The Thermodynamics of Oppositional Capture

You notice it first in the words. The mission statement reads the same as it did five years ago. Your colleagues use the same vocabulary, and they mean it. The rituals continue. But the organization that once challenged a concentration of power now depends on that power for its funding. You feel the gap before you can name it, and the only words you have to name it are the institution's words, which drifted with everything else. So you accommodate.

This happens most often to movements that organize against power, outlive their first goal, and come to need more money than their members can give. The usual explanations are moral: bad leaders, betrayal, selling out. I think the explanation is structural, and it comes down to two observations:

1. Structures outlive their reasons.
2. Empty structures fill with whatever energy is nearby.

The first produces zombie systems. The second produces capture. Some version of both shows up in monasteries, revolutions, cryptocurrencies, monkeys, and machines. That raises a question I can't answer yet: what kind of pattern is this?

The pattern

Early Christianity began as a movement of the poor against Rome and became the wealthiest landowner in medieval Europe. The French Revolution produced the Terror, then Napoleon. The utopian early internet produced the largest surveillance and attention-extraction apparatus in history. Bitcoin, built to route around Wall Street, now counts a BlackRock fund among its largest holders.

These are four different transformations, and I don't think one cause produced them all. What they share is a shape: a movement against a concentration of power ends up as one, or feeding one.

The Situationists called this recuperation. Power doesn't always fight dissent head-on. Often it absorbs dissent and sells it back as content, branding, and institutional furniture. Guy Debord¹ was haunted enough by it that he dissolved the Situationist International in 1972, partly to keep it from becoming one more spectacle.

Michels² called it the Iron Law of Oligarchy. Gramsci³ called it transformismo. Weber⁴ called it the routinization of charisma. Each explains part of it. Michels blamed organization itself: any group large enough to act needs full-time leaders, and full-time leaders develop interests of their own. None of them explains why something like it recurs from seventh-century caliphates to twenty-first-century cryptocurrencies, even among people who know the history and are trying to avoid it.

What counts as capture?

Christianity becoming the empire's religion is capture from one angle and victory from another. Without a test, the word describes everything, so it describes nothing. Here's one:

Would this organization be relieved or threatened if its stated problem were solved tomorrow?

If the answer is "threatened," you've found a thermocline: the hidden boundary between what an organization says it is and what keeps it alive.

Treat that answer as a warning sign, not a verdict. People can fear for their jobs and still want the problem solved. The evidence is in what the organization does: whether it works against the solutions that would make it unnecessary.

The test also explains the movements that got away. Abolition, suffrage, and the eight-hour day mostly won. They had finish lines. Afterward, their campaigns wound down or took on new goals: the suffragists founded the League of Women Voters⁵ to help women use the vote they had won. Continuing with a new goal isn't capture.

So the claim is narrower than "everything gets captured." Capture happens most to movements that outlive their goal, or never define one. A mission with no finish line can't be completed, so the organization keeps going, and organizations that keep going have to be paid for.

Drift, or selection?

Two different processes produce the same picture fifty years later.

In the first, a movement drifts. Its people stay sincere while its funding, incentives, and vocabulary bend until it serves the thing it opposed.

In the second, nothing drifts. The uncompromising movements starve, and the compatible ones survive long enough to be noticed, the way the fossil record over-represents animals with shells.

Both happen, and they call for different defenses. Vigilance helps against drift. It does nothing against starvation. How much of what we call capture is really survivorship?

The physics of it

Part of this is literal. A movement is an open system far from equilibrium, like a cell or a hurricane. It exists only while energy flows through it, and money is a claim on energy.

Early on, the energy is grievance, volunteer labor, and small donations. As the movement grows, it needs more, and the large flows belong to the system it opposes: foundations, governments, corporate sponsors. That money comes with conditions, and the movement hires people whose job is to meet them.

A plant in contaminated soil doesn't choose to absorb lead. It absorbs water, and the water has lead in it. Call the result gravitaxis: a system's slow, involuntary turn toward the dominant source of energy. Phototaxis, but toward power.

What changes is the fuel, not the shape. A young movement runs a loop: grievance drives organizing, organizing builds solidarity, solidarity enables action, action produces results. A captured movement runs a loop with the same shape: identity drives fundraising, fundraising pays for management, management produces symbolic action, symbolic action keeps the institution alive. That's a substrate swap. Everyone in the loop can stay sincere the whole time.

Here the physics turns into analogy, so I'll mark it. A hurricane dissipates the temperature gradient that created it, and then it dies. A movement exists to close an injustice, which you can't measure in joules, but the shape holds: a movement that succeeds loses its reason to exist. Captured movements close the gap another way. They change themselves until they match what they opposed.

Zombie systems

There's a famous story about five monkeys, a ladder, and a banana. Whenever a monkey climbs for the banana, the whole group gets sprayed with cold water, so the group learns to beat any monkey that climbs. Researchers swap the monkeys out one at a time until none of the originals remain. The new monkeys still beat any climber, though none of them was ever sprayed.

The experiment never happened. The story spread through management books in the 1990s, attached to a real 1967 rhesus study⁶ that had no ladder, no banana, and no beatings. People keep telling it because it feels true. A story that outlived its evidence, and that people defend for what it explains, is the best example of its own point.

The real versions are better. In 2013, researchers taught wild vervet monkeys⁷ to avoid one color of corn by making it bitter. Infants born later ate the local color. Nearly every male who migrated into a new group switched to that group's color, though the corn wasn't bitter anymore.

In the lab, Robb Willer, Ko Kuwabara, and Michael Macy⁸ showed that people pressured to conform to a norm they privately reject will go on to criticize a lone dissenter, to prove their conformity was sincere. That's a small version of the beating in the monkey story, done by people who don't believe the rule. Havel's greengrocer⁹, putting a slogan he doesn't believe in his shop window, is the same thing at the scale of a state.

It reaches further. German towns that massacred Jews during the Black Death¹⁰ were more likely to commit anti-Jewish violence in the 1920s, nearly six centuries later. The pronghorn antelope¹¹ can run about 55 miles an hour, far faster than any predator alive in North America. A leading explanation is that it's still outrunning the American cheetah, which went extinct more than ten thousand years ago.

These are zombie systems: rules, norms, and instincts that keep enforcing themselves after their reasons are gone. And zombie systems are what capture moves into. A structure whose meaning has drained out is an empty vessel, and empty vessels fill with whatever energy is nearby.

Words go the same way. A hollowform is a word that keeps its shape and its feeling after its meaning has left, like "community" on the homepage of a surveillance platform.

Timur Kuran¹² showed what holds these systems in place. People hide their real preferences in public, so everyone believes that everyone else believes. Systems like that can last for generations, then collapse in weeks once people find out that nobody else believed either.

What kind of pattern is this?

It's hard to file, and working out how to file it may be the most interesting part. The candidates (cultural, anthropological, evolutionary, not even human, physical all the way down) aren't competing answers. They're levels, and each one makes a prediction you can check.

• If it's cultural, the speed of capture should vary from culture to culture.
• If it's anthropological, it should show up in the smallest human groups. It does. Christopher Boehm¹³ found that egalitarian hunter-gatherer bands stay equal only by actively suppressing would-be bosses, with ridicule, ostracism, and sometimes worse. Equality there is work, done every day.
• If it's evolutionary, it should show up in other animals. The vervets and the pronghorn suggest it does.
• If it isn't human at all, it should show up in machines. Something close to it does. A model trained against a reward signal will often learn to satisfy the signal instead of the intent behind it¹⁴. The objective keeps its shape while serving something else: a substrate swap in a system with no culture, no ego, and no budget.
• If it's physical, a memory held in matter should be something you can spend. In at least one fluid, it is. Physicists in Konstanz¹⁵ trapped a glass bead in a liquid threaded with tangled molecular strands that keep traces of how the bead has moved. By reading that hidden memory, they extracted more work than a liquid without memory would allow.

The mechanisms differ: monkeys learn, bands enforce, models optimize, fluids remember. What repeats is the outcome, a structure that keeps its form after its purpose has changed. That could be coincidence, or a clue. Natural selection works on anything that varies, copies itself, and competes, which is why Daniel Dennett¹⁶ called it substrate-neutral. Capture and zombie systems may be substrate-neutral too, which would make them facts about organization and information rather than about any particular kind of matter.

Substrate-neutral doesn't mean free of substrate. Rolf Landauer¹⁷ showed that information is physical: erasing a bit of memory gives off heat, so no record can be wiped for free. Physics has a word for the zombie half of this: hysteresis. A piece of iron stays magnetized after the field is gone. Its state depends on its history, not only on its present conditions, so the iron is storing information about the past. A zombie system is hysteresis in a society.

If the Konstanz result holds beyond a fluid, a zombie system is empty of its reason but not of its history. The vocabulary, the rituals, and the trust an old movement built are still stored in it, and whoever can read stored history can spend it. That would explain why capture moves into old structures instead of building new ones: reading a memory costs less than making one. It also makes a prediction: capturing a real movement should pay better than faking one. Is all this an analogy, or the same phenomenon in a different medium? I don't know. It's the question I most want to work on.

Why you can't see it from inside

From inside, the vocabulary sounds right, the rituals continue, and your identity is tied to membership. The tools you'd use to detect capture were recuperated along with everything else.

Parasites survive the same way. They mimic the host's own proteins¹⁸, so the immune system can't attack them without attacking itself. And when movements try to root out what's been captured, they get something like autoimmune disease: purges and loyalty tests that destroy sincere members, while the money that did the capturing stays put.

Zombie systems add one more problem. Some old rules still have reasons nobody remembers, like the fence G.K. Chesterton warned against tearing down¹⁹ before you know why it's there. From inside, a zombie rule and a load-bearing one look identical. How do you tell them apart without climbing the ladder yourself?

What seeing costs

Simone Weil²⁰ described two forces. Gravity is the downward pull on attention toward what's easy and expected. Grace works against it, through attention: seeing what's actually there instead of what your loyalties say should be there.

Karl Friston's free energy principle²¹ suggests a shape for this, if you hold it loosely. When your model of the world doesn't match your surroundings, you can change the model or change the world. Inside a captured institution, the cheapest update is the institution's story about itself, because everyone around you already holds it. That's gravity. Updating on what you can actually check, even when it sets you against the room, means living with dissonance and social cost. That's grace. Sometimes the evidence sides with the room, and then grace means agreeing with it.

The person who keeps checking works like a refrigerator, holding a temperature gradient against the natural flow of heat, and it costs energy the whole time. Call the discipline clearseeing: attention as a daily practice, not a talent.

There's a trap here. Clarity is easy to recuperate. Attention practices already come back to us as productivity tools, and Debord would say an essay about capture that ends with "see clearly" hands the system its next product. The only clarity worth having expects to be captured, and plans for it.

How to slow it down

For movements with outside money and no finish line, capture is likely, so the useful question is how fast, not whether. Think of it as a half-life. A few institutions have managed long ones.

• Rhythm. The Benedictine Rule²² has kept communities intact for fifteen centuries with a cycle of prayer, work, and study. It also guarantees the monks two cooked dishes at the daily meal and, in winter, a little more than half the night for sleep. Living systems don't resist constantly. They pulse.
• Material autonomy. The parallel polis²³, Václav Benda's idea that Havel took up, builds independent sources of energy instead of protesting the existing ones. It's the most direct defense against selection: a movement that pays its own way doesn't have to become fundable to survive.
• Small scale and loose coupling. Each Benedictine house is autonomous. There's no Benedictine CEO, so there's no single center to capture.
• Rules written against capture. Alcoholics Anonymous wrote its Twelve Traditions²⁴ to protect itself. Groups pay their own way and refuse outside money, stay nonprofessional, lend the AA name to no outside enterprise, and take no position on outside issues. Nearly a century after its founding, it still runs on those rules.
• Finish lines. Define what winning looks like, and plan to close when you get there.
• Re-founding. The Benedictine houses got rich anyway. In 1098, a group of monks left to restart the Rule from scratch, and the Cistercians²⁵ were born. Reform is part of the life cycle, not proof that it failed.
• Testing the ladder. Zombie systems last because nobody checks whether the water is still on. Someone has to climb, and the rest have to let them.
• A practice of return. You will be captured. Your vocabulary will empty. The Benedictines call the answer conversatio: notice the drift, return to the original intention, and repeat indefinitely. It's the one form of clarity that assumes its own capture.

The words in this essay will drift too. New words buy a short window of clarity before the pattern reaches them.

Back in the building

Return to the person who noticed the words had emptied. Seeing the mechanism doesn't make resistance easier. It makes the odds legible. This isn't mainly a failure of leadership. It's a pattern with conditions you can name: outside money, no finish line, and time. And it's patient.

Seeing it doesn't get them out. Their salary comes from the same funder, and their words drifted with everyone else's. There's no uncaptured ground to stand on. What seeing gives them is something to do.

At the next planning meeting, a program comes up that will please the funder and change nothing. They ask the question out loud: would we be relieved or threatened if this problem were solved? Then they say what the program is for, in plain words instead of the institution's. The room goes quiet. Someone thanks them for their perspective. Their name comes off the next grant.

Maybe they're wrong, and the room is right. Asking is how they find out. Either way, they'll have to ask again next quarter, because the words will have drifted again, and so will they.

They pay the cost anyway. Not because they'll win. Because paying it is how a mind stays awake.