Is continuous map from covering space to itself homeomorphism assumed both cover and base path-connected and $pf=p$?

802 Views Asked by At

In my topology assignment I came across the following problem:

True or false? Let $E$ and $X$ be path-connected. For every covering map $p:E\rightarrow X$ and continuous map $f:E\rightarrow E$ such that $pf=p$ holds $f$ is homeomorphism.

I couldn't think of any counterexample, so my guess is that assertion is true. Since $X$ is not locally path-connected, I cannot apply lifting theorem to obtain $f$ and its inverse, so I guess I should manually prove properties of homeomorphism on $f$, using path lifting theorem several times.

Recently I had to deal with the operation of a fundamental group operation on a set $p^{-1}(x)$ for $x\in X$ such that $p^{-1}(x)\times \pi_1(X,x)\rightarrow \pi_1(X, x): [w]\cdot e=w(1)$, that is each point in $E$ maps to the endpoint of a unique lift of the path $w$ (similar to the one in Hatcher). I played around a little trying to prove injection / surjection property of $f$, but didn't got far. The only thing I proved is that $f([w]\cdot e) = [w]\cdot f(e)$ but nothing more.

So I ask what would be an adequate approach to solve this question.

Thanks in advance!

2

There are 2 best solutions below

7
On BEST ANSWER

Hint 1: Here's a hint for showing that $f$ is a local homeomorphism: The condition $pf=p$ implies that points in $e$ are carried to points in the fiber $p^{-1}(p(e))$. We can find neighborhoods $U$ of $e$ and $W$ of $f(e)$ such that $p$ restricts to a homeomorphism on each neighborhood and $f(U) \subset W$. See if you can use this to show that $f$ in fact carries $U$ homeomorphically into a subset of $W$.

Hint 2: [Courtesy of the OP's own suggestion in the comments below.] For both injectivity and surjectivity: The action of $\pi_1(X,x_0)$ on $p^{-1}(x_0)$ can be shown to satisfy $f(e) \cdot [\gamma]=f(e \cdot [\gamma])$. Playing around with this fact and a couple well-chosen paths between points in $p^{-1}(x_0)$ will give both injectivity and surjectivity.

0
On

Unfortunately for the OP, their topology assignment is wrong, even with the stronger hypothesis of $E$ and $X$ being connected and locally path connected.

There is a counter example for this in the paper of Lima called Fundamental Groups and Covering Spaces on page 179, example 7.20.

The mechanics of this counterexample is clearly described also in this MO question https://mathoverflow.net/questions/80105/why-do-the-definition-of-deck-transformations-requires-homeomorphism