The Most Profound Result in Physics
Bell’s theorem is not about a particular physical system or a specific prediction. It is a mathematical proof about the structure of reality itself. It establishes that no theory satisfying certain intuitive assumptions about the world can reproduce all the predictions of quantum mechanics. The experiments confirm quantum mechanics. The assumptions fail.
The EPR Paradox
In 1935, Einstein, Podolsky, and Rosen (EPR) published a paper arguing that quantum mechanics must be incomplete. Their argument was precise and, on its face, compelling.
Consider two particles created together in an entangled state, then separated by a great distance. Quantum mechanics predicts that measuring a property of one particle instantly determines the corresponding property of the other, regardless of the distance between them. Measure the spin of particle A along some axis and the spin of particle B along that same axis is immediately determined.
EPR’s reasoning: either (a) the measurement of A physically influences B instantaneously (violating special relativity), or (b) the outcomes were determined all along by some underlying variables that quantum mechanics does not describe. EPR considered option (a) unacceptable. Therefore, they concluded, quantum mechanics is incomplete. There must be “hidden variables” carrying the predetermined outcomes.
Hidden Variable Theories
The hidden variable program attempts to restore the classical picture: particles carry definite values for all observable properties at all times, and measurements simply reveal those pre-existing values. The randomness of quantum mechanics would then be like the randomness of a coin flip, arising from ignorance of initial conditions rather than any fundamental indeterminacy.
For nearly three decades, this seemed like a matter of philosophical preference rather than experimental consequence. Quantum mechanics and hidden variable theories appeared to make the same predictions. Then Bell proved they do not.
Bell’s Inequality
In 1964, John Stewart Bell derived an inequality that any local hidden variable theory must satisfy. The argument is remarkably simple.
Consider two observers, Alice and Bob, each receiving one particle from an entangled pair. Each can choose to measure their particle’s spin along one of several axes. For each measurement, the result is either $+1$ or $-1$.
Define the quantity $S$ as a specific combination of correlations between Alice’s and Bob’s measurements along different axes:
$$S = E(a, b) - E(a, b’) + E(a’, b) + E(a’, b’)$$
where $E(a, b)$ is the average product of Alice’s result (measuring along axis $a$) and Bob’s result (measuring along axis $b$), and $a, a’, b, b’$ are four chosen measurement axes.
Bell proved that if the measurement outcomes are determined by local hidden variables, then:
$$|S| \leq 2$$
This is Bell’s inequality (in the CHSH form, named after Clauser, Horne, Shimony, and Holt, who derived this experimentally testable version).
The Quantum Prediction
Quantum mechanics predicts that for appropriately chosen measurement axes on a maximally entangled pair:
$$|S| = 2\sqrt{2} \approx 2.828$$
This exceeds Bell’s bound by more than 40%. The violation is not marginal. It is a clear, unambiguous conflict between quantum mechanics and any local hidden variable theory.
Experimental Confirmation
Aspect’s Experiments (1982)
Alain Aspect and colleagues performed the first rigorous tests of Bell’s inequality using entangled photons. They measured correlations between photon polarizations at separated detectors and found violations consistent with the quantum prediction. Crucially, they used rapid switching of measurement settings to close the “locality loophole”: the measurement choice on one side was made too late for a signal traveling at the speed of light to reach the other side before its measurement was complete.
Loophole-Free Tests (2015)
Earlier experiments left open technical loopholes. The “detection loophole” allowed the possibility that the detected subset of particles was unrepresentative. The “locality loophole” required that measurement choices and outcomes be space-like separated.
In 2015, three independent groups (Delft, NIST, Vienna) performed loophole-free Bell tests, closing all known loopholes simultaneously. The results confirmed the violation of Bell’s inequality. Local hidden variable theories are experimentally ruled out.
What Bell’s Theorem Means
The theorem forces a choice. At least one of the following assumptions must be abandoned:
Locality
The outcome of a measurement on one particle does not depend on what measurement is performed on a distant particle. This is the assumption that no physical influence propagates faster than light.
Realism
Particles possess definite values for all measurable properties at all times, independent of whether those properties are measured. Measurements reveal pre-existing facts.
Most physicists abandon realism (the properties do not exist until measured) while preserving a form of locality (no usable signal travels faster than light). Some interpretations (pilot wave theory) abandon locality instead. No mainstream interpretation preserves both.
The No-Communication Theorem
Despite the non-local correlations, entanglement cannot be used to transmit information faster than light. This is a provable theorem within quantum mechanics, not merely an experimental observation.
The reason: Alice’s local measurement results are completely random, regardless of what Bob does. The correlations only become visible when Alice and Bob compare their results, which requires classical communication (limited to the speed of light). Bob cannot determine, from his results alone, which measurement Alice chose to perform.
Non-locality without signaling. The correlations are stronger than any classical theory allows, but they do not permit faster-than-light communication. This is one of the most counterintuitive aspects of quantum mechanics.
Implications for Quantum Technology
Quantum Key Distribution
Bell inequality violations provide a device-independent test of security for quantum key distribution. If Alice and Bob observe a Bell violation, they can be certain that no eavesdropper holds a copy of their measurement outcomes, because the violation proves the outcomes were not predetermined. This is security guaranteed by the laws of physics rather than computational assumptions.
Quantum Computing
Entanglement, the resource that produces Bell violations, is also the resource that gives quantum computers their advantage over classical ones. A quantum computer that cannot produce Bell-violating correlations can be efficiently simulated by a classical computer. Bell violations are a necessary (though not sufficient) signature of quantum computational advantage.
Common Misconceptions
“Entanglement allows faster-than-light communication”
It does not. The no-communication theorem is rigorous. Entanglement produces correlations, not signals. The correlations are only visible after classical communication, which is limited to light speed.
“Bell’s theorem proves quantum mechanics is correct”
Bell’s theorem proves that local hidden variable theories are incompatible with quantum predictions. The experiments confirm that nature violates Bell’s inequality, consistent with quantum mechanics. The theorem does not prove quantum mechanics is the final theory; it proves that any replacement must also be non-local or non-real.
“The EPR paradox shows quantum mechanics is paradoxical”
EPR is not a paradox within quantum mechanics. It is a conflict between quantum mechanics and the assumptions of local realism. Since Bell and the subsequent experiments, we know that local realism fails. The “paradox” is resolved by accepting that nature does not conform to classical intuitions about locality and predetermined outcomes.