Toplogy Notes

Topology: A Firt Course∗
M.Thamban Nair†
Visiting Professor
Department of Mathematics
BITS Pilani, KK Birla Goa Campus, Goa 403726
∗
Notes for the course MATH F311: Introduction to Topology, August-December 2023 at BITS
Pilani, Goa Campus.
†
Formerly Professor & Head of the Department of Mathematics at IIT Madras, Chennai.
.
Contents
1 Topology: Motivation, Definition and Examples 6
2 Basic notions properties 12

2.1 Hausdorff space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 More on weak and strong topologies . . . . . . . . . . . . . . . . . . . . 15
2.4 Limit points, closure, interior and boundary . . . . . . . . . . . . . . . 17
2.5 Subspace topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 Base and sub-base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.7 A topological proof for infinitude of primes . . . . . . . . . . . . . . . . 28
3 Countability axioms and separability 29
4 Functions and their properties 36

4.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2 Continuity and sequential continuity . . . . . . . . . . . . . . . . . . . 37
4.3 Continuity and separability . . . . . . . . . . . . . . . . . . . . . . . . 41
4.4 Restriction and expansion and composition theorems . . . . . . . . . . 43
4.5 Homeomorphism, open and closed maps . . . . . . . . . . . . . . . . . 48
5 Connected spaces and path connected spaces 55

5.1 Connectedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2 Totally disconnected space . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.3 Path connectedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
6 Compact spaces 66
6.1 Definition and implications . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.2 Comapct subsets of R . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.3 Locally compact spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.4 One-point compactification . . . . . . . . . . . . . . . . . . . . . . . . . 77
7 Product topology 78
7.1 On product of finite number of topological spaces . . . . . . . . . . . . 78
7.2 On product of infinite number of topological spaces . . . . . . . . . . . 81
.
5
1 Topology: Motivation, Definition and Examples
Recall from the course on Real Analysis that if Ω is a metric space with metric d, then
the collection G of all open sets in Ω has the following properties:
1. ∅ ∈ G and Ω ∈ G;
2. Union of any arbitrary collection of members of G is a member of G;
3. Intersection of any two members of G is a member of G.
We have seen that many notions introduced in Real Analysis can be expressed using
the notion of open sets and the above three properties of G, without explicit reference
to the metric. For example, look at the following:
• A set A ⊆ Ω is closed iff Ac := Ω \ A ∈ G.
• A sequence (xn ) in Ω converges to x ∈ Ω iff for every A ∈ G with x ∈ A, there

exists n0 ∈ N such that
xn ∈ A ∀ n ≥ n 0 .
• If S ⊆ Ω and x ∈ Ω, then x is a limit point (accumulation point) of S iff for

every A ∈ G with x ∈ A, A contains atleast one point from S other than x.
• If Ω̃ is another metric space (possibly different from Ω), then a function f : Ω → Ω̃

is continuous iff for every open set A in Ω̃, f −1 (A) is open in Ω.
• A subset A of Ω is compact iff every open cover of A contains a finite sub-

collection which also covers A.
• A subset A of Ω is connected iff A cannot be written as a union of two nonempty

disjoint open sets.
Motivated by the above, we introduce the notion of a collection of open sets in an

arbitrary set X, which is known as a topology1 on X.
1
The creater of the subject Topology can be considered as Riemann in the middle of nineteenth
century, while he was working on theory of functions and their geometric properties. The term
topology was introduced by J.B.Listing (a student of Gauss) in 1847, but the term became common
only after Fréchet used it in his thesis. The subject was called Analysis situs earlier.
6
Definition 1.1. (Topology) Let X be a set and T be a collection of subsets of X.
Then T is called a topology on X if it has the following properties:
1. The empty set ∅ and the whole set X are members of T ;
2. Intersection of any two members of T is a member of T .
3. Union of any collection of members of T is a member of T ;
A set X together with a topology T on it is called a topological space, and in that

case, the pair (X, T ) is also called a topological space. ♢
Definition 1.2. Let (X, T ) be a topological space. Then, members of T are called
open sets in X. A subset of X is said to be a closed set if its compliment in X is
open in X. ♢
Symbolically the conditions in the definition of a topological space can be written

as follows: Let X be a nonempty set and T ⊆ 2X . Then T is a topology iff
1. ∅ ∈ T , X ∈T;
2. A, B ∈ R ⇒ A ∩ B ∈ T ;
[
3. Λ ̸= ∅ and Aα ∈ T ∀ α ∈ Λ ⇒ Aα ∈ T .
α∈Λ
TnT ) be a topological space. Show that, if Ai ∈ T for i = 1, . . . n

Exercise 1.3. Let (X,
for some n ∈ N, then i=1 Ai ∈ T . Deduce that finite union of closed sets is closed. ♢
Example 1.4. Let X be any set.
(i) T := 2X , the collection of all subsets of X, is a topology on X, called the

discrete topology on X. With respect to this topology, every subset of X is open
and also closed.
(ii) T := {∅, X} is a topology on X, called the indiscrete topology on X. With

respect to this topology, a subset of X which is different from ∅ and X is neither open
nor closed. If X is a nonempty set, then, In this topological space, a subset of X which
is different from ∅ and X is neither open nor closed. Note that, if X itself is the empty
set, then T := {∅}. ♢
Note that, if T is any topology on a set X, and if T0 and T1 are the indiscrete
topology and discrete topology, respectively, on X, then
T0 ⊆ T ⊆ T 1 .
7
Definition 1.5. Let T1 and T2 be topologies on a set X. Then, T1 is said to be weaker
than T2 if T1 ⊆ T2 , and in that case, T2 is said to be stronger than T1 .
The topology T1 is said to be strictly weaker than T2 if T1 is a proper subcollection
of T2 , and in that case T2 is said to be strictly stronger than T1 . ♢
• On a set X, indiscrete topology is the weakest topology and discrete topology is

the strongest topology; both coincide if X is a single ton set.
We have already observed in the beginning of this section that the collection of all
open subsets of a metric space satisfies the requirements for a topology. In other words
the following example provides a class of examples.
Recall that, if Ω is a metric space with metric d, then a subset A of Ω is said to
be an open set if for every x ∈ A, there exists r > 0 such that Bd (x, r) ⊆ A, where
Bd (x, r) is the open ball centered at x and radius r, that is,
Bd (x, r) = {y ∈ Ω : d(x, y) < r}.
Thus, we have the following class of examples for a topology.
Example 1.6. Let Ω be a metric space with metric and Td be the collection of all
open subsets of Ω. Then Then Td is a topology on Ω. ♢
Definition 1.7. (Metric topology) Let Ω be a metric space with metric d. Then
the topology Td defined as in Example 1.6 is called the topology induced by the
metric d, and it is denoted by Td . ♢
Definition 1.8. (Metrizable space) A topological space (X, T ) is called metrizable
if there exists a metric d on X such that T = Td . ♢
Exercise 1.9. Let (Ω, d) be a metric space. Show that a subset A of Ω is open iff A
is a union of open balls in Ω. ♢
We may recall that, on Rk , the usual metric or the euclidean metric d is defined by
p
d(x, y) := (x1 − y1 )2 + · · · (xk − yk )2 ,
for x = (x1 , . . . , xk ) and y = (y1 , . . . , yk ) in Rk . Similarly, the usual metric on C, the
set of all complex numbers is defined by
d(z, w) = |z − w|, z, w ∈ C,
and on Ck , the usual metric d is defined by
p
d(z, w) := |z1 − w1 |2 + · · · |zk − wk |2 ,
for z = (z1 , . . . , zk ) and w = (w1 , . . . , wk ) in Ck .
8
Definition 1.10. On Rk (respectively, Ck ), the topology induced by the usual metric
is called the usual topology on Rk (respectively, Ck ). ♢
Theorem 1.11. The topologies induced by any two equivalent metrics are the same.
Proof. Suppose d and ρ are equivalent metrics on X. Then there exists c1 , c2 > 0 such
that
c1 d(x, y) ≤ ρ(x, y) ≤ c2 d(x, y) ∀ x, y ∈ X.
We have to show that Td = Tρ . So, let A ∈ Td . Let x ∈ A. Then there exists r > 0
such that Bd (x, r) ⊆ A; that is,
y ∈ X, d(x, y) < r ⇒ y ∈ A.
Note that
ρ(x, y) < c1 r ⇒ d(x, y) < r.
Thus, we obtain Bρ (x, c1 r) ⊆ Bd (x, r). In particular, Bρ (x, c1 r) ⊆ A. Thus, for each
x ∈ A, there exists ε := c1 r > 0 such that Bρ (x, ε) ⊆ A. This shows that A ∈ Tρ , thus,
proving Td ⊆ Tρ . Analogously, it can be shown that Tρ ⊆ Td .
Exercise 1.12. Consider the metrics d1 , d2 , d∞ on R2 defined by
d1 (x, y) = |x1 − y1 | + |x2 − y2 |,

d2 (x, y) = (|x1 − y1 |2 + |x2 − y2 |2 )1/2 ,
d∞ (x, y) = max{|x1 − y1 |, |x2 − y2 |},
respectivlely, for x, y ∈ R2 . Show that the topologies induced by d1 , d2 , d∞ are the

same. ♢
Let us first observe a simple fact:

Theorem 1.13. Let X be a topological space and A ⊆ X. Then A is open in X iff for
every x ∈ A, there exists an open set G such that x ∈ G ⊆ A.
Proof. Suppose A is open in X. Then, for every x ∈ A, we may take G = A.
S x ∈ A, there exists an open set Gx such that

Conversely, suppose that for every
x ∈ Gx ⊆ A. Then we see that A = x∈A Gx . Thus, being a union of open sets, A is
open.
Definition 1.14. (Neigbourhood) Given a topological space X and a point x ∈ X,
any open set containing x ∈ X called a neigbourhood of x or an open neigbour-
hood of x. . ♢
Remark 1.15. Let X be a set and S be a family of subsets of X, i.e, S ⊆ 2X .
9
• S is said to be closed under
Sn finite unions if for any finite number of members of
S, say A1 , . . . , An ∈ S, i=1 Ai ∈ S.
• S is said to be closed underTfinite intersections if for any finite number of members

of S, say A1 , . . . , An ∈ S, ni=1 Ai ∈ S.
• S is said to be closed
S under arbitrary unions if for any subcollection of S, say
A ⊆ S, we have A∈A A ∈ S.
• S is said to be closedTunder arbitrary intersections if for any subcollection of S,

say A ⊆ S, we have A∈A A ∈ S.
Thus, we see that, if (X, T ) is a topological space and if C is the collection of all closed
sets in X, then
• T is closed under arbitrary unions and finite intersections;
An immediate consequence of the definition of a topology is the following:
Theorem 1.16. Let (X, T ) be a topological space and C be the collection of all closed
sets in X. Then
(i) ∅ and X are members of C;
(ii) Union of any finite number of members of C is a member of C.
(iii) Intersection of any arbitrary collection of members of C is a member of C;
• Let C be as in the above theorem.Then C is closed under arbitrary intersections

and finite unions. ♢
Notation: Given a set X, if A is a collection of subsets of X, then we we may denote

such a collection also by {A ⊆ X : A ∈ A}. Thus, it is of the form {Aα ⊆ X : α ∈ Λ},
where Λ = A, by identifying Aα with A for α = A. Thus, an arbitrary collection of
subsets of X can always be written as
{Aα ⊆ X : α ∈ Λ}
for some set Λ, which is called an indexSset. Thus,Sif A =S

{Aα ⊆ X : α ∈ Λ}, then the
union of all members of A, written as A∈A A or A is α∈Λ Aα , that is,
[ [
A= Aα .
A∈A α∈Λ
10
T T T
Similarly, the intersection of all members of A, written as A∈A A or A is α∈Λ Aα ,
that is, \ \
A= Aα .
A∈A α∈Λ
Note that if A is a finite set, then it can be written as {A1 , . . . , An } or {Ai : i ∈

{1, . . . , n}} for some n ∈ N, and if A is a countably infinite set, then it can be written
as {A1 , A2 , . . .} or {An : n ∈ N}.
Let us give a few more standard examples which we may use in the due course.
Example 1.17. Let X be any set.
(i) The discrete topology on X is induced by the discrete metric on X.
(ii) Let T consist of ∅ and all those subsets of X such that Ac is finite, together
with ∅ and X. Then T is a topology on X, called the co-finite topology on X.
(iii) Let T consist of ∅ and all those subsets A of X such that Ac is countable.
Then T is a topology on X, called the co-countable topology on X. ♢
• On a finite set, co-finite topology and co-countable topology are the same. In
general, co-finite topology is weaker than co-countable topology; strictly weaker
if the underlying set is infinite.
Example 1.18. Let T be the collection of all subsets of R such that A ∈ T iff for
every x ∈ A, there exists r > 0 such that [x, x + r) ⊆ A. Then T is a topology on R,
called the lower limit topology on R. ♢
Notation: In the due course, we shall use R with lower limit topology by Rℓ .
Example 1.19. Let T be the collection of all subsets of R such that A ∈ T iff for
every x ∈ A, there exists r > 0 such that (x − r, x] ⊆ A. Then T is a topology on R,
called the upper limit topology on R. ♢
Remark 1.20. We have already noted that a set which is not open need not be closed;
a set which is not closed need not be open; a set can be both open and closed.
For instance, let R be with the topology induced by the usual metric. Then [0, 1)
is neither open nor closed, and the sets ∅ and R are both open and closed. We may
also note that there is no other set which is both open and closed. ♢
Exercise 1.21. Show that the indiscrete topology on a nonempty set X is induced by
a metric iff X is singleton, and in that case this topology is the discrete topology. ♢
Exercise 1.22. Show that the co-finite topology on a nonempty set X is induced by
a metric iff X is a finite set, and in that case this topology is the discrete topology. ♢
11
Exercise 1.23. Show that the co-countable topology on a nonempty set X is induced
by a metric iff X is a countbale set, and in that case this topology is the discrete
topology. ♢
2 Basic notions properties
2.1 Hausdorff space
One of the important property of a metric space Ω is its Hausdorff property, that is,
for every pair of distinct points x, y in Ω, there exists disjoint open sets containing x
and y respectively.
Definition 2.1. (Hausdorff space) A topological space X is said to be a Hausdorff2

space if for every pair of distinct points x, y in X, there exist disjoint open sets A and
B containing x and y, respectively. ♢
• A topological space (X, T ) is Hausdorff iff for every x, y ∈ X, if x ̸= y, then there

exist A, B ∈ T such that
A ∩ B = ∅, x ∈ A, y ∈ B.
We may say that: A topological space X is Hausdorff space iff every pair of distinct
points in X can be separated by disjoint open sets.
• If a topology on X is induced by a metric, then X is a Hausdorff space.
Example 2.2. Let X be a set. The following can be verified easily (Verify!):
1. The indiscrete topology on X is Hausdorrf iff #(X) ≤ 1, and in that case the
topology is the discrete topology.
2. The co-finite topology on X is Hausdorrf iff X is a finite set, and in that case
the topology is the discrete topology.
3. The co-countable topology on X is Hausdorff iff X is a countable set, and in that

case the topology is the discrete topology. ♢
From the statements in the above example, the following can be deduced (Exercise).
2
Felix Hausdorff (1868-1942) was a mathematician from Germany.
12
• If X contains more than one element, then the indiscrete topology on X is not
induced by any metric.
• If X is not a finite set, then the co-finite topology on X is not induced by any
metric.
• If X is an uncountable set, then the co-countable topology on X is not induced
by any metric.
Example 2.3. The lower limit topology and the upper limit topology on R are Haus-
dorff:
First consider Rℓ , the set R with lower limit topology. Let x, y ∈ Rℓ with x ̸= y.
Taking r = |x − y|/2, we see that the open sets A := [x, x + r) and B := [y, y + r)
disjoint open sets in Rℓ containing x and y, respectively. Thus, we have shown that Rℓ
is a Hausdorff space.
Similarly, it can be shown that R with upper limit topology is also Hausdorff. ♢
Remark 2.4. In Example 2.3, we have seen that the lower limit topology and the
upper limit topology on R are Hausdorff. However, they are not metrizable. - we shall
prove this fact after introducing some more notions (see Theorem 3.10). ♢
Example 2.5. For x = (x1 , x2 ) and y = (y1 , y2 ) in R2 , define
ρ(x, y) = |x1 − y1 |, x, y ∈ R2 .
Note that, ρ(·, ·) is a semi-metric or pseudo metric3 . For x ∈ X and r > 0, let
Bρ (x) := {u ∈ R2 : ρ(x, u) < ρ}.
Then
T := {A ⊆ R2 : ∀ x ∈ R2 , ∃ r > 0 such that Bρ (x) ⊆ A}
is a topology on R2 , which is not Hausdorff. ♢
Exercise 2.6. Verify the assertions in the above example. ♢
2.2 Convergence
Definition 2.7. (Convergence in topological space) Let (X, T ) be a topological

space. A sequence (xn ) in X is said to converge to x ∈ X if for every open set G
containing x, there exists n0 ∈ N such that xn ∈ G for all n ≥ n0 .
If (xn ) converges to x, then we write xn → x, and in that case x is called a limit
of (xn ). ♢
3
A function ρ : X × X → R is said to be a semi-metric or pseudo-metric if it satisfies all the
properties of a metric, except that ρ(x, y) = 0 need not imply x = y.
13
• A sequence can have more than one limit!
Example 2.8. In an indiscrete topological space X, every sequence converges to every

pint in X, as X is the only open set containing a point from X. In particular, if X
contains more than one point, then every sequence in X has more than one limit. ♢
Example 2.9. In a discrete space every convergent sequence is eventually constant:
Suppose X is a discrete space and (xn ) in X such that xn → x. Since {x} is an open
set containing x, there exists k ∈ N such that xn ∈ {x} for all n ≥ k. That is, xn = x
for all n ≥ k. ♢
Theorem 2.10. If X is a Hausdorff space, then every convergent sequence in X has
a unique limit.
Proof. Suppose X is a Hausdorff space and (xn ) is a convergent sequence in X. Suppose

x, y ∈ X such that xn → x and xn → y. We have to show that x = y.
Assume, on the contrary, that x ̸= y. Then by the Hausdorff property, there exist
disjoint open sets A and B such that x ∈ A and y ∈ B. Since xn → x and xn → y,
there exist n1 ∈ N and n2 ∈ N such that
xn ∈ A ∀ n ≥ n1 and xn ∈ B ∀ n ≥ n2 .
Let n0 := max{n1 , n2 }. Then we have
xn ∈ A and xn ∈ B ∀ n ≥ n0 .
This is not possible, since A and B are disjoint. Hence, our assumption that x ̸= y is
wrong. Thus, we have proved that x = y.
• Since every metrizable space is Hausdorff, every convergent sequence in a metriz-

able space has a unique limit.
Theorem 2.11. Suppose X is with co-countable topology. Then every convergent se-
quence in X is eventually constant.
Proof. If X is a countable set, then co-countable topology on X is the discrete topology,

and hence every convergent sequence in X is eventually constant. So, assume that X
is an uncountable set. Let (xn ) in X be such that xn → x. Let A := {xn : xn ̸= x}.
Since A is countable, Ac is an uncountable open set containing x. Hence, there exists
k ∈ N such that xn ∈ Ac for all n ≥ k. Thus, xn = x for all n ≥ k.
• A convergent sequence in a co-finite topological space need not be eventually

constant, as the following example shows.
14
Example 2.12. Let X = {0} ∪ { n1 : n ∈ N} with co-finite topology. Consider the
sequence (1/n). Let G be an open subset of X containing 0. Then X \ G is a finite set.
Hence, there exists k ∈ N such that n1 ∈ G for all n ≥ k. Hence, n1 → 0. This shows
that, a convergent sequence in a co-finite topological space need not be eventually
constant. ♢
Example 2.13. Consider the sequences (1/n) and (−1/n) in R. We observe that both
the sequences converge to 0 with respect to the usual topology and (1/n) converges to
0 in Rℓ . But (xn ) := (−1/n) does not converge in Rℓ : For any x ≥ 0, the open set
[x, x + 1) in Rℓ containing x does not contain any members of the sequence, so that
xn ̸→ x in Rℓ . Also, if x < 0, taking 0 < r < |x|, the open set [x, x + r) can contain
only a finite number of members of the sequence, so that xn ̸→ x in Rℓ . ♢
Example 2.14. Let X be the set of all real valued functions defined on an interval
I := [a, b]. Let T ⊆ 2X be such that A ∈ T iff for every f ∈ A, there exists r > 0
such that {g ∈ X : |g(a) − f (a)| < r} ⊆ A. Then, T is a topology on X, which is not
Hausdorff. With respect to this topology,
fn → f ⇐⇒ fn (a) → f (a). ♢
2.3 More on weak and strong topologies
If T1 and T2 are topologies on a set X, then T1 ∩ T2 is also a topology on X, which is

weaker than T1 and T2 . More generally, we have the following theorem, whose proof
follows by verifying the axioms for a topology.
Theorem
T 2.15. Let {Tα : α ∈ Λ} be a family of all topologies on a set X. Then
α∈Λ Tα is a topology on X which is weaker than Tα for every α ∈ Λ.
Example 2.16. On R, consider the usual topology Tu and the lower limit topology Tℓ .
• Tu is weaker than Tℓ .
To see this, let A be an open set with respect to the usual topology, and let x ∈ A.
Then there exists r > 0 such that (x − r, x + r) ⊆ A. This also implies [x, x + r) ⊆ A.
Hence, A is open with respect to the lower limit topology. ♢
Example 2.17. On R, consider the usual topology Tu and
T := {(a, ∞) : a ∈ R} ∪ {∅, R}.
We see that T is a topology on R: To see this, note that {∅, R} ⊆ T and for every
a, b ∈ R,
a < b ⇒ (a, ∞) ∩ (b, ∞) = (a, ∞) ∈ T .
15
Now consider a collection of sets from T , which is of the form {(a, ∞) : a ∈ S} for
some S ⊆ R. Then [
(a, ∞) = (a0 , ∞) ∈ T ,
a∈S
where a0 := inf S is either in R or it is −∞. Clearly,
• T is weaker than the usual topology on R. ♢

Theorem 2.18. Let T1 and T2 be topologies on X such that T1 is weaker than T2 ,
and let (xn ) be a sequence inX. If (xn ) converges to x with respect to T2 , then (xn )
converges to x with respect to T1 .
Proof. Suppose (xn ) converges to x with respect to T2 . We show that (xn ) converges
to x with respect to T1 . For this, let A ∈ T1 be such that x ∈ A. Since A ∈ T2 , there
exists n0 ∈ N such that xn ∈ A for all n ≥ n0 . This shows, by definition of convergence
in T1 , that (xn ) converges to x with respect to T1 .
• Convergence of a sequence with respect to a stronger topology does not imply

convergence with respect to a weaker topology.
Example 2.19. We know that discrete topology on a set is stronger than any other
topology on it. Now, consider the sequence (1/n) in R.
• (1/n) converges to 0 with respect to the usual topology on R, but it does not
converge with respect to the discrete topology on R.
To see this, assume for a moment that (1/n) converges to some point x0 ∈ R. Consider
the set {x0 }. It is open with respect to the discrete topology on R. But, this open set
contains only one point, and hence (1/n) does not converge to x0 . ♢
Example 2.20. We know that, on R, lower limit topology is stronger than the usual
topology. Consider the sequence (−1/n). In example 2.13, we have seen that the above
sequence converges to 0 with respect to the usual topology. But,
• (−1/n) not converge to any point with respect to the lower limit topology. ♢
Example 2.21. Let T be the topology on R defined as in Example 2.17. We have
seen that this topology is weaker than the usual topology. Thus, every sequence which
converges with respect to the usual topology converges with respect to T as well. We
note that:
• If (xn ) is any sequence of real numbers which is bounded below and if x ≥ x0 ,

where x0 = inf n xn , then xn → x with respect to T .
16
Indeed, if G is any open set containing x, then G is of the form G = (a, ∞) for some
a < x so that xn ∈ G for all n ∈ N.
• If (xn ) is any increasing sequence of real numbers and x ∈ R and k ∈ N such that
x ∈ [x1 , xk ], then xn → x with respect to T .
Indeed, if G is any open set containing x, then G is of the form G = (a, ∞) for some
a ∈ R, and hence xn ∈ G for all n ≥ k. ♢
The following theorem is a consequence of Theorem 2.15.
Theorem 2.22. Let S be a collection of subsets of X. Then the intersection of all

topologies on X containing S as a subcollection is a topology on X, and it is the weakest
topology containing S.
Proof. By Theprem 2.15, if T is the intersection of all topologies containing S, then T

is a topology on X such that S ⊆ T . If Te is any topology containing S, then T ⊆ Te ,
so that T is the weakest topology on X containing S.
Definition 2.23. Let S be a collection of subsets of X. The weakest topology on X

containing S is called the topology generated by S. ♢
Example 2.24. (i) The topology generated by the collection of all open intervals of
the form (a, b) with a < b is the usual toplogy on R.
(ii) The topology generated by the collection of all open intervals of the form (a, ∞)
and (−∞, b) is the usual toplogy on R. ♢
Example 2.25. Let S be the collection of all intervals of the form [a, b) with a < b.
Then the topology on R generated by S is the lower limit topology on R. ♢
Exercise 2.26. Show that the toplogy on R2 defined as in Example 2.5 is weaker than
the topology induced by the usual metric on R2 . ♢
2.4 Limit points, closure, interior and boundary
Definition 2.27. Let (X, T ) be a topological space and A ⊆ X.
1. A point x ∈ X is said to be a limit point of A if every open set containing x

contains at least one point of A other than x. The set of all limit points of A is
called the derived set of A, and it is denoted by A′ .
17
2. A point x ∈ X is said to be an interior point of A if there exists an open set
containing x and contained in A. The set of all interior points of A is called the
interior of A, and it is denoted by A◦ or int(A).
3. A point x ∈ X is said to be a closure point of A if every open set containing x

contains some point of A. The set of all closure points of A called the closure
of A, and it is denoted by Ā or cl(A).
4. A point x ∈ X is said to be a boundary point of A if open set containing x

contains some point of A and some point of Ac . The set of all boundary points
of A is called the boundary of A, and it is denoted by ∂A or bd(A).
5. A subset D of X is said to be dense in X if cl(D) = X. ♢
We observe (Exercise):
A◦ ⊆ A ⊆ Ā = A ∪ A′ = A ∪ ∂A = A◦ ∪ ∂A.
• It is also true that Ā = A◦ ∪ ∂A. To see this, first we observe that A◦ ∪ ∂A ⊆ ∂A.
To see the other way inclusion, let x ∈ Ā. Then, every open set G containing
x contains some some point of A. If G does not contain any point of Ac , then
G ⊆ A so that x ∈ A◦ . Thus, we have proved that if x ̸∈ ∂A, then x ∈ A◦ . in
other words, we have proved that Ā ⊆ A◦ ∪ ∂A.
Example 2.28. Let R be with usual topology. Then we have the following:
1. If A = (0, 1], then A◦ = (0, 1), Ā = [0, 1], A′ = [0, 1], ∂A = {0, 1}.
2. If A ∈ {Q, Qc }, then A◦ = ∅, Ā = R, A′ = R, ∂A = R.
3. If A ∈ {N, Z}, then A◦ = ∅, Ā = A, A′ = ∅, ∂A = ∅. ♢

Example 2.29. Let R2 be with usual topology, that is, the topology induced by the
usual metric d defined by
p
d(x, y) = (x1 − y1 )2 + (x2 − y2 )2
for x = (x1 , x2 ) and y = (y1 , y2 ) in R2 .

(i) Let S be the unit circle, that is,
S = {(x1 , x2 ) ∈ R2 : x21 + x22 = 1}.
Then (Verify), we have
S ◦ = ∅, S̄ = S, S ′ = S, ∂S = S.
18
(ii) Let A be the open unit unit disc, and B be the closed unit disc, that is,
A = {(x1 , x2 ) ∈ R2 : x21 + x22 < 1}, B = {(x1 , x2 ) ∈ R2 : x21 + x22 ≤ 1}.
Then (Verify), we have
A◦ = A, Ā = B, A′ = B, ∂A = S,
B ◦ = A, B̄ = B, B ′ = B, ∂B = S. ♢
Example 2.30. Let R be with co-countable topology. Note that, with respect to this
topology, any open subset has to be an uncountable set, and every countable set is
closed. Hence,
Q̄ = Q, Q◦ = ∅, Qc = R, (Qc )◦ = Q◦ . ♢
Example 2.31. Let X = {a, b, c} and T = {∅, X, {a}}. Then
1. If A = {a}, then A◦ = A, Ā = X, A′ = {b, c}.
2. If A = {b}, then A◦ = ∅, Ā = {b, c}, A′ = {c}.
3. If A = {c}, then A◦ = ∅, Ā = {b, c}, A′ = {b}.
4. If A = {a, b}, then A◦ = {a}, Ā = X, A′ = {b, c}.
5. If A = {a, c}, then A◦ = {a}, Ā = X, A′ = {b, c}.
6. If A = {b, c}, then A◦ = ∅, Ā = A = A′ . ♢
Theorem 2.32. Let (X, T ) be a topological space, and A and B be subsets of X such
that A ⊆ B. Then
A◦ ⊆ B ◦ , Ā ⊆ B̄.
Proof. Suppose x ∈ A◦ . Then there exists an open set G such that x ∈ G ⊆ A. Since
A ⊆ B, this shows that x ∈ B ◦ . Thus, A◦ ⊆ B ◦ .
Next, suppose x ∈ Ā. Then, for every open set G containing x, G ∩ A ̸= ∅, so that
we also have G ∩ B ̸= ∅, showing that x ∈ B̄. Thus, Ā ⊆ B̄.
Theorem 2.33. Let (X, T ) be a topological space and A ⊆ X.
(1) Ā is a closed set.
(2) A◦ is an open set.
(3) A is a closed set iff Ā = A.
19
(4) A is open iff A◦ = A.
Proof. (1). To show that (Ā)c is an open set. If Ā = X, then its compliment is ∅,
which is an open set. So, assume that X̄ ̸= X. Let x ∈ (Ā)c , that is, x is not a closure
point of A. Then there exists an open set Gx containing x, which does not contain any
point of A. Also, if y ∈ Gx , then y ̸∈ Ā, since the open set Gx containing y does not
contain any point of A. Thus, Gx ∩ Ā = ∅, so that Gx ⊆ (Ā)c . Therefore,
[
(Ā)c ⊆ Gx ⊆ (Ā)c .
x∈(Ā)c
Hence, (Ā)c is an open set.

(2). Let x ∈ A◦ . Then there exists an open set Gx containing x such that Gx ⊆ A.
This also implies that Gx ⊆ A◦ , since for every y ∈ Gx , Gx is an open set containing y
and contained in A. Hence, [
A◦ ⊆ Gx ⊆ A◦ .
x∈A◦
◦
Thus, A is an open set.
(3). By (1), if Ā = A, then A is a closed set. Conversely, suppose that A is a closed
set. We have to prove that Ā = A. Since A ⊆ Ā, it is enough to prove that Ā ⊆ A.
That is, to prove that x ̸∈ A⇒x ̸∈ Ā. Suppose x ̸∈ A, that is, x ∈ Ac . Thus, Ac is an
open set containing x which does not contain any point of A. Hence, x ̸∈ Ā.
(4). By (2), if A◦ = A, then A is an open set. Conversely, suppose that A is an
open set. We have to prove that A◦ = A. Since A◦ ⊆ A, it is enough to prove A ⊆ A◦ .
Let x ∈ A. Since A is an pen set containing x and contained in A, x is an interior
point of A. Thus, A ⊆ A◦ .
Theorem 2.34. Let X be a topological space A ⊆ X. The following results hold.
(1) A is closed iff A contains all its boundary points.

(2) A is open iff A does not contain any of its boundary points.
(3) A◦ is the largest open set contained in A.
(4) Ā is the smallest closed set containing A.
Proof. (1) Recall that Ā = A ∪ ∂A, and A is closed iff Ā = A. This shows that
A closed ⇐⇒ A ∪ ∂A = A ⇐⇒ ∂A ⊆ A.
(2) It is enough to show that A is not open iff A contains a boundary point of A.
First we recall that A is open iff A = A◦ .
20
Suppose that A is not open. Then there exists x ∈ A \ A◦ , which implies that
every open neighbourhood of x contains some point from Ac . This shows that x is a
boundary point of A.
Conversely, suppose A contains a boundary point x of A. Then every open neigh-
bourhood of x contains a point from Ac so that x ̸∈ A◦ = A, so that A is not open.
(3) We know that A◦ is an open set. Suppose G is an open set such that G ⊆ A.
Then we have G = G◦ ⊆ A◦ . Thus, A◦ is the largest open set contained in A.
(4) We know that Ā is a closed set. Suppose F is a closed set such that A ⊆ F .
Then we have Ā ⊆ F̄ = F . Thus, Ā is the smallest closed set containing A.
Example 2.35. Let X is with the discrete topology and A ⊆ X. Then A◦ = A = Ā.
♢
Definition 2.36. (Isolated point) Let X be a topological space and A ⊆ X. A point

x ∈ A is called an isolated point in A if it is not a limit point of A. ♢
Theorem 2.37. Let X be a topological space and A ⊆ X. Then a point x ∈ A is an

isolated point of A iff the singleton set {x} is open in A.
Proof. We observe that x ∈ A is an isolated point of A iff there exists an open set V
in X which does not conain any other point of A iff there exists an open set V in X
such that V ∩ A = {x} iff {x} is open is open in A
Corollary 2.38. Let X be a topological space. Then every point in X is an isolated

point iff X is a discrete space.
• Let X be a topological space and A ⊆ X. If (xn ) is sequence in A consisting of

distinct points which converges to x ∈ X, then x is limit point of A. However,
a limit point of A need not be a limit of any sequence in A, as the following
example shows.
Example 2.39. Let R be with co-countable topology. Then every point in Q is a

limit point of R \ Q. Howwever, in view of Theorem 2.11, a sequence in R \ Q cannot
converge to a rational number. ♢
Theorem 2.40. Suppose X is a Hausdorff space, A ⊆ X and x ∈ X Then x ∈ A′ iff

every open neighbourhood of x contains infinite number of elements from A.
Proof. Suppose x ∈ A′ . Assume for a moment that there exist an open neighbourhood
G of x such that G ∩ A contains only finite number of members. Since x ∈ A′ , we know
that G ∩ (A \ {x}) ̸= ∅. Let G ∩ (A \ {x}) = {x1 , . . . , xn } for some n ∈ N. Since X
21
is Hausdorff, for each i ∈ {1, . T
. . , n}, there exist disjoint open sets Ui abd Vi such that
x ∈ Ui and xi ∈ Vi . Let U := ni=1 Ui . Then we have
x ∈ U ∩ G, {x1 , . . . , xn } ∩ U = ∅,
so that U ∩ (G ∩ (A \ {x}) = ∅. This is a contradiction to the fact that the open set
U ∩ G must contain some point from A \ {x}.
The converse is obvious from the definition of A′ .
Definition 2.41. (Nowhere dense) Let X be a topological space and A ⊆ X. Then
A is said to be nowhere dense in X if (Ā)◦ = ∅. ♢
Example 2.42. Let X = R with usual topology. Then Z is nowhere dense in X, since
Z is closed and Z◦ = ∅, and Q is not nowhere dense, since (Ā)◦ = R◦ ̸= ∅. ♢
Theorem 2.43. Let X be a topological space and A be a closed set in X. Then A is
nowhere dense in X iff Ac is dense in X.
Proof. We observe that A is nowhere dense in X iff A◦ = ∅ iff A does not contain any
nonempty open set iff every nonempty open set contains some point from Ac iff Ac is
dense in X.
Remark 2.44. In Theorem 2.43, the condition that A is closed cannot be dropped.
To see this, let X = R with usual topology and A = Q. Then we know that Ac = R \ Q
dense, but (Ā)◦ = R◦ = R. ♢
2.5 Subspace topology
Let (X, T ) be a topological space and Y ⊆ X. Then the topology T induces a topology
on Y in a natural way as per the following theorem.
Theorem 2.45. Let (X, T ) be a topological space, and Y ⊆ X. Then
TY := {G ∩ Y : G ∈ T }
is a topology on Y .
Proof. Clearly ∅ ∈ TY . Since ∈ T , Y = X ∩ Y ∈ TY . Let A1 , A2 ∈ TY . Then there

exist G1 , G2 ∈ T such that
A1 = G1 ∩ Y, A2 = G2 ∩ Y
so that
A1 ∩ A2 = (G1 ∩ G2 ) ∩ Y ∈ TY .
22
Next, suppose {Aα : α ∈ Λ} ⊆ TY . Then, for each α ∈ Λ, there exists Gα ∈ T such
that Aα = Gα ∩ Y . Hence,
[ [
Aα = Gα ∩ Y ∈ TY .
α∈Λ α∈Λ
Thus, we have proved that TY is a topology on Y .
Definition 2.46. (Subspace topology) The topology TY obtained in Theorem 2.45

is called the subspace topology on Y or the topology on Y induced by T . ♢
• Let (X, T ) be a topological space and A ⊆ Y . Then A ∈ TY ⇐⇒ ∃ G ∈ T such

that A = G ∩ Y .
Convention: When we say that a subset A of Y is open in Y, it means that it is open

with respect to the subspace topology on Y .
Let X be a topological space and A ⊆ Y ⊆ X. If A is open in Y , then A need not
be open in X:
Example 2.47. Let R be with usual topology .

(i) let Y = [0, 1]. Then [0, 1/2) is open in Y , but [0, 1/2) is not open in R.
(ii) Let Y = Z. Then TY is the discrete topology on Z, since for every n ∈ Z, we

have {n} = Z ∩ (n − 21 , n + 21 ), so that every subset of Y is open in Y . Note that, ∅ is
the only subset of Y which is open in X. ♢
However, if Y ⊆ X is open in X, then every open set in Y is open in X. In fact,

we have the following theorem.
Theorem 2.48. Let X be a topological space and Y ⊆ X. Then,
TY ⊆ T ⇐⇒ Y ∈ T .
Proof. Suppose TY ⊆ T . Since Y ∈ TY , we obtain Y ∈ T . Conversely, suppose

Y ∈ T . If A ∈ TY , then A = G ∩ Y for some G ∈ T ; and since Y ∈ T , we obtain
A = G ∩ Y ∈ T . Thus, TY ⊆ T .
• Let X be a topological space and Y ⊆ X. By Theorem 2.48, every open set in

Y is open in X iff Y is open in X.
• An open set in Y can be open in X, without Y being open in X, as the following

example shows.
23
Example 2.49. Let X = R be with usual topology and Y = [0, 1] be with subspace
topology. Then (0, 1) is open in Y and X, though Y is not open in X. ♢
In fact, we have the following theorem.

Theorem 2.50. Let (X, T ) be a topological space, and Y ⊆ X. Then
{A ⊆ Y : A ∈ T } ⊆ TY .
Further, equality holds above iff Y is open in X, that is,
{A ⊆ Y : A ∈ T } = TY ⇐⇒ Y ∈ T .
In particular, if Y is open in X, then a subset of Y is open in Y iff it is open in X.
Proof. If A ⊆ Y and A ∈ T , then, since A = A ∩ G with G = A ∈ T , so that A ∈ TY .

Thus, we obtain {A ⊆ Y : A ∈ T } ⊆ TY .
Next, suppose that {A ⊆ Y : A ∈ T } = TY . Then, since Y ∈ TY , we obtain Y ∈ T .
Conversely, suppose that Y ∈ T and A ∈ TY . Then since A = Y ∩ G for some G ∈ T ,
we obtain A ∈ T . Hence, we have TY ⊆ {A ⊆ Y : A ∈ T }. Since we already have
{A ⊆ Y : A ∈ T } ⊆ TY , we obtain {A ⊆ Y : A ∈ T } = TY .
Exercise 2.51. Let X be a topological space, Y is open in X and A ⊆ Y ⊆ X. Show
that A is open in Y iff A is open X. ♢
Remark 2.52. Let (X, T ) be a topological space and Y ⊆ X. It can be easily seen
that
TeY := {Y } ∪ {A ⊆ Y : A ∈ T }}
is a topology on Y . By Theorem 2.50, we know that this topology is weaker than TY .
It can be too weak! For example, let X = R with usual topology and Y = Q. In this
case we know that the only subset of Q which is open in R is ∅. Thus, TeY = {∅, Q},
the indiscrete topology on Q, whereas TY contains all sets of the form G ∩ Q where G
is open in R. ♢
Theorem 2.53. Let X be a topological space, Y be a closed subset of X and A ⊆ Y .
Then A is closed in Y iff A is closed in X.
Proof. Suppose Y is closed in X and A is closed in Y . We observe that
X \ A = (X \ Y ) ∪ (Y \ A).
Since Y \ A is open in Y , there exists an open set G in X such that Y \ A = Y ∩ G.

Hence,
X \ A = (X \ Y ) ∪ (Y ∩ G) = Y c ∪ (Y ∩ G) = (Y c ∪ Y ) ∩ (Y c ∩ G) = Y c ∩ G.
24
Thus, X \ A is open in X.
Conversely, suppose Y is closed in X and A ⊆ Y is closed in X. Since Y \A = Y ∩Ac ,
where Ac is open in X, we obtain that Y \ A is open in Y , so that A is closed in Y .
Theorem 2.54. Let X be a topological space and Y ⊆ X. Let CX and CY denote the
set of all closed sets in X and Y , respectively. Then,
CY ⊆ CX ⇐⇒ Y ∈ CX .
In other words, every closed set in Y is closed in X iff Y is closed in X.
Proof. Suppose CY ⊆ CX . Since Y ∈ CY , we obtain Y ∈ CX . Conversely, suppose

Y ∈ CY . Then, by Theorem 2.53, we obtain CY ⊆ CX .
Finally we also note the following.
Theorem 2.55. Let X be a topological space and Y ⊆ X. Let A ⊆ Y . Then A is

closed in Y iff there exists a closed set F in X such that A = F ∩ Y .
Proof. Suppose A is closed in Y . Then, Y \ A is open in Y so that
Y \A=G∩Y
for some open set G in X. Now, we observe that
A = [(X \ Y ) ∪ (Y \ A)]c = [Y c ∪ (G ∩ Y )]c = [Y c ∪ G]c = Y ∩ Gc ,
where Gc is closed in X. Thus, there exists a closed set F in X such that A = F ∩ Y .

Conversely, suppose that there exists a closed set F in X such that A = F ∩ Y .
Then we have
Y \ A = Y \ (F ∩ Y ) = Y ∩ (F ∩ Y )c = Y ∩ (F c ∪ Y c ) = Y ∩ F c .
Thus, Y \ A is open in Y , and hence, A is closed in Y .
2.6 Base and sub-base
Bases are, in some sense, building blocks for topological spaces. Recall that, in a metric
space Ω, a subset A of is open iff for every x ∈ A, there exists an open ball B(x, rx )
for some rx > 0 such that B(x, rx ) ⊆ A. Also, we know that every open set is a union
of open balls. Motivated by this, we define the notion of a base for a topology.
25
Definition 2.56. (Base) Let (X, T ) be a topological space and B be a sub-collection
of T . Then B is called a base or basis for T if for every G ∈ T and x ∈ G, there
exists B ∈ B such that x ∈ B ⊆ G. . ♢
• Given a topological space (X, T ), the topology T itself is a base for T .
Example 2.57. Let B be the collection of all open intervals of the form (a, b) with
a < b. Then B is a base for the usual topology on R. ♢
Example 2.58. Let B be the collection of all intervals of the form [a, b) with a < b.
Then B is a base for the lower limit topology on R. ♢
Example 2.59. Let B be the collection of all intervals of the form (a, b] with a < b.
Then B is a base for the lower limit topology on R. ♢
Example 2.60. If X is a metric space with metric d, then the collection of all open
balls in X is a base for Td . ♢
Example 2.61. If X is a discrete topological space then the set of all singleton sets
is a base for the topology. ♢
Example 2.62. If X is an indiscrete topological space then {X} and {∅, X} are the
only bases for the topology. ♢
Here is a very useful characterization of a base.
Theorem 2.63. Let (X, T ) be a topological space and let B be a subcollection of T .

Then B is a base for T iff every nonempty open set is a union of some members of B.
Proof. Suppose B is a base for T . Let G ∈ T . To show that G is a union of members

of B. For this, let x ∈ G.
S Since B is a base for T , there exists Bx ∈ B such that
x ∈ Bx ⊆ G. Hence, G = x∈G Bx .
Conversely, suppose every open set is a union of members of B. Then, clearly, every
point in an open set G is contained in one of the members of B which is a subset of G.
Hence, B is a base for T .
Theorem 2.64. Suppose B is a base for topologies T1 and T2 on a set X. Then T1 = T2 .
Proof. Let A ∈ T1 . Then A is a union of members of B. Since B ⊆ T2 as well, it follows

that A ∈ T2 . Thus, T1 ⊆ T2 . Similarly, T2 ⊆ T1 .
Suppose X is a set. Then a given collection of subsets of X need not form a base
for some topology on X. However, we have the following two theorems.
26
Theorem 2.65. Let X be a set and B be a family of subsets of X such that B covers
X and B is closed under finite intersections. Then B is a base for a (unique) topology
on X.
Proof. Let T be the family of all unions of members of B together with the empty set.
Then, T is a topology for which B is a base. Further, by Theorem 2.64, this is the only
topology for which B is a base.
Recall that, if (Ω, d) is a metric space and B := {B(x, r) : x ∈ X, r > 0}, then the
metric topology that we defined is
Td := {A ⊆ Ω : ∀ x ∈ A, ∃ B ∈ B such that x ∈ B ⊆ A}.
Motivated by this, for any set X and for a collection B of subsets of X, let us consider
TB := {A ⊆ X : ∀ x ∈ A, ∃ B ∈ B such that x ∈ B ⊆ A}.
Under what conditions on B can we assert that TB is a topology on X?

The following theorem gives an answer to the above question.
Theorem 2.66. Let X be a set and B be a family of subsets of X such that B covers X
and for every B1 , B2 ∈ B and x ∈ B1 ∩ B2 , there exists B3 ∈ B with x ∈ B3 ⊆ B1 ∩ B2 .
Then
T := {A ⊆ X : for every x ∈ A, ∃ B ∈ B such that x ∈ B ⊆ A}.
is a topology on X and B is a (unique) base for T .
Proof. It is enough to show that T is a topology on X. Clearly ∅ ∈ T , as ∅ contains

no elements. Also, X ∈ T , since B covers X. Now, let A1 , A2 ∈ T . To show that
A1 ∩ A2 ∈ T . For this, let x ∈ A1 ∩ A2 . We know that there exists B1 , B2 ∈ B such
that x ∈ B1 ⊆ A1 and x ∈ B2 ⊆ A2 . Hence, x ∈ B1 ∩ B2 ⊆ A1 ∩ A2 . By assumption,
there exists B3 ∈ B such that x ∈ B3 ⊆ B1 ∩ B2 . Thus, x ∈ B3 ⊆ A1 ∩ A2 . This
shows
S that A1 ∩ A2 ∈ T . Next, Slet {Aα : α ∈ Λ} ⊆ Te for some Λ ̸= ∅. To show that
α∈Λ Aα ∈ T . For this, let x ∈ S β ∈ Λ such that x ∈ Aβ ,
α∈Λ Aα . Then there exists
and hence, there exists B ∈ B such that x ∈ B ⊆ Aβ ⊆ α∈Λ Aα .
Exercise 2.67. Deduce Theorem 2.65 from Theorem 2.66. ♢
Remark 2.68. We may observe that, in the context of a metric space (Ω, d), the
conditions in Theorem 2.66 are satisfied: Note that
[
Ω= B(x, r)
x∈Ω
27
for any r > 0. Now, consider open balls B(x, r) and B(y, s) for some x, y ∈ Ω and
r, s > 0. Let u ∈ B(x, r) ∩ B(y, s). Then, taking
t = min{d(x, u), r − d(x, u), d(y, u), s − d(y, u)},
it can be seen that B(u, t) ⊆ B(x, r) ∩ B(y, s). Thus, we have shown that the family
B := {B(x, r) : x ∈ X, r > 0}
satisfies the conditions in Theorem 2.66. ♢
Definition 2.69. (Sub-base) Let (X, T ) be a topological space. If S is a collection of

subsets of X such that T is the topology generated by S, then S is called a sub-base
for T . ♢
• Every base is a sub-base, but the converse is not true.
Example 2.70. Let S be the collection of all intervals of the form (a, ∞) and (−∞, b).
Then S is a sub-base for the usual topology Tu on R, but S is not a base for Tu . For
instance, (−1, 1) is an open set containing 0. But, there is no set in S containing 0
and contained in (−1, 1). ♢
Theorem 2.71. Let X be a nonempty set and S be a family of subsets of X. Let T

be the collection of all unions of finite intersections of members of S, together with the
sets ∅ and X. Then T is a topology on X, and it is the topology generated by S.
Proof. Clearly, T is a topology on X. Also, if Te is any topology containing S, then Te

must contain T .
Theorem 2.72. Let (X, T ) be a topological space, S be a family of subsets of X, and

B be the family of all finite intersections of members of S. Then S is a sub-base for T
iff B is base for T .
Proof. Suppose S is a sub-base for T . Then, by Theorem 2.71 and Theorem 2.63, B
is a base for T . Conversely, suppose B is a base for T . Then, by Theorem 2.63, S is a
subs-base for T .
2.7 A topological proof for infinitude of primes
We may recall that the set P of all prime numbers is infinite. One of the simplest
proofs of this fact is due to Euclid:
28
Suppose P is a finite set, say P = {p1 , . . . , pn } for some n ∈ N. Consider
the number
q = 1 + p 1 p2 · · · pn .
Then none of the pi divides q. Hence, q is a prime. But, q ̸∈ P, which is a
contradiction.
Now, we give a topological proof for the infinitude of primes, by using a topology on
Z as follows:
For (a, b) ∈ Z × N, consider the set
Na,b := {a + bn : n ∈ Z}.
Define
T := {A ⊆ Z : ∀ a ∈ A, ∃b ∈ N such that Na,b ⊆ A}.
The following can be verified (Exercise):
• T is a topology on Z;
• Na,b is both open and closed;
• Every A ∈ T is an infinite set.
Now, we observe that [

Z \ {1, −1} = N0,p .
p∈P
This shows that, if P is a finite set, then Z \ {1, −1} is a closed set, which implies the
finite set {1, −1} is an open set. Thus, we arrive at a contradiction.
Exercise 2.73. Show that the set B := {Na,b : a ∈ Z, b ∈ N} is a base for the topology
T above. ♢
3 Countability axioms and separability

Recall that, if (X, T ) is a topological space, then a sub-family of T is a base (or basis)
for the topological space X (or for the topology) T iff every nonempty set in T is a
union of some members of B. We have many examples of topological spaces and bases
for them. Also, we know that, the base can be a much smaller sub-collection than the
topology itself. For example, for a discrete space, the collection of all singleton sets is
a base. In this case, the topology can be an uncountable collection, whereas a base can
29
be a countable collection4 . We may also observe that, if a base for a topology contains
only a finite number of sets, then the topology itself contains a finite number of sets.
Thus, if a topology consists of infinite number of sets, then every base of such topology
must contain infinite number of sets.
In view of the above observations we introduce the following definition.
Definition 3.1. (Second countable space) A topological space or a topology is said

to be second countable if it has a countable base. ♢
• If a topological space X is second countable, then X is said to satisfy the second

axiom of countability.
We shall define a first countable space soon.
Example 3.2. The space Rk with usual topology is separable and second countable.
This follows from the above theorem since the countable set D := Q × Q is dense in
R2 . In fact, the family
{B(u, r) : u ∈ D, r ∈ Q+ }
is a countable base for Rk with usual topology. ♢
Example 3.3. A discrete space X is second countable iff X is a countable set. ♢
Remark 3.4. Suppose S is any countable family of subsets of a set X. In the following
if we shall write S = {Sn : n ∈ N}, with the understanding that if S is a finite collection,
say S = {S1 , . . . , Sk } for some k ∈ N, then we take Sn = Sk for all n ≥ k. ♢
For the next theorem, recall that a subset D of a topological space X is dense in X
if D̄ = X, that is, for every x ∈ X, there exists an open neighbourhood G of X such
that G ∩ D ̸= ∅.
Theorem 3.5. Every metric space with a countable dense subset is second countable.
Proof. Let Ω be a metric space with a metric d and let D be a countable dense subset
of Ω. Let (εn ) be a sequence of positive reals such that εn → 0. Consider the family
A := {B(u, εn ) : u ∈ D, n ∈ N}.
Clearly, A is a countable family of open sets. We show that A is a base for X. For
this, let A be an open set in Ω and x ∈ A. Let r > 0 be such that B(x, r) ⊆ A. Let
4
If X = Z with discrete topology, then the cardinality of the base of singleton sets is ℵ0 , whereas
the cardinality of the topology is c = 2ℵ0 , the continuum.
30
n ∈ N be large enough such that εn < r/2. Since D is dense in Ω, the ball B(x, εn )
contains some point u ∈ D. Then
x ∈ B(u, εn ) ⊆ B(x, r) ⊆ A.
Indeed, u ∈ B(x, εn ) implies x ∈ B(u, εn ) and if v ∈ B(u, εn ), then
d(v, x) ≤ d(v, u) + d(u, x) < εn + εn < r
so that B(u, εn ) ⊆ B(x, r) as well. Thus, we have proved that A is a countable base
for Ω.
Definition 3.6. (Separable space) A topological space or a topology is said to be

separable if it has a countable dense subset. ♢
Example 3.7. The metric space X = Rk with usual metric is separable. For example,
the set Qk is a countable dense subset of X. ♢
Example 3.8. Every indiscrete topological space X separable and second countable.
In fact, the only base of the indiscrete space X in {X}, and every nonempty subset of
X is dense in X, in particular, every countable subset of X is dense in X. ♢
Every separable metric space is second countable (Theorem 3.5).

An uncountable discrete space is neither separable nor second countable.
Theorem 3.9. Every second countable space is separable.
Proof. Let (X, T ) be a second countable space and let B = {Bn : α ∈ Λ} be a

countable base for T , where Λ is some subset of N. For each n ∈ Λ, let xn ∈ Bn and
let D := {xn : n ∈ Λ}. Clearly, D is a countable set. We claim that D is dense in
X. For this, let x ∈ X and let G be an open neighbourhood of x. Since B is a base
for X, there exists k ∈ Λ such that x ∈ Bk ⊆ G. In particular, xk ∈ Bk ∩ G, so that
D ∩ G ̸= ∅. Hence, D is dense in X.
We have proved that (see Theorem 3.5) every separable metric space is second
countbale.
• Is every separable topological space second countable? Not necessarily, as the

following theorem shows.
Theorem 3.10. The lower limit topology and the upper limit topology on R are sepa-
rable, but not second countable. In particular, they are not metrizable.
31
Proof. Let Rℓ := R with lower limit topology.
(a) It can be easily seen that Q is dense in Rℓ , and hence it is separable (Exercise).
(b) Let B be a base for Rℓ . Then for each x ∈ Rℓ , there exists Bx ∈ B such
that x ∈ Bx ⊆ [x, ∞). Clearly, inf Bx = x. Hence, x ̸= y implies Bx ̸= By . Hence,
{Bx : x ∈ Rℓ } is uncountable, and therefore, B is uncountable. Thus, Rℓ is not second
countable.
(c) Since Rℓ is separable and not second countable, by Theorem 3.5, it is not
metrizable.
We know that, with respect to the usual metric, R is separable. Also, as a subspace
of R, Q is separable. Is R\Q separable? The answer is in the affirmative, which follows
from the following general result.
Theorem 3.11. Every subspace of a separable metric space is separable.
Proof. Let (Ω, d) be a separable metric space and Ω0 be a subset of Ω. We show that
Ω0 is also separable. Clearly, if Ω0 is countable, then it is separable. So assume that Ω0
is uncountable. Then D has to be countably infinite (Why?). Let D = {xn : n ∈ N}.
Then, for each n ∈ N, we have
∞
[
Ω= B(xi , 1/n)
i=1
so that ∞
[
Ω0 = B(xi , 1/n) ∩ Ω0 ∀ n ∈ N.
i=1
Also, for every n ∈ N, there exists j ∈ N such that B(xj , 1/n) ∩ Ω0 ̸= ∅. Let
∆ := {(j, n) ∈ N × N : B(xj , 1/n) ∩ Ω0 ̸= ∅}.
For (j, n) ∈ ∆, let yj,n ∈ B(xj , 1/n) ∩ Ω0 . Let
D0 := {yj,n : (j, n) ∈ ∆}.
We show that D0 is dense in Ω0 . For this, let x ∈ Ω0 and ε > 0 be given. We have to
show that B(x, ε) ∩ D0 ̸= ∅. Note that, for every n ∈ N, there exists j ∈ N such that
x ∈ B(xj , 1/n) so that (j, n) ∈ ∆. Let n be large enough such that 2/n < ε. Then, we
have
1 1
d(x, yj,n ) ≤ d(x, xj ) + d(xj , yj,n ) < + < ε.
n n
Thus, B(x, ε) ∩ D0 ̸= ∅, and the proof is complete.
32
At this point, one may ask: Is subspace of every separable topological space sepa-
rable? The answer to this question is in the negative, as the following example shows.
Example 3.12. Let T consists of ∅ and those A ⊆ R2 such that for every (x, y) ∈ A,
there exists r > 0, s > 0 such that
[x, x + r) × [y, y + s) ∈ A.
It can be easily seen that T is a topology on R2 . Let X = R2 with this topology

T . As in the case of lower limit topology, X is separable. We show that the set
X0 = {(x, −x) : x ∈ R} is not separable. To see this, we observe that, for every x ∈ R,
{(x, −x)} = X0 ∩ ([x, x + r) × [−x, −x + r))
for any r > 0. Hence, every singleton subset of X0 is open in X0 , so that the subspace
topology on X0 is the discrete topology. Since an uncountable set with discrete topology
is not seprable, X0 is not seprable. ♢
Definition 3.13. Let (X, T ) be a topological space and x ∈ X. A family Bx of open
neighbourhoods of x is said to be a base at x or local base at x if every open set
containing x contains some member of Bx . ♢
Definition 3.14. (First countable space) A topological space X is said to be first
countable if it has a countable local base at every point in X. ♢
• If a topological space X is first countable, then X is said to satisfy the first

axiom of countability.
Example 3.15. (i) In R with usual topology, the set Bx := {(x−1/n, x+1/n) : n ∈ N}
is a local base at x ∈ R.
(ii) If R is with lower limit topology, then Bx := {[x, x + 1/n) : n ∈ N} is a local
base at x ∈ R for every x ∈ R.
(ii) In a discrete space X, the family Bx = {{x}} is a local base at x.
(iii) In an indiscrete space X, Bx = {X} is the only local base at x. ♢
Theorem 3.16. Every metric space is first countable.
Proof. Let (Ω, d) be metric space. Then, for every x ∈ Ω, the family Bx := {Bd (x, 1/n) :
n ∈ N} is a countable base at x: For every open set A containing x, there exists r > 0
such that Bd (x, r) ⊆ A. Taking n ∈ N such that 1/n < r, we have Bd (x, 1/n) ⊆
Bd (x, r) ⊆ A.
Theorem 3.17. Every second countable space is first countable, and the converse is
not true.
33
Proof. Let (X, T ) be a second countable space, and let B be a countable base for X.
Let x ∈ X, and let
Bx := {B ∈ B : x ∈ B}.
Then Bx is a countable base at x. Indeed, for every open set A containing x, there
exists Bx ∈ B such that x ∈ B ⊆ A. Clearly Bx ∈ Bx . This shows that Bx is a
countable base at x.
The converse of the above theorem is not true. To see this, it is enough consider an
uncountable discrete space X. This space is not second countable. However, for each
x ∈ X, the family Bx := {{x}} is a base at x.
Every metric space is first countable, but not necessarily separable and
second countable.
We observe that:
Theorem 3.18. If (xn ) is a sequence in a subset A of a topological space such that
xn → x, then x ∈ Ā.
• Converse is not true as the following example shows.
Example 3.19. Let X = R with co-countable topology and A = R \ Q. Then Ā = X,

since for every x ∈ R every open neighbourhood G of x is an uncountable set. However,
if x ∈ Q, then there is no sequence in A which converges to x, as every convergent
sequence in X is eventually constant.
More generally, let X be any uncountable set with co-countable topology and let
A = X \ B, where B is a nonempty countable set. Then, following the arguments as
in last paragraph, we see that Ā = X and a point in B is not a limit of any sequence
in A. ♢
However, if X is a first countable space, then converse to Theorem 3.18 does hold.
Theorem 3.20. Suppose X is a first countable space and A ⊆ X. Then x ∈ Ā iff
there exists a sequence (xn ) in A such that xn → x.
Proof. Let x ∈ Ā. Let Bx := {Bn : n ∈ N} be a countable base at x. Let Gn :=

T n
j=1 Bj . Since x ∈ Ā and Gn is an open set containing x, there exists xn ∈ Gn ∩ A
for all n ∈ N. We claim that xn → x. To see this, let G be an open set containing x.
Since Bx is a countable base at x, there exists k ∈ N such that Bk ⊆ G. Note that
Gn ∩ A ⊆ Bk ⊆ G ∀n ≥ k.
34
In particular, xn ∈ G for all n ≥ k. Thus, we have proved that xn → x.
Conversely, suppose (xn ) is a sequence in A such that xn → x. Let G be any open
set containing x. Then, by the definition of convergence, there exists k ∈ N such that
xk ∈ G. In particular, G ∩ A ̸= ∅, showing that x ∈ Ā.
An uncountable co-countable topological space is not first countable.

(see Theorem 3.20 and Example 3.19.)
• With respect to co-countable topology R, Q is a proper closed subset of R,

whereas R \ Q is dense in R.
Recall that:
• In a Hausdorff topological space, every convergent sequence has a unique limit.

• Is the converse true? The answer is in the negative, as the following example
shows.
Example 3.21. Consider the co-countable topology on an uncountable set X. We
know that this space is not Hausdorff, and every convergent sequence is eventually
constant, and hence has a unique limit (cf. Theorem 2.11). ♢
However, we have the following theorem.

Theorem 3.22. Suppose X is a first countable topological space. Then every conver-
gent sequence has a unique limit iff X is Hausdorff.
Proof. We have already proved that if X is Hausdorff, then every convergent sequence
in X has a unique limit.
Conversely, suppose X is first countable and every convergent sequence in X has
a unique limit. To show that X is Hausdorff. For a moment assume that X is not
Hausdorff. Then there exists u, v ∈ X with u ̸= v such that every pair of open sets
containing u and v, respectively, will intersect. Now, let Bu = {Un : n ∈ N} and
Bv = {Vn : n ∈ N} be countable bases at u and v, respectively. Let
n
\ n
\
An = Uj , Bn = Vj ∀ n ∈ N.
j=1 j=1
Then An and Bn are open sets containing u and v, respectively, and hence An ∩Bn ̸= ∅.
Let xn ∈ An ∩ Bn . We show that xn → u and xn → v, which would contradict the
35
hypothesis. For this, let G be an open set containing u. Since Bu is a base at u, there
exists k ∈ N such that Uk ⊆ G. Hence An ⊆ G for all n ≥ k, so that we also have
xn ∈ An ∩ Bn ⊆ An ⊆ G ∀ n ≥ k.
Thus, xn → u. Analogously, using the fact that Bv is a local base at v, we see that
xn → v. This completes the proof.
• The co-countable topology on an uncountable space is not first countable, and

hence, not second countable as well (See Example 3.21 and Theorem 3.22).
4 Functions and their properties
4.1 Motivation
From the theory of metric spaces, recall the following: Suppose (X, d) and (Y, ρ) are
metric spaces. Recall that a function f : X → Y is continuous at a point x0 ∈ X if for
every ε > 0, there exists a δ > 0 such that
x ∈ X, d(x, x0 ) < δ ⇒ ρ(f (x), f (x0 )) < ε.
In fact, we know the following result5 .

Theorem 4.1. Let (X, d) and (Y, ρ) be metric spaces and x0 ∈ X. Then a function
f : X → Y is continuous at x0 ∈ X iff for every open neighbourhood V of f (x0 ), there
exists an open neighbourhood U of x0 such that
x∈U ⇒ f (x) ∈ V, i.e., f (U ) ⊆ V.
Proof. Suppose f : Ω1 → Ω2 is continuous at a point x0 ∈ Ω1 . Let V be an open

neighbourhood of f (x0 ). Then there exists ε > 0 such that BY (f (x0 ), ε) ⊆ V . Since
f : Ω1 → Ω2 is continuous at a point x0 ∈ Ω1 , there exists δ > 0 such that
d(x, x0 ) < δ ⇒ ρ(f (x), f (x0 )) < ε.
That is,
x ∈ BX (x0 , δ) ⇒ f (x) ∈ BY (f (x0 ), ε) ⊆ V.
Thus, taking U = BX (x0 , δ), we obtain
x∈U ⇒ f (x) ∈ V, i.e., f (U ) ⊆ V.

5
Let X be a metric space. Then the ball in a metric space with cetre x0 and radius r will be
denoted by BX (x0 , r).
36
Conversely, suppose that for every open neighbourhood V of f (x0 ), there exists an
open neighbourhood U of x0 such that
x∈U ⇒ f (x) ∈ V, i.e., f (U ) ⊆ V.
To show that f is continuous at x0 . For this, let ε > 0 be given and V = B(f (x0 ), ε).
Then, by hypothesis, there exists an open neighbourhood U of x0 such that x ∈ U
implies f (x) ∈ V . Let δ > 0 be such that BX (x0 , δ) ⊆ U . Thus, we have
x ∈ BX (x0 , δ) ⇒ f (x) ∈ BY (f (x0 ), ε).
Thus, we have proved that for every ε > 0, there exists δ > 0 such that d(x, x0 ) < δ
imples ρ(f (x), f (x0 )) < ε, that is, f is continuous at x0 .
For our later references, let us recall the following results from real analysis.
Theorem 4.2. Let Ω and Ω e be metric spaces and f : Ω → Ω

e be a continuous function.
The following results hold.
(1) If Ω is compact, then f (Ω) is compact.
(2) If Ω is connected, then f (Ω) is connected.
4.2 Continuity and sequential continuity
Motivated by the above, we have the following definition of continuity of functions

between topological spaces.
Definition 4.3. Let X and Y be topological spaces. A function f : X → Y is said to

be continuous at a point x ∈ X if for every open neighbourhood V of f (x), there
exists an open neighbourhood U of x such that f (U ) ⊆ V . ♢
We also define the following.
Definition 4.4. Let f : X → Y be a function between topological spaces X and Y .
1. f is said to be continuous on a subset E of X if it is continuous at every point

in E.
2. f is said to be continuous if it is continuous on X. ♢
Exercise 4.5. 1. If X is a discrete space and Y is any topological space, then

every functions f : X → Y is continuous.
37
2. Let X be a nonempty set and T1 and T2 be topologies on X. Let X1 = X with
T1 and X2 = X with T2 . Then the identity map from X1 to X2 is continuous iff
T1 is stronger than T2 . ♢
Theorem 4.6. Let f : X → Y be a function between topological spaces X and Y .

Then f is continuous iff for every open set G in Y , f −1 (G) is open in X.
Proof. Suppose f is continuous. Let G be an open set in Y . To show that f −1 (G) is

open in X. For this, let x ∈ f −1 (G). Then f (x) ∈ G. Hence, using the contiuity of
f at x, there exists an open set U in X such that x ∈ U and f (U ) ⊆ G. This implies
that U ⊆ f −1 (G). Thus, we have proved that for every x ∈ f −1 (G), there exists an
open set U in X such that x ∈ U ⊆ f −1 (G). Hence, f −1 (G) is open.
Conversely, assume that for every open set G in Y , f −1 (G) is open in X. We have
to show that f is continuous on X. So, let x ∈ X and V be an open neighbourhood of
f (x) in Y . By assumption, U := f −1 (V ) is an open set in X. Since f (x) ∈ V , U is an
open neighbourhood of x in X. Note that
y∈U ⇒ f (y) ∈ V, i.e., f (U ) ⊆ V.
This shows that f is continuous.
Remark 4.7. In some of the books on topology, the continuity of a function is defined
in terms of the above characterization. ♢
Theorem 4.8. Let X and Y be topological spaces. Then f : X → Y is continuous iff

for every closed set B in Y , f −1 (B) is closed in X.
Proof. Follows by observing that for every B ⊆ Y , f −1 (B c ) = (f −1 (B))c .
The proof of the following theorem follows using the definitions of base and sub-
base for a topology: Recall that a family B of open sets is a base for a topology T if
every open set is a union of members of B, and a family S of open sets is a sub-base
for a topology T if the collection of all finite intersections of members of S form a base
for T .
Theorem 4.9. Let X and Y be topological spaces, B be a base and S be a sub-base for
the topology on Y . The, for a function f : X → Y , the following are equivalent:
(i) f is continuous.
(ii) f −1 (B) is open in X for every B ∈ B.
(iii) f −1 (S) is open in X for every S ∈ S.
38
Example 4.10. Let X be any topological space and R be with usual topology. Then
a function f : X → R is continuous iff
{x ∈ X : a < f (x) < b}
is open in X for any a, b ∈ R with a < b. ♢
Theorem 4.11. Let X and Y be topological spaces, and let f : X → Y be a continuous
function. Then, for any sequence (xn ) in X,
xn → x in X ⇒ f (xn ) → f (x) in Y.
Proof. Let (xn ) be a sequence in X such that xn → x in X. To show that f (xn ) → f (x).
For this, let A be an open set containing f (x). Since f is continuous, f −1 (A) is open
in X. Also, x ∈ f −1 (A). Since xn → x and f −1 (A) is an open set containing x, there
exists n0 ∈ N such that xn ∈ f −1 (A) for all n ≥ n0 . This implies that f (xn ) ∈ A for
all n ≥ n0 . Thus, f (xn ) → f (x).
Remark 4.12. By the above theorem. if we can find a sequence (xn ) in X such that
xn → x for some x ∈ X and f (xn ) ̸→ f (x), then f is not continuous. ♢
• Converse of Theorem 4.11 is not true. To see this, look at the following example.
Example 4.13. Let Rcc = R with co-countable topology and Rdis = R with discrete
topology. Consider the identity function f : Rcc → Rdis , that is, f (x) = x for every
x ∈ R. This function is not continuous. For example, {1} is an open set in Rdis , but
f −1 ({1}) is not open, as its compliment is not countable. However, for every sequence
(xn ) in R,
xn → x in Rcc ⇒ f (xn ) → f (x) in Rdis .
This is true, because, if xn → x in Rcc , then (xn ) is eventually constant, and hence
xn → x in Rdis . ♢
More generally we have the following example.

Example 4.14. Let X be a co-countable topological space and Y be any topological
space. Let f : X → Y be any function. We know that if (xn ) is a convergent sequence
in X, then it is eventually constant, and hence its image (f (xn )) is also eventually
constant. In other words,
xn → x ⇒ ∃ N ∈ N such that xn = x ∀ n ≥ N
⇒ f (xn ) = f (x) ∀ n ≥ N
⇒ f (xn ) → f (x).
However, this f need not be continuous. For example, if X is uncountable co-countable
space, Y is with discrete topology and if f is one-one, then f is not continuous. Because,
for y ∈ Y , the singleton set {y} is open in Y , but f −1 ({y}), which is a singleton set, is
not open in X. ♢
39
Definition 4.15. (Sequential continuity) Let X and Y be topological spaces. Then
a function is said to be sequentially continuous if for any sequence (xn ) in X,
xn → x in X ⇒ f (xn ) → f (x) in Y. ♢
Continuity implies sequential continuity. Converse does not hold.
In certain special cases, continuity and sequential continuity are equivalent.
Theorem 4.16. Let X and Y be topological spaces. If X is a metrizable space, then

continuity and sequential continuity of a function f : X → Y are equivalent.
Proof. We know that if f : X → Y is continuous, then f is sequentially continuous.

Now, suppose f : X → Y is sequentially continuous, and X is metrizable. In this case
we show that f is continuous.
Let d be a metric which induces the topology on X, and assume for a moment that
f is not continuous. Then there exists x0 ∈ X and an neighbourhood V of f (x0 ) such
that for any open neighbourhood U of x0 , there exists x ∈ U such that f (x) ̸∈ V .
In particular, for any n ∈ N, taking U as BX (x0 , 1/n), there exists xn ∈ BX (x0 , 1/n)
such that f (xn ) ̸∈ V for all n ∈ N. Thus, we obtain sequence (xn ) in X such that
xn → x0 , but f (xn ) ̸→ f (x0 ). This is a contradiction to the fact that f is sequentially
continuous.
Continuity and sequential continuity are equivalent if the domain space

is metrizable.
More generally, we show that, if X is first countable, then continuity of a function

f : X → Y is equivalent to sequentially continuous. For proving this, we make use of
the following characterization of continuity in a general setting.
Theorem 4.17. Let X and Y be topological spaces. Then f : X → Y is continuous

iff for every A ⊆ X, f (Ā) ⊆ f (A).
Proof. ⇒): Suppose f is continuous. Let x ∈ Ā. We have to show that f (x) ∈ f (A).
Let V be an open neighbourhood of f (x). To show that V contains some point from
f (A). Since f is continuous, we know that f −1 (V ) is an open neighbourhood of x, and
since x ∈ Ā, the open neighbourhood f −1 (V ) of x contains some point from A, say
u ∈ f −1 (V ) ∩ A. Then f (u) ∈ V ∩ f (A).
40
⇐): Suppose f (Ā) ⊆ f (A) for every A ⊆ X. We have to show that f is continuous.
By Theorem 4.8, it is enough to show that f −1 (B) closed for every closed B ⊆ Y .
That is, taking A = f −1 (B), we have to show that A = Ā. So, let x ∈ Ā. Then, by
hypothesis, f (x) ∈ f (A). But, since A = f −1 (B), we have f (A) ⊆ B, so that
f (x) ∈ f (A) ⊆ B̄ = B.
Thus, we have proved that x ∈ Ā imples x ∈ f −1 (B) = A, showing that Ā = A.
Let us recapitulate some of the results which we have already proved.
Let X and Y be topological spaces and f : X → Y . Then the following are

equivalent.
• f is continuous.
• For every open set V in Y , f −1 (V ) is open in X.
• For every closed set F in Y , f −1 (F ) is closed in X.
•For every set A ⊆ X, f (Ā) ⊆ f (A).
Now, we prove the promised theorem.
Theorem 4.18. Let X and Y be topological spaces. If X is a first countable space,

then continuity and sequential continuity of a function f : X → Y are equivalent.
Proof. Let X and Y be topological spaces and X be a first countable space. Suppose
f : X → Y is sequentially continuous. We have to show that f is continuous. By
Theorem 4.17, it is enough to show that f (Ā) ⊆ f (A) for every A ⊆ X. So, let A ⊆ X
and x ∈ Ā. Since X is first countable, by Theorem 3.20, there exists (xn ) in A such
that xn → x. By assumption f (xn ) → f (x). Since f (xn ) ∈ f (A), it follows that
f (x) ∈ f (A), and the proof is complete.
Continuity and sequential continuity are equivalent if the domain space is

first countable.
4.3 Continuity and separability
Lemma 4.19. Continuous image of a dense subset is dense.
41
Proof. Let X and Y be topological spaces and f : X → Y be a continuous function.
Let D be a dense subset of X. We show that f (D) is dense in f (X). For this, let
y ∈ f (X) and V be an open set in f (X) which contains y. Then V = G ∩ f (X), where
G is open in Y and y ∈ G. We have to show that f (D) ∩ V ̸= ∅.
Let x ∈ X be such that y = f (x). Since f is continuous, f −1 (G) is an open
neighbourhood of x. Since D is dense in X, D ∩ f −1 (G) ̸= ∅. Let u ∈ D ∩ f −1 (G).
Clearly, f (u) ∈ f (D) ∩ G ⊆ f (D) ∩ V . Thus, f (D) ∩ V ̸= ∅ and the proof is
complete.
As a corollary to the above lemma, we have the following theorem.

Theorem 4.20. Continuous image of a separable space is separabe.

Suppose X is separable and D be a countable dense subset of X. By Lemma 4.19,
f (D) is dense in X. Clearly, f (D) is a countable subset of f (X).
Example 4.21. Let Pk be the set of all polynomials of degree atmost k with real
coefficients and let P be the set of all polynomials with real coefficients. Let
d(f, g) = sup |f (t) − g(t)| for f, g ∈ P.
0≤t≤1
It can be easily seen that d(·, ·) is a metric on P. We show that
• Xk := Pk and X := P are separable with respect to the metric d(·, ·) above.
For thus, let Rk+1 be with the metric

ρ(u, v) := |u1 − v1 | + · · · + |uk − vk |.
consider the function φk : Rk+1 → Xk defined by
φ(a0 , a1 , . . . , ak ) = a0 + a1 t + · · · + ak tk .
Note that φ is a bijective function. I can be easily seen, using the denseness of Q in
R, that Qk+1 is dense in Rk+1 (Verify). Note that
|φ(a0 , a1 , . . . , ak )| ≤ |a0 | + |a1 | + · · · + |ak |.
From this, it follows that
d(φ(u), φ(v)) = |φ(u) − φ(v)| ≤ ρ(u, v).
Hence, φ is a continuous function. Hence, by Theorem 4.20 that φ(Qk+1 ) is a countable
dense subset of Pk .
Since , P = ∞
S
k=1 Pk , X is also separable. ♢
Exercise 4.22. Subspace of a separable space is separable. ♢
42
4.4 Restriction and expansion and composition theorems
The proofs of all the items in the following theorem follow from the definition of
continuity. However, we supply brief explanations.
Theorem 4.23. Let X and Y be topological spaces and f : X → Y be a continuous
function. Then the following results hold.
(1) (Domain restriction theorem) Let X0 be a subspace of X. Then the restric-

tion of f , namely, the function f0 : X0 → Y defined by
f0 (x) = f (x) ∀ x ∈ X,
is continuous.
(2) (Co-domain restriction theorem) Let Y0 be a subspace of Y with f (X) ⊆ Y0 .
Then the function g : X → Y0 defined by
g(x) = f (x) ∀ x ∈ X,
is continuous.
(3) (Co-domain expansion theorem) Let Z be a topological space such that Y is
a subspace of Z. Then the function h : X → Z defined by
h(x) = f (x) ∀ x ∈ X,
is continuous.
Proof. (1) Let A be any subset of Y . Then we see that

f0−1 (A) = f −1 (A) ∩ X0 .
In particular, for every open set V in Y , f0−1 (V ) = f −1 (V ) ∩ X0 is open in X0 .
(2) Let V be open in Y0 . Then V = G ∩ Y0 , where G is open in Y . Then, we have
g −1 (V ) = f −1 (V ) = f −1 (G) ∩ X = f −1 (G),
which is open in X.
(3) Let G be open in Z. Then V = G ∩ Y is open in Y and we have
f −1 (V ) = f −1 (G ∩ Y ).
But, for x ∈ X,
x ∈ f −1 (G ∩ Y ) ⇐⇒ f (x) ∈ G ∩ Y ⇐⇒ f (x) ∈ G ⇐⇒ h(x) ∈ G.
Thus, f −1 (V ) = h−1 (G),where f −1 (V ) is open in X. Thus, inverse image of every open
set in G under h is open in X, showing that h is continuous.
43
Theorem 4.24. (Compositiion thorem) Let X, Y and Z be topological spaces and
f : X → Y and g : Y → Z be continuous functions. Then the composition function
g ◦ f : X → Z is continuous.
Proof. Let V be an open set in Z. Then, we have
(g ◦ f )−1 (V ) = f −1 [g −1 (V )],
which is open in X.
Definition 4.25. Given a function f : X → Y and a subset X0 of X, the function
f0 : X0 → Y defined by
f0 (x) = f (x), x ∈ X0 ,
is called the restriction of f to X0 , and it is usually denoted by f |X0 , and in that
case, f is called an extension of f0 . ♢
• Restriction of every continuous function is continuous. But, the converse need

not be true, as the following example shows.
Example 4.26. Let X = R be with usual topology and X0 = [0, 1] be with the
subspace topology. Then f0 : X0 → X defined by
f0 (x) = x, x ∈ X0 ,
is continuous. Let f : X → X be define by

x, x ∈ X0 ,
f (x) =
0, x ∈ X \ X0 .
We note that f is an extension of f0 , and f is not continuous. ♢
• Does every continuous function defined on a subspace has a continuous extension

to all of X? Not necessarily, as the following example shows.
Example 4.27. Let R be with usual topology and ((0, ∞) is with subspace topology.
Let f0 : (0, ∞) → R be defined by
1
f0 (x) = , x ∈ (0, ∞).
x
Then f0 is continuous. But, it does not have a continuous extension to R (Why?). ♢
In the following theorem6 , we give a condition on the subspace which guarantees

continuous extension.
6
This theorem is not explicitly stated in the referred books.
44
Theorem 4.28. Let X, Y be topological spaces, X0 be a subspace of X and f0 : X0 → Y
be continuous. If X0 is open as well as closed, then f0 has a continuous extension to
all of X.
Proof. Let f1 : X0c → Y be any continuous function (e.g., we may take f1 (x) = y0 for
every x ∈ X0c for any given y0 ∈ Y ). Then define f : X → Y by

f0 (x) if x ∈ X0 ,
f (x) =
f1 (x) if x ∈ X0c ,
Now, let V be an open set in Y . Then we have
f0−1 (V ) = f −1 (V ) ∩ X0 , f1−1 (V ) = f −1 (V ) ∩ X0c .
Hence,
f −1 (V ) = f0−1 (V ) ∪ f1−1 (V ).
Since f0 and f1 are continuous, the sets f0−1 (V ) and f1−1 (V ) are open in X0 and X0c ,
respectively. Now, since X0 and X0c are open in X, the sets f0−1 (V ) and f1−1 (V ) are
open in X as well. Thus, f −1 (V ) is open in X. Thus, we have proved that f is a
continuous extension of f0 .
Exercise 4.29. Does R with usual topology have a nonempty proper subset which is
open and closed? Why? ♢
Theorem 4.28 is a particular case of the following theorem, called the pasting lemma.
Theorem 4.30. (Pasting lemma) Let X and Y be topological spaces, and A and B
be closed subspaces of X such that X = A ∪ B. If f1 : A → Y and f2 : B → Y are
continuous functions such that
f1 (x) = f2 (x) ∀ x ∈ A ∩ B,
then the function f : X → Y defined by

f1 (x) if x ∈ A,
f (x) =
f2 (x) if x ∈ B,
is a continuous function.
Proof. We observe that for every subset S of Y ,
f −1 (S) = f1−1 (S) ∪ f2−1 (S).
Now, if S is closed in Y , then f1−1 (S) closed in A and f2−1 (S) closed in B, and since A
and B are closed, f1−1 (S) and f2−1 (S) are closed in X as well. Thus, we have proved
that inverse every closed set under f is closed in X.
45
Exercise 4.31. Prove the pasting lemma (Theorem 4.30) by assuming that both A
and B are closed, instead of assuming that both are open. ♢
Continuous extension of a function is also possible under certain other conditions

on the spaces and the function (see Appendix below).
Appendix: Continuous extensioin
We show that continuous extension of a function is possible under certain conditions

on the spaces and the function. First, let us recall the definition of uniform continuity
of a function between metric spaces.
Definition 4.32. Let (X, d) and (Y, ρ) be metric spaces. Then a function φ : X → Y
is uniformly continuous on a subset S of X if for every ε > 0, there exists δ > 0
such that
x, y ∈ S, d(x, y) < δ ⇒ ρ(f (x), f (y)).
♢
Theorem 4.33. Let (X, d) and (Y, ρ) be metric spaces, and X0 be a dense subspace of
X. If Y is is complete, then uniformly continuous function f0 : X0 → Y has a unique
continuous extension f : X → Y .
Proof. Suppose Y is complete and f0 : X0 → Y is a uniformly continuous function. Let

x ∈ X. Since X0 is dense in X, there exists a sequence (xn ) in X0 such that xn → x.
Since f0 is uniformly continuous, it can be seen (Exercise) that the sequence (f0 (xn ))
is a Cauchy sequence in Y . Since Y is complete, (f0 (xn )) converges. Let
f (x) = lim f0 (xn ).

n→∞
We show that f is a well-defined continuous function from X to Y which is the unique

extension of f0 .
(i) f f is a well-defined function from X to Y : To show this we have to show that
f (x) independent of the sequence (xn ) which converges to x. For this, let (un ) be any
other sequence in X0 which converges to x. Then the sequence (f0 (un )) is a Cauchy
sequence in Y so that (f0 (un )) also converges. We show that f (un ) → f (x). Since
ρ(f (un ), f (x)) ≤ ρ(f (un ), f (xn )) + ρ(f (xn ), f (x))

≤ ρ(f0 (un ), f0 (xn )) + ρ(f0 (xn ), f (x)) ∀ n ∈ N
and ρ(f0 (xn ), f (x)) → 0, it is enough to show that ρ(f0 (un ), f0 (xn )) → 0.
Now, let ε > 0 be given. Since f0 is uniformly continuous, there exists δ > 0 such
that
d(v, w) < δ ⇒ ρ(f (v), f (w)) < ε. (1)
46
Since d(xn , x) → 0 and d(un , x) → 0, there exists N ∈ N such that
d(xn , x) < δ/2 and d(un , x) < δ/2 ∀ n ≥ N
so that
d(xn , un ) ≤ d(xn , x) + d(x, un ) < δ ∀ n ≥ N. (2)
Hence, by (1) and (2) above,
ρ(f (xn ), f (un )) < ε ∀ n ≥ N.
Thus, we have shown that ρ(f (un ), f (x)) → 0, and the the value f (x) is is uniquely
defined, making f : X → Y a function.
(ii) Note that
f (x) = f0 (x) ∀ x ∈ X0 .
Thus, f0 is an extension of f0 .
(iii) f is continuous: For this, let x ∈ X and ε > 0 be given. Since f0 is uniformly
continuous, there exists δ > 0 such that
v, w ∈ X0 , d(v, w) < δ ⇒ ρ(f0 (v), f0 (w)) < ε. (1)
For a δ0 > 0, let u ∈ X be such that d(x, u) < δ0 , and let (xn ) and (un ) in X0 such
that d(xn , x) → 0 and d(un , u) → 0. We note that
ρ(f (x), f (u)) ≤ ρ(f (x), f0 (xn )) + ρ(f0 (xn ), f0 (un )) + ρ(f0 (un ), f (u)) (2)
for all n ∈ N. By the definition of f (x) and f (u), we know that ρ(f (x), f0 (xn )) → 0
and ρ(f (un ), f0 (u)) → 0. Hence, there exists N1 ∈ N such that
ρ(f (x), f0 (xn )) < ε, ρ(f (un ), f0 (u)) < ε ∀ n ≥ N1 . (3)
Note that
d(xn , un ) ≤ d(xn , x) + d(x, u) + d(u, un ) ∀ n ∈ N.
Since d(xn , x) → 0 and d(u, un ) → 0, there exists N2 ∈ N such that
δ δ
d(xn , x) < , d(u, un ) < ∀ n ≥ N2 .
4 4
Hence, taking δ0 := δ/2, we have
δ
d(x, u) < ⇒ d(xn , un ) < δ ∀ n ≥ N2
2
so that by (1),
ρ(f (xn ), f (un )) < ε ∀ n ≥ N2 .
47
Therefore, by (2) and (3),
d(x, u) < δ/2 ⇒ ρ(f (x), f (u)) < 3ε ∀ n ≥ N := max{N1 , N2 }.
Thus, we have proved that f is continuous.
(iv) The function f in (iii) is the unique continuous extension of f0 : Suppose
g : X → Y is any continuous extension of f0 . Then for every x ∈ X, if (xn ) in X0 is
such that d(xn , x) → 0, then we have
g(x) = lim g(xn ) = lim f0 (xn ) = f (x).
n→∞ n→∞
Thus, g = f .
4.5 Homeomorphism, open and closed maps
Definition 4.34. Let X and Y be topological spaces, and f : X → Y . Then f is said

to be a homeomorphism if f is bijective, continuous, and its inverse f −1 : Y → X is
continuous,. ♢
• Let X and Y be topological spaces, and f : X → Y be a bijective function. Then

f : X → Y is a homeomorphism iff f −1 : Y → X is a homeomorphism
Definition 4.35. Let X and Y be topological spaces. If there is a homeomorphism
from X onto Y , then X is said to be homeomorphic with Y , and in that case we
write X ∼ Y . ♢
We observe that
X ∼ Y ⇐⇒ Y ∼ X.
• Let X be the family of all topological spaces. On X , then the relation ∼ is an

equivalence relation (Verify).
We shall represent the fact that X is not homeomorphic with Y by

X ̸∼ Y.
Example 4.36. (A homeomorphism) Let R be with usual topology. For any a, b ∈
R with a ̸= 0, the function f : R → R defined by
f (x) = ax + b, x ∈ R,
is a homeomorphism. It can be easily seen that f is a continuous function. Also, for
x, y ∈ R,
y−b
f (x) = y ⇐⇒ x = ,
a
so that f is bijective and f −1 is continuous. ♢
48
Example 4.37. (A continuous bijection which is not a homeomorphism) Let
X = R be with discrete topology and Y = R with usual topology. Then the identity
map f : X → Y defined by
f (x) = x, x ∈ X,
is continuous and bijective, but not a homeomorphism: Clearly, f is bijective, and
since X is with discrete topolgy, f is continuous as well. Note that, g := f −1 : Y → X
is defined by g(y) := f −1 (y) = y for every y ∈ Y , and the inverse image of the open
set {1} in X under the map g is not open in X. ♢
Example 4.38. (A homeomorphism between open intervals) For any a, b, c, d
with a < b and c < d, the open intervals (a, b) and (c, d) are homeomorphic with udual
topologies on them: To see this consider the function f : (a, b) → (c, d) be defined by
d−c
f (x) = c + (x − a).
b−a
Then we see that f is continuous, bijective, and its inverse is given by
b−a
f −1 (y) = a + (y − c),
d−c
which is also continuous.
Using the same arguments, [a, b] is homeomorphic with [c, d], [a, b) is homeomorphic
with [c, d) and (a, b] is homeomorphic with (c, d]. ♢
With respect to the usual topology on R,
• any two open intervals are homeomorphic.
• any two closed and bounded intervals are homeomorphic.
For intervals with end points a and b, one may ask the following questions:
• Is (a, b) ∼ [a, b]?
• Is (a, b) ∼ (a, b]?
• Is (a, b) ∼ [a, b)?
• Is [a, b] ∼ (a, b]?
• Is [a, b] ∼ [a, b)?
49
• Is [a, b) ∼ (a, b]?
We shall see that answer to only the last question is in the affirmative.
Example 4.39. (Non-homeomorphic spaces) We have following.
(1) With topologies induced by the usual metric, the space X = [a, b] is not home-
omorphic with any of the spaces Y1 = (a, b), Y2 = (a, b], Y3 = [a, b): Note that X
is compact, whereas Y1 , Y2 , Y3 are not compact. Recall that, continuous image of a
compact metric space is compact. Hence, there is no continuous onto function from X
to any of Y1 , Y2 , Y3 .
(2) With topologies induced by the usual metric, the space X = (a, b) is not home-
omorphic with any of the spaces Y1 = (a, b], Y2 = [a, b).
To see this, first consider X and Y1 . Assume for a moment that X and Y are
homeomorphic. Then there is a continuous onto function g : Y1 → X. Let x0 = g(b).
The
A = Y1 \ {b} = (a, b), g(A) = (a, x0 ) ∪ (x0 , b).
Note that A is a connected subset of Y1 whereas g(A) is disconnected, which is a
contradiction to the fact that, continuous image of a connected set has to be connected.
Thus, we have proved that X = (a, b) is not homeomorphic with Y1 = (a, b].
Using similar arguments, it can be shown (Verify) that X = (a, b) is not homeo-
morphic with Y2 = [a, b). ♢
Example 4.40. (Homeomorphic spaces) With topologies induced by the usual
metric, the space X = [a, b) is homeomorphic Y = (a, b]. To see this, let φ : (a, b] →
[−b, −a) be defined by
φ(x) = −x.
This function is a homeomorphism from (a, b] onto [−b, −a), that is [a, b) ∼ [−b, −a).
Hence, in view of Example 4.38, [−b, −a) ∼ [a, b). Hence, by the transitivity of the
relation ∼, we have [a, b) ∼ (a, b]. Since (a, b] ∼ (c, d], we also have [a, b) ∼ (c, d]. ♢
(a, b) ∼ (c, d), [a, b) ∼ [c, d), (a, b] ∼ (c, d], (a, b] ∼ [c, d), [a, b] ∼ [c, d]
(a, b) ̸∼ [a, b], (a, b) ̸∼ [a, b), (a, b) ̸∼ (a, b], (a, b] ̸∼ [a, b], [a, b) ̸∼ [a, b].
Remark 4.41. Bear in mind that, to show that two spaces are homeomorphic, it is
enough to find one homeomorphism from X onto Y , whereas to show that two spaces
are not homeomorphic, it is necessary to show that there is no homeomorphism from
X onto Y . ♢
50
Example 4.42. (A homeomorphism) Let (0, 1) and (1, ∞) be with usual topology,
and f : (0, 1) → (1, ∞) be defined by
1
f (x) = , x ∈ (0, 1).
x
We note that f is bijective with its inverse defined by
1
f −1 (y) = , y ∈ (1, ∞).
y
Further, both f and f −1 are continuous (using a result from Real Analysis). Thus, f
is a homeomorphism.
Now, let c ∈ R. Let fc : (0, 1) → (c, ∞) be defined by
1−x 1−x
fc (x) = c + −1=c+ , x ∈ (0, 1).
x x
Then fc is a homeomorphism from (0, 1) to (c, ∞) (Verify). We know that (see Example
4.38), the space (a, b) is homeomorphic with (0, 1). Hence, it follows that (a, b) is
homeomorphic with (c, ∞).
Next, let gc : (c, ∞) → (−∞, c) be defined by
g(x) = −x, x ∈ (c, ∞).
This is a homeomorphism from (c, ∞) to (−∞, c). Hence, it follows that (a, b) is
homeomorphic with (−∞, c).
Using similar arguments, it can be seen that [a, b) is homeomorphic with [c, ∞) and
to (−∞, c] (Verify). ♢
Example 4.43. (A homeomorphism) Let (−1, 1) and R be with usual topology,
and f : (−1, 1) → R be defined by
x
f (x) = , x ∈ (−1, 1).
1 − x2
This function is a homeomorphism: It can be easily seen that f is continuous (using a
result from Real Analysis). Also, for y ∈ R and x ∈ (−1, 1),
x 2y
f (x) = y ⇐⇒ 2
= y ⇐⇒ x = .
1−x 1 + (1 + 4y 2 )1/2
This shows that f is bijective and the function g : R → (−1, 1) defined by
2y
g(x) = , y∈R
1 + (1 + 4y 2 )1/2
is the inverse of f , and g is continuous (Verify).
We know that (see Example 4.38), the space (a, b) is homeomorphic with (−1, 1).
Hence, it follows that (a, b) is homeomorphic with R. ♢
51
Example 4.44. (A continuous bijection which is not a homeomorphism) Let
X = [0, 1) with usual topology and
Y = S 1 := {(a, b) : a2 + b2 = 1}
with subspace topology of R2 with usual topology. Let f : X → Y be defined by
f (t) = (cos 2πt, sin 2πt), t ∈ [0, 1).
Then f is bijective and continuous, but its inverse is not continuous, and hence it is
not a homeomorphism (Verify). ♢
Example 4.45. (Non-homeomorphic spaces) Let X = [0, 1] and Y = [0, ∞) be

with usual topologies, so that X and Y are metric spaces. These spaces are not
homeomorphic. In fact, there is no continuous onto function from X to Y , since X is
compact and Y is not compact. ♢
Example 4.46. (Non-homeomorphic spaces) Let X = R with usual topology and

Y = R with discrete topology. Then X and Y are not homeomorphic. In fact there
is no continuous bijection function from X to Y . This is seen as follows: Suppose
f : X → Y be a bijective function. Then inverse image of singleton sets under f are
singleton sets. Now, singleton sets in Y are open sets, whereas singleton sets are not
open in X. Hence, f cannot be continuous.
Also, if g : Y → X is a bijective function, then it is continuous, but its inverse is
not continuous, by the same reason as in the last paragraph. ♢
Let X ∈ {(c, ∞), (−∞, d), R} and Y ∈ {[c, ∞), (−∞, d]}. Then
(a, b) ∼ X, [a, b) ∼ Y, X ̸∼ Y .
Theorem 4.47. Let X and Y be topological spaces, and f : X → Y be a homeomor-

phism. Then we have the following.
(1) Image of every open set in X is open in Y .
(2) Image of every closed set in X is closed in Y .
Proof. We note that for every A ⊆ X,
(f −1 )−1 (A) = f (A).
52
Indeed, since f is bijective, for y ∈ Y ,
y ∈ (f −1 )−1 (A) ⇐⇒ f −1 (y) ∈ A ⇐⇒ y ∈ f (A).
Hence, using the fact that f −1 is continuous, it follows that f (A) is open if A is open,
and f (A) is closed if A is closed.
Functions having the properties (1) and (2) in Theorem 4.47 have special names.
Definition 4.48. Let X and Y be topological spaces, and f : X → Y . Then a function

f : X → Y is said to be
1. an open map if for every open set A in X, f (A) is open in Y, and
2. a closed map if for every closed set A in X, f (A) is closed in Y . ♢
• By Theorem 4.47, every homeomorphism is an open and closed map.
• If f : X → Y be a bijective continuous function between topological spaces, then

f is open iff f −1 : Y → X is continuous.
A bijective continuous function is a homeomorphism iff it is open.
A bijective continuous function is a homeomorphism iff it is closed.
Example 4.49. The identity map on a topological space is open, closed, and a home-
omorphism. ♢
Example 4.50. If X is a topological space and Y is a discrete space, then every

function from X to Y is open and closed. ♢
Exercise 4.51. Write details of the proofs of the above two theorems.
(Hint: for every A ⊆ X, (f −1 )−1 (A) = f (A). ) ♢
Now, we give examples of following types of maps:
• an open map which is not a closed map;
• a closed map which is not an open map;
• a map which is neither open and nor closed;
53
Example 4.52. Let X = (0, 1) and Y = R with usual metric topologies and leet
f : X → Y be the inclusion map. That is, f (x) = x, x ∈ X. This is an open map
which is not a closed map:
Any open set in X is also open in Y ; hence it is an open map. Note that, A = (0, 1/2]
is closed in X which is not closed in Y . Hence, f is not a closed map. ♢
Example 4.53. Let X = [0, 1] and Y = R with usual metric topologies and leet
f : X → Y be the inclusion map. That is, f (x) = x, x ∈ X. This is an closed map
which is not an open map:
Any closed set in X is also closed in Y ; hence it is an closed map. Note that,
A = [0, 1/2) is open in X which is not open in Y . Hence, f is not an oepn map. ♢
Example 4.54. Let X = [0, 1) and Y = R with usual metric topologies and let
f : X → Y be the inclusion map. That is, f (x) = x, x ∈ X. This is an neither an
open map nor a closed map: In this case [0, 1) is open and closed in X, but it is neither
open nor closed in Y . Hence, f is neither an open map nor a closed map. ♢
Example 4.55. Let X = (0, 1) and Y = (1, ∞), both with usual topology, and
f : X → Y be defined by f (x) = 1/x, x ∈ X. Then f is a homeomorphism,. ♢
Theorem 4.56. Let X be a topological space and X0 be a subset of X with subspace

topology T0 . Then the inclusion map f : X0 → X, defined by f (u) = u, u ∈ X0 , is
(i) open iff X0 is open in X;
(ii) closed iff X0 is closed in X.
Proof. The proof follows from Theorem 2.48 and Theorem 2.54.
Exercise 4.57. Let R and R2 be with usual topologies. Show that the map f : R → R2
defined by
f (x) = (x, 0), x ∈ R,
is a closed map, but not an open map. ♢
Exercise 4.58. Verify the following:
1. The sets A = {(x, y) ∈ R2 , x2 + y 2 < 1} and B = {(x, y) : |x| + |y| < 1}

are homeomorphic with respect to the subspace topologies induced by the usual
topology on R2 .
2. The sets A = {(x, y) ∈ R2 , 1 < x2 + y 2 < 2} and B = {(x, y) : 1 < |x| + |y| < 2}
are homeomorphic with respect to the subspace topologies induced by the usual
topology on R2 .
54
3. The sets A = {(x, y) ∈ R2 , x2 + y 2 < 2} and B = {(x, y) : 1 < |x| + |y| < 2} are
not homeomorphic with respect to the subspace topologies induced by the usual
topology on R2 .
4. The sets A = {(x, y) ∈ R2 , x2 + y 2 = 1} and B = {x ∈ R : 0 ≤ x ≤ 1} are

not homeomorphic with respect to the subspace topologies induced by the usual
topology on R2 and R, respectively. ♢
Appendix
Consider the Example 4.44. Since t 7→ cos 2πt and t 7→ sin 2πt are continuous on [0, 1),
the map f : t 7→ (cos 2πt, sin 2πt) defined from [0, 1) to R2 also continuous. Hence,
f : X → Y is continuous. Now, for s, t ∈ [0, 1), we have
(cos 2πt, sin 2πt) = (cos 2πs, sin 2πs) ⇐⇒ (cos 2πt − cos 2πs, sin 2πt − sin 2πs) = (0, 0)
⇒
cos 2π(s − t) = cos 2πs cos 2πt + sin 2πs sin 2πt = cos2 2πt + sin2 2πt = 1
⇒
2π(s − t) = 0 ⇒ s = t.
Thus, we have proved that f is one-one.
For (a, b) ∈ S 1 with a ̸= 0,
b 1
(cos 2πt, sin 2πt) = (a, b) ⇒ tan 2πt = ⇒t = tan−1 (b/a).
a 2π
If (a, b) ∈ S 1 with a = 0, we have b = 0 so that f (0) = (0, 1). Thus, we have shown
that f is bijective and continuous.
To see that its inverse is not continuous, consider the open set [0, 1/4) in X. Its
inverse image under f −1 is
{(cos 2πt, sin 2πt) : 0 ≤ t < 1/4}
which is not an open set in Y .
5 Connected spaces and path connected spaces

We have already learned the notions of connectedness and compactness in the context
of metric spaces in the course on Real Analysis. We extend these notions to topological
spaces.
55
5.1 Connectedness
Definition 5.1. A topological space X is said to be a connected space if it is not a

union of two disjoint nonempty open sets. A topological space which is not connected
is called a disconnected space. ♢
Definition 5.2. By a separation of a topological space X, we mean, a pair (A, B) of

disjoint nonempty open sets in X such that X = A ∪ B. ♢
• A topological space X is connected iff there is no separation of X.
• A topological space X is disconnected iff there is a separation of X.
Definition 5.3. A subset of a topological space is said to be connected if it is a

connected space as a subspace. ♢
Example 5.4. The following can be verified easily (Exercise).
1. A subset of a discrete space is connected iff it is a singleton set (Exercise).
2. If X is an uncountable set with co-countable topology, then X is connected.
3. If X is an infinite set with co-finite topology, then X is connected. ♢
Theorem 5.5. A nonempty topological space is connected iff the whole space is its only
nonempty subset which is both open and closed.
Proof. Let X nonempty topological space. Suppose X is connected, and E is a

nonempty subset which is both open and closed. Then E c is also both open and
closed. Since X is connected and X = E ∪ E c , we have E c = ∅, for otherwise, (E, E c )
would be a separation of X. Hence, E = X.
Conversely, suppose that X is the only nonempty subset which is both open and
closed. We have to show that X is connected. Assume on the contrary that X is not
connected, and (A, B) is a separation of X. Then A and B are different subsets which
are both open and closed, which is a contradction to the hypothesis.
Theorem 5.6. Let X be a topological space and E ⊆ X0 ⊆ X. Then E is a connected

subset of X0 iff E is a connected subset of X.
Proof. It is enough to show that E disconnected in X0 iff E is disconnected in X.

Suppose E disconnected in X0 . Hence, there exists open sets V1 and V2 in X0 such
that
E = (E ∩ V1 ) ∪ (E ∩ V2 ),
56
where E ∩ V1 and E ∩ V2 are nonempty disjoint sets. Since V1 and V2 are open in X0 ,
there exists open sets G1 and G2 in X such that V1 = X0 ∩ G1 and V2 = X0 ∩ G2 .
Hence,
E ∩ V1 = E ∩ X0 ∩ G1 = E ∩ G1 , E ∩ V2 = E ∩ X0 ∩ G2 = E ∩ G2 .
Thus, E ∩ G1 and E ∩ G2 are disjoint nonempty subsets of E such that
E = (E ∩ G1 ) ∪ (E ∩ G2 ),
where G1 and G2 are open in X. This shows that E is disconnected in X.
Conversely, suppose that E disconnected in X. Then there exists open sets G1 and
G2 in X such that
E = (E ∩ G1 ) ∪ (E ∩ G2 ),
where E ∩ G1 and E ∩ G2 are nonempty disjoint sets. Note that
E ∩ G1 = E ∩ X0 ∩ G1 = E ∩ V1 , E ∩ G2 = E ∩ X0 ∩ G2 = E ∩ V2 ,
where V1 = X0 ∩ G1 and V2 = X0 ∩ G2 . Hence,
E = (E ∩ V1 ) ∪ (E ∩ V2 ),
where E ∩ V1 and E ∩ V2 are nonempty disjoint sets with V1 and V2 are open in X0 .
This shows that E is disconnected in X0 .
Remark 5.7. In view of the above theorem, while talking about connected sets in
a topological space, we do not specify explicitly whether it is connected in the whole
space or in a subspace. ♢
Theorem 5.8. Continuous image of a connected set is connected.

Let E be a connected subset of X. To show that f (E) is connected subset of Y .
Assume for a moment that f (E) is not connected subset of Y . Then there exists
disjoint nonempty open sets V1 and V2 in f (E) such that f (E) = V1 ∪ V2 . Let G1 and
G2 be open sets in Y such that V1 = G1 ∩ f (E) and V2 = G2 ∩ f (E). Then we have
f (E) ⊆ G1 ∪ G2
so that
E ⊆ f −1 (G1 ) ∪ f −1 (G2 ),
and hence
E = [E ∩ f −1 (G1 )] ∪ [E ∩ f −1 (G2 )].
Note that, since V1 and V2 are disjoint nonempty sets implies that E ∩ f −1 (G1 ) and
E ∩ f −1 (G2 ) are are disjoint nonempty sets in E, contradicting the fact that E is
connected.
57
Example 5.9. Let X = R with usual topology and Y be a discrete space. Then the
only continuous function from X to Y is the constant function: Suppose f : X → Y
is a continuous function. Then f (X) is connected, as X is connected. Since Y is a
discrete space, f (X) has to be a singleton set. Thus f is a constant function. ♢
Theorem 5.10. Let A and B be nonempty connected subsets of a topological space X

such that A ∩ B ̸= ∅. Then A ∪ B is connected in X.
This is a particular case of the following more general result.
T If {Aα : α ∈ Λ} S
Theorem 5.11. is a family of connected subsets of a topological space
X such that α∈Λ Aα ̸= ∅, then α∈Λ Aα is connected.
T S
Proof. Let x0 ∈ α∈Λ Aα . Suppose X0 := α∈Λ Aα is not connected. Then there exists
disjoint nonempty open sets V1 and V2 in X0 such that X0 = V1 ∪ V2 . Let G1 and G2
be open sets in X such that V1 = G1 ∩ X0 and V2 = G2 ∩ X0 . Since V1 ∩ V2 = ∅, either
x0 ∈ V1 or x0 ∈ V2 , but not in both. Assume that x0 ∈ V1 .
Now, let α ∈ Λ. Then
Aα = Aα ∩ (V1 ∪ V2 ) = (Aα ∩ V1 ) ∪ (Aα ∩ V2 ), (∗)
where Aα ∩ V1 and Aα ∩ V2 are disjoint open sets in Aα . Since Aα is connected and

x0 ∈ Aα ∩ V1 , we obtain Aα ∩ V2 = ∅. Hence, Aα = Aα ∩ V1 ⊆ V1 . This is true for
every α ∈ Λ. Hence [
X0 = Aα ⊆ V1 ⊆ X0 .
α∈Λ
This is a contradiction to the fact that V2 ̸= ∅.
As an immediate corollary to Theorem 5.10, we have the following result.
Theorem 5.12. If E0 , E1 , . . . , En are connected

Sn subsets of a topological space X such
that Ei−1 ∩ Ei ̸= ∅ for i = 1, . . . , n, then i=0 Ei is connected.
A natural question one may ask is the following:
Suppose (En ) is a sequence of connected subsets

S of a topological space X
such that En ∩ En+1 ̸= ∅ for every n ∈ N. Is ∞n=1 En connected?
We shall answer this question affirmatively by making use of the following proposition.
Propositon 5.13. Suppose X is a disconnected topological space and (A, B) is a sep-

aration of X. If E is a connected subset of X, then either E ⊆ A or E ⊆ B.
58
Proof. Given that X = A ∪ B, where A and B are disjoint nonempty open subsets of
X such that A = A ∪ B. Then we have
E = (E ∩ A) ∪ (E ∩ B). (∗)
Note that E ∩ A and E ∩ B are open in E, and they are disjoint as well. Hence, by
the connectedness of E,
E ∩ A = ∅ or E ∩ B = ∅.
If E ∩ A = ∅, then from (∗), we see that E = E ∩ B, so that E ⊆ B. Similarly,
E ∩ B = ∅ implies E ⊆ A.
As a corollary to the above, we have the following.
Theorem 5.14. Let X be a topological space. Then X is connected iff every pair of
points in X is contained in a connected subset of X.
Proof. Clearly, if X is connected, then every pair of points in X is contained in the

connected set X.
Conversely, suppose that every pair of points in X is contained in an connected
subset of X. We have to show that X is connected. Assume on the contrary that X is
not connected. Let (A, B) be a separation of X. Let x ∈ A and y ∈ B. By hypothesis,
there exists a connected set E containing both x and y. By Proposition 5.13, either
E ⊆ A or E ⊆ B. In particular, {x, y} ⊆ A or {x, y} ⊆ B. It is a contradiction to the
fact that A and B are disjoint and the way x, y are chosen.
Now, we state and prove the promised result.
Theorem 5.15. (Chain theorem) Suppose (En ) is a sequence of connectedSsubsets

of a topological space X such that En ∩ En+1 ̸= ∅ for every n ∈ N. Then ∞
n=1 En
connected.
Proof. By Theorem 5.14, it is enough

S∞ to show that every pair of points in X is contained
in a connected subset of E := n=1 En . So, let x, y ∈ E. Then there exists nx , ny ∈ N
such that x ∈ Enx and y ∈ Eny . Let N = max{nx , ny }. Then
N
[
{x, y} ⊆ En .
n=1
By Theorem 5.12, N
S
n=1 En is connected. Thus, we have shown that there is a connected
set containing {x, y}, completing the proof.
59
Theorem 5.16. (Intermediate value theorem) Let X be a connected topological
space and f : X → R be a continuous function (with usual topology on R). Suppose
x1 , x2 ∈ X be such that f (x1 ) < f (x2 ). Then, for every γ ∈ R such that
f (x1 ) < γ < f (x2 ),
there exists x ∈ X such that

f (x) = γ.
Proof. Let γ ∈ R be such that f (x1 ) < γ < f (x2 ). Suppose there is no x ∈ X such
that f (x) = γ. Then we see that
X = {x ∈ X : f (x) < γ} ∪ {x ∈ X : f (x) > γ}.
Since f is continuous, the sets A = {x ∈ X : f (x) < γ} and B = {x ∈ X : f (x) > γ}

are open sets. Clearly,
x1 ∈ A, x2 ∈ B, A ∩ B = ∅.
Thus, we arrive at a contradiction to the fact that X is connected.
We recall from real Analysis that a subset of R is connected iff it is an interval.

Now, we prove an analogous result in Rn . First, let us recall that a subset E of Rn is
convex if the line segment joining any tow points in E is contained in E, that is, for
every x, y ∈ E,
{(1 − t)x + ty : 0 ≤ t ≤ 1} ⊆ E.
Theorem 5.17. Every convex subset of Rn is connected with respect to the usual
topology on Rn .
Proof. Let E be a convex subset of Rn . By Theorem 5.14, it is enough to prove that

every pair of points in E is contained in a connected subset of E. So, let x, y ∈ E. Let
f : [0, 1] → X be defined by
f (t) = (1 − t)x + ty, t ∈ [0, 1].
Then f is continuous and [0, 1] is connected, its range, namely, the set
f ([0, 1]) = {(1 − t)x + ty : 0 ≤ t ≤ 1}
is connected. Since E is convex, {(1 − t)x + ty : 0 ≤ t ≤ 1} ⊆ E. Thus, f ([0, 1]) is a

connected subset of E containing x, y.
Theorem 5.18. Closure of a connected subset is connected.
60
Proof. Let E be a connected subset of a topological space X. We have to show that
Ē is connected. Assume, on the contrary that Ē is not connected. Let (A, B) be a
separation of Ē. Hence,
E ⊆ Ē = A ∪ B = (Ē ∩ G1 ) ∪ (Ē ∩ G2 )
for some open sets G1 and G2 . Hence,
E = (E ∩ G1 ) ∪ (E ∩ G2 ). (∗)
Note that
Ē ∩ G1 ̸= ∅ ⇒ E ∩ G1 ̸= ∅, Ē ∩ G2 ̸= ∅ ⇒ E ∩ G2 ̸= ∅.
(If x ∈ Ē ∩ G1 , then the open set G1 which contains a closure pint x of E must
contain some point of E.) Thus, (∗) shows that E has a separation, contradicting the
connectedness of E.
More generally, we have the following.
Theorem 5.19. Suppose A is a connected subset of a topological space X and B ⊆ X

is such that
A ⊆ B ⊆ Ā.
Then B is connected.
Proof. Suppose B is not connected. Let (B1 , B2 ) be a separation of B. Then there

exist open sets G1 and G2 in X such that
B = (B ∩ G1 ) ∪ (B ∩ G2 ),
with B ∩ G1 and B ∩ G2 are nonempty and disjoint. This implies
A = (A ∩ G1 ) ∪ (A ∩ G2 ),
Note that
B ∩ G1 ̸= ∅ ⇒ Ā ∩ G1 ̸= ∅ ⇒A ∩ G1 ̸= ∅.
Similarly,
B ∩ G2 ̸= ∅ ⇒ Ā ∩ G2 ̸= ∅ ⇒A ∩ G2 ̸= ∅.
Hence, (A ∩ G1 , A ∩ G2 ) is s separation of A, contradicting the connectedness of A.
Theorem 5.20. A topological space is disconnected iff there exists a continuous func-
tion from X onto a two-point set with discrete topology.
61
Proof. Let X be a topological space. Suppose X is disconnected. Then there exists
disjoint nonempty open subsets A and B such that X = A ∪ B. Let {0, 1} be with
discrete topology and f : X → {0, 1} be defined by

0, x ∈ A,
f (x) =
1, x ∈ B.
Then f is onto. Since A and B are also closed sets, by pasting lemma, f is continuous.
Conversely, suppose f : X → {0, 1} be a surjective continuous function, where
{0, 1} is endowed with discrete topology. Let
A = {x ∈ X : f (x) = 0} and B = {x ∈ X : f (x) = 1}.
Then A and B are nonempty open sets and X = A ∪ B,. Hence, X is disconnected.
5.2 Totally disconnected space
Definition 5.21. A topological space is said to be totally disconnected if singleton

sets are its only connected subsets. ♢
Example 5.22. Every discrete space is totally disconnected. Indeed, if X is a discrete
space and E ⊆ X contains more than one elements, then for every x ∈ E, ({x}, E \{x})
is a separation of E. ♢
Example 5.23. The topological spaces Q and R \ Q are totally disconnected with
respect to the induced usual topology. To see that Q is totally connected, let E ⊆ Q
contains more than one elements. Let x, y ∈ E be such that x < y, and let r ∈ R \ Q
be such that x < r < y. Let G1 = (−∞, r) and G2 = (r, ∞). Then we have
E = (E ∩ G1 ) ∪ (E ∩ G2 ).
Note that E∩G1 and E∩G2 are disjoint open subsets E containing x and y, respectively.
Hence, E is disconnected.
Analogously, it can be shown that every subset of R \ Q containing more than one
elements is disconnected. ♢
5.3 Path connectedness
For proving that every convex subset of Rn is connected, what we have actually done
was that every pair of points in Rn can be joined by a straight line segment and used
theorem Theorems 5.8 and 5.14. Similar idea can be used prove connectedness of more
general sets in topological spaces, as the following theorem shows.
62
Theorem 5.24. Let X be a topological space such that for every x, y ∈ X, there exists
a continuous function f : [a, b] → X such that f (a) = x and f (b) = y. Then X is
connected.
Proof. Under the given assumption, we have to show that X is connected. By Theorem
5.14, it is enough to show that every pair of points in X is contained in a connected
subset of X. So, let x, y ∈ X. By the assumption on X, there exists a continuous
function f : [a, b] → X for some a, b ∈ R with a < b such that f (a) = x and f (b) = y.
Since [a, b] is connected and f is continuous, the set f ([a, b]) is connected in X with
x = f (a) ∈ f ([a, b]) and y = f (b) ∈ f ([a, b]), and the proof is complete.
Definition 5.25. Let X be a topological space and x, y ∈ X. By a path from x to y,
we mean, a continuous function f : [a, b] → X such that f (a) = x and f (b) = y, and
in that case, we say that x and y are joined by a path. ♢
We observe that:
• If f : [a, b] → X is a path from x to y, then g : [a, b] → X defined by

g(t) = f (a + b − t), t ∈ [a, b],
is a path from y to x.
Definition 5.26. A topological space X is said to be a path connected space if any
pair of points in X can be joined by a path in X. ♢
In view of the above two definitions, Theorem 5.24 can be re-stated as follows.
Theorem 5.27. Every path connected space is connected.
Exercise 5.28. Every convex subset of Rn is path connected with respect to the usual
topology. ♢
• Let x, y be in a topological space X. If f : [a, b] → X is a continuous function

such that f (a) = x and f (b) = y, then for any closed interval [c, d], there is a
continuous function g : [c, d] → X such that g(c) = x and g(d) = y. This is seen
as follows:
Consider the function h : [c, d] → [a, b] defined by
b−a
h(t) = a + (t − c), t ∈ [c, d].
d−c
Then, h is a continuous function such that h(c) = a and h(d) = b. Now, consider
the function g := f ◦ h, that is, g : [c, d] → X defined by
g(t) = f (h(t)), t ∈ [c, d].
63
Then g is continuous and it satisfies
g(c) = f (h(c)) = f (a) = x and g(d) = f (h(d)) = f (b) = y.
Thus, we have shown the following.
Propositon 5.29. Let x, y be in a topological space X. If f : [a, b] → X is a path from

x to y and if h : [c, d] → [a, b] defined by
b−a
h(t) = a + (t − c), t ∈ [c, d],
d−c
then g : [c, d] → X defined by
g(t) = f (h(t)), t ∈ [c, d],
is also a path from x to y.
Propositon 5.30. Let X be a topological space and x, y, z ∈ X. If there is a path from

x to y and a path from y to z, then there is a path from x to z.
Proof. Let f : [a, b] → X and g : [a, b] → X be paths from x to y and y to z,

respectively. Then, by Proposition 5.29, there are paths
f1 : [0, 1] → X and f2 : [1, 2] → X
from x to y and y to z, respectively. Next, define g : [0, 2] → X by

f1 (t) if t ∈ [0, 1],
g(t) =
f2 (t) if t ∈ [1, 2].
Since f1 and f2 are continuous functions defined on closed subsets of R (with respect
to the usual topology) and f1 (1) = f2 (1) = y, by pasting lemma, g is continuous. Also,
we have g(0) = x and g(2) = z. Thus, g is a path from x to z.
• Converse of Theorem 5.27 is not true, as the following example shows.
Example 5.31. (See Munkres: Example 6, Sec. 2.4, Chapter 3.)

Let X = R2 with usual topology, and let
E = {(x, sin(1/x)) : 0 < x ≤ 1}.
Note that E is connected, as it is the continuous image of an open interval. Hence, we

know that its closure Ē is connected. We show that Ē is not path connected.
64
Note that Ē = E ∪ F , where F = {(x, y) ∈ R2 : x = 0, −1 ≤ y ≤ 1}. We show
that there is no path from the origin to any point E. For this, assume on the contrary
that there is a continuous function f : [a, b] → Ē such f (a) is the origin and f (b) is a
particular point in E. Consider the the set
J = {t ∈ [a, b] : f (t) ∈ F }.
Since J is closed and bounded, it has a maximum in J, say c = max J. Now, consider
the restriction map f : [c, b] → Ē. Note that f (c) ∈ F and f (b) ∈ E. Let us assume,
without loss of generality, that [c, b] = [0, 1], and write
f (t) = (x(t), y(t)), 0 ≤ t ≤ 1.
Then, we have x(0) = 0 and for 0 < t ≤ 1, y(t) = sin(1/x(t)). Our idea is to find a
sequence (tn ) in the open interval (0, 1) such that tn → 0, but (x(tn )) does not converge,
which would contradict the continuity of f at 0. For this, for n ∈ N, let sn such that
0 < sn < x(1/n) and sin(1/sn ) = (−1)n . Since x(·) is continuous, there exists tn such
that
1
0 < tn < and x(tn ) = sn .
n
Thus, we have tn → 0 but (y(tn )) does not converge. ♢
Remark 5.32. The set Ē in Example 5.31 is called the topologist’s sine curve. ♢
Theorem 5.33. Let A and B be path connected subsets of a topological space such that
A ∩ B ̸= ∅. Then A ∪ B is path connected in X.
Proof. Let x, y ∈ A ∪ B. Suppose both x and y are either in A or in B. Since A are

path connected, it follows that there is a path joining x and y.
Next, suppose that x ∈ A and y ∈ B. Let z ∈ A ∩ B. Since x, z ∈ A, there is a
path from x to z, and since z, y ∈ B, there is a pat from z to y. Therefore, there is a
path from x to y.
T {Eα : α ∈ Λ} be a familySof path connected subsets of a topological

Corollary 5.34. Let
space X such that α∈Λ Eα ̸= ∅. Then E := α∈Λ Eα is path connected.
T S
Proof. Let x0 ∈ α∈Λ Eα , and let x, y ∈ E := α∈Λ Eα . Let β, γ ∈ Λ be such that
x ∈ Aβ and y ∈ Aγ . Since x0 , x ∈ Aβ and x0 , y ∈ Aγ , and since Aβ and Aγ are path
connected, there is a path in Aβ from x to x0 and there is a path in Aγ from x0 to y.
In particular, there is a path in E from x to x0 and there is a path in E from x0 to y.
Hence, there is a path in E from x to y.
The following two theorems are also immediate consequences of Theorem 5.33, and
the details of their proofs are left as exercises.
65
Theorem 5.35. Let E0 , E1 , . . . , En be path connected
Sn subsets of a topological space X
such that Ei−1 ∩ Ei ̸= ∅ for i ∈ {1, . . . , n}. Then i=0 Ei is path connected.
Theorem 5.36. Let (En ) be a sequence of path connected

S∞ subsets of a topological space
X such that En ∩ En+1 ̸= ∅ for every n ∈ N. Then i=1 Ei is path connected.
We have seen that R \ Q is totally disconnected with respect to the usual topology
(see Example 5.23). What about its counter part in R2 ? However, the set
E = {(x1 , x2 ) ∈ R2 : x1 , x2 ∈ R \ Q}
is connected; in fact, path connected. Here are the details:
Let x, y ∈ E with x ̸= y. Consider the straight line segment in R2 joining x

and y and the straight line L perpendicular to it. for each z ∈ L, consider
the path fz : [0, 1] → R2 joining x to y consisting of the straight line
segments joining x to z and z to y. Clearly, {fz : z ∈ L} is an uncountable
family of paths joining x to y. Since Q is countable, there is at least one
path, say fz0 for some zo ∈ L such that its range lies entirely in E.
Thus, we have shown that for every pair of points x, y ∈ E, there exists a path in E
joining x and y. In other words, E is path connected.
6 Compact spaces
6.1 Definition and implications
Definition 6.1. Let X be a topological space and E ⊆ X. A collection G of subsets

of X is called a cover of E if E is contained in the union of members of G. A cover
which consists of open sets in X is called an open cover of E. ♢
Definition 6.2. A topological space X is said to be a compact space if every open

cover of X contains a finite subcollection which is also a cover of X. ♢
Definition 6.3. A subset of a topological space is said to be compact if it is a compact

space as a subspace. ♢
Theorem 6.4. Let X be a topological space and E ⊆ X. Then E is a compact set iff
for every family of open sets in X which covers E contains a finite subcollection which
also covers E.
66
Proof. Suppose E is compact. Let {Gα : α ∈ Λ} be a family of open sets in X which
covers E, that is, [
E⊆ Gα .
α∈Λ
Then [ [
E=E∩ Gα = (E ∩ Gα ).
α∈Λ α∈Λ
Thus, {E ∩ Gα : α ∈ Λ} be a family of open sets in E which covers E. Since E is

compact, there exists α1 , . . . , αn in Λ such that
n
[
E= (E ∩ Gαi ).
i=1
Thus, E ⊆ ni=1 Gαi . Thus, we have proved that every family of open sets in X which
S
covers E contains a finite subcollection which also covers E.
Conversely, suppose every family of open sets in X which covers E contains a finite
subcollection which also covers E. We have to prove that E is compact. For this, let
{Vα : α ∈ Λ} be a family of open sets in X which covers E, that is,
[
E= Vα .
α∈Λ
Since each Vα is open in E, there exists open set Gα in X such that Vα = E ∩ Gα .

Thus, we have [ [ [
E= Vα = (E ∩ Gα ) ⊆ Gα .
α∈Λ α∈Λ α∈Λ
Sn
By assumption, there exists α1 , . . . , αn in Λ such that E ⊆ i=1 Gαi . Thus,
n
[ n
[ n
[
E=E∩ Gαi = (E ∩ Gαi ) = Vαi .
i=1 i=1 i=1
Thus, we have proved that every family of open sets in E which covers E contains a
finite subcollection which also covers E, showing that E is compact.
The following can be verified easily (Exercise):
• Every finite subset of a topological space is compact.
• The set Z is not compact in R with respect to the usual topology.
• With respect to the usual topology on R, the set E = {1/n : n ∈ N} is not

compact, whereas the set E ∪ {0} is compact.
67
• With respect to the usual topology on R, every open interval is non-compact.
Example 6.5. Every co-finite topological space is compact: To see this, let X be a
co-finite topological space and let G be an open cover of X. If X ∈ G, then there is
nothing to be proved. So, suppose that X ̸∈ G. Let G0 ∈ G. Then Gc0 is a finite set,
say Gc0 = {x1 , . . . , n0 } for some n0 ∈ N. Let G1 , . . . , Gn in G be such that xj ∈ Gj for
j = 1, . . . , n0 so that
n0
[
c
G0 = {x1 , . . . , n0 } ⊆ Gj .
j=1
Then we have n0
[
X = G0 ∪ Gc0 ⊆ G0 ∪ Gi .
j=1
Thus, we have shown that G has a finite subcollection which covers X. ♢

Theorem 6.6. Continuous image of a compact set is compact.
Proof. Let X and Y be topological spaces, f : X → Y be a continuous function, and

E be a compact set in X. To show that f (E) is compact. For this, let G be an open
cover of f (E). Then we see that
A = {f −1 (G) : G ∈ G}
is an open cover on E. Since E is compact, there exists G1 , . . . , Gn in G such that
n
[
E⊆ f −1 (Gi ).
i=1
This implies that

n
[
f (E) ⊆ Gi .
i=1
Thus, the open cover G of f (E) contains a finite sub-collection which also covers f (E).
Hence, f (E) is compact.
Corollary 6.7. Let X be a topological space and R be with usual topology. Let E be
a compact subset of X and f : X → R be a continuous function. Then, there exists
x1 , x2 ∈ E such that
f (x1 ) ≤ f (x) ≤ f (x2 ) ∀ x ∈ E.
Proof. Since X is compact, by Theorem 6.6, f (E) is a compact subset of R. In patic-

ular, f (E) is closed and bounded, and hence there exists u, v ∈ f (E) such that
u = inf f (E), v = sup f (E).
Since u, v ∈ f (E), there exists x1 , x2 ∈ E such that f (x1 ) = u and f (x2 ) = v, which
completes the proof.
68
Theorem 6.8. Closed subset of a compact space is compact.
Proof. Let X be a compact topological space and C be a closed subset of X. To show

that C is compact. For this, let G be an open cover of C. Since G0 := X \ C is open,
{G0 } ∪ {G : G ∈ G}
is an open cover of X. Since X is compact, there exists G1 , . . . , Gn in G such that

n
[
X = G0 ∪ Gi .
i=1
Hence,
n
[ c
Gc0 ∩ Gi = ∅,
i=1
so that n
[
C= Gc0 ⊆ Gi .
i=1
Thus, the open cover G of C contains a finite sub-collection which also covers C,
showing that C is compact.
• Is compact subset of every topological space closed?
The answer is in the negative. Here are a few examples:
1. In every indiscrete space with more than one elements, singleton sets are compact,
but not closed.
2. On the set X = {a, b, c}, consider the topology T = {∅, {a}, {a, b, c}}. We
observe that the set {b} is compact, but not closed.
3. On the set R, consider the topology T := {(a, ∞) : a ∈ R} ∪ {∅, R}. With

respect to this topology, singleton sets are compact, but not closed.
4. For x = (x1 , x2 ) ∈ R2 adn r > 0, let
Br (x) = {(y1 , y2 ) ∈ R2 : |x2 − y2 | < r}.
Let T be the family of all subsets A of R2 such that, for every x ∈ A, there exists
r > 0 such that Br (x) ⊆ A. Then, T is a topology on R2 (Verify). With respect
to this topology, singleton sets are compact, but they are not closed.
69
The topological spaces in the above listed examples are not Hausdorff. This raises
the following question:
• Do we have the affirmative answer to the last question if the topological space is
Hausdorff?
Answer to this question is in the affirmative, as the following theorem shows.
Theorem 6.9. Compact subset of a Hausdorff space is closed.
Proof. Let X be a Hausdorff space and K be a compact subset of X. To show that K

is closed. That is to show that K c is open. For this, it is enough to show that every
point in K c is an interior point. So, let x ∈ K c . Since X is Hausdorff, for every y ∈ K,
there exist open sets Uy , Vy such that
x ∈ Uy , y ∈ Vy , Uy ∩ Vy = ∅.
Note that {Vy : y ∈ K} is an open cover of K. Since K is compact, there exists

y1 , . . . , yn in K such that
[n
K ⊆ V := Vyi .
i=1
Then, we have
n
\
x ∈ U := Uyi .
i=1
Since Uyi ∩ Vyi = ∅, we have U ∩ Vyi = ∅. This is true for each i = 1, . . . , n. Hence,
U ∩V = ∅. This implies U ∩K = ∅ so that U ⊆ K c . Thus, for each x ∈ K c , we obtain
an open set containing x and contained in K c , showing that K c is an open set.
Corollary 6.10. Every compact subset of a metrizable space is closed.
While proving Theorem 6.9, we have actually proved the following theorem.
Theorem 6.11. Let X be a Hausdorff space, K be a compact set and x ∈ X such that
x ̸∈ K. Then there exist disjoint open sets containing x and K, respectively.
Theorem 6.12. If K1 and K2 are two disjoint compact sets in a Hausdorff space X,
then there exist disjoint open sets containing K1 and K2 , respectively.
Proof. By Theorem 6.11, for each x ∈ K1 , there exist disjoint open sets Ux and Vx such
that
x ∈ Ux , K2 ⊆ Vx .
70
T hen{Ux : x ∈ K1 } is an open cover of K1 . Since K1 is compact, there exist x1 , . . . , xn
in K1 such that
n
[
K1 ⊆ U := Uxi .
i=1
Tn
Let V = i=1 Vxi . Then U and V are disjoint open sets containing K1 and K2 , respec-
tively.
Theorem 6.13. Every compact subset of a metric space is bounded.
Proof. Let Ω be a metric space and E be a compact subset of Ω. For any given r > 0,
we have [
E⊆ B(x, r).
x∈E
Since E is compact, there exists x1 , . . . , xn in E such that

[
E⊆ B(xi , r).
i=1
Hence, for any x, y ∈ E, there exists k, ℓ ∈ {1, . . . , n} such that
x ∈ B(xk , r), y ∈ B(xℓ , r).
Hence,
d(x, y) ≤ d(x, xk ) + d(xk , xℓ ) + d(xℓ , y) ≤ r + d(xk , xℓ ) + r
Hence, d(x, y) ≤ 2r + max{d(xi , xj ) : i, j ∈ {1, . . . , n}}. Hence, E is bounded.
By Corollary 6.10 and Theorem 6.13, every compact subset of a metric space is
closed and bounded. However, a closed and bounded subset of a metric space need not
be compact. A simple example is the discrete metric space on an infinite set. Here is
another simple example.
Example 6.14. Let Ω = (0, 2] with usual metric. Then the set E = (0, 1] is closed
and bounded, but not compact. ♢
Theorem 6.15. Let X be a topological space and X0 be a subspace of X. Let E ⊆ X0 .

Then, E is compact in X0 iff E is compact in X.
Proof. Suppose E is compact in X0 . We show that E is compact in X. For this, let G

be a family of open sets in X that covers E. Let
G0 := {G ∩ X0 : G ∈ G}.
71
Then G0 is a family of open sets in X0 that covers E. By hypothesis, there exists
a finite subcollection of G0 that covers E. This finite subcollection is of the form
{G1 ∩ X0 , . . . , Gn ∩ X0 } for some G1 , . . . , Gn in G. Hence,
n
[ n
[
E⊆ (Gi ∩ X0 ) ⊆ Gi .
i=1 i=1
Thus, we obtained a finite subcollection of G that covers E.

Conversely, suppose E is compact in X. We show that E is compact in X0 . For
this, let V = {Vα : α ∈ Λ} be a family of open sets in X0 that covers E. We know that
for each α ∈ Λ, there exists an open set Gα in X such that Vα = Gα ∩ X0 . Then, we
see that, {Gα : α ∈ Λ} is a family of open sets inS X that covers E. Since E is compact
in X, there exist α1 , . . . , αn in Λ such that E ⊆ ni=1 Gαi . Hence,
n
[ n
[
E⊆ (Gαi ∩ X0 ) = Vαi .
i=1 i=1
Thus, we have proved that the cover V consisting of open sets in X0 has a finite
subscollection that also covers E. Consequently, E is compact in X0 .
We have seen examples of bijective continuous functions (see, e.g., Example 4.44
and Example 4.46) which are not homeomorphisms. However, if the domain space is
compact and the co-domain space is Hausdorff, then such maps are homeomorphisms.
More precisely, we hae the following theorem.
Theorem 6.16. Every bijective continuous function from a compact space onto a Haus-
dorff space is a homeomorphism.
Proof. Let X be a compact space and Y be a Hausdorff space, and let f : X → Y be

a bijective continuous function. It is enough to show that f is a closed map. For this,
let C be a closed set in X. Then, by Theorem 6.8, C is clompact in X. Hence, by
Theorem 6.6, f (C) is compact. Since X is Hausdorff, by Theorem 6.9, f (C) is closed
in Y . This completes the proof.
An application of Theorem 6.16
Let X and Y be topological spaces and F : X → Y be a continuous function.

Suppose for every y ∈ Y , there exists a unique x ∈ X such that
F (x) = y. (∗)
Now, suppose (yn ) is a sequence in Y such that yn → y for some y ∈ Y . For each
n ∈ N, let xn ∈ X be such that F (xn ) = yn . One would like to know whether xn → x.
72
Theorem 6.16 shows that the answer is in the affirmative if X is compact and Y is
Hausdorff.
If X and Y are metric spaces with metrics d and ρ, respectively, with X compact,
and if F : X → Y is a continuous function, then the above consideration shows that
small error in y can lead only lead to small error in the solution x. In other words, the
solution of the equation (∗) depends continuously on the data y.
Finite intersection property

Theorem 6.17. (Finite intersection property) Let A := {Kα : α ∈ Λ} be a family
of non-empty compact subsets of a Hausdorff topological
T space X such that intersection
of any finite subcollection of A is nonempty. Then α∈Λ Kα is non-empty.
T
Proof. Suppose α∈Λ Kα = ∅. Let α0 ∈ Λ and Λ0 = Λ \ {α0 }. Then we have
\ c ∞
[
K α0 ⊆ Kα = Kαc .
α∈Λ0 α∈Λ0
Since X is Hausdorff, Kα is closed for every α ∈ Λ, so that the above inclusion shows
that {Kαc : α ∈ Λ0 } is an open cover of the compact set Kα0 . Hence, there exist
α1 , . . . , αN in Λ0 such that
N
[ N
\ c
K α0 ⊆ Kαc i = K αi .
i=1 i=1
Hence,
N
\
K α0 ∩ Kαi = ∅.
i=1
This is a contradiction to the fact that intersection of any finite subcollection of A is

nonempty.
Corollary 6.18. Let (Kn ) be a sequence of non-empty compactTsubsets of a Hausdorff
topological space X such that Kn ⊇ Kn+1 for all n ∈ N. Then ∞n=1 Kn is non-empty.
Remark 6.19. We may recall from real analysis that if (Fn ) is a sequence of non-
emptyTclosed sets in a complete metric space X such that Fn ⊇ Fn+1 for all n ∈ N,
then ∞n=1 Fn ̸= ∅. ♢
Regular and normal spaces
Let X be a Hausdorff space. We have seen, in Theorem 6.11, that a point in X and
a compact set K which does not contain x can be separated by disjoint open sets, and
73
in Theorem 6.12, every pair of disjoint compact sets in X can be separated by disjoint
open sets. Spaces having these properties with compactness replaced by closedness
are called T3 -spaces and T4 -spaces, respectively. More precisely, we have the following
definitions.
Definition 6.20. A topological space X is said to be a regular space if for every

closed set F and a point x ̸∈ F , there exist disjoint open sets U and V such that F ⊆ U
and x ∈ V . ♢
Definition 6.21. A topological space X is said to be a normal space if for every

pair of disjoint closed sets F1 and F2 , there exist disjoint open sets U and V such that
F1 ⊆ U and F2 ⊆ V . ♢
Recall that, in a topological space X, if every pair of distinct points can be separated
by disjoint open sets, then the space is called a Hausdorff space. A topological space
satisfying a slightly weaker notion than Hausdorff condition is the so called T1 -space.
Definition 6.22. A topological space X is said to satisfy T1 -axiom, if for every pair
of distinct points x and y in X, there there exist open sets U and V such that x ∈ U \V
and y ∈ V \ U , and in that case X is called a T1 -space. ♢
We know that a cofinite-topological space on an infinite set and a co-countable

topological space on an uncountable set are not T2 spaces. However, it can be easily
shown that they are T1 spaces.
• A Hausdorff space is also called a T2 -space.
• A regular Hausdorff space is called a T3 -space.
• A normal Hausdorff space is called a T4 -space.
Definition 6.23. A topological space X is said to be a completely regular space if

for every closed set F and a point x ̸∈ F , there exists a continuous function f : X → R
such that
f (x) = 1 and f (y) = 0 ∀ y ∈ F.
A completely regular space which is also Hausdorff is called a T3 1 -space. ♢
2
It can be shown that
T4 ⇒ T3 1 ⇒ T3 ⇒ T2 ⇒ T1 ,
2
and the reverse implications need not hold.
74
6.2 Comapct subsets of R
We assume throughout this subsection that R is with usual topology.
Theorem 6.24. For a, b ∈ R with a < b, the closed interval [a, b] is compact.
Proof. Let G be an open covering of [a, b]. Let
S = {y ∈ [a, b] : [a, y] is covered by a finite subcollection of G}.
Note that a ∈ S and S is bounded above. Hence,
β := sup S < ∞.
We first show that β ∈ S. For this, first observe that there exists G ∈ G such that
β ∈ G. Hence, there exists α ∈ [a, β) such that (α, β] ⊆ G. Therefore, by the definition
of supremum, there exists y ∈ S such that
α < y ≤ β.
Since
[a, β] = [a, y] ∪ [y, β],
where y ∈ S and [y, β] ⊆ G, it follows that β ∈ S.
Now, it is enough to show that β = b. To show this, assume the contrary that
β < b. Let G0 ∈ G be such that β ∈ G0 , and let t ∈ [β, b) be such that [β, t] ⊆ G0 .
Since
[a, t] = [a, β] ∪ [β, t],
it follows that t ∈ S. This is a contradiction to the fact that β = sup S.
Theorem 6.25. Every closed and bounded subset of R is compact.
Proof. Let E be a closed and bounded subset of R. Since E is bounded, there exists
E ⊆ [a, b] for some a, b ∈ R with a < b. Since [a, b] is compact, and since closed subset
of a compact set is compact, we obtain that E is compact.
We already know that every compact subset of a Hausdorff space is closed. Also,
every compact subset of a metric space is bounded. Hence, we obtain the following.
Theorem 6.26. A subset of E of R is compact with respect to the usual topology iff it
is closed and bounded.
Remark 6.27. We shall see that, a subset of E of Rn is compact (with respect to the
usual topology) iff it is closed and bounded. ♢
75
6.3 Locally compact spaces
Definition 6.28. A topological space X is said to be a locally compact space if

every point in X has an open neighbourhood contained in a compact subset of X. ♢
• A topological space X is locally compact iff for every x ∈ X, there exists open
set V and a compact set K such that
x ∈ V ⊆ K.
Here is a characterization of locally compact spaces.

Theorem 6.29. Let X be a Hausdorff space. The X is locally compact iff every point
in X has an open neighbourhood whose closure is compact.
Proof. Suppose X is locally compact and x ∈ X. Suppose V is an open set and K is a

compact set such that x ∈ V ⊆ K. Since X is Hausdorff, K is closed so that V̄ ⊆ K.
Being a closed subset of a compact set, V̄ is compact.
The converse is obvious.
Example 6.30. The following statements can be easily verified (Exercise).
1. Every compact space is locally compact.
2. Every discrete space is locally compact.
3. With respect to the usual topology, the spaces R, Z, N are locally compact, but
not compact.
4. With respect to the usual topology, every nonempty open subset of Rn is locally
compact, but not compact. ♢
Example 6.31. With respect to the usual topology, the spaces Q and R\Q are neither
compact nor locally compact. First let us show that Q is not locally compact. To see
this, let x ∈ Q. Suppose there exists an open set V in Q such that x ∈ V and clQ (V)
is compact in Q. Let G be an open set in R such that V = G ∩ Q, and let a, b ∈ R \ Q
such that a < b and
x ∈ (a, b) ∩ Q ⊆ G ∩ Q ⊆ V.
Note that U := (a, b) ∩ Q is a closed subset of Q and clQ (U) ⊆ clQ (V). Hence,
clQ (U) = U is compact in Q. Hence, U is compact in R as well, which is not true, as
compact subset of R are closed and bounded in R, but U is not closed in R. ♢
The conclusion in the above example also follows from the following general result.
76
Theorem 6.32. Let X be a Hausdorff space and X0 be a locally compact dense subspace
of X. Then X0 is open in X.
Proof. Let x ∈ X0 . We show that there exists an open subset of X containing x and
contained in X0 , that will complete the proof.
Since X0 is locally compact, there exists a compact set K ⊆ X0 and an open set V
in X0 such that
x ∈ V ⊆ K.
Since V is open in X0 , there exists an open set G in X such that V = G ∩ X0 . We
show that G ⊆ X0 . In fact, we show that G ⊆ G ∩ X0 so that
G ⊆ G ∩ X0 = V̄ ⊆ K̄ = K,
where the last equality follows, since K is a compact subset of a Hausdorff space. For
showing G ⊆ G ∩ X0 , let y ∈ G and let H be an open subset of X containing y. Since
H ∩ G is open set containing y and X0 is denset in X, we have H ∩ G ∩ X0 ̸= ∅.
That is, the open neighbourhood H of y contains some point from G ∩ X0 . Hence,
y ∈ G ∩ X0 .
By the above theorem, we have the following:
• The dense subspaces Q and R \ Q of R (with respect to the usual topology) are
not open in R, and hence are not locally compact.
6.4 One-point compactification
We know that the topological space X = (0, 1] with usual metric is a locally compact
Hausdorff space which is not compact. Now, if we add one point to it, we obtain a
compact space Y = [0, 1] with the usual topology. Note that X is dense in Y .
Given a locally compact Hausdorff space X which is not a compact space, can we
obtain a compact space in which X is a dense subspace?
For example, consider the space X = (0, 1), which is a locally compact Hausdorff
space, which is not compact, with respect to the with usual topology. Let us consider
the set
Y = X ∪ {p}, p := {0, 1}.
Consider the topology on Y for which the base consists of all open sub-intervals of
(0, 1) together with sets of the form Y \ [a, b] for a, b ∈ (0, 1). Then, it can be seen
that, with this topology, Y is a compact Hausdorff space. Also, since every set of the
form Y \ [a, b] for a, b ∈ (0, 1) contains points from X, X is a dense subspace of Y .
77
The above method can be adopted to obtain a compact Hausdorff space out of a
ny locally compact Hausudorff space, as described in the following theorem.
Theorem 6.33. (One-point compactification) Let (X, TX ) be a locally compact
Hausdorff space which is not compact. Then there exists a compact Hausdorff space
Y ⊇ X such that Y \ X is a singleton set, X is a subspace of Y and X̄ = Y .
Let Y = X ∪ {p}, where p ̸∈ X, and let TY be the family of all those subsets A of
Y such that either A is open in X or A = Y \ K for some compact subset K ⊆ X.
Note that for a compact set K ⊆ X,
A = Y \ K ⇐⇒ A = {p} ∪ (X \ K).
Then TY is a topology on Y and (Y, TY ) is a compact Hausdorff space satisfying the

requirements as in Theorem 6.33.
Definition 6.34. The space Y in Theorem 6.33 is called the one-point compactifi-
cation of X. ♢
7 Product topology
7.1 On product of finite number of topological spaces
Let {(Xi , Ti ) : i = 1, . . . , n} be a family of non-empty topological spaces and and let

X be the cartesian product of X1 , . . . , Xn , that is,
X = X 1 × · · · × Xn .
We observe that the family
V := {V1 × · · · × Vn : Vi open in Ti , i = 1, . . . , n}
is closed under finite intersections. Hence, V is a base for a unique topology.

Definition 7.1. The topology for which
V := {V1 × · · · × Vn : Vi open in Ti , i = 1, . . . , n}
is a base is called the product

Qn topology on X = X1 × · · · × Xn of X1 , . . . , Xn . This
space is also denoted by i=1 Xi . ♢
• Let X1 , . . . , Xn be topological spaces. A subset A of X := ni=1 Xi is open with

Q
respect to the product topology on X iff for every x ∈ A, there exists open sets
Vi in Xi for i = 1, . . . , n such that x ∈ V1 × · · · × Vn ⊆ A.
78
The proof of the following theorem is obvious from the definition of a product topology.
Theorem 7.2. Let X1 , . . . , Xn be non-empty topological spaces and X be their product

space. Then, for each i = 1, . . . , n, the map πi : X → Xi defined by
πi (x) = xi , x := (x1 , . . . , xn ) ∈ X,
is continuous and surjective, and the product topology on X is the weakest topology
such that π1 , . . . πn are continuous.
Proof. Let us show the conclusions one by one.

(1) Each πi is surjective: Fix i ∈ {1, . . . , n} and let y ∈ Xi . For j ∈ {1, . . . , n}, let
xj ∈ Xj be such that xi = y. Then
πi (x1 , . . . , xn ) = y.
(2) Each πi is continuous: Fix i ∈ {1, . . . , n} and let G be open in Xi . Then
πi−1 (G) = V1 × · · · × Vn ,
where Vi = G and Vj = Xj for j ̸= i.

(3) The product topology on X is the weakest topology such that π1 , . . . πn are
continuous: Let T be any topology on X such that π1 , . . . πn are continuous. Then T
must contain sets of the form πi−1 (G), where G open in Xi and i ∈ {1, . . . , n}. Hence
the product topology on X is weaker than T .
Definition 7.3. The maps πi in the above theorem are called projection maps. ♢
Example 7.4. Let Xi =Q R for i = 1, . . . , n with usual topology. Then the product
topological space X := ni=1 Xi is nothing n
Qn but R with usual topology. For this
topology, the set of all sets of the from i=1 (ai , bi ) forms a base. ♢
Theorem 7.5. Product of a finite number of connected spaces is connected.
Proof. It is enough (why?) to show that product of two connected spaces is connected.
Let X and Y be connected topological spaces. We show that X × Y is connected.
For this, let us fix an element (x0 , y0 ) ∈ X × Y , and consider the subspaces X × {y0 }
and and for each x ∈ X, the subspace {x} × Y . Note that
S
1. X × Y = x∈X Zx , Zx := (X × {y0 }) ∪ ({x} × Y ),
2. (x, y0 ) ∈ (X × {y0 }) ∩ ({x} × Y ) ∀ x ∈ X,
79
3. (x0 , y0 ) ∈ Zx for every x ∈ X.
Now, since X × {y0 } and {x} × Y are homeomorphic with the connected spaces X
and Y , respectively, and (x, y0 ) is a common point in these spaces, the subspace Zx is
conencted for each x ∈ X. Also, since (x0 , y0 ) ∈ Zx for each x ∈ X, their union X × Y
is connected.
Theorem 7.6. (Special case of Tychnoff ’s theorem) Product of a finite number

of compact spaces is compact.
For its proof we shall make use of the following lemma, known as tube-lemma7 .
Lemma 7.7. (Tube lemma) Let X and Y be topological spaces and x0 ∈ X. If Y

is compact and if A is an open subset of X × Y containing the slice {x0 } × Y , then A
contains the tube G × Y , where G an open neighbourhood of x0 .
Proof of Theorem 7.6. It it enough to prove for the case of n = 2. So, let X and Y
be compact topological spaces, and X × Y be with product topology. Let G be an open
cover of X × Y . For each x ∈ X, since {x} × Y is compact, there exists G1 , . . . , Gk in
G such that
[k
{x} × Y ⊆ Gi .
i=1
By the tube-lemma, there exists a tube Vx × Y with Vx is an open neighbourhood of x

such that
[k
{x} × Y ⊆ Vx × Y ⊆ Gi .
i=1
Thus, Vx × Y is covered by a finite sub-collection of G. Now, the space {Vx : x ∈ X}

is an open covering of X. Since X is compact, there exists x1 , . . . , xℓ in X such that
ℓ
[
X⊆ Vxi .
i=1
Hence
ℓ
[
X ×Y = (Vxi × Y ),
i=1
where each Vxi × Y is covered by a finite number of members of G.

Thus, we have proved that the open cover G of X × Y contains a sub-collection
which also covers X × Y .
7
For its proof, see“Topology: A First Course” by J.R. Munkres
80
7.2 On product of infinite number of topological spaces
Next suppose we have a countably infinite collection of topological spaces, say (Xi , Ti )
for i ∈ N. Consider the cartesian product of X1 , X2 , . . ., that is, the set
∞
Y
X := Xi .
i=1
Q∞
• i=1 Xi is the set of all sequences of the form (x1 , x2 , . . .) with xn ∈ Xn for n ∈ N.
Q∞
For each n ∈ N, let πn : i=1 Xi → Xn defined by
πn (x) = xn , x := (x1 , x2 , . . .).
Let
S := {πn−1 (G) : G ∈ Tn , n ∈ N}.
Q∞
• The topology on i=1 Xi generated by S is the weakest topology such that each
πn is continuous.
Q∞
Definition 7.8. The topology on i=1 Xi generated by
S := {πn−1 (G) : G ∈ Tn , n ∈ N}
is called the product topology on ∞

Q
i=1 Xi . ♢
• Let V be the collection of all sets of the form ∞

Q
i=1 Vi , where each Vi is open in Xi
and Vi = Xi Qexcept for a finite number of i’s. Then V is a base for the product
topology on ∞ i=1 Xi .
Definition 7.9. Let Λ be a non-empty set and for S each α ∈ Λ, let Xα be a non-empty
set. Then the collection of all functions f : Λ → α∈Λ Xα such that f (α) ∈ Xα for each
α ∈ Λ is called
Q the cartesan productQof Xα , α ∈ Λ, and this cartesian peroduct is
denoted by α∈Λ Xα . An element f ∈ ∞ α∈Λ Vα is usually denoted by (xα ) or (xα )α∈Λ
with the understanding that xα = f (α) for α ∈ Λ. ♢
Now,Qsuppose that, for each α ∈ Λ, (Xα , Tα ) is a topological space. For each β ∈ Λ,

let πβ : α∈Λ Xα → Xβ be the function defined by
πβ (x) = xβ , x := (xα )α∈Λ .
This πβ is called a projection map onto the factor space Xβ .
81
Definition 7.10.
Q Let {(Xα , Tα ) : α ∈ Λ} be a family of topological spaces. Then the
topology on α∈Λ Xα generated by
S := {πβ−1 (G) : G ∈ Tβ , β ∈ Λ}
Q
is called the product topology on α∈Λ Xα . ♢
• The product topology on

Q
α∈Λ Xα is the weakest topology such that each πβ is
continuous.
• V be the collection of all sets of the form ∞

Q
α∈Λ Vα , where each Vα is open in Xα
and Vα = Xα except possible
Q for a finite number of α’s. Then V is a base for the
product topology on α∈Λ Xα .
Y 7.11. If {Xα : α ∈ Λ} is a family of connected spaces, then the product

Theorem
space Xα is connected.
α∈Λ
Theorem 7.12. (Tychonoff ’s Ytheorem) If {Xα : α ∈ Λ} is a family of compact

spaces, then the product space Xα is a compact space.
α∈Λ
Q
We may define another topology on α∈Λ Xα which seems more natural than the
product topology, the so called box topology.
Suppose that, for each α ∈ Λ,Q(Xα , Tα ) is a topological space, and let U be the
collectionQof all sets of the form ∞ α∈Λ Vα , where each Vα is open in Xα . Then U
∞
contains α∈Λ Xα andQit is closed under finite intersections, and hence U is a base for
a unique topology on α∈Λ Xα .
Q
Definition 7.13. The topology on α∈Λ Xα for which
nY o
U := Vα : Vα ∈ Tα
α∈Λ
Q
is a base is called the box topology on α∈Λ Xα . ♢
• Box topology is stronger than product topology.
• If Λ is a finite set, then product topology and box topology on the product set is
the same.
• Box topology on the cartesian product of an infinite number of connected spaces

need not be connected.
82
References
[1] J. R. Munkres, Topology, 2nd Edition, Pearson Edu. 3rd. Impression, 2007.
[2] K.D. Joshy, Introduction to General Toplogy (Revised), new Age International,
new Delhi, 1999.
[3] G.F. Simmons, Introduction to Topology and Modern Analysis, McGraw-Hill,

Tokyo, 1963.
[4] bibitemKP K. Parthasarathy, Topology: An Invitation, Springer, 2022.
[5] Sidney A. Morris, Topoolgy without Tears,

https://www.topologywithouttears.net/
83

Toplogy Notes

Uploaded by

Copyright:

Available Formats

Toplogy Notes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Toplogy Notes

Uploaded by

Copyright:

Available Formats

Topology: A Firt Course∗

2 Basic notions properties 12

3 Countability axioms and separability 29

4 Functions and their properties 36

5 Connected spaces and path connected spaces 55

2. Union of any arbitrary collection of members of G is a member of G;

3. Intersection of any two members of G is a member of G.

• A set A ⊆ Ω is closed iff Ac := Ω \ A ∈ G.

• A sequence (xn ) in Ω converges to x ∈ Ω iff for every A ∈ G with x ∈ A, there

• If S ⊆ Ω and x ∈ Ω, then x is a limit point (accumulation point) of S iff for

• If Ω̃ is another metric space (possibly different from Ω), then a function f : Ω → Ω̃

• A subset A of Ω is compact iff every open cover of A contains a finite sub-

• A subset A of Ω is connected iff A cannot be written as a union of two nonempty

Motivated by the above, we introduce the notion of a collection of open sets in an

1. The empty set ∅ and the whole set X are members of T ;

2. Intersection of any two members of T is a member of T .

3. Union of any collection of members of T is a member of T ;

A set X together with a topology T on it is called a topological space, and in that

Symbolically the conditions in the definition of a topological space can be written

TnT ) be a topological space. Show that, if Ai ∈ T for i = 1, . . . n

(i) T := 2X , the collection of all subsets of X, is a topology on X, called the

(ii) T := {∅, X} is a topology on X, called the indiscrete topology on X. With

• On a set X, indiscrete topology is the weakest topology and discrete topology is

d1 (x, y) = |x1 − y1 | + |x2 − y2 |,

respectivlely, for x, y ∈ R2 . Show that the topologies induced by d1 , d2 , d∞ are the

Let us first observe a simple fact:

Proof. Suppose A is open in X. Then, for every x ∈ A, we may take G = A.

S x ∈ A, there exists an open set Gx such that

• S is said to be closed underTfinite intersections if for any finite number of members

• S is said to be closedTunder arbitrary intersections if for any subcollection of S,

• T is closed under arbitrary unions and finite intersections;

An immediate consequence of the definition of a topology is the following:

(i) ∅ and X are members of C;

(ii) Union of any finite number of members of C is a member of C.

(iii) Intersection of any arbitrary collection of members of C is a member of C;

• Let C be as in the above theorem.Then C is closed under arbitrary intersections

Notation: Given a set X, if A is a collection of subsets of X, then we we may denote

for some set Λ, which is called an indexSset. Thus,Sif A =S

Note that if A is a finite set, then it can be written as {A1 , . . . , An } or {Ai : i ∈

(i) The discrete topology on X is induced by the discrete metric on X.

2 Basic notions properties

2.1 Hausdorff space

Definition 2.1. (Hausdorff space) A topological space X is said to be a Hausdorff2

• A topological space (X, T ) is Hausdorff iff for every x, y ∈ X, if x ̸= y, then there

• If a topology on X is induced by a metric, then X is a Hausdorff space.

3. The co-countable topology on X is Hausdorff iff X is a countable set, and in that

Definition 2.7. (Convergence in topological space) Let (X, T ) be a topological

Example 2.8. In an indiscrete topological space X, every sequence converges to every

Proof. Suppose X is a Hausdorff space and (xn ) is a convergent sequence in X. Suppose

Let n0 := max{n1 , n2 }. Then we have

• Since every metrizable space is Hausdorff, every convergent sequence in a metriz-

Proof. If X is a countable set, then co-countable topology on X is the discrete topology,

• A convergent sequence in a co-finite topological space need not be eventually

2.3 More on weak and strong topologies

If T1 and T2 are topologies on a set X, then T1 ∩ T2 is also a topology on X, which is

T := {(a, ∞) : a ∈ R} ∪ {∅, R}.

where a0 := inf S is either in R or it is −∞. Clearly,

• T is weaker than the usual topology on R. ♢