Arithmetic groups have rational representation growth. (Q640806)

Suppose that \(\Gamma\) is a finitely generated group. Denote by \(r_n(\Gamma)\) the number of irreducible \(n\)-dimensional complex representations of \(\Gamma\), up to equivalence. If the sequence \(r_n(\Gamma)\) is bounded by a polynomial in \(n\), then the representation zeta function \(\zeta_\Gamma(s)=\sum_n r_n(\Gamma)n^{-s}\) converges in a suitable half complex plane. \textit{A. Lubotzky} and \textit{B. Martin} [Isr. J. Math. 144, 293-316 (2004; Zbl 1134.20056)] proved that an arithmetic lattice in characteristic zero has the congruence subgroup property if and only if the sequence \(r_n(\Gamma)\) grows polynomially, or equivalently, if and only if the abscissa of convergence of \(\zeta_\Gamma(s)\) is finite. The main result in this paper is the following: if an arithmetic lattice \(\Gamma\) in characteristic zero satisfies the congruence subgroup property, then the abscissa of convergence of \(\zeta_\Gamma(s)\) is a rational number. The proof follows a general strategy of Igusa and Denef. \textit{M. Larsen} and \textit{A. Lubotzky} [J. Eur. Math. Soc. (JEMS) 10, No. 2, 351-390 (2008; Zbl 1142.22006)] proved that \(\Gamma\) contains a finite index subgroup \(\Delta\) such that the representation zeta function of \(\Delta\) has an Euler-like factorization. The author proves that the abscissa of convergence of \(\zeta_\Gamma(s)\) is unchanged when passing to a finite-index subgroup and previous results ensure that the abscissa of convergence is rational for every local factor of \(\zeta_\Delta(s)\). Therefore, in order to prove that the abscissa of convergence of the global zeta function is rational, it remains to understand the relation between the local zeta functions for different primes. Indeed, the paper is mainly an attempt to give an approximate formula to the local zeta functions, which is uniform in the prime \(p\).