Preferred Basis Problem
The preferred-basis problem asks why certain alternatives count as stable measurement records when a state vector admits many mathematical decompositions. The answer depends on the physical subsystem split, interaction and accessible observables. A basis chosen solely to diagonalize an instantaneous reduced state need not be the basis in which a real apparatus stores persistent information.
Helpful background. Measurement Problem establishes the premeasurement correlation; Entangled States supplies the subsystem structure.
A maximally correlated state has several decompositions
Section titled “A maximally correlated state has several decompositions”For two qubits,
Both expansions correlate orthogonal alternatives, and each subsystem has state . The reduced spectrum is degenerate, so diagonalization cannot identify whether the apparatus is recording or . Yet an actual device is built to couple to particular physical variables. Its Hamiltonian and calibration are additional information beyond the final abstract vector.
For unequal Schmidt weights the Schmidt basis is unique up to phases under the usual nondegeneracy assumptions. That mathematical uniqueness still does not guarantee robustness under later interactions. The foundational question concerns records that can survive and be read, not merely a convenient expansion at one time.
An interaction identifies what is monitored
Section titled “An interaction identifies what is monitored”Consider an apparatus observable coupled to an environment through . A eigenstate evolves without being mixed with the other eigenstate by this interaction; instead it imprints a conditional evolution on the environment. A superposition of eigenstates becomes entangled with the environment when those conditional environmental states become distinguishable.
For an initially pure environment, the reduced off-diagonal term is multiplied by the conditional-state overlap . Suppression requires ; the Hamiltonian’s form alone does not guarantee it. If the environment starts in a eigenstate, the conditional states differ only by phase, , and no decoherence occurs. Under effective monitoring, an eigenstate does not have the same stability as a eigenstate. If the apparatus self-Hamiltonian does not commute with , exact preservation fails and one must compare dynamical time scales. Real pointer states can be approximate wave packets rather than exact eigenvectors of one finite-dimensional operator.
The technical treatments of Pointer States and Environment-Induced Superselection develop the appropriate models. Their explanatory contribution is a dynamical reason for an effective record algebra, not a fundamental declaration that every conceivable decomposition describes a separate classical world.
Basis selection and outcome selection are different
Section titled “Basis selection and outcome selection are different”Suppose the dynamics convincingly identifies the basis and renders its interference inaccessible. The weights of the two alternatives are still present in the reduced state, and the global unitary state still contains both correlations. A stochastic-selection rule, extra ontology, or branch-relative account is needed to address the outcome problem.
This separation also avoids a false dismissal of decoherence. Explaining why a stable record basis exists is a substantial physical achievement even when it does not by itself choose one result. A theory of outcomes that ignores record stability would remain incomplete as an apparatus model.
Exercise: rotating the apparatus
Section titled “Exercise: rotating the apparatus”Let and prepare the apparatus in . Contrast its reduced evolution with an initial when the environment begins in .
Solution
For , only the environment acquires the phase rotation ; the apparatus remains pure . For , the two apparatus components produce opposite environmental rotations. Their overlap is , which multiplies the apparatus off-diagonal element. This finite environment recoheres periodically, so the example demonstrates basis-dependent stability without irreversible macroscopic decoherence.
References
Section titled “References”- Zurek, Wojciech H., “Pointer basis of quantum apparatus: Into what mixture does the wave packet collapse?,” Physical Review D 24, 1516–1525 (1981), doi:10.1103/PhysRevD.24.1516 — interaction-selected records.
- Zurek, Wojciech H., “Decoherence, einselection, and the quantum origins of the classical,” Reviews of Modern Physics 75, 715–775 (2003), doi:10.1103/RevModPhys.75.715 — dynamical stability and classicality.