1.

BONUS: LOCAL GEOMETRY 15

E XERCISE 1.4. Either prove the following, or find a counterexample: if K is a simplicial com-
plex and L ⊂ K a subcomplex with L 6= K, then the complement K − L is also a subcomplex
of K.
E XERCISE 1.5. Assume that L, L0 are two subcomplexes of a simplicial complex K with a
nonempty intersection. Show that this intersection L ∩ L0 is also a subcomplex of K.

E XERCISE 1.6. Let K be a k-dimensional simplicial complex, and for each dimension i in
{0, 1, . . . , k} let ni be the number of i-simplices in K. How many i-simplices does the barycen-
tric subdivision Sd K have for each dimension i?
E XERCISE 1.7. Let M be a finite metric subspace of an ambient metric space ( Z, d). Show,
for each e > 0, that the associated Čech complex Ce ( M ) is a subcomplex of the Vietoris-Rips
complex VR2e ( M). Then, show that – no matter what Z we had chosen – this VR2e ( M) is itself
a subcomplex of C2e ( M).

E XERCISE 1.8. Consider any homeomorphism from |∆(k )| to a closed k-dimensional disk
for k ≥ 1; where must this homeomorphism send the subspace |∂∆(k )|?

E XERCISE 1.9. Let M be a finite subset of points in Euclidean space Rn (with its standard
metric). As a function of n, can you find the smallest δ so that VRe ( M ) is always a subcomplex
of Čδ ( M)? [Here the Čech complex has been constructed with respect to the ambient Euclidean
space Rn ]

E XERCISE 1.10. If σ and τ are a pair of simplices in a simplicial complex K satisfying σ ≤ τ,

show that st(σ) ⊃ st(τ ) and lk(σ) ⊃ lk(τ ).

E XERCISE 1.11. Show that if the link lk(σ) of a simplex σ in a simplicial complex K is non-
empty, then lk(σ) is a subcomplex of K.

E XERCISE 1.12. Let σ be a simplex in a simplicial complex K. Show that a simplex τ lies in
lkK (σ) if and only if the following condition holds: the open stars of σ and τ have a non-trivial
intersection and σ and τ have no common faces.
16
 2. HOMOTOPY 17

2. H OMOTOPY
 2. BASIC DEFINITIONS 18

2.1 B ASIC D EFINITIONS

Given topological spaces X and Y, the set of all continuous functions from X to Y is typically
quite large and complicated even in relatively simple cases (e.g., when both X and Y are the unit
circle in R2 ). In order to study such functions, we are compelled to define interesting equivalence
relations on them and restrict attention to equivalence classes. Among the deepest and most
fruitful equivalence relations between functions X → Y is the notion of a homotopy.
D EFINITION 2.1. Two continuous functions F, G : X → Y between topological spaces X and
Y are homotopic if there is a third continuous function
θ : X × [0, 1] → Y
(called a homotopy) so that for all x in X, we have θ ( x, 0) = F ( x ) and θ ( x, 1) = G ( x ).

The requirement that θ also be continuous is absolutely essential here, since it is always possi-
ble to find discontinuous θ satisfying the requirements of this definition. Thus, the fundamental
idea behind this definition is to put two functions in the same equivalence class whenever you
can continuously interpolate from one to the other as a parameter t ∈ [0, 1] slides from 0 to 1.
The picture below illustrates the homotopy equivalence of two maps F, G when X is a circle and
Y is R3 . These are homotopic if we can find a continuous θ from the cylinder X × [0, 1] to R3
whose restriction to the lower boundary X × {0} coincides with F and restriction to the upper
boundary X × {1} coincides with G.

Homotopies between functions can be used in order to produce an equivalence relation on

topological spaces as well.
D EFINITION 2.2. Two topological spaces X and Y are homotopy equivalent if there are con-
tinuous maps F : X → Y and G : Y → X so that
(1) the composite F ◦ G is homotopic to the identity map on Y, while
(2) the composite G ◦ F is homotopic to the identity map on X.

A pair of continuous maps F and G satisfying the two conditions above are often called homo-
topy inverses of each other, although it is important to note that in general there is no uniqueness
of such inverses — the set of homotopy inverses for a given F might contain several maps.
Homotopy equivalence is a topological property that tends to be largely agnostic to metric infor-
mation. The two panels below are designed to illustrate this phenomenon: in the first case, the
2-dimensional thickened figure-8 is homotopy equivalent to the thinner 1-dimensional figure-8
 2. CONTRACTIBLE SPACES 19

in its interior. But if we perturb this thinner curve ever so slightly to create a single loop, then
homotopy equivalence no longer holds.

Two simplicial complexes K and L are said to be homotopy equivalent, or have the same ho-
motopy type, whenever their geometric realizations |K | and | L| are homotopy equivalent in the
sense of the definition above. It may not be immediately obvious that homotopy is an important
equivalence relation between topological spaces — absorbing this fact takes time and experi-
ence. What should be clearer even at this early stage is that homotopy equivalence is far less
rigid than homeomorphism: homeomorphic spaces are always homotopy equivalent, but the
converse does not hold.

2.2 C ONTRACTIBLE S PACES

The quest to study topological spaces up to homotopy equivalence has a natural starting point
— we begin by asking which spaces are the least complicated from a homotopical perspective.
D EFINITION 2.3. A topological space X is contractible if it is homotopy equivalent to the
one-point space.

You should check that X is contractible if and only if there exists some point p ∈ X so that the
identity map on X is homotopic to the constant map sending every point of X to p. In particular,
the empty set ∅ is not contractible.
E XAMPLE 2.4. Here are several families of contractible simplicial complexes:
(1) Solids: for each k ≥ 0 the solid k-simplex ∆(k ) is contractible.
(2) Cones: the cone over any simplicial complex K (see Definition 1.19) is contractible.
(3) Trees: a tree is a connected graph with no cycles; these are all contractible.
We will prove the contractibility of these after developing some helpful machinery. For now,
it is important to start building a mental database which contains as many contractible spaces as
possible. The next few sections contain a suite of extremely powerful tools for detecting homo-
topy equivalence, and all of these tools rely in one way or another on your ability to recognize
contractible spaces. The underlying reason for this dependence is the following vital result.
L EMMA 2.5. Let X be a topological space and k ≥ 1 an integer. If X is contractible, then any
continuous map F : |∂∆(k )| → X from the hollow k-simplex to X can be extended to a continuous map
F + : |∆(k )| → X from the solid k-simplex.

P ROOF. Even the case k = 1 is quite insightful, so we will go over it carefully. Since |∆(1)|
is homeomorphic to the unit interval [0, 1] and |∂∆(1)| consists of the endoints {0, 1}, we must