Theta functions for singular curves

Imagine you are trying to understand the "music" of a shape. In the world of mathematics, specifically geometry, a smooth, perfect shape (like a sphere or a donut) has a very well-understood song. Mathematicians have a special tool called a Theta Function that acts like a universal sheet music for these smooth shapes. It helps them write down every possible note (function) the shape can play.

However, what happens when the shape isn't perfect? What if it has a kink, a knot, or a sharp point? These are called "singular curves." The old sheet music breaks down because the shape is no longer smooth.

This paper by Indranil Biswas and Jacques Hurtubise is about writing a new piece of sheet music that works even when the shape is broken or knotted.

Here is the breakdown of their work using simple analogies:

1. The Problem: The Broken String

Think of a smooth curve as a perfect violin string. You can pluck it anywhere, and it sings a clear, predictable note. Mathematicians have a map (called the Jacobian) that tells them exactly where every note lives.

Now, imagine that string gets knotted or snapped. It's still the same string, but it's now "singular."

The Desingularization: To fix the string, you imagine "untying" the knot. You pull the string apart at the knot so it becomes smooth again. In math, this is called the desingularization ( $\tilde{X}$ ).
The Issue: When you untie the knot, you have two loose ends where the knot used to be. To get back to the original knotted string, you have to glue those two ends back together. But there are many different ways to glue them (you could twist them, stretch them, or just stick them flat).

The authors realized that the old "sheet music" (Theta function) only knows how to play the smooth, unknotted version. It doesn't know how to handle the specific way the ends are glued back together.

2. The Solution: A Universal Glue

The authors built a Generalized Theta Function. Think of this as a "Universal Glue" or a "Master Key."

The Old Way: On a smooth shape, if you slide your sheet music around (translate it), you can generate every possible song the shape can sing.
The New Way: The authors created a new sheet music that lives on a "compactified" version of the Jacobian.
- Analogy: Imagine the old map was a flat sheet of paper. The new map is that same paper, but with extra "floors" added to it (like a skyscraper) to account for all the different ways the knot can be tied.
- This new Theta function is a section of a line bundle. In plain English, it's a specific pattern drawn on this new, taller map.

3. How It Works: The "Universal Section"

The magic of this new function is that it acts as a Universal Section.

The Metaphor: Imagine you have a master stamp. If you press this stamp onto a piece of paper, it leaves a specific mark. If you move the stamp to a different spot and press it again, it leaves a slightly different mark.
The Result: By moving (translating) this new Theta function around the "taller map" (the Generalized Jacobian), the authors can generate every possible way to glue the ends of the knot back together.
When they pull this pattern back down to the actual knotted curve, it gives them a "universal section." This means they can now write down the "songs" (functions) for the knotted curve just as easily as they did for the smooth one.

4. The "Riemann Constant" and The Knot

In the smooth world, there is a famous rule (Riemann's Theorem) that says: "If you find the places where the music stops (the zeroes of the Theta function), you can figure out exactly where you are on the map."

The authors proved that this rule still works for knotted curves, but it's more complex.

The Knot's Memory: Because the knot has "loose ends" (the points where the curve was singular), the new Theta function has to remember how those ends were glued.
The Calculation: They showed that if you add up the locations where the new music stops, you get a formula that tells you exactly how the knot is tied. It's like looking at the silence in a song to figure out how the instrument was tuned.

5. Why It Matters (According to the Paper)

The paper mentions that these functions are useful for integrable systems (complex physics equations that describe waves and flows).

Solitons: Sometimes, a smooth wave breaks down into a sharp, solitary wave (a soliton). Mathematically, this looks like the smooth curve turning into a knotted one.
The Connection: The authors' new Theta function allows mathematicians to describe these "broken" or "knotted" waves using the same elegant language they use for smooth waves. It bridges the gap between the perfect world and the messy, singular world.

Summary

The Goal: Create a mathematical tool (Theta function) that works for shapes with knots and sharp points.
The Method: They built a "taller" version of the mathematical map (Generalized Jacobian) that accounts for all the ways a knot can be tied.
The Result: They found a "Universal Section" (a master pattern) that, when moved around, generates all possible solutions for these knotted shapes.
The Takeaway: Just as a universal translator can speak every language, this new Theta function can "speak" the geometry of both smooth and broken curves, allowing mathematicians to solve problems involving singular shapes using the same powerful techniques they use for smooth ones.

Technical Summary: Theta-Functions for Singular Curves

Problem Statement
The paper addresses the generalization of classical theta functions and their associated geometric constructions from smooth compact Riemann surfaces to irreducible singular algebraic curves. In the smooth case, theta functions on the Jacobian $J(X)$ provide a "universal section" for line bundles of a specific degree. Through the Abel map $A: X \hookrightarrow J(X)$ , translating the theta function yields sections of all line bundles of that degree, effectively reconstructing the function theory of the curve. This framework is crucial for integrable systems, particularly in solving flows associated with KP-type equations via spectral curves.

However, when the spectral curve $X$ is singular (e.g., nodal or cuspidal), the standard Jacobian is insufficient. The relevant object is the generalized Jacobian $J(X)$ , which fibers over the Jacobian $J(\tilde{X})$ of the desingularization $\tilde{X}$ . While the generalized Jacobian and its Abel map (defined on $\tilde{X}$ ) are well-established, a corresponding theta function that acts as a universal section for line bundles on the singular curve $X$ has not been fully constructed for the general case. Previous literature has treated specific cases (e.g., simple nodes or cusps), but a unified construction for arbitrary singularities was lacking.

Methodology
The authors construct a theta function on a compactification of the generalized Jacobian $J(X)$ . The methodology proceeds through the following steps:

Structure of the Generalized Jacobian: The paper analyzes the exact sequence $0 \to Q_S(X) \to J(X) \to J(\tilde{X}) \to 0$ , where $Q_S(X)$ represents the "collapsing" of line bundles over the preimages of singular points. $Q_S(X)$ is an iterated extension of copies of $\mathbb{C}$ and $\mathbb{C}^*$ , corresponding to the local trivializations of line bundles at the singular points.
Differential Forms and Periods: The authors define a basis of meromorphic one-forms on the desingularization $\tilde{X}$ $\tilde{X}$ that span the dual of the Lie algebra of $Q_S(X)$ $Q_{S} (X)$ . These include:
- Forms with simple poles at preimages of nodes (residues $\pm 1$ ).
- Forms with higher-order poles and zero residues at preimages of cusps.
- Holomorphic differentials from $J(\tilde{X})$ .
  These forms are normalized to have zero $A$ -periods, with specific $B$ -periods forming a period matrix that extends the standard period matrix of $\tilde{X}$ .
Compactification: To define a theta function that yields finite values at the singular points (which map to infinity under the Abel map), the authors compactify the fibers of $J(X)$ corresponding to $Q_S(X)$ . Factors of $\mathbb{C}^*$ are compactified to $\mathbb{P}^1$ (adding $0$ and $\infty$ ), and factors of $\mathbb{C}$ are compactified to $\mathbb{P}^1$ (adding $\infty$ ). The theta function is constructed as a section of a line bundle $\mathcal{O}(1, \dots, 1)$ on this product of projective spaces.
Construction of the Theta Function: The theta function is built using the standard theta function $\tilde{\theta}$ $\tilde{θ}$ associated with $\tilde{X}$ $\tilde{X}$ .
- For rational desingularizations ( $\tilde{X} = \mathbb{P}^1$ ), the function is constructed explicitly as a product of terms involving exponentials of integrals (for $\mathbb{C}^*$ factors) and linear terms (for $\mathbb{C}$ factors).
- For higher genus desingularizations, the construction involves differential operators acting on $\tilde{\theta}$ . Specifically, for cuspidal singularities, the authors define operators $D_{i,h}$ based on the $B$ -periods of the singular forms and construct $\theta$ as a sum of terms involving derivatives of $\tilde{\theta}$ multiplied by the coordinates of the singular fibers.
- For arbitrary singularities, the function is defined as a sum over subsets of indices, combining exponential terms, polynomial terms in the singular coordinates, and shifted arguments of the smooth theta function.

Key Contributions and Results

Generalized Theta Function: The paper provides an explicit construction of a theta function on a compactification of $J(X)$ for an arbitrary irreducible singular curve. This function satisfies quasi-periodicity relations analogous to the classical case, with automorphy factors determined by the period matrix of the generalized Jacobian.
Universal Section Property: By translating the theta function by elements of the generalized Jacobian, the authors obtain sections of line bundles of a specific degree over the singular curve $X$ . The pull-back of these translates to the desingularization $\tilde{X}$ generates the specific subsheaves required to define line bundles on the singular curve.
Inversion of the Abel Map: The authors establish a generalized Riemann theorem. They prove that the sum of the Abel map images of the zeroes of a translated theta function $\theta_{a,b,\lambda}$ is linearly related to the translation parameters $(a, b, \lambda)$ , up to a generalized Riemann constant. Specifically, for a curve with desingularization $\tilde{X}$ , the zeroes $q_\rho$ satisfy:
$\sum_{\rho} \Pi(q_\rho) = \text{Linear Shift}(a, b, \lambda) + \kappa$
where the linear shift involves logarithmic terms for the $\mathbb{C}^*$ components and linear terms for the $\mathbb{C}$ components.
Explicit Formulas for Singularities: The paper details the specific form of the theta function and the resulting inversion formulas for nodal curves, cuspidal curves, and curves with higher-order singularities. It highlights that while the correspondence between the translation parameters and the divisor zeroes is linear in the smooth part, it involves non-linear terms (logarithms and rational functions) in the singular part, reflecting the mixing of singular points in the theta function.

Significance and Claims
The authors claim that this work extends Riemann's classical result on the function theory of curves to the singular setting. The constructed theta function serves as a "universal section" for line bundles over singular curves, mirroring the role of theta functions in the smooth case. This is significant for the theory of integrable systems, as it provides a geometric framework to handle spectral curves that degenerate into singular varieties (e.g., soliton limits).

The paper is modest in its geometric claims regarding the compactification; the authors note that the specific compactification used is "elementary" and they are "not sure if it has much geometric meaning," but it is sufficient to yield the desired theta function and universal section. Furthermore, while the construction allows for the inversion of the Abel map, the authors observe that the relationship between the translation parameters and the resulting line bundles is not a simple linear identity map (as in the smooth case) but involves complex, non-linear interactions between the singular points, which they have not found a simple way to linearize. The work is presented as a constructive generalization, providing the necessary tools to solve equations in terms of theta functions even when the underlying spectral curve is singular.