1. Loop Space
Let I be the closed unit interval [0, 1] and X be a topological space. The space of paths in X is the set C(I, X) of continuous maps from I to X.
For each compact subset K of I and open subset U of X, we denote (K : U ) = {σ ∈ C(I, X) : σ(K) ⊂ U }.
The set {(K : U ) : K ⊂ I compact, U ⊂ X open} forms a subbase for a topology on C(I, X). We call this topology the compact-open topology on C(I, X). For simplicity, we denote C(I, X) by XI.
Proposition 1.1. The evaluation : XI× I → X defined by (σ, t) = σ(t) is continuous.
Let x0 be a point in X. We denote Ω(X, x0) = {σ ∈ XI : σ(0) = σ(1) = x0}. Then Ω(X, x0) is a closed subspace of XI. We call Ω(X, x0) the loop space.
Proposition 1.2. The fundamental group π1(X, x0) of (X, x0) is exactly π0(Ω(X, x0)).
Given x0 ∈ X, there is a natural based point ex0 on the loop space: let ex0 be the constant loop at x0. Then (Ω(X, x0), ex0) is a pointed topological space. We define
π2(X, x0) = π1(Ω(X, x0), ex0).
Inductively, we define the higher homotopy groups of the pointed space (X, x0) by πn(X, x0) = πn−1(Ω(X, x0), ex0).
Proposition 1.3. The homotopic groups πn(X, x0) are all commutative for n ≥ 2.
Let f : (X, x0) → (Y, y0) be a morphism of pointed spaces. We define Ω(f ) : (Ω(X, x0), x0) → Ω(Y, y0)
by setting Ω(f )(σ) = f ◦ σ for σ ∈ Ω(X, x0). Then Ω(f ) is a morphism of pointed spaces.
Hence we can inductively defined a homomorphism of groups:
πn(f ) : πn(X, x0) → πn(Y, y0).
Proposition 1.4. For each n ≥ 1, πn is a functor from the category of pointed topological spaces into the category of groups.
Corollary 1.1. If X is contractible, πn(X, x0) is trivial for all n.
Proposition 1.5. For all n, there is a canonical isomorphism πn(X × Y, (x0, y0)) ∼= πn(X, x0) × πn(Y, y0).
Theorem 1.1. If p : (E, e0) → (X, x0) is a covering space, then πn(p) : πn(E, e0) → πn(X, x0) are isomorphisms for all n ≥ 2.
The m-dimensional real projective space is the quotient space of Rm+1\ {0} modulo the relation ∼ defined by
x ∼ y iff x = λy for some λ 6= 0.
Corollary 1.2. Let m ≥ 2. For all n ≥ 2,
πn(RPm) ∼= πn(Sm).
Here RPn is the n-dimensional real projective space.
1
2
Proof. Notice that we have a surjective (quotient) map p from Sm into RPm given by x 7→ [x]. Then p is a 2-sheet covering map. We choose any point x0 on Sm as our fixed point and chose y0= p(x0) as the based point of RPm. Corollary 1.3. For any n ≥ 2, πn(S1) = {0}.
Proof. Let p : R → S1 defined by p(t) = e2πit. Then p is a covering over S1. Hence πn(R, 0) ∼= πn(S1, 1) for all n ≥ 2. Since R is contractible, by Corollary 1.1, πn(R, 0) = 0 for all n.