On congruences for the traces of powers of some matrices (Q735624): Difference between revisions

From the text: In a series of recent papers, V.I. Arnold studied many questions concerning the statistics and dynamics of powers of elements in algebraic systems. In particular, on the basis of experimental data, he proposed an Euler-type congruence for the traces of powers of integer matrices as a conjecture. The proof of this conjecture was deduced from the author's theorem (obtained at the end of 2004) on congruences for the traces of powers of elements in number fields [Math. Notes 79, No. 6, 783--796 (2006); translation from Mat. Zametki 79, No. 6, 840--855 (2006; Zbl 1136.11022)]. Recently, it turned out that there also exist other approaches to congruences for the traces of powers of integer matrices. In the present paper, the author's results of 2004 (loc. cit.) are strengthened and a survey of their relations to number theory, theory of dynamical systems, combinatorics, and \(p\)-adic analysis is given. The main conclusion of this survey is that all approaches considered here ultimately reflect different points of view on a certain simple but important phenomenon in mathematics. \dots Quite recently, the author received a letter from C. Smyth in which the latter wrote that in his review of Vinberg's paper in Math. Rev., he pointed out that the mentioned Smyth's theorem on congruences for the traces of integer matrices was proved earlier by \textit{W. Jänichen} in 1921 [Sitzungsber. Berl. Math. Ges. 20, 23--29 (1921; JFM 48.0131.01)] and by \textit{I. Schur} in 1937 [Compos. Math. 4, 432--444 (1937; Zbl 0017.00202; JFM 63.0107.05)]. The same theorem was deduced by \textit{È. B. Vinberg} in [``Fermat's Little Theorem and its generalizations,'' Mat. Prosveshch., Ser. 3, No. 12, 1--11 (2008), \url{http://mi.mathnet.ru/rus/mp/v12/s3/p43}]. Thus, the theorem discussed here was proved in different years, on completely different occasions, and independently by different authors. Therefore, we will call the statement simply the Gauss congruence for the traces of integer matrices.