Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck pass
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Distinct density operators are separated by the law of a bounded rank-one projection

Statement

Assume the Axiom of Countable Choice (The Axiom of Countable Choice (ACω)); the Axiom of Choice is used only where the spectral theorem for unbounded self-adjoint operators is invoked (The Axiom of Choice). Let H be a nonzero complex Hilbert space (Hilbert space) and let ρ,σ be density operators on H with ρ≠σ (Density operators and scalar spectral probability clause (a)). Then:

(a) There is a unit vector w∈H with ⟨ρw,w⟩≠⟨σw,w⟩.

(b) For the rank-one orthogonal projection P:=Tw,w of Rank-one operator calculus (The Hilbert orthogonal projection onto a closed subspace), whose spectral projection-valued measure assigns P to the Borel set {1}, the two scalar laws of Density operators and scalar spectral probability clause (b) satisfy μρP({1})=tr⁡(ρP)=⟨ρw,w⟩,μσP({1})=tr⁡(σP)=⟨σw,w⟩, so the bounded sharp observable P assigns different outcome probabilities to ρ and σ.

(c) For every unit vector w and every self-adjoint trace-class T one has ∣tr⁡(TTw,w)∣≤∥T∥1; in particular ∣⟨ρw,w⟩−⟨σw,w⟩∣≤∥ρ−σ∥1.

Facts & Assumptions

Given: The Axiom of Countable Choice and, for the spectral PVM of clause (b), the Axiom of Choice (The Axiom of Countable Choice (ACω), The Axiom of Choice); a nonzero complex Hilbert space H; density operators ρ,σ on H; unit vectors w∈H.

[F1]

Rank-one calculus. Tu,vx=⟨x,v⟩u defines a trace-class operator with ∥Tu,v∥1=∥u∥ ∥v∥ and tr⁡(Tu,v)=⟨u,v⟩; Tu,u is self-adjoint and positive with ⟨Tu,ux,x⟩=∣⟨x,u⟩∣2 (Rank-one operator calculus).

[F2]

One-dimensional projections. For u∈S(H) the operator Tu,u is the Hilbert orthogonal projection onto the one-dimensional subspace Cu; in particular Tu,u2=Tu,u=Tu,u∗ (Rays, unit vectors and rank-one Hilbert projections, The Hilbert orthogonal projection onto a closed subspace).

[F3]

Density operators and their law. A density operator τ is self-adjoint, positive and trace class with tr⁡τ=1, and admits an expansion τ=∑ntnTvn,vn with tn≥0, ∑ntn=1 and (vn) orthonormal; for every PVM E, B↦tr⁡(τE(B)) is a probability measure with tr⁡(τE(B))=∑ntn∥E(B)vn∥2 (Density operators and scalar spectral probability clauses (a) and (b)).

[F4]

Trace ideal and trace bounds. Products of a trace-class operator with a bounded operator are trace class and ∥ATB∥1≤∥A∥ ∥T∥1∥B∥; the trace is linear on trace-class operators with ∣tr⁡S∣≤∥S∥1 (Trace class is a two sided Banach operator ideal clause 2, Trace is absolutely convergent and basis independent clause 4).

[F5]

Quadratic form and norm. For a bounded self-adjoint T one has ∥T∥=sup⁡{ ∣⟨Tx,x⟩∣:∥x∥=1 }, and the value is 0 exactly when T=0 (Norm of a self adjoint operator from its quadratic form).

[F6]

Spectral PVM of a projection. Every orthogonal projection P is a self-adjoint operator with 0≤P≤I; by the spectral theorem it has a unique regular PVM EP on the Borel sets of R with P=∫λ dEP(λ) (Spectral theorem for unbounded self-adjoint operators (PVM form)).

Proof

1.1givenF3F4F5algebra

[Quadratic forms separate trace-class states] Put T:=ρ−σ, a nonzero self-adjoint trace-class operator by [F3] and [F4]. Suppose ⟨Tw,w⟩=0 for every unit vector w; then the supremum of [F5] is 0, so ∥T∥=0 and T=0, contradicting ρ≠σ. Hence there is a unit w with 0≠⟨(ρ−σ)w,w⟩=⟨ρw,w⟩−⟨σw,w⟩, proving (a).

1.2givenF1F3F4algebra

[The trace pairing of a rank-one projection] Let τ=∑ntnTvn,vn be the expansion of a density operator and let w be a unit vector. For each n, the product rule Tvn,vnTw,w=Tvn,⟨vn,w⟩w holds by direct evaluation on any x: Tvn,vnTw,wx=⟨x,w⟩⟨w,vn⟩vn=Tvn,⟨vn,w⟩wx; hence by [F1] its trace is ⟨vn,⟨vn,w⟩w⟩=∣⟨vn,w⟩∣2. Multiplying by tn, summing and using the trace ideal and trace bound [F4] together with ∑ntn∥vn∥2∥w∥2=1 gives the absolutely convergent identity tr⁡(τTw,w)=∑ntn∣⟨vn,w⟩∣2. On the other hand ⟨τw,w⟩=∑ntn⟨Tvn,vnw,w⟩=∑ntn∣⟨w,vn⟩∣2 by [F1] and continuity of the pairing against the norm convergent series. Since ∣⟨w,vn⟩∣=∣⟨vn,w⟩∣, the two series agree: tr⁡(τTw,w)=⟨τw,w⟩ for every density operator τ.

2.1step 1.1step 1.2F2F3F6algebra

[The projection separates the two states] Fix a unit w with ⟨ρw,w⟩≠⟨σw,w⟩, furnished by step 1.1, and put P:=Tw,w, an orthogonal projection by [F2]. The map E′(B):=11∈BP+10∈B(I−P) is a regular PVM on the Borel sets of R with ∫λ dE′(λ)=1⋅P+0⋅(I−P)=P: the values are orthogonal projections with E′(∅)=0 and E′(R)=I, multiplicativity holds because P(I−P)=0, and countable additivity holds on the two-point space {0,1}. By uniqueness in the spectral theorem [F6], E′ is the spectral PVM of P, so EP({1})=P and μτP({1})=tr⁡(τP) for every density operator τ by [F3]. Step 1.2 identifies these numbers as ⟨ρw,w⟩ and ⟨σw,w⟩, which differ by the choice of w; this proves (b).

3.1step 1.1step 1.2step 2.1F2F4algebra∎

[The uniform bound] Let w be a unit vector and let T be self-adjoint trace class. By [F2] and [F4], ∥Tw,w∥≤∥Tw,w∥1=1 and ∣tr⁡(TTw,w)∣≤∥TTw,w∥1≤∥T∥1; applying this to T=ρ−σ and using step 1.2 gives ∣⟨ρw,w⟩−⟨σw,w⟩∣=∣tr⁡((ρ−σ)Tw,w)∣≤∥ρ−σ∥1, which is (c); the existence statement of (a) and the separating probability identity of (b) are steps 1.1 and 2.1.

Depends on

Dependency tree · 2 levels

Sources