1
$\begingroup$

I understood gerbes as generalization of line bundle here.

In this, I am trying to understand how to generalize notion of trivialization of line bundle to the notion of trivialization of gerbes. I am following this notes of Hitchin for this.

It says that

We can’t point to a gerbe as a space in the same way that we can for a line bundle, but there is a fundamental notion which helps to understand them – a trivialization of a gerbe.

Further, it says,

Recall that for line bundles, trivialization is a non-vanishing section $s$ of $L$. A unitary trivialization $s/||s||$ is also a section of the principal $S^1$ bundle. In terms of transition functions, this is a collection of functions $f_\alpha:U_\alpha\rightarrow S^1$ such that on $U_\alpha\cap U_\beta$ we have $f_\alpha=g_{\alpha\beta}f_\beta$.

Given another trivialization $f_\alpha':U_\alpha\rightarrow S^1$ we have $f_\alpha'=g_{\alpha\beta}f_\beta'$ and so we see that $$\frac{f_\alpha'}{f_\alpha}=\frac{f_\beta'}{f_\beta}.$$ Thus, $f_\alpha'/f_\alpha$ is restriction of a global function to $U_\alpha$ and we see that the difference of two local trivializations of a line bundle is the global function $X\rightarrow S^1$.

Here, $L$ is a unitary line bundle or its principal $S^1$ bundle of unitary frames. I do not know what is a unitary bundle. I assume that $L$ is just a line bundle $\pi:L\rightarrow X$.

I understand that for line bundles, trivialization is same as giving a non vanishing section of line bundle.

He generalises notion of trivialization to that of gerbes.

In case of line bundles, this is collection of maps $f_\alpha:U_\alpha\rightarrow S^1$.

In case of gerbes it has to be on $1$ fold intersection extra i.e., $f_{\alpha\beta}:U_\alpha\cap U_\beta\rightarrow S^1$.

The condition $f_\alpha=g_{\alpha\beta}f_\beta$ in case of line bundles should have its similar generalization in case of gerbes. I fail to understand how this generalisation is the condition $$g_{\alpha\beta\gamma}=f_{\alpha\beta}f_{\beta\gamma}f_{\gamma\alpha}$$

Any suggestion on understanding this better is useful.

$\endgroup$
9
  • $\begingroup$ If you’re not already familiar with it, you should look up the description of the first chern class in terms of Cech cohomology. $\endgroup$ Apr 12, 2018 at 13:39
  • $\begingroup$ @AaronBergman Can you suggest me some place to see that and tell me how does it help to understand this better. $\endgroup$ Apr 12, 2018 at 15:11
  • $\begingroup$ Try Section 1.1 of Griffiths and Harris $\endgroup$ Apr 12, 2018 at 18:45
  • $\begingroup$ @AaronBergman That is useful, thanks a lot. I was thinking Principal bundle comes with a local trivialization... I was thinking Giving another trivialization is like giving another principal bundle structure on same tuple $(\pi:P\rightarrow M)$ just like giving another smooth structure on a manifold... Isnt't it the situation? $\endgroup$ Apr 13, 2018 at 6:52
  • $\begingroup$ A principal bundle does not come with a local trivialization -- that's extra structure. I'd say it's like a different atlas of charts on a manifold, not like a different smooth structure. What you'll find is that changing the local trivialization doesn't change the Cech cohomology class. $\endgroup$ Apr 13, 2018 at 12:59

0

Your Answer

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

Browse other questions tagged or ask your own question.