No label defined (Q4355801)

This paper is partly expository. It gives a very lucid and detailed account of (sometimes conjectural) relationships between special values of zeta functions of algebraic number fields, cohomology of arithmetic groups, and polylogarithms. One of the goals of the paper is to explain a generalisation, to arbitrary number fields \(F\), of the following theorem of Siegel and Klingen: if \(F/\mathbb{Q}\) is a totally real number field, \(m\geq 2\) is an even integer and \(\zeta_F\) is the zeta function \(F\), then \(\zeta_F (1-m)\) is a non-zero rational number. (The author indicates how this can be derived from a formula of Harder relating the Euler-Poincaré characteristic of \(Sp_{2n} (O_F)\) (\(O_F=\) integers in \(F\)) and the values of the zeta function \(\zeta_F\) at odd negative integers). Suppose now that \(F\) and \(M\in\mathbb{N}\) are arbitrary. Let \(d_m\) denote the multiplicity of the zero \(\zeta_F\) at the integer \(1-m\) (then \(d_m=0\) if and only if \(F\) is totally real and \(m\) is even if \(m\neq 1\)). Now \(GL_N(O_F)\) is a discrete subgroup of \(G_N= GL_N (F\otimes\mathbb{R})\). Let \(X_N\) be the associated symmetric space of \(G_N\) and let \(X_{N,u}\) be the compact dual of \(X_N\). The author first sets up an isomorphism \[ \alpha^m: PH^{2m-1} (X_{N,u},\mathbb{C})\to PH^{2m-1} (GL_N(O_F),\mathbb{C}), \tag{1} \] if \(2\leq m<\frac N4\). Here \(PH\) denotes the space of primitive cohomology classes. From (1) the author deduces that the multiplicity \(d_m\) is the dimension of the space \(PH_{2m-1} (GL_N(O_F),\mathbb{C})\) of primitive homology classes of \(GL_N(O_F)\) if \(N>4m-2\). Now both sides of (1) have natural \(\mathbb{Q}\)-structures, and the author indicates a proof that the regulator \(R(\alpha^m)\) of the isomorphism (1) is a non-zero rational multiple of \(D_F^{1/2} \zeta_F(m) (\pi i)^{-dm}\) (where \(d=\) degree of \(F\) over \(\mathbb{Q}\)). From the proof, he also recovers the Siegel-Klingen result. The author also reviews the relation between the regulator \(R(\alpha^m)\) and a related regulator \(R_m\) in the \(K\)-theory of the integer ring \(O_F\). He explains a conjecture of Zagier relating \(R_m\) to a determinant involving higher polylogarithms. If \(m=2\), Goncharov has shown that this conjecture essentially follows from the work of the author. For the entire collection see [Zbl 0864.00048].