Mahler Measure in Number Theory & Dynamics

• Mahler measure is an analytic invariant that quantifies the average logarithmic size of a polynomial’s values on the unit torus.
• It plays a central role in linking transcendental number theory, arithmetic geometry, and dynamical systems through explicit toral integrals and regulator interpretations.
• Recent advances reveal deep connections with L-values, regulator maps, and open problems like Lehmer’s question in Diophantine analysis.

The Mahler measure is a fundamental analytic invariant in transcendental number theory, algebraic geometry, and arithmetic dynamics. It encodes the average size and distribution of polynomials or algebraic numbers on the unit torus, and is tightly connected to special values of $L$-functions, $K$-theory, and arithmetic properties such as Pisot and Salem numbers, Lehmer’s problem, and regulator maps. The classical Mahler measure associates to a polynomial $P$ in $n$ variables a mean value of $\log|P|$ on the $n$-torus, but a wealth of generalizations, conjectural links, and dynamical perspectives have emerged in contemporary research.

1. Definition and Core Properties

Let $P(z_1, ..., z_n)$ be a nonzero Laurent polynomial in $n$ complex variables. Its (logarithmic) Mahler measure is

$$m(P) = \frac{1}{(2\pi)^n} \int_0^{2\pi} \cdots \int_0^{2\pi} \log|P(e^{i\theta_1}, ..., e^{i\theta_n})| \, d\theta_1 \cdots d\theta_n,$$

with the multiplicative version $M(P) = \exp(m(P))$ (Mossinghoff, 2024, Dobrowolski et al., 2016). For $n=1$, Jensen’s formula shows $m(P) = \log|a_0| + \sum_{j=1}^{d} \log^+|\alpha_j|$, for $P(z)=a_0 \prod_{j=1}^d (z-\alpha_j)$.

For polynomials in many variables, including nontrivial cases in two and three variables, Mahler measure cannot generally be expressed in terms of roots; however, it admits deep interpretations via toral integrals, regulators, and motivic symbols (Bornhorn, 2015). Key properties:

• $m(P)$ is subadditive: $m(PQ) \le m(P) + m(Q)$
• For algebraic $\alpha$, $m(\alpha^n) = |n|\, m(\alpha)$
• $m(\alpha) = 0$ iff $\alpha$ is a root of unity
• $M(P) \geq h(P)/2^{k-2}$ for $k$-term polynomials, where $h(P)$ is the height (Dobrowolski et al., 2016)

Lehmer’s question asks whether there exists $c>0$ such that $m(\alpha) \geq c$ for all nonzero non-root-of-unity $\alpha$; the minimal known such value is $m(L) \approx 0.16235$ for the Lehmer polynomial.

2. Mahler Measure in Arithmetic, Dynamics, and Extremal Problems

The Mahler measure is central in:

• Mahler’s problem for Littlewood polynomials: Determining $\sup_{f \in \mathcal{L}_n} M(f)/\Vert f \Vert_2$, where $\mathcal{L}_n$ is the set of degree-$n$ polynomials with $\pm 1$ coefficients. Recent work establishes that Turyn polynomials with quarter-degree shift yield $M(f)/\Vert f \Vert_2 \to 0.9511$ as degree grows, setting a new record and bounding any global gap $1-\epsilon$ by $\epsilon < 0.049$ (Mossinghoff, 2024). Previous extremal examples, such as Rudin–Shapiro polynomials, attained the lower bound $M(f)/\Vert f \Vert_2 \approx 0.8578$ (Erdelyi, 2014).
• Lower bounds for sparse and “almost reciprocal” polynomials: For any $k$-term polynomial, $M(f) \geq h(f)/2^{k-2}$; for $k$-nonreciprocal polynomials, explicit sharp quadratic bounds in terms of coefficient deviations are proved (Dobrowolski et al., 2016, Saunders, 2017).
• Closure and gaps: The set of Mahler measures for $k$-term integer polynomials is closed with 0 as an isolated point; only finitely many such polynomials have Mahler measure exactly 1 for fixed $k$ (Dobrowolski et al., 2016).
• Metric and dynamical Mahler measures: Iteration of $M(\cdot)$ on algebraic numbers defines a dynamical system exhibiting fixed, preperiodic, and wandering points, with full classification for all abelian number fields and degree $\le 5$. Generic unit orbits of degree $d \ge 4$ fall into classes with orbit size $1,2,$ or $\infty$; non-units admit algebraic integers of arbitrary finite orbit size (Fili et al., 2019, Fili et al., 2021, Samuels, 2017).

Table: Extremal Normalized Mahler Measures

Family	Limiting $M(f)/\Vert f \Vert_2$	Achieved by
Rudin–Shapiro	$\approx 0.8578$	Known explicit sequence
Turyn (quarter-shift)	$\approx 0.9511$	Turyn polynomials

3. Analytical, K-Theoretic, and Motivic Interpretations

Mahler measure admits profound connections to regulators and motives:

• Regulator integrals: For curves defined by $P(x, y) = 0$, $m(P)-m(P^*)$ equals the regulator pairing $\langle r_{\mathcal{D}}\{x, y\}, [A]\rangle$ in $K_2$ cohomology; this underlies the appearance of $L$-derivatives and special values in Mahler measure (Bornhorn, 2015, Brunault, 2015, Trieu, 2023).
• Beilinson’s conjecture: For genus-1/2/3 polynomials, Mahler measure is (conjecturally) a rational multiple of $L'(E, 0)$ for an associated elliptic curve $E$, or sums of such multiples for higher genus and split Jacobians (Bornhorn, 2015, Liu et al., 2019, Trieu, 2023).

A general exactness theorem in three variables provides Mahler measure formulas in terms of elliptic $L$-values and Bloch-Wigner dilogarithms, conditional on Beilinson's conjecture (Trieu, 2023).

4. Explicit Calculations, Special Families, and $L$-Value Identities

• Boyd’s conjectures and modular examples: For many parametric families, e.g. $P_{k}(x,y)=x+1/x+y+1/y+k$, Rodriguez Villegas and successors proved that $m(P_k)$ can be written as Kronecker–Eisenstein series and linked to $L$-values of newforms, further extended to 28 new parametric identities for CM points (Tao et al., 2022, Meemark et al., 2019).
• Genus 2 and 3 curves: For reciprocal tempered polynomials, two or three linearly independent $K_2$ elements yield rational multiples of $L'$-values of corresponding elliptic factors, with numerical identities between different Mahler measures classified (Liu et al., 2019).
• Multivariable and higher Mahler measures: Closed formulas in terms of polylogarithms and Dirichlet $L$-values (e.g., $\zeta(3)$, $L(\chi_{-3},s)$) are established for arbitrarily many variables (Nair, 2023, Mehrabdollahei, 2021); limits such as $$\lim_{d \to \infty} m(P_d) = \frac{9}{2\pi^2}\zeta(3)$$ obtained for increasing-degree families (Mehrabdollahei, 2021).
• Mahler measure on arbitrary tori: Extension of the integration to circles of radii $r_i$ yields new formulas parametrized by geometric deformations, bridging pure Mahler measure and regulator contributions (Lalin et al., 2017).

Table: Sample Mahler Measure–$L$-Value Identities

Polynomial $P(x, y)$	Mahler measure formula	Reference
$x+1/x+y+1/y+k$	$m(k) = C_k L(f_k, 2)$, $L$ of newforms over $\mathbb{Q}$ or quadratic fields	(Tao et al., 2022)
$y^2 + 2 x y - x^3 + x$	$m(P) = 2 L'(E, 0)$	(Lalin et al., 2017)
$z + (x+1)(y+1)$	$m(P) = -2 L'(E_{15a8}, -1)$ (conditional)	(Trieu, 2023)

5. Extensions: Generalizations and Higher Mahler Measures

• Multiple and higher Mahler measures: Definitions such as generalized, multiple, and order-$k$ Mahler measures involve integrals of maxima, products, or powers of $\log|P(z)|$ and possess limit theorems reducing them to one-variable cases (Issa et al., 2012).
• Metric Mahler measures: For number fields $K$, the $t$-metric Mahler measure $m_{K, t}(\alpha)$ considers infima over multiplicative decompositions and is a genuine metric on $K^\times/\mu(K)$; infima are shown to be attained for $K = \mathbb{Q}$ and imaginary quadratic fields of class number 1 (Samuels, 2017).
• Iteration and dynamical classification: Orbits under repeated Mahler measure reflect nontrivial arithmetic dynamics; explicit law for units (infinite orbit or stopping in 1-2 steps) and construction of examples with prescribed orbit size are proven (Fili et al., 2019, Fili et al., 2021).
• Limits and convergence: For polynomials with growing degree or number of variables, Mahler measure limits converge to constants times $\zeta(3)$ or Dirichlet $L$-values, with rigorous error bounds (Mehrabdollahei, 2021, Nair, 2023).

6. Research Directions and Open Problems

Current open questions and conjectures include:

• Supremal normalized Mahler measure: Is there a universal gap $1 - \epsilon$ in $M(f)/\|f\|_2$ for Littlewood polynomials of large degree? If so, $\epsilon \le 0.049$ (Mossinghoff, 2024).
• Classification of Mahler measures for higher-genus and multivariable families: Systematic computation of exact formulas as motivated by $K$-theory and Beilinson's conjecture; determinants of Mahler measures for elliptic curves over real quadratic fields (Tao et al., 2022).
• Extension to arbitrary tori and Regulator theory: Relation to deformation of cycles, modular and motivic interpretations, and connections to higher-rank $K$-groups (Lalin et al., 2017, Brunault et al., 2017).
• Arithmetic dynamics: Distribution and statistics of wandering points in Mahler measure iteration, backward orbits, and connection to Lehmer’s problem (Fili et al., 2019, Fili et al., 2021).

7. Significance and Impact

Mahler measure acts as a quantitative bridge between analysis (toral integrals, moments), algebraic number theory (heights, Lehmer’s problem, unit equations), arithmetic geometry (regulators, $K$-theory, mixed motives), and modular forms ($L$-values, modular curves). Its explicit computation is a laboratory for testing deep conjectures such as Beilinson’s, and for exploring extremal phenomena in both additive and multiplicative contexts. Recent advances have sharpened bounds, extended the dictionary to arbitrary genus and multivariable cases, illuminated its dynamical landscape, and forged links to special values and period identities, making it a central organizing concept in modern Diophantine analysis and arithmetic geometry.