On autoduality of Drinfeld modules and Drinfeld modular forms

This paper establishes that rank-two Drinfeld modules with a specific $\Gamma_1^\Delta(\mathfrak{n})$-structure are isomorphic to their Taguchi duals, a result that enables the derivation of a dual Kodaira–Spencer isomorphism for the Hodge bundle on the corresponding Drinfeld modular curve.

The Gist

Imagine you are an architect designing a city of numbers. In this city, there are two famous types of buildings: Elliptic Curves (the classic, well-understood skyscrapers of traditional math) and Drinfeld Modules (the exotic, futuristic structures built over a different kind of ground, specifically using polynomials instead of regular numbers).

For a long time, mathematicians knew a very special rule about the classic skyscrapers (Elliptic Curves): They are their own mirrors. If you look at an elliptic curve, you can find a perfect, natural way to flip it inside out and get the exact same shape back. This is called autoduality. It's like having a building that, when you look in a mirror, shows you the exact same building, not a reflection.

However, the exotic Drinfeld Modules were stubborn. They didn't have this mirror property. If you tried to flip them, you got a "dual" building that looked similar but wasn't quite the same. You couldn't just say, "This is the same as that." This missing mirror made it very hard to do certain calculations, specifically a type of blueprint conversion known as the Kodaira–Spencer isomorphism.

Think of the Kodaira–Spencer isomorphism as a translator. It translates the "shape" of the building (the Hodge bundle) into the "directions" of the city streets (differential forms).

• For the classic buildings: The translator works perfectly. One shape translates directly to one street direction.
• For the exotic buildings: Because they lacked the mirror, the translator was awkward. It had to mix the shape with its "dual" (the slightly different reflection) to make the translation work. This made the math messy and less elegant.

The Breakthrough: Finding the Mirror

In this paper, Shin Hattori solves the mystery. He proves that under specific conditions (using a special "level structure" called $\Gamma^\Delta_1(n)$), these exotic Drinfeld Modules do have a mirror. They are actually autodual!

The Analogy of the Magic Key:
Imagine the Drinfeld Module is a locked box. For years, we thought the key to opening it (the isomorphism to its dual) didn't exist. Hattori found a "Magic Key" hidden in the city's history. This key is a special mathematical function called Gekeler's $h$-function.

1. The Discovery: Hattori realized that if you use this $h$-function, it acts like a magical lens. When you look at the Drinfeld Module through this lens, the "dual" version suddenly snaps into place, looking exactly like the original.
2. The Proof: He showed that this magic key works everywhere in the city, even at the edges (the "cusps" or boundaries of the mathematical space), by using a technique called "gluing" (like patching together a quilt).

The Result: A Cleaner Blueprint

Once the mirror is found, the translator (the Kodaira–Spencer isomorphism) becomes much simpler and more beautiful.

• Before: The translator had to say, "To get the street directions, I need to mix the building's shape with its weird reflection."
• After: The translator can now say, "To get the street directions, I just need to square the building's shape."

This is a huge deal because it makes the theory of Drinfeld Modular Forms (the study of these exotic buildings) look much more like the well-understood theory of Elliptic Modular Forms. It unifies the two worlds.

Why Should You Care?

Even if you aren't a mathematician, this is a story about symmetry and elegance.

• Symmetry: Nature and math often prefer things that are balanced. Hattori showed that these exotic structures are more balanced than we thought.
• Simplicity: He took a complicated, messy rule and replaced it with a clean, simple one.
• Connection: He bridged a gap between two different areas of math, showing that the rules governing the "exotic" are actually very similar to the "classic" ones, once you find the right perspective.

In short, Hattori found the missing mirror for a strange new world of numbers, allowing mathematicians to see that world with much greater clarity and beauty.

Technical Summary

Here is a detailed technical summary of the paper "On Autoduality of Drinfeld Modules and Drinfeld Modular Forms" by Shin Hattori.

1. Problem Statement and Motivation

Context:
Drinfeld modular forms are function field analogues of classical elliptic modular forms. In the classical theory of elliptic curves, every curve $E$ is isomorphic to its dual $E^\vee$ (autoduality). This property is fundamental to the geometric theory, particularly in establishing the Kodaira–Spencer isomorphism, which relates the square of the Hodge bundle ($\omega_E^{\otimes 2}$) to the sheaf of differentials ($\Omega^1$).

The Gap:
In the theory of Drinfeld modules (rank 2 over $F_q[t]$), a dual object $E^D$ exists (defined by Taguchi), but a general isomorphism $E \cong E^D$ does not exist. Consequently, the standard Kodaira–Spencer isomorphism in the Drinfeld setting takes the form:
$$\omega_E \otimes \omega_{E^D} \xrightarrow{\sim} \Omega^1$$
rather than the more desirable and symmetric form found in the elliptic case:
$$\omega_E^{\otimes 2} \xrightarrow{\sim} \Omega^1$$
This lack of autoduality complicates the geometric study of Drinfeld modular forms, particularly when analyzing the Hodge filtration and the behavior of modular forms near cusps.

Objective:
The paper aims to construct a specific level structure (denoted $\Gamma^\Delta_1(n)$) under which rank 2 Drinfeld modules do become autodual. Using this, the author constructs a "dual" Kodaira–Spencer isomorphism of the form $\bar{\omega}^{\otimes 2} \cong \Omega^1(2\text{Cusps})$ on the Drinfeld modular curve, mirroring the classical elliptic case.

---

2. Methodology and Key Tools

The author employs a combination of arithmetic geometry, rigid analytic geometry, and the theory of $\tau$-sheaves.

A. The $\Gamma^\Delta_1(n)$-Structure

The paper introduces a refined level structure. Let $n \in F_q[t]$ be a monic polynomial with a prime factor of degree coprime to $q-1$. Let $\Delta \subset (A/nA)^\times$ be a subgroup such that the map $\Delta \to (A/nA)^\times / F_q^\times$ is bijective.
A $\Gamma^\Delta_1(n)$-structure on a Drinfeld module $E$ is a pair $(\lambda, \mu)$, where $\lambda$ is a standard $\Gamma_1(n)$-structure and $\mu$ is an additional datum derived from $\Delta$. This structure is crucial for defining the specific modular curve $Y^\Delta_1(n)$ where autoduality holds.

B. Gekeler's $h$-function and the $x$-expansion Principle

The core mechanism for proving autoduality relies on Gekeler's $h$-function, a Drinfeld modular form of weight $q+1$.

• Key Relation: The $h$-function satisfies $h^{q-1} = -g_2$, where $g_2$ is the discriminant function related to the coefficient $\alpha_2$ in the Drinfeld module action $\Phi_t = \theta + \alpha_1 \tau + \alpha_2 \tau^2$.
• Construction: A section $H$ of the line bundle $L_E^{\otimes (-1-q)}$ satisfying $H^{q-1} = -\alpha_2$ induces an isomorphism $E \to E^D$.
• Descent: Using the $x$-expansion principle (which allows descending modular forms from the analytic upper half-plane to the algebraic modular curve), the author shows that the $h$-function defines a global section on the modular curve $X^\Delta_1(n)$, thereby providing the global autoduality isomorphism.

C. Extension to Compactification (Beauville–Laszlo Gluing)

To handle the behavior at the cusps (where the modular curve is compactified), the author uses Tate–Drinfeld modules.

• The paper analyzes the local behavior of the Hodge bundle and the de Rham sheaf near the cusps using the Tate–Drinfeld module $TD(\Lambda)$.
• It employs Beauville–Laszlo gluing (Lemma 4.3) to extend sheaves from the open modular curve $Y^\Delta_1(n)$ to its compactification $X^\Delta_1(n)$. This involves constructing "modifications" of sheaves at the cusps that are compatible with the $F_q^\times$-action.

D. Correction of Previous Literature

The author identifies and fixes gaps in previous works (specifically [Hat1] and [Hat2]) regarding the base change compatibility of differential sheaves. This is resolved by utilizing the $F$-finiteness of the base rings (specifically that the Frobenius map is finite), ensuring that the formation of differentials behaves correctly under base change to the cusps.

---

3. Key Contributions and Results

The paper establishes four main theorems (summarized in Theorem 1.1):

1. Autoduality of Drinfeld Modules (Theorem 5.5)

For any scheme $S$ over $A[1/n]$, any rank 2 Drinfeld module $E$ over $S$ admitting a $\Gamma^\Delta_1(n)$-structure is isomorphic to its Taguchi dual $E^D$.

• Mechanism: The isomorphism is explicitly constructed using the pullback of the $h$-function.
• Significance: This resolves the fundamental asymmetry in the Drinfeld case for this specific level structure.

2. Extension of the Hodge Filtration (Theorem 5.16)

The Hodge filtration sequence on the open curve:
$$0 \to \omega_E \to H^1_{dR}(E) \to \omega_{E^D} \to 0$$
extends to an exact sequence of finite locally free sheaves on the compactified curve $X^\Delta_1(n)$:
$$0 \to \bar{\omega} \to \bar{H}^1_{dR} \to \bar{\omega}^\vee \to 0$$
Crucially, because of the autoduality isomorphism $E \cong E^D$, the quotient term becomes $\bar{\omega}^\vee \cong \bar{\omega}^\vee$ (identifying the dual of the dual with the original via the isomorphism).

3. Dual Kodaira–Spencer Isomorphism (Theorem 5.17)

The paper constructs an isomorphism:
$$\overline{KS}^\circ: (\bar{\omega}^\Delta_{un})^{\otimes 2} \xrightarrow{\sim} \Omega^1_{X^\Delta_1(n)/R}(2\text{Cusps})$$

• Contrast: Unlike the previous Gekeler isomorphism involving $\omega_E \otimes \omega_{E^D}$, this isomorphism involves the square of the Hodge bundle, exactly analogous to the classical elliptic modular curve case.
• Cusps: The isomorphism holds with a twist by the divisor of cusps ($2\text{Cusps}$), reflecting the behavior of the differential forms at the boundary.

4. Arithmetic de Rham Pairing (Theorem 5.19)

The author constructs a perfect, alternating, $O_X$-linear pairing on the extended de Rham sheaf:
$$\langle \cdot, \cdot \rangle^\Delta_{un}: \bar{H}^1_{dR, un} \otimes \bar{H}^1_{dR, un} \to O_{X^\Delta_1(n)}$$
This pairing induces the canonical pairing between $\bar{\omega}$ and its dual, confirming the symplectic structure of the de Rham cohomology in this setting.

---

4. Significance and Impact

1. Geometric Unification: The paper successfully bridges the gap between the geometric theory of elliptic modular forms and Drinfeld modular forms. By establishing autoduality under a specific level structure, it allows the transfer of classical intuition (like the $\omega^{\otimes 2}$ Kodaira–Spencer isomorphism) to the function field setting.
2. Refined Level Structures: The introduction and rigorous treatment of the $\Gamma^\Delta_1(n)$-structure provide a new tool for studying Drinfeld modular forms, particularly for analyzing their behavior at cusps and their relation to the $h$-function.
3. Technical Corrections: The paper clarifies technical issues in the literature regarding the compatibility of differential sheaves with base change and the $F$-finiteness of the base rings, ensuring the robustness of future geometric arguments in this field.
4. Applications: The resulting isomorphism $\bar{\omega}^{\otimes 2} \cong \Omega^1(2\text{Cusps})$ is essential for the study of the geometry of the modular curve, the construction of $p$-adic modular forms, and the understanding of the arithmetic properties of Drinfeld modular forms (e.g., in the context of the Langlands program over function fields).

In summary, Shin Hattori's work provides a foundational geometric framework where Drinfeld modules behave "classically" (autodual) under specific conditions, enabling a more powerful and symmetric geometric analysis of Drinfeld modular forms.