Physlib

PhyslibAlpha.AlgebraicFramework.HilbertSpace.Unbounded.CayleySpectralData.SpecTheorem

Bounded spectral data for a Cayley transform: the self-adjoint spectral theorem

Continues `CayleySpectralData/Construction.lean`: transports its bounded spectral certificate back through the Cayley transform to `cayleyRealSpectralMeasure`, proves the resulting measure is a weak spectral resolution of the original self-adjoint operator, and assembles `unboundedSpectralTheorem` — the public unbounded spectral theorem for a self-adjoint `LinearPMap` — together with its essentially-self-adjoint variant.

Restriction identities for PVM scalar measures

These identities are independent of the Cayley transform. They express the elementary fact that testing a projection-valued measure after applying one of its projections restricts the corresponding scalar measure. They are the bookkeeping lemmas needed when a bounded spectral moment is localized to a measurable spectral set.

Complex vector-measure density transport

The scalar measures of a complex PVM are complex vector measures, rather than positive measures. This is the reusable density-integral theorem needed by the inverse Cayley step.

The inverse-moment transport lemma

The bounded spectral reconstruction supplies the Cayley moment `z`. The genuinely unbounded step is the inverse-Cayley moment on the operator domain. The following theorem isolates the measure-transport part of that step: once the inverse moment is established on the bounded Cayley measure, it is transported automatically to the real spectral measure. This keeps the analytic domain argument separate from the bookkeeping for mapped vector measures.

20 declarations

theorem

The domain of a self-adjoint operator TT is contained in the square-moment domain of its spectral measure μT\mu_T

Let HH be a complex Hilbert space and TT be a self-adjoint partial linear operator on HH with domain D(T)D(T). Let μT\mu_T denote the real spectral measure of TT (the spectral measure on R\mathbb{R} obtained via the inverse Cayley transform). For any vector x∈D(T)x \in D(T), xx belongs to the square-moment domain of μT\mu_T. This means that the integral of the squared identity function with respect to the diagonal measure μx\mu_x is finite: ∫Rλ2dμx(λ)<∞ \int_{\mathbb{R}} \lambda^2 d\mu_x(\lambda) < \infty where μx(S)=⟨x,μT(S)x⟩\mu_x(S) = \langle x, \mu_T(S) x \rangle for measurable sets S⊆RS \subseteq \mathbb{R}.

theorem

μμ(S)x,y=μx,y∣S\mu_{\mu(S)x, y} = \mu_{x, y} |_S

Let HH be a Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on a measurable space α\alpha. For any measurable set S⊆αS \subseteq \alpha and vectors x,y∈Hx, y \in H, let μx,y\mu_{x, y} denote the complex scalar measure defined by μx,y(A)=⟨y,μ(A)x⟩\mu_{x, y}(A) = \langle y, \mu(A) x \rangle for measurable sets AA. Then the scalar measure associated with the projected vector μ(S)x\mu(S)x and the vector yy is equal to the restriction of the scalar measure μx,y\mu_{x, y} to the set SS: μμ(S)x,y=μx,y∣S\mu_{\mu(S)x, y} = \mu_{x, y} |_S where μx,y∣S\mu_{x, y} |_S denotes the restriction of the measure to SS (i.e., (μx,y∣S)(A)=μx,y(A∩S)(\mu_{x, y} |_S)(A) = \mu_{x, y}(A \cap S)).

theorem

μx,μ(S)y=μx,y∣S\mu_{x, \mu(S)y} = \mu_{x, y}|_S for WOT spectral measures

Let HH be a Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on a measurable space α\alpha. For any two vectors x,y∈Hx, y \in H, let μx,y\mu_{x, y} denote the complex scalar measure defined by μx,y(A)=⟨y,μ(A)x⟩\mu_{x, y}(A) = \langle y, \mu(A) x \rangle for every measurable set A⊆αA \subseteq \alpha. For any measurable set S⊆αS \subseteq \alpha, the scalar measure formed by projecting the second vector yy via μ(S)\mu(S) is equal to the restriction of the original scalar measure to SS: μx,μ(S)y=μx,y∣S\mu_{x, \mu(S)y} = \mu_{x, y}|_S

theorem

Complex Weak Integral with Left Projection μ(S)x\mu(S)x equals Integral of 1Sg\mathbb{1}_S g

Let HH be a Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on a measurable space α\alpha. For any measurable set S⊆αS \subseteq \alpha, any vectors x,y∈Hx, y \in H, and any complex-valued function g:α→Cg: \alpha \to \mathbb{C} that is integrable with respect to the scalar measure μx,y\mu_{x, y} (defined by μx,y(A)=⟨y,μ(A)x⟩\mu_{x, y}(A) = \langle y, \mu(A) x \rangle), the complex weak integral with the projected vector μ(S)x\mu(S)x satisfies: ∫αg(z) d⟨y,μ(z)(μ(S)x)⟩=∫α1S(z)g(z) d⟨y,μ(z)x⟩\int_{\alpha} g(z) \, d\langle y, \mu(z) (\mu(S)x) \rangle = \int_{\alpha} \mathbb{1}_S(z) g(z) \, d\langle y, \mu(z) x \rangle where 1S\mathbb{1}_S is the indicator function of the set SS.

theorem

∫g d⟨μ(S)y,μx⟩=∫Sg d⟨y,μx⟩\int g \, d\langle \mu(S)y, \mu x \rangle = \int_S g \, d\langle y, \mu x \rangle for WOT Spectral Measures

Let HH be a Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on a measurable space α\alpha. For any two vectors x,y∈Hx, y \in H, let μx,y\mu_{x, y} denote the complex scalar measure defined by μx,y(A)=⟨y,μ(A)x⟩\mu_{x, y}(A) = \langle y, \mu(A) x \rangle for every measurable set A⊆αA \subseteq \alpha. For any measurable set S⊆αS \subseteq \alpha and any complex-valued function g:α→Cg: \alpha \to \mathbb{C} that is integrable with respect to μx,y\mu_{x,y}, the complex weak integral satisfies: ∫αg(z) d⟨μ(S)y,μ(z)x⟩=∫α1S(z)g(z) dμx,y(z)\int_\alpha g(z) \, d\langle \mu(S)y, \mu(z) x \rangle = \int_\alpha \mathbb{1}_S(z) g(z) \, d\mu_{x,y}(z) where 1S\mathbb{1}_S is the indicator function of the set SS. This identity shows that projecting the test vector yy onto the subspace associated with SS is equivalent to restricting the integration of gg to the set SS.

theorem

The operator U=∫f dμU = \int f \, d\mu commutes with the spectral projections μ(S)\mu(S)

Let HH be a Hilbert space and α\alpha a measurable space. Let μ\mu be a weak operator topology (WOT) spectral measure on α\alpha acting on HH. Let f:α→Cf: \alpha \to \mathbb{C} be a complex-valued function that is integrable with respect to every scalar measure μx,y\mu_{x,y} defined by μx,y(S)=⟨y,μ(S)x⟩\mu_{x,y}(S) = \langle y, \mu(S) x \rangle for x,y∈Hx, y \in H. Suppose there exists a bounded linear operator UU on HH such that for all x,y∈Hx, y \in H, the relation ⟨y,Ux⟩=∫αf(z) d⟨y,μ(z)x⟩\langle y, U x \rangle = \int_\alpha f(z) \, d\langle y, \mu(z) x \rangle holds. Then, for any measurable set S⊆αS \subseteq \alpha, the projection μ(S)\mu(S) commutes with UU, i.e., μ(S)U=Uμ(S).\mu(S) U = U \mu(S).

theorem

∫(1−z) d⟨y,μx⟩=⟨y,x−Ux⟩\int (1 - z) \, d\langle y, \mu x \rangle = \langle y, x - Ux \rangle

Let HH be a complex Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on the complex plane C\mathbb{C} acting on HH. Let U:H→HU: H \to H be a bounded linear operator such that UU is the weak integral of the identity function, meaning ⟨y,Ux⟩=∫Cz d⟨y,μ(z)x⟩\langle y, U x \rangle = \int_{\mathbb{C}} z \, d\langle y, \mu(z) x \rangle for all vectors x,y∈Hx, y \in H. Assuming that the identity function z↦zz \mapsto z is integrable with respect to the complex scalar measures μx,y\mu_{x,y} (defined by μx,y(S)=⟨y,μ(S)x⟩\mu_{x,y}(S) = \langle y, \mu(S) x \rangle) and that these scalar measures have finite variation, then for any x,y∈Hx, y \in H, the complex weak integral of the function f(z)=1−zf(z) = 1 - z satisfies: ∫C(1−z) d⟨y,μ(z)x⟩=⟨y,x−Ux⟩\int_{\mathbb{C}} (1 - z) \, d\langle y, \mu(z) x \rangle = \langle y, x - U x \rangle

theorem

Domain Factorization of the Scalar Measure ⟨y,μ(S)(I−U)v⟩\langle y, \mu(S) (I-U) v \rangle

Let HH be a complex Hilbert space and μ\mu be a weak operator topology (WOT) spectral measure on C\mathbb{C} acting on HH. Let U:H→HU: H \to H be a continuous linear operator such that for all x,y∈Hx, y \in H, ⟨y,Ux⟩=∫Cz d⟨y,μ(z)x⟩\langle y, U x \rangle = \int_{\mathbb{C}} z \, d\langle y, \mu(z) x \rangle. Assume that for all x,yx, y, the identity function f(z)=zf(z) = z is integrable with respect to the complex scalar measure μx,y(S)=⟨y,μ(S)x⟩\mu_{x,y}(S) = \langle y, \mu(S) x \rangle and that these measures have finite variation. Further, assume that UU commutes with the projections μ(S)\mu(S) for all measurable sets S⊆CS \subseteq \mathbb{C}. Given vectors x,v,y∈Hx, v, y \in H such that x=v−Uvx = v - U v, then for any measurable set S⊆CS \subseteq \mathbb{C}, it holds that: ⟨y,μ(S)x⟩=∫C(1−z) d⟨y,μ(z)μ(S)v⟩\langle y, \mu(S) x \rangle = \int_{\mathbb{C}} (1 - z) \, d\langle y, \mu(z) \mu(S) v \rangle

theorem

∫g d(qμ)=∫gq dμ\int g \, d(q\mu) = \int gq \, d\mu for real-valued gg and complex density qq

Let α\alpha be a measurable space and μ\mu be a complex-valued vector measure on α\alpha. Let q:α→Cq: \alpha \to \mathbb{C} be a function that is integrable with respect to μ\mu, and let ν=qμ\nu = q\mu be the complex vector measure defined by the density qq with respect to μ\mu. Let g:α→Rg: \alpha \to \mathbb{R} be a real-valued function that is integrable with respect to ν\nu. Provided that the variation of the measure ν\nu satisfies ∣ν∣=∣q∣∣μ∣|\nu| = |q||\mu| (where ∣⋅∣|\cdot| denotes the variation of a vector measure), then the integral of gg with respect to ν\nu is equal to the integral of the product gqgq with respect to μ\mu: ∫g dν=∫g(x)q(x) dμ(x)\int g \, d\nu = \int g(x)q(x) \, d\mu(x) In this identity, the left-hand integral views gg as a real scalar acting on the complex values of ν\nu, while the right-hand integral treats the product gqgq as a complex-valued function.

theorem

μ(I−U)v,y=(1−z)μv,y\mu_{(I-U)v, y} = (1-z) \mu_{v,y} for U=∫z dμU = \int z \, d\mu

Let μ\mu be a weak operator topology (WOT) spectral measure on C\mathbb{C} acting on a Hilbert space HH. Let UU be a continuous linear operator on HH such that UU is the integral of the identity function with respect to μ\mu, i.e., ⟨y,Ux⟩=∫Cz d⟨y,μ(z)x⟩\langle y, U x \rangle = \int_{\mathbb{C}} z \, d\langle y, \mu(z) x \rangle for all x,y∈Hx, y \in H. Assume that: 1. The identity function z↦zz \mapsto z is integrable with respect to the complex scalar measures μx,y\mu_{x,y} defined by S↦⟨y,μ(S)x⟩S \mapsto \langle y, \mu(S) x \rangle. 2. The variation of each scalar measure μx,y\mu_{x,y} is finite. 3. UU commutes with the spectral projections μ(S)\mu(S) for every measurable set S⊆CS \subseteq \mathbb{C}. For any vectors x,v,y∈Hx, v, y \in H such that x=v−Uvx = v - Uv, the scalar measure μx,y\mu_{x,y} is equal to the scalar measure μv,y\mu_{v,y} with density (1−z)(1 - z). That is: μx,y=(1−z)μv,y\mu_{x,y} = (1 - z) \mu_{v,y} or equivalently, for any measurable set S⊆CS \subseteq \mathbb{C}: ⟨y,μ(S)x⟩=∫S(1−z) d⟨y,μ(z)v⟩\langle y, \mu(S) x \rangle = \int_S (1 - z) \, d\langle y, \mu(z) v \rangle

theorem

The Spectral Measure of the Cayley Transform Commutes with the Cayley Transform UU

Let HH be a complex Hilbert space and TT be a self-adjoint partial linear operator on HH. Let U=(T−iI)(T+iI)−1U = (T - iI)(T + iI)^{-1} be the bounded Cayley transform of TT, and let EE be the spectral measure (projection-valued measure) associated with UU. For any measurable set S⊆CS \subseteq \mathbb{C}, the spectral projection E(S)E(S) commutes with the Cayley transform UU: E(S)U=UE(S)E(S) U = U E(S)

theorem

⟨y,Tx⟩\langle y, Tx \rangle is the integral of the inverse Cayley transform against the Cayley spectral measure

Let HH be a complex Hilbert space and TT be a self-adjoint operator on HH with domain D(T)\mathcal{D}(T). Let EUE_U be the spectral measure associated with the bounded Cayley transform U=(T−iI)(T+iI)−1U = (T - iI)(T + iI)^{-1} of TT. Let f:C→Rf: \mathbb{C} \to \mathbb{R} be the inverse scalar Cayley transform defined by: f(z)={−Im z1−Re zif z≠10if z=1 f(z) = \begin{cases} -\frac{\text{Im } z}{1 - \text{Re } z} & \text{if } z \neq 1 \\ 0 & \text{if } z = 1 \end{cases} For any x∈D(T)x \in \mathcal{D}(T) and y∈Hy \in H, the function ff is integrable with respect to the complex scalar measure μx,y\mu_{x,y} defined by μx,y(S)=⟨y,EU(S)x⟩\mu_{x,y}(S) = \langle y, E_U(S) x \rangle, and the operator TT satisfies the following weak integral representation: ⟨y,Tx⟩=∫Cf(z) d⟨y,EU(z)x⟩ \langle y, Tx \rangle = \int_{\mathbb{C}} f(z) \, d\langle y, E_U(z) x \rangle

theorem

Cayley Inverse-Moment Property Implies Weak Spectral Resolution

Let HH be a complex Hilbert space and TT be a self-adjoint operator on HH with domain D(T)\mathcal{D}(T). Let EUE_U be the spectral measure associated with the bounded Cayley transform of TT, and let f:C→Rf: \mathbb{C} \to \mathbb{R} be the inverse scalar Cayley transform. Suppose that for every x∈D(T)x \in \mathcal{D}(T) and y∈Hy \in H, the function ff is integrable with respect to the complex scalar measure S↦⟨y,EU(S)x⟩S \mapsto \langle y, E_U(S) x \rangle and satisfies the identity: ⟨y,Tx⟩=∫Cf(z) d⟨y,EU(z)x⟩ \langle y, Tx \rangle = \int_{\mathbb{C}} f(z) \, d\langle y, E_U(z) x \rangle Then, the real spectral measure ERE_{\mathbb{R}} of TT (defined as the push-forward of EUE_U via ff) is a weak spectral resolution of TT. That is, for every x∈D(T)x \in \mathcal{D}(T) and y∈Hy \in H, the identity function λ↦λ\lambda \mapsto \lambda is integrable with respect to the scalar measure S↦⟨y,ER(S)x⟩S \mapsto \langle y, E_{\mathbb{R}}(S) x \rangle and satisfies: ⟨y,Tx⟩=∫Rλ d⟨y,ER(λ)x⟩ \langle y, Tx \rangle = \int_{\mathbb{R}} \lambda \, d\langle y, E_{\mathbb{R}}(\lambda) x \rangle

theorem

The Real Spectral Measure of a Self-Adjoint Operator TT is a Weak Spectral Resolution of TT

Let TT be a self-adjoint linear operator on a complex Hilbert space HH with domain D(T)\mathcal{D}(T). Let μT\mu_T be the real spectral measure associated with TT (constructed via the Cayley transform). Then μT\mu_T is a weak spectral resolution of TT. That is, for every x∈D(T)x \in \mathcal{D}(T) and y∈Hy \in H, the identity function λ↦λ\lambda \mapsto \lambda is integrable with respect to the complex scalar measure ⟨y,μT(⋅)x⟩\langle y, \mu_T(\cdot) x \rangle, and the following reconstruction formula holds: ⟨y,Tx⟩=∫Rλ d⟨y,μT(λ)x⟩ \langle y, Tx \rangle = \int_{\mathbb{R}} \lambda \, d\langle y, \mu_T(\lambda) x \rangle

theorem

Spectral Theorem for Self-Adjoint Operators via Cayley Transform

Let HH be a complex Hilbert space and TT be a self-adjoint operator on HH. Then TT satisfies the spectral theorem with respect to its real spectral measure μT\mu_T, which is the spectral measure on the real line R\mathbb{R} obtained by pulling back the spectral measure of the Cayley transform of TT to R\mathbb{R}.

theorem

TT is a restriction of its maximal spectral integral T≤∫λ dμTT \le \int \lambda \, d\mu_T

Let HH be a complex Hilbert space and T:D(T)⊆H→HT: D(T) \subseteq H \to H be a self-adjoint partially defined linear operator. Let μT\mu_T be the real spectral measure of TT obtained via the Cayley transform. Then TT is a restriction of the maximal spectral integral of μT\mu_T. That is, T≤∫Rλ dμT(λ) T \le \int_{\mathbb{R}} \lambda \, d\mu_T(\lambda) where the relation ≤\le for partially defined linear operators denotes that D(T)D(T) is contained in the domain of the maximal spectral integral and the operators agree on D(T)D(T).

theorem

A Self-Adjoint Operator TT Equals its Maximal Spectral Integral ∫λ dμT\int \lambda \, d\mu_T

Let HH be a complex Hilbert space and TT be a self-adjoint partial linear operator on HH. Let μT\mu_T be the real spectral measure associated with TT (obtained via the Cayley transform). Then TT is equal to the maximal spectral integral of the identity function with respect to μT\mu_T: T=∫Rλ dμT(λ) T = \int_{\mathbb{R}} \lambda \, d\mu_T(\lambda) This equality implies that both operators have the same domain and agree on all elements of that domain.

theorem

Spectral theorem for self-adjoint operators TT via the Cayley real spectral measure

Let HH be a complex Hilbert space and T:D(T)⊆H→HT: \mathcal{D}(T) \subseteq H \to H be a self-adjoint partial linear operator. Let μT\mu_T be the real spectral measure of TT (the spectral measure on R\mathbb{R} obtained by pulling back the spectral measure of its Cayley transform). Then TT satisfies the domain-aware spectral theorem with respect to μT\mu_T, which characterizes TT as the spectral integral of the identity function λ↦λ\lambda \mapsto \lambda and accounts for the correspondence between the domain D(T)\mathcal{D}(T) and the integrability conditions of the spectral measure.

theorem

Unbounded Spectral Theorem for Self-Adjoint Operators

Let HH be a complex Hilbert space and TT be a self-adjoint linear operator on HH (which may be unbounded). Then TT satisfies the spectral theorem with respect to its real spectral measure μT\mu_T (constructed via the Cayley transform). Specifically, this theorem provides a spectral decomposition where TT is represented as the spectral integral of the identity function with respect to μT\mu_T, formally satisfying the `DomainAwareSelfAdjointSpectralTheorem` interface which relates the operator TT to its measure μT\mu_T while accounting for the operator's domain.

theorem

Unbounded Spectral Theorem for Essentially Self-Adjoint Operators

Let HH be a complex Hilbert space and TT be a partially defined linear operator on HH. If TT is essentially self-adjoint, then its closure Tˉ\bar{T} is self-adjoint and satisfies the unbounded spectral theorem with respect to the real spectral measure μ\mu derived from the Cayley transform of Tˉ\bar{T}. This theorem provides a spectral resolution for Tˉ\bar{T} and establishes that the domain of the operator Tˉ\bar{T} is exactly the set of vectors in HH whose square-moment with respect to μ\mu is finite.