Actions of some simple compact Lie groups on themselves (Q2274780)

Bredon observed in his review of Wu Yi Hsiang's book that the action of the Lie group \({\mathrm{SU}}(3)\) on itself by \(A\cdot B := ABA^t\) is a counterexample to a conjecture of \textit{W. Y. Hsiang} [Cohomology theory of transformation groups. Berlin-Heidelberg-New York: Springer-Verlag (1975; Zbl 0429.57011), p. 146] which claimed that a nontransitive, nontrivial action of a compact connected simple Lie group \(G\) on itself is structurally similar to the adjoint action of \(G\), for instance, because the principal isotropy group of this action is not a maximal torus. In the paper at hand the authors give some indication that a weaker form of Hsiang's conjecture might be true: they show that for certain simple Lie groups \(G\), namely \({\mathrm{SO}}(n)\), \(n\geq 34\), and \({\mathrm{SU}}(3)\), the principal isotropy groups of any nontrivial action of G on itself with nonempty fixed point set are maximal tori. Hsiang's conjecture is, however, not shown for these classes of actions, as it additionally asks for the structure of the fixed point set of a maximal torus \(T\), its weight system of local \(T\)-representations, as well as the action of the Weyl group on the \(T\)-fixed point set.