Rationally speaking, is a presentation of New York City skeptics dedicated to promoting critical thinking, skeptical inquiry and science education. For more information, please visit us at NYC Skeptic's Doug. Welcome to, rationally speaking, the podcast, where we explore the borderlands between reason and nonsense, I am your host master YouTube. And with me, as always, is my co-host, Julia Gillard. Julia, what's today's topic? Masimo.

Today we're asking, is it possible to study history scientifically? Are there general laws and principles underlying how history unfolds? And can we infer them from historical data with a level of rigour matching that of the sciences?

Here with us today is our first live guest, Professor Peter Turchin, who's a professor at the University of Connecticut in the Department of Ecology and Evolutionary Biology and adjunct professor of mathematics. He works on the scientific study of historical dynamics, a field for which he has coined the term Clear Dynamics. He's written three books on the subject, including 26 Is War and Peace and War The Life Cycles of Imperial Nations. Welcome, Peter.

Thank you. So just to start off, I have to say my first reaction to the idea of studying history scientifically was that human societies are just too complex and history being a result of of thousands and thousands of different variables. And changing over time would just be it would just be too infeasible to study it with the kind of rigor that that we require in the sciences. I'm sure you've you've heard that a lot.

Well, in fact, that was my first reaction. You remember, I was trained as a biologist and I became, well, clear dynamicists only, you know, in the last 10 years. So it started really as a hobby. I thought I thought about some history of biology, the work of people such as Volterra and Alfred Lord, and how useful were their mathematical models that they designed, how useful they were in their study of population dynamics. So I thought, well, let's try and see what might happen then.

It was just a lark. But then as I started digging deeper into the subject, because I can't just work on mathematical models, I am immediately curious as to how these mathematical models might relate to data to reality. So I started looking for data to test the models. And to my great surprise, I started I found, first of all, that there is a lot of data. We can talk about that later if you wish, but more importantly, that there are very strong empirical patterns.

And when you have strong empirical patterns, that suggests that there may be some general principles, laws of history, if you wish.

So, Peter, underlie this battle. We'll get to a couple of examples soon so that our listeners can can get a better grasp of the kind of things that you that you're doing with this approach. But from what I hear, this is therefore significantly different from, say, another famous biologist who did some scientific approaches to history, which is George Diamond's work. So do you see that as a continuum? Your your work is a continuation of the kind of approach or as I suspect, actually something quite different?

Well, no, I see us working towards the same goal. Both he and I are looking for general patterns and also we are looking for general explanations that underlie this patterns. Jared Diamond's book is a wonderful piece of, you know, scientific writing.

It's where we should say that guns, germs and steel well collapse. Yes, I actually like guns, germs and steel better than Collapse of the Collapse is a wonderful book. To Diamond, on the other hand, does not push this approach as far as I and my colleagues have used to push it.

So he throws out a wonderful hypothesis and brings up some very suggestive data that may support explanations that he proposes. But he doesn't bring the whole power over the scientific method to the study of history.

I would say because is not quantitative enough, you think? Yes, well, it's more than that. So briefly, you know, the tried and old scientific method tried and true scientific method is you have to design alternative hypothesis, right. Competing hypotheses, competing explanations for how nature or societies operate. And then you go and you find data that allows you to distinguish between the predictions of these two alternative hypothesis. That's basically a scientific experiment in a nutshell.

Essentially, a scientific experiment is not about messing around with nature, although many of them are manipulative, so-called manipulative experiments. The experiment is really about distinguishing between two out of two or more alternative hypothesis on the basis of data.

But but an interesting objection that could be raised. That is, I don't want this discussion to go too far afield. And actually I want to go back to to explore one or two of your examples. But a reasonable answer. I guess objection to this approach could be that, well, there are even very well established fields in science where that method that you just described doesn't seem to work very well to the obvious and most controversial one these days is string theory, super string theory in physics.

My understanding is that there is something like two hundred general families of versions of this theory, none of which makes any sense, any empirically testable prediction, at least at the moment. Now, if if sometimes it's so difficult to make a precise empirical predictions in areas that are clearly scientific, such as fundamental physics, I can see somebody saying, well, and you want to do that in history. So how would you do that? How can we get to one example of how to do that?

Well, I'm pleased to hear that. It turns out that history, historical theories are easier to test than some physical theories, I think. Funny thing, huh? Yes. Well, I think that there could be a there could be no science without empirical testing of theories, empirical testing of theories. That's that's basically the heart of science. And well, the problem with string theory is that basically they need very high energies.

As I just said, I'm not a physicist myself. So in principle, it's still science because it's testable. Right. It's just that maybe they haven't figured out a way to test it in history. On the other hand, it is actually not that difficult to test theories. We do it all the time.

Can you give us a specific example of your work then? Then what is it that you found out that it's interesting for the general public, let's say?

And can you tell us how you actually go about testing a theory? Is it possible to get new data in history, given that it's all in the past? Does it rely on us discovering new things we hadn't already discovered?

Yes, absolutely. So basically, let's talk about scientific prediction. Scientific prediction doesn't have to be about the future. All right. In science, we aim at understanding. Prediction is basically an intermediate step that we use to test alternative hypotheses. We extract predictions which are different predictions. The predictions eglash about some aspect of reality from two alternative hypothesis. And we go and take a look and see what does the reality conform to reactions of one hypothesis or the other one.

So there is nothing that has you know, it doesn't have to happen in the future. And that means that in many historical fields of science, such as history, geology, evolutionary biology, astrophysics, these are legitimate branches of science. They don't allow manipulative experiments, but they allow what's called as manager of experiments, which is what I have described.

I think that's an interesting point. If you don't mind for a second, we should emphasize for our listeners, which is this is a common misconception about science. There are several common misconceptions about science, one, which is that it has to have to include experiments. And that's not the case, what you call manipulative experiments. You can do perfectly good science by observational testing of hypotheses. But the other very common misconception is that predictions are, in fact about the future.

Of course, in everyday language, that's the meaning. Usually that is the meaning of the term. You know, if you predict something, it's about something in the future. But you're right. In science, predictions actually can can be about past events that have already occurred. And they are therefore predictions about the discovery of particular kinds of evidence that are that are supportive or not of a particular. But as you mentioned, historical science, such as evolutionary biology and for instance, by ontology and palaeontology, does provide us with one of the best examples of historical science.

When people finally figured out that there was, in fact, a large impact 65 million years ago, that probably contributed quite a bit to the demise of the dinosaurs, as well as a lot of other species. But initially, this was just a hypothesis is a possibility. But in fact, several pieces of crucial evidence were found after the puzzle was formulated. And it is one of the best examples of historical science that has been established on hypothesis testing. So it's clear that that you can, in fact, test hypotheses about about the past.

What about human past then? But in fact, there is a word for it, as some people call it, a prediction, where you make a prediction about some factoring in the fact the best way to approach this is to actually predict something that you don't know and there is no chance, no possibility for you to know. So, for example, ideally, the way we would do it in the clear dynamics is that we will generate predictions until archaeologist's.

Well, this is an ideal situation to remember. OK, dig there and find out what is actually the case. That would be a completely ironclad example of scientific prediction because there was no way ahead of time for us to know what is under two metres of dirt in that particular location. All right.

So we haven't yet got to that point, but I don't see why we shouldn't in. The near future. So let me give you some examples of how you go about testing theories. Meanwhile, so well, let's for example, take there is one very general pattern in history, which is which we call secular cycles, secular cycles. So it's it happens. It turns out that historical societies are affected by recurrent waves of political instability that tend to happen every couple of centuries or so.

And so essentially, historical societies go through this very long term oscillations which have may divide this oscillation into two phases. The integrated phase and disintegrating phase.

The integrated phase about a century long or sometimes longer, is characterized by internal peace and stability, population growth and general optimistic outlook of the society. And disintegrating stage is the vice versa. We have a lot of civil war and other kinds of political instability of populations either stagnate or even decline, sometimes even collapse. And there is a lot of pessimistic ideologies that actually are present in the society. So you have this general, very general pattern. Predictions could be made at several different levels.

First of all, you could say, well, there's a general pattern that would affect any agrarian society, any society that that is a state which is large enough to develop under its own power, so to speak, due to its internal dynamics, should go through these cycles. And so the way to test that prediction at the empirical level is to look into other regions of the world compared to the ones where we developed this theory. Right. And see if something like that happens.

And that's actually how it has been done. For example, maybe we'll talk later about Southeast Asia region, about which I didn't know anything. And it turns out that there is a very interesting historian specializing in Southeast Asia who has discovered exactly the same pattern.

So, Peter, the question then that arises from the example you were just explaining is that I think plenty of people might grant that there are patterns that you can quantify, which certainly is something that historians themselves should be doing or interesting should be interested in doing.

But there's, of course, a difference between identifying patterns or quantifying patterns and in fact, testing causal hypothesis. That seems to me the more tricky step in the case of human history. Right? I agree.

So let's talk about there are several alternative explanations that explain why we have this repeating oscillations, by the way, they're not perfect cycles. They vary in period somewhat. But we can talk about that later.

So you have to authoritative hypotheses. One of them is the Malthusian model. It suggests that as population increases to the point when it starts pressing on its resources, popular immiseration increases. And that directly brings about instability, warfare, which decreases population, and the cycle can start again. So you have this cycle which is driven purely by demographic mechanisms. The alternative theory. It's called demographic structural theory. It also has the demographic growth as a component, but it assigns more importance to what happens to the other two components of the social system, namely the elites.

These are the people with power. So think about nobility of their medieval France and what happens to this state in the demographic structure theory, their way of instability happens not because the population is miserable, because peasants in those agrarian societies didn't have much military power and peasant rebellions were easily crushed as long as their elites were unified and the state was strong. What actually happens is that there is a problem called elite overproduction. There are too many elites who are fighting for the diminishing slice of the same pie and the state becomes weakened.

So it is when the state loses their ability to pay for their army, police and so on and so forth. And by that point, the elites have already got to the point where they are riven by rival factions. That is actually basically everything starts to break down. The state loses the ability to control their internal order and the elites, rival factions over the Head Start civil wars. All right. So there's two alternative theories. How can we test them?

Well, we can test them by looking at the actual sequence of events, you know, you take. Look at what happened in the medieval period in the late Middle Ages in, say, England or France. There was a very strong population growth during the most of the 13th century. And by the beginning of the 14th century, we had a very strong Malthusian situation, starvation, popular immiseration and everything.

But there was no really a breakdown of the state and there were no civil wars. It really took the additional push of the Black Death, which decreased the population as a result, the social pyramid, because the elites were much less affected by their mortality. It's sort of a pyramid became top heavy. There were too many elites for their sort of social support base. And so as a result of that, the whole society was thrown into disorder. And that is when actually you had all the civil war started and then vice versa.

Once there was an explanation that once the Black Death destroyed about half of the population, you should have seen the resumption of the demographic growth and the whole cycle to start again. But that's not what happened. For another 100 years or so, English society and French society were experiencing a series of civil wars and the population never grew. It actually declined and stagnated. So it really was that other two factors, the elites and the state that that were responsible for them this long period of civil wars such as.

So you would count that as sort of a falsification of the Malthusian hypothesis? That's right. And so this example but what I was talking about, where the data is, actually is supportive of the predictions of the demographic structural hypothesis, but it contradicts the predictions of the Malthusian model because the theory developed using that data.

Or is that new data? Because the key to making predictions is not as much as I said. It's not about data from the future. It's just data that we didn't already have when we were developing the theory that precisely.

So what we do next, we actually now apply the same theoretical predictions to other societies and what we find if you look at the Roman Republic late, the transitional period between the republic and the Empire, you'll find very similar pattern. Again, about a hundred years of civil wars. And in fact, the evidence for demographic pressure on resources isn't terribly strong, but there is evidence for elite overproduction is very strong. So we just go and test the theory in a variety of case studies like that.

So, Peter, as you might imagine, this kind of topic brings up a lot of comments. We usually post a teaser for the podcast before the actual the actual episode. And we have several interesting comments and we may get to discuss a couple of them. Julia, what did you find going through that?

Well, one commenter finally said that it's the the black swans, the really unpredictable events that, as far as he can tell, have the main the major impact on the future. He says historians are pretty good at finding plausible after the fact explanations of why things happened. But they nearly always miss the big factors when they extrapolate to the future, such as World War One, the collapse of the Soviet Union, the housing bubble, 9/11, etc., etc.

And those are the events that really changed the course of history. How does that fit into your description of sort of universal repeating patterns? Yes, obviously, black swans is something that happens all the time, I would not overemphasize the importance of such events because basically dynamical systems come in a variety of guises and shapes. Some of them are characterized by this runaway type of things like earthquakes and, you know, collapse in store for their stock exchange. They are characterized by black swans.

In other situations. You have more types of stable cycles and things like that.

But even the black swans, that's an interesting point. But I don't know if the black swans are not amenable, in fact, to the quantitative analysis. For instance, there is an analogy in in paleontology where abundant energies have identified a number of more or less recurring mass extinctions. And a mass extinction certainly is an unusual event. I mean, by definition, it happens only only every every several tens of millions of years. It has a huge impact and so on.

And yet since the historical record is long enough, we have actually observed several of these in the fossil record and there are patterns and that you can start making some interesting about this is about why these occasional events occur. So you cannot predict necessarily the next event when that's going to happen. You can perhaps pick up a pattern on that. Yeah, that's right. So, for example, let's take wars.

The the magnitude of war has a power law distribution. So it is a Tiepolo black swan because sometimes wars really get out of hand. All right. So but there were thousands and thousands of wars in human history. So first of all, we can to study them statistically. That's why, you know, the distribution of war intensity follows that parallel curve.

But also, even though we may not be able to predict the amplitude. Right. The intensity of any particular war. Right. Like we cannot predict ahead of time with the intensity of the next earthquake. Still, you can do lots of useful things, you know, for example, where earthquakes happen right along the fault lines. And so the same thing, whether a war is any society is likely to go into internal war or not. Now, geographic structure, Syria, I was describing to you that actually makes pretty strong predictions about their timing and and dynamics of such events during wars.

You mentioned the statistical distribution that you can derive. And one of our commenters, Costus, talked a lot about how how much data might be necessary in order to test the theories. And I'm wondering, you know, we hold scientific theories to high standards of rigor like they you know, scientific phenomena have to be statistically significant in order for us to accept them. So what what role does statistical significance play?

Is there is there really enough data played a crucial role? Because without statistical significance, you cannot really everything is you're talking about stochastic or at least event dynamics, which are affected by stochastic factors. And therefore, we have to use statistics to answer to test theories.

Great. Thanks, Peter. It's fascinating. Now, however, it's time to move on to our rationally speaking, PEX.

Welcome back. Every episode of the Russian speaking podcast features picks by Julia and myself, but this time we have a live guest in our studios and we decided to do it differently. We'll have our guest, Professor Peter Turchin of the University of Connecticut, doing the pick of of the episode.

Well, my suggestion was a book by a Michigan University. University of Michigan historian Ed Lieberman. It's called Strange Parallels. The reason I just finished reading this book, it actually comes in two volumes and I just finished reading volume two. And the reason I'm suggesting it is because this book talks about strange bedfellows and more success than the one with the original meaning of their title is that Victoria Lieberman is a specialist in Southeast Asia Asian history.

And he observed that societies that he studies, such as Thailand, Burma and Vietnam, went through a series of oscillations that are very similar to some of those solutions that we observe in other Eurasia and societies such as China, France, Russia and so on and so forth. And so he wrote these two volumes.

One of them was almost 1000 pages long, detailing in detail there are similarities that we observed between the dynamics of such different regions in the world. And the reason I thought this would be an interesting pick is because here is an example where somebody working very independently of this stuff that I and my colleagues do actually hits on a very similar both patterns and explanations of historical dynamics.

So this suggests to me that there is something, something real in the world of study piter.

So I notice something interesting in the description Liberman's work, which is that these cycles are also appeared to be synchronous, even among civilization populations that are known, that have contact with each other, which suggests some kind of internal dynamics rather than a mechanism that synchronizes things across populations. Now I notice that there is a panel there with ecology and evolutionary biology where there is always a debate about the relative importance of internal dynamic factors affecting dynamics versus external factors. Right.

So if there is, what is the idea here? What might be the kind of internal factors that might synchronize these cycles across civilizations? Well, there are two sort of alternatives. One of them is that there is an external driver that actually affects societies as far flung apart as Europe and Southeast Asia and that that external driver actually produces this synchronize cycles. The other possibility, and even more exciting one, is that each society goes through cycles because of endogenous internal dynamics, and then these cycles get synchronized by some external cues.

The difference being the difference being that in the first case, the external driver is actually cyclic. In a second case, the cyclicity comes internally. All right. But then different cycles get entrained because of shared environment. All right. So this is one of the theories, a set of two outargue hypothesis that we could actually start testing with data and we have started doing that.

It's an interesting experience that you that your theory was strengthened by the fact that you didn't know ahead of time about this research.

So I'm wondering if if this experience is an argument in favor of of extreme specialization in history, like should you really only study one particular field so that your the theories that you generate are are sort of uncontaminated by the data from the rest of the world or other period, the history. What do you think? What did you draw from this experience?

Well, Victor Lieberman is a very unusual historian. Actually, historians have largely repudiated the search for general patterns and laws of history. They think there is not nothing such a thing that is just one damn thing after sort. So they wouldn't say that. But they say that everything is different. Every society is different, every period is different. And in fact, well, most historians don't even call themselves scientists. They are humanists. Right. And humanists have a very different set of priorities.

So Lieberman is quite unusual in this respect. He, of course, is very careful here, actually appease you know, he talks about how each state is different, but then he comes over and drives at the similarities. And that's very unusual for historians.

Thanks, Peter. That's all the time we have for this episode of rationally speaking. Join us next time for more explorations on the borderlands between reason and nonsense.

The rationally speaking podcast is presented by New York City skeptics for program notes, links, and to get involved in an online conversation about this and other episodes, please visit rationally speaking podcast Dog. This podcast is produced by Benny Pollack and recorded in the heart of Greenwich Village, New York. Our theme, Truth by Todd Rundgren, is used by permission. Thank you for listening.