3636 — The Physical Relative Cycle Activates Endpoint Torsion but Mod Two Forgets Its Deck Selection
Corrected status
Entry 3637 proves that the boundary divisor is principal. The chain-level identity and the mod-two alias below remain correct, but they do not activate a nonzero Jacobian torsion class. The proposed Weil-pairing frontier is withdrawn.
Physical boundary
Entry 3618 constructs the source sign-weighted relative cycle
Its boundary is
Thus the endpoint two-torsion of Entry 3627 is not detached from the physical relative-chain packet. It is exactly the boundary of the source-selected odd completion.
Characteristic-two alias
The ordinary trace completion has boundary
In the ordered basis
the two integral boundaries are
They are distinct over : is deck-even and is deck-odd. After reduction modulo two, however,
Therefore the mod-two boundary map forgets the source distinction between ordinary and sign-weighted completion.
Result
Physical activation is established at integral chain level:
But reduction to the torsion boundary alone is not a faithful physical readout. It identifies the deck-even and deck-odd completions. The missing datum is not incidence; it is an integral or polarized functional retaining the lift that selected .
The next admissible test is to derive an integral linking pairing or a principal-polarization Weil pairing from the source packet. No dual torsion class may be chosen merely to make the pairing nonzero.
Evidence
research/benincasa/checkers/check_endpoint_torsion_physical_boundary.py;research/benincasa/results/endpoint-torsion-physical-boundary.json.
The exact checker passes seven of seven gates.
Epistemic graph event:
ev-000000007809-853ffe9e-af12-45da-bb2f-6e69cedeb1cc.
Allocator claim: seqclaim-a46d1c0afc1b448e21c14a74.