1

Sem-III
MAT 202: Linear Algebra-I
Unit-3 Linear transformations

Definition:- Linear transformation :-

Suppose U and V are vector spaces either both real or both complex.
Then the map T: U  V is said to be a linear map (transformation,
operator), if
(i) T (u1 + u2 ) = T (u1) + T (u2 ) for all, u1 , u2  U
(ii) T (  u ) =  T (u) for all, u  U and all scalars  .

Note:- A linear map T: U  U is said to be a linear map on U.

Whenever we say T: U  U is a linear map, then U and V shall be taken as
vector spaces over the same field of scalars.

Example:- Prove that the map T: V3  V3 define by T (x1,x2,x3) = (x1,x2,0)

linear map.
Solution:- Let  be any scalar and x, y V3 where x=(x1, x2, x3) and y = (y1,y2,y3)
 x + y =(x1, x2, x3) + (y1,y2,y3)= (x1+ y1, x2+ y2 , .x3+ y3)
And  x=(  x1,  x2,  x3)
Now T(x + y) = T(x1+ y1, x2+ y2 , .x3+ y3)
= (x1+ y1, x2+ y2, 0) ____________(1)
( by definition of T)
Now T(x) + T(y) = T(x1, x2, x3) + T(y1,y2,y3)
= (x1, x2, 0) + (y1,y2,0) ( by definition of T)
= (x1+ y1, x2+ y2, 0) ____________(2)
T(  x)= T(  x1,  x2,  x3) = (  x1,  x2, 0)
=  (x1,x2,0) _________(3)
 T(x)=  T (x1,x2,x3) =  (x1,x2,0)____________(4)
From (1) ,(2),(3) and (4)
T: V3  V3 linear map.

Note:- T: V3  V3 define by T (x1, x2, x3) = (x1, x2, 0) is called the projection on
x1x2 plane.

Example:- Prove that the map T: V3  V2 define by T (x1,x2,x3) = (x1 - x2, x1+x3)
linear map.
Solution:- Let  be any scalar and x, y V3 where x=(x1, x2, x3) and y = (y1,y2,y3)
 x + y =(x1, x2, x3) + (y1, y2, y3)= (x1+ y1, x2+ y2 , .x3+ y3)
And  x=(  x1,  x2,  x3)
Now T(x + y) = T(x1+ y1, x2+ y2 , .x3+ y3)
= (x1+ y1 - x2- y2, x1+ y1+ x3+ y3) ____________(1)
( by definition of T)
Now T(x) + T(y) = T(x1, x2, x3) + T(y1,y2,y3)
 2
= (x1 - x2, x1+x3) + (y1-y2, y1+y3) ( by definition of T)
= (x1+ y1 - x2- y2, x1+ y1+ x3+ y3) ____________(2)
T(  x)= T(  x1,  x2,  x3) = (  x1 -  x2 ,  x1 +  x3 )
=  (x1 - x2, x1+x3) _________(3)
 T(x)=  T (x1,x2,x3) =  (x1 - x2, x1+x3) ____________(4)
From (1) ,(2),(3) and (4)
T: V3  V2 linear map.

Example:- Examine the map T: V3  V1 define by T (x1,x2,x3) = ( x12  x22  x32 )

linear map or not.
Solution:- Let  be any scalar and x, y V3 where x=(x1, x2, x3) and y = (y1,y2,y3)
 x + y =(x1, x2, x3) + (y1, y2, y3)= (x1+ y1, x2+ y2 , .x3+ y3)
And  x=(  x1,  x2,  x3)
Now T(x + y) = T(x1+ y1, x2+ y2 , .x3+ y3)
= ((x1+ y1)2 + (x2+ y2)2 + (x3+ y3)2) ____________(1)
( by definition of T)
Now T(x) + T(y) = T(x1, x2, x3) + T(y1,y2,y3)
= ( x12  x22  x32 )+ ( y12  y22  y32 ) ( by definition of T)
____________(2)

From (1) and (2)

 T(x + y)  T(x) + T(y)
T: V3  V1 is not linear map.

Example:- Prove that the map T: U  V define by T (u) = 0v linear map.

Solution:- Let  be any scalar and x, y U
Now T(x + y) = 0v _________(1)
( by definition of T)
Now T(x) + T(y) = 0v + 0v = 0v ( by definition of T)
____________(2)
T(  x)= 0v _________(3)
 T(x)=  0v = 0v ____________(4)
From (1) ,(2),(3) and (4)
T: V3  V2 linear map.
Note:- the map T: U  V define by T (u) = 0v is called zero map.

Example:- Prove that the map T: U  U define by T (u) = u linear map.

Solution:- Let  be any scalar and x, y U
Now T(x + y) = x + y _________(1)
( by definition of T)
Now T(x) + T(y) = x + y ( by definition of T)
____________(2)
T(  x)=  x _________(3)
 T(x)=  x=  x ____________(4)
From (1) ,(2),(3) and (4)
T: U  U linear map.
 3
Note:- the map T: U  U define by T (u) = u is called identity map.

Example:- Prove that the map T: V2  V2 define by T (x1,x2) = (x1, - x2)

linear map.
Solution:- Let  be any scalar and x, y V2 where x=(x1, x2) and y = (y1,y2)
 x + y =(x1, x2) + (y1, y2)= (x1+ y1, x2+ y2 )
And  x=(  x1,  x2)
Now T(x + y) = T(x1+ y1, x2+ y2)
= (x1+ y1, - x2- y2) ____________(1)
( by definition of T)
Now T(x) + T(y) = T(x1, x2) + T(y1,y2)
= (x1 - x2) + (y1-y2) ( by definition of T)
= (x1+ y1, - x2- y2) ____________(2)
T(  x)= T(  x1,  x2) = (  x1,-  x2)
=  (x1, - x2) _________(3)
 T(x)=  T (x1,x2) =  (x1, - x2) ____________(4)
From (1) ,(2),(3) and (4)
T: V2  V2 linear map.

Note:- the map T: V2  V2 define by T (x1,x2) = (x1, - x2) is called the reflection
in the x1-axis.

Figure

Example:- Prove that the map D: Ç(1)(a,b)  Ç(1)(a,b) define by D (f) = f '
linear map. Where is the derivative of f.
Solution:- Let  be any scalar and f, g  Ç(1)(a,b)

Now T(f + g) = (f + g) '

= f '+ g '____________(1)
( by definition of T)
Now T(f ) + T(g) = f '+ g ' ____________(2)
( by definition of T)

T(  f)= (  f )' =  (f )' _________(3)

 T(f) =  (f )'____________(4)
From (1) ,(2),(3) and (4)
D: Ç(1)(a,b)  Ç(1)(a,b) linear map.

b
Example:- Prove that the map D: Ç(1)(a,b)  R define by I (f) =  f ( x)dx
a

linear map.
Solution:- Let  be any scalar and f, g  Ç(1)(a,b)
 4
b
Now T(f + g) =  ( f ( x)  g ( x))dx
a
b b
= 
a
f ( x)dx +  g ( x)dx ____________(1)
a

( by definition of T)
b b
Now T(f ) + T(g) =  f ( x)dx +  g ( x)dx ____________(2)
a a

( by definition of T)
b b
T(  f)=  f ( x)dx =   f ( x)dx _________(3)
a a
b
 T(f) =   f ( x)dx ____________(4)
a

From (1) ,(2),(3) and (4)

D: Ç(1)(a,b)  R linear map.

Example:- Prove that the map T: U  U define by T (x) = x + u0 is not linear

map. Where u0 is a fixed vector in U.
Solution:- Let  be any scalar and x, y U
Now T(x + y) = x + y + u0 _________(1)
( by definition of T)
Now T(x) + T(y) = x + u0 + y + u0 ( by definition of T)
____________(2)

From (1) and (2)

 T(x + y)  T(x) + T(y)
T: U  U is not linear map.
Note:- the map T: U  U define by T (x) = x + u0 is called translation by the
vector u0 .Where u0 is a fixed vector in U.
 The function f: R  R defined by f(x) = x + a (‘a’ fixed) is called a linear
function, because its graph in xy-plane is straight line. But it is not a linear map
from the vector space V1 to itself.

Theorem:- let T: U  V be a linear map, then

(a) T(0u) = 0v
(b) T(-u) = -T(u)
(c) T(  1 u1+  2 u2+  3 u3+ ….+  n un)
=  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(un).
i.e. A linear map T transforms the zero vector of U into the zero vector of V and
negative of every u of U into the negative of T(u) of V.
Proof:- (a) T(0u) = T(0.u) [ 0.u = 0, u  U]
= 0T(u) [ T is linear]
=0v
(b) T(-u) = T(-1.u) [ (-1).u = -u, u  U]
 5

= (-1)T(u) [ T is linear]
=-T(u)
(c) This result can be proved by mathematical induction.
Let p(n): T(  1 u1+  2 u2+  3 u3+ ….+  n un)
=  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(un).
Then p(1): T(  1 u1) =  1 T(u1)
Since T is linear this is obviously true.
So the result is true for n =1.
Assume that p(k) to be true

i.e. p(k): T(  1 u1+  2 u2+  3 u3+ ….+  n uk)

=  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(uk). is true.
We try to establish the result for n =k + 1
T(  1 u1+  2 u2+  3 u3+ ….+  k 1 uk+1)
=  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  k 1 T(uk+1).
By the hypothesis and linearity of T.
Since (i) p(1) is true.
(ii) p(k)  p(k+1)
The result is true for all n.

Theorem:- A linear transformation T is completely determined by the values of elements of

a basis. Precisely, if B={u1,u2,…,un} is a basis for U and v1,v2,…,vn be n vectors
( not necessarily distinct) in V, then there exists a unique linear transformation
T:U  V such that T(ui) = vi for i=1,2,…,n.

Proof:- Let u  U . since B={ u1,u2,…,un} is a basis for U ,any vector u in U can be written as
a unique linear combination of basis elements. Hence there exist scalars 1 ,  2 ,…,  n
satisfying
u= 1 u1+  2 u2+…+  n un.
We define mapping T: U  V by T(u)= 1 v1+  2 v2+…+  n vn.
We prove the following facts:
(i) T is linear transformation.
(ii) T(ui) =vi
(iii) Such mapping T is unique.

Proof of (i):- Let u,v  U . Then there are scalars 1 ,  2 ,…,  n and 1 ,  2 ,…,  n for
which
u= 1 u1+  2 u2+…+  n un.
v= 1 u1+  2 u2+…+  n un.
u + v = 1  1  u1+  2  2  u2+…+  n  n  un
By definition of T
T(u) = 1 v1+  2 v2+…+  n vn.
 6

T(v)= 1 v1+  2 v2+…+  n vn.

T(u+v)= 1  1  v1+  2  2  v2+…+  n  n  vn
It is clear that
T(u+v)= T(u) +T(v)
Also we can easily show that
T(  u)=  T(u)
For every scalar  and every vector u  U. This establishes that T is a
linear transformation from U to V.

(ii) Now ui  B, i=1,2,…,n

So, ui can be expressed in terms B as ui =0.u1 + 0.u2 + 0.u3 + …. +0.un
T(ui) = 0.v1 + 0.v2 + 0.v3 + …. +0.vn
= vi, i= 1, 2, 3, …, n.
(iii) Let S: U  V be any other linear transformation define by
S(ui) =vi i= 1, 2, 3, …, n.
Now S(u) = S( 1 u1+  2 u2+…+  n un)
=  1 S(u1)+  2 S(u2)+  3 S(u3)+ ….+  n S(un).
= 1 v1+  2 v2+…+  n vn
= T(u).
 S(u) = T(u).
This proved that such mapping T is unique.

Example:- If T is a linear transformation from V2  V2 define by

T (2,1) = (3, 4)
T (-3,4) = (0, 5)
then express (0, 1) as a linear combination of (2, 1) and (-3, 4). Hence find image of
(0, 1) under T.
Solution:- Let (0, 1) =  (2, 1) +  (-3, 4)
 (2  - 3  ,  +4  )=(0, 1)
 2  - 3  = 0,  +4  = 1
3 2
Solving these equation then we get  = , =
11 11
3 2
 (0, 1) = (2, 1) + (-3, 4).
11 11

3 2
 T(0, 1) =T( (2, 1) + (-3, 4))
11 11
3 2
= T(2, 1) + T (-3, 4)
11 11
3 2 1
= (3, 4) + (0, 5) = (9, 22)
11 11 11
1
Thus we get T(0, 1) = (9, 22)
11

Example:- If T is a linear transformation from R3  R3 define by

 7

T (e1) = e1 +e2 +e3 ,T (e2) = e2 + e3 and T(e3) = e2 – e3 where e1 ,e2 , e3 are unit
vector of R3 . Then (i)Determine the transformation of (2, -1, 3) And
(ii)describe explicitly the linear transformation T.
Solution:- Since e1 ,e2 , e3 are unit vector of R3
 e1 =(1, 0, 0),e2 =(0, 1, 0) , e3 =(0, 0, 1)
We have T (e1) = e1 +e2 +e3  T(1, 0, 0) = (1, 0, 0 ) +(0, 1, 0) +(0, 0, 1)
= (1, 1, 1 )
T (e2) = e2 + e3  T(0, 1, 0) = (0, 1, 0) +(0, 0, 1)
= (0, 1, 1 )
T(e3) = e2 – e3  T(0, 0, 1) = (0, 1, 0) - (0, 0, 1)
= (0, 1, -1 )
Since e1 ,e2 , e3 form basis for R3 .
3
 every vector of R can be uniquely expressed as a linear combination of e1 ,e2 , e3.
(i) Now (2, -1, 3) = 2(1, 0, 0 ) +(-1)(0, 1, 0) +3(0, 0, 1) = 2e1 +(-1)e2 +3e3
 T(2, -1, 3) = 2T(e1)+(-1)T(e2)+3T(e3)

= 2 (1, 1, 1) + (-1) (0, 1, 1 ) + 3(0, 1, -1 )

= (2, 4, -2)
(ii) (x, y, z)  R3.
Now (x, y, z) = x(1, 0, 0 ) + y(0, 1, 0)+ z(0, 0, 1) = xe1 +ye2 +ze3
 T(x, y, z) = xT(e1) +yT(e2) +zT(e3)
= x (1, 1, 1 ) +y (0, 1, 1 )+z (0, 1, -1 )
= (x, x + y + z, x+ y – z)
 T(x, y, z) = (x, x + y + z, x+ y – z)
Which is require linear transformation T.

Range and Kernel of a Linear map

Definition:-Kernel of a Linear map (null space)

Let T: U  V be a linear map. The Kernel (null space) of T is the set
N(T) = {u  U/ T(u) = 0}.
It is denoted as kerT.
OR N(T) is the set of all those elements in U that are mapped by T into the zero of V.

Definition:- Range of a Linear map (null space)

Let T: U  V be a linear map. The range of T is the set
R(T) = {T(u)  V/ u  U }.
It is denoted as kerT.

Example:- Let T: V3  V3 be a linear map define by T(x1,x2,x3)= (x1,x2,0) Find N(T) &
R(T) (OR) Find the range and kernel of T.
Solution:- Here ,R(T) = {(x1,x2,0) / x1,x2  R}
 8
R(T) is x1x2 plane.
T is not onto.
Since R(T) is a subset of co domain V3.
T is not one-one.
Since different vectors (1,0,2) and (1,0,5) have the same image (1,0,0).
N(T) = x3-axis.
Since any vector (0,0, x3) on the x3-axis will be taken onto zero to vector of V3.

Example:- Let T: U  U be an identity linear map then find N(T) & R(T).
(OR) Find the range and kernel of T.
Solution:- Here T: U  U be an identity linear map.
i.e. T(u) = u for u  U.
This is one – one and onto linear map.
 R(T) = U and N(T) = 0

Example:- Let T: U  U be zero linear map then find N(T) & R(T).
(OR) Find the range and kernel of T.
Solution:- Here T: U  U be zero linear map.
i.e. T(u) = 0 for u  U.
This is not one – one and onto linear map.
 R(T) = 0 and N(T) = U

Example:-Let T: V3  V2 be a linear map define by T (x1, x2, x3) = (x1 - x2, x1+x3) then find
N(T) & R(T) (OR) Find the range and kernel of T.
. Solution:- Here T: V3  V2 be a linear map define by T (x1,x2,x3) = (x1 - x2, x1+x3)
Let (a, b)  V2 such that T (x1,x2,x3) = (a, b)
 (x1 - x2, x1+x3) = (a, b)
 x1 - x2 = a , x1+x3 = b
Solving these equation then we get
x2 = x1 – a, x3 = b - x1
Hence T (x1, x1 – a, b - x1) = (a, b)
 R(T) = V2 ( every vector (a, b) V2 in R(T) )
So this is onto map.
Now for kernel of T
T (x1, x2, x3) = (0, 0)
 (x1 - x2, x1+x3) = (0, 0)
 x1 - x2 = 0 , x1+x3 = 0
Solving these equation then we get
 x1= x2 = -x3
i.e. all vectors of the form (x1,x1,-x1) will be mapped into zero.
 N(T) = { x1 (1,1,-1) / x1 any scalar} = [(1,1,-1)]

Example:- Let T: V2  V2 be a linear map define by T (x1, x2) = (x1, - x2) then find N(T) &
R(T) (OR) Find the range and kernel of T.
. Solution:- Solution:- Here T: V2  V2 be a linear map define by T (x1,x2) = (x1, - x2)
Let (a, b)  V2 such that T (x1,x2) = (a, b)
 9
 (x1 ,- x2) = (a, b)
 x1 = a , - x2 = b
Solving these equation then we get
x1 = a , x2 = -b
Hence T (a, – b ) = (a, b)
 R(T) = V2 ( every vector (a, b) V2 in R(T) )
So this is onto map.
Now for kernel of T
T (x1, x2) = (0, 0)
 (x1 ,- x2) = (0, 0)
 x1 = 0 , x 2 = 0
 N(T) = (0,0)}

Example:- Let the map D: Ç(1)(a,b)  Ç(1)(a,b) define by D (f) = f '

linear map. Where is the derivative of f. then find N(T) & R(T). (OR) Find the range
and kernel of T.

Solution:- Since every continuous function g on (a, b) possesses an antiderivative.

hence D is an onto map.
(1)
 R(D) = Ç (a,b) .
And N(D) is the set of all constant functions in Ç(1)(a,b) .

b
Example:- Let the map D: Ç (a,b)  R define by I (f) =  f ( x)dx linear map. Where is the
(1)

derivative of f. then find N(T) & R(T). (OR) Find the range and kernel of T.

Solution:- Since every real number can be obtained as the algebraic area under some
curve y = f(x) from a to b.
hence D is an onto map.
 R(D) = R .
And it is difficult to say anything about kernel i.e. N(D).

Note :- From above example we see that if T is one-one when N(T) is the zero subspace and
conversely.

Theorem:- Let T: U  V be a linear map. Then

(a) R(T) is a subspace of V.
(b) N(T) is subspace of U.
(c) T is one-one iff N(T) is the zero subspace, {0U}, of U.
(d) If [u1,u2,…,un] = U, then [T(u1),T(u2),…,T(un)]
(e) If U is finite- dimensional, then dimR(T)  dimU.
Proof:- (a) we want to prove that R(T) is a subspace of V.
For this , let v1,v2  R(T) such that T(u1) = v1 and T(u2) = v2 for u1,u2 U.
Now v1 + v2 = T(u1) + T(u2) = T(u1 + u2) [ T: U  V be a linear map]
Since U is a vector space .
 u1 + u2  U
 10
And T: U  V be a linear map
 T(u1 + u2) = v1 + v2  R(T)
Similarly ,  v1  R(T) then  v1 =  T(u1) = T(  u1)  R(T).
Thus R(T) is a subspace of V.

(b) we want to prove that N(T) is a subspace of U.

For this , let u1,u2  N(T) such that T(u1) = 0v and T(u2) = 0v for u1,u2  U.
Now T(u1 + u2)= T(u1) + T(u2) = 0v [ T: U  V be a linear map]
 T(u1 + u2) = 0v
 u1 + u2  N(T)
Similarly , for any scalar  T(  u1)=  T(u1) =  0v =0v  N(T).
  u1  N(T)
Thus N(T) is a subspace of V.

(c) Suppose T is one-one.

We want to prove that N(T) is the zero subspace, {0U}, of U.
Since T is one-one then T(u) = T(v)  u = v
If u  N(T) then T(u) = 0v = T(0U).
 u = 0U
i.e. no nonzero vector u of U can belong to N(T).
Since 0U in any case belongs to N(T).
i.e. N(T) contains only 0U and nothing else.
Hence, N(T) is the zero subspace, {0U}, of U.
Conversely,
Suppose N(T) = {0U}
We want to prove that T is one-one.
i.e. We want to prove that T(u) = T(v)  u = v
suppose T(u) = T(v)
then T(u-v) = T(u) - T(v) = 0V.
 u – v  N(T) = {0U}
 u – v = 0U.
i.e. u = v
i.e. T is one-one.
(d) let [u1,u2,…,un] = U
u U
then u can be expressed as a linear combination of vectors u 1,u2,…,un.
The T(u1),T(u2),…,T(un) are in R(T).
So [T(u1),T(u2),…,T(un)]  R(T).
Let v  R(T).
Then there exists a vector u  U such that T(u) = v.
Since u  U = [u1,u2,…,un], we have
u= 1 u1+  2 u2+…+  n un.
v = T(u) =T( 1 u1+  2 u2+…+  n un)
=  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(un).
So v  [ T(u1),T(u2),…,T(un)].
This proves that R(T) = [ T(u1),T(u2),…,T(un)].
 11

(e) Let U be finite dimensional and dim U = n.

So there can be at most n LI vectors in U.
Let { u1,u2,…,un} be the basis of U.
Then R(T) = [T(u1),T(u2),…,T(un)]
So that there can’t be more than n LI vectors inR(T).
So dimR(T)= n  dimU.

Definition:- Rank of T:-

Let T: U  V be a linear map. Then If R(T) is finite- dimensional, the dimension
of R(T) (i.e.dimR(T)) is called the rank of T and is denoted by r(T).

Definition:- nullity of T:-

Let T: U  V be a linear map. Then If N(T) is finite- dimensional, the dimension of
N(T) (i.e.dimN(T)) is called the nullity of T and is denoted by n(T).
Rank and Nullity

Theorem:- Let T: U  V be a linear map. Then

(a) If T is one-one and u1,u2,…,un. are linearly independent vectors of U, then
T(u1),T(u2),…,T(un) are LI.
(b)If v1,v2,…,vn are linearly independent vectors of R(T) and u1,u2,…,un are vectors of U
such that T(u1) = v1, T(u2) = v2, …,T(un) = vn then u1,u2,…,un. are linearly independent.
Proof:- (a) Let T is one-one and {u1,u2,…,un}are linearly independent vectors in U.
We want to prove that T(u1),T(u2),…,T(un) are LI.
Consider  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(un) = 0v
 T( 1 u1+  2 u2+…+  n un) = 0v [ T is linear map]
Also T(0u) = 0v
Since T is one-one is given.
 1 u1+  2 u2+…+  n un= 0u
Since u1,u2,…,un. are linearly independent vectors of U is given.
 1 =  2 =…=  n = 0
Thus T(u1),T(u2),…,T(un) are LI.
(b) Since v1,v2,…,vn are linearly independent vectors in V
and T(u1) = v1, T(u2) = v2, …,T(un) = vn where u1,u2,…,un U is given.
We want to prove that u1,u2,…,un. are linearly independent.
Consider 1 u1+  2 u2+…+  n un = 0
 T( 1 u1+  2 u2+…+  n un) = 0v [ T is linear map]
  1 T(u1)+  2 T(u2)+  3 T(u3)+ ….+  n T(un) = 0v
Since T(u1),T(u2),…,T(un) are LI
 1 =  2 =…=  n = 0
Hence u1,u2,…,un. are linearly independent vectors in U.

Theorem:- (Rank and Nullity Theorem):

Let T: U  V be a linear map and U a finite dimensional vector space. Then prove that
 12

dim R(T) + dim N(T) = dim U.

i.e r(T) + n(T) = dim U
(or) rank + nullity = dimension of the domain space.
Proof:- N(T) is a subspace of a finite dimensional vector space U. Then N(T) must be finite
dimensional.
Let dim N(T) = n(T) = n and dim U = p .
So, n  p
Let the basis for N(T) be { u1,u2,…,un}.
 { u1,u2,…,un}are linearly independent vectors in N(T)
 {u1,u2,…,un}are linearly independent vectors in U.
Now extend this set of n linearly independent vectors of U to the basis for u.
So we find the vectors un+1,un+2,…,up
So that the enlarged set { u1,u2,…,un , un+1,un+2,…,up } is a basis for U.
Since this set of p vectors generate vector space U.
R(T) =[ T(u1),T(u2),…,T(up)]
But ui  N(T), i = 1, 2, 3, …,n.
Hence T(u) = 0v, i = 1, 2, 3, …,n.
 R(T) = [T(un+1),T(un+2),…,T(up)]
Now we shall prove that A = { T(un+1),T(un+2),…,T(up)} is basis for R(T).
Since we already proved that R(T) = [T(un+1),T(un+2),…,T(up)]
So we have only prove that A is LI set.
Let us consider  n 1 T(un+1)+  n  2 T(un+2)+…+  p T(up) = 0
 T[  n 1 un+1+  n  2 un+2+…+  p up] = 0
Since T is linear
 n 1 un+1+  n  2 un+2+…+  p up  N(T).
But N(T) has a basis {u1,u2,…,un}
so  n 1 un+1+  n  2 un+2+…+  p up which is the elements of N(T) can be expressed as a
linear combination of basis {u1,u2,…,un} of N(T).
 there exists scalars 1 ,  2 ,…,  n such that
 n 1 un+1+  n  2 un+2+…+  p up = 1 u1+  2 u2+…+  n un
 1 u1+  2 u2+…+  n un +(-1)  n 1 un+1+ (-1)  n  2 un+2+…+ (-1)  p up = 0
Since {u1,u2,…,up} is a basis for t vector space U this set is LI.
So  n 1 =  n  2 =…=  p = 0
This prove that set A is LI.
 A = { T(un+1),T(un+2),…,T(up)} is basis for R(T).
Dim R(T) = number of elements in basis A
=p–n
= dim U - dim N(T)
Hence rank + nullity = dimension of the domain space.

Example:- Prove that the linear map T : V3  V3 define by T (e1) = e1 - e2 , T (e2) = 2e2 + e3
and T(e3) = e1 + e2 + e3 is neither one-one nor onto.
Solution:- Here R(T) = [T (e1), T (e2) , T(e3)]
= [e1 - e2 , 2e2 + e3 , e1 + e2 + e3]
 13

=[ e1 - e2 , 2e2 + e3]
Since e1 + e2 + e3 is linear combination of e1 - e2 , 2e2 + e3
R(T) has dimension 2.
R(T)  V3
 T is not onto.
Since N(T) consists those vectors (x1 , x2 , x3) V3 such that T (x1 , x2 , x3) = 0.
i.e T(x1e1+ x2 e2 + x3 e3) = 0
 x1 T(e1)+ x2 T(e2) + x3 T(e3) = 0
 x1 (e1 - e2) + x2 (2e2 + e3 )+ x3 (e1 + e2 + e3) = 0
 x1 + x3= 0 , -x1 +2 x2 +x3 = 0 and x2+ x3= 0
Solving these equation then we get x1 = x2 = -x3
 N(T) = { (x1 , x1 , -x1)/ x1 an arbitrary scalar}= [(1, 1,-1)].
 N(T) is not the zero subspace of V3.
Hence T is not one-one.

Example:- Let linear map T : V4  V3 define by T (e1) =(1, 1, 1), T (e2) = (1, -1, 1) , T(e3) =
(1, 0, 0) and T(e4) = (1, 0, 1) then verify that r(T) + n(T) = dim U(=V4) = 4.
Solution:- Here R(T) = [T (e1), T (e2) , T(e3), T(e4)]
 R(T) = [(1, 1, 1), (1, -1, 1) , (1, 0, 0) , (1, 0, 1)]
(1, 1, 1), (1, -1, 1) , (1, 0, 0) (1, 0, 1) is LD, because a set of four vectors of V3
(dim V3 = 3)is always LD.
1 1
Now find that (1, 0, 1) = (1, 1, 1) + (1, -1, 1) +0 (1, 0, 0)
2 2
Hence we discard the vector (1, 0, 1) so that
R(T) = [(1, 1, 1), (1, -1, 1) , (1, 0, 0)]
Now check R(T) is LI.
Let 1 ,  2 ,  3  R such that 1 (1, 1, 1)+  2 (1, -1, 1)+  3 (1, 0, 0) = 0
 ( 1 +  2 +  3 , 1 -  2 , 1 +  2 ) = 0
 1 +  2 +  3 = 0, 1 -  2 = 0, 1 +  2 = 0
From above equations we get
1 =  2 =  3 = 0
Hence {(1, 1, 1), (1, -1, 1) , (1, 0, 0)} is LI.
 dim R(T) = r(T) = 3
Now we find N(T).
Let us suppose that T(u) = 0 for u  V4 where u = (x1 , x2 , x3, x4),
Now u = x1e1+ x2e2+ x3e3+x4e4
 T(u) = T(x1e1+ x2e2+ x3e3+x4e4) = 0
 x1 T(e1)+ x2 T(e2) + x3 T(e3) = 0
 x1 [(1, 1, 1)+ x2 (1, -1, 1) + x3 (1, 0, 0) ++ x4 (1, 0, 1) = 0
 x1 + x2+ x3+ x4= 0 , x1 - x2 = 0 and + x1+x2+ x4= 0
Solving these equation then we get x1 = x2 = -x4/2, x3 =0
 N(T) = { (x1 , x1 , 0,-2x1)/ x1 an arbitrary scalar}= [(1, 1,0,-2)].
 dim N(T) = n(T) =1
Hence r(T) + n(T) = 3 +1 = 4 = dim U(=V4)
 14

Inverse of a linear transformation

Definition:- Nonsingular or Isomorphism :

A linear map T : U  V is said to be nonsingular if it is one – one and onto. Such a
map is also called an isomorphism.

Definition:- Inverse function:

Any function is called inverse function iff it is one – one and onto.

Note: a linear transformation is Nonsingular iff it has an inverse.

Example:- Define a linear map T : V2  V2 by T(x1 , x2) = (x1 , -x2). Prove that this map is
Nonsingular.
Solution:- Here the linear map T : V2  V2 define by T(x1 , x2) = (x1 , -x2)
Now for N(T)
T(x1 , x2) = 0
 (x1 , -x2) = 0
 x1 = 0 and x2 = 0
N(T) = 0
 This linear map T is one – one.
Now for R(T)
For every (y1 , y2)  V2 then there exists (x1 , x2) V2 such that
T(x1 , x2) = (y1 , y2)
 (x1 , -x2) =(y1 , y2)
 x1 = y1and x2 = y2
 R(T) = V2
This linear map T is onto
-1
 T (y1 , y2) = (y1 , -y2).
 it has inverse
 This map is Nonsingular.

Example:- Prove that a linear map T: U  U define by T (u) = u is Nonsingular.

(OR)Prove that an identity linear map is one-one and onto.
(OR) Find the inverse of a linear map T: U  U define by T (u) = u.
Solution:- Here the linear map T : U  U define by T(u) = (u)
Now for N(T)
T(u) = 0
u = 0
 N(T) = 0
 This linear map T is one – one.
Now for R(T)
For every (y )  U then there exists (u) U such that
T(u) = (y)
 (u) =( y)
 R(T) = U
 15

This linear map T is onto

-1
 T (y) = (y).
 it has inverse
 This map is Nonsingular.

Example:- Prove that a linear map T : P2  V3 define by T (  0 + 1 x +  2 x2) = (  0 , 1 ,  2 )is

Nonsingular. OR Prove that a linear map T : P2  V3 define by T (  0 + 1 x +  2 x2) =
(  0 , 1 ,  2 ) is isomorphism.
Solution:- Here the linear map T : P2  V3 define by T (  0 + 1 x +  2 x2) = (  0 , 1 ,  2 )
Now for N(T)
T (  0 +  1 x +  2 x2 ) = 0
 (  0 , 1 ,  2 ) = 0
  0 = 1 =  2 = 0
 N(T) = 0
 This linear map T is one – one.
Now for R(T)
For every ( 1 ,  2 ,  3 )  V3 then there exists ( 1 +  2 x +  3 x2) P2 such that
T( 1 +  2 x +  3 x2) = ( 1 ,  2 ,  3 )
 ( 1 ,  2 ,  3 ) is image of T.
 R(T) = V3
This linear map T is onto
 T ( 1 ,  2 ,  3 ) = ( 1 +  2 x +  3 x ).
-1 2

 it has inverse
 This map is Nonsingular.

Theorem:- Let T : U  V be nonsingular linear map. Then T-1 : V  U is a linear, one –

one and onto map.
Proof:- First to prove that T-1 is linear.
Let v1,v2  V.
Let T-1(v1)= u1 and T-1(v2) = u2 for u1,u2  U.
Since T is nonsingular linear map
 T is one – one and onto map.
 u1and u2 exists uniquely.
 v1 = T(u1) and v2 = T(u2)
 v1 + v2 = T(u1) + T(u2) = T(u1 + u2) [ T is linear]
-1 -1 -1
 T (v1 + v2)= u1 + u2 = T (v1) + T (v2)
-1 -1 -1
 we get T (v1 + v2) = T (v1) + T (v2)
Also  v1 =  T(u1) = T(  u1)
 T (  v1) =  u1 =  T (v1)
-1 -1

i.e. T-1(  v1) =  T-1(v1)

-1
 T is linear.
Now we want to prove that T-1 is one-one.
Let v1,v2  V such that T-1(v1) = T-1(v2) = u say for u  U.
 16

 v1 = v2= T(u)
Since image of u under T is unique.
 v1 = v2
Thus we get T-1(v1) = T-1(v2)  v1 = v2
-1
 T is one-one.
Now we want to prove that T-1 is onto.
Given any element u  U then there exists an element v V such that
T(u) = v.
-1
 u =T (v)
This show that T-1 is onto.
-1
 T : V  U is a linear, one – one and onto map.

Example:- Check that a linear map T: U  U ,where U is vector space, define by

T (x) = T({x1 , x2 , x3,…, xn, …}) = { x2 , x3,…, xn, …} is Nonsingular or not. Also
check the inverse of this linear map is exist or not.
Solution:- Here the linear map T : U  U define by
T({x1 , x2 , x3,…, xn, …}) = { x2 , x3,…, xn, …}
Now for N(T)
T (x)= 0
 T({x1 , x2 , x3,…, xn, …}) = 0
 { x2 , x3,…, xn, …} =0
Here let x1 = z which may not be zero.
 N(T) = {z, 0, 0, 0, ……}
 N(T)  0
 This linear map T is not one – one.
Now for R(T)
For every (y1 , y2 , y3,…, yn, … )  U then there exists (z, y1 , y2 , y3,…, yn, … ) U
such that T({z, y1 , y2 , y3,…, yn, … }) = (y1 , y2 , y3,…, yn, … )
 { z, y1 , y2 , y3,…, yn, … } is pre image of { y1 , y2 , y3,…, yn, … }
 R(T) = U
This linear map T is onto
Thus we get T is onto but not one-one.
 This map is not nonsingular.
 Also T has not inverse.

Example:- Check that a linear map T: U  U ,where U is vector space, define by

T (x) = T({x1 , x2 , x3,…, xn, …}) = {0, x1, x2 , x3,…, xn, …} is Nonsingular or not.
Also check the inverse of this linear map is exists or not.
Solution:- Here the linear map T : U  U define by
T({x1 , x2 , x3,…, xn, …}) = {0, x1, x2 , x3,…, xn, …}
Now for N(T)
T (x)= 0
 T({x1 , x2 , x3,…, xn, …}) = 0
 {0, x1, x2 , x3,…, xn, …} =0
 N(T) = {0, 0, 0, 0, ……}
 17

 N(T) = 0
 This linear map T is one – one.
Now for R(T)
let (1 , 1 , 1,…, 1, … ) U has no pre image in U
 R(T)  U
This linear map T is not onto
Thus we get T is not onto but one-one.
 This map is not nonsingular.
 Also T has not inverse.

Theorem:- If U and V are finite dimensional vector spaces of the same dimension, then a
linear map T: U  V is one-one iff it is onto.
Proof:- T is one –one  N(T) = {0v}
 n(T) = 0
 r(T) = dim U = dim V
[ By Rank and Nullity Theorem i.e r(T) + n(T) = dim U]
 T is onto.
Example:- Show that the linear map T : V3  V3 defined by
T(x1 , x2, x3) = (x1 + x2+ x3 ,x2 +x3, x3) is nonsingular and find its inverse.
Solution:- Here the linear map T : V3  V3 defined by T(x1 , x2, x3) = (x1 + x2+ x3 ,x2 +x3, x3)
Now for N(T)
T(x1 , x2, x3) = 0
 (x1 + x2+ x3 ,x2 +x3, x3)= 0
x1 + x2+ x3 = 0 ,x2 +x3 = 0, x3 = 0
 x1 = 0, x2 = 0 and x3 = 0
N(T) = 0
 This linear map T is one – one.
Since the dimension of domain space and dimension of co domain space are equal.
 T is onto.
Thus we get T is one-one and onto.
Hence T is nonsingular and the inverse of T exists. i.e.T-1 exists.
Now we derive the formula for T-1.
Let T-1(y1 , y2, y3) = (x1 , x2, x3).__________(1)
 (y1 , y2, y3) = T(x1 , x2, x3)
 (y1 , y2, y3) = (x1 + x2+ x3 , x2 +x3, x3)
 y1 = x1 + x2+ x3, y2 = x2 +x3 , y3= x3
Solving these equation then we get
x1 = y1 - y2, x2 = y2 -y3 and x3 = y3
put the values of x1 , x2, x3 in equation (1) then we get
T-1(y1 , y2, y3) = (y1 - y2, y2 -y3 , y3)

Example:- Show that the linear map T : V3  V3 defined by T (e1) = e1 + e2 , T (e2) = e2 + e3

and T(e3) = e1 + e2 + e3 is nonsingular and find its inverse.
Solution:- First we find the value of T.
Let (x1 , x2, x3) V3 such
 18

T(x1 , x2, x3) = T(x1e1 + x2 e2+ x3 e3)

= x1 T(e1) + x2 T (e2) + x3 T(e3)
= x1 (e1 + e2) + x2 (e2 + e3) + x3(e1 + e2 + e3)
= x1 (1,1,0) + x2 (0,1,1) + x3(1,1,1)
[ Where e1= (1,0,0) , e2 = (0,1,1)and e3= (0,1,1)]
= (x1 + x3 , x1+ x2 +x3, x2+ x3)
i.e. T(x1 , x2, x3) = (x1 + x3 , x1+ x2 +x3, x2+ x3)
Now for N(T)
T(x1 , x2, x3) = 0
(x1 + x3 , x1+ x2 +x3, x2+ x3)= 0
x1 + x3 = 0, x1 + x2+ x3 = 0 , x2 +x3 = 0,
 x1 = 0, x2 = 0 and x3 = 0
N(T) = 0
 This linear map T is one – one.
Since the dimension of domain space and dimension of co domain space are equal.
 T is onto.
Thus we get T is one-one and onto.
Hence T is nonsingular and the inverse of T exists. i.e.T-1 exists.
Now we derive the formula for T-1.
Let T-1(y1 , y2, y3) = (x1 , x2, x3).__________(1)
 (y1 , y2, y3) = T(x1 , x2, x3)
 (y1 , y2, y3) = (x1 + x3 , x1+ x2 +x3, x2+ x3)
 y1 = x1 + x3, y2 = x1+ x2 +x3, y3= x2+ x3
Solving these equation then we get
x1 = y2 – y3, x2 = y2 –y1 and x3 = y1– y2 + y3
put the values of x1 , x2, x3 in equation (1) then we get
T-1(y1 , y2, y3) = (y2 – y3, y2 –y1 , y1– y2 + y3)
Consequences of rank nullity theorem

Definition:-Isomorphic:-
Two vector spaces U and V are said to be isomorphic if there exists an isomorphism
from U to V.
If U and V are isomorphic then we write U ≈ V.

Theorem:- Every real (complex) vector space of dimension p is isomorphic to Vp( V pc )

Proof:- Let U be a real vector space of dimension p.
Let B = {u1 , u2 , u3,…, up} an ordered basis for U.
Let u  U
 u = 1 u1+  2 u2+…+  p up
i.e u is linear combination of set B.
 the coordinate vector of U relative to B is { 1 ,  2 ,…,  p }
Now define a mapping T : U  Vp by T(u) = ( 1 ,  2 ,…,  p )
We want to prove that T is linear map.
Let u, v  U where u = 1 u1+  2 u2+…+  p up and v = 1 u1+  2 u2+…+  p up
 19

 u + v = ( 1 + 1 )u1+(  2 +  2 )u2+…+(  p +  p )up

Now T(u + v) = T(( 1 + 1 )u1+(  2 +  2 )u2+…+(  p +  p )up)
= (( 1 + 1 ),(  2 +  2 ),…,(  p +  p )) _______(1) [ by defn of T]
And T(u) +T( v) = ( 1 ,  2 ,…,  p ) + ( 1 ,  2 ,…,  p )
= (( 1 + 1 ),(  2 +  2 ),…,(  p +  p )) _______(2)
From (1) and (2)
T(u + v) = T(u) +T( v) __________(3)
Similarly if  is any scalar and u  U then it can be proved that
T (  u) =  T(u) ____________(4)
From (3) and (4)
Hence T is linear map.
Now we want to prove that T is one –one.
Let u  N(T)
 T(u) = 0vp
 ( 1 ,  2 ,…,  p ) = 0
 1 = 0,  2 = 0,…,  p = 0
 u = 0u
N(T) is zero subspace of U.
 T is one –one.
Also T is onto.
Hence T is an isomorphism from U to Vp.
 U and Vp are isomorphic i.e U ≈ Vp .