Solvay 1927
The Argument That Took Fifty Years to Settle
Brussels, October 1927
In October 1927, twenty nine people gathered in Brussels for the fifth Solvay Conference. Seventeen of them either had or would receive a Nobel Prize. Curie, Einstein, Bohr, Planck, Dirac, Heisenberg, Schrödinger, Pauli, de Broglie, Born, Lorentz, Compton. The group photograph taken outside is probably the most reproduced image in the history of physics, and it is worth knowing what they were actually arguing about, because the argument is not what most retellings suggest.
They were not arguing about whether quantum mechanics works. Everyone in the room accepted that it predicted experiments beautifully. The argument was about whether it was the whole story, and it split the most capable group of physicists ever assembled into two camps that never reconciled.
Two Positions, Both Reasonable
Bohr’s position was that the theory is complete, and that the discomfort comes from expecting quantum objects to have properties when nobody is measuring them. On this view, asking what an electron’s position is between measurements is not a hard question. It is a question with no content. What exists is the experimental arrangement and the outcome, and the theory relates them perfectly.
Einstein’s position was that this is a description of our knowledge rather than of the world. If a theory only gives probabilities, either nature itself is genuinely undecided, or there are details the theory has not captured. He believed the second, and he believed it not out of nostalgia but because the alternative appeared to require influences that reach across space instantly.
It is worth being fair to both. Einstein was not confused about quantum mechanics; he had done as much as anyone to create it, and his objection was precise. Bohr was not dodging; his position was a serious philosophical claim about what physics is for. Neither was being stubborn in the way the story is often told.
Breakfast, and Then Dinner
The famous exchanges did not happen in the sessions. They happened at meals. Einstein would arrive at breakfast with a thought experiment designed to beat the uncertainty principle – some arrangement of slits, shutters, and screens that would measure two incompatible quantities at once. Bohr would spend the day taking it apart and present the answer by dinner. This went on for the length of the conference, and again three years later.
The best of them came in 1930. Einstein proposed a box full of light with a shutter on a clock, hanging from a spring. Open the shutter briefly, let one photon out, weigh the box before and after. Mass difference gives the photon’s energy exactly; the clock gives the time exactly; and energy and time are supposed to be a pair you cannot pin down together.
Bohr reportedly had a very bad night and an excellent morning. Weighing the box means letting it move in the gravitational field. In Einstein’s own general relativity, a clock’s rate depends on where it sits in that field. So the act of weighing introduces an unavoidable uncertainty in the clock’s reading, and it works out to exactly the amount required. Einstein’s argument was defeated with Einstein’s own theory, which is the kind of thing that gets a story retold for a century.
After this Einstein stopped attacking the theory’s consistency. He accepted that quantum mechanics does not contradict itself. He shifted to the far more dangerous claim that it is incomplete.
Einstein, Podolsky and Rosen
In 1935 Einstein, Podolsky and Rosen published the argument that mattered. Prepare two particles together so their properties are linked, then separate them by any distance you like. Measure one. The theory says the other’s corresponding property is instantly settled.
Their reasoning was tight. Either the measurement on the first particle physically affects the second across an arbitrary distance instantly, which conflicts with relativity’s insistence that nothing outruns light, or the second particle had that property all along and quantum mechanics simply does not track it. They regarded the first option as absurd, so they concluded the second: the theory is incomplete, and there are hidden details underneath.
Bohr replied within months, essentially arguing that the two particles are one indivisible system and it is a mistake to speak of the far particle as having separate properties at all. Most physicists considered the matter closed and went back to work. The theory kept predicting everything correctly, and the disagreement was filed as philosophy. That verdict held for nearly thirty years, and it was wrong.
Bell Turns Philosophy Into an Experiment
In 1964, John Bell found a way to put the disagreement to an experiment. He took the assumptions Einstein, Podolsky and Rosen had treated as obviously true – that the particles carry their properties with them, and that a measurement here cannot influence a result there – and derived a limit. Any theory obeying both assumptions must keep its correlations below a specific number. Quantum mechanics predicts correlations above it.
The importance of this is hard to overstate. A question that the entire profession had classified as untestable metaphysics turned out to have a number attached. You did not have to argue about what reality is. You could set two detectors at particular angles, count coincidences, and read off which side of the line nature sits on.
The experiments were hard. Clauser and Freedman managed a first version in 1972. Aspect’s experiments in the early 1980s switched the detector settings while the particles were already in flight, so no signal traveling at light speed could have carried the choice from one side to the other. Later work closed the remaining gaps one at a time, including using light from distant quasars to choose the settings, so that any conspiracy would have had to be arranged billions of years ago.
The Verdict
Nature exceeds Bell’s limit. It has done so in every properly conducted experiment, by wide margins, and the loopholes that once allowed a determined skeptic to object have been closed one after another. One of them cannot be closed, only pushed further away. If the detector settings were somehow fixed in advance along with the particles themselves, no experiment can exclude it. Pushing is what the quasar version does, and billions of years is about as far as pushing goes. Clauser, Aspect and Zeilinger shared the 2022 Nobel Prize for the work.
So Einstein’s conclusion was wrong. There is no set of hidden properties carried along by each particle that reproduces what is observed, at least not while also keeping influences local. One of the two assumptions in that 1935 paper has to go, and most physicists give up the idea that the particles carried definite properties independently all along.
But the framing is worth getting right, because Einstein comes out of this better than the scoreboard suggests. Bohr’s camp had declared the question meaningless. Einstein insisted it was a real question about the world. Bell then proved it was a real question by turning it into an experiment – and the experiment could only exist because someone had refused to accept that there was nothing there to ask. Einstein lost the answer and won the argument about whether the argument was worth having.
What Is Still Not Settled
It is easy to overstate what the Bell experiments closed. They rule out a specific and important class of explanations. They do not tell you which interpretation of quantum mechanics is correct, and physicists still genuinely disagree about that.
Nor does the result permit sending a signal faster than light. The correlations only become visible when the two sets of results are compared afterwards, and that comparison travels no faster than anything else. Relativity’s speed limit survives intact, which is a strange and specific kind of peace treaty between the two theories.
And the deeper worry that motivated Einstein has not gone away. Quantum mechanics and general relativity still do not fit together, and the tension the Solvay arguments circled – about what is real, and what a physical description is describing – shows up again wherever the two theories meet.
Handling a Question You Cannot Yet Answer
This is a story about how physics handles a question it cannot yet answer. For thirty years the profession’s consensus was that this one was empty, and the consensus was held by extremely good physicists for defensible reasons. It took one person taking the discarded question seriously enough to look for a testable consequence, and the consequence was there the whole time.
That is worth remembering when reading anything on this site marked as an open frontier. Some questions really are unanswerable. Others merely have not had their Bell yet, and telling the two apart in advance is not something anyone has ever been good at.
