Essential topology chapter 3

Goran Huygh
January 2024

1 Topological spaces
Whilst algebra studies the way elements of a space interact, topology tries to
study the fundamental structures and properties of the space itself and the ele-
ments. As such, one must craft a powerful and handy definition of a topological
space. Classically, the following definition is given.
Definition 1. (Topological space) A topological space (T, τ ) is an ordered pair
consisting of a set T and a collection of subsets τ ⊆ P(T ) such that:

∅, T ∈ τ (1)
[
Given L ⊆ τ,we have L∈τ (2)
n
\
Given n subsets li ,we have li ∈ τ (3)
i=1

We then say τ forms a topology over T . In short this boils down to a space with
a collection of subsets that are closed under arbitrary union and finite intersec-
tion and contains the space itself with the empty set.

We say a set is open if it’s an element of τ , on the other hand we call a set
S closed if and only if T \S is open. Notice that this is not a dual notion, sets
can be open and closed or neither closed nor open.

1.0.1 Some speculative motivation

I am by no means an expert in topology, but I will try to give some natural
motivation to why this the notion of space we want and why these structures
are interesting to study.

From what I know, it turns out that a lot of properties concerning the space
and it’s fundamental structure boils down to how neighboring elements behave
and how ”close” they are. In other words, developing a notion of closeness
and neighboring points allows us to speak of concepts such a Cauchy sequences
without actually defining a metric! In some sense it is exactly that that τ is

1
capturing. For the sake of intuition, I like to imagine that every element of τ
has essentially some mysterious and unknown ”closeness factor” c associated to
it. Like that, depending on the status of membership in the topology of the two
elements t1 and t2 we could say that they are c1 close but not c2 close. (If we
take a set S with t1 , t2 ∈ S we can take a sufficiently small subset such that
t1 ∈ S ′ but t2 ∈
/ S′)

If this explanation was a bit vague then just think about the way Tao discussed
ϵ-closeness in Analysis I

1.1 Basic and classical examples

We briefly cover some elementary examples

1.1.1 Discrete and indiscrete topology

One can easily verify that the sets {∅, T } and P(T ) form topologies over any set
T . We call P(T ) the discrete topology over T whilst we call {∅, T } the indiscrete
topology over T .

1.1.2 Metric spaces

The classic topology over metric spaces is the one of ε-balls. Later we will see
that the balls form a basis for the topology.

Usually this topology is introduced by first discussing a notion if interior points.

An interior points is informally a point in the set where there exists a ε such
that the ε-ball around the point is a subset of the set itself. If then the set of
interior points of a set is equal to itself then it’s said to be open.

Bε (x, y)
. (x, y)

2
1.2 Continuity in topological spaces
The next step would be to define continuous function so we can talk about con-
tinuous deformations and invariants. But as it turns out, talking about close
points is pretty hard when not equipped with a metric. How do we then capture
the same intuition of continuity in topological spaces?

If we consider the intuitive way of thinking of a topology and how it defines

a ”law of closeness”, like discussed earlier, then one can come up with a pretty
charming definition.
Definition 2. (Continuity in topological spaces) Given the topological spaces
(S, s) and (T, τ ), we say a function f : S → T is continuous if and only if

∀τi ∈ τ, f −1 (τi ) ∈ s

In other words, the preimage of an open set in T is an open set in S.

1.2.1 Some more speculative motivation

Finding a satisfying motivation for me was a really hard task. But I did end up
with a satisfying answer. Part of the motivation of classic continuity is being
able to estimate values of functions by looking at function values that are near
it. There is this very nice definition regarding sequences, which is if we have
sequences (xn )∞
n=m such that limn→∞ xn = a then the sequence (f (xn ))n=m
∞

must converge to f (a).

Now we want to transcribe these notions to a topological space, which can

be tricky since all of the previously discussed definitions are very metric de-
pendant. But once again, if you consider that topologies define closeness then
you would want values that are close by some factor in T , to be close by some
factor in S. But how do you then look even closer near a point ? Well just take
a subset, since Q ⊂ R =⇒ f −1 (Q) ⊂ f −1 (R) we have that taking a subset
(which intuitively has a ”closer” closeness factor) gives us a preimage of close
points which are even closer then the previous preimage.

Thus, we can get arbitrary close by taking smaller and smaller subsets without
requiring a metric. As such, we transcribed continuity from metric spaces to
topological spaces successfully.

1.2.2 More on continuity

We quickly develop some important notion concerning this continuity.

First of all, and most importantly, this notion of continuity is consistent with
the one of continuity in metric spaces.

3
Theorem 1. (Consistency of the topological definition of continuity) A function
f : X → Y , given two metric spaces X and Y , is continuous in the metric sense
if and only if it is continuous in the topological sense.
Proof. We will omit the proof, as we haven’t properly defined the metric space
topology in terms of it’s basis. Once that is done it becomes easier to proof the
result.
Besides this, one can easily also verify the following theorem using basic defini-
tions and notions of preimages.
Theorem 2. Given a function f : S → T where S and T are topological spaces,
if S is endowed with the discrete topology or T with the indiscrete one then f is
always continuous.
We also have the following theorem that can be used to proof some pretty tricky
theorems in a very easy manner.
Theorem 3. If f : X → Y and g : Y → Z where both f and g are continuous
then g ◦ f : X → Z is also continuous
The proof is once again really really straightforward. It’s a pretty neat theo-
rem since it allows us to show that if given any field (F, ⊕, ⊗) with a defined
topology on it and with a continuous addition and multiplication map then all
polynomials are always continuous.
n
X
P (x) = F
ai xi
i=0

4
2 Examples of topological spaces
Before we introduce some examples we want to develop one very common topol-
ogy used to define many others, and that is the subspace topology.

2.1 Subspace topology

This is a very common topology and a powerful one as well, since we ensure that
the topological properties of the subset are consistent with those of the larger
space (Later we see this in action with continuity).
Definition 3. (Subspace topology) Given a topological space (T, τ ) and a subset
S of T , we define the topological space (S, ΘT ) where ΘT is the topology such
that
θ ∈ ΘT ⇐⇒ θ = S ∩ τi
In other words a set is open in S only if it was essentially already open in T but
we just limited the elements by an intersection. Personally I like to denote this
topology with capital theta.

We now briefly go over some other examples of topological spaces.

2.2 n dimensional sphere S n

With the definition of the subspace topology we immediately define a very pop-
ular topological space. S n is the topological space of the n dimensional sphere.
Defined by the set

S n := {(xi )1≤i≤n+1 ∈ Rn+1 : d((xi )1≤i≤n+1 , (0)1≤i≤n+1 ) = 1}

with topology ΘRn+1 .

2.3 General linear group GL(n, F)

Another interesting topological space is GL(n, F) which is also very popular in
algebra. The topological space is defined by the set

GL(n, F) := {M ∈ Mn×n (F) : M is invertible}

with the topology ΘRn2 . The book also includes more examples with the or-
thogonal and special orthogonal group if your are interested.

2.4 Real projective space RP n

Another interesting topological space which isn’t defined by the subspace topol-
ogy is the one of the real projective space. This topological space is defined by
the set

RP n := {l ⊂ Rn+1 : l is a straight line through the origin}

5
where we consider a set S ⊂ RP n to open if and only if l∈S l\{0} is open
S
in Rn+1 \{0}. The reason we omit zero is because one can quickly check that
not doing so endows the space with the indiscrete topology, which is not very
interesting.

3 Continuity in (T ′ , ΘT )
Like briefly mentioned before, the subspace topology can help us make state-
ments about the continuity of maps. It turns out that if a map f is continuous
then so is a restriction of it.
Theorem 4. (Continuity in the subspace topology) Given a continuous function
f : S → T where S and T are topological spaces, every restriction f |Q : Q → R
where (Q, ΘS ) and (R, ΘT ) is continuous.
f
TX TY

∪ ∪
f |QX
QX QY

Proof. Let P ∈ ΘT , one can easily show that f −1 |Q (P ) = Q ∩ f −1 (P ) and since

P is open and we are in the subspace topology we have that there exists a open
set U ⊂ T such that P = R ∩ U . This gives us
f −1 |Q (P ) = Q ∩ f −1 (R ∩ U ) = Q ∩ f −1 (R) ∩ f −1 (U )
And since Q ⊂ f −1 (R) (Otherwise we would have a not well defined function,
f |Q being well defined was an obvious implicit assumption) we have
f −1 |Q (P ) = Q ∩ f −1 (U )
And since f is continuous we know that f −1 (U ) is open. But since we are in
the subspace topology Q ∩ f −1 (U ) is a open set and we have that the restricted
preimage of a open set in the restricted range is open.
A nice and small example is the continuity of the determinant map in GL(3, R).
The determinant map is continuous in GL(3, R) since
f
9 R
R

∪ ∪
det = f |GL(3,R)
GL(3, R) R

Where f (a, b, c, d, e, f, g, h, i) = aei − af h − bdi + bf g + cdh − ceg.

6
4 Bases of topological spaces
Just like in vector spaces we want to be able to talk about ”generated” topologies
and understand more about the components that define a topology.
Definition 4. (Base of topological space) Given a topological space (T, τ ) and
B ⊆ τ , we say B forms a basis for T if and only if
[
∀τi ∈ τ ∃B ′ ⊆ B : B ′ = τi

Just like vectors can be written as a linear combination of the basis vectors we
have that every open set can be written as a union of basic open sets.

A prominent example is the topology of metric spaces, where one could ver-
ify that it is equivalent to state that the topology is the one generated by all
possible ε-balls around every point.

By generating we mean that given a set of basis elements we can consider the
topology τ (B) which we define as
[
τ (B) := {U : U = B, for some F ⊆ B}
B∈F
S
Associating this to the set B gives us a topological space generated by B
[
( B, τ (B))

One can verify this is true since clearly ∅ and the entire set are in τ (B) and it
is closed under union and finite intersection (the latter not being hard to show
but just a bit tedious notation wise lol).

This being said we can also show that given f : S → T with both being topo-
logical spaces and T having basis B, if for all B ∈ B, f −1 (B) is open in S, then
f is continuous. The proof simply relies on the fact that
 
[ [
f −1  B = f −1 (B)
B∈F ⊆B B∈F ⊆B