Plain-language answer
Philosopher Karl Popper proposed that what distinguishes a genuinely scientific claim from a non-scientific one is not that it has been proven true, but that it could, in principle, be shown false by some possible observation. A claim compatible with any outcome whatsoever — one that would be “confirmed” no matter what actually happened — has not really made a testable prediction at all, whatever else its merits.(Popper, 1959)
Why it matters
Explanations for coincidences vary enormously in how falsifiable they are. “This happened by chance under a specified model” makes a testable prediction: given the stated model and reference class, an event of roughly this rarity should occur about this often, which can in principle be checked against a larger record of similar cases. By contrast, an explanation like “the universe arranges meaningful coincidences for people who are open to noticing them” is far harder to falsify — it is compatible with a coincidence happening, and equally compatible with one not happening (attributed to insufficient openness), leaving no observation that could count against it.
Worked example: two candidate explanations, two different tests
Suppose someone claims a specific object consistently “attracts” lucky coincidences to its owner. A falsifiable version of this claim would specify in advance what counts as a “lucky coincidence,” over what time period, and at what rate compared with a matched control group without the object — producing a prediction that a systematic test could confirm or refute.
An unfalsifiable version of the same claim adjusts after the fact: a run of good luck confirms the object’s power, and a run of bad luck is explained away (the object needs “recharging,” the owner wasn’t “open” enough, an unrelated negative influence interfered). Because every possible outcome is absorbed into the theory as consistent with it, no observation could ever count as evidence against it — which is precisely the feature Popper identified as the mark of an unfalsifiable claim, whatever its other appeal.(Popper, 1959)
This connects directly to post-hoc probability: a claim that only specifies its exact criteria for success after seeing the outcome is functioning as an unfalsifiable claim in practice, even if a falsifiable version could, in principle, have been stated in advance.
Common misconception
“That claim can’t be falsified, so it’s definitely false” is itself a misuse of the criterion. Falsifiability is a test for whether a claim is making a scientifically testable prediction, not a test for whether the claim is true or false. An unfalsifiable claim about coincidences might still be true — the criterion simply says that science, as a method, has no way to test it either way.
Limits and open questions
Falsifiability draws a useful line between claims and shows which ones science’s methods can engage with, but real theories are rarely purely falsifiable or purely unfalsifiable — auxiliary assumptions, measurement error, and reasonable hedging all complicate how cleanly a single failed prediction actually refutes a broader theory. Philosophers of science have continued to refine and debate Popper’s original criterion since 1959, and this page presents it as an influential, widely used starting point rather than a final, uncontested account of scientific method.
Related
- How science tests unusual claims situates falsifiability as one item within a broader practical checklist.
- Post-hoc probability covers the specific coincidence-relevant failure mode of specifying a claim’s success criteria only after the outcome is already known.
Key takeaways
- A scientific claim, on Popper’s influential criterion, is one that could in principle be shown false by some possible observation.
- An explanation that adjusts itself to be consistent with any possible outcome has not made a testable prediction, regardless of how compelling it feels.
- Falsifiability is a test of whether a claim is scientifically testable, not a direct test of whether it is true.