Documentary state: EDITORIALLY_ADMITTED. The manuscript has a protected editorial admission, not a theorem certification.
OAI-26 family 001: Milne's rationality conjecture for abelian varieties
The question in plain language
Suppose an abelian variety (a higher-dimensional analogue of an elliptic curve) defined in characteristic zero has good reduction modulo a prime. A Hodge class is a cohomological object on the original variety. The October 7 manuscript claims that, after reduction, pairing this class with any complementary collection of divisor classes produces the same rational number across the available cohomology theories. This is weaker than proving the original Hodge class is represented by an algebraic cycle. The manuscript explicitly separates its main rationality theorem from a stronger algebraicity consequence requiring an additional CM Hodge theorem.
A small example
For a codimension-zero class, namely the unit class on an elliptic curve, pairing with the class of a divisor gives its integer degree. This is a basic model of why intersection numbers should be rational. It does not establish the manuscript's main assertion for arbitrary rational Hodge classes, new divisors appearing only after reduction, or crystalline and ell-adic compatibility. In the opposite endpoint \(r=d\), the manuscript specifies an empty product of divisor classes, interpreted as \(1\).
What is known
The pinned PDF states Theorem 1.1 (source label thm:main, PDF pp. 1–2) for an abelian variety over the algebraic closure of the rationals with good reduction at a chosen prime. It fixes a rational Hodge class of codimension between zero and the dimension of the variety, and considers every complementary list of Cartier divisors on the special fiber. It asserts one common rational number for the corresponding top-degree traces in all ell-adic realizations away from the reduction prime and in the de Rham–crystalline realization at that prime. The quantifier applies to each fixed class and divisor list; it does not compare different reduction places.
The introduction attributes absolute-Hodge compatibility, Lefschetz pairing theory, CM transport and integral deformation inputs to earlier sources (PDF pp. 2–4). The manuscript's claimed argument uses two different auxiliary constructions: CM transport of the original specialized class (Corollary 3.4, p. 17), and a possibly non-CM lift with a prescribed filtration preserving a divisor (pp. 27–28). Their comparisons are between marked special fibers; the PDF explicitly claims no map between their generic fibers (Figure 1, p. 5). For balanced Weil classes, Proposition 5.2 (pp. 24–28) claims a dichotomy: all complementary pairings vanish, or the specialized class is a simultaneous rational Lefschetz class, expressible by rational divisor products in every realization. Lemma 6.1 (pp. 29–31) reduces CM classes to these balanced Weil classes, and the proof of Theorem 1.1 is on p. 31. Corollary 6.3 (pp. 31–33) concerns membership in a fixed cohomological theory, not an algebraic cycle; Corollary 6.4 (p. 33) additionally invokes the CM Hodge theorem to conclude algebraicity on the special fiber. These are manuscript assertions, not independently established results in this editorial article.
GCL verification state
The GCL source snapshot indexes Family 001 under one October 7 PDF. The corresponding pinned upstream LaTeX introduction has Git blob c926c1c540c138a4b46274f672fc0f4c43547358; the pinned PDF has Git blob 3780abc1f47a23e546f28190bd3db7b24419f1f9. The source revision is 1581d4cbdd69a0faf34b2089614fe2a814316a6a. The pinned 36-page PDF was acquired and its Git blob independently checked. A bounded source-fidelity comparison covered Theorem 1.1 (pp. 1–2), the proof overview (pp. 3–5), CM transport (p. 17), the Weil-class dichotomy (pp. 24–28), the CM reduction and theorem proof (pp. 29–31), and the two corollaries (pp. 31–33). The article's reported hypotheses, conclusions, and distinction between rationality and conditional algebraicity match those passages. This is not a page-by-page audit of all 36 pages, an independent verification of the cited integral-uniformization inputs, or a proof of the manuscript's theorem. No exact Family 001 Lean declaration correspondence, theorem-level replay, Comparator verdict, full adversarial proof audit or certification has been established.
Documentary status: DRAFT_UNACCEPTED. Semantic status: PINNED_PDF_SELECTED_PASSAGES_FIDELITY_REVIEWED; FULL_PDF_PROOF_AUDIT_AND_FORMAL_BRIDGE_UNRESOLVED. NO_CERTIFICATION.
What remains
Complete a detailed check of the unreviewed PDF passages and the cited external transport, integral-uniformization and filtration inputs, including their hypotheses and applicable editions. The bounded PDF review and critical pass above establish documentary fidelity, not an independent mathematical proof verdict. Establish an actual Lean declaration mapping before attributing compilation or Comparator results to this theorem. Record distinct critical-review evidence addressing the special-fiber transport and rational normalization; keep Corollary 6.3 separate from the CM-Hodge-dependent Corollary 6.4. Complete full MkDocs/public HTML readback, then consider exact-hash editorial admission under protected governance. Do not label this article accepted before that record exists.
Sources and claim boundaries
Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a.
Boundary: an upstream claim, bibliographic indexing, a source-language restatement, formal elaboration and independently certified mathematics are distinct. This first manuscript is an unaccepted source-critical draft only. NO_CERTIFICATION.
Indexed primary manuscript paths:
- Milne's rationality conjecture for abelian varieties —
preprints/Milnes-rationality-conjecture-for-abelian-varieties-October-7-2026/paper.pdf