# (Co)homology of the Eilenberg-MacLane spaces K(G.n)

Let $(G, n)$ be a pair such that $n$ is a natural number, $G$ is a finite group which is abelian if $n \geq 1$. It is well-known that $\pi_n((K,(G,n)) = G$ and $\pi_i (K(G,n)) = 0$ if $i \neq n$.

Also it is known that these spaces $K(G,n)$ play a very important role for cohomology. For any abelian group $G$, and any CW-complex $X$, the set $[X, K(G,n)]$ of homotopy classes of maps from $X$ to $K(G,n)$ is in natural bijection with the $n^{\mathrm{th}}$ singular cohomology group $H^n(X; G)$ with coefficients in $G$.

But what is known about the cohomology of the $K(G,n)$ themselves? It is interesting in the light of the above. Here I mean the singular cohomology with integral coefficients.

-
The (co)homology of $K(G,1)$ is well-known to equal the group (co)homology of $G$ with integer coefficients. I don't know what happens for $n>1$. –  Robin Chapman May 15 '10 at 14:23
See an introductory algebraic topology text like Hatcher or May. The (co)homology of Eilenberg-Maclane spaces are heavily studied. In a "stable range" this cohomology is called the Steenrod Algebra. –  Ryan Budney May 15 '10 at 14:29
Hatcher's spectral sequence text computes only cohomologies of Eilenberg-MacLane spaces with Z/2-coefficients, but her refers to J. P. May, A general approach to Steenrod operations, Springer Lecture Notes 168 (1970), 153–231 for an integral computation. –  Lennart Meier May 15 '10 at 21:18

Computing the integral cohomology of $K(\pi,n)$'s is feasible but a bit tricky. In fact the only reference I know is expos\'e 11 of H. Cartan's seminar, year 7 (available on Numdam). I'd be interested if there are other sources that cover that.

-
I just noticed this answer. I asked for other sources as a question: mathoverflow.net/questions/50417/… I'm trying to get hold of Schafer's thesis, to see if it fills this niche. –  Daniel Moskovich Jan 11 '11 at 13:37
@algori: Cartan seems to have also written up his calculations in some papers: "Sur les groupes d'Eilenberg-Mac Lane. I,II." (French) Proc. Nat. Acad. Sci. U. S. A. 40, (1954). I haven't looked at them yet, but presumably the presentation is cleaner. –  Mark Grant Aug 18 '11 at 14:25
The original papers of Eilenberg and MacLane have quite a few computations in them. Not so much of specific groups, but still quite a few things seem to be worked out, if you can get past their notation. –  Jon Beardsley Jul 10 at 21:12

Expanding slightly on Ryan's comment: it's an easy fact (often attributed to Serre) that the set of cohomology operations $H^k(-;G)\to H^r(-;H)$ (i.e. natural transformations) is in 1-1 correspondence with $[K(G,k),K(H,r)]=H^r(K(G,k);H)$ (for any abelian groups $G,H$). There are tons of these, some easy, some not so easy to understand, corresponding to how easy the calculation of $H^r(K(G,k);H)$ is. A nice subset are the stable operations (compatible with a certain suspension $[K(G,k),K(H,r)]\to [K(G,k+1),K(H,r+1)]$ which come in families, the most common family being the Steenrod algebra, corresponding to $G=H=Z/p$. There are non-stable operations also, eg the Pontryagin square $H^k(-;Z/2)\to H^{2k}(-;Z/4)$.

-

If you don't mind a little notational pain, you can look at the original papers where Eilenberg and Mac Lane worked many cases out: On the homology of groups $H(\pi,n)$ (I, II, III). Annals of Mathematics ~1953.

-
Scott, is the ring structure in the cohomology computed there as well? –  algori May 15 '10 at 15:26
I don't have the papers in front of me, but I will tentatively say "no". –  S. Carnahan May 15 '10 at 16:13
In particular, he gives an account of Cartan's results in Chapter 2, then describes a C++ program for computing integral (co)homology of certain ($2$-local) Eilenberg-Mac Lane spaces in Chapter 3. An appendix lists the integral (co)homology groups of $K(\mathbb{Z}_2,2)$, $K(\mathbb{Z}_2,3)$, $K(\mathbb{Z}_4,2)$ and $K(\mathbb{Z}_4,3)$ up to degree $200$.