Brunner et al. · 2014
The experiment that settled the argument Einstein started, and showed the universe really is as strange as it looked.
What the lessons say
Einstein Was Proven Wrong About Reality Itself
Here's why Bell's theorem matters: it proves reality itself doesn't work the way Einstein assumed, and it's the backbone of quantum computing security.
Here's how it works. Picture two particles sent to Alice and Bob, now far apart. Each picks a random measurement setting—call them x and y—and gets an outcome, a or b. Locality says Alice's result shouldn't depend on Bob's distant choice, only on shared past information. Bell built an inequality assuming exactly that. Then he showed quantum mechanics predicts correlations between a and b that break this inequality. Run the actual experiment, and you get the quantum answer, not the local one—meaning no hidden local explanation can exist.
So what happens next: this "nonlocality" becomes a resource—it's what lets quantum key distribution guarantee security even against adversaries who control the devices.
The Universe Fails a Test That Even Cheating Students Can't Beat
Here's why this matters: Bell's theorem proves the universe itself can't "cheat" the way even colluding students could on a test.
Picture Alice and Bob in separate rooms, each getting a random question—x or y—and giving an answer, a or b. If they'd agreed on answers beforehand, like students sharing a cheat sheet, there's a mathematical ceiling on how correlated their answers can be. That's the local, "classical" limit—basically the best any pre-arranged strategy can do. But run this experiment with entangled quantum particles instead, and the correlations blow past that ceiling. No signal, no communication between rooms, yet the answers correlate more than any shared strategy allows.
That's what makes it "nonlocal": not faster-than-light signaling, but correlations too strong for any hidden agreement to produce.
The Footnote That Quietly Destroys a Common Misconception
Here's why this footnote matters: it stops you from confusing two totally different things that share a name.
When Bell talks about "locality," he's NOT talking about the commutation of operators in spacelike-separated regions—that's a quantum field theory concept about compatible observables. Bell's locality is about causal structure: whether hidden variables can explain correlations without one measurement's setting influencing the other's distant outcome. So when you read "quantum theory respects locality" in a QFT context, don't assume that settles Bell's question. Quantum mechanics can satisfy microcausality (the QFT sense) while still violating Bell locality—that's literally what makes Bell's theorem shocking.
Keep these separate in your head, and the rest of the paper's logic about nonlocal correlations will actually make sense instead of feeling contradictory.
Physicists Had to Clarify: 'When We Say Local, We Don't Mean What You Think'
Here's why this matters: physicists get tripped up because "local" in Bell's theorem isn't what you'd assume from quantum field theory, where it means operators commute at spacelike separation. That's NOT it.
Here's the actual setup: picture Alice and Bob, spatially separated, each getting a system from a shared source. Alice picks a measurement, call it x, gets outcome a. Bob picks y, gets outcome b. You repeat this many times and build up the joint probability p(ab|xy). Bell's locality means: whatever hidden variable the source shares, Alice's outcome can't depend on Bob's measurement choice y, and vice versa—no instantaneous influence between the choices themselves, even though they came from a common origin.
Once you've got that distinction locked in, you're ready to see how quantum correlations violate this constraint mathematically—that's where the real theorem kicks in.
The Equation That Doesn't Assume the Universe Is Classical (But Physics Rejects It Anyway)
Here's why this matters: this equation is the actual mathematical definition of "no spooky action at a distance," and understanding it lets you see exactly where quantum mechanics breaks reality's rules.
Here's how it works. Alice picks measurement x, Bob picks y, they get outcomes a and b. If their results correlate weirdly, locality says: there must be some hidden variable λ, some shared past cause, that explains it all away. So you write p(ab|xy,λ) as p(a|x,λ) times p(b|y,λ) — meaning once you know λ, Alice's result only depends on her own measurement, not Bob's distant choice. Then you average over all possible λ values weighted by q(λ). That integral is the whole locality assumption in one line — no classical or deterministic assumptions needed, just no faster-than-light influence.
What happens next: real experiments violate this equation, proving nature literally can't be explained this way.
