Friday, November 23, 2007

War is not an exact science
[This is my latest muse column for]

General theories of why we go to war are interesting. But they'll never tell the whole story.

Why are we always fighting wars? That’s the kind of question expected from naïve peaceniks, to which historians will wearily reply “Well, it’s complicated.”

But according to a new paper by an international, interdisciplinary team, it isn’t that complicated. Their answer is: climate change. David Zhang of the University of Hong Kong and his colleagues show that, in a variety of geographical regions – Europe, China and the arid zones of the Northern Hemisphere – the frequency of war has fluctuated in step with major shifts in climate, particularly the Little Ice Age from the mid-fifteenth until the mid-nineteenth century [1].

Cold spells like this, they say, significantly reduced agricultural production, and as a result food prices soared, food became scarce – and nations went to war, whether to seize more land or as a result of famine-induced mass migration.

On the one hand, this claim might seem unexceptional, even trivial: food shortages heighten social tensions. On the other hand, it is outrageous: wars, it says, have little to do with ideology, political ambition or sheer greed, but are driven primarily by the weather.

Take, for example, the seventeenth century, when Europe was torn apart by strife. The Thirty Years War alone, between 1618 and 1648, killed around a third of the population in the German states. Look at the history books and you’ll find this to be either a religious conflict resulting from the Reformation of Martin Luther and Jean Calvin, or a political power struggle between the Habsburg dynasty and their rivals. Well, forget all that, Zhang and his colleagues seem to be saying: it’s all because we were suffering the frigid depths of the Little Ice Age.

I expect historians to respond to this sort of thing with lofty disdain. You can see their point. The analysis stops at 1900, and so says nothing about the two most lethal wars in history – which, as the researchers imply, took place in an age when economic, technological and institutional changes had reduced the impact of agricultural production on world affairs. Can you really claim to have anything like a ‘theory of war’ if it neglects the global conflicts of the twentieth century?

And historians will rightly say that grand synoptic theories of history are of little use to them. Clearly, not all wars are about food. Similarly, not all food shortages lead to war. There is, in historical terms, an equally compelling case to be made that famine leads to social unrest and potential civil war, not to the conflict of nation states. But more generally, the point of history (say most historians) is to explain why particular events happened, not why generic social forces sometimes lead to generic consequences. There is a warranted scepticism of the kind of thinking that draws casual parallels between, say, Napoleon’s imperialism and current US foreign policy.

Yet some of this resistance to grand historical theorizing may be merely a backlash. In particular, it stands in opposition to the Marxist position popular among historians around the middle of the last century, and which has now fallen out of fashion. And the Marxist vision of a ‘scientific’ socio-political theory was itself a product of nineteenth century mechanistic positivism, as prevalent among conservatives like Leo Tolstoy and liberals like John Stuart Mill as it was in the revolutionary socialism of Marx and Engels. It was Tolstoy who, in War and Peace, invoked Newtonian imagery in asking “What is the force that moves nations?”

Much of this can be traced to the famous proposal of Thomas Robert Malthus, outlined in his Essay on the Principles of Population (1826), that population growth cannot continue for ever on an exponential rise because it eventually falls foul of the necessarily slower rise in means of production – basically, the food runs out. That gloomy vision was also an inspiration to Charles Darwin, who saw that in the wild this competition for limited resources must lead to natural selection.

Zhang and colleagues state explicitly that their findings provide a partial vindication of Malthus. They point out that Malthus did not fully account for the economic pressures and sheer ingenuity that could boost agricultural production when population growth demanded it, but they say that such improvements have their limits, which were exceeded when climate cooling lowered crop yields in Europe and China.

For all their apparently impressive correlation indices, however, it is probably fair to say that responses to Zhang et al.’s thesis will be a matter of taste. In the end, an awful lot seems to hinge on the coincidence of minimal agricultural production (and maximum in food prices), low average temperatures, and a peak in the number of wars (and fatalities) during the early to mid-seventeenth century in both Europe and China. The rest of the curves are suggestive, but don’t obviously create a compelling historical narrative. At best, they provoke a challenge: if one cannot now show a clear link between climate/agriculture and, say, the Napoleonic wars from the available historical records themselves, historians might be forgiven for questioning the value of this kind of statistical analysis.

Yet what if the study helps us to understand, even a little bit, what causes war? That itself is an age-old question – Zhang and colleagues identify it, for example, in Thucydides’ History of the Peloponnesian Wars in the 5th century BC. Neither are they by any means the first in modern times to look for an overarching theory of war. The issue motivated the physicist Lewis Fry Richardson between about 1920 and 1950 to plot size against frequency for many recent wars (including the two world wars), and thereby to identify the kind of power-law scaling that has led to the notion that wars are like landslides, where small disturbances can trigger events of any scale [2-4]. Other studies have focused on the cyclic nature of war and peace, as for example in ecologist Peter Turchin’s so-called cliodynamics, which attempts to develop a theory of the expansion and collapse of empires [5,6].

Perhaps most prominent in this arena is an international project called the Correlates of War, which has since 1963 been attempting to understand and quantify the factors that create (and mitigate) international conflict and thus to further the “scientific knowledge about war”. Its data sets have been used, for example, in quantitative studies of how warring nations form alliances [7], and they argue rather forcefully against any notion of collapsing the causative factors onto a single axis such as climate.

What, finally, do Zhang and colleagues have to tell us about future conflict in an anthropogenically warmed world? At face value, the study might seem to say little about that, given that it correlates war with cooling events. There is some reason to think that strong warming could be as detrimental to agriculture as strong cooling, but it’s not clear exactly how that would play out, especially in the face of both a more vigorous hydrological cycle and the possibility of more regional droughts. We already know that water availability will become a serious issue for agricultural production, but also that there’s a lot that can still be done to ameliorate that, for instance by improvements in irrigation efficiency.

We’d be wise to greet the provocative conclusions of Zhang et al. with neither naïve acceptance nor cynical dismissal. They do not amount to a theory of history, or of war, and it seems most unlikely that any such things exist. But their paper is at least a warning against a kind of fatalistic solipsism which assumes that all human conflicts are purely the result of human failings.


1. Zhang, D. D. et al. Proc. Natl Acad. Sci. USA doi/10.1073/pnas.0703073104
2. Richardson, L. F. Statistics of Deadly Quarrels, eds Q. Wright and C. C. Lienau (Boxwood Press, Pittsburgh, 1960).
3. Nicholson, M. Brit. J. Polit. Sci. 29, 541-563 (1999).
4. Buchanan, M. Ubiquity (Phoenix, London, 2001).
5. Turchin, P. Historical Dynamics (Princeton University Press, 2003).
6. Turchin, P. War and Peace and War (Pi Press, 2005).
7. Axelrod R. & D. S. Bennett, Brit. J. Polit. Sci. 23, 211-233 (1993).

1 comment:

JimmyGiro said...

Gosh, and I thought I was the only solipsist.

I guess if they had historical data for local food availability to hand, then their climate analysis would have been redundant; or maybe unfunded due to lack of socio-political 'newsworthiness'.

As for the lack of 20th century analysis; could it not be argued that by modern times, communication was advanced enough to spread the rumours of food shortage via all political and religious rhetoric in a way not available before; hence the 'notion' of food shortage (by ones enemies) could correlate as much as the actual food supply?