Let $X$ be a simply connected finite CW-complex such that all but finitely many of its homotopy groups and its homology groups (with $\mathbb Z$ coefficients) are 0.

Is $X$ then necessarily contractible?

I do not really believe that this is true; but I was also not able to construct a counterexample.

  • 3
    $\begingroup$ Consider the rationalization of the $n$-sphere, call it $X_n$, defined as the mapping telescope of a sequence of maps $S^n \to S^n \to S^n \to \cdots$ in which the $i$-th map has degree $i$. The space $X_n$ has just one non-trivial reduced homology group. Moreover, the homotopy groups of $X_n$ can also be computed knowing that sequential homotopy colimits commute with homotopy groups. It follows that $X_n$ has non-trivial homotopy groups in at most two degrees, $n$ and perhaps $2n-1$, which correspond precisely to the homotopy groups of the $n$-sphere which are infinite. $\endgroup$ Jun 20, 2015 at 18:11
  • 2
    $\begingroup$ Alternatively, fix an odd natural number $n$, and consider the Eilenberg–MacLane space $K(\mathbb{Q},n)$; it has the same homotopy type as $X_n$ above. One can prove that $K(\mathbb{Q},n)$ has reduced homology concentrated in degree $n$ by applying inductively the Serre spectral sequence for the fibre sequences $K(\mathbb{Q},i-1) \to \ast \to K(\mathbb{Q},i)$. $\endgroup$ Jun 20, 2015 at 18:27
  • $\begingroup$ Thnaks for this comment. But in fact I was thinking about ${finite}$ CW-complexes, so David's answers was what I was looking for. $\endgroup$ Jun 21, 2015 at 19:03

1 Answer 1


By results of J.P. Serre (for $p=2$) and Y. Umeda (for odd $p$) we know that a 1-connected finite CW-complex $X$ with non-trivial cohomology mod $p$ has infinitely many non-trivial homotopy groups mod $p$.

In fact C.A. McGibbon and J.A. Neisendorfer have proved the existence of $p$-torsion elements in infinitely many dimensions for finite dimensional spaces using H. Miller's theorem on the Sullivan conjecture.


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.