Polynomials assuming only local prime powers

Przemysław Koprowski · 2025

Read Polynomials assuming only local prime powers by Przemysław Koprowski (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-2511-08100.txt.

Introduction

Przemyslaw Koprowski examines polynomials that send a local field into the subset of its prime power values, where the prime is the residue characteristic of that field. A local field here is a completion of a number field at a non Archimedean place. Koprowski characterizes exactly which polynomials have this mapping property and shows the class is always much broader than simple prime powers of polynomials themselves.

Catalog record

CREATORPrzemysław Koprowski
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-08100

Browse the full library · About the Tower of Babel

Full text

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

Polynomials assuming only local prime powers Przemysław Koprowski Abstract . We study the class of polynomials that map a local field (i.e., the completion of a number field at a non-Archimedean place) into the subset of its p p -th powers, where p p is the residue characteristic of the field in question. We present a characterization of such polynomials and show that this class is always much broader than the class of p p -th powers of polynomials. 1991 Mathematics Subject Classification Primary: 11S05, 11C08; Secondary: 12J25 1. Introduction It is a classical result (see, e.g., [ 20 , Problem 114] ) that if F\in\mathbb{Z}[x] is a polynomial with integer coefficients such that F(a) is a square for every rational argument a\in\mathbb{Q} , then F itself must be the square of another polynomial. There is also a multivariate generalization of this theorem due to Murty (see [ 18 ] ). This raises the natural question of whether similar statements remain valid over other fields occurring naturally in algebraic number theory. In a private conversation, Carlos D’Andrea asked the author if the above result extends to the setting of 2 -adic numbers. More precisely: does there exist a polynomial F\in\mathbb{Z}[x] such that F(a) is a 2 -adic square for every rational number a\in\mathbb{Q} , while F itself is not the square of any polynomial with 2 -adic coefficients? An example of such a polynomial is 4x^{4}+4x^{2}+9 . We omit the proof here, since a more general result will be established later (see Proposition 5.2 ). Nonetheless, this question motivated a systematic study of a class of polynomials— denoted \mathscr{C}(K_{\mathfrak{p}}) —with coefficients in rings of algebraic integers, whose values are always local prime powers, where the relevant exponents coincide with the residue characteristic of the local field. A precise definition of the class \mathscr{C}(K_{\mathfrak{p}}) is given in Section 2 . The class \mathscr{C}(K_{\mathfrak{p}}) has an interesting geometric interpretation. Recall (see, e.g., [ 7 ] ) that if K_{\mathfrak{p}} is a local field (i.e., the completion of a number field with respect to a non-Archimedean valuation) with residue characteristic p , then K_{\mathfrak{p}} can be partitioned into finitely many disjoint open sets of the form aK_{\mathfrak{p}}^{\times p} (so they are classes of p -th powers of nonzero elements of K_{\mathfrak{p}} ), together with the singleton \{0\} , which serves as their common boundary point. Thus, K_{\mathfrak{p}} can be visualized as a ‘fan’ of homeomorphic sets of the form aK_{\mathfrak{p}}^{\times p} , joined together at zero (see Figure 1 ). A polynomial F lies in \mathscr{C}(K_{\mathfrak{p}}) precisely when it collapses this ‘fan’ structure into a single ‘blade’, mapping the entire local field K_{\mathfrak{p}} into the set of p -th powers. One can think of the class \mathscr{C}(K_{\mathfrak{p}}) as—to some extent—analogous to the set of real polynomials that are non-negative everywhere. The field \mathbb{R} decomposes into two open half-lines, -1\mathbb{R}_{+} and 1\mathbb{R}_{+} , together with their common boundary point \{0\} . A polynomial that is nonnegative everywhere can be seen as folding these two half-lines onto one another. In this paper, we study the class \mathscr{C}(K_{\mathfrak{p}}) and several related classes. The paper is organized as follows. In Section 2 we introduce the relevant notation and, in particular, define the class \mathscr{C}(\mathscr{A}) for any subset \mathscr{A} of K_{\mathfrak{p}} . Section 3 presents a characterization of the polynomials that map the entire local field K_{\mathfrak{p}} into its set of p -th powers (see Proposition 3.7 ). The class \mathscr{C}(K_{\mathfrak{p}}) is decidable by the classical Ax–Kochen theorem; in Section 4 we give a simple explicit algorithm (Algorithm 4.9 ) for testing membership in this class. Finally, in Section 5 we return to the motivating question and show (see Proposition 5.2 ) that the class of polynomials mapping K_{\mathfrak{p}} into its set of p -th powers is always strictly larger than the class of p -th powers of polynomials—so the analogue of the classical result mentioned at the beginning never holds over local fields. In this paper, we study the class \mathscr{C}(K_{\mathfrak{p}}) and some of its related classes. The paper is organized as follows. In Section 2 we introduce all the relevant notation. In particular, we define the class \mathscr{C}(\mathscr{A}) for any subset \mathscr{A} of K_{\mathfrak{p}} . Next, in Section 3 we present the characterization of the polynomials that map the entire local field K_{\mathfrak{p}} into the set of its p -th powers (see Proposition 3.7 ). The class \mathscr{C}(K_{\mathfrak{p}}) is decidable by the classical Ax–Kochen theorem; in Section 4 we give a simple explicit algorithm (Algorithm 4.9 ) for testing membership in this class. Finally, in Section 5 we return to the motivating question and show (see Proposition 5.2 ) that the class of polynomials mapping K_{\mathfrak{p}} into its set of p -th powers is always strictly larger than the class of p -th powers of polynomials—so the analogue of the classical result mentioned at the beginning never holds over local fields. 2. Notation Throughout this paper, we use the following notation. Let K be a fixed number field and let \mathbb{Z}_{K} denote its ring of integers (i.e., the integral closure of \mathbb{Z} in K ). Next, let \mathfrak{p} be a prime of K lying above a rational prime p . By K_{\mathfrak{p}} we denote the completion of K at \mathfrak{p} . We write \ord_{\mathfrak{p}} for the unique normalized valuation K_{\mathfrak{p}}\to\mathbb{Z} , as well as its restriction to K . The associated valuation ring is \mathcal{O}_{\mathfrak{p}}=\{a\in K_{\mathfrak{p}}\;|\;\ord_{\mathfrak{p}}a\geq 0\} . ts maximal ideal is \mathfrak{p} , and the quotient \mathcal{O}_{\mathfrak{p}}/\mathfrak{p} is a finite field, called the residue class field of \mathfrak{p} . An element of valuation one is called a uniformizer of \mathfrak{p} . By the strong approximation theorem, one can always choose a uniformizer lying in \mathbb{Z}_{K} . It will be denoted \pi throughout the paper. Finally, we use the standard notation e=e(\mathfrak{p}\mid p) and f=f(\mathfrak{p}\mid p) for the ramification index (i.e., the valuation of p ) and the inertia degree (i.e., the degree of the residue class field over its prime field \mathbb{F}_{p} ), respectively. The quotient group K_{\mathfrak{p}}^{\times}/K_{\mathfrak{p}}^{\times p} of the multiplicative group of K_{\mathfrak{p}} modulo the p -th powers is an elementary p -group. By Artin’s theorem (see, e.g., [ 4 , Theorem 2.36] or [ 15 , Proposition II.6] ) its order is precisely p^{ef+2} if K_{\mathfrak{p}} contains a primitive p -th root of unity, and p^{ef+1} otherwise. As explained in the introduction, our main interest lies in polynomials whose values are local p -th powers, where p is the residue characteristic of the considered local field. For any nonempty subset \mathscr{A}\subseteq K_{\mathfrak{p}} , we define \mathscr{C}(\mathscr{A}):=\bigl\{F\in\mathbb{Z}_{K}[x]\;|\;F(a)\in K_{\mathfrak{p}}^{p}\text{ for every }a\in\mathscr{A}\bigr\}. We are particularly interested in the cases when \mathscr{A} is one of the rings \mathbb{Z}_{K} , \mathcal{O}_{\mathfrak{p}} , K , or K_{\mathfrak{p}} . \mathbb{Q}_{3}^{\times 3} 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 Figure 1. The field \mathbb{Q}_{3} as a union of classes of third powers. 3. Preliminaries The following proposition and its corollary will be used repeatedly throughout the paper. Although they are classical results, we state them explicitly for the reader’s convenience. Proposition 3.1 ( [ 7 , Corollary I.5.8.2] ) . For every k>\nicefrac{{ep}}{{(p-1)}} one has 1+\mathfrak{p}^{k}\subseteq K_{\mathfrak{p}}^{\times p} . Corollary 3.2 . Let a,b\in\mathcal{O}_{\mathfrak{p}} be two elements

Read the complete text