Currently I am studying for a topology exam next week and came across an exercise where I could need some hints (cf. here):

Let $X$ be a path-connected space with $\pi := \pi_1(X,*)$ abelian and contractible universal cover $\widetilde{X}$. Suppose $f \colon Y \rightarrow X$ is a continuous map so that the induced map $f^* \colon H^1(X, \pi) \rightarrow H^1(Y, \pi)$ is zero. Prove $f$ is null-homotopic.

Some thoughts I had so far: From Hurewicz and the fact that the fundamental group is already abelian we get that $\pi \cong \pi^{ab} \cong H_1(X;\mathbb{Z})$. Also from the universal coefficient theorem we get (since $H_0$ is always free and so the $Ext$-term vanishes, $G$ abelian group) an isomorphism $H^1(X; G) \rightarrow Hom_{\mathbb{Z}}(H_1(X;\mathbb{Z}), G)$. Similarly for $Y$, and we could take $G = \pi$ in the isomorphism above.

Since we somehow have to know something about the homotopy type of $f$ and we have the assumption that the universal covering is contractible, maybe we are looking for some lift

\begin{array} & & & \widetilde{X} \\ & \nearrow & \downarrow{p} \\ Y & \xrightarrow{f} & X \end{array}

of $f$ to the universal cover (then the map is nullhomotopic because it factors through a contractible space).

We can probably get that the map induced by $f$ on the fundamental groups is trivial and with more assumtions on $Y$ (path-connected and locally path-connected) that would give us the lift. But in the exercise Y seems to be a general topological space.

What approach would you take to this?


They mean to have the locally path-connected assumption, presumably, but didn't feel the need to state it (often "space" can be taken to mean "non-terrible space").

As a counterexample in general, take the Warsaw circle $X$. This has trivial homotopy groups and homology groups (exercise), but collapsing the 'bad part' gives us a non-null map $X \to S^1 =: Y$. This is Hatcher's exercise 1.3.7.

  • $\begingroup$ Thanks for your comment! How would you show that the map induced by $f$ on the fundamental group is trivial? Using the UCT and the vanishing of the Ext-terms I can get a diagram that shows the dual of f, $(f_*)^* \colon Hom(H_1(X;\mathbb{Z}), \pi) \rightarrow Hom(H_1(Y;\mathbb{Z}), \pi)$ is zero, how do you recover the information about $f_*$? $\endgroup$
    – ben300694
    Jul 21 '16 at 23:10
  • $\begingroup$ @ben If $i \in H^1(X;\pi)$ is the identity homomorphism, then try to prove that $f^*i$ is the induced homomorphism on fundamental groups when $Y$ is path-connected, and see if you can conclude the general non-path-connected case from this. $\endgroup$
    – user98602
    Jul 21 '16 at 23:13

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.