On the Arthur-Selberg trace formula for \(\mathrm{GL}(2)\) (Q664238): Difference between revisions

Generalizing Selberg's work, \textit{J. Arthur} proved a trace formula (see [Clay Math. Proc. 4, 1--263 (2005; Zbl 1152.11021)]) for the smooth action \(R_r(\varphi)\) on the right regular representation \([R_r(\varphi)g](x)=\int_{G}\varphi(y)g(xy)\,dy\) where \(G\) is a reductive Lie group, \(F\) a number field with adèle ring \(\mathbb A_F\), \(g\in L^2(G(F)\backslash G(\mathbb A_F))\) and \(\varphi\) a compactly supported smooth function on \(G(\mathbb A_F)^1\). The trace formula consists on the equality \(\tau_\sigma=\tau_{\text{geom}}\) of two distributions, \(\tau_\sigma\) related to the spectral decomposition of the regular representation \(R_r\) and \(\tau_{\text{geom}}\) given by a sum of geometric terms indexed by conjugacy classes in \(G(F)\): \textit{T. Finis, E. Lapid} and \textit{W. Müller} [Ann. Math. (2) 174, No. 1, 173--195 (2011; Zbl 1242.11036)] proved the continuity of the spectral distribution \(\tau_\sigma\) on the Schwartz space \({\mathcal C}(G(\mathbb A)^1;K)\) of \(K\)-biinvariant smooth functions for \(K\) a compact subgroup of \(G(\mathbb A_F)\). In this paper, the authors consider the geometric distribution \(\tau_{\text{geom}}\) in the particular case \(G=\mathrm{GL}(2)\) (and \(F=\mathbb Q\) merely for exposition clarity): they prove its continuity on the Schwartz space \({\mathcal C}(\mathrm{GL}(2)(\mathbb A)^1;K)\). The proof is based on the precise knowledge available in the \(G=\mathrm{GL}(2)\) case for the different summands (elliptic, hyperbolic or parabolic) entering in the distribution \(\tau_{\text{geom}}\). The authors introduce a new expression for the semisimple terms valid for \(K\)-biinvariant Schwartz functions and derive absolute convergence estimates using the seminorms involving derivatives of order at most 3. In another work [Compos. Math. 147, No. 3, 784--802 (2011; Zbl 1266.11072)], the authors succeeded to prove for any reductive group \(G\) the same continuity results on the semisimple geometrical part of the distribution \(\tau_{\text{geom}}\); \textit{W. Hoffmann} [Ann. Global Anal. Geom. 34, No. 3, 233--261 (2008; Zbl 1165.11051)] proved similar absolute convergence for the so-called coarse geometric expansion.