Cliodynamics and the Dangers of Prophesy
These ideas express one of the oldest dreams of mankind — the dream of prophecy, the idea that we can know what the future has in store for us, and that we can profit from such knowledge by adjusting our policy to it.
- Karl Popper, Conjectures and Refutations
The goal of cliodynamics is to transform history into an analytical, predictive science. It is an academic discipline that seeks to predict revolutions, war, and famine — to find the laws of history which govern the fate of our societies. Success in cliodynamics means correctly prophesying when nations will fall to dictatorship, when wars and disease will emerge, and when new technologies will be developed.
Strong empirical patterns arise because the dynamics of historical societies reflect the action of general social mechanisms. There are laws of history . . . Furthermore, successful case studies of scientific prediction, reviewed in this article, show that we are well on the way to identifying some of these laws.
- Toward Cliodynamics — an Analytical, Predictive Science of History
I’m highly skeptical of such approaches to history because they tend to discount the role of ideas in shaping the world. They also seem to take for granted that “strong empirical patterns” do in fact arise. History is full of … well, history. There’s no doubt you can find periods of higher violence, of more disease, of population booms and busts. But trends are not universal laws — trends break. Looking for the laws of history is begging the question. By almost all metrics, after all, the modern world looks entirely different than previous eras.
Moreover, exercises in predictive history don’t have a great track record. The desire to uncover the laws dictating human destiny is nothing new. Religious apocalypticism alleges that history will end with the return of a deity or a heavenly kingdom come to earth. Thomas Malthus believed that history was characterized by inescapable cycles of abundance and scarcity. Hindu cosmology contends that history is cyclical. Adolf Hitler believed that history led to an inevitable race war. Karl Marx that it led to a class war and ultimately, to a communist utopia.
While these older historical prophecies are now (mostly) ignored, cliodynamics has gained the attention of the Atlantic, the American Conservative, the New York Times, the Financial Times, Forbes, and others. It has been published in Nature and has its own journal. It has caught the eye of the Sante Fe Institute and is included in university curricula. Proponents of cliodynamics contend that the difference between their discipline and past prophecies is the use of advanced mathematics:
Mathematical approaches — modeling historical processes with differential equations or agent-based simulations; sophisticated statistical approaches to data analysis — are a key ingredient in the cliodynamic research program. But ultimately the aim is to discover general principles that explain the functioning and dynamics of actual historical societies.
Even the most sophisticated mathematics, however, does not let you see into the future. The Austrian philosopher Karl Popper spent much of his life arguing against the prophetic tendencies — which he called historicism — of some of the leading ideologies of his day, such as Marxism and Fascism. In his 1957 book, The Poverty of Historicism, he refuted the possibility of uncovering general laws of nature. He dedicated the book to the “memory of the countless men and women of all creeds or nations or races who fell victim to the fascist and communist belief in Inexorable Laws of Historical Destiny.”
Popper identifies five untenable aspects of historicism. One of these is:
It is logically impossible to know the future course of history when that course depends in part on the future growth of scientific knowledge.
Future knowledge is, by definition, unknowable in advance. It is due to knowledge creation that previous apocalyptic forecasts have been rendered mute. For instance, Paul and Anne Erlich predicted worldwide famine starting in the 1970s in their book The Population Bomb, which begins with the statement:
The battle to feed all of humanity is over. In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now. At this late date nothing can prevent a substantial increase in the world death rate.
This prediction may have come true had it not been for improvements in our knowledge of agricultural practices, genetically modified crops, and food storage. Such advances could not, of course, have been anticipated beforehand.
Historicism witnesses the predictive power achieved in the physical sciences and asks why history cannot do the same. If eclipses can be predicted, why not revolutions, wars, and political trends?
[Cliodynamic] theories are constructed and based on general principles and tested empirically with comprehensive databases. In short, we use the standard scientific method that worked so well in physics, biology, and many social sciences.
In attempting to uncover general trends of human societies, cliodynamics believes it has found various cyclical societal trends. The Malthusian idea of “secular cycles”, for instance, is central to their predictions:
[A secular cycle is] a process that takes place of years [sic] where the disparities in the division of land in agrarian societies increase vastly as peace persists (the Matthew Principle), leading to the emergence of a large and bulky elitist bureaucratic class and leads to a tipping point. At this tipping point, population would persistently decrease and remain relatively low until peace reigns. After that, the disparities being again, marking the beginning of another cycle. This cycle is observed to have a period of approximately 300 years.
These cycles are simply assumed to be inevitable: they are part of our past and will always be part of our future. Such cycles, however, are found and predictable in nature only to the extent that they are isolated, like solar systems. There, the effect of outside influence is so small that it can be ignored. Even then, the interactions must be relatively few and the initial conditions of a certain kind for reliable prediction to be possible (see, for example, the three body problem and chaos theory).
Clearly, human history is not an isolated system — it is influenced by ideas which, as Popper notes, are unknowable in advance. Whether we will continue reducing global poverty is not an inevitable outcome. It depends on the future problems we encounter and the ideas we generate to solve them, both of which are unforeseeable. It is for this reason that the ability to correct our mistakes is so crucial in a dynamic, open society. We cannot foresee the future; we don’t know what all the consequences of an idea will be a priori. Progress does not entail always making the right decision beforehand — an impossibility. Rather, it entails continually solving the inevitable problems with which we are confronted.
It is not the exercise of predicting the future per se that worries me the most. There are many academic disciplines engaged with ideas that I think are wrong or fruitless. But believing that history obeys general and inexorable historical laws can be actively dangerous. Discussing a hypothetical example based on Hitler’s Germany, Vaden Masrani explores the sociological consequences of such a belief:
Imagine that everyone around you starts to take the idea that there is an impending race war very seriously. You’re told that this is not a prophecy, it’s not like the Mayans circa 2012. It’s science. It’s a scientific certainty as inevitable as the sun. Given this inevitability the question you would have to ask yourself is: what are you going to do to prepare? Because there’s nothing you can do to stop it, it’s coming.
- Increments, Episode 14: Prediction, Prophecy, and Fascism. (Lightly edited with permission.)
It’s easy to see how such rhetoric, especially once adopted by a large portion of society, leads to frightening consequences. A belief in the inevitability of violence can make it so. Imagine the ease with which one can ignite a revolution or a war if one side believes it is their destiny to triumph.
Troublingly, cliodynamics makes similar predictions. It is claimed there are historical laws portending that violence occurs in fifty-year cycles:
[O]ur cliodynamic research has identified a number of periodic processes in historical dynamics. And one of them is the 50-year cycle in political violence . . .
In the United States we also see this cycle, which resulted in spikes of political violence spaced almost precisely 50 years apart: late 1960s–early 1970s, circa 1920, and in the 1860s–early 1870s . . . It is one of the reasons for my prediction that we will experience a peak of political violence in the early 2020s.
One can almost feel their confirmation bias being activated. “Well, we have had a particularly violent year, right? Perhaps cliodynamics is on to something after all. And if violence going to happen anyway, it might be wise to act preventatively …”
But patterns are easily found in all datasets. This, in fact, is a big cause of the current replication crisis plaguing the social sciences. If the research question is not specified beforehand and datasets are repeatedly analyzed in different ways, patterns are sure to emerge — even if the data are completely random. If a researcher is looking for any kind of cycle in a dataset, especially when that dataset is the entirety of human history, they’re sure to find it. If not violence, then weather patterns, or banana exports, or car manufacturing, or religious adherence, or political opinions, or so on ad infinitum. And after a pattern has been identified, we humans are very good at activating our confirmation bias and generating a post-hoc rationalization for its existence.
Moreover, this pattern of violence is based on only 1,590 incidents. While inferring future trends based on the past is impossible, analyzing past trends can of course be a legitimate endeavour. However, a few thousand incidents is much too small a number for any sort of robust analysis. There were 18,830 homicides in 2018 in the US alone. Peter Turchin, who ran the above analysis, claims that he’s only interested in “political violence”, and so filters all incidents that do not meet this criteria. But what exactly does political violence entail? Apparently not the American Civil War, which he excluded.
To his credit, Turchin, who may be considered the founder of cliodynamics, seems to be aware that his approach to history might be flawed. “It is possible that this new batch of theories will eventually end up on the same trash heap of history as Marxism and Social Darwinism. But I don’t think so,” he writes. It should also be said that there are legitimate aspects of cliodynamics mixed in with the prophecies. Applying novel statistical methods to historical records may shed light on previous periods of human history and help make discoveries thereof. But insofar as cliodynamics tries to predict the future, Popper’s critique is a devastating blow.
Note: Edited on Feb 3, 2022. I felt that the tone was overly harsh in the first draft and tried to dial it back a bit.
