Furstenberg systems of certain sequences of superpolynomial growth

Basic Information

• Paper ID: 2510.11957
• Title: Furstenberg systems of certain sequences of superpolynomial growth
• Authors: Andreu Ferré Moragues, Andreas Koutsogiannis
• Classification: math.DS (Dynamical Systems), math.NT (Number Theory)
• Submission Date: October 13, 2025
• Paper Link: https://arxiv.org/abs/2510.11957v1

Abstract

This paper investigates sequences defined by smooth functions of moderate growth and analyzes the Furstenberg systems that model their statistical behavior. In particular, it is proven that these systems are Bernoulli systems. This is achieved by studying exponential sums that reflect the strong equidistribution properties of the sequences. As a byproduct of the methodology, several convergence results are also obtained.

Research Background and Motivation

Problem Background

1. Applications of the Furstenberg Correspondence Principle: The correspondence principle introduced by Furstenberg in his proof of Szemerédi's theorem pioneered a new approach to studying combinatorial problems using ergodic theory, transforming the problem of finding patterns in large subsets into questions about recurrence properties on measure-preserving systems.
2. Dynamical Modeling of Sequence Statistics: The statistical properties of bounded sequences $a: \mathbb{N} \to \mathbb{C}$ can be studied through appropriate dynamical models, called Furstenberg systems.
3. Limitations of Existing Research:
   • Work by Frantzikinakis and others primarily addresses sequences with polynomial growth rates
   • For sequences of moderate growth (superpolynomial but subexponential), traditional methods such as van der Corput techniques no longer apply
   • The study of moderate growth sequences remains a broadly open problem

Research Motivation

This paper focuses on sequences of the form $(n\log^c n)_n$ and $(\alpha\lfloor n\log^c n\rfloor)_n$ (where $0 < c < 1/2$, $\alpha \in \mathbb{R} \setminus \mathbb{Z}$), which exhibit moderate growth: growing faster than any polynomial but slower than exponential functions.

Core Contributions

1. Characterization of Bernoulli Systems: It is proven that sequences of moderate growth $a(n) := e(G(n))$ and $b(n) := e(\alpha\lfloor G(n)\rfloor)$ possess a unique Bernoulli Furstenberg system, where $G(x) = x\log^c x$.
2. Strong Equidistribution Properties: The "goodness" property of sequences $((G(n+h_1), \ldots, G(n+h_r)))_n$ is established, which is a stronger concept than equidistribution.
3. Exponential Sum Estimates: Techniques for exponential sum estimates involving linear combinations of moderate growth functions are developed, generalizing the Brüdern-Perelli method.
4. Topological Dynamics Applications: Asymptotic orthogonality results for zero topological entropy systems are obtained.
5. von Neumann-type Convergence: $L^2$ convergence results in systems of commuting measure-preserving transformations are proven.

Detailed Methodology

Task Definition

Study the Furstenberg systems of sequences $a(n) = e(G(n))$ and $b(n) = e(\alpha\lfloor G(n)\rfloor)$, where:

• $G(x) = x\log^c x$, $0 < c < 1/2$
• $\alpha \in \mathbb{R} \setminus \mathbb{Z}$
• The goal is to determine the structure of these systems (particularly to prove they are Bernoulli systems)

Core Technical Framework

1. Furstenberg System Theory

According to Corollary 2.4, a sequence has a unique Bernoulli Furstenberg system if and only if: $\lim_{N\to\infty} \mathbb{E}_{n\in[N]} a^{\varepsilon_1}(n+h_1)\cdots a^{\varepsilon_r}(n+h_r) = 0$ holds for all $r \in \mathbb{N}$, distinct $h_1,\ldots,h_r \in \mathbb{Z}$, and all $(\varepsilon_1,\ldots,\varepsilon_r) \in \mathbb{Z}^r \setminus \{\vec{0}\}$.

2. Exponential Sum Estimation Techniques

Main Theorem (Theorem 3.1): For the function $F(x) = \sum_{j=1}^r \alpha_j G(x+h_j)$ under appropriate conditions: $\sum_{N+a \leq n \leq 2N-a} e(\alpha F(n)) \ll N\exp(-\kappa\log^{1-2c}N)$

Key Tools:

• Karacuba's lemma (for handling oscillatory integrals)
• Precise bounds on $G(x)$ and its derivatives
• Analysis of upper and lower bounds for derivatives of linear combinations $F(x)$

3. Technical Innovations in Derivative Estimation

Lower Bound Estimates (Proposition 3.5): Through analysis of finite difference operators, lower bounds of the form $\frac{F^{(s)}(x)}{s!} \geq d \cdot \frac{G(x)}{x^{s+\tau}}$ are established for $F^{(s)}(x)$, where constants $d,\tau$ depend only on the coefficients and shift parameters.

Upper Bound Estimates (Lemma 3.13): Using the Cauchy integral formula and analytic function theory: $\left|\frac{F^{(s)}(x)}{s!}\right| \leq C\left(\frac{2}{N}\right)^s G(3N)$

Technical Innovations

1. New Methods for Moderate Growth: Traditional van der Corput techniques fail for superpolynomial growth; this paper develops new techniques based on Karacuba's lemma.
2. Refined Derivative Analysis: By transforming the derivative problem for linear combinations using finite difference operators into more tractable forms, unified upper and lower bound estimates are established.
3. Proof of "Goodness": Strong equidistribution properties of shifted sequences are proven, which is a stronger concept than ordinary equidistribution.

Experimental Setup

This is a pure theoretical research paper, with results verified primarily through mathematical proofs. Key verification components include:

Theoretical Verification Framework

1. Parameter Range: $0 < c < 1/2$ ensures moderate growth properties
2. Sequence Types:
   • Continuous case: $a(n) = e(G(n))$
   • Discrete case: $b(n) = e(\alpha\lfloor G(n)\rfloor)$, $\alpha \in \mathbb{R}\setminus\mathbb{Z}$
3. Shift Parameters: Various integer shifts $h_1,\ldots,h_r \in \mathbb{Z}$

Main Results

Core Theorems

Theorem 1.1 (Characterization of Furstenberg Systems): For $0 < c < 1/2$, the sequences $a(n) := e(G(n))$ and $b(n) := e(\alpha\lfloor G(n)\rfloor)$ ($\alpha \in \mathbb{R}\setminus\mathbb{Z}$) possess a unique Bernoulli Furstenberg system.

Theorem 1.3 (Equidistribution Properties): The sequences $((G(n+h_1),\ldots,G(n+h_r)))_n$ and $((α\lfloor G(n+h_1)\rfloor,\ldots,α\lfloor G(n+h_r)\rfloor))_n$ are respectively equidistributed on $\mathbb{T}^r$ and corresponding spaces.

Theorem 1.4 (Goodness Property): The sequence $((G(n+h_1),\ldots,G(n+h_r)))_n$ is "good," meaning it exhibits strong equidistribution for any nontrivial linear combination.

Application Results

Corollary 1.2 (Topological Dynamics Application): For a topological dynamical system $(X,T)$ with zero topological entropy, asymptotic orthogonality holds: $\lim_{N\to\infty} \mathbb{E}_{n\in[N]} a(n)f(T^n x) = 0$

Theorem 1.5 (von Neumann-type Result): For systems of commuting measure-preserving transformations, $L^2$ convergence holds: $\lim_{N\to\infty} \left\|\mathbb{E}_{n\in[N]} T_1^{\lfloor G(n+h_1)\rfloor}\cdots T_r^{\lfloor G(n+h_r)\rfloor}f - Pf\right\|_2 = 0$

Related Work

Main Research Directions

1. Polynomial Growth Sequences: Frantzikinakis 7 determined the structure of Furstenberg systems for sequences with polynomial growth rates
2. Hardy Field Functions: Related work addresses iterations of smooth functions in Hardy fields
3. Number-Theoretic Applications: Connections to number-theoretic problems such as the Chowla conjecture and the Möbius function

Position of This Work

This paper fills the gap between polynomial and exponential growth, addressing the moderate growth case that previous techniques could not cover.

Conclusions and Discussion

Main Conclusions

1. It is proven that Furstenberg systems of moderate growth sequences exhibit maximal randomness (Bernoulli property)
2. A new technical framework for handling sequences of superpolynomial growth is established
3. Applications in topological dynamics and ergodic theory are obtained

Limitations

1. Growth Rate Restrictions: The method applies only to the case $0 < c < 1/2$
2. Technical Complexity: Refined analytic estimates are required, making generalization to more general cases difficult
3. Open Problems: For general moderate growth functions in Hardy fields, the problem remains open

Future Directions

The paper proposes two important conjectures:

Conjecture 1.6: For a measure-preserving system $(X,\mathcal{B},\mu,T)$ and function $f \in L^\infty(\mu)$, for almost every $x \in X$, the sequence $d(n) := f(T^{\lfloor G(n)\rfloor}x)$ has a unique Bernoulli Furstenberg system.

Conjecture 1.7: Concerning convergence properties of noncommuting transformations in zero-entropy systems.

In-Depth Evaluation

Strengths

1. Theoretical Innovation: First systematic treatment of Furstenberg systems for moderate growth sequences
2. Technical Breakthrough: New exponential sum estimation techniques are developed, overcoming limitations of traditional methods
3. Complete Results: Forms a complete theoretical system from foundational theory to concrete applications
4. Mathematical Rigor: Proofs are detailed and rigorous with refined technical handling

Weaknesses

1. Limited Scope: Results are restricted to specific growth rate ranges
2. Computational Complexity: Proofs involve substantial technical details that may hinder generalization
3. Practical Applications: Primarily theoretical results; practical applicability remains to be explored

Impact

1. Theoretical Contribution: Provides new tools for cross-disciplinary research between dynamical systems and number theory
2. Methodological Value: Techniques for handling moderate growth functions may apply to other problems
3. Open Problems: Proposed conjectures point directions for future research

Applicable Scenarios

1. Equidistribution problems in ergodic theory
2. Exponential sum estimates in analytic number theory
3. Randomness studies in dynamical systems
4. Pattern detection in combinatorial mathematics

References

The paper cites 22 related references, primarily including:

• Foundational work by Furstenberg 12,13
• Research by Frantzikinakis on Hardy field sequences 7,8
• Karacuba's exponential sum estimation techniques 17
• Foundational theory on Furstenberg systems by Bergelson et al. 1,2

Overall Assessment: This is a high-quality theoretical mathematics paper making important contributions at the intersection of dynamical systems and number theory. While technically demanding, it provides profound insights into the statistical properties of moderate growth sequences and possesses significant theoretical value.