4
$\begingroup$

I have come across this identity a few times and I have absolutely no idea why it holds.

$g^{-1}\exp(tX)g=\exp(t(\text{ad}_{g^{-1}}X))$

Would any one be able to explain exactly why this holds or point me to a resource that would explain.?

Thanks.

$\endgroup$
4
$\begingroup$

That's a special case of the naturality of the exponential map: If $\varphi: G\to H$ is a morphism of Lie groups with derivative $\text{d}\varphi:{\mathfrak g}\to{\mathfrak h}$, then the following diagram commutes: $$\begin{array}{ccc} G & \stackrel{\varphi}{\longrightarrow} & H \\ \uparrow &&\uparrow\\ {\mathfrak g} & \stackrel{\text{d}\varphi}{\longrightarrow} &{\mathfrak h}\end{array};$$ here, the left and right vertical maps are the exponential maps of $G$ and $H$, respectively.

Applying this to the conjugation automorphism $G\xrightarrow{\text{c}_{g^{-1}}: h\mapsto g^{-1}hg} G$, you have $\text{d}(\text{c}_{g^{-1}})\stackrel{\text{def}}{\equiv}\text{Ad}_{g^{-1}}: {\mathfrak g}\to{\mathfrak g}$, so $$\begin{array}{ccc} G & \stackrel{\text{c}_{g^{-1}}}{\longrightarrow} & G \\ \uparrow &&\uparrow\\ {\mathfrak g} & \stackrel{\text{Ad}_{g^{-1}}}{\longrightarrow} &{\mathfrak g}\end{array}$$ commutes, which is what you were asking for.

Note that $\text{d}\varphi$ is always a linear map, so you can move the scaling by $t$ outside in your equation. Also, one should write $\text{Ad}$ uppercase here, as $\text{ad}$ is usually reserved for the adjoint representation ${\mathfrak g}\to{\mathfrak g}{\mathfrak l}({\mathfrak g})$, which is itself the derivative of $G\stackrel{\text{Ad}}{\longrightarrow}\text{GL}({\mathfrak g})$.

$\endgroup$
2
$\begingroup$

First suppose $G$ is a matrix lie group. ad$_{g^{-1}}X=g^{-1}Xg\,,$ so $e^{t(ad_{g^{-1}}X)}=e^{tg^{-1}Xg}\,.$ So $\displaystyle e^{tg^{-1}Xg}=\sum_n \frac{(tg^{-1}Xg)^n}{n!}=\sum_n g^{-1}\frac{(tX)^n}{n!}g=g^{-1}e^{tX}g\,.$

For a general Lie group, $e^{tY}=\gamma(t)$ where is $\gamma(t)$ is the unique one parameter subgroup with $\gamma(0)=e$ and where $\gamma'(0)=Y\,.$ So it follows from the definition that $\frac{d}{dt}e^{tad_{g^{-1}}X}\vert_{t=0}=ad_{g^{-1}}X\,.$ Now from the definition of $ad_{g^-1},$ we have that $ad_{g^{-1}}X=\frac{d}{dt}g^{-1}e^{tX}g\vert_{t=0}\,.$ Hence they have the same derivative at $t=0\,,$ so it follows immediately that $g^{-1}e^{tX}g=e^{tad_{g^{-1}}X}\,.$

$\endgroup$

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.