If historians think their field is only about recording past events, there's no reason to listen to their opinions about current events. They're effectively saying they have no more expertise on the matter than my hair stylist.
I think no historian regards the field as only about recording (or creating a narrative record of) past events. Discerning causal factors, institutional norms and their constraints, cultural perceptions and their influence--things like these are elements of just about every historical exercise.
When we ask questions about politics and culture today, people who have spent a lot of time and attention looking at patterns in political and cultural developments in the past have additional resources to identify possible pattern emerging now. The question is whether they are deploying those resources responsibly--good historical work takes a lot of time, effort, and intellectual reflection. I think Noah's best point is that historians who don't devote equal care to analysis of the present, but just note that points of past/present resemblance may constitute a recurring pattern, are better regarded as pundits than as academic authorities.
I think the right point to start at is Popper. What claims can be falsified? Many prefer history as a qualitative wrapper for explanation/rationalization of their priors. Others who prefer quantitative wrapping choose economics. But the qualitative-quantitative divide (which is really just about the handling of uncertainty) is not solved by breadth of knowledge or statistics. Both fields can form bad habits of rewarding those who put up the tallest pile of junk to intimidate those from investigating.
Popper’s positivistist approach offers a way to sort the good from the bad. Is it testable? If the answer is yes test it. If the answer is no don’t invest too much authority in the explanations you encounter. Taken to its extreme it can lead us to focus on testable questions (eg looking for your lost keys only in places where there is light) but we’re pretty far from that pole. Our current problem is unnecessarily complex explanations for where keys might be from strangers with no stake in the outcome.
The skill set we should focus on is answering the question “is it testable”? If they answer is yes, the proponent is as vulnerable to a grad student with too much time on their hands as they are to a giant of the field or an influencer. If it is not testable, assume it’s wrong, but be open to useful pieces. Not everything in the junkyard is trash.
Perhaps the field of history can function as an empirical discipline providing the facts about reality against which theories can be tested. In other words, rooting out the facts of the past that may otherwise be lost to the present seems like a noble enterprise in and of itself. The fact that theories may then be tested against those facts is just an added benefit.
“Testable” is good. “Falsifiable” is insisting on an unreachable standard. You would think that Newtonian gravity is a good falsifiable model, but the history of 19th century astronomy, where anomalies in Newtonian gravitation led to the prediction first of the planet Neptune, and then of the planet Vulcan, show that its more complicated.
Could you expand on this, Mr. Easwaran? I think the fact that there are anomalies that are later explained within the paradigm doesn't seem to me to attack Popper's standard of falsifiability. All that's required is the potential to stipulate some conceivable way in which a theory could be tested in practice (with conceivable technology) and be shown false, not that the test actually be available, or show the theory to be false, or that an initial false result itself never be overturned.
As I recall the history, it wasn't anomalies in Newtonian gravitation that led to Neptune's discovery, but anomalies in the orbit of Uranus, as Newtonian physics predicted it. Newtonian laws were deployed to hypothesize an undetected planet and search for it (and similarly with Pluto).
But physics is not exactly my field--my last course in physics was in high school, well over fifty years ago--so this post may just be my latest instance of bloviating public self-sabotage.
Around the end of the 18th century, the orbit of Uranus was found not to agree with the orbit predicted by Newtonian gravitation, given the known planets. This was judged not to be a falsification of Newtonian gravitation, because the anomalies could be explained if one postulated another planet, farther from the sun, which was called Neptune. Soon afterward, a planet of the relevant mass was found at the relevant location.
Around the middle of the 19th century, the orbit of Mercury was found not to agree with the orbit predicted by Newtonian gravitation, given the known planets. This was judged not to be a falsification of Newtonian gravitation, because the anomalies could be explained if one postulated another planet, closer to the sun, which was called Vulcan. Soon afterward, no planet was observed at the relevant location. However, that location is so close to the sun that astronomers thought it might just be unobservable.
Eventually, in 1916, it was discovered that the difference between Newtonian gravitation and Einsteinian gravitation was exactly what was needed to explain the anomalies in the orbit of Mercury. In fact, this is now said to be one of the big three observations that confirmed Einstein’s theory (the other two being the Michelson-Morley experiment in the 1880s, that had failed to identify the speed that the solar system was moving through the cosmos, and the observation of gravitational lending of light during the 1919 solar eclipse).
From a modern perspective, it is natural to claim that the anomalous orbit of Mercury was the sort of prediction that could be used to falsify Newtonian gravitation (or that the Michelson-Morley experiment is the crucial experiment that could falsify Maxwell’s non-relativistic electromagnetism), but at the time they were not perceived as such. This is because it is impossible to falsify a theory without holding fixed certain auxiliary assumptions (like the number of planets in the solar system, or the way the luminiferous ether can be moved) and anything that is taken to falsify the theory can instead be said to falsify the auxiliaries. It is only when a better theory is available, that explains the observations without any strange auxiliary assumptions (like an invisible planet closer to the sun than Mercury) that the observations are retroactively called a falsification.
The cases of Neptune and Mercury are very different. One was solved within the Newtonian paradigm, the other had to await the development of an alternative paradigm for solution, and the anomalies were, as you nicely phrase it, assumed to be a function of "auxiliaries."
But Popper's theory of falsifiability seems to me entirely consistent with this history. Basically, what it says vis a vis Newtonian gravity is that it is a scientific theory because evidence of objects in the world not conforming to its predictions counts as evidence that the theory is false. That doesn't mean evidence constitutes disproof, but that it contributes against confirmation and towards disproof, and can, in principle, ultimately overturn the theory.
A weak theory will be quickly falsified because evidence against it will far outweigh confirming evidence. Because the Newtonian theory is an extremely strong theory, evidence will overwhelmingly tend to confirm it, but, to the degree it is imperfect, some evidence will confute it. As Kuhn's sociological model predicts (based on just the type of evidence you've laid out), as disconfirming evidence builds for a theory so strong as to underlie a scientific paradigm, it will at first be treated as the anomalous results of imperfect tests or imperfectly understood tests, but will gradually either undermine absolute confidence in the strong theory or be recognized as better satisfying an alternative theory.
That's all consistent with falsifiability. If Newton's theory weren't falsifiable, it wouldn't have been falsified--just as the theory that God created the universe cannot be falsified, and is therefore not a scientific theory, though "in theory" it may still be a true theory. (My claim that I can't recall what I dreamt last night is not scientific--it's not falsifiable, but it's still a true theory.)
This was an incredibly thought provoking, and at times uncomfortably piercing, piece about what historians can usefully add to the public discourse. As a historian, I definitely felt smug in my assumptions that in the battle between the social sciences and history, the context driven process of studying history was far superior. This was an important corrective so thank you!
Whether history is a ‘relevant’ topic is a very vexing one for historians, particularly given lower uptake of the subject at universities. It can lead historians down a dangerous path where we overextend it’s ability to reveal things about the world around us.
One distinction that I might venture is that the commentary that you cite here is one of ‘historical analogy’ - a deeply assumption driven and (in my opinion) dangerous exercise - with contextualisation. The latter is perhaps a more useful contribution for the likes of me to make!
In my experience, a strong grasp of the context within which a situation is unfolding (which includes deep historic trends and traditions), can be used effectively to inform when a predictive model is likely to have blind spots (as all must do, no matter how well evidenced and thought through). This is certainly not the sole prerogative of historians - lots of academic fields can add value there in a variety of aspects.
Sadly for historians looking to get into the political commentary game, that both requires knowledge of the models themselves (so we’ll have to rely on you darstardly economists), and specific knowledge of the context at hand. So built in to a historian’s expertise is their specific knowledge, an unhelpful factor for poor TV producers and editors looking for a reliable source of opinion that they can use for multiple events!
If social science wasn't such trash right now, I'd say that the main role for history w.r.t. prediction would be to provide the variables into their models.
Any direct predictions based on historical parallels need to be taken with a whole heaping barrel of salt.
Dear Noah Smith, excellent blog post with very high academic quality. First on Devereaux would be tyrants, the experience of Pakistan is that even democratic leaders have tried to cling to power. For example, the exiled leader Nawaz Sharif contested elections for the position of the prime minister 5 times since 1988 and won the majority seats 3 times. In 2017, though he was disqualified by the highest court on some say flimsy charges, he is yet looking for another 5 year term for himself. Same is true with Imran Khan. Most of the Pakistani think that Turky's Erodogan that is viewed by Western commentators to be less democratic, is the ideal case whereby one can cling to power for decades under democracies. Similarly the benevolent dictator Musharraf stayed in power from 1999 to 2008 and he wanted to continue his rule as is told by his acolytes based on some good economic growth years under his rule in Pakistan that he is indispensable for Pakistan. How would you translate his leaving of the office of president in 2008 not voluntarily but after years of struggle and protests by democratic forces within the country. Secondly the free media that he helped in creating by allowing private news channels was one of the most significant factor that informed people of Pakistan adequately for the country to have a democratic movements against his dictatorial rule.
Secondly you have again brought a very important empirical argument of causation and correlation in case of slave trade and earlier industrialization in the US. You are absolutely right that industrialization happened due to enterprising nature of American society that is raising the voice against slavery as early as 18nth century. So there is more of a correlation between slave trade and industrialization. And if anything industrialization has helped the American society to work on strengthening democracy in US whereas democracy has come with greater rights for African Americans and women and thus one may suggest that industrialization has actually lead to the abolishment of slave trade in America.
I have my BA in history, which I thankfully got prior to various forms of Theory completely capturing the discipline. I agree with the post generally, in that those in the business of making predictions ought to be subject to empirical evaluation. However I would say the assertion is less about history and more about punditry, which is actually what these people are doing. What's really going on here isn't an investigation of the past, but a statement about the present, and therefore not really history at all.
And anyway most of the kinds of claims in question tend to collapse, or at least become complicated, based on the historical record itself. A good historian that wants to engage in punditry will understand and at least attempt to account for the kinds of confounding facts and factors that crop up, frankly, everywhere. But where would the fun be in that?
Noah, while I find many parts of the piece thoughtful and useful to consider I must object to the characterization of Devereaux's piece. I believe that the quotation selected is misrepresentative to the level of intellectual dishonesty and the failure to properly analyze the piece serves to greatly undermine your argument. While you correctly quote from his introduction to a discussion on Peisistratos, you have omitted the conclusion in which he adds a qualifier to the prior "explicit law of human society," namely that Devereaux identifies that "Unless would-be tyrants *are made to face the consequences of their attempts to seize power*, they will keep trying until they succeed so thoroughly that justice is beyond recovery" (emphasis added). This not only fundamentally changes the law he is discussing, it makes the proposed counter-example of Richard Nixon no longer a counter-example: Nixon faced the most serious consequences of any president before or since and would almost certainly have faced criminal charges if not for Ford's pardon. If we are to critique bad theories by historians, we owe it to them to at least correctly identify these theories.
I also critique the link provided for the "historical list of dictators who stepped down." On said list are a mix of figures like Daniel Ortega and Indira Gandhi, who after losing an election sought to regain power through a more central regime (and in the case of Ortega, succeeded in doing so) and thus would be proof that would-be tyrants will keep trying. But also are figures like George Washington and Cincinnatus, who it would be quite a strain to argue qualify under the banner of "would-be tyrants". Again, we ought to critique bad theories by historians but not with shoddy scholarship that isn't really addressing the core claims.
Also, it feels bad. Economics still has foundations and funding and well paid positions in companies.
Historians, as you pointed out earlier, are disappearing. Are you going to spend the 5 years to learn classical Latin, Turkish, Greek, Manchu etc. necessary to do the history?
History doesn't survive. That's one of the first things historians learn.
We have many Greek and Latin texts that refer to "the history of x written by z" that are lost to time.
Also, he chose a poor example going after devereaux who is super good about talking about how contingent everything is. His tyranny bit is worth reading and he caveats it by saying this is based on Greek ancient cities.
You make history sound a lot like chemistry. Chemistry has much too large a search space for certainty. You don't know how a reaction is going to run until you dump everything together. When someone wants to synthesize some chemical or break down another, chemists reach into their bag of analogies. Sometimes they apply. Sometimes they don't. There are all sorts of heuristics one can apply as well. Chemists spend a lot of their time reading papers old and new in search of useful analogies. That doesn't make chemistry any less a science.
A big difference is that a chemist can try out a lot of things in a relatively short time. Astronomers and historians don't have this advantage. When they want to predict things, they have to look backwards for analogies, suggest things, and then, like the rest of us, wait to see what happens. For example, consider all the predictions that were made about what the Webb telescope or the LIGO gravitational wave detector would reveal. Just as with the outcomes of political events, there are some things one expects, e.g. black holes colliding and disturbing the space-time, and there are some things that are surprises, e.g. the statistics of the black holes' sizes.
When a historian calls ancient Greece a laboratory, that is just saying it is a good place to look for useful analogies. What makes a polity stable? How do different forms of government make decisions and how good are they? Are their common patterns in the starts of wars or peacemaking or in how tyrants attain power? Greece is particularly useful because there were so many city states and so much was written down and passed on to us. There have been a lot of tyrants since the classical era, and, as Mark Twain put it, history doesn't repeat itself, but it rhymes.
You can find examples of dozens of tyrants who failed in their first attempt to take power and who later succeeded. You can study how tyrants foment and exploit a crisis, their use of their dedicated following and the ways they can be stopped. Are you going to dismiss a warning that a tyrant may try to seize power again because you are unsatisfied with their dataset and definition of tyrant? You can't say that historical context isn't useful unless you absolutely hate spoilers.
One reason economics got trashed by so many after the Great Recession was that the government took a lot of advice from economists and that advice hasn't really delivered on its promises. You can argue that the government chose a bad batch of economists, but economic advice still seems stuck in the old rut. The 1930s, for example, led to major changes. Even now, the Federal Reserve is working hard to induce a serious recession in hope of eliminating jobs, cutting wages and worsening working conditions, all based on discredited economic theories. (Alternatively, it could be simple sadism, but I'll withhold judgement.)
The government never asks historians for advice, at least not in matters of policy. Politicians may cite different historians, but there is no Council of Historical Advisors or Office of Historic Analogy. As far as public interest goes, most people recognize that they can't influence which economists the government chooses to listen to, but they can choose their own historians. It's hard not to sense we are at an important point in history, when the decisions we make now will haunt us for quite a few years.
Personally, I'm glad historians have relatively little power. History shows us that history is frequently used for political ends. Wasn't there something in "1984" about this? There's a reason that political entities adopt founding myths and official chronicles. Those become history, and history justifies the present and guides the future. Asking for our public debate to pay more attention to the work of historians is not the same thing as placing them in the halls of power.
I'd rather see what I have been seeing, historians entering the public debate and being taken seriously. When faced with difficult political problems, a nation needs rethink it's history. This is as old as the Bible. Those "people's" histories in the 1970s, those New Deal histories in the 1930s, the glorification of the "Lost Cause" in the early 20th century, and even things like the 1619 Project were entries in debate about current politics and history. I'm glad that historians don't get to make the decisions, not that I'm always happy with what the politicians come up with.
Economists, in contrast, have been invited into the halls of power where they actually make political decisions. They write legislation. They structure markets. They set the course of the Federal Reserve. They participate in cabinet meetings. They do all these things in their capacity as economists. Naturally, I'd expect their expert advice to meet a higher standard than someone without such power.
This is a very thought provoking piece. I think it fails to appreciate the wide variety of historical genres that academics practice, but since it's basically aimed at a certain type of academic-turned-pundit, that's understandable.
"Theory" is a very broad term. Many theoretical works in "history" borrow theory from outside the discipline (Hegel, Marx, Weber, structuralists, post-structuralists, feminist theorists--are all examples of sources of historical theory that originate outside the field). To call any observation about recurrent historical regularities a "theory" isn't wrong, but it conflates very different types of historiography. What we usually call "theories" are about social and political regularities. The pundit-historian may be speaking about precedents, and like legal precedent, looking for historical precedents is a search for guides that apply to circumstances similar to, but rarely close to identical with past outcomes.
I wonder whether anyone who has studied the rise of the Third Reich did not have an uneasy sense of pattern resonance as the Proud Boys and Antifa battled in the streets, with the President at the time clearly speaking benignly of one and aggressively of the other. Nothing exact, but it seemed a good time to review the details we know about the past to look for insights about the present. That's something I did, but I'm not an historian of the Third Reich, and I would have welcomed having someone who was appear on some screen and convey where they may have seen common patterns, especially if they could also specify where the limits of that commonality lay. (All analogies have limited applicability, but there are useful analogies.) That might have made me realize how simplistic my own thinking was.
"History" isn't a discipline like economics or physics. Look at a textbook for a college 101 course in each and you can see the structural differences. "Would-be tyrants keep trying until they succeed" is not disproved by some would-be tyrants giving up or others never succeeding--and has nothing to do with "dictators" who do succeed sometimes giving up power. It has the force of saying that we know of enough historical instances of individuals aspiring to relatively absolute political power persisting beyond where ordinarily ambitious political figures tend to, that when someone who appears to fit that profile is defeated, it would be ignorant to assume they will not reassemble their power base and try again. Ignorant because it shows no awareness of precedent.
Like courts, it's possible and often useful to argue whether a precedent really applies, or whether there isn't one that applies more usefully; arguments are always selective, and partisans shape interpretation to serve their predisposition. But it can still be a productive enterprise.
During his life, Frederick Douglas, noted how in a free state, a man with an ox, supplied the energy needed to produce a product, while in a slave state, the same amount of energy took ten slaves.
He found it a poor use of labor and capital. That may not be an exact correlation to the idea that the Industrial Revolution could not have happened without slavery, so much as it slowed it down where slavery was practiced. This can also be seen by how quickly the North was able to rise to the occasion, with both troops, and manufactured supplies, something the South could not match.
In general, I agree with your questioning about using history to explain or predict the future.
I have one caveat: those who don't learn from history are doomed to repeat the same mistakes. I feel the problem comes from applying historical lessons in a one-to-one relationship to today's issues. This is as foolish as generals fighting today's wars with yesterday's tactics.
Applying Chamberlain's policy of appeasement with Hitler to today's situation in Ukraine is foolish and downright incorrect. However, as a principle, in the long run, appeasement of dictators never works.
I'm not sure how to test theories created by historians, even though they claim their ideas are not predictive. There is no way to set up an experiment and vary one input at a time to see how it will change the output. Even the great Hari Sheldon (Yes, I am a fan) wasn't correct all the time.
You have done a great job of explaining the problem; now, who will "Bell-the-Cat" to solve the problem?
The simplest step forward is to find as many examples as possible and then compare them all. Even if the comparison has subjective judgements it is still way more useful than just plucking out one or two situations from history and building a causal theory from them. A book like “why civil resistance works” looks through a huge number of examples to conclude that nonviolence is more effective than violence in overthrowing regimes for example
When Erik Loomis has this essay bought to his attention it would be wise for Noah to be prepared for a stream of invective, this concept (Historians bad, economist’s good and especially Historians could learn from economists is a concept almost designed in a lab to piss Erik off)
I just saw it and came here purely to get a screenshot of my comment to tweet to you and say ‘told you so’
The thing he always goes to with ‘humans aren’t like the third law of thermodynamics’ is the type of thing you expect stupid ppl who don’t know what science is to say, say what you will about Erik he’s not stupid which only says terrible things about the reason he uses arguments like that
Fairly heavy reading, thanks for going through all that. My reaction to the Sweet fiasco was to think about how both sides are trying to use the past now, ie history, to promote present political arguments. MAGA is essentially saying America was better in some imagined past era and we have to go back there, 1619 is saying America was far worse than we admit in some imagined past era, or narrative as NHJ puts it, and that justifies...something different now than what we are doing.
Personally I think they are both wrong and are simply using the past to justify their agenda's. The "past" has become a weapon in the culture wars and we are slowly losing any objective grip on what it actually was due to all this politicizing.
I remember your piece about how military coup attempts in Japan in the decades preceding WWII went largely unpunished and eventually succeeded in changing the course of the government there and took it as a nomothetic cautionary tale about treating too lightly events like Jan. 6. So perhaps nomothetics can be qualified as cautionary tales.
Also, the statement about tyrants can be improved by saying that they keep trying until they succeed (otherwise we'll never know them as other than tyrant wannabes) and, as in the case of Nixon, until they become former tyrants by whatever contingent events (including death) force them from power.
Finally, wasn't James Madison a nomothetic thinker who applied the cautionary lessons of history to the drafting of a constitution.
If historians think their field is only about recording past events, there's no reason to listen to their opinions about current events. They're effectively saying they have no more expertise on the matter than my hair stylist.
I think no historian regards the field as only about recording (or creating a narrative record of) past events. Discerning causal factors, institutional norms and their constraints, cultural perceptions and their influence--things like these are elements of just about every historical exercise.
When we ask questions about politics and culture today, people who have spent a lot of time and attention looking at patterns in political and cultural developments in the past have additional resources to identify possible pattern emerging now. The question is whether they are deploying those resources responsibly--good historical work takes a lot of time, effort, and intellectual reflection. I think Noah's best point is that historians who don't devote equal care to analysis of the present, but just note that points of past/present resemblance may constitute a recurring pattern, are better regarded as pundits than as academic authorities.
This is such a good article, I really appreciate how you took the time to explain the steps in your thought processes clearly.
Thanks!!
I think the right point to start at is Popper. What claims can be falsified? Many prefer history as a qualitative wrapper for explanation/rationalization of their priors. Others who prefer quantitative wrapping choose economics. But the qualitative-quantitative divide (which is really just about the handling of uncertainty) is not solved by breadth of knowledge or statistics. Both fields can form bad habits of rewarding those who put up the tallest pile of junk to intimidate those from investigating.
Popper’s positivistist approach offers a way to sort the good from the bad. Is it testable? If the answer is yes test it. If the answer is no don’t invest too much authority in the explanations you encounter. Taken to its extreme it can lead us to focus on testable questions (eg looking for your lost keys only in places where there is light) but we’re pretty far from that pole. Our current problem is unnecessarily complex explanations for where keys might be from strangers with no stake in the outcome.
The skill set we should focus on is answering the question “is it testable”? If they answer is yes, the proponent is as vulnerable to a grad student with too much time on their hands as they are to a giant of the field or an influencer. If it is not testable, assume it’s wrong, but be open to useful pieces. Not everything in the junkyard is trash.
Perhaps the field of history can function as an empirical discipline providing the facts about reality against which theories can be tested. In other words, rooting out the facts of the past that may otherwise be lost to the present seems like a noble enterprise in and of itself. The fact that theories may then be tested against those facts is just an added benefit.
Yes!!
“Testable” is good. “Falsifiable” is insisting on an unreachable standard. You would think that Newtonian gravity is a good falsifiable model, but the history of 19th century astronomy, where anomalies in Newtonian gravitation led to the prediction first of the planet Neptune, and then of the planet Vulcan, show that its more complicated.
Could you expand on this, Mr. Easwaran? I think the fact that there are anomalies that are later explained within the paradigm doesn't seem to me to attack Popper's standard of falsifiability. All that's required is the potential to stipulate some conceivable way in which a theory could be tested in practice (with conceivable technology) and be shown false, not that the test actually be available, or show the theory to be false, or that an initial false result itself never be overturned.
As I recall the history, it wasn't anomalies in Newtonian gravitation that led to Neptune's discovery, but anomalies in the orbit of Uranus, as Newtonian physics predicted it. Newtonian laws were deployed to hypothesize an undetected planet and search for it (and similarly with Pluto).
But physics is not exactly my field--my last course in physics was in high school, well over fifty years ago--so this post may just be my latest instance of bloviating public self-sabotage.
Around the end of the 18th century, the orbit of Uranus was found not to agree with the orbit predicted by Newtonian gravitation, given the known planets. This was judged not to be a falsification of Newtonian gravitation, because the anomalies could be explained if one postulated another planet, farther from the sun, which was called Neptune. Soon afterward, a planet of the relevant mass was found at the relevant location.
Around the middle of the 19th century, the orbit of Mercury was found not to agree with the orbit predicted by Newtonian gravitation, given the known planets. This was judged not to be a falsification of Newtonian gravitation, because the anomalies could be explained if one postulated another planet, closer to the sun, which was called Vulcan. Soon afterward, no planet was observed at the relevant location. However, that location is so close to the sun that astronomers thought it might just be unobservable.
Eventually, in 1916, it was discovered that the difference between Newtonian gravitation and Einsteinian gravitation was exactly what was needed to explain the anomalies in the orbit of Mercury. In fact, this is now said to be one of the big three observations that confirmed Einstein’s theory (the other two being the Michelson-Morley experiment in the 1880s, that had failed to identify the speed that the solar system was moving through the cosmos, and the observation of gravitational lending of light during the 1919 solar eclipse).
From a modern perspective, it is natural to claim that the anomalous orbit of Mercury was the sort of prediction that could be used to falsify Newtonian gravitation (or that the Michelson-Morley experiment is the crucial experiment that could falsify Maxwell’s non-relativistic electromagnetism), but at the time they were not perceived as such. This is because it is impossible to falsify a theory without holding fixed certain auxiliary assumptions (like the number of planets in the solar system, or the way the luminiferous ether can be moved) and anything that is taken to falsify the theory can instead be said to falsify the auxiliaries. It is only when a better theory is available, that explains the observations without any strange auxiliary assumptions (like an invisible planet closer to the sun than Mercury) that the observations are retroactively called a falsification.
The cases of Neptune and Mercury are very different. One was solved within the Newtonian paradigm, the other had to await the development of an alternative paradigm for solution, and the anomalies were, as you nicely phrase it, assumed to be a function of "auxiliaries."
But Popper's theory of falsifiability seems to me entirely consistent with this history. Basically, what it says vis a vis Newtonian gravity is that it is a scientific theory because evidence of objects in the world not conforming to its predictions counts as evidence that the theory is false. That doesn't mean evidence constitutes disproof, but that it contributes against confirmation and towards disproof, and can, in principle, ultimately overturn the theory.
A weak theory will be quickly falsified because evidence against it will far outweigh confirming evidence. Because the Newtonian theory is an extremely strong theory, evidence will overwhelmingly tend to confirm it, but, to the degree it is imperfect, some evidence will confute it. As Kuhn's sociological model predicts (based on just the type of evidence you've laid out), as disconfirming evidence builds for a theory so strong as to underlie a scientific paradigm, it will at first be treated as the anomalous results of imperfect tests or imperfectly understood tests, but will gradually either undermine absolute confidence in the strong theory or be recognized as better satisfying an alternative theory.
That's all consistent with falsifiability. If Newton's theory weren't falsifiable, it wouldn't have been falsified--just as the theory that God created the universe cannot be falsified, and is therefore not a scientific theory, though "in theory" it may still be a true theory. (My claim that I can't recall what I dreamt last night is not scientific--it's not falsifiable, but it's still a true theory.)
I just posted a reply to Noah right above this (in my display), and then found much of what I said at length better said with brevity here.
More like this Noah
This was an incredibly thought provoking, and at times uncomfortably piercing, piece about what historians can usefully add to the public discourse. As a historian, I definitely felt smug in my assumptions that in the battle between the social sciences and history, the context driven process of studying history was far superior. This was an important corrective so thank you!
Whether history is a ‘relevant’ topic is a very vexing one for historians, particularly given lower uptake of the subject at universities. It can lead historians down a dangerous path where we overextend it’s ability to reveal things about the world around us.
One distinction that I might venture is that the commentary that you cite here is one of ‘historical analogy’ - a deeply assumption driven and (in my opinion) dangerous exercise - with contextualisation. The latter is perhaps a more useful contribution for the likes of me to make!
In my experience, a strong grasp of the context within which a situation is unfolding (which includes deep historic trends and traditions), can be used effectively to inform when a predictive model is likely to have blind spots (as all must do, no matter how well evidenced and thought through). This is certainly not the sole prerogative of historians - lots of academic fields can add value there in a variety of aspects.
Sadly for historians looking to get into the political commentary game, that both requires knowledge of the models themselves (so we’ll have to rely on you darstardly economists), and specific knowledge of the context at hand. So built in to a historian’s expertise is their specific knowledge, an unhelpful factor for poor TV producers and editors looking for a reliable source of opinion that they can use for multiple events!
If social science wasn't such trash right now, I'd say that the main role for history w.r.t. prediction would be to provide the variables into their models.
Any direct predictions based on historical parallels need to be taken with a whole heaping barrel of salt.
And also stop overusing Anacyclosis! =)
Dear Noah Smith, excellent blog post with very high academic quality. First on Devereaux would be tyrants, the experience of Pakistan is that even democratic leaders have tried to cling to power. For example, the exiled leader Nawaz Sharif contested elections for the position of the prime minister 5 times since 1988 and won the majority seats 3 times. In 2017, though he was disqualified by the highest court on some say flimsy charges, he is yet looking for another 5 year term for himself. Same is true with Imran Khan. Most of the Pakistani think that Turky's Erodogan that is viewed by Western commentators to be less democratic, is the ideal case whereby one can cling to power for decades under democracies. Similarly the benevolent dictator Musharraf stayed in power from 1999 to 2008 and he wanted to continue his rule as is told by his acolytes based on some good economic growth years under his rule in Pakistan that he is indispensable for Pakistan. How would you translate his leaving of the office of president in 2008 not voluntarily but after years of struggle and protests by democratic forces within the country. Secondly the free media that he helped in creating by allowing private news channels was one of the most significant factor that informed people of Pakistan adequately for the country to have a democratic movements against his dictatorial rule.
Secondly you have again brought a very important empirical argument of causation and correlation in case of slave trade and earlier industrialization in the US. You are absolutely right that industrialization happened due to enterprising nature of American society that is raising the voice against slavery as early as 18nth century. So there is more of a correlation between slave trade and industrialization. And if anything industrialization has helped the American society to work on strengthening democracy in US whereas democracy has come with greater rights for African Americans and women and thus one may suggest that industrialization has actually lead to the abolishment of slave trade in America.
I have my BA in history, which I thankfully got prior to various forms of Theory completely capturing the discipline. I agree with the post generally, in that those in the business of making predictions ought to be subject to empirical evaluation. However I would say the assertion is less about history and more about punditry, which is actually what these people are doing. What's really going on here isn't an investigation of the past, but a statement about the present, and therefore not really history at all.
And anyway most of the kinds of claims in question tend to collapse, or at least become complicated, based on the historical record itself. A good historian that wants to engage in punditry will understand and at least attempt to account for the kinds of confounding facts and factors that crop up, frankly, everywhere. But where would the fun be in that?
Noah, while I find many parts of the piece thoughtful and useful to consider I must object to the characterization of Devereaux's piece. I believe that the quotation selected is misrepresentative to the level of intellectual dishonesty and the failure to properly analyze the piece serves to greatly undermine your argument. While you correctly quote from his introduction to a discussion on Peisistratos, you have omitted the conclusion in which he adds a qualifier to the prior "explicit law of human society," namely that Devereaux identifies that "Unless would-be tyrants *are made to face the consequences of their attempts to seize power*, they will keep trying until they succeed so thoroughly that justice is beyond recovery" (emphasis added). This not only fundamentally changes the law he is discussing, it makes the proposed counter-example of Richard Nixon no longer a counter-example: Nixon faced the most serious consequences of any president before or since and would almost certainly have faced criminal charges if not for Ford's pardon. If we are to critique bad theories by historians, we owe it to them to at least correctly identify these theories.
I also critique the link provided for the "historical list of dictators who stepped down." On said list are a mix of figures like Daniel Ortega and Indira Gandhi, who after losing an election sought to regain power through a more central regime (and in the case of Ortega, succeeded in doing so) and thus would be proof that would-be tyrants will keep trying. But also are figures like George Washington and Cincinnatus, who it would be quite a strain to argue qualify under the banner of "would-be tyrants". Again, we ought to critique bad theories by historians but not with shoddy scholarship that isn't really addressing the core claims.
Watching you tight with devereaux hurts my heart.
Also, it feels bad. Economics still has foundations and funding and well paid positions in companies.
Historians, as you pointed out earlier, are disappearing. Are you going to spend the 5 years to learn classical Latin, Turkish, Greek, Manchu etc. necessary to do the history?
Being in a dying field doesn't (and shouldn't) protect them from criticism when they insert their field into political disputes.
Also, history will survive, barring a civilizational collapse. It is kind of it's thing.
History doesn't survive. That's one of the first things historians learn.
We have many Greek and Latin texts that refer to "the history of x written by z" that are lost to time.
Also, he chose a poor example going after devereaux who is super good about talking about how contingent everything is. His tyranny bit is worth reading and he caveats it by saying this is based on Greek ancient cities.
I would consider the end of antiquity to be a 'civilizational collapse'...
We aren't going to lose our history just by a reduction in the number of new professional historians being minted.
Even if the numbers were reduced to 1/10th of the current baseline, our histories would survive and continue.
That is super naive. We only have 25% of the films made before 1930.
History gets discarded if no one keeps it.
You make history sound a lot like chemistry. Chemistry has much too large a search space for certainty. You don't know how a reaction is going to run until you dump everything together. When someone wants to synthesize some chemical or break down another, chemists reach into their bag of analogies. Sometimes they apply. Sometimes they don't. There are all sorts of heuristics one can apply as well. Chemists spend a lot of their time reading papers old and new in search of useful analogies. That doesn't make chemistry any less a science.
A big difference is that a chemist can try out a lot of things in a relatively short time. Astronomers and historians don't have this advantage. When they want to predict things, they have to look backwards for analogies, suggest things, and then, like the rest of us, wait to see what happens. For example, consider all the predictions that were made about what the Webb telescope or the LIGO gravitational wave detector would reveal. Just as with the outcomes of political events, there are some things one expects, e.g. black holes colliding and disturbing the space-time, and there are some things that are surprises, e.g. the statistics of the black holes' sizes.
When a historian calls ancient Greece a laboratory, that is just saying it is a good place to look for useful analogies. What makes a polity stable? How do different forms of government make decisions and how good are they? Are their common patterns in the starts of wars or peacemaking or in how tyrants attain power? Greece is particularly useful because there were so many city states and so much was written down and passed on to us. There have been a lot of tyrants since the classical era, and, as Mark Twain put it, history doesn't repeat itself, but it rhymes.
You can find examples of dozens of tyrants who failed in their first attempt to take power and who later succeeded. You can study how tyrants foment and exploit a crisis, their use of their dedicated following and the ways they can be stopped. Are you going to dismiss a warning that a tyrant may try to seize power again because you are unsatisfied with their dataset and definition of tyrant? You can't say that historical context isn't useful unless you absolutely hate spoilers.
One reason economics got trashed by so many after the Great Recession was that the government took a lot of advice from economists and that advice hasn't really delivered on its promises. You can argue that the government chose a bad batch of economists, but economic advice still seems stuck in the old rut. The 1930s, for example, led to major changes. Even now, the Federal Reserve is working hard to induce a serious recession in hope of eliminating jobs, cutting wages and worsening working conditions, all based on discredited economic theories. (Alternatively, it could be simple sadism, but I'll withhold judgement.)
The government never asks historians for advice, at least not in matters of policy. Politicians may cite different historians, but there is no Council of Historical Advisors or Office of Historic Analogy. As far as public interest goes, most people recognize that they can't influence which economists the government chooses to listen to, but they can choose their own historians. It's hard not to sense we are at an important point in history, when the decisions we make now will haunt us for quite a few years.
One can’t simultaneously ask for more power and demand to be held to lower evidentiary standards than other social sciences.
I'm definitely not doing that.
Personally, I'm glad historians have relatively little power. History shows us that history is frequently used for political ends. Wasn't there something in "1984" about this? There's a reason that political entities adopt founding myths and official chronicles. Those become history, and history justifies the present and guides the future. Asking for our public debate to pay more attention to the work of historians is not the same thing as placing them in the halls of power.
I'd rather see what I have been seeing, historians entering the public debate and being taken seriously. When faced with difficult political problems, a nation needs rethink it's history. This is as old as the Bible. Those "people's" histories in the 1970s, those New Deal histories in the 1930s, the glorification of the "Lost Cause" in the early 20th century, and even things like the 1619 Project were entries in debate about current politics and history. I'm glad that historians don't get to make the decisions, not that I'm always happy with what the politicians come up with.
Economists, in contrast, have been invited into the halls of power where they actually make political decisions. They write legislation. They structure markets. They set the course of the Federal Reserve. They participate in cabinet meetings. They do all these things in their capacity as economists. Naturally, I'd expect their expert advice to meet a higher standard than someone without such power.
Interesting. I wrote more in an article.
This is a very thought provoking piece. I think it fails to appreciate the wide variety of historical genres that academics practice, but since it's basically aimed at a certain type of academic-turned-pundit, that's understandable.
"Theory" is a very broad term. Many theoretical works in "history" borrow theory from outside the discipline (Hegel, Marx, Weber, structuralists, post-structuralists, feminist theorists--are all examples of sources of historical theory that originate outside the field). To call any observation about recurrent historical regularities a "theory" isn't wrong, but it conflates very different types of historiography. What we usually call "theories" are about social and political regularities. The pundit-historian may be speaking about precedents, and like legal precedent, looking for historical precedents is a search for guides that apply to circumstances similar to, but rarely close to identical with past outcomes.
I wonder whether anyone who has studied the rise of the Third Reich did not have an uneasy sense of pattern resonance as the Proud Boys and Antifa battled in the streets, with the President at the time clearly speaking benignly of one and aggressively of the other. Nothing exact, but it seemed a good time to review the details we know about the past to look for insights about the present. That's something I did, but I'm not an historian of the Third Reich, and I would have welcomed having someone who was appear on some screen and convey where they may have seen common patterns, especially if they could also specify where the limits of that commonality lay. (All analogies have limited applicability, but there are useful analogies.) That might have made me realize how simplistic my own thinking was.
"History" isn't a discipline like economics or physics. Look at a textbook for a college 101 course in each and you can see the structural differences. "Would-be tyrants keep trying until they succeed" is not disproved by some would-be tyrants giving up or others never succeeding--and has nothing to do with "dictators" who do succeed sometimes giving up power. It has the force of saying that we know of enough historical instances of individuals aspiring to relatively absolute political power persisting beyond where ordinarily ambitious political figures tend to, that when someone who appears to fit that profile is defeated, it would be ignorant to assume they will not reassemble their power base and try again. Ignorant because it shows no awareness of precedent.
Like courts, it's possible and often useful to argue whether a precedent really applies, or whether there isn't one that applies more usefully; arguments are always selective, and partisans shape interpretation to serve their predisposition. But it can still be a productive enterprise.
During his life, Frederick Douglas, noted how in a free state, a man with an ox, supplied the energy needed to produce a product, while in a slave state, the same amount of energy took ten slaves.
He found it a poor use of labor and capital. That may not be an exact correlation to the idea that the Industrial Revolution could not have happened without slavery, so much as it slowed it down where slavery was practiced. This can also be seen by how quickly the North was able to rise to the occasion, with both troops, and manufactured supplies, something the South could not match.
In general, I agree with your questioning about using history to explain or predict the future.
I have one caveat: those who don't learn from history are doomed to repeat the same mistakes. I feel the problem comes from applying historical lessons in a one-to-one relationship to today's issues. This is as foolish as generals fighting today's wars with yesterday's tactics.
Applying Chamberlain's policy of appeasement with Hitler to today's situation in Ukraine is foolish and downright incorrect. However, as a principle, in the long run, appeasement of dictators never works.
I'm not sure how to test theories created by historians, even though they claim their ideas are not predictive. There is no way to set up an experiment and vary one input at a time to see how it will change the output. Even the great Hari Sheldon (Yes, I am a fan) wasn't correct all the time.
You have done a great job of explaining the problem; now, who will "Bell-the-Cat" to solve the problem?
The simplest step forward is to find as many examples as possible and then compare them all. Even if the comparison has subjective judgements it is still way more useful than just plucking out one or two situations from history and building a causal theory from them. A book like “why civil resistance works” looks through a huge number of examples to conclude that nonviolence is more effective than violence in overthrowing regimes for example
When Erik Loomis has this essay bought to his attention it would be wise for Noah to be prepared for a stream of invective, this concept (Historians bad, economist’s good and especially Historians could learn from economists is a concept almost designed in a lab to piss Erik off)
Oh, Loomis already went apeshit. He said empirical economics is trash because...and I kid you not...Pinochet.
https://twitter.com/Noahpinion/status/1563561812422889473
I just saw it and came here purely to get a screenshot of my comment to tweet to you and say ‘told you so’
The thing he always goes to with ‘humans aren’t like the third law of thermodynamics’ is the type of thing you expect stupid ppl who don’t know what science is to say, say what you will about Erik he’s not stupid which only says terrible things about the reason he uses arguments like that
Fairly heavy reading, thanks for going through all that. My reaction to the Sweet fiasco was to think about how both sides are trying to use the past now, ie history, to promote present political arguments. MAGA is essentially saying America was better in some imagined past era and we have to go back there, 1619 is saying America was far worse than we admit in some imagined past era, or narrative as NHJ puts it, and that justifies...something different now than what we are doing.
Personally I think they are both wrong and are simply using the past to justify their agenda's. The "past" has become a weapon in the culture wars and we are slowly losing any objective grip on what it actually was due to all this politicizing.
I remember your piece about how military coup attempts in Japan in the decades preceding WWII went largely unpunished and eventually succeeded in changing the course of the government there and took it as a nomothetic cautionary tale about treating too lightly events like Jan. 6. So perhaps nomothetics can be qualified as cautionary tales.
Also, the statement about tyrants can be improved by saying that they keep trying until they succeed (otherwise we'll never know them as other than tyrant wannabes) and, as in the case of Nixon, until they become former tyrants by whatever contingent events (including death) force them from power.
Finally, wasn't James Madison a nomothetic thinker who applied the cautionary lessons of history to the drafting of a constitution.
Fortunately, no one thought that my Japan coup piece was backed by rigorous academic research... ;-)