Quasi-semisimple actions on reductive groups and buildings

Loren Spice (TCU)

 (joint with Adler and Lansky) In the early 21st century, Prasad and Yu proved that, for a field of characteristic exponent p \geq1, the identity component of the group of fixed points of a finite group \Gamma of prime-to-p order acting on a reductive group \tilde{G} is itself a reductive group; and that, for a local field of residual characteristic p, the building of the resulting fixed-point group is the fixed-point set for the action of \Gamma on the building of \tilde G.

Applications by Adler and Lansky to lifting require analogous statements for quasisemisimple actions, which are those for which there are a Borel subgroup \tilde{B} of \tilde{G}, and a maximal torus in \tilde{B}, that are preserved by the action of \Gamma.  The fixed-point group remains reductive in this case, at least after smoothing.  Unfortunately, the directly analogous statement about buildings is too strong; but, fortunately, the directly analogous statement about buildings is too strong, so that we can explore the interesting ways that the building of the fixed-point group can fail to fill out the full fixed-point set in the building of \tilde G, and conditions that recover the exact analogue of the Prasad–Yu statement.