Lehmer’s Polynomial & Mahler Measure

• Lehmer’s polynomial is a unique monic, non-cyclotomic integer polynomial distinguished by its minimal known Mahler measure exceeding 1.
• It features one real Perron root (≈1.17628) and nine complex roots within or on the unit circle, underscoring its extremal properties.
• The polynomial informs several fields, including algebraic number theory, spectral graph theory, and dynamical systems, guiding methods to bound Mahler measures.

Lehmer's polynomial is a central object in Mahler measure theory and algebraic number theory, distinguished by its extremal properties among monic, non-cyclotomic integer polynomials. Defined as $$L(x) = x^{10} + x^9 - x^7 - x^6 - x^5 - x^4 - x^3 + x + 1$$, it attains the minimal known Mahler measure greater than 1, commonly referred to as Lehmer's number $\tau_0 \approx 1.17628\ldots$. Lehmer’s problem, originally posed in 1933, asks whether there exists a strictly positive lower bound $C > 1$ for the Mahler measure $M(P)$ of any non-cyclotomic $P \in \mathbb{Z}[x]$, a question fundamentally linked to the arithmetic of algebraic integers, dynamical systems, and spectral graph theory.

1. Mahler Measure and Lehmer’s Conjecture

For a monic polynomial $P(x)$ with complex roots $\alpha_i$, the Mahler measure is

$M(P) = \prod_{i=1}^n \max\{1, |\alpha_i|\} \,,$

with the logarithmic variant $$m(P) = \log M(P) = \sum_{i=1}^n \max\{0, \log |\alpha_i|\}$$. The measure satisfies $M(P) \geq 1$, equality holding if all roots lie on or inside the unit circle. Lehmer’s conjecture requests a constant $C > 1$ so that $M(P) \geq C$ for all non-cyclotomic $P \in \mathbb{Z}[x]$. Lehmer’s polynomial, $L(x)$, exhibits one real root (the Perron root, numerically $\tau_0$) and nine nonreal roots of modulus $\leq 1$ (Greaves et al., 2012, Johannesson, 2023), providing the minimal example known to date.

2. Salem Numbers, Pisot Numbers, and Extremal Polynomials

Salem numbers are algebraic integers $\lambda > 1$ whose conjugates are either on the unit circle or satisfy $|\gamma| = 1$. Pisot numbers, in contrast, are totally real algebraic integers greater than 1, with all conjugates inside the unit circle. Smyth's theorem puts a lower bound $M(P) \geq \theta_0 \approx 1.32471\ldots$ for nonreciprocal polynomials ($\theta_0$ being the real root of $x^3 - x - 1$) (Johannesson, 2023). For Salem numbers, recent work establishes a lower bound: $\theta_{31}^{-1} = 1.08545\ldots$, the dominant root of $-1-z^{30}+z^{31}$, for the set of Salem numbers (Verger-Gaugry, 2019).

3. Lehmer’s Polynomial in Spectral Graph and Matrix Theory

Lehmer's polynomial emerges naturally as the minimal Mahler measure for reciprocal polynomials, with spectral-graph-theoretic interpretations. For Hermitian matrices over $\mathbb{Z}[i]$ (Gaussian integers) and $\mathbb{Z}[\omega]$ (Eisenstein integers), Greaves–Taylor prove that every reciprocal polynomial arising as a characteristic polynomial of such matrices is either cyclotomic ($M(P) = 1$) or satisfies $M(P) \geq \tau_0$ (Greaves et al., 2012). The spectral radius and interlacing properties (Cauchy–Fischer Theorem) provide the key technical tools, and computational techniques confirm no polynomials of smaller measure arise in this matrix class.

4. Dynamical Systems Perspective and Universal Lower Bounds

Recent major advances employ dynamical systems, specifically Parry upper functions and Artin–Mazur dynamical zeta functions, to encode the distribution of conjugates (Galois orbit measures) of algebraic integers. The $\beta$-shift and its expansion reveal “lenticuli”—arcs along which conjugate roots equidistribute near the unit circle, and methods from Rouché’s theorem and Poincaré asymptotics are used to bound Mahler measures. A universal minorant $\Lambda_r \mu_r \approx 1.154 > 1$ is constructed, and for reciprocal algebraic integers $\alpha$ with dynamical degree at least 259, the universal lower bound $\theta_{259}^{-1} > 1$ is established, settling Lehmer’s conjecture for this regime (Verger-Gaugry, 2019).

5. Integer Sequences Associated to Lehmer’s Problem

Auxiliary integer sequences allow reformulation and investigation of Lehmer's conjecture. Consider the family of “power-polynomials” $$P_n(x) = \prod_{j=1}^d (x - \alpha_j^n) \in \mathbb{Z}[x]$$ with discriminants $\Delta(P_n)$. Various bounds link Mahler measure and discriminants:

$$M(P) = \limsup_{n \to \infty} |\Delta(P_n)|^{1/(2nd)}\,,$$

and Möbius inversion generates the “essential sequence” $\delta_n(P)$, encoding arithmetic properties and divisibility (Johannesson, 2023). Resultant-sequences $U(n)$ and quotient-sequences $b_n$ provide further algebraic structure, with rational generating functions that encode Mahler measure through analytic properties.

6. Cyclic Quintic Fields Arising from Lehmer’s Polynomials

Emma Lehmer introduced a notable quintic parametric family,

$$P_n(x) = x^5 - n x^4 - (n^2 + 5n + 10)x^3 - (n^3 + 5n^2 + 15n + 25)x^2 - (n^4 + 5n^3 + 15n^2 + 25n + 25)x - (n^5 + \dots) \,,$$

which defines cyclic quintic fields $K_n = \mathbb{Q}(\theta)$, with cyclic Galois group of order 5. The existence of normal integral bases in tamely ramified cases ($5 \nmid n$) is guaranteed, and explicit generators in terms of roots and Lucas-type sequences are constructed, generalizing classical results (Hashimoto et al., 2022). The arithmetic and Galois module structure of $K_n$ offers insights into explicit base constructions.

7. Status and Open Problems

Comprehensive results confirm Lehmer’s number as the minimal known measure for reciprocal non-cyclotomic integer polynomials. No nontrivial $M(P) < \tau_0$ is known, and matrix, spectral, and dynamical techniques produce tight bounds in quadratic and cyclotomic integer rings (Greaves et al., 2012). For Salem numbers, the interval $$(\theta_{31}^{-1}, \theta_{12}^{-1})=(1.08545, 1.17295)$$ essentially contains only Lehmer’s example, and any new Salem number in this interval must possess degree at least 44 and additional lenticular conjugates—a phenomenon not observed so far (Verger-Gaugry, 2019). Sequences such as $\Delta(P_n)$, $\delta_n(P)$, and $b_n$ remain central in formulating Lehmer-type lower bounds and potential new approaches. Although the conjecture is resolved for large dynamical degree and key matrix classes, prospects for extending unconditional bounds beyond quadratic rings and to all integer polynomials persist as major avenues for research.

---

Object/Class	Minimal Known Mahler Measure	Reference/arXiv id
Monic non-cyclotomic $\mathbb{Z}[x]$	$\tau_0 \approx 1.17628\ldots$	(Greaves et al., 2012, Johannesson, 2023)
Salem numbers	$> \theta_{31}^{-1} = 1.08545\ldots$	(Verger-Gaugry, 2019)
Pisot numbers	$> \theta_0 = 1.32471\ldots$	(Johannesson, 2023)
Hermitian matrices over $\mathbb{Z}[i],\mathbb{Z}[\omega]$	$\geq \tau_0$	(Greaves et al., 2012)

Lehmer’s polynomial thus remains the fundamental extremal case in Mahler measure theory, informing the structure of algebraic integers, dynamical zeta functions, and the ongoing pursuit of universal lower bounds.