🔢 mathematics

Finiteness of pointed families of symplectic varieties: a geometric Shafarevich conjecture

This paper establishes a geometric Shafarevich conjecture for pointed families of primitive symplectic varieties by proving the finiteness of isomorphism classes of generic fibers and, under semi-ampleness assumptions, of projective families, while demonstrating that these results are optimal by constructing counterexamples when projectivity is relaxed.

Original authors: Lie Fu, Zhiyuan Li, Teppei Takamatsu, Haitao Zou

Published 2026-04-14

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Lie Fu, Zhiyuan Li, Teppei Takamatsu, Haitao Zou

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

The Big Picture: The "Traveling Shape" Problem

Imagine you have a very special, complex 3D shape (let's call it a Symplectic Variety). Think of it like a magical, multi-dimensional crystal that has a unique internal structure and a specific "fingerprint" (its geometry).

Now, imagine you are building a family of these crystals. You start with one specific crystal at a starting point (let's call it Point 0 on a road). As you travel down the road (the curve $B$ ), the crystal changes shape slightly, morphing into new forms. However, there are strict rules:

The crystal must remain "locally trivial," meaning if you zoom in on any tiny spot, it looks exactly the same as it did before; the changes are smooth and global, not chaotic.
When you arrive back at Point 0, the crystal must be identical to the one you started with.

The Question: If you follow these rules, how many different types of crystals can you encounter in the middle of the road (the "generic fibers") before you get back to the start?

The Answer: The authors prove that there are only a finite number of possibilities. You cannot create an infinite variety of unique shapes just by traveling down this road, provided you start and end with the same shape.

The Key Concepts (Translated)

1. The "Unpolarized" Challenge

In the past, mathematicians studied these shapes only when they were "polarized."

Analogy: Imagine the crystal is a house. "Polarized" means the house has a specific, pre-approved blueprint and a fixed amount of land (a bounded degree). This makes it easy to count how many houses exist because they all fit into a standard catalog.
The Problem: This paper looks at "unpolarized" families. These are houses where the owner can add rooms, change the roof, or expand the land without limit, as long as the foundation (the starting point) stays the same.
The Fear: Without a size limit, you might think you could build an infinite number of different houses. The authors prove that even without a size limit, the universe of possibilities is still finite.

2. The "Pointed" Constraint

The problem is "pointed" because we fix the starting point (Point 0).

Analogy: It's like a "Choose Your Own Adventure" book where you are forced to start at Page 1 and must return to Page 1 at the end. The question is: How many different plot twists can happen in the middle chapters?

3. The "Isotrivial" Trap (The Counter-Example)

The paper mentions a tricky exception. If the family is "isotrivial," it means the shape doesn't actually change; it just looks different because of how you are viewing it (like rotating a cube).

The Twist: The authors show that if you allow families that don't change shape (isotrivial) but are glued together in weird ways, you can create an infinite number of non-isomorphic families.
The Metaphor: Imagine you have a red ball. You can paint it, wrap it in paper, or put it in a box. If you just rotate it, it's the same ball. But if you glue a red ball to a blue ball in a specific way, then unglue it, you might create a "new" family that looks the same everywhere except at the start and end. The paper proves that if you forbid these "fake" changes, the number of real changes is finite.

How They Solved It: The Detective's Toolkit

The authors used three main "tools" to prove this finiteness:

Tool 1: The "Kuga-Satake" Translator (The Universal Adapter)

The Problem: Symplectic varieties are too complex to count directly.
The Solution: They used a mathematical "translator" called the Kuga-Satake construction.
Analogy: Imagine you have a secret language (Symplectic Varieties) that is hard to count. You have a universal translator that converts every word in that language into a specific type of Abelian Variety (which is like a very structured, well-behaved torus or donut shape).
The Magic: The authors proved that even though the original shapes are wild, their "donut translations" are limited. There are only finitely many ways to translate the starting shape into a donut. Since the translation is unique, the original shapes must also be finite.

Tool 2: The "Cone" Map (The Traffic Controller)

The Problem: Even if you know the shapes are finite, you need to prove that the families (the journey) are finite.
The Solution: They used the Cone Conjecture.
Analogy: Imagine the possible shapes of the crystal are arranged in a giant, multi-dimensional cone. The "automorphisms" (ways to rotate or flip the crystal) act like a traffic controller. The conjecture says that even though the cone is huge, the traffic controller can divide it into a finite number of "zones."
The Result: Because the zones are finite, there are only finitely many ways the crystal can morph and return to its original state.

Tool 3: The "Matsusaka-Mumford" Glue

The Problem: How do you ensure that a shape that looks the same at the start and end is actually the same family?
The Solution: They used a theorem that acts like super-glue.
Analogy: If two families of crystals look identical at the start and end, and they follow the same smooth path, this theorem proves they are actually the same family all the way through. It prevents "ghost" families that look the same but are secretly different.

Why This Matters

It's a "Geometric Shafarevich Conjecture": In number theory (arithmetic), a famous theorem says there are only finitely many curves of a certain type over a number field. This paper proves the geometric version of that theorem. It says: "In the world of shapes and geometry, if you fix the start and end, you can't wander off into infinity."
It Handles "Rough" Shapes: Previous results only worked for perfectly smooth crystals. This paper works even if the crystals have mild singularities (tiny cracks or sharp points), making the result much more powerful.
It Sets a Limit: It tells us that the universe of these geometric objects is not chaotic. Even without strict size limits, nature imposes a hard cap on how many variations exist.

Summary in One Sentence

The paper proves that if you start with a specific complex geometric shape, travel along a path where the shape changes smoothly, and return to the exact same shape, you will only encounter a finite number of unique intermediate shapes, provided you don't count "fake" variations that are just the same shape viewed differently.

1. Problem Statement

The paper addresses the Pointed Shafarevich Problem for families of primitive symplectic varieties (singular generalizations of hyper-Kähler manifolds).

Context: The classical Shafarevich conjecture (proven by Faltings) states that there are finitely many abelian varieties of a fixed dimension over a number field with good reduction outside a finite set of places. Geometric analogues (over function fields of curves) have been established for polarized varieties.
The Challenge: This paper investigates the unpolarized case. Specifically, given a pointed smooth curve $(B, 0)$ and a fixed primitive symplectic variety $X$ (the fiber over $0$), is the set of isomorphism classes of locally trivial families $\pi: \mathcal{X} \to B$ with fiber $\mathcal{X}_0 \cong X$ finite?
Obstacles:
1. Unboundedness: Without a fixed polarization degree, the families do not fit into a single moduli stack of finite type.
2. Non-separatedness: The moduli problem is highly non-separated; birational models can lead to infinitely many non-isomorphic families that are isomorphic over the punctured curve $B \setminus \{0\}$ .

2. Methodology and Strategy

The authors employ a multi-faceted approach combining Hodge theory, birational geometry, and moduli theory.

A. Weak Polarization and Period Maps

Since the families are not assumed to be projective, the authors introduce the concept of a weak polarization (Definition 4.8). This is a global section of the variation of Hodge structure that behaves like a polarization on a very general fiber.

They construct a period map from the base curve $B$ to a moduli stack of polarized primitive symplectic varieties ( $\mathcal{M}_{2n,d}$ ).
They prove that this moduli stack is a separated Deligne–Mumford stack of finite type (Theorem 4.6), relying on a new Matsusaka–Mumford theorem for locally trivial families of singular symplectic varieties (Proposition 4.2). This theorem ensures that birational maps between generic fibers extend to isomorphisms outside closed subsets, and with compatible polarizations, they extend globally.

B. Uniform Kuga–Satake Construction

To handle the lack of a uniform polarization degree, the authors adapt Zarhin's trick using the Uniform Kuga–Satake construction (Theorem 5.2).

They associate a family of primitive symplectic varieties with a family of abelian varieties.
Crucially, they prove that for a fixed base point $X$ , there exists a finite étale cover $\tilde{B} \to B$ and a uniform Kuga–Satake map independent of the specific family, provided the fiber at $0$ is fixed.
This reduces the problem to the finiteness of families of abelian varieties (which is known) and the finiteness of lattice embeddings.

C. Cone Conjectures and Birational Finiteness

To distinguish between isomorphism classes of families (rather than just birational classes), the authors utilize the Kawamata–Morrison Cone Conjecture.

They prove the cone conjecture for Q-factorial terminal primitive symplectic varieties over non-algebraically closed fields (Theorem 3.5).
This implies that the action of the automorphism group on the nef cone admits a rational polyhedral fundamental domain.
Combined with the semi-ampleness of isotropic nef divisors (the SYZ conjecture), this ensures that there are only finitely many small Q-factorial modifications (birational models) over the base curve.

3. Key Contributions and Results

Main Theorems

Finiteness of Generic Fibers (Theorem 1.1):
- Let $(B, 0)$ be a pointed curve and $X$ a Q-factorial terminal primitive symplectic variety.
- If $b_2(X) \neq 4$ , there are only finitely many isomorphism classes of generic fibers for families in the pointed Shafarevich set.
- If $b_2(X) = 4$ , the same finiteness holds for non-isotrivial families.
- Note: The case $b_2(X)=4$ is exceptional because isotrivial families can be infinite, but non-isotrivial ones are finite.
Finiteness of Projective Families (Theorem 1.2):
- If we restrict to projective families and assume the semi-ampleness of isotropic nef divisors (which holds for known deformation types), the set of isomorphism classes of such families is finite.
- This result is unconditional for smooth hyper-Kähler varieties of known deformation types: $K3^{[n]}$ , $Kum_n$ , $OG6$, and $OG10$ (Corollary 1.3).

Foundational Technical Results

Simultaneous Resolution: They establish the existence of simultaneous resolutions of singularities for locally trivial families of primitive symplectic varieties (Proposition 2.11), allowing the transfer of results from the smooth case to the singular case.
Matsusaka–Mumford for Singular Varieties: They extend the Matsusaka–Mumford theorem to locally trivial families of singular symplectic varieties, proving that isomorphisms of generic fibers extend to global isomorphisms if polarizations match (Corollary 4.3).
Cone Conjecture over Arbitrary Fields: They prove the Kawamata–Morrison cone conjecture for Q-factorial terminal primitive symplectic varieties over non-algebraically closed fields (Theorem 3.5), a necessary step for arithmetic applications.

4. Counter-Examples and Optimality

The authors demonstrate that their results are optimal by constructing counter-examples when assumptions are relaxed.

Proposition 6.2: They construct infinitely many pairwise non-isomorphic families of smooth hyper-Kähler varieties over a pointed curve $(B, 0)$ $(B, 0)$ such that:
1. They are all isomorphic over the punctured curve $B \setminus \{0\}$ .
2. They have isomorphic fibers over the base point $0$.
3. They are not projective (or lack a uniform polarization).
This highlights that the projectivity assumption (or the semi-ampleness condition) in Theorem 1.2 is essential; without it, the moduli space is not separated, and infinite families exist.

5. Significance

Geometric Shafarevich Conjecture: This work provides a definitive answer to the geometric unpolarized Shafarevich conjecture for symplectic varieties, extending Faltings' arithmetic results to the geometric setting for a broad class of varieties.
Singular Varieties: It successfully generalizes results from smooth hyper-Kähler manifolds to primitive symplectic varieties with mild singularities (Q-factorial terminal), which are crucial in the minimal model program and moduli theory (e.g., moduli of sheaves on K3 surfaces).
Moduli Theory: The construction of the separated moduli stack for polarized primitive symplectic varieties and the proof of the cone conjecture in the relative setting provide new tools for studying the birational geometry of these varieties.
Uniformity: The introduction of the "Uniform Kuga–Satake" map offers a robust method to bound families of symplectic varieties without relying on a pre-existing uniform polarization, a significant technical innovation.

In summary, the paper establishes that while the moduli of unpolarized symplectic families is generally non-separated and infinite, imposing a fixed fiber at a point and restricting to projective families (or non-isotrivial ones in specific dimensions) restores finiteness, confirming the geometric hyperbolicity of these moduli spaces.