MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

Can we classify all finite CW complexes $X$ such that for each $i$ there is some isomorphism $\pi_i(X) \rightarrow H_i(X)$? Note that it is not hard to classify all complexes for which each isomorphism is given by the Hurewicz map.

share|cite|improve this question
@Manuel, do you have a particular motivation for this problem? The answer is going to be some small class of CW-complexes but I'm not seeing a motivation for the question. In that regard the title is a little misleading because homotopy groups form a type of graded lie algebra and homology groups don't. Moreover, you're asking for a dimension-wise isomorphism of groups, not an equality of groups. – Ryan Budney Oct 8 '10 at 15:07
Would we have to know, a priori, all of the homotopy groups of a space to answer this question? – jd.r Oct 8 '10 at 15:11
The "right" viewpoint on invariants like $H_i$ and $\pi_i$ is that they are functorial invariants. It is somehow "wrong" to not think of them functorially. The Whitehead theorem gives some justification for this philosophy. – Kevin H. Lin Oct 8 '10 at 15:19
Yes, certainly the question sounds like it lacks of motivation, since crazy maps between homotopy and cohomology and not particularly useful. I was just trying to think what is happening geometrically when you have these "non natural" isomorphisms. – Manuel Rivera Oct 8 '10 at 15:29
@ Manuel - Just to give slightly different example : A space $X$ and the space $\prod K(\pi_i(X),i)$ have isomorphic homotopy groups. Unless you have a map from one to the other realizing these isomorphisms, there is no geometric meaning to these isomorphisms. The same goes for homology too - given a suitable $X$ you can cook up a product of Moore spaces which have homology isomorphic to $X$. But until you have a map one way or another that realizes these isomorphisms, it is not veru helpful. – Somnath Basu Oct 8 '10 at 16:21

I also don't see a motivation for this, but anyway: if $i=0$ is allowed, there is not a single example ($\pi_0$ is always finite, $H_0$ never. For the empty space, $\pi_0$ is empty and $H_0$ is not). Apart from $i=0$, the only connected finite CW-complexes with only finitely many (edit: abelian) homotopy groups are tori. (J.-P. Serre, Cohomologie modulo 2 des complexes d’Eilenberg-MacLane, Comment. Math. Helv. 27 (1953), 198-232.)

share|cite|improve this answer
Not just tori have finitely many non-trivial homotopy groups (e.g., surfaces of positive genus, etc.). Do you mean these are the only ones with abelian $\pi_1$? – Charles Rezk Oct 8 '10 at 22:08
Oops, right. Thanks. – Tilman Oct 9 '10 at 7:06

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.