🔢 mathematics

Projective subvarieties of Bogomolov-Guan manifolds and quasi-diagonals in products of elliptic curves

This paper classifies projective subvarieties in Bogomolov-Guan manifolds by introducing and analyzing quasi-diagonals in products of elliptic curves, demonstrating that for a general such manifold, any projective subvariety is contained within a fiber of the Lagrangian fibration.

Original authors: Leila Abubakarova, Alexandra Kuznetsova, Misha Verbitsky

Published 2026-06-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Leila Abubakarova, Alexandra Kuznetsova, Misha Verbitsky

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are exploring a vast, strange, and beautiful landscape made of mathematics. This landscape is called a Bogomolov-Guan manifold. Think of it as a high-dimensional, twisted version of a torus (like a donut shape), but with some very peculiar rules: it's "non-Kähler," which is a fancy way of saying it doesn't play by the standard rules of geometry that most shapes we know follow. It's a bit like a world where parallel lines might eventually meet, or where the concept of "area" behaves differently.

The authors of this paper, Leila Abubakarova, Alexandra Kuznetsova, and Misha Verbitsky, are trying to answer a simple but tricky question: What kinds of smaller shapes (subvarieties) can exist inside this weird landscape?

Here is a breakdown of their journey and findings, using everyday analogies.

1. The Map and the Territory

To understand this strange landscape, the authors first look at a simpler "precursor" or a "blueprint" for it. They call this the Bogomolov-Guan precursor.

Imagine the landscape is built from a giant grid of elliptic curves. An elliptic curve is just a fancy name for a donut shape (a circle with a hole, but in complex math). If you take two donuts and multiply them together, you get a 4D shape. If you take many, you get a huge grid.

The authors are looking for specific paths or curves that can live on this grid. They call these special paths "Quasi-diagonals."

2. What is a "Quasi-Diagonal"?

Imagine you are walking on a grid made of two streets (Street A and Street B).

A normal "diagonal" path would be a straight line where you walk the same distance on Street A as you do on Street B.
A Quasi-diagonal is a more flexible path. It's a curve where, if you look at how your path relates to the "twist" or "bundle" of the streets, the relationship is perfectly balanced in a specific mathematical way.

Think of it like a dance. If the two dancers (the two streets) are holding hands with a specific type of elastic band (the line bundle), a quasi-diagonal is a dance move where the tension in the elastic band cancels out perfectly, leaving the dancers in a state of "torsion" (a fancy math word for a loop that can be untied if you go around it enough times).

3. The Big Discovery: Scarcity

The most exciting part of the paper is the discovery about how many of these special paths exist.

You might think that on an infinite grid, you could find an endless, continuous river of these special paths. You could slide a ruler along the grid and find a new path at every millimeter.

The authors prove this is wrong.

They show that for any given setup, there are at most a countable number of these quasi-diagonals.

The Analogy: Imagine you are looking for a specific type of rare bird in a massive forest. You might expect to see them everywhere. But the authors prove that these birds only nest in specific, isolated trees. You can list them all: Bird #1, Bird #2, Bird #3... but you will never find a "continuous flock" where they are everywhere at once. They are discrete, isolated, and rare.

4. Why Does This Matter for the Landscape?

The authors use this discovery to solve the puzzle of the Bogomolov-Guan manifold.

They found that the "projective" shapes (the ones that look like standard, well-behaved geometric objects) inside this weird landscape are extremely limited.

The Rule: If you want to find a nice, projective shape inside this manifold, it must be stuck inside a single "fiber" (a slice) of the landscape's structure. It cannot stretch across the whole map.
The Reason: Because the "quasi-diagonals" (the paths that would allow a shape to stretch across) are so rare and isolated, a shape can't form a continuous bridge across the manifold unless it stays put in one small area.

5. Two Ways to Prove It

The authors didn't just guess this; they proved it in two different ways, like solving a mystery with two different detectives:

The Algebraic Detective: They used pure math equations and logic about how these shapes fit together to show that a continuous family of these paths would lead to a mathematical contradiction.
The Geometric Detective: They used tools from complex geometry (like measuring the "volume" of shapes) to show that if these paths existed in a continuous family, the whole landscape would have to change its fundamental nature, which it doesn't.

Summary

In simple terms, this paper explores a strange, twisted mathematical world. The authors discovered that the only "nice" shapes that can live inside this world are very restricted. They can't wander freely across the whole landscape; they are forced to stay in specific, isolated pockets. This is because the "bridges" (quasi-diagonals) that would allow them to travel are incredibly rare—there are only a few of them, not an infinite stream.

The paper is a rigorous proof of scarcity: in this specific mathematical universe, beautiful, well-behaved shapes are rare gems, not common stones.

Technical Summary: Subvarieties in Bogomolov-Guan Precursors

Problem Statement
The paper investigates the geometric structure of complex subvarieties within Bogomolov-Guan (BG) manifolds and their precursors. BG manifolds are compact, simply connected, holomorphically symplectic manifolds that are non-Kähler, constructed by D. Guan and analyzed by F. Bogomolov. While previous work (specifically [BKKY]) established that the algebraic reduction of a BG manifold is a Lagrangian fibration $P: BG \to \mathbb{P}^n$ and that subvarieties mapping to dimensions $\geq 2$ cannot be projective, the classification of projective (or Moishezon) subvarieties mapping to curves remained incomplete. The central problem is to characterize exactly which subvarieties in these non-Kähler settings are projective and to understand the scarcity of such objects.

Methodology
The authors employ a multi-faceted approach combining algebraic geometry, complex differential geometry, and the theory of nilmanifolds:

Precursor Analysis: Instead of working directly with the intricate construction of BG manifolds (which involve Hilbert schemes of Kodaira surfaces), the authors focus on "BG-precursors." These are simpler complex manifolds constructed as quotients of products of Kodaira surfaces by diagonal actions of elliptic curves. The paper establishes a generically finite morphism between the BG manifold and a finite quotient of its precursor, allowing results on precursors to be transferred to BG manifolds.
Quasi-Diagonals: The core geometric object introduced is the "quasi-diagonal." Defined in the product of elliptic curves $E^n$ , a curve $S$ is a quasi-diagonal over a pair $(E, L)$ (where $L$ is an ample line bundle) if the line bundle $p_i^*L \otimes p_j^*L^{-1}$ is torsion on $S$ for all projections $p_i, p_j$ . The classification of projective subvarieties in BG precursors is reduced to the classification of these quasi-diagonals.
Algebraic Rigidity: The authors utilize the theory of principal toric bundles and Chern classes. They apply Blanchard's theorem and its generalizations to show that for a subvariety to be projective (or Fujiki class C), the pullback of specific line bundles must be torsion. This leads to a reduction of the problem to studying families of quasi-diagonals.
Complex Analytic Tools: To prove the scarcity of quasi-diagonals, the paper employs two distinct methods:
- Algebraic: A rigidity argument using the properties of the universal family of curves and the behavior of line bundles under pullback, leading to a contradiction if an uncountable family of quasi-diagonals exists.
- Analytic: The authors demonstrate that BG precursors admit a pluriclosed (SKT) metric. By invoking Gromov's compactness theorem for Barlet spaces of curves on pluriclosed manifolds and Campana's coreduction theorem, they prove that the moduli space of such curves is compact, implying that continuous families of quasi-diagonals cannot exist.

Key Contributions and Results

Classification of Projective Subvarieties (Theorem 1.3 & 4.9): The paper provides a complete characterization of projective and Fujiki class C subvarieties in BG manifolds and their precursors.
- A subvariety $Z$ is projective if and only if its image under the Lagrangian fibration is either a point or a "restricted quasi-diagonal" (a quasi-diagonal lying in the kernel of the summation map).
- A subvariety is of Fujiki class C if and only if its image is a curve where all projections to the elliptic curve factors have the same degree.
- Crucially, the authors prove that any Moishezon subvariety in a BG manifold is necessarily projective.
Rigidity of Quasi-Diagonals (Theorem 1.4 & 2.22): The paper proves that for a fixed elliptic curve $E$ and ample line bundle $L$ , the set of quasi-diagonals in $E^n$ is at most countable. This is a significant rigidity result, showing that quasi-diagonals do not form continuous families.
- Two proofs are provided: one purely algebro-geometric (Section 2.5) and one using the coreduction theorem and SKT metrics (Section 5.5).
Existence and Scarcity (Proposition 1.5 & Corollary 1.6):
- While countable sets of quasi-diagonals always exist, "restricted" quasi-diagonals (those summing to zero) are rare. For a general line bundle $L$ , no restricted quasi-diagonals exist.
- Consequently, for a general BG manifold, all projective subvarieties are strictly contained within the fibers of the Lagrangian fibration. They do not project to curves in the base $\mathbb{P}^n$ .
SKT Structure (Theorem 5.16): The authors establish that BG precursors admit a pluriclosed (SKT) metric. This is a non-trivial result for non-Kähler manifolds and serves as the foundation for the analytic proof of the countability of quasi-diagonals.

Significance and Claims
The paper claims to refine the understanding of the algebraic geometry of non-Kähler holomorphically symplectic manifolds. By shifting focus to precursors and introducing the concept of quasi-diagonals, the authors provide a precise criterion for projectivity that was previously lacking.

The significance lies in the demonstration of "generic emptiness" of projective subvarieties in the base of the Lagrangian fibration. While earlier results showed that high-dimensional images are impossible, this work proves that even curve images are generically impossible for projective subvarieties. The paper positions the scarcity of quasi-diagonals as a fundamental geometric constraint on these manifolds, derived from both algebraic and complex-analytic rigidity. The work also highlights the utility of SKT metrics and Barlet spaces in studying the moduli of curves in non-Kähler settings, offering a new perspective on the classification of subvarieties in these spaces.