Belfast and Geneva, 1928–1990
John Stewart Bell
John Bell worked by day on particle accelerators at CERN, and in his spare time asked whether Einstein could have been right. In 1964 he proved that the answer could be tested, and experiment answered: not in the way Einstein hoped.
The view
Bell was impatient with the vagueness at the heart of quantum mechanics as taught. The theory divided the world into a quantum system and a classical measuring apparatus, with no rule for where the line fell. He called for a theory of "beables", things that really exist, from which the results of measurements would follow without measurement having to be mentioned in the fundamental laws.
Reading Bohm's 1952 paper convinced him that such theories were possible, and that the proofs claiming otherwise were wrong. But Bohm's theory was non-local. Bell set out to find whether that was avoidable.
After Bell (1964): if a hidden-variable theory is local, it will not agree with quantum mechanics.
The argument
Take the EPR set-up: two particles prepared together and sent to distant detectors. Each detector can be set at one of several angles, chosen at the last moment, and each records yes or no. Suppose, with Einstein, that each particle carries its answers with it, like a glove in a box, and that nothing at one detector affects the other. Then the correlations between the results must obey a certain inequality, whatever the hidden variables are. Quantum mechanics predicts that the inequality is broken.
Experiments by John Clauser in 1972, Alain Aspect in 1982 and many since, with the last loopholes closed in 2015, found the inequality broken by just the amount quantum mechanics predicts. The 2022 Nobel Prize in Physics went to Aspect, Clauser and Anton Zeilinger for this work.
Where he meets Deutsch
Bell and Deutsch agree that measurement should not be a basic term in physics, and that a theory must say what exists. Both regard the theorem as one of the deepest results in physics.
Where they part
Bell's theorem assumes, without stating it, that each measurement has a single outcome. Given that, the violation of the inequality means that nature is non-local, and Bell drew that conclusion. Everett's theory drops the assumption. Each detector's outcome is not single: the observer at each end comes to exist in versions that saw yes and versions that saw no. The correlations are established locally, when the records from the two ends are brought together.
So Deutsch reads Bell's theorem as one of the strongest arguments for the multiverse. You can have locality, or one world, but not both. Bell chose one world. Deutsch keeps locality.
| Bell | Deutsch | |
|---|---|---|
| What is ultimately real? | He wanted "beables": things that exist, rather than observables that are measured. | Whatever our best explanations refer to: the multiverse, and emergent and abstract things too. |
| How do we know what is real? | By theory and experiment, without privileged observers. Measurement should not be a primitive. | Conjecture and criticism. If our best explanation needs it, it exists: if it kicks back, it is real. |
| One world or many? | One world, he assumed, which forced the conclusion that nature is non-local. | Many. Our universe is one layer of a multiverse, revealed by quantum interference. |
| Is reality independent of observers? | Yes. He thought the role given to observers in quantum theory a scandal. | Yes. Observers are physical systems inside reality, governed by the same laws. |
Further reading
- J. S. Bell, Speakable and Unspeakable in Quantum Mechanics (1987; second edition 2004), all his essays on the foundations.
- J. S. Bell, "Against 'measurement'", Physics World (1990).
- David Deutsch and Patrick Hayden, "Information flow in entangled quantum systems" (2000).
- Bell's Theorem, Stanford Encyclopedia of Philosophy.