2
$\begingroup$

Consider a central extension of Lie groups $1\rightarrow S^1\rightarrow \hat{G}\xrightarrow{\pi} G\rightarrow 1$.

I understand that this mean $\pi:\hat{G}\rightarrow G$ is a surjective homomorphism of Lie groups (not sure if this has to be submersion) and that $S^1\subseteq Z(\hat{G})$. There is a local section $\sigma:U\rightarrow \hat{G}$ such that $\pi\circ \sigma=1_U$ where $U$ is an open nbd of $1\in G$. Correct me if I am missing some conditions.

Let $X$ be a manifold with an action of $G$ on it. Then we have the notion of quotient stack $[X/G]$.

There is an action of $\hat{G}$ on $X$ given by $(\hat{g},x)\mapsto \pi(\hat{g})\cdot x$.

We can then consider the quotient stack $[X/\hat{G}]$.

Given a manifold $Y$, objects of $[X/G](Y)$ are pairs $(P\rightarrow Y,P\rightarrow X)$ where $P\rightarrow Y$ is a principal $G$ bundle and $P\rightarrow X$ is a $G$-equivariant space (see that $G$ acts on $P$ and $X$).

"As locally any map $T\rightarrow G$ can be lifter to $\tilde{G}$", the map of stacks $[X/\hat{G}]\rightarrow [X/G]$ is a gerbe over stack.

I see that, locally any map $\theta: T\rightarrow G$ can be lifted to $\hat{G}$. As there is a section $\sigma:U\rightarrow \hat{G}$, we can consider $\theta^{-1}(U)\xrightarrow{\theta} U\xrightarrow{\sigma} \hat{G}$ and $\pi\circ (\sigma\circ \theta)=\theta$. Thus, any map $\theta:T\rightarrow G$ can be locally lifted to $\hat{G}$. But, I am not able to see why this imply $[X/\tilde{G}]\rightarrow [X/G]$ is a gerbe over stack.

Any comments are welcome.

$\endgroup$
26
  • $\begingroup$ You can consider the special case of $X=pt$, since the example you give is pulled back along $[X/G] \to [pt/G]$, and the pullback of a gerbe is a gerbe. $\endgroup$
    – David Roberts
    Jan 13, 2019 at 20:05
  • $\begingroup$ @DavidRoberts I am getting confused with English ... You are asking to take $X=pt$ and then saying consider $[X/G]\rightarrow [pt/G]$... Are you saying consider the obvious map of stacks $[X/G]\rightarrow [*/G]$ and pull back $[*/\hat{G}]\rightarrow [*/G]$ along $[X/G]\rightarrow [*/G]$ to get $[X/\hat{G}]\rightarrow [X/G]$?? As $[*/\hat{G}]\rightarrow [*/G]$ is a gerbe over stack, so is the pull back $[X/\hat{G}]\rightarrow [X/G]$?? Is this what you mean? $\endgroup$ Jan 13, 2019 at 20:12
  • $\begingroup$ Then, also, it should be easier to see why the stack $[pt/\hat{G}]$ of principal $\hat{G}$-bundles is a gerbe over $[pt/G]$. Given any $X\to [pt/G]$, that is, a principal $G$-bundle $P\to X$, there is a cover $U\to X$ such that $U\to X \to [pt/G]$ lifts to $[pt/\hat{G}]$: just take a trivialising cover for $P$. Thus $[pt/\pi]$ is an epimorphism of stacks. A similar type of thinking—unwinding the definition of the stack in terms of bundles—will help to show that $[pt/\hat{G}] \to [pt/\hat{G}] \times_{[pt/G]} [pt/\hat{G}]$ is also an epimorphism. $\endgroup$
    – David Roberts
    Jan 13, 2019 at 20:24
  • $\begingroup$ In response to your comment: yes. $\endgroup$
    – David Roberts
    Jan 13, 2019 at 20:25
  • $\begingroup$ @DavidRoberts Ok. Thanks for the clarification :) First I have to prove that $[*/\hat{G}]\rightarrow [*/G]$ is a gerbe over stack and then prove that pull back (fiber product) of gerbe over stack is a gerbe over stack (it may be obvious but I did not prove yet).. It looks like it has nothing to do with central extension... Any morphism of Lie groups $\hat{G}\rightarrow G$ such that there is a local section $U\rightarrow \hat{G}$ should give a gerbe over stack $[*/\hat{G}]\rightarrow [*/G]$... Is that the case? $\endgroup$ Jan 13, 2019 at 20:30

1 Answer 1

1
$\begingroup$

As discussed in the comments, $[X/\hat{G}] \to [X/G]$ is the pullback of $[pt/\hat{G}] \to [pt/G]$ along the canonical map $[X/G] \to [pt/G]$, so it suffices to show that $[pt/\hat{G}] \to [pt/G]$ is a gerbe. Since every principal $G$-bundle is locally trivial, it can be locally lifted to a principal $\hat{G}$-bundle, which is another way of saying that $[pt/\hat{G}] \to [pt/G]$ is an epimorphism of stacks. Using the fact that $\hat{G}\to G$ is surjective, then there is an equivalence of stacks $$ [pt/\hat{G}\times_G\hat{G}] \stackrel{\simeq}{\to} [pt/\hat{G}]\times_{[pt/G]}[pt/\hat{G}] $$ Thus the diagonal $[pt/\hat{G}] \to [pt/\hat{G}]\times_{[pt/G]}[pt/\hat{G}]$ is equivalent to $[pt/\hat{G}] \to [pt/\hat{G}\times_G\hat{G}]$, induced by the diagonal homomorphism $\hat{G} \to \hat{G}\times_G\hat{G}$. Such a map of stacks is an epimorphism by the same argument as before, and so we are done.

$\endgroup$
1
  • 1
    $\begingroup$ You mean “which is another way of saying” when you said “which is not another way of saying “ $\endgroup$ Feb 21, 2019 at 11:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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