Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have come heard that the cohomology of the classifying space of a compact torus $T$ is equal to the symmetric algebra over the dual of its Liealgebra $t^*$, where elements of the $t^*$ are of degree two:


While I am able to compute that both sides are isomorphic to a polynomial ring in $dim(T)$ generators (of deg 2), I don't see at all how to write down a canonical map between both sides.

So my questions are:

What is the above map? How to think about it?

share|cite|improve this question
up vote 4 down vote accepted

In your case, $BS^1$ is $\mathbb{CP}^\infty$. To see this, you can either use the fact that complex line bundles are classified by (homotopy classes of) maps into $\mathbb{CP}^\infty$ and that complex line bundles are the same as principal $S^1$-bundles (in one direction, choose a metric on your bundle and consider the sphere bundle), or you can use explicit cover $$S^{1} \to S^{\infty} \to \mathbb{CP}^\infty$$ (the infinite-dimensional Hopf map) to see that there is a $S^1$-bundle over projective space whose total space is contractible, which is the distinguishing property of the classifying space. The cohomology of $\mathbb{CP}^\infty$ can be computed explicitly (it's a polynomial ring on a generator in degree two).

If you work modulo torsion, i.e. with $\mathbb{R}$-coefficients, this map can be given by the Chern-Weil homomorphism (which has the advantage of generalizing to any $G$). If $G$ is any Lie group, then there is a morphism of algebras from the space of invariant polynomials on $\mathfrak{g}$ (i.e. $(\mathrm{Sym} \mathfrak{g}^*)^G$) to the cohomology ring $H^*(BG; \mathbb{R})$. Essentially, the idea is that cohomology classes of $BG$ are the same thing as "characteristic classes" of principal $G$-bundles (by the universal property of $BG$), so the point is to, for each space $M$ with a principal $G$-bundle $P \to M$, to assign a characteristic class in $H^*(M; \mathbb{R})$ measuring in some sense the nontrivialty of your bundle for each invariant polynomial $q: \mathfrak{g} \to \mathbb{R}$.

To construct a characteristic class on the manifold $M$ for the invariant polynomial $q$ and the bundle $P \to M$, one can work as follows: choose a connection on $P$. By (at least one) definition, this is a $\mathfrak{g}$-valued 2-form $\omega$ on $P$, with a certain equivariance property of the $G$-action. Then, consider $q(\omega)$ as an ordinary differential form on $M$. One can show that this descends to a form on $M$ (this comes from the invariance of the polynomial $q$), is closed, and that its cohomology class does not depend on the connection. (This is explained in Kobayashi-Nomizu's book on differential geometry.) This is the characteristic class you want, in de Rham cohomology. Note in particular that this gives you a way of computing Chern classes by "analytic" data (the first Chern class is the situation you asked about).

The Chern-Weil homomorphism is an isomorphism for a compact Lie group.

share|cite|improve this answer
Thanks very much Akhil. You mean the cohomology class of $q(\omega)$ does not depend on the connection right? – jan Sep 24 '11 at 8:01
@Jan: Yes, that's right. – Akhil Mathew Sep 24 '11 at 16:36

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.