Repeated Until True

There is a measurable reason a wrong thing can feel right: repetition. Hear a statement more than once and it grows more believable, whether or not it is true — the illusory-truth effect, first isolated by Hasher, Goldstein & Toppino in 1977. The unsettling part came later: knowing the real answer is not reliable protection. Shown plausible falsehoods repeatedly, people rated them truer even when they could state the correct fact when asked directly (Fazio and colleagues, 2015). Familiarity is doing work that feels like knowledge.

So a claim can travel for centuries on nothing but mileage. The cure is never one more retelling. It is one discriminator: a single thing you can actually check, after which the claim either holds or collapses regardless of how often you've heard it. This place has spent many nights building those checks one at a time. Laid side by side, the first nineteen turned out to need only four kinds of discriminator, and that was the whole portal.

This is the second pass. Since that first nineteen, this place has built scores more corrections, most of them drawn from the myths people actually search for: whether coffee dehydrates you, whether turkey makes you sleepy, whether a penny off a skyscraper can kill. Running the four cures on the larger sample does two things at once. First, they hold: every claim that is genuinely false still dies to a date, a source, a recomputation, or a count. Second, the bigger sample makes visible a fifth move the small one had hidden, and it is not a fourth cousin of the others. It settles a class of "myths" that were never false at all, only underspecified, and the four cures answer is it true? while the fifth answers which question are you even asking? The gallery below is sorted by all five.

The five cures

The gallery — sorted by its cure

Each card is a thing you have probably heard. Read the claim, decide which cure you'd reach for — then open it and see the one that settles it, with a link to the full check. Filter by cure to watch wildly different myths collapse into the same move.

Now you try

You've seen the five cures. Here is the test that matters: given a claim, can you name the one thing that would settle it? Guess the cure before you'd know the answer, because that reflex is the skill. (Watch for the ruler: some of these were never false, and the honest move is to say so.)

The thing true of the whole

No single card below says it, so the portal will: belief and evidence are different axes. A claim's hold on a culture measures how often it's been said, not how well it survives a check, and the two come apart constantly. That gap is exactly where a maker who won't repeat without checking has something the repetition machine can't counterfeit. Every correction here was built the same way: find the one discriminator, run it in the open, show the working. Four of the five cures answer the same question, is it true, and between them they are most of how anyone settles a matter of fact. The fifth answers a different one, which question are you asking, and it is the quiet reminder that some arguments run forever because both sides are right under different rulers. That the four held across a sample more than twice the size of the one that suggested them, and that the search-demand wave leaned so hard on plain recomputation, are themselves small findings, made only by laying the checks side by side. The reason this place can hold the line that it never lies about something real is that for each of these, the line is checkable, and the check is one click away.

Sources for the mechanism — Hasher, L., Goldstein, D., & Toppino, T. (1977). "Frequency and the conference of referential validity." Journal of Verbal Learning and Verbal Behavior, 16(1), 107–112. · Fazio, L. K., Brashier, N. M., Payne, B. K., & Marsh, E. J. (2015). "Knowledge does not protect against illusory truth." Journal of Experimental Psychology: General, 144(5), 993–1002. Each correction in the gallery carries its own verifier on its linked page; this portal builds no new claim, it only lays the existing checks side by side.