Randomness, Determinism, and Free Will

Quantum mechanics says the world is random at its core. A radioactive atom decays at no predictable time; a measured qubit gives 0 or 1 with only probabilities to guide you. But what kind of random? There is a world of difference between "random because we don't know the cause" (a coin flip — the coin already landed) and "random because there is no cause." Quantum mechanics claims the second. This lesson is about whether that claim survives scrutiny — and why the answer matters for cryptography, computing, and free will itself.

Einstein never bought it. His 1935 EPR argument with Podolsky and Rosen said quantum mechanics must be incomplete: particles must carry pre-set "instructions" — hidden variables — that determine outcomes in advance, and the randomness is just our ignorance. For thirty years this was philosophy. Then in 1964 John Bell did something extraordinary: he turned it into arithmetic. If particles carry local pre-set instructions, then certain measured correlations — the CHSH combination — can never exceed 2. Quantum mechanics predicts 2.83. Somebody had to be wrong, and the question could be settled in a lab.

The labs spoke. Clauser and Freedman first (early 1970s), Aspect decisively (1981–82, choosing measurement settings after the photons left), the last loopholes closed in 2015 — and the 2022 Nobel Prize (Aspect, Clauser, Zeilinger) "for experiments with entangled photons, establishing the violation of Bell inequalities and pioneering quantum information science." Nature's measured value: ~2.8, matching quantum mechanics, smashing the limit of 2.

Now state the verdict precisely, because pop science mangles it: Bell did not prove "the universe is nonlocal" or "Einstein was wrong about everything." He killed local realism — a package deal of three assumptions: realism (outcomes pre-determined), locality (distant settings can't influence results), and statistical independence (settings are chosen freely, uncorrelated with the particles' past). The experiments force you to abandon at least one — and each exit is a live philosophical position. Give up locality: that's Bohm's pilot wave. Give up realism: that's Copenhagen and QBism. Give up statistical independence: that's superdeterminism. Accept all outcomes: that's many-worlds. The corpse is local realism; the suspects all walked free.

Where does free will enter? Through a remarkable theorem. Conway and Kochen's Free Will Theorem (2006, strengthened 2009) says: if experimenters have free will — defined minimally as making choices "not a function of the past" — then elementary particles have the same freedom: their responses aren't fixed by the past either. Note the crucial subtlety: mere randomness doesn't qualify, because a random outcome could have been pre-written in a cosmic table. The theorem rules out deterministic and pre-scripted-random theories alike. But be honest about the price: it assumes human free will as an axiom. It proves nothing about whether you have it — and critics (Goldstein and co.) dispute the argument. Also note the circularity the field lives with: Bell tests themselves assume the experimenters' setting choices are free.

That assumption is exactly what superdeterminism attacks — and it was Bell himself who named the escape. In a 1985 BBC interview he said: "Suppose the world is super-deterministic... the universe, including particle A, already 'knows' what that measurement, and its outcome, will be." Modern defenders Gerard 't Hooft and Sabine Hossenfelder argue that since experimenters and particles share a common past stretching to the Big Bang, the "free" setting choices may be subtly correlated with the particles — a cosmic selection bias. The price: absolute determinism, no free choice at all. The objections are fierce: the loophole "cannot be closed by scientific methods" because science itself assumes experimenters can freely choose settings; others call it a fine-tuned conspiracy that makes science pointless. Hossenfelder's pitch is blunt: "Superdeterminism returns us to determinism."

Why should a computing student care? Because randomness is a resource — and quantum physics can certify it. A Bell-inequality violation proves the outcomes were unpredictable even to whoever built the device: no pre-written table, however clever, can reproduce the statistics. Pironio and colleagues (2010) showed this yields randomness "without any assumption on the internal working of the devices" — "impossible classically." The state of the art: in 2026 the Renner group at ETH Zurich ran 1.5 billion Bell tests on entangled qubits and extracted provably perfect randomness with no assumed bound on the adversary's power. Compare your laptop's random numbers: a deterministic algorithm (unpredictable only if you don't know the seed) versus Bell-certified bits (unpredictable even if the hardware vendor is your enemy). Cryptography runs on the difference.

Common myth: "Quantum randomness proves we have free will." It doesn't — the Free Will Theorem assumes experimenter freedom to conclude anything, and superdeterminism denies the premise entirely. Randomness is not freedom; a pre-written random table is still a script.