On Testing Theories in History

August 31, 2022

by Peter Turchin

cliodynamics, empirical tests, History, theories

Comments are off

Noah Smith wrote a post on his Substack, On the wisdom of historians, to which a historian Bret Devereaux wrote a rebuttal, On the wisdom of Noah Smith. Yesterday I wrote a Twitter thread on this debate, which turned out to be fairly long, so I decided that it’s worth posting a version of it on my own blog. What follows is a lightly edited “unroll” of my thread.

In his response to Noah Bret takes a swipe in passing at my work:

It is not clear to what he refers (“effort to find support for this hypothesis in the ancient world”), as my my main effort for empirically testing this hypotheses has centered on the US from 1789 to the present, with a huge emphasis on the contemporary America (from the 1970s on). Perhaps America in the late 20 century is an ancient country? The main source is Ages of Discord.

I also stuck my neck out and made a scientific prediction in 2010, which unfortunately was borne out by reality

The 2010 structural-demographic forecast for the 2010–2020 decade: A retrospective assessment This article revisits the prediction, made in 2010, that the 2010–2020 decade would likely be a period of growing instability in the United States and Western Europe

Cliodynamics colleagues and myself have also applied theory to other contemporary societies, as well as many historical ones. So I would say that the elite overproduction thesis has fared quite well in empirical tests, and dismissing it in passing is not good scholarship.

At the same time, there is much in Bret’s blog post that I agree with. There are multiple epistemologies, and I quite appreciate what historians do, and I am very supportive of History as a mature discipline. My work is often misinterpreted as an attack on History, but nothing could be further from truth. See here:

“The Mad Prophet of Connecticut” Graeme Wood penned a “long read” about cliodynamics, me, and the Age of Discord in which we are currently find ourselves. Graeme is a very intelligent journalist and his explanations of cliodynamics a…

I also agree with Bret about his critique of Max Roser’s “data” which is indeed “not charting global deaths in conflict, but rather in charting the rate at which evidence for battles is preserved over time”

However, this doesn’t mean that such data are useless; to properly extract insights from them we need to include the measurement process in the analysis. This takes thought and work, but is a better alternative than just throwing up arms in defeat.

Returning to the main theme of Noah’s post, I see value in both the work that historians do (which is not science) and the need to empirically test theories (which is science). The latter is the task for Cliodynamics. There is no need for the same researchers do both.

I also want to return to my first encounter with Bret’s critique, which was of the Seshat project’s first attempt to test the Big God theory. That was admittedly flawed, but we didn’t give up — we redesigned data collection, gathered more and better buttressed data, and used better analysis methods. The result was published in June in a special feature in Religion, Brain, and Behavior, together with 7 commentaries, our response, and a retrospective:

Explaining the rise of moralizing religions: a test of competing hypotheses using the Seshat Databank The causes, consequences, and timing of the rise of moralizing religions in world history have been the focus of intense debate. Progress has been limited by the availability of quantitative data t…

I would be interested in Bret’s comment/critique of that. I think that this effort, although by no means the final word, moves us a long way towards testing (and rejecting) evolutionary theories about the role of religion in the rise of complex human societies.

Finally, Noah calls for empirical tests of theories in history. I’d like to attract his attention (and everybody else) to #Seshat project’s massive effort to test 17 different theories about the evolution of social complexity across the past 10,000 years: