Twin buildings and applications to \(S\)-arithmetic groups (Q2563840): Difference between revisions

This lecture notes volume has a double purpose, namely to present results about finiteness properties of \(S\)-arithmetic groups and to give the necessary background on twin buildings. A twin building, as defined by Ronan and Tits, is a pair of buildings together with a relation between them called the opposition relation. A key example for this work is the following. Let \(\mathcal G\) be a Chevalley group defined over \(\mathbb{F}_q(t)\), the field of rational functions in one variable over the finite field \(\mathbb{F}_q\). To the group \(G={\mathcal G}(\mathbb{F}_q[t,t^{-1}])\) belong two buildings, namely the Bruhat-Tits buildings \(\Delta_-\) and \(\Delta_+\) associated to the groups \({\mathcal G}(\mathbb{F}_q((t)))\) and \({\mathcal G}_q(\mathbb{F}((t^{-1})))\), and they are twinned in such a way that there are opposite vertices whose stabilizer in \(G\) are \({\mathcal G}(\mathbb{F}_q[t])\) and \({\mathcal G}(\mathbb{F}_q[t^{-1}])\), respectively. This point of view gives additional insight into the action of \(\Gamma={\mathcal G}(\mathbb{F}_q[t])\) on the Bruhat-Tits building \(\Delta_+\). E.g., it gives a more transparent proof of a theorem of Soulé stating that there is a fundamental domain for the action of \(\Gamma\) on \(\Delta_+\) which is a quarter (quartier) in \(\Delta_+\). Thus the first chapter of this volume is develoted to group actions on twin buildings. Particularly noteworthy is the section on coprojections and coconvexity. These notions are the analogs of the familiar notions of projections and convexity for buildings where the co-notions use the codistance derived from the opposition relation in place of the distance of chambers which is used for the usual notions. The finiteness properties of groups alluded to above are finite generatedness (\(=F_1\)), finite presentability (\(=F_2\)) and analogous higher finiteness conditions \(F_n\). To prove that a group \(\Gamma\) has these properties one uses a criterion due to K. Brown, as follows. Suppose an action of \(\Gamma\) on a certain complex \(X\) is given, in our situation the Bruhat-Tits building. One has to find a filtration of \(X\) by \(\Gamma\)-invariant subcomplexes \(X_n\), \(n\in\mathbb{N}\), such that one can compute the map induced in homotopy by the inclusion \(X_n\hookrightarrow X_{n+1}\). This is done by showing that certain subcomplexes of the Tits-building of \(\mathcal G\) over \(\mathbb{F}_q\) (sic, not the Bruhat-Tits building of \(\mathcal G\) over the local field \(\mathbb{F}_q((t^{-1}))\)) are spherical, i.e. homotopy equivalent to a bouquet of spheres. The homotopy properties of these complexes are studied in chapter II. The conclusions for finiteness properties of \(\Gamma\) are presented in chapter III. One of the main results of this work is the following theorem. Theorem C. Let \(\mathcal G\) be an absolutely almost simple classical \(\mathbb{F}_q\)-group of \(\mathbb{F}_q\)-rank \(n\geq 1\). If \(q\) is big enough, depending on \(n\), (\(q\geq 2^{2n-1}\) suffices), then \({\mathcal G}(\mathbb{F}_q[t])\) is of type \(F_{n-1}\) but not to type \(FP_n\). The group \({\mathcal G}(\mathbb{F}_q[t,t^{-1}])\) is of type \(F_{n-1}\).