Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $p : Y \to X $ be a covering projection. Given a path $ u : I \to X $ and a point $ y \in Y $ with $p(y) = u(0) $, there exists a unique path $ \hat{u}:I \to Y $ with $p \hat{u} = u$ and $ \hat{u}(0) = y$.

The proof I have of this proceeds as follows:

Consider first the case where the image of $u$ is contained within an open set $U \subseteq X$ that's evenly covered by $p$. Let $h: p^{-1}(U) \to U \times D$ be a homeomorphism, where $D$ is a discrete space. Then we can write $ h(y) = (p(y),d) = (u(0),d) $ for some $d \in D$. If $\hat{u}$ is any lifting of $u$, then the composite $ \pi_2 h \hat{u} : I \to p^{-1}(U) \to U \times D \to D $ must be constant (it's a continuous map from a connected space to a discrete space). If in addition $ \hat{u}(0) = y$, then this must be the constant map with value $d$.

So the only possibility for $\hat{u}$ is the mapping $ t \mapsto h^{-1}(u(t),d)$.

I understand everything up to this last sentence. Why "$u(t)$"? Why not $u(g(t))$, where $g$ is some continuous function with $g(0) = 0$? It's probably something simple, but I can't see it. Any help would be greatly appreciated.


share|cite|improve this question
In fact, when the theorem says $p\hat{u} = u$, does that mean $p\hat{u}(t) = u(t) \ \forall t$? If so, then I can see why. I was taking $p\hat{u} = u$ to mean that their images are the same. – Martin P Nov 1 '11 at 19:19
Because $p(h^{-1}(x,d))=x$. That's the nature of the homeomorphism between $h:p^{-1}(U)\rightarrow X\times D$. – Thomas Andrews Nov 1 '11 at 19:28

To answer your comment: yes, $ p \hat{u} = u$ is an expression of equality as functions i.e. $ p\hat{u}(t) = u(t) $ for all $t$.

I'll expand a bit on Thomas Andrews' comment. Suppose $ p :Y \to X$ is a continuous map. Then a subset $U$ of $X$ is evenly covered by $p$ if there's a homeomorphism $ h: p^{-1}(U) \to U \times D$, where $D$ is a discrete space, such that $\pi_1 h = p$ (where $\pi_1$ is projection onto the first factor). So $ p(h^{-1}(x,d)) = \pi_1 h h^{-1}(x,d) = \pi_1 (x,d) = x$.

share|cite|improve this answer

The proof fails to state one last condition on $h$, namely, that $p$ on $p^{-1}(U)$ factors through $h$ - that is, the composition map $p^{-1}(U)\rightarrow U\times D \rightarrow U$ is $p_{|p^{-1}(U)}$, where $U\times D \rightarrow U$ is the natural projection.

So, $p(h^{-1}(x,d))=x$ for all $x\in U$.

So, if $p(\hat{u}(t)) = u(t)$ for all $t$, then $h(\hat{u}(t)) = (u'(t),d(t))$, for some $d:I\rightarrow D$. Since $D$ is discrete, though, $d(t)$ must be constant.

On the other hand, $u(t)=p(\hat{u}(t)) = p(h^{-1}(h(\hat{u}(t)))) = p(h^{-1}(u'(t),d)) = u'(t)$

So, $u'(t)=u(t)$, and $h(\hat{u}(t)) = (u(t),d)$

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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