ProofComplete

Covering Spaces - Key Proof

We prove the fundamental path lifting theorem, which is the foundation for all other lifting properties of covering spaces.

Proof

Theorem (Path Lifting): Let p:X~Xp : \tilde{X} \to X be a covering space, γ:[0,1]X\gamma : [0,1] \to X a path with γ(0)=x0\gamma(0) = x_0, and x~0p1(x0)\tilde{x}_0 \in p^{-1}(x_0). Then there exists a unique path γ~:[0,1]X~\tilde{\gamma} : [0,1] \to \tilde{X} with pγ~=γp \circ \tilde{\gamma} = \gamma and γ~(0)=x~0\tilde{\gamma}(0) = \tilde{x}_0.

Existence: We construct the lift using the Lebesgue number lemma. Since [0,1][0,1] is compact, γ1\gamma^{-1} of the cover of γ([0,1])\gamma([0,1]) by evenly covered neighborhoods has a Lebesgue number δ>0\delta > 0.

Partition [0,1][0,1] into subintervals Ik=[tk1,tk]I_k = [t_{k-1}, t_k] with Ik<δ|I_k| < \delta, so γ(Ik)\gamma(I_k) lies in an evenly covered neighborhood UkU_k.

Inductive construction:

  • Base: Start with γ~(0)=x~0\tilde{\gamma}(0) = \tilde{x}_0.
  • Step: Suppose γ~\tilde{\gamma} is defined on [0,tk][0, t_k]. The set p1(Uk+1)p^{-1}(U_{k+1}) is a disjoint union of open sets U~α\tilde{U}_\alpha, each mapped homeomorphically onto Uk+1U_{k+1} by pp.
  • Since γ~(tk)p1(γ(tk))p1(Uk+1)\tilde{\gamma}(t_k) \in p^{-1}(\gamma(t_k)) \cap p^{-1}(U_{k+1}), it lies in exactly one sheet U~α\tilde{U}_\alpha.
  • Extend γ~\tilde{\gamma} over Ik+1I_{k+1} using (pU~α)1γ(p|_{\tilde{U}_\alpha})^{-1} \circ \gamma on Ik+1I_{k+1}.

By induction, γ~\tilde{\gamma} is defined on all of [0,1][0,1], and by construction pγ~=γp \circ \tilde{\gamma} = \gamma.

Uniqueness: Suppose γ~1,γ~2\tilde{\gamma}_1, \tilde{\gamma}_2 are two lifts starting at x~0\tilde{x}_0. Let A={t[0,1]:γ~1(t)=γ~2(t)}A = \{t \in [0,1] : \tilde{\gamma}_1(t) = \tilde{\gamma}_2(t)\}.

We show AA is both open and closed in [0,1][0,1]:

  • Closed: If tnAt_n \in A with tntt_n \to t, then by continuity γ~1(t)=limγ~1(tn)=limγ~2(tn)=γ~2(t)\tilde{\gamma}_1(t) = \lim \tilde{\gamma}_1(t_n) = \lim \tilde{\gamma}_2(t_n) = \tilde{\gamma}_2(t), so tAt \in A.
  • Open: If tAt \in A, let UU be an evenly covered neighborhood of γ(t)\gamma(t). Then γ~1(t)=γ~2(t)\tilde{\gamma}_1(t) = \tilde{\gamma}_2(t) lies in a unique sheet U~\tilde{U} of p1(U)p^{-1}(U). For small ϵ\epsilon, both γ~1\tilde{\gamma}_1 and γ~2\tilde{\gamma}_2 map [tϵ,t+ϵ][t-\epsilon, t+\epsilon] into U~\tilde{U}. Since pp is injective on U~\tilde{U} and pγ~1=pγ~2=γp \circ \tilde{\gamma}_1 = p \circ \tilde{\gamma}_2 = \gamma, we have γ~1=γ~2\tilde{\gamma}_1 = \tilde{\gamma}_2 on [tϵ,t+ϵ][t-\epsilon, t+\epsilon]. Thus AA is open.

Since AA is non-empty (contains 0), open, and closed in the connected space [0,1][0,1], we have A=[0,1]A = [0,1]. Therefore γ~1=γ~2\tilde{\gamma}_1 = \tilde{\gamma}_2. ∎

Remark

The proof of uniqueness uses a standard "open and closed" argument: if a subset of a connected space is both open and closed, it must be empty or the whole space. This technique appears throughout topology.

Remark

The homotopy lifting theorem follows by applying the path lifting theorem to each path in a family parameterized by the homotopy. The key is that the Lebesgue number argument works uniformly over the parameter space.

From path lifting, all other properties follow: loops lift to paths (which may not be loops), homotopies lift uniquely, and the induced map pp_* on π1\pi_1 is injective. This makes covering spaces indispensable for computing fundamental groups.