I have the following problem:

Let $M$ be a connected closed $7$-manifold such that $H_1(M,\mathbb{Z}) = 0$, $H_2(M,\mathbb{Z}) = \mathbb{Z}$, $H_3(M,\mathbb{Z}) = \mathbb{Z}/2$. Compute $H_i(M,\mathbb{Z})$ and $H^i(M,\mathbb{Z})$ for all $i$.

I know that if $M$ is orientable, using Poincaré duality, the fact that $\chi(M)=0$ and the exact sequence for $H^i(M,\mathbb{Z})$ I can get the result.

But, I don't know how to prove that $M$ is orientable. I know that is the case if $M$ does not have $2$-torsion on $\pi_1(M)$, but I don't see why this $2$-torsion should descend to $H_1(M)$.

  • $\begingroup$ Yeah, sorry I will edit. $\endgroup$ – abrax Jul 19 '16 at 5:31
  • 1
    $\begingroup$ Just to point out the connection between the two answers: $Hom(\pi_1(M),\mathbb Z/2)\cong Hom(H_1(M),\mathbb Z/2)\cong H^1(M,\mathbb Z/2)$. Eric talks about a distinguished element on the left side, which under the natural isomorphisms corresponds to $\omega_1(M)$ on the right side, which is the side which Najib uses for his argument. $\endgroup$ – Daniel Valenzuela Jul 19 '16 at 14:19

For any connected manifold $M$, there is a homomorphism $\pi_1(M)\to\mathbb{Z}/2$ which sends a loop to $0$ if going around the loop preserves orientation and sends the loop to $1$ if going around the loop reverses orientation. This homomorphism is trivial iff $M$ is orientable. Since $\mathbb{Z}/2$ is abelian, this homomorphism factors through the Hurewicz map $\pi_1(M)\to H_1(M)$. In particular, this means that if $H_1(M)=0$, the homomorphism is trivial so $M$ is orientable.

(By the way, the statement that $\chi(M)=0$ does not require $M$ to be orientable--you can prove it using mod $2$ Poincare duality, for instance.)


A quick proof using Stiefel–Whitney classes: a manifold $M$ is orientable iff the first SW class $w_1(M) \in H^1(M;\mathbb{Z}/2\mathbb{Z})$ is zero. But by the universal coefficient theorem, $$H^1(M;\mathbb{Z}/2\mathbb{Z}) = \operatorname{Hom}_\mathbb{Z}(H_1(M;\mathbb{Z}), \mathbb{Z}/2\mathbb{Z}) = 0.$$

Of course under the hood I don't think there's anything more than what you can find in Eric Wofsey's answer, but this argument is quite simple and shows the power of characteristic classes.


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.