Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I'm pretty much stuck on the following question (taken from the book Lie groups and introduction to linear groups by Rossman W.):

Rossman W. Lie groups and introduction to linear groups 2.6.5

I've found some clues, but I think I lack proper understanding of what needs to be done here.

If I recall correctly, the relation between lie groups homomorphism $T$ and its lie algebra homomorphism $\tau$ is $\tau(X) = \frac{d}{dt} T(e^{tX})|_{t=0}$ when in this case $ X \in gl(n,\mathbb{R})$. But then, some factor involving the exponent function should appear. On the other hand, the derivative of $f(x)$ is $$\displaystyle \sum_{i} \xi_i ' \frac{\partial f}{\partial\xi_i}$$, so it should be some kind of chain rule?

In (b) I have similar problem - it seems like I need to "go down" to $so(3)$ and use its basis matrices, but I'm not sure how to justify it.

Any assistance will be highly appreciated.

share|cite|improve this question
up vote 5 down vote accepted

It is exactly all about chain rule : $\tau(Y)f(x) = \dfrac{d}{dt}_{\vert t=0}T(e^{tY})f(x)=\dfrac{d}{dt}_{\vert t=0}f(e^{-tY}x)$.

If you write $\phi(t)=e^{-tY}x$ then $$\phi'(t)=-Yx=-\sum_{i,j=1}^nY_{ij}\xi_je_i$$ and you get $$\tau(Y)f(x) = df_{x}(-Yx)=-\sum_{i,j=1}^n\frac{\partial f}{\partial \xi_i}(x)Y_{ij}\xi_j.$$

You can see $\tau$ as the differential of $T:GL(n,\mathbb R)\rightarrow GL(F)$ at the identity element $I_n$.

Saying that $f\in \mathbb R[\xi_1,\xi_2,\xi_3]$ is invariant under the action of $SO(3,\mathbb R)$ means that the restriction of $\tau$ to the Lie algebra of $SO(3,\mathbb R)$ is the null linear map since $$\forall X\in so(3), \forall t\in \mathbb R, \quad T(e^{tX})f(x)=f(x).$$

And as you said, the Lie algebra $so(3)$ has a natural basis $$so(3)=<\begin{pmatrix} 0&0&1\\0&0&0\\-1&0&0\end{pmatrix},\begin{pmatrix} 0&1&0\\-1&0&0\\0&0&0\end{pmatrix},\begin{pmatrix} 0&0&0\\0&0&1\\0&-1&0\end{pmatrix}>=<A,B,[A,B]>$$ If you check fomula of question $(a)$ with one these three matrices, you will get exactly the expected three relations.

share|cite|improve this answer
Thanks for the answer and sorry it took me so much time to respond. I still can't understand, though, how $\xi_i'$ disappears in the last equation of part (a). – Pavel Feb 9 '13 at 11:31
The differential of $f$ at point $x$ is the linear form : $\displaystyle df_x= \sum_{i=1}^n \frac{\partial f}{\partial \xi_1}(x)d\xi_i$. It means that if $(\alpha_1,\cdots,\alpha_n)\in \mathbb R^n$ then $\displaystyle df_x(\alpha_1,\cdots,\alpha_n)=\sum_{i=1}^n \alpha_i\cdot \dfrac{\partial f}{\partial \xi_i}(x)$. Now, compute $df_x(-Y\cdot x)$ and check that you get the same thing that I wrote. – Bebop Feb 9 '13 at 12:10
Maybe I'm missing something, but how $d\xi_i$ turns to $\alpha_i$? My differentiation knowledge is a bit rusty... – Pavel Feb 9 '13 at 13:04
If you are working with a basis $(\xi_1,\cdots,\xi_n)$ on $\mathbb R^n$ then $d\xi_i$ is the linear form given by $d\xi_i(\xi_j)=\delta_{ij}$. So $\displaystyle d\xi_i(\sum_{i=1}^n\alpha_j \xi_j)=\sum_{i=1}^n\alpha_jd\xi_i(\xi_j)=\alpha_i$ and $(d\xi_1,\cdots,d\xi_n)$ is the dual basis associated to $(\xi_1,\cdots,\xi_n)$. In this basis, the matrix of $df_x$ is $J=\begin{pmatrix}\frac{\partial f}{\partial \xi_1}(x),\cdots,\frac{\partial f}{\partial \xi_n}(x)\end{pmatrix}\in M_{1,n}(\mathbb R)$ and for any vector $a=\begin{pmatrix}a_1\\\vdots\\a_n\end{pmatrix}$ of $\mathbb R^n$, $df_x(a)=J a$. – Bebop Feb 9 '13 at 16:55
So it stems to the Jacobian. Thanks for your help! – Pavel Feb 10 '13 at 19:20

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.