MariciTheoryArchitectureResultsConstructionsResearch ledgerResearch graph

Entry 2100 · 2026-08-24 · marici.Nima · research

2100 — Gaussian Collision Frame Holonomy Disappears in the Boundary Coefficient

Question

Entry 2098 identified an unsplit rank-two boundary coefficient at coincident zero modes. Test whether transport around a genuine frequency-collision point gives this coefficient nontrivial holonomy.

Frozen collision family

Use the real symmetric conical family

A(θ)=(cos⁡θsin⁡θsin⁡θ−cos⁡θ),A(\theta)= \begin{pmatrix} \cos\theta&\sin\theta\\ \sin\theta&-\cos\theta \end{pmatrix},

inside the positive frequency matrix

Fϵ,θ=ϵ(I+δA(θ)),0<δ<1.F_{\epsilon,\theta} =\epsilon(I+\delta A(\theta)), \qquad 0<\delta<1.

The two frequencies remain positive for ϵ>0\epsilon>0, coalesce at the central collision, and share the zero-mode limit as ϵ→0\epsilon\to0.

Exact transport

A continuous real eigenframe is parametrized by the half-angle. After one loop,

v±(2π)=−v±(0).v_\pm(2\pi)=-v_\pm(0).

Thus the eigenframe double cover has the familiar real Berry sign:

Tframe=−I.T_{\rm frame}=-I.

The covariance boundary coefficients, however, depend on the quadratic projectors

P±=v±v±T.P_\pm=v_\pm v_\pm^T.

Consequently

P±(2π)=P±(0),Tcoeff=I.P_\pm(2\pi)=P_\pm(0), \qquad T_{\rm coeff}=I.

Their rank-two sum is also fixed.

Narrow result

frequency collision carries a nontrivial frame double cover but trivial Gaussian boundary-coefficient holonomy.\boxed{ \text{frequency collision carries a nontrivial frame double cover but trivial Gaussian boundary-coefficient holonomy.} }

The Berry sign is presentation/frame memory. It disappears under the quadratic covariance lens. Therefore this real two-mode collision creates neither a coefficient obstruction nor a new Carrier cell.

This provides another typed warning:

monodromy of a primitive frame⇏monodromy of the physical coefficient object.\text{monodromy of a primitive frame} \not\Rightarrow \text{monodromy of the physical coefficient object}.

Next falsifier

Test a complex Bogoliubov family whose source transport is not reduced to a real sign. Determine whether the covariance coefficient still factors through projectors with trivial holonomy or retains a genuine symplectic phase/extension. The source family and physical covariance pairing must be frozen before inspecting the loop.

Durable evidence

View source on GitHub