Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-22
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.

Norm of a self adjoint operator from its quadratic form

Statement

Assume the Axiom of Countable Choice (The Axiom of Countable Choice (ACω)). Let H be a real or complex Hilbert space (Hilbert space) and let T∈B(H) be a bounded self-adjoint operator (Self-adjoint, positive, unitary and normal operators, A bounded linear operator between normed spaces), with operator norm ∥T∥ (The operator norm as the least bound and as the unit-sphere or unit-ball supremum). Put q(x):=⟨Tx,x⟩ for x∈H and

M:=sup⁡{ ∣q(x)∣:∥x∥=1 },

with the convention that a supremum over the empty set of reals is 0; the empty case occurs only for H={0}, where the only operator is T=0. Then

∥T∥=M=sup⁡∥x∥=1∣⟨Tx,x⟩∣.

The identity holds over both scalar fields, and in the case H={0} both sides equal 0.

Facts & Assumptions

Given: A real or complex Hilbert space H, a bounded self-adjoint operator T, the quadratic form q(x)=⟨Tx,x⟩, and M=sup⁡{∣q(x)∣:∥x∥=1} with the empty-supremum convention.

[A1]

Self-adjointness is the identity ⟨Tx,y⟩=⟨x,Ty⟩ for all x,y: a self-adjoint operator satisfies T∗=T, and the Hilbert adjoint is characterised by ⟨Tx,y⟩=⟨x,T∗y⟩ (Self-adjoint, positive, unitary and normal operators, The Hilbert-space adjoint of a bounded operator, Hilbert-adjoint identities).

[A2]

Inner-product algebra. The pairing is linear in the first argument, conjugate-linear in the second, conjugate symmetric, and positive definite, and ∥v∥2=⟨v,v⟩ (Real and complex inner-product spaces and their induced length). Consequently for real t>0 and x,y∈H one has ⟨T(tx),t−1y⟩=t⋅t−1⟨Tx,y⟩=⟨Tx,y⟩, the expansion q(x±u)=q(x)±⟨Tx,u⟩±⟨Tu,x⟩+q(u) holds, and if ⟨Tx,u⟩ is real then self-adjointness gives ⟨Tu,x⟩=⟨Tx,u⟩‾=⟨Tx,u⟩, so that q(x+u)−q(x−u)=4⟨Tx,u⟩. For z≠0 with unit vector u=z/∥z∥ one has q(z)=∥z∥2q(u).

[A3]

Parallelogram law. ∥x+u∥2+∥x−u∥2=2∥x∥2+2∥u∥2 for all x,u∈H (Pythagoras, the parallelogram identity, and the real and complex polarisation identities).

[A4]

Operator norm and Cauchy–Schwarz. ∥Tv∥≤∥T∥ ∥v∥ and ∥T∥=sup⁡{∥Tv∥:∥v∥≤1} (The operator norm as the least bound and as the unit-sphere or unit-ball supremum, A bounded linear operator between normed spaces); ∣⟨v,w⟩∣≤∥v∥ ∥w∥ (Cauchy–Schwarz: ∣⟨x,y⟩∣≤∥x∥ ∥y∥, with equality exactly for dependent pairs), so the dual norm formula ∥z∥=sup⁡{∣⟨z,y⟩∣:∥y∥=1} holds for every z∈H by Cauchy–Schwarz and by testing y=z/∥z∥ when z≠0 (both sides are 0 at z=0).

[A5]

Choice. Countable Choice is the hypothesis under which this pair's Hilbert-space interface is stated; no choice is used inside the argument below (The Axiom of Countable Choice (ACω)).

Proof

technique · direct

Given: Countable Choice, a real or complex Hilbert space H, a bounded self-adjoint T, and M=sup⁡{∣q(x)∣:∥x∥=1}.

1.1A2algebra

The quadratic form is dominated by M. For z∈H, if z=0 then ∣q(z)∣=0=M∥z∥2, while if z≠0 then [A2] gives q(z)=∥z∥2q(u) for the unit vector u=z/∥z∥, hence ∣q(z)∣=∥z∥2∣q(u)∣≤M∥z∥2.

1.2A4algebra

M≤∥T∥. For every unit vector x, Cauchy–Schwarz and the norm bound give ∣q(x)∣=∣⟨Tx,x⟩∣≤∥Tx∥ ∥x∥≤∥T∥, so ∥T∥ is an upper bound of the set whose supremum is M; hence M≤∥T∥, and when H={0} both numbers are 0.

2.1step 1.1A1A2A3algebra

A sesquilinear bound. For all x,y∈H one has ∣⟨Tx,y⟩∣≤M2(∥x∥2+∥y∥2): if x=0 or y=0 then ⟨Tx,y⟩=0 and the right side is ≥0; otherwise put c:=⟨Tx,y⟩, and if c≠0 choose the unit scalar ω with ω‾c=∣c∣ (namely ω=c/∣c∣ over C, ω=sign⁡(c) over R) and set u:=ωy, so that ∥u∥=∥y∥ and ⟨Tx,u⟩=ω‾c=∣c∣≥0 is real; then [A2] gives 4∣⟨Tx,y⟩∣=4⟨Tx,u⟩=q(x+u)−q(x−u)≤∣q(x+u)∣+∣q(x−u)∣≤M(∥x+u∥2+∥x−u∥2) by [step 1.1], and the parallelogram law [A3] turns the last factor into 2∥x∥2+2∥u∥2=2∥x∥2+2∥y∥2.

3.1step 2.1A2algebra

Removing the norms. For x,y≠0 and every real t>0, [step 2.1] applied to the pair (tx,t−1y) together with the scaling identity of [A2] gives ∣⟨Tx,y⟩∣≤M2(t2∥x∥2+t−2∥y∥2); the right side is minimised at t2=∥y∥/∥x∥>0, where it equals M∥x∥ ∥y∥, so ∣⟨Tx,y⟩∣≤M∥x∥ ∥y∥ for all x,y∈H (the zero cases being trivial).

4.1step 3.1A4algebra

∥T∥≤M. For x≠0 the dual norm formula [A4] gives ∥Tx∥=sup⁡∥y∥=1∣⟨Tx,y⟩∣≤M∥x∥ by [step 3.1] with ∥y∥=1, and the inequality also holds at x=0; thus M is a uniform bound for T on the unit ball, so ∥T∥≤M by the unit-ball characterisation of the operator norm in [A4].

5.1step 1.2step 4.1A5∎

Conclusion. Steps 1.2 and 4.1 give ∥T∥=M; if H={0} then T=0, M=0 by the empty-supremum convention and ∥T∥=0, so the identity holds there as well.

Depends on

Dependency tree · 2 levels

Sources