🔢 mathematics

Higher modularity of elliptic curves over function fields

This paper introduces the concept of $r$ -modularity for elliptic curves over function fields via algebraic correspondences with stacks of $r$ -legged shtukas and proves that nonisotrivial elliptic curves over $\mathbf{F}_q(t)$ with conductor degree 4 are 2-modular, utilizing K3 surface theory and a new characterization of Kummer surface morphisms.

Original authors: Adam Logan, Jared Weinstein

Published 2026-05-06

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Adam Logan, Jared Weinstein

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: A New Kind of "Fingerprint"

Imagine you have a mysterious object, like a unique elliptic curve (a specific type of mathematical shape defined by an equation). In the world of numbers (like the rational numbers we use every day), mathematicians have long known that these objects have a "fingerprint" called modularity. This fingerprint is a special pattern found in a completely different world of shapes called modular curves. If you can match the object to this pattern, you understand its secrets.

This paper asks a bold question: What happens if we look at these shapes not over numbers, but over "function fields"?

Think of a function field as a world where the numbers are actually functions (like $x^2 + 1$ ) rather than static integers. In this world, the usual "fingerprint" (modular curves) isn't enough. The authors propose a concept called "Higher Modularity."

Instead of matching the object to a simple curve, they want to match it to a much more complex, multi-dimensional structure called a Stack of Shtukas.

The Analogy: The "Legs" of a Shtuka

To understand "Higher Modularity," you have to understand Shtukas.

The Standard Case ( $r=1$ ): Imagine a shtuka as a single-legged creature. In the standard world, matching an elliptic curve to a 1-legged shtuka is like matching a key to a single lock. This is the famous "modularity" theorem that proved Fermat's Last Theorem.
The "Higher" Case ( $r=2, 3, \dots$ ): The authors imagine shtukas with $r$ legs.
- A 2-legged shtuka is like a creature with two legs.
- A 3-legged shtuka has three legs.

The paper defines an elliptic curve as " $r$ -modular" if there is a deep, structural connection (an "algebraic correspondence") between the elliptic curve and the world of these $r$ -legged creatures.

The Main Claim: The authors prove that for a specific, very common class of elliptic curves (called extremal rational elliptic fibrations), this connection exists not just for 1 leg, but for 2 legs (and even 3 legs for a specific famous example).

The Journey: How They Proved It

Proving that a 2-legged creature connects to an elliptic curve is incredibly hard. The authors had to build a bridge using some very heavy mathematical machinery. Here is the step-by-step journey they took, simplified:

1. The "Coincidence Map" (The Shortcut)

The authors discovered a surprising shortcut. They found a way to relate a 2-legged shtuka to a 1-legged shtuka through a "coincidence map."

Analogy: Imagine you are trying to find a path through a dense forest (the 2-legged world). Instead of hacking through the trees, you find a hidden tunnel that leads you to a simpler, well-mapped path (the 1-legged world). This tunnel is the "coincidence map." It allows them to translate a problem about complex 2-legged shapes into a problem about simpler 1-legged shapes.

2. The K3 Surface (The Shape-Shifter)

When they used this shortcut, the 2-legged shtuka space turned out to look like a K3 surface.

Analogy: A K3 surface is a complex, 4-dimensional shape (in real terms) that is very rigid and beautiful. Think of it as a "shape-shifter." The authors showed that the complex world of the 2-legged shtuka can be flattened out into this K3 surface.

3. The Kummer Surface (The Target)

On the other side of the bridge, they looked at the elliptic curve they were studying. They realized that if you take two of these curves and multiply them together, you get a shape called a Kummer surface (which is also a type of K3 surface).

The Goal: They needed to prove that the K3 surface coming from the shtuka (the "Left Side") is essentially the same as the Kummer surface coming from the elliptic curves (the "Right Side").

4. The "Picard Lattice" (The DNA Test)

How do you know two complex shapes are the same? You check their "DNA." In math, this DNA is called the Picard Lattice. It's a grid of numbers that describes how the surface is built.

The Breakthrough: The authors proved a powerful new theorem: If two K3 surfaces have the same "DNA" (Picard lattice), then they are connected by a finite map.
They calculated the DNA of their shtuka-derived shape and the DNA of the Kummer surface. They matched perfectly. Therefore, the two shapes are connected. This connection proves the curve is 2-modular.

The 3-Legged Case (Calabi-Yau Threefolds)

The paper doesn't stop at 2 legs. They also tackled the case of 3 legs.

The Challenge: A 3-legged shtuka is even more complex. It doesn't look like a surface anymore; it looks like a Calabi-Yau threefold (a 6-dimensional shape, often used in string theory).
The Solution: They took a specific, famous elliptic curve (the Legendre curve) and built a 3-legged shtuka space. They then built a "Kummer threefold" (the 3D version of the Kummer surface) from three copies of that curve.
The Result: They found a birational map between these two 3D shapes. This means they are the same shape, just viewed from a different angle or with some parts rearranged. This proves the Legendre curve is 3-modular.

Summary of Results

Definition: They defined what it means for an elliptic curve to be " $r$ -modular" (connected to an $r$ -legged shtuka).
Main Theorem: They proved that every non-isotrivial, "tame" extremal rational elliptic curve (a specific, well-behaved class of curves) is 2-modular.
Side Theorem: They proved that if two K3 surfaces have the same "Picard lattice" (DNA), they are connected by a finite map. This was a crucial tool for the main proof.
Bonus Result: They proved that the famous Legendre elliptic curve is 3-modular.

Why Does This Matter? (According to the Paper)

The paper connects this to the Birch and Swinnerton-Dyer (BSD) conjecture, one of the biggest unsolved problems in math.

The BSD conjecture relates the number of solutions to an equation to the behavior of a complex function (the L-function).
The authors explain that if a curve is " $r$ -modular," it helps construct special points (called Heegner-Drinfeld cycles) on the curve.
These points act like "keys" that can unlock the secrets of the curve's rank (how many solutions it has).
Essentially, proving higher modularity gives mathematicians a new, powerful tool to potentially solve the BSD conjecture for function fields.

In short: The authors built a bridge between the chaotic world of multi-legged mathematical creatures (shtukas) and the structured world of elliptic curves. By showing these worlds are connected, they provided a new way to understand the deep secrets of these curves.

Technical Summary: Higher Modularity of Elliptic Curves over Function Fields

Problem Statement
The paper investigates a notion of "higher modularity" for nonisotrivial elliptic curves $E$ defined over a function field $K = \mathbb{F}_q(t)$ . While the classical modularity of elliptic curves over $\mathbb{Q}$ (Shimura-Taniyama) establishes a correspondence between $E$ and a modular curve $X_0(N)$ , and the function field analogue (Drinfeld) establishes a correspondence with Drinfeld modular curves (1-modularity), this work seeks to generalize this to "higher" levels. Specifically, the authors ask whether an elliptic curve $E$ is $r$ -modular for integers $r \ge 1$ .

An elliptic curve $E/K$ is defined as $r$ -modular if there exists an algebraic correspondence (a degree 0 cycle) between the moduli stack of $r$ -legged shtukas for $G = \mathrm{PGL}_2$ (denoted $\mathrm{Sht}^r_G$ ) and the $r$ -fold product $E^r$ . This correspondence must induce a surjective map on the relevant $\ell$ -adic cohomology groups, effectively realizing the motive $h^1(E)^{\otimes r}$ as a quotient of the cohomology of the shtuka stack.

Methodology
The authors employ a geometric strategy that shifts the problem from the direct construction of correspondences on high-dimensional stacks to the study of specific algebraic surfaces and threefolds.

The Coincidence Map: A central tool is the "coincidence map," a rational map $c: \mathrm{Sht}^r_G \dashrightarrow \mathrm{Sht}^1_G$ (relating $r$ -legged shtukas to 1-legged shtukas). This map factors through a moduli space of "coincidences" (modifications of vector bundles without Frobenius). For $r=2$ , this allows the authors to relate the stack $\mathrm{Sht}^2_G$ to a family of elliptic surfaces $Z_2(E)$ , which are constructed as fiber products involving the Drinfeld modular curve (the $r=1$ case) and the coincidence space.
Reduction to K3 Surfaces: The family $Z_2(E) \to U^2$ (where $U$ is the base curve minus the conductor support) is shown to be a family of K3 surfaces. The problem of proving 2-modularity reduces to finding a finite morphism from $Z_2(E)$ to a Kummer surface $K_m(E_1 \times E_2)$ associated with a product of elliptic curves.
Lattice Theory and Isogenies: The authors utilize the theory of K3 surfaces, specifically the relationship between the Picard lattice and the existence of finite morphisms. They prove that if the Picard lattice of a K3 surface $S$ is rationally isometric to that of a Kummer surface $K_m(E_1 \times E_2)$ (where $E_1, E_2$ are non-isogenous), then there exists a finite morphism $S \to K_m(E_1 \times E_2)$ . This relies on the Shioda-Inose structure and lattice embedding theorems.
Explicit Calculations: For specific classes of elliptic fibrations, the authors perform explicit algebraic geometry calculations. They identify elliptic fibrations on the constructed K3 surfaces $Z_2(E)$ that possess specific singular fiber configurations (e.g., two $I^*_0$ fibers), allowing them to construct the required isogenies to Kummer surfaces or Inose surfaces.
Calabi-Yau Threefolds: For the case $r=3$ , the authors study fiber products of elliptic fibrations, which yield Calabi-Yau threefolds. They establish a birational equivalence between the fiber product $Z_3(E)$ and a "Kummer threefold" (a resolution of a quotient of $E^3$ by a finite group), thereby proving 3-modularity for specific cases.

Key Contributions and Results

Definition of Higher Modularity: The paper formalizes the concept of $r$ -modularity via algebraic correspondences between shtuka stacks and powers of elliptic curves, linking it to the Tate conjecture and the global Langlands correspondence.
Theorem 1.2.4 (Main Result): Every nonisotrivial tame extremal rational elliptic fibration $E \to \mathbb{P}^1$ $E \to P^{1}$ over a finite field (with conductor of degree 4) is 2-modular.
- The proof involves classifying these fibrations (semistable and unstable cases) and constructing explicit isogenies between the associated K3 surface $Z_2(E)$ and a Kummer surface $K_m(E_1 \times E_2)$ .
- For semistable cases, the result follows from the uniqueness of isogeny classes for large $q$ and explicit calculations for small $q$ .
- For unstable cases, the authors use the theory of Shioda-Inose structures and Inose surfaces to construct the necessary finite morphisms.
Theorem 1.4.3: The Legendre elliptic curve $y^2 = x(x-1)(x-t)$ is 3-modular. This is proved by establishing a birational map between the fiber product $Z_3(E)$ and the Kummer threefold $K_m(E^3)$ .
Independent Result on K3 Surfaces (Theorem 1.4.1 / 3.3.6): The authors prove that a K3 surface $S$ admits a finite morphism to a Kummer surface $K_m(E_1 \times E_2)$ (with $E_1, E_2$ non-isogenous) if and only if the Picard lattice of $S$ is rationally isometric to the Picard lattice of such a Kummer surface. This provides an effective criterion for the existence of such morphisms, extending previous results by Buskin and Mukai.
Explicit Equations: The paper provides explicit equations for the spaces of shtukas and coincidences for $r=2$ and $r=3$ , and details the construction of the isogenies for various extremal fibrations (e.g., $I^*_2 I_2 I_2$ , $II^* I_1 I_1$ , etc.).

Significance
The paper provides the first concrete examples of elliptic curves over function fields that are $r$ -modular for $r > 1$ . While the analytic modularity (existence of automorphic forms) is known for function fields, the geometric modularity (existence of algebraic correspondences) for $r > 1$ was previously unproven.

The authors frame their work as a step toward verifying a conjecture (Conjecture 1.2.3) that all nonisotrivial elliptic curves are $r$ -modular for all $r$ . Their results demonstrate that the geometric correspondence predicted by the Tate conjecture and the Langlands program exists for a significant class of curves. The work also highlights the utility of "coincidence maps" in relating shtuka spaces of different leg numbers, offering a potential pathway to proving $r$ -modularity for arbitrary $r$ .

The paper is modest in its scope, focusing on extremal rational elliptic fibrations (which have Mordell-Weil rank 0 and conductor degree 4). It does not claim to prove the general conjecture for all elliptic curves or all $r$ , but rather establishes the foundational mechanism and provides the first non-trivial instances of higher modularity. The connection to Heegner-Drinfeld cycles and the Birch and Swinnerton-Dyer conjecture is discussed as a motivation, noting that $r$ -modularity allows the construction of algebraic cycles that span the determinant of the Selmer group.