🔢 mathematics

A Eudoxian study of discriminant curves associated to normal surface singularities

This paper generalizes a theorem by Gryszka, Gwoździewicz, and Parusiński to normal surface singularities by demonstrating that the initial Newton polynomial of a discriminant's defining series depends solely on the germs of the curves defined by the morphism components, utilizing a proof strategy inspired by the Eudoxian method of comparing magnitudes through their positive integral multiples.

Original authors: Evelia Rosa García Barroso, Patrick Popescu-Pampu

Published 2026-03-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Evelia Rosa García Barroso, Patrick Popescu-Pampu

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

Imagine you are a detective trying to solve a mystery, but instead of looking at fingerprints, you are looking at the "shadows" cast by complex mathematical shapes. This paper is about understanding how these shadows behave when the objects casting them are slightly distorted.

Here is the story of the paper, broken down into simple concepts and everyday analogies.

1. The Setup: The "Shadow" and the "Object"

Imagine you have a crumpled piece of paper (a Normal Surface Singularity). On this paper, you draw two distinct lines, let's call them Line F and Line G.

Now, imagine you have a special camera (a Finite Morphism) that takes a picture of these two lines and projects them onto a flat wall (the Target Plane, $\mathbb{C}^2$ ). Because the paper is crumpled, the projection isn't a perfect copy. Some parts of the lines might overlap, fold over, or create weird kinks.

The Discriminant is the "shadow" or the "map of trouble" on the wall. It shows exactly where the projection goes wrong—where the lines fold, cross, or get squashed. It's the boundary between the "normal" parts of the image and the "chaotic" parts.

2. The Big Question: Does the Shadow Change?

The authors ask a very specific question:

If I slightly wiggle the lines on the crumpled paper (changing the equations of Line F and Line G just a tiny bit), does the shape of the shadow on the wall change?

In math terms, they are looking at the Initial Newton Curve. Think of this as the "skeleton" or the "outline" of the shadow. It ignores the tiny, messy details and focuses on the main structure.

The Discovery: The authors prove that the skeleton of the shadow does NOT change, as long as the original lines (Line F and Line G) stay in the same general "neighborhood" on the paper. Even if you wiggle the paper or the lines slightly, the fundamental shape of the shadow remains the same. It only depends on the type of lines you started with, not the exact wiggles.

3. The "Eudoxian" Strategy: The Ancient Scale

The paper gets its cool name ("Eudoxian") from an ancient Greek mathematician named Eudoxus.

The Ancient Problem: Eudoxus wanted to compare two things (like a long rope and a heavy rock) without having a ruler or a scale. How do you say "Rope A is to Rock A" as "Rope B is to Rock B"?
Eudoxus' Solution: He said, "Don't compare them directly. Instead, take 100 copies of Rope A and 100 copies of Rock A. Then take 50 copies of Rope B and 50 copies of Rock B. If the ratios hold up across all these different multiples, the relationship is the same."

How the Authors Use This:
Instead of just looking at the pair $(f, g)$ , the authors look at every possible combination of powers: $(f^2, g^3)$ , $(f^5, g^2)$ , $(f^{100}, g^{99})$ , and so on.

They treat the functions $f$ and $g$ like the "Rope" and "Rock."
By checking how the shadows behave for all these different "multiples" (powers), they can prove that the underlying structure is stable.
It's like checking if a building is stable not by looking at the front door, but by shaking the foundation, the roof, and the basement simultaneously. If it holds up in all those scenarios, you know the structure is solid.

4. The Tools: Measuring the "Twist"

To prove their point, the authors use a few mathematical "rulers":

The Newton Polygon: Imagine the shadow as a cloud of points. If you draw the tightest possible rubber band around the "outer" points of this cloud, you get a shape called a Newton Polygon. This shape tells you the "slope" and "direction" of the shadow.
The Milnor Number: This is a way of counting how "twisted" or "knotted" a singularity is. Think of it as a "knot counter."
The Formula: The authors found a new formula that connects the "knots" in the original lines to the "knots" in the shadow. It's like a recipe: If you know how twisted the ingredients are, you can calculate exactly how twisted the cake will be.

5. The "Special" Points

The paper also talks about "Special Points" on the shadow. Imagine the shadow is a map. Most of the map is flat, but there are a few "mountain peaks" or "valleys" where the geometry is special.

The authors prove that these special peaks are determined entirely by the original lines.
Even if you change the lines slightly, the locations of these peaks on the map don't move. They are locked in place by the geometry of the original setup.

6. Why Does This Matter?

Before this paper, mathematicians knew how to describe these shadows for simple, smooth surfaces (like a flat sheet of paper). But real-world math often deals with "crumpled" or "singular" surfaces (like a cone or a crumpled ball).

This paper says: "It doesn't matter if the surface is crumpled or smooth. If you know the shape of the lines you drew, you can predict the shape of the shadow, even if the surface is weird."

Summary Analogy

Think of a puppet show.

The Puppeteer is the mathematician.
The Puppets are the functions $f$ and $g$ .
The Screen is the target plane.
The Shadow is the discriminant.

The authors prove that if you have two different puppet shows using the same types of puppets (even if the strings are tied slightly differently), the silhouette of the shadow cast on the screen will look exactly the same. You can't tell the difference just by looking at the outline of the shadow.

They used an ancient Greek trick (checking all possible multiples) to prove that the "soul" of the shadow is invariant, regardless of the minor details of the performance.

1. Problem Statement

The paper addresses the study of discriminant curves associated with finite morphisms $\phi = (f, g): (S, s) \to (\mathbb{C}^2, 0)$ , where $(S, s)$ is a normal surface singularity (a germ of a normal complex analytic surface).

The Object of Study: The discriminant divisor $\Delta_\phi \subset (\mathbb{C}^2, 0)$ , which is the image of the critical locus of $\phi$ .
The Specific Question: Does the initial Newton curve of $\Delta_\phi$ (the curve defined by the initial Newton polynomial of a defining series of the discriminant) depend on the specific choice of the functions $f$ and $g$ , or only on the geometric data of the divisors they define, $Z(f)$ and $Z(g)$ ?
Context: Previous work by Gryszka, Gwoździewicz, and Parusiński established this invariance for the case where the source $(S, s)$ is smooth. This paper aims to generalize that result to arbitrary normal surface singularities.

2. Methodology: The "Eudoxian Strategy"

The authors coin the term "Eudoxian strategy" to describe their core methodological approach, drawing an analogy to Eudoxus' theory of proportions.

The Concept: Instead of analyzing the pair $(f, g)$ directly, the authors study the family of morphisms $(f^b, g^a)$ for all pairs of coprime positive integers $(a, b)$ .
The Analogy: Just as Eudoxus compared magnitudes $M_1, M_2$ by comparing the sizes of their integer multiples $(bM_1, aM_2)$ to define a ratio without a pre-existing total order, the authors analyze the pair $(f, g)$ by simultaneously examining the entire family of pairs $(f^b, g^a)$ .
Implementation:
1. They consider the pencil of curves generated by $(f^b, g^a)$ .
2. They relate the intersection numbers of the discriminant $\Delta_\phi$ with binomial curves $B_{a,b,\lambda} = Z(u^b - \lambda v^a)$ to the Milnor numbers of the functions $f^b - \lambda g^a$ .
3. They utilize tropical geometry concepts (tropical functions and Newton polygons) to encode this data.

3. Key Theoretical Contributions and Tools

The proof relies on several new generalizations and the synthesis of existing theorems:

A. Generalization of Intersection Formulas (Theorem B)

The authors generalize a formula by Casas-Alvero (originally for smooth surfaces) to arbitrary normal surface singularities.

Formula: For a finite morphism $\phi: (S, s) \to (T, t)$ and a reduced function $F$ on the target:
$\Delta_\phi \cdot Z(F) = (\mu(\phi^*F) - 1) - \deg(\phi) \cdot (\mu(F) - 1)$
Significance: This relates the intersection of the discriminant with a curve to the Milnor numbers of the function and its pullback, utilizing the Riemann-Hurwitz formula for ramified covers of Riemann surfaces with boundary.

B. Characterization of Special Points (Theorem 3.6)

They utilize a theorem by Delgado and Maugendre which establishes an equivalence between two definitions of "special members" (distinguished points) in a pencil of curves associated with a finite morphism:

Geometric: Points where the line is tangent to the discriminant.
Resolution-theoretic: Points arising from the image of "dicritical" components or singular points in a good resolution that lifts the indeterminacies of the pencil.

C. Combinatorial Invariance of Jacobian Newton Polygons (Theorem C)

They prove that the Jacobian Newton polygon (the Newton polygon of the discriminant) depends only on the combinatorial type of the triple $(S, Z(f), Z(g))$ .

This is achieved by showing that the tropical function of the discriminant, evaluated at primitive vectors $(a, b)$ , depends only on the orders of vanishing of $f$ and $g$ along the exceptional divisors of a good resolution.

4. Main Results

Theorem A (Main Theorem)

Let $(S, s)$ be a normal surface singularity and $\phi = (f, g): (S, s) \to (\mathbb{C}^2, 0)$ a finite morphism. The initial Newton curve of the discriminant divisor $\Delta_\phi$ depends only on the divisors $Z(f)$ and $Z(g)$ , up to toric automorphisms of $\mathbb{C}^2$ .

Implication: If one replaces $f$ by $(1+\alpha)f$ and $g$ by $(1+\beta)g$ (where $\alpha, \beta$ vanish at the singularity), the initial Newton curve of the discriminant remains unchanged (up to toric scaling).
New Invariant: This allows the definition of a new invariant for pairs of principal effective divisors on normal surface singularities: their associated initial Newton curve.

Theorem C (Combinatorial Invariance)

The Jacobian Newton polygon $N(\Delta_\phi)$ is determined entirely by the combinatorial type of the triple $(S, Z(f), Z(g))$ . This generalizes a result by Gwoździewicz from smooth sources to normal sources.

5. Significance and Impact

Generalization of Smooth Results: The paper successfully extends deep results regarding discriminant curves from the well-understood smooth case to the much more complex setting of normal surface singularities.
Moduli vs. Discrete Invariants: Previous studies of finite morphisms focused on discrete invariants (like Milnor numbers or degrees). This paper introduces moduli (the shape of the initial Newton curve) into the study of singularities, showing that this continuous invariant is actually rigidly determined by the combinatorial geometry of the source divisors.
Methodological Innovation: The "Eudoxian strategy" provides a powerful framework for comparing geometric objects by analyzing their integer multiples. This approach unifies the study of pencils of curves and tropical geometry in the context of singularity theory.
Connection to Tropical Geometry: The paper bridges classical singularity theory (Milnor numbers, resolutions) with modern tropical geometry (tropical functions, Newton polygons), showing how tropical functions of discriminants can be computed via intersection numbers and Milnor numbers.
Clarification of Discriminant Definitions: The authors rigorously prove that for smooth varieties, the discriminant defined via differential forms (divisorial) coincides with the one defined via Fitting ideals (Teissier's definition), and that these definitions behave well under base change.

Summary of Proof Structure

Setup: Define the discriminant and the "Eudoxian" family of morphisms $(f^b, g^a)$ .
Intersection Theory: Use the generalized Casas-Alvero/Lê formula (Theorem B) to express intersection numbers of the discriminant with binomial curves in terms of Milnor numbers.
Tropical Link: Relate these intersection numbers to the tropical function of the discriminant and the roots of univariate polynomials associated with the Newton polygon faces.
Invariance: Show that the Milnor numbers of $f^b - \lambda g^a$ depend only on the combinatorial type of the resolution, implying the Newton polygon is invariant (Theorem C).
Deformation: Use a holomorphic deformation argument (varying coefficients of $f$ and $g$ ) to show that the roots and their multiplicities of the initial Newton polynomial are invariant under the transformation $(f, g) \to ((1+\alpha)f, (1+\beta)g)$ .
Conclusion: Conclude that the initial Newton curve is an invariant of the pair of divisors $(Z(f), Z(g))$ .