Matthew Pusey set out to help Albert Einstein and ended up finishing him off.
In 2011 Pusey was a graduate student at Imperial College London, working on theories in which the wave function is not a physical object but a state of incomplete knowledge — Einstein's own position, rewritten in modern notation. With Jonathan Barrett and Terry Rudolph he spent days trying to strengthen that picture. He got fed up and decided to prove it impossible instead. The result, published in 2012 and known ever since as PBR, is the subject of a Quanta Magazine feature published today — and it closes a century of bargaining over what quantum mechanics is allowed to mean.
The bargaining went like this. Bell's theorem, in 1964, showed that no theory which gives particles definite properties can also be local. That left two exits: keep reality and accept that influences travel faster than light, as Bohmian mechanics does; or keep locality and abandon objective reality, as Bohr's Copenhagen interpretation had already suggested. A third path seemed to survive, and it was the one Einstein appeared to favour. Particles really do have properties. The wave function is simply our information about them — a weather forecast, not the weather.
PBR shuts that path. Its argument is a card game.
A dealer takes the 26 red cards from one deck and the four aces from another. He picks two cards at random, one at a time, from either pile, and lays them face down. The player must name one combination of piles that the two cards cannot both have come from. Turn over the two of diamonds and the ace of clubs — a red non-ace and a black ace — and the answer is easy: the dealer cannot have drawn twice from the red pile, nor twice from the ace pile. Turn over the ace of diamonds and the ace of hearts, and the player has nothing. Both cards could have come from either pile, so any answer is a guess with three-in-four odds, and over enough rounds the player must lose.
That is the difference between a real state and a state of knowledge. "Ace of diamonds" is written on the card. "Drew from the red pile" lives only in the dealer's head. The player can lose precisely because some real states — the red aces — belong to two different states of knowledge at once. They overlap.
PBR poses the quantum version and asks whether quantum states can overlap like that. Its answer, under two assumptions, is no. The assumptions are that particles have real states at all, and that the real state of one particle does not depend on how the other was prepared — a condition called preparation independence, and a different demand from locality.
Given those, a quantum machine can always win the game. Scan two unknown quantum states with the right apparatus and you can always rule out at least one pile combination, without ever measuring them. A 2012 experiment confirmed it. Last autumn, a team led by Crispin Barnes at Cambridge played an expanded five-card version on an IBM quantum computer and won every round. The motivation was less philosophical than practical: the researchers were using a theorem from quantum foundations as a test that their hardware was genuinely quantum.
So the wave function is not a state of ignorance. Something in the world changes when it changes. Pusey's own summary: if anything is real, the quantum state is real.
But that is where the comfort stops, because PBR does not say reality is the thing that survives. It rules out exactly one arrangement — a real world described by a quantum state that is merely informational — and leaves both extremes intact. Either reality is objective and thoroughly surreal, with collapse corresponding to some physical event nobody can picture. Or nothing is real in Einstein's sense, including the wave function, and the redness of a billiard ball genuinely comes into being with the act of looking at it.
That is the question Einstein put to Abraham Pais — do you really believe the moon is not there when you are not looking at it — still unanswered, but with the middle ground now fenced off.
Some physicists have taken the first door. Others, like Jacques Pienaar at the Federal University of Rio de Janeiro, took the second: PBR, he has said, convinced him the programme of saving objective reality was dead, and he moved to QBism, which holds that quantum mechanics describes an agent's experience rather than the world. Lucien Hardy at the Perimeter Institute is among those picking at the theorem's fine print, asking whether a state of reality must be a single definite thing at a single moment, or whether it might be smeared across time. He does not yet have a proposal.
There is one more honest caveat. Almost nobody expected the Cambridge experiment to fail; the theorem was taken as given. What that experiment demonstrated is that PBR can be turned into a diagnostic for quantum hardware — a use that may outlast the metaphysics. Meanwhile the metaphysics sits exactly where it was in 1935, minus one option.
