the next post), I am very skeptical of current claims that purportedly establish a direct relationship between climate and the rise or collapse of civilizations. It is quite likely that climate change is one of the secondary factors affecting history, but a prime mover?

Despite its trappings of ‘hard science’ current research on the influences of climate on history lavishes all scientific rigor on the first part of the job, reconstructing past climates, and then turns ‘soft’ when it gets to the climate impact on historical societies.

As an example, there was another recent paper in Science (4 February 2011, this one a research article), “2500 Years of European Climate Variability and Human Susceptibility.” Here’s what the authors did when they got to the effect of climate on history:

Apparently the only approach used by the authors was to eyeball the graph and connect it to various historical events (I did not check the Supplementary Materials, perhaps they present a statistical analysis there?). You give me any series of random numbers, and I can guarantee you that I will come up with a neat story about how it explains history.

This reminds me of an anecdote, apparently told about some famous physicist. As he was talking to a colleague, a student rushed up to him with a graph depicting the latest experimental results. The famous physicist looked at the graph and said:

“Of course, it makes perfect sense. This is precisely what the theory predicts.”And then proceeded to explain how.

“Professor, you are holding the graph upside down…”

“Oh…[turning it right side up], but of course, it makes perfect sense – this is precisely what the theory predicts!”

And this is how much of current research on climate and history looks to me.

Returning to the case of the Mongol conquest, sudden eruptions of steppe nomads have often been explained by poor rainfall in the steppe. Faced with the prospect of starvation on the steppe, the nomads instead choose to attack the nearby farming societies. As the Science article mentions, Arnold Toynbee was one of the proponents of this theory.

It is very likely that this mechanism was at the root of the intractable conflict in Darfur. As the climate got drier during the last two decades the pastoralists, who had raised livestock in the plains of Darfur, started migrating towards the hilly region, which always got more rain and was where agricultural communities lived. This resulted in conflict between farmers and nomads. Nomads then created an alliance against the farmers (this is the origin of the infamous Janjaweed) and raided villages, while farmers created their own defensive alliance and allied with the SPLA. The tragic results (but not the causes of the conflict) have been amply covered by the media.

Once again returning to Mongolia, we learn, however, that lack of rain was not a problem during the Mongol eruption under Chinggis Khan in early thirteenth century. In fact, the opposite was true. Apparently, abundant rainfall was good because it increased the productive base of the nomad society and made their military conquests more possible.

So far so good, but when the period of abundant rainfall ended around 1258, the Mongolian empire did not collapse; in fact it continued to expand. As one of the researchers quoted in the article says, the Mongols did not collapse, “they restructured.”

OK, I believe I made my point. But to end on a positive note: I think that what the climatologists cited in the article do is great. I can’t wait until they publish the data they are currently collecting. And then we will need to do a much better job determining how climate affects human history. I’ll discuss this question in the next blog.

Notify of
Most Voted
Newest Oldest
Inline Feedbacks
View all comments
Gene Anderson

The big problem for a nomad is raising all those horses. You can’t raise a lot if there is a drought. You can’t if there is cold weather in Mongolia–it causes lots of dzuds (blizzards that ice up the grass to the point where the horses can’t paw the ice and snow off, so they all starve). The height of the Medieval Warm Period, a very well-known and well-demonstrated worldwide phenomenon, eliminated the dzud risk, allowed the Mongols to ride out in force, and thus allowed Genghis Khan to do his thing. Without good climate he’d have stayed home and never conquered anyone beyond his family. So, climate facilitated. But it could not produce a Genghis, or a Mongol society of segmentary lineages and horse raiding. You had to have all three of those things together.

Peter Turchin

But what does ‘facilitate’ mean? There were many other steppe empires before the Mongols – the Hunnu, Rouran, Turks – to name just a few. Was there abundance of rain during those ‘eruptions’? Also, horses are fairly resistant to dzud, because they can use their hoofs to break through the ice. And as John suggests below, they are more resistant to drought. So if you are a nomad and you have lost most of your herds, but you still have horses ‘in reasonable working conditions’, you’d better hit those farmers before your horses starve.


The Mongols could take only a limited number of horses with them when they left their homeland, and it was certainly not the shortage of horses that had prevented them from moving out earlier. The only plausible mechanism through which the more favorable climate could cause the Mongol expansion would be human population growth – but there wasn’t enough time for that.

Also, it’s worth mentioning that Mongolian steppes get plenty of summer rainfall even in normal years, and that winter cattle die-offs are relatively uncommon – major ones happen about once per decade. When they do happen, they kill mostly sheep, cows and to lesser extent camels, not horses. Even more importantly, warmer climate in Mongolia would mean a drought and more frequent dzuds (they result from freezes following winter thaws) rather than favorable conditions.

John Lilburne

Climate change generally in itself does not cause change but it precipitates it.
Take roman history there was a cooling around 0AD that probably precipataed the change from republic to empire and a cooling around 400AD that caused the fall of the western part of the empire. However the real cause was the internal rot that had already happened.
There are probably some events that are causative inthemselves. The dramatic short cooling around 513AD together with the plague that happened (the 2 are often associated) during Justinians reign was probably enough to completely derail the revival of the Byzantine empire.

It is also an illwind that brings nobody any good. Like el nino certain limited areas can benefit whilst the rest of the world suffers. Some have suggested that horses are relatively less drought susceptible compared to cattle and so drought can hit settled city states more harshley than their nomadic competitors

Peter Turchin

I generally agree, and that is what I meant by climate playing a secondary role. My hypothesis is that its effect is contingent on the structural-demographic phase in which a society finds itself (now talking about agrarian societies). If it is in the middle of an integrative phase, then poor harvests will cause temporary difficulties, perhaps a population decline, but then things will get back on track. But if there is a shock arriving when the society is already on the brink, then it can trigger a catastrophic collapse – but the main causes will be social, not climate. A good example is the Black Death – by early 14th c. western Europe was already on the brink, and a catastrophe ensued. Russia, on the other hand, had already experienced a catastrophe a hundred years ago, so the Black Death had a fairly minor effect, less than the plague epidemic of the fifteenth century.

Alexander Sadykov

You are right “climatic explanation for everything” is a current and irresistible trend in “big names” science, perhaps by serious political reasons. I am also skeptical on that extensive use of modern statistics give raises a new inside. Actually, quite opposite is more likely, “Give me a big enough data set and I will proof anything what you want!” it is informal motto of Bayesian statisticians today. So, the raise of science-like subjectivity is unavoidable, simply because a recheck of such “sophisticated statistical results” requires enormous amount of time and computer power. The further development of science makes us more and more dependent on the credibility of quite a few persons or collective irresponsibility of large groups. Ironically, all of these trends make modern science is very similar to its historical opponent i.e. religion. The latter, to a large extent explain why we see a clear trend of increasing number of authors per article, and the numbers of references. Even now, many of the articles in the so-called prestigious scientific journals look like a cemetery of forgotten references and Supplementary information (another big part of these journals is long captions under the beautiful pictures).

Let’s go back to the question. What can counter this trend?
Is it a new theory, a new set of empirical data, the new “less sophisticated” non-Bayesian statistical methods or some combination of all the above? This seems very unlikely, because it does not add a new quality to what is already there. The key problem is that the gap between theory and reality can not be filled only by statistical modeling of empirical data. The only scientific answer to this kind of problem is an experiment.
Here you could argue that the direct historical experiment obviously impossible, and so on again. Yes it is true, but not completely true; the fundamental mechanisms of the theory can be tested directly in microbiological experiments. Of course, there is a difference between the human society and the colony of microbes, but the survival mechanisms of large communities of individuals in the unpredictable environment can be universal. In any case, it can add a new and much-needed “experimental dimension” to the eco-demo-historical study.
It would be interesting to know your opinion on this Peter.

On the issue of the relationship between climate and population expansion; when microbiological colony is in a new environment, it begins to increase after a certain delay. The length of this delay is only slightly dependent on environmental conditions, but is highly dependent on the variation of these conditions (unpredictability of environment). Perhaps it may be useful clue.


Being skeptical about papers published in Nature and Science would be a good start. The quality of their editorial judgement is falling like temperatures in Younger Dryas 🙂

Science and Nature is no longer a scientific journal, now it’s “about science” journals. More like a scientific fashion magazines. These magazines serve a necessary function that is not directly related to science, but to make a political game around research funding a little bit more transparent.
Apparently, the last Younger Dryas period led to the beginning of agriculture, so we’ll see what will bring us a new Younger Dryas era in science 🙂


The author quips “Although I am very much in favor of reconstructing past climate and analyzing its effect on historical dynamics (more on this in the next post), I am very skeptical of current claims that purportedly establish a direct relationship between climate and the rise or collapse of civilizations. It is quite likely that climate change is one of the secondary factors affecting history, but a prime mover?”
~ ~ ~

Ah… and why not?

Sounds like you are projecting an Argument From Ignorance rather than some serious analysis.

Have you actually considered how dependent people are on food and farming and how dependent that is on climate/weather? To say nothing of the other impacts weather conditions have on living conditions?

  1. Home
  2. /
  3. Cliodynamica
  4. /
  5. Regular Posts
  6. /
  7. The Rise and Demise...

© Peter Turchin 2023 All rights reserved

Privacy Policy