Divergent torus orbits in homogeneous spaces of \({\mathbb Q}\)-rank two (Q2382332): Difference between revisions

The \(\mathbb Q\)-rank of a semisimple algebraic \(\mathbb Q\)-group \(G\) influences the geometry of arithmetic locally symmetric spaces \(\Gamma\backslash G\). \textit{G. Tomanov} and \textit{B. Weiss} [Duke Math. J. 119, No. 2, 367--392 (2003; Zbl 1040.22005)] studied the orbits of an \(\mathbb R\)-split torus \(T\), calling the orbit of the point \(\Gamma x\) divergent if \(t\mapsto\Gamma xt\) is a proper map on \(T_{\mathbb R}\), and asking if there could be a divergent \(T_{\mathbb R}\) orbit when the dimension of \(T\) exceeds the \(\mathbb Q\)-rank of \(G\). \textit{B. Weiss} [Geom. Funct. Anal. 14, No.1, 94-149 (2004; Zbl 1074.37004)] conjectured the answer is negative. The case of \(\mathbb Q\)-rank \(0\) (which makes \(\Gamma\backslash G_{\mathbb R}\) compact) or \(1\) is straightforward. In this paper the conjecture of Weiss is proved for \(\mathbb Q\)-rank \(2\), and a divergent orbit is shown to imply that the dimension of \(T\) is less than twice the \(\mathbb Q\)-rank of \(G\) in general. Since this work was done, \textit{B. Weiss} [Isr. J. Math. 152, 221--227 (2006; Zbl 1126.22007)] has gone on to show the conjecture in the general case.