Historical behavior of skew products and arcsine laws

Basic Information

• Paper ID: 2501.00266
• Title: Historical behavior of skew products and arcsine laws
• Authors: Pablo G. Barrientos, Raul R. Chavez
• Classification: math.DS (Dynamical Systems)
• Publication Date: December 31, 2024
• Paper Link: https://arxiv.org/abs/2501.00266

Abstract

This paper investigates the phenomenon of historical behavior for almost every point in skew product systems with one-dimensional fiber dynamics. Under appropriate ergodic conditions, the authors establish results showing that weak forms of the arcsine law lead to non-convergence of Birkhoff averages on almost every orbit. As an application, they prove that this phenomenon occurs in one-step skew product maps over Bernoulli shifts, where the random processes induced by fiber map iterations are conjugate to random walks. Furthermore, the authors revisit known examples of skew products exhibiting historical behavior almost everywhere, verifying that they satisfy the required ergodic and probabilistic conditions. This provides a unified general framework linking these behaviors to the arcsine distribution of orbits.

Research Background and Motivation

Core Problem

The core problem addressed in this paper is Takens' last problem: Does there exist a persistent class of dynamical systems whose set of irregular points has positive Lebesgue measure? Here, irregular points refer to points exhibiting historical behavior, i.e., points where there exists a continuous function such that the time average (Birkhoff average) does not converge.

Research Significance

1. Theoretical Importance: Historical behavior is a fundamental concept in dynamical systems theory, closely related to the existence of physical measures
2. Statistical Behavior: Understanding the statistical properties of dynamical systems is crucial for predicting long-term behavior
3. Probabilistic Connection: Establishing deep connections between dynamical systems and the arcsine law from probability theory

Limitations of Existing Methods

• Known examples of historical behavior are mostly isolated, lacking a unified theoretical framework
• Systematic characterization of historical behavior in skew product systems is lacking
• The mechanism of the arcsine law in dynamical systems remains unclear

Research Motivation

The authors' motivation is to establish a unified framework for understanding historical behavior in skew product systems, particularly:

1. Identifying structural conditions that lead to historical behavior
2. Revealing the intrinsic connection between the arcsine law and historical behavior
3. Providing unified theoretical explanations for known examples

Core Contributions

1. Establishment of Two Main Theorems:
   • Theorem A: For one-step skew products, proves historical behavior under weak arcsine law and ergodic conditions
   • Theorem B: For general skew products, proves historical behavior under reference measure ergodicity and weak arcsine law
2. Provision of Unified Framework: Unifies known examples of historical behavior under the framework of the arcsine law
3. Establishment of Random Walk Connection: Proves that skew product systems conjugate to random walks exhibit historical behavior
4. Characterization of Empirical Measure Limit Points: Describes the asymptotic behavior of empirical measure sequences in historical behavior systems

Methodology Details

Task Definition

Definition of Historical Behavior: For a skew product map $F: \Omega \times M \to \Omega \times M$, if for $(P \times \text{Leb})$-almost every point $(\omega, x)$, there exists a continuous function $\phi: M \to \mathbb{R}$ such that the limit $\lim_{n \to \infty} \frac{1}{n} \sum_{j=0}^{n-1} \phi(f_\omega^j(x))$ does not exist, then $F$ exhibits historical behavior at that point.

Core Theoretical Framework

1. Skew Product System Setup

Consider the skew product map: $F: \Omega \times M \to \Omega \times M, \quad F(\omega, x) = (\tau(\omega), f_\omega(x))$

where:

• $\tau: \Omega \to \Omega$ is a measure-preserving ergodic map
• $f_\omega: M \to M$ is a fiber map
• $M$ is a compact one-dimensional manifold

2. Weak Arcsine Law Condition

Fiber Weak Arcsine Law: There exist $x_0, x_1, \gamma_0, \gamma_1 \in (0,1)$ such that $\liminf_{n \to \infty} P\left(\frac{1}{n} \sum_{j=0}^{n-1} 1_{I_i(\gamma_i)}(f_\omega^j(x_i)) \leq \alpha\right) < 1$ holds for all $\alpha \in (0,1)$ and $i = 0,1$.

Weak Arcsine Law: There exist constants $\gamma_0, \gamma_1 \in (0,1)$ such that $\liminf_{n \to \infty} (P \times \text{Leb})\left(\sum_{j=0}^{n-1} 1_{I_i(\gamma_i)}(f_\omega^j(x)) \leq \alpha\right) < 1$

3. Ergodic Conditions

For one-step skew products, triviality of the tail $\sigma$-algebra is required: (H0) For each $x \in (0,1)$, the tail $\sigma$-algebra of the sequence $\{X_n^x\}_{n \geq 1}$ (where $X_n^x(\omega) = f_\omega^n(x)$) is trivial.

For general skew products, ergodicity of the reference measure is required: $P \times \text{Leb}$ is ergodic with respect to $F$.

Technical Innovations

1. Constancy of Lyapunov Function

Key Lemma: Under condition (H0), the upper and lower limit functions $U_\phi(\omega, x) = \limsup_{n \to \infty} \frac{1}{n} \sum_{j=0}^{n-1} \phi(f_\omega^j(x))$$L_\phi(\omega, x) = \liminf_{n \to \infty} \frac{1}{n} \sum_{j=0}^{n-1} \phi(f_\omega^j(x))$ are constant for $P$-almost every $\omega$.

2. Random Walk Conjugacy

When fiber dynamics are conjugate to random walks, the Erdős-Kac arcsine law directly yields the weak arcsine law condition.

3. Unified Analysis Framework

Occupation time conditions (OT1) and (OT2) connect the weak arcsine law to historical behavior.

Experimental Setup

Theoretical Verification Cases

1. Generalized $(T, T^{-1})$ Transformation

• Setup: $T$ is a Morse-Smale diffeomorphism on a compact one-dimensional manifold
• Skew Product Definition: $F_T(\omega, x) = (\tau(\omega), T^{\omega_0}(x))$
• Verification Method: Proves conjugacy with simple symmetric random walk

2. Skew Flow Transformation

• Setup: $F_{\phi,\phi}(\omega, x) = (\tau(\omega), \phi(\phi(\omega), x))$
• Conditions: $\phi$ is Hölder continuous with zero expectation and non-coboundary
• Verification Method: Via conjugacy with skew translations

3. Zero Schwarzian Derivative Case

• Fiber Map: $f_\omega(x) = \frac{a(\omega)x}{1 + (a(\omega) - 1)x}$
• Conditions: $\int \log a(\omega) dP = 0$, $\int (\log a(\omega))^2 dP < \infty$
• Verification Method: Proves conjugacy with zero-mean finite-variance random walk

4. Thaler Function

• Characteristics: Interval map with neutral fixed point
• Properties: Satisfies ergodicity and weak arcsine law
• Example: Symmetric Manneville-Pomeau function

Verification Methods

For each example, the authors verify:

1. Satisfaction of ergodic conditions
2. Validity of weak arcsine law
3. Occurrence of historical behavior
4. Characterization of empirical measure limit points

Experimental Results

Main Results

Application of Theorem A

Proposition I: For one-step skew products conjugate to random walks, $\lim_{n \to \infty} P\left(\frac{1}{n} \sum_{j=0}^{n-1} 1_{J_i(x)}(f_\omega^j(x)) \leq \alpha\right) = \frac{2}{\pi} \arcsin\sqrt{\alpha}$

Results for Specific Examples

1. Generalized $(T, T^{-1})$ Transformation:
   • Proves historical behavior holds almost everywhere
   • Characterizes limit point set: $L(\omega, x) = \{\lambda\delta_p + (1-\lambda)\delta_q : \lambda \in [0,1]\}$
2. Skew Flow Transformation:
   • Proves historical behavior under non-coboundary condition
   • Extends Ji-Molinek results to broader cases
3. Zero Schwarzian Derivative:
   • Completes Bonifant-Milnor proof
   • Provides complete technical details
4. Thaler Function:
   • Revisits Aaronson et al. results
   • Provides new proof within unified framework

Empirical Measure Behavior

For systems satisfying the conditions, the limit point set of empirical measure sequences has special structure: $L(\omega, x) = \{\lambda\delta_0 + (1-\lambda)\delta_1 : \lambda \in [0,1]\}$

This indicates that empirical measures in historical behavior systems oscillate between two Dirac measures.

Statistical Significance

All results hold in the sense of $(P \times \text{Leb})$-almost every point, which is the strongest result in the measure-theoretic sense.

Related Work

Historical Behavior Research

• Ruelle (2001): Introduces the concept of historical behavior
• Takens (2008): Proposes research problems and framework
• Bowen Eye Model: Classical example of historical behavior

Arcsine Law Research

• Lévy (1939): Arcsine law for Brownian motion
• Erdős-Kac (1947): Arcsine law for random walks
• Probabilistic Applications: Applications in dynamical systems

Skew Product Systems

• Molinek (1994): Early research on historical behavior in skew products
• Ji-Molinek (2000): Historical behavior in skew flows
• Crovisier et al. (2020): Applications to partially hyperbolic diffeomorphisms

Conclusions and Discussion

Main Conclusions

1. Unified Framework: Establishes unified theoretical framework based on weak arcsine law
2. Sufficient Conditions: Provides sufficient conditions for historical behavior in skew product systems
3. Example Unification: Unifies known examples from literature under single theory
4. Random Walk Connection: Reveals deep connections with random walks

Limitations

1. Sufficiency: Given conditions are sufficient but not necessarily necessary
2. Dimension Restriction: Primarily addresses one-dimensional fiber dynamics
3. Countability Assumption: Some results require countable alphabet assumption
4. Continuity: Does not require fiber map continuity, but this may limit applications

Future Directions

1. Necessary Conditions: Seek necessary conditions for historical behavior
2. Higher-Dimensional Generalization: Extend to higher-dimensional fiber dynamics
3. Natural Measures: Study existence of natural measures in historical behavior systems
4. Smooth Systems: Applications in smooth dynamical systems

In-Depth Evaluation

Strengths

1. Theoretical Innovation:
   • Establishes deep connections between dynamical systems and probability theory
   • Provides unified theoretical framework
   • Achieves important technical breakthroughs
2. Mathematical Rigor:
   • Complete and rigorous proofs
   • Proper handling of technical details
   • Reasonable condition settings
3. Broad Applicability:
   • Covers major examples in literature
   • Provides new analytical tools
   • Has potential for further applications
4. Clarity of Exposition:
   • Clear structure with rigorous logic
   • Sufficient motivation explanation
   • Complete technical details

Weaknesses

1. Technical Complexity:
   • Requires substantial background in probability theory and ergodic theory
   • Some technical conditions may be overly restrictive
   • Proofs are highly technical
2. Application Scope:
   • Primarily limited to one-dimensional case
   • Some assumptions restrict applications
   • Connection to practical systems needs further exploration
3. Open Problems:
   • Leaves some important open problems
   • Necessary conditions remain unclear
   • Natural measure problem remains open

Impact

1. Academic Contribution:
   • Provides new tools for dynamical systems theory
   • Establishes interdisciplinary connections
   • Resolves long-standing problems
2. Practical Value:
   • Provides methods for analyzing complex systems
   • Helps understand statistical properties of chaotic systems
   • May have applications in physics and engineering
3. Reproducibility:
   • Theoretical results are verifiable
   • Methods can be generalized to other systems
   • Provides foundation for subsequent research

Applicable Scenarios

1. Theoretical Research: Dynamical systems theory, ergodic theory, probability theory
2. Applied Mathematics: Complex system analysis, statistical physics
3. Computational Science: Numerical analysis of chaotic systems
4. Interdisciplinary Research: Mathematical physics, biomathematics, and related fields

References

This paper cites extensive relevant literature, primarily including:

• Foundational literature on dynamical systems theory
• Research on historical behavior
• Probability theory literature on arcsine law
• Research on specific examples of skew product systems

The citations are comprehensive and authoritative, reflecting the authors' deep understanding of the field.