The Measurement Problem: Why the Observer Won't Go Away

Quantum mechanics is the most precisely tested theory we have. It is also, at its foundations, strangely silent about the thing every experiment actually produces: a single, definite result.

Two rules that don't fit together

Textbook quantum mechanics runs on two rules. Between measurements, a system's state evolves according to the Schrödinger equation: smooth, deterministic and linear. Linearity means superpositions stay superpositions — if an electron can be "here" and "there," it evolves into a combination of both. But when we measure, a second rule takes over: the state collapses to one outcome, with probabilities given by the Born rule.

The trouble is that measuring devices, and the people reading them, are also made of atoms. Apply the first rule consistently and the detector itself should end up in a superposition of "clicked" and "didn't click." We never see that. So where, physically, does the second rule take over? The theory does not say. That is the measurement problem.

Wigner's friend

Eugene Wigner made the puzzle personal. Imagine a friend measuring a quantum system inside a sealed lab. For the friend, there is a definite outcome. For Wigner outside, who treats the whole lab as a quantum system, the friend and the result may still be in superposition. Who is right? Recent extended versions of this thought experiment show that a handful of very plausible assumptions about observers and outcomes cannot all hold at once — some intuition has to give.

The main ways out

• Copenhagen-style views treat the quantum state as a tool for predicting measurement outcomes, with a "cut" between quantum system and classical apparatus that the theory does not derive.
• Many-worlds (Everett, 1957) keeps only the Schrödinger equation: every outcome happens, each in its own decohered branch, and "collapse" is what branching looks like from the inside.
• Pilot-wave theory (de Broglie, Bohm) adds definite particle positions guided by the wavefunction, so outcomes are always definite.
• Objective-collapse models (such as GRW) modify the dynamics so that large superpositions spontaneously collapse — a proposal that experiments are actively testing.
• Observer-centred views, including QBism and relational quantum mechanics, take the observer's role at face value rather than trying to remove it.

Why it matters

All these options reproduce the same experimental predictions in ordinary situations, which is why the debate can feel philosophical. But they disagree profoundly about what exists, about what an observer is, and about whether facts are absolute or relative. They also point to different physics at the frontiers — in gravity, cosmology and very large quantum systems. Taking the measurement problem seriously is, to my mind, the most direct route to understanding how a single, objective-seeming reality arises from a theory whose basic equations never pick one.

Further reading: H. Everett III, "'Relative State' Formulation of Quantum Mechanics" (1957); E. P. Wigner, "Remarks on the Mind-Body Question" (1961); D. Frauchiger & R. Renner, "Quantum theory cannot consistently describe the use of itself" (2018).