Let $G$ be the group of rational points of a connected reductive group $\mathbb G$ defined over a non-archimedean local field $F$. For simplicity sake, I assume that $\mathbb G$ is split over $F$. Let $T$ the group of $F$-points of a maximal split torus of $\mathbb G$. Let $K$ be a good maximal compact subgroup of $G$ fixing a special vertex of the apartment of $T$. Let $\mu$ be the Haar measure on $G$ giving volume $1$ to $K$. My question is:

Is there a closed formula for the measure $\mu (KtK)$, $t\in T$ ?

In the case of ${\mathbb G}={\rm GL}(N)$, there is such a formula in Macdonald's book "Symmetric functions and Hall polynomials". But I need a reference in the general case.


Let $t=\varpi^\lambda$ where $\varpi$ is a uniformizer and $\lambda:\mathbb{G}_m\to T$ is a dominant weight. The assumption that $\lambda$ is dominant is harmless as we may conjugate by an appropriate representative $n\in K$ of a Weyl group element. Then $t$ determines a standard parabolic $P_\lambda\subset G$, and we may consider the partial flag variety $G/P_\lambda$.

If $q$ is the cardinality of the residue field, then Tits shows in 3.3.1 of his Corvalis article that $$ \mu(K\varpi^\lambda K) = \frac{\#(G/P_\lambda)(q)}{q^{\dim(G/P_\lambda)}}q^{\langle \lambda, 2\rho\rangle}. $$ Here $2\rho=\sum_{\alpha\in \Phi^+}\alpha$ is the sum of the positive roots. Gross shows in his article on the Satake isomorphism how these volumes are related to Kazhdan-Lusztig polynomials, which I would expect is related to the formulas of Macdonald. In this direction, this may also be expressed in terms of the extended affine Weyl group: let $l$ be the length function of the extended affine Weyl group $X_\ast(T)\rtimes W$, where $W$ is the Weyl group of $(G,T)$. Then $$ \mu(K\varpi^\lambda K) = \frac{\sum_{y\in W\lambda W}q^{l(y)}}{\sum_{w\in W}q^{l(w)}}.$$

  • $\begingroup$ Thanks a lot, this is exactly what I need. I also found a formula in Macdonald's "Spherical functions on a group of $p$-adic type" in terms of Poincaré polynomials. But of course this is the same formula. $\endgroup$ – Paul Broussous May 15 at 11:48
  • $\begingroup$ After checking Tits's Corvalis article, the formula that Gross states is not proved by Tits. [Tits](3.3.1) is just the Bruhat-Tits decomposition $G=IW^{\rm aff}I$ and the formula $\vert IwI/I\vert =q^{l(w)}$, where $I$ is the Iwahori subgroup, $W^{\rm aff}$ the affine Weyl group and $l$ the length function. Gross just writes that the formula is a simple consequence of the BT decomposition. I do not think that it is that simple. $\endgroup$ – Paul Broussous May 15 at 12:12
  • $\begingroup$ Ah, sorry about that. I didn't check the reference closely. I agree it's not immediate, but it is not too bad. If I have time later, I'll write up the argument. $\endgroup$ – WSL May 15 at 15:52

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.