A family of measures on symmetric groups and the field with one element (Q898819): Difference between revisions

In a previous paper [Res. Number Theory 1, Article ID 7, 30 p. (2015; \url{doi:10.1007/s40993-015-0006-6})], the author and \textit{B. L. Weiss} have introduced, for each positive integer \(n\), a family \(\nu _{n,z}^*\) of complex valued measures on the symmetric group \(S_n\), depending on a complex parameter \(z\). When the parameter takes a prime power value \(q\), \(\nu _{n,z}^*\) describes splitting probabilities of monic, square-free, degree \(n\) polynomials over \(\mathbb F_q\). The paper under review is devoted to a study of the measures \(\nu _{n,1}^*\) and \(\nu _{n,-1}^*\). In Theorem 3.1 it is shown that \(\nu _{n,1}^*\) can be expressed as the sum of two measures: \(\omega _n\), which is supported on the rectangular partitions \([b^a]\) with \(ab=n\), and \(\omega_{n-1}^*\), supported on the almost-rectangular partitions \([b^a,1]\) (\(ab=n-1\)) and obtainable from \(\omega_{n-1}\) in a very simple, explicit fashion. For all \(n\geq 3\), both summands are strictly signed measures with total (signed) mass 1. As consequence of an explicit description for \(\omega_n\) given in Theorem 3.2, in Theorem 3.3 it is seen that \(\omega_n\) has an internal product structure compatible with multiplication of integers. In Section 4, the absolute probabilistic measure \(| \omega _n|\) is described as the output of a probabilistic sampling method. An analogous description is given for the signed measures \(\omega _n\) for \(n\geq 2\). The information available on \(\omega _n\) is expanded in Section 5 by studying the measures from the standpoint of representation theory. Thus, \(n! | \omega _n|\) is the character of a representation of \(S_n\), completely described in Theorem 5.1. A similar description for \(-(2n)! \omega_{2n}\) is available in Theorem 5.2 and for \((-1)^n n! \omega_{n-1}^*\) in Theorem 5.4. All these results entail that \(n!\nu_{n,1}^*\) is the character of a virtual representation of \(S_n\). Similar results on the even simpler structure of \(\nu_{n,-1}^*\) are given in Section 6. This very well written paper ends with suggestions for extending the reported study by considering related questions.