Lehmer’s conjecture
Is the Mahler measure of every integer polynomial either exactly 1 or at least some fixed constant greater than 1?
Summary
Why it is open
The Mahler measure M(P) of a polynomial with integer coefficients is the absolute value of its leading coefficient times the product of the absolute values of its roots outside the unit circle; equivalently, it is the geometric mean of |P| on the unit circle [3]. It is at least 1, and by a theorem of Kronecker it equals 1 only when, up to sign, the polynomial is a product of cyclotomic polynomials, whose roots are roots of unity, and a power of the variable [2, 3].
D. H. Lehmer, who was looking for large primes, asked whether monic integer polynomials (leading coefficient 1) with measure greater than 1 could have measure arbitrarily close to 1 [3]. The "conjecture" that they cannot is named after him, "although Lehmer did not in fact make this conjecture", Chris Smyth notes [3]. For an algebraic number α of degree d, with M(α) the measure of its minimal polynomial, the Weil height is h(α) = log M(α)/d [3], and the conjecture says that h(α) > C/d for some absolute constant C > 0 whenever α is non-zero and not a root of unity [7]. A strong version asks whether the constant can be taken to be the measure of Lehmer's own polynomial [5]. No prize sets the bar: a proof of a uniform constant, or a sequence of polynomials whose measures fall towards 1, would settle the question.
Two features make it hard. Known lower bounds for all algebraic numbers tend to 1 as the degree grows, and Smyth wrote in 2008 that "no-one has been able to improve the dependence on the degree d" in Dobrowolski's 1979 bound [3]. And reciprocal polynomials, whose coefficients read the same in both directions, "seem to be the most difficult to work with, perhaps because cyclotomic polynomials are reciprocal" [3]; the polynomial with the smallest known measure is one of them [5].
What is known
Lehmer's polynomial (1933). Lehmer's polynomial z^10 + z^9 − z^7 − z^6 − z^5 − z^4 − z^3 + z + 1 has Mahler measure 1.176280818…, and "to this day no-one has found a smaller value" above 1, Smyth wrote in 2008 [3]. Peter Borwein, Edward Dobrowolski and Michael Mossinghoff in 2007, and a September 2026 preprint by Jean-Marc Sac-Épée and colleagues, still describe it as the smallest known [2, 8]. One of its roots, about 1.176, here called Lehmer's number, is a Salem number: a real algebraic integer greater than 1, of degree at least 4, whose other conjugates, apart from its reciprocal, all have absolute value 1 [5]. Lehmer found it before Salem numbers had been defined [5].
Bounds for all algebraic numbers (1979–2019). In 1979 Dobrowolski proved that a non-zero algebraic integer α of degree d that is not a root of unity has M(α) > 1 + (1 − ε)(log log d / log d)^3 when d is large enough in terms of ε, and that the constant 1/1200 in place of 1 − ε works for every d [1]. P. Voutier later obtained the constant 1/4 for all d ≥ 2 [3]. In 2019 Vesselin Dimitrov wrote that Dobrowolski's bound "remains the best known on the Lehmer problem in the case of Salem numbers" [9].
Classes where it is proved (1971–2025). For non-reciprocal algebraic integers α, those not conjugate to 1/α, Smyth proved in 1971 that M(α) is at least 1.3247…, the real root of z^3 − z − 1, and this constant is best possible [1, 3]. A. Schinzel showed that totally real algebraic integers α ≠ 0, ±1, whose conjugates are all real, have M(α) ≥ φ^(d/2), where φ is the golden ratio [3]. Francesco Amoroso and S. David settled the problem when Q(α) is a Galois extension (1999), and Amoroso and R. Dvornicich showed M(α) ≥ 5^(d/12) when Q(α) is abelian (2000) [3]. Borwein, Dobrowolski and Mossinghoff proved that a polynomial of degree n − 1 with all coefficients odd and no cyclotomic factor satisfies log M(f) ≥ ((log 5)/4)(1 − 1/n) [2]. A 2025 preprint by Anup Dixit and Sushant Kala, not peer reviewed, proves the conjecture for α of degree d with at least a constant times √(d log d) conjugates in a finite extension of the p-adic numbers [7].
Searches (1933–2008). Lehmer himself found the smallest measures for low degrees [3]. David Boyd found all reciprocal α with measure at most 1.3 up to degree 20, Mossinghoff extended this to degree 24, and V. Flammang, M. Grandcolas and G. Rhin proved that the list of 47 known Salem numbers below 1.3 is complete up to degree 40 [3]. In 2008 Mossinghoff, Rhin and Qiang Wu determined the smallest measure of a primitive irreducible noncyclotomic integer polynomial of each even degree up to 54, and all such polynomials with measure below 1.3 up to degree 44 [4].
Recent results. The Schinzel–Zassenhaus conjecture of 1965, that a non-zero algebraic integer of degree d that is not a root of unity has a conjugate of absolute value at least 1 + c/d for some absolute constant c > 0, is implied by Lehmer's conjecture [3]. Dimitrov posted a proof of it in December 2019: an integer polynomial of degree n with constant term 1 is either a product of cyclotomic polynomials or has a root in the disc |z| ≤ 2^(−1/(4n)) [9]. Annals of Mathematics accepted the paper on 27 August 2026 [9]. Because the implication runs from Lehmer's conjecture to this one, the proof does not settle Lehmer's.
Main approaches
Auxiliary functions and congruences. Dobrowolski used the fact that, for each prime p, the resultant of the minimal polynomials of α and α^p is an integer multiple of p^d [3]. Strongest evidence: bounds of this form are the best general ones [2, 9]. Main difficulty: the dependence on the degree has not been improved [3], and Dimitrov describes the case of Salem numbers as "the opposite of the Schinzel-Zassenhaus case" that his method treats [9].
Restricted classes. Prove the conjecture for classes defined by symmetry, arithmetic or coefficients [2, 3, 7]. Strongest evidence: results of "Lehmer strength" or stronger for the non-reciprocal, totally real, Galois, abelian and odd-coefficient cases [2, 3]. Main difficulty: reciprocal polynomials, which include the minimal polynomials of Salem numbers, seem the hardest [3, 5].
Computation. Search for small measures [3, 4], and for limit points of the set of measures, since the measure of a polynomial in several variables is a limit of one-variable measures [3]. Earlier searches for limit points used genetic algorithms; the 2026 preprint adapts methods of missing-data restoration [8]. Strongest evidence: no smaller measure than Lehmer's has been found [2, 3, 8], and lists of small measures are complete up to degree 44 [4, 5]. Main difficulty: each search covers only bounded degrees [4].
Geometry and reformulation. Salem numbers correspond to closed geodesics on certain arithmetic hyperbolic surfaces, a geodesic of length 2 log τ for the Salem number τ, so there is a smallest Salem number exactly when there is a shortest such geodesic [5]. Eriko Hironaka proved that among Coxeter groups generated by reflections in the sides of polygons in the hyperbolic plane, the slowest growth rate is Lehmer's number [5]; Lehmer's polynomial with z replaced by −z is the Alexander polynomial of the (−2, 3, 7) pretzel knot [5]; and Curtis McMullen proved that automorphisms of compact complex surfaces with positive entropy have entropy at least the logarithm of Lehmer's number [5]. Emmanuel Breuillard and Péter Varjú restated the conjecture as a counting problem modulo primes [6]. Strongest evidence: these results give Lehmer's number a role in geometry, topology and dynamics [5]. Main difficulty: the geometric statements concern only Salem numbers [5], and Breuillard and Varjú write that their result "should not be regarded as making any progress towards a positive answer" [6].
Open sub-questions
- Even degrees beyond 54. Does any primitive irreducible noncyclotomic integer polynomial of even degree from 56 to 64 have Mahler measure below 1.17628 [4]?
- Dimitrov's method for Salem numbers. Can the method that proved the Schinzel–Zassenhaus conjecture give a lower bound for Salem numbers stronger than Dobrowolski's [1, 9]?
- A lower p-adic threshold. Does Lehmer's conjecture hold for α of degree d with at least a constant times √d conjugates in a fixed finite extension of the p-adic numbers, below the √(d log d) of Dixit and Kala [7]?
- Counting wild primes. Lehmer's conjecture is equivalent to almost no prime being C-wild for some C > 1, in the sense of Breuillard and Varjú [6]. For small C, does the proportion of C-wild primes fall as the primes grow, as far as computation reaches?
- A shorter geodesic. Does some arithmetic hyperbolic surface of the kind that corresponds to Salem numbers have a closed geodesic shorter than 2 log 1.17628, which would give a smaller Salem number [5]?
Work on this in Papex
If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.
Key references
- Dobrowolski, E. (1979). On a question of Lehmer and the number of irreducible factors of a polynomial. Acta Arithmetica 34(4), 391–401. 10.4064/aa-34-4-391-401 · Checked 16 Oct 2026
- Borwein, P., Dobrowolski, E. & Mossinghoff, M. J. (2007). Lehmer's problem for polynomials with odd coefficients. Annals of Mathematics 166(2), 347–366. 10.4007/annals.2007.166.347 · Checked 16 Oct 2026
- Smyth, C. (2008). The Mahler measure of algebraic numbers: a survey. In J. McKee & C. Smyth (eds.), Number Theory and Polynomials, London Mathematical Society Lecture Note Series 352, Cambridge University Press, 322–349. 10.1017/CBO9780511721274.021 · Checked 16 Oct 2026
- Mossinghoff, M. J., Rhin, G. & Wu, Q. (2008). Minimal Mahler measures. Experimental Mathematics 17(4), 451–458. 10.1080/10586458.2008.10128872 · Checked 16 Oct 2026
- Smyth, C. (2015). Seventy years of Salem numbers. Bulletin of the London Mathematical Society 47(3), 379–395. 10.1112/blms/bdv027 · Checked 16 Oct 2026
- Breuillard, E. & Varjú, P. P. (2019). On the Lehmer conjecture and counting in finite fields. Discrete Analysis 2019:5, 8 pp. 10.19086/da.8306 · Checked 16 Oct 2026
- Dixit, A. B. & Kala, S. (2025). A p-adic criterion for Lehmer's conjecture. Preprint, arXiv:2507.20141 (version 1, 27 July 2025), not peer reviewed. Source · Checked 16 Oct 2026
- Sac-Épée, J.-M., El Otmani, S., Maul, A. & Rhin, G. (2026). Finding new limit points of Mahler measure by methods of missing data restoration. Preprint, arXiv:2609.20468 (version 1, 17 September 2026), not peer reviewed. Source · Checked 16 Oct 2026
- Dimitrov, V. (to appear). The Schinzel–Zassenhaus conjecture. Annals of Mathematics, received 25 December 2019, accepted 27 August 2026, listed as "to appear in forthcoming issues" with no volume or DOI on 2026-10-16; first posted as arXiv:1912.12545 (28 December 2019). Source · Checked 16 Oct 2026
Source
Open, as listed by Wikipedia · Unsolved problems in mathematics. Checked 08 Oct 2026.
The status is the source’s. Papex does not decide whether a problem is solved.
Records on this problem
No public record addresses this problem yet.