Quantum measurement problem

A photon is one quantum of light. Send one through a beam splitter, a device with two exits, and put a detector at each exit. In an ideal trial, only one detector clicks. With a balanced splitter, many trials give about half the clicks at each detector.

Quantum theory predicts those odds. But when the photon and detectors are treated using the same quantum rules, the description contains both possible records together. How does that description account for the definite result of a trial?[3] [7]

A detector reports one outcome in each ideal trial. If we apply the same quantum rules to the photon, detector and surroundings, why does the description contain several possible records? A simplified two-route model makes that tension explicit.

Write the incoming path state as

$$\begin{gathered}|\psi\rangle=\alpha|L\rangle+\beta|R\rangle,\\|\alpha|^2+|\beta|^2=1.\end{gathered}$$

Here $L$ and $R$ label the two paths, and $\alpha$ and $\beta$ are complex amplitudes. Their squared magnitudes give the Born probabilities for ideal path detection. For an equal splitter, both probabilities are one half. An amplitude describes a possible contribution to the quantum state; it does not mean half a detector click.

Unitary evolution is the ordinary quantum rule that preserves the state’s total probability. Write $|D_L\rangle$ and $|D_R\rangle$ for the left and right detector records. If evolution sends each path to its matching record, linearity sends their superposition to

$$\alpha|D_L\rangle+\beta|D_R\rangle.$$

These record states include the photon’s absorption. Adding the surroundings gives

$$\begin{aligned}|\Psi\rangle={}&\alpha|D_L\rangle|E_L\rangle\\&+\beta|D_R\rangle|E_R\rangle.\end{aligned}$$

The environmental states $|E_L\rangle$ and $|E_R\rangle$ carry information about the corresponding records. To calculate observations confined to the apparatus, we use its reduced density matrix, which sets aside the environment’s unobserved details. The term describing interference between the apparatus records is multiplied by their environmental overlap $\langle E_R|E_L\rangle$. When this is tiny, local interference is suppressed while the record weights remain $|\alpha|^2$ and $|\beta|^2$. That calculation does not remove either component from the total state or select one global outcome.[1] [3]

The possible records become linked to the surroundings. This hides interference—the effects that alternatives can have on each other—and helps explain why the records behave independently. This process is called decoherence.

With only the ordinary quantum rule, the total state still contains both records. Proposed accounts differ: some change the rule, some add physical variables, and others give the records a different meaning. Decoherence alone does not choose between them.[1] [3]

Physical collapse models replace universal unitary evolution with stochastic dynamics that localize the state. Bohmian mechanics supplements the wavefunction with an actual configuration of particles, including the apparatus. Everettian approaches retain the total superposition and treat each record as definite relative to its branch. Consistent histories assigns probabilities within consistent families of histories; its proponents regard measurement as resolved within that framework.

These accounts can agree on click frequencies while disagreeing on whether the wavefunction collapses, guides extra variables, or retains branches. A diagonal reduced state gives the same apparatus-only probabilities as a classical ensemble. Treating this equivalence as proof that exactly one global alternative exists adds a claim the partial trace does not establish.[1] [2] [3] [4] [5]

A detector is made of quantum matter too. Explaining its records connects quantum theory to the facts an experiment actually produces, rather than giving the detector an unexplained special role.

Some proposed answers change the physics and predict effects that experiments can look for, such as extra radiation. Other accounts give the same ordinary predictions while describing what happens differently. Understanding measurement would clarify both what the theory describes and which claims an experiment can test.[1] [6]

Measurement is where a quantum description must connect to physical records. An account of those records would clarify whether ordinary quantum dynamics is a universal physical rule or needs additional ingredients. Proposed changes can also lead to observable effects beyond standard quantum statistics.

In a March 2026 paper, XENONnT searched its xenon detector data for extra x-rays predicted by specified continuous spontaneous localization (CSL) and Diósi–Penrose (DP) collapse models. The tested models assume Markovian noise, uncorrelated in time. Neither signal showed a significant excess, placing limits on those models’ parameters.

That result constrains particular mechanisms; it does not establish Everett, Bohmian mechanics or any other interpretation. The test depends on the assumed collapse noise and radiation model. Accounts that reproduce the same ordinary click frequencies need other grounds for comparison.[1] [4] [5] [6]