Lecture Notes of Unit II

Prepared by
Dr. Dilawar Juneed Mir

1. Vector Space

Let (F, +, ·) be a field, whose elements are called scalars. Throughout, by a field
F we shall mean one of the standard fields: the rationals Q, the reals R, or the
complex numbers C.

Let V be a nonempty set whose elements are called vectors. Then V is said
to be a vector space over the field F if it satisfies the following postulates:

A. Under Vector Addition

The addition of vectors (denoted by +) is defined as an internal composition in V

satisfying the following:

1. Closure: For all α, β ∈ V, we have α + β ∈ V.

2. Commutativity: For all α, β ∈ V, α + β = β + α.

3. Associativity: For all α, β, γ ∈ V, α + ( β + γ) = (α + β) + γ.

4. Existence of Identity: There exists 0 ∈ V such that

α+0 = α = 0+α ∀α ∈ V.

The element 0 is called the zero vector in V.

5. Existence of Inverse: For each α ∈ V, there exists (−α) ∈ V such that

α + (−α) = 0 = (−α) + α.

The vector −α is called the additive inverse of α.

1
B. Under Scalar Multiplication

Scalar multiplication is defined as an external composition of V with F satisfying

the following:

1. Closure: For all a ∈ F and α ∈ V, we have aα ∈ V.

2. Distributivity w.r.t. vector addition: For all a ∈ F and α, β ∈ V,

a(α + β) = aα + aβ.

3. Distributivity w.r.t. scalar addition: For all a, b ∈ F and α ∈ V,

( a + b)α = aα + bα.

4. Associativity: For all a, b ∈ F and α ∈ V,

( ab)α = a(bα).

5. Existence of Multiplicative Identity: For all α ∈ V,

1 · α = α, where 1 ∈ F.

Remarks

• V (F) is a real vector space if F = R, the field of real numbers.

• V (F) is a complex vector space if F = C, the field of complex numbers.

• V (F) is a rational vector space if F = Q, the field of rational numbers.

2. Illustrative Examples

Example 1. Prove that R × R is a vector space over R under the operations

( x1 , x2 ) + ( y1 , y2 ) = ( x1 + y1 , x2 + y2 ), and α( x1 , x2 ) = (αx1 , αx2 ).

Solution. The binary operation (+) is commutative and associative, and (0, 0) is
the zero element. The inverse of ( x1 , x2 ) is (− x1 , − x2 ). Hence, (R × R, +) is an
abelian group.
Let
u = ( x1 , x2 ), v = (y1 , y2 ), α, β ∈ R.

2
 Now, we verify the vector space axioms:


α ( u + v ) = α ( x1 , x2 ) + ( y1 , y2 )
= α ( x1 + y1 , x2 + y2 )
= (αx1 + αy1 , αx2 + αy2 )
= (αx1 , αx2 ) + (αy1 , αy2 )
= α ( x1 , x2 ) + α ( y1 , y2 )
= αu + αv.

(α + β)u = (α + β)( x1 , x2 )


= ( α + β ) x1 , ( α + β ) x2
= (αx1 + βx1 , αx2 + βx2 )
= (αx1 , αx2 ) + ( βx1 , βx2 )
= α ( x1 , x2 ) + β ( x1 , x2 )
= αu + βu.


α( βu) = α β( x1 , x2 )
= α( βx1 , βx2 )
= (αβx1 , αβx2 )
= (αβ)( x1 , x2 )
= (αβ)u.

1u = 1( x1 , x2 ) = (1x1 , 1x2 ) = ( x1 , x2 ) = u.
Thus, R × R satisfies all the axioms of a vector space.

∴ R × R is a vector space over R.

Example 2. Show that the set M2 (R) of all 2 × 2 matrices is a vector space over R
under matrix addition and scalar multiplication defined by
" # " #
a b αa αb
α = .
c d αc αd
Solution. Addition (+) is a binary operation in the set M2 (R). Associative and
commutative laws hold for (+) in M2 (R).

3
 The matrix " #
0 0
∈ M2 (R)
0 0
is the identity element for (+).
The inverse of " #
x1 x2
∈ M2 (R)
x3 x4
is " #
− x1 − x2
∈ M2 (R).
− x3 − x4
Hence, M2 (R) is an abelian group under matrix addition.
Let " # " #
x1 x2 y1 y2
M1 = , M2 = ∈ M2 (R), α, β ∈ R.
x3 x4 y3 y4
Then, " #
x1 + y1 x2 + y2
α( M1 + M2 ) = α
x3 + y3 x4 + y4
" #
α ( x1 + y1 ) α ( x2 + y2 )
=
α ( x3 + y3 ) α ( x4 + y4 )
" # " #
αx1 αx2 αy1 αy2
= +
αx3 αx4 αy3 αy4

= αM1 + αM2 .
Also, " #
( α + β ) x1 ( α + β ) x2
(α + β) M1 =
( α + β ) x3 ( α + β ) x4
" #
αx1 + βx1 αx2 + βx2
=
αx3 + βx3 αx4 + βx4

= αM1 + βM1 .

4
 Next, " #
βx1 βx2
α( βM1 ) = α
βx3 βx4
" #
αβx1 αβx2
=
αβx3 αβx4

= (αβ) M1 .
Finally, " # " #
x1 x2 x1 x2
1 · M1 = 1 · = = M1 .
x3 x4 x3 x4
Hence, the set M2 (R) of all 2 × 2 matrices is a vector space over R under
addition and scalar multiplication.

√
Example 3. Prove that V = { a + ib : a, b ∈ R, i = −1} is a vector space
over reals under usual addition and scalar multiplication of complex numbers.

Solution: Let
V = { a + ib : a, b ∈ R}
with usual addition and real scalar multiplication defined by

( a + ib) + (c + id) = ( a + c) + i (b + d), α ( a + ib) = (αa) + i (αb) ( α ∈ R).

We verify the vector space axioms over the field R.

(A1) Closure under addition. If z1 = a + ib and z2 = c + id are in V, then

z1 + z2 = ( a + c ) + i ( b + d ),

and since a + c, b + d ∈ R, we have z1 + z2 ∈ V.

(A2) Commutativity of addition.

z1 + z2 = ( a + c ) + i ( b + d ) = ( c + a ) + i ( d + b ) = z2 + z1 ,

using commutativity in R.

5
(A3) Associativity of addition. For z3 = e + i f ,





( z1 + z2 ) + z3 = ( a + c ) + e + i ( b + d ) + f = a + c + e + i b + ( d + f ) = z1 + ( z2 + z3 ),

by associativity in R.

(A4) Additive identity. Let 0V = 0 + i0. Then for all z = a + ib,

z + 0V = ( a + 0) + i (b + 0) = a + ib = z.

(A5) Additive inverses. For z = a + ib, take −z = (− a) + i (−b) ∈ V. Then

z + (−z) = ( a − a) + i (b − b) = 0 + i0 = 0V .

(B1) Closure under scalar multiplication. If α ∈ R and z = a + ib ∈ V, then

αz = (αa) + i (αb) ∈ V,

since αa, αb ∈ R.

(B2) Distributivity over vector addition. For α ∈ R,

α(z1 + z2 ) = α ( a + c) + i (b + d) = (αa + αc) + i (αb + αd) = αz1 + αz2 .

(B3) Distributivity over scalar addition. For α, β ∈ R,

(α + β)z = (α + β) a + i (α + β)b = (αa + βa) + i (αb + βb) = αz + βz.

(B4) Compatibility (associativity) of scalar multiplication. For α, β ∈ R,

α( βz) = α ( βa) + i ( βb) = (αβa) + i (αβb) = (αβ)z.

(B5) Multiplicative identity. With 1 ∈ R,

1 · z = (1 · a) + i (1 · b) = a + ib = z.

All axioms for a vector space over R are satisfied. Hence V is a vector space
over R under the usual addition and (real) scalar multiplication of complex num-
bers.

√
Example 4. Prove that V = { a + b 2 : a, b ∈ Q} is a vector space over Q

6
under usual addition and scalar multiplication.

Solution. Let
√
V = { a + b 2 : a, b ∈ Q},
with addition and scalar multiplication given by the usual operations inherited
from R. We check the vector-space axioms over the field Q.

For x, y ∈ V write
√ √
x = a1 + b1 2, y = a2 + b2 2 ( a i , bi ∈ Q ) .
√
For α ∈ Q write αx = αa1 + αb1 2.

A1. Closure under addition.

√ √ √
x + y = ( a1 + b1 2) + ( a2 + b2 2) = ( a1 + a2 ) + (b1 + b2 ) 2.

Since a1 + a2 , b1 + b2 ∈ Q, we have x + y ∈ V.

A2. Commutativity of addition. Addition in R is commutative, so

√ √
x + y = ( a1 + a2 ) + (b1 + b2 ) 2 = ( a2 + a1 ) + (b2 + b1 ) 2 = y + x.

√
A3. Associativity of addition. For z = a3 + b3 2 ∈ V,
√
√
( x + y) + z = ( a1 + a2 ) + (b1 + b2 ) 2 + ( a3 + b3 2)
√
= ( a1 + a2 + a3 ) + (b1 + b2 + b3 ) 2 = x + (y + z),
using associativity in R.
√
A4. Additive identity. The element 0V = 0 + 0 2 = 0 ∈ V satisfies
√ √
x + 0 = ( a1 + 0) + (b1 + 0) 2 = a1 + b1 2 = x.

√
A5. Additive inverse. For x = a1 + b1 2 ∈ V, its additive inverse in R is
√
− x = (− a1 ) + (−b1 ) 2 ∈ V,

since − a1 , −b1 ∈ Q, and x + (− x ) = 0.

7
 √
B1. Closure under scalar multiplication. For α ∈ Q and x = a1 + b1 2 ∈ V,
√
αx = αa1 + (αb1 ) 2,

and αa1 , αb1 ∈ Q, so αx ∈ V.

B2. Distributivity of scalar over vector addition. For α ∈ Q,

√
√
α( x + y) = α ( a1 + a2 ) + (b1 + b2 ) 2 = α( a1 + a2 ) + α(b1 + b2 ) 2
√
= (αa1 + αa2 ) + (αb1 + αb2 ) 2 = αx + αy.

B3. Distributivity of field addition over scalar multiplication. For α, β ∈ Q,

√ √
(α + β) x = (α + β) a1 + (α + β)b1 2 = (αa1 + βa1 ) + (αb1 + βb1 ) 2

= αx + βx.

B4. Compatibility of scalar multiplication (associativity). For α, β ∈ Q,

√ √
α( βx ) = α( βa1 + ( βb1 ) 2) = (αβ) a1 + (αβ)b1 2 = (αβ) x.

B5. Identity scalar. Let 1 ∈ Q be the multiplicative identity of the field. Then
√ √
1 · x = 1 · a1 + 1 · b1 2 = a1 + b1 2 = x.

All vector-space axioms over Q are satisfied, hence V is a vector space over Q.

Example 5. Let V be a set of all real valued continuous functions defined in

a closed interval [ a, b ]. Then show that V is a vector space over R with addition
and scalar multiplication defined as ( f + g)( x ) = f ( x ) + g( x ), for all f , g ∈ V
and (c f )( x ) = c f ( x ), for all c ∈ R, f ∈ V.

Solution. Let V = C ([ a, b]) is a vector space over R, that is,

V = { f : [ a, b] → R | f is continuous on [ a, b]},

with operations defined pointwise by

( f + g)( x ) = f ( x ) + g( x ), (c f )( x ) = c f ( x )

8
for all f , g ∈ V, all scalars c ∈ R and all x ∈ [ a, b]. We check the vector-space
axioms (all equalities hold for every x ∈ [ a, b]).
(A1) Closure and well-definedness.
• If f , g ∈ V then f and g are continuous on [ a, b]. The sum of continuous
functions is continuous, so f + g is continuous on [ a, b], hence f + g ∈ V.

• If c ∈ R and f ∈ V, then c f is continuous on [ a, b] (constant multiple of a

continuous function), so c f ∈ V.
Thus the operations are closed in V.
(A2) Commutativity of addition. For all f , g ∈ V and all x ∈ [ a, b],

( f + g)( x ) = f ( x ) + g( x ) = g( x ) + f ( x ) = ( g + f )( x ).

So f + g = g + f .
(A3) Associativity of addition. For all f , g, h ∈ V and all x ∈ [ a, b],


( f + g) + h ( x ) = ( f + g)( x ) + h( x ) = ( f ( x ) + g( x )) + h( x )

and


f + ( g + h) ( x ) = f ( x ) + ( g + h)( x ) = f ( x ) + ( g( x ) + h( x )).
By associativity of real addition these two are equal for every x, hence ( f + g) +
h = f + ( g + h ).
(A4) Additive identity. Define the zero function 0 ∈ V by 0( x ) = 0 for all
x ∈ [ a, b]. It is continuous, so 0 ∈ V. For any f ∈ V and all x ∈ [ a, b],

( f + 0)( x ) = f ( x ) + 0 = f ( x ),

hence f + 0 = f .
(A5) Additive inverse. For each f ∈ V define (− f ) ∈ V by (− f )( x ) = − f ( x ).
It is continuous, so in V. For all x ∈ [ a, b],

( f + (− f ))( x ) = f ( x ) + (− f ( x )) = 0,

so f + (− f ) = 0.
(B1) Compatibility of scalar multiplication with field multiplication. For all
scalars c, d ∈ R and f ∈ V, for every x ∈ [ a, b],




c(d f ) ( x ) = c (d f )( x ) = c(d f ( x )) = (cd) f ( x ) = (cd) f ( x ).

9
So c(d f ) = (cd) f .
(B2) Identity element of scalar multiplication. For the scalar 1 ∈ R and any
f ∈ V, for every x ∈ [ a, b],

(1 · f )( x ) = 1 · f ( x ) = f ( x ),

hence 1 · f = f .
(B3) Distributivity of scalar multiplication over vector addition. For all
c ∈ R and f , g ∈ V, for each x ∈ [ a, b],




c( f + g) ( x ) = c ( f + g)( x ) = c( f ( x ) + g( x )) = c f ( x ) + cg( x ) = c f + cg ( x ).

Thus c( f + g) = c f + cg.
(B4) Distributivity of scalar addition over a vector. For all c, d ∈ R and f ∈ V,
for each x ∈ [ a, b],



( c + d ) f ( x ) = ( c + d ) f ( x ) = c f ( x ) + d f ( x ) = c f + d f ( x ).

So (c + d) f = c f + d f .
All vector-space axioms over the field R are satisfied. Therefore V = C ([ a, b])
with the given pointwise addition and scalar multiplication is a vector space over
R.

3. Linear Combination in a Vector Space

Definition: Let V be a vector space over a field F and let v1 , v2 , . . . , vn ∈ V. A

vector v ∈ V is called a linear combination of v1 , v2 , . . . , vn if there exist scalars
a1 , a2 , . . . , an ∈ F such that

v = a1 v1 + a2 v2 + · · · + a n v n .

Examples:

1. In R3 , the vector
(3, −1, 4)
is a linear combination of

(1, 0, 0), (0, 1, 0), (0, 0, 1)

10
 since
(3, −1, 4) = 3(1, 0, 0) + (−1)(0, 1, 0) + 4(0, 0, 1).

2. In the vector space of all real-valued functions C (R), the function

f ( x ) = 2e x − sin x

is a linear combination of the functions e x and sin x, because

f ( x ) = 2 · e x + (−1) · sin x.

4. Linear Span

Let V be a vector space over the field F and S be any non-empty subset of V.
Then the linear span of S is defined as the set of all linear combinations of finite
sets of elements of S. It is denoted by L(S).

L(S) = {λ1 u1 + λ2 u2 + · · · + λn un | ui ∈ S, λi ∈ F, i = 1, 2, . . . , n}

The set L(S) is said to be generated or spanned by the set S, and S is said to
be the set of generators of L(S).
Example: The set (" # )
λ1 0
L(S) = λ1 , λ2 ∈ F
0 λ2
of all 2 × 2 diagonal matrices is the linear span of
" # " #
1 0 0 0
and .
0 0 0 1

5. Linear Dependence and Independence

Let V be a vector space over F. A subset S of V is said to be linearly dependent

if there exist distinct vectors a1 , a2 , . . . , an ∈ S and scalars λ1 , λ2 , . . . , λn ∈ F, not
all zero, such that
λ1 a1 + λ2 a2 + · · · + λn an = 0.
Similarly, the vectors of S are said to be linearly independent if the relation

λ1 a1 + λ2 a2 + · · · + λ n a n = 0

11
implies that λ1 = λ2 = · · · = λn = 0.

Consequences
1. Any superset of a linearly dependent set is linearly dependent.

2. Any subset of a linearly independent set is linearly independent.

3. Any set which contains the 0 vector is linearly dependent.

4. If a vector can be expressed as a scalar multiple of another vector then the

vectors are linearly dependent.

Notation: Throughout this unit, by Vn (F) we mean the vector space Rn (F), where
F is a field either Q or R.

Example 1. Prove that, in V3 (R), the vectors (1, 2, 1), (2, 1, 0) and (1, −1, 2) are
linearly independent.
Solution. Let

λ1 (1, 2, 1) + λ2 (2, 1, 0) + λ3 (1, −1, 2) = (0, 0, 0).

This gives the equations:

λ1 + 2λ2 + λ3 = 0 (1)
2λ1 + λ2 − λ3 = 0 (2)
λ1 + 2λ3 = 0. (3)

Solving (1), (2), and (3), we get λ1 = λ2 = λ3 = 0. Hence the given set of
vectors is linearly independent.

Example 2. In V3 (R), prove that the vectors (1, 4, −2), (−2, 1, 3) and (−4, 11, 5)
are linearly dependent.
Solution. Let

λ1 (1, 4, −2) + λ2 (−2, 1, 3) + λ3 (−4, 11, 5) = (0, 0, 0).

12
 Therefore,

λ1 − 2λ2 − 4λ3 = 0 (1)

4λ1 + λ2 + 11λ3 = 0 (2)
−2λ1 + 3λ2 + 5λ3 = 0. (3)

From (1) and (2), we obtain

λ1 λ λ
= 2 = 3.
−18 −27 9
Hence,
λ1 λ λ
= 2 = 3 =k (say).
−2 −3 1
Therefore,
λ1 = −2k, λ2 = −3k, λ3 = k.
Taking k = 1, we get λ1 = −2, λ2 = −3, λ3 = 1, which also satisfy (3). Since
λ1 , λ2 , λ3 are non-zero scalars, the given set of vectors is linearly dependent.

6. Bases and Dimension

Let V be a vector space. A basis for V is a linearly independent set of vectors in

V which spans V. The number of elements (vectors) present in the basis of V is
known as the dimension of V, denoted by dim(V ).
The space V is finite-dimensional if it has a finite basis. The vector space V is
called n-dimensional if its basis contains n elements (vectors).
As an example, if

S = {(1, 0, 0), (0, 1, 0), (0, 0, 1)}

is a basis of the vector space R3 , then the dimension of R3 is 3, since its basis
contains three linearly independent vectors.
Notes.

1. A vector space which is not finitely generated is called an infinite-dimensional

space.

2. The vector space {0} is of zero dimension.

3. A vector space may have more than one basis, i.e., the basis of a vector space
is not unique.

13
 4. The set {e1 , e2 , . . . , en } of unit vectors is a basis for Rn , which is known as
the standard basis.

Example. Show that the vectors (3, 0, 2), (7, 0, 9) and (4, 1, 2) form a basis for R3 .

Solution. It is easy to prove that the given vectors are linearly independent.
Now, we have to show that they span R3 . To show this, we need to prove that for
any vector ( a1 , a2 , a3 ) ∈ R3 , it is possible to find scalars λ1 , λ2 , λ3 such that

( a1 , a2 , a3 ) = λ1 (3, 0, 2) + λ2 (7, 0, 9) + λ3 (4, 1, 2).

This expands to

( a1 , a2 , a3 ) = (3λ1 + 7λ2 + 4λ3 , λ2 , 2λ1 + 9λ2 + 2λ3 ).

Equating corresponding components gives:

3λ1 + 7λ2 + 4λ3 = a1 , λ2 = a2 , 2λ1 + 9λ2 + 2λ3 = a3 .

From these equations, real values of λ1 , λ2 , λ3 can be found. Hence, the given
vectors form a basis for R3 .

Example: Prove that the vectors (1, 1, 1), (1, 1, 0) and (1, 0, 0) form a basis of
R3 . Also prove that the vector (1, 3, 1) can replace any one of the three vectors to
form a new basis.
Solution. Let A be the matrix formed by these vectors:
 
1 1 1
A = 1 1 0 .
 

1 0 0

Its determinant is
1 1 1
| A| = 1 1 0 = −1 6= 0.
1 0 0
Hence, rank( A) = 3, so the vectors are linearly independent.
Now, for ( a1 , a2 , a3 ) ∈ R3 ,

( a1 , a2 , a3 ) = λ1 (1, 1, 1) + λ2 (1, 1, 0) + λ3 (1, 0, 0).

14
Expanding:
( a1 , a2 , a3 ) = ( λ1 + λ2 + λ3 , λ1 + λ2 , λ1 ).
Equating components:

λ1 + λ2 + λ3 = a1 , λ1 + λ2 = a2 , λ1 = a3 .

Thus, the scalars exist and the vectors span R3 . Hence, the given vectors form
a basis.
Now, check if (1, 3, 1) can replace one of them. Suppose:

(1, 3, 1) = µ1 (1, 1, 1) + µ2 (1, 1, 0) + µ3 (1, 0, 0).

This gives:
µ1 + µ2 + µ3 = 1, µ1 + µ2 = 3, µ1 = 1.
Solving:
µ1 = 1, µ2 = 2, µ3 = −2.
Since not all coefficients are zero, (1, 3, 1) can replace any vector of the basis to
form a new basis.

Example. Find the dimension of the subspace spanned by the vectors (1, 0, 2), (2, 0, 1), (1, 0, 1)
in R3 .
Solution. Let A be the matrix:
 
1 0 2
A = 2 0 1 .
 

1 0 1

1 0 2
| A| = 2 0 1 = 0.
1 0 1
Thus, A is singular and the given vectors are linearly dependent. So, they do
not form a basis.
Now, check the set {(1, 0, 2), (2, 0, 1)}:

λ1 (1, 0, 2) + λ2 (2, 0, 1) = (0, 0, 0).

This gives:
(λ1 + 2λ2 , 0, 2λ1 + λ2 ) = (0, 0, 0).

15
Equating components:

λ1 + 2λ2 = 0, 2λ1 + λ2 = 0.

The only solution is λ1 = λ2 = 0, hence the vectors are linearly independent.

Therefore, {(1, 0, 2), (2, 0, 1)} forms a basis of the subspace.

dim(subspace) = 2.

7. TEST YOUR KNOWLEDGE

1. Express (3, −5) as a linear combination of a = (1, 2), b = (−3, 1) in V2 (R).

2. Express the vector (1, 7, −4) as a linear combination of the vectors (1, −3, 2), (2, −1, 1)
in V3 (R).

3. Express (1, 1, 2) as a linear combination of (0, 2, 1), (2, 2, 4) in V3 ( R).

4. Show that the vector (3, −4, 2) cannot be expressed as a linear combination
of the vectors, (2, 1, 5) and (−1, 3, 2).

5. Show that the vector (0, 2, 1) cannot be expressed as a linear combination of

the vectors (1, 1, 2) and (2, 2, 4).
" #
2 0
6. Can the matrix M = be expressed as the linear combination of
4 −5
the matrices

" # " # " #

0 −3 0 0 2 3
P= ,Q = and R = ?
2 0 2 1 0 5

7. Show that the following vectors are linearly dependent.

(i) (4, 3), (8, 6)
(ii) 23 , 13 , 1 , (9, 2, 6)
(iii) (1, 1, 2), (2, 1, 0), (0, 4, 1)
(iv) (1, 2, 3), (4, 1, 5), (−4, 6, 2).
( ! ! !)
2 0 0 0 0 −1
8. Show that the set of matrices , , in M2 (R)
0 0 2 0 3 0
is linearly independent.

16
 9. Are the following sets of vectors linearly independent or dependent?
(i) {(2, −4), (1, 9), (3, 5)}
(ii) {(−1, 5, 0), (16, 8, −3), (−64, 56, 9)}
{(1, −1, 1),(1,1, −1), (−
(iii) n  1, 1, 1)(0, 1,
o0)}
1 1 1 1 1 1
(iv) 4 , 0, − 4 , 0, 2 , − 2 , 3 , − 3 , 0
(v) {(1, 9, 9, 8), (2, 0, 0, 3), (2, 0, 0, 8)}.

10. Show that the vectors (1, 2, 1), (2, 1, 0), (1, −1, 2) form a basis of the vector
space E3 over the field of real numbers.

11. Is the vector (2, −5, 3) in the subspace of R3 spanned by the vectors (1, −3, 2), (2, −4, −1)
and (1, −5, 7) ?

12. Show that the following sets of vectors form a basis of the vector space R3 .
(i) {(2, −1, 0), (3, 5, 1), (1, 1, 2)}
(ii) {(1, −3, 2), (2, 4, 1), (1, 1, 1)}
(iii) {(1, 2, −3), (2, 5, 1), (−1, 1, 4)}
(iv) {(2, −3, 1), (0, 1, 2), (1, 1, 2)}
(v) {(1, 2, 2), (1, −1, 2), (1, 0, 1)}.

13. Shor that the following sets of vectors do not form a basis of the vector space
R3 .
(i) {(3, 2, 1), (3, 1, 5), (3, 4, −7)}
(ii) {(1, 2, 3), (2, 3, 4), (3, 4, 5)}
(iii) {(1, 2, 1), (1, 3, 5), (−1, 0, 1), (1, −1, 2)}
(iv) {(1, 2, −3), (2, −3, 1)(−3, 1, 1)}.

14. Find the dimension of the subspace spanned by the following vectors in
V3 (R) :
(i) (1, 1, 1), (−1, −1, −1)
(ii) (1, 2, −3), (0, 0, 1), (−1, 2, 1).

Answers

1 (3, −5) = − 12
7 (1, 2) −
11
7 (−3, 1)

2 (1, 7, −4) = −3(1, −3, 2) + 2(2, −1, 1)

3 (1, 1, 2) = 0(0, 2, 1) + 21 (2, 2, 4)

6 No.

17
 9 (i) linearly dependent
(ii) linearly dependent
(iii) linearly independent
(iv) linearly dependent
(v) linearly independent

11 No

14 (i) 1
(ii) 3

18
 [Chapter 2]

Linear Transformation

Let V and W be vector spaces over the same field F. A mapping

T : V −→ W

that assigns to each vector X ∈ V a unique vector T ( X ) ∈ W is called a linear

transformation (or linear map) from V to W if the following conditions hold:

1. Additivity: T ( X1 + X2 ) = T ( X1 ) + T ( X2 ), ∀ X1 , X2 ∈ V.

2. Homogeneity: T ( aX ) = a T ( X ), ∀ a ∈ F, X ∈ V.

Conditions (1) and (2) can be combined into the single requirement

T ( a 1 X1 + a 2 X2 ) = a 1 T ( X1 ) + a 2 T ( X2 ) , ∀ a1 , a2 ∈ F, X1 , X2 ∈ V.

In other words, a mapping T : V → W is linear precisely when it preserves

the two fundamental operations of a vector space: vector addition and scalar
multiplication.

Determining Whether a Mapping is a Linear Transformation

To check whether a transformation T : V → W is linear, verify the following two
properties:

1. Additivity:
T ( X1 + X2 ) = T ( X1 ) + T ( X2 )

2. Homogeneity (scalar multiplication):

T ( aX ) = a T ( X )

Equivalently, these two conditions can be combined into a single statement:

T ( a 1 X1 + a 2 X2 ) = a 1 T ( X1 ) + a 2 T ( X2 )

for all scalars a1 , a2 ∈ F and all vectors X1 , X2 ∈ V.

19
 Illustrative Examples

Example 1. Let T : V → W be defined by T ( X ) = 0 for all X ∈ V. Show that T is

a linear transformation.
Solution. Let X1 , X2 ∈ V and let a1 , a2 ∈ F. Since V is a vector space,
a1 X1 + a2 X2 ∈ V. Then

T ( a 1 X1 + a 2 X2 ) = 0 (by definition of T)
= 0+0
= a1 0 + a2 0
= a 1 T ( X1 ) + a 2 T ( X2 ) .

Since the condition for linearity is satisfied, we conclude that T is a linear trans-
formation.
Remark 1. Sometimes we write this map succinctly as

T : V −→ W, X 7→ 0 (∀ X ∈ V ).

Remark 2. If T is any linear transformation from V to W, then

T (0) = T (0 + 0)
= T (0) + T (0)
=⇒ T (0) = 0,

where the 0 on the left is the additive identity of V and the 0 on the right is the
additive identity of W.
Example 2. Let T : V → W. If W = V, define T : V → V by T ( X ) = X for all
X ∈ V. This transformation is called the identity transformation. Show that the
identity transformation is linear.
Solution. Consider T : V → V defined by T ( X ) = X.
Let X1 , X2 ∈ V and a ∈ F. Then

T ( X1 + X2 ) = X1 + X2 (since T ( X ) = X)
= T ( X1 ) + T ( X2 ) ,
T ( aX ) = aX (since T ( X ) = X)
= a T ( X ).

Hence T preserves both addition and scalar multiplication, so T is a linear trans-

20
formation.
Alternative Method. Let X1 , X2 ∈ V and a1 , a2 ∈ F. Then

T ( a 1 X1 + a 2 X2 ) = a 1 X1 + a 2 X2 (since T ( X ) = X)
= a 1 T ( X1 ) + a 2 T ( X2 ) .

Thus the mapping T : V → V defined by T ( X ) = X is indeed a linear transforma-

tion.

Examples on Linear Transformations

Example 3. Determine whether the mapping

T : R2 → R3 , T ( x, y) = ( x, y, 0)

is linear.
Solution. Let X1 = ( x1 , y1 ) and X2 = ( x2 , y2 ) be elements of R2 . Then


T ( X1 + X2 ) = T ( x 1 , y 1 ) + ( x 2 , y 2 )
= T ( x1 + x2 , y1 + y2 )
= ( x1 + x2 , y1 + y2 , 0)
= ( x1 , y1 , 0) + ( x2 , y2 , 0)
= T ( X1 ) + T ( X2 ) . (1)

For any scalar a ∈ R,

T ( aX1 ) = T a( x1 , y1 )
= T ( ax1 , ay1 )
= ( ax1 , ay1 , 0)
= a ( x1 , y1 , 0)
= a T ( X1 ) . (2)

From (1) and (2), T is a linear transformation.

Example 4. Prove that the mapping

T : R3 → R2 , T ( x, y, z) = (y, z)

is linear.

21
 Solution. Let α = ( x1 , y1 , z1 ) and β = ( x2 , y2 , z2 ) in R3 , and a ∈ R. Then


T ( α + β ) = T ( x1 + x2 , y1 + y2 , z1 + z2 )
= ( y1 + y2 , z1 + z2 )
= ( y1 , z1 ) + ( y2 , z2 )
= T ( α ) + T ( β ). (1)

Also,

T ( aα) = T ( ax1 , ay1 , az1 )

= ( ay1 , az1 )
= a ( y1 , z1 )
= a T ( α ). (2)

Hence T is a linear transformation (a homomorphism). Alternatively, for a, b ∈ R,

T ( aα + bβ) = a T (α) + b T ( β).

Example 5. Show that T : R2 → R2 defined by

T ( a, b) = ( a2 , b)

is not linear.
Solution. Let X1 = ( a1 , b1 ) and X2 = ( a2 , b2 ) in R2 . Then

T ( X1 + X2 ) = T ( a1 + a2 , b1 + b2 )
= ( a1 + a2 )2 , b1 + b2

= a21 + a22 + 2a1 a2 , b1 + b2 .

(1)

Whereas

T ( X1 ) + T ( X2 ) = ( a21 , b1 ) + ( a22 , b2 )
= a21 + a22 , b1 + b2 .


(2)

Since the extra term 2a1 a2 appears in (1) but not in (2),

T ( X1 + X2 ) 6 = T ( X1 ) + T ( X2 ) ,

so T is not linear (except in the trivial case a1 a2 = 0).

22
Thus we have illustrated:

• T ( x, y) = ( x, y, 0) is linear,

• T ( x, y, z) = (y, z) is linear,

• T ( a, b) = ( a2 , b) is not linear.

Test Your Knowledge

1. Show that the mapping

f : V2 (R) −→ V3 (R), f ( a, b) = ( a + b, a − b, b)

is a linear transformation.

2. Show that the mapping

T : V3 (R) −→ V2 (R), T ( x, y, z) = ( x + z, y + z)

is linear.

3. Show that the mapping

f : V2 (R) −→ V3 (R), f ( a, b) = ( a, b, 0)

is a linear transformation from V2 (R) onto V3 (R).

4. Show that the mapping

f : V3 (R) −→ V2 (R),


f ( a1 , a2 , a3 ) = 3a1 − 2a2 + a3 , a1 − 3a2 − 2a3

is a linear transformation.

5. Show that the mapping

T : V2 (R) −→ V2 (R), T ( x, y) = x3 + y3

is not linear.

6. Show that the mapping

T : R3 −→ R2 , T ( x, y, z) = ( x + 3, y + z)

is not linear.

23
 7. Show that the mapping

T : R2 −→ R, T ( x, y) = | x − y|

is not linear.

8. Let {α1 , α2 , α3 } be a basis of V. Define T : V → V by

T ( α1 ) = α2 , T ( α2 ) = α3 , T ( α3 ) = α1 .

Show that T 3 (α) = α, i.e. T 3 = I.

9. Let {α1 , α2 , α3 } be a basis of V. Define T : V → V by

T ( α1 ) = α2 , T ( α2 ) = α3 , T (α3 ) = 0.

Show that T 3 (α) = 0, i.e. T 3 = 0.

10. Let T : F3 → F3 be defined by

T ( a, b, c) = a + b + 5c, a + 2c, 2b + 6c ,

where a, b, c ∈ F. Show that T is a linear transformation.

11. Let V be the vector space of all complex numbers over the field R, i.e.

V = { a + ib | a, b ∈ R}.

Define T : V → V by

(a) T ( a + ib) = a − ib,

(b) T ( x + iy) = ( x, y).

Kernel of a Linear Transformation or Homomorphism

Let T : VF → WF be a linear transformation between vector spaces over a field F.

The kernel of T is the set of all elements of V that are mapped to the zero vector of
W. It is denoted by ker T (or sometimes by K) and is defined as

ker T = { α ∈ V : T (α) = 0W },

where 0W is the zero vector in W.

24
 The dimension of ker T is called the nullity of T and is denoted by ν( T ):

ν( T ) = dim(ker T ).

Remark. If we view T as a vector space homomorphism from V into W, the

kernel is also referred to as the null space of T, written

N ( T ) = { α ∈ V : T ( α ) = 0W } .

Its dimension is likewise the nullity of T:

dim N ( T ) = ν( T ).

Image of a Linear Transformation

Let T : V → W be a linear transformation of a vector space V over a field F into a

vector space W over the same field. The image (or range) of T is the set

Im T = T (V ) = { T ( x ) : x ∈ V } ⊆ W.

For x ∈ V, the vector T ( x ) is called the image of x under T. If y ∈ W, any x ∈ V

satisfying T ( x ) = y is called an inverse image (or preimage) of y. The set of all
such x is denoted by

T −1 ( y ) = { x ∈ V : T ( x ) = y } .

In general, T −1 (y) need not be a single element because more than one vector of
V may map to the same y.
If V1 ⊆ V is any subset, the image of V1 is

T (V1 ) = { T (α) : α ∈ V1 }.

When T (V ) = W, we say that T is onto (or surjective).

To show that T (V ) is a subspace of W, let β 1 , β 2 ∈ T (V ) and a1 , a2 ∈ F. Then
there exist α1 , α2 ∈ V such that T (α1 ) = β 1 and T (α2 ) = β 2 . Because T is linear,

a1 β 1 + a2 β 2 = a1 T ( α1 ) + a2 T ( α2 )
= T ( a1 α1 ) + T ( a2 α2 )
= T ( a1 α1 + a2 α2 )
= T ( α ),

25
where α = a1 α1 + a2 α2 ∈ V.
Thus a1 β 1 + a2 β 2 ∈ T (V ), proving that T (V ) is indeed a subspace of W.

Range of a Linear Transformation

Let T : V ( F ) → W ( F ) be a linear transformation between vector spaces over the
field F. The range of T is the set of all vectors Y ∈ W for which there exists some
X ∈ V such that
T ( X ) = Y.
It is denoted by R( T ) or Im T. In symbols,

R( T ) = { Y ∈ W | ∃ X ∈ V with T ( X ) = Y }.

If V ( F ) is finite dimensional, the dimension of R( T ) is called the rank of T,

denoted by ρ( T ). Thus,

ρ( T ) = rank T = dim R( T ).

Sylvester’s Law of Nullity (Rank–Nullity Theorem)

Statement: Let V and W be vector spaces over a field F, and let T : V → W be a

linear transformation. If V is finite dimensional, then

rank( T ) + nullity( T ) = dim V.

Equivalently,
ρ( T ) + ν( T ) = dim V,
dim V = dim R( T ) + dim N ( T ),
dim V = n (finite).
Here:

• R( T ) denotes the range (image) of T,

• N ( T ) denotes the null space (kernel) of T,

• ρ( T ) is the rank of T, and

• ν( T ) is the nullity of T.

26
Example 1 Show that the transformation

T : V2 (R) −→ V3 (R), T ( a, b) = ( a + b, a − b, b), ∀ a, b ∈ R,

is linear. Find its null space, nullity, range, and rank.

Solution The transformation is given by

T ( a, b) = ( a + b, a − b, b), ∀ a, b ∈ R.

Let α1 = ( a1 , b1 ) and α2 = ( a2 , b2 ) be any two vectors in V2 (R).

(i) Linearity of T:



T c1 α1 + c2 α2 = T c1 ( a1 , b1 ) + c2 ( a2 , b2 )


= T c1 a1 + c2 a2 , c1 b1 + c2 b2
 
= c1 ( a1 + b1 ) + c2 ( a2 + b2 ), c1 ( a1 − b1 ) + c2 ( a2 − b2 ), c1 b1 + c2 b2
= c1 T ( a1 , b1 ) + c2 T ( a2 , b2 )
= c1 T ( α1 ) + c2 T ( α2 ).

Hence, T is a linear transformation.

(ii) Null space (kernel) of T:

ker T = ( a, b) ∈ V2 (R) : T ( a, b) = (0, 0, 0) .



Solving
a + b = 0, a − b = 0, b = 0,
gives a = 0, b = 0. Thus
ker T = {(0, 0)}.
The null space is therefore the zero subspace of V2 (R).
(iii) Nullity:
nullity( T ) = dim ker T = 0.

(iv) Range of T: Let the standard basis of V2 (R) be {(1, 0), (0, 1)}. Then

T (1, 0) = (1, 1, 0), T (0, 1) = (1, −1, 1).

Hence
Range( T ) = span{(1, 1, 0), (1, −1, 1)}.

27
To check linear independence, suppose

a(1, 1, 0) + b(1, −1, 1) = (0, 0, 0).

This gives the system

a + b = 0, a − b = 0, b = 0,

which implies a = b = 0. Thus the vectors are linearly independent.

Therefore, {(1, 1, 0), (1, −1, 1)} is a basis for Range( T ).
(v) Rank:
rank( T ) = dim Range( T ) = 2.
Example 2 Let T : V2 → V3 be defined by


T ( x1 , x2 ) = x1 , x1 + x2 , x2 .

Find the range, kernel, rank and nullity of T. Decide whether T is one-one and/or
onto.
Solution. The standard basis of V2 is S = {(1, 0), (0, 1)}. Therefore

Range( T ) = { T (1, 0), T (0, 1)}

= {(1, 1, 0), (0, 1, 1)}.

These two vectors are linearly independent, hence form a basis of Range( T ).

rank T = dim Range( T ) = 2.

The kernel is

ker T = ( x1 , x2 ) ∈ V2 : T ( x1 , x2 ) = (0, 0, 0)
= {(0, 0)}.

Thus nullity T = dim ker T = 0.

Since ker T = {0}, T is 1 − 1 . Because Range( T ) 6= V3 , T is not onto .

Example 3 Let T : R3 → R2 be defined by

T ( a, b, c) = ( a + b, b + c).

28
 1. Linearity: For X1 = ( a1 , a2 , a3 ), X2 = (b1 , b2 , b3 ) and c ∈ R,

T ( X1 + X2 ) = T ( X1 ) + T ( X2 ) , T (cX1 ) = c T ( X1 ),

hence T is linear.

2. Kernel and Nullity:


 x + x = 0,
1 2
T ( x1 , x2 , x3 ) = (0, 0) =⇒
 x2 + x3 = 0.

Let x2 = r ∈ R. Then x1 = −r, x3 = −r, so

ker T = {r (−1, 1, −1) : r ∈ R}, nullity T = 1.

3. Range and Rank: T (1, 0, 0) = (1, 0) and T (0, 0, 1) = (0, 1), which are linearly
independent. Hence

Range( T ) = R2 , rank T = 2.

4. Injectivity / Surjectivity: Since ker T 6= {0}, T is not one-one. Because

Range( T ) = R2 , T is onto (an epimorphism).

Example 4 Define T : R4 → R3 by

T ( x, y, s, t) = ( x − y + s + t, x + 2s − t, x + y + 3s − 3t).

Image.
T (1, 0, 0, 0) = (1, 1, 1),
T (0, 1, 0, 0) = (−1, 0, 1),
T (0, 0, 1, 0) = (1, 2, 3),
T (0, 0, 0, 1) = (1, −1, −3).
Row-reducing the matrix with these as rows gives rank 2, so

Range( T ) = span{(1, 1, 1), (0, 1, 2)}, rank T = 2.

29
Kernel. Solve 
 x − y + s + t = 0,


x + 2s − t = 0,


x + y + 3s − 3t = 0.


Let s, t be free. Then y = −s + 2t and x = −s − t + y = 2t − s − s − t = t − 2s

(simplify as you wish). Hence dim ker T = 2 with basis, for example,

{(2, 1, −1, 0), (1, 2, 0, 1)}.

Rank-nullity holds: 2 + 2 = 4 = dim R4 .

Test Your Knowledge

1. Let T : V → V be defined by


T ( x1 , x2 , x3 ) = 2x1 + x2 + 3x3 , 3x1 − x2 + x3 , −4x1 + 3x2 + x3 .

Find a basis for T (V ) and a basis for K ( T ).

2. Let T : R3 → R3 be the linear transformation

T ( x, y, z) = x + 2y − z, y + z, x + y − 2z .

Find a basis and the dimension of the image (rank) of T and of the null
space of T.

3. Let T : R3 → R3 be the linear transformation

T ( x, y, z) = x + 2y − z, y + z, x + y − 2z .

Find a basis and the dimension of the image and the rank of T.

4. Let T : V4 → V3 be the linear transformation defined by

T (e1 ) = (1, 1, 1), T (e2 ) = (1, −1, 1),

T (e3 ) = (1, 0, 0), T (e4 ) = (1, 0, 1).

30
 Verify that
ρ( T ) + v( T ) = dim V4 = 4,
where ρ( T ) and v( T ) denote the rank and nullity of T, respectively.

5. Define the rank and nullity of a linear transformation. Let T : R3 → R3 be

given by


T (α1 , α2 , α3 ) = 2α1 + 3α2 − 5α3 , α1 + α2 + α3 , α1 − α2 − α3 .

Verify that T is a linear transformation and then find its rank and nullity.

Illustrative Examples

Example Let T1 and T2 be linear operators on R2 defined by

T1 ( x, y) = (0, x ), T2 ( x, y) = (y, 0).

Find T12 , T22 , T1 T2 and ( T1 T2 )2 .

Solution.
T12 ( x, y) = T1 T1 ( x, y) = T1 (0, x ) = (0, 0).

Similarly,
T22 ( x, y) = T2 T2 ( x, y) = T2 (y, 0) = (0, 0).

Next,


T1 T2 ( x, y) = T1 T2 ( x, y) = T1 (y, 0) = (0, y).
Hence
( T1 T2 )2 ( x, y) = T1 T2 T1 T2 ( x, y) = T1 T2 (0, y) = (0, y).

Example Let T1 , T2 : R2 → R2 be given by

T1 ( x1 , x2 ) = ( x2 , x1 ), T2 ( x1 , x2 ) = ( x1 , 0).

Show that T1 T2 6= T2 T1 .

31
Solution.


T1 T2 ( x1 , x2 ) = T1 T2 ( x1 , x2 ) = T1 ( x1 , 0) = (0, x1 ),


T2 T1 ( x1 , x2 ) = T2 T1 ( x1 , x2 ) = T2 ( x2 , x1 ) = ( x2 , 0).

Since (0, x1 ) 6= ( x2 , 0) in general, we have T1 T2 6= T2 T1 .

Example Let T : R3 → R3 be defined by

T ( x, y, z) = (y, − x, 2z).

Find a formula for ( T 3 − 2T 2 + T − 2I )( x, y, z).

Solution.

I ( x, y, z) = ( x, y, z),
T 2 ( x, y, z) = T T ( x, y, z) = T (y, − x, 2z) = (− x, −y, 4z),

T 3 ( x, y, z) = T T 2 ( x, y, z) = T (− x, −y, 4z) = (−y, x, 8z).

Therefore

( T 3 − 2T 2 + T − 2I )( x, y, z) = (−y, x, 8z) − 2(− x, −y, 4z) + (y, − x, 2z) − 2( x, y, z)

 
= (−y) + 2x + y − 2x, x + 2y − x − 2y, 8z − 8z + 2z − 2z
= (0, 0, 0).

Example Let the linear transformations T : V2 → V2 and S : V2 → V2 be de-

fined by

T ( x1 , x2 ) = ( x1 + x2 , 0), S( x1 , x2 ) = (2x1 , 3x1 + 4x2 ).

Determine the linear transformation 3S + 7T.

32
Solution

(3S + 7T )( x1 , x2 ) = 3S( x1 , x2 ) + 7T ( x1 , x2 )



= 3 2x1 , 3x1 + 4x2 + 7 x1 + x2 , 0



= 6x1 , 9x1 + 12x2 + 7x1 + 7x2 , 0


= 6x1 + 7x1 + 7x2 , 9x1 + 12x2


= 13x1 + 7x2 , 9x1 + 12x2 .

Test Your Knowledge

1. Let S and T be two linear operators on R2 defined by S( x, y) = (0, x ) and

T ( x, y) = ( x, y). Show that TS = 0, but ST 6= 0, and that T 2 = T.

2. Let U and V be vector spaces over a field F, and let T1 , T2 : U → V be linear

transformations. Define a map M : U → V by

M ( X ) = T1 ( X ) + T2 ( X ), ∀ X ∈ U.

Equivalently, we can write

M = T1 + T2 , so that ( T1 + T2 )( X ) = T1 ( X ) + T2 ( X ), ∀ X ∈ U.

Prove that M is also a linear transformation from U to V.

Inverse of a Linear Transformation

Let T : V → W. The mapping T −1 : W → V is called the inverse of T if T is

invertible.
Suppose T : V → W is an invertible linear transformation. Then there exists a
linear transformation T −1 : W → V such that

TT −1 = T −1 T = I,

where I denotes the identity mapping on V or W (as appropriate). This T −1 is

called the inverse of T.

33
In Other Words

The mapping T : V → W is invertible if there exists a mapping T −1 : W → V such

that
T −1 T = IV and TT −1 = IW ,
where IV and IW are the identity mappings on V and W respectively. If T is
invertible, then T −1 is unique.
Let V be a vector space over the field R. If T : V → V is a linear operator, then T
is said to be invertible if it is both oneâĂŞone and onto.

Right and Left Inverses.

• T is called right-invertible if there exists a linear transformation S such that

TS = I.

• T is called left-invertible if there exists a linear transformation U such that

UT = I.

If T is both right-invertible and left-invertible and

TS = UT = I,

then U = S, and S is unique. In this case we write S = T −1 .

Thus, T is invertible (or regular) if it is both right-invertible and left-invertible.

Remark. A linear transformation is completely determined by its action on a

basis of V. Therefore, to determine the values of T −1 on the elements of a basis of
W, it suffices to specify the values of T on the elements of a basis of V.

Non-Singular and Singular Linear Transformations

A linear transformation T : V → W over a field F is said to be non-singular if

T ( x ) = 0 =⇒ x = 0.

Equivalently, the null space (kernel) of T is trivial:

ker( T ) = {0}.

34
In other words, only the zero vector of V maps to the zero vector of W.
Evidently, T is one-one (injective) if and only if T is non-singular. Thus,
non-singular linear transformations are precisely those that preserve linear inde-
pendence.
A linear transformation T : V → W over the same field F is said to be singular
if there exists some non-zero vector α ∈ V such that

T (α) = 0.

That is, the kernel of T contains a non-zero vector.

Theorem 7.1. Let V and W be finite-dimensional vector spaces over a field F with
dim V = dim W = n. If T : V → W is a linear transformation, the following statements
are equivalent:

(i) T is invertible.

(ii) T is non-singular.

(iii) T is onto; that is, the range of T is W.

(iv) For every basis {α1 , α2 , . . . , αn } of V, the set { T (α1 ), T (α2 ), . . . , T (αn )} is a basis
of W.

(v) There exists a basis {α1 , α2 , . . . , αn } of V such that { T (α1 ), T (α2 ), . . . , T (αn )} is
a basis of W.

Proof. (a) (i ) ⇔ (ii ): A linear transformation is invertible if and only if it is

non-singular (by Theorem 3).

(b) (ii ) ⇔ (iii ): Suppose T is non-singular and let {α1 , . . . , αn } be a basis of V.

Then { T (α1 ), . . . , T (αn )} is linearly independent in W. Since dim W = n,
this set is a basis of W. Hence for any β ∈ W, there exist scalars b1 , . . . , bn
such that !
n n
β= ∑ bi T ( α i ) = T ∑ bi α i ,
i =1 i =1

showing that β lies in the range of T; thus T is onto.

(c) (iii ) ⇔ (iv): If T is onto and {α1 , . . . , αn } is a basis of V, then { T (α1 ), . . . , T (αn )}
spans W. Because dim W = n, these vectors are automatically linearly inde-
pendent, hence form a basis of W.

35
 (d) (iv) ⇔ (v): This is immediate: statement (iv) requires the basis condition
for every basis, while (v) asserts it for at least one basis.

(e) (v) ⇔ (i ): Assume there is some basis {α1 , . . . , αn } of V such that { T (α1 ), . . . , T (αn )}
is a basis of W. Then R( T ) = W. If α = ∑in=1 ai αi lies in the kernel K ( T ), we
have !
n n
T ∑ ai αi = ∑ ai T (αi ) = 0.
i =1 i =1

Because { T (αi )} is linearly independent, all ai = 0, hence α = 0. Thus

K ( T ) = {0}, so T is injective and onto, i.e. invertible.

Corollary 7.2. Let U and V be finite-dimensional vector spaces over F, and let T : U →
V be linear. Then dim U = dim V if and only if T is non-singular.

Example 1. Prove that the linear transformation

T : V3 → V3 defined by T ( α1 ) = α1 + α2 , T ( α2 ) = α2 + α3 , T ( α3 ) = α1 + α2 + α3

is non-singular, where {α1 , α2 , α3 } is a standard basis of V3 , and its inverse is given

by
T −1 ( y 1 , y 2 , y 3 ) = ( y 2 − y 3 , y 2 − y 1 , y 1 − y 2 + y 3 ).
Solution. Let ( x1 , x2 , x3 ) ∈ V3 . Since {α1 , α2 , α3 } is a standard basis, we can write

( x1 , x2 , x3 ) = x1 α1 + x2 α2 + x3 α3 .

Applying T, we get

T ( x1 , x2 , x3 ) = T ( x1 α1 + x2 α2 + x3 α3 )
= x1 T ( α1 ) + x2 T ( α2 ) + x3 T ( α3 )
= x1 ( α1 + α2 ) + x2 ( α2 + α3 ) + x3 ( α1 + α2 + α3 )
= ( x1 + x3 ) α1 + ( x1 + x2 + x3 ) α2 + ( x2 + x3 ) α3
= ( x1 + x3 , x1 + x2 + x3 , x2 + x3 ).

To show that T is one-one: Suppose

T ( x1 , x2 , x3 ) = (0, 0, 0).

36
Then 
 x1 + x3 = 0,


x1 + x2 + x3 = 0,


x2 + x3 = 0.


Solving this system, we get x1 = x2 = x3 = 0. Hence,

N ( T ) = {0},

which implies that T is one-one.

To show that T is onto: A linear transformation from V3 to V3 is one-one if
and only if it is onto. Since T is one-one, it is onto. Therefore, T is non-singular.
To find T −1 : Let

T ( x1 , x2 , x3 ) = ( y1 , y2 , y3 ) = ( x1 + x3 , x1 + x2 + x3 , x2 + x3 ).

Solving for ( x1 , x2 , x3 ) in terms of (y1 , y2 , y3 ), we get

x1 = y2 − y3 , x2 = y2 − y1 , x3 = y1 − y2 + y3 .

Thus,
T −1 ( y 1 , y 2 , y 3 ) = ( y 2 − y 3 , y 2 − y 1 , y 1 − y 2 + y 3 ).
Example 2. Let T be a linear operator on R3 defined by

T ( x1 , x2 , x3 ) = (3x1 , x1 − x2 , 2x1 + x2 + x3 ).

Determine whether T is invertible. If yes, find a formula for T −1 .

Solution. Given T : R3 → R3 . To prove T is one-one, assume

T ( x1 , x2 , x3 ) = 0 and solve for x1 , x2 , x3 .

If the only solution is (0, 0, 0), then T is non-singular and hence invertible.

Example 3. Let T : R3 → R3 be the linear operator defined by

T ( x1 , x2 , x3 ) = (3x1 , x1 − x2 , 2x1 + x2 + x3 ).

Prove that T is invertible and find T −1 .

37
 Solution. Step 1: Show that T is one-one.
Let
T ( α ) = T ( β ) , α = ( x 1 , x 2 , x 3 ) , β = ( y 1 , y 2 , y 3 ) ∈ R3 .
Then

T ( x1 , x2 , x3 ) = T (y1 , y2 , y3 ) =⇒ (3x1 , x1 − x2 , 2x1 + x2 + x3 ) = (3y1 , y1 − y2 , 2y1 + y2 + y3 ).

Comparing components:

3x1 = 3y1 =⇒ x1 = y1 ,


x1 − x2 = y1 − y2 =⇒ x2 = y2 ,


2x1 + x2 + x3 = 2y1 + y2 + y3 =⇒ x3 = y3 .


Hence, α = β, so T is one-one.

Step 2: Show that T is onto.

Let R( T ) denote the range of T. Consider the standard basis vectors of R3 :

α1 = (1, 0, 0), α2 = (0, 1, 0), α3 = (0, 0, 1).

Then
T (α1 ) = T (1, 0, 0) = (3, 1, 2) = β 1 ,
T (α2 ) = T (0, 1, 0) = (0, −1, 1) = β 2 ,
T (α3 ) = T (0, 0, 1) = (0, 0, 1) = β 3 .
The range R( T ) is generated by { β 1 , β 2 , β 3 }. To check linear independence, let

aβ 1 + bβ 2 + cβ 3 = 0 =⇒ a(3, 1, 2) + b(0, −1, 1) + c(0, 0, 1) = (0, 0, 0).

This gives the system:


3a = 0 =⇒ a = 0,


a − b = 0 =⇒ b = 0,


2a + b + c = 0 =⇒ c = 0.


Hence { β 1 , β 2 , β 3 } is linearly independent and spans R3 . Therefore, R( T ) =

R3 , and T is onto.

Step 3: Conclude invertibility.

Since T is one-one and onto, T is invertible.

38
 Step 4: Find T −1 .
Let T ( x1 , x2 , x3 ) = (y1 , y2 , y3 ), i.e.,

y1 = 3x1 , y2 = x1 − x2 , y3 = 2x1 + x2 + x3 .

Solving for x1 , x2 , x3 in terms of y1 , y2 , y3 :

1 1
x1 = y1 , x2 = x1 − y2 = y1 − y2 , x3 = y3 − 2x1 − x2 = −y1 + y2 + y3 .
3 3
Hence, the inverse is
 
−1 1 1
T ( y1 , y2 , y3 ) = y1 , y1 − y2 , − y1 + y2 + y3 .
3 3

Example 4 Let T be a linear operator on R3 defined by

T ( x, y, z) = (2x, 2x − 5y, 2y + z).

Find T −1 .

Solution Let
T ( x, y, z) = ( a, b, c),
so that
( a, b, c) = (2x, 2x − 5y, 2y + z). (1)
Comparing components, we get:

a = 2x, b = 2x − 5y, c = 2y + z.

From the first equation:

a
x= .
2
From the second equation:
a


2x − b 2 2 −b a−b
b = 2x − 5y =⇒ y = = = .
5 5 5
From the third equation:

a−b 2a − 2b 5c − 2a + 2b
 
z = c − 2y = c − 2 = c− = .
5 5 5

39
Thus, from (1):

a a − b 5c − 2a + 2b
 
−1
T ( a, b, c) = , , .
2 5 5

In general, for ( x, y, z) ∈ R3 :

x x − y −2x + 2y + 5z
 
−1
T ( x, y, z) = , , .
2 5 5

40