Ces,

4. Vector Spaces
. External composition.
Let Uand s be two
ito be anerternal non-empty sets. A
mapping f:UxS ’S is
Iix S has a definite composition
of U with S. Each ordered
image f(u, s) in S. pair (u, s)
Examples.
et S be the set of all
real matrices of order m X n. The mapping
,-RxS ’ S defined by c*A = cA,
mDOsition of IR with S. For each real cE R, A ¬S is an external
mmxn matrix in S. * is called numberc and each A in S, cA is
ral numbers. 'multiplication of mxn matrices by
4Let Gbe a group under a
:ZxGG be defined by multiplicative
n*a = a, nEcomposition.
Z,a ¬ G.
Let a mapping
For each integer
and each a in G, ar is an element in G. *
iZ with G. is an external composition
4.2. Vector Space over a Field.
Let Vbe a non-empty set
and
be a binary composition on V. Let
E+.) be afield and let Obe an external composition of Fwith V.
Sa mapping from V xVto V. is a mapping from FxV to V
ad it
maps the ordered pair (c, a) of F x Vto a definite element cOa
Vis said to be
e a vector space (or alinear space ) over the field Fif
fVidl. oaße
wing conditions are satisfied.
Vfor all a, ßE V; dot
V2. aß= B afor all a, ßE V;
V3. a (B) =(a B) yfor all a, B, yEV;
V4. there exists an element in Vsuch that a 0=afor all a ¬ V;
V5. for each a in Vthere exists an element -a in Vsuch that
5(-a)=
V6. ; cOa¬V for all ce F, all a ¬ V;
eo(doa) (c.d) oa for all c, dE F, all a¬ V;
co (a p) ==(cO a) (eo) for all c¬ F, all a, Be V;
 HIGHER ALGEBRA
112

y9. (c+ d) Oa =(cO a) (d o a) for all c, d e F, all ¬V;

E V, being the identity
V10. 1 Oa=afor all o element
in F,
space is denoted ( E, t,., 0, O). The
by (V,
(The vector
are called
of
the ground feld (orthe field of scalars) of the vector space.) elements
vectors and the elements FF are called 8calars, Fis óf V
calet
Four symbols +,., O, O denote four ddifferent compositions
FF. :Fx F ’ F;0: Vx ’ V;O: Fx V’V. W
dispense with and use only +to denote both types of addition. We shalWel
and denote c.d in F by co
shall dispense with O and . both and denote
cOa in V by co.
Therefore a non-empty set Vis said to form a vector space Over a
held F if
) there is a binary composition + on V, called 'addition', satisfvine
the conditions
Vl. a t E V for all a, Be V,
V2. a +B=Btafor all a, B¬ V,
V3. a + (B+7) = (a+B)+y for all a, B, E V,
V4. there exists an element 0 in V such that a+0=a for all oin
V.
V5. for each a in V there exists an element -a in V such th¡t
a+ (-a) =0;
'multi
and (ii) there is an external comnposition of F with V, called
plication by the elements of F' satisfying the conditions -
V6. caEV for all c¬ F, all aE V,
V7. c(da) = (cd)a for all c, dE F, all a ¬ V,
V8. c(a+ B) = ca + cß for all c E F, all a, BE V,
V9. (c + d)a = ca + da for all c, dE F, all a ¬ V,
Vi0. la = ofor all a ¬ V, 1being the identity element in F. called
of F are
The elements of Vare called vectors and the elementscalled,'multipl
scalars. The external composition of F with Vis also
cation by scalars'. vector
a compler
In particular, Vis said to be a real vector space (or
space) if the field F be R ( or C). element
Since F is a field, it has the zero element 0 and the identity
the
identity
1/Since Vis a commutative group with respect to addition, to
Thisissaid
avoid
element with respect to addition is the zero element in V. order to
0. In vector zero
be the zerO vector or the null vector and is denoted byand
confusion, two zero elements (the scalar zero in F the contusion
inoccurs
) the
will appear But nosuch
with different symbols. when
symbol 0 will be used for both the zero elements.
 VECTOR SPACES 113

Realvectorspace.
non-emmpty set Vis said to form a real vector space (or a vector
NCO
Ae v e rthe field R) if
thereisa binary composition
(+) on V, called 'addition', satisfying
)
heonditions
atBE Vfor all a, ß E V;
VI. tB=/B+a for all a, B E V;
V2. a (B+
)= (a +B) +y for all a, B,y¬ V;
V3. a + exists an element 0in Vsuch that a+0= afor all a E V;
V4.thhere
each a in Vthere exists an element -a in Vsuch that
V5. for
at(-a)= 0; 'multi
called
(ii) there is an external composition of Rwith V,
plication by real
numbers' satisfying the conditions
a E V:
V6. ca E V for all c ¬ R, all
a e V;
V7. c(do)=(cd)a for all c, deR, all all
+ B) = ca + cB for all c E R, a, BE V:
V8. c(a
all a e V:
V9. (c + d)a = ca + da for all c, d ¬ R, element in R.
V10. la=a for all a¬ V, 1being the identity
the elements of R are called
The elements of V are called vectors and
salars. R is said to bethe ground field (or the field of scalars) of the
vector space V.

Bxamples.
I. Real vector space R".
: a; E R}.
Let V be the set of all ordered n-tuples {(a1, a2,... , an)
'addition', defined by
Let + be a composition on V,called
1, a2,., an) + (b1, b2,...,bn) = (aj + b1, az t b2
'multiplication by real
external composition of R with V, called
Muabers' be defined by
cE R.
da1, a2,,an) = (ca1, ca2, -. , can),satisfied. Therefore V is a real
Then the conditions VI-V10 are
vetor space and it is denoted by R". denoted by .
is the null vector of R" and it is
(0,0similar
Ina ,. ,0) manner the yector spaces R²,
R®, R,... are defined. The
Ritself forms a real vector space.
Real vector space C.
complexnumbers {a+ib: a ¬ R, bE R, i= /(-1)}.
Of all
LetLet + be a composition on C, called
'addition', defined by
(+ib) + (c+id) = (a +c) + i(b d);
 VECTOR SPACES 127

44. Linear dependence and Linear independence.

Afinitesct of vectors 1,09, n} of avector space Vover afield
saidtobe linearly dependent in Vif there exist
Fis in Fsuch that scalars C1, C2,,Cn
allzero,
C1a1t C22 t Cha, =0 (i)
The set is said to be linearly independent in Vif the equality (i) is
onty when C1 = C2
An arbitrary set S of vectors of a vector space V over afield Fis said
linearly dependent in Vif there exists a finite subset of S which is
jgcarly dependent in V,
Aset of vectors which is not linearly dependent is said to be linearly
independent.
Although the definition of linear dependence or linear independence
wfers to a set of vectors, we shall also state that the individual vectors
d,a2,.. , an are linearly dependent or independent.
Theorem 4.4.1.(A superset of a linearly dependent set of vectors in a
vector space V over a field Fis linearly dependent.
Proof. Case (i). Let S be a linearly dependent set of vectors containing
afinite number of elements a1, a2,..., an: Let T be a superset of S.
Sbeing a finite subset of T and being linearly dependent, T is linearly
dependent, by definition.
Case (ii). Let S be a linearly dependent set of vectors containing an
minite number of elements and T be a superset of S. Since S is linearly
dependent, there exists a finite subset P of S such that P is linearly
linearly
pendent. P being a linearly dependent finite subset of T, T is
dependent, by definition.
This completes the proof.
vectors in a
4.4.2. Asubset of a linearly independent set of
TheorenSpaCe V over a field F is linearly
independent.
subset
Prooj. Let S be a linearlyindependent set of vectors and P be a
o S. If P be linearly dependent then Sbeing a superset of the linearly
dependent, by Theorem 4.4.1l. But it
dependentso. This proves that Pis linearly independent.
set P, must be linearly

Note. The set o is linearly independent.

in a vector
the null vector
Theoren 4.4.3. Aset of vectors containing
Space Vover afield Fiis linearly dependent.
 HIGHER ALGEBRA
128

Proof. Let S = (0). The set Sislinearly dependent since the relation
holds for a non-ero scalat c.
co
arbitrary set of vectors containing the null vector .
be ana superset ofS, is linearly dependent by Theorem 44.1.
Let TTbeing
Then
Theorem 4.4.4. The set consisting of a single non-zero vector in a
over a field F is linearly independent.,
vector space V 4.2.1.(iv).
the Theorem
The proof follows from

Worked Examples.
if the set of vectors {(2, 1, 1)(1,2, 2), (1, 1, 1)} is line
k Exanmine
dépendent in R®
Let a = (2, 1, 1), = (1, 2, 2), y= (1, 1,1).
C2tc2Y = 6, where c1. Co. co ars
Let us consider the relation c1+
real nunmbers.
0).
Then c1(2, 1, 1) + cz(1, 2,2) + ca(1,1, 1) = (0,0,
0, C1 + 2c2 + C3 = 0.
Therefore 2c1 + c t cg =0, C1+ 2c2 +t c3 =
+2c2 + C3 =0.
or equivalently, 2c1 + C2 t Cg = 0, C1
where k is a real number.
The solution is C1 =-k, C2 =-k, Cz = 3k,
zero, such that ca+
Since k is arbitrary, there exist c1, C2, C3, not all
-3.
C2B+ c3Y =0. For example, cË = 1, cz = 1, Cg =
Therefore the given set of vectors is linearly dependent in R".
1,2), (2, 2, 1)} is linearly
2. Prove that the set of vectors {(1,2, 2), (2,
Andependent in R3.
Let a = (1,2,2), ß = (2, 1,2), y = (2,2, 1).
(2.C3
Let us consider the relation c1a + caB+ cay= 0, where C1,
real numbers.
Then c(1, 2, 2) + cz(2, 1,2) + ca(2, 2, 1) = (0, 0,0). +=0.
Therefore cË + 2e2 + 2c3 = 0, 2c1 t c2 + 2c3 == 0, 2c + 2c2
This is a homogeneous system of three equations in c2C
1 2 2 =5f0.
1 2
The co-efficient determinant of the system is
2
andthe
C3
By Cramner's rule, there exists a unique solution for c, C2,
solution is C1 = 0,C2 = 0, C3 = 0. inR
independent
This proves that the given set of vecters is linearly
 VECTOR SPACES 129

Theoremn4.4.5. If the set of vectors {a1, a2,...,an} in a vector space

loverafield F be linearly dependent, then at least one of the vectors of
can be
expressed as a linear combination of the remaining others.
set
the
Coprerselu. ifone of the vectors of the set {a1,a2,.,an} be alinear
pombination of the remaining others, the set is linearly
dependent.
Pw. Sincethe set {a1, a2,.,an} is linearly dependent, there exist
in F, not all zero, such
Nalars C1, C2, . ,Cp that
Ca1t C202t * t C n an = 0.

.20. ThenceFand c, 'c; = 1,1 being the identity element in

F. Cz2
c,a, =C101 - - Cj-1j-1 C;+10j+1 Cnan
Therefore a; =
C;-C1a1-C202Cj-1@j-1- Cj+10j+1-*-Cn an]
= dËa1 + dza2 t +dj-10j-1 t dj+1aj+1 t +dnan
bere d, =-c;;EF, i=1,2,.. ,j - 1,j +1,...,n.
This shows that aj is a linear combination of the vectors

Conversely, let one of the vectors, say aj, is a linear combination of the
other vectors of the set.

Then a, =T101+ T202t+r;-1@;-1+rj+1aj+1 ++rnan, for

Some scalars r; E F, i=1,2,...,j- 1,j + 1,...,n.
Therefore T1a1+ T2@2+trj-1aj-1+(-1)a; +ri+1aj+1t +
Tnln =..
Since the above equality holds for scalars , 2 , ,Tj-1,-l,
Ti+l,,n in F and since one of them at least is non-zero, the set
Ia1,2,.., a,} is linearly dependent.
Ihis completes the proof.
CorolBary. Two vectors a,B in a vector space Vover a field F are
hearly dependent if at least one of them is ascalar multiple of the
other.

Examples.he set Of vectors S= {a, 2a, 3} of a real vector space Vis linearly
dependent., since 2a E S and 2a( 2a + 08) is a linear combination of
Temai n ing vectors of S.
learet of vectors S= fa,B,7.B + } of a real vector space V is
since +yES and + Y(= Oat + ) is a linear
dependent,
Combination of the remaining vectors of S.
 VECTOR SPACES 131

Using(i) we see that each element of S is alinear combination of the

ectors
ofr
Therefore L(S) C L(T), by Theorem 4.3.8.
Hence L(T) = L(S) = V. Thus Vis spanned by a proper subset T of S.
completesthe proof.
This
Note. If Vbe the null space then S ={0} is a generating set of V. V
Canalso be considered as L{} and the set is an improper subset of S.
Theorem 4.4.8. (If S ={a1, a2,.. . , an} be alinearly independent set
generators of a vector space V, then no proper subset of S can be a
Sf V.
Sanningset of
Prof. Iff; possible, let a proper subset T ofS be aspanning set of V. Let
S- T.
some a; belong to
Since L(T) =Vanda, E V a, can be expressed as a linear combina
Ren ofthe vectorsof T. Therefore the set TUfa;} is linearly dependent.,
by Theorem 4.4.5.
The set S being a superset of TU{aj} becomes a linearly dependent
t and this is a contradiction which proves the theorem.

Worked Examples (continued).

KXet aj = (1,2, 0), az = (3, -1, 1), ag = (4,1, 1). Show that the set
Y=fa, a3, as} is linearly dependent. Apply Deletion Theorem to find
aproper subset of S that can generate L(S).
Let Cj1+ C22t Cz3=0, where c1, C2, C3 are real numbers.
Then c1 (1,2,0) + cz(3, -1, 1)+ ca(4, 1, 1) = (0,0,0).
T'herefore c + 3c, + 4ca = 0, 2c1 C2 t C3=0, C2 t c3 =0.
S gives C1 = C =-C3. Taking C1 = 1, we have cz =1,C3 =-1
and then
a1t 2
This shows that the set S is linearly dependent.
As a3 = a1 t a2, 3 can be deleted from the generating set of L(S),
etion theorem. That is, La1, a2} = La1, a2, a3f:
Aote. Since a1 can also be deleted by the theorem and
,a3} = a1, a2, ag}. Similarly, L{a1, a3} =L{a1, a2, as}.
A5.
LetBasiVs be a vector
and
dimension of a vector space.
space Over a field F. Vis said to be finitely
yenneerraatteidn.g, or finite dimensional if there exists a finite set of vectors in
V.
Otherwise, Vis said to be infinite
dimensional.
V
132 HIGHER ALGEBRA

The null space {0} is finite dimensional, since it is

empty set . generated by the
We shall be mainly concerned with finite dimensional vector
this treatise.
spaces in
Basis of a vector space.
Definition. Let V be a vector space over a field F. A set S
of
in V is said to be a basis of V if vector
(i) S is linearly independent in V, and
Li) Sgenerates V. i,e. 1S) V
Theorem 4.5.1. There exists a basis for every finitely generated vector
space.
Proof. Case 1. Let V be a finitely generated vector
the null space. space other than
Let S = {a1, a2, . . . ,an} be a finite set of generators of V. If S is
linearly independent in V, then S itself is a basis of V and the theorem
is done.
If S is linearly dependent, then we can delete, by
at least one vector fromn S and obtain a proper Deletion theorem.
subset S; (of S) spanning
the same space V. If S, is linearly independent in V, then S is a
of V and the theorem is done. basis
If S, be not linearly independent then we
repeat the process of dele
tion and finally obtain, after k(< ) steps of deletion, a subset Sk which
is linearly independent in V and which also
spans V.
This is possible, because S is a finite set of n elements and in the
extreme unfavourable case we can come, after n-1steps of deletion, to
a subset Sn-1 containing only one non-zero vector that generates Vand
that is linearly independent.
Therefore our assertion that Sk is alinearly independent set for some
k(<n) is true and hence it is a basis of V.
Case 2. Let V= {0}. Since the set is linearly independent and
L(0) ={0}, is a basis of V.
This completes the proof.

Examples.
KThe set E = {e1 =(1,0), ¬2 = (0, 1)} is a basis of the vector SpaceR
relation
To prove linear independence of the set E, 1let us consider the
C1¬1 + C2¬2 = b for real c1, C2.
 VECTOR SPACES 133

ci(1,0) +- cz(0,1) =(0,0).

Thengives C1 = 0, C2
This = 0 and this proves that the set E is linearly
ndependent.
Let =(a, b) be an arbitrary vector of R2, Then a, b ¬ Rand can
expressedas a(1, 0). + b(0, 1) = a¬j + bez.
This shows that &E L(E) and therefore R2 C L(E).
AgainEcR? and L(E) being the smallest subspace of R² containing
LE) CR2 It follows that L(E) R2.
Thus the set E fulfils both the conditions for a basis of R?.
Note.This basis E is said to be the standard basis of R.
The set E= {e1 = (1, 0, 0), ¬2 = (0, 1, 0), e3 = (0,0, 1)} is a basis of
R3 Eis the standard basis of R3.
he vector space
The set E ={¬1 = (1,0,0,... 0), ¬2 = (0, 1, 0, ...,0),... , ¬n =
o0..... 1)} is a basis of R". Eis the standard basis of R

Worked Example.
LPpove that the set S= {(0,1,1), (1, 0, 1), (1, 1, 0)} is a basis of R®.
Let aj = (0, 1, 1), ay = (1,0, 1), a3 = (1, 1, 0).
To examine linear independence of the set, let us consider the relation
Cja1 + C2a2 + C3a3 = , where C1, C2, C3 are real numbers.
Then c1(0, 1, 1) + c(1, 0, 1)+ c3(1, 1,0)= (0,0,0).
This gives Cz + c3 = 0, c1 + c3 == 0, C1 + cz = 0.
Therefore c = C = Ca = 0and this proves that the set S is linearly
Bdependent.
Let us examine if
= (a, b, c) be an arbitrary vector of R,
ELIS).
If possible, let F for real r1,r2, r3-
Then T2 +T3 F a, rË t r3 = b, T1 + r = C.
T3.
This is a system of three equations in r1, r2,
non-homogeneous 0 1
1 0 1 =2#0.
The Co-efficient determinant of the system is 1 1
By Cramer's rule, there exists a unique solution for ri,T2, T3.
This proves that E L(S) andtherefore R$ C L(S).
subspace of R containing
Again ScR and L(S) being thesmallest
= R3.
us the SetR.Sfulfls both theL(S)
Consequently, conditions for a basis of R3.