Cubic points on dynamical modular curves

John R. Doyle, Alexander Galarraga · 2025

Read Cubic points on dynamical modular curves by John R. Doyle, Alexander Galarraga (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-2511-11491.txt.

Introduction

John R. Doyle and Alexander Galarraga classify which dynamical modular curves tied to quadratic polynomial maps carry infinitely many cubic points. These curves package questions about repeated iteration of quadratic polynomials, a core topic in arithmetic dynamics.

Settling which curves have infinitely many cubic points yields a classification of preperiodic points, points that eventually repeat under iteration, for quadratic polynomials over cubic number fields. The result extends earlier work by Poonen, Faber, and Doyle with Krumm. For number theorists, a complete list of this kind closes a case-by-case problem and sharpens what is known about polynomial dynamics over small-degree fields.

Catalog record

CREATORJohn R. Doyle, Alexander Galarraga
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE95 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-11491

Browse the full library · About the Tower of Babel

Full text

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

Cubic points on dynamical modular curves John R. Doyle, Alexander Galarraga Abstract . We consider the family of dynamical modular curves associated to quadratic polynomial maps and determine precisely which of these curves have infinitely many cubic points. We use this to prove a classification statement on preperiodic points for quadratic polynomials over cubic fields, extending previous work of Poonen, Faber, and the first author and Krumm. 2020 Mathematics Subject Classification Primary 37P15, 37P35; Secondary 11G30, 14G05 1. Introduction Let K be a number field, let \overline{K} be its algebraic closure, and let f(x)\in K(x) be a rational function. Write f^{0} for the identity map on \mathbb{P}^{1} , and for each n\geq 1 set f^{n}:=f\circ f^{n-1} . We say that a point P\in\mathbb{P}^{1}(\overline{K}) is periodic for f if there exists n\geq 1 such that f^{n}(P)=P ; the minimal such n is the ( exact ) period of P . More generally, we say that P is preperiodic for f if there exist integers m\geq 0 and n\geq 1 such that f^{m+n}(P)=f^{m}(P) ; the minimal such m and n are the preperiod and eventual period of P , respectively. The preperiodic portrait for f (over K ) is the functional graph G(f,K) of the restriction of f to the set of K -rational preperiodic points for f , that is, the directed graph whose vertices are the K -rational preperiodic points for f and whose edges are all ordered pairs of the form (P,f(P)) . Throughout, we will use the term “portrait” to refer to a functional graph on a finite set. Northcott [ northcott:1950 ] showed that if f has degree at least 2 , then f has finitely many K -rational preperiodic points; in particular, the preperiodic portrait G(f,K) is a finite graph. Based on an analogy between preperiodic points for rational maps and torsion points on elliptic curves, Morton and Silverman posed the following dynamical analogue of the strong uniform boundedness conjecture for torsion points on elliptic curves (now a theorem of Merel [ merel:1996 ] ). Conjecture 1.1 (Morton-Silverman [ morton/silverman:1994 , p. 100] ) . Fix integers d\geq 1 and D\geq 2 . There exists a constant B=B(d,D) such that if K is a number field of absolute degree d and f(x)\in K(x) is a rational function of degree D , then the number of K -rational preperiodic points for f is at most B . Conjecture 1.1 is completely open—it is not known for any values of d or D —though we mention that Looper [ looper:2021 , looper ] has proven a version of the uniform bounedness conjecture for polynomials assuming a stronger variant of the abc conjecture. To emphasize the difficulty of Conjecture 1.1 , we note that just the case D=4 would suffice to prove Merel’s theorem; see [ silverman:2007 , §3.3] . 1.1. Preperiodic portraits for quadratic polynomials Perhaps the most well-studied family of rational functions in complex and arithmetic dynamics is that of quadratic polynomials. We say that two rational functions f and g are linearly conjugate over K if there exists \gamma\in\operatorname{PGL}_{2}(K) such g=\gamma^{-1}\circ f\circ\gamma . This is a natural notion of equivalence in dynamics since conjugation commutes with iteration; in particular, if f,g\in K(x) are linearly conjugate over K , then the portraits G(f,K) and G(g,K) are isomorphic. Every quadratic polynomial in K[x] is linearly conjugate over K to a unique polynomial of the form f_{c}(x):=x^{2}+c with c\in K ; see [ silverman:2007 , §4.2] , for example. Thus, the space of quadratic polynomials up to linear conjugation is a one-parameter family, which we view as analogous to the fact that elliptic curves up to isomorphism form a one-parameter family (parametrized by the j -invariant). Conjecture 1.1 asserts that the number of K -rational preperiodic points for a quadratic polynomial should be bounded above by a constant depending only on [K:\mathbb{Q}] . We rephrase this in terms of preperiodic portraits. Conjecture 1.2 . Fix an integer d\geq 1 . There exists a (minimal) finite set \overline{\Gamma}=\overline{\Gamma}(d) of portraits such that for every number field K of absolute degree d and every quadratic f(x)\in K[x] , the preperiodic portrait G(f,K) is isomorphic to a portrait in \overline{\Gamma} . Thus, the problem is to determine for which portraits \mathcal{P} there exists a degree- d number field K and a quadratic polynomial f(x)\in K[x] for which G(f,K)\cong\mathcal{P} . For the d=1 case, Poonen [ poonen:1998 ] has shown that if there are no quadratic polynomials over \mathbb{Q} with rational points of period larger than 3 , then \overline{\Gamma}(1) consists of exactly the twelve portraits labelled \rm\emptyset,\ 2(1),\ 3(1,1),\ 3(2),\ 4(1,1),\ 4(2),\ 5(1,1)a,\ 6(1,1),\ 6(2),\ 6(3),\ 8(2,1,1),\text{ and }8(3) in Appendix A . It is not known whether there exists a quadratic polynomial over \mathbb{Q} with rational points of period larger than 3 ; however, it is known that a quadratic polynomial in \mathbb{Q}[x] cannot have rational points of period 4 [ morton:1998 ] , period 5 [ flynn/poonen/schaefer:1997 ] , or, assuming the Birch and Swinnerton-Dyer conjecture for a certain abelian fourfold, period 6 [ stoll:2008 ] . Further computational evidence [ hutz/ingram:2013 ] suggests that 3 the largest possible period arising in this setting. For the d=2 case, based on a large amount of computational evidence together with techniques similar to those used in [ poonen:1998 ] , it was suggested in [ doyle/faber/krumm:2014 , p. 514] and explicitly conjectured in [ doyle:2018quad , Conjecture 1.4] that \overline{\Gamma}(2) consists of precisely 46 portraits, which can be found in [ doyle/faber/krumm:2014 , Appendix B] . 1.2. Portraits realized infinitely often Given d\geq 1 , a complete classification of those preperiodic portraits that can be realized by quadratic polynomials defined over number fields of degree d —that is, determining the set \overline{\Gamma}(d) —seems to be quite difficult. Instead, we consider the following variant: Informally, given d\geq 1 , which preperiodic portraits can be realized infinitely often by quadratic polynomials over degree- d number fields? As phrased, this is not quite the right question. Indeed, any preperiodic portrait realized by even one quadratic polynomial over K will be realized infinitely often, since linearly conjugate maps have isomorphic preperiodic portraits. Thus, we wish to know which portraits can be realized by infinitely many equivalence classes of quadratic polynomials over degree- d number fields. For this problem, the analogue of Conjecture 1.2 is known. The following statement was proven in [ doyle/krumm:2024 ] using the main results of [ doyle/poonen:2020 ] . For d\geq 1 , we define \mathbb{Q}^{(d)}:=\{\alpha\in\overline{\mathbb{Q}}:[\mathbb{Q}(\alpha):\mathbb{Q}]\leq d\} to be the set of algebraic numbers of degree at most d . Theorem 1.3 ( [ doyle/krumm:2024 , Theorem 1.11] ) . Fix an integer d\geq 1 . There exists a finite set \Gamma=\Gamma(d) of portraits such that, for every portrait \mathcal{P} , the following are equivalent: (a) There are infinitely many c\in\mathbb{Q}^{(d)} such that G(f_{c},K)\cong\mathcal{P} for some degree- d number field K containing c . (b) The portrait \mathcal{P} is isomorphic to a portrait in \Gamma . The sets \Gamma(1) and \Gamma(2) are known. Again referring to the labels of portraits in Appendix A , Faber showed in [ faber:2015 ] that \Gamma(1) consists of the seven portraits \emptyset , 4(1,1), 4(2), 6(1,1),

Read the complete text