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

Let $G$ be a linear algebraic group defined over $\mathbb Q$. Suppose that $G$ is isomorphic to $SL_n$ over $\mathbb R$. Suppose the $\mathbb Q$-rank of $G$ is fixed, say $m$. How many types are there for $G$ up to $\mathbb Q$-isomorphism? Are they finite especially for $m>2$?

share|cite|improve this question
The only such forms are $SL_{m+1}(D)$, where $D$ is an indefinite division algebra over $\mathbb{Q}$ of degree $n/(m+1)$. So the question becomes: are there finitely many isomorphism classes of indefinite division algebras of fixed degree $d$? This is far from being true. Translating into Brauer groups, you're essentially looking for elements of \bigoplus_{p}\mathbb{Z}/d\mathbb{Z}) whose co-ordinates add up to $0$, and which have order exactly $d$. – Keerthi Madapusi Pera Feb 28 '12 at 17:51
Those unitary groups, aren't they groups over $k$ itself? It seems to me that if $G$ is a unitary group for some hermitian form over some real quadratic extension of $\mathbb{Q}$, then it's a group over $\mathbb{Q}$ which is isomorphic to $\text{SL}_n$ over $\mathbb{R}$. – YCor Feb 29 '12 at 4:39
Yves--You are correct. These would also be $\mathbb{Q}$-forms of $SL_{n,\mathbb{R}}$. – Keerthi Madapusi Pera Feb 29 '12 at 6:09

I would refer to the very nice article by Jacques Tits in "Algebraic Groups and Discontinuous Subgroups" AMS Symposiain Pure Math Vol 9 (Boulder Conference) 1966. The title of the article by Tits is "Classification of Algebraic Groups", and gives the $k$-forms of simple algebraic groups over the separable algebraic closure of $k$ for any field $k$.

From the tables in this article it is not difficult to deduce that the only $\mathbb Q$-forms of $SL_n$ which over $\mathbb R$ become isomorphic to $SL_n({\mathbb R})$ are the following

(1) The groups $SL_m(D)$ where $D$ is a central division algebra over $\mathbb Q$ of degree $d$ such that $dm=n$ and $D\otimes {\mathbb R}=M_d({\mathbb R})$. The $\mathbb Q$-rank of $G$ is $m-1$. As Keerti Madapusi has observed, there are infinitely many such.

(2) The groups $SU_m(D,h)$ where $K/{\mathbb Q}$ is a real quadratic extension, $D$ is a degree $d$ central division algebra over $K$ with an involution of the second kind such that the involution restricted to $K/Q$ is the non-trivial element of the Galois group of $K/{\mathbb Q}$, and $h:D^m\times D^m \rightarrow D$ is Hermitian with respect to this involution. Furthermore, $D\otimes K_v$ is the matrix algebra $M_d(K_v)$ for both the archimedean embeddings $K_v$ of $K$. The $\mathbb Q$ rank of $G$ is the maximal number of (see Tits' article for a discussion) hyperbolic planes in $h$; it can again be proved that there are infinitely many with a fixed $\mathbb Q$-rank, by varying either the field $K$ or the hermitian form $h$.

I think these may even be worked out in Dave Witte's e-book on arithmetic groups (not very sure).

share|cite|improve this answer

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.