Skip to content
Shadow Theory

The unified field theorem

The joint operator calculus and recovery of spectral algebras from readouts.

Once the realization and symbol are fixed, the two responses belong to one spectral algebra. This chapter also gives the precise conditions under which a family of readouts recovers that algebra.

7. The Unified Field Theorem

Definition 7.1. Canonical Spectral-Curvature Field.

Given an admissible monotone curvature system with canonical generator KK and an admissible coherence symbol hh, the Canonical Spectral-Curvature Field is

FK:=( K,  h(K),  e−tK,  e−it h(K),  ΛT(K),  ΓT(h(K)),  DT(K),  BT(K) ).\mathfrak F_K := \big(\, K,\; h(K),\; e^{-tK},\; e^{-it\,h(K)},\; \Lambda_T(K),\; \Gamma_T(h(K)),\; \mathcal D_T(K),\; \mathfrak B_T(K)\, \big).

Theorem 7.2. Unified Spectral-Curvature Field.

Let K=KM≥0K=K_{\mathcal M}\ge0 be the Friedrichs representative of an admissible curvature system, and fix an admissible symbol hh. Then:

  1. The bounded operators h(K)h(K), e−tKe^{-tK}, e−ith(K)e^{-it h(K)}, ΛT(K)\Lambda_T(K), ΓT(h(K))\Gamma_T(h(K)) and CT\mathbf C_T belong to the abelian algebra W∗(K)W^*(K). The possibly unbounded KK is affiliated with it.

  2. These bounded operators commute. They act through the same projection-valued measure EKE_K, with scalar mode responses h(λ)h(\lambda), e−tλe^{-t\lambda}, e−ith(λ)e^{-it h(\lambda)}, fT(λ)f_T(\lambda), aT(h(λ))a_T(h(\lambda)) and qT(h(λ))q_T(h(\lambda)), where aT(u)=(1−e−iTu)/(iTu)a_T(u)=(1-e^{-iTu})/(iTu) and aT(0)=1a_T(0)=1.

  3. H0=ker⁡K\mathsf H_0=\ker K is fixed pointwise by StS_t, UtU_t and ΛT\Lambda_T.

  4. Wherever (D1)–(D3) apply, DT\mathcal D_T and BT\mathfrak B_T are unitary-invariant scalar functionals of the same spectral data, with the full resonance convention for BT\mathfrak B_T.

  5. A unitary transport K~=VKV∗\widetilde K=VKV^* carries every bounded readout to Vϕ(K)V∗=ϕ(K~)V\phi(K)V^*=\phi(\widetilde K) and preserves defined budgets. Thus a specified physical realization of (K,h)(K,h) realizes both faces on one carrier.

Proof. The bounded Borel functional calculus is a unital ∗*-homomorphism. It gives (i), commutation and the spectral integrals in (ii). At λ=0\lambda=0, h(0)=0h(0)=0 and fT(0)=1f_T(0)=1, proving (iii). The determinant products and trace-log expressions depend only on the spectral multiplicities, proving (iv). The spectral theorem gives EVKV∗(B)=VEK(B)V∗E_{VKV^*}(B)=VE_K(B)V^*, so all functional calculi and their domains transport as stated. This proves (v). ◻

The theorem identifies one operator architecture with two complementary responses. Physical identification of a measured relaxation generator with KK specifies the dissipative realization. Identification of a measured coherent generator with the nominated h(K)h(K) specifies a paired physical realization; the mathematical construction already supplies the coherent readout associated to that choice.

Corollary 7.3. Combined attenuating phase evolution.

For γ>0\gamma>0, the effective operator Heff=h(K)−iγKH_{\rm eff}=h(K)-i\gamma K is normal on domK\mathop{\mathrm{dom}}K, and its forward evolution is

E(t)=e−γtKe−ith(K),t≥0.E(t)=e^{-\gamma tK}e^{-it h(K)},\qquad t\geq0.

It is a contraction semigroup. On a finite retained sector, [Heff,Heff∗]=0[H_{\rm eff},H_{\rm eff}^*]=0 and ddt∥E(t)u∥2=−2γ⟨E(t)u,KE(t)u⟩\frac{d}{dt}\left\lVert E(t)u\right\rVert^2=-2\gamma\langle E(t)u,KE(t)u\rangle.

Proof. The complex spectral multiplier h(λ)−iγλh(\lambda)-i\gamma\lambda defines a normal operator with domain domK\mathop{\mathrm{dom}}K, since hh is bounded. Its exponential is e−γtλe−ith(λ)e^{-\gamma t\lambda}e^{-it h(\lambda)}, proving factorization and contractivity. In finite dimension, commutation and the derivative follow directly from the same spectral multipliers. ◻

Thus the common-carrier non-Hermitian realization is a normal subclass. Reciprocal induction and memory may determine its local carrier through a nonlinear physical model, as in Section 12; arbitrary noncommuting reversible–irreversible dynamics need not have this form.

7.1 Recovery of the spectral algebra from readouts

Theorem 7.4. Spectral algebra recovery from separating readouts.

Let K=K∗K = K^* be selfadjoint, let X:=σ(K)X := \mathop{\mathrm{\sigma}}(K), and let {mα:X→C}α∈A\{m_\alpha:X\to\mathbb C\}_{\alpha\in A} be bounded Borel functions. Assume that the sigma-algebra generated by the readouts is the full Borel sigma-algebra of XX:

σ(mα:α∈A)  =  B(X).\sigma(m_\alpha:\alpha\in A) \;=\; \mathfrak B(X).

Then

W∗(mα(K):α∈A)  =  W∗(K).W^*(m_\alpha(K):\alpha\in A) \;=\; W^*(K).

In particular, if a countable subfamily {mn}n≥1\{m_n\}_{n\ge 1} separates points of XX, then the hypothesis holds and

W∗(mn(K):n≥1)  =  W∗(K).W^*(m_n(K):n\ge 1) \;=\; W^*(K).

Proof. Let M:=W∗(mα(K):α∈A)\mathcal M := W^*(m_\alpha(K):\alpha\in A). Since each mα(K)m_\alpha(K) is a bounded Borel function of KK, M⊆W∗(K)\mathcal M\subseteq W^*(K).

Conversely, define

S  :=  {B∈B(X):EK(B)∈M}.\mathfrak S \;:=\; \{B\in\mathfrak B(X) : E_K(B)\in\mathcal M\}.

This is a sigma-algebra. For each α\alpha and each Borel set B⊆CB\subseteq\mathbb C,

EK(mα−1(B))  =  Emα(K)(B)  ∈  M.E_K(m_\alpha^{-1}(B)) \;=\; E_{m_\alpha(K)}(B) \;\in\; \mathcal M.

Hence S\mathfrak S contains the sigma-algebra generated by all mαm_\alpha. By hypothesis, S=B(X)\mathfrak S = \mathfrak B(X). Therefore every spectral projection EK(B)E_K(B) belongs to M\mathcal M. Since

W∗(K)  =  W∗(EK(B):B∈B(X)),W^*(K) \;=\; W^*(E_K(B):B\in\mathfrak B(X)),

W∗(K)⊆MW^*(K)\subseteq\mathcal M. Thus equality holds.

It remains only to justify the countable point-separating clause. Suppose {mn}n≥1\{m_n\}_{n\ge 1} separates points of XX. The map

m:X→CN,m(λ)=(m1(λ),m2(λ),… ),m:X\to\mathbb C^{\mathbb N},\qquad m(\lambda) = (m_1(\lambda), m_2(\lambda),\dots),

is Borel and injective. Since X⊆RX\subseteq\mathbb R is a standard Borel space, the Lusin–Souslin theorem implies that m(X)m(X) is Borel in CN\mathbb C^{\mathbb N} and that m−1:m(X)→Xm^{-1}:m(X)\to X is Borel. Hence every Borel subset of XX is the inverse image under mm of a Borel subset of m(X)m(X), and therefore belongs to σ(mn:n≥1)\sigma(m_n:n\ge 1). Thus σ(mn:n≥1)=B(X)\sigma(m_n:n\ge 1) = \mathfrak B(X), and the first part applies. ◻

Corollary 7.5. CSCF readout recovery.

Let K≥0K\ge 0 be selfadjoint.

  1. For every fixed T0>0T_0 > 0,

    W∗(ΛT0(K))  =  W∗(K).W^*(\Lambda_{T_0}(K)) \;=\; W^*(K).

    Indeed, fT0(λ)=T0−1∫0T0e−tλ dtf_{T_0}(\lambda) = T_0^{-1}\int_0^{T_0}e^{-t\lambda}\,dt is bounded Borel and strictly decreasing on [0,∞)[0,\infty), hence separates points of σ(K)\mathop{\mathrm{\sigma}}(K). This is an exact operator-algebraic statement; it does not imply that finitely sampled or noisy empirical estimates of ΛT0(K)\Lambda_{T_0}(K) recover KK without a separate inverse/stability analysis. Theorem 5.4 is in fact stronger than this fixed-scale algebra-recovery statement: it gives the explicit operator-level inversion K=ψT0(ΛT0(K))K = \psi_{T_0}(\Lambda_{T_0}(K)) rather than mere equality of generated von Neumann algebras.

  2. If the coherence symbol hh is injective on σ(K)\mathop{\mathrm{\sigma}}(K), then

    W∗(h(K))  =  W∗(K).W^*(h(K)) \;=\; W^*(K).

    This injectivity condition is sufficient. Algebra recovery is insensitive to changes on EKE_K-null sets, so pointwise injectivity on the entire topological spectrum is not a necessary condition.

  3. Let qT(u):=sinc2(Tu/2)q_T(u) := \mathop{\mathrm{sinc}}^2(Tu/2). If a countable family of coherent compression readouts {qTn∘h}n≥1\{q_{T_n}\circ h\}_{n\ge 1} separates points of σ(K)\mathop{\mathrm{\sigma}}(K), then

    W∗(ΓTn(h(K))∗ΓTn(h(K)):n≥1)  =  W∗(K).W^*(\Gamma_{T_n}(h(K))^*\Gamma_{T_n}(h(K)):n\ge 1) \;=\; W^*(K).
  4. More generally, any declared bounded Borel CSCF readout family among fTf_T, hh, qT∘hq_T\circ h, or their countable combinations, recovers W∗(K)W^*(K) whenever it generates B(σ(K))\mathfrak B(\mathop{\mathrm{\sigma}}(K)). More generally, recovery is governed by the generated measurable algebra modulo EKE_K-null sets. Readouts that identify spectral distinctions of positive spectral measure recover a corresponding proper observable subalgebra.

7.2 The common spectral calculus

The dissipative face reads the positive-real axis of a single spectral calculus. The coherent face reads the phase / imaginary axis of the same calculus. There is one carrier KK, two readouts.

For the unregularized generator, the semigroup e−zKe^{-zK} is holomorphic for Re⁡z>0\operatorname{Re}z>0 and has the unitary boundary group e−itKe^{-itK} as a strong limit. Thus a literal restriction to real and imaginary time gives

z=t∈R+:e−zK=e−tK(dissipative),z=it:e−itK(unregularized coherent).\begin{aligned} z=t\in\mathbb R_+ &: \quad e^{-zK}=e^{-tK} &&\text{(dissipative)},\\ z=it &: \quad e^{-itK} &&\text{(unregularized coherent)}. \end{aligned}

The strong boundary limit follows by dominated convergence in each vector’s spectral measure. A general symbol hh instead selects the spectral reparameterization C=h(K)C=h(K) within the same Borel calculus. Its unitary group is bounded-generator and norm-continuous. When h(λ)=λh(\lambda)=\lambda on a retained spectral window, it agrees there with the imaginary-time boundary group; otherwise the scalar phase frequencies are reparameterized, while the common spectral projections are preserved.

The physical realization and source-selection questions are treated separately from the common-carrier theorem.

Published source: The Canonical Spectral-Curvature Field: A Unified Operator Origin for Coherence, Dissipation, and Spectral Budgets. Read the original PDF alongside the web treatment.