Let Y be a topological space and $\pi: X\rightarrow Y$ a covering map. Take two lifts of the covering map $\tilde {\pi_1}$, $\tilde {\pi_2} : X\rightarrow X$ such that they agree on $x_0 \in X$.

Under what hypothesis can I conclude that the two lifts coincide?

I have always seen proofs of existence and uniqueness together but I can't exactly single out what is needed for uniqueness only. I know the condition is connectedness, but I am not sure if I have to suppose it for $X$ or $Y$.

If you could also give a sketch of the proof, I would very much appreciate it.


More generally, let $f:Z\to Y$ be a map, and $g,h : Z\to X$ two lifts of $f$ that agree on $z_0\in Z$.

Then $\{x\in Z \mid g(x)=h(x)\}$ is nonempty (it contains $z_0$). Moreover, it is open : if $h(x) = g(x)$, then consider a basic open set $U$, neighbourhood of $f(x)$ such that $\pi^{-1}(U) =\displaystyle\coprod_{i\in I}U_i$ with each $U_i$ mapped homeomorphically onto $U$.

Then $\pi(h(x)) \in U$ so $h(x) \in U_i$ for some $i$. Let $\pi_i$ be the restriction/corestriction of $\pi$ to $U_i\to U$. Then, restricted to $h^{-1}(U_i)$, $\pi\circ h = f$, so $\pi_i\circ h_{\mid h^{-1}(U_i)} = f_{\mid h^{-1}(U_i)}$, and so $h= \pi_i^{-1}\circ f$ on $h^{-1}(U_i)$.

Similarly, on $g^{-1}(U_i)$, $g= \pi_i^{-1}\circ f$. So on $h^{-1}(U_i)\cap g^{-1}(U_i)$ (which contains $x$), $g=h$. Therefore our set is open.

It is also closed. This either follows trivially if you're assuming $Y$ to be $T_2$; but even if it's not then if $g(x) \neq h(x)$, then you can perform a similar argument to show that on a small enough neighbourhood of $x$, $g$ lands in $U_i$ while $h$ lands in $U_j, j\neq i$, so $U_i\cap U_j = \emptyset$, so that on this small neighbourhood, $g(y)\neq h(y)$. If this isn't clear you should do this yourself.

Thus we have found a nonempty clopen subset of $Z$: if $Z$ is connected, it follows that it's $Z$.

So we have

If $Z$ is connected and $f:Z\to Y$ has two lifts $g,h : Z\to X$ that agree on a point $z_0\in Z$, then $g=h$.

In your situation (after the edit) it means that if you assume $X$ to be connected, then the uniqueness follows.

If $X$ is not connected, then there are some counterexamples.

  • $\begingroup$ For very simple counterexamples where $X$ is not connected, you could take $Y$ to be a point and $X$ to be any discrete space, so any map $X\to X$ at all is a lift. $\endgroup$ – Eric Wofsey Jan 1 at 19:49
  • $\begingroup$ Thank you both! $\endgroup$ – Angelo Brillante Romeo Jan 1 at 19:59

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.