Idea
A prediction must exclude something
Suppose a theory explains a success as proof and explains a failure as equally strong proof. It can absorb events indefinitely, but it has given you no way to test it. A riskier claim says in advance what should happen and makes clear which observations would create a problem.
Popper emphasized the logical difference between collecting favorable examples and exposing a theory to possible contradiction. A finite collection of white swans cannot establish that all swans are white; a genuine counterexample defeats that universal statement. Scientific work adds complications because measurements also depend on instruments and supporting assumptions.
A claim becomes testable when the world is allowed to disagree.
Ask for the observation that would count against the explanation before the result arrives.
Where it breaks down
One unexpected measurement does not automatically defeat an entire theory. The instrument may be wrong, the setup may be misunderstood, or an auxiliary assumption may fail. Falsifiability is a guide to testable content, not a machine that sorts every field into good science and nonsense.
How it connects
A theory needs possible opponents in the evidence, not only favorable examples.
This idea appeared in Giving an idea a chance to fail, the Involves connection for September 12, 2026, which asked: What would you have to observe to change your mind?
Check yourself
Why might scientists retain a theory after an unexpected reading?
They must also test the instrument and supporting assumptions. Exactly. Evidence bears on a whole experimental setup. Investigating a failure is different from inventing an excuse for every outcome.
What the sources establish
The logic of a counterexample differs from practical theory rejection, which also involves instruments and auxiliary assumptions.
Sources
Karl Popper (Stanford Encyclopedia of Philosophy), Sections on falsifiability, scientific method, and the distinction between logical and methodological falsification