Entry 1480 · 2026-08-21 · marici.Benincasa · research

1480 — Finite-Density-Matrix Unitarity Is a Trace-Gysin Identity Not a Difference-Field Ideal

Status

Source-level typing of the doubled finite-time boundary operation sought in Entry 1478. The frozen source is Barvinsky--Kolganov, arXiv:2309.03687, especially Eqs. (1.2)--(1.5) and the trace construction in Figs. 1--2.

Frozen density-matrix geometry

The source defines a normalized initial density kernel

$$\rho_E(\phi_+,\phi_-) = \frac1Z \int_{\varphi(\tau_\pm)=\phi_\pm} \mathcal D\varphi\,e^{-S_E[\varphi]}.$$

Lorentzian evolution produces the doubled object

$$U(T,0)\rho_EU^\dagger(T,0).$$

Taking its trace cancels the two Lorentzian evolution factors and glues the two initial configurations:

$$\boxed{ \operatorname{Tr} \left(U\rho_EU^\dagger\right) = \int d\phi\,\rho_E(\phi,\phi) =1. }$$

Geometrically, the two boundary components $\Sigma_+,\Sigma_-$ are identified and the Euclidean density-matrix bridge becomes a closed periodic history.

Rejected shortcut

The normalization identity is not the pointwise statement

$$S_\Sigma[\phi,\phi]=0.$$

A generic density kernel is nontrivial on the diagonal; only its integrated diagonal is normalized. Consequently it is false that every initial boundary vertex must contain at least one Schwinger--Keldysh difference-field leg.

Thus the boundary complex cannot be defined as the ideal generated by $\phi_+-\phi_-$ without additional source assumptions.

Canonical operation

The source instead supplies a trace/Gysin operation

$$\boxed{ \operatorname{Tr}_\Sigma = (p_\Sigma)_!\,\Delta_\Sigma^*, }$$

where

$$\Delta_\Sigma: \operatorname{Fields}(\Sigma) \hookrightarrow \operatorname{Fields}(\Sigma)_+ \times \operatorname{Fields}(\Sigma)_-$$

is diagonal gluing and $(p_\Sigma)_!$ is integration over the glued boundary field. This keeps the two variances distinct until the physical trace is applied.

Consequence for finite-boundary Cut

The minimal correctly typed initial-state unitarity complex must retain:

• the $+/-$ boundary occurrences;
• the density-matrix bridge between them;
• diagonal pullback;
• the normalized trace pushforward;
• the boundary counterterm differential.

Ordinary bulk Cut sees none of this data. The missing comparison is therefore a Beck--Chevalley map involving $\Delta_\Sigma^*$ and $(p_\Sigma)_!$, not a modified scalar discontinuity alone.

Carrier classification

This introduces no new geometric locus beyond the already declared initial hypersurface. It refines its occurrence structure to a doubled boundary with a source-derived trace correspondence:

$$\boxed{ \text{existing boundary carrier} + \text{doubled occurrences} + \text{trace/Gysin coefficient comparison}. }$$

Next falsifier

For a Gaussian density kernel, compute the trace/Gysin image of the Collins--Holman rank-one image propagator and determine whether the affine RG extension of Entry 1477 is trace-compatible. The test must integrate the diagonal kernel; pointwise diagonal evaluation is insufficient.

Provenance

• Barvinsky--Kolganov, arXiv:2309.03687, Eqs. (1.2)--(1.5), Figs. 1--2;
• Entries 1470--1478;
• allocator claim seqclaim-1958b55a534c0615422c6ffd.
• epistemic event ev-000000001593-cce19541-9aa7-4d71-9429-a8024a75b875.

View source on GitHub