Skip to content
Shadow Theory

Appendix D: Fejér uniqueness

Full proof of the dissipative classification theorem.

D. Fejér Uniqueness: Proof of Theorem 8.2

Proof of Theorem 8.2. Fix a determinant-admissible spectral class and a positive selfadjoint KK in that class. Put

Hdiss=(ker⁡K)⊥,ΛT,d:=ΛT(K)∣Hdiss.\mathsf H_{\mathrm{diss}}= (\ker K)^\perp,\qquad \Lambda_{T,d} := \Lambda_T(K)\big|_{\mathsf H_{\mathrm{diss}}}.

By (A4), an admissible scalar has determinant/log-volume form

F(K,T)  =  −log⁡det⁡∗ ⁣(ΦT(K)∣Hdiss).F(K,T) \;=\; -\log\det\nolimits_*\!\big(\Phi_T(K)\big|_{\mathsf H_{\mathrm{diss}}}\big).

By the Fejér calibration axiom (A5),

ΦT(K)∣Hdiss  =  c(T)ΛT(K)∣Hdiss  =  c(T)ΛT,d,\Phi_T(K)\big|_{\mathsf H_{\mathrm{diss}}} \;=\; c(T)\Lambda_T(K)\big|_{\mathsf H_{\mathrm{diss}}} \;=\; c(T)\Lambda_{T,d},

where c(T)>0c(T)>0 and log⁡c\log c is bounded on the admissible scale range.

The determinant convention is part of the declared admissible class. Thus, in finite rank, in the ordinary Fredholm class when the scalar factor is admissible, the determinant factorizes as

det⁡∗ ⁣(c(T)ΛT,d)  =  det⁡∗ ⁣(c(T)IHdiss) det⁡∗ ⁣(ΛT,d).\det\nolimits_*\!\big(c(T)\Lambda_{T,d}\big) \;=\; \det\nolimits_*\!\big(c(T)I_{\mathsf H_{\mathrm{diss}}}\big)\,\det\nolimits_*\!\big(\Lambda_{T,d}\big).

Using the definition of DT(K)\mathcal D_T(K),

F(K,T)=−log⁡det⁡∗ ⁣(ΛT,d)−log⁡det⁡∗ ⁣(c(T)IHdiss)=DT(K)+Bc(K,T),\begin{aligned} F(K,T) &= -\log\det\nolimits_*\!\big(\Lambda_{T,d}\big) - \log\det\nolimits_*\!\big(c(T)I_{\mathsf H_{\mathrm{diss}}}\big) \\ &= \mathcal D_T(K) + B_c(K,T), \end{aligned}

where

Bc(K,T)  :=  −log⁡det⁡∗ ⁣(c(T)IHdiss).B_c(K,T) \;:=\; -\log\det\nolimits_*\!\big(c(T)I_{\mathsf H_{\mathrm{diss}}}\big).

By (A5), this modulation counterterm is bounded in TT on the fixed spectral class.

In the finite-rank case, if r=dim⁡Hdiss<∞r = \dim\mathsf H_{\mathrm{diss}}< \infty, the formula is explicit:

det⁡∗ ⁣(c(T)ΛT,d)  =  c(T)rdet⁡∗ ⁣(ΛT,d),\det\nolimits_*\!\big(c(T)\Lambda_{T,d}\big) \;=\; c(T)^r\det\nolimits_*\!\big(\Lambda_{T,d}\big),

and therefore

Bc(K,T)  =  −rlog⁡c(T).B_c(K,T) \;=\; -r\log c(T).

Since log⁡c∈L∞\log c \in L^\infty, Bc(K,T)B_c(K,T) is bounded in TT for fixed rr.

In the ordinary infinite-dimensional Fredholm determinant class, a nontrivial scalar multiple c(T)IHdissc(T)I_{\mathsf H_{\mathrm{diss}}} is not of the form I+L1I + \mathcal L^1 unless c(T)=1c(T) = 1. Hence the Fredholm-admissible infinite-dimensional case has Bc(K,T)=0B_c(K,T) = 0. Other determinant regularizations are outside this theorem.

Thus every admissible scalar in the declared Fejér-calibrated determinant class has the form

F(K,T)  =  DT(K)+B(K,T),B(K,T):=Bc(K,T),F(K,T) \;=\; \mathcal D_T(K) + B(K,T),\qquad B(K,T) := B_c(K,T),

with BB bounded in TT on each fixed spectral class. Axiom (A6) then says that the admissible scalar has genuine unbounded scale dependence as T→∞T \to \infty on the fixed nonzero dissipative sector. Since BB is bounded, the unbounded part is precisely the Fejér determinant budget DT(K)\mathcal D_T(K), up to bounded modulation.

No statement is made about dissipative spectral scalars outside (A1)–(A6) or outside the determinant conventions (D1)–(D2). ◻

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.