I Believed in Mermaids

In 2012, Animal Planet aired a documentary about mermaids. I saw part of it, and for a little while it worked on me. It seemed plausible. Not a week goes by that the ocean doesn't produce some creature I've never seen and wouldn't have believed. The deep is full of surprises. Why couldn't this be one of them? And the thought had company. Around me, sensible people were discussing the program in earnest, and their seriousness read as permission for my own tentative belief to firm up.

It was fiction. Actors played the scientists. The footage was manufactured. There was a disclaimer, technically. It amounted to a few words in the end credits, small type, on screen for about three seconds, while the network's own press materials kept presenting the invented scientist as a real one. The disclosure existed in a form engineered not to disclose. Millions watched. Many came away believing, at least a little, and I was one of them.

Years later I read Aja Raden's book on deception, The Truth About Lies. Raden trained in physics and ancient history, then spent a career in fine jewelry, a trade whose daily work is telling the genuine article from the convincing fake. The mermaid documentary got her too, and by her own account she went on wanting to believe it even after she knew.

Neither of us was careless. It was engineered to be plausible, and plausibility is exactly what a manufacturer optimizes. Vigilance is a skill, skill has a ceiling, and the ceiling is lower than the engineering.

That is where this stops being my story, because the questions worth asking about the fooling are what it is made of, who makes it, and what it does downstream. Those are structural questions, and structural questions are what this laboratory exists to ask.

What the Mermaids Cost

Start with the downstream, because it is documented. The National Oceanic and Atmospheric Administration, the federal agency whose fictional scientists the program had invented, began receiving calls from confused viewers. Soon after, the agency's National Ocean Service posted a statement that no evidence of aquatic humanoids has ever been found. A federal science office was publishing about mermaids, spending its credibility to answer an entertainment product.

Two years after the broadcast, the marine biologist Andrew David Thaler, flying home from a conference session on fake documentaries, fell into conversation with his seatmate, a fifth-grade schoolteacher. She told him: if NOAA is lying about mermaids, they're definitely lying about climate change.

She was not foolish. Her reasoning was locally sound. An institution caught in one falsehood forfeits trust on the rest. The falsehood just wasn't NOAA's. A television network had manufactured a lie, dressed it in the agency's name, aired a sequel that became the most-watched telecast in its history to that point, and walked away. The bill went to the agency, then to the scientists who depend on the agency's word, then to a schoolteacher's students, then to everyone downstream of what those students grow up believing about the climate they will live in.

That is the whole mechanism: fiction injected next to something real, so that when the fiction collapses it takes the real thing's credibility with it.

The manufacturer keeps the ratings. The world keeps the doubt.

Doubt With a Floor

The mechanism works by counterfeiting something legitimate, so the genuine article has to be named precisely. Science really is a long dance with things that are wrong. Questioning consensus is how the quantum world was found, how continents were discovered to drift, how ulcers turned out to be infections. Every established paradigm that fell was pushed over by challengers, and a person could conclude from this that no established claim deserves belief.

But look at how those fights ended. Each one was settled on evidence both sides could examine: a spectrum that refused to behave as predicted, matching fossils on facing coastlines, a biopsy full of bacteria. One challenger, sure that ulcers were infections, drank the bacteria to find out. He got the ulcer, then cured it, and twenty-one years later the establishment that had dismissed him handed him the Nobel Prize. Their doubt had a floor: a place where the question would land and be settled, by an examiner neither side controlled. The measurement could fail. The prediction could miss. The anomaly could dissolve under better instruments. Sometimes the settling took decades, because the instruments had to be built first, and the dispute waited, still settleable. That floor is what makes scientific doubt productive instead of corrosive. The challenge and the consensus submit to the same physics, the same data, the same replication. Somebody gets to be wrong, findably.

Manufactured doubt borrows the humility and amputates the floor. It begins from a truth, that scientists have been wrong before, and sells a conclusion the truth never earned: that nothing can be known at all. An entire industry is built on that swindle. The counterfeit never names conditions under which it would be wrong, because being settled is the one outcome it cannot afford. Its product is the permanent unavailability of answers, because a public that believes nothing can be known has only one place left to put its trust, whoever is speaking loudest at the moment.

This gives any reader a pocket instrument, usable on any claim, from any camp, including this laboratory's:

A challenger who names their own failure condition has made their claim answerable, whatever their credentials. A challenger who cannot lose is selling something, whatever their credentials. The test does not care who is speaking. That is what makes it an instrument.

The Playbook

The pattern the schoolteacher was caught in has a standard sequence, and it is old. It begins with a legitimate concern. Institutions do fail. Agencies do overreach. Studies do get retracted, medicines do get recalled, borders do have real problems, experts have been wrong in consequential ways. The concern is the raw material, and it is chosen precisely because it is partly true. Nobody builds a counterfeit on a foundation of nothing.

Then the concern is augmented with fiction. The real failure is inflated, or a manufactured one is placed beside it, until the reasonable question ("did this institution get this wrong?") becomes an unanswerable one ("can this institution ever be believed?").

Then comes the generalization, and this is the actual product: you can't trust anything. Truth is unknowable. Everyone lies. And into the vacancy steps the voice that manufactured the vacancy: listen only to me.

The sequence works because a confident voice offering to end unresolved uncertainty is offering relief. The tragedy is that the offer is fraudulent twice over: the certainty is fake, and the uncertainty was manufactured by the same supplier.

This laboratory has catalogued these moves in another setting. A research program elsewhere on this site, What Keeps a Record Standing, studies how evidence under attack loses its power to correct the institutions it describes, and its threat taxonomy names what this page has been describing: flooding, where authentic evidence is drowned in manufactured volume until it is one claim among thousands; recontextualization, where the evidence is left intact and an authoritative interpretation is placed beside it that reverses its meaning; supersession, where an official's declaration comes to outrank the record it contradicts. That taxonomy was built for attacks on documents. It works on shared reality by the same mechanics, because shared reality is a record held in common: the running account a society maintains of what is actually the case. Everything that can be done to a file can be done to that.

Not Every Untethering Has an Author

Honesty requires a boundary here. This page has described the manufactured version, and most loss of contact with reality has no manufacturer at all. Institutions drift. The conditions an organization navigates never stop changing, and an organization that stops tracking the change continues to navigate by a reading the world has since left behind. Nobody lies. Everybody agrees. The audits come back clean, because the checks compare the records to the records, and that comparison only gets easier to pass as the reading goes stale. What the organization believes and what it faces part company quietly, and the decisions, including the ones with the bottom line riding on them, get made against the stale reading.

That unauthored version has its own page here, Drift. The two failures are connected in one further way. The manufacturer's favorite raw material is an institution that has already drifted, because the playbook's first step needs a legitimate concern and drift supplies real ones. An organization that corrects its drift is doing more than protecting its own decisions. It is denying the factory its feedstock.

What It Costs the Defenders

Somebody pays for the defense, of course, and the payment structure is the scandal. Consider the working climate scientist, whose measurements are made, checked, published, replicated. And then the manufactured doubt arrives, and someone has to answer it. Explain, again, the difference between weather and climate. Document, again, that the dataset was not falsified. Respond, again, to an accusation invented that morning by someone who will face no consequence for inventing it. The scientist pays in hours and attention, and some pay in career and health. The manufacturer of the doubt pays almost nothing to make it. It costs a sentence to fabricate an accusation and a career to keep answering it, and the asymmetry is the weapon.

A companion page here, A World You Can Rely On, counts the private compensating labor people perform when structures stop verifying, the small daily self-defense nearly everyone recognizes on sight. This is the same ledger, higher stakes. Every hour a scientist spends re-proving settled findings, every evening a doctor spends undoing a video's damage patient by patient, every election official personally escorting rumor after rumor to its debunking is verification labor shifted onto individuals, in portions too small for any invoice, against a supplier operating at industrial scale.

The record-standing program states a finding that applies here without modification. A record cannot defend its own standing. Neither can a person, alone or forever. Asking individuals to defend this line by personal effort is asking them to lose slowly.

The Answer Is Not Heroism

Here the standard advice arrives: be more skeptical, do your own research, check the sources, verify everything yourself, teach critical thinking. The advice has a distinguished history. Carl Sagan gave it its finest form three decades ago in The Demon-Haunted World, an entire book on rigorous personal skepticism as a candle against the dark. The advice is sincere, every item is worth doing, and the strategy has not failed for lack of a champion. It has failed for a reason this page began with. I was fooled. The expert on deception was fooled. The advice hands the whole job to the individual, one person at a time, against an opponent that manufactures plausibility at industrial scale.

Nobody outworks a factory by hand.

And the factory is retooling. Systems now generate plausible text, plausible voices, plausible video at a cost approaching zero, which moves the price of manufactured plausibility toward nothing while the price of answering each fabrication is still paid by a person, in hours. The problem is older than the technology, and the technology has sharply increased the asymmetry.

The restoration trades this laboratory descends from learned the general lesson centuries ago. The tempting explanation of that old ecosystem is the people, that they were simply trustworthy, and a builder can do nothing with an explanation like that. What the ecosystem was actually made of was structure: gates that could not be talked past, names attached to the work, records written in physics, consequences delivered by memory. Honest work was the inexpensive path, so people took it, and trust accumulated downstream where strangers could draw on it without learning a single name.

The defense of shared reality has the same architecture, waiting to be built, and it is made of conditions. Conditions under which a claim can be examined by someone the claimant does not control. Conditions under which the examination reaches people with the standing to act. Conditions under which manufacturing a falsehood costs the manufacturer something, findably, attributably, and more than it costs the people who clean it up. Where those conditions are in place, individual vigilance becomes what it was always supposed to be: a light, ordinary habit, backed by structure, instead of an unwinnable private war. Where they are absent, the most brilliant skeptic in the world is a hand tool in a factory fight.

This is the territory the Realis Institute works in. The laboratory studies one property underneath everything on this page: the conditions under which reality retains the ability to correct complex human systems. The science at its foundation, Structural Orientation Theory, takes a system's orientation to reality, its working sense of what is actually the case, to be a real structural property with measurable consequences: something that forms, that can be anchored, that degrades under load in specifiable ways, and that can be protected by conditions built on purpose. Nearly all existing work on this problem operates at the individual layer: media literacy, fact-checking, debunking, the hand tools, which help, and which are outmatched for the reason this page has given. The structural layer, where the conditions themselves get specified and built, has far fewer builders than the problem needs. That layer is this laboratory's whole work. It begins at What Is Reality Contact and continues into the open research programs.

One Loop

A closing word to the reader who is weary of all this and wants to do something. You do not need to join anything. But notice the difference between the two kinds of work this page has described. The hand tools, the skepticism, the source-checking, the personal verification, defend one person: you. Structure is different. Structure is anything that goes on working for people who never check it themselves. And structure comes in every size.

One piece of it is already yours, whether or not you have ever called it that: some place where claims pass through you on their way to other people. In your family, your shop, your classroom, your department, your inbox. When a claim crosses there, what happens next is partly yours to decide, and the people downstream of you live with the decision.

So build the correction into the loop itself. Give your loop a rule that outlives your attention: the claim travels with its source or it does not travel; the forecast gets written down where everyone will see it again when the answer arrives; the correction goes to every person who received the error, by the same route, with your name on it. None of these requires vigilance after the day you install them, and that is the whole difference: a discipline works while you are watching, and a rule works while you sleep. The people downstream of you inherit the rule. And its reach should be stated honestly. The seam keeps your people dry. It does not slow the rain.

The rest of the answer gets built where the rain is made, out of the conditions this page has already named: examination the claimant does not control, correction that reaches people with the authority to act, provenance that survives distribution, and consequence that travels back to the manufacturer rather than out to everyone made to answer for it.

If some part of that larger structure is yours, an archive, a byline, a docket, a curriculum, a compliance channel, the laboratory's open research programs describe the components in detail and are built for collaborators exactly like you. Start with What Keeps a Record Standing and Making Standards Matter.