10
$\begingroup$

Dear community,

assume that $f\colon X\to Y$ is a homology equivalence, i.e. induces isomorphisms on all integer homology groups. Is there a map $g\colon Y\to X$ which is also a homology equivalence?

Actually, I am only interested in the case if $f\colon X\to X^+$ is obtained by Quillen´s plus construction.

Thanks for any comments in advance.

$\endgroup$
5
  • 1
    $\begingroup$ In extreme generality no - take the circle and map it onto the non-Hausdorff space "$[-1, 1]$ with two-origins" (by projecting vertically downward, mapping the north and south poles to the separate origins). This gives an isomorphism on all homology groups, but any map the other way is homotopic to a constant since the origins collapse. $\endgroup$ Aug 30, 2012 at 15:31
  • 5
    $\begingroup$ My guess is that the $X =$ the Poincare homology sphere should provide a counterxample. If there were a map $S^3 \to X$ which induced a homology isomorphism, then this map is necessarily of degree one and we see that the fundamental class of $X$ is spherical. I don't see it at the moment, but this is very unlikely to be true. $\endgroup$
    – John Klein
    Aug 30, 2012 at 15:44
  • 4
    $\begingroup$ @John: Such a map would lift to the universal cover of $X$, so have degree divisible by 120. $\endgroup$ Aug 30, 2012 at 15:57
  • $\begingroup$ Yep. I was just posting it as an answer, but you beat me to the punch. $\endgroup$
    – John Klein
    Aug 30, 2012 at 16:00
  • $\begingroup$ Merci beaucoup. $\endgroup$ Aug 30, 2012 at 21:51

1 Answer 1

13
$\begingroup$

As I was writing this answer, Oscar beat me to the punch. I will keep it posted anyway.

Let $X^3$ be the Poincare homology sphere. Let $\tilde X \to X$ be the universal cover (note: $\tilde X$ is $S^3$). As in my comment, if there were a map $S^3 \to X$ inducing a homology equivalence, then that map must necessarily factor through $\tilde X$ (by covering space theory). But this is impossible since $\tilde X \to X$ has degree 120, whereas the composite $S^3 \to \tilde X \to X$ is supposed to have degree one.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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