The Fact That Isn't a Fact for Everyone (Wigner's Friend)

Imagine a friend steps into a sealed lab and measures a single particle. It registers "up." For her, that is now a fact — a definite result, written down, real. You are standing outside, and you take quantum mechanics at its word: with the door closed, you have to describe the whole lab — your friend included — as one quantum system, still hovering in a blur of possibilities, no result settled yet. She says the measurement is done. You say it hasn't happened. And here is the strange part: nothing in the math forces one of you to be wrong.

This is Wigner's friend, a puzzle the physicist Eugene Wigner posed in the nineteen-sixties. For decades it sat in the category of things interesting to argue about over drinks — untestable, a matter of taste. Then it grew teeth.

In twenty eighteen, two physicists turned the puzzle into a theorem. They built a scenario where several observers each use quantum mechanics, correctly, to reason about each other — observers reasoning about observers. And they proved that if quantum theory applies to everything, including people using quantum theory, those observers can be led to flatly contradictory conclusions. One concludes the answer is definitely yes; another, reasoning just as carefully, concludes definitely no. Their paper's title says it without flinching: quantum theory cannot consistently describe the use of itself.

That was on paper. Then it went into the lab. Physicists ran a version of Wigner's friend using photons as the "friends" — tiny quantum systems standing in for observers, each making a measurement inside its own sealed corner of the experiment. They compared the outcomes the way you would run a Bell test, against a limit that any theory has to obey if measurement results are absolute facts, the same for everyone. The limit was broken. A second team tightened the logic, closed more of the escape routes, and broke it again.

So what has to give? Line up the assumptions. One: quantum mechanics works for the big and the small alike, observers included. Two: when someone measures something, they get one real outcome — a fact that is a fact for everybody. Three: no faster-than-light conspiracies, and experimenters can freely choose what to measure. These results say you cannot keep all three. At least one has to break — and most of the ways out are stranger than the problem.

Now the honest fine print, because this is where it gets oversold. The "friends" in these experiments are photons, not people. Whether a single particle counts as an observer making a genuine measurement is exactly the assumption under debate — and a respectable position is that it doesn't, that the paradox only truly bites for full-blown conscious observers no one can yet seal in a box. This is not a proof that reality is whatever you believe, or that looking at things conjures them into being, or that your mind rewrites the world. It is narrower, and stranger for being precise: the idea that there is one shared set of facts, identical for every observer, is not something quantum mechanics is willing to guarantee.

We move through life certain of one thing above almost all others — that what happened, happened, and it happened the same way for everyone who was there. Two people might remember a moment differently, but we assume there was one real event underneath. Wigner's friend is the quiet suggestion that, at the level where the world is actually assembled, even that may not hold. The measurement was done. The measurement was not done. Both, on the record, correct.

Sources

• Frauchiger & Renner, "Quantum theory cannot consistently describe the use of itself," Nature Communications 9, 3711 (2018). doi:10.1038/s41467-018-05739-8
• Proietti et al., "Experimental test of local observer independence," Science Advances 5(9), eaaw9832 (2019). doi:10.1126/sciadv.aaw9832
• Bong et al., "A strong no-go theorem on the Wigner's friend paradox," Nature Physics 16, 1199 (2020). doi:10.1038/s41567-020-0990-x
• Wigner, "Remarks on the Mind-Body Question," in The Scientist Speculates (ed. Good, 1961)