The order in which diffeomorphisms should be applied

49 Views Asked by At

I’m reading ‘Bombay Lectures on Highest Weight Representations of Infinite Dimensional Lie Algebras’ by Kac and Raina. And there is a following paragraph in the first chapter.

The Lie algebra $\operatorname{Vect}(S^{1})$ can be considered as the Lie algebra of the group $G$ of orientation preserving diffeomorphisms of $S^{1}$. If $\zeta_{1}$, $\zeta_{2}$ are two elements of $G$ then their product is defined by composition: $$\zeta_{1}\cdot\zeta_{2} = \zeta_{1}\big(\zeta_{2}(z)\big)$$ for each $z = \exp (i\theta)$ on $S^{1}$. If $f(z)$ is an element of the vector space of smooth complex-valued functions on $S^{1}$, then $\gamma\in G$ acts on $f(z)$ by $$\pi(\gamma) f(z) = f\big(\gamma^{-1}(z)\big).$$ This clearly defines a representation of $G$.

Probably this is a notation subtlety I missed on Lie Groups representations, but I can’t see why $\pi$ is a representation, as it switches the order of multiples: \begin{multline} \forall\xi_{1},\xi_{2}\in G: \pi(\xi_{1}\cdot\xi_{2})f(z) = f\big((\xi_{1}\cdot\xi_{2})^{-1}(z)\big) = f\left(\left(\xi_{2}^{-1}\cdot\xi_{1}^{-1}\right)(z)\right) = \\ = f\left(\xi_{2}^{-1}\left(\xi_{1}^{-1}\left(z\right)\right)\right) = \pi(\xi_{2})f\left(\xi_{1}^{-1}(z)\right)= \big(\pi(\xi_{2})\cdot\pi(\xi_{1})\big)f(z) \end{multline} Therefore $\pi(\xi_{1}\cdot\xi_{2}) = \pi(\xi_{2})\cdot\pi(\xi_{1})$ so it kind of works like an ‘anti’-representation. However if we denote \begin{equation} (\zeta_{1}\cdot\zeta_{2})(z) = \zeta_{2}\big(\zeta_{1}(z)\big) \end{equation} everything works as it should \begin{multline} \forall \xi_{1},\xi_{2}\in G : \pi(\xi_{1}\cdot\xi_{2})f(z) = f\big((\xi_{1}\cdot\xi_{2})^{-1}(z)\big) = f\left(\left(\xi_{2}^{-1}\cdot\xi_{1}^{-1}\right)(z)\right) = \\ = f\left(\xi_{1}^{-1}\left( \xi_{2}^{-1}(z)\right)\right) = \pi(\xi_{1})f\left(\xi_{2}^{-1}(z)\right) = \left(\pi(\xi_{1})\cdot\pi(\xi_{2})\right)f(z) \end{multline} therefore $\pi(\xi_{1}\cdot\xi_{2}) = \pi(\xi_{1})\cdot\pi(\xi_{2})$.

I’m not sure if it is a typo as I saw similar notation in other places and I do recall that a similar discussion was dedicated to the order of multiples in the case of regular representations for finite groups. But there it was more obvious. If $f:G\to \mathbb{C}$ we had three representations \begin{gather} \lambda_{g}f(h) = f(gh)\\ \rho_{g}f(h) = f\left(hg^{-1}\right)\\ \pi_{g}f(h) = f\left(ghg^{-1}\right) \end{gather} and it is straightforward why all of them are well-defined.

The more I think in which order should I multiply $\xi_{i}$, $\xi_{i}^{-1}$ and $\pi(\xi_{i})$ the more clueless I get. Could anyone clarify this concept, please!

1

There are 1 best solutions below

1
On BEST ANSWER

Introducing extra notation for the intermediate steps of the computation and worrying about the point evaluations at $z\in S^1$ separately might make things clearer.

Let $g:=\pi(\xi_2)f$. Applying the definition of the action, we obtain at each $z\in S^1$ $$\pi(\xi_1)g(z) = g(\xi_1^{-1}(z)).$$ On the other hand, by definition of the action we have at each point $w\in S^1$ $$g(w) = f(\xi_2^{-1}(w)).$$ Evaluating at $w=\xi_1^{-1}(z)$, we obtain $$\pi(\xi_1)(\pi(\xi_2)f)(z) = \pi(\xi_1)g(z) = g(\xi_1^{-1}(z)) = f(\xi_2^{-1}(\xi_1^{-1}(z))),$$ so we see that the order gets flipped correctly.

You can also avoid the point evaluations altogether by thinking in terms of composition of functions, so the group product $\xi_1\cdot\xi_2$ is composition $\xi_1\circ\xi_2$ and the action is $\pi(\gamma)f = f\circ\gamma^{-1}$. Then the required computation takes the form $$\pi(\xi_1\circ\xi_2)f = f\circ(\xi_1\circ\xi_2)^{-1} = f\circ\xi_2^{-1}\circ\xi_1^{-1} = (\pi(\xi_2)f)\circ \xi_1^{-1} = \pi(\xi_1)(\pi(\xi_2)f).$$