Documentary state: DRAFT_UNACCEPTED. This draft has NOT been editorially admitted. Mathematical claims require independent checking.
OAI-26 family 342: Taming and compatibility in four-dimensional geometry
The question in plain language
In symplectic geometry, a symplectic form can tame an almost-complex structure when it evaluates positively on the pair consisting of a nonzero direction and its image under that structure. Compatibility asks for something stronger: an invariant symplectic pairing that also defines a positive metric. The indexed manuscript addresses when taming can lead to compatibility in dimension four. A local example is easy; the claimed global implication is much harder.
A small example
On ordinary four-dimensional space, identify coordinates with two copies of the complex plane. Let \(J\) rotate each \((x_j,y_j)\) coordinate plane by a quarter-turn, and take
For nonzero \(v\) we have \(\omega_0(v,Jv)=\lVert v\rVert^2>0\), and \(\omega_0(Jv,Jw)=\omega_0(v,w)\). This is a concrete compatible (hence taming) example. It does not prove that arbitrary taming structures are compatible or deformable to compatible ones on a curved four-manifold.
What is known
The relevant background is almost-complex and symplectic geometry on four-manifolds. The upstream collection contains a September 23 manuscript and an October 6 edition with a substantive correction to a conditional cone identity and a strict-inclusion four-torus example. Treating the older and newer manuscripts as one interchangeable statement would conceal a mathematical change.
GCL verification state
GCL's source inventory identified the exact September formalization-source path missing from the October catalogue. The source-risk audit records distinct pinned Git blobs for the two PDFs. Earlier bounded Family 342 Lean replays and semantic observations exist, but the corrected cone theorem, reverse model correspondence and complete current-edition equivalence are not certified.
Documentary status: DRAFT_UNACCEPTED. Source status: VERSION_MISMATCH. Mathematical full claim: UNRESOLVED. NO_CERTIFICATION.
What remains
Write the September-to-October hypotheses and conclusions side by side; identify the corrected cone identity and strict-inclusion witness; trace the relevant definitions into the exact Lean import graph. Determine whether the formalized main existence claim actually implies the latest manuscript's claims, and preserve any surviving unmatched result as a separate obligation. Do not accept an obsolete source edition merely because its Lean code compiles.
Sources and claim boundaries
- GCL Family 342 pinned-edition risk register.
- Exact September/October PDF blob identities.
- MATHFORGE WP04 adversarial source questions.
Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a.
Boundary: the four-dimensional coordinate example is elementary geometry, not evidence for a global existence theorem. Preliminary Lean replay is distinct from complete paper equivalence and mathematical certification. NO_CERTIFICATION.
Indexed primary manuscript paths:
- Taming implies compatibility on four-manifolds —
preprints/Taming-implies-compatibility-on-four-manifolds-October-6-2026/paper.pdf