Quadratic counts of highly tangent lines to hypersurfaces

Stephen McKean, Giosuè Muratore, Wern Juin Gabriel Ong · 2025

Read Quadratic counts of highly tangent lines to hypersurfaces by Stephen McKean, Giosuè Muratore, Wern Juin Gabriel Ong (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-2504-21664.txt.

Introduction

Stephen McKean, Giosuè Muratore, and Wern Juin Gabriel Ong wrote a math paper in 2025 about lines that touch a hypersurface very closely at a single point. For such highly tangent lines they study the local type. A piece of information that feeds into a count valued in quadratic forms and tied to an Euler number that depends on a chosen orientation.

The authors give two shape-based readings of that local type. One uses the Wronski map, and the other uses the basic forms of the hypersurface itself. The two views connect an algebra-based counting problem with the surface's own geometry.

Catalog record

CREATORStephen McKean, Giosuè Muratore, Wern Juin Gabriel Ong
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-21664

Browse the full library · About the Tower of Babel

Full text

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

Quadratic counts of highly tangent lines to hypersurfaces Stephen McKean, Giosuè Muratore, Wern Juin Gabriel Ong Abstract . We give two geometric interpretations for the local type of a line that is highly tangent to a hypersurface in a single point. One interpretation is phrased in terms of the Wronski map, while the other interpretation relates to the fundamental forms of the hypersurface. These local types are the local contributions of an quadratic form-valued Euler number that depends on a choice of orientation. 2020 Mathematics Subject Classification Primary: 14N10. Secondary: 14G27. 1. Introduction Over an algebraically closed field, the number of roots of a polynomial, when counted with multiplicity, is given by its degree. In contrast, the number of real roots is not determined by the degree. In order to give a count of real roots that depends only on the degree of the polynomial, the roots must be counted with a sign rather than a multiplicity (see Figure 1 ). - + + - + - + - - 0 + Figure 1. Signed counts of real roots For polynomials of even degree, this signed count of real roots is always 0. For odd polynomials, the signed count is \pm 1 . This sign ambiguity can be interpreted as an artifact of non-orientability: the roots of a polynomial correspond to the intersection of two plane curves, and Bézout’s theorem is not an orientable problem over non-closed fields [ McK21 ] . By highly tangent line we mean a line meeting a hypersurface in \mathbb{P}^{n} at a point with contact order 2n-1 . In this article, we will investigate the geometric weight with which to count highly tangent lines to the hypersurface. Analogous to the case of real roots of odd polynomials, this enumerative problem is not orientable in many cases of interest, so the total count of highly tangent lines depends on the choice of hypersurface. Despite the lack of invariant total count, we find that this counting weight admits two interesting descriptions. Theorem 1.1 (Wronskian interpretation) . Let k be a field. Let X=\mathbb{V}(F) by a smooth, general hypersurface in \mathbb{P}^{n}_{k} of degree d\geq 2n-1 . Let \Phi be the flag variety of pointed lines in \mathbb{P}^{n}_{k} , and let \beta\colon\Phi\to\mathbb{P}^{n}_{k} send a pointed line to its marked point. Let L be a line such that the intersection multiplicity of L and X at p is 2n-1 . Then the local index associated to L is given by the Wronskian determinant (with respect to the local parameter of L at p ) of the gradient \nabla\beta^{*}F . Theorem 1.2 (Fundamental form interpretation) . Let X be a smooth, general, projective plane curve, and assume \mathrm{char}(k)\neq 2,3 . Let \inf(X) denote the locus of inflectional points on X . Let II and III denote the second and third fundamental forms of X , respectively. For each p\in\inf(X) , we have \mathrm{ind}_{p}\makebox[8.71106pt]{\hbox to0.0pt{\vrule width=8.71106pt,height=0.3014pt,depth=0.3014pt\hss}\hbox to0.0pt{\vrule width=8.71106pt,height=7.13472pt,depth=-6.5319pt\hss}\kern 0.5pt{II}\kern 0.5pt}=\mathrm{Tr}_{k(p)/k}\langle 3\cdot\makebox[12.56659pt]{\hbox to0.0pt{\vrule width=12.56659pt,height=0.3014pt,depth=0.3014pt\hss}\hbox to0.0pt{\vrule width=12.56659pt,height=7.13472pt,depth=-6.5319pt\hss}\kern 0.5pt{III}\kern 0.5pt}(p)\rangle. Moreover, if X is of even degree d and admits a theta characteristic, then \sum_{p\in\inf(X)}\mathrm{Tr}_{k(p)/k}\langle\makebox[12.56659pt]{\hbox to0.0pt{\vrule width=12.56659pt,height=0.3014pt,depth=0.3014pt\hss}\hbox to0.0pt{\vrule width=12.56659pt,height=7.13472pt,depth=-6.5319pt\hss}\kern 0.5pt{III}\kern 0.5pt}(p)\rangle=\frac{3d(d-2)}{2}\mathbb{H}, where \mathbb{H}:=\langle 1\rangle+\langle-1\rangle . We will make these theorems more precise in Sections 4 and 5 , respectively. Our work fits into the enriched enumerative geometry program, which consists of quadratic form-valued counts of geometric objects. These counts are usually valid over more general base fields than just algebraically closed or real fields. Quadratic forms arise as a result of the tools necessary to work over arbitrary fields — motivic homotopy theory provides suitable substitutes for Euler numbers and the local Brouwer degree, but these invariants are valued in the Grothendieck–Witt ring of the base field rather than in \mathbb{Z} . 1.1. Outline The enumerative problem of counting highly tangent (projective) linear spaces to hypersurfaces can be formulated in terms of a bundle of principal parts over a flag variety of pointed linear spaces (see e.g. [ EH16 , §11.1] ). In Section 2 , we will recall the construction of this flag variety, which serves as the parameter space of objects that we will count. We will also discuss standard coordinates on the flag variety. In Section 3 , we will recall the bundle of principal parts, as well as convenient local trivializations of this bundle. We will also make a few remarks about Euler numbers in this section. As previously mentioned, the problem we are treating is not orientable. To resolve this, we need to choose an orienting divisor in our parameter space, which we do in Section 3.1 . We then prove Theorem 1.1 in Section 4 . An alternative formulation of the problem of counting highly tangent lines involves fundamental forms. An advantage of this formulation is that the problem is relatively orientable in cases where the bundle of relative principal parts over the flag variety is not relatively orientable. We will set up the necessary background and prove Theorem 1.2 in Section 5 . Note that most computations of local indices in the enriched enumerative geometry literature involve a parameter space whose Nisnevich coordinates are actually isomorphisms, because the parameter space is covered by affine spaces. In contrast, our computation of the relevant Euler class in the second approach uses non-trivial Nisnevich coordinates. So our second approach adds a bit of novelty to the literature. 2. The parameter space Consider the Grassmannian \mathbb{G}\mathrm{r}(r,n) of projective r -planes in \mathbb{P}^{n} , which can also be thought of as the space of affine (r+1) -planes through the origin in \mathbb{A}^{n+1} . Our parameter space is a flag variety that naturally occurs as an incidence variety over \mathbb{G}\mathrm{r}(r,n) . Definition 2.1 . Define the flag variety of pointed planes as the incidence variety \Phi_{r,n}:=\{(H,p)\in\mathbb{G}\mathrm{r}(r,n)\times\mathbb{P}^{n}:p\in H\}. Note that the variety \Phi_{r,n} is isomorphic to the projective bundle \pi\colon\mathbb{P}\mathcal{S}\to\mathbb{G}(r,n) , where \mathcal{S}\to\mathbb{G}(r,n) is the tautological bundle over the affine Grassmannian. We will also use \pi when discussing the pair of projections {\lx@inpgf@ignorespaces\Phi_{r,n}} {\lx@inpgf@ignorespaces\mathbb{G}\mathrm{r}(r,n)} {\lx@inpgf@ignorespaces\mathbb{P}^{n},} \scriptstyle{\lx@inpgf@ignorespaces\pi} \scriptstyle{\lx@inpgf@ignorespaces\beta} which should not cause confusion as \Phi_{r,n}\cong\mathbb{P}\mathcal{S} . Before discussing coordinates for \Phi_{r,n} , we mention a few facts about \Phi_{r,n} that we will need later. All of these facts are standard computations for projective bundles. Lemma 2.2 . We have \dim\Phi_{r,n}=r(n-r)+n . Proof. Since \Phi_{r,n}\cong\mathbb{P}\mathcal{S} , we can compute its dimension as a projective bundle over \mathbb{G}(r,n) . This gives us \displaystyle\dim\mathbb{P}\mathcal{S} \displaystyle=\dim\mathbb{G}(r,n)+\operatorname{rank}\mathcal{S}-1 \displaystyle=(r+1)(n-r)+r+1-1 \displaystyle=r(n-r)+n.\qed We fix the following notation for the subsequent three lemmas. Notation 2.3 . Let \Phi:=\Phi_{r,n} and G:=\mathbb{G}(r,n) . Let \Omega_{\Phi/G} denote the relative cotangent bundle of \pi\colon\Phi\to G . Let \omega_{X} denote the canonical bundle of a scheme X . Lemma 2.4 . We have \omega_{\Phi}\cong\mathcal{O}_{\Phi}(-r-1)\otimes\pi^{*}\mathcal{O}_{G}(-n) . Proof. For any vector bundle V\to X of rank v with associated projective bundle \pi\colon\mathbb{P}V\to X , there is an isomorphism \omega_{\mathbb{P}V}\cong\mathcal{O}_{\mathbb{P}V}(-v)\otimes\pi^{*}\det V^{\vee}\otimes\pi^{*}\omega_{X}. The result now follows from \det\mathcal{S}\cong\mathcal{O}_{G}(-1) and \omega_{G}\cong\mathcal{O}_{G}(-n-1) , which are standard facts about Grassmannians and their tautological bundles. ∎ Lemma 2.5 . We have \operatorname{rank}\Omega_{\Phi/G}=r . Proof. Since \Phi and G are both smooth, we have \displaystyle\operatorname{rank}\Omega_{\Phi/G} \displaystyle=\operatorname{rank}\Omega_{\Phi}-\operatorname{rank}\Omega_{G} \displaystyle=\dim\Phi-\dim G \displaystyle=r(n-r)+n-(r+1)(n-r) \displaystyle=r.\qed Lemma 2.6 . We have \det\Omega_{\Phi/G}\cong\mathcal{O}_{\Phi}(-r-1)\otimes\pi^{*}\mathcal{O}_{G}(1) . Proof. By the relative Euler sequence 0\to\mathrm{T}_{\Phi/G}\to\mathrm{T}_{\Phi}\to\pi^{*}\mathrm{T}_{G}\to 0, we have \det\mathrm{T}_{\Phi/G}\cong\det\mathrm{T}_{\Phi}\otimes\pi^{*}\det\mathrm{T}^{\vee}_{G} . Taking duals and applying Lemma 2.4 , we find \displaystyle\det\Omega_{\Phi/G}

Read the complete text