Section17: Closed set and Limit points

Definition:
Let 𝐴 be a topological space. 𝐴 subset 𝐴 of 𝑋 is said to be closed if
𝑋 - 𝐴is open, i.e., the complement of 𝐴is open in 𝑋 .

Definition:
Let 𝑌 be a subspace of a topological space 𝑋 . 𝐴 subset 𝐴 of Y is said
to be Closed if 𝐴 is Closed in the subspace topology of𝑌 , i.e., 𝑌 - 𝐴is
open in 𝑌 .

Example:
(1) The subset [ 𝑎, 𝑏 ] of 𝑅 is closed because its complement
= ( ∞ , 𝑎) ∪ ( 𝑏 ,+∞ ) is open. Similarly, [ 𝑎,+ ∞ ) is closed because its
𝑅 − [ 𝑎 ,𝑏 ]
complement ( − ∞ , 𝑎 ) is open. The subset R is neither open nor closed.
(2) In the plane 𝑅2, the set { 𝑥 × 𝑦 ; 𝑥 ≥ 0 𝑎𝑛𝑑 𝑦 ≥ 0 } is closed because its
complement is the union of the two sets ( − ∞ , 0 ) × 𝑅 and 𝑅 × ( − ∞ ,0 ), each
of which is a product of open sets of 𝑅 and is therefore open in 𝑅2.
(3) In the finite complement topology on a set 𝑋 , the closed sets
consist of 𝑋 itself and all finite subsets of 𝑋 .
(4) In the discrete topology on the set 𝑋 ,every set is open; it follows
that every set is closed as well.
(5) Let us consider the subset of the real line 𝑌 =[ 0 , 1 ] ∪ ( 2 ,3 ) in the
subspace topology.
In this space, the set [ 0 , 1 ] is open since it is the intersection of the open
( 1 3
)
set − 2 , 2 of 𝑅 with 𝑌 . Similarly, ( 2 , 3 ) is open as a subset of 𝑌 ; it even
open as a subset of 𝑅.
 Since [ 0 , 1 ] and ( 2 , 3 ) are complements in 𝑌 of each other, we conclude
that both [ 0 , 1 ] and ( 2 , 3 ) are closed as subset of 𝑌 .

Theorem (5):
Let 𝑋 be a topology space, then the following conditions hold:
(i) 𝜙and 𝑋 are closed
(ii) Arbitrary intersections of closed sets are closed
(iii) Finite unions of closed sets are closed

Proof:
(i) Since 𝜙 and 𝑋 are the complements of the open sets 𝑋 and 𝜙
respectively, 𝜙 and 𝑋 are closed.
(ii) Let { 𝐴𝛼 }𝛼𝜖 𝑗 be a collection of closed sets. By applying De
Morgan’s law, we have 𝑋 −∩𝛼𝜖 𝑗 𝐴𝛼=∪𝛼𝜖 𝑗 ( 𝑋 − 𝐴𝛼 ) Since the sets 𝑋 − 𝐴𝛼 are
open, is open and so 𝑋 −∩ 𝛼𝜖 𝑗 𝐴𝛼 is open.
Consequently ∩𝛼𝜖 𝑗 𝐴𝛼 is closed, i.e., arbitrary intersections of closed
sets are closed.
(iii) Let{ 𝐴𝑖 }𝑖=1 be a collection of closed sets. By De Morgan’s law,
𝑛

we have 𝑋 −∩𝑛 𝑖=1 𝐴𝑖=∪𝑛 𝑖=1 ( 𝑋 − 𝐴𝑖)

Since the sets 𝑋 − 𝐴𝑖 are open, ∩𝑛𝑖=1 ( 𝑋 − 𝐴𝑖 ) is open and so 𝑋 −∩𝑛𝑖=1 𝐴𝑖 is
❑ ❑

open. Consequently, ∩𝑛𝑖=1 𝐴𝑖 is closed, i.e., finite unions of closed set

❑

are closed.

Theorem (6):
Let 𝑌 be a subspaces of a topological space 𝑋 . Then a set 𝐴 is
closed in 𝑌 if and only if it equals the intersection of a closed set of 𝑋
with 𝑌 , i.e., there exists a closed set 𝐶 in 𝑋 such that 𝐴

Proof:
Let 𝑌 be subspace of a topological spaces 𝑋 . . Suppose a set 𝐴 is
closed in 𝑌 .
Then 𝑌 − 𝐴 is open in 𝑌 and so by definition of the subspace topology
𝑌 − 𝐴=𝑌 ∩𝑈 for some open set 𝑈 of 𝑋 .
 Since 𝑈 is open in 𝑋 , is closed in 𝑋 .

Now 𝑌 ∩ ( 𝑋 − 𝑈 )=( 𝑌 ∩ 𝑋 ) − ( 𝑌 ∩𝑈 )
¿ 𝑌 − (𝑌 ∩ 𝐴)

¿𝐴

Hence 𝐴 equals the intersection of a closed set of 𝑋 with 𝑌 .

Conversely suppose for some closed set 𝐶 in 𝑋 .

Since 𝐶 is closed in 𝑋 , 𝑋 − 𝐶is open in 𝑋 .

Now ( 𝑋 − 𝐶 ) ∩𝑌 =( 𝑋 ∩𝑌 ) − ( 𝐶 ∩𝑌 )
¿𝑌 − 𝐴

Since 𝑋−𝐶 is open in 𝑋 , be definition of the subspace topology,

( 𝑋 − 𝐶 ) ∩𝑌 is open in 𝑌 and so 𝑌 − 𝐴is open in 𝑌 .

Hence 𝐴 is closed in 𝑌 .

Theorem 7:

Let 𝑌 be a subspace of a topological space 𝑋 . If 𝐴 is closed in 𝑌

and 𝑌 is closed in 𝑋 , then is closed in 𝑋 .

Proof:

Let 𝑌 be a subspace of a topological space 𝑋 . Suppose 𝐴is closed

in 𝑌 and 𝑌 is closed in 𝑋 .
 Since 𝐴is closed in 𝑌 , be the theorem (6). There exists a closed set
𝐶 in 𝑋 such that 𝐴=𝐶 ∩ 𝑌 .

Since 𝐶 is closed in 𝑋 and 𝑌 is closed in 𝑋 , 𝐴 is closed in 𝑋 .

Definition:

Let 𝑋 be a topological space. Let 𝐴 be a subset of 𝑋 .

The Interior of 𝐴 is defined as the union of all open sets contained
in 𝐴. The interior of 𝐴 is denoted by Int 𝐴
and 𝐴𝑜.
The closure of 𝐴❑ is defined as the intersection of all closed sets
containing 𝐴❑. The closure of 𝐴❑ is denoted by cl 𝐴❑ and 𝐴.

Note:
(i) 𝐴𝑜 is open
(ii) 𝐴 is closed
(iii) 𝐴𝑜 ⊆ 𝐴⊆ 𝐴
(iv) 𝐴 is open if and only if 𝐴= 𝐴𝑜
(v) 𝐴 is closed if and only if 𝐴= 𝐴

Theorem (8):
Let 𝑌 be a subspace of 𝑋 . Let 𝐴 be a subset of 𝑌 . Let 𝐴 denote
the closure of 𝐴 in 𝑋 . Then the closure of 𝐴 in 𝑌 equals 𝐴∩𝑌 .

Proof:
Let 𝑌 be a subspace of 𝑋 and let 𝐴 be subset of 𝑌 .
Let 𝐴 denote the closure of 𝐴 in 𝑋 and let 𝐵 denote the closure
of 𝐴 in 𝑌 .
Since 𝐴 is closed in 𝑋 , by the theorem (6) 𝐴∩𝑌
Since 𝐴⊆ 𝐴 and 𝐴⊆ 𝑌 , we have 𝐴⊆ 𝐴∩𝑌 .
 Since 𝐴∩𝑌 is a closed set in 𝑌 Containing 𝐴 and 𝐵 is the
intersection of all closed subsets of𝑌 Containing, we have 𝐵⊆ 𝐴 ∩𝑌
On the other hand, since 𝐵is closed in 𝑌 , by the theorem (6), 𝐵=𝐶 ∩𝑌
for some closed set𝐶 in 𝑋 .
Then 𝐶 is a closed set of 𝑋 containing 𝐴. Since 𝐴 is the intersection
of all closed sets containing 𝐴 , we conclude that 𝐴⊆𝐶 .
Hence 𝐴∩𝑌 ⊆ 𝐶 ∩𝑌 =𝐵
Form (1) and (2), 𝐵= 𝐴∩ 𝑌 .

Definition:
A set 𝐴is said to intersect a set 𝐵 if 𝐴∩ 𝐵 ≠ 𝜙 .

Definition:
Let 𝑋 be a topological space and let 𝑥 𝜖 𝑋 .Any open set containing
𝑥 is called a neighbourhood of 𝑥 .

Theorem(9)
Let 𝐴 be a subset of a topological space 𝑋 .

(i) 𝑥𝜖 𝐴 if and only if every open set𝑈 containing intersects A,

i.e., every neighbourhood of 𝑥intersects 𝐴.

(ii) If 𝛽 is a basis for the topology of 𝑋 , then 𝑥 𝜖 𝐴 if and only if

every basis element containing 𝑥 intersects 𝐴.

Proof:
Let 𝐴 be a subset of a topology space x.
(i) Suppose 𝑥 𝜖 𝐴. We shall show that every open set
containing 𝑥 intersects 𝐴.

Suppose there exists an open set𝑈 containing 𝑥 which dose not

intersects 𝐴.
Then 𝐴⊆ 𝑋 − 𝑈
Since𝑈 is open, 𝑋 −𝑈 is closed
Hence 𝑋 −𝑈 is a closed set containing 𝐴.
 Since 𝐴 is the intersection of all closed sets containing 𝐴,
Since 𝑥 𝜖 𝑈 , 𝑥 ∉ 𝑋 −𝑈 and so 𝑥 ∉ 𝐴, which is a contradiction.
Thus, every open set 𝑈 containing 𝑥 intersects 𝐴.
Conversely suppose every open set containing 𝑥 intersects 𝐴.
Now, we have to prove that 𝑥 𝜖 𝐴.suppose 𝑥 ∉ 𝐴 .
Then 𝑥 𝜖 𝑋 − 𝐴. Since 𝐴 is closed, 𝑋 − 𝐴 is open.
Hence 𝑋 − 𝐴 is an open set containing 𝑥.
Since 𝐴⊆ 𝐴, ( 𝑋 − 𝐴) ∩ 𝐴=𝜙 .
Hence 𝑋 − 𝐴 is an open set containing 𝑥 but not intersecting 𝐴.
This is an contradiction to our assumption that 𝑥 ∉ 𝐴.
Hence 𝑥 ∉ 𝐴.
(ii) Suppose 𝑥 𝜖 𝐴.
By (i), every open set containing 𝑥 intersects 𝐴
Since every basis element is open , every basis element containing 𝑥
intersects 𝐴.
Conversely suppose every basis element containing 𝑥 intersects 𝐴.
We shall show that 𝑥 𝜖 𝐴.
By(i), it suffices to show that every open set containing 𝑥
intersects 𝐴.
Let v be an open set containing 𝑥. then 𝑉 =𝑈 𝛼𝜖 𝐽 𝐵𝛼 , where 𝐵𝛼 𝜖 𝛽.
Since 𝑥 𝜖 𝑉 , 𝑥 𝜖 𝐵𝛼 for some 𝛼 0.
0

Hence 𝐵𝛼 is a basis element containing 𝑥.

0

By our assumption, 𝐵𝛼 ∩ 𝐴≠ 𝜙 . Consequently 𝑉 ∩ 𝐴 ≠ 𝜙 .

0

Since 𝑉 is an arbitrary open set containing 𝑥 , every open set

containing 𝑥 intersects A. Hence by (i), 𝑥 𝜖 𝐴.

Definition:
Let 𝐴 be a subset of a topological space 𝑋 . A point 𝑥 𝜖 𝑋 is Sai to
be a limit point of 𝐴 (or cluster point or accumulation pion) if every
neighbourhood 𝑈 of 𝑥 intersects 𝐴 in some point other than 𝑥 itself,
i.e., every neighbourhood of 𝑥 contains a point of 𝐴 different from 𝑥,
( 𝑈 ∩ 𝐴 ) { 𝑥 } ≠ 𝜙 . In otherwords 𝑥 is a limit point of 𝐴 if it belongs to the
closure of 𝐴− { 𝑥 }. The point 𝑥 may lie in 𝐴 or not.

Example:
 (i) Let us consider the real line 𝑅 . If 𝐴=( 0,1 ] , then the point 0 is a
limit point of 𝐴, but no other point of 𝑅 is limit point do 𝐴.
(ii) If 𝐵=¿, then 0 is the only limit point of 𝐵. Every other point 𝑥 of
𝑅 has a neighbourhood that either does not intersects of 𝐵 at all, or it
intersects 𝐵 only in the point 𝑥itself.
(iii) If 𝐶= { 0 } ∪ ( 1,2 ), then the limit point of 𝐶 are points of the interval
[ 1,2 ].
(iv) If 𝑄 is the set of rational numbers, then every point of 𝑅 is a limit
open of 𝑍 +¿ ¿.
(v) If 𝑍 +¿ ¿ is the set positive integers, then no point of 𝑅 is a limit
point of 𝑍 +¿ ¿.
(vi) If 𝑅+¿ ¿ is the set of positive reals, then every point of { 0 } ∪ 𝑅+¿¿ is a
limit point of 𝑅+¿ ¿.

Theorem 10:
Let 𝐴 be a subset of a topological space 𝑋 and let 𝐴′ be the set of
all limit points of 𝐴. Then 𝐴= 𝐴⊏ 𝐴′.
Proof:
Let 𝐴 be a subset of a topological space 𝑋 and let 𝐴′ be the set of
all limit points of 𝐴.
Let 𝑥 𝜖 𝐴′. Then every neighbourhood of 𝑥 intersects 𝐴 in a point
different from 𝑥.
By the theorem (9), 𝑥 𝜖 𝐴. Hence 𝐴′ ⊏ 𝐴.
Moreover, since 𝐴⊏ 𝐴, it follows that 𝐴∪ 𝐴′ ⊏ 𝐴 ----------(1)
For the reverse inclusion, let 𝑥 𝜖 𝐴.
We shall show that 𝑥 𝜖 𝐴∪ 𝐴′. If 𝑥 𝜖 𝐴, then 𝑥 𝜖 𝐴∪ 𝐴′
Suppose that 𝑥 ∉ 𝐴
Since 𝑥 ∉ 𝐴, every neighbourhood 𝑈 of 𝑥 intersects 𝐴 , ie.,
𝑈∩ 𝐴≠𝜙.
Since 𝑥 ∉ 𝐴, the set 𝑈 must intersect 𝐴 in a point different from 𝑥,
ie., 𝑈 ∩ 𝐴 must contain a point of 𝐴 which is not 𝑥.
Then 𝑥 𝜖 𝐴′ and so 𝑥 𝜖 𝐴∪ 𝐴′. So in the both cases, 𝑥 𝜖 𝐴∪ 𝐴′ -(2)
Thus from (1) and (2), 𝐴= 𝐴∪ 𝐴′.
Corollary 1:
 A subset of a topological space is closed if and only if it contains
all its limit points.

Proof:
Let 𝐴be a subset of a topological space 𝑋 .
The set 𝐴 is closed ⟺ 𝐴= 𝐴
⟺ 𝐴⊂ 𝐴
Definition:
A topological space 𝑋 is called a Hausdorff space if for each pair
𝑥1, 𝑥2 of distinct points of 𝑋 , there exist neighbourhoods 𝑈 1 and 𝑈 2 of 𝑥1
and 𝑥2 respectively such that 𝑈 1 and 𝑈 2 are disjoint, ie., 𝑈 1 ∩𝑈 2=𝜙.

Theorem 11:
Every finite point set in a Hausdorff space 𝑋 is closed.

Proof:
It suffices to show that every one point set { 𝑥 0 } is closed.
If 𝑥 is a point of 𝑋 different from 𝑥0 , then 𝑥 and 𝑥0 have disjoint
neighbourhood 𝑈 and 𝑉 repectively.
Since the neighbourhood 𝑈 of 𝑥 does not intersect { 𝑥 0 }, the point 𝑥
is not a limit point of { 𝑥 0 } and so the point 𝑥 cannot belong to the
closure of the set { 𝑥 0 }.
Consequently, the closure of the set { 𝑥 0 } is { 𝑥 0 } itself and so { 𝑥 0 }
is closed.

Result 2:
Every finite set in a Hausdorff space is closed.

Proof:
Since every finite set is the finite union of singleton sets and
every singleton set is closed by the theorem (11), we see that finite
sets are closed.

Note:
The condition that finite point sets are closed, is in fact weaker
 than the Hausdorff condition. For example, the real line 𝑅 in the
finite complement topology is not a Hausdorff space, but it is a space
in which finite point sets are closed.

The condition that finite point sets are closed is called the
𝑇 1 - axiom.
Theorem 12:
Let 𝑋 be a space satisfying the 𝑇 1 – axiom and let 𝑥 𝜖 𝑋 .
Let 𝐴 be a subset of 𝑋 . Then the point x is a limitpoint of 𝐴 if and
only if every neighbourhood of 𝑥 contains infinitely many points of
𝐴.
Proof:
If every neighbourhood of 𝑥 contains infinitely many points of
𝐴, ie., every neighbourhood of 𝑥 intersects 𝐴 in infinitely many
points, it certainly intersects 𝐴 in some point other than 𝑥 itself.
Hence 𝑥 is a limit point of 𝐴.
Conversely suppose that 𝑥 is a limit point of 𝐴. Let 𝑈
be a neighbourhood of 𝑥.

We have to show that 𝑈 cantains infinitely many points of 𝐴.

Suppose that 𝑈 contains only a finite number of points of 𝐴 , ie., 𝑈
intersects 𝐴 in only finitely many points.
Than 𝑈 also intersects 𝐴− { 𝑥 } in finitely many points.

Let { 𝑥 1 , 𝑥2 , … … … … , 𝑥 𝑚 } be the points of 𝑈 ∩ ( 𝐴− { 𝑥 }).

Since the finite point set { 𝑥 1 , 𝑥2 , … … … … , 𝑥 𝑚 } is closed, the set

𝑋 − { 𝑥 1 , 𝑥 2 , … … … … , 𝑥𝑚 } is an open set of 𝑋 containing 𝑥 .

Hence 𝑈 ∩ ( 𝑋 − {𝑥 1 , 𝑥2 , … … … … , 𝑥 𝑚 }) is a neighbourhood of 𝑥 containing

no point of 𝐴 other than 𝑥

This implies that 𝑥 is not a limit point of 𝐴.

This is a contradiction that 𝑥 is a limit point of 𝐴.
Since 𝑈 is an arbitray neighbourhood of 𝑥 , every
neighbourhood of 𝑥 contains infinitely many points of 𝐴.
 Theorem (13) :
If 𝑋 is Hausdorff space, then a sequence
of points of 𝑋 converges to atmost one point of 𝑋 .

Proof:
Let be a Hausdorff space.
𝑋
Suppose that ¿ 𝑥𝑛 > ¿ is a sequence of point of 𝑋
converging to two different points x and y.

Let 𝑈 and 𝑉 be disjoint neighourhoods of x and y respectively.

Since 𝑈 contains for all but finitely many values of n, the 𝑉
cannot.
This contradicts that 𝑉 is a neighbourhood of 𝑥.
Hence ¿ 𝑥𝑛 >¿ can not converge to two different pointa 𝑥 and 𝑦 .

Definition :
If a sequence ¿ 𝑥𝑛 >¿ of points of a Hausdrorff space 𝑋 converges
to a point 𝑥 of 𝑋 , we often write 𝑥𝑛 → 𝑥 and we say that 𝑥 is the limit
of the sequence ¿ 𝑥𝑛 > ¿.

Theorem (14) :
Every simply ordered set is a Hausdorff space in the order
topology. The product of two Hausdroff space is a Hausdrorff space.
A subspace of a Hausdorff space is a Hausdorff space.

Proof: