Unfalsifiable

An Essay on Theories That Can Never Be Wrong

In 2002, a 26-year-old woman named Lydia Fairchild applied for welfare benefits in Washington State to help raise her two children on her own. The application required a maternity test. A few weeks later, she was called into a meeting with social services and accused of not being the mother of her children.¹

The DNA said so.

Fairchild showed them photographs of herself pregnant. Her mother testified. The children’s father testified. Her obstetrician testified. She had carried these children inside her body and delivered them. Social services didn’t care. DNA is 100% foolproof, a social worker told her. It doesn’t lie. So who are you?

A state prosecutor launched an investigation. Could she have kidnapped them? Could she be a surrogate who kept the children? After three hearings in court, Fairchild was facing the real possibility of losing her children to the state. Every day, she said later, felt like it was going to be the last day she’d see them. She called every lawyer in the phone book. None of them believed her. It was her word against DNA. It was her against everyone else.

Fairchild was pregnant with her third child at the time, and the judge ordered that both mother and baby be tested immediately after birth. The baby emerged from Fairchild’s womb — witnessed, documented — and the DNA test said this child was not hers either.

That should have ended the matter. Not the investigation — the theory. A test that tells you a woman is not the mother of a child you just watched her deliver has been falsified in the most definitive way possible. The gold standard of maternity — watching a baby come out of a woman’s body — contradicted the gold standard of genetic identity. One of them had to give.

It wasn’t DNA.

Support Independent Investigative Journalism and Research

This work remains free because paid subscribers make it possible. If you find value here, consider joining them.

What paid subscribers get: Access to all books I’ve written plus 1-2 new ones per month. e.g.

The DMSO Book
The Kitchen Remedies Guide
Chlorine Dioxide
The PSA Trap
Breast Cancer
Drilling for Profit
What Your Vet Can’t Tell You

Plus: Access to the Deep Dive Audio Library — 180+ in-depth discussions (30-50 min each) exploring the books behind these essays. New discussions regularly added. That’s 100+ hours of content for less than the price of a single audiobook.

[Upgrade to Paid – $5/month or $50/year]

unbekoming@outlook.com

The Architecture of an Unfalsifiable Claim

Dr. Tom Cowan, in a March 2026 webinar, laid out a series of cases that share a common architecture — cases from genetics, virology, and diagnostic medicine where the fundamental claims have been made immune to falsification.³ The Fairchild case was one of them. His analysis draws on a principle that the philosopher Karl Popper spent decades refining: the line between science and non-science.² His answer was falsifiability. A scientific claim is one that could, in principle, be shown to be wrong. If no possible observation could ever contradict your theory, you don’t have a scientific theory. You have a belief. The strength of a scientific claim lies not in the volume of evidence for it, but in the fact that it specifies what evidence would count against it — and that evidence has not appeared.

Put simply: a real scientific claim tells you in advance what would prove it wrong. An unfalsifiable claim is one where nothing counts as proof that it’s wrong — every result, including a contradicting one, gets folded into the theory as confirmation.

This principle is so foundational that most scientists would agree with it in the abstract. They teach it in philosophy of science courses. They invoke it when dismissing claims they find distasteful. And then they violate it systematically in their own disciplines, often without noticing.

The violation works like this. A claim is made. Evidence accumulates in its favour. The claim hardens into orthodoxy. Then a contradicting result appears — not a marginal anomaly but a direct, unambiguous falsification. At this point, the scientific response should be to re-examine the claim. What actually happens is that the contradiction gets absorbed into the theory. A new sub-category is invented. An exception is declared. The theory expands to accommodate the very evidence that should have destroyed it, and in doing so, it becomes unfalsifiable.

The result looks like science. It uses the language of science, publishes in the journals of science, and receives the funding of science. But it has crossed the line into something else.

DNA Chimerism: When the Test Fails, Blame the Mother

Back to Lydia Fairchild. The scientific establishment did not respond to her case by questioning whether DNA testing is as definitive as claimed, or by investigating what the test is actually measuring, or by asking whether the entire framework of genetic identity might contain assumptions that haven’t been adequately tested. Instead, they invented a category: DNA chimerism.¹

The explanation goes like this. In some people — and we don’t know how many — DNA from an absorbed twin becomes incorporated into certain tissues during foetal development. The result is that different cells in the same body contain different DNA. So a woman can give birth to a child that doesn’t match her blood DNA because the relevant reproductive tissues carry a different genetic signature.

Focus on what this does to the original claim. The original claim was: DNA testing can determine with 100% certainty whether you are the parent of a child. When the test confirms maternity, it proves maternity. When the test denies maternity — even in a case where the birth was directly observed — the denial is not treated as evidence against the test. It’s treated as evidence for a previously unknown condition in the mother.

The claim has become unfalsifiable. If the DNA matches, that proves DNA determines parentage. If the DNA doesn’t match, that also proves DNA determines parentage — it’s just that your DNA is unusual. The test is never wrong. Only you are.

Notice the direction of the reasoning. When confronted with a direct, observed falsification — a mother gives birth, the test says she’s not the mother — the institution does not re-examine the test. It does not ask whether the chemicals used in DNA testing might be producing non-specific results. It does not question whether the whole model of DNA as identity is built on assumptions that haven’t been independently verified. It invents a new condition — chimerism — to explain why the falsification isn’t really a falsification. And then the theory rolls on, absorbing the contradiction into itself like an amoeba engulfing a foreign particle.

The Fairchild case is not an isolated anomaly that the system eventually corrected. It’s a demonstration of how the system responds to falsification: not with curiosity, but with patch management. The question that should follow a case like Fairchild’s is uncomfortable but obvious: if DNA testing can fail this spectacularly in a maternity case where we have physical proof of the birth, how many times has it failed in cases where we don’t? How many paternity disputes, criminal convictions, or immigration decisions have rested on a test that we now know produces categorically wrong results — results that the system then explains away rather than investigating?

Nobody is studying this. The chimerism explanation was enough to preserve the theory, and preserving the theory was the point.

There is also a quieter problem that rarely gets discussed. DNA ancestry services claim to match you to distant relatives and ancestors. But those distant ancestors never underwent DNA testing.³ There is no reference sample from your great-great-great-grandmother. The match is to a statistical model of what her DNA should have looked like based on population-level assumptions — assumptions that themselves rest on the theory that DNA is the hereditary material and that it behaves in predictable ways across generations. The reasoning is circular. The theory validates the test, and the test validates the theory. There’s nothing outside the loop.

The Virus That Doesn’t Kill Cells

In early 2026, a paper appeared in the journal Recent Infection on the human metapneumovirus, or hMPV — described as a significant cause of respiratory infections worldwide, particularly in children, the elderly, and the immunocompromised, first identified in 2001.⁴

The paper follows a familiar structure. There is a section on diagnostics. There are genome comparisons with RSV and influenza. There are the customary diagrams. And then there is the section on viral isolation — the method by which virologists claim to demonstrate that a virus exists and to characterise it.

The method described is cell culture. You take material from a sick person, place it on a cell line (in this case LLC-MK2), and wait. Within three to seven days, the cells break down. This breakdown is called the cytopathic effect (CPE), and it is the gold standard of virology. The cytopathic effect is what tells you a virus is present. It is the proof of isolation, the basis for pathogen characterisation, and the starting point for vaccine development.⁴

The paper says this directly: culture in cell lines followed by detection is the gold standard for pathogen characterisation and supports vaccine development.

Then the paper notes a problem. The hMPV virus, it says, produces cytopathic effects rarely.⁴

The contradiction is plain. The gold standard for proving a virus exists is that it kills cells in culture. This particular virus almost never kills cells in culture. And yet it has been isolated, characterised, and confirmed as a significant pathogen. Both of those things cannot be true simultaneously. If the cytopathic effect is the proof of viral presence, then a virus that rarely produces the cytopathic effect has rarely been proven to be present. If the virus can be present without producing the cytopathic effect, then the cytopathic effect is not proof of viral presence — and the foundation of the entire isolation method collapses.

The paper does not address this contradiction. It simply presents both claims — the CPE is the gold standard, and this virus rarely produces a CPE — side by side, as though one does not annihilate the other.

This is the same structure as the DNA chimerism case. If the cells die, the virus did it. If the cells don’t die, the virus is still there — it’s just slow-growing, or the effect is delayed, or it takes longer than expected. The test proves the theory when it works, and the theory survives the test when it doesn’t work. Heads I win, tails doesn’t count.

The cell culture method itself exposes the problem. Researchers take a sample from someone they believe is infected — nasal fluid, sputum, something from the respiratory tract. They place this unpurified material onto a cell line, typically derived from animal kidney tissue. They reduce the nutrients available to those cells and add antibiotics and antifungal agents to the culture. Then they wait. If the cells break down over the following days, that breakdown — the cytopathic effect — is attributed to the virus in the sample.

The cells could just as easily be dying from starvation, chemical stress, or the antibiotics themselves. This is what controls are for. You run the same procedure without the patient sample and see if the cells die anyway. When this control has been done, the results have been devastating for the theory.

The hMPV paper skips this question entirely. It acknowledges that the gold standard method doesn’t work for this particular virus — the cytopathic effect is rare — and proceeds as though this is a logistical inconvenience rather than a falsification of the method itself. The virus is slow-growing, we’re told. It’s labor-intensive to culture. These are framed as practical challenges, not as evidence that the method has failed to demonstrate what it claims to demonstrate.

What you end up with is a situation where a virus that has never reliably produced the one effect that virologists say proves a virus is present is nonetheless catalogued, sequenced, compared to other viruses, and used to justify public health warnings. The entire downstream apparatus — the diagnostics, the genome comparisons, the epidemiological modelling — rests on an isolation step that, by the paper’s own admission, almost never succeeds.

A Pattern, Not an Anomaly

These two cases — DNA testing and viral isolation — come from different branches of science, rely on different methodologies, and involve different institutions. They share a single structural feature: the claim has been arranged so that no result can count against it.

The method is consistent. It has a recognisable shape. First, a test or procedure is established and declared to be the gold standard. Second, the gold standard produces a result that contradicts the theory it’s supposed to support. Third, rather than questioning the theory, a new category or exception is created that explains why the contradicting result doesn’t count. Fourth, the theory continues unchanged, now with an additional layer of insulation against future falsification.

This is not a rare phenomenon. It recurs across the biomedical sciences.

John Enders won the Nobel Prize in 1954 for successfully growing poliovirus in tissue culture. His method — placing patient material on cell lines and observing the cytopathic effect — became the template for all subsequent viral isolation. When Enders applied this same method to measles, he conducted what should have been a routine control experiment: the same cell culture procedure, but without adding any material from a sick person. The cells died anyway. The cytopathic effect and the resulting particles, he wrote, “could not be distinguished with confidence from the viruses obtained from measles.”⁵ A second virologist, Ruckle, independently reported isolating an agent from monkey kidney cells that was “indistinguishable from human measles virus.”⁶

The father of modern viral isolation, in his own foundational paper, reported that his control produced the same result as his experiment. Cells that were never exposed to any patient material broke down in the same way and produced the same particles as cells that were. This is not a marginal finding. In any other field of science, a control that produces the same outcome as the experimental condition would invalidate the experiment. It would mean the method can’t distinguish between the thing it’s looking for and background noise.

Enders noted this. He published it. He flagged it as a concern. And then the scientific community built an entire discipline on his method while quietly ignoring the control that showed it didn’t work. Enders won his Nobel Prize. The cell culture method became the universal standard for viral isolation. Every subsequent claim about a new virus — from measles to RSV to SARS-CoV-2 — traces its methodological lineage back to a procedure whose inventor said the control couldn’t be distinguished from the experimental result.

In 2016, the German Federal Supreme Court ruled in a case brought by biologist Stefan Lanka, who had offered €100,000 to anyone who could prove the measles virus exists. A young doctor, David Bardens, took up the challenge and provided six studies as evidence. Lanka argued the studies were insufficient — that they misidentified ordinary cell constituents as viral particles. The lower court initially ruled in Bardens’ favour. Lanka appealed. The Federal Supreme Court reversed the decision, ruling that the six studies did not constitute proof of the virus’s existence. The plaintiff was ordered to bear all procedural costs.⁷

Lanka’s argument was not that measles as a clinical presentation doesn’t exist. It was that decades of consensus-building had created a model of a measles virus that has never been found — not in a human, not in an animal — as an actual structure corresponding to the model. The Supreme Court found this argument scientifically sound. The ruling received almost no mainstream coverage.

In 2020, a paper in the peer-reviewed journal Kidney360 reported finding particles in kidney cells that were “morphologically indistinguishable” from what was being called SARS-CoV-2 — except these particles were found in patients who tested negative for COVID-19, and in kidney biopsies from the pre-COVID era.⁸ Structures that look exactly like the virus existed in tissue samples from before the virus supposedly appeared. This, too, was absorbed without altering the fundamental claims about viral identification.

Each of these cases follows the same arc. A method is established. The method produces a result. The result is treated as proof. Then the method fails — or a control reveals that it produces the same result without the alleged cause — and the failure is either explained away or ignored entirely. The claim moves further from falsifiability with each iteration.

The Fairy Dust Bucket

There’s a related tactic that shows up when unfalsifiable claims are challenged. Defenders shift the argument to a domain where they are technically correct, then behave as though that technical correctness validates the original claim.

Cowan uses an example.³ Someone tells you they have a bucket made of fairy dust. You ask about its characteristics. They say it’s two inches by two inches by two inches — a volume of eight cubic inches. You say you doubt the bucket exists and that it’s made of fairy dust. They respond: are you saying that two times two times two isn’t eight? Are you that bad at maths?

The maths is correct. Two by two by two does equal eight cubic inches. That is not in dispute. What’s in dispute is whether you have a bucket and whether it’s made of fairy dust. The mathematical accuracy doesn’t prove the existence of the thing being measured. It only characterises something that must first be independently verified.

This move — using the accuracy of a secondary measurement to validate a primary claim that was never proven — appears constantly. A genome is sequenced, and the sequence is accurate, therefore the virus it came from must exist. A protein structure is modelled, and the model is mathematically consistent, therefore the gene that codes for it must function as described. The maths is real. The thing the maths describes may not be.

Explaining This to a Six-Year-Old

Imagine your friend says there’s a monster under his bed. You look under the bed. No monster. Your friend says the monster is invisible. So you sprinkle talcum powder on the floor to catch its footprints. No footprints. Your friend says the monster floats. So you hold out food to lure it. Nothing happens. Your friend says the monster isn’t hungry right now.

Every time you test it, the monster gains a new power that explains why your test didn’t work. You can never prove the monster isn’t there. But your friend can never show you that it is.

That’s what unfalsifiable means. The monster is set up so that no test can ever catch it. Not because the monster is clever. Because the story about the monster keeps changing to dodge every check.

Now imagine your school says you have to take monster medicine every year. Imagine your mum almost loses you because a monster test says she’s not your real mum. Imagine whole countries spend billions of dollars on monster detectors — all because of a monster that was designed from the start so that nothing could ever show it wasn’t real.

That’s what this essay is about. Except the monster is a DNA test that told a mother she wasn’t the mother of the child she’d just given birth to. The monster is a virus that doesn’t produce the one effect that’s supposed to prove it exists. And the reason nobody questions them is the same reason your friend’s monster survives every test: the claims have been built so that no result — not even a directly contradicting one — is allowed to count against them.

Why This Matters

DNA tests are used in courtrooms to send people to prison or to separate children from their parents. Lydia Fairchild nearly lost custody of children she had given birth to because a test said they weren’t hers and the institutional faith in that test outweighed the direct evidence of her body. Viral isolation methods are used to justify vaccine development, public health mandates, school closures, and the allocation of billions of dollars in research funding. If the methods don’t measure what they claim to measure, every decision built on them is compromised.

The pattern extends well beyond genetics and virology. A drug trial fails to show survival benefit, so the endpoint is changed to a surrogate marker — tumour shrinkage, antibody levels, some measurable proxy — and the drug is approved on that basis. A genetic test predicts a disease that never materialises, so the concept of “incomplete penetrance” is invoked — the gene causes the disease, except when it doesn’t. A diagnostic tool produces false positives at a high rate, so the false positives are reclassified as “subclinical” cases — the test was right after all, the disease is just silent. Each of these moves performs the same function: it takes a result that should count against the claim and reframes it as consistent with the claim.

The cumulative effect is a body of medical knowledge that has never been seriously tested against its own failures. Not because the failures don’t exist, but because each failure, as it arises, is metabolised into the theory and disappears. The theory grows larger with each contradiction it absorbs. It becomes more complex, more qualified, more hedged — and less falsifiable with every iteration.

Unfalsifiable claims feel like the strongest kind of knowledge. A theory that nothing can disprove seems proven beyond all doubt. Popper’s insight was the opposite: a theory that nothing can disprove has told you nothing about the world. It has only told you about its own internal structure. A sealed room with no windows.

The test of a theory is not whether it can survive every challenge. The test is whether it has specified what would constitute a challenge — and faced it honestly.

By that measure, some of the most foundational claims in modern medicine are not strong. They are untested. And the difference between untested and proven is the entire difference between science and faith.

---

References

¹ Lisa Barne-Naod, Hidden Gas: Migrating Cells and the New Science of Microchimerism is Redefining Human Identity, as discussed in Dr. Tom Cowan’s Wednesday Webinar, March 11, 2026.

² Karl Popper, The Logic of Scientific Discovery (London: Hutchinson, 1959). Originally published as Logik der Forschung, 1934.

³ Dr. Tom Cowan, Wednesday Webinar, March 11, 2026.

⁴ “The Human Metapneumovirus (hMPV): The Virus Who Came with the Common Cold,” Recent Infection 54 (2026): 1–13, as discussed in Dr. Tom Cowan’s Wednesday Webinar, March 11, 2026.

⁵ John F. Enders and Thomas C. Peebles, “Propagation in Tissue Cultures of Cytopathogenic Agents from Patients with Measles,” Proceedings of the Society for Experimental Biology and Medicine 86, no. 2 (1954): 277–286, as discussed in Thomas S. Cowan, Breaking the Spell: The Scientific Evidence for Ending the Covid Delusion (2023).

⁶ Ibid.

⁷ German Federal Supreme Court (BGH), ruling on the “measles virus trial,” February 16, 2016, as reported in Thomas S. Cowan and Sally Fallon Morell, The Contagion Myth (New York: Skyhorse Publishing, 2020).

⁸ Cassol et al., “Appearances Can Be Deceiving — Viral-like Inclusions in COVID-19 Negative Renal Biopsies by Electron Microscopy,” Kidney360 (August 2020), as discussed in Thomas S. Cowan, Breaking the Spell (2023).