How Science Changes Its Mind

Critics sometimes say science cannot be trusted because "they keep changing their minds." The observation is true; the conclusion is backwards. A system that never revised its views in the face of new evidence would not be reliable — it would be dogma.

Understanding how science changes its mind lets you tell healthy self-correction from mere flip-flopping, and lets you weigh a shifting field without either clinging to the old view or lurching after every headline.

Consensus is earned and provisional

Scientific consensus is not a vote or a fashion; it is the settled weight of many independent lines of evidence pointing the same way. It can be extremely strong — and it remains, in principle, revisable if the evidence turns. Strong consensus and openness to revision are not in tension; they are how a durable field behaves.

Normal science

Most science, most of the time, is what the philosopher Thomas Kuhn called normal science: careful puzzle-solving within an accepted framework. Researchers refine measurements, extend a theory to new cases, and tidy loose ends. Results shift a little, error bars shrink, details get corrected. This is the ordinary, incremental self-correction that a working field runs on.

Paradigm shift

Occasionally the anomalies — observations the framework cannot absorb — pile up until the framework itself gives way and is replaced by a new one. Kuhn called this a paradigm shift: not a tweak but a change in the basic picture, after which old problems look different and new questions become askable. Shifts are rarer and harder-won than normal science, and they are where the phrase "science changed its mind" really applies.

When a long-accepted result is overturned, ask which kind of change it is. A recalibrated measurement or a corrected effect size is normal science doing its job — expect small, frequent adjustments. A wholesale replacement of the underlying explanation is a paradigm shift — expect these to be rare, contested, and eventually settled by a large body of converging evidence rather than a single dramatic experiment. Sorting a news story into the right bucket tells you how much to update.

"Science changes, so nothing is settled" does not follow. That everything is revisable does not make everything equally uncertain. The age of the Earth, the germ origin of infectious disease, and common descent are not on the verge of reversal; they are supported by mountains of independent evidence. Openness to revision is a standing policy, not a prediction that any particular well-founded conclusion will fall.

In 1912 the German meteorologist Alfred Wegener proposed that the continents had once been joined and had drifted apart, citing the jigsaw fit of coastlines and matching fossils and rock formations across oceans. He expanded the idea in his 1915 book The Origin of Continents and Oceans. Most geologists rejected it for decades, largely because Wegener had no convincing mechanism for how continents could move. He died on a Greenland expedition in 1930, his hypothesis still out of favor. Then, in the 1950s and 1960s, new evidence — seafloor spreading and the striped pattern of magnetic reversals recorded in ocean-floor rock — supplied the missing mechanism, and the theory of plate tectonics took hold. A rejected idea was vindicated because the evidence finally arrived. That arc, and Kuhn's The Structure of Scientific Revolutions (1962) which named the pattern, is what a paradigm shift looks like.

• "Scientists keep changing their minds, so science is unreliable" — revision on evidence is exactly what makes it reliable.
• Every reversal is a revolution — most change is incremental normal science, not a paradigm shift.
• A paradigm shift happens overnight — the drift-to-tectonics change took roughly half a century.
• Because views can change, all views are equally uncertain — well-supported conclusions are revisable in principle but not fragile.

Sort these into "normal science" or "paradigm shift": a refined estimate of a physical constant; the move from an Earth-centered to a Sun-centered solar system; a corrected drug dosage after new trials; the shift from fixed continents to plate tectonics. For each, name the tell that placed it in that category.

Think Like a Maester: When science reverses a claim, ask whether the evidence changed — if it did, the reversal is the method working, not failing.

1. Why is a willingness to revise conclusions a strength rather than a weakness?
2. What distinguishes normal science from a paradigm shift?
3. Why was Wegener's continental-drift hypothesis rejected for decades?
4. What new evidence in the 1950s-60s led to acceptance of plate tectonics?
5. Why does "science changes, so nothing is settled" not follow?

Science changes its mind on purpose. Most of the time it does so incrementally — normal science refining and correcting within an accepted framework. Occasionally anomalies accumulate until the framework itself is replaced in a paradigm shift, as Kuhn described in 1962. Wegener's continental drift, dismissed for want of a mechanism and then vindicated once seafloor evidence arrived, shows both the patience and the power of the process. Revision tracks evidence; that is the feature, not the flaw.

• Peer Review and Replication
• Science and Society