Specifically, I'm somewhat confused by the answer to this question here, which I reproduce below:

Let $g$ denote the metric of $G$. The right invariance of $g$ guarantees that $$R^*_{\exp (tX)}g=g, \forall t \in \mathbb{R}. $$ On the other hand, $t\to R_{exp(tX)}$ is the flow of X, we have $\mathcal{L}_Xg=0$, where $\mathcal{L}_Xg$ denotes the Lie derivative of $g$ in the direction X. $$(\mathcal{L}_Xg)(U,V)=X(g(U,V))-g(\mathcal{L}_XU,V)-g(U,\mathcal{L}_YZ) \\ =-\langle[X,U],V\rangle-\langle U,[X,V]\rangle$$

where we use the invariance of $g$ at last equal to conclude that $g(U,V)$, seen as function of $G$, is constant in each connected component of $G$.

My question is, why can we assume that $\mathcal{L}_X g = 0$?


Given a vector field $X$ with flow $\varphi_t^X$, the definition of the Lie derivative $\mathcal{L}_X g$ is

$$ \mathcal{L}_X g = \frac{d}{dt} (\varphi_t^X)^{*}(g)|_{t = 0}. $$

In your case, $\varphi_t^X = R_{\exp(tX)}$ and the second line shows that $R_{\exp(tX)}^{*}g$ is constant (doesn't depend on $t$) so the derivative with respect to $t$ is zero.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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