Classification of Covering Spaces

Covers of a space are subgroups of its fundamental group, and nothing less.

The idea

The classification theorem for covering spaces matches the connected covering spaces of a space $X$ with the subgroups of its fundamental group, so that finding every cover of $X$ becomes an algebra problem.

A covering space is a map $p \colon \tilde{X} \to X$ that lays evenly stacked sheets over every small enough open set $U \subseteq X$ (Covering Spaces). Each connected cover records a subgroup. Take a loop in $X$ at the basepoint $x_{0}$ and lift it, starting at a chosen point $\tilde{x}_{0}$ of the fibre $p^{-1}(x_{0})$: the lifted path either closes up into a loop or ends at another point of the fibre. The classes of loops whose lifts close up form the subgroup $H = p_{*}\pi_{1}(\tilde{X}, \tilde{x}_{0})$ of $\pi_{1}(X, x_{0})$.

Theorem (Classification of covering spaces).

Let $X$ be path-connected, locally path-connected, and semilocally simply connected, with basepoint $x_{0}$. Sending a basepointed connected cover $p \colon (\tilde{X}, \tilde{x}_{0}) \to (X, x_{0})$ to the subgroup $p_{*}\pi_{1}(\tilde{X}, \tilde{x}_{0})$ is a bijection between basepointed connected covers of $X$ and subgroups of $\pi_{1}(X, x_{0})$. Forgetting basepoints, connected covers correspond to conjugacy classes of subgroups.

Every cell complex satisfies the hypotheses. Under this correspondence the number of sheets is the index of $H$, the trivial subgroup gives the simply connected universal cover, and a normal subgroup gives a cover whose symmetries permute the sheets transitively, with the quotient $\pi_{1}(X, x_{0})/H$ as the group of those symmetries. The correspondence has the same shape as Galois theory, and borrows its name.

Ways to work on it

Not sure where to start? Take the ten-question placement test.