3
$\begingroup$

For Riemannian symmetric spaces, we have a nice result that the simply-connected ones are products of irreducible symmetric spaces, of which we have a list of ten infinite families and some exceptional ones (see e.g. wikipedia). In Paul Garrett's answer to this MSE question, after giving the non-compact forms of the infinite families of irreducible Riemannian symmetric spaces, he says

Note that in all cases the three $\mathbb{R}$-algebras $\mathbb{R}$,$\mathbb{C}$,$\mathbb{H}$ play roles. In fact, a completely parallel thing happens for classical groups over $p$-adic fields and in other cases, as in A. Weil's "Algebras with involutions and classical groups", Indian (not Indiana) J. Math. 1960.

I flipped through Weil's paper but wasn't enlightened. What I didn't find and am looking for in this question is an explicit list of p-adic homogeneous spaces $G/K$, where $K$ is presumably the fixed points of some involution, which is the analogue in whatever suitable sense of the irreducible Riemannian symmetric spaces; additionally I would like an explanation of what this 'suitable sense' is. Any help is appreciated!

$\endgroup$
4
  • 3
    $\begingroup$ At least for the analogue of symmetric spaces of noncompact type, the answer is probably "Euclidean building", as developed by Bruhat and Tits in the late 60's. The problem with $G/K$ is that it's just a discrete space. But the point is that it's (close to be) the 1-skeleton of a nice complex. $\endgroup$
    – YCor
    Commented Mar 20, 2020 at 15:36
  • 2
    $\begingroup$ To complement Yve's comment: For compact symmetric spaces an analogue would be Tits buildings. Their classification (in the compact case) is well-known and can be found in any book on buildings, e.g. Brown or Ronan. $\endgroup$ Commented Mar 20, 2020 at 17:20
  • $\begingroup$ @YCor: are there pairs $(G,K)$ associated to these Euclidean buildings or to compact buildings, or are they something more complicated? (sorry, I know almost nothing about buildings). The fact that $G/K$ is a discrete space isn't a problem for my purposes. I'm really after a list of pairs of groups $(G,K)$, as I'm coming from a random matrix theory viewpoint. Can such a list be extracted from the classification of buildings? It looked like Ronan's treatment of the classification is using an axiomatic approach to buildings and not referencing p-adic groups too much. Thanks! $\endgroup$ Commented Mar 20, 2020 at 18:43
  • $\begingroup$ In the Riemannian case, the point is that you can define on something purely local (Riemannian symmetric space $X$ of non-positive sectional curvature) which essentially characterizes the pair $(G,K)$ (say, taking the isometry group of the universal covering and one point stabilizer). In the $p$-adic case, $G/K$ being discrete, one needs something new, and the new feature is the building structure. (Here I'm only addressing your sentence "the fact that $G/K$ is discrete isn't a problem for my purposes". I don't know what a good reference for Bruhat-Tits buildings is.) $\endgroup$
    – YCor
    Commented Mar 20, 2020 at 19:26

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.