Grothendiecks Section Conjecture in Anabelian Geometry

Grothendiecks Section Conjecture in Anabelian Geometry

OpenAI's claim

OpenAI presents the following claims relevant to this question, with manuscript numbers following paper order in Sources. In manuscript 1 (family 019; 2026-09-24), the asserted result is that sections of proper hyperbolic curves over finite extensions of $\mathbb Q_p$ come from points, with a conditional global application through finite-cover descent.

Mathematical statement of the claimed result

Manuscript 1.

For every prime $p$, finite extension $k/\mathbb Q_p$ take a proper smooth curve $X/k$ that remains connected over $\bar k$ and has genus $g(X)\ge2$. consider the exact sequence of profinite groups \[ 1\longrightarrow\pi_1(X_{\bar k})\longrightarrow\pi_1(X) \longrightarrow G_k\longrightarrow1. \] The claimed point-to-section map is a bijection \[ X(k)\ \longrightarrow\ \{s:G_k\to\pi_1(X):s\text{ continuous homomorphism},\; (\pi_1(X)\to G_k)s=\operatorname{id}\}/\pi_1(X_{\bar k}), \] where sections are identified by conjugation by the full geometric group. The cited modular-curve corollary asserts the same bijectivity over $k=\mathbb Q$ for each smooth projective $X_0(N)$ or $X_1(N)$ of genus at least two, for every positive integer $N$ satisfying that genus condition.

Outline of the argument

For manuscript 1 (family 019; 2026-09-24), two stages of auxiliary covers construct roots of arbitrary order on collars while controlling horizontal ramification. Skeletal cycles and Jacobian torus characters complete the universal covering tower, after which the completion criterion identifies point sections.

Formal verification

No Lean formalization is listed at this snapshot for manuscript 1 (family 019; 2026-09-24).

This site has not executed the Lean code. Described statement coverage is not an independent proof check.

Status caveat

The manuscripts were produced by an internal OpenAI model. OpenAI cautions that results without formalization can contain mistakes. This site has not independently checked the proofs; the label records the scope of OpenAI's claim.

Sources

• OpenAI, The p-adic section conjecture, 2026-09-24; citation key OAI:The-p-adic-section-conjecture-September-24-2026. PDF (GitHub); PDF (direct); Manuscript page and citation; Latest version; Statement source: thm:section; Statement source: cor:global-modular-curves.
• OpenAI catalogue entry.

Grothendiecks Section Conjecture in Anabelian Geometry

Definitions and mathematical statement

For a smooth projective geometrically connected curve $C/K$ of genus at least two, choose a geometric base point and consider \[ 1\longrightarrow\pi_1(C_{\bar K})\longrightarrow\pi_1(C) \longrightarrow G_K\longrightarrow1. \] A section is a continuous homomorphism $s:G_K\to\pi_1(C)$ whose composition with the projection is the identity. Sections are identified under conjugation by $\pi_1(C_{\bar K})$. A rational point supplies such a conjugacy class through functoriality of the étale fundamental group. The conjecture says that the map from $C(K)$ to these section classes is bijective.

Formulation and concept references: [S1].

Short English statement

Do rational points on a hyperbolic number-field curve correspond exactly to the sections of its arithmetic fundamental-group sequence?

Research attempt: nonabelian cocycles and a concrete lifting obstruction

Status. We give an exact description of all sections conditional on one existing section, and exhibit why abelianized tests cannot be sufficient in general. These are reductions of the section problem, not a proof that all sections of a curve are point-theoretic. The primary arithmetic treatment [R1] concerns particular birational section results with additional hypotheses, which are not silently removed here.

1. Sections relative to a chosen section

Write the exact sequence as $1\to N\to E\to G\to1$ and suppose a continuous section $s_0$ is given. For instance, a known $K$-point supplies one in the target setting. Define the continuous action $g\cdot n=s_0(g)ns_0(g)^{-1}$. Every section has a unique expression \[s(g)=c(g)s_0(g),\qquad c:G\to N\text{ continuous}.\] Multiplying two such expressions shows that $s$ is a homomorphism exactly when \[c(gh)=c(g)(g\cdot c(h)).\tag{1}\] Conversely, any continuous function satisfying $1$ defines a continuous section. Conjugating $s$ by $n\in N$ changes the cocycle to \[c'(g)=n c(g)(g\cdot n^{-1}).\tag{2}\] Hence section classes are exactly the nonabelian continuous cohomology pointed set $H^1(G,N)$ for this action. This is a bijection obtained from the equations, not merely an injection into a cohomological invariant.

For $A=N^{\mathrm{ab}}$, passage to the abelianization sends each such class to $H^1(G,A)$. In additive notation, $1$ becomes $a(gh)=a(g)+g\cdot a(h)$ and $2$ differs by a coboundary. Thus failure of an abelianized necessary condition can obstruct a proposed section. Satisfaction of that condition need not produce a nonabelian lift.

2. An explicit finite obstruction to lifting abelian data

For an odd prime $p$, let $H$ be the group of upper unitriangular $3\times3$ matrices over $\mathbb F_p$, written as triples $(a,b,c)$ with multiplication \[(a,b,c)(a',b',c')=(a+a',b+b',c+c'+ab').\] The map $H\to\mathbb F_p^2$, $(a,b,c)\mapsto(a,b)$, has central kernel. Suppose the identity homomorphism of $\mathbb F_p^2$ lifted to a homomorphism $\ell:\mathbb F_p^2\to H$. The two coordinate basis vectors commute, so their lifts would commute. But any lifts have forms $(1,0,c)$ and $(0,1,d)$, and their commutator is $(0,0,1)\ne1$. Thus no such lift exists.

The quotient is abelian and the proposed quotient-level map is as simple as possible, yet a central commutator prevents a lift. All groups are finite, so continuity creates no extra issue in this example. This is an elementary model of the nonlinear obstruction lost by abelianization; it is not a counterexample to Grothendieck's conjecture for arithmetic fundamental groups.

3. What the arithmetic problem becomes

If $C(K)$ is nonempty and $s_0$ comes from a chosen point, the conjectured surjectivity says every class in the explicitly defined $H^1(G_K,\pi_1(C_{\bar K}))$ is represented by some rational point. This requires identifying a very special geometric subset of a nonabelian cohomology set. Merely showing that a class has plausible images in abelian or finite quotients does not identify that subset. Compatibility and lifting across quotients can carry further obstructions.

If $C(K)$ is empty, no point supplies $s_0$. The semidirect-product parametrization then cannot be invoked without first proving that some section exists. In that case the conjecture predicts that none exists. Assuming a splitting as a harmless choice of notation would already assume the central disputed fact.

4. Audit and remaining gap

The conjugation equivalence is by the geometric subgroup $N$, exactly as in the target, not by arbitrary elements of $E$. The cocycles and sections are continuous, as required for profinite groups. The finite lifting calculation was checked directly from the multiplication formula. What remains is the arithmetic assertion that all continuous sections are induced by points and that distinct rational points give distinct classes; the general cocycle parametrization proves neither.

Sources

• S1 A. Grothendieck, Brief an G. Faltings, LMS Lecture Notes 242:49-58, 1997. https://doi.org/10.1017/CBO9780511758874.005. Formulation source.
• S2 Polylogarithmic Chabauty--Kim loci over number fields. https://arxiv.org/abs/2608.20615. Further reference; not a proof endorsement.
• S3 On the section conjecture over fields of finite type. https://onlinelibrary.wiley.com/doi/full/10.1002/mana.70049. Further reference; not a proof endorsement.
• R1 Jakob Stix, On the birational section conjecture with local conditions (primary research preprint). https://arxiv.org/abs/1203.3236.