John Bell · The inequality
Can a philosophical dispute be settled by experiment?
That time, that place
John Bell's day job was designing accelerators at CERN in Geneva. The interpretation of quantum mechanics was not his assignment, nor was it treated as a serious subject in the field.
The mood was: the numbers work, don't ask what they mean. A graduate student raising such questions would be steered away. Bell found this stifling, and took the problem up during a sabbatical year in the United States from 1963.
Why this question
He was not trying to refute Einstein. He began by granting Einstein's demand.
Assume just two things.
First, that the two particles leave already carrying their answers — ready for whatever is asked.
Second, that what you measure on one side does not affect the other.
What follows? Bell varied the orientations of the two detectors and worked out how often the two outcomes agree.
A ceiling appeared. If those two assumptions hold, a particular combination of the agreement rates at several angles can never exceed a certain number.
He then computed the same combination from quantum mechanics. At some angles, it exceeded that number.
What was found
For the first time the two positions gave different numbers.
For thirty years they had predicted identical results, which is why nothing could decide between them. But look at the correlations across several detector angles and they part company. Now you can run the experiment.
Bell had not found the answer. He had built the means of deciding it — returning to the laboratory a question that had been exiled to philosophy.
The paper appeared in a newly founded journal that folded not long after, and went almost uncited for years.
Whether hidden variables exist is a matter of taste, not something an experiment can settle
The calculation showing that the ceiling implied by local hidden variables differs from quantum mechanics' prediction
The first experiment came eight years later, and over thirty years its loopholes were closed one by one