When the Evidence Can’t Decide

Logic Truth and the Limits of Reason · Open board

The map of how arguments reach their conclusions. This lesson takes the arrow that runs from a theory to a prediction and asks what you are allowed to conclude when the prediction misses.

Before you start. Logic, Truth & the Limits of Reason showed that a rule built from past cases is never proved, only made likely, because the next case can break it. Its second read gave Karl Popper’s answer: a theory counts as scientific if some observation could prove it wrong, and you keep it only while it survives honest attempts to break it. This lesson tests that answer against two real planets, one that was there and one that was not.

The idea in 30 seconds

A theory never makes a prediction on its own; it always needs extra assumptions, such as “the instruments work” and “nothing unseen is interfering”.
When a prediction fails, logic only says that something in that bundle is false, not which part, so the evidence alone cannot force you to drop the theory.
The same rescue move, blaming a hidden cause, found the planet Neptune in 1846 and invented a planet that never existed in 1859.
Most everyday science works inside a shared framework and treats a misfit result as an unsolved puzzle; frameworks get replaced only when misfits pile up.
A rescue earns its keep when it predicts something new that is then observed, and loses it when it only explains away the failure it was invented for.
No single test separates science from opinion, but science states what would count against it, exposes its assumptions, and is checked by people who want it to be wrong.

Read the story · Skip to the numbers