Identifiability Limits in Experimental Tests of Quantum Collapse

Abstract

This record contains version 3 of the preprint “Identifiability Limits in Experimental Tests of Quantum Collapse.” It supersedes version 2 deposited under DOI 10.5281/zenodo.22071445. Loss of interference visibility does not by itself identify a collapse mechanism. This preprint formulates objective-collapse experiments as an operational identifiability problem. It combines a finite-process dilation boundary, visibility-factorisation non-identifiability and access-dependent record/coherence limits with a constructive multi-channel inference rule. Version 3 generalises the Fisher-information treatment to a full joint likelihood with shared and channel-specific nuisance parameters and correlated channels. It adds a reproducible, normalised two-channel CSL benchmark in which coherence and heating are individually rank-deficient but jointly identify the shared collapse parameters locally. It also treats the no-collapse hypothesis as a non-regular boundary problem: the collapse strength is one-sided and the correlation length is unidentified at zero strength, so a seeded parametric bootstrap is supplied instead of assuming Wilks’ theorem. The contribution is a falsifiable experimental decision framework and an explicit boundary on what accessible observations can establish. It is not presented as an ontological solution of the quantum measurement problem. Files in this record: Main preprint PDF. Reproducibility ZIP containing the editable DOCX, Markdown source, six original figures, deterministic calculation and figure-generation code, machine-readable numerical results, citation metadata and SHA-256 checksums. The numerical example is synthetic and dimensionless. The included calculations are deterministic mathematical and statistical checks; they are not experimental validation or an apparatus-specific sensitivity forecast.