Instructional Module for the Course

MATH 2250-LINEAR ALGEBRA

Module 5
Linear Transformations and Matrices
Overview

Module 5 discusses function mapping one vector space to another vector

space, one-to-one and onto linear transformations, the kernel and range of a
linear transformation and how it can be used to determine if the linear
transformation is one-to-one or onto, and the matrix of a linear transformation.

Objectives

After going through this module, you are expected to be able to do the
following:
• Determine whether a function mapping one vector space to another is a
linear transformation.
• Differentiate between one-to-one linear transformation and onto linear
transformation.
• Find the kernel, the range, the bases and dimension for the kernel and
range of a linear transformation.
• Using kernel and range, distinguish between one-to-one linear
transformation and onto linear transformation.
• Find the matrix of a linear transformation with respect to some bases.
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Learning Activities

Definition 5.1.1: Definition of Linear Transformation (Kolman & Hill, 2008,

p. 363)

Let 𝑉 and 𝑊 be vector spaces. A function 𝐿: 𝑉 → 𝑊 is called a linear

transformation of 𝑉 into 𝑊 if

a. 𝐿(𝐱 + 𝐲) = 𝐿(𝐱) + 𝐿(𝐲), for every vector 𝐱 and 𝐲 in 𝑉,

b. 𝐿(𝑐𝐱) = 𝑐𝐿(𝐱), for any vector 𝐱 in 𝑉 and every scalar 𝑐.

If 𝑉 is equal to 𝑊, the linear transformation 𝐿: 𝑉 → 𝑊 is also called a linear

operator on 𝑉.

Example 1. Let L: 𝑅3 → 𝑅3 be defined by

𝐿(𝑥, 𝑦, 𝑧) = (𝑥, 𝑦, 0)
Determine whether 𝐿 is a linear transformation.

Solution: 𝐿 is a linear transformation if the two conditions of Definition 5.1.1

are satisfied.

a) Let 𝐱 = (𝑥1 , 𝑦1 , 𝑧1 ) and 𝐲 = (𝑥2 , 𝑦2 , 𝑧2 ) ∈ 𝑅3 . Then

𝐿(𝐱 + 𝐲) = 𝐿(𝑥1 + 𝑥2 , 𝑦1 + 𝑦2 , 𝑧1 + 𝑧2 )
= (𝑥1 + 𝑥2 , 𝑦1 + 𝑦2 , 0)
= (𝑥1 , 𝑦1 , 0) + (𝑥2 , 𝑦2 , 0)
= 𝐿(𝐱) + 𝐿(𝐲)
Hence, 𝐿(𝐱 + 𝐲) = 𝐿(𝐱) + 𝐿(𝐲), for all 𝐱, 𝐲 ∈ 𝑅3 .

b) Let 𝐱 = (𝑥1 , 𝑦1 , 𝑧1) ∈ 𝑅3 and 𝑐 ∈ 𝑅. Then

𝐿(𝑐𝐱) = (𝑐𝑥1 , 𝑐𝑦1 , 0)

= 𝑐(𝑥1 , 𝑦1 , 0)
= 𝑐𝐿(𝐱)
Hence, 𝐿(𝑐𝐱) = 𝑐𝐿(𝐱) for all 𝐱 ∈ 𝑅3 and 𝑐 ∈ 𝑅.

Page 2 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Therefore, 𝐿 is a linear transformation.

𝑥1 2𝑥1 + 3𝑥2
Example 2. Let 𝐿: 𝑅 → 𝑅 be defined by 𝐿 [ 𝑥2 ] = [
3 3 0 ]. Is 𝐿 a linear
𝑥3 𝑥1 + 4𝑥3 − 3
transformation?

Solution: 𝐿 is not a linear transformation since

2 7 21
3𝐿 [1] = 3 [ 0 ] = [ 0 ]
3 11 33
while

2 6 21
𝐿 (3 [1]) = 𝐿 [3] = [ 0 ].
3 9 39

The second condition of Definition 5.1.1 is not satisfied.

Example 3. Let 𝐿: 𝑀22 → 𝑀22 be defined by 𝐿(𝐴) = 𝐴𝐶 − 𝐶𝐴 where 𝐶 is a fixed

2 × 2 matrix over the set of real numbers. Is 𝐿 a linear
transformation?

Solution: Let 𝐴, 𝐵 ∈ 𝑀22 . Then

(a) 𝐿(𝐴 + 𝐵) = (𝐴 + 𝐵)𝐶 − 𝐶(𝐴 + 𝐵)

= 𝐴𝐶 + 𝐵𝐶 − 𝐶𝐴 − 𝐶𝐵 Distributive property
and
𝐿(𝐴) + 𝐿(𝐵) = (𝐴𝐶 − 𝐶𝐴) + (𝐵𝐶 − 𝐶𝐵)
= 𝐴𝐶 + 𝐵𝐶 − 𝐶𝐴 − 𝐶𝐵 Associative property

Hence, 𝐿(𝐴 + 𝐵) = 𝐿(𝐴) + 𝐿(𝐵) for all 𝐴, 𝐵 ∈ 𝑀22

Page 3 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

(b) Let 𝐴 ∈ 𝑀22 and 𝑐 ∈ 𝑅. Then

𝐿(𝑐𝐴) = (𝑐𝐴)𝐶 − 𝐶(𝑐𝐴)
= 𝑐(𝐴𝐶 − 𝐶𝐴)
= 𝑐𝐿(𝐴)

Hence, 𝐿(𝑐𝐴) = 𝑐𝐿(𝐴) for all 𝐴 ∈ 𝑀22 and 𝑐 ∈ 𝑅

Therefore, 𝐿 is a linear transformation.

Example 4. Let 𝐿: 𝑅2 → 𝑅2 be defined by

𝑥 𝑥2 + 𝑥
𝐿 [𝑦 ] = [ ]
𝑦 – 𝑦2
𝑥1 𝑥2
Solution: Let 𝐱 = [𝑦 ], 𝐲 = [𝑦 ] be vectors in 𝑅 2 . 𝐿 is not a linear transformation
1 2
since
𝑥 +𝑥
𝐿(𝐱 + 𝐲) = 𝐿 [𝑦1 + 𝑦2 ]
1 2

(𝑥1 + 𝑥2 )2 + 𝑥1 + 𝑥2
=[ ]
𝑦1 + 𝑦2 − (𝑦1 + 𝑦2 )2

while

𝑥1 2 + 𝑥1 𝑥2 2 + 𝑥2
𝐿(𝐱) + 𝐿(𝐲) = [ ] + [ ]
𝑦1 – 𝑦1 2 𝑦2 – 𝑦2 2

𝑥1 2 + 𝑥2 2 + 𝑥1 + 𝑥2
=[ ]
𝑦1 + 𝑦2 – 𝑦1 2 − 𝑦2 2

Thus, 𝐿(𝐱 + 𝐲) ≠ 𝐿(𝐱) + 𝐿(𝐲) for some choice of 𝐱 and 𝐲.

Page 4 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Theorem 5.1.1: (Kolman & Hill, 2008, p. 367)

Let 𝐿: 𝑉→ 𝑊 be a linear transformation. Then

a. 𝐿(𝑶𝑉 ) = 𝑶𝑊
b. 𝐿(−𝐱) = − 𝐿(𝐱) for every 𝐱 ∈ 𝑅 𝑛
c. 𝐿(𝐱 – 𝐲) = 𝐿(𝐱) – 𝐿(𝐲)

Proof of Theorem 5.1.1.a.

a. 𝐿(𝑶) = 𝐿 (𝑶 + 𝑶)
𝐿(𝑶) = 𝐿(𝑶) + 𝐿(𝑶)
𝐿(𝑶) – 𝐿(𝑶) = 𝐿(𝑶) + 𝐿(𝑶) – 𝐿(𝑶)
𝑶 = 𝐿(𝑶)

Proof of Theorem 5.1.1.b.

b. 𝐿(−𝐱) = −𝐿(𝐱)
𝐿(−𝐱) = 𝐿(−1⦁ 𝐱)
= −1𝐿(𝐱)
= −𝐿(𝐱)

Proof of Theorem 5.1.1.c.

c. 𝐿(𝐱 − 𝐲) = 𝐿(𝐱 + (−1)𝐲)
= 𝐿(𝐱) + 𝐿(−1⦁ 𝐲)
= 𝐿(𝐱) − 𝐿(𝐲)

Example 5. Let 𝐿: 𝑅3 →𝑅2 be defined by 𝐿(𝑥, 𝑦, 𝑧) = (𝑥, 𝑦) be a linear

transformation. It can be verified that

𝐿(0,0,0) = (0, 0)

That is, the zero vector in the domain (𝑅3 ) is mapped to the zero vector in the
codomain (𝑅2 ).

Page 5 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Also, let 𝐱 = (𝑥1 , 𝑦1 , 𝑧1 ) and 𝐲 = (𝑥2 , 𝑦2 , 𝑧2 ) be any vectors in 𝑅 3 then

𝐱 − 𝐲 = (𝑥1 − 𝑥2 , 𝑦1 − 𝑦2 , 𝑧1 − 𝑧2 ).

Thus,
𝐿(𝐱 − 𝐲) = (𝑥1 − 𝑥2 , 𝑦1 − 𝑦2 )
= (𝑥1 , 𝑦1 ) − (𝑥2 , 𝑦2 )
= 𝐿(𝐱) – 𝐿(𝐲)

Example 6. Let 𝑇: 𝑅2 → 𝑅2 be the “translation mapping” defined by

𝑇(𝑥, 𝑦) = (𝑥 + 1, 𝑦 + 2)

By Theorem 5.1.1 (a), 𝑇 is not a linear transformation since

𝑇(0, 0) = (0 + 1, 0 + 2) = (1, 2) ≠ (0, 0)

(The zero vector in the domain is not mapped into the zero vector in the codomain)

Note: In Example 6, since 𝐿(𝑶𝑉 ) ≠ 𝑶𝑊 , it was not necessary anymore to check

the two conditions of Definition 5.1.1. Remember that 𝐿(𝑶𝑉 ) = 𝑶𝑊 is a
consequence of these conditions. However, if the zero vector in the domain
is mapped to the zero vector in the codomain, then it is necessary to check
the two conditions of Definition 5.1.1. See the next example.

Example 7. Let 𝐿 be the mapping defined in Example 4. We see that 𝐿 is not a

linear transformation even though

0 0
𝐿[ ] = [ ]
0 0

Theorem 5.1.2: If 𝐿: 𝑉 → 𝑊 is a linear transformation, then

𝐿(𝑐1𝐱1 + 𝑐2 𝐱 2 + ⋯ + 𝑐𝑘 𝐱 𝑘 ) = 𝑐1𝐿(𝐱1) + 𝑐2 𝐿(𝐱 2) + ⋯ + 𝑐𝑘 𝐿(𝐱 𝑘 )

for any vectors 𝐱1, 𝐱 2 , ⋯ , 𝐱 𝑘 in 𝑉 and any scalars 𝑐1 , 𝑐2 , … , 𝑐𝑘 .

Page 6 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

SAQ 5.1
Which of the following are linear transformations?

𝑥– 𝑦
𝑥
(a) 𝐿: 𝑅 → 𝑅 defined by 𝐿 [𝑦] = [ 0 ]
2 3

2𝑥 + 3
𝑥 2𝑥 − 3𝑦
(b) 𝐿: 𝑅3 → 𝑅3 defined by 𝐿 ([𝑦]) = [ 3𝑦 − 2𝑧 ]
𝑧 2𝑧
1
(c) 𝐿: 𝑃𝑛 → 𝑅 defined by 𝐿(𝑝(𝑥 )) = ∫0 𝑝(𝑥 )𝑑𝑥

Do the exercises on your own before looking at the solution.

Theorem 5.1.3: (Kolman & Hill, 2008, p. 368)

Let 𝐿: 𝑉 → 𝑊 be a linear transformation of an 𝑛-dimensional vector space 𝑉 into

a vector space 𝑊. Also let 𝑆 = { 𝐱1 , 𝐱 2 , … , 𝐱 𝑛 } be a basis for 𝑉. If 𝐱 is any vector
in 𝑉, then 𝐿(𝐱) is completely determined by {𝐿(𝐱1 ), 𝐿(𝐱 2 ), … , 𝐿(𝐱 𝑛 ) }.

Example 8. Let 𝐿: 𝑃2 → 𝑃1 be a linear transformation for which we know that

𝐿(𝑥 2 + 𝑥 ) = 2𝑥 + 1, 𝐿(𝑥 − 1) = 𝑥 + 2 and 𝐿(𝑥 + 1) = 𝑥. Let 𝐩(𝐱) =
2𝑥 2 − 3𝑥 + 1.

a. Find 𝐿(𝐩(𝐱)) b) What is 𝐿(𝑎𝑥 2 + 𝑏𝑥 + 𝑐)?

Solution: It can be verified that {𝑥 2 + 𝑥, 𝑥 − 1, 𝑥 + 1} is a basis for 𝑃2 . Next, we

express 𝐩(𝐱) = 2𝑥 2 − 3𝑥 + 1 as a linear combination of the basis vectors.

2𝑥 2 − 3𝑥 + 1 = 2(𝑥 2 + 𝑥 ) − 3(𝑥 − 1) − 2(𝑥 + 1)

Page 7 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Now, we apply 𝐿 to both sides of the equation

𝐿 (2𝑥 2 − 3𝑥 + 1) = 𝐿(2(𝑥 2 + 𝑥 ) − 3(𝑥 − 1) − 2(𝑥 + 1))
Applying Theorem 5.1.2, we have
𝐿((2𝑥 2 − 3𝑥 + 1)) = 2𝐿(𝑥 2 + 𝑥 ) − 3𝐿(𝑥 − 1) − 2𝐿 (𝑥 + 1)
= 2(2𝑥 + 1) − 3(𝑥 + 2) − 2𝑥
= −𝑥 − 4

(b) Express 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 as a linear combination of the basis vector.

1 1
𝑎𝑥 2 + 𝑏𝑥 + 𝑐 = 𝑎(𝑥 2 + 𝑥) + (−𝑎 + 𝑏 − 𝑐 )(𝑥 − 1) + (−𝑎 + 𝑏 + 𝑐)(𝑥 + 1)
2 2
Applying Theorem 5.1.2, we have

1 1
𝐿(𝑎𝑥 2 + 𝑏𝑥 + 𝑐 ) = 𝑎𝐿(𝑥 2 + 𝑥) + 2 (−𝑎 + 𝑏 − 𝑐 )𝐿(𝑥 − 1) + 2 (−𝑎 + 𝑏 + 𝑐)𝐿(𝑥 + 1)
1 1
= 𝑎(2𝑥 + 1) + (−𝑎 + 𝑏 − 𝑐 )(𝑥 + 2) + (−𝑎 + 𝑏 + 𝑐)𝑥
2 2

= (𝑎 + 𝑏)𝑥 + (𝑏 − 𝑐)

5.2 The Kernel and Range of a Linear Transformation

Definition 5.2.1: A linear transformation 𝐿: 𝑉 → 𝑊 is called one-to-one if for

all 𝐱1, 𝐱 2 in 𝑉, 𝐱1 ≠ 𝐱 2 implies that 𝐿(𝐱1) ≠ 𝐿(𝐱 2 ). An equivalent statement is that
𝐿 is one-to-one if for all 𝐱1, 𝐱 2 in 𝑉, 𝐿(𝐱1 ) = 𝐿(𝐱 2 ) implies that 𝐱1 = 𝐱 2 .

Example 1. Let 𝐿: 𝑅2 → 𝑅2 be a linear transformation defined by

𝐿( 𝑥, 𝑦 ) = ( 𝑥 + 𝑦, 𝑥 + 2𝑦 )

Determine whether 𝐿 is one–to–one.

Solution:

Let 𝐱1 = (𝑥1 , 𝑦1 ) and 𝐱 2 = (𝑥2 , 𝑦2 ) be vectors in 𝑅2 . We have to show that if

𝐿(𝐱1 ) = 𝐿(𝐱 2 ) then 𝐱1 = 𝐱 2 .

Page 8 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Assume that
𝐿(𝐱1) = 𝐿(𝐱 2 )
then
(𝑥1 + 𝑦1 , 𝑥1 + 2𝑦1 ) = (𝑥2 + 𝑦2 , 𝑥2 + 2𝑦2 )

Equating the corresponding parts, we have

𝑥1 + 𝑦1 = 𝑥2 + 𝑦2
𝑥1 + 2𝑦1 = 𝑥2 + 2𝑦2

If we subtract the first equation from the second, we get 𝑦1 = 𝑦2 . This implies that
𝐱1 = 𝐱 2 . Thus 𝐿 is one-to-one.

Example 2. Let 𝐿: 𝑅3 → 𝑅3 be a linear transformation defined by

𝐿(𝑥, 𝑦, 𝑧) = (𝑥, 𝑦, 0)

Determine if 𝐿 is one-to-one or not.

Solution:

Let 𝐱1 = (4, 5, −3) and 𝐱 2 = (4, 5, 2) be vectors in 𝑅3 . We see that 𝐱1 ≠ 𝐱 2 but

𝐿(4, 5, −3) = 𝐿(4, 5, 2) = (4, 5, 0).

Hence 𝐿 is not one-to-one.

Definition 5.2.2: (Kolman & Hill, 2008, p. 376)

Let 𝐿: 𝑉 → 𝑊 be a linear transformation of a vector space 𝑉 into a vector space

𝑊. The kernel of 𝐿, ker𝐿, is the subset of 𝑉 consisting of all vectors 𝐱 of 𝑉 such
that 𝐿(𝐱) = 𝑶𝑤

Note that ker𝐿 is not empty since by Theorem 5.1.1 we know that if 𝐿: 𝑉→ 𝑊 is
a linear transformation then the zero vector in 𝑉 is mapped to the zero vector in
𝑊. Thus 𝑶𝑉 ∈ ker𝐿.

Page 9 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Example 3. Let 𝐿: 𝑅4 → 𝑅3 be a linear transformation defined by

𝐿(𝑥, 𝑦, 𝑧, 𝑤) = (𝑥 + 𝑦, 𝑧 + 𝑤, 𝑥 + 𝑧)

The vector (1, −1, −1, 1) is in ker𝐿 since 𝐿(1, −1, −1, 1) = (0, 0, 0) while the vector
(1, 2, 3, −4) is not in kerL because 𝐿(1, 2, 3, −4) = ( 3, 1, 5 ) ≠ (0, 0, 0).

Example 4. Let 𝐿 be the linear transformation of Illustration 1. The ker𝐿

consists of all vectors 𝐱 in 𝑅4 such that 𝐿(𝐱) = 0. If we let
𝐱 = (𝑥, 𝑦, 𝑧, 𝑤) then

𝐿(𝑥, 𝑦, 𝑧, 𝑤) = (0, 0, 0)
(𝑥 + 𝑦, 𝑧 + 𝑤, 𝑥 + 𝑧) = (0, 0, 0)

Equating the corresponding parts we obtain the homogeneous system

𝑥+𝑦 =0
𝑧 + 𝑤=0
𝑥 +𝑧 =0

The augmented matrix in reduced echelon form (verify) is

1 0 0 −1 0
[0 1 0 1 0]
0 0 1 1 0

This implies that 𝑤 is a free variable and can take on any value. Hence

𝑥=𝑟
𝑦 = −𝑟
𝑧 = −𝑟
𝑤 = 𝑟 where 𝑟 ∈ 𝑅

So the ker 𝐿 consists of all vectors of the form (𝑟, −𝑟, −𝑟, 𝑟) where 𝑟 is any real
number. That is

ker𝐿 = { (𝑟, −𝑟, −𝑟, 𝑟): 𝑟  𝑅}

= { 𝑟(1, −1, −1, 1): 𝑟  𝑅}
= span { (1, −1, −1, 1)}

Page 10 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Since (1, −1, −1, 1) is linearly independent then it forms a basis for ker 𝐿. Thus,
dim(kerL) = 1.

Example 5. Let 𝐿: 𝑃3 → 𝑀22 be a linear transformation defined by

𝑑+𝑐 𝑑 − 2𝑏
𝐿(𝑎𝑡 3 + 𝑏𝑡 2 + 𝑐𝑡 + 𝑑 ) = [ ]
𝑎 𝑐−𝑎

Find ker𝐿.

Solution: Let 𝑎𝑡 3 + 𝑏𝑡 2 + 𝑐𝑡 + 𝑑 ∈ ker𝐿. Then

0 0
𝐿(𝑎𝑡 3 + 𝑏𝑡 2 + 𝑐𝑡 + 𝑑 ) = [ ]
0 0

𝑐+𝑑 𝑑 − 2𝑏 0 0
[ ]=[ ]
𝑎 𝑐−𝑎 0 0

Equating the corresponding parts, we obtain the homogeneous system

𝑐+𝑑 =0
𝑑 − 2𝑏 = 0
𝑎=0
𝑐−𝑎 = 0

The augmented matrix in reduced echelon form (verify) is

1 0 0 0 0
0 1 0 0 0
[ ]
0 0 1 0 0
0 0 0 1 0

The system has solution 𝑎 = 𝑏 = 𝑐 = 𝑑 = 0. Therefore,

ker𝐿 = 0 (the zero polynomial)

and dim(ker𝐿) = 0.

Page 11 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Theorem 5.2.1: If 𝐿: 𝑉 → 𝑊 is a linear transformation, then ker 𝐿 is a subspace

of 𝑉.

Proof: We need to show that ker 𝐿 is closed under addition and scalar
multiplication.

(a) Let 𝐱 and 𝐲 be vectors in the ker 𝐿. Then

𝐿(𝐱) = 𝐿(𝐲) = 𝑶𝑾 Definition of ker 𝐿

We have

𝐿(𝐱 + 𝐲) = 𝐿(𝐱) + 𝐿(𝐲) 𝐿 is a linear transformation

= 𝑶𝑾 + 𝑶𝑾
= 𝑶𝑾
This implies that 𝐱 + 𝐲 ∈ ker 𝐿.

(b) Let 𝐱 be a vector in ker 𝐿 and let 𝑐 ∈ 𝑅. Then

𝐿(𝑐𝐱) = 𝑐𝐿(𝐱) 𝐿 is a linear transformation

= 𝑐𝑶𝑾
= 𝑶𝑾

This implies that 𝑐𝐱 ∈ ker 𝐿.

Therefore, ker𝐿 is a subspace of 𝑉.

Definition 5.2.3: (Kolman & Hill, 2008, p. 378)

If 𝐿: 𝑉 → 𝑊 is a linear transformation of a vector space 𝑉 into a vector space 𝑊,

then the range of 𝐿 or image of 𝑉 under 𝐿, denoted by range 𝐿, consists of all
those vectors in 𝑊 that are images under 𝐿 of vectors in 𝑉. Thus a vector 𝐲 is in
range 𝐿 if there exists some vector 𝐱 in 𝑉 such that 𝐿(𝐱) = 𝐲. If range 𝐿 = 𝑊, we
say that 𝐿 is onto.

Example 6. Let 𝐿: 𝑅3 → 𝑅2 be a linear transformation defined by

𝐿(𝑥, 𝑦, 𝑧) = (𝑥 + 𝑦, 𝑦 − 𝑧)

Page 12 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

The range or image of 𝐿 (Im𝐿) is:

Im 𝐿 = { (𝑥 + 𝑦, 𝑦 − 𝑧) }
= { 𝑥 (1, 0) + 𝑦 (1, 1) + 𝑧 (0, −1) ∶ 𝑥, 𝑦, 𝑧  𝑅 }
= span {(1, 0), (1, 1), (0, −1)}

By Theorem 4.5.1 {(1, 0), (1, 1), (0, −1)} is linearly dependent. Since (0, −1) is a
linear combination of the other vectors such as

(0, −1) = 1 (1, 0) + (−1)(1, 1)

then we can delete (0, −1) from the set and the remaining vectors still span Im𝐿.
The set {(1, 0), (1, 1)} is linearly independent therefore forms a basis for Im𝐿.
Hence dim(Im𝐿) = 2.

Theorem 5.2.2: Let 𝐿: 𝑉 → 𝑊 be a linear transformation

(i) 𝐿 is a monomorphism (one-to-one) if and only if dim(ker𝐿) = 0.

(ii) 𝐿 is an epimorphism (onto) if and only if dim(Im𝐿) = dim 𝑊.

Example 7. Let 𝐿: 𝑅2 → 𝑅2 be a linear transformation defined by

𝑢 2𝑢 − 5𝑣
𝐿[ ] = [ ]
𝑣 4𝑢 + 3𝑣

Show that 𝐿 is a one-to-one linear transformation.

𝑢
Solution: Let 𝐱 = [ ] ∈ ker 𝐿, then
𝑣

2𝑢 − 5𝑣 0
[ ]=[ ]
4𝑢 + 3𝑣 0

Which is equivalent to the system,

2𝑢 − 5𝑣 = 0
4𝑢 + 3𝑣 = 0

Page 13 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

0
This system has solution 𝑢 = 𝑣 = 0. Hence, ker 𝐿 = [ ] and dim(ker𝐿) = 0
0
By Theorem 5.2.2 (i), 𝐿 is one-to-one.

Example 8. Determine whether the linear transformation 𝐿: 𝑀22 → 𝑅3 defined

𝑎+𝑏
𝑎 𝑏
by 𝐿 [ ] = [𝑏 + 𝑑 ] is onto.
𝑐 𝑑
𝑐+𝑎

Solution:
𝑎+𝑏
Im 𝐿 = [𝑏 + 𝑑 ]
𝑐+𝑎

1 1 0
= 𝑎 [0] + 𝑏 [1] + 𝑐 [0]
1 0 1

1 1 0
= span {[0] , [1] , [0]}
1 0 1

1 1 0
The determinant of [0 1 0] ≠ 0 hence, the set of vectors in the spanning set is
1 0 1
linearly independent and by Theorem 4.6.1 (a), it is a basis for the Im 𝐿. Since
dim(Im𝐿) = 3 = dim𝑅3 , then 𝐿 is onto.

Theorem 5.2.3: (Kolman & Hill, 2008, p. 378)

If 𝐿: 𝑉 → 𝑊 is a linear transformation of a vector space 𝑉 into a vector space 𝑊,

then range 𝐿 is a subspace of 𝑊.

Proof: We need to show that the range 𝐿 is closed under addition and scalar
multiplication.

(a) Let 𝐰𝟏 and 𝐰𝟐 be vectors in the range 𝐿. Then there exist vectors 𝐯1 and 𝐯2
in 𝑉 such that

𝐿(𝐯1 ) = 𝐰1 and 𝐿(𝐯2 ) = 𝐰2

Page 14 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

We have

𝐿(𝐯1 + 𝐯2 ) = 𝐿 (𝐯1 ) + 𝐿(𝐯2 ) 𝐿 is a linear transformation

= 𝐰1 + 𝐰2
This implies that 𝐿(𝐯1 + 𝐯2 ) is in range 𝐿.

(b) Let 𝐰𝟏 be a vector in range 𝐿 and 𝑐 ∈ 𝑅. Then

𝐿(𝑐𝐯1 ) = 𝑐𝐿(𝐯1 ) 𝐿 is a linear transformation

= 𝑐𝐰1

which shows that 𝐿(𝑐𝐯1 ) is in range 𝐿. Therefore, range 𝐿 is a subspace of 𝑊.

Theorem 5.2.4: (Kolman & Hill, 2008, p. 381)

If 𝐿: 𝑉 → 𝑊 is a linear transformation of an 𝑛 dimensional vector space 𝑉 into a

vector space 𝑊, then

dim ker𝐿 + dim range 𝐿 = dim 𝑉

Example 9. Verify Theorem 5.2.4 by using the linear transformation of Example

6, Section 5.2.

Solution:

ker𝐿 = { (𝑥, 𝑦, 𝑧) ∶ (𝑥 + 𝑦, 𝑦 − 𝑧) = (0, 0)}

Equating the corresponding parts we get:

𝑥+𝑦 = 0
𝑦−𝑧= 0

The augmented matrix of this system is

1 1 0 0
[ ]
0 1 −1 0

Page 15 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

which is row equivalent to

1 0 1 0
[ ]
0 1 −1 0

The system has solution

𝑥 = −𝑟, 𝑦 = 𝑟, and 𝑧 = 𝑟 where 𝑟  𝑅

Thus,
ker𝐿 = { (−𝑟, 𝑟, 𝑟) ∶ 𝑟  𝑅}
= { 𝑟(−1, 1, 1) ∶ 𝑟  𝑅}
= span {(−1, 1, 1)}

Since (−1, 1, 1) is linearly independent then it forms a basis for ker𝐿. Thus
dim(ker𝐿) = 1.

From the previous example, we know that dim(Im𝐿) = 2. Therefore

dim(ker𝐿) + dim(Im𝐿) = dim 𝑅3

1 + 2 = 3

Remark: Let 𝐿: 𝑉 → 𝑊 be a linear transformation. Then the rank of 𝑳 is defined

to be the dimension of its image, and the nullity of 𝑳 is defined to be the
dimension of its kernel.

rank (𝐿) = dim(Im𝐿) and nullity (𝐿) = dim(ker𝐿)

Thus, the preceding theorem yields the following formula for 𝐿 when 𝑉 has a finite
dimension:
rank(𝐿) + nullity(𝐿) = dim 𝑉.

Page 16 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Corollary 5.2.5: (Kolman & Hill, 2008, p. 384)

If 𝐿: 𝑉 → 𝑊 is a linear transformation of a vector space 𝑉 into a vector space 𝑊

and dim 𝑉 = dim 𝑊, then the following statements are true:
a. If 𝐿 is one-to-one, then it is onto.
b. If 𝐿 is onto, then it is one-to-one.

Example 10. Determine whether the linear transformation 𝐿: 𝑅 2 → 𝑅2 defined

by
𝑥 𝑥 + 2𝑦
𝐿 ([𝑦]) = [ ]
2𝑥 − 𝑦

is one-to-one and onto.

Solution:
𝑥 + 2𝑦
Im𝐿 = [ ]
2𝑥 − 𝑦
1 2
= 𝑥[ ]+𝑦[ ]
2 −1
1 2
= span {[ ] , [ ]}
2 −1

1 2
Since the determinant of [ ] ≠ 0, then the vectors in the spanning set is
2 −1
linearly independent and dimension of Im𝐿 is 2. By Corollary 5.2.5 (b), 𝐿 is one-
to-one and onto linear transformation.

Page 17 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

SAQ 5.2

1. Let 𝐿: 𝑅3 → 𝑅 4 be defined by

𝐿(𝑥, 𝑦, 𝑧) = (𝑥 + 𝑦 + 𝑧, 𝑥 + 2𝑦 – 3𝑧, 2𝑥 + 3𝑦 – 2𝑧, 3𝑥 + 4𝑦 – 𝑧)

(a) Find a basis for and the dim(ker𝐿),

(b) Find a basis for and the dim(Im𝐿),
(c) Verify Theorem 5.2.4.

2. Let 𝐿: 𝑃2 → 𝑃2 be the linear transformation defined by

𝐿(𝑎𝑡 2 + 𝑏𝑡 + 𝑐) = (𝑎 + 𝑐)𝑡 2 + (𝑏 + 𝑐)𝑡.

(a) Is 𝑡 2 – 𝑡 – 1 in ker𝐿? (d) Find a basis for ker𝐿.

(b) Is 𝑡 2 + 𝑡 − 1 in ker𝐿? (e) Find a basis for Im𝐿.
(c) Is 2𝑡 2 – 𝑡 in range 𝐿?

Do the exercises on your own before looking at the solution.

Page 18 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

If you have no further

questions, you can now
answer the assignment.

End of Module 5

References

Anton, H. & Rorres, C. (2005). Elementary Linear Algebra (9th ed.). John Wiley &
Sons, Inc.
https://www.sku.ac.ir/Datafiles/BookLibrary/55/Howard%20Anton,%2
0Chris%20Rorres%20-
%20Elementary%20Linear%20Algebra%20with%20Applications-
Wiley%20(2005).pdf

Kolman, B., & Hill, David R. (2008). Elementary Linear Algebra with Applications
(9th ed.). Upper Saddle River, NJ: Prentice Hall.

Larson, R. & Falvo, David C. (2009). Elementary Linear Algebra (6th ed.).
Houghton Mifflin Harcourt Publishing Company.
https://www.kau.edu.sa/Files/0009138/Subjects/Larson%20-
%20Edwards%20y%20Falvo%20Elementary%20Linear%20Algebra.p
df

Page 19 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

Answers to Self-assessment Questions

ASAQ 5.1
𝑥 +𝑥
a) 𝐿(𝐱 + 𝐲) = 𝐿 [𝑦1 + 𝑦2 ]
1 2

(𝑥1 + 𝑥2 ) − (𝑦1 + 𝑦2 )
=[ 0 ]
2(𝑥1 + 𝑥2 ) + 3
𝑥1 + 𝑥2 − 𝑦1 − 𝑦2
=[ 0 ]
2𝑥1 + 2𝑥2 + 3

and
𝑥1 − 𝑦1 𝑥2 − 𝑦2
𝐿(𝐱) + 𝐿 (𝐲) = [ 0 ] + [ 0 ]
2𝑥1 + 3 2𝑥2 + 3
𝑥1 + 𝑥2 − 𝑦1 − 𝑦2
=[ 0 ]
2𝑥1 + 2𝑥2 + 6

Since 𝐿(𝐱 + 𝐲) ≠ 𝐿(𝐱) + 𝐿(𝐲) then 𝐿 is not a linear transformation.

𝑥1 𝑥2
b) Let 𝐱 = [𝑦1 ] and 𝐲 = [𝑦2 ] be vectors in 𝑅3 and 𝑐 ∈ 𝑅. Then
𝑧1 𝑧2

𝑥1 + 𝑥2 2(𝑥1 + 𝑥2 ) − 3(𝑦1 + 𝑦2 )
(i) 𝐿(𝐱 + 𝐲) = 𝐿 ([𝑦1 + 𝑦2 ]) = [ 3(𝑦1 + 𝑦2 ) − 2(𝑧1 + 𝑧2 ) ]
𝑧1 + 𝑧2 2(𝑧1 + 𝑧2 )
2𝑥1 − 3𝑦1 2𝑥2 − 3𝑦2
= [ 3𝑦1 − 2𝑧1 ] + [ 3𝑦2 − 2𝑧2 ]
2𝑧1 2𝑧2

= 𝐿(𝐱) + 𝐿(𝐲), for every vector 𝐱 and 𝐲 in 𝑅3 .

Page 20 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

𝑐𝑥1 2𝑐𝑥1 − 3𝑐𝑦1

ii) 𝐿 (c𝐱) = 𝐿 ([𝑐𝑦1 ]) = [ 3𝑐𝑦1 − 2𝑐𝑧1 ]
𝑐𝑧1 2𝑐𝑧1

2𝑥1 − 3𝑦1
= 𝑐 [ 3𝑦1 − 2𝑧1 ]
2𝑧1

= 𝑐𝐿(𝐱), for every vector 𝐱 in 𝑅3 and 𝑐 ∈ 𝑅.

Therefore, 𝐿 is a linear transformation.

c) Let 𝑝(𝑥 ), 𝑞(𝑥) ∈ 𝑃𝑛 , 𝑐 ∈ 𝑅. Then

1
i) 𝐿(𝑝(𝑥 ) + 𝑞(𝑥 )) = ∫0 (𝑝(𝑥 ) + 𝑞(𝑥 ))𝑑𝑥
1 1
= ∫0 𝑝(𝑥 )𝑑𝑥 + ∫0 𝑞(𝑥 )𝑑𝑥

= 𝐿(𝑝(𝑥 )) + 𝐿(𝑞(𝑥 )), for every vector 𝑝(𝑥) and 𝑞(𝑥) in 𝑃𝑛 .

1
ii) 𝐿(𝑐𝑝(𝑥 )) = ∫0 𝑐𝑝(𝑥 )𝑑𝑥
1
= 𝑐 ∫0 𝑝(𝑥 )𝑑𝑥

= 𝑐𝐿(𝑝(𝑥) for every vector 𝑝(𝑥) in 𝑃𝑛 and 𝑐 ∈ 𝑅.

Therefore, 𝐿 is a linear transformation.

Page 21 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

ASAQ 5.2

(a) (𝑥 + 𝑦 + 𝑧, 𝑥 + 2𝑦 – 3𝑧, 2𝑥 + 3𝑦 – 2𝑧, 3𝑥 + 4𝑦 – 𝑧) = (0,0,0,0)

Equating the corresponding parts, we obtain the homogeneous system:

𝑥 + 𝑦 + 𝑧=0
𝑥 + 2𝑦 – 3𝑧 = 0
2𝑥 + 3𝑦 – 2𝑧 = 0
3𝑥 + 4𝑦 – 𝑧 = 0

The augmented matrix in reduced echelon form (verify) is

1 0 5 0
0 1 −4 0
[ ]
0 0 0 0
0 0 0 0

This implies that 𝑧 is a free variable and can take on any value. Hence

𝑥 = −5𝑟
𝑦 = 4𝑟
𝑧 = 𝑟 where 𝑟 ∈ 𝑅

Thus, ker𝐿 consists of all vectors of the form (−5𝑟, 4𝑟, 𝑟) where 𝑟 is any real
number. That is

ker𝐿 = { ((−5𝑟, 4𝑟, 𝑟): 𝑟  𝑅}

= { 𝑟(−5,4,1): 𝑟  𝑅}
= span {(−5,4,1)}

Since (−5,4,1) is linearly independent then it forms a basis for ker𝐿. Thus,
dim(ker𝐿) = 1.

Page 22 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

(b) Im 𝐿 = { (𝑥 + 𝑦 + 𝑧, 𝑥 + 2𝑦 – 3𝑧, 2𝑥 + 3𝑦 – 2𝑧, 3𝑥 + 4𝑦 – 𝑧) }

= { 𝑥 (1,1,2,3) + 𝑦 (1,2,3,4) + 𝑧 (1, −3, −2, −1) ∶ 𝑥, 𝑦, 𝑧  𝑅 }
= span {(1,1,2,3), (1,2,3,4), (1, −3, −2, −1)}

To determine linear independence of the vectors in the spanning set, we form the
equation

𝑐1(1,1,2,3) + 𝑐2 (1,2,3,4) + 𝑐3 (1, −3, −2, −1) = (0,0,0,0)

The equivalent system is

𝑐1 + 𝑐2 + 𝑐3 = 0
𝑐1 + 2𝑐2 − 3𝑐3 = 0
2𝑐1 + 3𝑐2 − 2𝑐3 = 0
3𝑐1 + 4𝑐2 − 𝑐3 = 0

The augmented matrix is

1 1 1 0
1 2 −3 0
[ ]
2 3 −2 0
3 4 −1 0

which is row equivalent to

1 0 5 0
0 1 −4 0
[ ].
0 0 0 0
0 0 0 0

Since the leading 1s appear in columns 1 and 2, then

{(1,1,2,3), (1,2,3,4)}

is a basis for Im𝐿. Therefore the dim(Im𝐿) = 2.

Note: The augmented matrix used to solve for the basis for ker𝐿 is the same as
the augmented matrix used to solve for the basis for Im𝐿.

Page 23 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

(c) dim(ker𝐿) + dim(Im𝐿) = dim 𝑅3

1 + 2 = 3

2. Let 𝐿: 𝑃2 → 𝑃2 be the linear transformation defined by

𝐿(𝑎𝑡 2 + 𝑏𝑡 + 𝑐) = (𝑎 + 𝑐)𝑡 2 + (𝑏 + 𝑐)𝑡.

(a) 𝐿(𝑡 2 – 𝑡 – 1) = −2𝑡. Therefore, 𝑡 2 – 𝑡 – 1 is not in ker𝐿.

(b) 2𝑡 2 – 𝑡 is in range 𝐿 if we can find a vector 𝑎𝑡 2 + 𝑏𝑡 + 𝑐 such that

𝐿(𝑎𝑡 2 + 𝑏𝑡 + 𝑐) = 2𝑡 2 – 𝑡

(𝑎 + 𝑐)𝑡 2 + (𝑏 + 𝑐)𝑡 = 2𝑡 2 – 𝑡

This leads to the system

𝑎+𝑐 =2
𝑏 + 𝑐 = −1
The solution is
𝑎 =2−𝑟
𝑏 = −1 − 𝑟
𝑐 = 𝑟, where 𝑟 ∈ 𝑅.

Since a solution exists, therefore, 2𝑡 2 – 𝑡 is in range 𝐿.

(c) Let 𝑎𝑡 2 + 𝑏𝑡 + 𝑐 ∈ kerL. Then

𝐿(𝑎𝑡 2 + 𝑏𝑡 + 𝑐 ) = 0
(𝑎 + 𝑐 )𝑡 2 + (𝑏 + 𝑐 )𝑡 = 0

This equation leads to the homogeneous system

𝑎 + 𝑐=0
𝑏 + 𝑐=0

Page 24 of 25
 MATH 2250 (Linear Algebra)
Prepared by: Angelita V. Seeping

The solution is
𝑎 = −𝑟
𝑏 = −𝑟
𝑐 = 𝑟, where 𝑟 ∈ 𝑅.
and the ker𝐿 is

ker𝐿 = {(−𝑟𝑡 2 − 𝑟𝑡 + 𝑟): 𝑟 ∈ 𝑅}

= {𝑟(−𝑡 2 − 𝑡 + 1): 𝑟 ∈ 𝑅}
= span {(−𝑡 2 − 𝑡 + 1)}

Since −𝑡 2 − 𝑡 + 1 is linearly independent, then it forms a basis for ker𝐿.

(d) The Im𝐿 = (𝑎 + 𝑐)𝑡 2 + (𝑏 + 𝑐)𝑡

= 𝑎(𝑡 2 ) + 𝑏𝑡 + 𝑐(𝑡 2 + 1)
= span {𝑡 2 , 𝑡, 𝑡 2 + 1}

To determine linear independence of the vectors in the spanning set, we form the
augmented matrix

1 0 1 0
[ ]
0 1 1 0

Thus, {𝑡 2 , 𝑡} forms a basis for Im𝐿.

Page 25 of 25