Measurement

The “collapse” here is an update of the description, and the interpretation is split on whether anything physical corresponds to it. Bohr required a classical side to every experiment: apparatus described in laboratory language, quantum system described by amplitudes, with the boundary set by the experimental arrangement rather than chosen at will. The sweep line in this figure is that cut in action — and the standing criticism is that the theory never says where the sweep must happen, only that by the time you read the dial, it has.

Decoherence does the work collapse was invented for: once the detector's trillions of degrees of freedom correlate with the system, the two branches stop interfering for all practical purposes. Each branch contains a consistent record — apparatus, environment, memory — so from inside, history looks single. The price is probability: if both outcomes happen, the Born weights need a new justification, and whether the standard arguments succeed is still contested.

In de Broglie–Bohm mechanics the guiding wave develops separated packets during measurement, and the particle ends up in exactly one of them — which one being fixed by where it started. The empty packet persists but stops mattering to the particle's future. Quantum randomness becomes classical-style ignorance over initial conditions (quantum equilibrium). The guidance is explicitly nonlocal: entangled partners respond instantly, which is why this theory sits uneasily beside special relativity.

GRW-style models add rare, random localization events to the dynamics; CSL makes the process continuous. The collapse rate scales with the size of the entangled system, which turns the quantum–classical boundary into a number. Because the modified dynamics differ slightly from standard quantum mechanics, matter-wave interferometry with ever larger molecules genuinely constrains these models — a rare case where interpretation edges into testable physics. This scene shows a generic, simplified collapse; the specific theories differ in mechanism and rate.

QBism reads the Born rule as a normative constraint on any agent's gambles — a rule of coherence, like logic, not a law of motion. Measurement outcomes are personal experiences, and two agents may lawfully assign different states to the same system until they compare notes — comparing notes being one more measurement.

Rovelli's proposal treats every interaction as a measurement for the systems involved, with no super-observer to keep a universal account. How far fact-relativity can be pushed is an active debate — “Wigner's friend” scenarios, where one observer measures another, are exactly the pressure tests, and their interpretation remains unsettled.

The consistency condition guarantees that a family's history probabilities add up classically, with no interference between them. Within the pointer-reading family, measurement is unmysterious. Across families, descriptions cannot be combined — the “single framework rule” — and critics press on what that rule says about which events actually occur.