A family of Non-Weierstrass Semigroups

This paper introduces a new method utilizing syzygies to demonstrate that certain numerical semigroups are not Weierstrass, notably providing the first known examples of non-Weierstrass semigroups with multiplicity 6 and genus 13.

The Gist

Here is an explanation of the paper "A Family of Numerical Semigroups That Are Not Weierstrass Semigroups" by Eisenbud and Schreyer, translated into everyday language with analogies.

The Big Picture: The "Impossible" Club

Imagine you are at a party where everyone is trying to join an exclusive club called the Weierstrass Club.

• The Members: The members of this club are special groups of numbers (called Numerical Semigroups).
• The Rule: To be a member, a group of numbers must be able to represent the "pole orders" (a fancy way of saying "how badly a function breaks") of functions on a smooth, perfect, round surface (like a sphere or a donut).
• The Mystery: For over 100 years, mathematicians wondered: Can every possible group of numbers join this club? Or are there some groups that simply don't fit the rules of the geometry of these surfaces?

In 1892, a mathematician named Hurwitz asked this question. For a long time, people thought maybe all groups could join. But in 1980, a mathematician named Buchweitz proved that some groups are "imposters"—they look like they belong, but they can't actually exist on a smooth surface.

The Problem: Before this paper, the only known "imposters" were very large, complicated groups (like a group with a minimum number of 13). No one had found an imposter that was small and simple.

The Breakthrough: Eisenbud and Schreyer found a new way to spot imposters. They discovered the smallest, simplest imposter group ever found. It's a group built from the numbers {6, 9, 13, 16}.

---

The Analogy: The "Blueprint" and the "Crack"

To understand how they found this, let's use a construction analogy.

1. The Blueprint (The Semigroup)

Think of a numerical semigroup as a blueprint for a building. It tells you what materials you have and how they can be combined.

• If the blueprint is valid, you should be able to build a smooth, perfect house (a smooth curve) that fits this blueprint perfectly.
• If the blueprint is fake, you might be able to build a house, but it will have a crack, a hole, or a weird bump that ruins the smoothness.

2. The Old Way (Buchweitz's Method)

Previously, to prove a blueprint was fake, you had to do a massive calculation involving the "weight" of the building materials. It was like trying to prove a house is unstable by counting every single brick. It worked, but it only worked for huge, complex houses.

3. The New Method (The "Special Resolution")

Eisenbud and Schreyer invented a new way to look at the blueprint. They looked at the internal structure of the blueprint itself, specifically how the numbers relate to each other in a "chain of dependencies" (mathematicians call this a resolution).

They found a specific pattern, which they call a "Special Resolution."

• Imagine the blueprint has a hidden weak spot or a crack in its foundation.
• In their specific group {6, 9, 13, 16}, the blueprint has a very rigid structure. It forces the building to have a "section" (a specific line of support) that is singular (broken/cracked) in every possible variation of the building.
• The Logic: If you try to build a smooth house using this blueprint, the blueprint forces a crack to appear. Since a Weierstrass semigroup must represent a smooth house, this blueprint cannot belong to the club.

The "Degree-Special" Discovery

The authors define a group as "Degree-Special" if its blueprint has this specific, rigid structure that forces a crack.

• The Magic Number: They found that the group generated by {6, 9, 13, 16} is "Degree-Special."
• Why it matters: This is the smallest possible group (multiplicity 6) that can be an imposter. Before this, people thought you needed a minimum number of at least 8 or 13 to be an imposter. They proved that even with the smallest possible starting number (6), you can still create a group that is impossible to build smoothly.

The "Family" of Imposters

The paper doesn't just stop at one example. They found a whole family of these imposters.

• They showed that if you take the numbers {6, 9, g, g+3} (where g is a number like 13, 14, 15, etc., but not divisible by 3), you get a whole new set of imposters.
• They also showed that you can take a known imposter and "double" it to create even bigger imposters (using a method related to "Torres covering").

The "Smooth" vs. "Rough" Test

The authors use a clever trick involving deformations.

• Imagine you have a clay sculpture (the building). You want to know if it's perfectly smooth.
• You try to wiggle the clay (deform it) to see if it stays smooth.
• For a "Weierstrass" group, you should be able to wiggle it and keep it smooth.
• For these "Degree-Special" groups, no matter how you wiggle the clay, there is always a specific line on the sculpture that remains rough or broken.
• Because the "roughness" is forced by the math of the blueprint itself, the group cannot represent a smooth surface.

Summary: Why Should You Care?

1. Solving a 130-year-old puzzle: They answered a question Hurwitz asked in 1892 by finding the smallest possible counter-example.
2. A New Tool: They didn't just find one number; they gave mathematicians a new "detector" (the Degree-Special method) to find many more imposters in the future.
3. The "Lowest" Record: They broke the record for the smallest "impossible" group. It's like finding a tiny, simple puzzle piece that doesn't fit in the box, proving the box isn't as perfect as we thought.

In a nutshell: The paper says, "We found a tiny, simple group of numbers that looks like it should belong to the club of smooth geometric shapes, but we proved it's a fraud because its internal structure forces a permanent crack. And we found a whole family of them!"

Technical Summary

Here is a detailed technical summary of the paper "A Family of Numerical Semigroups That Are Not Weierstrass Semigroups" by David Eisenbud and Frank-Olaf Schreyer.

1. Problem Statement

The paper addresses the Weierstrass Semigroup Problem, originally posed by Adolf Hurwitz in 1892.

• Context: A numerical semigroup $S$ is a subset of non-negative integers closed under addition with a finite complement (the "gaps"). The size of the complement is the genus ($g$), and the smallest non-zero element is the multiplicity ($m$).
• Definition: A semigroup is Weierstrass if it arises as the set of pole orders of rational functions regular everywhere except at a single point $p$ on a smooth projective curve $X$ of genus $g$.
• The Question: Do all numerical semigroups occur as Weierstrass semigroups for some curve and point?
• Historical Context:
  • It was known that all semigroups with $m < 6$ or $g < 8$ are Weierstrass.
  • In 1980, Buchweitz provided the first counterexample ($m=13, g=16$) using the dimension of quadratic differential forms.
  • Prior to this paper, no non-Weierstrass semigroup was known with genus $< 16$ or multiplicity $< 8$.

2. Methodology

The authors introduce a new method based on deformation theory and homological algebra (specifically syzygies of the semigroup ideal) to prove that certain semigroups cannot be Weierstrass.

A. Pinkham's Theorem and Deformations

The approach relies on a result by Henry Pinkham (1974): A semigroup $S$ is Weierstrass if and only if its semigroup algebra $k[S]$ admits a homogeneous 1-parameter smoothing (a flat family of algebras where the special fiber is $k[S]$ and the generic fiber corresponds to a smooth curve).

• Strategy: The authors argue that if a specific structural feature of the resolution of $k[S]$ forces the existence of a "distinguished section" in any flat deformation, and this section corresponds to a singular point in every fiber, then the semigroup cannot be Weierstrass (since Weierstrass points on smooth curves must be smooth).

B. Degree-Special Semigroups and Special Resolutions

The core of the method involves analyzing the minimal free resolution of the semigroup ring $P/I$ (where $P$ is a polynomial ring and $I$ is the semigroup ideal).

• Format {1, 6, 8, 3}: The authors focus on semigroups whose resolutions have the format $0 \to P^3 \xrightarrow{\phi_3} P^8 \xrightarrow{\phi_2} P^6 \xrightarrow{\phi_1} P \to 0$.
• Definition (Degree-Special): A semigroup is degree-special if its resolution admits a specific "subcomplex" of format $\{1, 4, 4, 1\}$ that persists under flat graded deformations. This subcomplex is defined by specific zero-patterns in the matrices $\phi_2$ and $\phi_3$ (specifically, certain entries have negative formal degrees and thus vanish).
• The Singularity Argument:
  1. In a degree-special semigroup, the first column of $\phi_3$ (restricted to the subcomplex) defines a regular sequence of 4 elements.
  2. These elements define a subscheme in the spectrum of the algebra.
  3. The authors prove that in any flat deformation, the Jacobian matrix of the ideal has minors lying in the ideal generated by these 4 elements.
  4. Consequently, the variety $\text{Spec}(P/I)$ is singular along this subscheme in every fiber of the deformation.
  5. Since a Weierstrass semigroup requires a smoothing to a smooth curve, the existence of this persistent singularity proves the semigroup is not Weierstrass.

3. Key Contributions and Results

A. First Examples with Minimal Multiplicity and Genus

The paper provides the first known examples of non-Weierstrass semigroups with the lowest possible parameters:

• Multiplicity: $m = 6$ (previously the lowest known was $m=8$).
• Genus: $g = 13$ (previously the lowest known was $g=16$).
• Specific Example: The semigroup generated by $\{6, 9, 13, 16\}$ is proven to be non-Weierstrass.

B. Infinite Families

The authors construct infinite families of non-Weierstrass semigroups:

• General Form: Semigroups generated by $\{6, 9, g, g+3\}$ for any $g \ge 13$ where $g \not\equiv 0 \pmod 3$.
• Kunz-Waldi Connection: They relate these to Kunz-Waldi semigroups, noting that while some degree-special semigroups are Kunz-Waldi, others (like $\{6, 9, 14, 17\}$) are not, showing the method's broader applicability.

C. Computational Verification

• The authors provide a Macaulay2 program (in auxiliary files) verifying that all numerical semigroups of genus $< 11$ are Weierstrass. This tightens the boundary of the problem significantly.
• They list all degree-special semigroups up to genus 20, showing they exist for every genus from 13 to 20 (except 18).

D. Extension via Covering Constructions

Using a theorem by Torres (1994), the authors show how to generate further non-Weierstrass semigroups from existing ones. If $L$ is non-Weierstrass, then a semigroup $L'$ generated by $2L$and sufficiently large odd generators is also non-Weierstrass. This allows the construction of non-Weierstrass semigroups for **every multiplicity$m \ge 6$**.

4. Significance

• Resolution of a Long-Standing Gap: The paper closes the gap between the known bounds for Weierstrass semigroups ($m<6, g<8$) and the first known counterexamples ($m=13, g=16$). It establishes that the threshold for non-Weierstrass behavior is much lower than previously thought.
• New Theoretical Tool: The method of using syzygy structures and degree-special resolutions to detect singularities in deformations is a novel approach in algebraic geometry. It moves beyond Buchweitz's cohomological dimension argument to a more structural, homological obstruction.
• Completeness: By proving that all semigroups of genus $< 11$ are Weierstrass, the authors provide a near-complete classification of the "small" cases, leaving the exact boundary for $11 \le g < 13$as the remaining open question (though the paper suggests$g=13$ is the first failure).
• Counter-Intuitive Results: The paper demonstrates that the existence of a "section" in a family of deformations does not guarantee smoothness; in these specific cases, the section tracks a singularity, which is the key to the non-Weierstrass proof.

In summary, Eisenbud and Schreyer fundamentally lower the known bounds for non-Weierstrass semigroups and introduce a powerful homological technique for proving that certain algebraic structures cannot arise from smooth curves.