1
$\begingroup$

$\DeclareMathOperator\U{U}\DeclareMathOperator\GL{GL}$Let $E/F$ be a quadratic extension of local fields and $G=U(V)$ a unitary group associated to hermitian space $V$ over $E/F$. We fix a minimal parabolic subgroup $P_0$ of $G$ and call $P=NM$ a standard parabolic subgroup of $G$ if it contains $P_0$. Write $M=\GL_{a_1}(E) \times \dotsb \times \GL_{a_r}(E) \times \U(W)$, where $W$ is a subspace of $V$.

Let $A_0$ be the maximal split torus of $G$ and $N_G(A_0)$ and $Z_G(A_0)$ its normalizer and centralizer respectively. If $P'=N'M'$ is another standard parabolic subgroup (i.e., there is a Weyl group element $w \in N_G(A_0)(F) / Z_G(A_0)(F)$ such that $w \cdot P=P'$), then I am wondering whether $M'$ should be of the form $\GL_{b_1}(E) \times \dotsb \times \GL_{b_r}(E) \times \U(W)$ and $\{b_1,\dotsc,b_r\}$ is a permutation of $\{a_1,\dotsc,a_r\}$?

And if $\rho=\rho_1 \boxtimes \dotsb \boxtimes \rho_r \boxtimes \tau$ is a representation of $M$, then $w\cdot \rho$ is also the form of $\rho_{S(1)} \boxtimes \dotsb \boxtimes \rho_{S(r)} \boxtimes \tau$, where $S$ is a permutation of $\{1,2,\dotsc,r\}$?

$\endgroup$
5
  • 4
    $\begingroup$ Dear Monty, I suggest that you edit your question. Make an effort to explain your notation (and maybe to correct grammar and typos ... ). What is "another associated standard parabolic subgroup"? Associated to what? What are standard parabolic subgroups if you have not chosen a maximal torus and a Borel subgroup? Where does your element $w$ live? In $F$-points or in $\bar F$-points? The answer to your question might depend on all this. $\endgroup$ Commented Apr 18, 2020 at 17:24
  • $\begingroup$ @Mikhail, Oh, I have read your comment so late. I am very sorry for not explaing the notations in detail. I corrected my question more precisely. I would appreciate if you see it again. $\endgroup$
    – Monty
    Commented Apr 22, 2020 at 8:09
  • $\begingroup$ The question is still not clear. First, $w\cdot P$ is not standard. Second, you write: "then I am wondering whether $M'$ should be of the form ${\rm GL}_{b_1}(E) \times \dotsb \times {\rm GL}_{b_r}(E) \times U(W)$..." What do you mean by "of the form"? $\endgroup$ Commented Apr 22, 2020 at 14:06
  • $\begingroup$ Anyway, if you state your question clearly and read the description of the Weyl group on page 1272 of the reference in the answer of "Not a grad student", you can answer your question yourself, at least in the case when $G$ is quasi- split. Concerning the general case, you can ask a question at Math.StackExchange.com for the description of the Weyl group and the root system of your group $G$. $\endgroup$ Commented Apr 22, 2020 at 16:41
  • $\begingroup$ @Mikhail, thank you! The all I wanted for is on the reference of Goldberg. $\endgroup$
    – Monty
    Commented Apr 23, 2020 at 23:49

1 Answer 1

1
$\begingroup$

Not quite, since the Weyl group of the quasi-split unitary group is not the symmetric group. See page 1272 of Goldberg, "R-Groups and Elliptic Representations of Unitary Groups," http://www.jointmathematicsmeetings.org/proc/1995-123-04/S0002-9939-1995-1224616-6/S0002-9939-1995-1224616-6.pdf

$\endgroup$
5
  • $\begingroup$ The reference in the answer has no page 1272. $\endgroup$ Commented Apr 18, 2020 at 18:46
  • $\begingroup$ Thanks. There are several papers by David Goldberg with similar titles. $\endgroup$ Commented Apr 18, 2020 at 19:05
  • $\begingroup$ This seems to answer the converse of the question. $\endgroup$
    – LSpice
    Commented Apr 18, 2020 at 19:30
  • $\begingroup$ Thank you very much! The reference you informed me is the very one I am searching for. There is another action like taking contragredient representation. But I am still wondering why for $GL(n)$, there is no such involution action but only permutation. Do you know some intuitive reason for why there does not Weyl group element corresponding to taking $g^{-1}$ to each components of $GL(n_i)$? $\endgroup$
    – Monty
    Commented Apr 25, 2020 at 17:41
  • $\begingroup$ @May I ask you one more? Is there a similar involution Weyl group action for other classical group like symplectic group? $\endgroup$
    – Monty
    Commented Apr 26, 2020 at 6:02

Not the answer you're looking for? Browse other questions tagged .