Why do civilizations collapse? This question bears not only on safeguarding our society’s future but also makes sense of our present. The answer relies on some of the same technē that humanity needed to build civilization in the first place: we have to evaluate the perceptions that mint facts and theory, not merely peruse the body of theories handed down to us.

Institutional failure comes as a surprise because organizations try to hide their shortcomings. They lean on other, more functional organizations in order to keep up appearances. During civilizational collapse, no organization can properly hide its own inadequacy, since the whole interdependent ecosystem of institutions is caving in on itself. States, religions, material technologies, and ways of life that once seemed self-sustaining turn out to have been dependent on the invisible subsidy of just a few key institutions. The environment of societal collapse reveals much of the otherwise obscured inner workings of crucial social technologies. After all, to analyze something is to break it apart!

Despite being an excellent epistemic opportunity, civilizational collapse seldom inspires introspection among thinkers living through it. Mayan or Roman thinkers don’t seem to have reflected on their ongoing collapse. As institutions turn to cannibalizing each other, there is little patronage or emotional energy going towards accurately describing the wider process. The notable exception that proves the rule of civilizational delusion is the Zhou Dynasty of ancient China. It is an encouraging example, since it shows a societal failure arrested and reversed by an intellectual golden age called the Hundred Schools of Thought. Confucianism, Legalism, and Taoism could only come into being with this kind of epistemic opportunity.

In the West today, we operate under the influence of our own key philosophy, which we can call scientism: the tendency to rely on scientific claims to describe the functioning of society, even when there is no empirical reason to assume that they apply. We act as if we are already living in a scientifically-planned society, immune to collapse on a time scale that any of us have to worry about. This is very far from the truth. We are certainly living in socially-engineered societies, but they are not scientifically planned in any straightforward way. Our organs of economic management do not secretly know how the economy really works. Our systems of political regulation are operating on the fumes of their institutional inheritance from two or three generations ago—the last spurt of institutional growth in Western societies happened roughly during the 1970s. At this time in the United States, new federal bodies such as the Department of Energy and Education were created and organizations such as NASA reached their modern form. Concurrently, the United Kingdom dispensed with organized labor as a political force in favor of an expanded administrative apparatus, and France saw the resignation of Charles de Gaulle, the architect of the Fifth Republic; neither country’s political economy has evolved much since.

Civilizational collapse always looms on the horizon. Though we usually think of collapse as a slow process, it can in fact happen very quickly, as was the case with the Late Bronze Age collapse. The old dictum “gradually, then suddenly” is cliché, but accurate. To ascertain whether or not we are headed for collapse, we must first analyze the functionality of our own society and pinpoint where things go wrong.

Mechanisms of Collapse

Our society is dominated by large bureaucracies. These bureaucracies break down the processing of physical goods and information into discrete tasks, such as how a factory worker puts doors on a car, or a stock trader buys futures contracts. These tasks are shorn of their context and executed in a systematized environment whose constraints are quite narrow: put the car door in, increase the portfolio value. Our society is thoroughly compartmentalized. This compartmentalization isn’t driven by the division of labor, but rather by the need to make use of misaligned talent without empowering it. By radically limiting employees’ scope of action, you make office politics more predictable. By fragmenting available knowledge, you can leverage information asymmetries to the intellectual or material advantage of the center. Some of this is necessary for scaling organizations beyond what socially connected networks can manage—but move too far towards compartmentalization, and it becomes impossible to accomplish the original mission of the organization.

Such large bureaucratic systems do not emerge organically; they require design and implementation. Empirically, we can know this simply by examining the intent of the original founders of these systems. If you want to know, say, why the FBI exists, you can find the answer in the documents of its founder, J. Edgar Hoover.  You could do the same for the IRS, or for Amazon, or for any other number of institutions.

It is very difficult, though, to apply this analysis to the construction of society. No matter how large or how small, institutions always coexist in a symbiotic relationship with other institutions. There is no Amazon without the United States government, no U.S. government without—at least—some parts of the U.S. economy. Each of these institutions depends on the others in an intricate mesh. Society is not a single institution, after all, but an ecosystem of interdependent institutions.

In addition to this complexity, non-functional institutions are the rule. Our institutions today rarely function in accordance with their stated purpose. Individuals within a given society are often very bad at judging institutional functionality. Some people spend their entire lives ruthlessly profiting from the misery of others, or greatly contributing to the prosperity of others, without even knowing that they are doing so. People who try to effect change are most often frustrated. Countless people spend their lives wrestling with a societal problem, slaving over papers for publication in academia or the nonprofit world. They act as if there is some sort of metaphorical wall which they throw their papers over, with some responsible person on the other side taking the output of their disinterested scientific study and translating it into policy, medical practice, or industrial production.

More often than not, there is nobody on the other side of that wall. Since society is so deeply compartmentalized, it rarely functions as a whole with a single purpose. Note that dysfunctionality is not a normative distinction; it often boils down to the simple reality of whether or not anyone ever follows up on key actions within the institution. It is also a question of whether or not there is a multiplier—be it individual, bureaucratic, oligarchic—behind that metaphorical wall.

Institutions often become non-functional due to the loss of key knowledge at critical junctures. Take, for example, the recent failure of the National Nuclear Security Administration (NNSA) to reproduce a niche classified material known as FOGBANK that is necessary for manufacturing nuclear weapons. It took the NNSA ten years and millions of dollars to re-engineer a material that their staff in the 1980s knew how to make. That knowledge never should have been lost in the first place, but in a dysfunctional society, such loss of knowledge becomes the rule. Attempts at reverse engineering do not always succeed, if they are even made.

Civilizational collapse, then, looks like this dynamic at the scale of an entire civilization: a low-grade but constant loss of capabilities and knowledge throughout the most critical parts of our institutions, that eventually degrades our ability to perpetuate society. There might be a sudden point where the superstructure gives way dramatically, such as occurred during the Bronze Age Collapse, or there might be slow accommodation to this convergence to zero, as with the Byzantine Empire.

The key dynamic here is the loss of the subtle social technologies that allow us to solve the succession problem. Running a large and complex institution requires skills which are often difficult to fully pass on. How can a successful founder ensure a successor who leads as competently as they did? The succession problem is the central obstacle to transferring the ownership and knowledge of institutions from generation to generation. In the case of the Nobel Committee, for example, the goal of succession is to produce a new chairman with similar faculties of judgment to the original chairman. In the case of ancient Egypt, it would have been making sure that the Pharaoh’s son knows how to interact with all of the powerful people in Egypt and has an intuitive feel for the subtleties of public order, diplomacy, and famine prevention. In addition to knowledge succession, there is also power succession. The son of the Pharaoh may be just as skilled as his father, but if he does not inherit his base of power, the son will be vulnerable to usurpation or invasion.

The succession problem is especially important when transferring secrets. In ancient Egypt, accurate measurement of the Nile river was a state secret, in order to allow the state to monopolize agricultural production and resource flows. This was crucial to the functionality of Egyptian civilization—it was the legitimating story of the state. By design, it was not clear to the Egyptian public how they would go about running their society if it weren’t for the expert knowledge of the state.

The failure to maintain implicit traditions of knowledge speaks to the extreme difficulty of transferring secrets between generations. Often the problem is that the kids “don't get the joke”: if you create an institution with a false premise in order to mislead society as to your true goals, the people you hire into it might be fooled by the propaganda themselves. This is why claims about multigenerational conspiracies are always highly suspect: such organizations are plagued by succession failures in knowledge.

Avoiding collapse is so difficult because succession failure is often opaque. If the Institute of Pottery lost the ability to make good pots—to mold people into skilled pot makers—would they declare it to the world? Of course not—institutions are very rarely self-abolishing. The same holds true throughout crucial niches of our society, from social engineering to science and philosophy. All of these areas could be in profound crisis today, and we wouldn’t even know it. The intellectual apocalypse is invisible if there are no true intellectuals around. Again, institutional failure typically comes as a surprise.

And yet, clearly some functional institutions still exist, or our society would not function at all. At the end of the day, you can still go online and call a cab or go to a dealership and buy a car. This car will have doors bolted on by a worker you’ve never met, and these doors will seem to work. The same cannot always be said of stock portfolios.

History Guides Analysis of Decline

We can define civilizational collapse as a process wherein most recognizable large-scale institutions of a society vanish, coupled with a drop in material wealth, a drop in the complexity of material artifacts and social forms, a reduction in travel distance and physical safety of the inhabitants, and a mass reduction in knowledge.

Loss of knowledge is especially damaging, since it accelerates the other aspects of collapse and ensures that they will be long-lasting. Nearly all of the written evidence we have of societal decline comes from elites. Historically, literacy was restricted to the traditional elite class of a society, as they were the only ones with any use for reading and writing. This accounts for the total disappearance of writing after the Late Bronze Age collapse, since Bronze Age societies had a very small literate class. The result was a wholesale loss of civilizational knowledge. When writing reappeared in the eastern Mediterranean centuries later, it was based on the new Phoenician alphabet, rather than the old hieroglyphic system that gave birth to the cuneiform of the Assyrians or the Linear B of the Minoans. Such losses of knowledge are a constant throughout human history: as with FOGBANK, or as with the state of New Jersey recently scrambling to find a COBOL programmer with the ability to overhaul their legacy information systems.

Despite how difficult it can be to gather historical data, it’s still a far better way to understand societal collapse than purely theoretical models. Rather than picking and choosing our preferred explanations of collapse beforehand, we should first recognize that there are simply too many causal variables to control for. The best we can hope for is rigorous cross-comparison with the historical record, using sets of natural experiments between past societies. A broad historical literature of collapse does exist, especially on the Late Bronze Age collapse and the fall of the Roman Empire. But the scholars that pose these questions often have particular—and popular—answers in mind as to what causes collapse: environmental fragility, moral decline, an overloading of systemic complexity, and so on. The morality play is written first, the facts are found second, and this often results in a shoddy final product of a theory. Thus, the relevance of history for investigating our own society’s potential collapse is also obvious: without comparing the present to other civilizations, we can’t say much of anything useful about it.

It is hard to come to a consensus on historical cause and effect. In geology, we didn’t build another planet to discover the Earth’s plate tectonics, but rather dug among the rocks on which we found ourselves. In our macro-study of history and civilizations, we too must rely on in-depth exploration of historical examples.

That exploration is still itself theory-driven. Good historians and theoreticians explicitly acknowledge the theses they work with, so I will do the same. My theory of history is great founder theory: I propose that social technologies do not evolve out of mass action, but rather are devised by a tiny subset of institutional designers. Looking at history, we see that new organizations and social forms often arise within a single generation, showing jumps in social complexity far too rapid to be explained away by collective action or evolution. This would be the equivalent of expecting a tornado tearing through a junkyard to assemble a Boeing 747 or a Tesla Cybertruck.

Designing complex objects through collective action, or perhaps through an intermittent individual strategy similar to the open software approach, is tempting. However, unowned commons tend to be raided, and individual visions tend to differ massively. It often takes an exceptional individual with exceptional vision to create a new social or material technology. It’s hard to remember nowadays that the smartphone once had to be devised as a combination of the cell phone, the tablet, and the camera, and did not merely emerge out of mass market sentiments. It took a single individual, Steve Jobs, to see that while a combination of the car, the airplane, and the submarine would produce an inferior version of all three, the opposite case would be true in the creation of the smartphone. And then that individual had to implement the vision.

The result is usually one or more institutions, created by the individual to carry out their goals. Institutions are not naturally self-documenting. The descriptions of themselves that they provide can be misleading. Suppose you were watching the birth of a mystical movement like the Franciscans in the 13th century. At the time, you might describe them as the cult of a new god. But an observer in the 15th century would, according to the institutional information available at the time, describe them as a movement firmly within the Catholic Church. In theory, the Franciscans have always been good Catholics, but this only gets recognized after a struggle has played out. In the Middle Ages, you could often believe any heresy you wanted as long as you formally declared your loyalty to the Pope.

A different example might be the United States government today. A keen observer would examine the way that laws are made today and conclude that we have witnessed the emergence of a new legislative body all but in name, with Congress reduced to a vestigial organ of this governing structure. Law today is made mostly by the Supreme Court, or the civil service when it chooses what to implement and how, or occasionally via Presidential executive order. Yet very few people today come to such a conclusion, as the ideology of American government dictates that law is made in Congress, and does not make room for the development of new federal legislative bodies. If no one believes a hypothesis, the evidence for it remains unnoticed, even when such evidence is abundant.

Such thinking requires going beyond both public appearances and official narratives. The popular or expert definition of “law” that we use in 2020 can’t help us here. Rather, we should use the term “law” in the same sense that one would use it to empirically describe the formalized customs of medieval Iceland, or of the function of the Twelve Tables in ancient Rome, or Lycurgus’ Laws in Sparta, or Sharia law in modern-day Somalia. In none of these cases would the official self-documentation of the institutions give you an accurate picture of real conditions. Institutions are similar to individuals in this way. Deep theory-building enabled by thorough scholarship or, better yet, high quality anthropological fieldwork would be needed.

Something as seemingly objective as the occurrence of a battle can only be inferred from scattered artifacts that other people have found, or the writings of long-dead strangers. Moreover, battles are relatively low-complexity—can you imagine trying to parse through Obama’s emails to figure out what his key agenda was during the course of his administration? Could you even do this with your own emails from last year? Establishing historical knowledge is difficult. Narrative fills the gaps; stories are told both by you and Obama and FDR, and by Julius Caesar. These are always a mix of accuracy and self-interest, which is, in fact, what history is.

Material Signs of Collapse

Material evidence can provide something closer to objectivity—at least sometimes. The archaeological record shows that many large Roman cities were depopulated over the course of the 4th and 5th centuries AD. As indicators go, this is a fairly clear and obvious sign of a high urban society’s decay. Indirect evidence corroborates this, such as the reduction in atmospheric lead pollution generated by Roman mining activity leading up to the collapse of the empire. If we assume that mining activity is related to economic production, this is a good indicator of economic decline.

But even material evidence can be unreliable, since understanding it requires a high degree of contextual knowledge. The interesting question for the prospective collapse of our own society is this: if you were a late imperial Roman, and someone told you about the ongoing decline in atmospheric lead, how would you process this information? Today, if we saw a drop in lead pollution, our first assumption might be that this is due to the advent of greener technology. Economic decline wouldn’t naturally come to mind. Victory has many fathers, but defeat is an orphan. We find it hard to believe that we were once more capable of intentionally affecting the world than today. After the collapse of the Roman Empire, wanderers among the ruins of aqueducts concluded that they must have been built by giants. The classical Greeks examined the massive stone ruins of Mycenaean civilization and assumed that the great walls were built by a race of Cyclopes.

Our persistent failure to understand the monumental achievements of the past speaks not to the historical prevalence of pyramid-loving extraterrestrials, but rather to the fact that we often lose the knowledge of the social technology on which such material artifacts rested. It is easier for us to conceptualize an extraterrestrial force that constructed the pyramids than it is to conceive of the political and economic system that made such architectural feats possible. Such loss of knowledge gives us an idea of our limited ability to maintain advanced social systems over generations.

If we compare the roughly twelve identifiable Dark Ages following civilizational collapse on the Eurasian continent—the collapse of the Bronze Age civilizations, the end of Mohenjo Daro, the decline of the Roman Empire, Han China and so on—we always find that nearly all material technology is not self-perpetuating, but rather rests on foundations of social technology. The only material technologies that routinely survive collapse are small-scale agriculture and small-scale metallurgy, likely because the social technologies needed to sustain such smaller communities can arise organically.

Since collapse in material technology is always preceded by collapse in the practice of social technology, Dark Ages are always preceded by Intellectual Dark Ages. Knowledge of these social technologies is highly compartmentalized and, as a result, they are not understood explicitly by all parts of society. This means that a society undergoing an Intellectual Dark Age doesn’t realize it is going through one at all—all the people who would notice are long-gone, and those who remain are miseducated, role-playing the forms left behind by their predecessors without realizing that they’ve lost the substance. Often not just the knowledge, but the socioeconomic niche that once fostered the creation of new social technology has been obliterated in all but name.

Today, our material technology is far superior to that of Rome, but our social technology may not be. Take the Industrial Revolution for example: surely the most interesting thing that has happened within the last 500 years, and a process that most currently assume is still ongoing. But if the Industrial Revolution was over, what would we expect to see? Much as we see a late Roman drop in lead pollution, today we see drops in pollution in the West. The standard explanation is gains in efficiency and greener technology. But if we take a more global perspective, it seems that we outsourced not just production, but also the pollution associated with production to China. The economists’ argument here is that we have intentionally outsourced our industries to China, obeying the industry-agnostic logic of gains from trade. It is worth considering the economists might be wrong if the promised gains from trade haven’t materialized.

One could hypothesize the American worker and manager have, over time, lost the social technology that enabled them to run the assembly lines in the first place and that, now, our support for outsourcing isn’t so much due to greed as it is an adaptation to inability. Europe, arguably, has adapted to its inability to fight wars anymore with a narrative of intentional pacifism. We should seriously consider the possibility that we are a post-industrial society not in a positive sense, but in the sense that in our society the Industrial Revolution has stopped.

Such a hypothesis is strikingly hard to defeat. A civilizational collapse under conditions of advanced material technology might look very much like what we have now. Our society is the product of what were once advanced, rational, self-catalyzing systems of production, but we have now reverted to a more customary system, where things are simply done as they were 40 or 50 years ago. We have the same bureaucratic and economic institutions as we did then, with some marginal tweaks. Thanks to narrow progress in the CPU industry, most of which has left the United States, we are now able to have Zoom calls. Unfortunately, there are few other reasons for optimism.

Collapse Is Silent

When collapses occur more slowly, it is even more difficult to find anyone acknowledging the process at all. This was true of the late Roman Empire, where one finds letters exchanged between patricians complaining that the roads were often unsafe this time of year, but little acknowledgement of the fundamental changes taking place. The collapse of the Roman Empire was much less the constant burning of cities so much as it was GDP-equivalent shrinking by about 1% per year, while remaining more or less the same on the books, for two hundred years in a row.

In the context of adaptation to the COVID-19 pandemic, markets seem to be fairly stable and even slightly better than they were a few months ago, even though common sense tells us that production has fallen massively. If such a huge drop in economic activity can be papered over with government and private sector intervention, can we imagine how many crucial slower-moving changes are going unnoticed? If our actual wealth per capita, say, has been declining 1% per year for the last 20 years, how would we even know? Civilizational collapse happens on similar or even slower time scales—even though coronavirus hasn’t left everything in flames, we may still be on the long road to collapse. Absent reform, I think we are in for a slow century of decline starting in about 2030 or so.

Intentional and, more importantly, successful reform of society is very rare. It does happen, but macro-scale social engineering is exceptionally difficult: Augustus Caesar truly did save the Roman Republic from tearing itself to shreds through unsustainable warfare. His imperial system was in turn torn to shreds through warfare after long-standing economic and intellectual decay 300 years later. The sheer difficulty of reform, coupled with the accumulation of social and cultural technical debt, provides a fairly solid explanation for why civilizations collapse.

The United States is well-positioned to attempt such civilizational reforms, since it has a remarkable ability to integrate exceptional talent from all over the world and has put that talent to work on some of the most successful institutional projects in history, including the Manhattan Project and the Apollo Program. America is, for now, in an unavoidable period of relative decline, and in 2030 or 2040 the largest economy in the world will almost certainly be that of China. But absolute decline is reversible—2060 is still an open question. A deep pragmatism runs through this country, and if reimagined, the 21st century could see another explosion in American economic, social, and cultural development.

The solution lies with a small number of people who can independently judge the generative minds behind the facts, rather than merely minding the integrity of the established body of theories and observations. If there is such a thing as a technē of civilization—the skill of managing the institutions of society and culture—it exists in very narrow corners of society. Engineering society to be self-perpetuating is an extremely difficult challenge, and we can devise all sorts of machinery to do so, but this is the bottom line. Such people are extremely rare, but if we create a socioeconomic niche for them, our civilization can rewrite its own future for the better.