4
$\begingroup$

The notion of bundle gerbe is a categorification of line bundles/principal $U(1)$-bundles, and comes in two presentations: a linear version (with $Line_\mathbb{C}$-enriched underlying groupoid) and a principal version (using $U(1)$-torsors). It is really the first one that deserves the name "2-line bundle" (or line 2-bundle), and the second one "principal 2-bundle".

The concept of trivialisation of a gerbe is very much like the idea of a global section of a principal $U(1)$-bundle, and one also sees the analogue in the linear setting. However, line bundles may have global sections that vanish at certain points. Hence there should be a notion of section of a 2-line bundle (which would itself be some bundle-like object) than exhibits such a behaviour. Away from the points where the "section" vanishes, one should have a trivialisation of the gerbe by that data, but at the vanishing points the data should somehow be degenerate.

After that preamble, the only place I have seen such a definition is in a video of a talk at a German university (perhaps Hamburg) I started watching but gave up after prolonged buffering issues. The problem is, I cannot remember the speaker or the title or the conference! I think it was a graduate student speaking, but cannot after extensive searching figure out who it was. So my question is: what is the definition of a general section of a 2-line bundle, and what was the talk?

$\endgroup$
8
  • $\begingroup$ The notion of noninvertible morphisms between bundle gerbes is due to Konrad Waldorf, see “More morphisms between bundle gerbes”. $\endgroup$ Commented Jul 29, 2017 at 16:37
  • $\begingroup$ If you think that a line bundle is the associated bundle construction $\mathbb C \times^{U(1} M \to M$, then can you think of a gerbe as a principal $BU(1)$ bundle and use some representation of $BU(1)$? (I don't know how to think about those.) $\endgroup$ Commented Jul 29, 2017 at 17:44
  • $\begingroup$ @dmitri no, it was a rank one thing. As I said, outside the vanishing locus, it's a trivialisation. $\endgroup$
    – David Roberts
    Commented Jul 29, 2017 at 23:26
  • $\begingroup$ @allen no, it was nothing like that. More like a circle bundle where the circles degenerate. $\endgroup$
    – David Roberts
    Commented Jul 29, 2017 at 23:27
  • 1
    $\begingroup$ @DavidRoberts: I see. Start with coherent sheaves of rank at most 1, suspend, and sheafify. The resulting stack contains bundle gerbes and their isomorphisms (because line bundles are coherent sheaves of rank at most 1), but also contains morphisms with nontrivial vanishing locus, corresponding to coherent sheaves of rank at most 1 that are not line bundles. $\endgroup$ Commented Jul 30, 2017 at 10:06

0

You must log in to answer this question.

Browse other questions tagged .