chapter chapter 2 vector spaces 2.1 introduction in various practical and theoretical problems, we...

Chapter 2Vector Spaces

2.1 INTRODUCTIONIn various practical and theoretical problems, we come across a set V whose elements may bevectors in two or three dimensions or real-valued functions, which can be added and multi-plied by a constant (number) in a natural way, the result being again an element of V . Suchconcrete situations suggest the concept of a vector space. Vector spaces play an important rolein many branches of mathematics and physics. In analysis infinite dimensional vector spaces(in fact, normed vector spaces) are more important than finite dimensional vector spaces whilein linear algebra finite dimensional vector spaces are used, because they are simple and lineartransformations on them can be represented by matrices. This chapter mainly deals with finitedimensional vector spaces.

2.2 DEFINITION AND EXAMPLES

VECTOR SPACE. An algebraic structure (V, F, ⊕, �) consisting of a non-void set V , a field F, abinary operation ⊕ onV and an external mapping � : F×V →V associating each a ∈F,v ∈Vto a unique element a� v ∈V is said to be a vector space over field F, if the following axiomsare satisfied:

V -1 (V, ⊕) is an abelian group.V-2 For all u, v ∈V and a, b ∈ F, we have

(i) a� (u⊕ v) = a�u⊕a� v,(ii) (a+ b)�u = a�u⊕b�u,

(iii) (ab)�u = a� (b�u),(iv) 1�u = u.

The elements of V are called vectors and those of F are called scalars. The mapping � iscalled scalar multiplication and the binary operation ⊕ is termed vector addition.

115

116 • Theory and Problems of Linear Algebra

If there is no danger of any confusion we shall say V is a vector space over a field F ,whenever the algebraic structure (V, F, ⊕, �) is a vector space. Thus, whenever we say thatV is a vector space over a field F , it would always mean that (V, ⊕) is an abelian group and� : F ×V →V is a mapping such that V -2 (i)–(iv) are satisfied.

V is called a real vector space if F = R (field of real numbers), and a complex vector spaceif F = C (field of complex numbers).

REMARK-1 V is called a left or a right vector space according as the elements of a skew-fieldF are multiplied on the left or right of vectors in V . But, in case of a field these two conceptscoincide.

REMARK-2 The symbol ‘+’ has been used to denote the addition in the field F. For the sakeof convenience, in future, we shall use the same symbol ‘+’ for vector addition ⊕ and additionin the field F. But, the context would always make it clear as to which operation is meant.Similarly, multiplication in the field F and scalar multiplication � will be denoted by the samesymbol ‘·’.

REMARK-3 In this chapter and in future also, we will be dealing with two types of zeros. Onewill be the zero of the additive abelian group V , which will be known as the vector zero andother will be the zero element of the field F which will be known as scalar zero. We will use thesymbol 0V to denote the zero vector and 0 to denote the zero scalar.

REMARK-4 Since (V,⊕) is an abelian group, therefore, for any u,v,w ∈ V the followingproperties will hold:

(i) u⊕ v = u⊕w ⇒ v = w(ii) v⊕u = w⊕u ⇒ v = w

}

(Cancellation laws)

(iii) u⊕ v = 0v ⇒ u = −v and v = −u(iv) −(−u) = u(v) u⊕ v = u ⇒ v = 0v

(vi) −(u⊕ v) = (−u)⊕ (−v)(vii) 0v is unique

(viii) For each u ∈V, −u ∈V is unique.

REMARK-5 If V is a vector space over a field F, then we will write V (F).

ILLUSTRATIVE EXAMPLES

EXAMPLE-1 Every field is a vector space over its any subfield.

SOLUTION Let F be a field, and let S be an arbitrary subfield of F . Since F is an additiveabelian group, therefore, V -1 holds.

Taking multiplication in F as scalar multiplication, the axioms V -2 (i) to V -2 (iii) arerespectively the right distributive law, left distributive law and associative law. Also V -2 (iv) isproperty of unity in F .

Vector Spaces • 117

Hence, F is a vector space over S.Since S is an arbitrary sub field of F , therefore, every field is a vector space over its any

subfield.

REMARK. The converse of the above Example is not true, i.e. a subfield is not necessarilya vector space over its over field. For example, R is not a vector space over C, becausemultiplication of a real number and a complex number is not necessarily a real number.

EXAMPLE-2 R is a vector space over Q, because Q is a subfield of R.

EXAMPLE-3 C is a vector space over R, because R is a subfield of C.

EXAMPLE-4 Every field is a vector space over itself.

SOLUTION Since every field is a subfield of itself. Therefore, the result directly follows fromExample 1.

REMARK. In order to know whether a given non-void set V forms a vector space over a fieldF, we must proceed as follows:(i) Define a binary operation on V and call it vector addition.(ii) Define scalar multiplication on V , which associates each scalar in F and each vector in

V to a unique vector in V .(iii) Define equality of elements (vectors) in V .(iv) Check whether V-1 and V-2 are satisfied relative to the vector addition and scalar

multiplication thus defined.

EXAMPLE-5 The set Fm×n of all m× n matrices over a field F is a vector space over F withrespect to the addition of matrices as vector addition and multiplication of a matrix by a scalaras scalar multiplication.

SOLUTION For any A,B,C ∈ Fm×n, we have(i) (A + B)+C = A +(B +C) (Associative law)

(ii) A + B = B + A (Commutative law)(iii) A + O = O + A = A (Existence of Identity)(iv) A +(−A) = O = (−A)+ A (Existence of Inverse)

Hence, Fm×n is an abelian group under matrix addition.If λ∈F and A∈Fm×n, then λA∈Fm×n and for all A,B∈Fm×n and for all λ,µ∈F , we have

(i) λ(A + B) = λA + λB(ii) (λ+ µ)A = λA + µA

(iii) λ(µA) = (λµ)A(iv) 1A = A

Hence, Fm×n is a vector space over F .


EXAMPLE-6 The set R m×n of all m× n matrices over the field R of real numbers is a vectorspace over the field R of real numbers with respect to the addition of matrices as vector additionand multiplication of a matrix by a scalar as scalar multiplication.

SOLUTION Proceed parallel to Example 5.

REMARK. The set Q m×n of all m× n matrices over the field Q of rational numbers is not avector space over the field R of real numbers, because multiplication of a matrix in Qm×n anda real number need not be in Q m×n. For example, if

√2 ∈ R and A ∈ Q m×n, then

√2A ∈/Qm×n

because the elements of matrix√

2A are not rational numbers.

EXAMPLE-7 The set of all ordered n-tuples of the elements of any field F is a vector spaceover F.

SOLUTION Recall that if a1, a2, . . . , an ∈ F , then (a1,a2, . . . , an) is called an orderedn-tuple of elements of F . Let V = {(a1, a2, . . . ,an) : ai ∈ F for all i ∈ n} be the set of allordered n-tuples of elements of F .

Now, to give a vector space structure to V over the field F , we define a binary operation onV , scalar multiplication on V and equality of any two elements of V as follows:

Vector addition (addition on V ): For any u = (a1, a2, . . . ,an), v = (b1, b2, . . . ,bn) ∈ V , wedefine

u+ v = (a1 + b1, a2 + b2, . . . ,an + bn)

Since ai + bi ∈ F for all i ∈ n. Therefore, u+ v ∈V .Thus, addition is a binary operation on V .

Scalar multiplication on V : For any u = (a1, a2, . . . ,an) ∈V and λ ∈ F , we define

λu = (λa1, λa2, . . . ,λan).

Since λai ∈ F for all i ∈ n. Therefore, λu ∈V .Thus, scalar multiplication is defined on V .

Equality of two elements of V : For any u = (a1, a2, . . . ,an),v = (b1, b2, . . . ,bn)∈V , we define

u = v ⇔ ai = bi for all i ∈ n.

Since we have defined vector addition, scalar multiplication on V and equality of any twoelements of V . Therefore, it remains to check whether V -1 and V -2(i) to V -2(iv) are satisfied.


V -1. V is an abelian group under vector addition.

Associativity: For any u = (a1,a2, . . . ,an),ν = (b1, b2, . . . ,bn), w = (c1,c2, . . . ,cn) ∈ V, wehave

(u+ ν)+ w = (a1 + b1, . . . ,an + bn)+ (c1, c2, . . . ,cn)

= ((a1 + b1)+ c1, . . . ,(an + bn)+ cn)

= (a1 +(b1 + c1), . . . ,an +(bn + cn)) [By associativity of addition on F]= (a1, . . . ,an)+ (b1 + c1, . . . ,bn + cn)

= u+(ν+ w)

So, vector addition is associative on V .

Commutativity: For any u = (a1, a2, . . . ,an), ν = (b1, . . . ,bn) ∈V, we have

u+ ν = (a1 + b1, . . . ,an + bn)

= (b1 + a1, . . . ,bn + an) [By commutativity of addition on F]= ν+ u

So, vector addition is commutative on V .

Existence of identity (Vector zero): Since 0 ∈ F , therefore, 0 = (0, 0, . . . ,0) ∈V such that

u+ 0 = (a1, a2, . . . ,an)+ (0, 0, . . . ,0)

⇒ u+ 0 = (a1 + 0, a2 + 0, . . . ,an + 0) = (a1, a2, . . . ,an) = u for all u ∈V

Thus, 0 = (0, 0, . . . ,0) is the identity element for vector addition on V .

Existence of inverse: Let u = (a1, a2, . . . ,an) be an arbitrary element in V . Then,(−u) = (−a1, −a2, . . . ,−an) ∈V such that

u+(−u) =(

a1 +(−a1), . . . ,an +(−an))

= (0, 0, . . . ,0) = 0.

Thus, every element in V has its additive inverse in V .Hence, V is an abelian group under vector addition.

V -2 For any u = (a1, . . . ,an), ν = (b1, . . . ,bn) ∈V and λ, µ∈ F, we have

(i) λ(u+ ν) = λ(a1 + b1, . . . ,an + bn)

= (λ(a1 + b1), . . . ,λ(an + bn))

= (λa1 + λb1, . . . ,λan + λbn)

[

By left distributivity of multiplicationover addition in F

]

= (λa1, . . . ,λan)+ (λb1, . . . ,λbn)

= λu+ λν


(ii) (λ+ µ)u =(

(λ+ µ)a1, . . . ,(λ+ µ)an)

= (λa1 + µa1, . . . ,λan + µan) [By right dist. of mult. over add. in F]= (λa1, . . . ,λan)+ (µa1, . . . ,µan)

= λu+ µu.

(iii) (λµ)u =(

(λµ)a1, . . . ,(λµ)an)

=(

λ(µa1), . . . ,λ(µan))

[By associativity of multiplication in F]= λ(µu)

(iv) 1u = (1a1, . . . ,1an) = (a1, . . . ,an) = u [∵ 1 is unity in F]

Hence, V is a vector space over F . This vector space is usually denoted by Fn.

REMARK. If F = R, then Rn is generally known as Euclidean space and if F = C, then Cn iscalled the complex Euclidean space.

EXAMPLE-8 Show that the set F [x] of all polynomials over a field F is a vector spaceover F.

SOLUTION In order to give a vector space structure to F[x], we define vector addition, scalarmultiplication, and equality in F[x] as follows:

Vector addition on F[x]: If f (x) = ∑i

aixi, g(x) = ∑i

bixi ∈ F[x], then we define

f (x)+ g(x) = Σi(ai + bi)xi

Clearly, f (x)+ g(x) ∈ F[x], because ai + bi ∈ F for all i.

Scalar multiplication on F[x]: For any λ ∈ F and f (x) = ∑i

aixi ∈ F[x] λ f (x) is defined as the

polynomial ∑i(λai)xi. i.e,

λ f (x) = Σi(λai)xi

Obviously, λ f (x) ∈ F[x], as λai ∈ F for all i.

Equality of two elements of F[x]: For any f (x) = ∑i

aixi,g(x) = ∑i

bixi, we define

f (x) = g(x) ⇔ ai = bi for all i.

Now we shall check whether V -1 and V -2(i) to V -2(iv) are satisfied.


V -1 F[x] is an abelian group under vector addition.

Associativity: If f (x) = ∑i

aixi, g(x) = ∑i

bixi, h(x) = ∑i

cixi ∈ F[x], then

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

(

Σi(ai + bi)xi

)

+ Σi

cixi

= Σi[(ai + bi)+ ci]xi

= Σi[ai +(bi + ci)]xi [By associativity of + on F]

= Σi

aixi + Σi(bi + ci)xi

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

So, vector addition is associative on F[x].

Commutativity: If f (x) = ∑i

aixi, g(x) = ∑i

bixi ∈ F[x], then

f (x)+ g(x) = Σi(ai + bi)xi = Σ

i(bi + ai)xi [By commutativity of + on F]

⇒ f (x)+ g(x) = Σi

bixi + Σi

aixi = g(x)+ f (x)

So, vector addition is commutative on F[x].

Existence of zero vector: Since 0̂(x) = ∑i

0xi ∈ F[x] is such that for all f (x) = ∑i

aixi ∈ F[x]

⇒ 0̂(x)+ f (x) =Σi

0xi + Σi

aixi =Σi(0+ ai)xi

⇒ 0̂(x)+ f (x) =Σi

aixi [∵ 0 ∈ F is the additive identity]

⇒ 0̂(x)+ f (x) = f (x)Thus, 0̂(x)+ f (x) = f (x) = f (x)+ 0̂(x) for all f (x) ∈ F [x].

So, 0̂(x) is the zero vector (additive identity) in F[x].

Existence of additive inverse: Let f (x) = ∑i

aixi be an arbitrary polynomial in F[x]. Then,

− f (x) = Σi(−ai)xi ∈ F[x] such that

f (x)+ (− f (x)) = Σi

aixi + Σi(−ai)xi = Σ

i[ai +(−ai)]xi = Σ

i0xi = 0̂(x)

Thus, for each f (x) ∈ F[x], there exists− f (x) ∈ F[x] such that

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

So, each f (x) has its additive inverse in F[x].Hence, F[x] is an abelian group under vector addition.


V -2 For any f (x) = ∑i

aixi, g(x) = ∑i

bixi ∈ F[x] and λ, µ∈ F, we have

(i) λ[ f (x)+ g(x)] = λ(

Σi(ai + bi)xi

)

= Σi

λ(ai + bi)xi

= Σi(λai + λbi)xi [By left distributivity of multiplication over addition]

= Σi(λai)xi + Σ(λbi)xi

= λ f (x)+ λg(x)

(ii) (λ+ µ) f (x) = Σi(λ+ µ)aixi

= Σi(λai + µai)xi [By right distributivity of multiplication over addition]

= Σi(λai)xi + Σ

i(µai)xi

= λ f (x)+ µf (x).

(iii) (λµ) f (x) = Σi(λµ)aixi

= Σi

λ(µai)xi [By associativity of multiplication on F]

= λΣi(µai)xi

= λ(µf (x))

(iv) 1 f (x) = Σi(1ai)xi

= Σi

aixi [∵ 1 is unity in F]

= f (x)

Hence, F[x] is a vector space over field F .

EXAMPLE-9 The set V of all real valued continuous (differentiable or integrable) functionsdefined on the closed interval [a, b] is a real vector space with the vector addition and scalarmultiplication defined as follows:

( f + g)(x) = f (x)+ g(x)(λ f )(x) = λ f (x)

For all f , g ∈V and λ ∈ R.


SOLUTION Since the sum of two continuous functions is a continuous function. Therefore,vector addition is a binary operation on V . Also, if f is continuous and λ ∈ R, then λ f iscontinuous. We know that for any f , g ∈V

f = g ⇔ f (x) = g(x) for all x ∈ [a,b].

Thus, we have defined vector addition, scalar multiplication and equality in V .It remains to verify V -1 and V -2(i) to V -2(iv).

V-1 V is an abelian group under vector addition.

Associativity: Let f , g, h be any three functions in V . Then,

[( f + g)+ h](x) = ( f + g)(x)+ h(x) = [ f (x)+ g(x)]+ h(x)= f (x)+ [g(x)+ h(x)] [By associativity of addition on R]= f (x)+ (g+ h)(x)= [ f +(g+ h)](x) for all x ∈ [a, b]

So, ( f + g)+ h = f +(g+ h)

Thus, vector addition is associative on V .

Commutativity: For any f , g ∈V , we have

( f + g)(x) = f (x)+ g(x) = g(x)+ f (x) [By commutativity of + on R]⇒ ( f + g)(x) = (g+ f )(x) for all x ∈ [a, b]

So, f + g = g+ f

Thus, vector addition is commutative on V .

Existence of additive identity: The function 0̂(x) = 0 for all x ∈ [a, b] is the additive identity,because for any f ∈V

( f + 0̂)(x) = f (x)+ 0̂(x) = f (x)+ 0 = f (x) = (0̂+ f )(x) for all x ∈ [a, b].

Existence of additive inverse: Let f be an arbitrary function in V . Then, a function − f definedby (− f )(x) = − f (x) for all x ∈ [a, b] is continuous on [a, b], and

[ f +(− f )](x) = f (x)+ (− f )(x) = f (x)− f (x) = 0 = 0̂(x) for all x ∈ [a, b].

Thus, f +(− f ) = 0̂ = (− f )+ f

So, each f ∈V has its additive inverse − f ∈V .Hence, V is an abelian group under vector addition.


V-1. V1 ×V2 is an abelian group under vector addition:

Associativity: For any (u1, u2), (v1, v2), (w1, w2) ∈V1 ×V2, we have

[(u1, u2)+ (v1, v2)]+ (w1, w2) =(

(u1 + v1), (u2 + v2))

+(w1, w2)

=(

(u1 + v1)+ w1, (u2 + v2)+ w2)

=(

u1 +(v1 + w1), u2 +(v2 + w2))

By associativity ofvector addition on V1and V2

= (u1, u2)+ (v1 + w1, v2 + w2)

= (u1, u2)+ [(v1, v2)+ (w1, w2)]

So, vector addition is associative on V1 ×V2.

Commutativity: For any (u1, u2), (v1, v2) ∈V1 ×V2, we have

(u1, u2)+ (v1, v2) = (u1 + v1, u2 + v2)

= (v1 + u1, v2 + u2) [By commutativity of vector addition on V1 and V2]= (v1, v2)+ (u1, u2)

So, vector addition is commutative on V1 ×V2.

Existence of additive identity: If 0V1 and 0V2 are zero vectors in V1 and V2 respectively, then(0V1 , 0V2) is zero vector (additive identity) in V1 ×V2. Because for any (u1,u2) ∈V1 ×V2

(u1, u2)+ (0V1, 0V2) = (u1 + 0V1, u2 + 0V2)

= (u1, u2) = (0V1 , 0V2)+ (u1, u2) [By commutativity of vector addition]

Existence of additive Inverse: Let (u1, u2) ∈V1 ×V2. Then, (−u1, −u2) is its additive inverse.Because,

(u1, u2)+ (−u1, −u2) =(

u1 +(−u1), u2 +(−u2))

= (0V1, 0V2) = (−u1, −u2)+ (u1, u2)

[

By commutativity of vectoraddition

]

Thus, each element of V1 ×V2 has its additive inverse in V1 ×V2.Hence, V1 ×V2 is an abelian group under vector addition.

V-2. For any (u1, u2),(v1, v2) ∈V1 ×V2 and λ, µ∈ F, we have

(i) λ[(u1, u2)+ (v1, v2)] = λ(u1 + v1, u2 + v2)

=(

λ(u1 + v1), λ(u2 + v2))

= (λu1 + λv1, λu2 + λv2) [By V -2(i) in V1 and V2]= (λu1, λu2)+ (λv1, λv2)

= λ(u1, u2)+ λ(v1, v2)


(ii) (λ+ µ)(u1,u2) =(

(λ+ µ)u1, (λ+ µ)u2)

=(

λu1 + µu1, λu2 + µu2)

[By V -2(ii) in V1 and V2]= (λu1, λu2)+ (µu1, µu2)

= λ(u1, u2)+ µ(u1, u2)

(iii) (λµ)(u1, u2) =(

(λµ)u1,(λµ)u2)

=(

λ(µu1),λ(µu2))

[By V -2(iii) in V1 and V2]= λ(µu1,µu2)

= λ{µ(u1,u2)}

(iv) 1(u1,u2) = (1u1,1u2)

= (u1,u2) [By V -2(iv) in V1 and V2]

Hence, V1 ×V2 is a vector space over field F .

EXAMPLE-11 Let V be a vector space over a field F. Then for any n ∈ N the function spaceV n = { f : f : n →V} is a vector space over field F.

SOLUTION We define vector addition, scalar multiplication and equality in V n as follows:

Vector addition: Let f , g ∈V n, then the vector addition on V n is defined as

( f + g)(x) = f (x)+ g(x) for all x ∈ n.

Scalar multiplication: If λ ∈ F and f ∈V n, then λ f is defined as

(λ f )(x) = λ( f (x)) for all x ∈ n.

Equality: f , g ∈V n are defined as equal if f (x) = g(x) for all x ∈ n.Now we shall verify V n-1 and V n-2(i) to V n-2 (iv).

V n −1. V n is an abelian group under vector addition:

Associativity: For any f , g, h ∈V n, we have

[( f + g)+ h](x) = ( f + g)(x)+ h(x) = [ f (x)+ g(x)]+ h(x) for all x ∈ n= f (x)+ [g(x)+ h(x)] [By associativity of vector addition on V ]= f (x)+ (g+ h)(x)= [ f +(g+ h)](x) for all x ∈ n

∴ ( f + g)+ h = f +(g+ h)

So, vector addition is associative on V n.


Commutativity: For any f , g ∈V n, we have

( f + g)(x) = f (x)+ g(x) for all x ∈ n⇒ ( f + g)(x) = g(x)+ f (x) [By commutativity of vector addition on V ]⇒ ( f + g)(x) = (g+ f )(x) for all x ∈ n∴ f + g = g+ f

So, vector addition is commutative on V n.

Existence of additive identity (zero vector): The function ̂0 : n → V defined by ̂0(x) = 0vfor all x ∈ n is the additive identity, because

( f +̂0)(x) = f (x)+̂0(x) = f (x)+ 0V

⇒ ( f +̂0)(x) = f (x) for all f ∈V n and for all x ∈ n [∵ 0V is zero vector in V ]

Existence of additive inverse: Let f be an arbitrary function in V n. Then − f : n → Vdefined by

(− f )(x) = − f (x) for all x ∈ n

is additive inverse of f , because

[ f +(− f )](x) = f (x)+ (− f )(x) = f (x)− f (x) = 0V = ̂0(x) for all x ∈ n.

Hence, V n is an abelian group under vector addition.

V n-2. For any f , g ∈V n and λ,µ∈ F, we have

(i) [λ( f + g)](x) = λ[( f + g)(x)] = λ[ f (x)+ g(x)]⇒ [λ( f + g)](x) = λ f (x)+ λg(x) [By V -2(i) in V ]⇒ [λ( f + g)](x) = (λ f + λg)(x) for all x ∈ n∴ λ( f + g) = λ f + λg.

(ii) [(λ+ µ) f ](x) = (λ+ µ) f (x)⇒ [(λ+ µ) f ](x) = λ f (x)+ µf (x) [By V -2(ii) in V ]⇒ [(λ+ µ) f ](x) = (λ f )(x)+ (µf )(x)⇒ [(λ+ µ) f ](x) = (λ f + µf )(x) for all x ∈ n∴ (λ+ µ) f = λ f + µf

(iii) [(λµ) f ](x) = (λµ) f (x)⇒ [(λµ) f ](x) = λ

(

µf (x))

[By V -2(iii) in V ]⇒ [(λµ) f ](x) = λ

(

(µf )(x))


9. Let U and W be two vector spaces over a field F . Let V be the set of ordered pairs (u,w)where u ∈ U and w ∈ W . Show that V is a vector space over F with addition in V andscalar multiplication V defined by

(u1, w1)+ (u2, w2) = (u1 + u2, w1 + w2)

and a(u, w) = (au, aw)

(This space is called the external direct product of u and w).10. Let V be a vector space over a field F and n ∈ N. Let V n = {(v1, v2, . . . ,vn) : vi ∈ V ;

i = 1, 2, . . . ,n}. Show that V n is a vector space over field F with addition in V and scalarmultiplication on V defined by

(u1, u2, . . . ,un)+ (v1, v2, . . . ,vn) = (u1 + v1, u2 + v2, . . . ,un + vn)

and a(v1, v2, . . . ,vn) = (av1, av2, . . . ,avn)

11. Let V = {(x, y) : x, y ∈ R}. Show that V is not a vector space over R under the additionand scalar multiplication defined by

(a1, b1)+ (a2, b2) = (3b1 + 3b2, −a1 −a2)

k(a1, b1) = (3kb1, −ka1)

for all (a1, b1), (a2, b2) ∈V and k ∈ R.12. Let V = {(x, y) : x,y ∈ R}. For any α = (x1, y1),β = (x2, y2) ∈V, c ∈ R, define

α⊕β = (x1 + x2 + 1, y1 + y2 + 1)

c�α = (cx1, cy1)

(i) Prove that (V,⊕) is an abelian group.(ii) Verify that c� (α⊕β) �= c�α⊕ c�β.

(iii) Prove that V is not a vector space over R under the above two operations.13. For any u = (x1, x2, x3), v = (y1, y2, y3) ∈ R3 and a ∈ R, let us define

u⊕ v = (x1 + y1 + 1, x2 + y2 + 1, x3 + y3 + 1)

a�u = (ax1 + a−1, ax2 + a−1, ax3 + a−1)

Prove that R3 is a vector space over R under these two operations.14. Let V denote the set of all positive real numbers. For any u, v ∈V and a ∈ R, define

u⊕ v = uva�u = ua

Prove that V is a vector space over R under these two operations.15. Prove that R3 is not a vector space over R under the vector addition and scalar multipli-

cation defined as follows:(x1, x2, x3)+ (y1, y2, y3) = (x1 + y1, x2 + y2, x3 + y3)

a(x1, x2, x3) = (ax1, ax2, ax3)


16. Let V = {(x, 1) : x ∈ R}. For any u = (x, 1), v = (y, 1) ∈V and a ∈ R, define

u⊕ v = (x+ y,1)

a�u = (ax,1)

Prove that V is a vector space over R under these two operations.

17. Let V be the set of ordered pairs (a,b) of real numbers. Let us define

(a, b)⊕ (c, d) = (a+ c, b+ d)

and k(a, b) = (ka, 0)

Show that V is not a vector space over R.

18. Let V be the set of ordered pairs (a,b) of real numbers. Show that V is not a vector spaceover R with addition and scalar multiplication defined by

(i) (a, b)+ (c, d) = (a+ d, b+ c) and k(a, b) = (ka, kb)

(ii) (a, b)+ (c, d) = (a+ c, b+ d) and k(a, b) = (a, b)

(iii) (a, b)+ (c, d) = (0, 0) and k(a, b) = (ka, kb)

(iv) (a, b)+ (c, d) = (ac, bd) and k(a, b) = (ka, kb)

19. The set V of all convergent real sequences is a vector space over the field R of all realnumbers.

20. Let p be a prime number. The set Zp = {0, 1, 2, . . . ,(p−1)} is a field under addition andmultiplication modulo p as binary operations. Let V be the vector space of polynomialsof degree at most n over the field Zp. Find the number of elements in V .[Hint: Each polynomial in V is of the form f (x) = a0 + a1x + · · · + anxn, wherea0, a1, a2, . . . ,an ∈ Zp. Since each ai can take any one of the p values 0, 1, 2, . . . ,(p−1).So, there are pn+1 elements in V ]

2.3 ELEMENTARY PROPERTIES OF VECTOR SPACES

THEOREM-1 Let V be a vector space over a field F. Then

(i) a ·0V = 0V for all a ∈ F

(ii) 0 · v = 0V for all v ∈V

(iii) a · (−v) = −(a · v) = (−a) · v for all v ∈V and for all a ∈ F

(iv) (−1) · v = −v for all v ∈ F

(v) a · v = 0V ⇒ either a = 0 or v = 0V


PROOF. (i) We have,

a ·0V =a · (0V + 0V ) for all a ∈ F [∵ 0V + 0V = 0V ]⇒ a ·0V =a ·0V + a ·0V for all a ∈ F [By V -2(i)]∴ a ·0V + 0V =a ·0V + a ·0V for all a ∈ F [0V is additive identity in V ]

⇒ 0V =a ·0V for all a ∈ F[

By left cancellation law for additionon V

]

(ii) For any v ∈V , we have

0 · v = (0+ 0) · v [∵ 0 is additive identity in F]⇒ 0 · v = 0 · v+ 0 · v [By V -2(ii)]∴ 0 · v+ 0V = 0 · v+ 0 · v [∵ 0V is additive identity in V ]⇒ 0 · v = 0V [By left cancellation law for addition in V ]

So, 0 · v = 0V for all v ∈V .(iii) For any v ∈V and a ∈ F , we have

a · [v+(−v)] = a · v+ a · (−v) [By V -2(i)]⇒ a ·0V = a · v+ a · (−v) [∵ v+(−v) = 0V for all v ∈V ]⇒ 0V = a · v+ a · (−v) [By (i)]⇒ a · (−v) = −(a · v) [∵ V is an additive group]

Again, [a+(−a)] · v = a · v+(−a) · v [By V -2(ii)]⇒ 0 · v = a · v+(−a) · v [∵ a+(−a) = 0]⇒ 0V = a · v+(−a) · v [By (ii)]⇒ (−a) · v = −(a · v) [∵ V is an additive group]

Hence, a · (−v) = −(a · v) = (−a) · v for all a ∈ F and for all v ∈V.

(iv) Taking a = 1 in (iii), we obtain

(−1) · v = −v for all v ∈V .

(v) Let a · v = 0V , and let a �= 0. Then a−1 ∈ F .Now

a · v = 0V

⇒ a−1 · (a · v) = a−1 ·0V

⇒ (a−1a) · v = a−1 ·0V [By V -2(iii)]⇒ (a−1a) · v = 0V [By (i)]⇒ 1 · v = 0V

⇒ v = 0V [By V -2(iv)]


Again let a ·v = 0V and v �= 0V . Then to prove that a = 0, suppose a �= 0. Then a−1 ∈ F .Now

a · v = 0V⇒ a−1 · (a · v) = a−1 ·0V⇒ (a−1a) · v = a−1 ·0V [By V -2(iii)]⇒ 1 · v = 0V [By (i)]⇒ v = 0V , which is a contradiction. [By V -2(iv)]

Thus, a · v = 0V and v �= 0V ⇒ a = 0.Hence, a · v = 0V ⇒ either a = 0 or v = 0V .

Q.E.D.

It follows from this theorem that the multiplication by the zero of V to a scalar in F or bythe zero of F to a vector in V always leads to the zero of V .

Since it is convenient to write u− v instead of u + (−v). Therefore, in what follows weshall write u− v for u+(−v).REMARK. For the sake of convenience, we drop the scalar multiplication ‘·’ and shall write avin place of a · v.

THEOREM-2 Let V be a vector space over a field F. Then,(i) If u, v ∈V and 0 �= a ∈ F, then

au = av ⇒ u = v(ii) If a, b ∈ F and 0V �= u ∈V , then

au = bu ⇒ a = bPROOF. (i) We have,

au = av⇒ au+(−(av)) = av+(−(av))⇒ au−av = 0V

⇒ a(u− v) = 0V

⇒ u− v = 0V [By Theorem 1(v)]⇒ u = v

(ii) We have,au = bu

⇒ au+(−(bu)) = bu+(−(bu))

⇒ au−bu = 0V

⇒ (a−b)u = 0V

⇒ a−b = 0 [By Theorem 1(v)]⇒ a = b Q.E.D.


EXERCISE 2.2

1. Let V be a vector space over a field F . Then prove that

a(u− v) = au−av for all a ∈ F and u,v ∈V.

2. Let V be a vector space over field R and u, v ∈V . Simplify each of the following:

(i) 4(5u−6v)+ 2(3u+ v) (ii) 6(3u+ 2v)+ 5u−7v

(iii) 5(2u−3v)+ 4(7v+ 8) (iv) 3(5u+2v)

3. Show that the commutativity of vector addition in a vector space V can be derived fromthe other axioms in the definition of V .

4. Mark each of the following as true or false.

(i) Null vector in a vector space is unique.(ii) Let V (F) be a vector space. Then scalar multiplication on V is a binary operation

on V .(iii) Let V (F) be a vector space. Then au = av ⇒ u = v for all u, v ∈ V and for all

a ∈ F .(iv) Let V (F) be a vector space. Then au = bu ⇒ a = b for all u ∈ V and for all

a, b ∈ F .

5. If V (F) is a vector space, then prove that for any integer n, λ ∈ F, v ∈ V, prove thatn(λv) = λ(nv) = (nλ)v.

ANSWERS

2. (i) 26u−22v (ii) 23u+ 5v(iii) The sum 7v+8 is not defined, so the given expression is not meaningful.(iv) Division by v is not defined, so the given expression is not meaningful.

4. (i) T (ii) F (iii) F (iv) F

2.4 SUBSPACESSUB-SPACE Let V be a vector space over a field F. A non-void subset S of V is said to be asubspace of V if S itself is a vector space over F under the operations on V restricted to S.

If V is a vector space over a field F , then the null (zero) space {0V} and the entire space Vare subspaces of V . These two subspaces are called trivial (improper) subspaces of V and anyother subspace of V is called a non-trivial (proper) subspace of V .

THEOREM-1 (Criterion for a non-void subset to be a subspace) Let V be a vector space overa field F. A non-void subset S of V is a subspace of V iff or all u, v ∈ S and for all a ∈ F

(i) u− v ∈ S and, (ii) au ∈ S.


PROOF. First suppose that S is a subspace of vector space V . Then S itself is a vector spaceover field F under the operations on V restricted to S. Consequently, S is subgroup of theadditive abelian group V and is closed under scalar multiplication.

Hence, (i) and (ii) hold.Conversely, suppose that S is a non-void subset of V such that (i) and (ii) hold. Then,

(i) ⇒ S is an additive subgroup of V and therefore S is an abelian group under vectoraddition.

(ii) ⇒ S is closed under scalar multiplication.

Axioms V -2(i) to V -2(iv) hold for all elements in S as they hold for elements in V .Hence, S is a subspace of V . Q.E.D.

THEOREM-2 (Another criterion for a non-void subset to be a subspace) Let V be a vectorspace over a field F. Then a non-void subset S of V is a subspace of V iff au + bv ∈ S for allu, v ∈ S and for all a, b ∈ F.

PROOF. First let S be a subspace of V . Then,

au ∈ S, bv ∈ S for all u,v ∈ S and for all a,b ∈ F [By Theorem 1]⇒ au+ bv ∈ S for all u,v ∈ S and for all a,b ∈ F[∵ S is closed under vector addition]

Conversely, let S be a non-void subset of V such that au + bv ∈ S for all u, v ∈ S and for alla, b ∈ F .

Since 1, −1 ∈ F , therefore

1u+(−1)v ∈ S for all u, v ∈ S.

⇒ u− v ∈ S for all u, v ∈ S (i)

Again, since 0 ∈ F . Therefore,

au+ 0v ∈ S for all u ∈ S and for all a ∈ F.

⇒ au ∈ S for all u ∈ S and for all a ∈ F (ii)

From (i) and (ii), we getu− v ∈ S and au ∈ S for all u, v ∈ S and for all a ∈ F .Hence, by Theorem 1, S is a subspace of vector space V .

Q.E.D.


EXAMPLE-1 Show that S = {(0, b, c) : b, c ∈ R} is a subspace of real vector space R3.

SOLUTION Obviously, S is a non-void subset of R3. Let u = (0, b1, c1), v = (0, b2, c2) ∈ Sand λ, µ∈ R. Then,


λu+ µv = λ(0, b1, c1)+ µ(0, b2, c2)

⇒ λu+ µν = (0, λb1 + µb2, λc1 + µc2) ∈ S.

Hence, S is a subspace of R3.

REMARK. Geometrically-R3 is three dimensional Euclidean plane and S is yz-plane whichis itself a vector space. Hence, S is a sub-space of R3. Similarly, {(a, b, 0) : a, b ∈ R},i.e. xy-plane, {(a,0,c) : a,c ∈ R}, i.e. xz plane are subspaces of R3. In fact, every planethrough the origin is a subspace of R3. Also, the coordinate axes {(a,0,0) : a ∈ R}, i.e. x-axis,{(0,a,0) : a ∈ R}, i.e. y-axis and {(0,0,a) : a ∈ R}, i.e. z-axis are subspaces of R3.

EXAMPLE-2 Let V denote the vector space R3, i.e. V = {(x,y,z) : x,y,z ∈ R} and S consistsof all vectors in R3 whose components are equal, i.e. S = {(x,x,x) : x ∈ R}. Show that S is asubspace of V .

SOLUTION Clearly, S is a non-empty subset of R3 as (0, 0, 0)∈ S.Let u = (x,x,x) and v = (y,y,y) ∈ S and a,b ∈ R. Then,

au+ bv = a(x,x,x)+ b(y,y,y)⇒ au+ av = (ax+ by,ax+ by,ax+ by) ∈ S.

Thus, au+ bv ∈ S for all u, v ∈ S and a,b ∈ R.Hence, S is a subspace of V .

REMARK. Geometrically S, in the above example, represents the line passing through theorigin O and having direction ratios proportional to 1,1,1 as shown in the following figure.

S

YO

X

Z


If there is no danger of any confusion we shall say V is a vector space over a field F ,whenever the algebraic structure (V, F, ⊕, �) is a vector space. Thus, whenever we say thatV is a vector space over a field F , it would always mean that (V, ⊕) is an abelian group and� : F ×V →V is a mapping such that V -2 (i)–(iv) are satisfied.

V is called a real vector space if F = R (field of real numbers), and a complex vector spaceif F = C (field of complex numbers).

REMARK-1 V is called a left or a right vector space according as the elements of a skew-fieldF are multiplied on the left or right of vectors in V . But, in case of a field these two conceptscoincide.

REMARK-2 The symbol ‘+’ has been used to denote the addition in the field F. For the sakeof convenience, in future, we shall use the same symbol ‘+’ for vector addition ⊕ and additionin the field F. But, the context would always make it clear as to which operation is meant.Similarly, multiplication in the field F and scalar multiplication � will be denoted by the samesymbol ‘·’.

REMARK-3 In this chapter and in future also, we will be dealing with two types of zeros. Onewill be the zero of the additive abelian group V , which will be known as the vector zero andother will be the zero element of the field F which will be known as scalar zero. We will use thesymbol 0V to denote the zero vector and 0 to denote the zero scalar.

REMARK-4 Since (V,⊕) is an abelian group, therefore, for any u,v,w ∈ V the followingproperties will hold:

(i) u⊕ v = u⊕w ⇒ v = w(ii) v⊕u = w⊕u ⇒ v = w

}

(Cancellation laws)

(iii) u⊕ v = 0v ⇒ u = −v and v = −u(iv) −(−u) = u(v) u⊕ v = u ⇒ v = 0v

(vi) −(u⊕ v) = (−u)⊕ (−v)(vii) 0v is unique

(viii) For each u ∈V, −u ∈V is unique.

REMARK-5 If V is a vector space over a field F, then we will write V (F).


EXAMPLE-1 Every field is a vector space over its any subfield.

SOLUTION Let F be a field, and let S be an arbitrary subfield of F . Since F is an additiveabelian group, therefore, V -1 holds.

Taking multiplication in F as scalar multiplication, the axioms V -2 (i) to V -2 (iii) arerespectively the right distributive law, left distributive law and associative law. Also V -2 (iv) isproperty of unity in F .


λu+ µv = λ(0, b1, c1)+ µ(0, b2, c2)

⇒ λu+ µν = (0, λb1 + µb2, λc1 + µc2) ∈ S.








S

YO

X

Z


EXAMPLE-3 Let a1,a2,a3 be fixed elements of a field F. Then the set S of all triads (x1,x2,x3)of elements of F, such that a1x1 + a2x2 + a3x3 = 0, is a subspace of F3.

SOLUTION Let u = (x1,x2,x3),v = (y1,y2,y3) be any two elements of S. Then x1,x2,x3,y1,y2,y3 ∈ F are such that

a1x1 + a2x2 + a3x3 = 0 (i)and, a1y1 + a2y2 + a3y3 = 0 (ii)

Let a,b be any two elements of F . Then,

au+ bv = a(x1,x2,x3)+ b(y1,y2,y3) = (ax1 + by1,ax2 + by2,ax3 + by3).

Now,

a1(ax1 + by1)+ a2(ax2 + by2)+ a3(ax3 + by3)

= a(a1x1 + a2x2 + a3x3)+ b(a1y1 + a2y2 + a3y3)

= a0+ b0 = 0 [From (i) and (ii)]

∴ au+ bv = (ax1 + by1,ax2 + by2,ax3 + by3) ∈ S.Thus, au+ bv ∈ S for all u,v ∈ S and all a,b ∈ F . Hence, S is a subspace of F3.

EXAMPLE-4 Show that the set S of all n× n symmetric matrices over a field F is a subspaceof the vector space Fn×n of all n×n matrices over field F.

SOLUTION Note that a square matrix A is symmetric, if AT = A.Obviously, S is a non-void subset of Fn×n, because the null matrix 0n×n is symmetric.Let A,B ∈ S and λ,µ∈ F . Then,

(λA + µB)T = λAT + µBT

⇒ (λA + µB)T = λA + µB [∵ A,B ∈ S ∴ AT = A,BT = B]

⇒ λA + µB ∈ S.

Thus, S is a non-void subset of Fn×n such that λA + µB ∈ S for all A,B ∈ S and for allλ,µ∈ F .

Hence, S is a subspace of Fn×n.

EXAMPLE-5 Let V be a vector space over a field F and let v ∈V . Then Fv = {av : a ∈ F} is asubspace of V .

SOLUTION Since 1 ∈ F , therefore v = 1 v ∈ Fv. Thus, Fv is a non-void subset of V .Let α,β ∈ Fv. Then, α = a1v, β = a2v for some a1,a2 ∈ F .


Let λ,µ∈ F . Then,

λα+ µβ = λ(a1v)+ µ(a2v)⇒ λα+ µβ = (λa1)v+(µa2)v [By V -2(iii)]⇒ λα+ µβ = (λa1 + µa2)v [By V -2(ii)]⇒ λα+ µβ ∈ Fv [∵ λa1 + µa2 ∈ F]

Thus, Fv is a non-void subset of V such that λα + µβ ∈ Fv for all α,β ∈ Fv and for allλ,µ∈ F .

Hence, Fv is a subspace of V .

EXAMPLE-6 Let AX = O be homogeneous system of linear equations, where A is an n× nmatrix over R. Let S be the solution set of this system of equations. Show S is a subspace of Rn.

SOLUTION Every solution U of AX = O may be viewed as a vector in Rn. So, the solution setof AX = O is a subset of Rn. Clearly, AX = O. So, the zero vector O ∈ S. Consequently, S isnon-empty subset of Rn.

Let X1,X2 ∈ S. Then,X1 and X2 are solutions of AX = O⇒ AX1 = O and AX2 = O (i)Let a, b be any two scalars in R. Then,

A(aX1 + bX2) = a(AX1)+ b(AX2) = aO + bO = O [Using(i)]⇒ aX1 + bX2 is a solution of AX = O⇒ aX1 + bX2 ∈ S.

Thus, aX1 + bX2 ∈ S for X1,X2 ∈ S and a,b ∈ R.

Hence, S is a subspace of Rn.

REMARK. The solution set of a non-homogeneous system AX = B of linear equations in nunknowns is not a subspace of Rn, because zero vector O does not belong to its solution set.

EXAMPLE-7 Let V be the vector space of all real valued continuous functions over the field Rof all real numbers. Show that the set S of solutions of the differential equation

2d2ydx2 −9

dydx

+ 2y = 0

is a subspace of V.

SOLUTION We have,

S =

{

y : 2d2ydx2 −9

dydx

+ 2y = 0}

, where y = f (x).


Now,

(aa1 + ba2)+ (ab1 + bb2)+ 2(ac1 + bc2)

= a(a1 + b1 + 2c1)+ b(a2 + b2 + 2c2)

= a×0+ b×0 = 0 [Using (i)]∴ au+ bv = (aa1 + ba2, ab1 + bb2, ac1 + bc2) ∈ S

Thus, au+ bv ∈ S for all u,v ∈ S and a,b ∈ R.Hence, S is a subspace of R3.

EXAMPLE-9 Let S be the set of all elements of the form (x+ 2y, y, −x+ 3y) in R3, wherex,y ∈ R. Show that S is a subspace of R3.

SOLUTION We have, S = {(x+ 2y, y, −x+ 3y) : x,y ∈ R}Let u,v be any two elements of S. Then,

u = (x1 + 2y1,y1,−x1 + 3y1),v = (x2 + 2y2,y2,−x2 + 3y2)

where x1,y1,x2,y2 ∈ R.Let a,b be any two elements of R. In order to prove that S is a subspace of V , we have

to prove that au + bv ∈ S, for which we have to show that au + bv is expressible in the form(α+ 2β,β,−α+ 3β).

Now,

au+ bv = a(x1 + 2y1,y1,−x1 + 3y1)+ b(x2 + 2y2,y2,−x2 + 3y2)

⇒ au+ bv =(

(ax1 +bx2)+ 2(ay1 + by2),ay1 + by2,−(ax1 +bx2) + 3(ay1 + by2))

⇒ au+ bv = (α+ 2β,β,−α+ 3β) , where α = ax1 +bx2 and β = ay1 + by2.

⇒ au+ bv ∈ S

Thus, au+ bv ∈ S for all u,v ∈ S and a,b ∈ R.Hence, S is a subspace of R3.

EXAMPLE-10 Let V be the vector space of all 2×2matrices over the field R of all real numbers.Show that:

(i) the set S of all 2×2 singular matrices over R is not a subspace of V .(ii) the set S of all 2×2 matrices A satisfying A2 = A is not a subspace of V .

SOLUTION (i) We observe that A =

[

0 a0 0

]

, B =

[

0 0b 0

]

, a �= 0 b �= 0 are elements of S.

But, A + B =

[

0 ab 0

]

is not an element of S as |A + B|= −ab �= 0.

So, S is not a subspace of V .


(ii) We observe that I =

[

1 00 1

]

∈ S as I2 = I. But, 2I =

[

2 00 2

]

∈/S, because

(2I)2 =

[

2 00 2

][

2 00 2

]

=

[

4 00 4

]

�= 2I.

So, S is not a subspace of V .

SMALLEST SUBSPACE CONTAINING A GIVEN SUBSET. Let V be a vector space over a fieldF, and let W be a subset of V . Then a subspace S of V is called the smallest subspace of Vcontaining W, if

(i) W ⊂ Sand (ii) S ′ is a subspace of V such that W ⊂ S ′ ⇒ S ⊂ S ′.The smallest subspace containing W is also called the subspace generated by W or subspacespanned by W and is denoted by [W ]. If W is a finite set, then S is called a finitely generatedspace.

In Example 5, Fv is a finitely generated subspace of V containing {v}.

EXERCISE 2.3

1. Show that the set of all upper (lower) triangular matrices over field C of all complexnumbers is a subspace of the vector space V of all n×n matrices over C.

2. Let V be the vector space of real-valued functions. Then, show that the set S of allcontinuous functions and set T of all differentiable functions are subspaces of V .

3. Let V be the vector space of all polynomials in indeterminate x over a field F and S bethe set of all polynomials of degree at most n. Show that S is a subspace of V .

4. Let V be the vector space of all n×n square matrices over a field F . Show that:

(i) the set S of all symmetric matrices over F is a subspace of V .(ii) the set S of all upper triangular matrices over F is a subspace of V .

(iii) the set S of all diagonal matrices over F is a subspace of V .(iv) the set S of all scalar matrices over F is a subspace of V .

5. Let V be the vector space of all functions from the real field R into R. Show that S is asubspace of V where S consists of all:

(i) bounded functions.(ii) even functions.

(iii) odd functions.

6. Let V be the vector space R3. Which of the following subsets of V are subspaces of V ?

(i) S1 = {(a,b,c) : a+ b = 0}(ii) S2 = {(a,b,c) : a = 2b+ 1}


12. Which of the following sets of vectors α = (a1,a2, . . . ,an) in Rn are subspaces of Rn?(n ≥ 3).

(i) all α such that a1 ≥ o

(ii) all α such that a1 + 3a2 = a3

(iii) all α such that a2 = a21

(iv) all α such that a1a2 = 0

(v) all α such that a2 is rational.

13. Let V be the vector space of all functions from R→ R over the field R of all real numbers.Show that each of the following subsets of V are subspaces of V .

(i) S1 = { f : f : R → R satisfies f (0) = f (1)}

(ii) S2 = { f : f : R → R is continuous }

(iii) S3 = { f : f : R → R satisfies f (−1) = 0}

14. Let V be the vector space of all functions from R to R over the field R of all real numbers.Show that each of the following subsets of V is not a subspace of V .

(i) S1 ={

f : f satisfies f (x2) = { f (x)}2 for all x ∈ R}

(ii) S2 = { f : f satisfies f (3)− f (−5) = 1}

15. Let Rn×n be the vector space of all n×n real matrices. Prove that the set S consisting ofall n×n real matrices which commutate with a given matrix A in Rn×n form a subspaceof Rn×n.

16. Let C be the field of all complex numbers and let n be a positive integer such that n ≥ 2.Let V be the vector space of all n× n matrices over C. Show that the following subsetsof V are not subspaces of V .(i) S = {A : A is invertible} (ii) S = {A : A is not invertible}

17. Prove that the set of vectors (x1,x2, . . . ,xn) ∈ R n satisfying the m equations is a subspaceof R n(R):

a11x1 + a12x2 + · · ·+ a1nxn = 0a21x1 + a22x2 + · · ·+ a2nxn = 0

......

...am1x1 + am2x2 + · · ·+ amnxn = 0,

aij ’s are fixed reals.


If S1 ⊂ S2, then S1 ∪S2 = S2, which is a subspace of V . Again, if S2 ⊂ S1, then S1 ∪S2 = S1which is also a subspace of V .

Hence, in either case S1 ∪S2 is a subspace of V .Conversely, suppose that S1 ∪ S2 is a subspace of V . Then we have to prove that either

S1 ⊂ S2 or S2 ⊂ S1.If possible, let S1 ⊂/ S2 and S2 ⊂/ S1. Then,

S1 ⊂/ S2 ⇒ There exists u ∈ S1 such that u ∈/S2

and, S2 ⊂/ S1 ⇒ There exists v ∈ S2 such that v ∈/S1.

Now,u ∈ S1,v ∈ S2

⇒ u,v ∈ S1 ∪S2

⇒ u+ v ∈ S1 ∪S2 [∵ S1 ∪S2 is a subspace of V ]⇒ u+ v ∈ S1 or, u+ v ∈ S2.

If u+ v ∈ S1

⇒ (u+ v)−u ∈ S1 [∵ u ∈ S1 and S1 is a subspace of V ]

⇒ v ∈ S1, which is a contradiction [∵ v ∈/S1]

If u+ v ∈ S2

⇒ (u+ v)− v ∈ S2 [∵ v ∈ S2 and S2 is a subspace of V ]

⇒ u ∈ S2, which is a contradiction [∵ u ∈/S2]

Since the contradictions arise by assuming that S1 ⊂/ S2 and S2 ⊂/ S1.Hence, either S1 ⊂ S2 or, S2 ⊂ S1.

Q.E.D.

REMARK. In general, the union of two subspaces of a vector space is not necessarily a sub-space. For example, S1 = {(a,0,0) : a ∈ R} and S2 = {(0,b,0) : b ∈ R} are subspaces ofR3. But, their union S1 ∪S2 is not a subspace of R 3, because (2,0,0),(0,−1,0) ∈ S1 ∪S2, but(2,0,0)+ (0,−1,0) = (2,−1,0) ∈/S1 ∪S2.

As remarked above the union of two subspaces is not always a subspace. Therefore, unionsare not very useful in the study of vector spaces and we shall say no more about them.

THEOREM-4 The inter section of a family of subspaces of V (F) containing a given subset Sof V is the (smallest subspace of V containing S) subspace generated by S.

PROOF. Let {Si : i ∈ I} be a family of subspaces of V (F) containing a subset S of V . LetT = ∩

i∈ISi. Then by Theorem 2, T is a subspace of V . Since Si ⊃ S for all i. Therefore,

T = ∩i∈I

Si ⊃ S.Hence, T is a subspace of V containing S.


Now it remains to show that T is the smallest subspace of V containing S.Let W be a subspace of V containing S. Then W is one of the members of the family

{Si : i ∈ I}. Consequently, W ⊃ ∩i∈ I

Si = T .Thus, every subspace of V that contains S also contains T . Hence, T is the smallest sub-

space of V containing S. Q.E.D.

THEOREM-5 Let A = {v1,v2, . . . ,vn} be a non-void finite subset of a vector space V (F).

Then the set S =

{

n∑

i=1aivi : ai ∈ F

}

is the subspace of V generated by A, i.e. S = [A].

PROOF. In order to prove that S is a subspace of V generated by A, it is sufficient to show thatS is the smallest subspace of V containing A.

We have, vi = 0 · v1 + 0v2 + · · ·+ 0vi−1 + 1 · vi + 0vi+1 + · · ·+ 0vnSo, vi ∈ S for all i ∈ n⇒ A ⊂ S.Let u =

n∑

i=1aivi, v =

n∑

i=1bivi ∈ S, and let λ,µ∈ F . Then,

λu+ µv = λ(

nΣ

i=1aivi

)

+ µ(

nΣ

i=1bivi

)

⇒ λu+ µv =nΣ

i=1(λai)vi +

nΣ

i=1(µbi)vi [By V -2(i) and V -2(iii)]

⇒ λu+ µv =nΣ

i=1(λai + µbi)vi ∈ S [∵ λai + µbi ∈ F for all i ∈ n]

Thus, λu+ µv ∈ S for all u,v ∈ S and for all λ,µ∈ F .So, S is a subspace of V containing A.Now, let W be a subspace of V containing A. Then, for each i ∈ n

ai ∈ F,vi ∈ A⇒ ai ∈ F,vi ∈W [∵ A ⊂W ]

⇒ aivi ∈W [∵ W is a subspace of V ]

⇒nΣ

i=1aivi ∈W

Thus, S ⊂W .Hence, S is the smallest subspace of V containing A. Q.E.D.

REMARK. If ϕ is the void set, then the smallest subspace of V containing ϕ is the null space{0V}. Therefore, the null space is generated by the void set.

SUM OF SUBSPACES. Let S and T be two subspaces of a vector space V (F). Then their sum(linear sum) S+ T is the set of all sums u+ v such that u ∈ S,v ∈ T .

Thus, S+ T = {u+ v : U ∈ S,v ∈ T}.


THEOREM-6 Let S and T be two subspaces of a vector space V (F). Then,

(i) S+ T is a subspace of V .(ii) S+ T is the smallest subspace of V containing S∪T , i.e. S+ T = [S∪T ].

PROOF. (i) Since S and T are subspaces of V . Therefore,

0V ∈ S, 0V ∈ T ⇒ 0V = 0V + 0V ∈ S+ T.

So, S+ T is a non-void subset of V .Let u = u1 + u2,v = v1 + v2 ∈ S+ T and let λ,µ∈ F . Then, u1,v1 ∈ S and u2,v2 ∈ T .Now,

λu+ µv = λ(u1 + u2)+ µ(v1 + v2)

⇒ λu+ µv = (λu1 + λu2)+ (µv1 + µv2) [By V -2(i)]⇒ λu+ µv = (λu1 + µv1)+ (λu2 + µv2) [Using comm. and assoc. of vector addition]

Since S and T are subspaces of V . Therefore,

u1,v1 ∈ S and λ,µ∈ F ⇒ λu1 + µv1 ∈ Sand, u2,v2 ∈ T and λ,µ∈ F ⇒ λu2 + µv2 ∈ T.

Thus,λu+ µv = (λu1 + µv1)+ (λu2 + µv2) ∈ S+ T.

Hence, S + T is a non-void subset of V such that λu + µv ∈ S + T for all u,v ∈ S + T andλ,µ∈ F . Consequently, S+ T is a subspace of V .

(ii) Let W be a subspace of V such that W ⊃ (S∪T ).Let u+ v be an arbitrary element of S+ T . Then u ∈ S and v ∈ T .Now,

u ∈ S,v ∈ T ⇒ u ∈ S∪T,v ∈ S∪T ⇒ u,v ∈W [∵ W ⊃ S∪T ]⇒ u+ v ∈W [∵ W is a subspace of V ]

Thus, u+ v ∈ S+ T ⇒ u+ v ∈W

⇒ S+ T ⊂W .Obviously, S∪T ⊂ S+ T .Thus, if W is a subspace of V containing S∪ T , then it contains S + T . Hence, S + T is thesmallest subspace of V containing S∪T . Q.E.D.

DIRECT SUM OF SUBSPACES. A vector space V (F) is said to be direct sum of its two subspacesS and T if every vector u in V is expressible in one and only one as u = v+w, where v ∈ S andw ∈ T .

If V is direct sum of S and T , then we write V = S⊕T .


EXAMPLE-4 Consider the vector space V = R3. Let S = {(x,y,0) : x,y ∈ R} andT = {(x,0,z) : x,z ∈ R}. Prove that S and T are subspaces of V such that V = S + T , butV is not the direct sum of S and T .

SOLUTION Clearly, S and T represent xy-plane and xz-plane respectively. It can be easilychecked that S and T are subspaces of V .

Let v = (x,y,z) be an arbitrary point in V . Then,v = (x,y,z) = (x,y,0)+ (0,0,z)

⇒ v ∈ S+ T [∵ (x,y,0) ∈ S and (0,0,z) ∈ T ]∴ V ⊂ S+ TAlso, S+ T ⊂V.

Hence, V = S+ T i.e. R3 is a sum of the xy-plane and the yz-plane.

Clearly, S∩T = {(x,0,0) : x ∈ R} is the x-axis.∴ S∩T �= {(0,0,0)}.

So, V is not the direct sum of S and T .

COMPLEMENT OF A SUBSPACE. Let V (F) be a vector space and let S be a subspace of V . Thena subspace T of V is said to be complement of S, if V = S⊕T.

INDEPENDENT SUBSPACES. Let S1,S2, . . . ,Sn be subspaces of a vector space V.Then S1,S2, . . . ,Snare said to be independent, if

Si ∩nΣj=1j �=i

S j = {0V} for all i = 1,2, . . . ,n.

THEOREM-8 Let S1,S2, . . . ,Sn be subspaces of a vector space V (F) such thatV = S1 + S2 + · · ·+ Sn. Then, the following are equivalent:

(i) V = S1 ⊕S2 ⊕·· ·⊕Sn.(ii) S1,S2, . . . ,Sn are independent.

(iii) For each vi ∈ Si,n∑

i=1vi = 0V ⇒ vi = 0V .

PROOF. Let us first prove that (i) implies (ii).

Let u ∈ Si ∩n∑j=1j �=i

S j. Then there exist v1 ∈ S1,v2 ∈ S2, . . . ,vi ∈ Si, . . . ,vn ∈ Sn such that

u = vi = v1 + v2 + · · ·+ vi−1 + vi+1 + · · ·+ vn

⇒ u = v1 + v2 + · · ·+ vi−1 + 0V + vi+1 + · · ·+ vn = 0v1 + 0v2 + · · ·+ 1vi + · · ·+ 0vn

⇒ v1 = 0V = v2 = · · · = vi = · · · = vn [∵ V = S1 ⊕S2 ⊕·· ·⊕Sn]⇒ u = 0V


Thus, u ∈ Si ∩n∑j=1j �=i

S j ⇒ u = 0V .

Hence, Si ∩n∑j=1j �=i

S j = 0V . Consequently, S1,S2, . . . ,Sn are independent.

(ii) ⇒(iii).Let v1 + v2 + · · ·+ vn = 0V with vi ∈ Si, i ∈ n. Then,

⇒ vi = −(v1 + v2 + · · ·+ vi−1 + vi+1 + · · ·+ vn)

⇒ vi = −nΣj=1j �=i

v j.

Since vi ∈ Si and −n∑j=1j �=i

v j ∈n∑j=1j �=i

S j. Therefore,

vi ∈ Si ∩nΣj=1j �=i

S j for all i ∈ n.

⇒ vi = 0V for all i ∈ n [∵ S1,S2, . . . ,Sn are independent]

(iii) ⇒ (i)Since V = S1 +S2 + · · ·+Sn is given. Therefore, it is sufficient to show that each v ∈V has

a unique representation as a sum of vectors in S1,S2, . . . ,Sn.Let, if possible, v = v1 + · · ·+ vn and v = w1 + · · ·+ wn be two representations for v ∈ V .

Then,

v1 + v2 + · · ·+ vn = w1 + w2 + · · ·+ wn

⇒ (v1 −w1)+ · · ·+(vn −wn) = 0V

⇒ v1 −w1 = 0V , . . . ,vn −wn = 0V [Using (iii)]⇒ v1 = w1, . . . ,vn = wn.

Thus, each v ∈V has a unique representation. Hence, V = S1 ⊕S2 ⊕·· ·⊕Sn.

Consequently, (iii) ⇒(i).Hence, (i) ⇒ (ii) ⇒ (iii) ⇒ (i) i.e. (i) ⇔ (ii) ⇔ (iii)

Q.E.D.

LINEAR VARIETY. Let S be a subspace of a vector space V (F), and let v ∈V . Then,

v+ S = {v+ u : u ∈ S}

is called a linear variety of S by v or a translate of S by v or a parallel of S through v.S is called the base space of the linear variety and v a leader.


THEOREM-9 Let S be a subspace of a vector space V (F) and let P = v+ S be the parallel ofS through v. Then,

(i) every vector in P can be taken as a leader of P, i.e u+ S = P for all u ∈ P.

(ii) two vectors v1,v2 ∈V are in the same parallel of S iff v1 − v2 ∈ S

PROOF. (i) Let u be an arbitrary vector in P = v+ S. Then there exists v1 ∈ S such that

u = v+ v1. ⇒ v = u− v1.

Let w be an arbitrary vector in P. Then,

w = v+ v2 for some v2 ∈ S.

⇒ w = (u− v1)+ v2 [∵ v = u− v1]

⇒ w = u+(v2 − v1) [∵ v2 − v1 ∈ S]

⇒ w ∈ u+ S

Thus, P ⊂ u+ S.

Now let w′ ∈ u+ S. Then,

w′ = u+ v3 for some v3 ∈ S.

⇒ w′ = (v+ v1)+ v3 [∵ v = u− v1]

⇒ w′ = v+(v1 + v3)

⇒ w′ ∈ v+ S [∵ v1 + v3 ∈ S]

Thus, u+ S ⊂ P. Hence, u+ S = PSince u is an arbitrary vector in P. Therefore, u+ S = P for all u ∈ P.

(ii) Let v1,v2 be in the same parallel of S, say u+ S. Then there exist u1,u2 ∈ S such that

v1 = v+ u1,v2 = v+ u2

⇒ v1 − v2 = u1 −u2 ⇒ v1 − v2 ∈ S [∵ u1,u2 ∈ S ⇒ u1 −u2 ∈ S]

Conversely, if v1 − v2 ∈ S, then there exists u ∈ S such that

v1 − v2 = u⇒ v1 = v2 + u⇒ v1 ∈ v2 + S. [∵ u ∈ S]

Also, v2 = v2 + 0V where 0V ∈ S⇒ v2 ∈ v2 + S.

Hence, v1,v2 are in the same parallel of S.Q.E.D.


EXERCISE 2.4

1. Let V = R3, S = {(0, y, z) : y, z ∈ R} and T = {(x, 0, z) : x, z ∈ R}Show that S and T are subspaces of V such that V = S + T . But, V is not the direct sumof S and T .

2. Let V (R) be the vector space of all 2×2 matrices over R. Let S be the set of all symmetricmatrices in V and T be the set of all skew-symmetric matrices in V . Show that S and Tare subspaces of V such that V = S⊕T .

3. Let V = R3, S = {(x, y, 0) : x, y ∈ R} and T = {(0,0,z) : z ∈ R}.Show that S and T are subspaces of V such that V = S⊕T .

4. Let S1, S2, S3 be the following subspaces of vector space V = R3:

S1 = {(x, y, z) : x= z, x, y, z ∈ R}S2 = {(x, y, z) : x+ y+ z = 0, x, y, z ∈ R}S3 = {(0, 0, z) : z ∈ R}

Show that: (i) V = S1 +S2 (ii) V = S2 +S3 (iii) V = S1 +S3 When is the sum direct?5. Let S and T be the two-dimensional subspaces of R3. Show that S∩T �= {0}.6. Let S, S1, S2 be the subspaces of a vector space V (F). Show that

(S∩S1)+ (S∩S2) ⊆ S∩ (S1 + S2)

Find the subspaces of R2 for which equality does not hold.7. Let V be the vector space of n × n square matrices. Let S be the subspace of upper

triangular matrices, and let T be the subspaces of lower triangular matrices.Find (i) S∩T (ii) S+ T .

8. Let F be a subfield of the fieldC of complex numbers and let V be the vector space of all

2×2 matrices over F . Let S =

{[

x yz 0

]

: x, y, z ∈ F}

and T =

{[

x 00 y

]

: x, y ∈ F}

.

Show that S and T are subspaces of V such that V = S+ T but V �= S⊕T .[

Hint: S∩T =

{[

x 00 0

]

: x ∈ F}

∴ S∩T �=

{[

0 00 0

]}]

9. Let S and T be subspaces of a vector space V (F) such that V = S+T and S∩T = {0V}.Prove that each vector v in V there are unique vectors v1 in S and v2 in T such thatv = v1 + v2.

10. Let U and W be vector spaces over a field F and let V = U ×W = {(u, w) : u ∈U andw∈W}. Then V is a vector space over field F with addition in V and scalar multiplicationon V defined by

(u1, w1)+ (u2, w2) = (u1 + u2, w1 + w2)

and, k(u, w) = (ku, kw).Let S = {(u, 0) : u ∈U} and T = {(0, w) : w ∈W}. Show that


Since S is a subgroup of additive abelian group V . Therefore, for any u+ S,v+ S ∈V/S

u+ S = v+ S ⇔ u− v ∈ S

Thus, we have defined vector addition, scalar multiplication on V/S and equality of any twoelements of V/S. Now we proceed to prove that V/S is a vector space over field F under theabove defined operations of addition and scalar multiplication.

THEOREM-1 Let V be a vector space over a field F and let S be a subspace of V . Then, the set

V/S = {u+ S : u ∈V}

is a vector space over field F for the vector addition and scalar multiplication on V/S definedas follows:

(u+ S)+ (v+ S) = (u+ v)+ Sand, a(u+ S) = au+ S for all u+ S, v+ S ∈V/S and for all a ∈ F.

PROOF. First we shall show that above rules for vector addition and scalar multiplication are

well defined, i.e. are independent of the particular representative chosen to denote a coset.

Let u+ S = u ′ + S,where u,u ′ ∈V and, v+ S = v ′ + S,where v,v ′ ∈V.

Then,

u+ S = u ′ + S ⇒ u−u ′ ∈ S and, v+ S = v ′ + S ⇒ v− v ′ ∈ S.

Now,

u−u ′ ∈ S,v− v ′ ∈ S⇒ (u−u ′)+ (v− v ′) ∈ S [∵ S is a subspace of V ]⇒ (u+ v)− (u ′ + v ′) ∈ S⇒ (u+ v)+ S = (u ′ + v ′)+ S⇒ (u+ S)+ (v+ S) = (u ′ + S)(v ′ + S)

Thus, u+ S = u ′ + S and v+ S = v ′ + S⇒ (u+ S)+ (v+ S) = (u ′ + S)+ (v ′+ S) for all u, u ′, v, v ′ ∈V .

Therefore, vector addition on V/S is well defined.Again,

λ ∈ F, u−u′ ∈ S ⇒ λ(u−u′) ∈ S ⇒ λu−λu′ ∈ S ⇒ λu+ S = λu′ + S


Therefore, scalar multiplication on V/S is well defined.V −1 : V/S is an abelian group under vector addition:

Associativity: For any u+ S,v+ S,w + S ∈V/S, we have

[(u+ S)+ (v+ S)]+ (w + S) = ((u+ v)+ S)+ (w + S)

= {(u+ v)+ w}+ S= {u+(v+ w)}+ S [By associativity of vector addition on V ]= (u+ S)+ ((v+ w)+ S)

= (u+ S)+ [(v+ S)+ (w + S)]

So, vector addition is associative on V/S.

Commutativity: For any u+ S,v+ S ∈V/S, we have

(u+ S)+ (v+ S) = (u+ v)+ S= (v+ u)+ S [By commutativity of vector addition on V ]= (v+ S)+ (u+ S)

So, vector addition is commutative on V/S.

Existence of additive identity: Since 0V ∈V therefore, 0V + S = S ∈V/S.Now,

(u+ S)+ (0V + S) = (u+ 0V )+ S = (u+ S) for all u+ S ∈V/S∴ (u+ S)+ (0V + S) = (u+ S) = (0V + S)+ (u+ S) for all u+ S ∈V/S.

So 0V + S = S is the identity element for vector addition on V/S.

Existence of additive inverse: Let u+ S be an arbitrary element of V/S. Then,u ∈V ⇒ −u ∈V ⇒ (−u)+ S ∈V/S

Thus, for each u+ S ∈V/S there exists (−u)+ S ∈V/S such that

(u+ S)+ ((−u)+ S) = [u+(−u)]+ S = 0V + S = S⇒ (u+ S)+ ((−u)+ S)=0V + S=(−u)+ S+(u+ S) [By commutativity of addition on V/S]

So, each u+ S ∈V/S has its additive inverse (−u)+ S ∈V/S.Hence, V/S is an abelian group under vector addition.


V −2 : For any u+ S, v+ S ∈V/S and λ,µ∈ F, we have

(i) λ[(u+ S)+ (v+ S)] = λ[(u+ v)+ S]

= [λ(u+ v)]+ S= (λu+ λv)+ S [By V -2(i) in V]= (λu+ S)+ (λv+ S) [By definition of addition on V/S]= λ(u+ S)+ λ(v+ S) [By def. of scalar multiplication on V/S]

(ii) (λ+ µ)(u+ S) = (λ+ µ)u+ S [By def. of scalar multiplication on V/S]= (λu+ µu)+ S [By V -2(ii) in V]= (λu+ S)+ (µu+ S) [By definition of addition on V/S]= λ(u+ S)+ µ(u+ S)

(iii) (λµ)(u+ S) = (λµ)u+ S= λ(µu)+ S [By V -2(iii) in V]= λ(µu+ S)

= λ[µ(u+ S)]

(iv) 1(u+ S) = 1u+ S = u+ S [By V -2(iv) in V]

Hence, V/S is a vector space over field F .Q.E.D.

QUOTIENT SPACE. Let V be a vector space over a field F and let S be a subspace of V . Thenthe set V/S = {u+ S : u ∈V} is a vector space over field F for the vector addition and scalarmultiplication defined as

(u+ S)+ (v+ S) = (u+ v)+ Sa(u+ S) = au+ S

For all u+ S,v+ S ∈V/S and for all a ∈ F.This vector space is called a quotient space of V by S.

EXERCISE 2.5

1. If S is a subspace of a vector space V (F), then show that there is one-one correspondencebetween the subspaces of V containing S and subspaces of the quotient space V/S.

2. Define quotient space.


2.6 LINEAR COMBINATIONS

LINEAR COMBINATION. Let V be a vector space over a field F. Let v1,v2, . . . ,vn be n vectorsin V and let λ1,λ2, . . . ,λn be n scalars in F. Then the vector λ1v1 + · · ·+ λnvn(or

n∑

i=1λivi) is

called a linear combination of v1,v2, . . . ,vn. It is also called a linear combination of the setS = {v1,v2, . . . ,vn}. Since there are finite number of vectors in S, it is also called a finite linearcombination of S.

If S is an infinite subset of V , then a linear combination of a finite subset of S is called afinite linear combination of S.


EXAMPLE-1 Express v = (−2,3) in R2(R) as a linear combination of the vectors v1 = (1,1)and v2 = (1,2).

SOLUTION Let x,y be scalars such that

v = xv1 + yv2

⇒ (−2,3) = x(1,1)+ y(1,2)

⇒ (−2,3) = (x+ y,x+ 2y)⇒ x+ y = −2 and x+ 2y = 3⇒ x= −7,y = 5

Hence, v = −7v1 + 5v2.

EXAMPLE-2 Express v= (−2,5) in R2(R) as a linear combination of the vectors v1 = (−1,1)and v2 = (2,−2).

SOLUTION Let x,y be scalars such that

v = xv1 + yv2

⇒ (−2,5) = x(−1,1)+ y(2,−2)

⇒ (−2,5) = (−x+ 2y,x−2y)⇒ −x+ 2y = −2 and x−2y = 5

This is an inconsistent system of equations and so it has no solution.Hence, v cannot be written as the linear combination of v1 and v2.

EXAMPLE-3 Express v= (1,−2,5) in R3 as a linear combination of the following vectors:

v1 = (1,1,1),v2 = (1,2,3),v3 = (2,−1,1)


This is equivalent to

1 2 10 2 −20 −5 5

xyz

=

21

−1

Applying R2 → R2 + 3R1,R3 → R3 +(−2)R1

or,

1 2 10 2 −20 0 0

xyz

=

213/2

Applying R3 → R3 +52

R2

This is an inconsistent system of equations and so has no solution. Hence, v cannot be writtenas a linear combination of v1,v2 and v3.

EXAMPLE-5 Express the polynomial f (x) = x2 +4x−3 in the vector space V of all polynomialsover R as a linear combination of the polynomials g(x) = x2 − 2x+ 5,h(x) = 2x2 − 3x andφ(x) = x+ 3.

SOLUTION Let α,β,γ be scalars such thatf (x) = u g(x)+ v h(x)+ w φ(x) for all x ∈ R

⇒ x2 + 4x−3 = u(x2 −2x+ 5)+ v(2x2 −3x)+ w(x+ 3 for all x ∈ R⇒ x2 + 4x−3 = (u+ 2v)x2 +(−2u−3v+ w)x+(5u+ 3w) for all x ∈ R (i)⇒ u+ 2v = 1,−2u−3v+ w = 4,5u+ w = −3 (ii)The matrix form of this system of equations is

1 2 0−2 −3 1

5 0 3

uvw

=

14

−3

This system of equations is equivalent to

1 2 00 1 10 −10 3

uvw

=

16

−8

Applying R2 → R2 + 2R1, R3 → R3 −5R1

or,

1 2 00 1 10 0 13

uvw

=

1652

Applying R3 → R3 + 10R2

Thus, the system of equations obtained in (i) is consistent and is equivalent tou+ 2v = 1,v+ w = 6 and 13w = 52 ⇒ u = −3,v = 2,w = 4

Hence, f (x) = −3g(x)+ 2h(x)+ 4φ(x).

REMARK. The equation (i) obtained in the above solution is an identity in x, that is it holds forany value of x. So, the values of u,v and w can be obtained by solving three equations whichcan be obtained by given any three values to variable x.


SOLUTION Let x,y be scalars such thatu = xv1 + yv2

⇒ (1,k,4) = x(1,2,3)+ y(2,3,1)

⇒ (1,k,4) = (x+ 2y,2x+ 3y,3x+ y)⇒ x+ 2y = 1,2x+ 3y = k and 3x+ y = 4

Solving x+ 2y = 1 and 3x+ y = 4, we get x=75

and y = −15

Substituting these values in 2x+ 3y = k, we get k =115

EXAMPLE-8 Consider the vectors v1 = (1,2,3) and v2 = (2,3,1) in R3(R). Find conditions ona,b,c so that u = (a,b,c) is a linear combination of v1 and v2.

SOLUTION Let x,y be scalars in R such thatu = xv1 + yv2

⇒ (a,b,c) = x(1,2,3)+ y(2,3,1)

⇒ (a,b,c) = (x+ 2y,2x+ 3y,3x+ y)⇒ x+ 2y = a,2x+ 3y = b and 3x+ y = cSolving first two equations, we get x= −3a+ 2b and y = 2a−b.

Substituting these values in 3x+ y = c, we get3(−3a+ 2b)+ (2a−b) = c or, 7a−5b+ c = 0 as the required condition.

EXERCISE 2.61. Express v = (3,−2) in R2(R) as a linear combination of the vectors v1 = (−1,1) and

v2 = (2,1).2. Express v = (−2,5) in R2(R) as a linear combination of the vectors v1 = (2,−1) and

v2 = (−4,2).3. Express v = (3,7,−4) in R3(R) as a linear combination of the vectors v1 = (1,2,3),

v2 = (2,3,7) and v3 = (3,5,6).4. Express v = (2,−1,3) in R3(R) as a linear combination of the vectors v1 = (3,0,3),

v2 = (−1,2,−5) and v3 = (−2,−1,0).5. Let V be the vector space of all real polynomials over field R of all real numbers. Ex-

press the polynomial p(x) = 3x2 + 5x− 5 as a linear combination of the polynomialsf (x) = x2 + 2x+ 1,g(x) = 2x2 + 5x+ 4 and h(x) = x2 + 3x+ 6.

6. Let V = P2(t) be the vector space of all polynomials of degree less than or equal to 2 andt be the indeterminate. Write the polynomial f (t) = at2 + bt + c as a linear combinationof the polynomials p1(t) = (t −1)2, p2(t) = t −1 and p3(t) = 1.

7. Write the vectors u = (1,3,8) in R3(R) as a linear combination of the vectors v1 = (1,2,3)and v2 = (2,3,1).


8. Write the vector u = (2,4,5) in R3(R) as a linear combination of the vectors v1 = (1,2,3)and v2 = (2,3,1).

9. Consider the vector space R2×2 of all 2× 2 matrices over the field R of real numbers.

Express the matrix M =

[

5 −67 8

]

as a linear combination of the matrices

E11 =

[

1 00 0

]

,E12 =

[

0 10 0

]

,E21 =

[

0 01 0

]

and E22 =

[

0 00 1

]

ANSWERS1. v =

−73

v1 +13

v2 2. v is not expressible as a linear combination of v1 and v2.

3. v = 2v1 −4v2 + 3v3 4. v is not expressible as a linear combination of v1,v2,v3.

5. p(x) = 3 f (x)+ g(x)−2h(x) 6. f (t) = ap1(t)+ (2a+ b)p2(t)+ (a+ b+ c)p3(t)

7. u = 3v1 − v2 8. Not Possible 9. M = 5E11 −6E12 + 7E21 + 8E22

2.7 LINEAR SPANS

LINEAR SPAN OF A SET. Let V be a vector space over a field Fand let S be a subset of V . Thenthe set of all finite linear combinations of S is called linear span of S and is denoted by [S].

Thus,

[S] =

{

n∑

i=1λivi : λ1,λ2, . . . ,λn ∈ F, n ∈ N and v1,v2, . . . ,vn ∈ S

}

.

If S is a finite set, say S = {v1,v2, . . . ,vn}, then

[S] =

{

nΣ

i=1λivi : λi ∈ F

}

.

In this case [S] is also written as [v1,v2, . . . ,vn].If S = {v}, then [S] = Fv as shown in Example 5 on page 136.Consider a subset S = {(1,0,0),(0,1,0)} of real vector space R3. Any linear combination

of a finite number of elements of S is of the form λ(1,0,0)+µ(0,1,0) = (λ,µ,0). By definition,the set of all such linear combinations is [S]. Thus, [S] = {(λ,µ,0) : λ,µ∈ R}. Clearly [S] isxy-plane in three dimensional Euclidean space R3 and is a subspace of R3. In fact, this is truefor every non-void subset of a vector space. The following theorem proves this result.

THEOREM-1 Let S be a non-void subset of a vector space V (F). Then [S], the linear span ofS, is a subspace of V .

PROOF. Since 0V = 0v1 +0v2 + · · ·+0vn for every positive integer n. Therefore, 0V is a finitelinear combination of S and so it is in [S].

Thus, [S] is a non-void subset of V .


Let u and v be any two vectors in [S]. Then,u = λ1u1 +λ2u2 + · · ·+λnun for some scalars λi ∈ F , some u’s

i ∈ S, and a positive integer n.and, v = µ1v1 + µ2v2 + · · ·+ µmvm for some µ’s

i ∈ F , some v’si ∈ S, and a positive integer m.

If λ,µ are any two scalars in F , then

λu+ µv = λ(λ1u1 + λ2u2 + · · ·+ λnun)+ µ(µ1v1 + · · ·+ µmvm)

⇒ λu+ µv = (λλ1)u1 + · · ·+(λλn)un +(µµ1)v1 + · · ·+(µµm)vm

Clearly, λu+ µv is a finite linear combination of vectors in S and so it is in [S].Thus, λu+ µv ∈ [S] for all u,v ∈ [S] and for all λ,µ∈ F .Hence, [S] is a subspace of V .

Q.E.D.

THEOREM-2 Let S be a subspace of a vector space V (F). Then [S] is the smallest subspace ofV containing S.

PROOF. By Theorem 1, [S] is a subspace of V . Let v be an arbitrary vector in S. Then,

v = 1v [By V -2(iv)]⇒ v is a finite linear combination of S.⇒ v ∈ [S].

Thus, S ⊂ [S].

Now it remains to prove that [S] is the smallest subspace of V containing S. For this, weshall show that if there exists another subspace T containing S, then it also contains [S]. So letT be a subspace of V containing S. Let u be an arbitrary vector in [S]. Then,

u = λ1v1 + · · ·+ λnvn for some λ’si ∈ F , some v’s

i ∈ S and a positive integer n.Since S ⊂ T and vi ∈ S. Therefore,

vi ∈ T for each i ∈ n⇒ λ1v1 + λ2v2 + · · ·+ λnvn ∈ T [∵ T is a subspace of V ]⇒ u ∈ T

Since u is an arbitrary vector in S. Therefore,

u ∈ T for all u ∈ [S].

⇒ [S] ⊂ T.

Hence, [S] is the smallest subspace of V containing S.Q.E.D.

REMARK-1 Since the null space {0V} is the smallest subspace of V containing the void set ϕ,Therefore, by convention, we take [ϕ] = {0V}.


REMARK-2 It is clear from the foregoing discussion that if S is a non-void subset of a vectorspace V (F), then S itself is smaller than [S]. In fact, [S] �= {0V} always contains an infinitenumber of vectors whatever be the number of vectors in S.

THEOREM-3 Let S = {v1,v2, . . . ,vn} be a finite subset of a vector space V (F). Then for any0 �= a ∈ F

(i) [v1,v2, . . . ,vk, . . . ,vn] = [v1,v2, . . . ,avk, . . . ,vn](ii) [v1,v2, . . . ,vk, . . . ,vn] = [v1 + avk,v2, . . . ,vk, . . . ,vn]

(iii) If v =n∑

i=1aivi, where ai ∈ F for all i ∈ n, then [v1,v2, . . . ,vn] = [v1,v2, . . . ,vn,v].

PROOF. (i) Let u∈ [v1,v2, . . . ,vk, . . . ,vn]. Then u is a linear combination of vectors v1,v2, . . . ,vk, . . . ,vn.Consequently, there exist scalars λ1,λ2, . . . ,λn in F such that

u = λ1v1 + · · ·+ λkvk + · · ·+ λnvn

⇒ u = λ1v1 + · · ·+ λka−1(avk)+ · · ·+ λnvn [∵ 0 �= a ∈ F ∴ a−1 ∈ F]⇒ u is a linear combination of v1,v2, . . . ,avk, . . . ,vn

⇒ u ∈ [v1,v2, . . . ,avk, . . . ,vn]

Thus, u ∈ [v1,v2, . . . ,vk, . . . ,vn] ⇒ u ∈ [v1,v2, . . . ,avk, . . . ,vn]

∴ [v1,v2, . . . ,vk, . . . ,vn]⊂ [v1,v2, . . . ,avk, . . . ,vn] (i)Now let u be an arbitrary vector in [v1,v2, . . . ,avk, . . . ,vn]. Then there exist scalarµ1,µ2, . . . ,µk, . . . ,µn in F such that

u = µ1v1 + µ2v2 + · · ·+ µk(avk)+ · · ·+ µnvn

⇒ u = µ1v1 + µ2v2 + · · ·+(µka)vk + · · ·+ µnvn

⇒ u is a linear combination of v1,v2, . . . ,vk, . . . ,vn

⇒ u ∈ [v1,v2, . . . ,vk, . . . ,vn]

Thus, u ∈ [v1,v2, . . . ,avk, . . . ,vn] ⇒ u ∈ [v1,v2, . . . ,vk, . . . ,vn]

∴ [v1,v2, . . . ,avk, . . . ,vn]⊂ [v1,v2, . . . ,vk, . . . ,vn] (ii)From (i) and (ii), we obtain

[v1,v2, . . . ,vk, . . . ,vn] = [v1,v2 . . . ,avk, . . . ,vn]

(ii) Let u be an arbitrary vector in [v1,v2, . . . ,vk, . . . ,vn]. Then, there exist scalars λ1,λ2, . . . ,λn,such that

u =n

∑i=1

λivi

⇒ u = λ1(v1 + avk)+ λ2v2 + · · ·+(λk −aλ1)vk + · · ·+ λnvn

⇒ u is a linear combination of v1 + avk,v2, . . . ,vk, . . . ,vn

⇒ u ∈ [v1 + avk,v2, . . . ,vk, . . . ,vn]


(iii) We know that [S] is a subspace of V . Therefore, by (ii) it follows that[

[S]]

= [S].(iv) Let u be an arbitrary vector in [S∪T ]. Then,

u = λ1u1 +λ2u2 + · · ·+λmum +µ1v1 + · · ·+µnvn for some u’si ∈ S, v’s

i ∈ T, λ’si , µ’s

i ∈ F ,and positive integers m and n.

Now, u’si ∈ S and λ’s

i ∈ F ⇒ λ1u1 + · · ·+ λmum ∈ [S]

and, v’si ∈ T and µ’s

i ∈ F ⇒ µ1v1 + · · ·+ µnvn ∈ [T ].

∴ λ1u1 + · · ·+ λmum + µ1v1 + · · ·+ µnvn ∈ [S]+ [T ]

⇒ u ∈ [S]+ [T ]

Thus, u ∈ [S∪T ] ⇒ u ∈ [S]+ [T ]. Consequently,[S∪T ] ⊂ [S]+ [T ].

Now, let v be an arbitrary vector in [S]+ [T ]. Then,

v = u+ w for some u ∈ [S], w ∈ [T ].

Now, u ∈ [S] ⇒ u is a finite linear combination of S.

and, w ∈ [T ] ⇒ w is a finite linear combination of T.

Therefore, v = u+ w is a finite linear combination of S∪T and so it is in [S∪T ].

Thus, v ∈ [S]+ [T ] ⇒ v ∈ [S∪T ]

Consequently, [S]+ [T ] ⊂ [S∪T ]

Hence, [S∪T ] = [S]+ [T ]

Q.E.D.

2.7.1 ROW AND COLUMN SPACES OF A MATRIX

Let A = [ai j] be an arbitrary m×n matrix over a field F . The rows of A areR1 = (a11, a12, a13, . . .a1n), R2 = (a21, a22, a23, . . .a2n), . . . ,Rm = (am1, am2, am3, . . . , amn)

These rows may be viewed as vectors in Fn. So, they span a subspace of Fn, which is knownas the row space of matrix A and is denoted by rowsp(A). That is,

rowsp(A) = [R1, R2, . . . ,Rm]

Thus, row space of a matrix is the subspace of Fn spanned the rows of the matrix.Similarly, the column space of A is the subspace of Fm spanned by the columnsC1 = (a11, a21, . . . ,am1), C2 = (a12, a22, . . . ,am2) . . . , Cn = (a1n, a2n, . . . , amn) and is

denoted by colsp(A).Clearly, colsp(A) = rowsp(AT ).

THEOREM-1 Row equivalent matrices have the same row space.

PROOF. Let A be a matrix and B be the matrix obtained by applying one of the followingrow operations on matrix A:


(i) Interchange Ri and R j (ii) Replace Ri by kRi (iii) Replace R j by R j + kRi

Then each row of B is a row of A or a linear combination of rows of A. So, the row space ofB is contained in the row space of A. On the other hand, we can apply the inverse elementaryrow operation on B to obtain A. So, the row space of A is contained in the row space of B.Consequently, A and B have the same row space.

Q.E.D.

THEOREM-2 Let A and B be row canonical matrices. Then A and B have the same row spaceif and only if they have the same non-zero rows.

EXAMPLE Let v1 = (1,2,−1,3),v2 = (2,4,1,−2),v3 = (3,6,3,−7),v4 = (1,2,−4,11) andv5 = (2,4,−5,14) be vectors in R4(R) such that S = {v1,v2,v3} and T = {v4,v5}. Show that[S] = [T ].

SOLUTION There are two ways to show that [S] = [T ]. Show that each of v1,v2,v3 is a linearcombination of v3 and v4 and show that each of v4 and v5 is a linear combination of v1,v2,v3.But, this method is not very convenient. So, let us discuss an alternative method. Let A be thematrix whose rows are v1,v2,v3 and B be the matrix whose rows v4 and v5. That is,

A =

1 2 −1 32 4 1 −23 6 3 −7

⇒ A ∼

1 2 −1 30 0 3 −80 0 6 −16

Applying R2 → R2 −2R1,R3 → R3 −3R1

⇒ A ∼

1 2 −1 30 0 3 −80 0 0 0

Applying R3 → R3 −2R2

B =

[

1 2 −4 112 4 −5 14

]

⇒ B ∼

[

1 2 −4 110 0 3 −8

]


⇒ B ∼

[

1 2 −1 30 0 3 −8

]

Applying R1 → R1 + R2

Clearly, non-zero rows of the matrices in row canonical form are identical. Therefore,rowsp(A) = rowsp(B).

Hence, [S] = [T ]



EXAMPLE-1 Prove that any vector v = (a,b) in R2(R) is a linear combination of v1 = (1,1)and v2 = (−1,1).

SOLUTION Let v = xv1 + yv2 for some x,y ∈ R. Then,

(a,b) = x(1,1)+ y(−1,1)

⇒ (a,b) = (x− y,x+ y)⇒ x− y = a and x+ y = b

Clearly, this system of equations is consistent with unique solution as the determinant of the

coefficient matrix is non-zero. The values of x and y area+ b

2and

a−b2

respectively.

Hence, every vector in R2 is a linear combination of v1 and v2.

EXAMPLE-2 Show that the vectors v1 = (1,1) and v2 = (1,2) span R2(R).

SOLUTION In order to show that vectors v1 and v2 span R2(R), it is sufficient to show that anyvector in R2 is a linear combination of v1 and v2. Let v = (a,b) be an arbitrary vector in R2(R).Further, let

v = xv1 + yv2

⇒ (a,b) = x(1,1)+ y(1,2)

⇒ (a,b) = (x+ y,x+ 2y)⇒ x+ y = a and x+ 2y = bThis is a consistent system of equations for all a,b ∈ R. So, every vector in R2(R) is a linearcombination of v1 and v2.

EXAMPLE-3 Let V = Pn(x) be the vector space of all polynomials of degree less than or equalto n over the field R of all real numbers. Show that the polynomials 1,x,x2, . . . ,xn−1,xn span V .

SOLUTION Any polynomial f (x) of degree less than or equal to n can be written as

f (x) = a0 + a1x+ a2x2 + · · ·+ anxn, where a0,a1,a2, . . . ,an ∈ R

⇒ f (x) is a linear combination of 1,x,x2, . . . ,xn−1,xn

Hence, polynomials 1,x,x2, . . . ,xn−1,xn span V .

EXAMPLE-4 Prove that the polynomials 1,1 + x,(1 + x)2 span the vector space V = P2(x) ofall polynomials of degree at most 2 over the field R of real numbers.

SOLUTION Let f (x) = a + bx+ cx2 be an arbitrary polynomial in V , where a,b,c ∈ R.In order to prove that the polynomials 1,1+ x,(1+ x)2 span V , it is sufficient to show that


This system of equations is inconsistent, because the rank of the coefficient matrix is not equalto the rank of the augmented matrix.

EXAMPLE-10 Let v1 = (−1,2,0),v2 = (3,2,−1) and v3 = (1,6,−1) be three vectors in R3(R).Then show that [v1,v2] = [v1,v2,v3].

SOLUTION Clearly,[v1,v2] = {λ1v1 + λ2v2 : λ1,λ2 ∈ R}

and, [v1,v2,v3] = {µ1v1 + µ2v2 + µ3v3 : µ1,µ2,µ3 ∈ R}.Let v3 = (1,6,−1) = α1v1 + α2v2. Then,

(1,6,−1) = (−α1 + 3α2,2α1 + 2α2,−α2)

⇒ −α1 + 3α2 = 1,2α1 + 2α2 = 6,−α2 = −1.

⇒ α1 = 2,α2 = 1∴ (1,6,−1) = 2(−1,2,0)+ 1(3,2,−1)

⇒ µ3(1,6,−1) = 2µ3(−1,2,0)+ µ3(3,2,−1)

⇒ µ3v3 = 2µ3v1 + µ3v2

Now, [v1,v2,v3] = {µ1v1 + µ2v2 + 2µ3v1 + µ3v2 : µ1,µ2,µ3 ∈ R}⇒ [v1,v2,v3] = {(µ1 + 2µ3)v1 +(µ2 + µ3)v2 : µ1,µ2,µ3 ∈ R}⇒ [v1,v2,v3] = {λ1v1 + λ2v2 : λ1,λ2 ∈ R}⇒ [v1,v2,v3] = [v1,v2].

EXAMPLE-11 Let v1 = (1,2,−1),v2 = (2,−3,2),v3 = (4,1,3) and v4 = (−3,1,2) be fourvectors in R3(R). Then show that [v1,v2] �= [v3,v4].

SOLUTION Suppose [v1,v2] = [v3,v4]. Then for given λ3,λ4 ∈R there exist λ1,λ2 ∈R such that

λ1v1 + λ2v2 = λ3v3 + λ4v4

⇒ λ1(1,2,−1)+ λ2(2,−3,2) = λ3(4,1,3)+ λ4(−3,1,2)

⇒ (λ1,2λ1,−λ1)+ (2λ2,−3λ2,2λ2) = (4λ3,λ3,3λ3)+ (−3λ4,λ4,2λ4)

⇒ (λ1 + 2λ2,2λ1 −3λ2,−λ1 + 2λ2) = (4λ3 −3λ4,λ3 + λ4,3λ3 + 2λ4)

⇒ λ1 + 2λ2 = 4λ3 −3λ4 (i)2λ1 −3λ2 = λ3 + λ4 (ii)−λ1 + 2λ2 = 3λ3 + 2λ4 (iii)

Solving (i) and (iii), we get λ1 = 12(λ3 −5λ4),λ2 = 1

4(7λ3 −λ4)

These values of λ1 and λ2 should satisfy (ii) but they do not satisfy (ii).Hence, [v1,v2] �= [v3,v4].


EXAMPLE-12 Let R2×2 be the vector space of all 2 × 2 matrices. Show that the matrices

E11 =

[

1 00 0

]

,E12 =

[

0 10 0

]

,E21 =

[

0 01 0

]

,E22 =

[

0 00 1

]

form a spanning set of R2×2.

SOLUTION Let M =

[

a bc d

]

be any matrix in R2×2, where a,b,c,d ∈ R.

Clearly, M = aE11 + bE12 + cE21 + dE22Thus, any matrix M in R2×2is expressible as a linear combination of E11,E12,E21,E22.

Hence, {E11,E12,E21,E22} forms a spanning set of R2×2.

EXAMPLE-13 Find one vector in R3(R) that spans the intersection of subspaces S and T whereS is the xy-plane, i.e. S = {(a,b,0) : a,b ∈ R} and T is the space spanned by the vectorsv1 = (1,1,1) and v2 = (1,2,3).

SOLUTION Since S and T are subspaces of R3(R). Therefore, S∩T is a subspace of R3(R).Let u = (a,b,c) be a vector in S∩T . Then,

u ∈ S and u ∈ T⇒ c = 0 and u is linear combination of v1 and v2 [∵ S = {(a,b,0) : a,b ∈ R}]⇒ u = xv1 + yv2 for some x,y ∈ R⇒ (a,b,0) = x(1,1,1)+ y(1,2,3)

⇒ (a,b,0) = (x+ y, x+ 2y, x+ 3y)⇒ x+ y = a, x+ 2y = b, x+ 3y = 0Solving first two equations, we get x= 2a−b,y = b−aSubstituting these values in x+ 3y = 0, we get a = 2b∴ v = (2b,b,0),b ∈ R.

EXAMPLE-14 Is the vector v = (3,−1,0,−1) in the subspace of R4(R) spanned by the vectorsv1 = (2,−1,3,2),v2 = (−1,1,1,−3) and v3 = (1,1,9,−5)?

SOLUTION If v is expressible as a linear combination of v1,v2 and v3, then v is the subspacespanned by v1,v2 and v3, otherwise not. So, let

v = xv1 + yv2 + zv3 for some scalars x,y,z.⇒ (3,−1,0,−1) = x(2,−1,3,2)+ y(−1,1,1,−3)+ z(1,1,9,−5)

⇒ (3,−1,0,−1) = (2x− y+ z,−x+ y+ z,3x+ y+ 9z,2x−3y−5z)⇒ 2x− y+ z = 3, −x+ y+ z = −1, 3x+ y+ 9z = 0, 2x−3y−5z = −1This system of equations can be expressed in matrix form as follows:

2 −1 1−1 1 1

3 1 92 −3 −5

xyz

=

3−1

0−1


This is equivalent to

0 1 3−1 1 1

0 4 120 −1 −3

xyz

=

1−1−3−3

Applying R1→R1 + 2R2, R3→R3 + 3R2, R4→R4 + 2R2

or,

0 0 0−1 0 −2

0 0 00 −1 −3

xyz

=

−2−4−15−3

Applying R1→R1 + R4, R2 → R2 + R4, R3 → R3 + 4R4

Clearly, this is an inconsistent system of equations. So, v is not expressible as a linear combi-nation of v1,v2,v3. Hence, v ∈/ [v1,v2,v3].

EXERCISE 2.7

1. Show that the vectors e(3)1 = (1,0,0),e(3)

2 = (0,1,0) and e(3)3 = (0,0,1) form a spanning

set of R3(R).

2. Show that the vectors v1 = (1,1,1),v2 = (1,1,0) and v3 = (1,0,0) form a spanning setof R3(R).

3. Consider the vector space V = Pn(t) consisting of all polynomials of degree ≤ n inindeterminate t. Show that the set of polynomials 1, t, t2, t3, . . . , tn forms a spanning setof V .

4. Show that the vectors v1 = (1,2,3),v2 = (1,3,5) and v3 = (1,5,9) do not spam R3(R).[Hint: (2,7,8) cannot be written as a linear combination of v1,v2,v3].

5. If v1,v2 and v3 are vectors in a vector space V (F) such that v1 + v2 + v3 = 0V , then showthat v1 and v2 span the same subspace of V as v2 and v3.

[Hint: Let S1 and S2 be subspaces spanned by vectors v1,v2 and v2,v3 respectively.i.e, S1 = [{v1,v2}] and S2 = [{v2,v3}]. We have to show that S1 = S2.

Let v be an arbitrary vector in S1. Then,

v ∈ S1

⇒ v = λ1v1 + λ2v2 for some λ1,λ2 ∈ F⇒ v = λ1(−v2 − v3)+ λ2v2 [∵ v1 + v2 + v3 = 0V ]

⇒ v = (λ1 −λ2)v2 +(−λ)v3

⇒ v ∈ S2

∴ S1 ⊂ S2.


REMARK-4 Note that the set {0V} in a vector space V(F) is always l.d., because λ0V = 0V forall non-zero λ ∈ F. On the other hand, a singleton set containing a non-zero vector v is alwaysl.i., because λv = 0V ⇒ λ = 0.

THEOREM-1 Any list of vectors containing the null vector is always l.d.

SOLUTION Let v = (v1,v2, . . . ,vn) be a list of vectors in a vector space V (F) such that vk = 0V .Then for any 0 �= λ ∈ F

0v1 + 0v2 + · · ·+ λvk + 0vk+1 + · · ·+ 0vn = 0V + 0V + · · ·+ 0V = 0V

This shows that a non-trivial linear combination of v1,v2, . . . ,vn equals to the null vector.Hence, v is a l.d. list.

Q.E.D.

REMARK. It follows from the foregoing discussion that the nature of the null vector is so strongthat its presence in the set always makes it l.d.

THEOREM-2 A list v = (v1,v2, . . . ,vn) of non-zero vectors in a vector space V (F) is l.d. iffsome one vector vk of the list is a linear combination (l.c.) of the previous vectors of the list.Also, if for some j,1 ≤ j ≤ n, the list (v1,v2, . . . ,v j) is a list of linearly independent vectors,then j < k.

PROOF. First suppose that the list (v1,v2, . . . ,vn) of vectors in V (F) is linearly dependent.Then there exist scalars λ1,λ2, . . . ,λn ∈ F not all zero such that

λ1v1 + λ2v2 + · · ·+ λnvn = 0V

Let k be the largest positive integer such that λk �= 0, so thatλ1v1 + λ2v2 + · · ·+ λkvk = 0V

We claim that k > 1. For, if k = 1λ1v1 = 0V and λ2 = λ3 = · · · = λn = 0

⇒ λ1 = 0 = λ2 = · · · = λn [∵ v1 �= 0V ∴ λ1v1 = 0V ⇒ λ1 = 0]

which contradicts the fact that (v1,v2, . . . ,vn) is a l.d. list.Now, λ1v1 + λ2v2 + · · ·+ λkvk = 0V

⇒ vk = (−λ1λ−1k )v1 + · · ·+(−λk−1λ−1

k )vk−1 (i)⇒ vk is a linear combination of the previous vectors of the list.

Hence, some one vector of the list is a linear combination of the previous vectors of the list.Conversely, suppose that a vector vk (say) of the list (v1,v2, . . . ,vk, . . . ,vn) is a linear combi-

nation of the previous vectors v1,v2, . . . ,vk−1. Then there exist scalars λ1,λ2, . . . ,λk−1 such thatvk = λ1v1 + λ2v2 + · · ·+ λk−1vk−1

⇒ λ1v1 + λ2v2 + · · ·+ λk−1vk−1 +(−1)vk = 0V⇒ λ1v1 + λ2v2 + · · ·+ λk−1vk−1 +(−1)vk + 0vk+1 + · · ·+ 0vn = 0V⇒ A non-trivial linear combination of the list (v1,v2, . . . ,vk, . . . ,vn) is zero vector


Hence, the list is linearly dependent.Q.E.D.

If possible, let the list (v1,v2, . . . ,v j) be a list of independent vectors and j ≥ k. Then, thelist of vectors v1,v2, . . . ,vk is linear independent. But,

vk = (−λ1λ−1k )v1 + · · ·+(−λk−1λ−1

k )vk−1 [From (i)]⇒ vk is a linear combination of v1,v2, . . . ,vk−1

⇒ v1,v2, . . . ,vk−1,vk are linearly dependent.

This is a contradiction. Hence, j < k.

THEOREM-3 Let S = {v1,v2,v3, . . .vk} be a linearly independent set of vectors in a vectorspace V (F) and v be a non-zero vector in V (F). Then, S1 = S∪{v} is linearly independent iffv ∈/ [S].

PROOF. First, let S1 be linearly independent. Suppose that v ∈ [S]. Then, there exist scalars

λ1,λ2, . . . ,λk in F such that v =k∑

i=1λivi

⇒ S1 = S∪{v} is linearly dependent.This contradicts the hypothesis that S1 is linearly independent. Hence, v ∈/ [S].Conversely, let v ∈/ [S]. Then, we have to prove that S1 is linearly independent. If possible,

let S1 be linearly dependent. Then, there exists a finite non-empty subset B (say) of S1 that islinearly dependent.

Now,B ⊂ S1 = S∪{v} and S is linearly independent

⇒ B contains v and some vectors in S = {v1,v2, . . . ,vk}

⇒ v is a linear combination of some vectors in S⇒ v is a linear combination of vectors in S⇒ v ∈ [S]

This is a contradiction. Hence, S1 is linearly independent.Q.E.D.

EXAMPLE-2 Show that the set S = {(1,2,4), (1,0,0), (0,1,0), (0,0,1)} is a linearly depen-dent set of vectors in R3(R).

SOLUTION Clearly,(1,2,4) = 1(1,0,0)+ 2(0,1,0)+ 4(0,0,1)

⇒ (1,2,4) is linear combination of (1,0,0), (0,1,0) and (0,0,1).Hence, S is a linearly dependent set of vectors.


THEOREM-4 Each subset of a linearly independent set is linearly independent.

PROOF. Let S = {v1,v2, . . . ,vm} be a linearly independent set of vectors in a vector spaceV(F). Let S ′ = {v1,v2, . . . ,vn} n � m, be an arbitrary finite subset of S. Then,

λ1v1 + λ2v2 + · · ·+ λnvn = 0V

λ1v1 + λ2v2 + · · ·+ λnvn + 0vn+1 + · · ·+ 0vm = 0V

⇒ λ1 = λ2 = λ3 = · · · = λn = 0 [∵ S is l.i.]

Hence, S ′ is a linearly independent set.Q.E.D.

THEOREM-5 Any super set of linearly dependent set of vectors in a vector space is linearlydependent.

PROOF. Let S = {v1,v2, . . . ,vm} be a linearly dependent set of vectors in a vector space V (F),and let S′ = {v1,v2, . . . ,vm,vm+1} be a super set of S.

Now,S is a linearly dependent set

⇒ There exist scalars λ1,λ2, . . . ,λm not all zero such that λ1v1 + λ2v2 + · · ·+ λmvm = 0V

⇒ λ1v1 + λ2v2 + · · ·+ λmvm + λm+1vm+1 = 0V , where λm+1 = 0

⇒ A non-trivial linear combination of v1,v2, . . . ,vm,vm+1 is zero vector[

∵ At leastone λi �= 0

]

⇒ S′ is a linearly dependent set.

Q.E.D.

THEOREM-6 Let S be any linearly independent subset of a vector space V (F) and letS = S1 ∪S2 such that S1 ∩S2 = φ.Prove that (i) [S] = [S1]+ [S2] (ii) [S1]∩ [S2] = φ i.e. [S] isthe direct sum of [S1] and [S2].

PROOF. (i) By Theorem 4(iv) on page 169, we have

[S1 ∪S2] = [S1]+ [S2] i.e, [S] = [S1]+ [S2]

(ii) Let S1 = {v1,v2, . . . ,vm} and S2 = {u1,u2, . . . ,un}.Let v be an arbitrary vector in [S1]∩ [S2]. Then,

v ∈ [S1]∩ [S2]

⇒ v ∈ [S1] and v ∈ [S2]

⇒ v =mΣ

i=1λivi and v =

nΣ

i=1µju j for some λi ∈ F and µj ∈ F


⇒mΣ

i=1λivi =

nΣ

j=1µju j

⇒mΣ

i=1λivi +

nΣ

j=1(−µj)uj = 0V

⇒ λi =0 and µj =0 for all i ∈ m and j ∈ n[

∵ S = S1 ∪S2 = {v1,v2, . . . ,vm,u1,u2, . . . ,un} is l.i.

]

∴ v = 0V

Since v is an arbitrary vector in [S1]∩ [S2]. Therefore, [S1]∩ [S2] = φ.Q.E.D.


EXAMPLE-1 Let F be a field with unity 1. Show that the vectors e(n)1 = (1,0,0, . . . ,0),

e(n)2 = (0,1,0, . . . ,0),e(n)

3 = (0,0,1, . . . ,0) . . . ,e(n)n = (0,0,0, . . . ,1) are linearly independent in

the vector space Fn(F).

SOLUTION Let λ1,λ2, . . . ,λn be scalars such that

λ1e(n)1 + λ2e(n)

2 + λ3e(n)3 + · · ·+ λne(n)

n = 0⇒ (λ1,λ2,λ3, . . . ,λn) = (0,0,0, . . . ,0)

⇒ λ1 = λ2 = λ3 = · · · = λn = 0

Hence, e(n)1 ,e(n)

2 , . . . ,e(n)n are linearly independent.

Particular Cases:

(i) e(2)1 = (1,0),e(2)

2 = (0,1) are linearly independent vectors in R2(R).

(ii) e(3)1 = (1,0,0),e(3)

2 = (0,1,0),e(3)3 = (0,0,1) are linearly independent vectors in R3(R).

EXAMPLE-2 Show that the vectors v1 = (1,1,0),v2 = (1,3,2),v3 = (4,9,5) are linearly de-pendent in R3(R).

SOLUTION Let x,y,z be scalars i.e., real numbers such that

xv1 + yv2 + zv3 = 0⇒ x(1,1,0)+ y(1,3,2)+ z(4,9,5) = (0,0,0)

⇒ (x+ y+ 4z,x+ 3y+ 9z,2y+ 5z) = (0,0,0)

⇒ x+ y+ 4z = 0, x+ 3y+ 9z = 0, 0x+ 2y+ 5z = 0


This is a homogeneous system of linear equations. The determinant of the coefficientmatrix A is

|A| =

∣

∣

∣

∣

∣

∣

1 1 41 3 90 2 5

∣

∣

∣

∣

∣

∣

= 0

So, the system has non-trivial solutions. Hence, given vectors are linearly dependent in R3(R).

Aliter In order to check the linear independence or dependence of vectors, we may follow thefollowing algorithm:

ALGORITHM

Step I Form a matrix A whose columns are given vectors.Step II Reduce the matrix in step-I to echelon form.Step III See whether all columns have pivot elements or not. If all columns have pivot

elements, then given vectors are linearly independent. If there is a column not havinga pivot element, then the corresponding vector is a linear combination of the preced-ing vectors and hence linearly dependent.

In Example 2, the matrix A whose columns are v1,v2,v3 is

A =

1 1 41 3 90 2 5

∴ A ∼

1 1 40 2 50 2 5

Applying R2 → R2 −R1

⇒ A ∼

1© 1 40 2© 50 0 0


Pivots in the echelon form of matrix A have been encircled. We observe that the third columndoes not have a pivot. So, the third vector v3 is a linear combination of the first two vectors v1and v2. Thus, the vectors v1,v2,v3 are linearly dependent.

EXAMPLE-3 Show that the vectors v1 = (1,2,3),v2 = (2,5,7),v3 = (1,3,5) are linearly inde-pendent in R3(R).

SOLUTION Let x,y,z be scalars such thatxv1 + yv2 + zv3 = 0

⇒ x(1,2,3)+ y(2,5,7)+ z(1,3,5) = (0,0,0)

⇒ (x+ 2y+ z,2x+ 5y+ 3z,3x+ 7y+ 5z) = (0,0,0)

⇒ x+ 2y+ z = 0, 2x+ 5y+ 3z = 0, 3x+ 7y+ 5z = 0


The determinant of the coefficient matrix A of the above homogeneous system of equations isgiven by

|A| =

∣

∣

∣

∣

∣

∣

1 2 12 5 33 7 5

∣

∣

∣

∣

∣

∣

= 1 �= 0

So, the above system of equations has trivial solution only, i.e. x= y = z = 0.Hence, the given vectors are linearly independent in R3(R).

EXAMPLE-4 Show that the vectors v1 = (1,1,2,4),v2 = (2,−1,−5,2),v3 = (1,−1,−4,0) andv4 = (2,1,1,6) are linearly dependent in R4(R).

SOLUTION Let x,y,z, t be scalars in R such that

xv1 + yv2 + zv3 + tv4 = 0⇒ x(1,1,2,4)+ y(2,−1,−5,2)+ z(1,−1,−4,0)+ t(2,1,1,6) = 0⇒ (x+ 2y+ z+ 2t,x− y− z+ t,2x−5y−4z+ t,4x+ 2y+ 0z+ 6t) = (0,0,0,0)

⇒ x+ 2y+ z+ 2t = 0, x− y− z+ t = 0, 2x−5y−4z+ t = 0, 4x+ 2y+ 0z+ 6t = 0

The coefficient matrix A of the above homogeneous system of equations is

A =

1 2 1 21 −1 −1 12 −5 −4 14 2 0 6

⇒ A ∼

1 2 1 20 −3 −2 −10 −9 −6 −30 −6 −4 −2

Applying R2 → R2 −R1, R3 → R3 −2R1, R4 → R4 −4R1

⇒ A ∼

1 2 1 20 −3 −2 −10 0 0 00 0 0 0

Applying R3 → R3 −3R2, R4 → R4 −2R2

Clearly, rank of A = 2 < Number of unknowns. So, the above system has non-trivial solutions.Hence, given vectors are linearly dependent in R4(R).Aliter The matrix A whose columns are v1,v2,v3,v4 is

A =

1 2 1 21 −1 −1 12 −5 −4 14 2 0 6


The echelon form of matrix A is

1 2 1 20 −3 −2 −10 0 0 00 0 0 0

We observe that the third and fourth columns does not have pivots. So, third and fourth vectorsare linear combinations of first two vectors v1 and v2. Thus, the vectors are linearly dependent.

EXAMPLE-5 Determine whether the vectors f (x) = 2x3 + x2 + x+ 1,g(x) = x3 + 3x2 + x− 2and h(x) = x3 + 2x2 − x+ 3 in the vector space R[x] of all polynomials over the real numberfield are linearly independent or not.

SOLUTION Let a,b,c be real numbers such that

a f (x)+ bg(x)+ ch(x) = 0 for all x

⇒ a(2x3 + x2 + x+ 1)+ b(x3 + 3x2 + x−2)+ c(x3 + 2x2 − x+ 3) = 0 for all x

⇒ (2a+ b+ c)x3 +(a+ 3b+ 2c)x2 +(a+ b− c)x+(a−2b+ 3c) = 0 for all x

⇒ 2a+ b+ c = 0, a+ 3b+ 2c = 0, a+ b− c = 0, a−2b+ 3c = 0

The coefficient matrix A of the above system of equations is

A =

2 1 11 3 21 1 −11 −2 3

⇒ A ∼

1 1 −11 3 22 1 11 −2 3

Applying R1 ↔ R3

⇒ A ∼

1 1 −10 2 30 −1 30 −3 4

Applying R2 → R2 −R1,R3 → R3 −2R1, R4 → R4 −R1

⇒ A ∼

1 1 −10 0 90 −1 30 0 −5

Applying R2 → R2 + 2R3, R4 → R4 −3R3


⇒ A ∼

1 1 −10 0 90 −1 30 0 0

Applying R4 → R4 +59

R2

⇒ A ∼

1 1 −10 −1 30 0 90 0 0

Applying R2 ↔ R3

Clearly, rank of A is equal to 3 which is equal to the number of unknowns a,b,c. So, the systemhas trivial solution only, i.e. a = b = c = 0.

So, f (x),g(x),h(x) are linearly independent in R[x].

EXAMPLE-6 Let V be the vector space of functions from R into R. Show that the functionsf (t) = sin t,g(t) = et ,h(t) = t2 are linearly independent in V .

SOLUTION Let x,y,z be scalars such that

x f (t)+ y g(t)+ z h(t) = 0 for all t ∈ R

Putting t = 0,π andπ2

, respectively we get

x f (0)+ y g(0)+ zh(0) = 0x f (π)+ y g(π)+ z h(π) = 0

x f(π

2

)

+ y g(π

2

)

+ z h(π

2

)

= 0

⇒ y = 0, y eπ + z π2 = 0,x+ y eπ2 + h

π2

4= 0

⇒ y = 0,z = 0,x= 0

Thus,x f (t)+ yg(t)+ zh(t) = 0 ⇒ x= y = z = 0

Hence, f (t),g(t),h(t) are linearly independent.

EXAMPLE-7 Show that in the vector space F[x] of all polynomials in an indeterminate x, theset S = {1,x,x2, . . .} is linearly independent.

PROOF. Let S′ = {xn1 ,xn2 , . . . ,xnk} be a finite subset of S, where n1,n2, . . . ,nk are distinctnon-negative integers.Let λ1,λ2, . . . ,λk ∈ F be such that

λ1xn1 + λ2xn2 + · · ·+ λkxnk = 0(x) [0(x) is zero polynomial over F]⇒ λ1 = λ2 = · · · = λk = 0 [By the definition of equality of two polynomials]

So, S′ is a linearly independent set.


Since S′ is an arbitrary finite subset of S. Therefore, every finite subset of S is linearlyindependent. Consequently, S is linearly independent.

Q.E.D.

EXAMPLE-8 Show that the vectors (matrices) A1 =

[

1 11 1

]

, A2 =

[

1 00 1

]

, A3 =

[

1 10 0

]

in

R 2×2 are linearly independent.

SOLUTION Let λ1,λ2,λ3 ∈ R be such that

λ1A1 + λ2A2 + λ3A3 = O 2×2 (Null matrix).

⇒ λ1

[

1 11 1

]

+ λ2

[

1 00 1

]

+ λ3

[

1 10 0

]

=

[

0 00 0

]

⇒

[

λ1 + λ2 + λ3 λ1 + λ3λ1 λ1 + λ2

]

=

[

0 00 0

]

⇒ λ1 + λ2 + λ3 = 0, λ1 + λ3 = 0, λ1 = 0, λ1 + λ2 = 0⇒ λ1 = λ2 = λ3 = 0.

Hence, A1,A2,A3 are linearly independent vectors in R 2×2.

EXAMPLE-9 If v1,v2 are vectors in a vector space V (F) and λ1,λ2 ∈ F, show that the set{v1,v2,λ1v1 + λ2v2} is linearly dependent set.

SOLUTION Since the vector λ1v1 +λ2v2 in the set {v1,v2,λ1v1 +λ2v2} is a linear combinationof other two vectors. Therefore, by Theorem 2 on page 180 the set is linearly dependent.

Aliter. Clearly, (−λ1)v1 +(−λ2)v2 +1(λ1v1 +λ2v2) = 0V . Therefore, a non-trivial linear com-bination of v1,v2,λ1v1 + λ2v2 equals to the null vector. Hence, the given set is linearly depen-dent.

2.8.1 LINEAR DEPENDENCE AND ECHELON MATRICESConsider the following echelon matrix A, whose pivots have been circled:

A =

0 2© 3 4 5 6 70 0 4© 3 2 3 40 0 0 0 7© 8 90 0 0 0 0 6© 70 0 0 0 0 0 0

We observe that the rows R2,R3,R4 have 0’s in the second column below the non-zero pivot inR1, and hence any linear combination of R2,R3 and R4 must have 0 as its entry as the secondcomponent. Whereas R1 has a non-zero entry 2 as the second component. Thus, R1 cannot bea linear combination of the rows below it. Similarly, the rows R3 and R4 have 0’s in the third


column below the non-zero pivot in R2, and hence R2 cannot be a linear combination of therows below it. Finally, R3 cannot be a multiple of R4 as R4 has a 0 in the fifth column belowthe non-zero pivot in R3. Thus, if we look at the rows from the bottom and move upward, wefind that out of rows R4,R3,R2,R1 non row is a linear combination of the preceding rows. So,R1,R2,R3,R4 are linearly independent vectors R7(R).

The above discussion suggests the following theorem.

THEOREM-1 The non-zero rows of a matrix in echelon form are linearly independent.

EXAMPLE-1 Let v1,v2,v3 be vectors in a vector space V (F), and let λ1,λ2 ∈ F. Show that theset {v1,v2,v3} is linearly dependent if the set {v1 + λ1v2 + λ2v3,v2,v3} is linearly dependent.

SOLUTION If the set {v1 + λ1v2 + λ2v3,v2,v3} is linearly dependent, then there exist scalarsλ,µ,ν ∈ F (not all zero) such that

λ(v1 + λ1v2 + λ2v3)+ µv2 + νv3 = 0V

⇒ λv1 +(λλ1 + µ)v2 +(λλ2 + ν)v3 = 0V (i)

The set {v1,v2,v3} will be linearly dependent if in (i) at least one of the scalar coefficients isnon-zero.

If λ �= 0, then the set will be linearly dependent whatever may be the values of µ and ν.But, if λ = 0, then at least one of µ and ν should not be equal to zero and hence at least one ofλλ1 + µ and λλ2 + ν will not be zero (since λ = 0 ⇒ λλ1 + µ= µ and λλ2 + ν = ν).

Hence, from (i), we find that the scalars λ, λλ1 +µ, λλ2 +ν are not all zero. Consequently,the set {v1,v2,v3} is linearly dependent.

EXAMPLE-2 Suppose the vectors u,v,w are linearly independent vectors in a vector spaceV (F). Show that the vectors u+ v,u− v,u−2v+ w are also linearly independent.

SOLUTION Let λ1,λ2,λ3 be scalars such that

λ1(u+ v)+ λ2(u− v)+ λ3(u−2v+ w) = 0V

⇒ (λ1 + λ2 + λ3)u+(λ1 −λ2 −2λ3)v+ λ3w = 0V

⇒ λ1 + λ2 + λ3 = 0, λ1 −λ2 −2λ3 = 0, 0λ1 + 0λ2 + λ3 = 0[

∵ u,v,w are linearlyindependent

]

The coefficient matrix A of the above system of equations is given by

A =

1 1 11 −1 −20 0 1

⇒ |A| = −2 �= 0

So, the above system of equations has only trivial solution λ1 = λ2 = λ3 = 0.Thus, u+ v, u− v, u−2v+ w are linearly independent vectors.


EXERCISES 2.8

1. Mark each of the following true or false.(i) For any subset S of a vector space V (F), [S+ S] = [S]

(ii) If S and T are subsets of a vector space V (F), then S �= T ⇒ [S] �= [T ].(iii) If S and T are subsets of a vector space V (F), then S ⊂ T ⇒ [S] ⊂ [T ]

(iv) Every subset of a linearly independent set is linearly independent.(v) Every superset of a linearly independent set is linearly independent.

(vi) Every subset of a linearly dependent set is linearly dependent.(vii) Every superset of a linearly dependent set is linearly dependent.

(viii) A subset of a linearly dependent set can never be linearly dependent.(ix) The void set is linearly dependent.(x) If a non-void subset of a vector space is not linearly dependent, then it is linearly

independent.(xi) Any set of vectors containing the null vector is linearly dependent.

(xii) Every singleton set consisting of a non-zero vector is linearly independent.(xiii) Intersection of two linearly independent subsets of a vector space is linearly inde-

pendent.(xiv) Union of two linearly independent subsets of a vector space is linearly independent.

2. Which of the following subsets S of R 3 are linearly independent?(i) S = {(1,2,1),(−1,1,0),(5,−1,2)}

(ii) S = {(1,1,0),(0,0,1),(1,5,2)}

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

(iv) S = {(1,3,2),(1,−7,8),(2,1,1)}

(v) S = {(1,5,2),(1,0,0),(0,1,0)}

3. Which of the following subsets S of R[x] are linearly independent?(i) S = {1,x,x2 ,x3,x4}

(ii) S = {1,x− x2,x+ x2,3x}(iii) S = {x2 −1, x+ 1, x−1}(iv) S = {x, x− x3, x2 + x4, x+ x2 + x4 + 1

2}

(v) S = {1, 1+ x, 1+ x+ x2, x4}

(vi) S = {1, x, x(1− x)}(vii) S =

{

1, x, 1+ x+ x2}


4. Let f (x), g(x), h(x), k(x) ∈ R[x] be such that f (x) = 1, g(x) = x, h(x) = x2 and k(x) =1+ x+ x2 for all x ∈ R. Show that f (x), g(x), h(x), k(x) are linearly dependent but anythree of them are linearly independent.

5. Which of the following subsets of the space of all continuous functions on R are linearlyindependent?

(i) S = {sinx, cosx, sin(x+ 1)}

(ii) S = {xex, x2ex, (x2 + x−1)ex}(iii) S = {sin2 x, cos2x, 1}(iv) S = {x, sinx, cosx}(v) S = {x, x2, e2x}

6. Prove that the set S = {1, i} is linearly independent in the vector space C of all complexnumbers over the field R of real numbers while it is linearly dependent in the vector spaceC of all complex numbers over the field of complex numbers.

7. If the set {v1, v2, v3} is linearly independent in a vector space V (F), then prove that theset {v1 + v2, v2 + v3, v3 + v1} is also linearly independent.

8. Prove that the vectors v1 = (1, 1, 0, 0), v2 = (0, 0, 0, 3), and v3 = (0, 1, −1, 0)in F4(F) are linearly independent if F is a field of characteristic zero and are linearlydependent if F is of characteristic 3.

9. If {v1, v2, . . . ,vn} is a linearly independent set of vectors in a vector space V (F) and{v1, v2, . . . ,vn, u} is a linearly dependent set in V . Then show that u is a linear combi-nation of v1, v2, . . . ,vn.

10. If the set S = {v1, v2, . . . ,vn} is a linearly independent set of vectors in a vector spaceV (F) and u∈[S], then show that representation of u as a linear combination of v1, v2, . . . ,vnis unique.

11. Show that the vectors v1 = (1+ i, 2i) and v2 = (1, 1+ i) are linearly dependent in C2(C)but linearly independent in C2(R).

12. If v1, v2, v3 are linearly dependent vectors of V (F) where F is any subfield of the fieldof complex numbers, then so also are v1 + v2, v2 + v3, v3 + v1.

13. Show that the functions f (t) = sin t, g(t) = cos t, h(t) = t from R to R are linearlyindependent.

14. Find a maximal linearly independent subsystem of the system of vectors:v1 = (2,−2,−4), v2 = (1,9,3), v3 = (−2,−4,1), v4 = (3,7,−1).

ANSWERS1. (i) F (ii) F (iii) T (iv) T (v) F (vi) F (vii) T (viii) F (ix) F (x) T

(xi) T (xii) T (xiii) T (xiv) F2. (i), (ii), (iv), (v) 3. (i), (iii), (iv), (v), (vi), (vii) 5. (ii), (iv), (v)

14. {v1, v2}


2.9 BASIS AND DIMENSION

BASIS. A non-void subset B of a vector space V(F) is said to be a basis for V , if(i) B spans V , i.e. [B] = V .

and, (ii) B is linearly independent (l.i.).In other words, a non-void subset B of a vector space V is a linearly independent set of vectorsin V that spans V .

FINITE DIMENSIONAL VECTOR SPACE. A vector space V(F) is said to be a finite dimensionalvector space if there exists a finite subset of V that spans it.

A vector space which is not finite dimensional may be called an infinite dimensional vectorspace.

REMARK. Note that the null vector cannot be an element of a basis, because any set containingthe null vector is always linearly dependent.

For any field F and a positive integer n, the set B = {e(n)1 , e(n)

2 , . . . ,e(n)n } spans the vector

space Fn(F) and is linearly independent. Hence, it is a basis for Fn. This basis is calledstandard basis for Fn.

REMARK. Since the void set ϕ is linearly independent and spans the null space {0V}. There-fore, the void set ϕ is the only basis for the null space {0V}.

Consider the subset B = {e(3)1 , e(3)

2 , e(3)3 }, where e(3)

1 = (1, 0, 0), e(3)2 = (0, 1, 0),

e(3)3 = (0, 0, 1), of the real vector space R3. The set B spans R3, because any vector (a, b, c)

in R3 can be written as a linear combination of e(3)1 , e(3)

2 and e(3)3 , namely,

(a, b, c) = ae(3)1 + be3

2 + ce(3)3

Also, B is linearly independent, because

λ e(3)1 + µ e(3)

2 + νe(3)3 = (0, 0, 0)

⇒ λ(1, 0, 0)+ µ(0, 1, 0)+ ν(0, 0, 1) = (0, 0, 0)

⇒ (λ, µ, ν) = (0, 0, 0) ⇒ λ = µ= ν = 0

Hence, B is a basis for the real vector space R3.

Now consider the subset B1 = {(1, 1, 0), (1, 0, 1), (0, 1, 1)} of R3.This set is also a basis for R3, because any vector (a,b,c) ∈ R3 can be written as

(a, b, c) =

(

a+ b− c2

)

(1, 1, 0)+

(

a1 + c−b2

)

(1, 0, 1)+

(

b+ c−a2

)

(0, 1, 1)

and, λ(1, 1, 0)+ µ(1, 0, 1)+ ν(0, 1, 1) ⇒ λ = µ = ν = 0.

REMARK. It follows from the above discussion that a basis for a vector space need not beunique. In fact, we shall show that corresponding to every non-zero vector in a vector space Vone can obtain a basis for V .


EXAMPLE-1 Show that the set B = {1, x, x2, . . . ,xn} is a basis for the vector space of all realpolynomials of degree not exceeding n.

SOLUTION Let λ0, λ1, . . . ,λn ∈ R be such thatλ0 ·1+ λ1 · x+ · · ·+ λnxn = 0(x) (zero polynomial)

Then,

λ0 ·1+ λ1 · x+ · · ·+ λnxn = 0(x) (zero polynomial)⇒ λ0 + λ1x+ λ2x2 + · · ·+ λnxn = 0+ 0x+ 0x2 + · · ·+ 0xn + . . .

⇒ λ0 = λ1 = · · · = λn = 0

Therefore, the set B is linearly independent.Also, the set B spans the vector space Pn(x) of all real polynomials of degree not exceeding n,because every polynomial of degree less than or equal n is a linear combination of B.Hence, B is a basis for the vector space Pn(x) of all real polynomials of degree not exceeding n.

EXAMPLE-2 Show that the infinite set {1, x, x2, . . .} is a basis for the vector space R[x] of allpolynomials over the field R of real numbers.

SOLUTION The set B = {1, x, x2, . . .} is linearly independent. Also, the set B spans R[x],because every real polynomial can be expressed as a linear combination of B.

Hence, B is a basis for R[x]. Since B is an infinite set. Therefore, R[x] is an infinite dimen-sional vector space over R.

EXAMPLE-3 Let a, b ∈ R such that a < b. Then the vector space C[a, b] of all real valuedcontinuous functions on [a, b] is an infinite dimensional vector space.

EXAMPLE-4 The field R of real numbers is an infinite dimensional vector space over its subfieldQ of rational numbers. But, it is a finite dimensional vector space over itself.

SOLUTION Since π is a transcendental number, that is, π is not a root of any polynomial overQ. Thus, for any n ∈ N and ai ∈ Q(i = 0, 1, 2, . . . ,n)

a0 + a1π+ a2π2 + · · ·+ anπn �= 0

⇒ 1, π, π2, π3, . . . ,πn, . . . are linearly independent over Q.Hence, R is an infinite dimensional vector space over Q.

EXAMPLE-5 The set {1, i} is a basis for the vector space C of all complex numbers over thefield R of real numbers.

We have defined basis of a vector space and we have seen that basis of a vector spaceneed not be unique. Now a natural question arises, does it always exist?. The answer is in theaffirmative as shown in the following theorem.


THEOREM–1 Every finite dimensional vector space has a basis.

PROOF. Let V be a finite dimensional vector space over a field F . If V is the null space,then the void set ϕ is its basis, we are done. So, let V be a non-null space. Since V is finitedimensional. So, there exists a finite subset S = {v1, v2, . . . ,vn} of V that spans V . If S isa linearly independent set, we are done. If S is not linearly independent, then by Theorem 2on page 180, there exists a vector vk ∈ S which is linear combination of the previous vectors.Remove vk from S and let S1 = S−{vk}. By Theorem 3 on page 167, S1 spans V .

If S1 is a linearly independent set, the theorem is proved. If not we repeat the above processon S1 and omit one more vector to obtain S2 ⊂ S1. Continuing in this manner, we obtainsuccessively S1 ⊃ S2 ⊃ S3, . . . , where each Si spans V .

Since S is a finite set and each Si contains one vector less than Si−1. Therefore, we ulti-mately arrive at a linearly independent set that spans V . Note that this process terminates beforewe exhaust all vectors in S, because if not earlier, then after (n−1) steps, we shall be left with asingleton set containing a non-zero vector that spans V . This singleton set will form a basis forV because each singleton set containing a non-zero vector in V forms a linearly independentset. Hence, V has a basis.

Q.E.D.

REMARK. This theorem remains valid if the word finite is omitted. In that case, the proofrequires the use of Zorn’s lemma. This lemma involves several concepts which are beyond thescope of this book.NOTE. In future unless otherwise mentioned a vector space will mean a finite dimensional vectorspace.

THEOREM-2 The representation of a vector in terms of the basis vectors is unique.

PROOF. Let V be a vector space over a field F , and let B = {b1, b2, . . . ,bn} be a basis for V .Let v be an arbitrary vector in V . We have to prove that v has a unique representation in

terms of vectors in B.Let v = λ1b1 + · · ·+ λnbn and v = µ1b1 + · · ·+ µnbn be two representations for v ∈ V as a

linear combination of basis vectors. Then,

λ1b1 + · · ·+ λnbn = µ1b1 + · · ·+ µnbn

⇒ (λ1 −µ1)b1 + · · ·+(λn −µn)bn = 0V

⇒ λ1 −µ1 = 0, . . . ,λn −µn = 0 [∵ B is a linearly independent set]⇒ λ1 = µ1, λ2 = µ2, . . . ,λn = µn

Hence, v has a unique representation in terms of vectors in B. Q.E.D.

THEOREM-3 Let S be a subset of a vector space V(F) such that S spans V . Then S contains abasis for V .


PROOF. By Theorem 1, the basis for V is obtained by removing those vectors from S, whichare linear combinations of previous vectors in S. Hence, S contains a basis for V .

Q.E.D.

THEOREM-4 A subset B of a vector space V (F) is a basis for V if every vector in V has aunique representation as a linear combination of vectors of B.

PROOF. First suppose that B is a basis for V . Then B spans V and so every vector in V is alinearly combination of vectors of B. To prove the uniqueness, we consider the following twocases:

Case I When B is finite, say B= {b1, b2, . . . ,bn}.Let v be an arbitrary vector in V and suppose that

v = λ1b1 + λ2b2 + · · ·+ λnbn and, also v = µ1b1 + µ2b2 + · · ·+ µnbn,

where λi, µi ∈ F for all i ∈ n. Then,

λ1b1 + λ2b2 + · · ·+ λnbn = µ1b1 + µ2b2 + · · ·+ µnbn

⇒ (λ1 −µ1)b1 + · · ·+(λn −µn)bn = 0V

⇒ λ1 −µ1 = 0, . . . ,λn −µn = 0 [∵ B is linearly independent]⇒ λ1 = µ1,λ2 = µ2, . . . ,λn = µn.

Thus, v ∈V has a unique representation as a linear combination of vectors of B.Case II When B is infinite.

Let v ∈V and suppose that v = ∑b∈B

λb b and, also v = ∑b∈B

µb b

where λb,µb ∈ F and each sum contains only finite number of terms (i.e. λb,µb arezero in all except a finite number of terms). Then,

Σb∈B

λb b = Σb∈B

µb b

⇒ Σb∈B

(λb −µb)b = 0V

⇒ λb −µb = 0 for all b ∈ B [∵ B is linearly independent]⇒ λb = µb for all b ∈ B

This proves that v ∈V has a unique representation as a linear combination of vectors of B.Conversely, suppose that every vector in V has a unique representation as a linear combi-

nation of vectors of B. Then, B spans V . Further suppose that

λ1b1 + λ2b2 + · · ·+ λnbn = 0V for some λ1,λ2, . . . ,λn ∈ F, b1,b2, . . . ,bn ∈ B.


Then,

λ1b1 + λ2b2 + · · ·+ λnbn = 0b1 + 0b2 + · · ·+ 0bn

⇒ λ1 = λ2 = · · · = λn = 0[

∵ Every vector in V has a unique representationas a linear combination of vectors in B

]

Thus, the only linear combination of vectors of B that equals to the null vector is the triviallinear combination. Hence, B is a basis for V .

Q.E.D.

The following theorem proves an extremely important result that one cannot have morelinearly independent vectors than the number of vectors in a spanning set.

THEOREM-5 Let V be a vector space over a field F. If V is spanned by the set {v1,v2, . . . ,vn}

of n vectors in V and if {w1,w2, . . . ,wm} is a linearly independent set of vectors in V , thenm � n. Moreover, V can be spanned by a set of n vectors containing the set {w1,w2, . . . ,wm}.

PROOF. We shall prove both the results together by induction on m.First we shall prove the theorem for m = 1.Let {w1} be a linearly independent set. Then w1 �= 0V .

Now,

w1 ∈V⇒ w1 is a linear combination of v1,v2, . . . ,vn [∵ {v1,v2, . . . ,vn} spans V ]⇒ {w1,v1,v2, . . . ,vn} is linearly dependent set [By Theorem 2 on page 180]⇒ There exist vk ∈ {w1,v1, . . . ,vn} such that vk is a linear combination of the

preceeding vectors

Let us rearrange the vectors w1,v1, . . . ,vn in such a way that vn is a linear combination of theprevious vectors. Removing vn from the set {w1,v1, . . . ,vn} we obtain the set {w1,v1, . . . ,vn−1}

of n vectors containing w1 such that it spans V and n−1 � 0 ⇒ n � 1 = m.Note that the vector removed is one of v’s because the set of w’s is linearly independent.Hence, the theorem holds for m = 1.Now suppose that the theorem is true for m. This means that for a given set {w1, . . . ,wm}

of m linearly independent vectors in V , we have

(i) m � nand, (ii) there exists a set of n vectors w1, . . . ,wm,v1, . . . ,vn−m in V that spans V .

To prove the theorem for m + 1, we have to show that for a given set of (m + 1) linearlyindependent vectors w1,w2, . . . ,wm,wm+1 in V , we have

(i) m + 1 � nand, (ii) there exists a set of n vectors containing w1, . . . ,wm,wm+1 that spans V .


we repeat the above process. Continuing in this manner we remove, one by one, every vectorwhich is a linear combination of the preceding vectors, till we obtain a linearly independent setspanning V . Since S is a linearly independent set, vi is a linear combination of the precedingvectors. Therefore, the vectors removed are some of bi’s. Hence, the reduced linearly indepen-dent set consists of all vi’s and some of bi’s. This reduced set {v1,v2, . . . ,vm,b1,b2, . . . ,bn−m}

is a basis for V containing S.Q.E.D.

This theorem can also be stated as

“Any l.i. set of vectors of a finite dimensional vector space is a part of its basis.”

COROLLARY-1 In a vector space V (F) of dimension n

(i) any set of n linearly independent vectors is a basis,

and, (ii) any set of n vectors that spans V is a basis.

PROOF. (i) Let B={b1,b2, . . . ,bn} be a set of n linearly independent vectors in vector spaceV . Then by Theorem 7, B is a part of the basis. But, the basis cannot have more than n vectors.Hence, B itself is a basis for V .

(ii) Let B = {b1,b2, . . . ,bn} be a set of n vectors in V that spans V . If B is not linearlyindependent, then there exists a vector in B which is linear combination of the preceding vectorsand by removing it from B we will obtain a set of (n− 1) vectors in V that will also span V .But, this is a contradiction to the fact that a set of (n−1) vectors cannot span V . Hence, B is alinearly independent set. Consequently, it is a basis for V .

Q.E.D.

COROLLARY-2 Let V (F) be a vector space of dimension n, and let S be a subspace of V . Then,

(i) every basis of S is a part of a basis of V .

(ii) dimS � dimV .

(iii) dimS = dimV ⇔ S = V .

and, (iv) dimS < dimV , if S is a proper subspace of V .

PROOF. (i) Let B ′ = {b1,b2, . . . ,bm} be a basis for S. Since S is a subspace of V . Therefore,B ′ is a linearly independent set of vectors in V and hence it is a part of a basis for V .(ii) Since basis for S is a part of the basis for V . Therefore, dimS � dimV.

(iii) Let dimS = dimV = n. Let B = {b1,b2, . . . ,bn} be a basis for V so that [B] = S.


Since B is a linearly independent set of n vectors in V and dimV = n. Therefore, B spans V ,i.e. [B] = V .Thus, S = [B] = V ⇒ S = V.Conversely, let S = V. Then,

S = V⇒ S is a subspace of V and V is a subspace of S⇒ dimS � dimV and dimV � dimS [By (ii)]⇒ dimS = dimV.

(iv) Let S be a proper subspace of V . Then there exists a vector v ∈V such that v∈/S. Therefore,v cannot be expressed as a linear combination of vectors in B ′, the basis for S. That is, v ∈/ [S].Consequently, the set {b1,b2, . . . ,bm,v} forms a linearly independent subset of V . Therefore,the basis for V will contain more than m vectors.

Hence, dimS < dimV .

Q.E.D.

THEOREM-8 Let V (F) be a finite dimensional vector space, and let m be a positive integer.Then the function space V m = { f : f : m →V} is also finite dimensional and

dim(V m) = m.dimV

PROOF. Let dimV = n, and let B = {b1,b2, . . . ,bn} be a basis for V . By Example 11 on page126, the function space V m is a vector space over field F . For any i ∈ m, j ∈ n define a mapping

ϕi j : m →V given byϕi j(k) = δikb j for all k ∈ m.

Here δik is Kronecker delta

Let B ′ ={

ϕi j : i ∈ m, j ∈ n}

. Clearly O(B ′) = mn.

We shall now show that B ′ is a basis for the function space V m.

B ′is l.i.: Let ai j ∈ F, i ∈ m, j ∈ n be such that

m∑

i=1

n∑j=1

ai jϕi j = ̂0 (zero function)


Then,(

mΣ

i=1

nΣ

j=1ai jϕi j

)

(k) = ̂0(k) for all k ∈ m

⇒mΣ

i=1

nΣ

j=1ai jϕi j(k) = 0V

⇒mΣ

i=1

nΣ

j=1ai jδikb j = 0V

⇒nΣ

j=1

(

mΣ

i=1ai jδik

)

bj = 0V

⇒nΣ

j=1ak jb j = 0V [∵ δik = 1 for i = k]

⇒ ak j = 0 for all j ∈ n [∵ B = {b1, . . . ,bn} is a basis for V ]

Thus, ak j = 0 for all k ∈ m and for all j ∈ n.

So,m∑

i=1

n∑j=1

ai jϕi j = ̂0 ⇒ ai j = 0 for all i ∈ m and for all j ∈ n.

Therefore, B ′ is a linearly independent set.

B ′spans V m: Let f : m →V be an arbitrary mapping in V m. Then for each k ∈ m, f (k) ∈V .

Since B is a basis for V . Therefore, f (k) ∈ V is expressible as a linear combination ofvectors in B. Let

f (k) =nΣ

j=1λk jb j (i)

For any k ∈ m, we have(

mΣ

i=1

nΣ

j=1λi jϕi j

)

(k) =mΣ

i=1

nΣ

j=1λi jϕi j(k) =

mΣ

i=1

nΣ

j=1λi jδikb j

⇒

(

mΣ

i=1

nΣ

j=1λi jϕi j

)

(k) =nΣ

j=1

(

mΣ

i=1λi jδik

)

bj

⇒

(

mΣ

i=1

nΣ

j=1λi jϕi j

)

(k) =nΣ

j=1λk jb j = f (k) [From (i)]

∴ f =mΣ

i=1

nΣ

j=1λi jϕi j


Hence, B1 is a basis for S+ T . Consequently, we havedimS+ T = number of vectors in B1 = p+q−r=dimS+dimT−dim(S∩T ).

Q.E.D.

COROLLARY. If S and T are subspaces of a finite-dimensional vector space V (F) such thatS∩T = {0V}, then

dim(S+ T ) = dimS+ dimT.

PROOF. We have,S∩T = {0V}

⇒ dim(S∩T ) = 0Thus, dim(S+ T ) = dimS+ dimT −dim(S∩T )

⇒ dim(S+ T ) = dimS+ dimT.

Q.E.D.

In three-dimensional Euclidean space R3, let S = the xy-plane and T = the yz-plane. Clearly,S and T are subspaces of R3 and dimS = 2, dim T = 2. Clearly, S∩T = y-axis, whose dimen-sion is 1, and S+T = R3. Therefore, dimS+T = dimR3 = 3 and dimS+dimT −dim(S∩T ) =2 + 2 − 1 = 3. So, dim(S + T ) = dimS + dimT − dim(S ∩ T ). This verifies the result ofTheorem 9. On the other hand, if S = the xy-plane and T = z-axis, then S∩ T = {0V} andS + T = R3. Also, dimR3 = dim(S + T ) = 3 = 2 + 1 = dimS + dimT . This verifies the resultof corollary to Theorem 9.

THEOREM-10 If a finite dimensional vector space V (F) is direct sum of its two subspacesS and T , then

dimV = dimS+ dimT.

PROOF. Since V is direct sum of S and T , therefore, V = S+ T and S∩T = {0V}.

Now, S∩T = {0V} ⇒ dim(S∩T ) = 0By Theorem 9, we have

dim(S+ T ) = dimS+ dimT −dim(S∩T )⇒ dim(S+ T ) = dimS+ dimT.

Q.E.D.

THEOREM-11 Every subspace of a finite dimensional vector space has a complement.

PROOF. Let V be a finite dimensional vector space over a field F , and let S be a subspace ofV . Further, let dimV = n and dimS = m.

Let B1 = {b1,b2, . . . ,bm} be a basis for S. Then B is a linearly independent set of vectors inB. So it can be extended to a basis for V . Let B = {b1,b2, . . . ,bm,v1,v2, . . . ,vn−m} be extendedbasis for V . Let T be a subspace of V spanned by the set B2 = {v1,v2, . . . ,vn−m}


We shall now show that T is the complement of S.Let v be an arbitrary vector in V . Then,

v =mΣ

i=1λibi +

n−mΣ

i=1µivi for some λi,ui ∈F [∵ B is a basis for V ]

Since B1 and B2 are bases for S and T respectively. Therefore,mΣ

i=1λibi ∈ S and,

n−mΣ

i=1µivi ∈ T

⇒ v =mΣ

i=1λibi +

n−mΣ

i=1µivi ∈ S+ T

Thus, v ∈V ⇒ v ∈ S+ T∴ V ⊂ S+ TAlso, S+ T ⊂V .Hence, V = S+ T .Let u ∈ S∩T . Then, u ∈ S and u ∈ T .Now,

u ∈ S ⇒ There exists λ1, . . . ,λm ∈ F such that u =mΣ

i=1λibi [∵ B1 is a basis for S]

and,

u ∈ T ⇒ There exists µ1, . . . ,µn−m ∈ F such that u =n−mΣ

i=1µivi [∵ B2 is a basis for T ]

∴m∑

i=1λibi =

n−m∑

i=1µivi

⇒ λ1b1+, . . . ,+λmbm +(−µ1)v1 + · · ·+(−µn−m)vn−m = 0V⇒ λ1 = λ2 = · · · = λm = 0 = µ1 = · · · = µn−m [∵ B is a basis for V ]⇒ u = 0b1 + · · ·+ 0bm = 0V

Thus, u ∈ S∩T ⇒ u = 0V∴ S∩T = {0V}

Hence, V = S⊕T . Consequently, T is complement of S.Q.E.D.

THEOREM-12 Let V (F) be a finite dimensional vector space, and let S be a subspace of V .Then, dim(V/S) = dimV −dimS.

PROOF. Let dimV = n,dim S = m. Then n � m. Let B1 = {v1,v2, . . . ,vm} be a basis for S.Then, B1 is a linearly independent set of vectors in V . So it can be extended to a basis for V ,say B2 = {v1, . . . ,vm,b1, . . . ,bn−m}.


We shall now show that the set B = {b1 + S,b2 + S, . . . ,bn−m + S} is basis for V/S.To prove that B is a basis for V/S, we have to show that B is a linearly independent set and

B spans V/S.

B is a l.i set in V/S: Let λ1, . . . ,λn−m be scalars in F such that

λ1(b1 + S)+ · · ·+ λn−m(bn−m + S) = S (zero of V/S)

Then,

(λ1b1 + S)+ · · ·+(λn−mbn−m + S) = S [By def. of scalar multiplication on V/S]

⇒ (λ1b1 + · · ·+ λn−mbn−m)+ S = S⇒ λ1b1 + · · ·+ λn−mbn−m ∈ S [∵ S is additive subgroup of (V,+) ∴ S+ u = S ⇔ u ∈ S]

⇒ λ1b1 + · · ·+ λn−mbn−m is a linear combination of B1 = {v1, . . . ,vm} [∵ B1 is a basis for S]

⇒ λ1b1 + · · ·+ λn−mbn−m = µ1v1 + · · ·+ µmvm for some µi ∈ F⇒ λ1b1 + · · ·+ λn−mbn−m +(−µ1)v1 + · · ·+(−µm)vm = 0V

⇒ λ1 = λ2 = · · · = λn−m = 0 = µ1 = · · · = µm [∵ B2 is linearly independent]

Thus, only the trivial linear combination of vectors of B equals to the null vector of V/S.Hence, B is a linearly independent set in V/S.

B spans V/S: Let u+ S be an arbitrary vector in V/S. Then,

u ∈V

⇒ u is a linear combination of B2 [∵ B2 is a basis for V ]

⇒ there exist λ1, . . . ,λm, µ1, . . . ,µn−m in F such that

u = λ1v1 + · · ·+ λmvm + µ1b1 + · · ·+ µn−mbn−m

⇒ u = v+ µ1b1 + · · ·+ µn−mbn−m, where v = λ1v1 + · · ·+ λmvm ∈ S [∵ B1is a basis for S]

⇒ u = (µ1b1 + · · ·+ µn−mbn−m)+ v [∵ vector addition is commutative on V ]

⇒ u+ S = (µ1b1 + · · ·+ µn−mbn−m)+ v+ S

⇒ u+ S = (µ1b1 + · · ·+ µn−mbn−m)+ S [∵ v ∈ S ∴ v+ S = S]

⇒ u+ S = (µ1b1 + S)+ · · ·+(µn−mbn−m + S)

⇒ u+ S = µ1(b1 + S)+ · · ·+ µn−m(bn−m + S)

⇒ u+ S is a linear combination of B.

Since u+ S is an arbitrary vector in V/S. Therefore, every vector in V/S is a linear combi-nation of B. Consequently, B spans V/S,


Hence, B is a basis for V/S.∴ dim(V/S) = Number of vectors in B = n−m = dimV −dim S.

Q.E.D.

CO-DIMENSION. Let V (F) be a finite dimensional vector space and let S be a subspace of V .Then codimension of S is defined to be the dimension of the quotient space V/S.

Clearly, Codimension S = dimV −dimS. [By Theorem 12]

MAXIMAL LINEARLY INDEPENDENT SET. Let V (F) be a vector space. Then a linearly indepen-dent set of vectors in V is said to be a maximal linearly independent set if every set of vectorsin V containing it as a proper subset is a linearly dependent set.

THEOREM-13 If S is a maximal linearly independent set of vectors in a finite dimensionalvector space V (F), then S spans V , i.e. [S] = V .

PROOF. Since V is a finite dimensional vector space. Therefore, S is a finite set. LetS = {v1, . . . ,vn}, and let v =

n∑

i=1λivi be an arbitrary vector in V . Since S is a maximal linearly

independent set. Therefore, the larger set S′ = {v1, . . . ,vn,v} must be linearly dependent.Hence, we can find scalars µ1,µ2, . . . ,µn,µ not all zero such that

µ1v1 + · · ·+ µnvn + µv = 0V

We claim that µ �= 0, becauseµ = 0

⇒ µ1v1 + · · ·+ µnvn = 0V where at least one µi �= 0.

⇒ S is a linearly dependent⇒ a contradiction.

Now, µ1v1 + · · ·+ µnvn + µv = 0V

⇒ µv = −nΣ

i=1µivi

⇒ v =nΣ

i=1(µiµ−1)vi [∵ µ �= 0 ∴ µ−1 ∈ F]

⇒ v is a linear combination of SHence, S spans V .

Q.E.D.

THEOREM-14 Let V (F) be a finite dimensional vector space. Then a set of n linearly inde-pendent vectors in V is maximal iff n= dimV .

PROOF. Let S = {v1, . . . ,vn} be a set of n linearly independent vectors in V .


Hence, B1 ∪B2 is a basis for V .

Now we shall show that V = S1 ⊕S2. For this, it is sufficient to show that S1 ∩S2 = {0V}.

Let v ∈ S1 ∩S2. Then, v ∈ S1 and v ∈ S2.

Since B1={v1, . . . ,vm} is a basis for S1. Therefore, there exist scalars λ1, . . . ,λm∈F such that

v = λ1v1 + · · ·+ λmvm.

Since B2={u1, . . . ,un} is a basis for S2 and v ∈ S2. Therefore, there exist scalars µ1,µ2, . . . ,µn∈

F such that

v = µ1u1 + · · ·+ µnun

∴ λ1v1 + · · ·+ λmvm = µ1u1 + · · ·+ µnun

⇒ λ1v1 + · · ·+ λmvm +(−µ1)u1 + · · ·+(−µn)un = 0V

⇒ λ1 = · · · = λm = 0 = µ1 = · · · = µn [∵ B1 ∪B2 is a basis for V ]⇒ v = 0v1 + · · ·+ 0vm = 0V

Thus, v ∈ S1 ∩S2 ⇒ v = 0V

∴ S1 ∩S2 = {0V}

Hence, V = S1 ⊕S2

Q.E.D

THEOREM-17 Let S and T be subspaces of a finite dimensional vector space V such thatV = S⊕T . Let B1 and B2 be bases for S and T respectively. Then B1 ∪B2 is basis for V anddimV = dimS+ dimT .

PROOF. Since V is a finite dimensional vector space over a field F , therefore, S and T arealso finite dimensional vector spaces over field F . Let dimS = m and dimT = n. Further letB1 = {b1,b2, . . . ,bm} be a basis for S and B2 = {bm+1,bm+2, . . . ,bm+n} be a basis for T . LetB = B1 ∪B2 = {b1,b2, . . . ,bm,bm+1, . . . ,bm+n}.

We shall now establish that B is a basis for V .

B is linearly independent: Let λ1,λ2, . . . ,λm,λm+1, . . . ,λm+n be scalars in F such that

λ1b1 + λ2b2 + · · ·+ λmbm + λm+1bm+1 + · · ·+ λm+nbm+n = 0V⇒ λ1b1 + · · ·+ λmbm = −(λm+1bm+1 + · · ·+ λm+nbm+n) (i)

⇒ λ1b1 + · · ·+ λmbm ∈ T[

∵ −nΣ

i=1λm+ibm+i ∈ T

]


THEOREM-20 A vector space V (F) is the direct sum of its n(≥ 2) subspaces S1,S2, . . . ,Sn i.e.V = S1 ⊕S2 ⊕S3 ⊕·· ·⊕Sn iff

(i) V = S1 + S2 + · · ·+ Sn (ii) Si ∩n∑j �=ij=i

S j = {0V} for each i = 1,2, . . . ,n

PROOF. First suppose that V = S1 ⊕S2 ⊕·· ·⊕Sn. Then, V = S1 + S2 + · · ·+ Sn

Let v be an arbitrary vector in Si ∩n∑j=1j �=i

S j. Then,

v ∈ Si ∩n∑j=1j �=i

S j ⇒ v ∈ Si and v ∈n∑j=1j �=i

S j

Now, v ∈ Si

⇒ v = vi for some vi ∈ Si

⇒ v = 0V + 0V + · · ·+ vi + 0V + · · ·+ 0V [∵ 0V ∈ Sj for all j]

and, v ∈nΣj=1j �=i

S j

⇒ There exist vectors v j ∈ Sj, j = 1,2, . . . ,n and j �= i such that

v = v1 + v2 + · · ·+ vi−1 + vi+1 + · · ·+ vn

But, V = S1 ⊕ S2 ⊕·· ·⊕ Sn. So, each vector in V has unique representation as the sum ofvectors in S1,S2, . . . ,Sn.

∴ v1 = 0V = v2 = · · · = vi−1 = vi = vi+1 = · · · = vn

⇒ v = 0V

Since v is an arbitrary vector in Si ∩n∑j=1j �=i

S j.

∴ Si ∩n∑j=1j �=i

S j = {0V}

Conversely, let S1,S2, . . . ,Sn be subspaces of V (F) such that (i) and (ii) hold. We have toprove that V is the direct sum of its subspaces S1,S2, . . . ,Sn.

Let v be an arbitrary vector in V . Then,

v ∈ S1 + S2 + · · ·+ Sn [∵ V = S1 + S2 + · · ·+ Sn]⇒ v = v1 + v2 + · · ·+ vn, where vi ∈ Si; i = 1,2, . . . ,n


B spans V: Let v be an arbitrary vector in V . Since V is the direct sum of S1, . . . ,Sk. Therefore,there exist unique v1 ∈ S1, . . . ,vk ∈ Sk such that

v = v1 + · · ·+ vk.

Since for each i ∈ k, Bi is a basis for Si. Therefore, each vi can be expressed as a linearcombination of vectors in Bi. Consequently, v is expressible as a linear combination of vectorsin

k∪

i=1Bi = B. Hence, B spans V .

B is a l.i. set in V: To show that B is a linearly independent set, let λ11, . . . ,λ1

n1 ,λ21, . . . ,λ2

n2 , . . . ,λi

1, . . . ,λini , . . . ,λ

kk, . . . ,λ

knk

be scalars in F such that(

λ11b1

1 + · · ·+ λ1n1

b1n1

)

+ · · ·+(

λi1bi

1 + · · ·+ λini

bini

)

+ · · ·+(

λkkbk

1 + · · ·+ λknk

bknk

)

= 0V

⇒(

λ11b1

1 + · · ·+ λ1n1b1

n1

)

+ · · ·+(

λkkbk

1 + · · ·+ λknk

bknk

)

= 0V + 0V + · · ·+ 0V︸︷︷︸

k−terms

(i)

Since V is the direct sum of S1, . . . ,Sk. Therefore, each vector in V is uniquely expressible as asum of vectors one in each Si. Consequently, (i) implies that

λi1bi

1 + · · ·+ λini

bini

= 0V for all i = 1,2, . . . ,k⇒ λi

1 = · · · = λini = 0 for all i = 1,2, . . . ,k [∵ Bi is a basis for Si]

Hence, B is a linearly independent set in V .Thus, B is a basis for V .(ii) ⇒ (i).

Let B =k∪

i=1Bi be a basis for V . Then we have to show that V is the direct sum of S1,S2, . . . ,Sk.

Let v be an arbitrary vector in V . Then,

v =kΣ

i=1

(

λi1bi

1 + · · ·+ λinib

ini

)

[∵ B =k∪

i=1Bi is a basis for V ]

⇒ v =kΣ

i=1vi ,where vi = λi

1bi1 + · · ·+ λi

nibi

ni∈ Si [∵ Bi is a basis for Si]

Thus, v is expressible as a sum of vectors one in each Si. Since v is an arbitrary vector in V .Therefore, each vector in V is expressible as a sum of vectors one in each Si. Consequently,V = S1 + S2 + · · ·+ · · ·+ Sk.

Now we shall show that the expression of each vector in V as a sum of vectors one in eachSi is unique. If possible, let v ∈V be such that

v =kΣ

i=1

(

λi1bi

1 + · · ·+ λinib

ini

)

and v =kΣ

i=1

(

µi1bi

1 + · · ·+ µinib

ini

)

.


Then, kΣ

i=1

(

λi1bi

1 + · · ·+ λinib

ini

)

=kΣ

i=1

(

µi1bi

1 + · · ·+ µinib

ini

)

⇒kΣ

i=1

{

(λi1 −µi

1)bi1 +(λi

2 −µi2)bi

2 + · · ·+(λini−µi

ni)bi

ni

}

= 0V

⇒ λi1 −µi

1 = 0, λi2 −µi

2 = 0, . . . , λini−µi

ni= 0 for i = 1,2, . . . ,k [∵ B =

k∪

i=1Bi is l.i.]

⇒ λi1 = µi

1, . . . ,λini

= µini

for i = 1,2, . . . ,kThus, v is uniquely expressible as a sum of vectors one in each Si. Therefore, every vector in Vis expressible as a sum of vectors one in each Si.

Hence, V is the direct sum of S1,S2, . . . ,Sk.Q.E.D.

THEOREM-22 Let S and T be two subspaces of a vector space V (F) such that S∩T = {0V}.If B1 and B2 be bases of S and T respectively, then show that B = B1 ∪B2 is a basis of S + Tand B1 ∩B2 = ϕ.

PROOF. Let B1 = {b1,b2, . . . ,bm} and B2 = {b′1,b′2, . . . ,b′n} be bases of S and T respectively.Then, B1 and B2 are linearly independent subsets of S and T respectively. Therefore, B1∩B2 isalso linearly independent and B1 ∩B2 ⊆ S∩T .∴ B1 ∩B2 ⊆ {0V}

⇒ B1 ∩B2 = ϕ [∵ B1 and B2 are l.i. and any l.i. set cannot have null vector in it]We shall now show that B = B1 ∪B2 is a basis for S+ T .B = B1 ∪B2 is a l.i. subset of S+T: Clearly, S and T are subsets of S+ T . So, S∪T ⊆ S+ T .∴ B1 ∪B2 ⊆ S∪T ⊆ S+ T.

Let λi(i = 1,2, . . . ,m) and µj( j = 1,2, . . . ,n) be scalars in F such thatmΣ

i=1λibi +

nΣ

j=1µjb ′

j = 0V

⇒mΣ

i=1λibi = −

nΣ

j=1µjb′j = u (say)

Clearly,mΣ

i=1λibi ∈ S and −

nΣ

j=1µjb′j ∈ T

⇒ u ∈ S and u ∈ T

⇒ u ∈ S∩T

⇒ u = 0V [∵ S∩T = {0V}]

∴mΣ

i=1λibi = 0V and, −

nΣ

j=1µjb′j = 0V

⇒ λi = 0(i = 1,2, . . . ,m),µj = 0( j = 1,2, . . . ,n) [∵ B1 and B2 are linearly independent]


SOLUTION (i) Since dimR3 = 3. So, a basis of R3(R) must contain exactly 3 elements.Hence, B1 is not a basis of R3(R).(ii) Since dimR3 = 3. So, B2 will form a basis of R3(R) if and only if vectors in B2 are

linearly independent. To check this, let us form the matrix A whose rows are the given vectorsas given below.

A =

1 1 11 2 32 −1 1

Since non-zero rows of a matrix in echelon form are linearly independent. So, let us reduce Ato echelon form.

A ∼

1 1 10 1 20 −3 −1

Applying R2 → R2 −R1,R3 → R3 −2R1

⇒ A ∼

1 1 10 1 20 0 5


Clearly, the echelon form of A has no zero rows. Hence, the three vectors are linearly indepen-dent and so they form a basis of R3.

(iii) Since (n+1) or more vectors in a vector space of dimension n are linearly dependent.So, B3 is a linearly dependent set of vectors in R3(R). Hence, it cannot be a basis of R3.

(iv) The matrix A whose rows are the vectors in B4 is given by

A =

1 1 21 2 55 3 4

Since non-zero rows of a matrix in echelon form are linearly independent. So, let us reduce Ato echelon form.

A ∼

1 1 20 1 30 −2 −6

Applying R2 → R2 −R1,R3 → R3 −5R1

⇒ A ∼

1 1 20 1 30 0 0


The echelon form of A has a zero row, hence the given vectors are linearly dependent and soB4 does not form a basis of R3.

EXAMPLE-2 Let V be the vector space of all 2× 2 matrices over a field F. Prove that V hasdimension 4 by finding a basis for V .


SOLUTION Let E11 =

[

1 00 0

]

, E12 =

[

0 10 0

]

, E21 =

[

0 01 0

]

and E22 =

[

0 00 1

]

be four

matrices in V , where 1 is the unity element in F .Let B = {E11,E12,E21,E22}.We shall now show that B forms a basis of V .B is l.i.: Let x,y,z, t be scalars in F such that

xE11 + yE12 + zE21 + tE22 = 0

⇒

[

x yz t

]

=

[

0 00 0

]

⇒ x= y = z = t = 0

So, B is l.i.

B spans v: Let A =

[

a bc d

]

be an arbitrary matrix in V . Then,

A = aE11 + bE12 + cE21 + dE22

⇒ A is expressible as a linear combination of matrices in B.So, B spans V . Thus, B is a basis of V .Hence, dimV = 4.

EXAMPLE-3 Let v1 = (1, i,0),v2 = (2i,1,1),v3 = (0,1 + i,1− i) be three vectors in C3(C).Show that the set B = {v1,v2,v3} is a basis of C3(C).

SOLUTION We know that dimC3(C) = 3. Therefore, B will be a basis of C3(C) if B is a linearlyindependent set. Let x,y,z ∈C such that

xv1 + yv2 + zv3 = 0⇒

(

x+ 2iy,xi+ y+ z(1+ i),y+ z(1− i))

= (0,0,0)

⇒ x+ 2iy = 0,xi+ y+ z(1+ i) = 0,y+ z(1− i) = 0⇒ x= y = z = 0.

Hence, B is a basis of C3(C).

EXAMPLE-4 Determine whether (1,1,1,1),(1,2,3,2),(2,5,6,4),(2,6,8,5) form a basis ofR4(R). If not, find the dimension of the subspace they span.

SOLUTION Given four vectors can form a basis of R4(R) iff they are linearly independent asthe dimension of R4 is 4.

The matrix A having given vectors as its rows is

A =

1 1 1 11 2 3 22 5 6 42 6 8 5


Since non-zero rows of a matrix in echelon form are linearly independent. So, let us reduce Ato echelon form

A ∼

1 1 1 10 1 2 10 3 4 20 4 6 3

Applying R2 → R2 −R1,R3 → R3 −2R2,R4 → R4 −2R1

⇒ A ∼

1 1 1 10 1 2 10 0 −2 −10 0 −2 −1


⇒ A ∼

1 1 1 10 1 2 10 0 −2 −10 0 0 0


The echelon matrix has a zero row. So, given vectors are linearly independent and do not forma basis of R4. Since the echelon matrix has three non-zero rows, so the four vectors span asubspace of dimension 3.

TYPE II ON EXTENDING A GIVEN SET TO FORM A BASIS OF A GIVEN VECTOR SPACE

EXAMPLE-5 Extend the set {u1 = (1,1,1,1),u2 = (2,2,3,4)} to a basis of R4.

SOLUTION Let us first form a matrix A with rows u1 and u2, and reduce it to echelon form:

A =

[

1 1 1 12 2 3 4

]

⇒ A ∼

[

1 1 1 10 0 1 2

]

We observe that the vectors v1 = (1,1,1,1) and v2 = (0,0,1,2) span the same space as spannedby the given vectors u1 and u2. In order to extend the given set of vectors to a basis of R4(R),we need two more vectors u3 and u4 such that the set of four vectors v1,v2,u3,u4 is linearlyindependent. For this, we chose u3 and u4 in such a way that the matrix having v1,v2,u3,u4 asits rows is in echelon form. Thus, if we chose u3 = (0,a,0,0) and u4 = (0,0,0,b), where a,bare non-zero real numbers, then v1,u3,v2,u4 in the same order form a matrix in echelon form.Thus, they are linearly independent, and they form a basis of R4. Hence, u1,u2,u3,u4 also forma basis of R4.

EXAMPLE-6 Let v1 = (−1,1,0),v2 = (0,1,0) be two vectors in R3(R) and let S = {v1,v2}.Extend set S to a basis of R3(R).


SOLUTION Let A be the matrix having v1 and v2 as its two rows. Then,

A =

[

−1 1 00 1 0

]

Clearly, A is in echelon form. In order to form a basis of R3(R), we need one more vector suchthat the matrix having that vector as third row and v1,v2 as first and second rows is in echelonform. If we take v3 = (0,0,a), where a(�= 0) ∈ R, then matrix having its three rows as v1,v2,v3is in echelon form.Thus, v1,v2,v3 are linearly independent and they form a basis of R3(R).

REMARK. Sometimes, we are given a list of vectors in the vector space Rn(R) and we want tofind a basis for the subspace S of Rn spanned by the given vectors, that is, a basis of [S]. Thefollowing two algorithms help us for finding such a basis of [S].

ALGORITHM 1 (Row space algorithm)Step I Form the matrix A whose rows are the given vectorsStep II Reduce A to echelon form by elementary row operations.Step III Take the non-zero rows of the echelon form. These rows form a basis of the subspace

spanned by the given set of vectors.In order to find a basis consisting of vectors from the original list of vectors, we usethe following algorithm.

ALGORITHM 2 (Casting-out algorithm)Step I Form the matrix A whose columns are the given vectors.Step II Reduce A to echelon form by elementary row operations.Step III Delete (cast out) those vectors from the given list which correspond to columns

without pivots and select the remaining vectors in S which correspond to columnswith pivots. Vectors so selected form a basis of [S].

TYPE III ON FINDING THE DIMENSION OF SUBSPACE SPANNED BY A GIVEN SET OF VEC-TORS

EXAMPLE-7 Let S be the set consisting of the following vectors in R5:

v1=(1,2,1,3,2),v2 =(1,3,3,5,3),v3 = (3,8,7,13,8),v4 = (1,4,6,9,7),v5 = (5,13,13,25,19)

Find a basis of [S] (i.e the subspace spanned by S) consisting of the original given vectors.Also, find the dimension of [S].

SOLUTION Let A be the matrix whose columns are the given vectors.Then,

A =

1 1 3 1 52 3 8 4 131 3 7 6 133 5 13 9 252 3 8 7 19


Let us now reduce A to echelon form by using elementary row operations.

A ∼

1 1 3 1 50 1 2 2 30 2 4 5 80 2 4 6 100 1 2 5 9

Applying R2 → R2 −2R1,R3 → R3 −R1,R4 → R4 −3R1 and R5 → R5 −R1

⇒ A ∼

1 1 3 1 50 1 2 2 30 0 0 1 20 0 0 2 40 0 0 3 6

Applying R3 → R3 −2R2,R4 → R4 −2R2,R5 → R5 −R1

⇒ A ∼

1© 1 3 1 50 1© 2 2 30 0 0 1© 20 0 0 0 00 0 0 0 0


We observe that pivots (encircled entries) in the echelon form of A appear in the columnsC1,C2,C4. So, we “cast out” the vectors u3 and u5 from set S and the remaining vectorsu1,u2,u4, which correspond to the columns in the echelon matrix with pivots, form a basis of[S]. Hence, dim[S] = 3.

EXAMPLE-8 Let S be the set consisting of following vectors in R4:

v1 = (1,−2,5,−3),v2 = (2,3,1,−4),v3 = (3,8,−3,−5).

(i) Find a basis and dimension of the subspace spanned by S, i.e. [S].(ii) Extend the basis of [S] to a basis of R4.

(iii) Find a basis of [S] consisting of the original given vectors.

SOLUTION (i) Let A be the matrix whose rows are the given vectors. Then,

A =

1 −2 5 −32 3 1 −43 8 −3 −5

Let us now reduce A to echelon form. The row reduced echelon form of A is as given below.

A ∼

1 −2 5 −30 7 −9 20 0 0 0


Let B denote the matrix whose rows are v1,v2,v3. Then, B =

1 −1 −33 −2 −82 1 −3

⇒ B =

1 −1 −33 −2 −82 1 −3

⇒ B ∼

1 −1 −30 1 10 3 3


⇒ B ∼

1 −1 −30 1 10 0 0


⇒ B ∼

1 1 −10 1 10 0 0


Clearly, A and B have the same row canonical form. So, row spaces of A and B are equal.Hence, [S] = [T ].

EXERCISE 2.9

1. Mark each of the following true or false.

(i) The vectors in a basis of a vector space are linearly dependent.(ii) The null (zero) vector may be a part of a basis.

(iii) Every vector space has a basis.(iv) Every vector space has a finite basis.(v) A basis cannot have the null vector.

(vi) If two bases of a vector space have one common vector, then the two bases arethe same.

(vii) A basis for R3(R) can be extended to a basis for R 4(R).(viii) Any two bases of a finite dimensional vector space have the same number of

vectors.(ix) Every set of n+ 1 vectors in an n-dimensional vector space is linearly dependent.(x) Every set of n+1 vectors in an n-dimensional vector space is linearly independent.

(xi) An n-dimensional vector space can be spanned by a set n−1 vectors in it.(xii) Every set of n linearly independent vectors in an n-dimensional vector space is a

basis.


(xiii) A spanning set of n vectors in an n-dimensional vector space is a basis.(xiv) [v]B is independent of B.(xv) If S and T are subspaces of a vector space V (F), then

dimS < dimT ⇒ S � T .

2. Find three bases for the vector space R2(R) such that no two of which have a vector incommon.

3. Determine a basis for each of the following vector spaces:

(i) R2(R)

(ii) R(√

2) = {a+ b√

2 : a, b ∈ R} over R(iii) Q(

√2) = {a+ b

√2 : a, b ∈ Q} over Q

(iv) C over R(v) C over itself

(vi) Q(i) = {a+ ib : a, b ∈ Q} over Q

4. Which of the following subsets B form a basis for R3(R)?

(i) B = {(1,0,0),(1,1,0),(1,1,1)}

(ii) B = {(0,0,1),(1,0,1)(1,−1,1),(3,0,1)}

(iii) B = {(1,2,3),(3,1,0),(−2,1,3)}

(iv) B = {(1,1,1),(1,2,3),(−1,0,1)}

(v) B = {(1,1,2),(1,2,5),(5,3,4)}

5. Which of the following subsets B form a basis for the given vector space V ?

(i) B = {(1,0),(i,0),(0,1),(0, i)},V = C2(R).(ii) B = {(1, i,1+ i),(1, i,1− i),(i,−i,1)},V = C3(C).

(iii) B = {(1,1,1,1),(2,6,4,5),(1,2,1,2),(0,3,2,3)},V = R 4(R).(iv) B = {1, sinx, sin2 x, cos2 x},V = C[−π,π]

(v) B = {x−1, x2 + x−1, x2 − x+ 1},V = Space of all real polynomials of degree atmost two.

6. Extend the set S = {(1,1,1,1),(1,2,1,2)} to a basis for R 4(R).7. Extend the set S = {(3,−1,2)} to two different bases for R3(R).8. Consider the set S of vectors in R3(R) consisting of v1 = (1,1,0),v2 = (0,2,3),

v3 = (1,2,3),v4 = (1,−2,3).

(i) Prove that S is linearly dependent.(ii) Find a subset B of S such that B is a basis of [S].

(iii) Find dim[S].


9. Let S be the subspace of R5 spanned by the vectors. v1=(1,2,−1,3,4),v2=(2,4,−2,6,8),v3 = (1,3,2,2,6),v4 = (1,4,5,1,8),v5 = (2,7,3,3,9). Find a subset of these vectors thatform a basis of S.

10. Prove that the set S = {a + 2ai : a ∈ R} is a subspace of the vector space C(R). Find abasis and dimension of S. Let T = {a−2ai : a ∈ R}. Prove that C(R) = S⊕T .

11. Extend each of the following subsets of R3 to a basis of R3(R):(i) B1 = {(1,2,0)} (ii) B2 = {(1,−2,0),(0,1,1)}

12. Find a basis of the vector space R4(R) that contains the vectors (1,1,0,0),(0,1,0,−1).13. Find four matrices A1,A2,A3,A4 in R2×2(R) such that A2

i = Ai for 1 ≤ i ≤ 4 and{A1,A2,A3,A4} is a basis of R2×2.

14. Let S be the set of all matrices A in R2×2 which commute with the matrix[

1 12 2

]

. Prove

that S is a subspace of R2×2. Find dimS.15. Let V = R3×3,S be the set of symmetric matrices and T be the set of skew-symmetric

matrices in V . Prove that S and T are subspaces of V such that V = S ⊕ T . Finddimensions of S and T .

16. Express R5(R) as a direct sum of two subspaces S and T with dimS = 2,dimT = 3.17. If S and T be two subspaces of a finite dimensional vector space V (F) such that

dimV = dimS+ dimT and S∩T = {0V}.

Prove that V = S⊕T.

18. Let V (F) be a finite dimensional vector space, and let S1,S2, . . . ,Sk be subspaces of Vsuch that

V = S1 + S2 + · · ·+ Sk and dimV = dimS1 + dimS2 + · · ·+ dimSk.

Prove that V = S1 ⊕S2 ⊕·· ·⊕Sk.19. Show by means of an example that under certain circumstances, it is possible to find three

subspaces S1,S2,S3 of a finite dimensional vector space V (F) such that V = S1 ⊕ S2 =S2 ⊕S3 = S3 ⊕S1. What does this imply about the dimension of V ?

20. Give an example of an infinite dimensional vector space V (F) with a subspace S suchthat V/S is a finite dimensional vector space.

21. Construct three subspaces S1,S2,S3 of a vector space V such that V = S1 ⊕S2 = S1 ⊕S3but S2 �= S3

22. Give an example of a vector space V (R) having three different non-zero subspacesS1,S2,S3 such that

V = S1 ⊕S2 = S2 ⊕S3 = S3 ⊕S1.

23. Let S = {(a,b,0) : a,b ∈ R} be a subspace of R3. Find its two different complements inR3.


24. Determine a basis of the subspace spanned by the vectors:v1 = (1, 2, 3), v2 = (2, 1, −1), v3 = (1, −1, −4), v4 = (4, 2, −2)

25. Show that the vectors v1 = (1, 0, −1), v2 = (1, 2, 1), v3 = (0, −3, 2) form a basis forR3. Express each of the standard basis vectors e(3)

1 , e(3)2 , e(3)

3 as a linear combination ofv1, v2, v3.

ANSWERS

1. (i) F (ii) F (iii) T (iv) F (v) T (vi) F (vii) F (viii) T (ix) T (x) F (xi)F (xii) T (xiii) T (xiv) F (xv) F

2. B1 = {(1,0),(0,1)} ,B2 = {(1,−1),(2,3)} ,B3 = {(−2,1),(2,−3)}.Vectors (a1,b1) and (a2,b2) form a basis of R2(R) iff a1b2 �= a2b1.

3. (i) B = {(1,0),(0,1)} (ii) B = {1,√

2} (iii) B = {1,√

2} (iv) B = {1, i}(v) B = {1, i} (vi) B = {1, i}

4. (i)

5. (ii), (iii)

6. {(1,1,1,1),(1,2,1,2),(0,0,1,0),(0,0,0,1)}

7. {(3,−1,2),(0,1,0),(0,0,1)},{(3,−1,2),(0,−1,0),(0,0,2)}

8. (ii) B = {(1,1,0),(0,2,3)} (ii) dim[S] = 2

9. {v1,v3,v5}

10. {1+ 2i}, 1

11. (i) {(1,2,0),(0,1,0),(0,0,1)} (ii) {(1,−2,0),(0,1,1),(0,0,1)}

12. {(1,1,0,0),(0,1,0,−1),(0,0,1,0),(0,0,0,1)}

13. A1 =

[

1 00 0

]

,A2 =

[

0 00 1

]

,A3 =

[

1 10 0

]

,A4 =

[

0 01 1

]

15. dimS = dimT = 3

16. S = {(a,b,0,0,0) : a,b ∈ R}, T = {(0,0,a,b,c) : a,b,c ∈ R}

19. dimV is an even integer.

21. V = R2(R),S1 = {(a,0) : a ∈ R},S2 = {(0,a) : a ∈ R} and S3 = {(a,a) : a ∈ R}

23. (i) S1 = {(0,0,a) : a ∈ R} (ii) {(a,a,a) : a ∈ R}

24. {v1, v2}.

25. e(3)1 =

710

v1 +3

10v2 +

15

v3, e(3)2 = −

15

v1 +15

v2 −15

v3, e(3)3 =

−310

v1 +3

10v2 +

15

v3


(ii) If v = (a,b), thenv = xv1 + yv2

⇒ (a,b) = (x+ 2y,x+ 3y)⇒ x+ 2y = a,x+ 3y = b⇒ x= 3a−2b,y = −a+ b

Hence, the coordinate matrix [v]B of v relative to the basis B is given by [vB] =

[

3a−2b−a+ b

]

EXAMPLE-2 Find the coordinate vector of v = (1,1,1) relative to the basis B = {v1 = (1,2,3),v2 = (−4,5,6), v3 = (7,−8,9)} of vector space R3.

SOLUTION Let x,y,z ∈ R be such that

v = xv1 + yv2 + zv3

⇒ (1,1,1) = x(1,2,3)+ y(−4,5,6)+ z(7,−8,9)

⇒ (1,1,1) = (x−4y+ 7z, 2x+ 5y−8z, 3x+ 6y+ 9z)⇒ x−4y+ 7z = 1, 2x+ 5y−8z = 1, 3x+ 6y+ 9z = 1

⇒ x=7

10, y =

−215

, z =−130

Hence [v]B =

710−215−130

is the coordinate vector of v relative to basis B.

EXAMPLE-3 Find the coordinates of the vector (a,b,c) in the real vector space R3 relative tothe ordered basis (b1,b2,b3), where b1 = (1,0,−1), b2 = (1,1,1), b3 = (1,0,0).

SOLUTION Let λ,µ,γ ∈ R be such that

(a,b,c) = λ(1,0,−1)+ µ(1,1,1)+ γ(1,0,0).

⇒ (a,b,c) = (λ+ µ+ γ,µ,−λ+ µ)

⇒ λ+ µ+ γ = a, µ = b, −λ+ µ= c⇒ λ = b− c, µ = b, γ = a−2b+ c.

Hence, the coordinates of vector (a,b,c) ∈ R3 relative to the given ordered basis are (b −

c, b, a−2b+ c).


EXAMPLE-4 Let V be the vector space of all real polynomials of degree less than or equal totwo. For a fixed c ∈ R, let f1(x) = 1, f2(x) = x+ c, f3(x) = (x+ c)2. Obtain the coordinates ofc0 + c1x+ c2x2 relative to the ordered basis ( f1, f2, f3) of V .

SOLUTION Let λ,µ,γ ∈ R be such that

c0 + c1x+ c2x2 = λ f1(x)+ µf2(x)+ γ f3(x).

Then,

c0 + c1x+ c2x2 = (λ+ µc+ γc2)+ (µ+ 2γc)x+ γx2

⇒ λ+ µc+ γc2 = c0, µ+ 2γc = c1, γ = c2 [Using equality of two polynomials]⇒ λ = c0 − c1c+ c2c2, µ= c1 −2c2c,γ = c2.

Hence, the coordinates of c0 + c1x+ c2x2 relative to the ordered basis ( f1, f2, f3) are(c0 − c1c+ c2c2,c1 −2c2c, c2).

EXAMPLE-5 Consider the vector space P3[t] of polynomials of degree less than or equal to 3.(i) Show that B = {(t −1)3,(t −1)2,(t −1),1} is a basis of P3[t].(ii) Find the coordinate matrix of f (t) = 3t3 −4t2 + 2t −5 relative to basis B.

SOLUTION (i) Consider the polynomials in B in the following order:

(t −1)3,(t −1)2,(t −1),1

We see that no polynomial is a linear combination of preceding polynomials. So, thepolynomials are linearly independent, and, since dimR3[t] = 4. Therefore, B is a basis of P3[t].

(ii) Let x,y,z,s be scalars such that

f (t) = x(t −1)3 + y(t −1)2 + z(t −1)+ s(1)

⇒ f (t) = xt3 +(−3x+ y)t2 +(3x−2y+ z)t +(−x+ y− z+ s)⇒ 3t3 −4t2 + 2t −5 = xt3 +(−3x+ y)t2 +(3x−2y+ z)t +(−x+ y− z+ s)⇒ x= 3,−3x+ y = −4,3x−2y+ z = 2,−x+ y− z+ s = −5⇒ x= 3,y = 13,z = 19,s = 4

Hence, the coordinate matrix of f (t) relative to the given basis is [ f (t)]B =

313194

.

EXAMPLE-6 Let F be a field and n be a positive integer. Find the coordinates of the vector(a1,a2, . . . ,an) ∈ Fn relative to the standard basis.


λu+ µv = λ(0, b1, c1)+ µ(0, b2, c2)

⇒ λu+ µν = (0, λb1 + µb2, λc1 + µc2) ∈ S.








S

YO

X

Z


EXAMPLE-3 Let a1,a2,a3 be fixed elements of a field F. Then the set S of all triads (x1,x2,x3)of elements of F, such that a1x1 + a2x2 + a3x3 = 0, is a subspace of F3.

SOLUTION Let u = (x1,x2,x3),v = (y1,y2,y3) be any two elements of S. Then x1,x2,x3,y1,y2,y3 ∈ F are such that

a1x1 + a2x2 + a3x3 = 0 (i)and, a1y1 + a2y2 + a3y3 = 0 (ii)

Let a,b be any two elements of F . Then,

au+ bv = a(x1,x2,x3)+ b(y1,y2,y3) = (ax1 + by1,ax2 + by2,ax3 + by3).

Now,

a1(ax1 + by1)+ a2(ax2 + by2)+ a3(ax3 + by3)

= a(a1x1 + a2x2 + a3x3)+ b(a1y1 + a2y2 + a3y3)

= a0+ b0 = 0 [From (i) and (ii)]

∴ au+ bv = (ax1 + by1,ax2 + by2,ax3 + by3) ∈ S.Thus, au+ bv ∈ S for all u,v ∈ S and all a,b ∈ F . Hence, S is a subspace of F3.

EXAMPLE-4 Show that the set S of all n× n symmetric matrices over a field F is a subspaceof the vector space Fn×n of all n×n matrices over field F.

SOLUTION Note that a square matrix A is symmetric, if AT = A.Obviously, S is a non-void subset of Fn×n, because the null matrix 0n×n is symmetric.Let A,B ∈ S and λ,µ∈ F . Then,

(λA + µB)T = λAT + µBT

⇒ (λA + µB)T = λA + µB [∵ A,B ∈ S ∴ AT = A,BT = B]

⇒ λA + µB ∈ S.

Thus, S is a non-void subset of Fn×n such that λA + µB ∈ S for all A,B ∈ S and for allλ,µ∈ F .

Hence, S is a subspace of Fn×n.

EXAMPLE-5 Let V be a vector space over a field F and let v ∈V . Then Fv = {av : a ∈ F} is asubspace of V .

SOLUTION Since 1 ∈ F , therefore v = 1 v ∈ Fv. Thus, Fv is a non-void subset of V .Let α,β ∈ Fv. Then, α = a1v, β = a2v for some a1,a2 ∈ F .


⇒

[

2 34 −7

]

=

[

x+ y+ z+ t x− y− zx+ y x

]

⇒ x+ z+ t = 2,x− y− z = 3,x+ y = 4,x= −7⇒ x= −7,y = 11,z = −21, t = 30

Hence, the coordinate matrix of A relative to the basis B2 is [A]B2 =

711

−2130

.

EXERCISE 2.10

1. Find the coordinate matrices of vectors u = (5,3) and v = (a,b) relative to the orderedbasis B = {(1,−2),(4,−7)} of vector space R2(R).

2. The vectors v1 = (1,2,0),v2 = (1,3,2),v3 = (0,1,3) form a basis of R3(R). Find thecoordinate matrix of vector v relative to the ordered basis B = {v1,v2,v3}, where

(i) v = (2,7,−4) (ii) v = (a,b,c)

3. B ={

t3 + t2, t2 + t, t + 1,1}

is an ordered basis of vector space P3[t]. Find thecoordinate matrix f (t) relative to B where:

(i) f (t) = 2t3 + t2 −4t + 2 (ii) f (t) = at3 + bt2 + ct + d

4. B=

{[

1 11 1

]

,

[

1 −11 0

]

,

[

1 10 0

]

,

[

1 00 0

]}

is an ordered basis of vector space R2×2(R).

Find the coordinate matrix of matrix A relative to B where:

(i) A =

[

3 −56 7

]

(ii) A =

[

a bc d

]

5. Consider the vector space R2(t) of all polynomials of degree less than or equal to 2.Show that the polynomials f1(t) = t − 1, f2(t) = t − 1, f3(t) = (t − 1)2 form a basis ofP2[t]. Find the coordinate matrix of polynomial f (t) = 2t2−5t +9 relative to the orderedbasis B = { f1(t), f2(t), f3(t)}.

6. Find the coordinate vector of v = (a,b,c) in R3(R) relative to the ordered basis B ={v1,v2,v3}, where v1 = (1,1,1),v2 = (1,1,0),v3 = (1,0,0).


ANSWERS

1. [u]B =

[

−4111

]

, [v]B =

[

−7a−4b2a+ b

]

2. (i) [v]B =

−1113

−10

(ii) [v]B =

7a−3b+ c−6a+ 3b− c

4a−2b+ c

3. (i) [ f (t)]B =

2−1−2

2

(ii) [ f (t)]B =

ab− c

a−b+ c−a+ b− c+ d

4. (i) [A]B =

dc−d

b+ c−2da−b−2c+ 2d

(ii) [A]B =

7−1−13

10

5. [ f (t)]B =

3−4

2

6. [v]B =

cb− ca−b


⇒

[

2 34 −7

]

=

[

x+ y+ z+ t x− y− zx+ y x

]

⇒ x+ z+ t = 2,x− y− z = 3,x+ y = 4,x= −7⇒ x= −7,y = 11,z = −21, t = 30

Hence, the coordinate matrix of A relative to the basis B2 is [A]B2 =

711

−2130

.

EXERCISE 2.10

1. Find the coordinate matrices of vectors u = (5,3) and v = (a,b) relative to the orderedbasis B = {(1,−2),(4,−7)} of vector space R2(R).

2. The vectors v1 = (1,2,0),v2 = (1,3,2),v3 = (0,1,3) form a basis of R3(R). Find thecoordinate matrix of vector v relative to the ordered basis B = {v1,v2,v3}, where

(i) v = (2,7,−4) (ii) v = (a,b,c)

3. B ={

t3 + t2, t2 + t, t + 1,1}

is an ordered basis of vector space P3[t]. Find thecoordinate matrix f (t) relative to B where:

(i) f (t) = 2t3 + t2 −4t + 2 (ii) f (t) = at3 + bt2 + ct + d

4. B=

{[

1 11 1

]

,

[

1 −11 0

]

,

[

1 10 0

]

,

[

1 00 0

]}

is an ordered basis of vector space R2×2(R).

Find the coordinate matrix of matrix A relative to B where:

(i) A =

[

3 −56 7

]

(ii) A =

[

a bc d

]

5. Consider the vector space R2(t) of all polynomials of degree less than or equal to 2.Show that the polynomials f1(t) = t − 1, f2(t) = t − 1, f3(t) = (t − 1)2 form a basis ofP2[t]. Find the coordinate matrix of polynomial f (t) = 2t2−5t +9 relative to the orderedbasis B = { f1(t), f2(t), f3(t)}.

6. Find the coordinate vector of v = (a,b,c) in R3(R) relative to the ordered basis B ={v1,v2,v3}, where v1 = (1,1,1),v2 = (1,1,0),v3 = (1,0,0).


ANSWERS

1. [u]B =

[

−4111

]

, [v]B =

[

−7a−4b2a+ b

]

2. (i) [v]B =

−1113

−10

(ii) [v]B =

7a−3b+ c−6a+ 3b− c

4a−2b+ c

3. (i) [ f (t)]B =

2−1−2

2

(ii) [ f (t)]B =

ab− c

a−b+ c−a+ b− c+ d

4. (i) [A]B =

dc−d

b+ c−2da−b−2c+ 2d

(ii) [A]B =

7−1−13

10

5. [ f (t)]B =

3−4

2

6. [v]B =

cb− ca−b

chapter chapter 2 vector spaces 2.1 introduction in various practical and theoretical problems, we...

Documents