A geometric approach to complete reducibility. (Q2484074)

Let \(G\) be a connected reductive linear algebraic group. A subgroup \(H\) is called \(G\)-completely reducible provided that, whenever \(H\) is contained in a parabolic subgroup \(P\) of \(G\), it is contained in a Levi subgroup \(L\) of \(P\). This notion is due to \textit{J.-P.~Serre}, who studied it using the theory of Tits buildings [Astérisque 299, 195-217, Exp.~No.~932 (2005; Zbl 1156.20313)]. There are also the related notions \(G\)-irreducible and \(G\)-indecomposable subgroup. Now let \(S\) be a maximal torus of the centralizer \(C_G(H)\) of a subgroup \(H\) of \(G\). Then \textit{R.~W.~Richardson} has called \(H\) strongly reductive in \(G\) provided that \(H\) is not contained in any proper parabolic subgroup of \(C_G(H)\) [Duke Math. J. 57, No. 1, 1-35 (1988; Zbl 0685.20035)]. Now suppose that \(H\) is topologically generated in \(G\) by \(h_1,\dots,h_n\). Then Richardson showed that \(H\) is strongly reductive in \(G\) if and only if the \(G\) orbit of \((h_1,\dots,h_n)\) under the simultaneous conjugation action on \(G^n\) is closed in \(G^n\). The main result of the present paper is that \(H\) is strongly reductive in \(G\) if and only if it is \(G\)-completely reducible. The arguments are based on geometric invariant theory. The authors deduce a wealth of consequences. They also develop the theory for the case that \(G\) is reductive but maybe not connected.