MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider a reductive group over a local field. What is the normalizer of a maximal compact subgroup?

If this is to general, what is the normalizer of $GL(n, \mathbb{Z}_p)$ in $GL(n, \mathbb{Q}_p)$, $U(n)$ in $GL(n, \mathbb{C})$, and $O(n)$ in $GL(n, \mathbb{R})$?

share|cite|improve this question
This is related to previous questions:…… – Alain Valette Dec 17 '11 at 21:16
up vote 4 down vote accepted

Hints :

-- For $K= {\rm GL}(n,{\mathbb Z}_p )$. Make $G={\rm GL}(n,{\mathbb Q}_p )$ acts on ${\mathbb Z}_p$-lattices of ${\mathbb Q}_p^n$. Prove that the lattices stabilized by $K$ are the $p^k {\mathbb Z}_p^n$, $k\in {\mathbb Z}$. Observe that the normalizer $\tilde K$ of $K$ permutes these lattices and conclude that ${\tilde K}={\mathbb Q}_p^\times K$.

-- For $K=O(n)$ (or $U(n)$). Do someting similar by making $G$ act on the set of positive definite symmetric (hermitian) matrices via $(X,A)\mapsto XA{\bar X}^{t}$.

In general, for a non-archimedean base field, you can make $G$ act on the extended Bruhat-Tits building, but the answer is going to be technical according to whether $G$ has a center or not, is simply connected or not. Over archimedean fields, I guess you have to use symmetric spaces.

share|cite|improve this answer
Seems like you know the answer to every question, I can come up with;) Thx. – Marc Palm Dec 18 '11 at 10:13

$SO(n)$ is not only maximal compact in $SL_n(\mathbb{R})$, it is even a maximal subgroup! Same for $SU(n)$ in $SL_n(\mathbb{C})$. This follows e.g. from Exercise A.3 in Chapter VI in S. Helgason, ``Differential geometry, Lie groups and symmetric spaces'', Academic Press, 1978. From this the normalizers of $O(n)$ in $GL_n(\mathbb{R})$, and of $U(n)$ in $GL_n(\mathbb{C})$, are easily determined.

To establish maximality of $K=SL_n(\mathbb{Z}_p)$ as a subgroup of $G=SL_n(\mathbb{Q}_p)$, there is a cute argument using the Howe-Moore theorem, telling you that, if $\pi$ is a unitary representation of $G$ without non-zero fixed vectors, then all coefficients of $\pi$ go to $0$ at infinity on $G$. So assume that $U$ is a proper subgroup of $G$, containing $K$. Since $K$ is open, $U$ is open too. By simplicity of $G$, the subgroup $U$ cannot have finite index. Let $\pi$ be the quasi-regular representation of $G$ on $\ell^2(G/U)$, it has no non-zero fixed vector; let $\delta_U\in\ell^2(G/U)$ be the characteristic function of the base-point of $G/U$, corresponding to the trivial coset of $U$. Then the coefficient $g\mapsto<\pi(g)\delta_U|\delta_U>$ is constant and equal to 1 on $U$, so by Howe-Moore $U$ must be compact. Since $K$ is maximal compact, we have $U=K$. From this you also deduce the normalizer of $GL_n(\mathbb{Z}_p)$ in $GL_n(\mathbb{Q}_p)$.

Ref: R.E. Howe and C.C. Moore. Asymptotic properties of unitary representations. J. Funct. Anal. 32, 72-96, 1979.

EDIT: For the maximality of $SO(n)$ in $SL_n(\mathbb{R})$, the strategy of the proof is nicely explained in

share|cite|improve this answer
Both answers are very useful. I checked the other simply because it was first. Thanks. – Marc Palm Dec 18 '11 at 10:12
Of course, one must be careful not to rely too much on this sort of luck; $\operatorname{PGL}_n(\mathbb Z_p)$ is not a maximal subgroup of $\operatorname{PGL}_n(\mathbb Q_p)$! – L Spice Jan 13 '15 at 20:35

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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