1
$\begingroup$

Let $X$ be a path-connected CW-complex. If $\omega: [0,1] \to X$ is a loop in $X$ and $\partial$ the boundary operator in simplicial homology, then $\partial(\omega)=\omega(1)-\omega(0)=0$, i.e. $\omega$ is a cycle and represents a homology class $[\omega] \in H_1(X)$. Moreover, each homology class in $H_1(X)$ is represented by a loop (as is shown in the usual proof of $H_1(X) \cong \Pi_1(X)_{ab})$.

Denote by $X^2$ the 2-skeleton of $X$. It's well-known that the inclusion $X^2 \hookrightarrow X$ induces an isomorphism $H_1(X^2) \xrightarrow{\cong} H_1(X)$. Hence, as noted above, $\omega$ is homologous to a loop $\gamma: [0,1] \to X^2$.

Question: Is it possible to describe/construct $\gamma$ in an explicit way ?

$\endgroup$
3
  • 2
    $\begingroup$ More than homologous, $\omega$ is homotopic through loops to a loop in $X^2$, in fact to a loop in $X^1$. The homotopy is explicit in terms of $\omega$. To start, $image(\omega)\subset X^k$ for some $k$ (by compactness). Now do a tiny homotopy so that $image(\omega)$ misses the midpoint of each $k$-cell; this works as long as $k\ge 2$. Then compose $\omega$ with a deformation retraction from $X^k−$( midpoints of $k$−cells) to $X^{k−1}$. Then proceed by induction. $\endgroup$
    – Lee Mosher
    Aug 15, 2012 at 23:23
  • 1
    $\begingroup$ The inclusion of the 2-skeleton also induces an isomorphism on $\pi_1$, by the cellular approximation theorem. That well-known theorem produces $\gamma$ out of $\omega$. $\endgroup$ Aug 15, 2012 at 23:58
  • $\begingroup$ Lee, Fernando, thanks for your answers. You can also post them as answers - I would be glad to accept them (however, it seems that the question will be closed soon). $\endgroup$
    – Ralph
    Aug 16, 2012 at 9:23

0

Your Answer

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

Browse other questions tagged or ask your own question.