2.

Groups

2.1. Binary composition.

Let A be a non-empty set. A binary
tion on A is a mapping f: AxA> A. composition (or a binary opera-
Therefore a binary composition
f assigns a definite element of A to each ordered
This mapping f is generally denoted by the
pair of elements of A.
symbol o. For a pair of el-
èments a, b in A, the image of (a, b) under the
binary composition o is
denoted by aob. The image of the element (b, a) is
obviously boa.
The symbols like *, +,., D, O are also used to
denote a binary com-
position.

Examples.
1. On the set Z, let o stand for the binary composition 'addition'. Then
203 5 , 4o 4 0.
2. On the set Z, let o stand for the
Then 203 = 6, 300 = 0.
binary composition 'multiplication'.
3. On the set Z, let stand for the binary
o
composition 'subtraction'.
Then 302 =1, 103=-2.
4. Let a binary composition o be defined on the set Z by aob = a + 2b,
a,bE Z. Then 203 = 8, 300 =3.
5. Let a binary composition * be defined the set Q by a *b
on =
zab,
a,beQ. Then 2 5 = 5, 3*8 = 12.
A binary composition is said to be defined on a non-empty set A if
o

aob E A for all a, b in A. In this case the set A is said to be closed under
(or closed with respect to) the binary composition o.
For example, the set N is closed under 'addition', since a E N,b ¬
N>a+b E N. But the set N is not closed under 'subtraction', because
a-b does not belong to N for some a, b in N.

Definition. Let o be a binary composition on a set A.

o is said to be commutative if aob =
boa for all a, b E A.
o is said to be associative if ao(boc) ==(aob)oc for all a, b,c E A.

HA-5
66 HIGHER ALGEBRA

Examples ( continued).
6. Addition on the set R is both commutative and associative. Minle:
cation on the set R is both commutative and associative, but
on the set R is neither commutative nor associative. subtract
7. Let S be a non-empty set and P(5) be the power set of S. Then
(union), n (intersection) and A (symmetric ditference) are binary com
positions on P(S) and each of these is commutative and associative o
on
P(S).
8. Let M2(R) be the set of all 2 x 2 real matriccs. Let o stand
for
multiplication of matrices. Then o is associative but not commutative.

9. Let n be a positive integer and let us consider thc

p-equivalence classes
of the relation p on Z defined by "a pb if and only if a -b
is divisible by
n for a, b E Z. There are rn classes cl(0), cl(1), cl(2), .. . , cl(n-1). These
are also called the classes of residues of integers modulo n. We use the
notation to denote the class cl(a). Let Z, be the set of residue classes
{0,1,2,,n-1).
We define a binary com1position +, called addition modulo n, on the
set Zn by +b a +6. =

In order that this definition/may be valid we must check that it is

well defined, i.e., it is independent of the choice of ropresentatives of
the equivalence classes. Therefore we have to show that if a, a', 6,b are
integers such that cl(a') = cl(a) and cl(8) = cl(6) then a +b = a +
a a a - a = kn for some integer k,
b = b = b - b = pn for some integer p.

Therefore (a+b) - (a' +b) = tn, where t(=k +p) is an integer.

Consequently, a + b= a ' +&.

This proves that 'addition modulo n' is a well defined binary compo-
sition on the set Zn
In like manner, we define a binary composition, called multiplication
modulo n, on the set Zn by .b = ab and we can prove similarly that
it is a well defined composition on the set Zn, ie., if cl(a') = cl(a) and
cl(6) =
cl(b) then ab = a'b.
Both these compositions are commutative as well as associative, be
cause

a+b a =
+b= b+ a =
b+ for all , be Zn;
a.b ab = = ba b Zn: =
b. for all , e
and a+(6 +)= a+b+e =a+ (b+c) (a +6) +c =

a+6+T = ( +b) +T for all a. b, E Zni

 GROUPS 67

a(b.c) = a.bc a(bc) =(ab)c

ab.c =(.5).c for all , b,e Zn
Composition table.
When A is a non-empty finite set, a binary composition o on the set
can be defined by a table, called the composition table. If the number
of elements in A be n, the table has n rows and n columns, one for each
element of the set. The elements of the set are listed on the topmost row
and the leftmost column in the same order.
I fA = {a1, a 2 , . , ans then a;oaj appears on the table in the ith row
and yth column. The n^ entries of the table are all elements of A, since
A is closed under o.

If the table be symmetric about the principal diagonal (i.e., if a;oa,

a, oa) then is commutative.
o

For example, the table for the binary composition of 'addition modulo
3 on the set Z3 is

Here the composition is commutative.

Exercises 6

1. Examine whether the composition o defined on the set is (i) commutative,

) associative.
(a) o on Z defined by aob =
a +b+1,a,bE Z;
(b) o on Q defined by aob =ab+1, a,bEQ
(c) o on R defined by aob a + 2b,a, b¬ R;
=

d) o on R defined by aob |ab|, a, b E R;

=

(e) o on Z x Z defined by (a,b)o(ce,d) (a

c,b-d), (a, b), (c,d) e
=
- Zx Z
A, B E Ma(R).
4) oon M2 (R) defined by AoB (AB BA),
-

on a set S. Let T be a subset of

composition
.Let o be a n associa.tive binary
defined by T {a E S:a or =r oa for
=
all a ¬S)}. Prove that T is closed

under o.
How many different binary compositions
3. Let S be set of two elements.
a
can be defined on S? How many different
commutative binary compositions
can be defined on S?
 GROUPS 75

2.6. Groups.

A non-empty set G is said to form a

group with rospect to a binary
composition o, if

(i) G is closed under the composition «

(ii) o is associative,
(iü) there exists an element e in G such that eoa = aoe = a for all a
in G.
(iv) for each element a in G, there exists an clement o' in G such that
a oa = aoa = e.

The group is denoted by the symbol (G, o).

The element e is said to be an identity element in the group. We shall
Drove that there is only one such element in the group and therefore e
will be said to be the identity element.
The element a' is said to be an inverse of a. We shall prove that each
the
element a has only one inverse and therefore a' will be said to be
nverse of a.

Definition. A group (G, o) is said to be a commutative group or an

abelian group (after the name of Norwegian mathematician N. Abel) if o
is commutative.

identity element.
Theorem 2.6.1. A group (G, o) contains only one

Proof. Let e, f be two identity elements in the group

Then eoa =aoe a and foa aof = a for all a in G.

=
=

of e;
We have eof f, by the property =

e, by the property
of f.
and also eof =

and this proves uniqueness

of identity element
Therefore e =
f
Note. The element in (G, o) is denoted by eg
identity
each element has only one inverse.
Theorem 2.6.2. In a group (G, o)
a', a" be two inverses of a. Then a'oa
=
aoa =
e
Proof. Let a E G and
element.
and a"oa = aoa" =
e, e being the identity
is associative.
since o
We have a'o(aoa") = (a'oa)oa",
a' and (c'oa)oa" eoa" a".= =

a'o(aoa") a'oe
=
But =

proves that the inverse

of a is unique.
Therefore a' a'" and this =

is denoted by a , Therefore for each a in G,

Note. The inverse of a

aoa =a-loa = e holds.

 HIGHER ALGEBRA
76

Theorem 2.6.3. Cancellation

laws.

In a group(G,o), for all a, b, c E G,

cancellation law);
(i) aob aoc implies b=c (left
=

cancellation law).
(ii) boa =coa implies b=c (right
Proof. Since a e G, aleG.
() aob aoc =
ao(aob) =alo(aoc)
(aloa)ob (aoa)»c, = since o is associative
eob = eoc, e being the identity element
b= C.
(boa)oa = (coa)oa
(ii) boa = coa

bo(aoa) co(aoa), = since o is associative

boe = coe, e being the identity element
b= C.

Theorem 2.6.4. In a group (G, o), for all a, b in G,

each of the equations aox = b and yoa = b has a unique solution in G.

Proof. Since a, b e G, aob e G.

We have ao(a"ob) = (aoa)ob, since o is associative
= egob =b.

This shows that aob is a solution of the equation aor = b. We prove

that this solution is unique.
Let there be two solutions 71,T2 in G ofthe equation aor =b. Then
aor = aor2. By left cancellation law, this implies 1 = T2 and this
proves uniqueness of the solution.
Again, boaeG. And (boa-1)oa = bo(a-loa) = b.
This shows that boa is a solution of the equation yoa = b. The
uniqueness of the solution follows from right cancellation law.

Theorem 2.6.5. In a group (G, o), (aob)-l =bloal

for all a, b E G.
Proof. Let a,b E G. Then a,b-1, aob, b-loa-l all belong to G.
Now (boa)o(aob)=bo(aoa)]ob, since o is associatin
(6oeg)ob = b-lob = eG.
Also, (aob)o(boa) =lao(bob-)]oa-1, since is associativ o 18 associativ

We have (oa-1)o(aob) (aoec)oa aoa eg.

= =

(bloa-)o(aob)=(aob)o(b-loa-l)
ave

= ea.
It follows that boa"l is the inverse of aob. That
b-oa-1. is, (aob)=
 GROUPS 77

Examples.
The set Z forms a commutative group with respect to addition.
Let a,b E Z. Then a +bE Z. This shows that the set Z is closed
under addition.

(i) Addition is associative on the set R. Z being a subset of R,

addition is associative on Z.
(ii) 0 E 2 and 0 + a a + 0 =a for all a in Z. Therefore 0 is the
identity element.

(iv) Let a EZ. Then -a E Z. -a is the inverse of a, since -a + a=

a+(-a) = 0.

(v) Addition is commutative on R. Z being a subset of R, addition is

commutative on Z.
Therefore (Z, +) is a commutative group.
2. (Q.+) is a commutative group.

3. (R.+) is a commutative group.

4. (C,+) is a commutative group.

5. Let m be a positive integer. Then mZ denotes the set of all integral
multiples of m. mZ= {0, tm,t2m, +3m,... ...

For example, 22 = {0,t2,t4, +6,...

3Z= {0,+3,t6,+9,...
We prove that (2Z, +) is a commutative group.

Let a, b E 2Z. Then a 2p.b p, q ¬ Z. Therefore

2q for some
(1) = =

atb= 2(p + g) ¬ 2Z, since p +q E Z. This shows that the set 2Z is

closed under +.
associative on Z and 22 is a subset of Z. Therefore
i) Addition is
addition is associative on 22.
for all a ¬ 2Z. Therefore 0 is the
(ii) 0 E 2Z and 0+ a =a+0 = a

identity element.
Then - a ¬ 2Z and (-a) +a =
a+(-a) = 0.
(iv) Let a E 2Z.
Tho
CTetore-a is the inverse of a.

() Addition is a commutative binary composition on Z. 22 being a

uDSet of Z, addition is commutative on 22.

herefore (2z, +) is a commutative group.

mZ, +) is a commutative group, m being a positive integer.
78 HIGHER ALGEBRA

7. (Z.) is not a group.

(Z,.) is a commutative monoid, 1 being the identity element. Th
inverse of no element other than 1 and -1 in Z exists in Z.

8. (Q..) is not a group.

(Q..) is a commutative monoid, 1 being the identity element. Th-
inverse of 0 does not exist in Q.

9. (R,.) is not a group.

(R..) is commutative monoid, 1

a being the identity element. Th
inverse of 0 does not exist in R.

10. (C,.) is not a group.

11. Let Q* = Q - {0. The set Q" forms a commutative group with

respect to multiplication.
1 is the identity element and the inverse of an element a in Q* is
in Q.
12. (R°,.) and (C*,.) are commutative groups.
13. Let M2R) be the set of all 2 x 2 matrices whose elements are real
numbers. We prove that the set M2(R) forms a commutative group with
respect to 'matrix addition (+).
i) Let A, BE M2(R). Then A + Be M2(R).
Therefore the set M2 (R) is closed under t
(ü) Matrix addition is associative on M2(R).
(ii) The null matrix Og ¬ Ma(R) and O +A A+O2 = = A for all
A in Ma(R). Therefore Og is the identity element.
(iv) Let A e Ma(R). Then -A e M2(R) and (-A)+A =A+(-A)
O2. Therefore -A is the inverse of A.
(v) Matrix addition is commutative on M2(R).
Hence (M2 (R), +) is a commutative group.
14. The set M2(R) does not form a gtoup under matrix multiplication.
(M2(R),-) is a monoid, I2 being the identity element.
Let A be a singular matrix in M2(R). Then there does not exist a
matrix B in M2(R) such that A.B = B.A= I2. Therefore A has no
inverse.
Therefore (M2 (R),.) is not a group.
 GROUPS 79

15. Let S be the set of all 2 x 2 non-singular matrices whose elements

15.
are real
numbers. S i1s a proper subset of M, (R).
We prove that S forms a non-commutative group under 'matrix mul-
tiplication ()'.
Then det A # 0, det
)Let A,BES. B # 0.
det (A.B) = (det A)(det B) # 0 and this implies A.B E S.
Therefore S is closed under matrix multiplícation.

() Matrix multiplication is associative on S.

(i) I ¬ S since det l2 #0. And I2.A = A.I = A for all A in S.
Therefore Ia is the identity, element.
(iv) Let A e S. Then det A # 0 and hence the matrix A-l exists
and det A-l = 1/ det A # 0. Therefore the matrix A-1 e S. Since
AA-1 = A-1A = Ia, it follows that the matrix A-l is the inverse of A
in S.
Therefore (5,.) is a group.
This group is a non-commutative group because, for two arbitrary
matrices A, B E S, A.B # B.A, in general.
This group is called the general inear group of degree 2 over R and
is denoted by GL(2, R).
Similarly, the group GL(7, R) is the group of all non-singular n x n
real matrices under matrix multiplication.

Theorem 2.6.6. Let (GC, o) be a semigroup and for any two elements
a, b in G, each of the equations aor = b and yoa = b has a solution in G
Then (C,) is a
group
Proof. Since (G,o) is semigroup, G is closed with respect to o and o is
associative.
Let e be an element in G satisfying the equation aor = a and e' be
an element in C satisfying the equation yoa = a.

Then aoe =a and e' oa = a.

Let e be an element in G and p be a solution in G of the equation
007 C and q be a solution in G of the equation yoa =c.
Then aop =c and qoa= C.
Now coe= (goa)oe
go(aoe), since o is associative
qoa= C.

Dince is arbitrary, for all in G

c aoe = a a
()
 HIGHER ALGEBRA
86

2.7. Finite groups.

A group (G, o) is said to be a finite group if G contains a finite numh
aber
of elements.
The order of a finite group (G, o) is the number of elements of G. Th
e
order of the group G is denoted by o(G) or by |G|.

Theorem 2.7.1. If (G,o) be a fînite group then in every row (or column
of the composition table each element of G appears exactly once.
Proof. Let (G, o) be a finite group oforder n and let aj, a2,-,Gn be the
list of elements of G in the order in which they appear in the topmost
row and the leftmost column.
Let us consider any arbitrary row of the table, say, the row of aj. The
entries in the row are the elements a;oa1, a;0a2, ,a;0@n
These all belong to G and no two of these are equal, because
aoa = ;0a, a r = as, by left cancellation law.
Since these are n in number, these are all the elements of G in some
order. In other words, the row of elements is a permutation of the ele
ments of G.
Therefore each element of G appears only once in the row of aj.
By similar arguments, each element of G appears exactly once in each
column of the composition table.
This completes the proof.
Examples.
1. Let S =
{1,w, w*} where w = 1. Then S is an abelian group with
espect to multiplication.
The composition table for the set is

|1 W w
1 w w
W 1
w2w2 1 W

i) It appears from the table that

the set S is closed under
cation. multipl-
i) Multiplication is associative on C and S is a
multiplication is associative on S. subset of C. Theretore
(i) It appears from the row of 1 in the table that 1 is a
It appears from the column of 1 in the left identity.
Therefore 1 is the table that 1 is a right identiy
identity element.
iv) The inverse of 1 is the
.
Therefore the inverse of1,each inverse of w is w, the inverse of w 13
element in S exists in S.
 GROUPSs
87

()The table is symmetric about the principal diagonal. Therefore

is commutative on S.
multiplication
(S,) is an abelian group.
Hence
Tet S= {1,i, -1, -i} where i" =1. Then (5,.) is an abelian group.
Let. S={z ¬C:z" =1}. S is the set of n distinct nth roots of unity.
is an abelian group.
Then (S,.)
Proof. (i) Let 21,22 E S. Then z = 1,22 =1. Clearly (z1-2)" = 1 and
this implies z1.22 E S. Therefore S is closed under multiplication.
(ü) Multiplication is associative on the set C and S is a subset of C.
Therefore multiplication is associative on S.
(ii) 1 E S and 1.z = z.l = z for all z in S. Therefore 1 is the identity

element.
(v)Let z E S. Then z" = 1.()" = = 1 * e S , and z =
z.! = 1 for all z E S. Therefore is the inverse of z in S.
is commutative on the set C and S is subset of
(v) Multiplication a

C. Therefore multiplication is commutative on S.

Hence (S,.) is an abelian group.
Note. For each positive integer n, there exists an abelian group of order
Tn. The group is (S,.) where S is the set ofn distinct nth roots of unity.
For n 2, we have the abelian group ({1,-1},.). For n = 1, we have

the trivial group containing the single element 1.

4. The set Z3, the classes of residues of integers modulo 3, forms an
abelian group with respect to +, addition (modulo 3).
The elements of Z3 are 0, 1,2. The composition table is

+|0 1 2
O0 1

() It appears from the

table that the set is closed under +.

(ii)+is associative.
left identity as well as a right
(1ii) It appears from the table that 0 is a
aentity. Therefore Õis the identity element.
inverse of 2 is I.
(iv) The inverse of Õis 0, the inverse of I is 2, the
Derefore the inverse of each element belongs to theset
Therefore
() The table is symmetric about the principal diagonal.
 GROUPS 91

Thus the set S forms a non-abelian group with respect to multiplica

tion.The group is called the symmetric group of degree 3 and is denoted
order of the group is 6.
by S3. The

10. Alternating group An

The set of all even permutations on the set {1,2,...,n} forms a group
with respect to multiplication of permutations.
and is denoted
The group is called the alternating group of degree n
non-commutative group for
by An. An
contains n! elements. An is a

n4.
is the of all even
Ag is the alternating group of degree 3. It the group are
elements of
permutations on the set {1,2,3. The group

2
and Pa=(3 i 2
It is a commutative group of order 3.

11. Klein's 4-group.

S
Let S =
{e, a, b, c} and let o be the binary composition defined on

eoe = aoa = bob =

boe b, eoc = coe =

,
a, eob
= =
aoe =
by eoa =

boc cob a.
coc= e, aob = boa c,= aoc = coa =
b, = =

The composition table is given below.

a C
ee a c
a e c

6 b c e a

le a e
and
abelían of order 4. It is called Klein's 4-group
(5, o) is an group
mathematician and V comes from
is denoted by V. Klein was a German
the German word Viergruppe.
is that every element of the group
An important property of the group
is its own inverse.

12. Symmetries of an equilateral triangle.

Let S be the set of all points in a Euclidean space. (A line is a 1
Space, a plane is a 2-space). An isometry
of the space is a bijection of S
onto S that preserves distance between two points in S.
A symmetry of a geometrical figure in a Euclidean space is an isometry
that keeps the figure as a whole unchanged.
 AIGHD

96
indices. m and n.
integers
Laws of Then
for
Theorem
2.8.1.
of a group (G, o). =a,
(a")=
be an
element
= a"n, (ii)
Let a (am)"
=a"mtn, (i) positive.
Then

i) a oa and n be both

Case 1.
Let m oa(n factors))
Proof. (i)
f a c t o r s ) ) o ( a o a o . .

associative
(aoaooa(m since o is
factors),
aoa=

aoaooa(m +n
amtn
=

= -T
<0.
m> 0 and n
Case 2. Let a(aoaooa)o(a"oalo.
oa oa-1)
then a"oa"
Ifm > r , -
r factors),
aoaooa(m
associative
since o is
a -T=qmtn

a"oa"
a oa(aoaooa)o(aoao...oal)
then
Ifm <r,
-
m factors)
aoao...oa(r

am-T = am+n
aT-m) =

<0 and n > 0.

Case 3. Let m = -s 2.
is similar to case
This case

-r <0. Then
<0 and
=
n
-s
Case 4. Let
=
m

aoaT factors))
a oa factors))o(alo..oa(r
(aooa(s
=
aoao.. oa(s +rfactors ),
associative
since o is
a-(8+T)=a-8-r = am+n.

the theorem is obvious.

or n or both be zero,
Case 5. If either
m

left exercise.
Proofs of (ii) and (ii) are as

additive group, a" is denoted by na.

Note. In an

element.
2.9. Order of an
be
Let (G,o) be a group
and let a an elément of G. a is said to ec
eG,
exists a positive integer n such that a"
=

of finite order if there

a is the least positive integer
being the identity element. The order of
such that a" = eg and is denoted by o(a).
of ab
ais said to be of infinite order (or of order zero), if the order
not finite.
 GROUPS 97

Examples.
Tn the group (5,.) of Example 1, Art.2.7, o(u) = 3, ofw2) = 3.
2. In the group (Za, +),ot+) =6,o)=3,o(3) =2.o) = 3,of5) =6.
3. Inthe group Sa, o(pa) = 3,o(pa) = 3,otPa) = o(pa) = olps) = 2.

4. In the group V, ofa) ofb) o(e) = 2.

= =

5. In the group Us = {1,3,5, 7), o(3) = o(5) = o(7)=2.

6. In the group U1o = {1,3, 7,9}, o(3) = 4, o(7) = 4,o(9) = 2.

7. In the group Da, o(ri) =

o(rs) =
4,o(ra) =
olh) =
o(v) =
o(d) =

o(d')=2.
8. In the group (Z, +), the order of each non-zero element is infinite.

group is 1 and other

Note. The order of the identity element in a no

element in a group is of order 1.

be element of a group (G, o). Then

Theorem 2.9.1. Let a an

(i) ola) = o(a);

and a"" e, then n is a divisor of m;

(ii) if o(a)
=
= n

(ii) if ofa) = n then a, a*,...,a"(=e) are distinct

elements of G;
n
(iv) if ofa) = n, then for a positive integer m, o(a") = gcd(m,n

if and only if p is prime to n;

(v) if o(a) =
n, then o(aP) = n

is infinite.
(vi) if o(a) is infinite and p is a positive integer, then o(aP)
Then a" =
e, where n is the least
Proof. (i) Case 1. Let ofa) = n.

positive integer. (a)" =a-" Therefore

=
(a")l =e" =e.

< n such that («-)" = e.

Let there be another positive integer
m

Then a-m = e.
<n, this
that a"-" = e and since n - m
a and a-" = e imply
=
e
Therefore o(a")=e.
Contradicts that o(a) =n.
infinite.
We assert that ofa-:) is
Case 2. Let ola) be infinite.
= e . i.e., (a")-=e.
o(a) m. Then (a)" There-
Let =

a is of finite order,
a contradiction.
implies a" e, i.e.,
This
=

fore o(a-1) is infinite.

This completes the proof.
such that a" = e. By
is the least positive integer
Since o(a) n, n exist
=

integers q and r such

that m = qn + ,
there
avision algorithm, atT =(a")?oa" = eoa"
a', =

wuere 0 r <n. Then e a = =

98 HIGHER ALGEBRA

This relation holds only when r = 0, because, otherwise it will con.

cOn-
tradict that o(a) = n.
Therefore m = qn and the theorem is proved.

(iii) If possible, let a" = a* for some positive integers r, s such th¡t
r < s n . Then a oaT = e, or, a*-T: .
Since 0<s - r <n, this contradicts the assumption that o(a) = n.
This establishes that a,a*,...,a,a" are all distinct.

(iv) Let o(am) = k. Then am =e. o(a) = n > n\mk.

Let gcd(m, n) = d. Then m = du, n = du, where gcd(u, v) = 1.

nmk >dolduk > vluk

vlk, since ged(u,v) = 1 . )
Again (am)" = (atu)" = atuv = (a")" = e.
o(a") = k and (a")" =e klv .. (i)

From (i) and (i) we have k = v, i.e., k =

Therefore o(a") =ged(m,n)'

(v) Let p be prime to n. Then gcd(n,p) = 1.
Since ofa)=n, o(a") gcdp,n) by Theorem 2.9.1.(iv).
Therefore o(aP) = n, since gcd(n, p) = 1.
Conversely, let o(a") = n.
Since o(a) n
=
n, o(aP) gcd(p,n)* Therefore acdp.nJ= n. This gives
ged(n, p) = 1. That is, p is prime to n.

This completes the proof

(vi) If possible, let o(aP) be finite, say, m.
Then o(aP) m. This
=
implies aP= e and this shows that
finite order, a contradiction. Therefore o(aP) is infinite.
a is of
Note. If o(a) is infinite and p be a negative integer then also o(aP) is
infinite, because o(aP) =
o(a").
Theorem 2.9.2. Each element of a finite
group is of finite order.
Proof. Let a be an element of a finite group (G, o). Then a, a2, a*,... are
all elements of G. Since G is finite, these
elements are not all distinct.
Therefore a" a" must hold for some
=

Therefore a"o(a")-1 eg, or,

positive integers m,n (m > T)
This proves that
=
a"- =
eg.
a is of finite order.
Note. The order of an element in a finite
of the group. This follows group cannot exceed the order
from the Theorem 2.9.1.(iii).
 GROUPS 101

To illustrate,
=m let o(a)
90, o(b) n =50. =
=

1m = 3.25, n=3°.2*.5. [m, n] = 32.21.52 = 450.

Let u= a3, v =b. Then o(u) =ged(90,5)

18, ofv) gcd(50,2) 25.

olu) and o(v) are relatively prime. Therefore o(uv) = o(u).o(u) =

=

god(s0225.
AK0, uU a°b". uv EG and o(uv) [o(a),
= =
o(b)].
Note. If G be a non-comutative group and a E G, b E G with o(@) = m,
ofb) = n, god(m, n) = 1, then there may not exist an element c in G such

that o(c)= m, n.
For example, in the non-comutative group S3, o(pi) = 3, o(p3) = 2.
But there does not exist an element of order 6 (2,3]) in S3.
5. In a group (G,o), a is an element of order 30. Find the order of a8.
Since o(a) = 30, a3 = e. Let o(a8) = m. Then al8m = e, where m
is the least positive integer.
Since o(a)= 30, 30 is a divisor of 18m. Therefore 5 is a divisor of
Sm. Since m is the least positive integer, m = 5. Therefore o(a5) = 5.

Alternative method.
Here o(a) = 30. o(a8) = gcd(18,30)
30 =5.

6. Find all elements of order 8 in the group (Z24, +).

0,1,2,.. . , 23. o(0) 1 and o(1) = 24.
The elements of the group are
=

Let o(m) = 8, where 0< m < 24.

o(1) = 24. o(m) = o(ml) = 24
ged24,m)

o(m) =
8 gcd(24,m) 3. Thereforeand are prime to each
=

other, i.e., is less than 8 and prime to 8, i.e., =1,3,5, 7.

Hence the elements of order 8 are 3,9,15, 21.
b" [gcd(m,n) 1] hold for elements =

group G, a" = b and

7. In a a" =

a,b in G. Prove that a = b.

z, y such that ma + ny = 1.
Since ged(m, n) 1, = there exist integers
(8")® (67)y bmzt+ny = b.
a =
a at+ny amaaY =
=
(a"m)"(a")" == =

1 for all
b"a"", a"b" = b"a" with gcd(m, n)
=
8. In a
group G, a"bm
=

a,6 in G. Prove that the group G is abelian.

commutes with 8" and a" commutes
with "",
We first prove that a
1e., a =ba" and a"b" =ba" for all a,b in G.

(ab)=a"(b"a")"a-m
("am)"aam, by the given condition
=
 104 HIGHER ALGEBRA

2.10. Subgroups.
Let (G, o) be a group and H be a
tobe stable (closed) under o
non-empty subset of G. H is 8aia
if a H, be H> aob E H. If H is stable
E
under o then the restriction of o to H x H is a
to H. This restriction, mapping from H x H
say *, is a composition on H and is
a b aob for all a,b ¬ H. * is called the induced defined by
composition on H.
Definition. Let (G, o) be a group and H be a
(H, o) is a non-empty subset of G. If
group where o is the induced composition, then
to be a
subgroup of (G, o). (H,o) is said
This says that the subset H of G forms
the induced composition o.
a group in its own right under

Examples.
1. Let
of
(G,0) be a group. G being a subset of G, (G,o) is a subgroup
(G,o). This subgroup (G,o) is said to be the
(G,o). improper subgroup of
Let be the identity element in the
e

forms a group under the induced group. The singleton set {e}
is said to be the trivial composition o. The subgroup ({ef,o)
subgroup of (G, o).
The subgroups other than
(G, o) and ({e}, o) are said to be non-trivial
proper subgroups of (G,o).
2. (Q,+) is a group. Z is non-empty subset of Q and (Z,+)
a
is
Therefore (Z, +) is a
subgroup of the
a group.
(Q,+). group
3. (Q,+) is a
group. Q" Q-{0} is a subset of Q and (Q*..) isa
=

group. But (Q*,.) is not a subgroup of the

group (Q,+).
4. Let (G,o) be an
abelian group and
Then (H, o) is an abelian (H,o) is a subgroup of (G,o).
group, since o, being commutative on G,
commutative on H. is also
5. The symmetric group S3 is a non-commutative
group A3 is group. The alternating
a commutative subgroup of the
non-commutative group S3.
Theorem 2.10.1. Let
(G,o). Then (G, o) be a
group and (H,o) be a
subgroup of

) the identity element of the subgroup

of the
group (G,o); (H, o) is the identity element
(i) if a E
H, then the inverse of a in the
the inverse of a in the
group (G,o). subgroup (H,o) is same as
 GROUPS 105

Proof. (i) Let en be the identity element in (H, o) and eg be the identity
element in (G,°). Then epoh = hoe = h for all h in H.
Also egoh = hoeg= , considering h as an element of G.

It follows that hoep = hoeg in G.

Therefore eH= eG, by left cancellation law in (G,o).
G) Let a' be the inverse of a in (H,o) and a be the inverse of a in
iG. o. Then a'oa = aoa' = eH, since (H,o) is a group
Also a o a = aoa = eG. It follows that a'oa = aloa in (G, o), by (i)
Therefore a = a , by right cancellation law in (G,o).

Note. Every subgroup of (G, o) must contain the element eg. Therefore
there cannot be two disjoint subgroups of a group.

Theorem 2.10.2. Let (G,o) be a group. A non-empty subset H of G

forms a subgroup of (G, o) if and only if
(i) a E H, beH>aob ¬ H and (ii) a e H a l e H
Proof. Let (H, o) be a subgroup of (G, o).
Since (H,o) is a group, (i) and (i) are satisfied.
Conversely, let H be a non-empty subset of G satiskying (i) and (i).
Since (1) holds, H is closed under o.
Since H C C and o is associative on G, o is then associative on H.
Let a E H. Then by (ii) a (the inverse of a in G) ¬ H. And since
a,aE H, (i) implies aoa =e EH.
Since e E H, a (the inverse of a in C) is also the inverse of a in H
Therefore a EH implies the inverse of a in H belongs to H.
Therefore (H, o) is a group and hence (H, o) is a subgroup of (G,o).
Theorem 2.10.3. Let (G,o) be a group. A non-empty subset H ofG
forms a subgroup of (G, ) if and only if
a EA, b¬H>aob" E H.
Proof. Let (H, o) be a subgroup of (G, o).
Since (H, ) is a group, bE H » b ¬ Hand therefore a E H and
bEH> aob-l e H.
OTvErsely, let H be a non-empty subset of G such that
aEH, b¬H>aob- E H.
Let a e H. Then a E H, a E H aoa E H, i.e., e EH.
Therefore H contains the identity element.
eEH, a E H eoaeH, by the condition, i.e., al eH.
 106 HIGHER ALGEBRA

Hence a E H implies a-le H. Therefore the inverse of each elemo

in H exists in H. nens

Let a e H, be H. Then a e H and b e H; and by the giv.

er
condition ao(b-l)=l ¬ H, ie., aob ¬H.
Hence a E H, beH* aob e H. Therefore H is closed undero.
Since H is a non-empty subset of G and o is associative on G.o:
associative on H.
Therefore (H,o) is a group and hence (H,°) is a subgroup of (G,o).
Theorem 2.10.4. Let (G, o) be a group and H be a
non-empty finite
subset of G. Then (H, o) is a subgroup of (G,o) if and only if a e
H aob E H.
H, be

Proof. Let (H, o) be a subgroup of (G, o). Since (H, o) is a group, H i

closed under o. Therefore a E H, bE H = aob E H.
Conversely, let H be a non-empty finite subset of G such that a E H, b
H aob EH.\
Let h E H. Then h,
h*,h3,... are all in H, since H is closed under o.
But H being a finite set, these elements
cannot be all distinct.
Let h h for some positive
=

these as elements in G, h"-* =e.

integers r and s, T> s. Considering
Since r- s is a positive
the identity element.
integer, h"-* E H and therefore H contains-
Again r8-120 since r - s21.
Therefore ha-l being a
h-oh = e.
member of {e, h, h2,...} belongs to
This proves that H and.
h-(=h-8) belongs to H.
Since H is a
non-empty subset of G and o is associative G,o
associative on H. on

Therefore (H, o) is
group and hence H forms a
a

Note. The theorem does not subgroup of (G,°).

hold if H be an infinite
example, let G (Z, +) and H N.
= subset of G. Fo
infinite subset of the set Z and (Z, +) is=
a
group. The set N is au
(N, +) is not a group.
Theorem 2.10.5. Let
(G,o).
Then (G,o) be a
group and H, K
H nk is a
subgroup of (G, o)
are
subgroups
Proof. Hnk is a
non-empty subset
K, e
being the identity element. of G, since e belongs to both H a d
Let a,bE HNK.
Then a, b E H and
Since H is a a, bE K.
subgroup, a,b ¬ H
Since K is a
subgroup, a,b ¬ K> aob EH.
aobE K.
 GROUPS 107

Therefore a ¬HNk, beHnK » aob-e HnK and this proves

that HnK is a subgroup of (G,o).
Note 1. The union of two subgroups of a group G is not necessarily a
subgroup of G.
Let us consider the groupG= (Z, +) and the subgroups H = (2Z, +),
K (32, +).
2EHUK, 3 ¬ HUK but 2 +3 gHUK. As HUK is not closed
under addition, HUK does not form a subgroup of G.
If however, we consider two subgroups H (2Z,+) and K = (42, +)

ofthe group G =
(Z, +), then HUK(=H) is a subgroup of G.
Note 2. The intersection of an arbitrary family of subgoups of a group
G is a subgroup of G.
of
Theorem 2.10.6. Let (G,o) be a group and H, K are subgroups
if either
(G,o). Then H U K forms a subgroup of (G,o) if and only
HCK, or K c H.
We shall prove that either H K or
subgroup.
-

Proof. Let HUK be a

K- H is empty.
H be non-empty .(1)
Ifnot, let both of H K and K
-
-

Let a E H -K andbeK-H.
Then a E H but a gK; be K but b g H.
Therefore a E HUK and b ¬ HUK.

subgroup, it follows that aob e

HUK... (i)
Since H UK is a
element of at least one of H and K.
This implies aob is an

But a E H and aob EH ao(aob) ¬ H, since H is a subgroup

bE H, a contradiction.

Also b EK and aob EK (aob)obEK, since K is a subgroup

a EK, a contradiction

belongs to K and this contradicts

Therefore aob neither belongs to H nor

(1). therefore either

(i) is not tenable and
It follows that the assumption
HCKorKcH. if K c H then HU K = H.
HUK =
K, or
Conversely, if H c K then
n any case HUK is a subgroup.
his completes the proor
subgroups.
be the union of two proper
Oe. A group G cannot
 GROUPS
111

JK =
L, KJ =
-L, KL =
J, LK =
-J, LJ =K, JL = -K.
The composition table for the set H is
J K -K
J K
I -J -K K -L L
J J
L -L -K
J -J K
-I -L
K K -K K -K
-L L -I I J -J
-K-K K L L I -I J
LL -L K J
-L-L L -K K J -J
I
I -I
The finite set H is closed under
matrix multiplication. Therefore
(H,.)is a subgroup of CL(2,C).
H is a subset of the set of all
complex matrices
u,vEC, where k is the
complex conjugate of u.
The eight elements of H
correspond tou= 1, v = 0; u= -1, v =
0; u =i, v 0; u = -i, v 0; u 0, v
=
=

0, v= i, u= 0, v =-i.
=1; u =
=0, v =-1, u=

The group (H,.) is said to be the

denoted by Q8.
group of unit quaternions and is

Each of the subsets

{I,-I}, {I,-1,J,-J}, {I-1,K, -K},
{I,-I,L, -L} is closed under multiplication.
Therefore each forms a subgroup of Qs.

Some important subgroups of a group.

1. The centre of a group.
Let (G,o) be
group and H be the subset defined by
a

H {z E G:zog =gor for all ge G}.

=

H is a subset of G containing those elements of G that commute with

every element of G.
We prove that (H,o) is a subgroup of (G,o).
H is a non-empty subset of G, because e ¬ H.
Let p E H, q¬ H. Then pog = 9oP, 9o9 = goq for all g in G.

Pog)og po(qo9) = po(gog) = (pog)oq = (gop)oq = go(pog) for all 9

in G. This shows that poq EH.
112 HIGHER ALGEBRA

Therefore pe H, q E H> poq E H .

()
Let p E H. Then pog gop for all g in G.
=

Hence po(pog)op-l =

in G. This shows that p ' E H.

plo(gop)op,
i.e., gop =
pog for all1
Therefore peH~p e H (i)
From (i) and (ii) it follows that (H, ) is a subgroup of (G, o).
Note 1. This subgroup is called the centre of the group and is
denoted
by Z(C). The elements of Z(C) are called the central elements of G.
Note 2. Z(G) is a commutative subgroup of G. If, however, G be a
commutative group, then Z(G) = G.

2. The centraliser of an element in a

group.
Let (G, o) be a group and let a eG. Let H be the subset of G defined
by H = {r ¬G:zoa = aot}.
H is a subset of G containing those elements of G that commute with
the particular element a.
We prove that (H,o) is a subgroup of (C, o).
H is a non-empty subset of G, since a E H.
Let p E H, q E H. Then poa = aop, qoa = aoq.
And (pog)oa = po(qoa) = po(ao9)= (poa)oq = (aop)oq = ao(pog).

This shows that poqEH.

Thereforep E H, q¬H> poq EH . )
()
Let p E H. Then poa = aop.
Hence po(poa)op = p""o(aop)op, i.e., aop"' =p*oa.
This shows that pl e H.
Thereforep
pEH~pl eH... (i)
(ii)
From (i) and (i) it follows that (H, o) is a subgroup.

element and is
Note 1. This subgroup is called the centraliser of the a

denoted by C(a).
3. Cyclic subgroup generated by an element.
Let (G, o) be a group and a be an element of G. Let H be the subset
of G defined by H = {a" : n ¬ Z}
H is the subset of G containing all integral powers ofa.
We prove that (H, o) is a subgroup of (G, o).
H is a non-empty subset of G, since a E H.
Let pE H, qE H. Then p = a", q = a for some integers r, 3.
 GROUPS 113

pog a* EH, since r +s is an integer;

nd 1 a" ¬ H, since - r is an
and =
integer.
Therefore p ¬H, q¬ H> poq ¬ H ()
(1)
and pE H>p¬ H. (iü)
...

From (i) and (i) it follows that (H, o) is a

subgroup of (G,o).
Note 1. This subgroup is said to be the cyclic
by the element a and is denoted by (a). subgroup of G generated
Note 2. Thesubgroup (a) is the
smallest
cOntaining the element a, i.e., if (K,o) be subgroup
of the o)
(G,o)(G,
any subgroup ofgroup such
that a E K, then (a) cK.
Proof. Since a E K and (K,o) is a subgroup, a" e K for every n ¬ Z.
Therefore (a) C K and this proves that (a) is the smallest subgroup of
(C,o) containing the element a.

Note 3. (a) is a commutative subgroup of the group (G, o).

Proof. Let p ¬ (a),q ¬ (a). Then p =a",q=' for some integers r, s.

poqaoa =a**, qop = a'oa" = a*.
Since and s
r are integers, r +s s+T. Therefore poq
= =
qor.
This proves that (a) is a commutative subgroup.
Note 4. Let G be an additive group and a e G. The cyclic subgroup
(a) is defined by (a) = {na : n e Z},.
For example, in the group (Z,+)
() the cyclic subgroup generated by 1 is the group (Z, +) itself;
(i) the cyclic subgroup generated by 2 is the group (2Z, +);
ii) the cyclic subgroup generated by 3 is the group (3Z, +).

Worked Examples.
1. Find all cyclic
subgroups of the group (S,.), where S={1,i,-1,-4).
(1) = {1};

i)= {1,i,-1,i}, since i2=-1,= -i,i = 1;

(-1)= {1,-1}
)={1,i,-1,-i}, since (-i)"= -1,(-i)=i,(-1 =1.
Note. The cyclic subgroup generated by 1 (the identity element ) is the
val subgroup. The cyclic subgroup generated by i (and also generated
Dy-i) is the group itself. The cyclic subgroup generated by -1 is a
proper subgroup.

HA-8
 114 HIGHER ALGEBRA

2. Find all cyclic subgroups of the group (Zs,+).

The elements of the group are 0, 1, 2, 3, 4.
(0) = {Õ};

()= {0. 1, 2, 3, }, since 21 =2, 3* =3, 41 =7, 51 =0,

(2)= {0, I, 2, 3, }, since 22 =4, 32 =1, 42 =3, 52 =0;
(3)= {0, 1, 2, 3, ), since 23-1, 33=4, 43= 2, 53 = õ;
(4)= {0, I, 2, 3, 4}, since 2 -3, 3 =2, 4 = I, 54 =ô.
Note. The cyclic subgroup generated by the
identity element is the
trivial subgroup. The cyclic subgroup generated by each
non-identity
element is the group itself. All cyclic subgroups are the trivial
the
subgroup
or
improper subgroup of the group (Z5,+).
3. Find all cyclic subgroups of the symmetric group S3.
The elements of the group S3 are
po, P1, P2, P3 P4» P5.
(Po) = {po}i

(P1)=1Po, Pa Since pi P2, Pi =Po

P1.
Pa)= {po.P1. Pa), since pá=P1. P
Po =

p3)= {Po», Pa}, since P =Po%

(pa)= {Po. Pa}, since P= Po%
P5)= {Po. Ps}, since P =Po.
Note. All cyclic
subgroups are proper subgroups of S3.
4. Find all
cyclic subgroups of the group V.
The elements of the
group V are e, a, b,c.
(e) = {e}

(a) {e, a}, since a2 e

=
=

(6) {e, b}, since b=e;

=

(c) = {e, c), since c* =e.

Note. All cyclic

of the group V.
subgroups are different and each is a
proper subgroOP
Observation. We observe that there are
cycic groups which are identical w
subgroups generated by one or more
elements of the oup
groups in
examplesl and 2 are such. These
and are dealt with in groups are of speciante
the following
chapter.
 116 HIGHER ALGEBRAA

2.11. Cyclic groups.

Definition. A group (G,o) is said to be a cyclic group if there exists
an
element a in G such that G {a" :n e Z}, i.e., G (a). a is said
to be
= =

a
generator of the cyclic group.
In additive notation, G {na:n E Z}
= =
(a).

Examples.
1. (Z,+) is
cyclic group generated by 1. -1 is also a generator.
a

2. (Z4,+) is a cyclic group generated by 1. 3 is also a generator.

3. Let S =
{1,i,-1,-i}. Then (S,.) is a cyclic group generated by i.
-i is also a
generator.
4. Let S= {1,-1}.
Then (S,.) is
the only generator of the
a cyclic group generated by -1. -1is
group.
5. Klein's 4-group V is not a cyclic group. The elements of V are e, a, b,c
and none of (e), (a), (6) and (<) equals V. Therefore no element of V
can generate the whole group V.
Theorem 2.11.1. Let (G,o) be
ais also a generator.
a cyclic group generated by a. Then

Proof. Since a is a
generator, G {a"
Let H
=
: n e Z}.
=
{(a)" : n E Z}.
Let p E G.
Then p=a for some integer r.
p can be
expressed as (a)- and since - r is an integer, p EH.
Thus p e GpeH and therefore GCH. (i)
Let g E H. Then q
(a)" for some integer s.
g can be expressed as as and
=

since - s is an integer, q E G.
Thus q e
H>q¬Cand therefore H cG
From (i) and (i), G H. That (i)
is, G = {(a)" ¬
=

a is a generator of the group (G,o).Z).

This proves that :n

Theorem 2.11.2. Every cyclic

group is abelian.
Proof. Let (G, o) be a cyclic group
Let p,q E G. generated by a.
Thenp =
a", q =
a" for some integers r and
poq =
aoa" =
a"*t8; a'oa"
s.
Since r + 8
qop = =
as+.
= 8 +T, poq qop. =

Because poq =
qop for all p, q E G, the group (G, o) is abelian.
Note. An abelian
group is not
Klein's 4-group V is abelian
but it is not
necessarily a cyclic group. For example
cyclic.
 GROUPS 117

Examples (continued).
6. The symmetric group S3 is not cyclic, since it is not abelian.
7, The dihedral group D4 is not cyclic, since it is not abelian.
Theorem 2.11.3. Let (G,o) be a finite cyclic group generated by a.
Then o(C) =n if and only if ofa) =n.
Proof. Let o(a) = n. Then a, a,... ,a"(= e) are distinct elements of G.

Therefore {a,a*,...a"} cG... (i)

Let p be an arbitrary element of G. As G= {a" :n e Z}, p = a for
some integer m. By division algorithm, there exist integers q and r such
that m =gn +r where 0 r <n.
Thereforep = a" == a*T = (a")"oa" = eoa" = a", since a" = e.
As p = a" for some r satisfying 0 r < n , p¬ {a°,a,a?,..a-1}.
That is, p E {a,0°,... ,a*,a"(se)f.

Therefore
Gc{a,a,...a-,a"}.. (i)
From(i) and (i), G = {a, a",...,a"-,a"} and therefore o(G) = n.

Conversely, let o(G) = n. Since G is a finite group, every element of G

is of finite order.
Let o(a) = k. Then a,a?,... ,ak-1,a*(= e) are distinct elements of
G. Since G contains n elements, k n .
But k is not less than n, because by the foregoing argument, o(a) = k
implies that o(G) = k, a contradiction.

Therefore ofa) =n and this complets the proof.

Corollary. If G = (a) and ofa) = n then G = {a,a2,... ,a"(= e)}.

Theorem 2.11.4. Let (G,oj be a cyclic group generated by a. Then G

is infinite if and only if ola) is infinite.
Proof. Let o(a) be infinite. Then the set {a,a,a",...}is an infinite set
of distinct elements.
lf not, let a" = a' for some positive integers r, s where r > s.

and this implies that ofa) is finite, a contradiction.

Therefore ar-* = e
But fa,a2, a3,...} c G. Hence o(G) is infinite.

Conversely, let o(G) be infinite.

lf possible, let o(a) be finite. Then o(G) is finite by Theorem 2.11.3
And this is a contradiction. Hence o(a) is infinite.

This complets the proot.