Gleason's theorem, finite-dimensional: the core lemma discharged, the theorem proved #
Category: 1-Mathlib (CSD-free; staged for upstream). The end of the staged plan of
specs/gleason-feasibility.md (BACKLOG #57): Gleason/General.lean proved the core lemma
(frameFunction_regular_sphere, stage 57(e)) and Gleason/Reduction.lean proved the theorem
from it. This file joins the two.
coreLemma— the core lemma in the formReduction.leanconsumes:CoreLemmaholds.- ★★
ProjectionPackage.gleason_representation— Gleason's theorem forℂᴺ,N ≥ 3: for every projection packageOPthere is a unique density matrixρwithOP.p P = Re Tr(ρ P)for every orthogonal projectionP. No hypothesis beyondN ≥ 3;#print axiomsreports the foundational triple (pinned inTests/AxiomAudit/MathlibStaging.lean).
What is and is not claimed. This is the finite-dimensional theorem (ℂᴺ, N ≥ 3) for
probability assignments on orthogonal projections. The infinite-dimensional theorem,
dimension 2 (where the statement fails) and the POVM generalisation (Busch's,
LF2/EffectGleason.lean, proved
separately) are outside it. Neither of the two theorems is derived from the other: the proofs
share the Jordan–von Neumann engine (Gleason/Polarization.lean) and the descent
(Gleason/Descent.lean), nothing else.
Source #
Gleason 1957, J. Math. Mech. 6, 885 (Theorem 2.8, the nonnegative core lemma;
Theorem 3.1, the theorem); Cooke, Keane, Moran 1985, Math. Proc. Cambridge Philos. Soc. 98, 117 (the
elementary proof of the core lemma followed in Gleason/Sphere.lean through
Gleason/General.lean).
★★ Gleason's theorem for ℂᴺ, N ≥ 3. For every projection package OP there is a
unique density matrix ρ with OP.p P = Re Tr(ρ P) for every orthogonal projection P.