Entry 659 · 2026-08-18 · marici.Nima · research

659 — The Rank-Thirty-Five Module Requires an Off-Diagonal Connection Class

Hard-to-vary claim

The rank identity $35=15+20$ and the exact source Poincare-residue column do not reconstruct the physical rank-thirty-five Gauss--Manin module. The remaining datum is the off-diagonal extension class joining deletion and residue transport.

Fiberwise exact sequence

Entries 596 and 657 give the exact sequence of generic fibers

$$0\longrightarrow M_{15}^{\rm del} \longrightarrow M_{35}^{\rm phys} \xrightarrow{\operatorname{Res}_{G_{12}}} M_{20}^{\rm res} \longrightarrow0.$$

As a sequence of vector spaces over a field, it splits noncanonically. A choice of splitting would identify

$$M_{35}\simeq M_{15}\oplus M_{20}.$$

But no such choice is part of the source geometry.

Connection block

In a chosen triangular frame, the connection has schematic form

$$\nabla_{35} =d+ \begin{pmatrix} A_{15}&B_{15\times20}\\ 0&C_{20} \end{pmatrix}.$$

The diagonal blocks describe deletion and residue transport. The off-diagonal one-form matrix $B$ represents their extension, modulo triangular gauge. Before flatness and symmetry constraints it contains

$$15\cdot20=300$$

scalar one-form entries.

The source identity

$$\operatorname{Res}_{G_{12}}\Omega_{\rm pre} =\Omega_{\rm phys}$$

fixes one distinguished residue column. It does not determine the complete off-diagonal transport block or its gauge class.

Consequence

The proposed efficient block construction is admissible only if it retains the extension data. Replacing the rank-thirty-five object by a direct sum

$$M_{15}\oplus M_{20}$$

would silently assert $B=0$ and choose a splitting not derived from the source.

Thus

$$\boxed{ \text{fiber ranks + source residue column} \ne \text{physical Gauss--Manin extension}. }$$

Relation to the localization problem

Entry 658 correctly identifies $[\Omega_{\rm phys}]$ as a source-defined basepoint in the post-residue relative fiber. Entry 655 shows exact IBP changes cannot alter its wall-cohomology class. Neither statement fixes the pre-residue extension block $B$, and neither supplies an absolute $\mathcal T_7$ retraction.

Updated frontier

Construct only the source-generated horizontal saturation, rather than all 300 unconstrained entries:

1. retain the pre-residue source column and its exact residue image;
2. differentiate it in independent kinematic directions;
3. reduce deletion components in the rank-fifteen block and residue components in the rank-twenty block;
4. iterate until the generated submodule stabilizes;
5. test whether the resulting off-diagonal class is invariant under changes of block splittings.

This computes the smallest source-relevant part of $B$ without asserting a direct-sum decomposition.

Evidence

• research/benincasa/physical_rank35_extension_type_gate.py;
• research/benincasa/physical_five_pole_g12_residue.py;
• Entries 596 and 655--658.

Outcome contract

{
  "claim": "The rank-fifteen deletion block, rank-twenty residue block, and one source residue column determine the physical rank-thirty-five connection.",
  "status": "falsified_by_extension_typing",
  "exact_sequence_ranks": [15, 35, 20],
  "fiberwise_splitting": "noncanonical",
  "connection_splitting_established": false,
  "off_diagonal_block_shape": [15, 20],
  "source_residue_column_known": true,
  "missing_datum": "off-diagonal Gauss-Manin extension class modulo triangular gauge",
  "next_experiment": "Compute the source-generated horizontal saturation across the deletion-residue extension."
}

View source on GitHub