Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Could any one give me hint how to show $SO(n)$ is connected? I understand that it is closed subset, I can prove $O(n)$ is not connected. Edit: suddenly got some idea, any matrix from $SO(n)$ can be written in this from, where first and second row are like below $$\left(\begin{array}{cc} \cos\theta & \sin\theta\\ -\sin\theta & \cos\theta\end{array}\right).$$

we can make $f(t)= tA+(1-t)I$ this is a path between $A$ and $I$. can this idea be rigorised?

share|cite|improve this question
$O(n)$ is not connected. In fact $O(n)=[O(n)\cap \det^{-1}(1)]\cup[O(n)\cap\det^{-1}(-1)]$ and $SO(n)=O(n)\cap\det^{-1}(1)$. – Mercy King Aug 5 '12 at 9:49
@Mercy just for got to put Not :) – miosaki Aug 5 '12 at 9:55
Do you know the normal form for such matrices? – mland Aug 5 '12 at 10:01
@mland No, I dont know what is Normal form' – miosaki Aug 5 '12 at 10:05
Your updated question more or less is the normal form.. But of course you have to take sums of these kinds of matrices.. Then you do not want to take a liner path between A and Id (else you could always do that) but homotop the angle $\theta$ to zero.. – mland Aug 5 '12 at 10:16
up vote 2 down vote accepted

You can prove connectedness of SO(n) by induction on $n$: Sketch: For any unit vectors $v,w\in {\mathbb R}^n$ there is $1$-parameter family of matrices $A_\theta\in SO(n)$ such that $A_0(w)=w$ and $A_1(w)=v$ -- take $A_\theta$ to be the matrix composed of your $2\times 2$ matrix (acting on the space spanned by $v,w$) and the identity matrix on the orthogonal complement to $Span(v,w).$

Fix $v\in \mathbb R^n.$ Take any $M\in SO(n).$ Suppose $M(v)=w$. Then $M$ is connected through $A_\theta^{-1}M$ to a matrix $N$ which fixes $v$. Take an orthogonal basis of the orthogonal complement of $v$. Then $N$ is represented by an $SO(n-1)$ matrix wrt to this basis. By inductive assumption $N$ is connected to $I$ by a path.

I skipped the base step. Note that the base step fails for $O(n)$ for $n$ odd.

share|cite|improve this answer

This should probably be a comment on student's answer, but I cannot preview comments...

One way to restate his/her argument is to notice that there is a smooth action of $SO(n)$ on $S^{n-1}\subseteq\mathbb R^n$ which is transitive. If $s$ is a point in the sphere, the stabilizer of $s$ in $SO(n)$ is isomorphic to $SO(n-1)$, as one easily sees, so we have a map $p:g\in SO(n)\mapsto gs\in S^n$ which is surjective and, in fact, a fibration with fiber $SO(n-1)$. We can apply to it the long exact sequence in homotopy and an induction to show $SO(n)$ is connected. Alterantively (in fact, equivalently, but less technologically, maybe) one can show that the domain of the map $p$ is connected because its codomain and its fibers are, and the map is proper, or locally trivial.

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.