Let $\tilde{X} \to X$ be the universal cover of a connected, locally path-connected and semi-locally simply connected topological space $X$. Is it always true that the orbit space $$ \tilde{X} \;/\; \pi_1(X) $$ is naturally isomorphic to $X$? Or do we need any extra assumptions?

This is the example I have in mind: the torus $T^2$ and it's universal cover $\mathbb{R}^2$. The fundamental group is $\mathbb{Z}^2$, and we indeed have $$ \mathbb{R}^2\;/\; \mathbb{Z}^2 \cong T^2 $$


Let $p\colon\tilde{X}\rightarrow X$ be a covering map. The orbit space is the space of all orbits under the action of $\pi_1(X)$ on $\tilde{X}$ (the action given by lifts of loops at a point). Note however that, if a covering has a deck transformation group which is transitive on fibers, then it is regular, and every universal cover is a regular cover.

The orbit of a point $x\in\tilde{X}$ is then precisely the fiber of the point $p(x)\in X$. So, we have a natural homeomorphism between the quotient space $\tilde{X}/\pi_1(X)$ and $X$ which sends fibers over $p(x)\in X$ (equivalently orbits of the point $x\in\tilde{X}$) to the point $p(x)$. Note, this behaves well because the action of $\pi_1(X)$ on $\tilde{X}$ is properly discontinuous and free.

This is rather succinctly explained here.


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.