The warning came buried in a handbook designed to help people tell the truth. Stephan Lewandowsky and John Cook had spent years studying misinformation. They understood the mechanics of false belief better than almost anyone alive. And what they found, when they looked hard enough, was this: correcting a lie can make people believe it more. Not always. Not even usually. But sometimes, the act of debunking backfires — and the myth walks away stronger than before.
That is the paradox sitting at the center of one of humanity's oldest intellectual traditions.
Debunking is not a modern invention. It is not a YouTube genre or a fact-checking column. It is something closer to an instinct — the compulsion to say *that isn't true* and then prove it. The word itself is barely a century old, but the practice reaches back to before the Roman Empire. The question that has haunted it ever since is deceptively simple: does it work?
---
To understand why this question matters now, you have to understand what the internet did to misinformation. Before social media, a false claim had friction. It spread through word of mouth, through tabloids, through chain emails that your uncle forwarded. The speed was manageable. The volume was not infinite. Then platforms like Facebook, Twitter, and YouTube removed the friction entirely, and the volume became something no one had designed for. Viral hoaxes could lap a correction before the correction finished loading. The debunkers — the Snopes editors, the MythBusters crew on Discovery Channel, the skeptics running websites from their home offices — found themselves in a race where the starting gun had already fired.
Communities formed around this work. Mick West built Metabunk, a forum dedicated to methodical, evidence-based investigation of conspiracy theories. Susan Gerbic founded Guerrilla Skepticism on Wikipedia, organizing volunteers to improve the sourcing on articles about pseudoscience and the paranormal. Brian Dunning ran Skeptoid as a podcast. Phil Plait ran Badastronomy.com. Captain Disillusion — Alan Melikdjanian — built a YouTube channel around dismantling viral video fakery with the production values of the very content he was exposing. These were not institutions. They were individuals and loose collectives, self-funded, motivated by something that looked a lot like genuine frustration with the spread of nonsense.
James Randi was the most famous of them. A magician by training, Randi spent decades exposing faith healers and people claiming paranormal powers, eventually founding the James Randi Educational Foundation. His method was theatrical — he understood that the performance of debunking mattered as much as the evidence itself. Penn and Teller understood this too. So did Adam Conover, whose television format turned the correction into entertainment. The Committee for Skeptical Inquiry published the Skeptical Inquirer. The Skeptics Society, founded by Michael Shermer, gave the movement academic credibility. Stephen Barrett built Quackwatch to target medical misinformation specifically. The infrastructure of professional skepticism, by the early 2000s, was real.
What that infrastructure was fighting had its own ancient pedigree. In 44 BCE, Cicero wrote *De Divinatione*, a systematic takedown of the Roman practice of reading omens and entrails for prophecy. Around 160 CE, Sextus Empiricus was dismantling the claims of astrologers. A decade later, Lucian wrote *Alexander the False Prophet* — a detailed, contemptuous exposé of Alexander of Abonoteichus, a cult leader whose god Glycon turned out to be an elaborate snake puppet operated through hidden mechanisms, with hired thugs available for anyone who asked too many questions. The targets change. The structure of the con does not.
---
The word "debunk" itself has a specific birthday. In 1923, American journalist W. E. Woodward published a novel called *Bunk* and, in doing so, coined the verb: to debunk, meaning to take the bunk out of things. The American Heritage Dictionary confirms that "bunk," "debunk," and "debunker" all entered American English that year, though "bunk" itself traces back to 1828, when North Carolina representative Felix Walker gave a speech so tedious and irrelevant during the 16th United States Congress that it became synonymous with empty political noise — he was speaking, he explained, for Buncombe County. Buncombe became bunkum. Bunkum became bunk. And bunk became the thing that debunkers existed to remove.
The modern challenge, though, was never about finding the bunk. The bunk was everywhere, obvious, abundant. The challenge was what happened after you pointed at it.
Lewandowsky and Cook's *Debunking Handbook* — first published and then significantly revised in 2020 — laid out the problem in clinical terms. The backfire effect: the documented phenomenon where presenting someone with a correction to a false belief causes them to hold that belief more firmly, not less. The psychological mechanism is cognitive dissonance. The correction threatens not just the belief but the identity built around it. The mind, under that pressure, sometimes doubles down.
The 2020 revision complicated the picture. Backfire effects, the updated handbook stated, occur only occasionally, and the risk is lower than the original research had suggested. But "only occasionally" is not the same as "never." And in an environment where a single viral post reaches millions of people, occasional is still a very large number.
Cook and Lewandowsky theorized something else, too — something about the structure of human cognition that debunkers had been learning empirically for decades without having the language for it. People, they argued, abhor an incomplete explanation. If you remove a false narrative and offer nothing in its place, the mind reaches back for what it had. The debunker's job is not just to disprove. It is to replace. A better story has to be available, or the old one returns.
---
What investigators in the skeptical community confirmed, through decades of documented work, is that the mechanics of specific hoaxes are almost always findable. The Glycon puppet. The faith healer's earpiece. The controlled demolition claims about the World Trade Center, which the National Institute of Standards and Technology addressed directly. *Popular Mechanics* ran the numbers. MythBusters — Jamie Hyneman and Adam Savage — built the experiments on camera. The evidence, in most cases, was not the problem.
What remained contested was whether the evidence changed minds at the scale that mattered. The psychological literature on this is genuinely unsettled. The 2020 *Debunking Handbook* moved the consensus toward cautious optimism, but the debate between researchers about methodology and effect size has not resolved cleanly.
The community came to believe — and this is speculative, drawn from the theorizing of Lewandowsky and Cook rather than hard outcome data — that the manner of debunking matters as much as the content. Affirming a person's identity before challenging their belief. Providing the alternative narrative. Avoiding the repetition of the false claim in ways that inadvertently reinforce it. These are craft questions as much as scientific ones, and the answers are still being worked out.
---
Today, the infrastructure of professional skepticism remains active. Metabunk still runs. Snopes still publishes. The *Debunking Handbook* is available as a free PDF. The James Randi Educational Foundation's legacy continues through the communities it trained. Captain Disillusion still posts videos.
The paradox Lewandowsky and Cook identified has not been resolved — only refined. Debunking works, sometimes, under the right conditions, delivered in the right way, to people who are not yet fully committed to the belief being challenged. That is a narrower window than anyone in the business would prefer.
The question that lingers is the oldest one. Lucian wrote his exposé of Alexander of Abonoteichus around 170 CE. The Glycon cult outlived him by decades. The puppet was exposed. The believers kept believing. And somewhere in that gap — between the proof and the persuasion — the real mystery lives.
SOURCES