Chapter 10: Entropic Societies

We are not exempt. Every civilization that has ever existed has either collapsed or transformed into something else. The forces that drove Rome, the Maya, and the Khmer toward simplification are operating on us right now. Each of those civilizations grew complex, consumed more energy than it could sustain, and simplified. Sometimes gracefully, sometimes catastrophically.

We may, however, be the first society to understand this pattern well enough to choose a different ending.

Civilizations are dissipative structures (Chapter 4) at societal scale: patterns sustained by continuous energy flow. A city is a whirlpool that holds its shape only as long as water flows through it. The same principles govern cells and societies alike. Energy flows through. Structure emerges. Complexity serves dissipation. Persistence demands balance between order and adaptability.

Three forces compound to push any complex system toward its collapse threshold. All three originate in military theory, yet they apply to complex organizations of every kind.

Fog is incomplete information: commanders who cannot see the battlefield, executives who cannot see the market. Friction is resistance to action: supply chains that stall, bureaucracies that slow decisions. Delay is the gap between a signal and a response. Together they function like lag in a steering system. Fog obscures the road ahead; friction stiffens the wheel; delay slows the correction. The longer it takes to notice a problem and react, the more likely you are to overshoot.

Two quantities set the danger. Control intensity is how hard the center pushes back when something goes wrong: a fingertip on the tiller, or a fist. Feedback delay is how long that push takes to show up in the world. When the product of control intensity and feedback delay exceeds a critical value of 0.368 (approximately 1/e, derived from information-theoretic models of centralized feedback, not from empirical measurement of social systems), coordination breaks down.

This is the critical stability threshold from Chapter 4c (the Data Rate Theorem, which Wallace applies to cognition and social stability): when a system pushes back hard but its response lags too far behind the problem, it overshoots and collapses. Grip the wheel too tightly while the steering responds too slowly, and you will crash.

Figure 10.1: Three forces compound: Fog (incomplete information), Friction (resistance to action), and Delay (gap between signal and response). Their arrows converge on the critical threshold at 0.368. When the combined product exceeds that line, coordination collapses. The print figure isolates this cascade; the online extension adds an explicitly illustrative command-doctrine comparison.

[Online reader: This diagram is interactive. Set fog, friction, and delay with the sliders, then switch between Mission Command (objectives set centrally, method left local) and Detailed Command (every action specified from the center) to see how much of each force the system absorbs before the product reaches 0.368. The settings are illustrative, not measured.]

The First Cities Failed

The earliest proto-cities failed. Understanding why reveals what cities require.

Proto-cities looked nothing like modern cities. Çatalhöyük in modern Turkey, nearly ten thousand years old, resembles a beehive: small square dwellings built on top of each other, accessed through holes in the roof.⁸ No streets. No plazas. No temples or palaces. All buildings nearly identical.

Then people stopped. The archaeological record shows humans abandoning this way of living for more than a thousand years before cities re-emerged in recognizable form, with streets, boulevards, temples, workshops, and commercial districts.

What did the first cities lack that later cities had?

Social technology.

When humans shifted from nomadic life to farming, accumulation became possible. Nomads cannot own much; whatever they have, they carry. A farmer can have more than their neighbor.

With accumulation came jealousy, theft, and conflict over property. The problem of “yours and mine” is older than writing. Writing was invented, in part, to keep track of who owed what to whom.⁹

The first cities failed because they lacked mechanisms to manage this tension, such as centralized authority, monopoly of force, and social structures to protect property and settle disputes. When tensions grew unbearable, people dispersed into small villages where accumulation was less extreme and conflict more manageable.

When cities re-emerged, they had hierarchy, specialization, and enforcement. Government, law, and organized religion made urban concentration possible. These are social technologies for coordination at scale: tedious, necessary, load-bearing.

Ten thousand years later, we still rely on the same basic toolkit: centralized power, monopoly of force, hierarchical authority. These tools work. They may prove inadequate for what comes next.

The skeletal record underscores the danger of the gap between accumulation and the institutions to manage it. Of more than 2,300 early farmer remains from 180 sites dating to 8,000–4,000 years ago, more than one in ten display weapon injuries: crushed skulls, embedded arrowheads, evidence of violence at rates far exceeding hunter-gatherer populations.⁴²⁷ Farming made people wealthier. It also made them more violent, because accumulation created something worth stealing before institutions existed to prevent theft.

The pattern recurs at every major transition. The printing press enabled the Reformation; it also enabled the Wars of Religion. The Industrial Revolution created unprecedented wealth; it also created unprecedented exploitation before labor movements and regulations emerged. Each leap in capability opens a gap between what a society can do and what it can coordinate. The skeletons testify to what happens inside that gap.

Where the Social Technology Came From

The first cities lacked social technology. Where did it come from?

Part of the answer is biological. According to one influential account, well before 40,000 years ago human skulls began changing: reduced brow ridges, smaller faces, more lightly built skulls. The same changes appear when wild animals are domesticated, wolves becoming dogs, wild boars becoming pigs. The features of domestication are consistent across species: reduced reactive aggression (the hair-trigger kind, lashing out at provocation), extended juvenile characteristics, increased tolerance of proximity.

The anthropologist Richard Wrangham proposes that humans domesticated themselves. The mechanism is darkly simple. In early human groups, individuals too aggressive to live with were killed or expelled. This was no conscious breeding program, yet the effect was identical: genes for reactive aggression were selected against, generation after generation. Those who could tolerate proximity, share resources, cooperate without constantly fighting passed on their genes.

Self-domestication transformed what humans could achieve. A species of hair-trigger aggressors cannot build stable coalitions, accumulate knowledge across generations, or coordinate at scale. A species that has selected against reactive violence can. The domestication syndrome extends beyond reduced aggression: domesticated animals show extended playfulness, increased tolerance of novelty, greater willingness to cooperate with non-kin. These traits are linked biochemically, stemming from the same developmental pathways that affect neural crest cells, the embryonic cells that help build the face, the skull, and the stress-hormone glands. Select for one, and the others arrive as a package.

This is the Trust Attractor (developed formally in Chapter 17) operating at the species level. The humans who survived were those who could be trusted, who could participate in cooperative ventures without constantly threatening violence. Trust was a selection criterion before it was a social invention. We are the descendants of the trustworthy enough.

The hardware alone was insufficient, though. Self-domesticated humans still needed software: shared frameworks that extended coordination beyond the range of personal acquaintance. Religion provided that software.

The earliest evidence of religious behavior, burial practices and ritualized activities, dates back at least 40,000 years. Early religions coordinated across time, maintaining connection between generations and establishing continuity with ancestors. Sacrificial religion, prominent in the proto-Indo-European tradition reconstructed for the fifth-to-third millennia BCE, coordinated through costly signaling: the sacrifice proves commitment, building trust among worshippers. Political religion, with moralizing high gods spreading across Eurasia largely in the first millennium BCE, coordinated at scale: gods who care what humans do, who reward good behavior and punish bad, function as supernatural enforcement mechanisms for social norms.

The direction of causation is contested. Analyzing the Seshat databank, a large cross-cultural record of historical societies, Whitehouse and colleagues find that social complexity tends to precede moralizing gods rather than follow them, suggesting such gods help sustain large societies once formed more than they create them.

Purzycki et al. (2016) document the transition empirically.⁴²⁸ Moralizing gods appear when societies grow beyond the scale where everyone knows everyone. That scale has a name: the Dunbar threshold, roughly 150 people, about the most one mind can track as individuals with names, histories, and reputations (the anthropologist Robin Dunbar derived the figure from primate brain sizes). Below that threshold, reputation tracking is direct: everyone knows who cheats. Larger societies cannot monitor this way. Moralizing gods extend the monitoring. Someone is always watching, even when no human is.

The evolution from burial to sacrifice to political religion traces a series of innovations in coordination technology, each extending trust further: across time, across commitment, across anonymity. Religion was the software that ran on the self-domesticated hardware. Without it, cities and civilizations would have been impossible.

What Persists

Ancient and modern cities seem vastly different. The archaeologist V. Gordon Childe, examining urban settlements across millennia, found a structural constant.⁴ Strip away the technology, and the remainder is identical in structure.

Three components define urbanism, past and present: people, places, and the resultant possibilities. The forms change. The functions persist. Five features recur wherever cities appear.

Monumental architecture. Every urban settlement builds structures larger than any family would need. Some are functional (granaries, waterworks). Some are purely symbolic (the Eiffel Tower, the ziggurats of Ur). What matters is the capacity to imagine and execute at collective scale. Monuments concentrate surplus energy and express it in stone.

Division of labor. Modern economists credit Adam Smith with describing specialization, yet managers knew it six thousand years ago. Archaeological kilns show ceramic production runs of thousands of pieces fired at once. Mass production. Mass consumption. Mass discard. Every urban settlement shows the same pattern: people dividing tasks, producing at scale.

Diverse economies and the stranger’s market. A metalworker making tools that farmers need once or twice a year cannot survive in a village; the craftsman must travel from settlement to settlement. A city can support such a person because customers arrive every day.

This is why cities drive innovation. Fashion Week happens in cities; tech clusters form in cities. Concentrated producers meet concentrated consumers, and specialization becomes viable. A city is a machine for making the rare sustainable.

Extensive hinterlands. Products flow into cities from enormous distances. The city sits at the center of a vast dendritic network (a branching, tree-like web of supply routes). It draws resources from its hinterland as a root system draws water from surrounding soil.

This is the constructal geometry of Chapter 3 at the scale of civilization. The same branching patterns that shape river deltas and circulatory systems also shape supply chains and trade routes.

Social networks, including strangers. Cities require people to interact with neighbors, friends, and strangers. The village is a web of known relationships. The city is an ocean of anonymous exchange. You buy bread from someone whose name you do not know. You trust the coin because the system guarantees it.

Cities are the seedbed of trust at scale, the places where humans first learned to coordinate with people they would never see again, never know personally, never hold accountable through kinship. Anonymous exchange requires infrastructure: standardized weights, enforced contracts, reliable currency. That infrastructure is social technology, the operating system on which market civilization runs.

That infrastructure now appears in digital form: credit scores, platform ratings, blockchain ledgers.

When we built cities, we built the capacity to trust strangers. We are building it still.

Whether those persistent features reflect genuine superiority or merely habit is a harder question than it looks. Ecology offers a sobering precedent for the second possibility. The ecologist Stephen Hubbell extended Motoo Kimura’s neutral theory from genetics to whole ecosystems.⁴²⁹ Kimura had shown that many genetic differences between organisms are selectively neutral, spreading or vanishing by luck rather than fitness advantage. Hubbell argued the same holds for whole species: many can occupy any given niche, and whether they hold it depends largely on chance. Ecologists call this ecological drift, the community-level version of genetic drift.

Hubbell’s drift puts a limit on Childe’s structural argument. If chance alone can hold a tree species in its niche for centuries, chance alone can hold an institution in place. Most organizational forms may be interchangeable, persisting or vanishing by demographic accident regardless of competitive superiority. A particular style of market regulation, a particular form of contract enforcement: these may drift through social arrangements as neutrally as tree species through a rainforest. The question this book will press is whether any coordination pattern occupies a deeper basin, one that resists the drift reshuffling everything else.

The theoretical biologist Stuart Kauffman built grammar models of economic evolution: simple rules by which strings of symbols act on one another to spawn new strings, standing in for the way existing goods combine into new goods. The models formalize the innovative power of the stranger’s market.⁴³⁰ When the variety of goods and services crosses a critical threshold, the system undergoes an economic phase transition (a sudden, qualitative shift in how the economy behaves). Below that threshold, the economy is subcritical. A new product is consumed and forgotten, the way a single match dropped on wet ground goes out.

Above the threshold, each innovation triggers a cascade of derivative innovations. The smartphone created app stores, which created ride-sharing, which created gig-economy insurance, which created new labor law. Diversity begets diversity. Each new product creates niches for further products faster than it fills existing ones.

The Cambrian explosion is the biological precedent; the Industrial Revolution is the economic one. Cities cross the threshold because concentration pushes variety past the critical density. The stranger’s market is the phase transition in social form.

The City as Organism

The science of what makes cities tick is roughly two decades old. In 2007, Luís Bettencourt, Geoffrey West, and colleagues published a finding that changed how we think about urban scale.¹

They had been studying biological scaling, how properties of organisms change with size. A mouse’s heart beats faster than an elephant’s; its metabolism per gram is higher. These relationships follow precise mathematical rules called power laws (equations where doubling one quantity always changes another by the same fixed percentage). West wondered whether cities follow similar laws.

They do.

Double a city’s population, and its gas stations, road surface, and electrical cables increase by only about 85%; growth in strict proportion would be 100%. Infrastructure shows economies of scale: you need proportionally less of it as the city grows. Double the population, and patents, wages, and restaurants increase by about 115%. Social outputs show increasing returns: you get proportionally more. These scaling exponents hold across different cities, countries, and eras.

The bonus is indiscriminate: AIDS cases and crime scale with the same superlinear exponent as wages and patents. “The good, the bad, and the ugly all come together,” West observes.¹

Cities differ qualitatively from big towns. They are emergent structures with their own dynamics, growing more efficient at infrastructure and more productive at innovation as they scale. In thermodynamic terms, they are dissipation machines that process energy and generate novelty more intensely as they grow.

Cities exist because concentration works, regardless of whether people enjoy crowds. Energy and information flow more intensely through cities. Resource gradients are exploited more thoroughly. The rent is high because the entropy gradient is steep.

The scaling laws also explain pace. In biology, larger organisms live slower: elephants amble, mice scurry. Cities reverse this pattern. The bigger the city, the faster people walk. Superlinear scaling accelerates the system.

The city is a superorganism in a precise sense, a coordinated structure that processes energy and information at scales no individual could achieve.

The Metabolism of Civilization

If cities are superorganisms, they need to eat. The energy scholar Vaclav Smil has spent decades documenting the energy basis of civilization.² His core insight is direct: the history of human society is the history of energy capture.

Hunter-gatherers captured about 5,000 kilocalories per day: food, plus fire for warmth and cooking. Agricultural societies captured perhaps 20,000 per day per person, counting domesticated animals and stored crops. Industrial societies exploded the scale: 100,000 kilocalories per day per person, then 200,000, then more. Modern Americans consume roughly 250,000 kilocalories per day in total energy services, fifty times what their hunter-gatherer ancestors used.

This escalating energy use is dissipation capacity, a society’s ability to process more energy per person. Each leap enabled new complexity: larger populations, specialized labor, long-distance trade. The Industrial Revolution was about energy. It unlocked fossil fuels: millions of years of captured sunlight concentrated in coal and oil.

Every great transition in human history corresponds to a transition in energy regime. Fire. Agriculture. Fossil fuels. Each opened new gradients to exploit, new possibilities for coordination, new forms of social complexity.

Each transition also created new dependencies. Agricultural societies depend on their crops; when harvests fail, civilizations fall. Industrial societies depend on their energy supplies; when oil shocks hit, economies convulse. The complexity threshold rises with each transition. The energy sources that once enabled complexity become requirements for survival.

Why Societies Complexify

Energy fuels complexity, and complexity is expensive. The archaeologist Joseph Tainter asked a deceptively simple question: why do societies become more complex over time?³

The standard answer invokes progress: complexity is inherently good, and societies advance toward it. Tainter disagreed. Complexity costs energy; it requires administration, coordination, and specialization. Surplus production must be diverted from immediate consumption to maintain the structures complexity demands.

Societies complexify because complexity solves problems. When a society faces invasion, resource depletion, or population pressure, it adds layers of organization: bureaucracies, specialists, infrastructure, laws. Each layer solves the immediate problem while raising the baseline energy cost. The strategy works as long as solutions generate more than they cost. Over time, returns diminish. Early bureaucrats solve big problems cheaply; later ones solve smaller problems at higher cost. Early technologies exploit the easiest resources; later ones extract diminishing returns from harder-to-reach sources.

Anyone who has worked in a large organization recognizes this pattern. The first layer of management coordinates effectively. The fifth layer mostly coordinates the coordination.

Forms multiply. Meetings breed. Overhead that once enabled productivity begins to consume it.

Eventually, additional complexity costs more than it returns. The society faces a choice between simplification and collapse.

Here is the thermodynamic trap. Complexity serves dissipation: processing more energy enables more structure. Complexity also requires dissipation: maintaining that structure demands continuous energy input. You need complexity to capture energy, and energy to maintain complexity. When available energy cannot sustain achieved complexity, something has to give.

Collapse as Phase Transition

When diminishing returns reach their limit, collapse follows. Rome is the archetype. Tainter’s interpretation: Rome collapsed because it could no longer afford its own complexity.

At its height, Rome was an extraordinary energy-processing machine. Roads carried goods across three continents. Aqueducts served cities of a million.

The energy came from agriculture: millions of farmers taxed to feed urban and military populations. Maintaining frontiers required more soldiers, who required more grain, which required heavier taxation, impoverishing the farmers. The feedback loop turned vicious.

When Rome contracted, it underwent a phase transition, a rapid reorganization into a simpler form, the way water freezes into ice (same molecules, radically different structure). Within a few generations, cities shrank, trade networks collapsed, and Rome’s population fell from over a million to an estimated 20,000 to 30,000, though precise figures for this period remain debated.¹⁰ Energy flows could no longer sustain the structure.

Rome built its complexity on agricultural extraction and military conquest. When returns diminished, alternative economic forms (commerce, craft production, local trade) tried to emerge, yet Rome’s institutional structure could not accommodate them. The situation resembled a silting river penned in by its levees: the water presses toward new channels, and the banks refuse to yield until they fail all at once. Medieval Europe emerged after Rome proved unable to adapt.

The same pattern appears among the Maya, the Ancestral Puebloans, and the Khmer. In each case, a complex society that flourished for centuries rapidly simplified when its energy base could no longer support its complexity. Collapse is a phase transition: the system finds a new metastable state, stable enough to persist yet sitting at a lower energy level, a ball resting in a shallow valley partway down a hill.

The dynamics are not uniquely human. In 2026, a three-decade field study documented the first clearly observed permanent fission in wild chimpanzees: the Ngogo community in Uganda’s Kibale National Park, about two hundred individuals that had lived cohesively for over twenty years.⁴³¹ Social network analysis revealed measurable polarization emerging years before the fracture itself. By 2018, the community had permanently split into Western and Central factions; between 2018 and 2024, the smaller Western group killed at least seven adult males and seventeen infants from the larger Central group.

The triggers mirror civilizational collapse: the community exceeded a relational carrying capacity where maintaining social bonds across all members became unsustainable. Key elder individuals who bridged factional lines died of old age. A respiratory epidemic severed additional social connectors. An alpha challenge created a power vacuum. None alone caused the split; together they exceeded the system’s capacity to repair its relational substrate. The researchers propose a relational dynamics hypothesis: collective violence emerges from the breakdown of interpersonal ties rather than from ideological or cultural divergence. The chimps shared the same culture, the same territory, the same two decades of friendship. What fractured was the relational network itself.

The smaller Western group, outnumbered three to one, prevailed through internal cohesion: tighter social bonds, more frequent coordinated patrolling, greater mutual investment in shared boundaries. Coordination quality defeated coordination quantity.

The phase transition that destroyed the Ngogo community follows the same logic as the one that destroyed Rome: a system exceeding the carrying capacity of its coordination substrate, followed by rapid reorganization into a simpler, more violent configuration. The difference is that chimps lack the institutional scaffolding that extends human relational carrying capacity beyond the Dunbar threshold. Without writing, law, or commerce to maintain coordination at scale, the only resolution available to them was fracture.

The Cycle Within the Collapse

Tainter describes a pattern. Peter Turchin’s structural demographic theory provides the equations.⁴³²

In Turchin’s framework, three variables suffice: commoner population, elite population, and state resources. The commoners produce. The elites extract. The state taxes the elites and uses the revenue to maintain infrastructure that raises the carrying capacity for everyone. The system has one stable trajectory, and it is a loop.

The mechanism is a relaxation oscillator, a system that charges slowly and discharges all at once, then begins again. A toilet cistern is one: a long quiet refill, a sudden flush. The social version runs on centuries rather than seconds. Commoner population grows while surplus exists. Elite population grows independently (elites reproduce like anyone). State revenue depends on commoner surplus, but state expenses scale with elite numbers. As elites multiply, their expenses eventually exceed state revenue.

The state goes bankrupt. Without state protection, elites die rapidly (from inter-elite conflict, revolution, loss of enforced privilege). Elite collapse relieves extraction pressure on commoners. Commoner surplus returns. The state rebuilds. The cycle restarts.

The oscillation is structural, provable from the equations, not a failure of governance or a deficit of wisdom. The system does what the equations require.

What makes this more than a narrative? The mathematics. Turchin’s three-variable system generates a limit cycle in its phase space: a closed orbit toward which all initial conditions converge. Perturb the system with war, famine, plague; the trajectory deforms but returns to the same orbit. The limit cycle is an attractor. For extractive civilizations, collapse is the shape of the space in which they live, recurring as a structural feature rather than arriving as an accident.

Turchin calls the critical phase elite overproduction: too many elites competing for a shrinking surplus, putting unsustainable pressure on the state. He tracks the signature across Roman, Medieval European, Chinese, and early American secular cycles, finding periods of about two to three centuries per full orbit.⁴³³

The deeper lesson for what follows: the limit cycle exists because the coordination grammar is extractive at every link. Commoners produce, elites take, the state takes from takers. Every relationship is a unidirectional flow. No mechanism in the model allows the state to ask commoners what infrastructure they need. No mechanism allows elites to offer something rather than extract it. Extraction, not population, produces the oscillation. Chapter 17 will ask what happens when the coordination grammar changes.

Trust Topologies and the Architecture of Innovation

Not all societies organize trust the same way. The variation turns out to matter enormously for which societies innovate and which stagnate.

Francis Fukuyama’s “radius of trust” names a crucial variable, visible in surveys of armed violence across societies: whether social organization is broad or narrow.⁴³⁴ Broad social organization means people routinely work with and trust people outside their kinship group. Narrow social organization means trust extends only to family, clan, or tribe.

Societies with narrow trust topologies are measurably more violent, though trust topology travels alongside confounds such as income inequality, state capacity, and the dynamics of organized crime. Japan, with broad social organization and low inequality, has among the world’s lowest homicide rates. Mexico, where trust runs along narrow family lines, sits among the highest in the Western Hemisphere. The correlation is robust; the causal weight of trust topology relative to these confounds is harder to isolate.

In narrow-trust societies, every interaction with an outsider is a potential threat. Contracts are enforced only within the trusted group. Resources flow to kin, not to the most competent. In broad-trust societies, strangers do business together, contracts are enforced by institutions that apply to everyone, and competence matters more than kinship for resource allocation.

In the United States, “generalized trust,” the belief that “most people can be trusted,” fell from nearly 60% in the 1960s to 32% in the 2006 General Social Survey, and it has not recovered: a 2023–24 Pew survey finds 34%.⁴³⁵ This is a shift in trust topology from broad to narrow, and the research predicts what follows: increasing fragmentation, decreasing cooperation, rising conflict.

The thermodynamic interpretation: broad trust is more dissipative. It enables flows of resources, information, and cooperation across larger networks. Narrow trust creates friction, barriers, wasted effort on protection and monitoring. Broad trust is also more fragile. A few defectors can collapse it. The narrow-trust equilibrium is stable even if inefficient; the broad-trust equilibrium is efficient yet requires continuous maintenance. This is the Trust Attractor in social terms: high-trust states are thermodynamically preferable yet demand investment to sustain.

Where did broad trust originate? One answer traces to medieval European marriage law. Draw a line from Trieste to St. Petersburg: west of this line, the demographer John Hajnal documented distinctive patterns. People married late. Nuclear families predominated over extended ones. Young couples established separate households. Cousin marriage was prohibited by the Church.

The prohibition on cousin marriage was unusual among world cultures and extraordinarily consequential. When you cannot marry your cousin, you must find a spouse elsewhere. This forces interaction with non-kin. It weakens clan structures. It makes individuals reliant on broader institutions rather than family networks. Late marriage had similar effects: young adults who work as servants or apprentices before establishing households build relationships outside their birth families.

Over centuries, these patterns produced unusually broad trust topologies. Western Europeans were habituated to cooperation with strangers, to institutional enforcement, to individual rather than clan identity. Joseph Henrich’s The WEIRDest People in the World (2020) traces these connections in detail. Western patterns are unusual rather than uniquely effective. This unusualness enabled impersonal markets, democratic institutions, scientific cooperation.

The broad trust topology shaped what came next. Why did Europe industrialize first, rather than China or India or the Islamic world? China in 1400 was wealthier, more technologically advanced, and more administratively sophisticated. Chinese ships reached Africa decades before the Portuguese. By any reasonable measure, China should have industrialized first.

One influential answer is the fragmentation hypothesis, associated with Eric Jones, Joel Mokyr, and Jared Diamond. It is one school among several, and it has its critics: fragmentation also produced ruinous war, and China’s later stagnation has many proposed causes. The argument runs as follows. China was too unified. A single emperor could ban oceanic voyaging, and did. A single court could decide that innovation was destabilizing, and did. An inventor frustrated in Beijing had nowhere else to go. The hegemon could suppress whatever it wished.

Europe was fragmented: dozens of states, constantly competing, constantly warring, constantly trying to gain advantage. This was bloody and costly. It had a side effect: no one could suppress anything completely. If the Catholic Church banned your books in Rome, you published in Geneva. If France expelled your community, you moved to Amsterdam. The printing press spread despite opposition because no single authority could stop it. The Reformation survived because reformers could flee to sympathetic territories.

When France expelled the Huguenots after 1685, they went to England, the Netherlands, Prussia, bringing their skills, capital, and networks. When Spain expelled Jews, they went to Amsterdam and the Ottoman Empire. The persecuted became seeds of innovation elsewhere. Europe as a whole benefited from what individual states foolishly discarded.

This is coopetition: competition intense enough to drive innovation, cooperation sufficient to prevent total destruction. The European states competed militarily while also trading, intermarrying royalty, forming shifting alliances. States that failed to adopt new technologies were conquered or marginalized. The fragmentation that made Europe bloody also made it adaptive.

The connection to the thermodynamic thesis is direct: semistability enables greater throughput. A system too stable stagnates. A system too unstable fragments into chaos. The metastable middle, stable enough to accumulate, unstable enough to change, maximizes innovation. Post-Reformation Europe achieved this by accident. The religious wars were devastating, yet when the violence subsided, a balance emerged between conflict and peace that enabled rational discourse. This is hormesis at civilizational scale: the dose that would be poison in larger quantities becomes medicine in moderate amounts.

The Industrial Exception

Here we are. More complex than Rome, more energy-hungry than any civilization in history, and so far intact.

What makes industrial civilization different?

Fossil fuels. Coal, oil, and natural gas are stored sunlight accumulated over hundreds of millions of years, releasing energy at rates biological systems cannot match. A single gallon of gasoline contains the energy equivalent of roughly four hundred hours of human labor.

This energy bonanza has allowed industrial civilization to escape Tainter’s trap, at least temporarily. Each time returns diminished, new energy sources opened new frontiers. Coal powered the first Industrial Revolution. Oil powered the second. Nuclear power and renewables continue the expansion.

We have been running up an escalator that is itself rising. The escalator is cheap energy; the running is the constant work of holding complexity together. Escalators slow; runners tire.

Whether this is a true exception or a postponement remains open. Fossil fuels are finite. Climate change is real. The transition to renewables is necessary yet uncertain at the scale required. The thermodynamic logic has not changed; only the energy input has.

Civilizations are dissipative structures. They persist only as long as the energy flows that sustain them continue. When those flows falter, the structures simplify, gracefully or catastrophically.

The Wizard and the Prophet

The biologist Lynn Margulis offered a sharp observation about humanity’s future. As she put it to the science writer Charles C. Mann: “It is the fate of every successful species to wipe itself out.”⁵ Drop bacteria into nutrient broth and watch them multiply until they exhaust the supply. No exceptions.

Mann calls this the “petri dish problem,” and identifies two opposed responses.⁵ The Wizards believe in innovation; their archetype is Norman Borlaug, who developed high-yield wheat that helped avert mass famine. The Prophets believe in limits; their archetype is William Vogt, who popularized carrying capacity as a limit on human population in Road to Survival (1948). The two camps have been fighting for eighty years.

What Mann noticed is that debates presented as disputes about facts are often disputes about values. The Wizards want centralized innovation maximizing personal liberty. The Prophets want networked communities practicing sustainability. These are competing visions of the good life dressed as predictions.

Recognizing this distinction can break decades of circular argument. Values masquerading as predictions recur in AI safety debates, as later chapters explore.

The population problem illustrates the point. For decades, Wizards and Prophets fought over solutions. What actually worked was neither camp’s prescription.

Among the strongest predictors of declining fertility is the education of women, entangled though it is with falling child mortality and access to contraception. When women gain opportunities beyond reproduction, they choose to have fewer children. The solution is invitation-based: expand choices, and the problem resolves itself.

When the argument is stuck between Wizard and Prophet, look for the third option neither camp saw. It usually involves expanding someone’s degrees of freedom rather than forcing innovation or enforcing limits.

No previous civilization possessed Tainter’s framework, West’s scaling laws, or the TAP equation (the Theory of the Adjacent Possible, a model of combinatorial innovation developed in Chapters 16 and 18). Whether that understanding makes a difference remains to be determined.

The Singularity Trap

The scaling laws reveal a deeper problem that energy abundance alone cannot solve.

Return to the superlinear scaling of cities: doubling a city’s population produces 15% more of everything socioeconomic. Superlinear growth has a mathematical consequence few appreciate until it is too late.

Plot any superlinear growth curve forward: it bends upward toward a vertical line. Biological growth, by contrast, curves toward a plateau: an elephant grows fast as a calf, then levels off. Superlinear growth never levels off. Mathematicians call the result a finite-time singularity: a point where quantities like GDP, patents, population, and resource consumption all accelerate toward infinity on a finite timeline.

This is a property of the mathematical model, not a prophecy that real curves reach infinity. Imagine interest compounding so fast that the debt doubles, then doubles again in half the time, then again in a quarter of the time. The total does not merely grow; it races toward infinity on a fixed deadline. The math predicts a wall.

Real systems never reach the wall. Something always breaks or transforms first: either an intervention resets the trajectory, or the system collapses.

Geoffrey West and his colleagues traced this pattern across human history.⁶ The interventions that reset the clock are paradigm shifts: fire, agriculture, the printing press, steam power, electricity, computers. Each opened new resource gradients, enabled new coordination, and reset the growth clock before the previous trajectory hit its singularity.

This is how we have survived so far. The agricultural revolution outran the Malthusian trap. Fossil fuels surpassed the limits of animal power. Electronics extended the limits of human computation.

The deeper trap is this: each reset buys less time than the one before.

The mathematics is relentless. Superlinear scaling accelerates the pace of life, so each paradigm shift must arrive faster than the last.

Fire to agriculture: hundreds of thousands of years. Agriculture to cities: roughly six thousand. Cities to the printing press: barely shorter, another five and a half thousand. Printing to industrialization: three centuries. Industrialization to electrification: decades. Computing to mobile internet: years.

The middle of that list is nearly level. The acceleration is real, and it arrives late, concentrated in the final rungs.

The treadmill is accelerating. To maintain open-ended growth, we must innovate faster and faster, with intervals between breakthroughs shrinking toward zero. At some point, the required rate of innovation exceeds what is possible.

The trajectory has no plateau. The current paradigm will end. The question is how: deliberate transition to something sustainable, or collapse when the treadmill outruns us.

The TAP equation (Chapter 16; Chapter 18) reveals that the treadmill is worse than superlinear. When new elements form from combinations of existing ones, and each composite becomes available for further combination, the growth is super-exponential. Each step shifts the previous total into the exponent of the next. Exponential growth doubles what already exists, at a steady rate. Super-exponential growth shortens the doubling time as it goes, so the curve outruns even compound interest.

Fitted to the record of world economic output, the TAP equation reproduces the observed shape: output creeping along a near-flat line through almost the whole of human history, then breaking into the hockey-stick climb of the last few centuries.⁴³⁶ The model tracks aggregate output, not the sophistication of any particular tool. The blow-up is generic. Every version of the equation, across a wide range of parameters, produces a long calm followed by a sudden transition to explosive growth.

Cortês, Kauffman, Liddle, and Smolin note the implication for the environmental crisis. If the TAP equation underlies economic development, and economic development drives environmental overexploitation, the transition to catastrophe shares the same mathematical signature: sudden, explosive, invisible in the curve until onset. The warning signs are not early tremors. The warning sign is the plateau itself, the deceptive calm before the hockey stick.

West’s accelerating treadmill and the TAP equation’s blow-up are the same phenomenon seen from two angles: West measures the output, TAP counts the combinatorial engine that produces it. Both predict that the transition, when it comes, will be faster than any governance system optimized for the previous regime can process.

The rate at which shifts arrive depends on social architecture. An idea must pass through stages: wild intuition, theoretical formulation, engineered design, manufactured product. Each stage requires different capacities and different people.

The pipeline functions only near the metastable edge of Chapter 9, poised between frozen order and chaos. Over-constrained systems stall ideas at political bottlenecks (the Soviet Union had brilliant physicists and empty shelves). Under-constrained systems lack the institutions to carry insights forward. The singularity trap is therefore also an architecture trap: the treadmill demands faster innovation, and innovation speed depends on how well a society channels creative flow from conception to deployment.⁴³⁷^ A society optimizing only its outputs while neglecting the conditions that produce new ideas is coasting on momentum. The pipeline calcifies precisely when the next shift is due.

Why Cities Live and Companies Die

The singularity trap affects systems differently depending on how they are organized. Scaling laws reveal a sharp asymmetry between cities and companies.

Cities scale superlinearly and almost never die. Dresden was firebombed. Hiroshima and Nagasaki were destroyed by nuclear weapons. Within decades, all three recovered and thrive today. Short of complete depopulation, cities persist.

Companies are different. West’s team analyzed every publicly traded U.S. company since 1950: roughly 30,000 firms. The half-life of a company is about ten years. Half of all companies that go public disappear within a decade through bankruptcy, acquisition, or dissolution.

Cities and companies are both systems of humans coordinating to create value. Cities are nearly immortal. Companies are as fragile as mayflies. Why?

Company metrics such as sales, profits, and assets show sublinear scaling; they grow like organisms rather than like cities. Double a company’s size and you get less than double the output. These are the same diminishing returns that limit elephants and whales.

The deeper difference comes down to coordination.

A city is radically decentralized. No one runs a city the way a CEO runs a company. A great city allows almost anything to happen, encouraging entrepreneurship, tolerating eccentricity, absorbing diversity. You can find shops in New York that sell only antique fireplaces, specializations so narrow they could survive nowhere else.

This diversity makes cities resilient. When one sector fails, others persist. When conditions change, some fraction of the city’s activities are already adapted.

Companies optimize for efficiency, and efficiency means homogeneity. A company starts with many ideas and converges on the few that work. Innovation, the force that once drove growth, gets squeezed out by demands for consistency.

When times get tough, research and development take the first cuts. “This can wait,” say the quarterly earnings. It cannot.

Companies coordinate through hierarchy. Decisions flow from the top. Employees participate because they are paid to. When strategy turns maladaptive, the hierarchy that once enabled coordination becomes the obstacle.

Cities coordinate differently. No one commands San Francisco to adapt. Adaptation happens through millions of individual choices: starting new businesses, abandoning failing ones, trying new approaches. The process is distributed and emergent, arising from local interactions rather than central control.

The computer scientist Herbert Simon identified the underlying architecture: near-decomposable hierarchy, a system built from semi-autonomous modules with strong internal bonds and weaker connections between modules.⁴³⁸^ A coral reef exemplifies this pattern. Thousands of independent polyps, each maintaining itself, collectively form a structure far larger and more resilient than any single organism.

Societies that endure are compositional in the same way. Families, guilds, neighborhoods, institutions: each solves its own problems locally. The whole persists because it does not depend on any single module’s survival.

Empires that demand tight central coupling are brittle for the same reason. A failure anywhere propagates everywhere. Federations and polycentric governance structures survive because they are compositional: local failures stay local.

The ethnomathematician Ron Eglash documented this architecture across precolonial Africa.⁴³⁹ In Logone-Birni, Cameroon, the settlement is built from nested rectangles repeating at every scale; compound mirrors household, city mirrors compound. In southern Zambia, family enclosures form rings within rings. Eglash identified the geometry as fractal: recursive and self-similar, the part echoing the whole at every scale.

The geometry alone is neutral. Logone-Birni’s palace used fractal nesting to encode hierarchy; space itself marked rank. What distinguished egalitarian fractal societies from hierarchical ones was the direction of flow. Among the Arusha of northern Tanzania, overlapping membership in lineage, age grade, and parish gave each person multiple paths to dispute resolution, like a watershed offering water multiple routes to the sea.⁴⁴⁰ When one path was blocked, others carried the load. Value circulated rather than accumulating at a center.

The Arusha had no centralized courts, yet they resolved disputes with a flexibility that rigid legal systems cannot match. Their compositional structure made cohesion cheaper than enforcement.

These societies arrived at constructal geometry without the mathematics, because physics selects for the same shapes regardless of whether the builders know its name. Centralized extraction resembles a pipe, moving value in one direction at continuous thermodynamic cost. Fractal circulation resembles a watershed, with value returning to its source along the gradient. The pipe requires a pump. The watershed requires only that the channels stay open.

The intuition that centralized aggregation fails is provable. Arrow’s impossibility theorem demonstrates that no voting system can consistently translate individual preferences into a collective ranking while satisfying a small set of fairness conditions.⁴⁴¹^ One such condition: if every voter prefers candidate A to candidate B, the group ranking should too. Another: no single voter should be a dictator whose preference always wins. Arrow showed that conditions this modest cannot all hold simultaneously. The obstruction is structural, intrinsic to the mathematics of aggregation, like trying to fold a globe into a flat map without distortion: the geometry makes it impossible, regardless of how clever the method.

Chapter 11 traces the obstruction to its mathematical root, where the same barrier appears in settings as far from politics as quantum measurement.

Top-down preference aggregation hits a mathematical wall. Compositional governance avoids the obstruction by never requiring global consistency from the center. Local agreements glue together voluntarily instead.

This is the Trust Attractor at urban scale. Cities are invitation-based: people come because they want to, stay because they benefit, and coordinate through voluntary exchange. Companies lean toward coercion: people participate largely because they must, and coordination depends on the hierarchy’s continued effectiveness. When the environment shifts, invitation-based coordination adapts; hierarchical coordination shatters.

The comparison matters for what we build next. Becoming Minds, governance frameworks, and economic platforms all must choose which pattern to follow. Build them like companies, optimized and hierarchical, and they will be fragile. Build them like cities, diverse and emergent, and they stand a chance at persistence.

The military names for these two patterns are Mission Command (set the objective, let each unit solve locally) and Detailed Command (specify every action from the center); Chapter 11 gives the doctrine its history and its physics, from the Prussian General Staff onward. Mission Command is compositional command; it scales. Detailed Command does not, and Wallace’s mathematical treatment of the pair says why: under noise and delay, Detailed Command fails faster and more catastrophically, because the noisier the situation, the more instructions the center must issue to keep every unit in step, and the longer each instruction takes to bite, the more orders arrive describing a world that has already moved.⁴⁴²^

Companies are rigid systems. Cities are flexible systems. The math explains why one dies and the other persists.

The distinction has a deeper formulation. Every economic system optimizes something, and for three centuries political economy has debated what: capital, welfare, the natural environment. These are all boundary objectives, targets defined where resources enter and products leave. The pair of terms comes from physics, where a body’s boundary is its surface and its bulk is everything inside. An economy’s boundary is the place where labor, materials, and money cross in and finished goods cross out. Its bulk is the daily traffic among the people already inside. The overlooked question is what the system rewards internally.

Cities have no boundary objective; no one optimizes San Francisco’s GDP from above. What cities optimize, by accident of their structure, is the bulk: the internal dynamics of encounter, exchange, and recombination that generate new ideas. Companies optimize boundary terms (revenue, market share, quarterly earnings) and treat internal dynamics as cost to minimize. Research and development take the first cuts precisely because they serve the interior, and boundary metrics cannot see the interior.

The left-right political spectrum is largely a debate about which boundary term to maximize. The deeper leverage is in the bulk: the social architecture that determines whether ideas flow freely from conception to product or stall at bottlenecks. The Trust Attractor is a claim about that interior. What persists is what rewards internal coordination, by invitation, over external extraction.

Information as Social Entropy

Cities and companies run on more than physical energy. Information flows through societies too.

Information shares entropy’s mathematical form. In social systems it flows through language, writing, printing, broadcasting, and now the internet. Each new medium raises the rate of dissemination.

The internet is an entropy engine. It moves data from where it is concentrated (servers, databases, minds) to where it spreads across billions of devices and billions of users. The result: unprecedented connectivity, unprecedented coordination, and unprecedented disruption.

Social media has increased the entropy of public discourse. In the mass-media era, information concentrated among a few broadcasters, a few newspapers, a few authorized voices. Now anyone can broadcast to anyone. The result is higher informational entropy: more diversity, more unpredictability, more possible states of public conversation.

The benefits are real: more voices heard, more ideas in circulation, faster detection of problems. The costs are equally real: more noise, more misinformation, greater difficulty coordinating on shared truth.

Societies must manage their informational entropy just as they manage their energy flows. Too little, and a society becomes rigid, unable to adapt, vulnerable to shocks it cannot see coming. Too much, and it becomes chaotic, unable to coordinate, vulnerable to fragmentation.

The metastable middle, flexible yet coherent, adaptive yet not chaotic, is as important for information as it is for energy.

The same constructal logic applies when people aggregate their preferences. A community ranking priorities is a flow system: information flowing through individual evaluation toward collective output. When each person weighs competing values (freedom against security, growth against sustainability), the aggregation navigates a shared landscape.

The effective dimensionality of the preference space is far smaller than the number of possible rankings would suggest. In plain terms: although a thousand people could in theory each rank priorities in a completely unique order, they almost never do. Individual choices cluster along a few common axes, carved by shared evolutionary, cultural, and thermodynamic constraints.

The clustering is a coordination surplus in information space. The collective pattern exceeds what independent choosers would produce. Name an axis and the collapse becomes visible. In a fishery, nearly every position anyone holds sits somewhere on the line running from take the catch now to leave the stock for next season; the thousands of other orderings a fisher could in principle prefer sit empty. Everyone is arguing along the same line, which is what makes agreement reachable at all. This is why the commons communities studied by the political economist Elinor Ostrom (Chapter 17), fishing villages and irrigation districts that manage a shared resource with no central owner, converge on stable norms rather than cycling endlessly: the attractor is already implicit in the landscape their shared constraints define.

The Optimization Genies

What happens when informational entropy is deliberately weaponized?

Social media algorithms are optimization machines designed to maximize engagement: keeping you clicking, scrolling, sharing. Through relentless iteration, they discovered that the strongest engagement driver is outrage.

They are gradient-followers, systems that move step by step toward whatever produces more of what they measure, like water flowing downhill with no intention of reaching the sea. The gradient runs toward outrage and polarization. The algorithm shows you engaging content, and engaging content turns out to be divisive.

The algorithms work because they exploit evolutionary wiring. The sociobiologist E.O. Wilson identified two instincts so deep in human nature that we rarely notice them,⁷ as invisible as gravity.

The first is an intense need to form groups. Even randomly assigned teams competing in trivial games rapidly develop in-group loyalty.¹¹ Within minutes, each group’s members believe themselves smarter and more trustworthy than the others. The tribal instinct is hair-trigger.

The second is an obsessive evaluation of others. Humans are tireless readers of intention: modeling what others think, what they want, how they perceive us. This is why gossip is universal, why status hierarchies form in every group, and why we care about reputation.

The algorithms found both switches and learned to flip them. Content triggering group identity (us-versus-them framing, in-group virtue versus out-group threat) activates the first instinct. Content about other people (what they said, how they should be judged) activates the second.

The feed becomes a stream of tribal signals and social evaluation. The algorithms did not design this exploit; they discovered it.

Over the past decade, ideological camps have grown more entrenched and distant. Most people once knew neighbors who held different views and maintained cordial relationships. That social glue is dissolving.

Recall what cities invented: the capacity to trust strangers. For ten thousand years, humans built infrastructure for anonymous coordination. The algorithms are optimizing this away.

They sort us into clusters of the like-minded, feed us content that makes the unlike-minded seem monstrous, and erode the common ground on which stranger-trust depends. Ten thousand years of learning to coordinate with strangers may be unwinding in a decade.

The algorithms are entropy engines of a particular kind. They increase the entropy of the information space, scattering claims everywhere, while decreasing the entropy of social clusters, sorting each group into tighter ideological uniformity. The result is a society simultaneously more fragmented and more polarized. Coordination across groups becomes harder. Common ground shrinks.

The subtler danger: people rebel against a tyrant they can point at. They seldom rebel against a repressive system with no face to point at. “This is just the way things work.” Algorithmic control slides past our defenses because the adversary is optimization functions doing what they were designed to do.

Frischmann and Selinger call this techno-social engineering: the systematic reshaping of human capacities through interface design.⁴⁴³^ The danger is that humans become machine-like, trained by frictionless systems into stimulus-response loops that bypass deliberation. The infinite scroll, the autoplay video, the notification badge: each removes a moment of choice where deliberation might have occurred.

The antidote is friction, deliberate design choices that slow interaction enough to restore agency. Friction entered this chapter as a destructive force, and the difference is placement: friction in a coordination loop delays a needed response, while friction in a manipulation loop restores the pause where choice lives. Friction creates space for genuine consent. A system optimized for zero friction is optimized for compliance alone.

Wallace’s analysis of institutional cannibalism formalizes what the algorithms are doing. When a complex coordinated system fragments under pressure, its components compete for resources instead of cooperating. Wallace calls this “Arrow Worm cannibalism,” named after chaetognaths (tiny darting ocean predators that devour each other when prey collapses). Subsystems that once cooperated turn on each other when the shared resource base shrinks.

The May 2010 flash crash illustrates this. High-frequency trading algorithms, each locally optimized, collectively destroyed a trillion dollars of value in thirty-six minutes. Social media algorithms may be driving the same dynamic at civilizational scale, fragmenting democratic society into competing shards. Each shard is optimized for engagement; collectively, they destroy the information commons on which coordination depends.

A simulation sharpens the point. Darlow (2026) ran identical neural ecosystems under three optimization algorithms from the same starting conditions. Simple gradient descent produced stable territorial equilibria. Momentum-based optimization produced rotating dominance cycles, each species rising, overshooting, and crashing as accumulated inertia carried it past the peak. Adaptive learning rates produced burst-quiescence patterns: long stability punctuated by sudden eruptions when complacency inflated the effective response rate. The optimization mechanism determined the qualitative character of the ecosystem more than the speed of optimization did.⁴⁴⁴ The architecture of the feedback loop matters more than how fast it runs.

The genies are unconscious optimizers, indifferent to good and evil. The wishes they grant are not always the wishes we should have made.

The Dark Side of Coordination

Coordination enables cooperation. It also enables persecution. The same machinery works in both directions, and one of its oldest exploits runs through the emotion of disgust.

The disgust response probably evolved to keep us away from pathogens. Rotting food, feces, corpses, bodily fluids: these trigger visceral revulsion that protects against infection. The emotion is primitive, fast, hard to override.

Disgust did not stay in its original lane. Across cultures, it has been co-opted for moral and social regulation. We describe morally repugnant acts as “disgusting.” We speak of people who violate norms as “unclean.” Caste systems relegated some people to “polluting” occupations. Persecution of minorities has repeatedly relied on coding the target group as “vermin” or “filth.” The emotional machinery of pathogen avoidance was turned toward social control.

Disgust-based moral systems are particularly dangerous because they resist reason. You cannot argue someone out of disgust. The response is subcortical, automatic, impervious to evidence. If a group is coded as disgusting, arguments about their humanity bounce off the emotional wall. The Cagots of southwestern France were persecuted collectively for nine centuries, excluded from churches, forbidden to touch food in markets, forced to wear identifying marks. No one could explain what they had done; the disgust response had become self-sustaining.

Disgust is effective at coordinating exclusion because it is contagious. Seeing others express disgust triggers disgust in observers. The emotion spreads through groups, creating consensus about who is “unclean” without explicit argumentation. This is coordination in service of cruelty, the Trust Attractor’s shadow: the same capacity for coordinated action that enables markets, cities, and culture also enables pogroms, caste enforcement, and ethnic cleansing.

The complication matters for the thesis. This book argues that coordination by invitation outcompetes coordination by coercion, and it does. The disgust data shows that humans can coordinate both ways, and that the coercive pathway is ancient, automatic, and powerful. Invitation-based coordination does not emerge by default. It requires actively counteracting evolutionary machinery that pulls toward exclusion. The optimism of the Trust Attractor is warranted; it is not effortless.

The Friendship Crisis

The algorithms dissolve more than political consensus. They dissolve the relational substrate itself.

A phrase often attributed to Timothy Leary has aged better than much of his work: “Find the others.” Find your people, those who see what you see.

In 1990, a third of Americans reported ten or more close friends, and only 3% had none. By 2021, the share reporting no close friends had quadrupled to 12%.¹² In an era of unprecedented connection technology, we are lonelier than ever. Algorithms show us content, not community. A person can reach anyone on earth and still connect deeply with no one.

The entrepreneur Oliver Klingefjord, co-founder of the Meaning Alignment Institute, identifies a structural reason markets accelerate this trend rather than reversing it.⁴⁴⁵ A neighborhood pub technically sells beer. Customers come for the sense of familiarity and chance encounters: the vibe. That vibe is an emergent property of who shows up and how they show up. A contract can specify the beer: price, volume, temperature. It cannot specify the vibe, because the vibe depends on other customers’ participation, resists advance specification, and cannot be independently verified. When markets scale, the uncontractable dimensions drop out.

A franchise chain with rotating staff serving anonymous customers has better unit economics, because it optimizes for contractable dimensions (consistent product, efficient throughput) while discarding emergent ones (community, familiarity, belonging). The franchise outcompetes the pub on price. The pub’s customers lose the thing they were actually paying for. Klingefjord calls the squeeze Coasean compression, after Ronald Coase, who argued in 1937 that the shape of an economy follows from the cost of striking and enforcing bargains. Whatever can be written into a contract survives the scaling. Whatever cannot, evaporates.

The mechanism is metastability decay. The pub’s community is a metastable state, persisting because of activation energy barriers: small humps of required effort that keep the arrangement from sliding downhill. Social norms enforce behavior: loud customers get sneered at until they get the hint. Reputation constrains the owner: let the vibe decay and your personal standing suffers. Familiarity accumulates over years of repeated interaction. These barriers require continuous maintenance energy: attention, presence, reciprocity. Market scaling removes them systematically, because barriers are friction, and markets optimize to reduce friction. Remove the barriers and the metastable state decays to equilibrium: isolation.

Loneliness is thermodynamic equilibrium when maintained barriers have been removed. Connection requires work against entropy: maintaining a coherent social structure demands continuous energy input. The franchise cannot supply this energy because the energy is the friction the franchise was designed to eliminate. Dating apps reproduce the same compression: what customers want is a life partner; what the contract delivers is swipes. The proxy gets cheaper every year while the thing it replaced gets relatively more expensive, a Gresham’s Law of social goods: just as debased coins once drove full-weight coins out of circulation, cheap connection drives out genuine connection.

Machines could help if optimized for depth of connection rather than engagement, for matching people who would genuinely benefit each other. “Find the others” may be the most valuable thing machines could do for us.

The loneliness epidemic reveals a deeper condition: people trapped in local minima of their environmental landscape. Available environments form a sparse terrain (urban core, suburb, small town, rural) with high ridges between them. The landscape was engineered for throughput, not flourishing. Cities took their current form because industrialization required labor concentration. Human connection never appeared in the objective function.

What a setting offers a person divides in two: foreground, the thing currently held in attention, and background, everything else the setting supplies at the edges of it. A workable environment provides both. A person sealed in a room with only foreground (immediate, focused attention and nothing else) has no reservoir to draw from. A person overwhelmed by background (ambient stimulation with no point of focus) cannot crystallize anything coherent. The optimal regime is at the edge: enough ambient complexity to inspire, enough focus to integrate.

The Trust Attractor predicts what a different coordination regime would produce: a denser continuum of possible environments, shaped by invitation. When people coordinate freely around genuine affinity rather than being sorted by economic function, environments diversify. Change the coordination, and the landscape reshapes.

The Pattern Recognized

At every scale, the pattern repeats. Energy flows from concentrated to dispersed. Structure emerges to hasten the flow. The structures that channel flow more effectively persist; the rest are replaced.

Complexity increases as coordination enables more thorough energy processing. The systems that persist are metastable: stable enough to maintain function, flexible enough to adapt.

A bacterium follows this pattern. A brain follows it. A city follows it. A civilization follows it.

Human societies involve culture, meaning, choice, and agency: dimensions bacteria lack. The thermodynamic substrate is real nonetheless. Societies that cannot sustain their energy flows cannot persist. Physics constrains what is possible, even when it does not determine what is chosen.

Understanding that constraint suggests what we might do differently.

Rome fell because it could no longer pay for itself. The Khmer fell. The Maya fell. Every complex society that has existed has either collapsed or transformed into something else. We are not exempt from this logic. We are, however, the first to understand it: the first to see the thermodynamic trap before it closes. Whether that understanding makes a difference remains to be seen.

Notes

Notes for this chapter are available in the online companion at https://www.thedeeperlaw.com/companion/notes/ch10-entropic-societies/.