Original text: CC0 1.0. Prerequisite proofs and component terms.

Comparing weights through their finite-energy vectors

OA-MOD-DW-01 — Conventions and the exact comparison problem

Inner products are linear in the first variable. Fix a weight φ:M+→[0,∞]\varphi:M_+\to[0,\infty], and write

nφ={x:φ(x∗x)<∞},mφ=span⁡{y∗x:x,y∈nφ}. \mathfrak n_\varphi=\{x:\varphi(x^*x)<\infty\},\qquad \mathfrak m_\varphi=\operatorname{span}\{y^*x:x,y\in\mathfrak n_\varphi\}.

The construction in OA-MOD-WG supplies a positive linear extension of φ\varphi to mφ\mathfrak m_\varphi, a Hilbert space HφH_\varphi, a dense-range map Λφ\Lambda_\varphi, and a unital representation πφ\pi_\varphi with

⟨Λφ(x),Λφ(y)⟩=φ(y∗x),πφ(a)Λφ(x)=Λφ(ax). \langle\Lambda_\varphi(x),\Lambda_\varphi(y)\rangle=\varphi(y^*x),\qquad \pi_\varphi(a)\Lambda_\varphi(x)=\Lambda_\varphi(ax).

In particular mφ+=mφ∩M+={a∈M+:φ(a)<∞}\mathfrak m_\varphi^+=\mathfrak m_\varphi\cap M_+=\{a\in M_+:\varphi(a)<\infty\}, and every element of mφ\mathfrak m_\varphi is a complex linear combination of this cone. These are algebraic domains, not norm closures.

Let Dφ\mathcal D_\varphi be the cone of positive linear forms ℓ:mφ→C\ell:\mathfrak m_\varphi\to\mathbb C such that, for some finite c≥0c\geq0,

0≤ℓ(a)≤cφ(a)(a∈mφ+). 0\leq\ell(a)\leq c\varphi(a)\quad(a\in\mathfrak m_\varphi^+).

Positivity refers to the inherited cone above. It does not assert that ℓ\ell is bounded for the operator norm, or that it has already been extended to a normal weight on all of M+M_+. The order on these forms is positivity of their difference on this cone.

OA-MOD-DW-02 — A factorization inside the algebra

Lemma. If x,y∈Mx,y\in M and y∗y≤x∗xy^*y\leq x^*x, there is a unique contraction v∈Mv\in M which vanishes on xH‾⊥\overline{xH}^{\perp} and satisfies y=vxy=vx, in any faithful concrete realization M⊆B(H)M\subseteq B(H).

Proof. On xHxH define xξ↦yξx\xi\mapsto y\xi. The inequality says both that this is well defined and that its norm is at most one. Extend it continuously to xH‾\overline{xH}, and set it equal to zero on the orthogonal complement. This proves existence and uniqueness in B(H)B(H). Every unitary u∈M′u\in M' preserves xH‾\overline{xH}; because x,yx,y commute with uu, both vv and uvu∗uvu^* have the prescribed properties. Thus vv commutes with every unitary of M′M'. Every element of a unital C*-algebra is a linear combination of unitaries: a self-adjoint contraction bb is the real part of b+i(1−b2)1/2b+i(1-b^2)^{1/2}. Consequently v∈M′′=Mv\in M''=M. This uses bounded continuous functional calculus and the bicommutant theorem, recorded foundation contracts. □\square

For a,b≥0a,b\geq0 put d=a+bd=a+b and p=s(d)p=s(d), the projection onto dH‾\overline{dH}. Applying the lemma to a1/2,d1/2a^{1/2},d^{1/2} and b1/2,d1/2b^{1/2},d^{1/2} gives contractions v,w∈Mv,w\in M with

a1/2=vd1/2,b1/2=wd1/2,v=vp,w=wp,v∗v+w∗w=p. a^{1/2}=vd^{1/2},\quad b^{1/2}=wd^{1/2},\quad v=vp,\quad w=wp,\quad v^*v+w^*w=p.

To verify the last identity, its quadratic form agrees with that of pp on d1/2Hd^{1/2}H, by a+b=da+b=d; continuity extends the equality to pHpH. Both sides vanish on (1−p)H(1-p)H. This argument needs neither an inverse of dd nor a positive lower bound for it.

OA-MOD-DW-03 — The commutant correspondence

Theorem. There is an additive, positively homogeneous order isomorphism

πφ(M)+′⟷Dφ,T⟼ℓT, \pi_\varphi(M)'_+\longleftrightarrow\mathcal D_\varphi,\qquad T\longmapsto\ell_T,

uniquely characterized by

ℓT(y∗x)=⟨TΛφ(x),Λφ(y)⟩(x,y∈nφ). \ell_T(y^*x)=\langle T\Lambda_\varphi(x),\Lambda_\varphi(y)\rangle \quad(x,y\in\mathfrak n_\varphi).

Moreover T≤cIT\leq cI if and only if ℓT≤cφ\ell_T\leq c\varphi on mφ+\mathfrak m_\varphi^+. Thus the least admissible comparison constant is ∥T∥\|T\|. The assertion includes the zero Hilbert space, where both cones consist only of zero.

Proof, from forms to operators. For ℓ∈Dφ\ell\in\mathcal D_\varphi, positivity applied to (x+zy)∗(x+zy)(x+zy)^*(x+zy), for every z∈Cz\in\mathbb C, gives

∣ℓ(y∗x)∣2≤ℓ(x∗x)ℓ(y∗y)≤c2∥Λφ(x)∥2∥Λφ(y)∥2. |\ell(y^*x)|^2\leq\ell(x^*x)\ell(y^*y) \leq c^2\|\Lambda_\varphi(x)\|^2\|\Lambda_\varphi(y)\|^2.

For completeness, when ℓ(y∗y)>0\ell(y^*y)>0, minimize this quadratic polynomial in zz; when that diagonal term is zero, varying the magnitude and phase of zz forces the cross term to vanish. The inequality shows that the expression is independent of null representatives and defines a bounded sesquilinear form on Λφ(nφ)\Lambda_\varphi(\mathfrak n_\varphi). It extends uniquely to HφH_\varphi. The Hilbert-space representation theorem for bounded sesquilinear forms gives a unique positive operator TT with 0≤T≤cI0\leq T\leq cI.

For a∈Ma\in M and x,y∈nφx,y\in\mathfrak n_\varphi, associativity gives

⟨Tπφ(a)Λφ(x),Λφ(y)⟩=ℓ(y∗ax)=⟨TΛφ(x),πφ(a∗)Λφ(y)⟩=⟨πφ(a)TΛφ(x),Λφ(y)⟩. \begin{aligned} \langle T\pi_\varphi(a)\Lambda_\varphi(x),\Lambda_\varphi(y)\rangle &=\ell(y^*ax)\\ &=\langle T\Lambda_\varphi(x),\pi_\varphi(a^*)\Lambda_\varphi(y)\rangle\\ &=\langle\pi_\varphi(a)T\Lambda_\varphi(x),\Lambda_\varphi(y)\rangle. \end{aligned}

The test vectors form a dense subspace, so Tπφ(a)=πφ(a)TT\pi_\varphi(a)=\pi_\varphi(a)T. This proves membership in the commutant without any appeal to a modular group.

Proof, from operators to forms. Fix T∈πφ(M)+′T\in\pi_\varphi(M)'_+. For finite-weight positive aa, define

fT(a)=⟨TΛφ(a1/2),Λφ(a1/2)⟩. f_T(a)=\langle T\Lambda_\varphi(a^{1/2}),\Lambda_\varphi(a^{1/2})\rangle.

We must prove additivity; merely writing a formula on products would not prove independence of their decompositions. Let a,b∈mφ+a,b\in\mathfrak m_\varphi^+, and use d,p,v,wd,p,v,w from OA-MOD-DW-02. Here d1/2∈nφd^{1/2}\in\mathfrak n_\varphi. Put ζ=Λφ(d1/2)\zeta=\Lambda_\varphi(d^{1/2}). Then πφ(p)ζ=ζ\pi_\varphi(p)\zeta=\zeta and

fT(a)+fT(b)=⟨Tπφ(v)ζ,πφ(v)ζ⟩+⟨Tπφ(w)ζ,πφ(w)ζ⟩=⟨Tζ,πφ(v∗v+w∗w)ζ⟩=fT(d). \begin{aligned} f_T(a)+f_T(b) &=\langle T\pi_\varphi(v)\zeta,\pi_\varphi(v)\zeta\rangle +\langle T\pi_\varphi(w)\zeta,\pi_\varphi(w)\zeta\rangle\\ &=\langle T\zeta,\pi_\varphi(v^*v+w^*w)\zeta\rangle =f_T(d). \end{aligned}

Homogeneity follows from the square root of a scalar. Thus fTf_T extends to a real linear form on mφ,sa\mathfrak m_{\varphi,\mathrm{sa}}: assign fT(a)−fT(b)f_T(a)-f_T(b) to a−ba-b. If a−b=a′−b′a-b=a'-b', the equality a+b′=a′+ba+b'=a'+b and additivity prove independence. Complexification gives a positive linear form ℓT\ell_T on mφ\mathfrak m_\varphi. Also

0≤fT(a)≤∥T∥φ(a). 0\leq f_T(a)\leq\|T\|\varphi(a).

If x∈nφx\in\mathfrak n_\varphi, take its polar decomposition x=u∣x∣x=u|x| in MM. Then Λφ(x)=πφ(u)Λφ(∣x∣)\Lambda_\varphi(x)=\pi_\varphi(u)\Lambda_\varphi(|x|), and u∗uu^*u fixes ∣x∣|x|. Commutation with TT yields

⟨TΛφ(x),Λφ(x)⟩=fT(x∗x). \langle T\Lambda_\varphi(x),\Lambda_\varphi(x)\rangle=f_T(x^*x).

Polarization now proves the formula for y∗xy^*x. Uniqueness holds because these products span mφ\mathfrak m_\varphi, and their GNS vectors are dense. The two constructions are inverse. Addition and scalar multiplication follow from the defining pairings. Finally positivity of ℓT2−ℓT1\ell_{T_2}-\ell_{T_1} is equivalent to nonnegativity of the quadratic form of T2−T1T_2-T_1 on a dense subspace, hence on all of HφH_\varphi. Apply this to cI−TcI-T to obtain the bound and the optimal constant. □\square

OA-MOD-DW-04 — Comparing two weights and the precise target space

Theorem. Let ψ\psi be another weight on MM, with ψ≤cφ\psi\leq c\varphi on M+M_+ for a finite c>0c>0. There is a unique bounded map

Cψ∣φ:Hφ⟶Hψ,Cψ∣φΛφ(x)=Λψ(x)(x∈nφ), C_{\psi\mid\varphi}:H_\varphi\longrightarrow H_\psi, \qquad C_{\psi\mid\varphi}\Lambda_\varphi(x)=\Lambda_\psi(x) \quad(x\in\mathfrak n_\varphi),

with norm at most c\sqrt c. It intertwines the representations, and

Tψ∣φ=Cψ∣φ∗Cψ∣φ∈πφ(M)+′,ψ(y∗x)=⟨Tψ∣φΛφ(x),Λφ(y)⟩. T_{\psi\mid\varphi}=C_{\psi\mid\varphi}^*C_{\psi\mid\varphi}\in\pi_\varphi(M)'_+, \quad \psi(y^*x)=\langle T_{\psi\mid\varphi}\Lambda_\varphi(x),\Lambda_\varphi(y)\rangle.

The range closure is exactly K=Λψ(nφ)‾⊆HψK=\overline{\Lambda_\psi(\mathfrak n_\varphi)}\subseteq H_\psi. If C=UT1/2C=UT^{1/2} is its polar decomposition, then UU is a unitary from s(T)Hφs(T)H_\varphi onto KK, and is zero on ker⁡T\ker T. These two subspaces reduce the corresponding representations, and UU intertwines their restrictions.

Proof. Domination gives nφ⊆nψ\mathfrak n_\varphi\subseteq\mathfrak n_\psi and ∥Λψ(x)∥2≤c∥Λφ(x)∥2\|\Lambda_\psi(x)\|^2\leq c\|\Lambda_\varphi(x)\|^2. Consequently the displayed assignment is well defined even for nonfaithful weights and extends from a dense domain. Its range closure is KK by construction. On the dense GNS domain,

Cπφ(a)Λφ(x)=Λψ(ax)=πψ(a)CΛφ(x). C\pi_\varphi(a)\Lambda_\varphi(x) =\Lambda_\psi(ax)=\pi_\psi(a)C\Lambda_\varphi(x).

Boundedness extends this identity to all vectors. Taking adjoints and using the identity for a∗a^* shows that C∗C^* intertwines in the reverse direction, so C∗CC^*C commutes with πφ(M)\pi_\varphi(M). The pairing formula is immediate. It agrees with OA-MOD-DW-03 applied to ψ∣mφ\psi|_{\mathfrak m_\varphi}.

The closed subspaces ker⁡C\ker C and KK are invariant under the representations and their adjoints. Therefore their orthogonal projections commute with these representations. Since T1/2T^{1/2} also commutes with πφ(M)\pi_\varphi(M), the equality Uπφ(a)T1/2ξ=πψ(a)UT1/2ξU\pi_\varphi(a)T^{1/2}\xi=\pi_\psi(a)UT^{1/2}\xi holds first on ran⁡T1/2\operatorname{ran}T^{1/2}, and then by continuity on its closure s(T)Hφs(T)H_\varphi. Both sides vanish on its orthogonal complement. The initial and final subspaces of a polar decomposition give the claimed unitary. □\square

Do not replace KK by HψH_\psi in this level of generality. If φ\varphi is zero at zero and infinite at every nonzero positive element, then nφ={0}\mathfrak n_\varphi=\{0\}, while every weight ψ\psi satisfies ψ≤φ\psi\leq\varphi. The comparison map is zero even when Hψ≠0H_\psi\neq0. This also shows why restriction to a finite domain need not determine the whole weight.

Editable source · Proof dependencies and component terms