Theorem: Let Y=I. Fix x0∈X and z∈p−1(x0). Let α,β:I→X be equivalent loops based at x0, meaning α≊β.
α has a unique lift α~z:I→Z (a path, occasionally a loop) such that α~(0)=z.
If F:I2→X is the path-homotopy realizing α≊β, it has a unique lift F~:I2→Z such that F~0=α~z (the bottom edge of I2).
Proof:
One can partition I into closed intervals [aj,aj+1] such that α([aj,aj+1]) lies in an open set U covered by sheets. α~ is defined on each in succession by bijectivity.
This follows from a similar argument by dividing the square I2 into rectangles Rjk, each mapped into a U, then working upwards from F~0 to define F~ on each column. ■
F~ must be constant on the vertical lines s=0,1 so F~(0,t)=z and F~1=β~z. Thus:
β~z(1)=F~(1,1)=F~(1,0)=α~z(1).
Corollary: The endpoint α~z(1) of Z depends only on the class [α] of α in π1(X,x0), so fixing z determines a map π1(X,x0)→p−1(x0).
Groups Acting on Sets
Let G be a group (finite or infinite, with identity element e), and Ω any set.
Definition: A right action of G on Ω is a mapping:
Ω×G→Ω,(z,g)↦z⋅g,
such that:
z⋅e=z,
z⋅(gh)=(z⋅g)⋅h for all z∈Ω and g,h∈G.
If Ω is finite of size n, this simply means that there is a group homomorphism G→Sn.
Examples:
Let Ω=R3={(x1,x2,x3)},G=SO(3)={A∈R3,3:ATA=I,detA=1}. G acts on Ω on the right. It also acts on S2 on the right. Let C be a cube centred at the origin. The subgroup of G mapping C onto C is known to be isomorphic to S4; it contains 2-cycles, 3-cycles, 4-cycles (how do each of these rotate the cube?).
Lemma: Fix z∈Ω. Then Gz={g∈G:z⋅g=z} is a subgroup of G, called the stabilizer of z, and g↦z⋅g identifies the set {Gzg:g∈G} of right cosets with Ω. Note that G acts on this set of cosets on the right, so we can dispense with Ω.
Action of π1 on a Fibre
Let p:Z→X be a covering map. Fix x0∈X. Set:
G=π1(X,x0),Ω=p−1(x0),the ’fibre’ over x0.
Theorem:
Setting z⋅[α]=α~z(1) defines a right action of G on Ω.
If z∈Ω, the induced homomorphism p∗:π1(Z,z)→π1(X,x0) is injective.
The stabilizer {[α]:z⋅[α]=z} is precisely the subgroup p∗(π1(Z,z)) of G.
Proofs:
It is obvious that z⋅[ε]=z. To prove that z⋅([α][β])=(z⋅[α])⋅[β], observe that αβ~z=α~zβ~z′, where z′=α~z(1) (since α~zβ~z′ is a lift of αβ from z).
If γ is a loop at z such that p∘γ≊Fε, then the unique lift F~ gives γ≊εz.
If α~z=z, then α~z is a loop and:
[α]=[p∘α~z]=p∗[α~z].
Conversely, if [α]=p∗[γ] with γ a loop based at z, then γ=α~z and z⋅[γ]=z. ■
Fundamental Group of the Circle
Let p:Z→X be a covering map, fix x0∈X and z∈p−1(x0). Set G=π1(X,x0) and Ω=p−1(x0) as before.
Lemma: The map G→Ω defined by [α]↦z⋅[α] is:
surjective if Z if path-connected,
bijective if Z is simply-connected (meaning path-connected and π1(Z)={e}).
Proof of Surjectivity: Given z′∈Ω, choose a path σ from z to z′ in Z. Then α=p∘σ is a loop based at x0. By uniqueness, α~z=σ, so α~z(1)=z′. Proof of Injectivity: If z⋅[α]=z⋅[β] then z⋅g=z where g=[α][β]−1. But the stabilizer of z is trivial by part (3) of the previous theorem, so [α]=[β].
We can apply the lemma to X=S1⊂C, x0=1,Z=R,p(z)=e2πis. Then Ω=Z. The path s↦sn in R projects to the loop αn:s↦exp(2πisn) in S1, and 0⋅[αn]=n.
Corollary:π1(S1,1) is isomorphic to (Z,+), i.e. the infinite cyclic group F1.
Double Coverings
We previously defined RPn to be the quotient of Sn obtained by identifying each point v=(x0,x1,...,xn) with −v=(−x0,−x1,...,−xn), its antipodal point.
Recall that RP2 is homeomorphic to the surface P originally defined from a square. The square model showed that P is also a quotient of a circular disk D in which opposite points on the boundary ∂D are identified. Now deform D into the southern hemisphere of S2 and retain ∂D as its equator. Then any equivalence class {v,−v} (defining a point on RP2) has a unique representative in P.
Let's explain why the group SO(3) of rotations fixing the origin is homeomorphic to RP3. Any non-identity rotation is determined by an axis (i.e. a unit vector v), and an angle θ∈[0,π]. Map this to the point (πθ)v in R3. The only ambiguity is that an angle π around v is the same rotation as π around −v. So we are identifying antipodal points of the boundary S2=∂B of the solid unit ball B. Since B is just a 3-dimensional analogue of D above, the resulting space is RP3 by the same argument.
Torus to Klein Bottle in Squares
Represent the torus T by the square I2 with boundary aba−1b−1, giving qT:I2→T. Represent the Klein bottle K by I2 with boundary cd−1c−1d−1, giving qK:I2→K.
is a 2:1 covering map (the quotient map P→P^ is denoted by q rather than π).
If we regard a,c as loops in T,K respectively, then p∘a=cc, whilst p∘b=d.
T to K in Letters and Words
Fix z=qT(0,1) as a basepoint z∈T, and x0=qK(0,1)∈K. The interior of qK(I2) allows us to define a homotopy cd−1c−1d−1≊ε with basepint x0.
Abusing notation to identify loops with classes in π1(K,x0), write cd−1c−1d−1=e:
⟹cd−1=dc,dc−1=c−1d−1,⟹c2d−1=cdc=d−1c2.
If we now set a=c2 and b=d−1 (justified since p∗:π1(T,x0)→π1(K,x0) is injective) we recover the relation ab=ba or aba−1b−1=e in π1(T,z).
We shall see that π1(K,x0) is the group generated by c and b subject only to the relation cbc−1b=e, induced by the boxed homotopy that shrinks the loop to a point. It follows that π1(T,z) is isomorphic (via p∗) to the abelian subgroup generated by a=c2 and b. Each element of π1(T,z) can be written uniquely as ambn for some m,n∈Z. Converting to additive notation:
π1(T,x0)≅Z⊕Z≅Z2.
Summary
Topologically the 2-torus T=T2 is the product S1×S1, and there is a 'combined' covering map R2→T defined by:
(s,t)↦(e2πis,e2πit).
A theorem about π1 of a product of spaces confirms the isomorphism π1(T)≅Z×Z.