MathOverflow is a question and answer site for professional mathematicians. 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

Suppose $G$ and $A$ are abelian groups (I'm setting $G$ abelian to keep the discussion simple, though there are analogues for non-abelian $G$) with $G$ acting trivially on $A$. By the universal coefficients theorem, we have short exact sequences for all positive integers $n$:

$$0 \to \operatorname{Ext}^1(H_{n-1}(G;\mathbb{Z});A) \to H^n(G;A) \to \operatorname{Hom}(H_n(G;\mathbb{Z}),A) \to 0$$

In the case $n = 2$, this becomes:

$$0 \to \operatorname{Ext}^1(G;A) \to H^2(G;A) \to \operatorname{Hom}(H_2(G;\mathbb{Z}),A) \to 0$$

The surjection $H^2(G;A) \to \operatorname{Hom}(H_2(G;\mathbb{Z}),A)$ has a nice interpretation:

$H_2(G;\mathbb{Z})$ is the Schur multiplier of $G$, which (since $G$ is abelian) is the exterior square of $G$, so $\operatorname{Hom}(H_2(G;\mathbb{Z}),A)$ is the alternating bilinear maps from $G$ to $A$. The mapping $H^2(G;A) \to \operatorname{Hom}(H_2(G;\mathbb{Z}),A)$ is the "skew" mapping; it sends a 2-cocycle $f:G \times G \to A$ to the function:

$$(x,y) \mapsto f(x,y) - f(y,x)$$

Even without prior knowledge of the universal coefficients theorem, this short exact sequence makes sense: one can check that the skew of a 2-cocycle for an abelian group is alternating and bilinear, and the mapping descends to cohomology classes because any 2-coboundary is symmetric. The kernel of the mapping corresponds to symmetric cohomology classes, which corresponding to the abelian extension groups, given precisely a $\operatorname{Ext}^1(G;A)$. In group extension terms, the image of a given element of $H^2(G;A)$ under the skew map describes the commutator map of the extension group.

My question is: what's an analogous concrete interpretation for higher $n$?

I'm listing below some references with generalizations of the above to non-abelian $G$ and varietal generalizations:

John Burns and Graham Ellis. On the nilpotent multipliers of a group. Math. Zeitschr., Volume 226, Page 405 - 428(Year 1997) (Official gated copy (PDF))

C. R. Leedham-Green and Susan McKay. Baer-invariants, isologism, varietal laws, and homology. PDF online. Pages 37-41 (135-139 in the original) describe a generalization of universal coefficients to varietal laws (where the usual universal coefficients is with respect to abelian groups inside groups)

share|cite|improve this question
Have you tried assuming that $G$ is finite cyclic, so that cohomology is $2$-periodic? The isomorphism $H^q\cong H^{q+2}$ is given by an explicit cup-product so you might deduce something...but that's only a guess. – Filippo Alberto Edoardo Jun 28 '12 at 10:56
Noting here: the original version of the question had the wrong statement of the universal coefficients theorem for general n (correct for n = 2 though) -- I fixed it today morning. I had said "Ext^{n-1}(G^{ab},A)" and the correct term is "Ext^1(H_{n-1}(G;Z),A)" – Vipul Naik Jun 29 '12 at 3:12

This is an extended-comment, but I think no:

In homology, we know that $H_1$ is abelianization, $H_2$ is Schur multiplier, but $H_{n\ge 3}$ is ???
In cohomology, $H^1(G;A)$ is split extensions, and this fits well with $Hom(G,A)$ in the UCT. As you mention, $H^2(G,A)$ corresponds to group extensions of $G$ by $A$, and this fits well with $Hom(H_2G,A)$ in the UCT. But as we see, this all ties in to knowing $H^n(G;A)$ and $H_n(G;A)$ very well...

If we look at $H^3(G;A)$ we get crossed module extensions $0\to A\to N\to E\to G\to 0$, and these are cumbersome. We have no nice interpretation for $H^n(G;A)$ for $n>3$, except more crazy-looking exact sequences. This is why I don't expect a "nice" map (i.e. "interpretation" of it) in the UCT to arise.

share|cite|improve this answer
Let me say that what you describe as 'cumbersome' and 'crazy' looks to me (and to many others) as very elegant and natural constructions. – Fernando Muro Jun 29 '12 at 16:08

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.