1

COURSE CONTENT

1. Topological Spaces (Definitions)

2. Bases and Subbases

3. Continuity and Topological Equivalence

4. Countability and Separation Axioms

5. Compactness

6. Connectedness

REQUIRED BOOKS

(1) General Topology by Seymour Lipschitz

(2) Introduction to Modern Analysis and Topology by Simon

(3) General Topology by J. L. Kelley

Chapter 1

TOPOLOGICAL SPACES

Definition 1.0.1. (Topology): Let X be a non-empty set. A collection (or class) J of

open subsets of X is a topology on X if J satisfies the following axioms:

(Y1) X, φ ∈ J .

(Y2) The union of any number of sets in J belongs to J . That is, if {Gi }i∈I is a
collection of elements of J , then ∪i∈I Gi ∈ J .

(Y3) The intersection of any two (or finite) sets in J belongs to J . That is, if
G1 , · · · , Gn ∈ J then ∩ni=1 Gi ∈ J .

Definition 1.0.2. ( J -Open Sets): The members of J are then called J -open sets or
open sets.

Definition 1.0.3. (Topological Space): The set X together with J , that is, the pair
(X, J) is called a topological space.

Example 1.0.4. (a) Indiscrete Topology: The collection J = {X, φ} is itself a

topology on X . it is called the indiscrete topology. The pair (X, J) is called an
indiscrete topological space (or indiscrete space).

(b) Discrete Topology: Let J be the collection of all subsets of X . Then J is a

topology on X . This topology is called the discrete topology. The pair (X, J) is
called the discrete topological space (discrete space).

(c) Usual Topology: Let X = < and J be the collection of all open sets of < . Then
J is a topology on < . It is called the usual topology on < .

(d) Cofinite-Topology: Let J denote the collection of all subsets of X whose compli-
ments are finite together with the empty set φ . Then J is a topology on X . It is
called the cofinite topology. That is, J = {Gi ⊂ X|Gci is finite} .

2
 3

(e) Let X = {a, b, c, d, e} , J1 = {X, φ, {a}, {c, d}, {a, c, d}, {b, c, d, e}} , J2 =
{X, φ, {a}, {c, d}, {a, c, d}, {b, c, d, }} , J3 = {X, φ, {a}, {c, d}, {a, c, d}, {a, b, , d, e}} .
It is easy to show that J1 is a topology on X but J2 and J3 are not.

Theorem 1.0.5. :

(a) Let J1 and J2 be topologies on X . Then J1 ∩ J2 is a topology on X .

(b) If {Ji }i∈I is a collection of topologies on X , then ∩i∈I Ji is a topology on X .

Proof. : (a) (Y1 ) is satisfied, since X, φ ∈ J1 and J2 implies that X, φ ∈ J1 ∩ J2 .

(Y3 ) If G, H ∈ J1 ∩ J2 , then G, H ∈ J1 and G, H ∈ J2 , . Since J1 and J2 are topologies
then G ∩ H ∈ J1 and G ∩ H ∈ J2 , hence G ∩ H ∈ J1 ∩ J2 .
(Y2 ) Let {Gi }i∈I be a collection of elements of J1 ∩ J2 . Then {Gi }i∈I are elements of
each J1 , J2 . But J1 , J2 are topologies. Therefore, ∪i∈I Gi belongs to each J1 , J2 and so
∪i∈I Gi ∈ J1 ∩ J2 . Hence the proof that J1 ∩ J2 is a topology on X is complete.

Theorem 1.0.6. Let J1 and J2 be topologies on X , then J1 ∪ J2 need not be a topology

on X .

Proof. By a counter example. Let X = {a, b, c} , J1 = {X, φ, {a}} and J2 = {X, φ, {b}} .
It is clear that J1 and J2 are topologies on X . Whereas J1 ∪ J2 = {X, φ, {a}, {b}} and
{a}, {b} ∈ J1 ∪ J2 but {a} ∪ {b} = {a, b} ∈/ J1 ∪ J2 .

Definition 1.0.7. (Open Neighborhood): Let A be an open set containing a point p ∈

X , then A is called an open neighborhood of p . Also, A without p , that is, A \ {p} is
called a deleted open neighborhood of p .

Definition 1.0.8. (Accumulation Point, that is, limit point or cluster point
or derived point). Let X be a topological space. A point p ∈ X is an accumula-
tion point of a subset A of X if and only if every open set G containing p contains a
point of A different from p , that is, G open and p ∈ G ⇒ (G \ {p}) ∩ A 6= φ .

Definition 1.0.9. (Derived Set): The set of all accumulation points of A denoted by
Ap is called the derived set of A .

Definition 1.0.10. (Closed Set): Let X be a topological space. A subset A of X is

a closed set if its complement Ac is an open set.

Theorem 1.0.11. : In a top space X , a subset A of X is open if and only if its

complement is Closed.

Proof. : S.T.V. (Note that Acc = A ).

Note: We can also define a topological space on X in terms of its closed sets (Hint:
(Y1 ) remains, while (Y2 ) ∪i=n
i=1 Hi ∈ Jc and (Y3 ) ∩i∈I Hi ∈ Jc ) .
 4

Theorem 1.0.12. (Definition of topology in terms of closed subsets

of X ): Let X be a topological space. Then the collection J of closed subsets of X
is a topology on X if: (i) X and φ ∈ J ; (ii) Given (Hi )i∈I where each Hi ∈ J then
i=n
∩i=1 Hi ∈ J ; (iii) For H1 , H2 , · · · , Hn ∈ J , ∪i=1 Hi ∈ J .

Proof. : (Exercise).

Note that the elements of J are closed sets.

Definition 1.0.13. (Closure of a Set): Let A be a subset of a topological space X .

The closure of A , denoted by Ā is the intersection of all closed supersets of A , that is,
if {Fi , i ∈ I} is the class of all closed subsets of X containing A , then Ā = ∩i∈I Fi .

Theorem 1.0.14. :

(i) Ā is closed

(ii) If F is a closed superset of A , then A ⊂ Ā ⊂ F

(iii) A is closed if and only if A = Ā .

Proof. : (Ex)

Theorem 1.0.15. : Ā = A ∪ Ap

Proof. : Since A ⊂ Ā and Ā is closed, Ap ⊂ (Āp ) ⊂ Ā and hence A ∪ Ap ⊂ Ā . But

A ∪ Ap is closed set containing A , so A ⊂ Ā ⊂ A ∪ Ap . Thus Ā = A ∪ Ap .

Definition 1.0.16. (Closure Point): A point p ∈ X is called a closure point (or

adherent point) of A ( A ⊂ X ) if p belongs to the closure of A , that is, p ∈ Ā .

Theorem 1.0.17. : p ∈ X is a closure point of A ⊂ X if and only if p ∈ A or p is a

limit point of A .

Proof. : (Ex).

Definition 1.0.18. (Dense Set): A subset A of a topological space X is said to be

dense in B ⊂ X , if B is contained in the closure of A , that is, B ⊂ Ā . In particular
A is dense in X if and only if Ā = X .

Definition 1.0.19. (Interior Point): Let A be a subset of a topological space X . A

point p ∈ A is called an interior point of A if p belongs to an open set G contained in
A . That is, if p ∈ G ⊂ A where G is open, then p is an interior point of A .

Definition 1.0.20. (Interior): The set of all interior points of A , denoted by int(A)
is called the interior of A .
 5

Theorem 1.0.21. The interior of a set A is the union of all open subsets of A . Fur-
thermore,

(i) int(A) is open;

(ii) int(A) is the largest open subset of A , that is, if G is an open subset of A then
G ⊂ int(A) ⊂ A ;

(iii) A is open iff A = int(A) .

Proof. : Let {Gi } be the class of all open subsets of A . If x ∈ int(A) , then x belongs
to an open subset of A , that is, there exists n ∈ I such that x ∈ Gn . Hence x ∈ ∪i Gi
and so int(A) ⊂ ∪i Gi . On the other hand, if y ∈ ∪i Gi , then y ∈ Gm for some m . Thus
y ∈ int(A) and ∪i Gi ⊂ int(A) . Accordingly, int(A) = ∪i Gi .
Now: (i) int(A) = ∪i Gi is open since it is the union of open sets.
(ii) If G is an open subset of A then G ∈ {Gi } ; so G ⊂ ∪i Gi = int(A) .
(iii) If A is open then A ⊂ int(A) ⊂ A , that is, A = int(A) . If A = int(A) then A is
open since int(A) is open.

Theorem 1.0.22. : A subset A of a topological space X is closed iff A contains each

of its accumulation points.(Or A is closed iff Ap ⊂ A ).

/ A . Then p ∈ Ac and Ac is open. Hence p ∈

Proof. : Let A be closed and p ∈ / Ap for
Ac is an open set 3 p ∈ Ac and Ac ∩ A = φ . Thus Ap ⊂ A if A is closed.
Now, let Ap ⊂ A ; we show that Ac is open. Let p ∈ Ac the p ∈
/ Ap , so ∃ an open set
G 3 p ∈ G and (G \ {p}) ∩ A = φ . But p ∈ / A ; hence G ∩ A = (G \ {p}) ∩ A = φ . So
G ⊂ A . Thus p is an interior point of A and so Ac is open.
c c

Definition 1.0.23. (Exterior): The exterior of A , written ext(A) , is the interior of

the complement of A , that is, int(Ac ) = ext(A) .

Definition 1.0.24. (Boundary): The boundary of A , written bdry(A) , is the set of

points which do not belong to the interior or exterior of A .

Theorem 1.0.25. Let A be any subset of a topological space X . Then the closure of A
is the union of the interior and boundary of A . That is, Ā = int(A) ∪ bdry(A)

Proof. : Since X = int(A) ∪ ext(A) ∪ bdry(A) , then ext(A) = {int(A) ∪ bdry(A)}c

and it suffices to show that (Ā)c = ext(A) . Let p ∈ ext(A) , then ∃ an open set
G 3 p ∈ G ⊂ Ac which ⇒ G ∩ A = φ . So p is not a limit point of A , that is, p ∈
/ Ap ,
and p ∈/ A . Hence p ∈/ Ap ∪ A = Ā . In other words ext(A) ⊂ (Ā)c .
Now assumed p ∈ (Ā)c = (A ∪ Ap )c . Thus p ∈/ Ap , so ∃ an open set G 3 p ∈ G and
/ A , so G ∩ A = φ and p ∈ G ⊂ Ac . Thus p ∈ ext(A) .
(G \ {p}) ∩ A = φ . But also p ∈
Hence (Ā)c = ext(A) .
 6

Definition 1.0.26. (Nowhere Dense): A subset of a topological space X is said to be

nowhere dense in X if the interior of the closure of A is empty, that is, int(Ā) = φ .

Definition 1.0.27. (Neighborhood and Neighborhood Systems): Let p be a point in

a topological space X . A subset N of X is a neighborhood of p iff N is a superset of
an open set G containing p ; that is, p ∈ G ⊂ N , where G is an open set. The class of
neighborhoods of p ∈ X denoted by Np is called the neighborhood system of p .

Theorem 1.0.28. A set G is open iff it is a neighborhood of each of its points.

Proof. : Suppose G is open. Each point p ∈ G belongs to the open set G contained in
G . Hence G is a neighborhood of each of its points.
Conversely, suppose G is a neighborhood of each of its points. So for each point
p ∈ G, ∃ an open set Gp 3 p ∈ Gp ⊂ G . Hence G = ∪{Gp : p ∈ G} and is open since
it is a union of open sets.

Theorem 1.0.29. (Neighborhood Axioms):

(i) Np is not empty and p belongs to each member of Np .

(ii) The intersection of any two members of Np belongs to Np .

(iii) Every superset of a member of Np belongs to Np .

(iv) Each member N ∈ Np is a superset of a member G ∈ Np where G is a neighborhood

of each of its points, that is, G ∈ Ng for every g ∈ G .

Proof. : (i) If N ∈ Np , then ∃ an open set G 3 p ∈ G ⊂ N ; hence p ∈ N . Note

X ∈ Np since X is an open set containing p . So Np 6= φ .
(ii) Let N and M be neighborhoods of p . So ∃ open sets G, H 3 p ∈ G ⊂ N and
p ∈ H ⊂ M . Hence p ∈ G ∩ H ⊂ N ∩ M and G ∩ H is open, thus M ∩ N is a
neighborhood of p .
(iii) Let N ∈ Np and M a superset of N . Since N ∈ Np , ∃ an open set G 3 p ∈ G ⊂
N . By hypothesis, N ⊂ M , so p ∈ G ⊂ N ⊂ M which ⇒ p ∈ G ⊂ M and hence M
is a neighborhood of p .
(iv) If N ∈ Np , then N is a neighborhood of p so ∃ an open set G 3 p ∈ G ⊂ N . But
by Theorem (1.0.28) G ∈ Np and G is a neighborhood of each of its points.

Definition 1.0.30. (Convergent Sequences): A sequence {an } of points in a topologi-

cal space X converges to a point a ∈ X iff for each open set G containing a∃ a positive
no ∈ N 3 ∀n > no , an ∈ G . Write limn→∞ an = a or an → a . This means that every G
contains almost all the an0 s except for a finite number.

Exercise 1.0.31. (i) Let J be the class of subsets of < consisting of < , φ and all
open infinite intervals Ea = (a, ∞) with a ∈ < . Show that J is a topology on < .
 7

(ii) Show that if A ⊂ B then Ap ⊂ B p , where A and B are subsets of a topological

space.

(iii) Let A and B be subsets of a topological space. Show that (A ∪ B)p = Ap ∪ B p .

(iv) Show that if Np is finite, then ∩{N : N ∈ Np } belongs to Np .

(v) Prove that if Ā ∩ B̄ = φ , then bdry(A ∪ B) = bdry(A)U bdry(B) .

Definition 1.0.32. (Coarser and Finer Topologies): Let J1 and J2 be topologies

on a non-empty set X . Suppose that each J1 -open subset of X is also a J2 -open subset
of X . That is, suppose that J1 is a subclass of J2 ( J1 ⊂ J2 ). Then we say that J1 is
coarser or smaller or weaker than J2 or that J2 is finer or larger or stronger than J1 .

Proposition 1.0.33. The collection J = {Ji } of all topologies on X is partially ordered

by class inclusion. Note: We can write J1 ≤ J2 for J1 ⊂ J2 .

Definition 1.0.34. We shall say that two topologies on X are not comparable if neither
is coarser than the other.

Definition 1.0.35. (Relative Topologies, Subspaces): Let A be a non-empty sub-

set of a topological space (X, J) . The class JA of all intersections of A with J -open
subsets of X is a topology on A ; it is called the relative topology on A or the relativiza-
tion of J to A . The topological space (A, JA ) is called a subspace of (X, J) .
Note: A subset H of A is a JA -open set (i.e open relative to A ) iff ∃ a J -open subset
G of X 3 H = G ∩ A .
Chapter 2

BASES AND SUBBASES

Definition 2.0.36. Let (X, J) be a topological space. A class B of open subsets of X ,

is a base for the topology J iff every open set G ∈ J is the union of members of B .

Note that B ⊂ J . Equivalently B ⊂ J is a base for J iff for any p belonging to an

open set G, ∃Bp ∈ B with p ∈ Bp ⊂ G .
Theorem 2.0.37. Alternative definition of a base: Let B a class of subsets of a
non-empty set X . Then B is a base for some topology on X iff it possess the following
two properties.

(i) X = ∪B : B ∈ B} ;
(ii) For any B, B ∗ ∈ B , B ∩ B ∗ is the union of members of B . [or equivalently if
p ∈ B ∩ B ∗ then ∃Bp ∈ B 3 p ∈ Bp ⊂ B ∩ B ∗ ].

Proof. Let B be a base for a topology J on X . Since X is open, X is the union of

all the members of B , that is, X = ∪B : B ∈ B} . Furthermore, if B, B ∗ ∈ B , then in
particular, B and B ∗ are open. Hence the intersection of B ∩ B ∗ is also open and since
B is a base for J , it is the union of members of B . Thus (i) and (ii) are satisfied.
Conversely, suppose B is a class of subsets of X which satisfy (i) and (ii) above. Let
J be the class of all subsets of X which are union members of B . We claim that J is a
topology on X . Observe that B ⊂ J will be a base for this topology.
By (i) X = ∪{B : B ∈ B} ; so X ∈ J . Note that φ is the union of the empty subclass of
B . That is, φ = ∪{B : B ∈ φ ⊂ B} ; hence φ ∈ J and so (Y1) is satisfied. Now let {Gi }
be a class of members of J , each Gi is the union of members B ; hence ∪{Gi } is also
the union of B and so belong to J . Thus J satisfies (Y2). Lastly, suppose G, H ∈ J .
We need to show that G ∩ H also belongs to J . By definition of J , ∃ two subclass
{Bi : i ∈ I} and {Bk : k ∈ I} of B 3 G = ∪Bi and H = ∪Bk . Then by distributive
law G ∩ H = (∪Bi ) ∩ (∪Bk ) = ∪{Bi ∩ Bk : i, k ∈ I} But by (ii) Bi ∩ Bk is the union of
members of B , hence G ∩ H is also the union of members of B and so belongs to J .
Therefore, (Y3) is satisfied. Hence J is a topology on X with base B .

8
 9

Definition 2.0.38. Subbase: Let (X, J) be a topological space. A class S of open

subsets of X , that is, S ⊂ J , is a subbase for the topology J on X iff finite intersections
of members of S form a base for J .

Theorem 2.0.39. Any class A of subsets of a non-empty set X is the subbase for a
unique topology J on X .

Proof. We show that the class B of finite intersections of A satisfies the two conditions
in Theorem (2.0.37). Now X ∈ B , since X by definition is the empty intersection of
members of A , so X = ∪{B : B ∈ B} . Furthermore, if G, H ∈ B then G and H
are finite intersections of members of A . Hence G ∩ H is also a finite intersection of
members of A and therefore belongs to B . Accordingly, B is a base for a topology J
on X for which A is a subbase.
Let S be a subbase for two topologies J and J ∗ on X , then show that J = J ∗ .
Suppose G ∈ J . Since S is a subbase for J, G = ∪{Si1 ∩ · · · ∩ Sik } where Sik ∈ S .
But S is also a subbase for J ∗ and so S ⊂ J ∗ , hence each Sik ∈ J ∗ . Since J ∗ is a
topology Si1 ∩ · · · ∩ Sik ∈ J ∗ and hence G ∈ J ∗ . Thus J ⊂ J ∗ . Similarly, J ∗ ⊂ J and
so J = J ∗ .

Theorem 2.0.40. Let A be a class of subsets of a non-empty set X . Then the topology
J on X generated by A is the intersection of all topologies on X which contain A .

Proof. Let {Ji } be the collection of topologies on X containing A and let J ∗ = ∩Ji .
Note that A ⊂ J ∗ . We want to prove that J = J ∗ . Since J is a topology containing A
and J ∗ is the intersection of all such topologies, we have J ∗ ⊂ J .
On the other hand, let G ∈ J then by the definition of J , G = ∪i {Si1 ∩ Si2 ∩ · · · ∩ Sik }
where Sik ∈ A . But A ⊂ J ∗ , hence each Sik ∈ J ∗ . Accordingly Si1 ∩ · · · ∩ Sik ∈ J ∗
and so G = ∪i {Si1 ∩ · · · ∩ Sik } ∈ J ∗ . This shows that for G ∈ J then G ∈ J ∗ , hence
J ⊂ J ∗ . Consequently J = J ∗ .

Definition 2.0.41. Local base: Let p be any arbitrary point in a topological space X .
A class βp of open sets containing p is called a local base at p iff for each open set G
containing p ∃Gp ∈ βp with the property that p ∈ Gp ⊂ G .

Theorem 2.0.42. Let B be a base for a topology J on X and let p ∈ X . Then the
members of the base B which contains p form a local base at the point p .

Proof. STV

Theorem 2.0.43. A point p in a topological space X is an accumulation point of A ⊂ X

iff each member of some local base βp at p contains a point of A different from p .

Proof. Recall p ∈ X is an accumulation point of A iff (G \ {p}) ∩ A 6= φ ∀ G ∈ J 3

p ∈ G . But βp ⊂ J so in particular (B \ {p}) ∩ A 6= φ ∀ B ∈ βp .
 10

Conversely, suppose (B \ {p}) ∩ A 6= φ ∀ B ∈ βp and let G be any open subset of

X containing p . Then ∃ B0 ∈ βp for which p ∈ B0 ⊂ G . But then
(G \ {p}) ∩ A ⊃ (B0 \ {p}) ∩ A 6= φ . So (G \ {p}) ∩ A 6= φ , that is, p is an accumulation
of A .

Theorem 2.0.44. A sequence {an } of points in a topological space X converges to

p ∈ X iff each member of some local base βp at p contains almost all of the terms of the
sequence.

Proof. Recall that an → p iff every open set G ∈ J containing p contains almost all the
terms of the sequence. But βp ⊂ J , so in particular each B ∈ βp contains almost all the
terms of the sequence.
On the other hand, suppose every B ∈ βp contains almost all the terms of the sequence,
and let G be any open set containing p . Then ∃ B0 ∈ βp for which p ∈ B0 ⊂ G .
Hence G also contains almost all the terms of the sequence and so {an } converges to
p.

Corollary 2.0.45. Let B be a base for a topology J on X .

1. p ∈ X is an accumulation point of A ⊂ X iff each open base set B ∈ B containing

p contains a point of A different from p .

2. A sequence {an } of points in X converges to p ∈ X iff each open base set B ∈ B

containing p contains almost all of the terms of the sequence.

Proof. STV

Definition 2.0.46. Nested Local Base: Let βp be a countable local base at the point
p ∈ X . Then we can index the members of βp by N , that is, we can write
βp = {B1 , B2 , · · · } (we permit repetitions in the case that βp is finite). If in addition
B1 ⊃ B2 ⊃ B3 ⊃ · · · then we call βp a nested local base at p .
Chapter 3

CONTINUITY AND
TOPOLOGICAL EQUIVALENCE

Definition 3.0.47. Continuous Functions

Let (X,J) and (Y, J ∗ ) be topological spaces. A function f from X into Y is continuous
relative to J and J ∗ or J − J ∗ continuous or simply continuous iff the inverse image
f −1 [H] of every J ∗ -open subset H of Y is a J− open subset of X . That is ; iff H ∈ J ∗
implies f −1 [H] ∈ J .

Theorem 3.0.48. A function f : X → Y is continuous iff the inverse of each member

of a base B for Y is an open subset of X .

Proof. Let function f : X → Y , where X and Y are topological spaces. Let B be a

base for the topology on Y . Suppose for each member B ∈ B, f −1 [B] is an open subset of
X ; then f is a continuous function. For let H be an open subset of Y ; then H = ∪Bi ,
a union of members of B . But f −1 [H] = f −1 [∪Bi ] = ∪f −1 [Bi ] and each f −1 [Bi ] is open
by hypothesis, hence f −1 [H] is the union of open sets and is therefore open. Accordingly,
f is continuous.
Conversely, let f : X → Y be continuous, where X and Y are top spaces. Let B be
a base for the topology on Y . Let H be an open subset of Y ; then H = ∪Bi , Bi ∈ B .
Since f is continuous, then f −1 [H] is an open subset of X . Thus f −1 [H] = f −1 [∪i Bi ] =
∪i f −1 [Bi ] ⇒ ∪i f −1 [Bi ] is an open subset of X . Therefore f −1 [Bi ] is an open subset of
X.

Theorem 3.0.49. Let S be a subbase for a top space Y . Then a function f : X → Y

is continuous iff the inverse of each member of S is an open subset of X .

Proof. Suppose f −1 [S] ∈ J for every S ∈ S . Let G ∈ J ∗ . By definition of subbase,

G = ∪i (Si1 ∩ · · · ∩ Sin ) , where Sik ∈ S . Hence, f −1 [G] = f −1 [∪i (Si1 ∩ · · · ∩ Sin )] =
∪i f −1 [Si1 ∩ · · · ∩ Sin ] = ∪i (f −1 [Si1 ] ∩ · · · ∩ f −1 [Sin ]) . But Sik ∈ S ⇒ f −1 [Sik ] ∈ J , for

11
 12

each i, k . Hence f −1 [G] ∈ J , since it is the union of finite intersections of open sets.
Accordingly, f is continuous.
Conversely, if f is continuous, then the inverse of all open sets, including the members
of S are open.
Theorem 3.0.50. A function f : X → Y is continuous iff the inverse image of every
closed subset of Y is a closed subset of X

Proof. Suppose f : X → Y is continuous and let F be a closed subset of Y . Then F c

is open and so f −1 [F c ] is open in X . But f −1 [F c ] = (f −1 [F ])c . Therefore f −1 [F ] is
closed.
Conversely, assume F closed in Y ⇒ f −1 [F ] closed in X . Let G be an open subset
of Y . Then Gc is closed in Y , and so f −1 [Gc ] = (f −1 [G])c is closed in X . Accordingly,
f −1 [G] is open and therefore f is continuous.
Definition 3.0.51. Let X be a topological space. A point p ∈ X is said to be arbitrarily
closed to a set A ⊂ X if either
(i) p ∈ A or
(ii) p is an accumulation point of A .

Note :
(i) Recall that Ā = A ∪ Ap ; so the closure of A consists precisely of those points in X
which arbitrarily closed to A .
(ii) Recall also that Ā = int(A) ∪ bdry(A) : hence p is arbitrarily close to A if p is
either an interior or a boundary point of A .
Theorem 3.0.52. A function f : X → Y is continuous iff for any p ∈ X and A ⊂ X ,
p arbitrarily close to
A ⇒ f (p) arbitrarily to f [A] or p ∈ Ā ⇒ f (p) ∈ f [A] or f [Ā] ⊂ f [A]

Proof. Suppose f : X → Y is continuous. Now f [A] ⊂ f [A] , so A ⊂ f −1 [f [A]] ⊂

f −1 [f [A]] . But f [A] is closed, and so f −1 [f [A]] is also closed, hence A ⊂ Ā ⊂ f −1 [f [A]]
and therefore f [Ā] ⊂ f [A] = f [f −1 [f [A]] .
Conversely, suppose f [Ā] ⊂ f [A] for any A ⊂ X and let F be a closed subset of Y .
Set A = f −1 [F ] ; we wish to show that Ā = A . Now f [Ā] = f [f −1 [F ]] ⊂ f [f −1 [F ]] =
F̄ = F. Hence Ā ⊂ f −1 [f [Ā]] ⊂ f −1 [F ] = A . But A ⊂ Ā , so Ā = A and f is
continuous.
Definition 3.0.53. (Continuity at a point): A function f : X → Y is continuous
at p ∈ X iff the inverse image f −1 [H] of every open set H ⊂ Y containing f (p) is
a superset of an open set G ⊂ X containing p or equivalently iff the inverse image of
every neighborhood of f (p) is a neighborhood of p , that is, N ∈ Nf (p) ⇒ f −1 [N ] ∈ Np .
 13

Theorem 3.0.54. : Let X and Y be topological spaces. Then a function X → Y is

continuous iff it is continuous at every point of X .

Proof. Assume f is continuous, and let H ⊂ Y be an open set containing f (p) . But
then p ∈ f −1 [H] and f −1 [H] is open. Hence f is continuous at p .
Now suppose f is continuous at every point p ∈ X and let H ⊂ Y be open. For every
p ∈ f −1 [H], ∃ an open set Gp ⊂ X 3 p ∈ Gp ⊂ f −1 [H] . Hence f −1 [H] = ∪{Gp : p ∈
f −1 [H]} a union of open sets. Accordingly, f −1 [H] is open and so f is continuous.

Definition 3.0.55. (Sequential Continuity at a point): A function f : X → X

is sequentially continuous at a point p ∈ X iff for every sequence {an } in X converging
to p , the sequence {f (an )} in X converges to f (p) . That is, an → p ⇒ f (an ) → f (p) .

Theorem 3.0.56. : If a function f : X → X is continuous at p ∈ X , then it is

sequentially continuous at p .

Proof. We need to show that any neighborhood N of f (p) contains almost all the terms
of the sequence {f (an )} . Let N be a neighborhood of f (p) . By hypothesis, f is
continuous at p , hence M = f −1 [N ] is a neighborhood of p . If the sequence {an }
converges to p , then M contains almost all the terms of the sequence {an } , that is,
an ∈ M for almost all n ∈ N . But an ∈ M ⇒ f (an ) ∈ f [M ] = f (f −1 [N ]) = N .
Hence f (an ) ∈ N for almost all n ∈ N , and so the sequence {f (an )} converges to f (p) .
Accordingly, f is sequencially continuous at p .

Remark 3.0.57. : The converse of Theorem (3.0.56) is not true. Consider, for
example the topology J on the real line < consisting of ∅ and the complements
of countable sets. Note that a sequence {an } converges to p iff it has the form
{a1 , a2 , · · · , an , p, p, p, · · · }. Then for any function f : (<, J) → (<, J ∗ ) , {f (an )} =
{f (a1 ), f (a2 ), · · · , f (an ), f (p), f (p), f (p), · · · } converges to f (p) . In other words, ev-
ery function in (<, J) is sequentially continuous. On the other hand, the function
f : (<, J) → (<, J ∗ ) defined by f (x) = x where J ∗ is the usual topology is not J − J ∗
continuous since f −1 [(0, 1)] = (0, 1) is not J− open subset of < .

Definition 3.0.58. (Open Function): A function f : X → Y is called an open function

if the image of every open set is open.

Definition 3.0.59. (Closed Function): A function g : X → Y is called a closed

function if the image of every closed set is closed.

Definition 3.0.60. (Homeomorphic or Topologically Equivalent Spaces): Two

topological spaces X and Y are called homeomorphic or topologically equivalent if ∃
a bijective function f : X → Y 3 f and f −1 are continuous. The function is called a
homeomorphism.
 14

Definition 3.0.61. (Bicontinuity): A function f is called bicontinuous if f is open

and continuous.
Note: f : X → Y is a homeomorphism iff f is bicontinuous and bijective.
Theorem 3.0.62. : The relation in any collection of topological spaces defined by “ X is
homeomorphic to Y ” is an equivalence relation.

Proof. : Exercise.

Note: By the fundamental Theorem on equivalence relations, any collection of topolog-

ical spaces can be partitioned into classes of topological equivalent spaces.
Definition 3.0.63. (Topological Invariant): A property p of sets is called topologi-
cally invariant if whenever a topological space (X, J) has p then every space homeomor-
phic to (X, J) also has p .
Definition 3.0.64. (Topologies Induced By Functions): Let {Yi , Ji } be any collec-
tion of topological spaces and for each Yi let there be given a function fi : X → Yi defined
on some arbitrary non-empty set X . Consider the following class of subsets of X
ξ = ∪i {fi−1 [H] : H ∈ Ji } .
The topology J on X generated by ξ is called the topology induced (or generated) by the
functions fi .

Exercise: Check that ξ is a topology on X .

Theorem 3.0.65. : The main properties of J in Definition (3.0.64) are:

(i) All the functions fi are continuous relative to J .

(ii) J is the intersection of all the topologies on X with respect to which the functions
fi are continuous.
(iii) J is the coarsest topology on X with respect to which the functions fi are contin-
uous.
(iv) ξ is a subbase for the topology J .

Proof. : (i) For any function fi : (X, J) → (Yi , Ji ) , if H ∈ Ji then fi−1 [H] ∈ ξ ⊂ J .
Hence all the fi are continuous with respect to J .
(ii) Let {Ji } be a collection of topologies on a set X . If a function f : X → Y is
continuous w.r.t each Ji , then f is continuous w.r.t. the intersection topology Jn = ∩Ji .
All the functions fi are also continuous w.r.t. Jn . Hence ξ ⊂ Jn and since J is the
topology generated by ξ then J ⊂ Jn . On the other hand, J is one of the topologies
w.r.t which the fi are continuous, hence Jn ⊂ J and so J = Jn .
(iii) Follows from (ii).
(iv) Follows from the fact that any class of sets is a subbase of the topology it generates.
Chapter 4

COUNTABILITY AND
SEPARATION AXIOMS

Definition 4.0.66. (First Countable Space): A topological space X is called a first

countable space if it satisfies the first axiom of countability, that is, for each point p ∈ X ,
∃ a countable class βp of open sets containing p 3 every open set G containing p also
contains a member of βp .
Theorem 4.0.67. : Let X satisfy the 1st axiom of countability, then f : X → Y is
continuous at p ∈ X iff it is sequentially continuous at p .

Proof. : It suffices to show that if f is sequentially continuous at p then f is continuous

at p , since converse had been proved (see Theorem (3.0.56)) for an arbitrary topological
space.
Now suppose f is not continuous at p , then we need to show that f is not sequentially
continuous at p . Let βp = {B1 , B2 , · · · } be a nested local base at p and suppose f is
not continuous at p . Then ∃ an open set H of Y 3 f (p) ∈ H but Bn 6⊂ f −1 [H]
for every n ∈ N . Hence for every n ∈ N , ∃ an ∈ Bn 3 an 6∈ f −1 [H] which implies
f (an ) 6∈ H . Noting that the sequence {an } Converges to p , but the sequence {f (an )}
does not converge to f (p) , since the open set H containing f (p) does not contain any
of the terms of the sequence, {f (an )} . Accordingly, f is not sequentially continuous at
p.
Definition 4.0.68. (Second Countable Space): A topological space X is called a sec-
ond countable space if it satisfies the second axiom of countability. That is, ∃ a countable
base B for the topology J on X .
Theorem 4.0.69. : A second countable space is also first countable.

Proof. : If B is a countable base for a space X, and if βp consists of the members of B

which contain the point p ∈ X , then βp is a countable local base at p .

15
 16

Definition 4.0.70. (Cover): Let A ⊂ X and let A be a class of subsets of X such

that A ⊂ ∪{E : E ∈ A} , then A is called a cover (or covering) of A , or A is said to
cover A .

Note: (i) If each member of A is an open subset of X , then A is called an open cover
of A .
(ii) If A contains a countable (finite) subclass which also is a cover of A , then A is said
to be reducible to a countable (finite) cover or A is said to contain a countable (finite)
subcover.
Theorem 4.0.71. (Lindelof): Let A be any subset of a 2nd countable space X . If G
is an open cover of A , then G is reducible to a countable cover.

Proof. Let B be a countable base for X . Since A ⊂ ∪{G : G ∈ G} , then for every p ∈ A
∃ Gp ∈ G 3 p ∈ Gp . Also B is a base for X , implies that for every p ∈ A , ∃ Bp ∈ B
3 p ∈ Bp ⊂ Gp . Hence A ⊂ ∪{Bp : p ∈ A} . But {Bp : p ∈ A} ⊂ B , so it is countable,
hence {Bp : p ∈ A} = {Bn : n ∈ N} , where N is a countable index set. For each n ∈ N
choose one set Gn ∈ G 3 Bn ⊂ Gn . Then A = ∪{Bn : n ∈ N} ⊂ ∪{Gn : n ∈ N} and
so {Gn : n ∈ N} is a countable subcover of G .
Theorem 4.0.72. (Lindelof): Let G be a base for 2nd countable space X . Then G
is reducible to a countable base for X .

Proof. Since X is 2nd countable, X has a countable base B = {Bn : n ∈ N} . G is

also a base for X , so for each n ∈ N , Bn = ∪{G : G ∈ Gn } with Gn ⊂ G . So Gn
is an open cover of Bn and by Theorem (4.0.71) is reducible to a countable cover Gn∗ ;
that is, for each n ∈ N , Bn = ∪{G : G ∈ Gn∗ } with Gn∗ ⊂ G and Gn∗ is countable. But
G ∗ = {G : G ∈ Gn∗ , n ∈ N} is a base for X since B is. Furthermore, G ∗ ⊂ G and G is
countable.
Definition 4.0.73. (Lindelof Space): A topological space X is called a Lindelof space
if every open cover of X is reducible to a countable cover. For example, every 2nd
countable space is a Lindelof space.
Definition 4.0.74. (Separable Space): A topological space X is said to be separable
if it contains a countable dense subset, that is, ∃ A ⊂ X 3 Ā = X .
Theorem 4.0.75. If X satisfies the 2nd axiom of countability, then X is separable.

Proof. Since X satisfies the 2nd axiom of countability, then X has a countable base
B = {Bn : n ∈ N} . For each n ∈ N , choose a point an ∈ Bn . Then the set A = {an :
n ∈ N} is also countable. We show that Ā = X or equivalently, that each point p ∈ Ac
is an accumulation point of A . Let G be an open set containing p . Then G contains
at least one set Bn ∈ B . Hence an0 ∈ Bn0 ⊂ G . Now an0 is different from p , since
p ∈ Ac and an0 ∈ A . Accordingly, p is an accumulation point of A , since every open
set G containing p also contains a point of A different from p .
 17

Definition 4.0.76. ( T1 -Space): A top space X is a T1 -space iff given any pair of
distinct points a, b ∈ X ∃ open sets G and H 3 a ∈ G , b 6∈ G and b ∈ H , a 6∈ H .
Note: G and H need not be disjoint.

Theorem 4.0.77. A topological space X is a T1 -space iff every singleton subset {p} of
X is closed.

Proof. Suppose X is a T1 -space and p ∈ X . We show that {p}c is open. Let x ∈ {p}c .
Then x 6= p , and so by Definition (4.0.76) ∃ an open set Gx 3 x ∈ Gx but p 6∈
Gx .Hence x ∈ Gx ⊂ {p}c and this implies that {p}c = ∪{Gx : x ∈ {p}c } . Accordingly
{p}c a union of open sets is open and {p} is closed.
Conversely, suppose p is closed for every p ∈ X . Let a, b ∈ X with a 6= b . Now
a 6= b ⇒ b ∈ {a}c , hence {a}c is an open set containing b but not containing a .
Similarly {b}c is an open set containing a but not containing b . Accordingly, X is a
T1 -space.

Definition 4.0.78. (Hausdorff = T2 Space): A topological space X is a Hausdorff

space or a T2 -space iff each pair of distinct points belong respectively to disjoint open
sets.

Theorem 4.0.79. If X is a Hausdorff space then every convergent sequence in X has

a unique limit.

Proof. Suppose {an } converges to a and b , and suppose a 6= b . Since X is Hausdorff ∃

open sets G and H 3 a ∈ G , b ∈ H and G ∩ H = φ . By hypothesis, {an } converges
to a , hence ∃ n0 ∈ N 3 ∀ n > n0 , an0 ∈ G . That is, G contains all except a
finite number of the terms of the sequence. But G and H are disjoint, hence H can
only contain those terms of the sequence which do not belong to G and there are only
a finite number of these. Accordingly, {an } cannot converge to b . But this violates the
hypothesis, hence a = b .

Theorem 4.0.80. Let X be a first countable space. Then the following are equivalent:

(i) X is Hausdorff.

(ii) Every convergent sequence has a unique limit.

Proof. By Theorem (4.0.79), (i) ⇒ (ii) ; hence we need only show that (ii) ⇒ (i) .
Suppose X is not Hausdorff. Then ∃ a, b ∈ X , a 6= b with the property that every open
set containing a has a non-empty intersection with every open set containing b . Now
let {Gn } and {Hn } be nested local bases at a and b respectively. Then Gn ∩ Hn 6= φ
for every n ∈ N , and so ∃ {an } 3 a1 ∈ G1 ∩ H1 , a2 ∈ G2 ∩ H2 , · · · . Accordingly,
{an } converges to both a and b . In other words (ii) ⇒ (i) .
 18

Definition 4.0.81. (Regular Spaces): A topological space X is regular iff it satisfies

the following: If F is a closed subset of X and p ∈ X does not belong to F , then ∃
disjoint open sets G and H 3 F ⊂ G and p ∈ H .

Definition 4.0.82. ( T3 -Space): A regular T1 -space is called a T3 -space.

Definition 4.0.83. (Normal Space): A topological space X is normal if F1 and F2

are disjoint closed subsets of X , then ∃ disjoint open sets G and H 3 F1 ⊂ G and
F2 ⊂ H .

Theorem 4.0.84. A topological space X is normal iff for every closed set F and open
set H containing F ∃ an open set G 3 F ⊂ G ⊂ Ḡ ⊂ H .

Proof. Let X be normal and F ⊂ H with F closed and H open. Then H c is closed
and F ∩ H c = φ . But X is normal, hence ∃ open sets G, G∗ 3 F ⊂ G , H c ⊂ G∗ and
G ∩ G∗ = φ . So G ∩ G∗ = φ ⇒ G ⊂ G∗c and H c ⊂ G∗ ⇒ G∗c ⊂ H . Note that G∗c is
closed. Hence F ⊂ G ⊂ Ḡ ⊂ G∗c ⊂ H .
Conversely, let F1 and F2 be disjoint closed sets. Then F1 ⊂ F2c and F2c is open. By
the assumption there exists an open set G 3 F1 ⊂ G ⊂ Ḡ ⊂ F2c . But Ḡ ⊂ F2c ⇒ F2 ⊂
Ḡc and G ⊂ Ḡ ⇒ G ∩ Ḡc = φ . Furthermore, Ḡc is open. Thus F1 ⊂ G and F2 ⊂ Ḡc
with G, Ḡc disjoint open sets. Hence X is normal.

Definition 4.0.85. ( T4 -Space): A normal T1 -space is called a T4 -space.

Definition 4.0.86. (Relationship Between Some Spaces):

Chapter 5

COMPACTNESS

Definition 5.0.87. A subset, A , of a topological space X is compact if every open cover

of A is reducible to a finite cover. [That is, if A is compact and A ⊂ ∪i Gi , where the Gi
are open sets, then one can select a finite number of the open sets, say, Gi1 , Gi2 , · · · , Gin
3 A ⊂ Gi1 ∪ Gi2 ∪ · · · ∪ Gin .]

Theorem 5.0.88. Continuous images of compact sets are compact.

Proof. Let f : X → Y be continuous and A a compact subset of X . We want to

show that f (A) is a compact subset of Y . Suppose G = {Gi } is an open cover of
f (A) , that is, f (A) ∪i (Gi ) , then A ⊂ f −1 (f (A)) ⊂ f −1 (∪i Gi ) = ∪i f −1 (Gi ) . Hence
L = {f −1 (Gi )} is a cover of A . Now f is continuous and each Gi is an open set, so
each f −1 (Gi ) is also open. In other words L is an open cover of A . But A is compact,
so L is reducible to a finite cover, say A ⊂ f −1 (Gi1 ) ∪ · · · ∪ f −1 (Gin ) . Accordingly,
f (A) ⊂ f [f −1 (Gi1 ) ∪ · · · f −1 (Gin )] ⊂ Gi1 ∪ · · · ∪ Gin ⇒ f (A) is compact.

Theorem 5.0.89. Let A be a subset of a topological space (X, J) . Then A is compact

with respect to J iff A is compact with respect to the relative topology JA on A .

Proof. Suppose A is compact with respect to J . Let {Gi } be a JA -open cover of A . By

definition of relative topology ∃ Hi ∈ J 3 Gi = A ∩ Hi ⊂ Hi . Hence A ⊂ ∪i Gi ⊂ ∪i Hi
and therefore {Hi } is a J -open cover of A . By the hypothesis, A is J -compact, so
{Hi } contains a finite subcover, say Hi1 , · · · , Hin . That is, A ⊂ Hi1 ∪ · · · ∪ Hin ; where
Hik ∈ {Hi } . But

A ⊂ A ∩ (Hi1 ∪ · · · ∪ Hin )
= (A ∩ Hi1 ) ∪ · · · ∪ (A ∩ Hin )
= Gi1 ∪ · · · ∪ Gin .

Thus {Gi } contains a finite subcover {Gi1 , · · · , Gin } and (A, JA ) is compact.

19
 20

Conversely, let {Hi } be a J -open cover of a set A . Set Gi = A ∩ Hi ; then A ⊂ ∪i Hi

⇒ A ⊂ A∩(∪i Hi ) = ∪i (A∩Hi ) = ∪i Gi . But Gi ∈ JA , so {Gi } is a JA -open cover of A .
By hypothesis, A is JA -compact, thus {Gi } contains a finite subcover {Gi1 , · · · , Gin } .
Accordingly,
A ⊂ Gi1 ∪ · · · ∪ Gin = (A ∩ Hi1 ) ∪ · · · ∪ (A ∩ Hin ) = A ∩ (Hi1 ∪ · · · ∪ Hin ) ⊂ Hi1 ∪ · · · ∪ Hin .
Thus {Hi } is reducible to a finite cover {Hi1 , · · · , Hin } and therefore A is compact with
respect to J .
Theorem 5.0.90. A subset of a compact space need not be compact.

Proof. By an example. The closed unit interval [0, 1] is compact by Heine-Borel Theorem.
However, the open interval (0, 1) which is a subset of it, is not compact. How? Let
1
A = (0, 1) with the usual topology. Consider the class of open intervals G = {( n+2 , n1 )} =
{( 31 , 1), ( 41 , 12 ), · · · } . We note that A = ∪∞
n=1 Gn , where Gn ∈ G is an open cover of A .
But G contains no finite subcover. For let G ∗ = {(a1 , b1 ), · · · , (an , bn )} be any finite
subclass of G . If ∈= min(a1 , · · · , an ) then ∈> 0 and (a1 , b1 ) ∪ · · · ∪ (an , bn ) ⊂ (∈, 1) .
But (0, ∈] and (∈, 1) are disjoint; hence G ∗ is not a cover of A and so A is not
compact.
Theorem 5.0.91. Let F be a closed subset of a compact space X . Then F is also
compact.

Proof. Let G = {Gi } be an open cover of F , that is, F ⊂ ∪i Gi . Then X = (∪i Gi ) ∪ F c ;

that is, G ∗ = {Gi } ∪ {F c } is a cover of X . But F c is open since F is closed, so G ∗ is
an open cover of X . By the hypothesis X is compact; hence G ∗ is reducible to a finite
cover of X , say, X = Gi1 ∪ · · · ∪ Gin ∪ F c , where Gik ∈ G . But F and F c are disjoint,
hence F ⊂ Gi1 ∪ · · · ∪ Gin . With this, we have shown that any open cover {Gi } of F
contains a finite subcover, that is, F is compact.
Definition 5.0.92. (Finite Intersection Property): A class {Ai } of sets is said
to have the finite intersection property if every finite subclass {Ai1 , · · · , Ain } has a non-
empty intersection, that is, Ai1 ∩ · · · ∩ Ain 6= φ .
Lemma 5.0.93. The following statements are equivalent:

(i) X is compact.

(ii) For every class {Fi } of closed subsets of X , ∩i Fi = φ ⇒ {Fi } contains a finite
subclass {Fi1 , · · · , Fin } with Fi1 ∩ · · · ∩ Fin = φ .

Proof. (i) ⇒ (ii): Suppose ∩i Fi = φ . Then by De Morgan’s law,

X = φc = (∩i Fi )c = ∪i Fic . So {Fic } is an open cover of X since each Fi is closed. But
by hypothesis, X is compact, hence ∃ Fi1c , · · · , Fin
c
∈ {Fic } 3 X = Fi1c ∪ · · · ∪ Fin
c
. Thus
c c c c cc cc
by De Morgan’s law, φ = X = (Fi1 ∪ · · · ∪ Fin ) = Fi1 ∩ · · · ∩ Fin = Fi1 ∩ · · · ∩ Fin and
we have shown that (i) ⇒ (ii)
 21

(ii) ⇒ (i): Let {Gi } be an open cover of X , that is, X = ∪i Gi . By De Morgan’s

law, φ = X c = (∪i Gi )c = ∩i Gci ; since each Gi is open, {Gci } is a class of closed sets
and by above, has empty intersection. Hence by the hypothesis, ∃ Gci1 , · · · , Gcin ∈ {Gci }
3 Gci1 ∩ · · · ∩ Gcin = φ . Thus by De Morgan’s law, X = φc = (Gci1 ∩ · · · ∩ Gcin )c =
Gcc cc
i1 ∪ · · · ∪ Gin = Gi1 ∪ · · · ∪ Gin . Accordingly, X is compact and so (ii) ⇒ (i).

Theorem 5.0.94. A topological space X is compact iff every class {Fi } of closed subsets
of X which satisfies the finite intersection property, has itself a non-empty intersection.

Proof. Exercise.

Lemma 5.0.95. Let A be a compact subset of a Hausdorff space X and suppose p ∈

X \ A . Then ∃ open sets G, H 3 p ∈ G , A ⊂ H , and G ∩ H = φ .

Proof. Let a ∈ A . Since p 6∈ A , p 6= a . By hypothesis, X is Haursdorff; hence ∃ open

sets Ga , Ha 3 p ∈ Ga , a ∈ Ha and Ga ∩ Ha = φ . Hence A ⊂ ∪{Ha : a ∈ A} , that
is, {Ha : a ∈ A} is an open cover of A . But A is compact, so ∃ Ha1 , · · · , Han ∈ {Ha }
3 A ⊂ Ha1 ∪ · · · ∪ Han . Now let H = Ha1 ∪ · · · ∪ Han and G = Ga1 ∩ · · · ∩ Gan . Thus
H and G are open since they are respectively the union and finite intersection of open
sets. Furthermore, A ⊂ H and p ∈ G since p belongs to each Gai individually. Lastly
we claim that G ∩ H = φ . Note first that Gai ∩ Hai = φ ⇒ G ∩ Hai = φ . Thus, by
the distributive law, G ∩ H = G ∩ {Ha1 ∪ · · · ∪ Han } = (G ∩ Ha1 ) ∪ · · · ∪ (G ∩ Han ) =
φ ∪ ··· ∪ φ = φ.

Corollary 5.0.96. Let A be a compact subset of a Hausdorff space X , if p 6∈ A , then

there is an open set G 3 p ∈ G ⊂ Ac .

Proof. By Lemma (5.0.95), ∃ open sets G and H 3 p ∈ G, A ⊂ H and G ∩ H = φ .

Hence G ∩ A = φ and p ∈ G ⊂ Ac .

Theorem 5.0.97. If A is a compact subset of a Hausdorff space X , then A is closed.

Proof. We prove equivalently, that Ac is open. Let p ∈ Ac ; that is, p 6∈ A . Then by

Corollary (5.0.96) ∃ an open set Gp 3 p ∈ Gp ⊂ Ac . Hence Ac = ∪{Gp : p ∈ Ac } .
Thus Ac is open as it is the union of open sets. Hence A is closed.

Exercise 5.0.98. 1. Let B and B ∗ be bases, respectively, for topology J and J ∗ on a

set X . Suppose that each B ∈ B is the union of members of B ∗ . Show that J is
coarser than J ∗ .

2. Let S be a subbase for a topology J on X and let A be a subset of X . Show that

the class SA = {A ∩ S : S ∈ S} is a subbase for the relative topology JA on A .

3. Show that a topology J on X is finite iff J has a finite base.

 22

4. Show that if S is a subbase for a topology J on X , then S \ {X, φ} is also a subbase

for J .
Theorem 5.0.99. Let J be the cofinite topology on any set X . Then (X, J) is a
compact space.

Proof. Let G = {Gi } be an open cover of X . Choose G0 ∈ G . Since J is the cofinite

topology, Gc0 is a finite set, say Gc0 = {a1 , · · · , an } . Since G is a cover of X , for each
ak ∈ Gc0 ∃ Gik ∈ G 3 ak ∈ Gik . Hence Gc0 ⊂ Gi1 ∪ · · · ∪ Gin and X = G0 ∪ Gc0 =
G0 ∪ Gi1 ∪ · · · ∪ Gin . Thus X is compact.
Theorem 5.0.100. Let A and B be disjoint compact subsets of a Hausdorff space X .
Then ∃ disjoint open sets G and H 3 A ⊂ G and B ⊂ H .

Proof. Let a ∈ A . Thus a 6∈ B , for A and B are disjoint. By hypothesis, B is

compact; hence by Lemma (5.0.95), ∃ open sets Ga and Ha 3 a ∈ Ga , B ⊂ Ha and
Ga ∩ Ha = φ . Since a ∈ Ga , {Ga : a ∈ A} is an open cover of A . Thus A is compact,
implies that we can select a finite number of the open sets, say, Ga1 , · · · , Gan ; such that
A ⊂ Ga1 ∪ · · · ∪ Gan . Furthermore, B ⊂ Ha1 ∩ · · · ∩ Han , since B is a subset of each
set. Now let G = Ga1 ∪ · · · ∪ Gan and H = Ha1 ∩ · · · ∩ Han . Observe by the above, that
A ⊂ G and B ⊂ H . In addition, G and H are open as they are the union and finite
intersection respectively of open sets. It now remains to show that G and H are disjoint.
Observe that for each i , Gai ∩ Hai = φ ⇒ Gai ∩ H = φ . Hence, by the distributive law,
G ∩ H = (Ga1 ∪ · · · ∪ Gan ) ∩ H = (Ga1 ∩ H) ∪ · · · ∪ (Gan ∩ H) = φ ∪ φ · · · ∪ φ = φ .
Corollary 5.0.101. Every compact Hausdorff space is normal.

Proof. Suppose X is both Hausdorff and compact and F1 and F2 are disjoint closed
subsets of X . By Theorem (5.0.91), F1 and F2 are compact and by Theorem (5.0.99),
F1 and F2 are subsets, respectively of disjoint open sets.
Theorem 5.0.102. Let f be a one-one continuous function from a compact space X
into a Hausdorff space Y . Then X and f (X) are homeomorphic.

Proof. Now f : X → f (X) is onto and by hypothesis 1-1 and continuous so f −1 :

f (X) → X exists. We must show that f −1 is continuous. Recall that f −1 is continuous
if for every closed subset F of X , (f −1 )−1 (F ) = f (F ) is a closed subset of f (X) . By
Theorem (5.0.91), the closed subset F of the compact space X is also compact. Since
f is continuous, then by Theorem (5.0.88), f (F ) is a compact subset of f (X) . But the
subspace f (X) of the Hausdorff space Y is also Haudorff, hence by Theorem (5.0.97),
f (F ) is closed. Accordingly, f −1 is continuous, so f : X → f (X) is a homeomorphism
and, X and f (X) are homeomorphic.
Definition 5.0.103. (Sequentially Compact Sets): A subset A of a topological space
X is sequentially compact iff every sequence in A contains a subsequence which converges
to a point in A .
 23

Theorem 5.0.104. A continuous image of a sequentially compact set is sequentially

compact.

Proof. Let f : X → Y be a continuous function and let A be a sequentially compact

subset of X . We want to show that f (A) is a sequentially compact subset of Y . Let
{bn } be a sequence in f (A) . Then ∃ a1 , a2 , · · · ∈ A 3 f (an ) = bn , ∀n ∈ N . But
A is sequentially compact, so the sequence {an } contains a subsequence {ain } which
converges to a point a0 ∈ A . Now f is continuous and hence sequentially continuous, so
{f (ain )} = {bin } converges to f (a0 ) ∈ f (A) . Thus f (A) is sequentially compact.

Definition 5.0.105. (Countably Compact Sets): A subset A of a topological space X

is countably compact iff every infinite subset B of A has an accumulation point of A .

Theorem 5.0.106. If X is compact, then X is countably compact.

Proof. Let A be a subset of X with no accumulation point in X . Then each point

p ∈ X belongs to an open set Gp which contains at most one point of A . Observe that
the class {Gp : p ∈ X} is an open cover of the compact space X and, hence, contains
a finite subcover, say, {Gp1 , · · · , Gpm } . Hence A ⊂ X ⊂ Gp1 ∪ · · · ∪ Gpm . But each
Gpi contains at most one point of A , hence A , a subset of Gp1 ∪ · · · ∪ Gpm can contain
at most m points, that is, A is finite. Accordingly, every infinite subset contains an
accumulation point in X , that is, X is countably compact.

Theorem 5.0.107. Let X be sequentially compact, then X is also countably compact.

Proof. Let A be any infinite subset of X . Then ∃ a sequence {an } in A with distinct
terms. Since X is sequentially compact, the sequence {an } contains a subsequence
{ain } (also with distinct terms) which converges to a point p ∈ X . Hence every open
neighborhood of p contains an infinite number of terms of the convergent subsequence
{ain } . But the terms are distinct, hence every neighborhood of p contains an infinite
number of points of A . Accordingly, p ∈ X is an accumulation point of A . In other
words, X is countably compact.

Theorem 5.0.108. Let A ⊂ X be sequentially compact. Then any countably open cover
of A is reducible to a finite cover.

Proof. We may assume A is infinite, for otherwise the proof is trivial. Suppose ∃ a
countable open cover {Gi : i ∈ N} with no finite subcover. We define the sequence
{an } As follows: Let n1 be the smallest positive integer 3 A ∩ Gn1 6= φ . Choose
a1 ∈ A ∩ Gn1 . Let n2 be the least positive integer larger than n1 3 A ∩ Gn2 6= φ .
Choose a2 ∈ (A ∩ Gn2 ) \ (A ∩ Gn1 ) . Such a point always exists, for otherwise Gn1 covers
A . Continuing in this manner we obtain the sequence {an } with the property that for
every i ∈ N , ai ∈ A ∩ Gni , ai 6∈ ∪n−1
j=1 (A ∩ Gnj ) and ni > ni−1 . We claim that {ai } has
no convergent subsequence in A . Let a ∈ A . Then ∃ Gi0 ∈ {Gi } 3 p ∈ Gi0 . Now
 24

A ∩ Gi0 6= φ , since p ∈ A ∩ Gi0 ; hence ∃ j0 ∈ N 3 Gj0 = Gi0 . But the choice of the
sequence {a} , i > J0 ⇒ ai 6∈ Gi0 . Accordingly, since Gi0 is an open set containing p ,
no subsequence of {ai } converges to p . But p was arbitrary, so A is not sequentially
compact.
Chapter 6

CONNECTEDNESS

Definition 6.0.109. (Separated Sets): Two subsets A and B of a topological space

X are said to be separated if:

(i) A and B are disjoint sets, and

(ii) neither contains an accumulation point of the other.

OR A and B are separated if A ∩ B̄ = φ and Ā ∩ B = φ .

Definition 6.0.110. (Disconnection): A subset, A , of a topological space X is dis-

connected if ∃ open subset G and H of X 3 A ∩ G and A ∩ H are disjoint non-empty
sets whose union is A . A set is connected if it is not disconnected.

Note: (i) Observed that A = (A∩G)∪(A∩H) iff A ⊂ G∪H and φ = (A∩G)∩(A∩H)

iff G ∩ H ⊂ Ac . Therefore G ∪ H is a disconnection of A iff A ∩ G 6= φ , A ∩ H 6= φ ,
A ⊂ G ∪ H and G ∩ H ⊂ Ac .
(ii) The empty set φ and singleton sets {p} are always connected.

Theorem 6.0.111. If A and B are non-empty separated sets, then A ∪ B is discon-

nected.

Proof. Since A and B are separated, A ∩ B̄ = φ and Ā ∩ B = φ . Let G = B̄ c and

H = Āc . Then G and H are open, and (A ∪ B) ∩ G = A and (A ∪ B) ∩ H = B are
non-empty disjoint sets whose union is A ∪ B . Thus G and H form a disconnection of
A ∪ B and so A ∪ B is disconnected.

Theorem 6.0.112. Let G ∪ H be a disconnection of A , then A ∩ G and A ∩ H are

separated sets.

Proof. Now A∩G and A∩H are disjoint, hence we need only show that each set contains
no accumulation point of the other. Let p be an accumulation point of A∩G and suppose

25
 26

p ∈ A ∩ H Then H is an open set containing p and so H contains a point of A ∩ G

distinct from p , that is, (A ∩ G) ∩ H 6= φ . But (A ∩ G) ∩ (A ∩ H) = φ = (A ∩ G) ∩ H .
Accordingly p 6∈ A∩H . Similarly, if p is an accumulation point of A∩H , then p 6∈ A∩H .
Thus A ∩ G and A ∩ H are separated.

Theorem 6.0.113. A set is connected iff it is not the union of two non-empty separated
sets.

Proof. We show, equivalently, that A is disconnected iff A is the union of two non-
empty separated sets. Suppose A is disconnected and let G ∪ H be a disconnection of
A . Then A is the union of non-empty sets A ∩ G and A ∩ H which are separated by
Theorem (6.0.112). On the other hand A is the union of two non-empty sets, then A is
disconnected by Theorem (6.0.111).

Theorem 6.0.114. Let G ∪ H be a disconnection of A and let B be a connected subset

of A . Then either B ∩ H = φ or B ∩ G = φ and so either B ⊂ H or B ⊂ G .

Proof. Now B ⊂ A , and so A ⊂ G∪H ⇒ B ⊂ G∪H and G∩H ⊂ Ac ⇒ G∩H ⊂ B c .

Thus if both B ∩ G and B ∩ H are non-empty, then G ∪ H is a disconnection of B .
But B is connected hence the conclusion.

Theorem 6.0.115. If A and B are connected sets which are not separated, then A ∪ B
is connected.

Proof. Suppose A ∪ B is disconnected and suppose G ∪ H is a disconnection of A ∪ B .

Since A is a connected subset of A ∪ B , either A ⊂ G or A ⊂ H by Theorem (6.0.114).
Similarly, either B ⊂ G or B ⊂ H . Now if A ⊂ G and B ⊂ H (or B ⊂ G and A ⊂ H ),
then by Theorem (6.0.112), (A ∪ B) ∩ G = A and (A ∪ B) ∩ H = B are separated sets.
But this contradicts the hypothesis; hence either A ∪ B ⊂ G or A ∪ B ⊂ H , and so
G ∪ H is not a disconnection of A ∪ B . In order words, A ∪ B is connected.

Theorem 6.0.116. Let A = {Ai } be a class of connected subsets of X 3 no two

members of A are separated. Then B = ∪i Ai is connected.

Proof. Exercise.

Theorem 6.0.117. Let A = {Ai } be a class of connected subsets of X with a non-empty

intersection. Then B = ∪i Ai is connected.

Proof. Exercise.

Theorem 6.0.118. Let A be a connected subset of X and let A ⊂ B ⊂ Ā . Then B is

connected and hence in particular, Ā is connected.
 27

Proof. Suppose B is disconnected and suppose G ∪ H is a disconnection of B . Now A

is a connected subset of B and so, by Theorem (6.0.114) either A∩H = φ or A∩G = φ ;
say A ∩ H = φ . Then H c is a closed super set of A and therefore A ⊂ B ⊂ Ā ⊂ H c .
Consequently, B ∩ H = φ . But this contradicts the fact that G ∪ H is a disconnection
of B ; and hence B is connected.

Theorem 6.0.119. Let X be a topological space, then the following conditions are equiv-
alent:

(i) X is disconnected.

(ii) There exists a non-empty proper subset of X which is both open and closed.

Proof. (i) ⇒ (ii) : Suppose X = G ∪ H , where G and H are non-empty and open.
Then G is a non-empty proper subset of X and, since G = H c , G is both open and
closed.
(ii) ⇒ (i) : Suppose A is a non-empty proper subset of X which is both open and
closed. Then Ac is also a non-empty and open, and X = A ∪ Ac . Accordingly, X is
disconnected.

Theorem 6.0.120. Let A be a subset of a topological space (X, J) and let JA be the
relative topology on A . Then A is J -connected iff A is JA -connected.

Proof. Suppose A is disconnected with G ∪ H forming a J -disconnection of A . Now,

G, H ∈ J and so A ∩ G , A ∩ H ∈ JA . Accordingly, A ∩ G and A ∩ H form a JA -
disconnection of A ; hence A is JA -disconnected.
On the other hand, suppose A is JA -disconnected, say G∗ and H ∗ form a JA -
disconnection of A . Then G∗ , H ∗ ∈ JA and so ∃ G, H ∈ J 3 G∗ = A∩G H ∗ = A∩H .
But A ∩ G∗ = A ∩ A ∩ G = A ∩ G and A ∩ H ∗ = A ∩ A ∩ H = A ∩ H . Hence G ∪ H is
a J -disconnection of A and so A is J -disconnected.

Theorem 6.0.121. Continuous images of connected sets are connected.

Proof. Let f be a continuous function from a connected space X into a topological space
Y . Thus f : X → f (X) is continuous. We show that f (X) is connected. Suppose f (X)
is disconnected, say G and H form a disconnection of f (X) . Then f (X) = G ∪ H and
G ∩ H = φ , and so X = f −1 (G) ∪ f −1 (H) and f −1 (G) ∩ f −1 (H) = φ . Since f is
continuous f −1 (G) and f −1 (H) are open subsets of X and hence form a disconnection
of X , which is impossible. Thus if X is connected, so is f (X) .

Definition 6.0.122. (Component): A component E of a topological space X is a max-

imally connected subset of X ; that is, E is connected and E is not a proper subset of
any connected subset of X .

Theorem 6.0.123. Every component E is closed.

 28

Proof. Now E is connected and so by Theorem (6.0.118), Ē is connected. E ⊂ Ē . But

E , a component, is a maximally connected set; hence E = Ē , and so E is closed.

Theorem 6.0.124. Let p ∈ X and A = {Ai } be the class of connected subsets of X

containing p . Furthermore, let Cp = ∪i Ai , then:

1. Cp is connected.

2. If B is a connected subset of X containing p , then B ⊂ Cp .

3. Cp is a maximally connected subset of X , that is, a component.

Proof. (i) Since each A ∈ A contains p , p ∈ ∩i Ai and so by Theorem (6.0.117) Cp is

connected.
(ii) If B is a connected subset of X containing p , then B ∈ Ap and so B ⊂ Cp =
∪i {Ai : Ai ∈ Ap } .
(iii) Let Cp ⊂ D , where D is connected. Then p ∈ D and hence, by (ii) D ⊂ Cp , that
is, Cp = D. Therefore Cp is a component.

Theorem 6.0.125. The component of X form a partition of X . Every connected subset

of X is contained in some components.

Proof. Consider the class C = {Cp : p ∈ X} , where Cp is as defined in Theorem

(6.0.124). We claim that C consists of the components of X . By Theorem (6.0.124),
each Cp ∈ C is a component. On the other hand, if D is a component, then D contains
some point p0 ∈ X and so D ⊂ Cp0 . But D is a component; hence D = Cp0 .
We now show that C is a partition of X . Clearly, X = ∪{Cp : p ∈ X} ; hence we
need only show that distinct components are disjoint or equivalently, if Cp ∩ Cq 6= φ ,
then Cp = Cq . Let a ∈ Cp ∩ Cq . Then Cp ⊂ Ca and Cq ⊂ Ca , since Cp and Cq are
connected sets containing a . But Cp and Cq are components, hence Cp = Ca = Cq .
Lastly, if E is a non-empty connected subset of X ; then E contains a point p0 ∈ X ,
and so E ⊂ Cp0 , by Theorem (6.0.124). If E = φ ; then E is contained in every
component.

Definition 6.0.126. (Path): Let I = [0, 1] , be the closed unit interval. A path from a
point a to a point b in a topological space X is a continuous function f : I → X with
f (0) = a and f (1) = b . Here a is called the initial point and b is called the terminal
point of the path.

Definition 6.0.127. (Arcwise Connected Sets): A subset A of a topological space X

is said to be arcwise connected if for any two point a, b ∈ A ∃ a path f : I → X from a
to b which is contained in A ; that is, f (I) ⊂ A . The maximal arcwise connected subsets
of X are called arcwise connected components, and they form a partition of X .

Theorem 6.0.128. Let f : I → X be any path in X . Then f (I) is connected.

 29

Proof. I = [0, 1] is connected and f is continuous, hence by Theorem (6.0.121), f (I) is

connected.

Theorem 6.0.129. Continuous images of arcwise sets are arcwise connected.

Proof. Let E ⊂ X be arcwise connected and let f : X → Y be continuous. We claim

that f (E) is arcwise connected. For let p, q ∈ f (E) , then ∃ p∗ , q ∗ ∈ E 3 f (p∗ ) = p
and f (q ∗ ) = q . But E is arcwise connected and so ∃ a path g : I → X 3 g(0) = p∗ ,
g(1) = q ∗ and g(I) ⊂ E . Now the composition of continuous functions is continuous and
so f ◦ g : I → Y is continuous. Furthermore, f ◦ g(0) = f (p∗ ) = p , f ◦ g(1) = f (q∗) = q
and f ◦ g(I) = f (g(I)) ⊂ f (E) . Thus f (E) is arcwise connected.

Theorem 6.0.130. Every arcwise connected set is connected.

Proof. If A is empty, then A is connected. Suppose A is not empty; say, p ∈ A . Now

A is arcwise connected and so, for each a ∈ A , there is a path fa : I → A from p to
a . Furthermore, a ∈ fa (I) ⊂ A and so A = ∪{fa (I) : a ∈ A} . But p ∈ fa (I) for every
a ∈ A ; hence ∩{fa (I) : a ∈ A} is non-empty. Moreover, each fa (I) is connected and so
by Theorem (6.0.117), A is connected.