MUDGIL ACADEMY OF MATHEMATICS

Contact No. +91- 8607383607 Email : [email protected]

MSc. Pre (Topology) Unit - I Page 31

Theorem : Let (X,  ) be a topological space. Let b(A) be the boundary of A, then

(i) b(A) = A  (X  A) (ii) b(A) = A – A o

(iii) X – b(A) = A o  (X  A)o (iv) A = A  b(A)

(v) A o = A – b(A) (vi) A is closed iff b(A)  A.

(vii) A is open iff b(A)  A =  (viii) A is both open and closed iff b(A) = 

(ix) b(A  B)  b(A)  b(B) (x) b(A  B)  b(A)  b(B)

(xi) b(A) = b(X – A) (xii) If A is an open set, then b(A) = A – A.

Proof : (i) x  b(A) iff x  A o and x  ( X  A)o

iff x  X – A o and x  X – ( X  A)o

iff x  (X  A) and x  X  ( X  A)

iff x  (X  A) and x  A

iff x  A  (X  A)

Hence, b(A) = A  (X  A)

(ii) By (i), we have , b(A) = A  (X  A)

 b(A)  A

By def. of b(A) , b(A)  A o = 

 b(A)  A – A o ......(*)

Conversely, let x  A – A o  x  A and x  A o

 x  A and x  X – A o

 x  A and x  ( X  A)

 x  A  (X  A) = b(A)

 A – A o  b(A) ......(**)

By (*) and (**), b(A) = A – A o .

(iii) Let x  X – b(A) iff x  X and x  b(A)

iff x  A o or x  ( X  A)o

iff x  A o  ( X  A)o

So X – b(A) = A o  ( X  A)o

(iv) By (i), we have : b(A) = A  (X  A)

so A  b(A) = A  [ A  (X  A) ]
 32

= [A  A ]  [ A  (X  A) ] ......(1)

Now, A  (X  A)  A  (X – A) = X

But A  (X  A)  X , so A  (X  A) = X.

Using in (1), we get : A  b(A) = A  X = A .

(v) We have : A o  A and A o  b(A) = 

Thus, A o  A – b(A) ......(2)

Conversely, let x  A – b(A)  xA and x  b(A)

Now, x  b(A)  x  Ao or x  ( X  A)o .

If x  ( X  A)o , then x  X – A [ ( X  A)o  X – A ]

 x  A, a contradiction,

so x  ( X  A)o  x  A o , so A – b(A)  A o ......(3)

By (2) and (3), we get : A o = A – b(A)

(vi) Let A be closed. Now by (i), we have b(A) = A  (X  A)

But A is closed, so A = A, so b(A) = A  (X  A)

 b(A)  A

Conversely, suppose that b(A)  A  A - Ao  A [By (ii)]

 ( A - Ao )  Ao  A  Ao = A [ Ao  A]

 A  A

but A  A , so A = A i.e. A is closed.

(vii) If A is open , then Ao = A. Also b(A) = A - Ao . So b(A)  A = ( A - Ao )  Ao =  .

Conversely, suppose that b(A)  A =  i.e. ( A - Ao )  A = 

 A  (X- A )  A =  [ A-B = A  (X-B)]

 A  A  (X- Ao ) = 

 A  (X- Ao ) =  [ A  A ]

 A- Ao = 

 A = Ao
 A is open .

(viii) Let A be both open and closed , then A = A and Ao = A.

Now b(A) = A  Ao = A  A =   b(A) =  .

Conversely, let b(A) =  .

Now by (iv) , Ao =A  b(A) = A   = A. So A is closed .

Also by (v), Ao = A b(A) = A  =A. So A is open.

MUDGIL ACADEMY OF MATHEMATICS
Contact No. +91- 8607383607 Email : [email protected]

MSc. Pre (Topology) Unit - I Page 33

(ix) We know that b(A) = A   X-A

 b(A  B) =  A  B   X-(A  B)

= [ A  B ]  (X-A)  (X-B) [By De Morgan’s law]

 [ A  B ]  [  X-A   X-B ]

= [ A   X-A   X-B ]  [ B   X-A   X-B ]

= [b(A)   X-B ]  [b(B)   X-A ] [ b(A)= A   X-A ]

 b(A)  b(B)
 b(A  B)  b(A)  b(B)

(x) We know that b(A) = A   X-A

 b(A  B) = A  B   X-(A  B)

= A  B  (X-A)  (X-B) [By De Morgan’s law]

= A  B  [  X-A   X-B ]

 [ A  B ]  [  X-A   X-B ]

=[ A  B   X-A ]  [ A  B   X-B ]

=[b(A)  B ]  [b(B)  A ]
 b(A)  b(B)
 b(A  b)  b(A)  b(B).

(xi) We have : b(X-A) =  X-A   X-(X-A) [By property (i)]

=  X-A  A = A   X-A = b(A).

(xii) Let A be an open set so that Ao = A. Now we have:

b(A) = A - Ao = A -A ( Ao = A). This completes the proof.

Example : Show that in a discrete topological space, every set has empty boundary .
Solution : Let (X,  ) be a discrete topological space and let A be any subset of X.

Now b(A) = A   X-A .

Since in a discrete topological space each set is closed, so A =A and  X-A =X-A.

Thus we have: b(A)= A  (X-A) =  .

Thus, each subset of X has empty boundary.

Bases and Subbases

 34
Base for a topology : Let (X,  ) be a topological space. A family  of  is called base for  iff each

member of  can be written as union of members of  .

Remark : It is not necessary to include the empty set in a base for a topology , since

= { B

:   } i.e.  is the union of empty family of members of  .

Examples : (1) Let X={a,b,c,d} and  ={  ,{a},{b},{a,b},{c,d},{a,c,d},{b,c,d},X} . Then the subfamily ,

 ={{a},{b},{c,d}} is a base for  .

(2) Let X={a,b}and  ={  ,{a},{b},{a,b},X}, then the subfamily  of  , where  is a base for  .

(3) Let (X,  ) be discrete topological space , then the subfamily  ={{x}:x  X} of  is for  .

Theorem : Let (X,  ) be a top space , then a subfamily  of  is a base for G   and any x  G,

there exist a B   s.t. x  B  G.

Proof : Suppose  is a base for  and let G   and x  G. Then by definition of base , G can be written

as union of members of  i.e. G= B 

 : B 

But x  G, so x  B

 : B   There exist some B   s.t. x    G.

Conversely , Suppose that given conditions holds . We shall prove that  is base for  . Let G   be

any arbitrary member .Now let X  G , then by given condition, there exist a set Bx   s.t.

x  Bx  G   {x}   B
xG xG
x G

 G B
xG
x G

 G= B
xG
x

i.e. G is union of members of  . Hence  is a base for  .

Remark : (1) A topological space (X,  ) can have more than one base but we are always interested in
abase which contains a few members as possible .
(2) A given family  of subsets of X can be a base for one and only one topology on X.

Proof : Let  and  ' be topologies on X for which  is a base . We shall prove that  =  ' .

Let G   be an arbitrary member , then G is union of members of  . But since  is base for  ' so  is

subfamily of  ' and therefore G is the union of members of  ' , which in turn , shows that G   ' . Hence
   ' . Similarly ,  '   .

Note : Can every family of subsets of X be a base for some topology ? The answer is negative. For
example, if X={a,b,c} and  ={{a,b},{c,d}}, then this  can not be a base for a topology on X because if 

is a base for some topology on X, then {a} must belong to  , since {a, b} and {a,c}
MUDGIL ACADEMY OF MATHEMATICS
Contact No. +91- 8607383607 Email : [email protected]

MSc. Pre (Topology) Unit - I Page 35

are open sets and by axiom [O2] , their intersection {a} must be open and hence be member of  .

Now, question arises that is there some condition for a family to be a base. In the next
theorem, we have a discussion on the same .

Theorem: A family  of sets is a base for a topology for the set X=  {B : B   } iff for any B1, B2

  and every x  B1  B2 , i.e. the intersection of any two members of  is a union of members of

.

Proof : Suppose  is a base for the topology  on X . Let B1 , B2   as base is always a subfamily of

 . Then by [O2] , B1  B2   and so B1  B2 is the union of members of  as  is a base for  .

Hence the intersection of any two members of  is a union of members of  .

Conversely, Suppose  is a family of sets satisfying the conditions of theorem . Let  be the collection

of all subsets of X which are union of members of  i.e.

 = {G : G is union of members of  }

Since each member of  is a union of  , then the union of any number of members of  is a union of

members of  and so belongs to  . Thus [O1] is satisfied .

Now, suppose that G1 ,G2   be any two members of  .We shall prove that G1  G2   .Let x  G1  G2,

then x  G1 and x  G2. Now by definition of  , G1 and G2 are members of  and so  B1 and B2 belonging

to  s.t. x  B1  G1 and x  B2  G2  x  B1  B2  G1  G2. By the given condition , there exist Bx

 G1  G2  
xG1G2
x  
xG1G2
Bx   G1  G2

 G1  G2   Bx  G1  G2  G1  G2 = 
xG1G2
Bx
xG1G2

i.e. G1  G2 is the union of members of  and so G1  G2   . Thus [O2] is satisfied.

Now,  = B

 :    i.e.  is the union of empty family of members of  and    . Thus [O3] is

satisfied .Finely we note that the set X was defined to be the union of all members of  and so is member

of  ,and axiom [O4] is satisfied. Hence  defines a topology on X and clearly  is a for  .

Theorem : Show that two bases  and  * generate the same topology for the set X iff

(i) for every x  B   , there exist a B*   * s.t. x  B*  B and

(ii) for every x  B*   * , there exist a B   s.t. x  B  B* .

Proof : Suppose that the base  and  * generate the same topology , say  for a set X. We shall prove

the required conditions :

 36
Let x  B   . Since  is base for  , so B   and therefore B is the union of members of  * as  * is

also a base for  . Hence  a set B*   * so that x  B*  B. Hence condition (i) is satisfied. Condition

(ii) can be proved similarly .

Conversely, Suppose the base  and  * satisfy the given conditions . Let  and  * are the

topologies generated by  and  * respectively. We shall prove that  =  * .

Let G   and x  G. Since  is a base for  , G is union of members of  .Then there exist a B   s.t.

x  B  G. Now by condition (i) ,  B*x   * s.t. x  B*x  B  G  x  B*x  G

  x   B
xG xG
*
x G  G B
xG
*
x G  G= B
xG
*
x

i.e. Gis the union of members of  * . But  * is the base for  * so G   *    * .

Similarly ,we can show that  *   . Hence  =  * .

Subbase : Let (X, ) be a topological space . A subfamily S of  is called a subbase for  iff the family
of all finite intersection of members of S is a base for .
Remark : It is not necessary to include X on a subbase for a topology , since
X= 

S :    i.e. X can be viewed as intersection of empty family of members of S .

Example : Let X = {a, b, c, d} and  = { , X, {a},{a, c}, {a, d}, {a, c, d}} then the subfamily
S = {{a,c},{a,d}} of  is a subbase for  since the family  of finite intersection of S is given by
 = {{a},{a, c},{a, d},X} which is a base for .

Theorem : Let S be the family of sets , then the family  of finite intersection of members of S is a base
for a topology on a set X = S : S  S and prove that topology is the smallest topology containing S .
Proof : Suppose S is the given family of sets and , the family of finite intersection of S members of S,
then if B1 , B2   then
n
B1 = S
i 1
i for some positive integer n where Si  S  i = 1, 2, 3,…., n.

m
B2 = S
j 1
j for some positive integer m and S j  S  j = 1, 2, 3,….., m.

n m
 B1  B2 =  i 1
Si  S
j 1
j which is nothing but finite intersection of members of S and so

B1  B2  . Therefore to each x  B1  B2 , there exist B  B1  B2   such that x  B  B1  B2 .

Hence  is a base for some topology  on the set
B : B = S : S  S =X
Now , it remains to prove that  is the smallest topology containing S. Now, since S   and   , so
let * be any other topology containing S and we shall show that   *.
We note that S  * and * is a topology . So all finite intersection of members of S and in *
i.e.   *. Again all union of members of  are in *, it follows that   *. Hence  is the smallest
topology containing S.
Remark : By above theorem , it is clear that every non empty family of sets is a subbase for some
topology .

Subbase and relative topology :

MUDGIL ACADEMY OF MATHEMATICS
Contact No. +91- 8607383607 Email : [email protected]

MSc. Pre (Topology) Unit - I Page 37

Def : Let (X, ) be a topological space and X*  X. the relative topology or induced topology for X*
w.r.t.  is the collection * of all subsets of X* which intersection of X* with members of . In other

words *= G : G  X  G 
for some G   . In order that our definition can be meaning full, we
must prove that * is in deed is a topology on X*.

Theorem : Let (X, ) be a topological space on X*  X. then * defined above is a topology on X*.
Proof :  
[O1] , let G

be any arbitrary family of members of *. Then for every    , there exist

a set G   s.t. G = G  X*. So G



  *, which proves [O1] for *.

For [O2] , let G1 , G2  *, then there exist two sets G1 and G2   s.t.
G1 = G1  X* and G2 = G2  X*
 G1  G2 = (G1  X*)  (G2  X*)
= (G1  G2)  X*
Since G1  G2  *, so (G1  G2)  X*  *. That is G1  G2  *
So * satisfies [O2] axiom . Also X* = X*  X and X   so X*  *.  = X*   and   .so
  *. Hence * defines a topology on X*.

Subspace : Let (X, ) be a topological space and X*  X. Then (X*,*) is called a subspace of (X, )
iff * is relative topology.

Example : Let X={a, b, c, d, e} and  = { , X, {a}, {c, d}, {a, c, d}, {b, c, d, e}} be the topology on X.
*={  X*, X  X*,{a}  {a,d,e},{c,d}  {a,d,e},{a,c,d}  {a,d,e},{b,c,d,e}  {a,d,e}}
={, X*,{a},{d},{a,d},{d,e}
Remark : If (X*,*) is a subspace of (X, ) , then a subset of X* may be open or closed w.r.t. (,) but not
w.r.t. (X, ). To avoid the confusion , sets which are open w.r.t. the subspace (X*,*) will be referred as
*- open, X*- open or open in X*. Similarly, open sets w.r.t. the subspace (X, ) will be called as - open,
X-open or open in X.
In the above example, we see that {a} is both - open and *- open ; {d} is open in
X* but not open in X ; {c,d} is X-open but not X*- open and {b,c} is neither - open nor *-open. Same
conventions are used for closed sets.

Theorem : Let (X*,*) be a subspace of (X, ) then

(i) A subset G*F* of X* is open in X*iff G*= G  X* for some open set in X.
(ii) a subset F* of X is closed in X* iff F*= F  X* for some closed set F in X
Proof : (i) this proof is clear from the definition of subspace .
(ii) F*is closed in X*iff X*- F* is open in X*
iff X*- F* = G  X* for some G  
iff F* = X*- (G  X*) for some G  
iff F*= (X*- G)  ( X*- X*) for some G  
iff F*= (X*- G)   for some G  
iff F*= X*  (X - G) for some G  
iff F* = X*  F where F = X - G and F is closed in X.
Remark : By (ii) , if it is clear that the family of closed sets of (X*,*) is the collection of sets which
are intersection of X* with closed sets in (X, ) .

Theorem : If E is subset of a subspace (X*,*,) of a topological space (X, ), then

(i) C*(E) = X*  C(E) where C* denotes the closer w.r.t. *and C denotes the closer w.r.t. .
 38
(ii) D* (E) = X*  D(E) where D* denotes the derived set w. r. t. * and D denotes the derived set w.r.t.
.
(iii) A subset M of X* is a * neighbourhood of a point x  X* iff M = X*  N for some -neighbourhood
N of x.
Proof : (i) X*  C(E) = X*  [  {F : F is  closed and E  F}]
=  [ X*  F : F is  closed and E  F]
( A  (B  C)=(A  B)  (A  C))
=  [ F* : F* is * closed and E  F* ]
(where X*  F= F* and E  F, E  X*  E  F* )
= C* (E)
(ii) x  d* (E) iff x  X* and x is * limit point of E
iff x  X* and E  G* {x}    G*  *s.t. x  G
iff x  X* and E  (G  X*)- {x}    G   s.t. x  G
iff x  X* and E  G-{x}    G   s.t. x  G
[ E  X* so E  G-{x}=E  (G- X*)-{x}]
iff x  X* and x is a -limit pont of E
iff x  X* and x  d(E)
iff x  X*  d(E) . Hence d*(E)= X*  d(E)
(iii) Let M is *-neighbourhood of x, then there exist an *-open set G* s.t. x  G*  M. G* is *-
open , so G*= X*  G for some G  . So x  X*  G  M ......(1)
Now, let N=M  G. We shall prove that N is  neighbourhood of x and M=N  X*. We see that G is -
open set and x  G  N. This implies that N is a  neighbourhood of x.
Further , N  X* = (M  G)  X*=(M  X*)  (G  X*) = M  (G  X*) = M
[ by (1), G  X*  M]
Conversely, let M=N  X*for some -neighbourhood N of x. We shall prove that M is *-
neighbourhood of x. Since N is -neighbourhood of x,  an -open set G s.t. x  G  N
 x  G  X*  N  X*
 x  G*  M and G* is an *-open set .
 M is a * neighbourhood of x.

Theorem : If (X*,*) be a subspace of (X, ) then prove that :

(i) *   iff X*   i.e. every open set in X* is open in X iff X* is open in X .
(ii) *  iff X*   i.e. every closed set in X* is closed in X iff X* is closed in X.
Proof : Suppose *  , then since X*  * , it follows that X*  .
Conversely , suppose X*   and let G*  * be any arbitrary member, then G*=G  X* for some G  .
Now G  , X*   implies that G  X*   (by O2)
 G  . Hence *  .
(ii) Suppose *  , then since X*  *, it follows that X*  .
Conversely, Suppose X*   and let F*  * be any arbitrary member ,then F* =F  X* for some F   .
Now F   and X*   implies that F  X*   [by (C1)]

Theorem : Let (X*,*) be subspace of (X, ) and  be a base for . Then *={B  X*: B  } is a
base for *.
Proof : Let G*  * and x  G* be any point .To prove that * is a base for *, it is sufficient that
there exist a member B*  * s.t. x  B*  G*
Now G*  *  G* = G  X* for some G   and x  G*  x  G, x  X* but  is given to be a base
for , so  a member B   s.t. x  B  G
 x  B  X*  G  X* [ x  X*]
 x  B*  G* where B*  *
Hence * is a base for *.
MUDGIL ACADEMY OF MATHEMATICS
Contact No. +91- 8607383607 Email : [email protected]

MSc. Pre (Topology) Unit - II Page 1

CONNECTEDNESS
Definition : Two subsets A and B are said to be form a separation of a set E in a topological space (X,  )
if and only if

E = AB ; A  , B   ; A B = , A B = 

In other words, we say that E is separated and write A│B.

Definition : A set E in a topological space (X,  ) is said to be connected or E is  -connected if and

only if E is not separated in X. Clearly an equivalent formulation of this definition is that a set is
connected if and only if whenever

E = A  B ; A   , B   then either A  B   , A  B  

Remark : The two requirements A  B =  and A  B =  in the definition of a separated set are

sometimes replaced by a single condition

(A  B )  ( A  B) =  .

This condition is known as Hausdorff Lennes separation condition. Also, a set E is connected if whenever

E = A  B ; A   , B   then (A  B )  ( A  B)   .

Examples : (i)  is connected in every topological space

(ii) Singleton set {x} is connected.

(iii) In case of discrete topological space, the set consisting of at least two points are not connected.
For, if
E = {x, y} , then E = {x}  {y} where {x}   , {y}   .

Also, {x}  { y } = {x}  {y} = 

{ x }  {y} = {x}  {y} = 

Hence E is separated i.e. E is not connected.

Theorem : Let E be a subset of a subspace of (X*,  *) of a topological space (X,  ), then E is  *-

connected if and only if E is  - connected.
Proof : Suppose that E = A  B where A   , B   , then A, B  X*  X

Now ( A  B * )  ( A *  B) = [A  ( X*  B )]  [( A  X* )  B]

= [(A  X* )  B ]  [ A  ( X*  B)]

= (A  B )  ( A  B) [ A, B  X*]

where A * and A are closure of A w.r.t.  * and  respectively.

Now E is  *- connected if and only if (A  B *)  ( A *  B)   .

 2

if and only if (A  B )  ( A  B)   (Proved above)

if and only if E is  - connected.

Definition : A topological space (X,  ) is connected if X is connected as a subset of X.

Theorem : If C is connected subset of a topological space (X,  ), which has a separation X =

A│B, then either C  A or C  B.

Proof : Since X = A│B, so X = A  B ; A   , B   ; A  B =  , A  B = 

Now C = C  X = C  (A  B) = (C  A)  (C  B)

and (C  A)  (C  B)  A  B = 

(C  A)  (C  B)  A  B = 

If C  A   and C  B   , then C  A and C  B form a separation of C but since C is connected so it

cannot be separated. Therefore either C  A =  or C  B = 

 C  X–A or C  X–B
 C  B or C  A

Corollary 1 : If C is connected in a topological space (X,  ) and C  E  C , then E is connected.

In particular, C is connected.
Proof : Let, if possible, E is not connected, then it must have a separation E = A│B i.e.

E = AB ; A  , B   ; A B = , A B = 

Now C  E and E has a separation A│B, so by above theorem,

C  A or C  B  C  A or C  B

 E  A or E  B [ E  C ]

If E  A , then E  B  A  B = 

 E B = 

 (A  B)  B = 

 B =  , which is a contradiction.

Similarly if E  B , then we get A =  . Thus we get a contradiction in both cases.

Hence E must be connected. In particular, if E = C , then C is connected.

Corollary 2 : If every two points of a set E are contained in some connected subset of E, then E
must be connected.
Proof : Let, if possible, E is not connected, then it must have a separation E = A│B i.e.

E = AB ; A  , B   ; A B = , A B = 

Choose a  A and b  B, then by hypothesis {a, b} is contained in some connected subset say C of E.
Since C is connected subset of E and E = A│B, so by above theorem,