Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

I'm doing some constructions with the universal central extension $\widehat{\Omega G}$ of the loop group $\Omega G$ (here $G$ is a matrix group), where a priori the loops involved are just smooth, but in fact turn out to be rational functions. This got me thinking to see if everything will in fact land inside the algebraic loop group $G\left(\mathbb{C}((t))\right)$. However, I would like to know how one constructs the central extension in that case, as for some reason I can't seem to find a decent discussion of this (I'm probably being stupid in my searching...). In particular, is the central extension something like an ind-affine algebraic group? Given the cocycle describing the extension, how do you get said extension? (the method I know gives it as a quotient of a split central extension of $P\Omega G$, see links above)

Now my intended aim is to package this into something like the crossed module $\widehat{\Omega G} \to PG$ representing the String 2-group, but using more algebraic ingredients. For instance, replace the Frechet manifold $PG$ of based paths in $G$ with the space of polynomial or rational connections on the trivial $G$-bundle on $\mathbb{C}^\times$. The part I don't know is the central extension as indicated above.

EDIT: let me add that I would be most interested in knowing whether the central extension of the loop group is something like an algebraic group or if it is in some sense 'inherently transcendental' (for instance, the cocycle one uses to build it uses a residue).

share|improve this question
I don't know a lot about this story, but there's a survey paper by Tits that is often referenced; see here: people.mpim-bonn.mpg.de/zagier/files/doi/10.1007/BFb0084581/… He discusses various aspects of constructing these extensions. –  Chuck Hague Oct 1 '13 at 16:18
The central extension of the algebraic loop group is indeed a group object in the category of ind-varieties. See mathoverflow.net/questions/24845/… The string 2-group doesn't exist in the world of algebraic geometry. The closest thing that does exist is the central extension of G by $K_2$. –  André Henriques Oct 2 '13 at 13:10
@André - ok, thanks for that. I'd read that answer but didn't glean that fact. Homotopy theoretic aspects aside, there must be something inherently transcendental about the construction of the crossed module which I'd like to pin down. –  David Roberts Oct 2 '13 at 22:34
The central extension is non-trivial U(1) bundle over $LG$ and does therefore not even submit a continuous cocycle. In the following two papers it was considered as a quotient of a bigger trivial U(1) bundle, but if this is algebraic I don't actually know: Mickelsson, J.: Kac-Moody groups, topology of the Dirac determinant bundle and fermionization. Commun. Math. Phys. 110, 173-183 (1987) Mickelsson, J.: In: Current algebras and groups. New York: Plenum Press 1989, cf. Gabbiani, Fröhlich: "Operator Algebras and Conformal Field Theory", Commun. Math. Phys. 155, 569-640 (1993) p.593 –  Marcel Bischoff Dec 19 '13 at 8:55
You might have a look at §4 of Beauville-Laszlo, Commun. Math. Phys. 164, 385-419 (1994). –  abx Jan 18 at 10:29
show 2 more comments

1 Answer

Lie algebra cocycles are always algebraic. But if you try to integrate them up to group cocycles you meet homotopy obstacles, and you solve ODE's (twice) which give transcendental answers except in very degenerate cases.

share|improve this answer
But in the usual case of the smooth central extension of the loop group you don't get a group cocycle, just a Segal-Mitchison-Brylinski cocycle, which is is determined by a cocycle on the local group given by a neighborhood of the identity of G. Not to say that your comment doesn't apply in that setting... –  David Roberts Oct 20 '13 at 10:20
add comment

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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