The Barnes-Hurwitz zeta cocycle on \(\operatorname{PGL}_2(\mathbb{Q})\) (Q2167491)

The author uses Barnes' double zeta-function given for positive real \(z,\omega_1,\omega_2\) and complex \(s\) with \(\Re(s) >2\) by \[ \zeta_2(z,\binom{\omega_1}{\omega_2},s) := \sum_{m=0}^\infty \sum_{n=0}^\infty (z + m\omega_1 + n\omega_2)^{-s}, \] and meromorphically extended (in \(s\)) to the complex plane in order to define a certain non-vanishing 1-cocycle on \(G=\mathrm{PGL}_2(\mathbb{Q})\) with values in a certain module \(\mathcal{D}\) of distributions. More precisely, the distributions under consideration are certain maps defined on triples \((x,\omega,s)\) with \(x\in \mathbb{Q}^2 / \mathbb{Z}^2,\) \(\omega = \binom{\omega_1}{\omega_2}\) with incommensurable real \(\omega_1,\omega_2,\) and which depend meromorphically on \(s,\) have values in \(\mathbb{P}^1(\mathbb{C})\) and obey several other conditions. If \(\gamma\) is an integer primitive \(2\times 2\)-matrix with non-zero determinant and class \([\gamma]\in G,\) then for \(\nu\in \mathcal{D}\) a new distribution \([\gamma]\cdot \nu\) is defined by \[ ([\gamma]\cdot \nu)(x,\omega,s) := \operatorname{sgn}(\det(\gamma))\cdot \sum_{\mu\in \mathbb{Z}^2 / \mathbb{Z}^2\gamma^\top} \nu((x+\mu)(\gamma^\top)^{-1}, \gamma\top \omega , s). \] This definition gives the \(G\)-module structure on \(\mathcal{D}\). The author then defines a 1-cocycle \(\mathfrak{Z} \in Z^1(G,\mathcal{D})\) which turns out to be parabolic and to have non-zero cohomology class. The values of the distribution \(\mathfrak{Z}\left[\begin{matrix} 0 & -1 \\ 1 & 0 \end{matrix}\right]\) at \((x,\omega,s)\) are calculated to be (if \(x_1,x_2\not\in \mathbb{Z}\)) \[ \zeta_2(\langle x_1\rangle \omega_1 + \langle x_2\rangle \omega_2, \omega , s), \] where \(\langle r\rangle\) denotes the smallest positive element in \(r+\mathbb{Z}.\) If \(x_1\) or \(x_2\) are integers, there is an extra summand involving the Hurwitz zeta function. Lastly the author is able to show that there are no non-zero \(G\)-invariant elements in \(\mathcal{D},\) using elementary facts about real quadratic fields.