One of the major reasons I enjoy having a blog is the many excellent comments that the readers make on various ideas I float here. Sometimes they make me change my views, sometimes they force me to explain things better, or clarify my own thinking on the subject. All of this happened as a result of exchange following the previous blog (The Rise of the West: Science and Ideology).
I started by arguing that you cannot do science that is solely directed at explaining a unique event, such as the Great Divergence. At the same time, I was unwilling to roundly denounce all literature on this topic as unscientific. There has been a number of books that I found very illuminating (to give three examples, The Great Divergence by Ken Pomeranz, Why Europe? by Jack Goldstone, and Why the West Rules – For Now by Ian Morris; there are more). But thinking and responding to comments made on the previous blog made me realize that you can do it. So here’s what I think now.
All historical events are unique, but a valid explanation of any particular event must involve a mixture of unique and generic features. Let’s use as an example the Cretaceous–Paleogene extinction event that killed off most of the dinosaurs. Currently the best explanation of this mass extinction is the Alvarez impact theory (Schulte et al. 2010). The explanation is based on a unique event: a huge asteroid hitting the Earth ~65.5 million years ago. How have natural scientists built their case?
They start with unique features – the asteroid impact itself, how big it was and where it hit. Also, that there was an Earth and it had a certain kind of biota – it is easily forgotten that all such factors must be included in the account. The next step, however, is to start building models that link the impact to extinction. The unique features provide ‘boundary and initial conditions’ (using the modeling jargon) while general theories allow investigators to build dynamic models for postulated processes. For example, a big issue is what were the environmental consequences of the asteroid impact, and climate simulations suggest that the impact was followed by a long period of global darkness and cooling.
Finally, the models generate predictions that were then tested against the data in the fossil record. The conclusions of Schulte et al. (2010) have not be universally accepted by all scientists (see the scientific correspondence associated with their article). However, nobody can deny that the huge scientific progress took place since the Alvarez hypothesis was proposed in 1980. This is a fine example of what Randall Collins calls rapid-discovery science (although he is rather pessimistic about whether social sciences are capable of it) and it deals with a unique historical event.
Nothing prevents us from applying the same approach in historical social science. For example, the Great Divergence, or the beginning of the Industrial Revolution in Great Britain. Again, the unique features of Great Britain (geography, demography, social structure, etc.) provide the initial and boundary conditions for dynamic models. For example, the realtive proximity to Americas provided an opportunity to loot the societies there. Alternatively, North America provided millions of arable acres that helped England to get out of the Malthusian trap.
Next, it is important to realize that the Industrial Revolution was a complex event that involved rapid changes in many distinct spheres, although these were connected by feedback loops. Thus, we need to have separate models addressing such questions as: How was agricultural productivity improved? Why did England escape the Malthusian trap? What were the causes of institutional change, leading to government becoming more responsive to populace? Why did the pace of scientific and technological change accelerate? And many others. Each of these questions can be modeled, with models used to make predictions to be tested against the historical record (in the same way as the climate models and the fossil record were used to test the Alvarez hypothesis).
There is one big difference between explaining the Cretaceous–Paleogene extinction event and explaining the Great Divergence. Climate models have really got quite good during the last few decades, and while we don’t have complete understanding of how and why climate changes, we do have reasonable models with which to simulate the environmental effects of an asteroid impact. In this respect historical social sciences lag behind historical natural sciences. In a few cases, we have reasonable models (e.g. demography), but usually this is not the case. For example, how do we model technological change?
And now for a dash of realism. Although questions as to why the Great Divergence occurred are fascinating and important, and eventually we will be able to address them, right now the state of historical social science does not yet permit us to develop a rigorous research program comparable to the one on the Alvarez hypothesis.
On the other hand, fifty years ago geologists, climatologists, and evolutionary biologists were in a similar state – they did not have the tools to do what they are doing now on dinosaur extinctions. The first order of business for us, then, is to develop general theories of social change that will give us the tools to answer questions about unique events such as the Great Divergence, the Industrial Revolution, why the Roman Empire fell, and many others. Building and empirically testing such general theories about history seems to me to be a much more productive way of utilizing our collective energies and talents (instead of arguing fruitlessly, as is the general tendency now).
And that is, of course, what Cliodynamics is all about.
I’ve had inklings of these thoughts before—I’m sure many others have as well. Why do you think it is that a more rigorous, quantitative approach to studying history hasn’t gained traction? Is it just a matter of time until it does, or are there cultural or institutional factors that militate against the discipline evolving in this direction?
Personally I think adverse selection might be part of the problem (the humanities seem to draw heavily from the math-phobic), but perhaps not for much longer, if recent trends continue.
In any case, I’m glad I stumbled across your blog. Looking forward to getting caught up 🙂
It (quantitative, rigorous approach to history) is gaining traction – lots of people are joining in on the fun, and many more are discovering it independently. The conditions are right, and it had to happen. And it is happening.
What of the role and contribution of enslaved labor?
Slightly off topic, but the geologist Walter Alvarez (of the Alvarez impact theory) is a proponent and practitioner of Big History, which has some overlap with cliodynamics. One difference, though, (it seems to me) is that Big History is more oriented to teaching than research. Eventually cliodynamics will need textbooks and other teaching tools.
“For example, how do we model technological change?”
This is a question which probably deserves its own blog post. One could ask an even more fundamental question: “what exactly is technology and what are its key characteristics”?
To give a (not so) brief survey of how the problem has been approached in economics:
* Roughly in what may be called “Classical Economics”, up to WW2, there is very little distinction between the process of technological improvement and capital accumulation. The two were treated as the same or only the latter is addressed/emphasized. This description of the thinking is of course a big oversimplification but I think it captures what differentiates these views from what came later
* And that was the Solow model, where technological progress is modeled simply as a continuous and exogenous increase in labor productivity that is not due to increases in capital accumulation (lots of cans with lots of worms inside here). Indeed, the main point of Solow, in my view, is that it is precisely this continuous, exogenous, and unexplained technological progress which is responsible for observed improvements in standards of living, what Solow called “the measure of our ignorance” and later economists christened the “Solow residual”
* While Solow tells us that in the end it’s NOT capital accumulation, but this technological “dark matter” that is responsible for income growth, in that model technological improvements fall like manna from heaven. Obviously this is not satisfactory and this led to the so called “New Growth Theory”. This had antecedents in some microeconomic models from the 50’s and 60’s (Ken Arrow’s “learning by doing” or Marvin Frankel’s “knowledge and capital spillovers”) – in fact I remember reading somewhere that Solow said this was all just old stuff dressed up in newer fancier math – but was mainly developed by people like Paul Romer in the 1980’s. This view basically treats the invention of better technology and new ideas in the same way that production is always treated; you got your inputs, maybe some randomness, you combine it all and out comes output, i.e. new technology. Romer focused on purposeful R&D activity by firms (and government) while other folks, like Robert Lucas, have stressed the role of education in knowledge production.
* At the end of the day though the “New Growth Theory”, in my view at least, has failed to really address the fundamental question here, which is why some economies experience high rates of technological progress/adoption, while others don’t. This kind of distinction between “proximate” causes of growth, like technological improvement and “fundamental” causes, like the actual reason why it happens or not, has been stressed by folks like Daron Acemoglu (whom you mention in the previous blog entry). And that sort of led back to the discussion of “Institutions” as key. But this view too has been criticized (defining and measuring “institutions” is tricky, their influence has been over stated, there’s circular reasoning going on, and institutions themselves are endogenous given enough time), most notably by Greg Clark in his Farewell to Alms.
So I think this area too has essentially seen a lot of “clearing of debris” but has not been able to provide a satisfactory account.
For a more historical take, Joel Mokyr has the wonderful book “Lever of Riches” as well a few more recent works that I’ve been too busy to sit down and work through. Some people have began trying to employ some of the methods and insights from the New Growth Theory to historical economies (for example http://www.nber.org/papers/w14484), while others, like Oded Galor have put the demographic transition in a more central place. Avner Grief and others have fleshed out the role of institutions in supporting technological innovation. Going a bit further back there’s of course Ester Boserup’s idea that technological growth is driven by necessity.
I would say that even after 100 years of thinking about it, we’re still in “early stages” of understanding what technological growth is and where it actually comes from.
Hi, Radek – I think this is very interesting and does deserve a separate blog. So how about doing a guest blog on this issue for the SEF? Send me an e-mail to discuss.
“I started by arguing that you cannot do science that is solely directed at explaining a unique event, such as the Great Divergence.”
Unless your explanation makes the event not unique.
China, India and parts of the Middle-east weren’t all at the same level as Europe and then Europe suddenly took off. China, India and parts of the Middle-East were *ahead* of most of Europe for a very long time. Europe eventually caught up and then some time later Europe ran ahead.
If the explanation is that relative wealth and power are the product of relative technology gaps then the Great Divergence isn’t unique because over the preceding period China, India and parts of the Middle-east had been Greatly Diverging to various degrees with everywhere else, including Europe, as the technology gaps varied over time.
So if the grand explanation is relative technology gaps then the underlying explanation would be what causes technological innovation?
And that couldn’t be unique to the Western Divergence because the earlier Chinese Divergence had happened centuries before so if this explanation is correct China must have had a lot of what causes technological innovation at one time but gradually lost it while Europe had a lot of it at a later time. So according to this idea the question then becomes what was *similar* about China or India or Europe during (or just before) periods of great technological innovation.
Agreed. In fact I make the same point in the previous blog:
There was never much divergence actually between agricultural societies of whatever time period. Even the most advanced peoples produced hardly above the subsistence level. This is pretty much agreed upon. Quote Milanovic (An Estimate of Average Income and Inequality in Byzantium Around Year 1000; from 2006): “A realistic maximum income that could be envisaged for the pre-industrial societies might be a bit more than twice the subsistence minimum, or around $PPP 1000 (at 1990 international prices)”.
So even the wealthiest pre-industrial society could not produce more than twice as much as the poorest peoples who merely hung on to their lifes (by contrast, the modern divergence between industrial and agricultural societies could be easily 20-30 times the subsistence level).
Once you realize how small the productive margin everywhere actually was, the mainly quantitative California approach cited above immediately loses much of its appeal, even if we do not take into account the feeble empirical basis it actually builts its case upon (there is little hard data for ancient commodity prices). We should rather turn our gaze back to the qualitative dimension because while the Athenian rower might not have earned much more than a Persian shepherd of the same low class, he was instrumental in creating democracy while the latter left nothing comparable behind.
Peter, thought you might be interested in my follow up post (yep, I know it is a month late…) on this series:
“The Rise of the West: Asking the Right Questions”
T. Greer. The Scholar’s Stage. 7 July 2013.
Thanks for the link. One month is not late! All good things need a thinkover.