Science Puzzle
The Professor Said So
A world-famous professor with forty years of eminence states that a certain reaction cannot occur. An unknown first-year student produces careful, reproducible measurements showing that it does. The professor offers no data, only his judgement and his record.
How should this disagreement be settled, and does that mean expertise is worthless?
The Answer
The evidence settles it. If the student's measurements are sound and reproducible, the reaction occurs, and no amount of eminence changes that. Treating the professor's authority as though it were itself evidence is the appeal to authority fallacy, and it is a fallacy precisely because authority is not a measurement.
But the second half of the question is the more interesting one, and the overcorrection is a real trap. "Expertise counts for nothing" is not the lesson, and people who reach that conclusion end up believing anything, which is a worse epistemic position than deferring to professors.
Expertise is a very good guide to prior probability. The professor has seen a thousand students find a spectacular result that turned out to be contamination, a miscalibrated instrument, or a spreadsheet error, and he is right that this is the overwhelmingly common explanation. That is a legitimate and useful judgement. What it is not is a refutation. It tells the student where to look for a mistake; it does not establish that she made one.
So the correct behaviour from both sides is the same: go and look. The student should hunt hard for the error the professor expects, because he is usually right. The professor should ask what result would change his mind and then examine the data against it. If the measurements survive that, the reaction occurs, and forty years of eminence has to yield, which is exactly what happened to a great deal of settled physics in the early twentieth century.
The clean formulation: authority is a heuristic about where truth is likely to be found, and evidence is what determines where it actually is. Using the first to shortcut the second is the fallacy. Discarding the first entirely is not scepticism, it is just noise.
The principle: Appeal to authority. Credentials are a guide to which claims deserve attention and where errors are likely, but they are not evidence. Only evidence settles a claim, and rejecting expertise entirely is the opposite error.