Family 062 — Contact geometry forces a homogeneous model

← All digests

Family 062 · Algebraic and complex geometry
Importance 5/5 · Provisional editorial rating
Sources consulted 2026-10-12

1. The question

A contact structure chooses a hyperplane of allowed tangent directions at every point, but arranges those hyperplanes so that they twist as much as possible. A Fano manifold has a different kind of positivity, encoded by its anticanonical bundle. Together, do these conditions force the manifold to be one of the highly symmetric spaces arising from simple Lie algebras?

The contact-Fano homogeneity conjecture says yes. Through twistor geometry, it is closely connected to the LeBrun–Salamon conjecture: every compact positive quaternionic-Kähler manifold should be a symmetric Wolf space.

Family 062 claims both conclusions. Its algebraic theorem preserves the given contact distribution, and its metric conclusion identifies the original metric up to constant scaling. The proposed mechanism constructs enough global contact vector fields to move any point in any tangent direction, thereby converting local contact geometry into global symmetry.

2. Background and essential definitions

A maximally twisting hyperplane field

Let $X$ have complex dimension $2n+1$. A contact structure is a rank-$2n$ subbundle $F\subset T_X$ with line-bundle quotient $L$:

$$0\longrightarrow F\longrightarrow T_X \xrightarrow{\theta}L\longrightarrow0.$$

Locally, after choosing a frame for $L$, the quotient map is a one-form $\alpha$. The contact condition is

$$\alpha\wedge(d\alpha)^n\ne0$$

at every point. Equivalently, the bracket of vector fields induces a nondegenerate skew pairing $F\otimes F\to L$. This contrasts with an integrable hyperplane field, whose tangent directions close under brackets and form local leaves.

A concrete local model is $\alpha=dz-\sum_{j=1}^n y_j\,dx_j$. Moving in permitted directions can generate a displacement in the missing $z$-direction through commutators. This is a useful local picture; the classification problem concerns whether compact global positivity forces a particular homogeneous realization.

A smooth projective variety is Fano when $-K_X$ is ample, meaning some positive tensor power embeds it in projective space. For a contact manifold,

$$-K_X\simeq L^{\otimes(n+1)}.$$

Thus the contact line bundle is ample in the Fano case.

Homogeneous spaces and quaternionic metrics

A homogeneous manifold has an automorphism group acting transitively on its points. A simple complex Lie algebra $\mathfrak g$ has no nontrivial proper ideals and is nonabelian. Its adjoint variety is the unique closed orbit in $\mathbb P(\mathfrak g)$, equivalently the projectivization of its smallest nonzero nilpotent adjoint orbit. It carries a canonical contact structure.

A Riemannian manifold of real dimension $4m$, with $m\ge2$, is quaternionic-Kähler here if its holonomy lies in $\operatorname{Sp}(m)\operatorname{Sp}(1)$. Holonomy describes the transformations of tangent vectors obtained by parallel transport around loops. The name does not assert that the manifold is a complex Kähler manifold with one globally chosen complex structure.

“Positive” means positive scalar curvature, an averaged curvature quantity. It does not mean every two-dimensional tangent plane has positive sectional curvature. The compact symmetric examples are called Wolf spaces. Two metrics are homothetic if an isometry identifies one with a positive constant multiple of the other.

3. Development history and previous partial results

Wolf related quaternionic symmetric spaces to homogeneous complex contact spaces. Salamon and, independently, Bérard-Bergery developed the positive twistor construction: a positive quaternionic-Kähler manifold gives a contact Fano twistor space. LeBrun’s converse identifies which contact Fano manifolds arise in this way through existence of a Kähler–Einstein metric. LeBrun–Salamon’s rigidity and finiteness program provides the broader conjectural setting. Wolf, LeBrun, LeBrun–Salamon.

The algebraic problem had major reductions. Kebekus–Peternell–Sommese–Wiśniewski and Demailly reduced projective contact classification to the Fano case and projectivized cotangent bundles. Within the Fano case, classical positivity theorems identify projective-space and incidence-space alternatives. Kebekus developed contact lines, and Buczyński studied their two-line loci and associated tensors. These provide the starting geometry for M1. Projective contact manifolds, Buczyński’s corrected contact-line paper.

Low-dimensional classifications include real dimension eight on the quaternionic side and complex dimensions three and five on the contact side. Buczyński–Wiśniewski obtained real dimensions twelve and sixteen and further contact results with reductive automorphism group. Other theorems use bounds on torus actions or assumptions on the contact linear system. Brendle–Semmelmann established symmetry under nonnegative sectional curvature. M1 claims to remove such extra assumptions in its stated settings. Buczyński–Wiśniewski, Brendle–Semmelmann.

The source also acknowledges previous broad proof claims. Kobayashi’s preprint claims the Riemannian classification under its complete locally irreducible hypotheses. Yasukura’s earlier unrestricted assertions were subsequently revised; the author’s version 12 explicitly records false statements in version 11 and retains a theorem assuming an immersive contact line bundle. These are distinct historical claims and corrections, not a basis for independently assessing the present proof. Kobayashi’s cited version, Yasukura’s revised statement.

4. What the new manuscripts claim

Contact-Fano classification with the distribution retained

M1 claims that every smooth connected complex projective contact Fano manifold of dimension $2n+1\ge3$ is contact-isomorphic to the adjoint variety of a simple complex Lie algebra. The isomorphism carries the specified $F$ to the canonical contact distribution and identifies $L$ with the hyperplane bundle of the adjoint embedding.

Two familiar cases require care with the polarization. For $X=\mathbb P^{2n+1}$, the contact line is $\mathcal O(2)$; the adjoint embedding is quadratic. For

$$X=\mathbb P(T^*\mathbb P^{n+1}),$$

the space is the incidence variety of points and hyperplanes, and the contact line is the restriction of $\mathcal O(1,1)$. The remaining proof works in the primitive Picard case $\operatorname{Pic}(X)=\mathbb Z[L]$.

A consequence is a positive-scalar-curvature Kähler–Einstein metric for the given complex structure of every contact Fano manifold. Kähler–Einstein means its Ricci curvature tensor is a constant multiple of the metric.

General projective contact manifolds

Combining the classification with the established reductions, M1 states the following alternatives for any smooth connected projective contact manifold of dimension at least three. If $b_2(X)=1$, it is contact-isomorphic to an adjoint variety. If $b_2(X)\ge2$, its underlying manifold is biholomorphic to $\mathbb P(T^*Z)$ for a smooth connected projective manifold $Z$.

In the second alternative, the conclusion identifies the underlying manifold without asserting that the displayed isomorphism preserves the initially given contact distribution. Here projectivization parametrizes lines in cotangent spaces, equivalently hyperplanes in tangent spaces. Each listed underlying model admits a contact structure.

The original quaternionic metric is symmetric

For every closed connected smooth quaternionic-Kähler manifold $(M^{4m},g)$, with $m\ge2$ and positive scalar curvature, M1 claims

$$\nabla\operatorname{Rm}_g=0,$$

and that $(M,g)$ is homothetic to a compact symmetric Wolf space. The displayed equation means the curvature tensor is parallel. The result is a metric classification, not just a diffeomorphism or a description of a finite cover.

The source does not assume simple connectivity. It obtains it using the contact Fano twistor space and its sphere-bundle structure before recovering the normalized metric from the complex twistor model.

5. Why the advance matters

Importance: 5/5. If correct, the theorem would settle two central classification predictions connecting algebraic contact geometry and Riemannian holonomy. It would show that positivity and contact structure force the full symmetry of a Lie-theoretic model in every allowed dimension.

The strength is visible in what is retained. The algebraic result remembers the given contact distribution, and the Riemannian result remembers the metric up to scale. Merely proving that some manifold in the same topological class has a symmetric metric would be much weaker.

The proposed mechanism could also matter beyond classification. It turns the behavior of rational curves near a diagonal into global first-jet generation, then into transitive contact automorphisms. The manuscript states this jet criterion separately, without a Fano assumption once its tensor hypothesis is supplied.

6. The main idea

Smooth two contact lines into a flexible conic

A contact line is a map from $\mathbb P^1$ whose pullback of $L$ has degree one. A conic has contact degree two. These names specify degree, not an assumed embedding.

After the known Picard alternatives are removed, the source chooses two contact lines through a general point with tangent vectors having nonzero Levi pairing. Their union can be smoothed to a conic while its outer points move. The nonzero pairing supplies the missing deformation direction for two-point evaluation.

The resulting conic is very free: every summand of its pulled-back tangent bundle has positive degree. This permits independent variations of positions and velocities, essential for the later local expansion.

A two-point tensor descends to the whole square

On a smooth rational source curve, the expression

$$(z-w)^2\partial_z\otimes\partial_w$$

is coordinate-independent in the appropriate tensor bundle. The manuscript applies it to a proper compactification of the conic family, with two variable evaluation points. At a nodal degeneration into two lines, vanishing of the contact differential cancels possible poles.

Two vector-valued lifts then give a common contact value. By testing on contact lines, the source excludes branching of that value and descends it to a symmetric global section

$$B\in H^0(X\times X,L\boxtimes L).$$

The external product $L\boxtimes L$ means the tensor product of the line bundle pulled back from each factor. Descending this section is stronger than constructing separate values only along individual conics.

Diagonal derivatives generate symmetry

Near the diagonal $x=y$, the source identifies the leading quadratic behavior of $B$ with the square of the contact form. Symmetry determines the relevant cubic behavior as well. These relations imply that global sections of $L$ generate every first jet: one can prescribe both a value and its first derivatives at any point.

Contact Hamiltonians associate sections of $L$ with vector fields preserving the contact distribution. First-jet generation makes those fields span every tangent space. Their flows therefore give a transitive contact automorphism group, and the contact moment map identifies the manifold with a simple adjoint variety.

For the metric conclusion, the positive twistor construction produces such a contact Fano space from $M$. Wolf’s correspondence supplies the symmetric model, and LeBrun’s metric uniqueness theorem recovers the normalized original metric from the complex twistor space. Undoing the normalization gives homothety.

7. Scope, limitations, and open questions

The algebraic theorem requires smoothness, projectivity, contact nondegeneracy everywhere, and Fano positivity. It is not a classification of singular contact spaces or of every noncompact complex contact manifold.

The Riemannian statement starts in real dimension eight and assumes positive scalar curvature. It does not classify zero- or negative-scalar-curvature quaternionic-Kähler manifolds, nor does it address the separate conventions needed in real dimension four.

The general projective contact corollary has a weaker compatibility statement in the projectivized-cotangent case. The finite-dimensional Lie-theoretic classification in the Fano case should not be used to restrict the arbitrary projective base $Z$ in that other branch.

Further problems include singular and nonprojective extensions and understanding when the symmetric-tensor criterion can be constructed without Fano positivity. The manuscript’s all-dimensional claim is reported as stated; its curve deformation and tensor descent arguments are not independently verified here.

8. Sources and further reading

M1’s Theorem 1.1 is the contact classification; Corollaries 1.2–1.4 give Kähler–Einstein existence, the projective-contact alternatives, and the Riemannian conclusion. Sections 3–6 construct and descend the conic tensor, Section 7 establishes the jet criterion, and Sections 8–9 identify the algebraic and metric models.

LeBrun’s twistor correspondence explains the bridge between the two geometric settings. The primary historical links in Section 3 describe the established reductions, low-dimensional results, curvature-based partial theorem, and corrected earlier claims.

Source manuscripts

• M1. OpenAI, Contact Fano manifolds and the LeBrun–Salamon conjecture (September 23, 2026).

Sources consulted on 12 October 2026 at repository revision fd4aeeb. Download the manuscript citations.