Inversion of the Abel--Prym map for real curves with involutions

This paper provides a detailed presentation of the inversion of the Abel--Prym map for real algebraic curves with holomorphic involutions, specifically addressing the previously unexamined case of non-separating real curves and formulating the symmetry of the associated Prym theta function.

The Gist

The Big Picture: Unraveling a Knot

Imagine you have a very complex, knotted piece of string (a mathematical object called a curve). Mathematicians have a special tool called the Abel Map that takes this knotted string and flattens it out onto a smooth, flat surface (called a Jacobian). This is great because it turns a messy knot into a neat point on a map.

However, the real challenge is the Inversion Problem: If I give you a specific point on that flat map, can you tell me exactly how the string was knotted to get there? In other words, can you reconstruct the original shape from the flattened point?

This paper is about solving that puzzle for a very specific, tricky type of string: one that has symmetry (it looks the same if you flip it) and is real (it exists in our physical world, not just in imaginary math land).

---

Key Characters and Concepts

1. The Curve with a "Mirror" (The Involution)

Think of the curve as a piece of fabric.

• Holomorphic Involution ($\sigma$): Imagine folding the fabric perfectly in half. If you fold it, the top layer matches the bottom layer perfectly. This is a "holomorphic" symmetry.
• Anti-holomorphic Involution ($\tau$): Imagine looking at the fabric in a mirror. The mirror image is the "real" version of the curve. In math, this is what makes the curve "real" (like the numbers on a ruler) rather than "complex" (like numbers with imaginary parts).

The paper studies curves that have both of these symmetries at the same time.

2. The "Prym" Variety: The Shadow of the Knot

When you fold that fabric in half (using $\sigma$), some parts of the string overlap, and some parts are unique to the fold.

• The Jacobian is the map of the entire string.
• The Prym Variety is a special, smaller map that only tracks the parts of the string that are "anti-symmetric" (the parts that cancel each other out when you fold). It's like looking at the shadow of the knot rather than the knot itself.

The author introduces a concept called the isoPrym variety. Think of this as a "universal cover" or a "zoomed-in version" of the shadow. It's a slightly larger space that makes the math easier to handle, like unfolding a map to see all the tiny details.

3. The "Real" Problem

Most math papers deal with curves that exist in a purely abstract, imaginary world. This paper focuses on Real Curves.

• Separating Curves: Imagine a rubber band floating in space. If you cut it along its "real" parts (the ovals), it falls apart into two separate pieces.
• Non-Separating Curves: Imagine a figure-eight. If you cut it along the real parts, it stays in one piece, just with a hole in it.

The paper is unique because it solves the inversion problem for both types, especially the "non-separating" ones, which had been largely ignored or considered too messy to solve completely before.

---

The Solution: How to Reconstruct the Knot

The paper provides a recipe (an algorithm) to go from the flat map point back to the knotted string. Here is the step-by-step logic:

Step 1: The "Riemann Vanishing" Clue

In the old days, mathematicians used a theorem (Riemann's Vanishing Theorem) to find the knots. It said: "If you have a point on the map, the original knot is made of the places where a special function equals zero."

• The Problem: For these symmetric curves, this method gives you too many points. It's like getting a list of 10 suspects when you only need to find 5. The list is messy and redundant.

Step 2: The "Symmetry" Filter

The author realizes that because the curve has a mirror ($\tau$) and a fold ($\sigma$), the solution must respect these symmetries.

• If the point on the map has a specific symmetry (e.g., it looks the same when reflected), then the original knot must also be symmetric.
• The paper proves that if you look for knots that are invariant (unchanged) under these symmetries, the "messy list" of points shrinks down to the exact, correct solution.

Step 3: The "Theta Function" Calculator

To actually find the points, the author uses a special mathematical tool called the Theta Function.

• Think of the Theta function as a metal detector.
• You set the detector to a specific frequency (based on your point on the map).
• The detector beeps at the locations of the "knots" (the zeroes).
• The paper shows that for real curves, the metal detector has a special pattern. If the point on the map is "real" (symmetric), the beeps will happen in pairs that mirror each other perfectly.

The "Aha!" Moment

The paper's main breakthrough is showing that for these specific real curves:

1. You don't need to solve a massive, impossible equation. Instead, you can calculate a few "symmetric sums" (like adding up the coordinates of the points).
2. The symmetry acts as a shortcut. By knowing the curve is "real," you know the solution must be "real" too. This cuts the work in half and guarantees the solution is physically meaningful.

Summary Analogy

Imagine you are trying to reconstruct a shattered vase (the curve) from a single photo of its shadow (the map point).

• Old Method: You try to guess every possible way the shards could fit. It takes forever, and you get many wrong answers.
• This Paper's Method: You notice the vase has a specific pattern (it's a "real" vase with a fold). You realize that if the shadow is symmetrical, the vase must be symmetrical.
• The Result: You only have to look for symmetrical arrangements of shards. This instantly eliminates 90% of the wrong guesses and gives you the exact shape of the vase.

Why Does This Matter?

This isn't just about abstract shapes. These mathematical curves are used to model integrable systems—complex physical systems that behave in predictable ways, like waves in the ocean, light in fiber optics, or particles in quantum mechanics.
By making it easier to "invert" the map (solve the equations), this paper gives physicists and engineers better tools to predict how these complex systems will behave in the real world.

Technical Summary

1. Problem Statement

The paper addresses the Jacobi inversion problem specifically for real algebraic curves (curves equipped with an antiholomorphic involution $\tau$) that also possess a holomorphic involution $\sigma$.

• Context: In classical algebraic geometry, the Abel map establishes a correspondence between divisors on a curve $\Sigma$ and points on its Jacobian variety $Jac(\Sigma)$. The inverse problem (recovering the divisor from a point in the Jacobian) is solved via the Riemann Vanishing Theorem, which expresses the divisor as the zero set of a Riemann $\theta$-function.
• The Specific Challenge: When a curve has a holomorphic involution $\sigma$, one considers the Prym variety (a sub-variety of the Jacobian consisting of $\sigma$-anti-invariant points). The corresponding map is the Abel–Prym map.
  • Unlike the classical Jacobian case, the inversion for the Prym map is more complex. The Riemann Vanishing Theorem yields a divisor $\zeta$ of degree $2h$ (where $h = \dim(\text{isoPrym})$), but these points are not free; they satisfy the relation $\zeta + \sigma\zeta \sim D$ (where $D$ is a constant divisor).
  • Furthermore, for real curves (where $\tau$ is an antiholomorphic involution commuting with $\sigma$), the literature lacks a complete description of the inversion theorem, particularly for non-separating real curves (where the set of fixed points of $\tau$ does not disconnect the curve).
• Goal: To provide a rigorous formulation and proof of the inversion theorem for the Abel–Prym map on real curves, covering both separating and non-separating topological types, and to establish the symmetry properties of the associated Prym $\theta$-functions.

2. Methodology

The author employs a combination of algebraic geometry, complex analysis, and the theory of Abelian varieties. The methodology proceeds in three main stages:

A. Topological and Algebraic Setup

• Real Bases: The paper defines "real bases" of homology cycles $\{a_j, b_j\}$ for real curves. These bases are adapted to the action of the antiholomorphic involution $\tau$.
  • Separating Type ($\epsilon=1$): The curve is disconnected by the fixed points (ovals) of $\tau$.
  • Non-separating Type ($\epsilon=0$): The curve remains connected after removing the fixed points.
• Period Matrices: The author analyzes the Riemann matrix $B$ and the Prym matrix $\Pi$. By utilizing the commutativity of $\sigma$ and $\tau$, specific symmetries and reality conditions for these matrices are derived.
  • Key Lemma: The action of $\tau$ on the normalized holomorphic differentials induces a permutation $t$ of indices, leading to relations like $\tau^* \omega = -t\omega$.

B. Symmetry Analysis of $\theta$-Functions

• Prym $\theta$-Function: Defined using the Prym matrix $\Pi$. The paper investigates how this function behaves under the action of $\tau$ and $\sigma$.
• Symmetry Lemmas: The author proves that for separating curves, $\theta(z) = \theta(tz)$, and for non-separating curves, specific shifts involving a vector $\lambda$ (related to the periods of the non-oval cycles) preserve the function. These symmetries are crucial for determining the invariance of the zero divisors.

C. The Inversion Procedure

• Abel–Prym Map: Defined as $A': \Sigma \to P_0$ (the isoPrym variety), mapping a point $P$ to the integral of $\sigma$-anti-invariant differentials.
• Auxiliary Function: The function $F_z(P) = \theta(A'(P) - \epsilon^{-1}z)$ is constructed, where $\epsilon$ is a diagonal matrix scaling the Prym coordinates.
• Zero Divisor Analysis: The zero divisor $\zeta$ of $F_z(P)$ is analyzed. The paper utilizes the Veselov–Novikov condition ($A(\zeta) + A(\sigma\zeta) = 2\Delta$) and the properties of symmetric functions of the roots.
• Effective Calculation: Following Dubrovin's approach, the paper outlines a method to calculate the symmetric functions of the points in the divisor $\zeta$ using derivatives of the $\theta$-function, reducing the inversion to solving a degree $2h$ algebraic equation.

3. Key Contributions

1. Comprehensive Treatment of Non-Separating Real Curves:
   Previous literature (e.g., Fay, Sheinman's earlier works) focused heavily on separating curves or specific cases (like hyperelliptic curves). This paper provides the first detailed presentation of the inversion theorem for non-separating real curves with a holomorphic involution.

2. Formulation of the Inversion Theorem for Real Curves:
   The paper establishes a precise correspondence between points in the isoPrym variety and divisors on the curve. It proves that the inverse image of specific real subvarieties of the isoPrym variety corresponds to divisors that are invariant under specific involutions ($\tau$ or $\sigma\tau$).

3. Symmetry of the Prym $\theta$-Function:
   The author derives explicit symmetry relations for the Prym $\theta$-function on real curves.

   • For separating curves: $\theta(z) = \theta(tz)$.
   • For non-separating curves: $\theta(z) = \theta(z + \lambda)$ (under specific conditions where all $\sigma$-invariant cycles are ovals).
     These symmetries are essential for proving the invariance of the zero divisors.

4. Clarification of the "Real" Structure:
   The paper distinguishes between the actions of $\tau$ (antiholomorphic) and $\sigma$ (holomorphic) and how their combination ($\sigma\tau$) affects the divisor classes. It resolves the ambiguity regarding the existence of $\tau$-invariant divisors in the non-separating case (showing they do not exist due to contradictory period conditions).

4. Main Results

Theorem 4.5 (Inversion Theorem for Real Curves)

Let $F_{\epsilon^{-1}z}(P)$ be the auxiliary $\theta$-function. The zero divisor $\zeta = A'^{-1}(z)$ satisfies the following invariance properties based on the curve type and the point $z$:

• Case 1: Separating Curves
  • If $z + tz = 0$, then $\zeta$ is $\tau$-invariant (i.e., $\tau(\zeta) = \zeta$).
  • If $z - tz = 0$, then $\zeta$ is $\sigma\tau$-invariant.
• Case 2: Non-Separating Curves
  • If $z - \bar{z} = -\epsilon\lambda$, then $\zeta$ is $\sigma\tau$-invariant.
  • Note: The paper explicitly states that for non-separating curves, a $\tau$-invariant divisor $\zeta$ does not exist because the required condition $z + \bar{z} = -\lambda$ is contradictory (as $\lambda$ is purely imaginary and non-zero).

Effective Inversion (Theorem 4.4)

The paper confirms that finding the divisor $\zeta$ is equivalent to solving a degree $2h$ algebraic equation. The coefficients of this equation are given by the symmetric functions $\sigma_k(z)$, which can be computed explicitly via derivatives of the Prym $\theta$-function:
$$\sigma_k(z) = c - \sum_{q \in \pi^{-1}(\infty)} \text{res}_q f \, d \ln F_{\epsilon^{-1}z}$$
This provides a constructive algorithm for the inversion problem, avoiding direct solution of the transcendental equation $\theta(A'(P) - \epsilon^{-1}z) = 0$.

5. Significance

• Integrable Systems: The Jacobi inversion problem is a cornerstone of the theory of integrable systems (e.g., KdV, sine-Gordon, Schrödinger equations). Real solutions to these equations often correspond to real algebraic curves. This paper provides the necessary geometric machinery to construct real solutions for systems associated with curves possessing involutions.
• Completeness of Theory: By addressing the "non-separating" case, the paper fills a significant gap in the literature regarding real Prym varieties. This is crucial for applications where the topology of the real part of the curve is non-trivial (e.g., in soliton theory on non-orientable surfaces or specific boundary value problems).
• Algorithmic Application: The derivation of the $\theta$-function formula for symmetric functions offers a practical computational tool. Instead of numerically solving for roots of a transcendental function, one can compute derivatives of the $\theta$-function to generate the coefficients of an algebraic equation, which is often more stable and efficient.

In summary, Sheinman's work extends the classical theory of Abel-Jacobi inversion to the more complex setting of real curves with involutions, providing a unified framework for both separating and non-separating topological types and establishing the precise symmetry conditions required for the inversion to yield real-invariant divisors.