1

Does anyone have an example of two spaces which have the same homology groups, the same cohomology groups, but have different cohomology rings?

is it possible?

rhl
  • 133
  • 2
    sounds like a homework problem to me. – Ian Agol Mar 02 '11 at 17:59
  • it isn't. I am teaching myself the material. – rhl Mar 02 '11 at 19:40
  • Slightly different example I quite like: the two partial flag manifolds for $SO(5)$ have the same cohomology groups (free abelian, in even degrees), the same rational cohomology rings, and different integer cohomology rings. – Allen Knutson Mar 04 '11 at 04:07

2 Answers2

8

A very standard example would be $S^2\vee S^4$ and $\mathbf{C}P^2$.

Neil Strickland
  • 55,003
  • hm, OK. but these have different homotopy groups,

    Looks like \pi_5(S^2 \wedge s^4) = Z_2 x Z_2 but \pi_5(CP^2) = Z

    correct?

    It's not clear to me that the cohomology ring is actually a finer invariant than cohomology. When I read, all sources say 'the structure is richer than homology,' but is it useful? maybe mathoverflow is not an appropriate forum for this question?

     – rhl Mar 02 '11 at 19:44
  • math.stackexchange.com might be better for you. Also, this is the first example I learned. It satisfies all of the requirements listed in your question. The ring structure on $S^2 \vee S^4$ is trivial, but not on $\mathbb{C}P^2$. – Sean Tilson Mar 02 '11 at 21:43
  • oh, and cohomology is easier to compute than homotopy for finite dimensional cell/CW complexes. – Sean Tilson Mar 02 '11 at 21:44
3

There are also standard examples in which both spaces are compact manifolds. For instance, if $n \geq 1$ is an integer and $Q_n \subset \mathbb{P}^{2n+2}$ is a non-singular quadric, then $Q_n$ has the same integral homology and cohomology groups as $\mathbb{P}^{2n+1}$, but the cohomology rings are different.

damiano
  • 5,107