I am completely stuck on the following algebraic topology exercise:

Let $X$ and $Y$ be CW complexes and $\alpha \in H_p(X)$, $\beta \in H_q(Y)$, $p, q > 0$, homology classes such that the homology cross product $\alpha \times \beta \neq 0$ in $H_{p+q}(X \times Y).$ Show that $\alpha \times \beta$ is aspherical, i.e. there is no map $\phi \colon S^{p+q} \to X \times Y$ such that $\phi_*[S^{p+q}] = \alpha \times \beta$, where $[S^{p+q}]$ denotes the fundamental class of $S^{p+q}$.

I am quite sure that the Hurewicz theorem will play a prominent rule in the solution, but I just don't see how.

Maybe it could be useful to play around with the definition of the cellular homology of $X$ and $Y$, since it involves maps from disks and spheres to $X$ and $Y$, but again I don't see how this might actually lead to a solution.

(The exercise is taken from the book Stöcker/Zieschang - Algebraische Topologie and is not a homework problem.)


I don't see how to use Hurewicz's theorem since we don't know anything about lower-dimensional classes. I would try to use the product structure in cohomology, e.g. if $a$ is a cohomology class dual to $\alpha$ and $b$ is a cohomology class dual to $\beta$ then $a\cup b$ will be a nonzero class. If it pairs nontrivially with a homology class represented by a sphere $S^{p+q}$ then you can pull back the classes $a$ and $b$ to the sphere and get nontrivial (co)homology classes there that are "forced" on you by the naturality of the cup product, yielding a contradiction. The technical details may not work out exactly as I stated but I am confident this direction has a better chance of succeeding than Hurewicz.

| cite | improve this answer | |
  • $\begingroup$ Sorry, but I don't understand the details of your solution. I would be very glad, if you could try to give a more in-depth account of your argument. $\endgroup$ – AlgTop Aug 5 '13 at 6:19
  • $\begingroup$ The idea is to exhibit a $p+q$-dimensional cohomology class in $X\times Y$ which is a cup product of cohomology classes of dimensions $p$ and $q$. Since the product structure in cohomology is natural with respect to pullback of maps, the sphere $S^{p+q}$ would also inherit such a product decomposition for the pullback class, which is impossible. $\endgroup$ – Mikhail Katz Aug 5 '13 at 7:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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