Proportion of periodic points in reduction of polynomials
Santiago Radi · 2026
Read Proportion of periodic points in reduction of polynomials by Santiago Radi (2026) 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-2603-21620.txt.
Introduction
Radi completes a classification problem in arithmetic dynamics: how many periodic points survive when a polynomial is reduced modulo a prime ideal. A 2014 question by Juul and co-authors asked for the limit of that proportion as the prime grows. Earlier work with Fariña-Asategui settled the Chebyshev-like cases, where the limit is zero. Radi now handles all remaining polynomial cases. The result gives a complete answer for polynomials of degree at least two over number fields.
Catalog record
| CREATOR | Santiago Radi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50364 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.21620
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Proportion of periodic points in reduction of polynomials Thanks: The author is supported by Grigorchuk’s Simons Foundation Grant MP-TSM-00002045 and the department of Mathematics of Texas A&M University. Santiago Radi Abstract . In 2014, Juul, Kurlberg, Madhu and Tucker asked the following: given K K a number field and f f a rational function with coefficients in K K , if f 𝔭 f_{\mathfrak{p}} denotes the reduction of f f modulo a prime ideal 𝔭 \mathfrak{p} in the ring of integers of K K , what is the limit inferior of the proportion of periodic points of f 𝔭 f_{\mathfrak{p}} when the norm of 𝔭 \mathfrak{p} goes to infinity? Recent results of Fariña-Asategui and the author show that when f f is a polynomial of degree d ≥ 2 d\geq 2 non-linearly conjugate over ℂ \mathbb{C} to a Chebyshev polynomial then the limit is zero. In this article, we address the remaining cases to give a complete classification of the problem in the case of polynomials. 2020 Mathematics Subject Classification Primary: 37P25, 37C25; Secondary: 11R45, 37F10, 13A35 1. Introduction Let K be a number field, \mathcal{O}_{K} its ring of integers and f a polynomial with coefficients in K . For all but finitely many prime ideals \mathfrak{p} in \mathcal{O}_{K} , we can consider the reduction map f_{\mathfrak{p}} acting on the residue field \mathbb{F}_{\mathfrak{p}}:=\mathcal{O}_{K}/\mathfrak{p} . As \mathbb{F}_{\mathfrak{p}} is a finite field, a point x\in\mathbb{F}_{\mathfrak{p}} is either periodic or strictly preperiodic for f_{\mathfrak{p}} . How does the proportion of periodic points varies as the norm of the prime \mathfrak{p} goes to infinity? The first problem in this direction was formulated by Schur in 1923, who asked for a classification of all polynomials with coefficients in \mathbb{Q} whose reduction is a bijection for infinitely many primes Schur1923 (note that if the reduction is a bijection, every point must be periodic and consequently the proportion of periodic points is 1 ). The question was answered in 1970 by Fried over any number field K , showing that a polynomial whose reduction is a bijection for infinitely many primes must be a composition of linear polynomials with coefficients in K and Dickson polynomials, where Dickson polynomials are defined by f(x+a/x)=x^{n}+a^{n}/x^{n} with a\in K ; see Fried1970 . In 2002, Guralnick, Müller and Saxl generalized Fried’s result to rational functions over any number field, using the theory of arithmetic and geometric iterated Galois groups GuralnickMullerSaxl2003 . These maps are referred in GuralnickMullerSaxl2003 as arithmetically exceptional. Although we find infinitely many primes where the proportion of periodic points is 1 for arithmetically exceptional maps, the proportion of periodic points for the other primes might be arbitrarily small. In fact, this is what we expect to happen in a random map. In FlajoletOdlyzko1990 , Flajolet and Odlyzko proved that if f:S\rightarrow S is a random map, where S is finite, the number of periodic points is asymptotically \sqrt{\frac{\pi\cdot\#S}{2}} when \#S\rightarrow+\infty . Therefore, based on this heuristic, we would expect to find prime ideals \mathfrak{p} where the reduction f_{\mathfrak{p}} has a proportion of periodic points arbitrarily small. Unfortunately, the assumption that the reductions of a polynomial behave randomly is far from being true. This leads to the following question: Question 1.1 ( ( Juul2014 , Question 1.1) ) . Given K a number field and f\in K[x] a polynomial with coefficients in K . Denote \Per(f,\mathbb{F}_{\mathfrak{p}}):={\left\{\alpha\in\mathbb{F}_{\mathfrak{p}}:\alpha\text{ is periodic for }f_{\mathfrak{p}}\right\}} . What is the value of \displaystyle\Per_{\inf}(f,K):=\liminf_{N(\mathfrak{p})\rightarrow+\infty}\frac{\#\Per(f,\mathbb{F}_{\mathfrak{p}})}{N(\mathfrak{p})}? 1.1 was firstly formulated in 2014 by Juul, Kurlberg, Madhu and Tucker Juul2014 , who asked for a classification of the values of \Per_{\inf}(f,K) for every number field K and any rational function f\in K(x) ; ( Juul2014 , Question 1.4) . In Juul2014 , the authors proved that \Per_{\inf}(x^{d}+c,K)=0 for any d\geq 2 and c\in K such that 0 is not preperiodic, allowing them to give a complete classification of the problem in the case of quadratic polynomials over \mathbb{Q} ; ( Juul2014 , Theorem 1.5) . In ( Juul2014 , Theorem 1.3) , they also proved that if f has two critical points whose forward orbits do not intersect, then \Per_{\inf}(f,K)=0 . It is expected that in general \Per_{\inf}(f,K)=0 and this has been confirmed in most of the examples known. In accordance with this, in ( Juul2014 , Theorem 1.2) authors showed that given K a number field and numbers \epsilon>0 and d\geq 2 , there is an open Zariski dense set U_{d,\epsilon} in the moduli space of rational functions such that \Per_{\inf}(f,K)\leq\epsilon for all f\in U_{d,\epsilon} . The only cases known where \Per_{\inf}(f,K)>0 correspond to the Chebyshev polynomials. Recall that the Chebyshev polynomial of degree d , denoted T_{d} , is defined as the unique polynomial with coefficients in \mathbb{Z} satisfying (1) \displaystyle T_{d}\left(x+\frac{1}{x}\right)=x^{d}+\frac{1}{x^{d}}. In ( Juul2014 , Example 7.2) is proved that if d\geq 2 then (2) \displaystyle\Per_{\inf}(T_{d},\mathbb{Q})=\left\{\begin{matrix}1/2&\mbox{if $d$ is an odd prime power,}\\ 1/4&\mbox{if $d$ is an even prime power,}\\ 0&\mbox{otherwise.}\end{matrix}\right. In this article, we give a complete answer to 1.1 . The cases where f is a polynomial of degree d=0 or 1 are straightforward and will be solved in Lemma 3.1 . When d\geq 2 , we rely on the tools of arboreal representations, particularly, the results related to iterated Galois groups and fixed-point proportion. In 1985, Odoni initiated the study of iterated Galois groups and fixed-point proportion to solve problems in number theory Odoni1980 . In his pioneering article, he developed these tools to prove that the Dirichlet density of prime numbers dividing at least one term in the Sylvester sequence (defined recursively as w_{0}=2 and w_{n+1}=1+w_{1}\dots w_{n} ) is zero. His idea was then generalized to solve other problems in Number theory. Specifically, in 2014 Juul, Kurlberg, Madhu and Tucker proved that (3) \displaystyle\Per_{\inf}(f,K)\leq\FPP(G_{\infty}(\mathbb{C},f,t)), where \FPP is the fixed-point proportion, G_{\infty}(\mathbb{C},f,t) is the iterated Galois group of f over \mathbb{C} and t is transcendental over \mathbb{C} (see Section 3 for the unexplained terms here and Jones2014 for a survey about the topic). The recent breakthrough of Fariña-Asategui and the author completely classifies the fixed-point proportion for geometric iterated Galois groups of polynomials ( Fari~naRadi2026FPP , Theorem 2) . As \mathbb{C} is algebraically closed, the result applies to this case obtaining that \FPP(G_{\infty}(\mathbb{C},f,t))=0 if f is not linearly conjugate over \mathbb{C} to \pm T_{d} . Therefore \Per_{\inf}(f,K)=0 if f is not linearly conjugate over \mathbb{C} to \pm T_{d} . In the case where f is linearly conjugate over \mathbb{C} to \pm T_{d} , then the fixed-point proportion of G_{\infty}(\mathbb{C},f,t) is positive and therefore the upper bound in Equation 3 is not enough to calculate \Per_{\inf}(f,K) . For these cases, we will work out \Per_{\inf}(f,K) directly. We will prove that if f is linearly conjugate over \mathbb{C} to \pm T_{d} then (4) \displaystyle\Per_{\inf}(f,K)=\Per_{\inf}(T_{d},K). Although this step can look obvious, it is not obvious if the conjugation does not have coefficients in K . We shall see in Section 5 that when the conjugation is not in K , it lies in a field extension K^{\prime}/K of degree 2 . If \mathfrak{p} is a prime ideal in \mathcal{O}_{K} that is inert in \mathcal{O}_{K^{\prime}} , then the periodic points of f in \mathbb{F}_{\mathfrak{p}} are in bijection with the periodic points of T_{d} . But if \mathfrak{p} is