Properties of Linear Transformation

Theorem: Let T : V W be a linear transformation, then for all

vectors u, v , v1 , . . . , vn V and scalars 1 , . . . , n F:
(i) T (0) = 0
(ii) T (u v ) = T (u) T (v )
(iii) T (1 v1 + + n vn ) = 1 T (v1 ) + + n T (vn ).
Proof: (i) T (0) = T (0 + 0) = T (0) + T (0) T (0) = 0.

(ii) T (u v ) = T (u + (1)v ) = T (u) + T (1(v ))

= T (u) + (1)T (v ) = T (u) T (v ).
(iii) Apply induction on n.

Consider the differential map, D : P3 P2 , defined by

D(p(X )) = p 0 (X ).
In this map, we know that
D(X 3 ) = 3X 2 ; D(X 2 ) = 2X ; D(X ) = 1 and D(1) = 0 and use
the linearity of differential operator to obtain D(p(X )) for any
polynomial p(X ) P3 .
Suppose T : R2 R be a linear map such that T (1, 0) = 2 and
T (0, 1) = 1, then what is T (2, 3)?
What are the information required to describe a linear map T ?
Theorem: Let V be a finite dimensional vector space with basis
B = {v1 , . . . , vn }. If T1 and T2 are two linear maps from V to
another vector space W such that T1 (vi ) = T2 (vi ) for all
i = 1, . . . , n, then T1 (v ) = T2 (v ) for all v V .

Proof of the theorem

Let v V . {v1 , . . . , vn } is a basis of V v = 1 v1 + + n vn
for some i s F.
Then
T1 (v ) = T1 (1 v1 + + n vn )
= 1 T1 (v1 ) + + n T1 (vn ) (using property 3)
= 1 T2 (v1 ) + + n T2 (vn ) (using the hypothesis)
= T2 (1 v2 + + n vn )
= T2 (v ).

Remark: This theorem tells us that if T : V W is linear and V

is finite dimensional, then we need to know only what T does to a
basis vectors in V . This determines T completely.

Example Let T : R3 R2 be linear such that T (1, 0, 0) = (2, 3),

T (0, 1, 0) = (1, 4) and T (0, 0, 1) = (5, 3). Then
T (3, 4, 5) = 3T (1, 0, 0) + (4)T (0, 1, 0) + 5T (0, 0, 1)
= 3(2, 3) + (4)(1, 4) + 5(5, 3)
= (35, 22).
Exercise: Let T : R2 R2 be a map such that T (1, 1) = (1, 1),
T (0, 1) = (1, 1) and T (2, 1) = (1, 0). Can T be a linear
transformation?
Question: Suppose we take a basis {v1 , . . . , vn } of V and
randomly chosen vectors w1 , . . . , wn W . Does there exists a
linear map T : V W such that T (vi ) = wi ?

Theorem: Let V be a finite dimensional vector space with basis

B = {v1 , . . . , vn }. Let W be a vector space containing the n
vectors w1 , . . . , wn . Then, there exists a unique linear map
T : V W such that T (vi ) = wi for i = 1, . . . , n.
Proof: We need to construct a map from V to W and prove that
this map is linear and is unique.
Let v V v = 1 v1 + + n vn for some 1 , . . . , n F.
Define T (v ) = 1 w1 + + n wn . Since, {v1 , . . . , vn } is a basis,
given a vector v , the scalars 1 , . . . , n are unique. Therefore, this
map is well-defined.
Linearity: Let u, v V . Then u = 1 v1 + + n vn and
v = 1 v1 + + n vn
u + v = (1 + 1 )v1 + + (n + n )vn .

Proof continued. . .
Therefore
T (u + v ) = (1 + 1 )w1 + + (n + n )wn
= (1 w1 + + n wn ) + (1 w1 + + n wn )
= T (u) + T (v ).
Similarly,
T (u) = T (1 v1 + + n vn )
= 1 w1 + + n wn = (1 w1 + + n wn )
= T (u).
Therefore T is a linear transformation.
Uniqueness follows from the previous theorem.

Caution: The vectors w1 , . . . , wn appearing in the previous

theorem, need not be distinct or not even be linearly independent.
Example: Construct a linear map T : R2 W , where
W = {(x1 , x2 , x3 ) | x1 x2 x3 = 0}. Describe the map
completely.
Solution: Start with a basis {v1 = (1, 0), v2 = (0, 1)} of R2 .
Choose any two vectors in W , for example w1 = (1, 1, 0) and
w2 = (1, 0, 1). We want T (1, 0) = (1, 1, 0) and T (0, 1) = (1, 0, 1).
Then define T (x1 , x2 ) = x1 (1, 1, 0) + x2 (1, 0, 1) = (x1 + x2 , x1 , x2 ).
Therefore, this is a linear map from R2 to W .
Exercise: Find another linear map from R2 to W .
[Lect - 12]

1 1

1
Let A = 0
2 1
T : R3 R3 given

1
2 . Consider the linear transformation
0
by T (x) = Ax for every x R3 .

Then Ae1 = (1, 0, 2); Ae2 = (1, 1, 1) & Ae3 = (1, 2, 0).
Note that (1, 0, 2) = 1 e1 + 0 e2 + 2 e3 .
Let V be a vector space and B = {v1 , . . . , vn } be a basis of V . If
v V , then there exist unique scalars 1 , . . . , n F such that
v = 1 v1 + + n vn .
In other words, all the information about v is contained in the
scalars 1 , . . . , n .

We express this fact by saying that

the
matrix of v with respect to
1

the basis B is the column matrix ... , denoted by [v ]B .
n
Example 1: Let B = {(1, 1), (1, 0)}. Find [(0, 1)]B .



1
Solution: (0, 1) = 1(1, 1) + 1(1, 0). [(0, 1)]B =
.
1
Example 2: Let B = {1, 1 + X , 1 + X 2 } P2 . Is B a basis of P2 ?
Find [1 + X + X 2 ]B .
Note here that the matrices would be different if we alter the
positions of the basis vectors.i.e., the matrices w.r.t.
{1, 1 + X 2 , 1 + X } and {1, 1 + X , 1 + X 2 } are different.

i.e., the position of vectors in a basis is important when we

compute the matrix of a vector.
An ordered basis is a basis with the positions of the basis vectors
fixed.
As bases of R3 , {e1 , e2 , e3 } and {e2 , e1 , e3 } are same, but as
ordered bases they are NOT THE SAME.
Let V and W be F-vector spaces and B1 = {v1 , . . . vn } and
{w1 , . . . , wm } be bases of V and W respectively. Let T : V W
be a linear transformation.
If u V , then u = 1 v1 + + n vn
T (u) = 1 T (v1 ) + + n T (vn ).

 Matrix of T (u) can be obtained once we know the matrices of

T (v1 ), . . . , T (vn ).
Let [T (vi )]B2 denote the matrix of T (vi ) w.r.t. B2 .
We define the matrix of T with respect to the ordered bases B1 and
2
B2 to be [T ]B
B1 := [[T (v1 )]B2 . . . [T (vn )]B2 ]].
Suppose T (vj ) = 1j w1 + + mj wm . Then

11 12 1n
21 22 2n

2
[T ]B
=
..
..
..
..
B1
.
.
.
.
m1 m2

mn

Example 1: Let T : R2 R3 be given by

T (x1 , x2 ) = (2x1 x2 , x1 + x2 , x2 x1 ), B1 = {e1 , e2 } and
B2 = {e1 , e2 , e3 }. Then
T (e1 ) = (2, 1, 1) = 2e1 + 1e2 + 1e3
T (e2 ) = (1, 1, 1) = 1e1 + 1e2 + 1e3 .

2 1
2
1
1 .
Therefore [T ]B
B1 =
1 1
Note that if A =

2
[T ]B
B1 ,

 
 
x1
x
then T
=A 1 .
x2
x2

Example 2: Let D : P3 P2 be the map given by D(p) = p 0 . Let

2
3
2
A = {1, X
, X , X } and
B = {1, X , X }. Then
0 1 0 0
0 0 2 0 .
[D]B
=
A
0 0 0 3
Let B = {1, 1 + X , 1 + X 2 }. Then compute [D]B
A

0 1 2 3
B

0 .
Ans: [D]A = 0 0 2
0 0 0
3
RX
Example 3: Let T : P2 P3 be the map T (p(X )) = 0 p(t)dt.
2
2
2
3
Let A = {1, 1 +
X , X + X } and
B = {1, X , X + X , X + X }.
0 0
0

1 1/2 1/6 .
[Lect - 13]
Then [T ]B
A = 0 1/2
1/6
0 0
1/3

Properties of matrices of linear maps

How are the matrices of x, T and T (x) related?
Theorem: Let V , W be finite dimensional vector spaces. Let
A = {v1 , . . . , vn } be an ordered basis of V and B = {w1 , . . . , wm }
be an ordered basis of W . Let T : V W be a linear
transformation. Then for every x V
[T (x)]B = [T ]B
A [x]A .
i.e., every linear transformation can be realized as a matrix
multiplication.
Theorem: Let V , W be finite dimensional vector spaces. Let
A = {v1 , . . . , vn } be an ordered basis of V and B = {w1 , . . . , wm }
be an ordered basis of W . If T1 and T2 are linear transformations
from V to W and F, then
B
B
1. [T1 + T2 ]B
A = [T1 ]A + [T2 ]A .
B
B
2. [T1 ]A = [T1 ]A .

Theorem: Let V , W , Z be finite dimensional vector spaces with

A, B, C their respective ordered bases. Let T1 : V W and
T2 : W Z be linear transformations. Then
[T2 T1 ]CA = [T2 ]CB [T1 ]B
A.
Definition: Let V and W be vector spaces and T : V W be a
linear transformation. Then
1. Kernel of T = N(T ) = {v V : T (v ) = 0}.
2. Range of T =
R(T ) = {w W : w = T (v ) for some v V }.

Note that ker T 6= , since T (0) = 0. For the same reason

Im T 6= .
Let u, v ker T and F. Then
T (u + v ) = T (u) + T (v ) = 0 u + v ker T
ker T is a subspace of V .
Similarly, if u, v Im T and F, then there exist x, y V such
that T (x) = u and T (y ) = v . Therefore,
T (x +y ) = T (x)+T (y ) = u +v u +v Im T Im T
is a subspace of W .
Theorem: Let T : V W is a linear transformation. Then
1. ker T is a subspace of V .
2. Im T is a subspace of W .
N(T ) is also called the null space of T .

Example: Let T : R3 R2 be defined by

T (x1 , x2 , x3 ) = (x1 + x2 , x1 x3 ). Find a basis for N(T ) and a
basis for R(T ).
Solution: T (x1 , x2 , x3 ) = (0, 0) x1 = x2 and x2 = x3 .
Therefore
N(T ) = {(x1 , x2 , x3 ) : x1 = x2 = x3 } = span({(1, 1, 1)}).
T (0, 1, 0) = (1, 0) and T (0, 0, 1) = (0, 1). Since R(T ) is a
subspace containing (1, 0) and (0, 1), R(T ) = R2 .
Example: Let T : R2 R3 be defined by
T (x1 , x2 ) = (x1 + x2 , x1 x2 , 0). Find R(T ) and N(T ).