SMSTC (2024/25)

Algebraic topology
LECTURE NOTES 0: Background on metric and
topological spaces
Gwyn Bellamya

www.smstc.ac.uk

0.1 Background
These first notes are concise background notes on metric and topological spaces. They were written by
Brendan Owens. Many proofs are left as exercises – for the most part these are simply a matter of
applying definitions. More detail can be found in standard undergraduate point-set topology books, such
as

ˆ J. R. Munkres, Topology: a first course, Prentice-Hall (1975).

Calculus gives us the concepts of open sets in Euclidean space and continuous functions, as well as
properties such as connectedness, compactness and others. In this lecture we will see how these concepts
can be generalised first to metric spaces and then to topological spaces.
Notation: In these lectures ⊆ means ‘is a subset of’, and the symbol ⊂ is reserved for ‘is a proper subset
of’.

0.1.1 Metric spaces

The standard n-dimensional Euclidean space consists of the set

Rn = {(x1 , . . . , xn ) | xi ∈ R}

together with the distance function d : Rn × Rn → R

p
d((x1 , . . . , xn ), (y1 , . . . , yn )) = (x1 − y1 )2 + · · · + (xn − yn )2 .

This satisfies the following properties:

(1) d(x, y) ≥ 0, and d(x, y) = 0 ⇐⇒ x = y

(2) d(x, y) = d(y, x)
(3) d(x, z) ≤ d(x, y) + d(y, z).

More generally, any set M together with a function d : M × M → R satisfying the above three properties
is called a metric space. The following “ϵ-δ = epsilon-delta” definition may be familiar from calculus or
analysis.

Definition 0.1 Let f : (M, dM ) → (N, dN ) be a function from one metric space to another. We say f
is continuous at x ∈ M if for any ϵ > 0 there exists δ > 0 such that

dM (x, y) < δ =⇒ dN (f (x), f (y)) < ϵ.

We say f is continuous if it is continuous at all points x ∈ M .

a [email protected]

0–1
SMST C: Algebraic topology 0–2

Example 0.2 The function f : R → R defined by

(
x if x ≤ 0,
f (x) =
x+1 if x > 0,

is continuous on R \ {0}, but not at 0. (Verify this using ϵ-δ ideas.)

The concept of an open set will enable us to give a friendlier looking definition of continuity. Roughly
speaking a subset U of a metric space is open if for any point x in U , all sufficiently nearby points are
also in U .

Definition 0.3 The open ball centred at x ∈ M with radius r > 0 is the subset

B(x, r) = {y ∈ M | d(x, y) < r} ⊆ M.

Example 0.4 (Manhattan metric) Let M = R2 with

d((x1 , y1 ), (x2 , y2 )) = |x1 − x2 | + |y1 − y2 |.

Sketch the unit ball B(0, 1) about the origin.

Example 0.5 (Chessboard metric) Let M = R2 with

d((x1 , y1 ), (x2 , y2 )) = max(|x1 − x2 |, |y1 − y2 |).

Sketch the unit ball B(0, 1) about the origin.

Example 0.6 (Railway metric) Let M = R2 with

p p
d((x1 , y1 ), (x2 , y2 )) = x12 + y12 + x22 + y22

provided (0, 0), (x1 , y1 ) and (x2 , y2 ) are not collinear; if they are collinear, set d((x1 , y1 ), (x2 , y2 )) to be
the usual Euclidean distance. Sketch the unit ball B(0, 1) about the origin.

Definition 0.7 Let (M, d) be a metric space.

A subset U ⊆ M is open if for any x ∈ U there exists r > 0 with B(x, r) ⊆ U .
A subset U is closed if its complement M \ U is open.

Exercise 0.8 Verify that a subset U of a metric space is open if and only if U is a union of open balls.

Example 0.9 The Euclidean, Manhattan and Chessboard metrics all give the same open sets in R2 . The
Railway metric has open sets which are not open using any of the previous three metrics. (Verify this.)

If (M, d) is a metric space and A ⊆ M is any subset, then A inherits the distance function d. Note that
a set V in A is open if and only if V = U ∩ A for some open set U in M .
Let f : M → N be a function between metric spaces. For a subset U ⊆ N , the preimage of U under f is

f −1 (U ) = {x ∈ M | f (x) ∈ U }.

Proposition 0.10 Let f : M → N be a function between metric spaces. Then f is continuous if and
only if the preimages of open sets are open. Equivalently, f is continuous if and only if the preimages of
closed sets are closed, since f −1 (N \ U ) = M \ f −1 (U ).

Proof Exercise. □
SMST C: Algebraic topology 0–3

0.1.2 Topological spaces

Proposition 0.11 If (M, d) is a metric space, then
(1) M and ∅ are open sets;
(2) arbitrary unions of open sets are open;
(3) finite intersections of open sets are open.

Proof Exercise. □
Note that infinite intersections of open sets may not be open. For example, in R with its usual metric,
∞  
\ 1 1
− ,
n=1
n n

is not open.

Definition 0.12 A topological space is a set X together with a collection of subsets called open sets
satisfying the three properties in Proposition 0.11:
(1) X and ∅ are open sets;
(2) arbitrary unions of open sets are open;
(3) finite intersections of open sets are open.
The complements of open sets are called closed sets.

Examples 0.13 Any set X can be given the discrete topology, in which every subset is open, or the
indiscrete topology, in which the only open sets are X and ∅.

Example 0.14 Given a subset A ⊆ X of a topological space X, the subspace topology on A is formed
by taking V ⊆ A to be open if and only if V = U ∩ A for some open set U in X.

Proposition 0.10 now suggests the definition of continuity for functions between topological spaces.

Definition 0.15 Let X and Y be two topological spaces. A function f : X → Y is called continuous (or
a continuous map, or later just a map) if the preimage of every open set if open, or equivalently if the
preimage of every closed set is closed.

If A ⊆ X is given the subspace topology, it is straightforward to check that a continuous function

f : X → Y gives rise to a continuous restriction f |A : A → Y .
It is often convenient to specify a topology by giving a base; this is a collection of open sets B such that
every open set is a union of elements of B. For example open balls are a base for the topology in any
metric space, as are open balls with rational radii.

Definition 0.16 Two topological spaces X and Y are homeomorphic (written X ∼ = Y ) if there exists a
function f : X → Y which is a continuous bijection with a continuous inverse. Such an f is called a
homeomorphism.

This defines a notion of equivalence between topological spaces which satisfies the three properties of an
equivalence relation. Also note that for a fixed space X, the set Homeo(X) of homeomorphisms from X
to itself is a group under composition.

Examples 0.17 Give all subsets of Rn the subspace topologies.

(1) The open interval (0, 1) is homeomorphic to (0, 2) via x 7→ 2x, and similarly any open interval (a, b)
is homeomorphic to (0, 1).
(2) The open interval (−π/2, π/2) is homeomorphic to R via x 7→ tan x, and hence any open interval
(a, b) ∼
= R.
(3) The sets {0, 1, 2, 3} and {0, 1} (with any choice of topology) are not homeomorphic since there is no
bijection between these sets.
(4) The sets
X = S 1 = {(x, y) ∈ R2 | x2 + y 2 = 1}
and Y = (0, 1) are not homeomorphic: details will be given later.
SMST C: Algebraic topology 0–4

Example 0.18 Let X be any set. The cofinite topology is the topology in which the empty set and
complements of finite sets are open. If X = R this is also called the Zariski topology.

Example 0.19 The Zariski topology on Rn (or on kn for any field k): closed sets are intersections of zero
sets of polynomials in n variables. Equivalently: a base for the Zariski topology is given by complements
of zero sets of polynomials.

Examples 0.20 Other important examples of topological spaces: S n (the unit sphere in Rn+1 ), orientable
surfaces Σg , nonorientable surfaces Ng , matrix groups,. . .

0.1.3 Compactness
Let XS be a topological space. An open cover of X is a collection {Uα }α∈A of open subsets of X with
X = α∈A Uα .
For example take X = (0, 2] and Un = (1/n, 2], then {Un }n∈N is an open cover of X. If we let Y = [0, 2]
and U0 = [0, 1/2) then {Un }n≥0 is an open cover of Y . Note that {Un }n≥100 also covers X; we say it is
a subcover of {Un }n∈N . Any such subcover is infinite whereas the cover {Un }n≥0 of Y admits the finite
subcover {U0 , U3 }.

Definition 0.21 A topological space is compact if every open cover admits a finite subcover.

We recall the following fundamental theorems about compactness in Euclidean space. Note a subset of
Rn is bounded if it is contained in B(0, R) for some R > 0.

Theorem 0.22 (Heine-Borel) Closed intervals [a, b] are compact subsets of R.

Theorem 0.23 (Heine-Borel) A subset U of Rn is compact if and only if it is closed and bounded.

Proposition 0.24 The image of a compact space under a continuous map is compact. Thus compactness
is a topological property (preserved under homeomorphism).

Proof Exercise. □
∼ 1
Thus if X = Y , then X is compact if and only if Y is compact. This shows that S is not homeomorphic
to (0, 1).
Another useful notion of compactness is defined using sequences. A sequence x1 , x2 , . . . in a topological
space X converges to x ∈ X if for any open set U containing x there exists N ∈ N with xn ∈ U for all
n ≥ N . A space X is sequentially compact if every sequence in X has a convergent subsequence. For
subsets of metric spaces, sequentially compact is equivalent to compact.

0.1.4 Connectedness, path-connectedness and π0

Definition 0.25 A space X is connected if it cannot be written X = U1 ∪ U2 with U1 , U2 nonempty,
open and disjoint.

Example 0.26 The real line R is connected. This is surprisingly subtle, but here is a sketch. Suppose
that R = U1 ∪ U2 with U1 , U2 open and disjoint and 0 ∈ U1 . Let R = sup{r > 0 | B(0, r) ⊆ U1 }. If R is
finite then B(0, R) is contained in U1 but one of ±R is in U2 , which contradicts openness of U2 . If R is
infinite then U2 is empty.

Proposition 0.27 The continuous image of a connected set is connected.

Proof Exercise. □

Example 0.28 The matrix group GL(n, R) of n × n real matrices with nonzero determinants (with
2
subspace topology from Rn ) is disconnected, as the determinant gives a continuous surjection onto R∗ =
R \ {0}, which is disconnected.
SMST C: Algebraic topology 0–5

It follows from Example 0.26 and Proposition 0.27 that the open interval (0, 1) is connected, since it is
homeomorphic to R. It is then an easy exercise to show that the closed interval I = [0, 1] is connected.
A (continuous) path in a space X is a continuous map f : I → X. Two points x, y ∈ X are said to be
joined by a path if there is a path f : I → X with f (0) = x and f (1) = y.

f (t)
X x = f (0) • • f (1) = y

Definition 0.29 A space X is path-connected if any two points in X may be joined by a path.

Proposition 0.30 The continuous image of a path-connected set is path-connected.

Proof Exercise. □

Proposition 0.31 For a topological space X,

X path-connected =⇒ X connected.

Proof Assume X is path-connected and X = U1 ∪ U2 with U1 , U2 nonempty, open and disjoint. Choose
x ∈ U0 , y ∈ U1 and a path f : I → X with f (0) = x, f (1) = y. Connectedness of f (I) is contradicted by

f (I) = (f (I) ∩ U0 ) ∪ (f (I) ∩ U1 ).

□
n n 2
Using Proposition 0.31 we can easily show that many examples such as R , S , T etc. are connected,
since it is fairly easy to see that they are path-connected. For example two points x, y ∈ Rn are joined
by the path f (s) = x + s(y − x).

Example 0.32 The circle S 1 is not homeomorphic to the closed interval [0, 1].

Proof Suppose f : [0, 1] → S 1 is a homeomorphism. Let y ∈ (0, 1). Then [0, 1] \ {y} is not connected.
The homeomorphism f restricts to give a homeomorphism on [0, 1] \ {y} whose image is S 1 \ {f (y)} which
is path-connected, giving a contradiction. □
Note that connectedness does not always imply path-connectedness as the following example shows.

Example 0.33 The subspace

X = {(0, y) | y ∈ R} ∪ {(x, sin(1/x)) | x > 0} ⊆ R2

is connected but not path-connected.

Given two paths f, g : I → X with f (1) = g(0), their concatenation is the path f ∗ g : I → X given by
(
f (2s) if 0 ≤ s ≤ 1/2,
f ∗ g(s) =
g(2s − 1) if 1/2 ≤ s ≤ 1.

(Continuity of f ∗ g follows from the Gluing Lemma which is proved in the exercises.)
We can now define a relation on the points of a space X by x ∼ y if x and y are joined by a path.

Claim 1 This is an equivalence relation.

Proof Exercise. □
Equivalence classes are called path components. The set of path components of X is denoted π0 (X). We
will see that π0 is a functor from the category TOP of topological spaces and continuous maps to the
category SET of sets with functions. What this means is the following: each continuous map f : X → Y
induces a function π0 (f ) : π0 (X) → π0 (Y ) given by [x] 7→ [f (x)] with the following properties:
SMST C: Algebraic topology 0–6

(1) π0 (IdX ) = Idπ0 (X) ;

(2) if f : X → Y and g : Y → Z are continuous then
π0 (g ◦ f ) = π0 (g) ◦ π0 (f ).
This functoriality is a very useful property. In particular it shows that the cardinality of π0 is a topological
invariant.
Proposition 0.34 If f : X → Y is a homeomorphism then π0 (f ) : π0 (X) → π0 (Y ) is a bijection.
Proof Exercise. □

Example 0.35 Let X = R, Y = R \ {0}, Z = {0, 1} ⊆ R. Then π0 (X) = {[0]}, π0 (Y ) = {[−1], [1]},
π0 (Z) = {[0], [1]}, This shows that X is not homeomorphic to either Y or Z. The spaces Y and Z have
bijective π0 so we need to use other means to show that they are not homeomorphic (for example there is
no bijection between them).
The converse of Proposition 0.31 is not true in general, but the following result is often useful.
Definition 0.36 A topological space X is locally path-connected if for every x ∈ X and for every open
set V containing x there is a path-connected open set U ⊆ X with x ∈ U ⊆ V .
Proposition 0.37 For a topological space X,
X connected and locally path-connected =⇒ X path-connected.
Proof The basic idea is to show that each path component is open, hence there can be only one by
connectedness. □
For example, manifolds are locally path-connected, hence a connected manifold is also path-connected.

0.1.5 Hausdorff spaces

For topological spaces which are not metric spaces, there are various separability conditions that have
important consequences. The Hausdorff condition is one of the most commonly encountered.
Definition 0.38 A topological space X is Hausdorff if for every pair of distinct elements u, v ∈ X, there
are disjoint open subsets U ⊆ X and V ⊆ X with u ∈ U and v ∈ V .
It is clear that every metric space is Hausdorff. To see an example of a non-Hausdorff space, consider the
set with 2 elements X = {a, b} and open sets
∅, {a}, X.
Then X is not Hausdorff with this topology. Another example is provided by taking any infinite set Y
with the cofinite topology: open sets are subsets with finite complements together with ∅. The Zariski
topology on kn for a field k is also non-Hausdorff.
Here is a general fact about Hausdorff spaces.
Proposition 0.39 Let X be a Hausdorff space. Then for each u ∈ X, the subset {u} ⊆ X is closed.
Proof For each v ∈ X with v ̸= u, choose disjoint open sets Uu,v , Vu,v with u ∈ Uu,v and v ∈ Vu,v .
Then [
X \ {u} = Vu,v ,
v∈X\{u}

so X \ {u} is open, hence {u} is closed. □

Proposition 0.40 Let X be a Hausdorff space. Then every compact subset of X is closed.
Proof Exercise. □
Finally, here is an important result, heavily used in algebraic topology.
Proposition 0.41 Let X be a compact space and let Y be a Hausdorff space. Then every continuous
bijection f : X → Y is a homeomorphism.
Proof Exercise: the proof involves showing the inverse f −1 is continuous. □
There are other separation properties which are commonly encountered such as normality.
SMST C: Algebraic topology 0–7

0.1.6 One-point compactifications

Compactness is a good property for a topological space to possess, but many spaces are not compact.
There are various ways to enlarge a space to obtain a compact space in which it sits as a “dense” subspace.
The (Alexandroff ) one-point compactification is particularly important.
Let X be a topological space. Let X ∗ = X ∪ {∞} where ∞ is supposed to be a new point lying outside
of X. Define the open sets of a topology on X ∗ by taking all open subsets U ⊆ X and all sets W ⊆ X ∗
containing ∞ for which X \ (W ∩ X) ⊆ X is closed and compact.

Proposition 0.42 X ∗ with the above notion of open sets is a compact topological space. X ∗ is Hausdorff
if and only if X is Hausdorff and locally compact (meaning every point has a compact neighbourhood).

Proof Exercise. □
Here is an important family of examples.

Example 0.43 Consider X = Rn with its usual metric topology. Then the one-point compactification
X ∗ = (Rn )∗ is homeomorphic to the unit n-sphere S n ⊆ Rn+1 .

One explicit homeomorphism can be produced using stereographic projection of S n with a point deleted
to Rn . For R it is intuitively clear that the circle S 1 is obtained by adjoining a single “point at infinity”,
but this construction requires care to make the topology precise.