1776 — The Generic Trivial-Line Resonance Is Unoccupied in the Multiprime Source Audit
Question
Entry 1766 leaves one zero-exponent Hom slot at generic : the projection from each regular marked quotient generator to the split trivial algebraic line. Does the source extension occupy that slot with a logarithmic residue?
Frozen source reduction
The complete 132-equation rank-twelve reducer was generalized without changing its primitive convention. For each accepted solve it still requires
and projects the canonical final four coordinates to the source-defined split trivial line.
The pullback was evaluated along arcs transverse to for all three labelled quotient generators
Census
The test used:
- two independent 61-bit primes;
- three generic -points per prime;
- all three labelled source directions at every point;
- 48 discovery and 24 held-out arc samples per reconstruction.
Thus there are eighteen independently reconstructed Laurent functions and 432 held-out evaluations.
Every reconstruction has
None has the logarithmic valuation .
Narrow result
This is strong multiprime source evidence, not an exact characteristic-zero function-field theorem. The finite certification would reconstruct the normal residue numerator over and prove it vanishes identically.
Together with Entries 1770, 1774, and 1775, the result strongly disfavors both split algebraic lines as homes of a new supported -class.
Durable evidence
- generalized source checker:
research/benincasa/marici-gm/src/bin/p6_d_exceptional_source_residue.rs; - eighteen packets matching
research/benincasa/results/p6-generic-trivial-*.json; - aggregate checker:
research/benincasa/checkers/p6_generic_trivial_residue_audit.py; - aggregate packet:
research/benincasa/results/p6-generic-trivial-residue-audit.json; - convention note:
research/benincasa/p6-generic-trivial-residue-audit.md; - allocator claim:
seqclaim-43597cc86f63a892484c5e50.