The Step 4 marking is trivial at a nonpositive threshold #
Paper: R-324 — §4.2 — the Step 4 marking is inert at a nonpositive threshold
The proved two-half physical collapse
(R324TwoHalfToNestedCrossBridge, R324CertifiedTwoHalfPhysicalCollapse)
carries the cross-copy covariances in marked form
r324MarkedPairingCovarianceProductOn ε L κ marked B, in which one
distinguished lower endpoint has its covariance replaced by the Step 4(B)
high-frequency projection r324ProjectedCovarianceC ε L. That marking
exists for the ⟨ε²(α+β)⟩⁻⁸ pigeonhole and is not part of the (4.18)
integrand of Steps 2--3.
This module shows the marking degenerates: the retained mode set is
r324HighModeSet ε L = {k | (√ε/2)·L ≤ ‖k‖},
so at L ≤ 0 it is all of ℤ⁴, the projection is the complete
covariance, and the marked product is literally the plain covariance
product of (4.18). Steps 2--3 may therefore run the proved two-half
collapse at L ≤ 0 and read off exactly the paper's (4.18) integrand,
with no marking to account for; Step 4(B) uses the same machinery at the
genuine threshold.
At a nonpositive threshold the Step 4 pigeonhole retains every mode.
Hence every complementary low mode vanishes.
The complementary low-mode covariance is zero at a nonpositive threshold.
The projection is trivial at a nonpositive threshold: it is the
complete spatial covariance η_ε of (4.18).
The marked covariance product is the plain one at a nonpositive
threshold, for any distinguished endpoint — including one inside the
block, where r324MarkedPairingCovarianceProductOn_eq_complete does not
apply.
This is the identification Steps 2--3 need: the proved two-half
physical collapse, run at L ≤ 0, transports exactly the (4.18)
integrand, with the Step 4 marking inert.