Pointwise ergodic theorem along primes of the form x 2 + n ​ y 2 x^{2}+ny^{2}

Jan Fornal · 2025

Read Pointwise ergodic theorem along primes of the form x 2 + n ​ y 2 x^{2}+ny^{2} by Jan Fornal (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-2508-15466.txt.

Introduction

Jan Fornal proves a pointwise ergodic theorem along special primes in two thousand twenty-five. The primes take the form of one square plus n times another. Averages of a function along those prime values are at stake. Bourgain's classic work frames the whole line of study.

The Hardy-Littlewood circle method carries the proof here. Major and minor arc estimates for prime ideals are the new core. Pointwise convergence holds for the class of functions he treats. It cannot stretch to the full L1 class, he also shows. That bar marks the sharp edge of the result.

Catalog record

CREATORJan Fornal
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE115 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-15466

Browse the full library · About the Tower of Babel

Full text

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

Pointwise ergodic theorem along primes of the form x 2 + n ​ y 2 x^{2}+ny^{2} Jan Fornal Abstract . This paper resolves the question of pointwise convergence for ergodic averages of a single function along the set of polynomial values of primes of the form x 2 + n ​ y 2 x^{2}+ny^{2} . Following the influential paper of Bourgain [ Bou89 ] , we employ the Hardy-Littlewood circle method where major arc and minor arc estimates for the set of prime ideals constitute the main novelty of the paper. We also prove that our convergence results cannot be extended to class of L 1 L^{1} functions. 1. Introduction Pointwise ergodic theory dates back to the work of Birkhoff who proved the pointwise ergodic theorem for ordinary ergodic averages. Motivated by a question of Furstenberg, Bourgain started investigating subsequential ergodic averages. These take the form: (1.1) \frac{1}{M}\sum_{k\leq M}f(T^{a_{k}}x). More precisely, he published in Publications Mathématiques de l’IHÉS [ Bou89 ] the highly creative and celebrated proof of the following: Theorem 1.1 . Let P be a polynomial with integer coefficients. Fix q>1 , a measure preserving system (X,\mu,T) . If f\in L^{q}(X) , then: (1.2) \frac{1}{m}\sum_{k\leq m}f(T^{P(k)}x) converges for almost all x\in X . In 1988, building off work of Bourgain, specifically the L^{2} formulation of the below result, Máté Wierdl in his paper [ Wie88 ] managed to prove the following: Theorem 1.2 . Let \mathbb{P} be the set of prime numbers. Fix q>1 and suppose that (X,\mu,T) is measure-preserving system. If f\in L^{q}(X) , then: (1.3) \frac{1}{|\{p\in\mathbb{P}:p\leq m\}|}\sum_{p\leq m}f(T^{p}x) converges for almost all x\in X . In the previous two theorems, by a measure-preserving system, we mean a \sigma -finite measure space equipped with a measure-preserving transformation; throughout (X,\mu,T) will denote such a measure-preserving system. These two papers made a significant impact in the field of pointwise ergodic theory and in the author’s opinion they changed the general perception of this area of analysis. Later on, mathematicians gave plenty of modifications regarding convergence of nonstandard ergodic averages. Our focus is the class of primes: (1.4) \mathbb{P}_{n}=\{p\in\mathbb{P}:\text{ there exists }u,v\in\mathbb{Z}\text{ so that }p=u^{2}+nv^{2}\}, where n is fixed for the rest of the paper. These primes have drawn the attention of various eminent mathematicians such as Fermat, Euler, Gauss, Hecke and Hilbert. They comply with various interesting principles coming from both analytic number theory and algebraic number theory. Using these primes, we will construct new types of ergodic averages and the main goal of this paper is to fully address the issue of pointwise convergence of these averages. The first result is qualitative: Theorem 1.3 . Let n be positive integer. Fix q>1 , a measure preserving system (X,\mu,T) and a polynomial P(x)\in\mathbb{Z}[x] . If f\in L^{q}(X) , then: (1.5) A_{m}^{n}f(x):=A_{m}f(x)=\frac{1}{|\{p\in\mathbb{P}_{n}:p\leq m\}|}\sum_{\begin{subarray}{c}p\in\mathbb{P}_{n}\\ p\leq m\end{subarray}}f(T^{P(p)}x) converges for almost all x\in X . We further address quantitative convergence statistics for these averages, namely r -variation and the jump counting function, introduced by Lépingle in the context of martingales ( [ Lep76 ] ) and imported to the ergodic-theoretic context by Bourgain (see for instance [ Bou06 ] ); we also address Bourgain’s oscillation seminorm. We will introduce the precise definitions of r -variation, the jump counting function and the oscillation seminorm in Section 7 . Much work has been devoted to addressing the interplay between these operators and classical questions in harmonic analysis and ergodic theory (see for instance [ Cam+00 ] , [ Zor15 ] or [ Kra22 ] ). Our next result proves that ergodic averages along primes of the form x^{2}+ny^{2} contribute another example to the aforementioned collection: Theorem 1.4 . Let n be a positive integer. Fix r>2,q>1 , a polynomial P(x)\in\mathbb{Z}[x] , a finite increasing sequence I=(I_{j})_{j\in[1,M]} and a positive real number \lambda . For all measurable functions f on a measure-preserving system (X,\mu,T) , the following inequalities are satisfied: (1.6) \left\lVert\mathcal{V}^{r}\Bigl(A_{m}f(x):m\in\mathbb{Z}\Bigr)\right\rVert_{L^{q}(X)}\lesssim_{P,q,n}\frac{r}{r-2}\left\lVert f\right\rVert_{L^{q}(X)} (1.7) \left\lVert\text{Osc}_{I}\Bigl(A_{m}f(x):m\in\mathbb{Z}\Bigr)\right\rVert_{L^{q}(X)}\lesssim_{P,q,n}\left\lVert f\right\rVert_{L^{q}(X)} (1.8) \left\lVert\lambda N_{\lambda}^{1/2}\Bigl(A_{m}f(x):m\in\mathbb{Z}\Bigr)\right\rVert_{L^{q}(X)}\lesssim_{P,q,n}\left\lVert f\right\rVert_{L^{q}(X)} It is worthwhile mentioning explicitly that in the second and third inequalities, the implied constants does not depend on the choice of sequence I nor the number \lambda , respectively. Remark 1.5 . Lemma 2.12 from [ MSZ20b ] together with interpolation methods justifies why ( 1.6 ) follows from ( 1.8 ). However, we will give the unified proof of Theorem 1.4 for all operators above. The final result of our work addresses the L^{1} -endpoint. For the most popular nonconventional ergodic averages i.e. along primes and polynomial images, convergence was established only for functions in L^{p}(X) with p>1 , in contrast to Birkhoff’s theorem. For a long time it was a serious question to investigate convergence only under the assumption that f\in L^{1} . In celebrated work, Buczolich and Mauldin (in [ BM10 ] ) established the divergence for averages of the form: (1.9) \frac{1}{N}\sum_{n=1}^{N}f(T^{n^{d}}x) for general f\in L^{1}(X) and d=2 . They used Sawyer’s principle (more precisely, equation (1) from [ Saw66 ] ) regarding the link between weak-type maximal inequalities and almost everywhere convergence/divergence of ergodic averages for functions in L^{1}(X) . Later on, LaVictoire invented a more general construction for showing that certain ergodic averages diverge (see Theorem 3.1 in [ LaV11 ] ), using it to show that ergodic averages along primes and ergodic averages diverge ( 1.9 ) for d\geq 3 . We will use that construction to justify the following: Theorem 1.6 . Let n be a positive integer. For every measure preserving system (X,\mu,T) , there exists a function f\in L^{1}(X) so that: (1.10) A_{m}f(x)=\frac{1}{|\{p\in\mathbb{P}_{n}:p\leq m\}|}\sum_{\begin{subarray}{c}p\in\mathbb{P}_{n}\\ p\leq m\end{subarray}}f(T^{p}x) diverges for almost every x\in X . In particular, the theme of this paper is to provide additional input to the program of study of nonconventional ergodic averages introduced by Bourgain: (1) Bourgain’s investigations on ergodic averages with polynomial values appeared in a sequence of papers beginning with [ Bou06 ] and culminating with the famous [ Bou89 ] . (2) The case of primes was resolved by Wierdl using simpler methods (see [ Wie88 ] ). (3) Weighted ergodic averages were investigated by Cuny and Weber in [ CW17 ] . (4) A new paradigm was introduced in last decade by Mirek, Stein and collaborators using Ionescu-Wainger theory (respectively brought up in Section 7 ). In the series of papers [ MST19 ] , [ MST ] they established estimates for variation and square function of multidimensional polynomial ergodic averages. (5) Variational estimates for primes are covered in [ Zor21 ] and [ MTZ17 ] . (6) Estimates for the jump counting function and oscillation seminorms for multidimensional ergodic averages for polynomial values/primes are discussed by Mirek, Stein, Zorin-Kranich and by Mehlhop and Słomian in [ MSZ20a ] , [ MS24 ] . (7) Notions of oscillation proved to be particularly useful in [ Bou+23 ] , giving substantial progress on the linear version of the Furstenberg-Bergelson-Leibman conjecture (i.e. Conjecture 1.22. from the aforementioned paper). The strategy for establishing quantitative and qualitative results in pointwise ergodic theory goes back to Bourgain, as he provided an adaptation of the analytic-number-theoretic Hardy-Littlewood circle method to accommodate the needs of pointwise ergodic theory. By the Caldéron’s transference principle, one may focus on the integer system (\mathbb{Z},x\to x-1) . In that context,

Read the complete text