A Brill-Noether theorem for (toric) surfaces

Alessio Cela and Carl Lian · 2025

Read A Brill-Noether theorem for (toric) surfaces by Alessio Cela and Carl Lian (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-18905.txt.

Introduction

Alessio Cela and Carl Lian carry a classic theorem about curves into the world of surfaces, in this 2025 mathematics paper. The Brill-Noether theorem says a general curve's maps into space come in families of the expected size, unless the image is trapped in a flat slice. The authors observe that standard results on Severi varieties yield the same statement for maps onto any smooth projective surface.

The condition becomes a degree requirement: the image curve must score at least four against the surface's anti canonical measure. For toric surfaces, the borderline cases, degrees three and below, get an elegant description through special contractions of the surface. The authors close by asking whether the theorem extends to toric varieties in higher dimensions, where the geometry is richer and the counting harder.

Catalog record

CREATORAlessio Cela and Carl Lian
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-18905

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Brill-Noether theorem for (toric) surfaces Alessio Cela and Carl Lian Abstract . The classical Brill-Noether theorem states that a map from a general curve to a projective space deforms in a family of expected dimension as long as its image does not lie in any hyperplane. In this note, we observe, as a direct consequence of standard results on Severi varieties, an analogous statement for maps from a general curve to any smooth, projective surface. Namely, a non-constant map deforms in a family of expected dimension as long as its image has anti-canonical degree at least 4. In the case of toric surfaces, curves of anti-canonical degree at most 3 admit a particularly elegant description in terms certain toric contractions. We raise the question of whether a Brill-Noether theorem could hold for toric varieties of higher dimension. 1. Introduction Let C be a general curve over \mathbb{C} , by which we mean a general point of {\mathcal{M}}_{g} . Classical Brill-Noether theory concerns maps from C to projective spaces. One formulation of the Brill-Noether theorem is the following. Theorem 1.1 . Suppose that f:C\to\mathbb{P}^{r} is a morphism of degree d which is non-degenerate , that is, the image of f does not lie in any hyperplane. Then, the map f moves in a family of expected dimension. That is, the moduli space of maps {\mathcal{M}}_{d}(C,\mathbb{P}^{r}) of maps C\to\mathbb{P}^{r} of degree d is pure of dimension (r+1)d+r(1-g), the expected, in a neighborhood of f . In fact, the Gieseker-Petri theorem asserts further that the moduli space of maps is smooth at f . Indeed, a map f whose image does not lie in a hyperplane is given, up to automorphisms on \mathbb{P}^{r} , by a base-point free, rank r linear series, that is, a line bundle \mathcal{L} on C and a (r+1) -dimensional space of sections V\subset H^{0}(C,\mathcal{L}) . The moduli space \mathcal{G}^{r}_{d}(C) of rank r , degree d linear series is smooth of dimension \rho(g,r,d)=g-(r+1)(g-d+r) [ ACGH85 , Theorem V.1.6] . In this note, we prove the following analogous statement for maps to a smooth, projective surface. Theorem 1.2 . Let S be a smooth, projective surface. Let f:C\to S be a morphism with f_{*}[C]=\beta\neq 0 , and let C^{\prime}=f(C) be its scheme-theoretic image. Suppose that [C^{\prime}]\cdot K_{S}^{\vee}\geq 4 . Then, the map f moves in a family of expected dimension. That is, the moduli space of maps {\mathcal{M}}_{\beta}(C,S) is pure of dimension \beta\cdot K^{\vee}_{S}+2(1-g), the expected, in a neighborhood of f . The moduli space {\mathcal{M}}_{\beta}(C,S) is furthermore generically smooth in a neighborhood of f . When S=\mathbb{P}^{2} , the curves of anti-canonical degree at most 3 are exactly lines, so Theorem 1.2 recovers the classical Brill-Noether theorem when r=2 . Theorem 1.2 is proven in § 2 . As we will see, it is a consequence of well-known transversality results for Severi varieties of embedded curves on S ; the work is to compare dimensions for moduli of maps which may or may not be birational onto their image. This comparison is also standard, see for example [ Vak00 , §3] . If S is a toric surface, then the loci on {\mathcal{M}}_{\beta}(C,S) that must be removed in order to obtain a moduli space that is pure of expected dimension, namely the loci of multiple covers of curves of anti-canonical degree at most 3, admit a particularly simple description. The following is proven in § 3 . Theorem 1.3 . Let S be a smooth, projective toric surface. Let C^{\prime}\subset S be an integral curve with [C^{\prime}]\cdot K_{S}^{\vee}\leq 3 , and assume that C^{\prime} is not a component of the toric boundary. Then, one of the following holds: (i) [C^{\prime}]\cdot K_{S}^{\vee}=2 , in which case there exists a toric morphism \pi:S\to\mathbb{P}^{1} contracting C^{\prime} to a point, or (ii) [C^{\prime}]\cdot K_{S}^{\vee}=3 , in which case there exists a fake toric projective plane S^{\prime} (Definition 3.1 ) and a birational toric morphism \pi:S\to S^{\prime} induced by deletion of rays of the fan of S , such that \pi(C^{\prime}) lies in the primitive linear system of S^{\prime} . We suspect that Theorems 1.2 and 1.3 are known in some form to experts, but we were not able to locate these precise statements in the literature. Combining these statements, one obtains a “Brill-Noether theorem” for toric surfaces. Corollary 1.4 . Let S be a smooth, projective toric surface. Let f:C\to S be a morphism with f_{*}[C]=\beta\neq 0 , and let C^{\prime}=f(C) be its scheme-theoretic image. Assume that C^{\prime} is not contained in the toric boundary of S , and furthermore that C^{\prime} satisfies neither (i) nor (ii) of Theorem 1.3 . Then, the map f moves in a family of expected dimension. The moduli space {\mathcal{M}}_{\beta}(C,S) is furthermore generically smooth in a neighborhood of f . In other words, if f fails to move in a family of expected dimension, then the failure is explained by the degeneracy of f after post-composing with a certain toric morphism to either a projective line or to a fake toric projective plane. Question 1.5 . Does Corollary 1.4 admit a generalization to smooth, projective toric varieties of any dimension? Classical Brill-Noether theory concerns sections of line bundles on curves, and maps to toric varieties are also governed by sections of line bundles on the source [ Cox95 ] , so this seems to be a natural generalized setting for Brill-Noether theory. While the methods of this note are highly specific to the case of surfaces, some hope that an extension to higher dimension could be possible is afforded by the following result of Farkas. Theorem 1.6 . [ Far08 , Theorem 0.1] Let b:Y\to\mathbb{P}^{r} be the blow-up along a (torus-invariant) linear space. Let f:C\to Y be a morphism with f_{*}[C]=\beta . Suppose that b(f(C))\subset\mathbb{P}^{r} is not contained in any hyperplane. Then, the moduli space of maps {\mathcal{M}}_{\beta}(C,Y) is pure of dimension c_{1}(K_{Y}^{\vee})\cdot\beta+r(1-g), the expected, in a neighborhood of f . Farkas’s result is stated in the language of linear series with secants, but the formulation above is equivalent. See also [ Cot11 , Ung21 , Far23 ] for closely related results. Recall that the Gieseker-Petri theorem asserts further that the moduli space of non-degenerate maps from a general curve to \mathbb{P}^{r} is smooth. Equivalently, if f:C\to\mathbb{P}^{r} is non-degenerate, then it is unobstructed, that is, we have H^{1}(C,f^{*}T_{\mathbb{P}^{r}})=0 . We do not know whether this remains the case in the settings of Corollary 1.4 and Theorem 1.6 . Question 1.7 . Let f:C\to S (resp. f:C\to Y ) be a map as in Corollary 1.4 (resp. Theorem 1.6 ). Is the moduli space {\mathcal{M}}_{\beta}(C,S) (resp. {\mathcal{M}}_{\beta}(C,Y) ) smooth at f ? Whereas Corollary 1.4 gives that {\mathcal{M}}_{\beta}(C,S) is generically smooth in a neighborhood of f , generic smoothness of {\mathcal{M}}_{\beta}(C,Y) in the setting of Theorem 1.6 does not seem to follow directly from Farkas’s proof. Note that Questions 1.5 and 1.7 are only interesting when C has positive genus. Indeed, a rational curve generically contained in the interior of a smooth, projective toric variety is free, and therefore unobstructed. (Rational curves contained in the boundary are more naturally viewed as maps to the appropriate toric stratum.) 1.1. Acknowledgments We thank Renzo Cavalieri, Gavril Farkas, François Greer,

Read the complete text