2
$\begingroup$

Let $G=SO(n,1)$ and let $G=KAN$ be an Iwasawa decomposition of $G$. Let $M$ be the centralizer of $A$ in $K$. In this case, we have $K≃SO(n)$, $A≃\Bbb R$(this is the maximal diagonalizable subgroup), $N≃\Bbb{R}^{n−1}$ and $M≃SO(n−1)$. Let $\mathfrak k$ be the Lie algebra of $K$ and $\mathfrak m\subseteq\mathfrak k$ be the lie algebra of $M$. Let $\mathfrak h$ be the orthocomplement to $\mathfrak m$ in $\mathfrak k$. That is $\mathfrak k= \mathfrak m\oplus \mathfrak h$. Let $H$ be the Lie hgroup associated to generate by $exp(\mathfrak h)$. What can we say about $HM$? When is it true that $HM=K$?

I believe $K/HM$ is discrete and finite. In this case there exists $e=\omega_1,\omega_2,\dots,\omega_l$ s.t. $K=HM\sqcup\omega_2 HM\sqcup\dots\sqcup\omega_l HM$. Can I explicitly calculate these $\omega_i$?

$\endgroup$
4
  • $\begingroup$ ${\mathfrak h}$ is not a Lie algebra, what do you mean by $H$? $\endgroup$
    – Misha
    Feb 22, 2013 at 4:18
  • $\begingroup$ the issue is one of connected components; so one must be careful about the groups $K$ and $M$. your $K$ is not $SO(n)$ but $O(n)$. Your $M$ is not $SO(n-1)$ but $O(n-1)$. In this case $K/HM$ is connected. $\endgroup$ Feb 22, 2013 at 4:27
  • $\begingroup$ this is independent of what $\mathfrak h$ may be. As Misha says, $\mathfrak h$ is not a Lie algebra. $\endgroup$ Feb 22, 2013 at 4:51
  • $\begingroup$ @Misha. Thank you. Indeed, $\mathfrak h$ is not a Lie Algebra. For my purpose it is enough to consider the group generated by $exp(\mathfrak h)$. In this case, what can I say about the cosets of $K/HM$? $\endgroup$
    – Davis
    Feb 22, 2013 at 5:35

2 Answers 2

1
$\begingroup$

I believe that your set $\exp(\mathfrak{h})$ will always act transitively on $S^n$ and will contain a stabilizer of at least one point of $S^n$, which should imply that $\exp(\mathfrak{h})$ alone generates $K$. Or am I missing something?

$\endgroup$
1
$\begingroup$

The pair (SO(n), SO(n-1))) is a symmetric pair, hence in this case M is a maximal subgroup of SO(n). Thus the connected component of your subgroup must the whole group.

$\endgroup$

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.