I have just concluded a very intense ‘micro-workshop’ (only five participants) that I convened in Storrs over the weekend. We have been brainstorming to develop approaches to estimating crop productivities in historical societies going all the way back to the rise of agriculture. Ultimately, we would like to construct a historical GIS of crop productivities. This is a question of utmost importance in social evolution.

There is no question that agriculture is a necessary condition for the evolution of truly complex societies. Yes, some societies of hunter-gatherers living in very productive environments (for example, the Pacific Northwest) have achieved a remarkable degree of cultural complexity.

Northwest Indians: culturally complex, yet small-scale societies (photograph by the author)

However, these groups numbered in thousands (at most). Large-scale societies, those encompassing millions of individuals and more, can exist only on the basis of intensive agriculture. There are no exceptions.

But if agriculture is a necessary condition, one may ask whether it is also a sufficient one. If food-producing technology becomes sophisticated enough, is the rise of large-scale societies – macrostates and empires – simply a matter of time? Is it in some sense inevitable? This may be called the ‘bottom-up’ hypothesis: the complexity of a society is ultimately determined by the productivity of agriculture.

Most archaeologists hold this view implicitly, although you will not necessarily get them to admit it out loud. This theory is implicit in the works of such luminaries as Lewis Morgan and Leslie White, Elman Service and Marshal Sahlins, and, most recently, Jared Diamond.

Diamond’s view is stated very clearly in his recent review of Why Nations Fail by Daron Acemoglu and James Robinson ((see also response by Acemoglu and Robinson). I haven’t yet read this important book, but will blog about it when I do.

In his review, Jared Diamond states:

The various durations of government around the world are linked to the various durations and productivities of farming that was the prerequisite for the rise of governments. For example, Europe began to acquire highly productive agriculture 9,000 years ago and state government by at least 4,000 years ago, but subequatorial Africa acquired less productive agriculture only between 2,000 and 1,800 years ago and state government even more recently. Those historical differences prove to have huge effects on the modern distribution of wealth.

This is a very strong statement of the hypothesis, and one that can be, and should be tested empirically. And that is why we were brainstorming in my lab last weekend:

(photograph by the author)

Although I agree that agriculture is a necessary condition, within agricultural areas the durations and productivities of farming seem to do a poor job of explaining the rise of large states. Let’s take Europe, the example that Diamond uses. But it wasn’t ‘Europe’ that acquired productive agriculture beginning 9,000 years ago. It was the Mediterranean belt.

As Diamond himself proposed in Guns, Germs, and Steel, cultivars and agriculture in general spread much more easily in east-west directions, where climates and ecological communities are similar. Spreading north or south, across different climatic and ecological zones, is much more difficult.

While agriculture spread West from the Fertile Crescent and reached Spain quite readily, moving North into central and northern Europe was much more difficult. Plant cultivars had to adapt to very different climates and soils and that takes time.

For many millennia the economy of central and northern Europe was based on agropastoralism, with a heavy emphasis on animal husbandry supplemented, probably, by hunting and gathering. In eastern Europe agriculture appeared only during the first millennium CE, a mere 1.5 thousand years ago. Even in northwestern Europe agriculture was still very unproductive in the first millennium. Historical evidence from Carolingian manses (eighth and ninth century CE) points to really appalling yields, two seeds for each planted one. As recently as in the Middle Ages the yields increased only to 3–4:1, while in medieval Russia, which has even worse climate, they still were closer to 2:1. Compare this to yields of around 10:1 for Mediterranean regions in antiquity – not surprising, as wheat and other cereals were pre-adapted to such environments.

Cereals growing wild near Ephesus, Turkey (photograph by the author)

But today the most effective large-scale societies are found not in the Mediterranean Europe. Greece is essentially a basket case, and everybody is waiting with bated breath when Spain will enact its own Greek tragedy. Effective governments are found in areas that only two thousand years ago did not have productive agriculture – Germany and Scandinavia.

Going back in history, if we look where really large-scale societies appeared, we observe Russia and Mongolia – at certain times the largest empires on Earth (in terms of territory controlled). Yet, Ulan Bator and Moscow are the numbers 1 and 2 in the sad list of national capitals ranked from the least productive environments (for growing crops) up.

Now I agree that such ‘point comparisons’ are not the best way to test scientific hypotheses. One can always be accused of selecting cases to prove one’s point. Fine. So we need a systematic dataset of durations and productivities of farming for the whole world, or at least those regions where agriculture is in principle possible.

An appeal to colleagues: if you know of any datasets or techniques that could be useful in this research, we would really appreciate learning about it.