2
$\begingroup$

I understand that via the Borel density theorem given a finite dimensional (polynomial) representation of the simple non-compact Lie groups $SL_n \mathbb R$ or $Sp_n \mathbb R$, I get an irreducible representation when I restrict to $SL_n \mathbb Z$ or $Sp_n \mathbb Z$, respectively.

I wonder if the restriction is injective on the set of isomorphism classes of irreducible polynomial representations. I.e. is it true that if $V$ and $W$ are irreducible polynomial $SL_n \mathbb R$-representations and $$Res_{SL_n \mathbb Z} V \cong Res_{SL_n \mathbb Z} W$$ then $V\cong W$?

$\endgroup$
1
  • 4
    $\begingroup$ I wouldn't call the fact that $\text{SL}_n(\mathbb{Z})<\text{SL}_n(\mathbb{R})$ is Z-dense a result of the Borel density theorem (even though this fact indeed follows from the BDT). This is almost a triviality. $\endgroup$
    – Uri Bader
    Aug 5, 2016 at 13:54

1 Answer 1

8
$\begingroup$

(This is too long for a comment.) This follows from the fact that $\text{SL}_n\mathbb{Z}$ is Zariski dense in $\text{SL}_n\mathbb{R}$. Let $\phi:V\to W$ be a linear isomorphism such that for every $g\in \text{SL}_n\mathbb{Z}$, the "conjugate" $\phi^g = g^{-1}\cdot\phi(g\cdot -)$ equals $\phi$. The subset of $\text{SL}_n\mathbb{R}$ of all $g$ such that $\phi^g$ equals $\phi$ is the simultaneous zero locus of finitely many polynomial functions of the matrix entries of $g$: just look at the matrix entries of $\phi^g - \phi$ as functions of $g$. Thus, since these polynomials vanish on the Zariski dense subset $\text{SL}_n\mathbb{Z}$, they vanish on all of $\text{SL}_n\mathbb{R}$.

$\endgroup$
1
  • $\begingroup$ This is exactly what I hoped for. $\endgroup$ Aug 5, 2016 at 13:32

Your Answer

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

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