On algebraically coisotropic submanifolds of holomorphic symplectic manifolds

Basic Information

• Paper ID: 2205.07958
• Title: On algebraically coisotropic submanifolds of holomorphic symplectic manifolds
• Authors: Ekaterina Amerik (Université Paris-Sud & HSE University), Frédéric Campana (Université Lorraine)
• Classification: math.AG (Algebraic Geometry), math.CV (Complex Analysis)
• Publication Date: 2023, published in Épijournal de Géométrie Algébrique special issue (in memory of Claire Voisin)
• Paper Link: https://arxiv.org/abs/2205.07958

Abstract

This paper investigates algebraically coisotropic submanifolds $X$ in holomorphic symplectic projective manifolds $M$. Building on the authors' previous results in the hypersurface case, the paper poses a central question: when $X$ is not unirational, does there exist a product decomposition $(Z\times Y, N\times Y)$ in the sense of finite étale covers, where $N, Y$ are holomorphic symplectic manifolds and $Z\subset N$ is a Lagrangian submanifold? The authors prove this conclusion holds when $M$ is an abelian variety, and provide partial answers when the canonical bundle $K_X$ is semiample. In particular, when $K_X$ is nef and big, $X$ is Lagrangian in $M$. The paper also notes that, in contrast to the irreducible hyperkähler case, sufficiently general abelian varieties do not admit Lagrangian submanifolds.

Research Background and Motivation

Core Problem

This paper studies the structural problem of algebraically coisotropic submanifolds in holomorphic symplectic manifolds. Coisotropic submanifolds are an important concept in symplectic geometry: for a submanifold $X$ in a holomorphic symplectic manifold $(M,\sigma)$, if at each point $x\in X$, the corank of the restriction $\sigma|_X$ equals the codimension of $X$, then $X$ is called coisotropic.

Research Motivation

1. Generalizing known results: The authors characterized algebraically coisotropic divisors in AC17; this paper attempts to extend these results to higher codimension cases
2. Structural classification problem: Understanding whether non-unirational algebraically coisotropic submanifolds can always be decomposed as products of Lagrangian submanifolds with holomorphic symplectic manifolds
3. Geometric intuition: Lagrangian submanifolds are the higher-dimensional analogue of curves in holomorphic symplectic surfaces; studying their existence and structure is of significant importance

Limitations of Existing Research

• Hwang-Viehweg HV10 proved that smooth algebraically coisotropic hypersurfaces of general type are curves in holomorphic symplectic surfaces
• Structural theory for higher codimension cases has received less attention
• A unified classification framework is lacking

Core Contributions

1. Complete classification in the abelian variety case: Proves that Question 1.4 has an affirmative answer when $M$ is an abelian variety (Theorem 1.11)
2. Structure theorem under canonical bundle conditions: When $K_X$ is semiample, the characteristic fibration is isotrivial (Theorems 1.7-1.8)
3. Lagrangian criterion: When $K_X$ is nef and big, $X$ must be Lagrangian (Corollary 1.9)
4. Existence results: Proves that Hodge-general abelian varieties do not admit Lagrangian submanifolds (Corollary 5.5)
5. Concrete example construction: Provides examples of Lagrangian submanifolds in the non-projective case

Detailed Methodology

Basic Definitions and Concepts

Definition 1.1 (Coisotropic submanifold): Let $M$ be a complex projective manifold equipped with a holomorphic symplectic form $\sigma$, and $X$ an irreducible complex submanifold of $M$. If at each point $x\in X$, the corank of $\sigma|_X$ equals $\text{codim}(X)$, then $X$ is called coisotropic.

Algebraically coisotropic: If the characteristic foliation on a coisotropic submanifold $X$ is algebraic (i.e., the leaves are algebraic submanifolds), then $X$ is called algebraically coisotropic.

Core Technical Methods

1. Characteristic Fibration Analysis

For an algebraically coisotropic submanifold $X$, its characteristic foliation defines a fibration $f: X \to B$, where the fibers are the leaves of the foliation. Key lemma:

Lemma 2.1 (Sawon): There exists a holomorphic symplectic form $\eta$ on $B$ such that $f^*\eta = \sigma|_X$.

2. Geometry of the Base

Proposition 2.2: The base $B$ of the characteristic fibration has the following properties:

• No multiple fibers in codimension 1
• $B$ has only canonical singularities
• $\kappa(B) = 0$ (Kodaira dimension equals zero)

3. Isotriviality Criterion

Using results on special bases from AC18:

Theorem 1.7: If $K_X$ is semiample, then the characteristic fibration $f: X \to B$ is isotrivial, and $\kappa(X) = \kappa(F)$.

Analysis in the Abelian Variety Case

For algebraically coisotropic submanifolds $X$ in an abelian variety $M$, using Ueno's classification theorem:

1. Structural decomposition: $X = p^{-1}(Z)$, where $p: M \to L = M/A$, $A$ is an abelian subvariety, $Z \subset L$ is a subvariety of general type
2. Product decomposition: Via Poincaré complete reducibility, obtain the decomposition $M = D \times C \times N \times P$
3. Orthogonality analysis: Analyze orthogonality relations between factors using the coisotropic condition
4. Symplectic form structure: Determine the restriction properties of $\sigma$ on each factor

Main Theorems and Proof Strategies

Theorem 1.11 (Abelian Variety Case)

Statement: Let $M$ be an abelian variety and $X \subset M$ an algebraically coisotropic submanifold. Then after a finite étale cover, there exist sub-tori $D,N,C,P$ such that:

1. $M = D \times C \times N \times P$
2. $X = D \times C \times Z$, where $Z \subset N$
3. $Z$ is Lagrangian in $N$

Proof outline:

1. Use Ueno's theorem to obtain the basic structure of $X$
2. Achieve product decomposition via Poincaré complete reducibility
3. Analyze the Künneth decomposition of the symplectic form: $\sigma = \sigma_D \oplus \sigma_N \oplus \sigma_P \oplus s$
4. Use the coisotropic condition to determine orthogonality relations between factors
5. Prove that $\sigma_N$ is symplectic and $Z$ is $\sigma_N$-coisotropic
6. Prove $Z$ is Lagrangian through dimension calculations

Theorems 1.7-1.8 (Isotriviality)

Proof strategy:

1. Use Proposition 2.2 to determine the geometric properties of the base $B$
2. Apply the specialness results from AC18 or Taj23
3. Analyze the relative canonical bundle to obtain $\kappa(X) = \kappa(F)$

Important Applications and Corollaries

Corollary 1.9 (Lagrangian Criterion)

When $K_X$ is nef and big, $X$ must be Lagrangian. This generalizes the Hwang-Viehweg theorem to higher dimensions.

Corollary 1.10 (Simple Abelian Varieties)

Algebraically coisotropic submanifolds in simple abelian varieties must be Lagrangian.

Corollary 5.5 (Non-existence)

Hodge-general abelian varieties do not admit Lagrangian submanifolds (when dimension $>2$).

Technical Innovations

1. Künneth decomposition analysis of symplectic forms: Refined analysis of the structure of symplectic forms on abelian varieties
2. Geometric control of characteristic fibrations: Use specialness properties of the base to control the fibration structure
3. Application of Hodge theory: Use properties of Hodge groups to prove non-existence results
4. Dimension calculation techniques: Determine geometric structures through dimensional constraints from the coisotropic condition

Comparison with Related Work

Relationship to Existing Results

• AC17: This paper generalizes the divisor case results to higher codimension
• HV10: Generalizes the Hwang-Viehweg criterion for general type
• Saw09: Uses and develops Sawon's foliation theory
• Voi92, Bea11: Provides contrast in the irreducible hyperkähler case

Points of Innovation

1. First complete classification of algebraically coisotropic submanifolds in abelian varieties
2. Establishes deep connections between canonical bundle properties and geometric structure
3. Reveals essential differences between abelian varieties and hyperkähler manifolds regarding Lagrangian submanifold existence

Concrete Examples and Applications

Example 1 (Schoen Construction)

Lagrangian surfaces exist in certain simple abelian fourfolds, given by Schoen's construction.

Example 2 (Non-projective Case)

Proposition 6.1: There exist a 2-dimensional complex torus $T$ and an automorphism $g$ such that for any symplectic form $\sigma$, we have $g^*\sigma = \lambda\sigma$, where $\lambda$ is not a root of unity.

This provides an example of a Lagrangian surface on $T \times T$.

Open Problems

Question 1.4 (Main Conjecture)

For non-unirational algebraically coisotropic submanifolds in general holomorphic symplectic manifolds, can one always realize a product decomposition?

Related Problems

1. Problem 5.1: Find examples of Lagrangian submanifolds in simple abelian varieties of dimension $\geq 6$
2. Question 5.8: What restrictions exist on the fundamental group of middle-dimensional submanifolds in simple abelian varieties?

In-Depth Evaluation

Strengths

1. Theoretical depth: Establishes important foundations for the theory of algebraically coisotropic submanifolds
2. Technical innovation: Cleverly combines tools from algebraic geometry, complex geometry, and symplectic geometry
3. Completeness: Provides complete classification in the abelian variety case
4. Comparative analysis: Reveals essential differences between different types of holomorphic symplectic manifolds

Technical Contributions

1. Develops geometric theory of characteristic fibrations
2. Provides refined analysis of algebraic properties of symplectic structures on abelian varieties
3. Establishes connections between Hodge theory and geometric existence

Limitations

1. Main results are limited to the abelian variety case; the general case requires further development
2. Some technical conditions (such as semiampleness) may be too restrictive
3. Relatively few constructive examples

Impact Assessment

This paper establishes important theoretical foundations for the study of algebraically coisotropic submanifolds, particularly:

1. Provides affirmative answers to Question 1.4 in special cases
2. Establishes standard technical methods in the field
3. Reveals core problems requiring further research

Applicable Scenarios

1. Submanifold classification problems in algebraic geometry
2. Lagrangian submanifold theory in symplectic geometry
3. Geometric research on abelian varieties and hyperkähler manifolds

References

The paper cites important literature in the field, including classical works by Voisin, Sawon, Hwang-Viehweg, and others, as well as the authors' own prior results in AC17, AC18. Particularly noteworthy is the application of recent results by Taji Taj23, reflecting the cutting-edge nature of this research.

This paper represents an important theoretical contribution to algebraic geometry, providing profound insights into understanding the geometric structure of submanifolds in holomorphic symplectic manifolds, particularly establishing a complete classification theory in the abelian variety case.