Variations on a Theme of Kronecker

From MaRDI portal





In this survey article, the author gives a summary of the present state of affairs connected with the following problem: Let \(\theta_1,\ldots,\theta_n\) be the roots of the polynomial \(P(z) = z^n +c_1z^{n-1}+ \ldots+c_n\) where \(c_1,\ldots, c_n\) are integers with \(c_n\ne 0\) and put \(M(P) := \max_i \vert\theta_i\vert\) and \(\Omega(P) := \prod_i \max(\vert\theta_i\vert,1)\). Then \(M(P)\le 1\) implies that each \(\theta_i\) is a root of unity (Kronecker, 1857). The following conjectures are still unsettled: (1) (D. H. Lehmer, 1933) If a constant \(\varepsilon >0\) exists, independent of \(n\), such that \(\Omega(P) <1+\varepsilon\), then \(\theta_1,\ldots,\theta_n\) are roots of unity. (2) (Schinzel-Zassenhaus, 1965) If \(M(P) <1 + c/n\) for a certain constant \(c\), then \(\theta_1,\ldots,\theta_n\) are roots of unity. C. J. Smyth [Bull. Lond. Math. Soc. 3, 169--175 (1971; Zbl 0235.12003)] answered both questions completely in case \(P\) is a so-called non-reciprocal polynomial. Also connections with Pisot-Vijayaraghavan numbers and Salem numbers are discussed.











This page was built for publication: Variations on a Theme of Kronecker

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4173458)