Suppose $G$ is a finite group with a split $BN$-pair, satisfying the commutator relations, and such that representatives in $N$ of elements of the Weyl group can be chosen in a nice way: if $W\ni w=s_{i_1}\cdots s_{i_r}$, then $N\ni\dot{w}=\dot{s}_{i_1}\cdots\dot{s}_{i_r}$. (I'm essentially assuming $G$ behaves like a finite group of Lie type $G^F$.)

Let $(W,S)$ denote the corresponding Coxeter system, and $T$ be a maximal torus. I'm assuming all characters are complex characters in what follows. Suppose for $J\subseteq S$, you take an irreducible character $\chi\in\operatorname{Irr}(L_J)$ on the standard Levi $L_J$ such that the generalized restriction/Harish-Chandra restriction $^\ast R^{L_J}_T(\chi)\neq 0$. For $J\subseteq I\subseteq S$, are the irreducible constituents of the generalized induction character $R^{L_I}_{L_J}(\chi)$ again characters $\lambda\in\operatorname{Irr}(L_I)$ such that $^\ast R^{L_I}_T(\lambda)\neq 0$?

Taking $\mu\in\operatorname{Irr}(L_I)$ such that $^\ast R^{L_I}_T(\mu)=0$, I wondering if $(\mu,R^{L_I}_{L_J}(\chi))=0$, but the various formulas I know like the Mackey Intertwining Number don't seem to lend themselves well here.


By that adjointness we have that $(\mu,R^{L_I}_{L_J}\chi)=({}^*R^{L_I}_{L_J}\mu,\chi)$. If this value is not zero, then it means that ${}^*R^{L_I}_{L_J}\mu$ contains a non-zero multiple of $\chi$, say it is $m\chi$. Then by the transitive property we see that ${}^*R^{L_I}_{T}\mu$ contains $m({}^*R^{L_J}_{T}\chi)\neq0$, a contradiction.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.