I asked this question on math.SE, but even with a bounty, there were no answers/comments. I hope this is not too low-level for this site.

Suppose I have a covering map $\pi:E\rightarrow B$, and a path in $B$, which is just a map $f$ from $I=[0,1]$ to $B$; then I know I can lift this path to a map $\hat{f}:I\rightarrow E$. I was wondering about the following generalization.

If we define the subset $X\subset E\times \mathrm{Map}(I,B)$ as $$ X = \{(e,f)\mid f(0)=\pi(e)\}$$

Then path-lifting says there exists a map $p:X\rightarrow\mathrm{Map}(I,E)$. In this formulation, $p(e,f)$ is the lift $\hat{f}$ with $\hat{f}(0)=e$.

My question: is $p$ continuous?

I'm thinking the $\textrm{Map}$ spaces have the compact-open topology, but maybe that's not the right one for this question.

I guess this could be even further generalized: if $(Y,q)$ is a simply-connected space with basepoint, then we can still define $$ X=\{(e,f)\mid f(q)=\pi(e)\}\subset E\times\mathrm{Map}(Y,B)$$ and we still get a map $p:X\rightarrow\mathrm{Map}(Y,E)$, and we can still ask: is $p$ continuous?


Yes, $p$ is continuous. In proposition 3.7 of this paper it is shown that the lifting map $Map((I,0),(B,b))\to Map((I,0),(E,e))$, $f\mapsto \tilde{f}$ is continuous (and therefore a homeomorphism) when you fix $p(e)=b$. Essentially the same proof will tell you that your map $p$ is continuous, you just don't fix the initial point. Recall the compact open topology is generated by subbasic sets $\langle K,U\rangle=\{f|f(K)\subseteq U\}$ where $K$ is compact and $U$ is open. The key in the proof is to use basic open neighborhoods of the form $$\bigcap_{j=1}^{n}\Big\langle \left[\frac{j-1}{n},\frac{j}{n}\right],U_j\Big\rangle\cap \bigcap_{j=1}^{n-1}\Big\langle \left\{\frac{j}{n}\right\},U_j\cap U_{j+1}\Big\rangle$$ where $n\geq 3$ and $U_j$ evenly covers $\pi(U_j)$.

By using the exponential law, your generalization also has a "yes" answer when $Y=[0,1]^n$ but I think it is not so clear if it holds more generally, for instance, if $Y$ is not a compact metric space.

  • $\begingroup$ Why do we need $n\ge3$? $\endgroup$ – Steve D May 31 '18 at 17:57
  • $\begingroup$ @SteveD Well, $n= 2$ could be included for this particular proof but this is not really an important feature. You could always take finer partitions and consider only sets of the form given where $n$ is greater than some fixed number. Allowing $n=1$ would just require an unnecessary extra case to write. $\endgroup$ – Jeremy Brazas May 31 '18 at 18:38
  • $\begingroup$ Right, I think I followed the argument, but then seeing this requirement made me second-guess it. Sure, we can always refine, but my question was more "why do we have to?" $\endgroup$ – Steve D May 31 '18 at 18:52
  • 1
    $\begingroup$ You don't have to exclude $n=1$...I guess it's a preference. But the argument absolutely requires the form above for $n\geq 2$. Why would you want to consider $n=1$ as something special when you don't have to? My instinct to start with $n\geq 3$ is related to covers of simplicial complexes...that is all. $\endgroup$ – Jeremy Brazas May 31 '18 at 23:23
  • 1
    $\begingroup$ Thank you for your answer! I think I fully understand how this works now. Do you want to post a similar answer on math.SE? I could also post my understanding of your answer there. Again, thanks! $\endgroup$ – Hempelicious Jun 7 '18 at 2:34

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.