Table of Contents
Fetching ...

Hodge loci associated with linear subspaces intersecting in codimension one

Remke Kloosterman

TL;DR

The paper investigates when the Noether-Lefschetz loci NL(Pi1,Pi2) and NL(Pi1+λPi2) coincide near a smooth hypersurface X in P^{2k+1} containing two k-planes intersecting in codimension one. It translates Hodge-type conditions into tangent-space computations via the Jacobian ring S/J and the associated ideal I(γ), and analyzes the two-class deformation [Y1]+λ[Y2] using left kernels and Gram-matrix methods (Theorem thmTsp). The results separate the cases k=1 and k≥2: for k=1 and d≥5, NL(Pi1+λPi2) is nonreduced for all λ≠0,1, while for k≥2 there exist X for which NL(Pi1+λPi2)=NL(Pi1,Pi2) for all but finitely many λ (and potentially {0,1} when d≥6), with generic X yielding equality near X for λ≠0,1. The paper also provides explicit constructions to demonstrate the method and discusses embedded components arising from special split configurations, thereby clarifying when Noether-Lefschetz phenomena persist or degenerate in higher codimension.

Abstract

Let $X\subset \mathbb{P}^{2k+1}$ be a smooth hypersurface containing two k-dimensional linear spaces $Π_1,Π_2$ intersecting in codimension one. In this paper we study the question whether the Hodge loci $NL([Π_1]+λ[Π_2])$ and $NL([Π_1],[Π_2])$ coincide. This turns out to be the case in a neighborhood of $X$ if $X$ is very general on $NL([Π_1],[Π_2])$, $k>1$ and $λ\neq 0,1$. However, there exists a hypersurface $X$ for which $NL([Π_1],[Π_2])$ is smooth at $X$, but $NL([Π_1]+λ[Π_2])$ is singular for all $λ\neq0,1$. We expect that this is due to an embedded component of $NL([Π_1]+λ[Π_2])$. The case $k=1$ was treated before by Dan, in that case $NL([Π_1]+λ[Π_2])$ is nonreduced.

Hodge loci associated with linear subspaces intersecting in codimension one

TL;DR

The paper investigates when the Noether-Lefschetz loci NL(Pi1,Pi2) and NL(Pi1+λPi2) coincide near a smooth hypersurface X in P^{2k+1} containing two k-planes intersecting in codimension one. It translates Hodge-type conditions into tangent-space computations via the Jacobian ring S/J and the associated ideal I(γ), and analyzes the two-class deformation [Y1]+λ[Y2] using left kernels and Gram-matrix methods (Theorem thmTsp). The results separate the cases k=1 and k≥2: for k=1 and d≥5, NL(Pi1+λPi2) is nonreduced for all λ≠0,1, while for k≥2 there exist X for which NL(Pi1+λPi2)=NL(Pi1,Pi2) for all but finitely many λ (and potentially {0,1} when d≥6), with generic X yielding equality near X for λ≠0,1. The paper also provides explicit constructions to demonstrate the method and discusses embedded components arising from special split configurations, thereby clarifying when Noether-Lefschetz phenomena persist or degenerate in higher codimension.

Abstract

Let be a smooth hypersurface containing two k-dimensional linear spaces intersecting in codimension one. In this paper we study the question whether the Hodge loci and coincide. This turns out to be the case in a neighborhood of if is very general on , and . However, there exists a hypersurface for which is smooth at , but is singular for all . We expect that this is due to an embedded component of . The case was treated before by Dan, in that case is nonreduced.
Paper Structure (4 sections, 12 theorems, 32 equations)

This paper contains 4 sections, 12 theorems, 32 equations.

Key Result

Theorem 1.1

Let $d\geq 3$ be an integer. Moreover, suppose that $k>4$ if $d=3$, that $k>2$ if $d=4,5$ and $k\geq 2$ if $d\geq 6$. Then there exists a hypersurface $X\subset \mathbf{P}^{2k+1}$ of degree $d$ containing two $k$-planes $\Pi_1,\Pi_2$ such that $\dim \Pi_1\cap \Pi_2=k-1$, and a finite set $\Sigma$ su

Theorems & Definitions (33)

  • Theorem 1.1
  • Remark 2.2
  • Example 2.3
  • Definition 2.4
  • Remark 2.5
  • Lemma 2.6
  • proof
  • Definition 2.7
  • Remark 2.8
  • Proposition 2.9
  • ...and 23 more