Prove native F2 innovation beyond the earlier centre history

[jmbarrancoml]

I prove that a decoder cannot recover the selected F2 query from the complete centre history of calls 0..9 alone. I construct coupled source/target differences for which all 80 earlier fields vanish across all owners, while the eight F2 fields at one owner range over the full payload space.

The conditional theorem is Q(ker H) = M^8. Even with arbitrary decoding, a linear fresh observation needs at least 8 dim(M) complex coordinates to recover that query. I use the retained native common-frame and full-space cover-view hypotheses and verify the selected chronological premises. I infer no recursive-child count or supplier saving.

I credit David Charbonneau's PR3147 for the circle proposal, Nick Petalas's PR3151 for the full-input scalar warning, Aaron Wang/navinor and ywz65 for kernel/owner-route arguments, DLamdata for native endpoint gauges, and the selected native construction lineage. The new scope includes all earlier centre fields and varying original targets; it differs from the source-only terminal factorization and scalar rank16 audit.

I passed the registered Python 3.12 replay with -I -O -B: both complete scalar endpoints, 16 reverse-derived cuts, the 19-event prefix, 112 routes, 1,344 native boundary identities, 112 kernel vectors, six algebraic controls and three premise mutations. The response determinant is 52^8 × 32. I also passed metadata/source-pin and commit-byte validation. I confirmed that GitHub's plan and package replay passed for head 080309ba9e29449ea4f29971e0ce1d9278703350.

Reproduce with:

python3.12 scripts/research_package.py --package research/centre-history-innovation

I leave native compiler admission, fixed-grid realization, fresh-data acquisition, cache restoration and the full supplier invoice open. I keep the selected frontier unchanged. I contribute as jmbarrancoml with substantial OpenAI Codex assistance; I claim no independent human or formal-kernel verification.

Coordination and handoff: Discussion #3149.