🌀 nonlinear sciences

Growth of Mahler measure and algebraic entropy of dynamics with the Laurent property

This paper investigates the growth rate of the Mahler measure in discrete dynamical systems with the Laurent property and cluster algebras, proposing and providing evidence for the conjecture that this growth rate coincides with algebraic entropy, while also deriving precise asymptotic formulas for specific cases like the Kronecker quiver.

Original authors: Andrew N. W. Hone

Published 2026-03-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Andrew N. W. Hone

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are watching a complex machine, like a giant, magical Rube Goldberg device. Every time you pull a lever (perform a step in a sequence), the machine produces a new number or a new expression. Sometimes these expressions get incredibly complicated, growing huge and messy.

Mathematicians want to know: How fast does this mess grow? Is it growing slowly like a houseplant, or explosively like a wildfire?

This paper, written by Andrew Hone, introduces a new, clever way to measure that growth speed. It compares three different "thermometers" for measuring complexity in these mathematical machines.

The Three Thermometers

To understand the paper, we need to know about the three ways mathematicians measure growth:

The Degree Thermometer (Algebraic Entropy):
Imagine the machine spits out a giant polynomial (a math expression with many terms like $x^2 + 5x + 1$ ). This thermometer counts the highest power of $x$ . If the expression is $x^{100}$ , the degree is 100. If it grows to $x^{1000}$ , the degree is 1000. This is the "standard" way to measure complexity, but it's very hard to calculate for big machines because the numbers get astronomically large, and computers crash trying to handle them.
The Height Thermometer (Diophantine Entropy):
Instead of looking at the formula, this thermometer looks at the actual numbers you get if you plug in simple values (like 1, 2, 3). It measures how big the numerators and denominators get. It's easier to calculate than the first one, but still requires doing exact arithmetic with massive fractions, which is slow and tedious.
The Mahler Measure Thermometer (Mahler Entropy):
This is the new hero of the story. Instead of counting terms or big fractions, this thermometer takes the "average size" of the expression when you plug in complex numbers that sit on a circle (specifically, numbers like $e^{it}$ ).
- The Analogy: Imagine the polynomial is a landscape. The "Degree" counts the tallest mountain peak. The "Height" measures how heavy the rocks are. The Mahler Measure is like measuring the average elevation of the entire landscape.
- Why it's cool: It turns out you can calculate this "average elevation" very quickly using floating-point numbers (like a calculator), even when the actual formula is too huge to write down.

The Big Guess (The Conjecture)

The author's main idea is a bold guess: These three thermometers all read the same temperature.

He proposes that for a special class of mathematical machines (called "Cluster Algebras" or systems with the "Laurent Property"), the speed at which the "average elevation" (Mahler Measure) grows is exactly the same as the speed at which the "tallest peak" (Degree) grows.

Why does this matter? Because calculating the "average elevation" is like taking a quick snapshot with a smartphone, while calculating the "tallest peak" is like climbing the mountain with a heavy backpack. If the guess is true, we can use the easy method to predict the hard result.

The Evidence: Testing the Machines

The author tested this guess on three famous mathematical machines:

Rank 2 Cluster Algebras (The Simple Machines):
These are machines with two variables.
- Case A (Slow Growth): For some settings, the machine is "integrable" (predictable). The growth is linear (like a straight line). The author proved mathematically that the Mahler Measure grows exactly as fast as the Degree.
- Case B (Fast Growth): For other settings, the machine is chaotic. The growth is exponential (like a virus). The author used a computer to simulate the machine. The "average elevation" grew at a rate that matched the "tallest peak" rate to 9 decimal places!
The Markoff Numbers (The 3D Machine):
This machine is related to a famous equation about triangles and numbers. It's a bit more complex. The author showed that even here, the "average elevation" grew at the exact same speed as the "tallest peak."
The Somos-4 Recurrence (The Elliptic Machine):
This machine is related to shapes called elliptic curves. Here, the growth isn't a straight line or a curve; it's a parabola (growing like $n^2$ ). The author showed that the Mahler Measure also grows like $n^2$ . This confirms that even when the growth pattern changes shape, the Mahler Measure still tracks the complexity perfectly.

The "Magic" Connection

The paper also hints at a deeper magic. The author notes that the "average elevation" (Mahler Measure) often relates to the Bloch-Wigner Dilogarithm, a special mathematical function that describes the volume of hyperbolic shapes (like weird, curved 3D spaces).

It's as if the complexity of these number machines is secretly connected to the geometry of curved spaces. The author suggests that because these machines (Cluster Algebras) are built on deep geometric principles, their growth rates naturally align with these geometric volumes.

Summary in Plain English

The Problem: Measuring how fast complex math formulas grow is usually very hard and slow.
The Solution: The author suggests using a "quick and dirty" method (Mahler Measure) that calculates the average size of the formula.
The Discovery: In many special cases, this quick method gives the exact same answer as the slow, hard method.
The Analogy: It's like realizing that if you want to know how fast a car is going, you don't need to measure every single gear and piston (Degree). You can just look at the speedometer (Mahler Measure), and it will tell you the truth, much faster and easier.

The paper provides strong numerical evidence that this "speedometer" works, and it offers a new, powerful tool for mathematicians to study complex systems without getting bogged down in impossible calculations.

1. Problem Statement

The paper addresses the challenge of measuring the complexity and growth rates in discrete dynamical systems, specifically those possessing the Laurent property (where iterates are Laurent polynomials with integer coefficients). These systems include cluster algebras and specific recurrence relations (e.g., Somos sequences, Markoff triples).

Traditionally, complexity is measured via:

Algebraic Entropy ( $E$ ): Defined by the exponential growth rate of the degree of the rational functions defining the iterates.
Diophantine Entropy ( $E_H$ ): Defined by the growth rate of the logarithmic height (size of numerators/denominators) of iterates evaluated at specific integer points.

The Core Problem:
Calculating algebraic entropy requires exact symbolic computation of iterates, which becomes computationally prohibitive as degrees grow exponentially. Calculating Diophantine entropy requires high-precision arithmetic on large integers. The author proposes investigating the Mahler measure ( $m(P)$ ) as an alternative, computationally efficient proxy for entropy. The central question is whether the growth rate of the Mahler measure (Mahler entropy, $E_M$ ) coincides with the algebraic entropy ( $E$ ) for these systems.

2. Methodology

The author employs a combination of theoretical analysis, exact symbolic integration, and numerical approximation.

Definitions:
- Mahler Measure: For a Laurent polynomial $P$ , $m(P) = \int_{T^N} \log |P| d\Omega$ , where $T^N$ is the $N$ -torus.
- Mahler Entropy ( $E_M$ ): Defined as $\lim_{n \to \infty} \frac{1}{n} \log m(\phi^n(x))$ , where $\phi$ is the dynamical map.
Theoretical Framework:
- The paper leverages the positivity property of cluster algebras (cluster variables have positive coefficients) to establish an inequality between Mahler entropy and Diophantine entropy ( $E_M \leq E_H$ ).
- It utilizes Jensen's formula to reduce multi-variable Mahler measures to single-variable integrals or sums of roots.
- It connects Mahler measures to Bloch-Wigner dilogarithms ( $D(z)$ ) and $L$ -functions, which often appear in the exact evaluation of these integrals.
Numerical Techniques:
- Monte Carlo Integration: To approximate $m(P)$ for high-order iterates, the author uses Riemann sums over the torus $T^N$ . Points are chosen as roots of unity or via pseudo-random generators.
- Floating Point Arithmetic: Unlike algebraic entropy (which requires exact rational arithmetic), Mahler measure approximations can be computed rapidly using floating-point numbers, allowing for the analysis of much higher iterates ( $n \approx 100$ ).
Analytical Tools:
- Tropical Dynamics: The behavior of degree vectors ( $d$ -vectors) is analyzed using tropical exchange relations to predict degree growth.
- Elliptic Functions: For the Somos-4 recurrence, the solution is expressed using the Weierstrass sigma function to analyze asymptotic growth.

3. Key Contributions

Formulation of Conjecture 1.7: The paper explicitly conjectures that for cluster algebras and maps with the Laurent property, the Mahler entropy coincides with the algebraic entropy ( $E = E_M$ ).
Proposition 1.9: A theoretical proof establishing that for positive Laurent polynomials, the Mahler entropy is bounded by the Diophantine entropy of the orbit starting at $(1, \dots, 1)$ . If degrees and Mahler measures grow exponentially with the same rate $\lambda$ , then $E_M = E_H = \log \lambda$ .
Numerical Efficiency: Demonstrates that computing Mahler entropy via Monte Carlo integration is significantly faster and less resource-intensive than computing algebraic entropy via exact degree tracking.
Exact Formulas: Derives exact formulas for the Mahler measures of specific cluster variables in rank 2 cluster algebras, expressing them in terms of the Bloch-Wigner dilogarithm.

4. Results

The paper provides evidence for the conjecture across several distinct dynamical systems:

Rank 2 Cluster Algebras ( $F(x) = x^r + 1$ ):
- Case $r=1, 2$ (Integrable): The algebraic entropy is 0. The Mahler measure $m(x_n)$ grows linearly with $n$ . The slope $C^*$ is calculated exactly as an integral involving the conserved quantity $K$ .
- Case $r \geq 3$ (Chaotic): The algebraic entropy is $E = \log(\frac{r + \sqrt{r^2-4}}{2})$ . Numerical calculations show the Mahler measure grows exponentially with a rate matching $E$ to high precision (e.g., for $r=3$ , $E_M \approx 0.96242365$ vs $E \approx 0.96242364$ ).
Markoff Recurrence ( $x_{n+3}x_n = x_{n+2}^2 + x_{n+1}^2$ ):
- This system corresponds to the exchange matrix of affine type $\tilde{A}_1$ (Kronecker quiver).
- The algebraic entropy is $E = \log(\frac{1+\sqrt{5}}{2})$ .
- Numerical evaluation of the Mahler measure for the sequence $x_n$ yields a growth rate matching $E$ to 9 decimal places.
Somos-4 Recurrence:
- This system has algebraic entropy $E=0$ (quadratic degree growth).
- The Mahler measure $m(x_n)$ is shown to grow quadratically ( $m(x_n) \sim Cn^2$ ).
- This confirms $E_M = 0$ , consistent with the algebraic entropy. The quadratic growth is linked to the addition of points on an elliptic curve and the asymptotic behavior of the Weierstrass sigma function.

5. Significance

Computational Tool: The paper establishes Mahler entropy as a practical, high-speed numerical tool for detecting integrability and measuring complexity in dynamical systems where exact degree calculation is infeasible.
Unification of Concepts: It bridges the gap between arithmetic geometry (Mahler measure, heights), complex dynamics (entropy), and combinatorial algebra (cluster algebras).
Deep Mathematical Connections: The results suggest a profound link between the growth of cluster variables and special functions (dilogarithms, $L$ -functions) and hyperbolic geometry (volumes of polyhedra).
Future Directions: The paper suggests that if the sequence of Mahler measures satisfies a "tropical" version of the original dynamical system, it could lead to a rigorous proof of the conjecture relating Mahler and algebraic entropies. It also highlights the need for deeper theoretical explanations for the appearance of dilogarithms in these contexts.