Probability as Closure in General Probabilistic Theories: Uniqueness and Rigidity of State Assignment from Auditability Principles Paper 39 of the NEMS Suite
Year: 2026
Type: preprint
Abstract: In any finite-dimensional General Probabilistic Theory (GPT), closure-style constraints (context-independence, additivity, normalization, convex mixing) force probability assignment to a unique affine state functional. This paper proves that theorem and shows the Born rule is the matrix-ordered instance. The development is mechanized in Lean 4 as the GPTClosure library in nems-lean. The abstract core has 0 sorrys in QuantumFinite (all three prior gaps — PSD cone pointedness, Born-rule nonnegativity, wiring to busch_gleason_unique — are now fully proved). 0 sorry in all other GPTClosure modules. The proofs of PSD cone pointedness and Born-rule nonnegativity rely on re_trace_psd_mul_psd_nonneg, an explicit axiom stating Re(Tr(AB)) >= 0 for positive-semidefinite Hermitian A and B; this is a standard fact from matrix analysis whose Lean proof is pending a bridge to Mathlib's Matrix.PosSemidef. No other custom axioms.
Author: Spivack, Nova
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7150080244
Cites:
Cited by:
Related to: