1
$\begingroup$

Suppose $\Gamma$ is a finitely presented group that has a torsion element. Can the classifying space $K(\Gamma,1)$ be homotopic to a finite-dimensional manifold? If yes, what is the simplest example?

$\endgroup$
2
  • 4
    $\begingroup$ The answer is no, because the presence of torsion forces $\Gamma$ to have infinite cohomological dimension. See e.g. Ken Brown's "Cohomology of groups", Chapter VIII.2. $\endgroup$
    – Mark Grant
    Commented Mar 29, 2017 at 15:57
  • 3
    $\begingroup$ Note that there is a version of classifying spaces $\underline{E}G$ where $G$ acts so that compact (i.e. finite) subgroups have (weakly) contractible fixed point sets. Sometimes groups with torsion can have finite-dimensional $\underline{E}G$. ams.org/mathscinet-getitem?mr=1292018 $\endgroup$
    – Ian Agol
    Commented Mar 29, 2017 at 17:40

1 Answer 1

8
$\begingroup$

This is not possible. Let $B$ be a space of type $K(\Gamma,1)$ and of dimension $d<\infty$, and let $E$ be the universal cover, which is contractible. Then for any $A\leq\Gamma$ we see that $E/A$ is a $K(A,1)$ and has finite dimension, so the group cohomology $H^i(A)$ vanishes for $i>d$. However, for $n>1$ the group cohomology of $C_n$ is nonzero in all even degrees, so $\Gamma$ cannot contain $C_n$, so $\Gamma$ must be torsion-free.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .