Non-simple abelian varieties in a family: arithmetic approaches

Yu Fu · 2024

Read Non-simple abelian varieties in a family: arithmetic approaches by Yu Fu (2024) 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-2412-11048.txt.

Introduction

Yu Fu, writing alone in 2024, works on a question in pure math about abelian varieties. These are shapes defined by equations that also carry a group law. In a family of such shapes that shifts with one dial, the typical member may be simple, meaning it does not split into smaller parts.

Fu asks how far that simple trait spreads to other members of the family. Prior work used analytic tools on a narrow set of curve families. Fu widens the result to all families with large symmetry action, using number-based tools instead. The method leans on bounds for covers of the shapes and tunes those covers to get sharp counts.

Catalog record

CREATORYu Fu
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-11048

Browse the full library · About the Tower of Babel

Full text

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

Non-simple abelian varieties in a family: arithmetic approaches Yu Fu Abstract . Inspired by the work of Ellenberg, Elsholtz, Hall, and Kowalski, we investigate how the property of the generic fiber of a one-parameter family of abelian varieties being geometrically simple extends to other fibers. In [ 6 ] , the authors studied a special case involving specific one-parameter families of Jacobians of curves using analytic methods. We generalize their results, particularly Theorem B, to all families of abelian varieties with big geometric monodromy, employing an arithmetic approach. Our method applies Heath-Brown-type bounds on certain covers with level structures and optimizes the covers to derive the desired results. 1. Introduction Let \mathcal{A}_{g} be the moduli space of principally polarized abelian varieties of dimension g . Let K be a number field and let X\subset\mathcal{A}_{g} be a rational curve over K parameterizing a one-dimensional family of abelian varieties. This is equivalent to saying X is birational to \mathbb{P}^{1} over K . For a K -point x\in X , let A_{x} denote the specialization of this family over x . Let A_{\eta} be the generic fiber of this family. Assume that the geometric monodromy of A_{\eta} modulo \ell is the full symplectic group \operatorname{Sp}_{2g}(\mathbb{F}_{\ell}) for almost all primes \ell . In particular, this implies that A_{\eta} is geometrically simple. We are interested in understanding how the property of being geometrically simple extends to other fibers. To be precise, we ask the following question: Question 1. What can we say about the number of x\in X(K) of bounded height such that A_{x} is geometrically non-simple? For t\in X(K) , let H(t) denote the multiplicative projective height of t . Define S(B) as the set S(B)=\{t\in K\mid H(t)\leq B\text{, and the fiber $A_{t}$ is non-simple.}\} One may ask whether S(B) is finite. If it is finite, can we obtain an upper bound in terms of B ? The finiteness of S(B) comes from the Northcott property in K . Moreover, in [ 6 , Theorem A] , the authors proved that S(B) is "small" and is bounded by a constant depending on f , under certain assumptions on the family that ensure large monodromy. In particular, these assumptions concern a particular type of family of abelian varieties, namely the family A_{f}\to\mathbb{A}^{1} of Jacobians of the hyperelliptic curves defined by affine equations y^{2}=f(x)(x-t)\text{, t $\in$ $\mathbb{A}^{1}$,} for some fixed square-free polynomial f\in\mathbb{Z}[X] of degree 2g , g\geq 1 . The authors of [ 6 ] also provided analytic methods, based on sieve theory, to obtain upper bounds for the number of non-simple varieties with height at most B . Theorem 1.1 . [ 6 , Theorem B] There exist constants C\geqslant 0 and D\geqslant 1 , independent of f , such that we have (1.1) \displaystyle|S(B)|\leqslant C\left(g^{2}D(\log 2B)\right)^{11g^{2}} for all B\geqslant 1 . However, an issue lies in the proof of [ 6 , Theorem 24] , where they did not take into account the possibility that the reduction of the squarefree polynomial f modulo a prime ideal might not be squarefree. So in which case they cannot appeal to [ 6 , Proposition 22] , which causes the constant C and D in 1.1 to depend on the largest prime that divides the discriminant of f and the genus of the family. In this article, we aim to generalize these results to arbitrary rational one-dimensional families of abelian varieties over a number field, without knowing the explicit defining equation of the family. Notably, compared to [ 6 , Theorem B] , the logarithmic terms in our theorem do not depend on g or K . To state our theorem, we need the following definitions. Definition 1.2 . Let n be the dimension of \mathcal{A}_{g} over K and let N=n+1 . Define \iota:X\to\mathbb{P}^{N} to be the restriction of the Hodge embedding on \mathcal{A}_{g} to X , which is a birational map from \mathbb{P}^{1} to its image (see Section 2.1 for details). We see that specifying the image of \iota is the same as specifying the family X . In the definition of S(B) , H(t) is defined as the height of t\in X(K) as an element of K . Let P_{t}=\iota(t) . Alternatively, we can view t as an element of \mathbb{P}^{N} and define its height as the multiplicative height of P_{t} . In this perspective, we introduce the notion of S_{X}(B) as follows: Definition 1.3 . Define S_{X}(B) as the set S_{X}(B)=\{t\in K\mid H(P_{t})\leq B\text{, and the fiber $A_{t}$ is non-simple}\}. Our main theorem is as follows. Theorem 1.4 . Let K be a number field of degree d_{K} over \mathbb{Q} . Let X\subset\mathcal{A}_{g} be a rational curve over K parameterizing a one-dimensional family of abelian varieties. Let A_{\eta} be the generic fiber of this family and assume that the geometric monodromy of A_{\eta} modulo \ell is the full symplectic group \operatorname{Sp}_{2g}(\mathbb{F}_{\ell}) for almost all primes \ell . Let \iota be the map defined in Definition 1.2 and let d denote the degree of X with respect to \iota . Then there exists an absolute constant C_{K,g} depending on K and g , B_{0} depending on X , g and \iota , and an absolute constant \kappa such that (1.2) \displaystyle|S_{X}(B)|\leq C_{K,g,\iota}d^{\kappa}(\log B)^{\kappa} for all B\geq B_{0} . If we change the height of t from H(P_{t}) to H(t) and assume that H(t)\leq B , then there exists a constant C_{K,g,\iota} depends on d_{K} g and \iota , B_{0} depends on X and g , and an absolute constant \kappa such that (1.3) \displaystyle|S(B)|\leq C_{K,g,\iota}d^{\kappa}(\log B)^{\kappa} for all B\geq B_{0}. Remark 1.5 . The bound in Theorem 1.4 is uniform in the sense that the constant C_{K,g,\iota} in the bound depends only on the embedding and the degree of the family. The method does not require the curve X to have a specific form or be defined by specific equations. The approach also makes it easier to work in generality than in [ 6 ] . Being non-simple is a rare phenomenon. The theorem shows that most of the fibers carry the generic property of being simple, while the non-simple fibers exhibit subpolynomial growth with respect to height. Indeed, a general point in \mathcal{A}_{g} has Picard number 1 , and being non-simple implies that the Picard number is at least 2 , therefore the abelian variety lies in the Noether-Lefschetz locus, which is a countable union of closed and reduced subvarieties of \mathcal{A}_{g} (see [ 4 ] ). Considering S(B) as the intersection of a rational curve with \mathcal{A}_{g} , our result implies that most intersection points are irrational. The strategy for proving Theorem 1.4 is totally different from [ 6 ] . Although their approach relies on analytic number theory, our methods are arithmetic in nature. Specifically, we apply Heath-Brown-type bounds, which was first proved by Heath-Brown in [ 11 ] and then improved by a series of works [ 21 , 2 , 16 , 1 ] , on certain covers with level structure and optimize these covers. More precisely, we will make good use of Theorem 2.2 . This approach, used in previous work by the author [ 8 ] and Ellenberg-Lawrence-Venkatesh [ 5 ] , remains underexplored. We hope that

Read the complete text