I. Complexity Has a Cost, and the Bill Comes Due
II. The Hidden Cost of Hierarchy
III. Fragility, Overshoot, and the Brittleness of the Optimised
IV. Elites, Extraction, and the Wealth Pump
V. The Psychology of Living Inside It
VI. Are We Already Inside It?
VII. Societal Collapse 101
VIII. What This Means
IX. Cross-Links
Why the structures we built to solve our problems eventually become the problem.
So far, we’ve traced entropy through cells, bodies, and minds. Now it scales all the way up to civilisations, because the same logic that governs a dissipative structure governs a society: an improbable order, held together against the pull toward disorder only by a continuous flow of energy, and destined to come apart when the energy can no longer maintain stasis. This page looks at how societies come apart, what the warning signs are, and why the comforting assumption that ours is exempt is exactly the assumption every previous collapsing society also made.
The analysis here is genuinely uncomfortable for anyone invested in the idea that large, centralised, hierarchical order is the natural and permanent state of things. Hierarchy and centralisation are expensive technologies, useful in some conditions and ruinous in others, and a look at collapse keeps surfacing an old and serious idea that the dominant culture treats as unthinkable: that much of the elaborate top-down structure we mistake for civilisation itself is, energetically and humanly, a bad deal that persists mainly because those it serves can enforce it.
Societies are problem-solving systems. Faced with a challenge, hunger, threat, coordination, and scale, they respond by adding complexity: new layers of administration, specialisation, infrastructure, hierarchy, and control. And complexity works, at first. The early returns are enormous: irrigation, writing, law, bureaucracy, and trade each pay back far more than they cost.
But complexity runs on energy, and it is subject to diminishing marginal returns. Each new layer solves a little less and costs a little more, until a society is spending staggering amounts of energy simply maintaining a complexity that no longer pays for itself, an ever-larger administrative and managerial apparatus producing ever-thinner benefit, every problem met by adding yet another layer that becomes next year’s burden. Eventually, the system is so loaded with the cost of its own complexity that it has nothing left in reserve, and then an ordinary shock, a drought, a war, a plague, a financial spasm, that a leaner society would have absorbed, topples it. In Tainter’s reading, collapse is a rapid simplification: a society shedding the complexity it can no longer afford, often experienced as catastrophe by those living through it, but understood structurally as a system returning to a sustainable level of order after overshooting.
You can feel the sense that everything has become more complicated, more layered with administration and compliance and process, while delivering less; that it takes more effort each year to get the same result; that the institutions meant to solve problems now mostly generate procedure. Some call it bureaucracy. Others call it inflation. Both feed the beast that can never be satiated.
Not all complexity is equal. A rich, interwoven system is complex, the kind a forest or a healthy community has, and then there is complicatedness: the proliferation of hierarchy, control, and centralised management, which is often complexity’s most expensive and least productive form. A great deal of what societies pour their energy into is this second kind: the layers of command, the concentration of decision-making, the apparatus required to make a population legible and controllable from the top.
The anarchist tradition, in its serious political-philosophy form rather than the chaotic caricature, made this point long before complexity theory arrived, and it is worth acknowledging because the dominant culture dismisses it reflexively. Thinkers in the lineage of Kropotkin and, more recently, the anthropologist James C. Scott argued that centralised hierarchy is frequently parasitic rather than productive. The elaborate apparatus of top-down control consumes enormous energy to extract, administer, and defend, while much of the actual problem-solving, the cooperation, the local knowledge, the day-to-day keeping-things-running, happens horizontally, among people coordinating directly. Second, that humans are not naturally helpless without rulers: mutual aid, voluntary cooperation and reciprocal support are not a utopian fantasy but a well-documented feature of how human (and animal) communities survive. Kropotkin’s observation that cooperation is at least as central to survival as competition has aged far better than nature as pure ruthless struggle.
A large share of the complexity that collapses under its own weight is the hierarchical, extractive kind, the centralised control structures that grow heavier and more fragile over time, defended by those they enrich long after they have stopped contributing. The reflexive opinion that order with centralised hierarchy is safer than decentralisation with chaos gets things backwards: highly centralised systems are often the fragile ones, efficient in good times and catastrophic in bad, while decentralised, federated, locally-rooted arrangements are frequently the resilient ones.
Modern civilisation has optimised relentlessly for efficiency, and in doing so has traded away resilience, and the distinction is the difference between a system that runs beautifully until it snaps and one that can bend under stress. An efficient system runs lean: no redundancy, no slack, global just-in-time supply chains, hyper-specialisation, everything tuned to extract maximum output under expected conditions. A resilient system keeps reserves, redundancy, local capacity, and diversity, paying a constant efficiency tax in exchange for the ability to absorb shocks. Our civilisation has systematically chosen the first and called it progress, which is why a single blocked canal, a single failed supplier, a single novel virus can now ripple into global disruption. We have built a system that is spectacularly productive and spectacularly brittle, the textbook setup for the kind of shock-triggered collapse Tainter describes, and we have mistaken the absence of recent catastrophe for proof of safety.
A population or a civilisation overshoots when it draws down its resource base faster than the base can regenerate, temporarily supporting a level of complexity and population the underlying system cannot sustain. The ecological version is overplanting the same crop and wiping nutrients from soil, water pollution, overfishing, and mass producing cheap concentrated energy faster than they renew, which is the literal energy-and-entropy story at planetary scale: a civilisation running on stored and borrowed gradients. Overshoot can look like prosperity right up until the reserve runs low, which is what makes it so dangerous and so easy to deny.
The historian Peter Turchin’s findings, drawn from data across many societies, are that political breakdown tends to follow two intertwined trends: the immiseration of ordinary people (stagnating living standards, rising inequality) and, crucially, elite overproduction, the generation of far more people aspiring to elite status, wealth, power, and position than there are elite slots to hold them. The mechanism Turchin calls the wealth pump: arrangements that siphon wealth upward, enriching a swelling elite while the broad population’s share stagnates or falls. The surplus of frustrated elite aspirants then turns into destabilising counter-elites, ambitious, capable, credentialed, and locked out, who mobilise mass discontent against the established order. Across history, by Turchin’s count, the indicators of this are inequality, intra-elite conflict, declining wellbeing, and polarisation. These tend to peak together in periods of upheaval, and most have been moving the wrong way in the West for decades.
Collective morals set the incentives that reward the ruthless. A system whose rules pump wealth and power upward will, predictably, concentrate both, hollow out the middleclass, breed resentment below and rivalry above, and grow steadily more unstable because the incentive structure selects for extraction. Centralised, extractive structures are not just energetically expensive; they are socially destabilising, manufacturing the very discontent that eventually tears them apart.
Collapse psychologically unwinds the people living through it, and often the blame for that feeling is placed on whatever is most convenient to them and their lives. Life getting hard? It must be the immigrants, the neighbouring countries’ religion, the men, the women, or even aliens. There is a recognisable pre-traumatic stress, a low, chronic dread about the future (sharpest in the young, who will inherit it), a grief for a future that was promised and now seems unavailable. There is the societal grief of watching familiar institutions, certainties, and shared realities lose their solidity. And there are the predictable nervous-system responses, the same fear and hypervigilance the manual traces in individuals, now ambient across a culture: the retreat into tribes, the scapegoating, the oscillation between frantic doomscrolling and numb distraction, the learned helplessness of people who sense something is deeply wrong and feel powerless to affect it.
This is worth treating gently, because dread is itself a kind of dysregulation, and marinating in collapse-anxiety tends to feed the very apathy it laments. Instead, we should probably take the analysis seriously enough to act on what we can personally affect and refuse to drown in the fear. The point is to take responsibility for what is in your control instead of wallowing in the things you can’t.
Collapse is rarely an event you wake up to, and far more often a process you are already living through. The cinematic image, the sudden fall, the barbarians at the gate, is largely a myth; historically, most collapses unfold over decades, even generations, of decline that those inside experience as a slow worsening, a series of crises each absorbed and normalised, a steady ratcheting down of expectations. People living through the long unravelling of past civilisations mostly did not know they were “in a collapse”; they knew that things were harder than before, that the institutions worked less well, that the future felt less assured, and they adapted to each step down until the cumulative descent was undeniable.
Diminishing returns on staggering complexity, a brittle hyper-optimised global system, ecological overshoot, the wealth pump and elite overproduction, fracturing shared reality and collapsing institutional trust, polarisation, and the ambient dread. This does not mean the end is imminent or certain; prediction is a fool’s game, and resilience and renewal are equally real forces. It means the honest question is not “will collapse happen?” but “what stage of a long, uneven process are we in, and what is still ours to shape?” And that reframe is the doorway out of paralysis: if collapse is a process rather than an event, it is something that can still be steered, mitigated, and rebuilt through.
Civilisations are dissipative structures, and they obey the same law as every other: order persists only while the energy sustains it, and the more is spent holding up expensive, brittle, extractive complexity, the closer the system drifts to the edge. Collapse is not a monster at the gate but a process of simplification, usually slow, that a society undergoes when its complexity outruns its returns, and many of its signatures, fragility, overshoot, the wealth pump, and fracturing trust are present and observable now, without any need for prophecy. The most subversive lesson is that much of what we have been taught to revere as order is in fact expensive, fragile, centralised control that survives more by enforcement than by merit, while the resilient, regenerative forms of order tend to be the cooperative, decentralised, bottom-up ones that hierarchy tries to suffocate. That is not a counsel of despair, and it is not a call to burn anything down; it is the ground for rebuilding the rest of this section.