5
$\begingroup$

The question is in the title : Can we find spaces $A$ and $B$, each non contractible, such that their smash product $A \wedge B$, i.e. the homotopy cofibre of $A \vee B \to A \times B$, is a contractible space ?

$\endgroup$
2
  • 4
    $\begingroup$ Maybe. Take $A=S^1$ and $B$ the classifying space of a group with trivial homology, but I don't know if these exist. $\endgroup$ Sep 23, 2015 at 15:14
  • 4
    $\begingroup$ @FernandoMuro: there are many groups with trivial homology. There is a general construction as part of the proof of the Kan-Thurston theorem, for example. $\endgroup$ Sep 23, 2015 at 15:31

2 Answers 2

8
$\begingroup$

Oh yes ! Fernando gave the hint ! Take $A = S^1$ and $B$ = the Epstein's space. Then $A \wedge B \simeq \Sigma B$ is contractible but $B$ is not !

$\endgroup$
3
  • $\begingroup$ What's the Epstein's space? $\endgroup$ Sep 26, 2015 at 18:42
  • $\begingroup$ Here is the reference for Epstein's space : D.B.A.Epstein, "A group with zero homology", Mathematical Proceedings of the Cambridge Philosophical Society 64(03):599 - 602 · July 1968 $\endgroup$
    – jpaul
    Sep 30, 2015 at 10:31
  • $\begingroup$ Thank you. From your comment, I understand that Epstein's space is the classifying space for a certain group. $\endgroup$ Sep 30, 2015 at 20:47
6
$\begingroup$

I just saw this in Ravenel's "orange book"! Let $X$ be any simply connected CW complex whose reduced homology is all torsion. Let $X_{(p)}$ be the p-localization for a prime $p$. Then for primes $p\neq q$, $X_{(p)}\wedge X_{(q)}$ is contractible. But neither $X_{(p)}$ nor $X_{(q)}$ is contractible if $H_*(X)$ has $p$ and $q$ torsion.

$\endgroup$
2
  • 6
    $\begingroup$ In the same spirit, just take the smash product of two CW complexes that are Moore spaces for finite cyclic groups of relatively prime orders. $\endgroup$ Sep 23, 2015 at 23:50
  • $\begingroup$ Oh yeah, that's much simpler :-) $\endgroup$ Sep 24, 2015 at 2:08

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.