🔢 mathematics

Smoothness results for the schemes of special divisors on general k-gonal curves

This paper establishes smoothness results for specific open subsets of Brill-Noether degeneracy schemes $\overline{\Sigma}_{\overrightarrow {e}}(C,f)$ on general $k$ -gonal curves, demonstrating that these schemes (and consequently certain irreducible components of $W^r_d(C)$ ) are smooth along these subsets unless they lie at the intersection of multiple components, while also characterizing invertible sheaves with injective Petri maps and clarifying the structure of the singular locus of $W^r_d(C)$ .

Original authors: Marc Coppens

Published 2026-03-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Marc Coppens

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 cartographer trying to map a vast, mysterious landscape called Algebraic Geometry. In this world, the "land" consists of shapes called curves (think of them as twisted, knotted loops of string floating in space).

Mathematicians are obsessed with understanding the "special spots" on these curves. These spots are defined by how many ways you can draw lines or shapes that fit perfectly onto the curve. This is the realm of Brill-Noether Theory.

For a long time, mathematicians knew that if you pick a "generic" curve (a completely random, typical one), these special spots form a smooth, well-behaved landscape. However, the author of this paper, Marc Coppens, is looking at a specific, slightly more "twisted" type of curve called a $k$ -gonal curve.

The Analogy: The "Twisted Rope" vs. The "Straight String"

The Generic Curve: Imagine a straight, smooth piece of string. If you try to wrap a rubber band around it in specific ways, the rules are simple and predictable. The "map" of all possible rubber band positions is perfectly smooth.
The $k$ -gonal Curve: Now, imagine a rope that has been twisted and folded over itself $k$ times before being tied off. It's still a single rope, but it has a specific "fold" or "structure" (called a morphism to a line). Because of this twist, the rules for wrapping rubber bands (divisors) become more complex.

The Problem: Where is the Map Smooth?

In this mathematical landscape, the "map" is a scheme called $W^r_d(C)$ .

Smooth points: These are places where the map is a nice, flat plain. If you stand here, everything makes sense, and you can predict what's happening next.
Singular points: These are the "cliffs," "canyons," or "cracks" in the map. If you stand here, the rules break down, and the geometry gets messy.

For a long time, mathematicians thought the only "cracks" in the map happened when you had too many rubber bands (a condition called $W^{r+1}_d$ ). They thought: "If you aren't in the 'too many' zone, the map is smooth."

Coppens' discovery is that this isn't always true for twisted ropes ( $k$ -gonal curves).

The Key Concepts Explained Simply

1. The "Petri Map" (The Stress Test)

Imagine you have a rubber band (a line bundle) on your twisted rope. You want to know if it's "stable."

The Petri Map is like a stress test. It checks if the rubber band is "tight" in a unique way or if it's loose and wobbly.
The Rule: If the stress test passes (the map is injective), the spot on the map is smooth. If it fails, you might be standing on a cliff.
The Twist: For these special twisted ropes, the stress test doesn't just depend on the rubber band itself; it depends on how the rubber band interacts with the twist of the rope (the line bundle $M$ ).

2. The "Splitting Sequence" (The Fingerprint)

Every rubber band on a twisted rope has a unique "fingerprint" called a splitting sequence. It's a list of numbers describing how the rubber band wraps around the different folds of the rope.

Coppens realized that by looking at this list of numbers, you can predict exactly where the smooth spots are and where the cracks are.
He found that even if you aren't in the "too many rubber bands" zone, you can still hit a crack if your fingerprint (splitting sequence) interacts badly with the rope's twist.

3. The "Components" (The Islands)

Sometimes, the map $W^r_d(C)$ isn't just one big plain; it's an archipelago of islands (irreducible components).

The Big Surprise: You can stand on a spot that is smooth on your specific island, but because that spot is also touching a different island, the whole map is considered "cracked" (singular) at that point.
Coppens proved that for these twisted ropes, you can have a spot that is the only island it touches, yet it's still a crack in the map. This happens because the "stress test" (Petri map) fails due to the rope's twist, not because you're touching another island.

The Main Takeaways (The "So What?")

Refining the Map: Coppens didn't just say "it's broken." He drew a new, more detailed map. He identified specific open areas (like $V^r_{d,l}$ ) where the map is guaranteed to be smooth, even for these tricky twisted ropes.
The "Twist" Matters: He showed that the smoothness of the map depends entirely on how the rubber band interacts with the rope's twist. If the rubber band avoids the twist in a specific way, the map is smooth. If it hugs the twist too tightly, the map gets rough.
A New Rule for Cracks: He proved that the old rule ("cracks only happen if you have too many rubber bands") is wrong for these curves. Cracks can happen for other reasons, specifically when the "stress test" fails because of the curve's geometry.

Why Should You Care?

Think of this like engineering. If you are building a bridge (the mathematical theory), you need to know exactly where the stress points are.

If you assume the bridge is smooth everywhere except the obvious weak spots, you might build a bridge that collapses.
Coppens' paper is like a new blueprint that says: "Hey, for this specific type of bridge material (k-gonal curves), there are hidden stress points you didn't know about. Here is exactly how to find them and how to ensure the bridge stays standing."

In short, this paper takes a complex, abstract map of mathematical shapes and adds a layer of detail that explains exactly where the ground is solid and where it might crumble, specifically for shapes that have a "twist" in them. It uses the "stress test" (Petri map) as a compass to navigate these rough terrains.

1. Problem Statement

The paper addresses the geometry of Brill-Noether schemes $W^r_d(C)$ , which parameterize line bundles of degree $d$ with at least $r+1$ global sections on a smooth irreducible projective curve $C$ of genus $g$ .

While the classical Brill-Noether theory (for general curves) establishes that $W^r_d(C)$ is smooth and has the expected dimension $\rho = g - (r+1)(g-d+r)$ , the situation is more complex for $k$ -gonal curves (curves admitting a degree $k$ map $f: C \to \mathbb{P}^1$ ) where $k < \lceil \frac{g+3}{2} \rceil$ .

The Core Issue: For general $k$ -gonal curves, the scheme $W^r_d(C)$ can be singular. The singular locus is not simply the union of $W^{r+1}_d(C)$ and the intersections of distinct irreducible components, as is often the case for general curves.
Specific Gap: The paper investigates the smoothness of specific irreducible components of $W^r_d(C)$ associated with splitting types of the line bundle $L$ with respect to the $k$ -gonal map $f$ . It seeks to determine exactly where these schemes are smooth and how the Petri map injectivity relates to this smoothness.

2. Methodology

The author employs a combination of degeneracy locus theory, Fitting ideals, and Petri map analysis to analyze the local structure of these schemes.

Splitting Loci and Degeneracy Schemes:
- The paper utilizes the Brill-Noether splitting loci $\Sigma_{\vec{e}}(C, f)$ , defined by the splitting type $\vec{e} = (e_1, \dots, e_k)$ of the pushforward $f_*L \cong \bigoplus \mathcal{O}_{\mathbb{P}^1}(e_i)$ .
- It distinguishes between the set-theoretic closure $\Sigma_{\vec{e}}(C, f)$ and the scheme-theoretic degeneracy locus $\Sigma_{\vec{e}}(C, f)$ , which is defined by the intersection of Brill-Noether schemes $W^{r(n)}_{d+nk}(C)$ via Fitting ideals.
Reduction of Fitting Ideals:
- A major technical hurdle is that the scheme structure of $\Sigma_{\vec{e}}(C, f)$ is defined by an infinite intersection of Fitting ideals.
- Key Technical Step: The author proves (Proposition 2) that locally at certain points, the scheme structure is defined by a finite number of Fitting ideals. This is achieved by restricting the necessary indices $n$ based on the splitting sequence $\vec{e}$ and utilizing an extension of the base point free pencil trick (Lemma 1).
Petri Map Analysis:
- The smoothness of $W^r_d(C)$ at a point $L$ is linked to the injectivity of the Petri map $\mu_L: H^0(L) \otimes H^0(K-L) \to H^0(K)$ .
- The author uses the injectivity of Petri maps for specific line bundles (often related to $L \pm M$ , where $M$ is the $k$ -gonal bundle) to deduce smoothness properties of the ambient schemes.
Inductive Arguments:
- The proofs often proceed by induction on the "distance" between splitting types, reducing complex degeneracy loci to simpler ones where smoothness is already known.

3. Key Contributions and Results

A. Smoothness of Splitting Degeneracy Loci (Theorem C / Theorem 1)

The paper proves that for a general $k$ -gonal curve, the scheme $\Sigma_{\vec{e}}(C, f)$ is smooth along a naturally defined open subset $V^r_{d,l}(C, f)$ .

This subset $V^r_{d,l}(C, f)$ strictly contains the set of line bundles with the exact splitting type $\vec{e}$ .
This result generalizes previous smoothness results (which were limited to specific types of curves or required characteristic 0) to general $k$ -gonal curves in arbitrary characteristic.

B. Characterization of Petri Map Injectivity (Theorem A / Theorem 2)

The author provides a necessary and sufficient condition for the injectivity of the Petri map $\mu_L$ for a line bundle $L$ on a general $k$ -gonal curve:
$\mu_L \text{ is injective} \iff h^0(L - M) \cdot h^0(K - L - M) = 0$

This means the Petri map fails to be injective if and only if $L$ "interacts" with the $k$ -gonal bundle $M$ in both directions (i.e., $L$ has sections after subtracting $M$ , and $K-L$ has sections after subtracting $M$ ).
This result is crucial because the injectivity of the Petri map is equivalent to the smoothness of $W^r_d(C)$ at $L$ (when $L \notin W^{r+1}_d(C)$ ).

C. Smoothness of $W^r_d(C)$ and Component Intersections (Theorem B / Theorem 3)

The paper establishes a precise criterion for the smoothness of the Brill-Noether scheme $W^r_d(C)$ at a point $L$ belonging to a specific component $Z$ :

Theorem: $W^r_d(C)$ is smooth at $L \in Z$ if and only if $L$ belongs to exactly one irreducible component of $W^r_d(C)$ .
Significance: This refines the classical understanding. Even if $L$ is not in $W^{r+1}_d(C)$ , $W^r_d(C)$ can be singular at $L$ if $L$ lies in the intersection of two different components.
Counter-intuitive Finding: The paper provides examples (Proposition 6, Example 7) where $W^r_d(C)$ is singular at a point $L$ even if $L$ belongs to a unique irreducible component. This occurs when the component itself is singular at that point due to the specific geometry of the $k$ -gonal curve (specifically when $h^0(L+M)$ is too large).

D. Classification of Singularities

The paper classifies singular points in $W^r_d(C)$ into two types:

Trivial Singularities: Points lying in the intersection of two or more irreducible components.
Intrinsic Singularities: Points lying in a unique component where the component itself is singular. This happens when the splitting sequence $\vec{e}$ contains specific values (like $-1$ or $-2$) that cause the Petri map to fail injectivity despite the uniqueness of the component.

4. Significance

Refinement of Brill-Noether Theory: The paper moves beyond the "general curve" case to provide a detailed, stratified understanding of the geometry of special divisors on curves with low gonality.
Resolution of Open Questions: It settles questions regarding the smoothness of the schemes $V^r_{d,l}(C, f)$ introduced in recent literature (specifically referencing [10]), proving they are smooth in arbitrary characteristic for general $k$ -gonal curves.
Geometric Insight: By linking the smoothness of the scheme directly to the injectivity of the Petri map and the intersection properties of components, the paper offers a concrete geometric explanation for singularities that were previously less understood.
Technical Advancement: The method of reducing infinite Fitting ideal definitions to finite sets via the "stepwise description" of splitting loci provides a powerful tool for future research in degeneracy loci and moduli spaces of vector bundles.

In summary, Coppens demonstrates that for general $k$ -gonal curves, the smoothness of Brill-Noether schemes is governed by a delicate interplay between the splitting type of the line bundle, the injectivity of the Petri map, and the intersection of irreducible components, providing a complete local description of these schemes.