HIGHER ALGEBRA

ABSTRACT AND LINEAR

FOR DECREE HONOURS COURSE]

SADHAN KUMAR MAPA

Reader in Mathematics (retired)
Presidency College, Calcutta

LEVANT
A
Levant Books
Kolkata, India
 CONTENTS
PART I

PREFACE TO THE FIRST EDITION

1. Sets 3
1.1. Basic concepts
This book is designed to serve the need of Mathematics Honours
1.2. Subset
students and those who require basic knowledge in Abstract and Linear on sets
1.3. Algebraic operations 11
algebras. In this introductory volume the topics have been discussed Set of sets
1 4
at an elementary level with due consideration to modern notions and Partition of a set 13
1.5.
notations. A good number of examples worked.out in the book will help Cartesian product of sets 14
1.6.
the beginner learn and understand the subject. Some useful applications Exercises 1 15
of Linear algebra to Geometry have been shown in order to stimulate Relationn 19
1.7. Equivalence relation
the reader's interest. Exercises have been arranged with well-graded 1.8. 27
Partial order relation 25
problems in which the reader will find the application of his learning and .9.
Exercises 2 32
understanding. 1.10. Mapping 37
The author expresses his indebtedness to the authors of some ad- Composition of mappings 41
.11.
vanced texts which have been consulted during the preparation of this Inverse mapping 45
.12.
volume. A bibliography of such texts is at Direct and inverse images 48
given the end. 1.13.
author also conveys his sincere thanks to the publisher and thhe Exercises 3 50
The
their in the publication of this book. Permutation 52
printer for co-operation 1..14.
of best s o m e mistakes and misprints have crept
in. An 1.15. Cycle 57
Inspite efforts,
is given at Exercises 4 57
errata showing correction for mistakes and major misprints number
sets and cardinal
1.16. Equipotent 58
the end. countable sets
17. Enumerable and 64
Any correction and any suggestion for improvement of the book will Exercises 5
be highly appreciated.
2. Groups 65
Binary composition 67
CALCUTTA, 2.1.
S. K. Mapa Exercises 6 68
July 14, 1984
2.2. Groupoid 70
Semigroup 71
2.3.
2.4. Monoid T2
2.5. Quasigroup 74
Exercises 7 75
2. 6. Groups 84
Exercises 8 86
2.7. Finite groups 94
Exercises 9 95
of a n element 96
2.8. Integral powers
Order of a n elemnent
2.9.
 Exercises 10 4. Boolean Algebra
10. 102
Subgroups 104 4.1. Introduction 301
Exercises 11 310
2.1 115 4.2. Order relation
Cyclic groups 116 315
Exercises 12 4.3 Subalgebra
123 Atom 318
12. Cosets 4.4 323
124 Exercises 26
Exercises 13 324
134 4.5. Boolean functions
2.1. Normal subgroups 334
135 Exercises 27
2.1 Quotient group 142 4.6. Switching circuits 334
Exercises 14 339
148 Exercises 28
2.15. Homomorphism 150
2.16. Isomorphism 159 PART II
2.17.
Automorphism 171
Exercises 15
175
2.1 Conjugacy relation 177
1. Matrices (I)
Exercises 16 1.1. Introduction
181
19. Direct product of groups 1.2. on matrices
182 Algebraic operations
Exercises 17 1.3. of a matrix 10
189 Transpose
1.4. Symmetric and skew symmetric matrices 12
3. Rings and Fields 1.5. Block multiplication of matrices 14
3.1. 1. 6. Some special matrices 17
Ring 191 18
Exercises 18 Exercises 1
208
.2. Integral domain 210
3.3. Skew field 214
2. Determinants
3.4. Field 218 Introduction 21
2.1 35
Exercises 19 224 2.2. Cofactors and minors
3.5. 2.3. Complementary minor 43
Subring 225
3.6. Subield 2.4. Multiplication of determinants 46
230
Exercises 20 231 2.5. Cramer's rule 54
3.7. Ideals of a ring 2.6. Symmetric and skew symmetric determinants 58
232
3.8. Exercises 2 62
Quotient ring 245
Exercises 21 250
3.9. Divisibility in an integral domain 251 3. Matrices (II)
3.10. Unique factorisation domain 258 3.1. Adjoint of a matrix 69
Exercises 22 261 3.2. Inverse of a matrix 71
3.11. Ideals of an integral domain 262 3.3. Orthogonal matrices 77
.12. Principal ideal domain 262 3.44. Complex matrices 79
3..13. Euclidean domain 268 Exercises 3 83
Exercises 23 274 3. 5. Rank of a matrix* 86
Homomorphism of rings 275 3. 6. Elementary operations 87
3.14.
Exercises 24 284 3. 7. Elementary matrices 98
Field of quotients Exercises 4 103
3.15. 285
3.8. Congruence operations 104
3.16. Polynomial rings 290
Exercises 5 110
Exercises 25 299
4. Vector Spaces
4.1. External composition 111
4.2 Vector space over a field 111
4.3 Subspaces 116
Exercises 6 125
4.4. Linear dependence and linear independence 127
4.5. Basis and dimension 131
4.6. Co-ordinatisation of vectors 143
Exercises 7 144
4.7. Complement of a subspace 146
8. Quotient space 150 PARTI
Exercises 8 152
and column space of a matrix 153 ABSTRACT ALGEBRA
4.9. Row space
Exercises 9 161
4.10. System of linear equations 162
Exercises 10 177
4.11. Application to Geometry 179
Exercises 11 183
.12. Euclidean spaces 184
4.13. Orthogonal complement of a subspace 196
Exercises 12 198
4.14. Matric polynomials 200
.15. Characteristic equations 202
.16. Eigen values of a matrix 204
4.17. Eigen vectors of a matrix 206
Exercises 13 215
4..18. Diagonalisation of matrices 217
4.19. Orthogonal diagonalisation 223
Exercises 14 226
4.20. Real quadratic form 226
Exercises 15 234
237
4.21. Linear mappings
Exercises 16 249
Exercises 17 258
.22. Linear space of linear mappings 266
. .23. Linear operators 276
Exercises 18 287
4.24. Linear functionals 289
.25. Orthogonal mappings on Euclidean spaces 292
Exercises 19 295
4.26. Application to Geometry 295
Exercises 20 315
317
Answers
Bibliography 328
Index 329
 1. Sets

1.1. Basic concepts.

The theory of sets in its intuitive form was developed by G. Cantor,
a German mathematician, in the last part of the nineteenth century.
Logicians analysed the theory in later years and built the theory on an
axiomatic foundation. Here we shall introduce the naive theory of sets
as was developed by Cantor. According to bhim 'A set is a well-defined
collection of distinct objects of our perception or of our thought, to be
conceived as a whole'.
Commonly we shall use capital letters C,..
A, B, and
todenote sets
small letters a, to denote objects (or We sum-
b, c,... elements) of a set.
marise briefly some of the features of a set given in the definition.
(i) A set S is a collection of objects (or elements) which is to be
regarded as a single entity.
A set S is of distinct objects (elements) and be an
(i) comprised if a
object of S, we denote this by a ¬ S (read as 'a belongs to S').
(ii) A set is well defined, meaning that if S be a set and a be an
object, then either a is definitely in S (a e S) or a is definitely not in S,
denoted by a g S ( read as - ' a does not belong to S").
A set may be described by the characterising property of its elements.
If p be the defining property of the elements of S S expressed as
then is
S {T: r has the The brace notation for
propertyp}. S expresses the
fact that S is the collection of all those elements c such that t has the
property p.

Examples.
as S =
1. The set S of all integers is expressed {z:r is aninteger.
Here 3 E S but g S.
2. The set T of all prime numbers less than 20 as
is expressed
T={z : t is a prime less than 20}. Here 3 ET but 4 gT.
A set may also be described by listing all elements of the collection.
Thus the set in Example 2 can also be described as
T={2,3, 5,7, 11, 13, 17, 19}.
 HIGHER ALGEBRA
SETS

Throughout this text we shall use some accepted notations for the
familiar sets of numbers. Examples.
20 is a subset of the set N.
1. The set of all natural numbers less than
N is the set of all natural numbers It is denoted by {n e N:n < 20.
is the set of all integers
a subset of the set Z. It is
Z* is the set of all positive integers 2. The set of all integers greater than 5 is
is the set of all rational numbers denoted by {z E Z:r> 5}.
Q is the set of all positive rational numbers a subset of the set Itis
3. The set of all positive rational numbers is Q.
R is the set of all real numbers
denoted by {z e Q:r>0}.
is the set of all positive real numbers
C is the set of all complex numbers. a subset of the
4. The set of all real numbers lying between 1 and 2 is
set R. It is denoted by {z E R: 1<s < 2}.
1.2. Subset.
unit modulus is a subset of the set
5. The set of all complex numbers of
Let S be a set. A set T is said to be a subset of S if z eT>z E S. C. It is denoted by {z ¬C: |2l = 1}.
This means that each element of T is an element of S. This is expressed
symbolically by Tc S signifying that T is contained in S.
If T be a subset of S, S is said to be a superset of T and this is 1.3. Algebraic operations on sets.
expressed by S T.
In this section we shall discuss several ways of cormbining different
Z
For examnple, c Q, Q cR. sets and develop some properties among them. For this purpose we shall
We conceive of the existence of a set containing no element. This is consider several sets, in a particular discussion, as subsets of a single
called the nuil set, or the empty set, or the void set and is denoted by ó. fixed set, called the universal set in relation to its subsets. A universal
For logical consistency, the null set is taken to be a subset of every set. set is generally denoted by U.
Therefore for every set S, S c S and o c5. S and o are said to be
Let U be the universal set and A, B, C,.. . be the subsets of U. We
the impToper subsets of S. Any other subset of S is said to be a proper define the following operations on the class of subsets of U.
subset of S.
Definition. A set is said to be a finite set if either it is empty or it (a) Union. The union (or join) of two subsets A and B is a subset of
contains a finite number of elements; otherwise it is said to be an infinite U, denoted by AU B and is defined by
set. AUB =
{r:r ¬A or z E B}.
Definition. Two sets S and T are said to be equal (expressed by S =T)
if S is a subset of T and T is a subset of S. Two equal sets contain Therefore AUBis the set of all those elements which belong either
to A, or to B, or to both. It follows that A c AUB, Bc AUB.
precisely the same elements.
and A =

The relation C is called set inclusion. The main properties of the For example,let U= {1,2,3,. 10) be the universal set
=
{1,3,5,7, B= {5,7,9). Then A UB
{1,3,5,7,9}.
inclusion relation are the following:
SCS for every set S (reflexivity) (b) Intersection. The intersection (or meet) of two subsets A and B is
i)
a
(ii) ScT and T C S =S=T (antisymmetry) subset of U, denoted by AnB and is defined by
(ii) SCT and T C V S c V (transitivity). AnB= z e
{z:c e Aand B}
There may, however, exist sets S and T such that neither S c T nor
is the set
those elements which belong to both
TCS. Such a pair is said to be incomparable.
ofsets ThereforeAnB ofall
A and B. It follows that AnB c A, AnB CB.
example, let S be the set of all integral multiples of 2, T be the
For
set of all integral multiples of 3. Then neither S c T nor T c S. For example, let U = {1,2,3,. .. , 10} be the universal set and A =
(1,3,5,7), B = {5, 7,9). Then AnB ={5,7}
 T

C t t t 1t
W

tt tt
 HIGHER ALGEBRA SETS
9

Let y be an element of Q. yEQ2 yE A' and yEB' (e) Symmetric Difference. The syininetric difference of two subsets
y A and y f B A and B is
a subset of U, denoted by AAB and is defined
by
-
y AUB AAB = U -

yE (AU B)!
(A B) (B A).
Therefore QcP (ii) AAB is the set of all those elements which either to A
belong or to
B, but not to both.
and we =
Combining (i) (ii), have P Q, i.e., =
(A UB) A'nB'.
AAB is also -

Proof. 7b. Let us consider two subsets A' and B'. expressed as (AUB) (AnB).
AAB can also be expressed as (AnB') U (A'nB).
By 7a, =
(A' UB'Y (A'Y n(B'Y
=
It follows from the definition that AA$
or, (A'UB') = AnB, by property 6 AAA = o for all subsets A.
A for all subsets A and
=
or, {(A'UB'Y} (AnB), taking complements Note. (A- B) n (B - A) =
Or, AUB' = (AnB), by property 6.
n = =
(AnB') (BnA') An(B'nB) nA' = =
(An¢)n 4' ¢nA' ¢.
The independent proof of 7b is left as exercise. Therefore AAB can be considered as the union of two
-
A- B and B disjoint subsets
A, provided A - B and B - A are both
(d) Difference. The difference of two subsets A and B is a subset of non-empty.
-
U, denoted by A B and is defined by
Properties of symmetric difference.
A B =
{r e A: c
g B}. 9. AAB= BAA.
A- B is a subset of A and is the set of all those elements of A (Commutativity ).
which 10. =
are not in B. AA(BAC) (AAB)AC. (Associativity).
= -
-
In = -
BAC -

particular, A B ¢ A B = A =
Proof. (B C)u (C
if AcB and if AnB ¢. B)
=
- u
The subset A B is also called the complement of B relative to A. (BnC') (B'nC).
The difference A- B can be expressed in terms AA(BAC)
as A - B = An B'.
of the complement =
{An [(Bnc) u(B'ncy}u{A'n
=
{(Bnc) u (B' nci}
Properties of difference. {An [(B'UC) n (B UC')]} U(A'nBnC') u
(A' nB'nc)
- -
8a. =
A-(Bnc) (A B) U(4A C). ={An [(B' nC')u(CnB)]} U(A'
8b. A (BUc) = - -
=
NBnc)u(A' nB'nc)
(A B) n (A C) (AnB' nc) u (xnBnc)u (A'nBnC') u (A'n B'nc)
Proof. 8a. A- (Bnc) = An (Bncy = (ANBnc)u(AnBnC)u (A'nBnc)u(A'NB®nc)
= An(B' UC")
The right hand
= expression remains same if
(AnB') u (AnC) A and C are
= - -
interchanged.
(A B) u (A C). herefore =
AA(BAC) CA(BAA)
Proof of 8b is left as exercise. =
(BAA)AC, by 9
=
A-(BUC) is the relative complement of BUCin A and A-(Bnc) (AAB)AC, by 9.
is the relative complement of BnC in A. The set -

operations union, intersection,

These are the
and complementation, difference
properties generalisation of De Morgan's Laws. symmetric difference can be visualised from the diagrammatic
10 iiGHER ALGEBRA SETS
11

ropresentation of sets, called the Vern diagram 1.4. Set of sets.

We lave «lefined a set as a collection of its cloments. If the clemcnts

he scts theuselves, then we have a family of sets, or a set of sets.
thhe subscts of a non-enmpty sot S is a
t
Por cxample,
of sets. This
collection of all
sct is said to be the power set of S and is donotod
by
P(S).
If S rontains n clenents, then P(S) contains 2 subscts, bccausc
a of P(5) is cither or a subset containingr clements of
S,
subset
r= 1,2, , 72.
AUB A B A B AAB
Lot I be the finite set {1,2,.. . , 72} and F be the family of n scts
A A2,. An is oxprcssed as
The rectangular region represcuts the universal set U and the circular
t
subsets A and B. The shaded portion rcpresents tlhe sot E
rogionsthe F = {Aa 1}
named below the diagram.
I is called the inder set of the family F. The members of F arc
The propertics of differeut set operatious cau be verified fron the being
indexed by [.
Venn diagram.
Let I be the set N. Then the family of scts F
={Aa : a E I} is an
Worked Examples. infinite collection of the sets Aa, whero cach Aa corrcspouds to a natural
1. If A, B, C be subscts of a universal set U, prove that nber. The family is also cquivalently cxpressocd as

(AUB)n (BuC) (CUA) = (AnB) u ( B n c ) u (CnA). F= {A,n : n E N}.

(4UB) n (BUC) =(BUA) n (BUC) Let I be an
{Aa : x E
is au
arbitary sct. Then thc family of scts F= I}
BU(Ac) distributive law] arbitrary collcction of sots Aa ndexod by I.
Bu(Cn A).
a
Let I bo a n indcx sot and {Aa a E
: I} be a family of subscts of
L.H.S. (AUB)n(BuC) n(CUA) universal set S.
[BU (Cn A)] n (CUA) Tlen the union of the family of these subsets is defiued
by
Bn (CU A)] U [(OnA)n(CUA)] [distributive law]
I(Bnc)u(BnA)] u(CnA) [CnA c CUA] A {r E S : E Aa for at lcast onc a in 1}.
E
(AnB)u (Bnc)u(CnA) = R.H.S. (proved).
The intersection of the family of tlhese subsets is defined
2. Prove that An(BAC) = (ANB)A(Anc), wherc A, B, C are subscts by
of a univcrsal sct U.
A - { * e S:ae A for all a e l}.
R.H.S. =(ANB)A(Anc)
In particular, for a finite F =
=
family of subsets {A1, A2.. , A,,}
=
{(AnB) n (Anc)u[(ANB)' n (AnC (i) U A, {*: E A; for at least onc
=
i};
[(An B) n (A' UC')]u [(A'UB') n(AnC] (ii) , A; = {* : æe A, for cach i};}
=
[(A nBnA')u(AnBnc)]u [(A'nAnC) u (B' n AnC)]
= =
(AnBnc)u(An B'nc) [AnBnA' ). A'n AnC= (ii) 4.-A
=
An[(Bnc') u (B' nc)]
=
A
An (BAC)= L.H.S. (provecd). (4.) -
12 HIGHER ALGEBRA SETS 13

Examples. be four subsets of Z defined by

4. Let Co, C1, C2, C3
1. Let S ={1,2,3}. Then the power set of S is given by {0,4, t8,},
P(S) = {¢, {1}, {2}, {3}, {1,2}, {2,3}, {1,3},S}. Co {4n: Tn is an integer } {1,1t4, 1 +8,}
= and
C1 {4n +1: n is an integer
2. For each n ¬ N, let I, [0, 1/n]. Then I1 I 2 I3 . {2,2t4,2 +8,
C2 {4n +2:n is an integer
nEN {0} C3 {4n+3;n is an integer } {3,3 t4,3 +8,
are such that each is non-empty and
= . and Then the subsets Co, Ci, C2, Ca
3. n E N, let In (0,1/n). Ia Is
Foreach Then I1 U u = Z and Cn C o,i #j.
nEN
Co UC Ca Ca
form a partition of the set Z.
Therefore the subsets Co, C1, C2, C3
1.5. Partition of a set.
Let S be a non-empty set. A family of non-empty subsets {Sa : a E 1.6. Cartesian product of sets.

I}, I being the index set, is said to form a partition of S if Let A and B be non-empty sets. The
Cartesian product of A and B,
) U Sa=S denoted by A x B, is the set defined by
aEI
and (ii) Sa n Sg = o for a, BE I and a #B.
A x B= {(a,b) : a E A,b E B}.
a E is
The condition (ii) says that the family of subsets {Sa I the first component being
A x B is the set of all ordered pairs (a, b),
pairwise disjoint. an element of A and the second being an element of B.
In particular, a finite family of non-empty subsets S1,S2,.. . , Sn of a
Let A1, A2,.. , An be a finite collection of non-empty sets. The Carte-
a partition of S if
non-empty set S is said to form A1 X A2 x.. x An,is the set
= sian product of the collection, denoted by
1) S1 U S2U.U S, S defined by
=
and (ii) Sns; d for i # j.
x. x = :
x Az An
A {(c1,72,..-,Tn) zi E Aif.
= =
Examples. -
In particular, if Aj Az =* An, the Cartesian product of the
1. Let A and B be non-empty sets such that the sets A- B, B
A and
collection of sets, denoted by A", is the set of all ordered n-tuples
-

AnB are each Then the sets A B, B -A and AnB form

non-empty. E A}.
a partition of the set AU B, because (1,2..,Tn): o;
- =

4) (A B) u (B A) U (AnB) AUB, and Examples.

- - subsets of A U B.
(ii) A B, B A and An B are mutually disjoint 1. Let A = {1,3,5}, B = {2,4}.
Then A x B {(1,2), (1,4), (3, 2), (3, 4), (5,2), (5, 4)};
2. Let S be a
non-empty set and A be a proper subset of S. Then the
BxA = {(2, 1),(2,3), (2,5), (4, 1), (4,3),(4,5)}.
two subsets A and A' are such that each is non-empty and
(i) AUA' = S, (ii) AnA' = o. 2. ZxZ is the set of all ordered pairs {(z,v): c E Z,y ¬ Z}.
Therefore the subsets A and A' form a partition of S.
3. Rx R( or R^) is the set of all R, y :a ¬ E
ordered pairs{(, y) R}
3. Let S be a subsets of S such
non-empty set and A,B are non-empty 4. R is the set of all ordered triplets {(r1,T2, #3) : oi E R}.
that each of ANB,ANB', A'n B,A' nB' is non-empty.
= = = = A'NB'. 5. R" is the set of all ordered n-tuples {(T1,F2,...,Tr): ci E R}.
Let Mi AnB,M2 AnB, M3 A' nB, M4
Then Mi U Ma U Mg UMa = S and Mn M, = ¢d,i #j. 6. Let A = {» ¬ R : 0 r s 1}, B = {r ¬ R : 1 < r < 2}. Then
AxB = :0 zs1,1< ys
Therefore the subsets M1, M2, Ms, M4 form a partition of S. {(r,v) ¬ R 2}. A xB is the square region
in R (Cartesian plane) bounded by the lines z = 0, z = 1,y = 1,y = 2.
 14
HIGHER ALGEBRA
SETS

Exercises 1 7. Let A, B be subsets of a universal set. Prove that A = B if and only if

AAB = o.
1. Let A =
{z e Z:0<<10},
{z e Z:z2 5} D E2:5
Bf={z ¬Z:r S15), 8. Let A, B,C be subsets of a universal set I and AAB = C. Prove that
(a) Find (i) AUB, (ii) An
15} (i) A = BAC, (ii) B = CAA,
B, (ii) B C, (iv) A- D.
(b) Verify that (ii) An (BUC) = A, (iv) B n (CU A) = B, (v) Cn(AUB) = C.
(i) Au(Bnc) =
(AUB)n(AUC), (i) = 9. (i) Find the number of . .
An(BUD) (ANB)U(AnD), subsets of the set S = {1,2,3,. ,20} such that the
(iil) (AUCy = A' nc', product of the elements in each subset is even.
(iv) (An D)' = A'UD'.
2.
(ii) Find the number of subsets of the set S {1,2,3,... , 25} such that
A, B,C are subsets of a universal set S. Prove that the product of the elements in each subset is a
multiple of 5.
=
10. Let A, B,C be subsets of a
() [An(BUC]n [A' U(B'
nC)) ¢, universal set S. Prove that
= =
=
(i) (An B) u (AnB') u (A' n B) u (A' n B') S,
(i) Ax (Bnc) (A x B)n(Ax C), (ii) Ax (BUC) (Ax B)U(Ax C),
(iii) (A U B)n (AUB) n(A'UB) AnB, =
(ii) (A- B) xC =
=(AxC)- (B xC), (iv) Ax (B-C) (AxB) -(AxC).
=
A,
(iv) (AnBnc)u (AnBnC') u (An B'
nc)u(AnB'nC')
=
(v) (AU BUC)n(AUBUC')
n(AUB'UC)n(A' UBUC) (AU B)n 1.7. Relation.
(BUC) n(CUA).
3. A, B,C are
subsets of a universal set S.
Let S and T be two non-empty sets.
Intuitively a relation p between
Find the complement of the
following sets in simplified form. S and 1 is a rule that associates some or all of the elements of S with
elements of T.
G) (AUBUC)n(AUB'UC'), (ii) (AUB'UC') n (A'UB' UC),
(ii) (A' nE nC) u (A'nB'nc)u(A' nBnC)u(A' nBnc). Definition. Let S and T be two non-empty sets. A
binary relationp
between S and T is a subset of Sx T.
4. A, B,C are subsets of a universal set S. Prove that If the ordered pair (s,
(i) (A - C ) u (B - C) = (AUB) - C,
t) E p, then the element s of the set S is said
to be related to the element t of the set 1T
by the relation P.
-

= -
(il) An(B- C) (AnB) (Anc), If (s, t) E (Sx T) p, then s is said to be not related to t by the
-
relation P.
= -
(ii) (A B) U (B-C) u (C - A) (AUBUC) (ANBnC),
-

(iv) AAB =(AU B) (An B) (A UB)A(An B), =

(v) AUB = (AnB)A(AAB),

Examples.
1. Let S be the set of all students and T be the set of all
teachers
(vi) An B = (AUB)A(AAB), of Presidency College. Let p be a relation between S and T defined by
(vii) An(BAC) = (AN B)A(Anc). specifying that a student s in S is related to a teacher t in T, if s happens
to be a student of t. It
may be that a particular student, say s1 of S is
5 A, B,C are subsets of a universal set S. a student of two teachers
ti and t2 of T but not a student of the teacher
(1) If A UB = AUC and AnB = AnC, prove that B = C. ta of T. Then (s1,t1) E p, ($1, t2) E p but (s1,ta) F p.
(ii) If AUB = AUC and AUB = A' uC, prove that B = C.
2. Let S = {2,3,4, 5},T = A relation p between S and
(iii) If AAB = AAC, prove that B =C.
{11, 12, 13, 14).
Tis defined by specifying that an element s in S is related to an element
t in T if s is a divisor of t. =
6. Let S be a universal set and A be a fixed subset of S. Then p {(2, 12), (2, 14), (3, 12), (4, 12)}.
(2,11) E S xT but (2, 11) p, since 2 is not a divisor of 11.
(i) If AUB = B holds for all subsets B, prove that A = d. Here 2 in S is related to two elements 12 and 14 of
(il) If AnB = B holds for all subsets B, prove that A = S. T, but 5 in S is
related to no element of T by the relation P.
 SETS 17
16 HIGHER ALGEBRA

* =
Definition. Let p be a relation between the sets A and B. Example 6. Let A {2, 3, 5,6), B {a, b, c, d} and p is defined by
g = {(3,a), (5, a), (5, c). (6, d)}. Then
The donain of p, denoted y D(p), is defined by p= {(2,c), (3, a), (5, c), (6, a)},
=
D(p) = {a: a ¬ A and (a, b) ¬ P}. =
pUa {(2, c), (3, a). (5,a), (5,c),(6,a), (6, d)}i pno {(3, o), (5,c)}.
The image of p, denoted by Im(p), is defined by relation p on S is a
Im(p) = {b:be B and (a, b) E p}. Definition. Let S be a non-empty set. A binary
subset of the Cartesian product S x S.
Example 3. Let A = {2,3, 5, 6}, B = {8, 10, 13, 20} and p is detined by
apb if and only if a is a divisor of b for a ¬ A,b E B. Then If be an element of S xS and (a, 6) ¬ p then a is said to be
(a,b)
= = related to b by the relation p. This is expressed by the symbol apb.
p={(2,8), (2, 10), (2, 20), (5, 20)}. D(p)
(5,10), {2,5} and Im(p)
{8, 10, 20. If (p, g) be an element of S x S but g) £ p then p is said to be not
(p,
related to q the relation p. This is expressed by the symbol p p q.
by
Definition. Let p be a relation between the sets A and B. case p
=
S x S then a pb for all a,b in S. In this
The inverse of the relation P, denoted by p*i, is a relation between In particular, if p
=
is said to be the universal relation on S.
the sets B and A and is defined by p-l {(6, a) : (a, b) E is said to be the
p} = then a þb for all a,b in S. In this case p
If p ó
Example 4. Let A = {2,3,5,6), B = {8, 10, 13, 20} and p is Tnull relation on S.
detined by p = {(2,8), (2, 20), (3, 13), (5, 13), (6, 20)} Then p-l=
(8, 2), (20, 2), (13, 3),. (13, 5), (20, 6)} Examples (continued).
= = 7. Let S =
p be the subset of
the set S x S given by
Note. D(p-1) Im(p); Imlp-1) D(P). {a, b, c} and
p= {(a, a), (b, b), (6, c), (c, c)}.
Composition of relations. a
pc, ba, cpa, cpb.
Then a p a, bp b, bpc, cpc, a pb,
= x S:
S
Definition. Let p be a relation between the sets A and B and o be aa a
college and p {(z,v) E
8. Let S be the set of all students of
relation between the sets B and C. Then the composite relation oop is "r on S defined by p y if and
r is a friend of y}. Then p is a relation
a relation between the sets A and C and is defined by ¬ S. c, y
only if is a friend of y" for
= E B such =
dop {(a, c) : a ¬ A,c E C} if there exists an element b 9. Let p be the subset of the set 2 x Z given by p {(z,y) E Z x Z:
that (a, b) E p and (6, c) E a. on Z defined by "c p y if and only
z+y is e v e n }. Then p is a relation
Note. D(rop) is a subset of D(p) and Im(aop) is a subset of Im(o). if z+y is even" for z,y E Z. Here (1,3) E p, (2,3) 2 P.
=

Example 5. Let A = {2,3,5, 6}, B= {8, 10, 13, 20}, C= {a, b, e,d) and 10. Let p be the subset of the set N x N given by p {(m, n) E N x N:
m is a divisor of n}. Then p is a relation on N defined by "m p n if and
p is defned by p = {(2, 8), (2, 20),. (3, 10), (5, 10), (6, 20)}, a is defined by
n" for E N. Here (2,4) E p, (4,2) p.
= {(8, b), (8, c), (10, a), (13, d). (20, c)}. Then only if m is a divisor of m, n
= ¬ E x E :
11. Let E be the set of all even integers and p {(z,v)
gop= {(2, 6), (2, c), (3, a), (5, a), (6, e)}. on E defined "r py if and only
t+y is e v e n f. Then p is a relation by
=
l {(8,2), (20,2), (10,3), (10,5), (20,6)) if r +y is even" for c, y E E.
element of E by the relation
op= {(2, 2), (3,3), (3, 5), (5, 3), (5,5),(6,6)}; Every element of E is related to every
= E x E. p is a universal relation on E.
pop= {(8,8), (8, 20), (20,8), (20, 20), (10, 10)}. p. Here p
=
Note that (2,8) E P. (8, 8) E a 2 p 8, 8 a b 2 dop b (2,6) E oop. 12. let D be the set of all odd integers and p {(z, y) E D x D:
D defined by "c p y if and only
o
be relations between the sets A and B. Then
r +y is odd }. Then p is a relation on
Definition. Let p and if c +y is odd" for z,y E D.
=

p and a are subsets of A x B. No two elements of D are related by the relation p.

Here p p. p is
pUd is defined to be the union of the relations p and o. a null relation on D.
pna is defined to be the intersection of the relations p and o.
18 HIGHER ALGEBRA SETS 19

a relation on a set A. The domain of p, denoted 1.8. Equivalence relation.

Definition. Let p be
by D(p), is defined by Let S be a non-empty set and p be a binary relation on S.
D(p) = {a: a E A and (a, b) E p}.
The The relation p is said to be refiezive if (a, a) E p for all a in S, i.e.,
image of p, denoted by Im(p), is defined by
Im(p) = {b:b¬ A and (a,b) E p}. a in S.
apa holds for all
The relation p is said to be symmetric if for any two elements a, bin
Example 13. Let A {1,2,3,4,5} and p = (1,1),(2,3), (3,4),
(3,5),(4,5). Then D(P) = {1,2,3, 4} and Im(e) = {1,3, 4, 5}. S, (a, b) Ep (8,a)) E p, i.e., apb~bpa.
The relation p is said to be transitive if for any three elements a, b,c
Let be a set A. The inverse of the relation in S, (a,6) ep and (b, c) Ep> (e,c) E p, i.e., a pb and b p c>apc.
Definition. p relation on a
P, denoted by p1, is defined by p l = {(b, a) : (a, 6) ¬ p}. or an
The relation p on S is said to be an equvalen.ce relation on s,
Example 14. Let p be the relation of Ex.13. Then RST relation on 5, if p is reflexive, symmetric and transitive.
P = {(1, 1), (3,2), (4,3),(5, 3), (5, 4))
Note. = =
Worked Examples.
DP-) Im(P): Imlpi) D(p). 1. A relation p is defined on the set Z by "apbif and only if a -b is
Composition of relations. divisible by 5" for a, b ¬ Z. Examine if p is an equivalence relation on Z.
-

Definition. Let p and o be relations on a set A. Then the composite (i) Let a E Z. Then a a is divisible by 5. Therefore a pa holds for
relation oop is defined by all a in Z and p is reflexive.
-
aop {(a, c) : a E A,c E A} if there exists an element b in (ii) Let a, b e Z and a p b hold. Then a b is divisible by 5 and
A such that (a, 6) Ep and (b, c) E o. therefore b - a is divisible by 5.

Note. is a subset of and is a subset of Thus a p b>bp a and therefore p is symmetric.

D(dop) D(P) Im(Top) Im(o).
-

Let a
hold. Then b and b-c a
Example 15. Let A = {2,3,5,6,7} and p = {(2,3), (2, 5), (3,3), (iii)
are both divisible by 5. Therefore a - c
a,b,c ¬ Z and p b, b pc both(a-b) + (b-c) is divisible by
o = Then
(5,5), (6, 6)}, {(3, 2), (5, 3), (6,5), (6,6) }. 5.
dop = {(2, 2), (2, 3), (3, 2), (5, 3), (6, 5), (6, 6))
Thus a p b and b p c > a p c and therefore p is transitive
poa = {(3, 3), (3, 5), (5,3), (6, 5), (6, 6)};
Since p is reflexive, symmetric and transitive, p is an equivalence
l={(3, 2), (5, 2), (3,3), (5, 5), (6, 6)}. relation on the set Z.
op={(2,2), (2, 3), (2, 5), (3, 2), (3, 3), (5, 2), (5, 5), (6, 6) Note 1. This relation p is said to be the relation of congruence (mod 5)
pop-1= {(3, 3), (3, 5), (5,3), (5, 5), (6, 6)}. on the set Z. If a, b E Z and a pb holds, a is said to be congruent to b
Note that (2,5) E P., (5,3) E a 2 p 5 , 5o 3 2 dop 3 >(2,3) E dop. (mod 5), and it is expressed as 'a = b (mod 5)'.

Definition. Let p and a be relations on a set A. Then p and o are For example, 5 = 0 (mod 5), 16 =1 (mod 5), -21 = 4 (mod 5).
subsets of A x A. Note 2. Let m be a positive integer and a relation p is defined on the
pUa is defined to be the union of the relations p and o. set Z by "a p b if and only ifa - b is divisible by m" for a,b E Z. Then
pno is defined to be the intersection of the relations p and o. p is an equivalence relation on the set Z.
Example 16. Let A = {2,3,5,6} and the relations p ando are defined 3. Let S be the set of all lines in 3-space. A relation p is defined on S
by p {(2, 5), (3, 2), (5, 6), (6,6)}, a = {(3,5), (5, 2), (5, 6), (6, 6)}. Then
by "l p m if and only if l lies on the plane of m" for l, m e S. Examine
pUa = {(2, 5), (3,2), (3, 5), (5, 2), (5,6), (6, 6)};: pna = {(5, 6), (6, 6)}. if p is (i) reflexive, (ii) symmetric, (ii) transitive.
 SETS
20 HIGHER ALGEBRA
21

It appears from the argument that a symmetric and transitive relation

(i) Let le S. Then l is coplanar with itself. Therefore l p l holds for
a set is an equivalence relation. But the argument
is fallacious.
on
allin S and p is reflexive.
p is reflexive if a p a holds for every a E A. 1E p is symmetric
(ii) Let 1, m ¬ S' and l p m hold. Then l lies on the plane of m. and transitive, apa holds on the assumption that there exists another
Therefore m lies on the plane of l. Thus lp m >mpl and therefore
p element b in A such that apb holds.
is symmetric. It may happen that for a particular element a in A there exists no
(ii) Let l, m,p ¬ S and p m , m pp both hold. Then lies on the element b in A such thatapb holds. In that case apadoes not hold.
For instance, let us consider Ex. 4. and transitive
plane of m and m lies on the plane of p. This does not always imply Here pis symmetric
not hold. This situation arises
that lies on the plane of p. That is, I p m and m p p do not but p fails to be reflexive since 0 p 0 does
necessarily because there is no other element b in Z such that 0 p b holds.
imply &pp. Therefore p is not transitive.

3. A relation p on the set N is given by "p = {(a,b) e Nx N:a| b}. Theorem 1.8.1. Let p be a relation on a set A. Then p is reflexive if
if p is (i) reflexive, (ii) symmetric, (ii) transitive. and only if 6a C p, where 6a = {(a, a) : a E A}, the set of all diagonal
Examine
a l b means a is a divisor of b] elements of the set A x A.
(i) Let m e N. Then m is a divisor of m. Let p be reflexive. Let (a, a) e QA for some a E A. Since p is
Proof.
Hence (n,m) E p for all m E N. So p is reflexive reflcxive, (a, a) E p. Therefore 6a C p.
(ii) Let m, n E N and (m, n) Ep. Then m is a divisor of n. This may Conversely, let 6a C p. Then (a, a) E p for all a E A. This implies p is
not imply that n is a divisor of m. reflexive.
Therefore (m, n) Ep does not necessarily imply (n,m) E p. This completes the proof.
So p is not symmetric.
Theorem 1.8.2. Let p be a relation on a set A. Then p is symmetric if
(ii) Let m, n,p E N and (m, n) E p, (n,p) E p. Then m is a divisor and only if p*i = p.
of n and rn is a divisor of p and this implies m is a divisor of p.
Proof. Let p be symmetric. Let (a, b) E p for some a,b E A. Since p is
Therefore (m, n) E p and (n, p) E p > (m, p) E p.
symmetric, this implies (b, a) Ep and this again inplies (a, b) E p-.
So p is transitive Thus (a,b) E p implies (a,b) E p . Therefore pCp .. (i)
4. A relation p is defined on the set Z by "a pb if and only if ab >0" Let (a,6) E p , Then (b, a) E p and since p is symmetric, (a,b) E p.
for a, b E Z. Examine if p is (i) reflexive, (ii) symmetric, (ii) transitive. Thus (a, b) E p-l implies (a, b) Ep. Therefore p-l cp (ii)
(i) Let a e Z. Then a.a >0 provided a z# 0. From (i) and (ii) we have p = p.
Therefore apa does not hold for all a in Z. So p is not reflexive.
Conversely, let p-l=p.
Let
(ii) a,b E Z and a pb hold. Then ab > 0 and therefore ba > 0. Let (a, b) E p. Then (b, a) ¬ p , by the definition of inverse. Since
Thus a pb =>bpa. So p is symmetric. p = p, this implies (b, a) E p.
(ii) Let a, b,c E Z and a p b, b pc both hold.
Then ab> 0 and be > 0. Thus (a,b) E p implies (b, a) Ep. Therefore p is symmetric.
We have (ab)(bc) > 0. This implies ac> 0 since b2 > 0. This completes the proof.
Thus a p b and b pe>apc. So p is transitive.
Theorem 1.8.3. Let p be a relation on a set A. Then p is transitive if
A fallacy. and only if pop C p.
Let p be a symmetric as well as a transitive relation on a non-empty Proof. Let p be transitive and let (a, c) ¬ pop for some a,c E A. Then
set A. Let a,b E A and a p b. there exists some bE A such that (a, b) E p and (b, c) ¬ p.
Since p is symmetric, a p b > bpa. Then a p b and bpa together Since p is transitive, this implies (a, c) E p.
imply a pa. Thus (a, c) E pop implies (a, c) E p. Therefore pop C p.
So p is reflexive. Consequently, p is an equivalence relation'
22 HIGHER ALGEBRA
SETS
23

Conversely, let pop Cp. Let

(a,b) E P, (b, c) ¬ p. Then
definition. Since (a, c) E pop by Theorem 1.8.7. The union of two reflexive relations on a set is a
pop C p, (a, c) ¬ p.
Thus (a,b) E p and reflexive relation.
(b, c) Ep (a, c) E p. Therefore p is transitive.
This completes the
proof. Proof. Let p, o be two reflexive relations on a set S.
Theorem
Let a E S. Since p and o are both reflexive, (a, a) E p and (a, a) E a.
1.8.4. Let p be a relation on a set A. Then p is an Therefore (a, a) E pUa. So pUa is reflexive.
relation on A if and only if equivalence
p, = Theorem 1.8.8. The union of two
(i) da C where dA {(a,
a): a E A}, symmetric relations on a set is a
ii) p =p and (iii) pop Ep. symmetric relation.
This is a combination of the Theorems Proof. Let p, o be two symmetric relations on a set S.
1.8.1, 1.8.2 and 1.8.3.
Let (a,b) E pUa. Then
an
Theorem 1.8.5. The inverse of (a, b) belongs to at least one of p and o.
equivalence relation is an
relation. equivalence Without loss of generality, let (a, b) E p. Since p is symmetric,
(b, a) E
p and therefore (b, a) E pUa.
Proof. Let p be an equivalence relation on a set S. Then is (a, b) E p U d (b, a) E pUo. So pUo is symmetric.
p reflexive,
symmetric and transitive.
(i) Let a E S. Since p is reflexive, (a, a) E p and therefore Note. The union of two transitive relations on a set S is not
(a, a) ¬ P, necessarily
by the definition of inverse. a transitive relation on S.
Thus a ES is reflexive.
(a, a) E p*. Therefore p-i Let us consider p, the relation of congruence
(mod 3) on the set Z; and
(ii) Let (a, b) ¬ for some a, b E S. o, the relation of congruence (mod 5) on the set Z.
p- Then (b, a) ¬ p. Since p is (2,5) E p, (5, 10) E a.
symmetric, (6, a) Ep > (a,6) E p and Therefore (2,5), (5, 10) E pUa. In order that the relation p U o
therefore (b, a) E p . may be
Thus (a, b) E p* (6, a) Ep. Therefore p* is symmetric. transitive, (2, 10) must belong to pUo. But (2,10) neither belongs to
Let p nor belongs to and therefore (2,
(ii) (a, b) E p*, (6,c) E p. Then (6, a) ¬E p. (c, b) E p. Since p shows that pU a is not transitive.
10) does not belong to pUa. This
transitive, (c, b) E p and (b, a) E p is
(c,a) E p and therefore (a, c) E p*
Thus (a, b) E p, (6, c) ¬ Let us consider p, the relation of
transitive.
p (a, c) E p. Therefore p is and o, the relation of congruence
congruence (mod 3) on the set Z
(mod 6) on the set Z.
-

-
Consequently, p- is an equivalence relation. Let (a,b) E o. Then a b is divisible by 6 and therefore a b is also
and =
divisible by 3. So (a, b) E p. This implies a
Cp therefore pUd p.
Theorem 1.8.6. The intersection of two equivalence relations on a set In this case p Uo is transitive.
is an equivalence relation.
These two examples establish that the union of two transitive relations
on a set is not
Proof. Let p, o be two equivalence relations on a set S. necessarily a transitive relation on the set.
Then p and o
are both reflexive,
symmetric and transitive.
E
Definition.
p
(i) Let a ¬ S. Since p and o are both reflexive, (a, a) and,
(a,a) E o. Therefore (a, a) Epno. So pno is reflexive. Let p be an
equivalence relation on a set S. Let a E S. Let be
a subset of S defined by cl(a)
(ii) Let (a,b) E pna. Then (a, b) EP and (a, b) E a. Since p and o
are both symmetric,
(b, a) E p and (b, a) Eo and therefore (b, a) E pNo. cl(a) = {a e S: rpaf.
So pna is symmetric. cl(a) is the set of those elements a of S such that r pa holds.
to
(iii) Let (a,6) ¬ pno, (b, c) E pno. Then (a, b), (b, c) E p and cl(a) is a non-empty subset of S since a E cl(a). cl (a) is said
be the p-equivalence class of a and each element of
be
(a, b), (6, c) E o. Since p and o are both transitive, (a, c) E p and (a, c)
¬E cl(a) is said to
and therefore (a, c) E pna. So pna is transitive. p-equivalent to a.
Consequently, p n a is an equivalence relation on S.
For example, in Ex.1 of =
1.8, cl{0) {0, +5, +10,...}
24 HIGHER ALGEBRA SETS 25

Theorem 1.8.9. Let p be an equivalence relation on a set S and a, b E S. Example (continued).

Then cl(a) = ei(6) if and only if a p . 7. Find the equivalence classes determined by the equivalence relation p
on 2 defined by "a pb if and only if a - b is divisible by 5" for a, b E Z.
Proof. Let cl(a) = ci(6) and let r E el(a). Then z ¬ cl(b) also.
E d{a)>*pa E cl(6) > cpb. There are five distinct equivalence classes. They are
and z
Tpa and r pb * a p r and c p b, since p is symmetric cl(0) {0, +5, t10, +15, .} {5n: n is an integer ,
apb, since p is transitive. cl(1) {1,1t5,1 t 10, ..} {5n +1: n is an integer
Therefore =
cl (2) {2,2 +5,2+10, ..} 5 n +2: Tn is an integer
el(a) el{b) >ap b. 3,3 5 , 3 t 10, } {5n +3:n is an integer
cl(3)
Conversely, let a p b. Let z E cl(a). Then c pa holds. = {5n +4 :n is an integer }.
cl(4)= {4,45,4 +10, }
zE cl{a) and a p b t p a and apb Note. Let a E 2. By division algorithm there exist integers q and r such
r p b , since p is transitive that a=5q +r, 0 r <5. r is called the least non-negative residue of
E cl(b). a (mod 5). a belongs to cl(7).
This proves that cl(a) c d(b). Similarly cl(b) c cl(a). For example, 34 E cl(4) since 4 is the remainder when 34 is divided by 5.
It follows that cl(a) = cl(b).
Thus a p b > d(«) = el(6) and this proves the theorem. The classes_ci(0),cl(1), el(2), el(3), d(4) are also called the classes of
residues of Z (mod 5).
1.8.10. be an equivalence relation on a set S and
Theorem Let p
a,bE S. If a pb then el(a) and ci(6) are disjoint. Theorem 1.8.12. An equivalence relation p on a set S determines a
partition of S. Conversely, each partition of S yields an
Proof If possible, let cl(a) n cl(ö) # ó and p e ct(a) n d{t). equivalence
relation on S.
pE cl(a) Ppa, pE c{6) Ppb.
Ppa and p pb a pp and p pb, since p is symmetric Proof. Each element of S belongs to a
p-equivalence class, because if
apb, since p is transitive. p¬S then p E cl(p). If a,b be elements of S then the
p-equivalence
This is a contradiction. Therefore cl(a) n cl(6) = ¢. classes cl(a) and cl{6) are either disjoint or
equal.
If we consider the family of distinct
an p-equivalence classes we observe
Theorem L.8.11. Let p be equivalence relation on a set S and that (i) each class in the family is
a,b E S. Then the classes cl(a) and are
cl(6) either equal or disjoint. non-empty
Proof. Since a, b e S, (a,b) ¬ Sx S. (ii) the union of the family of classes is the set S, and
Since p is a binary relation on S, either (a, b) E p or (a, b) £ p. (iii) the classes are pairwise disjoint.
Let (a, b) E p. Then a p b holds.
of
Therefore the family of distinct p-equivalence classes form a
We have cl(a) = cl(6), by Theorem 1.8.1. the set S. partition
Let (a,b) gp. Then cl(a) and cl(6) are disjoint, by Theorem 1.8.2.
Consequently, either cl(a) = el{b) or cl(a) n cl(6) = d. Conversely, let there be a partition P of the set S
into subsets.
Let us define a relation
p on the set S to mean that a p b holds if a
an
and b belong to one and the same
If p be equivalence relation on a set S then the family of distinct subset of the partition P.
p-equivalence classes is such that Let a E S. Then a p a holds
since a and a belong to one and the
i) each class in the family is non-empty, same subset of the
partition P. Therefore p is reflexive.
i ) the union of the family of classes is the set S and Let a, b E S and a p b. Then a and b
(iii) the classes are pairwise disjoint. subset of the partition P and therefore b belong to one and the same
and a belong to the same subset
Therefore the distinct p-equivalence classes form a partition of the set S. of P. That is, apb>bpa. Therefore p is symmetric.
26 HIGHER ALGEBRA
SETS
27
a pb,
Let a,b, c E S and
and the
bpc both hold. Then a and b belong to one 1.9. Partial order relation.
same subset, say S' of P; b and e belong to one and the same
subset, say S" of P. Let S a on is said
Definition. be non-empty set. A relation p the set S
to be antisymmetric if a p b and b p a a = b for a , b E S.
partition, of a
either identical or nust be
Sand S'" being subsets
disjoint. Since b E S ' n S", it follows that S' = S' and consequently,
a and e belong to one and the same subset of P. That is, a p b and Examples.
bpcapc. Therefore p is transitive. 1. The relation p defined on the set R by "a pPy if and only if z for
an on
T, y E R is antisymmetric.
Thus p is equivalence relation S. This completes the proof.
2. Let X bea non-empty set. The relationp defined on P(X) by "A p B
Worked Examples. if and only if A is a subset of B" for A, B E P(X) is antisymmetric.
1. A on the set A =
relation R is defined {1,2,3,4} by 3. The relation p defined on the set Z by "a p b if and only if a | 6" is
R={(1,1), (1,3), (2, 2), (2,4), (3, 1), (3,3),(3,4),(4,2),4,3)}. Examine not antisymmetric, because 1p(-1) and (-1) p1 but 14 -1.
if (i) R is reflexive, (i) R is symmetric, (iii) R is transitive.
i) 6A = {(1,1), (2,2), (3, 3), (4, 4)}. Let be
Definition. a non-empty set. A relation p on S is said to be
S
a partial order relation if p is reflexive, antisymmetric and transitive.
oA is not a subset of R. So R is not reflexive.
A relation of partial order is often denoted by '<, even if it is not
i) R-= {(1,1), (3, 1),. (2,2), (4,2),(1,3), (3,3), (4,3), (2,4),(3,4)}. "less than".
R = R. So R is symmetric.
Poset. A non-empty set S together with a relation of partial order
() RoR={(1,1),(1,3).(1, 4), (2, 2), (2, 3), (2, 4), (3, 1), (3, 2), (3, 3), onS is called a poset (partially ordered
(3,4), (4, 1), (4,2),(4,3), (4,4)}. set) and is denoted by (S, <).
RoR is not a subset of R. So R is not transitive.
Examples (continued).
2. A relation R is defined on the set A = {1,2,3, 4, } by 4. a c
(R, S) is poset, where y means "c is less than or equal to y"
R= {(1, 1), (1, 3), (2,2), (2,4), (3, 1), (3, 3), (4,2), (4,4)}. Show that R is for c, y in R.
an equivalence relation. Describe the R-equivalence classs.
5. Let X be a set and
non-empty P(X) be the power set of X. (P(X), <)
(i) SA= {(1,1), (2, 2), (3, 3), (4, 4)}. Sa is a subset of R. So R is is a poset where A <B means "A is a subset of B" for A, B E P(X).
reflexive.
6. Let X be a non-empty set and S be the set of
all proper subsets of X.
(i) R-- {(1,1), (3,1),(2,2), (4,2), (1,3),(3,3), (2,4),(4,4)). (S,S) is a poset where A B means "A is a subset of B" for
= R. So R is A, B E S.
R symmetric.
n
7. (N, S) is a poset, where m
< means "m is a divisor of n" for
(ii) RoR ={(1,1), (1,3), (2, 2), (2,4), (3, 1), (3, 3), (4,2), (4,4)}. RoR m, n E N.
is a subset of R. So R is transitive.
8. Let S be the set of all
Consequently, R is an equivalence relation. positive divisors of 72. (S, <) is a
poset, where
a<b means "a is a divisor
of 6" for a,b in S.
cl(1) = {r E A: (1,r) E R} ={1,3}.
ct(2) {r E A: (2,z) E R} ={2,4} Worked Example.
= E A: E
1. Let ,S) be a poset. Define a relation
cl(3) {r (3,z) R} ={3,1). 2 on S by "a 2 b if and only
cl(4) = {z E A: (4,z) E R} ={4,2). if b a", for a,b E S. Show that (S, 2) is a poset.
Since a a for alla in S, a2a for all a in S. Therefore
The R-equivalence classes are ci(1) = cl(3) = {1, 3}, cl(2) = d(4) = 2 is reflexive.
and a
Let a,b ES and a2 b,b a . Then
{2,4). b<a < b.
This implies b =
a, since < is antisymmetric.
 SETS 29
28 HIGHER ALGEBRA

So a 2 b and b2 a = a = b. Therefore 2 is antisymmetric. can contain at most o n e element a which

Definition. A poset (S, <)
a ¬ S. Because be another such, say a',
if there
Let a,b,c E S and a b,b 2 c. Then b a and c s b, i.e., c satisfics a < a for all
as well as a' a
must hold, whence a =a' by antisymmetry.
b and ba . then a a the least element of S.
Such an element a, if it exists, is called
This implies c a, since is transitive. That is, a c.
similar
arguments, a
poset (S, <) c a n contain at most o n e element
So a 2 b and b2 c a 2 c. Therefore> is transitive. By
satisfiesz for all z E S. Such an element b, if it exists, is
a poset b which b
Hence is a partial order relation on S and (S,2)
called the greatest element of S.
Definition. The poset (S, 2) is called the dual of the poset (5, ) . More element of any subset X of S is an element
generally, the least
all z E X. The greatest element of any subset
a EX such thatac for
Covering relation.
X of S is a n element b EX such that t < b for all z E X.
Let (S, ) be a poset. Define a relation < on S by "a < b if and only
be a An element a E S is said to be a
but a # b", a,b e S. Definition. Let (S, <) poset.
ifab b is said to cover a if a < b and there is no element e in S such that minimal element if z < a for no z E S. An element b E S is said to be a

a<c<bholds. mazimal element if b <z for no a E S.

For example, in the poset (N, S) of Ex.7, 6 covers 2, 10 covers 2; but Clearly, the least element in a poset is a minimal element and the
8 does not cover 2 although 2 < 8, because 2<4<8. greatest element in a poset is a maximal element but the converse is not
true.
The covering relation can be used to give a graphical representation
of the poset S. Small circles are drawn to represent the elements of S.
A line segment is drawn upwards from a tob when b covers a. Examples (continued).
9. Let X = {1,2, 3} and S be the set of all proper subsets of X. Then
Let S be the set of all positive divisors of 12. S = {1,2,3, 4,6, 12}. (S,) is a poset where A B means "A is a subset of B"for A, B E S.
The covering diagram of the poset (S, <) is given by the figure 1. The poset (5, ) contains no greatest element and no least element.

30 There are three minimal elements {1}, {2}, {3} and three maximal el-
ements {1,2}, {2,3}, {1,3}.
10. The poset (N, <) of Ex.7 contains no greatest element and no m a x
imal element. The least element is 1 and 1 is the
10 only minimal element.
6
Definition. Let (S, <) be a poset and a, b be two elements of S.
An element c E S is said to be an
upper bound of the pair a, b if a < c
and b C.
An element u E S is said to be a least
upper bournd (lub) of a and b if
(i) u is an upper bound of a and b, i.e., a u, b u;
and (ii) if cES be an upper bound
Figure I Figure 2 of the pair a, b then u C.

a
An element d e S is said to be
Let S be the set of all positive divisors of 30. S= {1,2, 3, and d b.
lower bound of the pair a,b
if d <a
5,6,10,15,30). The covering diagram of the poset (5, S) is given by
the figure 2. An element v E S is said to be a greatest lower bound (glb) of a and
if
Note. The dual poset (S, 2) is obtained simply by turning the covering
diagram of (S, S) upside down.
(i) v is a lower bound of a and b, i.e., va,v<b; and
(ii) if d E S be a lower bound of the pair then d a, b v.
 30 HIGHER ALGEBRA
SETS
31
Theorem 1.9.1. Let (S, <) be poset.a
Definition. Lattice. A poset (L, 3) is called a lattice if every pair of
i) If a, b E S have a least upper bound then that is unique. elements of L has a least upper bound and a greatest lower bound.

(ii) If a,b E S have a greatest lower bound then that is unique.

Proof. (i) Let u1 and u2 be two lub's of a, b. Examples (continued).
13. The poset (P(X), <) of Ex.5 is a lattice where for any two elements
Since u1 is a
upper bound a u1,b u1. =
of the pair a, b; A, B of P(X), AVB AUB and An B AnB.
Since u2 is a
least upper bound, u2
1 14. The poset (N, ) of Ex.7 is a lattice where for any two elements
Interchanging the roles of u1 and u2, u1 u2. a,b
of N, a V b= lcn(a, b) and a A b = gcd(a, b).
u1 u2 and u2 u1 > u1 =u2, by antisymmetTy.
(ii) Proof left to the r.
of
rea 15. For X -{1,2,3) the poset Ex.6 is not a lattice because
(S, s)
the pair of elements {1, 2} and {2, 3} has no lub and the pair of elements
Note. The unique lub of the elerments a, b is denoted by a Vb (a join b). 1 and 2 has no glb.
The unique glb of the pair a,b is denoted by a ^b (a meet b). 16. Let P be the set of al real valued functions defined on the closed
This is to note that if the elements a,b in a poset
(S, <) have an upper interval(0,1. Let f,g EP and f sg mean that f(r) < g(r), z e [0,1].
bound (a lower bound) then they
may not have a least upper bound (a a lattice with fVg as the lub and
greatest lower bound). Then(P ) is a poset. Theposet is
fAg as the glb, where fVg and fAg are defined on [0,1] by
fV g() = max {f(r), g(r)}, r ¬ [0, 1]
Examples (continued). fA g(r) = min z E
11. In the poset (P(X), <) of Ex.5, the lub of A and B is AUB and the {f(z), 9(r)}, [0, 1]
glb of A and B is AnB.
12. In the poset (N, <) of Ex.7, the lub of a and b is the lcm of a, b and Exercises 2
the glb of a andb is the gcd of
a, b. 1. Determine the nature of the on the set 2.
following realtions p
(1) a pb if and only if a E Z, b e Z and a b,
Worked Example (continued).
2. Let n eN and S be the set of all (i) a pbif and only if a E Z, b E Z and ab 2 0,
positive divisors of n. In the poset
(5, )where a b means a is a divisor of b for a, b e S, prove that (iil) a p b if and
only if a e Z,b E Z and a2 +b2 is a multiple of 2,
aAb=gcd(a, b), aVb= lem(a, b). (iv) a p b if and only if a e Z, be Z and 2a +3b is divisible by 5.
Let d= gcd(a, b). Then dja and d|b; and if p be common divisor of a 2. Examine if the relation p on the set S is (i) reflexive, (ii) symmetric, (ii)
and b then pld. transitive.
Therefore d a, d< b; and p< a,p<b>p<d. (a) S is the set of all lines on a plane and p is defined on S by 1 p m if
and only if l is perpendicular to m'" for 1, m E S.
d a a, b; and p is a lower bound of the
Hence is
pair a, b implies p
lower bound of the pair
d. Consequently, d = anb. That is, ged(a, b) = a^b. (b) S = Z and p is defined on Z by "a p b if and only if a - b < 3" for
a,b E Z.
= Then and
all b|l; and if g be common multiple of a
Let lem(a,b). (c) S = Z x Z and p is defined on Z x Z by "(a, b) p (c, d) if and only if
and b then lq. a+d b +c' for (a, b), (c, d) e Zx Z.
Therefore a<l,b< i and a
g,b S q l < q. (d) S = Z x Z and p is defined on Z x Z by "(a, b) p (c, d) if and only if
Hence l is an upper bound of the pair a, b; and q is an ad = be" for (a, b) , (c, d) ¬ Z x 2.
upper bound of
=
the pair a, b implies lsq. Consequently, =aVb. That is, lcm(a, 6) (e) S= N x N and p is defined on N x N by "(a,b) p (c, d) if and only if
aVb. ad = bc" for (a, b), (c, d) e Nx N.
32 HIGHER ALGEBRA SETS 33

3. Examine if the relation p on the set Z is an eqivalence relation. of f have the same first eomponent. If (z, y) E f, f is said to assign y to

(a) p = {(a, b) E Z x Z: 3a + 4b is divisible by 7} the element z of A.

(b) p = {(a, b) E Z x Z:la - b i 3). Definition. Let A and B be two non-empty sets. f fromA mapping A
to B is a rule that to each element z of A a definite element y in
4. A relation is defined o n Z by " y if and only if r2 -
assigns
is divisible by
y
for
B.
5 , y E Z. Prove that is an equivalence relation o n Z. Show that there
A is said to be the domain of f and B is said to be the co-domain
are three distinct equivalence classes.
of f. The mapping f with the domain A and co-domain B is displayed
= =
5. Let S {a, 8, c, d} and p {(a, a), (b, b), (c, e), (d, d), (a, b), (b, a), (b, c), symbolically by f: A~ B.
(C, b)}. Show that the relation p is reflexive and symmetric but not transitive.
We can imagine f as a kind of agent that carries (or transforms, or
6. A relation R is defined on S = {1,2,3,4} by R = {(1, 1), (1,2), (2, 1)
(2,2),(3,3), (3, 4), (4,3), (4, 4)}. Show that R is an equivalence relation on S.
maps) each element e of A to a unique element y in B.
A mapping f is also called a function, or a transformation, or a map.
7. (i) Let S be a finite set containing two elements. How many different binary
relations can be defined on S? How many of these are reflexive? Let f: A ^ B be a mapping and « E A. Then the unique element
(ii) Show that the number of different reflexive relations on a set of n y of B that corresponds
to (is associated with) z by the mapping f is
elements is 2 - called the f-image of z (or the image of z under f) and is denoted by
f(r). If f(*) = y, we often say that ' f maps z to y'.
[Hint. (ii) The total number of subsets of A x A is 2*. A reflexive relation on A
is a subset of A x A that contains all elements of the set {(a, a) : a E A}. The total
The set of all f-images, i.e., the set {S(«) : z E A} is denoted by
number is - -
S(A) and is said to be the image set of f (denoted by im S) or the range
(n2 n)co t+ (n2 n)e, ++(n2 - n)c.a_,l set of f.
8. Define a relation p on C by "(a + ib) p (c + id) if and only if a c and
domain of f is denoted by D() and the range of
b ' for (a + ib), (c + id) E C. Show that p is a partial order relation. Iu some texts the
f is denoted by R(f).
9. Let S be the set of all positive divisors of 72. Define a relation on S by
y if and only if r is a divisor of " for z, y E S. Prove that (S, <) is a
poset. Draw the covering diagram of the poset.
Examples.
1 Let S -{1,2,3,4},T = {a, b, c, d}. Let us examine the following
10. Do the same as in Ex.8, where S is the set of all positive divisors of 60. relations f1, f2, fs. f4 between S and T.
11. Let S = {1,2,3, 4,5,6,7, 8, 9, 10). Let z y mean that z is a divisor of y. =
i) Si {(1, a), (1,6), (2, c), (3, c), (4, d)}.
Find the maximal element/elements in the poset (5, ) . (i) Ja = {(1, ) , (2, 5), (3, c).

(ii) Ss={(1,6), (2, 8), (3, c), (4, d)}.

1.10. Mapping.
Si is not a mapping from S to T since the element 1 in S is related
Let A and B be two non-empty sets. If f is a relation between A and to two different elements of T by the relation.
B then an element z of A may be related to one element or no element
or many elements of B by the relation f. S2 is not a mapping from S to T since the element 4 in S is not related
to any element of T by the relation.
A relation f with the property that each element c of A is related to
exactly one element y of B, is said to be a mapping from A to B. fa is a mapping from S to T. Here the image set is {6, c, d} and it is
a proper subset of the co-domain set T.
Since a relation f between A and B is a subset of A xB, f is a mapping
2. Let =
from A to B if each element z of A appears as the first component exactly f {(z, y) E R xR: y Let us examine if f is a
=>f mapping
in one of the ordered pairs of f. Therefore no two distinct ordered pairs from R to R.
 HIGHER ALGEBRA SETS 35
34

7. Let f : R ^ R be defined by f(r) = 2c, r E R. For an element y in

The element 0 in the domain set R is not related to any element of
the co-domain set. Therefore f is not a mapping from R to R. the co-domain set R, f-(y) == {v}, a one-element subset of the domain
=
set R.
E Sx R: y = is a
Let S = R - {0}. Then f {(z,v) } mapping
from S to R. It is displayed symbolically as Injective mapping. Surjective mapping. Bijective mapping.
"f: S R is defined by f(r) = , rE S"
A mapping f : A B is said to be injective (or one-to-one) if for
Into mapping. Onto mapping. each pair of distinct elements of A, their f-images are distinct.
A mapping f : A B is said to be an into mapping if f(A) is a A mapping f: A-~ B is said to be surjective (or onto) if f(A) =B.
proper subset of B. In this case we say that f maps A into B. A mappingf:A ~ B is said to be bijective if f is both injective and
A mapping f : A » B is said to be an onto mapping if f(A) = B. In surjective.
this case we say that f maps A onto B. An injective mapping is called an injection, a surjective mapping is
called a surjection and a bijective mapping is called a bijection.
Examples (continued).
3. Let = Thus f: A- B is injective if a1 # t2 in A implies f(z1) # f(z2) in
f Z - Z be defined by f(z) 2a, « e Z. Then f is an into
mapping because f (Z) (the set of all even integers) is a proper subset of B. In this case, each element of B has at most one pre-image.
the co-domain set Z. If f is surjective, each element of B has at least one pre-image.
4. Let f: Z » z be defined by f(r) = l|, z e Z. Then f is an into
If f is bijective, each element of B has eractly one pre-image.
mapping because f(Z) (the set of all non-negative integers) is a proper
subset of the co-domain set Z. In Example 3, f is an injective mapping since for two distinct elements
=
5. Let f: Z z be defined by f(r) c++1, z e Z. Then every element T1, T2 in the domain of f, # f(z2). f is not surjective because
f (c1)
-

in the co-domain set Z has a pre-image y 1 in the domain set Z. f(Z) is a proper subset of the co-domain set Z.
y
Therefore f(Z) = Z and f is an onto mapping. In Example 4, f is not injective because two distinct elements 1 and
IEf: A B be a mapping and z e A, then f(r) is a definite element -1 in the domain of f have the same f-image. f is not surjective because
in B. The element z is said to be a pre-image (or inverse image) of f(r). f(Z) is a proper subset of the co-domain set Z.
It may happen that an element y in the co-domain set B has only one In Example 5, f is injective since for two distinct elements c1, t2 in
the domain of f, f(ri) # f(r2). f is surjective because f(Z) == the
pre-image, n o pre-image o r many pre-images in A.
In Example 4, 0 in the co-domain set Z has only o n e pre-image in the co-domain set Z. Therefore f is a bijection.
domain set; 1 in the co-domain set Z has two pre-images in the domain
set;-2 in the co-domain set Z has no pre-image. When f is a bijection from A to B, f sets up a one-to-one correspon-
a of dence between the elements of A and the elements of B. Each element z
Thus the pre-images of an element y in B form subset A, which
may be the null set, or a singleton set (a set containing one element only), of A is put in correspondence with the single element f(z) of B and each
set (or the element y of B is put in correspondence with the single element f-(y)
or a set containing more than one elements. The pre-image
inverse image set) of y is denoted by f-^(y). ofA.

Definition. A mapping f: A B is said to be a constant mapping9

Examples (continued). (a constant if
function) f maps each element of A to one and the samme
= a
6. Let f: R- R be defined by f(r) sin a, E R. Here the image element of B, i.e., f(A) is a singleton set.
set of f {zis E R: -1 < S 1}. Every element in the image set has
= z ¬ R is
infinite numberof pre-images. For example, the pre-image set of 1 is For example, the mapping f :R R defined by f(r) 2,
a constant =
{(4n+1)5 : n is an integer} mapping. Here f(R) {2}.
 SETS 37
36 HIGHER ALGEBRA

=
Let f : R^ R be defined by f(T) r, c e R.
Definition. A mappingf: A A is said to be the identity mapping on
12. real numbers). The restriction
=
A if = Let S Rf(the set of all positive
f(z) z,c E A. The =
identity mapping on A is denoted by i4. is defined by S/S() ' , r E S.
Note. The mapping f/S : S R
restriction mapping f/S is injective.
identity mapping on A is clearly a bijection on A. Here f is not injective, but the
z, z E R. This mapping f
Definition. Equality of mappings. Two mappings f: A B and 13. Let f: R - R be defined by f(z) =sin
g: A C are said to be =
equal if f(c) g(z) for all z ¬ A. For the is neither surjective n o r injective.
= e
R: - 1 < z < 1}, then the
equality of two mappings f and g the reduce the co-domain to T {z
following conditions must hold: Ifwe =
but
T defined sin c, c E R is surjective,
(i) f and g have the same domain D; and mapping f :R by Í(E)
for all z e =
not injective. mapping
(ii) D, f() g(=). r Then the restriction
Let S = {r ¬ R :- Z}.
= sin c, t ¬ S is a bijection.
Examples (continued). g:S^T defined by g(c)
8. Let S =
{r E R: c 0). Let f: S
R be defined by f(c)
= Worked Examples.
E S and g: S Examine if is
= R. f
R be defined by g(«)
1, z e S. Then f = 9. 3z +1,z E
9. Let S =
1. Let f : R > R be defined by f(c)
{z E R: 1 r < 2}. Let f S :
= R be defined by (1) injective, (ii) surjective.
f(c) z E
a - c, =
z
S andg:S^R be defined
by g(c) 1-a, E S. elements z1,T2 in R, the domain of f.
Then f=9. Let us take two distinct
(i)
=
10. = 31 +1, S(z2) 3t2 +1.
Let S {z E R: 0 a S Z}. S f(1) = since 1 r2.
by a) = cos t - sinc, c E S and g: S et
- Rf :be defined be defined
R by g(z) = S(o1)- f(c2) 3(71 -c2) # 0,
1 - sin 2c, a ¬ S. Since ti # T2f(T1) # f(r2), f is injective.
Here f and g have the same domain. co-domain of f;
f(z) Let us take an
arbitrary element y in the set R, the
9(c) forre [0, 4. But ii) the domain of f.
and let us examine if y has a pre-image t in
S(r) # g(z) for z E (}, ~. Therefore f # g. =
= or, z=
of a
Then f(c) y and therefore 3c +1 =y
Restriction mapping.
E E R.
Therefore y has a pre-image in the
Let f: A B be a mapping Since y R,
and let D be a
non-empty subset of A. domain of f. arbitrary, each
Since y is element in the co-domain of f has
Then the mapping g: D B defined by g(c) = f(), a ¬ D is
be the restriction of f to D. This is said to a pre-image under f. Therefore f is surjective.
denoted by f/D.
In this case, f is said to be an
ertension of g to A. An 2. Let f: R » R be defined by f(r) = z2, a e R. Examine if f is (i)
f: A B of the
mapping g: D B is not unique,
extension injective, (ii) surjective.
-
of the elements of A D may be
since the f-images
chosen. arbitrarily =
4, f(-2) = 4. f is not injective, since two distinct elements 2
When a () S(2)
non-bijective
mapping f is given, sometimes it is to
and -2 in the domain of f have the same image.
have a bijective restriction of possible
f. (ii) f is not surjective, since -1 in the co-domain of f has no pre-image
Examples (continued). in the domain of f.
11. Let =
f:R>R be defined by f(r) 1, ifz be rational
=
1.11. Composition of mappings.
0, if z be irrational.
Then the restriction Let f: A B and g : C ^ D be two mappings such that f(A) is a
mapping f/Q: Q R is =
for all zE Q given by f/Q(7) 1 subset of C.
The restriction mapping f/ R is Let E A. Then f maps z to an element y in f(A) C B and since
= -
given by f/(R-
Q)() 0 for all z ¬ (R Q). (R-Q): (R-Q) yEf(A) c C, g maps y to an element z in D.
 SETS
NiHER ALEBRA 39

=
fog: R is dotined by fog(r) = f(x + 5) r +6, a e R;
e can meie à: A D detint by a(r) g(f()),
t a mapping gof R Ris defined by gof(r) = g(r + 1) = r +6, rE R.
The mapping A D is said to be the composite (or the
E od and s and is dewtai by g»f (or by gf). The composite Since = for all x e we =
gof (r) fog(r) R, have gof fog.
A D is dotinni ony it f{) is a subset of the domain ofg.
It is quite
For the mappings: A and g: B - C the composite gof
is
clear from the examples that the composition of mappings
That
i s deñni. For the mappins f:A-B nd g:B»A both the not commutative. is, for two mappings f and g their composites
gof and fog are, in general, not equal. In fact, when one of them is
nites gof A Aand fog: B are derined.
detined. the other may not be detined at all. In particular cases, however,
the equality holds.
Exampies.
1. Let A = {1,2,S}, B = {p.g,r}. and let Although the composition of mappings is not commutative, it is as-
sociative. That is, for three mappings f,g and h, ho(gof) = (hog)of,
A Bbe detineri =
by f{i) wlen botl1 sicles are defined mappings.
A be deinei by gip) = 3 . g(g) = 2, g(r) =
=p. f(2) 9, f(3) = 1;ri
B A be detined by h(P) = 1.h(g) = 2 , h{r) = 3. Theorem 1.11.1. Let f : A -B, g: B- C, h: C D be three
=
mappings. Then ho(gof) = (hog) of.
fAA is deined by gof(1) 1.
=3,g»f(2) =2.gof(3) Proof. Here the composite mappings gof, hog are defined because the
fog: B - B is defined by foo(p) = r, fog(a) = q.fog(r) = p.
range C dom g and the range g C dom h. The composite mappings
= = 3.
àf:A-A is deined by hof(1) 1, h>f(2) =2, hof(3)
= =
he(gof). (hog)of are defined because the range gof c dom h and the
= r. range f c dom hog.
foh: B B is deined by foh(p) p. foh(a) q.foh (r)
= = We shall now prove the equality of : A D
Here gef #fog, hof iA. foh ig. the mappings ho(gof)
and (hog)of: A -- D.
: Z - Q and g : Q - Qbe deined by f(z) = z , z e Z and
Let = = =
y, g(y) wo.
be an element of A and let f(c) z, h(z)
s) Then gof(r) = g(y) = z, hog(y) = h(z) = w.
= z eZ.
gof:Z- Q is deined by gof(r) =s(j#) }*, holgos) :A^ D s defined by ho(gof) (T) = h(z) = w, z E A.
= = = =
(hog)of :A - D is defined by (hog)f(c) = hog(y) = w, a E A.
For example, gef(1) =sf(1) I() gof(2) gf(2))
g{l) = 1. Since =
ho(gof) (r) for all z e A, we have =

Here feg is not defned since the range of g is not a subset of the
(hog)of (z) ho(gof)
(hog)of.
domain of f.
Theorem 1.11.2. : A B and g :
3. Let f: R -R and g: R R be defined by f(z) =z+ 1, ER and If f B - C be both injective
mappings then the composite mapping gof: A - C is injective.
g(c) = 3z, z E R
Proof. Let c1,F2 be two distinct elements of A.
Here gof and fog are both defined.
Let f(1) = V1,
gof R R is defned by gof() = g(z + 1) = 3z +3, z ¬ R;
f (T2) = y2
Since f is injective, y1 and y2 a r e distinct elements of B.
= =

fog: R R is defined by fog(z) S(3z) 3z +1, r E R. Let glyi) =

Z1, g(y2)= 22.
oth the mappings gof and fog have the same domain but gof(z) f Since g is injective z1 and z2 are distinct
elem of C.
fog(r), z e R. Therefore gof # fog. Now =
gof
(z1) = g(y1) = z1, gof(c2) = 9(v2) z2 and z1 # r2 in A
z1 Z2 in C. is injective.
Therefore gof
4. Let f: R ~ R be defined by f(r) = z + 1, z E Rand g: R R be
defined by g{z) = z +5, z E R. Note. The converse of the theorem is not true. However if
gof is injective
then f is injective (while g need not
Here both gof and fog are defined. be).
 40 HIGHER ALGEBRA
SETS 41

Theorem 1.11.3. If f: A- B and g: B - C

that two mappings such be Theorem 1.11.6. If f : A B be both
gof: A - C is injective then f is injective. and g: B^C bijective then
=
thecomposite mapping gof: A C is bijective.
=
Proof. Let f(zi) f(z2) for some i , #2 in A. Then
This is a
=
since g is a mapping. So g[f(r1)] g[f(#2)], combination of the Theorems 1.11.2 and 1.11.4.
gof(71) gof(r2).
=
=
Because gof is injective, gof(T1) Note. The converse of the theorem is not true. However if gof is bijective
gof(c2) implies zi T2
=
Thus f(ri) = then f is injective and g is surjective.
f(z2) implies ?1 T2 and therefore f is injective.
This follows from the Theorems 1.11.3 and 1.11.5.
Note. In order that gof may be
injective it is not necessary that g is In order that gof may be bijective, it is neither
injective. necessary that f is
surjective nor necessary that g is injective.
For example, let f: R - =
R be defined by f(s) e", z E
g:R be defined Rand The example given in the note of the previous theorem establishes
R by g(z) =r', c ER. the assertion.
Here gof : =
z ¬ R.
R-R is deined by gof(r) es,
gej is injective but g is not injective. 1.12. Inverse mapping:

Theorem 1.11.4. Definition. Let f:A+ B be a

mapping. f there exists a mapping
If f: A - B and g: B C be both surjective then g:B A such that gof = iA then g is said to be a left inverse of f. If
zhe composite mapping gof: A~C is surjective. there exists a mapping h :B A such that foh = ip thenh is said to
Proof. Let z be an element of C. be a right inverse of f.
Since g is surjective, there is at least one pre-image of z in B.
one such be y. Then y E B and g(y) = z.
Let Definition. Let f : A ^ B be a mapping. f is said to be invertible if
there exists a mapPping g: B A such that =
gof iA and fog = ip.
Since f is surjecive and y EB, there is at least one pre-image of y In this case g is said to be an inverse of f.
in A Let one such be z. Then z E A and f(z) = y.
g i l = g(y) = z. This implies that z has a pre-image r in A under Theorem 1.12.1. If f: A B be an invertible then its
inverse is unique.
mapping
he mapping gef. Since z is arbitrary, gof is surjective.
Proof. Since f: A- B is invertible, there exists : B -A
Note. The cozverse of the theorem is not true. However if gof is a mapping g
sur such that = =
jeccive then g is surjective (while f need not be). gof iA and fog iB.
If possible, let there exist another mapping h : B A such that
Theorem 1.11.5. If f : A - B and g: B -- C be two mappings such hof = iA and foh = ip.

g f : A-C is surjective then g is surjective. ho(fog) = (hof)og, since composition of mappings is associative.
= =
Therefore hoiß iA®g, i.e., h g. This proves that g is
Proof.LEt z be an element of C. Since gef is surjective there is
an unique.
lement z in A such that gof(z) = z. Therefore glf(c)) = z. Note 1. g is said to be the inverse of f and is denoted by f-1. Therefore
f f = iA and fof = ip.
This stows that z has a pre-image f{z) in B under the mapping g.
Since z is arbitrary g is surjective. 2. If f: A =
A be an invertible mapping then f-of fof-i =i,
Note. In order that
where i is the identity mapping on A.
gof nay be surjective it is not necessary that f is
ritive.
For eanple, let f: Z - Z = z E Z
Examples.
be defined by flz) 2z, and R- Z = z
1. Let f by
be defined f(r) ), ¬
:ZZ , z e Z. [a) denotes the greatest defined by g(e) = z +2, ze Z2.
R and g: Z - R be
z.
ie definsi by glz)s =defned
by g-f(z)= E Z.
ezer Then gef:Z-2 z,z Sog: Z Z is defined by fog(z) = f(z +2) = [z + l =z, z ¬ Z.
9j = iz ard ia, threiore, surjective; but f is not surjective.
gof:RR is defned by gof(z) = 9([z]) = [z]+ , zER.
42 HIGHER ALGEBRA
SETS
43
Here fog =
iz, gof # iR.
Then = =
f(y) =
Therefore g is a right inverse of f, but not a left inverse of ff(c)] fof(r) t, since fof =
f. iA
z
This shows that y is a
pre-image of
=
Let h: Z - R be defined under the mapping f . There-
by h(z) c + a ¬Z. fore f is surjective.
Then foh =

as
iz and therefore h is a right inverse of f. well as
Since f is injective
surjective, f is a
Note. There are many bijection.
right inverses of f. Since is not a f bijection, f Second part.
is not invertible.
Since f A
=
iA
» B is a bijection, f1: B A exists and f-of
2. Let R ^ R be defined =
and
f : by f(c) 3r, z e R and g: R » fof =ip. Since f : B B
A is a
defined by g(c) == tER. R be exists andf-o(f-1)7 =
bijection, (f-1)-1: A =
iA and (f)-of ipB.
R Let c E A =
=
gof: R is defined z
and f (z) y.
by gof(r) =g(3«) r, ER
-
fog R is defined = = =
:R by fog(z) a, z ¬R.
f(G) f(u)= ff(T)] = fof(r) z, since flof = iA.
Here gof = = f =
fog iR. Threfore g is the inverse of f. = (f*)-f(u))= (f*1)=of-1(y)
-1)-of r)= ip. y, since
Theorem 1.12.2. A Therefore =
mapping f: A- B is is f(r) (f-1)-1(z) for all z E A.
=
a bijection. invertible if and only if f Consequently, f (f-)=1. This completes the proof.
Proof.Let f: A B be invertible.
"Then there exists a mapping Theorem 1.12.4. Let f : A - B and g: B - C be
g: B A such that gof = iA and =
Then the
both bijective
fog iB. mappings. =
Since iA is injective and gof = iA, f is an mapping gof: A C is invertible and (gof)-
injection. og
Since iB is surjective and fog =
iB, f is a surjection. Proof. Since f: A B
Therefore f is a bijection. and g: B-C are both bijective, the
mapping gof: A C isa bijection, by Theorem 1.11.6. composite
Conversely, let f : A> B be a bijection. Hence
(gof)-l : C-A exists, by Theorem 1.12.2.
Let y E B. Since f is a bijection, y has a Since f: A- B is
unique pre-image in A. a bijection,
=
Define a mapping g: B*A by
g(y) c
(the pre-image of y under
f : B A exists.
= =
Since g: B-C is a bijection, g : C B exists.
S), y E B. Then gof(c) g(y) t, t E A and fog(y) = =

B.
f(z) y, y E Then the composite mapping f-log:C»
z
A exists.
= =
Let E C. Let y be the
Here gof iA, fog ip and therefore f is invertible. pre-image of under the z
z
and be the bijective mapping g
pre-image of y under the bijective
= mapping f.
z.
Note. The theorem gives a clue how to determine To each elemen
f. Then gof(c) Since is gof
y in B, f assigns the pre-image of y under f. invertible, (gof)*"(z) =c.
=
Since f is invertible and
f(z) y, =
f**(y) z.
=
Theorem 1.12.3. Let f Since g is invertible and
: A> B be a bijective mapping. Then the g(y) 2,g(2) = y.
mapping f-: B Therefore =
= =
A is also a bijection
and (S-)-1 =f. Thus
fog-l(z) f-lg()]
=
f-(y) z.
is for all z E C.
Proof. Since f: A B a bijection, f-: B - A exists and
(gof)-1(z) (fog-)(E)
iA and fof-l = ip. f-1of =. Consequently, (gof)-1 =
f-log. This completes the proof.
=
=
Let 1,y2 be two distinct elements in B and
f-(y1) T1, f-(y2) T2. Worked Examples.
Then f(ci) = = =
fif(y1)) fof-"(y1) V1 and f(c2) =
f f ( y 2 ) = fof-'(v2) = y2, since fof-l = ip. 1. Let A, B be
a
both finite sets of n elements and A- BB
Since f is a mapping, y1 #
is injective. Prove that
f is a bijection. mapping f:
y2 t 1 # T2.
Let A =
is
That is, y1 # y2 >f"(v1) #f*(v2) and therefore f-i injective. B. is
Since f
{«1,a2,... , an}. Then f(a1), f(a2),..., all belong to
.
To establish that f-1 is =
surjective, let r E A. Let fla) y. of B. As
injective, f(a1), f(a2),.. ,flan) are allf(an)
distinct
they are n in elements
number, they are all the elements of B.
 SETS 45
44 HIGHER ALGEBRA

1.13. Direct and inverse images.

Let b e B. Then b = f(a,), for some as e A and this shows that f is Let f: A B be a mapping. Let P be a non-empty subset of A.
surjective. Since f is injective as well as surjective, f is a bijection. Then the direct image of P under f is the subset f(P) of B given by
= f(P) = {f(r) : z e P)
2. Let f:R - R be defined by f(=) 37 + 1, a e R. Prove that f is
invertible. Find f .
Let S be a non-empty subset of B. Then the inverse image of S
As f is injective and surjective (worked Ex.1, 1.9), f is a bijection under f is the subset f-(S) of A given by f"(S) = {z : f(c) ¬ S}
and therefore f is invertible.
We emphasize that f need not be a bijection and so the inverse map-
Each y in the co-domain of f has a pre-image unique 3 A need not exist. If, however, f-l exists as the inverse
f-1 R ~ R is defined by f ( ) = 3 , y e R; or equivalently ping f : B as
defined by f-(T) = z , z¬R. mapping» f(S) c a n be viewed as the direct image of S under f*i
well as the inverse image of S under f.
3. Let S =
{z e R : -1 and f: RS be defined by
= t E R. Show that is a Determine f-.
<af < 1}bijection.
f(T) e Examples.
=
f(z) < 1. 1. Let f: R ^R be defined by f(a) = z2, z eR.
Let s > 0. Then f(a) 1#z and 0 <
Let a = 0. Then f(z) = 0. Let P {# eR :0%z $ 2). Then the direct image f(P) = {y e
= 1 and -1 < f(r) < 0.
Let s < 0. Then f(z) = R : 0 y s 4}. The inverse image of the set f(P), ie., f-(f(P) =F
Let 1 , 2 belong to R and flci) = f(m). Then 141:
{rER:-2a2.
This implies T1 and c2 are either both positive, or both negative. Here f(f(P)) # P.
Let 1 0 and T2 > 0. Let S { r ¬ R : -1 S r s 4}. Then f1(S) = {a e R: -2 r S
Then f(T1) = f(r2) 4 = 1+2 7 1 *2 2}
2) and f f ( S ) = {z e R:0 <4}.
Let i < 0 and t2 <0. Here
ana-i(s) # S.
Then f(1) =f(ra) 2. Let f: R>R be defined by f(r) = 3x, r ¬ R.
in R So f is injective.
It follows that *1 # r2 f(ri) # fra). Let P ={r e R :0<r< 2}.
an element in S and let 0 < y<1. Let us examine if r(> 0}
Let y be Then the direct image f(P) = {y e R 0 y 6}. The inverse
in R be a pre-image of y. Then i Y image of the set f(P), i.e., f*(f(P)) = {z E R : 0 r < 2 } .
R. Therefore a pre-image of y.;
or, E R, since y E is Therefore f(f(P)) =P.
y < 0. Let us examine
Let y be an element in S and let -1 < Here f is a bijection and therefore the inverse mapping f - exists.
(<0) in R be a pre-image of y. Then
=
is The direct image of the set f(P) under f l is {r e R:0r< 2}.
or, z ER, since yER. Therefore a pre-imageof y..
Let y= 0 E S. Then a = 0 is a pre-image of y. Therefore the direct image of the set S = {y e R:0<y<6} under
a
f i s same as the inverse image of the set S under f.
It follows that each y in S has pre-image in R. So f is surjective.
is a bijection. Theorem 1.13.1. Let f: A B be a mapping and P, Q be non-empty
Since f is injective as well as surjective, f
in S has a unique pre-image. subsets of A. Then
Since f is a bijection, each y
For y 0, the pre-image is ( a ) P cQ>f(P) c f(9);
For y<0, the pre-image is 14= (b) f(PUQ) = f(P) u f();
For y = 0, the pre-image is 0 = Hp
(c) f(PnQ) c f(P) ns(o);
It follows that for each y in S the pre-image is T
(d) P n ) = f(P)n f ( ) , if f is injective;
= ¬ S.
Hence f1:S- R defined by f(«)is
 46 HIGHER ALGEBRA
SETS

(e) f(P)l C S(P), if f is surjective; Since f

() SP¥ =S(P), is
(f) is surjective, [f(P)I' Cf(P).
if f bijective. Let y E S(P"). Since f is
injective, y has a unique pre-image, say c
Proof. (a) Let y E f(P). Then there is an
and since y E f(P),c ¬ P' and therefore c z P
element r in P such
f(a) = y. that This implies f(r) f(P), i.e., y E [f(P)I. Therefore, f(P') C
EPreQ. Therefore f(r) E f(Q), i.e., f(P).
Consequently, f(P) C S(Q). y E f(). Consequently, [f(P)) = f(P').
(b) PC PUQ. QcPUQ. : A
Theorem 1.13.2. Let f B be an onto
By (a), f(P) c f(Pu) and subsets of B. Then
mapping and S, T be
f () C SPuQ).
Consequently, f(P) u f(Q) c f(PU).. (G) (a) S cT f(S) cf1(T);
Let y E f(PUQ). Then there exists an =
element r e
z PUQ
y=
(b) f(SUT) f\(S) Uf=1(TT):
f(c). ¬ PUQ>c ¬P or z
¬Q
such that
=
¬ or
(c) f-1(SnT) f-(S) nf-l(T);
f(a) f(P) f(a) ¬ f(Q) =

f(a) ¬ f(P) Uf(Q). (d) f-(5') /-(5)]'.

Therefore y ¬ f(PUQ) > r
y¬ Proof. (a) Let E f(S). Then f(r) E S.
U AQ)
Consequently, f(P UQ)C f(P)f(P)Uf(). (ii) f(s) E T. Therefore z ¬
Since S c T, f(c) E S>
From (i) and (i) it follows that =
f-(T).
f(PUQ) f(P) U f(Q). Thus r Ef-(S) > z Ef-(T). Consequently, f-1(S)
(c) PnQCP and Pnac (b) SC SUT,T C SUT.
Cf*1 (T).
Q.
By (a), f(Pn) c f(P) and
f(PnQ) C By (a),f-(S) c f-1(SUT) and f-1(T) E f-1(SUT).
Consequently, f(Pn) c s(P) n f(). f(Q). Consequently, f*i(S) Uf=(T) C f-"(SUT).. (i)
Note. fPne) # f(P) Let E
n f(Q), in general.
For example, let f: R» R be f(T) E T. z
f-(SUT). Then f(T) E SUT and this implies f (r) E S or
That is,
=
e e R.
defined by f(z) **,
Ef-(S) or z E f-(T), i.e., z ¬ f(S)Uf(T).
P= Let Therefore z E f-(SUT) r ¬
{1,2, 3, 4},Q={-1,-2,3,4). f*(S)
= . .
n =
Then Pn Consequently, f-(SUT) C f(S) U Uf-(T).
{3,4}. f(P) f() {1,4,9, 16). f-(T) (ii)
=
=
Here f(Pno) {9, 16) # f(P) From(i) and (ii) it follows that T)
n f(Q). f1(S U f-1(S) Uf-(T)
(d) By (c), f(PnQ) C (c) SnTcS and SnTeT.
f(P)n f(Q) (i)
Let y E f(P) n f(Q). Then ¬
y f(P) and y E
By (a), f-(SnT) cS-1(S) and
Since f is injective, y has a f(). f-(S n T)Cf(T).
So r E
unique pre-image, say c, in P and Q both. Consequently, f-1(SnT) cf-'(S)
PnQ. Let r
nf-1(T).. (G)
Then z
E f-1(S)
This implies f(r) E ns=1(T). E f(S) and z
f(Pne), i.e., y E Ef(S) > f(s) E S; s E f(T) f(s) ET. E f-\(T).
Consequently, f(P)n f() c f(PnQ) f(Pn). . (i) So z E f(S)
=
From (1) and (ii) it follows that nf*'(T) > f(r) ¬ SnTr¬f(SnT).
f(PnQ) f(P) n f(Q). Consequently, f-1(S) nf-1(T)
(e) Let y E [f(P)]I. Then y E f(P). Since f is cf-(SnT).. (i)
y has at least one
From (i) and (ii) it follows that =
P.
pre-image. Since y f(P),y has no pre-image inonto, f-(SnT) f-1(5) nf-(T).
Let s be a (d) Let r ¬ f-1(S"). Then f(r) ¬
pre-image of y. Then z E P and therefore f(r) ¬ z
S', i.e., f(T) g S. This
That is, y E
f(P') f(P). gf-l(s), i.e., ¬
. .
-1(s)¥. Therefore f-(S") C [Í-1(S)}' implies
Consequently, [f(P)Y cf(P). Let e s
a
z ,
[f-(S)}. Then g f-1(S), So f(z) g S, i.e f(r) E (6)S'.
This implies E
. .
f-1(S'). Therefore [f-1(S)|' C f-(S")
(ii)
 SETS 49
HIGHER ALGEBRA

5. Show that the mapping f is surjective but not injective.

From (1) and (i) it follows that [f-1(S)Y = f ( S ) .
N = n ¬ N,
Alternative method. SUS' = B f-1(SUS') = f ( B ) = A and ()f:N defined by f(n) [2],
(i) f: R Z by f() = [e), z e R,
Sns' = d f-1(Sns') = f ( 6 ) = d.
Thercfore f- (S) uf-l(S') = A and f-(S) nf-l(S") = . R R =z* a ER.
(ii) f: defined by f(«) -z,
Consequently, f-1(S)} = f-(S'). 6. Show that the mapping f is a bijection. Determine f*
(i) f: R -R defined by f(r) = 2a +3, z eR,
Worked Example. (it) f: R R defined by f() =*', ze R,
1. Let f: A - B and S c B. Prove that = R: -1 <
(i) f: S - R defined by f(r) TE S, where S = {z E
i) Sf-(S) c s, (ii) ff-(S) = S if f be surjective.
r<1},
(i) Let y E ff-1(S). Then y has a pre-image, say z, in f-i(S). (iv) f: R - { 1 ) -R - {1} defined by f(r) = 2},z e R - {1}.
r Ef-(S) f(c) E S, i.e., yE S.
7. Find gof and fog, if
Thus y e ff-"(S) =y¬S and therefore ff-1(S)cs.
(i) f: Z Z is defined by f(n) = n^, n eZ and g : Z Z is defined by
(ii) Let y eS. Since f is surjective, y E S implies y has a pre-image, say gT)= 2n, n E Z;
, in f'(S).
(ii) f : Z2 2 is defined by f(n) = (-1)", n e Z and g : Z Z is defined
aES-1(S) f(r) E ff-1(S), i.e., y E ff-1 (S). by 9(n) = 2n, n E Z;

Therefore Scff-\(s). : is = r » R is
(ii) f R- R defined by f(r) lel +z, ¬ R and g: R
Using (i) we have ff-1(5) = S if f be surjective. defined by g(«) =
|=| - z, z E R.

8. A =
- B,g: B-C,h: B C be mappings such that gof hof
and Prove that
Letisf:surjective.
g = h.
Exercises 3
Hint. B and = = = =
f(z) y,z E A. gof hof gof(=) hof(r) 9(y) h(y).
Let y E
1. Let S be Prove that a mapping f:S S is injective 9. A
a non-empty finite set. Let g: B,h: A B,f: B C be mappings such that
if and only if it is surjective. and f is injective. Prove that g = h.
fog = foh
2. Give an infinite set S and a mapping f: S - S such that [Hint. Let r ¬ A. fog = foh fog(z) = foh(r) g(æ) = h(r), since f is
example of an
(1) f is injective but not surjective; injective.]
(ii) f is surjective but not injective. 10. Let ^
f: A B and g: B-C be
two mappings. Prove that
3. Show that the mapping f is neither injective nor surjective. (i) if gof is injective and f is surjective then g is injective;
f() z e R, (ii) if gof is surjective and g is injective then f is surjective.
) f : RR deined by =sin a,
by f(r) z ¬ R, Hint. (i) Let y ¬ B and 9(u) = z,z E C. Since gof is surjective, gof(r) = z for
(ii) S: R - R defined =Iz|,
some E A. =

defined by f(r) +, z e R, g(f()) g(u) f(:) = y, since g is injective.]

(is) f: R R =r
(iv) f : R R dofined by f(z) = E R 11. Let f : A^ B be a mapping. A relation p is defined on A by "t p y if
and only if =
f(r) f(v), z,y e A". Show that p is a n equivalence
relation.
4. Show that the mapping f is injective but not surjective 12. Let f: A - Band P C A. Prove that
Z definod by f(r) z E Z,
0) S:Z =3r, G) P c f f(P), (ü) P = fr'f(P) if f be injective.
- =
(ii) f : N N defined by f(n) n+1, n eN,
Give an example where
(ii) f: Z - Q defined by flr) = 2, r e Z.. P ff(P).

HA-4
 50 HIGHER ALGEBRA SETS
51

13. Let A be a set of n elements and B be a set of m a1

elements. Show that a2 an
then fg= and
i) the total number of from A to B is flg(as)l flg(a2)]
mappings m"; flo(an)
(ii) if n m, the total number a1 a2 an
of injective mappings from A to B is 9f= glfla)) slf(as)]9fla,)]))
m
14. Find the total number of
surjections from A onto B, where A is a set of 4 Since the
elements and B is a set of 3 elements. composition of mappings is not in
commutative, fg # gf,
general.
[Hint. Let the total number of surjections from a set of n elements to a set of Since the
mn composition of mappings is associative,
elements be sn(m) (n 2 m). The total number of
mappings from A to B is 34. multiplication of
= -
permutations is associative. Thus if f,g, h be three
=
Therefore s4(3) 34 3e 34(1) 3ea S4 (2),
s4 (2) 24 2c $4 (1), s4(1) 1.] = f(gh) = (fo)h. permutations on S,

Inverse of a a
1.14. Permutation. permutation. Let f be permutation on S. Since f is
a bijective mapping, it admits of a unique inverse f-l : S> S and f
Let S be a non-empty is also bijective. and =
finite set. A bijective mapping f:S* S is Sof- is a permutation on S i. If
ff-1= f-1f
said to be a permutation on S. by f,a aj, then by f*,a;ai.
Let S =

{a1,a2, .,an}. Then the number of bijections from S a1 an

onto S is n!. Let one such
bijection be f that maps a; to f(ai). This then
f(a2) J(an)
Therefore if f=f(a1)
permutation f is denoted by the symbol
- = ((a1) f(a) fla,)
a2 a2 an
fla1) f(c2). f(an)
The identity mapping is is also a bijective mappping. is is said Example.
be the to
=
=
identity permutation and it is denoted by i. 1. Let S and
{1,2,3,4} f G 342).
i=d a1 a2 an

Let S 1,2,3,4) and a permutation f be defined by f(1) =

g-(2 1),h-(4 13
2, f(2)=4, f(3) = 1, f(4) = 3. Then f is expressed as Then =
fg since =
fo(1) f(2) 3 , f9(2) =

f(3)= 4 etc.
(33)..
or
(2
in many
or as23 3).oras (3 33). since gh(1) =
g(4) = 1 etc.
other ways by interchanging the columns. But the standard gh-(i2):
form for f
the natural order.
is2 413where the elements in the top row arein ro)-(1 3 )(G3i )-(1 i3 :)
Multiplication of permutations. Let f: S S, g: S » : be two uh-( )(G3}i)-(1i3i)
permutations on S. Since
Therefore =
range f= dom g, the composite mapping
gof S S is defined. Since f and g are both bijective, gof is f(gh) (fg)h.
bijective. Since = =
ff=i,ff(1) =
Therefore gof is a permutation on S. Similarly fog is a
on S.
permutation 1,f-f(2) 2,f(3) 3,/-1f(4) =4.
=
The products 9f ( or 9.f) and fg ( or f.9) are defined
by the Therefore = =
f(1) 1, f(3) 3, =
composites gof and fog respectively. If 2,f(4) f\(2) 4 and
a2 an a2
f-(a) and g-la1) s(oa) 9(an) -(12 )-(i 2)
 HIGHER ALGEBRA 53
52 SETS

rows of f and =
Note that f- is obtained just by interchanging the
The index laws (ii) (fm)" fm
(4) S.fn = f*n and
then by expressing it in the standard form. where m, n are integers.
hold for all m, n, in general.
= does not hold, since fg # gf,
But (f.g)" f".g"
1.15. Cycle. a or
on a finite set is either cycle
Let S = {a1, a2,.. . , an}. A permutation f: S S is said to be a Theorem 1.15.1. Every permutation
if there are T elements în of disjoint cycles.
of r, or an r-cycle aij, aia1 .,ai, it c a n be expressed as a product
cycle length us
= = a permutation o n S. Let
=
S such that f(a) = iganflai,) =
aig f(ai,) ai, f(a,) a Proof. Let S
=
{a1, a2,... ,an}. Let f be Al these can not be . . .
distinct,
and f(a) aj.j # i1,t2...,ir consider the elements a1, f(a1), f°(a1),
The cycle is denoted by or and S is a finite set.
di)s
,i order. since all of them belong to S
the elements Then =
(a4,,i29.*, Gi,), by (aia,Gisp*cyclic
appear in a fixed
or in any other form provided such that f"(a1) a1.
Let r be the least positive integer
are said to be the elements of the cycle. because if
ai,Qi2 , ai, . . .
are r distinct elements of S,
=
Two cycles f and g on the same set S are said to be disjoint if they a1, f(a1), S*(a1), ,fT="(a1) such that 0 < p <g < r then f9-P (a1)
fP(a1) = f°(a1) for some p,q
have no common elements. contradicts that r is the least positive integer
°(a1) =a1 holds and this
is commutative. satisfying f"(a1) = G1.
Multiplication of two disjoint cycles
= . . -

Let us consider the r-cycle p1 (a1, f(a1), f*(a1), ,fT-(a1).

Ifr = n then f = P1 and the theorem is proved.
Example. . Such that an
3 If r < , let an be the first element among a2,. , an
1. does not belong to the cycle p1.
Let S (1,2,3,4} and f=i 34 2
. . .
2 3 Let us considerthe elements am, f(am), f (an), Neither of these
=
9-(i 2).h-G 214 belong to Pi, because if f(a1) f (am) for integers i,j, then some

Here f is a 3-cycle. f = (2,3, 4). f can also be expressed as (3,4,2) =

or as (4, 2,3). g is a 2-cycle. g = (2,4). h is a 2-cycle. h = (1,3).

fi-i(a) am, a contradiction.
a =
arrive we at cycle pa
Arguing as before, (am, f(am), S(am),...
and h are not disjoint.
9 and h are disjoint cycles. f f- (an)) of length s.
12 3 4 1 2 3 4 Ifr+8 = n, then f is the product of disjoint cycles pi and p2.
2 3 4
gh =
4 3 4 Ifr+s < n, let ak be the first element among a2, a3, a4. , an Which
i 2 3 2 1 does not belong to pi or p2.
4
3 413 2
1 2
hg 3 4 4 3 2 We proceed in a similar manner. Since S is a finite set, this process
= It can be verified that
2. fg # gf, fh # hf. terminatos after a finite number of steps and we arrive at the decompo-
Clearly, gh hg.
sition of f as the product p1-pa. .pe of disjoint cycles.
Integral powers. Note 1. Since multiplication of disjoint cycles is commutative, the order
Let f be a permutation on a set S. Then f.f is also a permutation
factors ...
on S. Since multiplication is associative, ( f f ) . f = f.(ff). We can of the Pe in which they appear in the docomposition of
Pi,P2
i s not unique.
eliminate the paranthesis and denote each by f.ff.
= factors for all n E N.
Note 2. The identity permutation i = 2 is the prod-
We define f" f.f.f.... f(n ) a1 a2 an
Since f is a permutation on S, S is also a permutation on S. üct of n disjoint cycles (a1),(a2),... ,(an), each of length 1.
permutations are on S.
sf1,f-1f-1f-1,.. Order of a permutation.
We define f-7 = f - f f . . . f ( n factors ) for all n e N and we
Let f be a permutation on a finite set S. The order of f is the least
define f i . Thus f" is defined for all n, where n is an integer.
positive integer n such that f = i, the identity permutation on S.
 54 HIGHER ALGEBRA SETS
55

an
Theorem 1.15.2. The order of is r. (ar, a,)(an, an) (am, an). So there are many ways in which a cycle, and
r-cycle
a can be as a
Proof. = therefore permutation, expressed
Let S = {a1,a2,... and p
,an} be an
r-cycle
product of transpositions.
on S. =
(a1,02,.. . ,ar) The number of factors in the of
= =
Then p(a1) = decompostion
=
a permutation into trans-
a2, p°(a1) p(a2) a3,... ,P(a1) P(ar) 1.
=
Similarly, p"(a2) a2,p"(a3)= a3, . positions thus loses its uniqueness.
=
,p"(ar)=ar
=
But an important property of a permutation is that the number of
=
.
=r+1,.. ,n. factors in its decomposition into transpositions is either always odd or
Also p(as) as for s r+1,... ,n and so p" (as) a for s
It follows that = =
p"(ak) ak for k always even.
1,2,... ,n and therefore p is thhe
identity permutation.
r is the least Definition. A permutation is said to be even if it can be expressed as
= i
positive integer such that p" =i. Because if p
for some the product of an even number of transpositions, and odd if it can be
so.
positive integer m <rthen p"(a1) must be a1 which is not
Therefore the order of p is r. expressed as the product of an odd number of transpositions.
-
Theorem 1.15.3. The order of a
Since an r-cycle can be expressed as the product of r 1 transposi-
of the
permutation on a finite set is the l.c.m. tions, an r-cycle is an odd or even permutation according as r is even or
lengths of its disjoint cycles.
odd.
Proof. Let f be a permutation on the set S {a1, =
a2,.. .,an} and let The identity permutation i can be as the of two
f be expressed as the product of
disjoint cycles P1,P2, ,Pm of lengths expressed product
. transpositions (ar, as).(2s, ar). So i is even.
r1,T2. , Tn respectively. "Then f
=PiP2 Pm
Thus f" = The product of two even permutations is even; the product of two
P? p...p for each positive integer n, since multiplica
tion of is
disjoint cycles commutative. odd permutations is even; the product of an even
permuatation and an
.
odd permutation is odd.
P= i, P2= i,.. ,P = i, i being the identity permutation.
Let s be a The inverse of an
common multiple of odd permutation is odd and the inverse of an even
r1, T2,. ., Tn. Then f3 permutation is even.
Pi P Pm=:
Obviously, the least positive = i
integer r for which f" holds, must Theorem 1.15.5. The number of even
be the least value of s. In other words, r is the l.c.m. of . . .
permutations on afinite set (con-
r1,T2, , Tn taining at least two
elements) is equal to the number of odd permutations
Therefore the order of f is the l.c.m. of .
on.it.
ri, r2,.. , Tm.
=
Transposition. A 2-cycle is called a transposition. Proof. Let S {a1, a2,... , an), n 2 2. Let A be the set of all even
permutations and B be that of all odd
A 1-cycle is the identity and it can be
expressed as the of the
permutations. Then both A
product and B are non-empty, because the
identity permutation i belongs to A
transpoitions (ap, G,) and (a7, ag). and the transposition (a1, a2)
belongs to B. Let t be the transposition
A 2-cycle is itself a
transposition. (a1,a2). Let us define a mapping d : A - B by o(f) = tf, f E A. Since
A fEA,fis even and therefore tf is odd and tf e B.
3-cycle (a1, a2, a3) can be expressed as the product (a1, a3) (a1, a2).
Let A. =
.
An r-cycle fi, fa ¬ tfi tf2 t(tf1) =t1(t/2)
(a1, a2,. ,a) can be expressed as the product
( 1 , ) ( a 1 ,ar-1)...(a1, 3)(a1, az). Ir - 1 transpositions. ( t t ) f 1 = (t-lt)f2 > fi = f2.
Therefore fi # fa th # tf. This proves that
Theorem 1.15.4. Every permutation on a finite set (containing at least
o is injective.
Let g be an
two can
element of B. Then tg being even, belongs to A and
elements) be expressed as a product of transpositions.
= =
This follows from the Theorem 1.15.1. o(tg) t (t9) t"g = 9, since t = i. This shows that tg is a pre-image
of g under the mapping o and therefore
o is surjective.
An identity permutation can be expressed as a product of tranas Therefore d is a bijection and since A and B are finite sets, they have
= = the same number of elements.
positions in many ways. For example, i (r, as)(ar, a,) (4,, a,) This completes the proof.
 HIGHER ALGEBRA SETS 57
56

Exercises 4
Worked Examples.
on the set {1, 2,3). Find their respective
1. Describe all the permutations 1. Find fg. gf. f-i and g* where
orders. 12 3 4 5 6
on the set. They are
There are six permutations -(2 4 s56 dg-(3 45 S)
of Ex.1 are odd or even.
2. Examine whether the permutations f,g
(123)- (G 3i)-a.2.. ( i i)-a.32, 3.
and 4,
Find the images of the elements 3
2 3 4 5
(33)-(2..( )-a.». (} ? 3)-a.» be an odd permutation,

The even permutations are i, (1,2,3), (1, 3, 2).

(4 23 4 5 6
be an even permutation.
are (2,3), (1,3), (1,2). 4 3
The odd permutations
The order is 1, since i =i. 4. Find the order of the permutations
of i
is 3, since each is a
The order of each of (1,2,3) and 1,3, 2) cycle of ).c (12345 6 7 )
length 3.
a 5. Express a as a cycle or as a product of disjoint cycles.
The order of each of (2,3), (1,3) and (1,2) is 2, since each is cycle
a =
a= a (ii) (1234) (456).
of length 2. (i) (123) (34), (i) =(135)(124),
set of all permutations on
f on the set {1,2,3,4, 5,6} is 6. Find the number of distinct 3-cycles in Ss (the
2. Determine whether the permutation
the set {1, 2, 3, 4, 5})
odd or even, where f = (3 3 6 1 4 7. Let a1, a2,.., ani b1,b2,...,bn be two permutations of 1,2,, n and n is
5
3 the of two disjoint cycles odd. Prove that (a1 -b1)*(a2-b:).(an -b.)° is an even number.
can be expressed as product
(1,2,3,5)(4, 6). 8. Let Sa be the set of all permutations on the set {1,2, 3, 4}. Prove that there
as
can be expressed (1,5)(1, 3) (1, 2).
The cycle (1,2,3,5) exists at least one element f in S4 such that f cannot be expressed as g for
of fou
=
a s f is the product
Therefore f (1,5)(1,3)(1, 2)(4, 6) and any g E S4.
transpositions, f is an even permutation. Hint. The mapping F: Sa S4 defined by F(a) = a4 for a E S4 is not injective
and since Ss is a finite set, F is not surjective.]
on the set {1,2,3,4}.
even permutations
3. Describe all
are
on the set. They
There are 12 even permutations
1.16. Equipotent sets.
G34)-(2 i s)-a,2)(3,4), Let A and B be two non-empty sets. A is said to be equipotent with
Bif there exists a bijective mapping f from A to B. We write A ~ B. If
(G2i4)- 1.(2,0, ( 32i) - a.902,. A is equipotent with B, then B is equipotent with A, since the existence
of a bijective mapping f: A B implies the existence of the inverse
i ) -a3.0. (132 3)=e4, mapping f-1: B A which is also bijective. Thus two sets A and B
( are equipotent with each other.
(3 i) -a39. ( 4 ii)-(1.4.3), Let us examine the relation of equipotence on the set S of all non
empty subsets of a universal set U.
(4 23 i) =a.2.0. ( i3 )-1,4.2), t1() Let AES. A ~ A because the identity mapping i : A A isa
bijective mapping. So the relation is reflexive.
(G 3 ) -a.2.».( i 2 )-a32.
 HIGHER ALGEBRA
SETS
59
(ii) Let A, B E S and A B. Then there exists a bijective
mapping A set wlhich is cither finite or cnumerable is said to be
A B . The inverse mapping f - : B ^ A is also a bijection. countable (or
~
at most enumerable).
Therefore A B ~
B A, showing that the relation is symmetric. Enumerable sets are also called
Let E S
countubly infinite sets.
(iii) A,B,C
and A~ B,B C . Then there exists a bijective When a set is finite and contains
n elements, its
mapping f : A B and another bijective mapping
g:B C. The elements can be
described as a1,a2,.. ,an the elements being indexed by the finite set
composite mapping gof : A C is also a bijective mapping.
Therefore A~ B and B {1,2,, n}.
C A~C, showing that the relation is
transitive. When a set A is enumerable, there is a
bijective mapping f : N A
and f assigns to each n E N an element in A. Thus the ele-
Thus the equipotence of subsets is an equivalence relation on S and ments of A can be described as
f(n)
so S is or as
partitioned into classes of equipotent sets. .
f(1), f(2), f(3),.. .,f(n),... a1,
a2, a3,. , Gny.. showing that the elements are
Two indexed by the set N.
equipotent sets are
said to have the same potency, or the
cardinal number. The cardinal number of a set A is denoted by card
same
=
(A). Examples.
Therefore card (A) card (B) if and only if A and B are equipotent. 1. The set N is enumerable, because the
N- N defined
Two non-empty finite sets A and B
mapping f: by
belong to the same class if f(n) = n, n E N is a bijection.
only if they have the same number of elements.
and
2. The set S =
{2,4,6,8, 10,...} is enumerable, because the mapping
The cardinal number assigned to the equipotence class of finite sets f: N Sdefined by f(7) = 2n, n e N is a bijection.
each with n elements is n.
3. The set Z is enumerable, because the
The cardinal number assigned to the nullset is 0. mapping f: N- Z defined by
if n be even
The cardinal number of an infinite set is said to be a transfinite car- i f n be odd, is a bijection.
dinal number.
The cardinal number of the set N is denoted by d. Theorem 1.17.1. Every infinite set contains an enumerable subset.
The set R has a different cardinal number and it is denoted by c. Proof. Let A be an infinite set. Let us define
a mapping h: N A
=
There are many other cardinal numbers. by h(1) one element of A, say a
(2)= one element of A-{a1}, say a2
=
-

Examples. h(3) one element of A [{aa}u{a2}|, say a3

.
1. The set S = { 1 . . . } is equipotent with set N, because the map
= -
h(n) one element of A
ping f: N Sdefined by f(n) =,n E N is a bijection. [{a1}U{a2}U...U{an-1}], say an
2. The set A = {z E R : 0 r 1} is equipotent with the set
Since A is -

B E R: 03r < because infinite,A

{r 3}, the mapping f : A B defined by {{a1} U {az} U.. U {an-1}] is infinite for all
f(a) = 3z, a E A is a bijection. n1. So h(n) is well defined for all n eN.
Also for
3. The set A = {z e R 0 z 1 } is equipotent with the sei p, g ¬ N,p #q h(p)# hlg).
Therefore h is an
B {z E R : a r < because the mapping f: A B defined by injective mapping of N into A.
f(r) = a + ( b - a)x, c ¬ A is a bijection. Let B be the subset
B is a {h(1), h(2), ..., h(n), ...} then BCA and h: N
bijection. Therefore B is enumerable.
This completes the proof.
1.17. Enumerable sets.
A set A is said to be enumerable Note. That at each step we can choose one
(or denumerable) if A is equipotent U... U
element of the set A-[{a1}u
with the set N. a2} {an-1}}, for each n 2 2, is
as the atiom of choice. guaranteed by an axiom, known
 SETS 61
HIGHER ALGEBRA
60

Theorem 1.17.4. The union of two enumerable sets is enumerable.

an enumerable set is enumerable.
Theorem 1.17.2. An infinite subset of =
= - .
Let A
Proof. {b1,b2,
{aj1, a2,..,an,. .}, B .} be two e n u -
. ,bn
A be an enumerable set and B be an infinite subset of A.
Proof. Let merable sets.
Since A is enumerable, its elements can be described as a1,a 2 , a n s
Case 1. Let An B = d.
B contains infinite number of a's and the suffixes of the elements of
B form an infinite subset P of the set of all natural numbers. P, being Let us define a mapping f : N AUB by
S72) a(n+1)/2, if n be odd
a subset of N contains a least element, say 41
Let B1 = B - {a,u,}. Then B1 is an infinite set and the suffixes of =bn/2, if n be e v e n .
Then f is a bijection and so AUB is enumerable.
the elements of B1 form an infinite subset Pi of N. Pi contains a least
Case 2. Let An B # p.
element, say 2 Let A1 = A, B1 = B - A. Then A1 U Bi = AUB and A1nB1 = ¢.
-

Let B2 = B {u1, pa Following the same argument with

Now Bi being a subset of B, is either finite o r enumerable.
B2, B3,.. . we get the elements aus aus If B be enumerable, A1 UB) is enumerable, by case 1. If B1 is finite,
Let us define a mapping f:N>B by A UB1 is enumerable by Theorem 1.17.3.
S(1) = a 1 f(2) = a@u2, S{n)= apn1 Therefore AUB is enumerable.
Let p,g E N and let p < q. Then E This completes the proof.
f(a) ¬ Ba-1 = B-{a1 ugs.., apg-î}:
f(p) { Gpa..,auq-tand
Theorem 1.17.5. The union of an enumerable number of enumerable
Therefore f(a) # f (p). So f is injective.
sets is enumerable.
Let us take a n element a, E B. Since r is a positive integer, there are Proof. Let A1, A2,..., Ans... be an enumerable family of enumerable
at mostr- 1 elements in B whose suffixes are less than r. So a, is one
sets.
of f(1), f(2),.. . ,f(r - 1), f(r) and hence f is surjective.
Therefore f is a bijection and B is enumerable. Let A1 a11,a12.. , a1n .
A2 a21, a22 , a2n,. -
This completes the proof.
or
Corollary 1. A subset of an enumerable set is either finite enumer- An= an1,an2 ann,.
**
able.
Case 1. Let A, nA = p for all i,j.
a countable set is countable.
Corollary 2. A subset of
a finite set and an enumerable set is Let B=U A. Each element of B is of the type amrn where m, n e N.
Theorem 1.17.3. The union of
enumerable Let us define a mapPping f: B N by f(amn) = 27.3", amn E B. f
enumerable set and B be an finite set containingm is injective, because for two elements amn, Gpq E B,
an
Proof. Let A be amn apa (m,n) # (p, 9) > 2m3" 4 2P3",.
elements Let A =
b1,b2,.,bm. {ai, a2, G3,.. S(B) is a proper subset of N, because there are elements in N (for
Case 1. AnB =
Let us define a mapping f : N AUB by example 5, 7, 13,..) which have no pre-image in B. Let f(B) =Ni.
S(i) = b;, i = 1,2,.. . ,m; f(m+i) = a4, i = 1,2, 3,... Then f: B - Ni is a bijection.
Then f is a bijective mapping. Therefore AUB is enumerable. Since Ni is an infinite subset of the set N, it is enumerable. Thus B
is equipotent with an enumerable set N1 and so B is enumerable.
Case 2. AnB # .
Let A = A - B. Then Aj U B = AUB and A1 nB = d. Case 2. Let the sets {A:} be not pairwise disjoint.
Now A1 is an infinite subset of A and therefore A1 is enumerable. Let us define sets {B} such that
By case 1, A1 UB is enumerable and therefore A UB is enumerable. Bi = A1, B2 = A2 - A1, Bs As - (A1 UA2),...
This completes the proof.
 62 HIGHER ALGEBRRA
SETS
63
= -

B Ak (41 C
UA2U...UAx-1),... Then Bk Ax for all .
Theorem 1.17.6. The open interval (0, 1) =
i=l i=1
B= j A, and B,nB, =¢ for all i,j. enumerable. { ER:0<r<1} is not
Since Bk C Ak, is finite or Proof. If
B enumerable. By case 1, U B, is enu- possible, let the set be
set can be described as
enumerable. Then the elements of the
merable. So U A is enumerable. This completes the
a, a2,,n..
proof.
a
Every irrational number in the set has
an infinite decimal. A rational unique
representation as
Worked Examples. nunber in the set has either a
deciinal representation or a finite decinal recurring
1. Prove that the set of
all representation.
positive rational numbers is enumerable. A fiuite
The set can a
decimal can be expressed
be described as the union of the sets uniquely as
{Ak}1, where curring) decimal by using 9's. For example, .125 non-terminating (re-
=

representation by using 0's (e.g., .125 .12499..., the other

=.1250000...) is being ignored.
.
Thus every real number
a
in the set can have
decimal representation. unique non-terminating
Am -
.
( m )
.
Let a =

Since each Ak is 0.a11@12a13

enumerable, U, Ak is enumerable. a2= 0.a21@22@23
k=l

2. Prove that the set of all

rational numbers is enumerable. = 0.an1@n20n3.
Let P be the set of all positive
..
rational numbers and P be the set of
all negative rational numbers. The
sets P and P are
equipotent because ,where0 aij 9.
=
there exists a bijective
mapping f: P - P ' given by f(z) -z, a ¬P. By our
So P is enumerable, since P is so. assumption, every real number in the
have a place in the open interval must
.
(0,1)
The set
Q of all rational numbers
enumeration {a1,
a2,.,Gn.
is the union of PUP and the Let us
.}
singleton set {0}. So Q is enumerable. b =
consider the real number
0.b1b2d3..., where
be=1 if akk #1
3. Prove that the set N x N is enumerable. =
=
2 if agk 1.
Let us
=
consider the
for (n, n) E N x N.
mappingf:NxN- N defined by f(m, n) 2"3" Then 0 <b<1 and b is
different from each ak, because b
ap inthe kth decimal differs from
S is injective, because f(m, n) =
=
place. Thus b does not belong to the
f(p, g) >2m3" 2P3 tiona1 a2. ,an. .. So our assumption is wrong and the enumera-
2 - P =39- enumerable. set is not
This the
m - p = 0,g-n =0 completes proof.
m P, T =

=
Theorem 1.17.7. The set R is not
(m, n) (p, g). enumerable.
The image set B of f is a
proper subset of the set N, because there are Proof. If not, let R be
is an
infinite subset
enumerable. The set {r e R : 0
< z < 1}
. .
elements in the set N (for example, 5,7,10, of R and so this must be enumerable.
.) which have no pre-image This isa
in Nx N. Therefore f: Nx N- B is a contradiction by Theorem 1.17.6 and this
bijection. proves the theorem.
Since B is an infinite subset of the set N, B is an
enumerable set. As Note. The set R has a cardinal
the set N xNis equipotent with number different from that of the set N.
an enumerable set, it is enumerable. The cardinal number of R
is denoted by c.
 HIGHER ALGEBRA
64

Exercises 5

1. Prove that the following sets are enumerable.

(i) The set of all integral multiples of 5. 2. Groups
(ii) The set of all integral powers of 2.
m ¬ Z,n e Z}.
(i) The set of all ordered pairs {(m,n):
of the form {r ¬ R 2.1. Binary composition.
2. Prove that the set of all closed and bounded intervals
a r 6 } with rational end points a, b is enumerable. Let A be a non-empty set. A binary composition (or a binary opera
3. Prove that the set of all circles in a plate having rational radii and centres tion) on A is a mapping f : A x A-A. Therefore a binary composition
with rational co-ordinates is enumerable. of elements of A.
f assigns a definite element of A to each ordered pair
is denoted the o. For a pair of el-
This mapping f generally by symbol
4. Prove that the Cartesian product of two enumerable sets is enumerable. o is
ements a, b in A, the image of (a, b) under the binary composition
(Hint. Let A = {z1,z2 , Tn.}, B = {y1.2, Vn..- aob. The of the element (b, a) is obviously boa.
deñoted by image
. .
= .}. Then A1 is enumerable.
Let Ai {(T1, V1), (71, y2), (T1, v3), The symbols like *, +, , D,O are also used to denote a binary com-
AxB= U Ak
k=1 position.
5. Let S be an enumerable subset andT be an infinite non-enumerable subset
of R. Prove that
Examples.
(i) SUT is non-enumerable, (ii) SnT is at most enumerable, 1. On the set Z, let o stand for the binary composition 'addition'. Then
4o- 4 = 0.
(ii) S - T is at most enumerable, (iv) T - S is non-enumerable. 203 5,
6. Prove that the sets A and B are equipotent. 2. On the set Z, let o stand for the binary composition 'multiplicationn'.
Then 203 = 6, 300 = 0.
(i) A = {r e R:0r<1}, B = {z E R:0Sr < 1}.
ii) A = {= e R:0r<1}. B = {z e R:a<r b}. 3. On the set Z, let o stand for the binary composition 'subtraction'.
Then 302= 1, 103 =-2.
(ii) A = {z ¬R:0%r$ 1}), B={z E R:0<r <1)
4. Let a binary composition o be defined on the set Z aob = a +
B by 2b,
(iv) A ={z e R:a21}, {r ¬R:r>1]}. a,be Z. Then 203 = 8, 300 = 3.
Hint. () Let f: A B be defined by f ( ) = H n=1,2,3,.. and f(o) =
5. Let a binary composition * be defined on the set =
if z E [0,1] and z # n=1,2,3,. Q by a *b ab,
a,beQ. Then 2 * 5 = 5, 3*8 12.
(iv) Let f : A - B be defined by f(æ) = z+1, if s be an integer 2 1 and
fla) = z, if z be not an integer, r E A. be defined on a if
non-empty set A
case
A binary composition o is said to the
aob E A for all a, b in A. In set A is said to
this be closed under
(or closed with respect to) the binary composition o.
For 'addition', since a e E
example, theset N is closed under N,b
Na+bE N. But the set N is not closed under 'subtraction', because
6 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.
=
E A.
ois said to be associative if ao(boc) (aob)oc for all a, b, c

HA-5
 HIGHER ALGEBRA GROUPS
66 67

Examples continued). .(b.e) a.bc a(bc) =(ab)e

=
6. Addition on the set R is both commutative and associative. Multipli. ab.c (.b). for all a, b, E Z
cation on the set R is both commutative and associative, but subtraction
on the set R is neither commutative nor associative. Composition table.

7. When A is a non-empty finite set, a binary composition o on the set

Let S n be a non-empty set and P(S) be the power set of S. Then u A can be defined by a table, called the composition table. If the number
(union), (intersection) and A (symmetric difference) are binary com
positions on P(5) and each of these is commutative and associative on of elements in A be n, the table has n rows and n columns, one for each
element of the set. The elernents of the set are listed on the topmost row
P(S). and the leftmost column in the same order.
8.
on
Let M2(R) be the sct of all 2x 2 real matriccs. Let o stand for the table in the ith row
multiplication of matrices. Then o is associative but not commutative. IfA=a1,G2 The, a) then a-oa, appears
n entries of the table are all elements of A. since
and jth column.
A is closed under o.
9. Let n
be a positive integer and let us consider the p-equivalence
classes,
of the relation p on Z defined by "a pb if and only if a - b is divisible by =
symmetric about the principal diagonal (i.e., if a,oa,
o
Ifthethentableisbecommutative.
n" are n ,
a,oa)
for a.b e Z. There
are also
classes cl(0),cd(1),ci(2),.. ct(n -1). These
classes of residues of integers modulo n. We use the For example, the table for the of 'addition modulo
called the binary composition
notation to denote the class cl(a). Let Z, be the set of residue classes 3 onthe set Z3 is
{0, 1.2,...n -1}. +|0
We detine a binary_composition +, called addition modulo n, on the
set +b =
Zn by a +b.
In order that this definitionmay be valid we must chcck that it is
well defined. i.e., it is independent of the choice of representatives
of Here the composition is commutative.
the equivalence classes. Thercfore we have to show that if a, a'.b, b' are
integers such that cl{«') = cl(a) and cl(8') = cl(b) then a + b= a +b.
= a a - a' = kn for some integer k, Exercises 6
b = b = b- b = pn for some integer p.
o
1. Examine whether the defined on the set is
Therefore
composition (i) commutative,
(a t(= k + p) is an integer. (ii) associative.
+b)-(a' +b') =tn, where
Consequently, a +b = a +U.
(a)o on Z defined by aob = a +b+ 1,a,be Z;
This proves that 'addition modulo n' is a well defined binary compo- (b) o on =

sition on the set Z Q defined by aob ab + 1, a, b E Q

(c) o on R defined by aob = a +
In like manner, we define a binary composition, called multiplication 2b, a,b¬ R;
=
(d) on R defined by aob = |ab|,a, b ¬ R:
inodulo n, on the set Z, by a.b ab and we can prove similarly that.
on -
it is a well defined conposition on the set Z i.e., if cl(a') = cl(a) and (e)o Zx Z defined by =
(a, b)o(c, d) (a c,b- d),
(a, b),(c, d) e Zx Z
cl{b) = cl(b) then ab = a'b. (f) on -

Ma (R) defined by AoB= (AB BA), A, BE Ma(R).

Both these compositions are commutative as well as associative, be 2. Let o be an associ&.tive binary composition on a set S. Let T be a subset of
cause T = =
S defined by {a e S: a o r * oa for all r ¬ S}). Prove that T is closed
ä +b = a +b = b t a = b + ä for all ä, be Z,: under o,.
.b= ab = ba = b.a for all ä,be Z,,; 3. Let S be a
set of two elements. How many different
binary compositions
and a + (b + ë) = ä t b +C = a t (b+ c) = (a + b) +c cn on S7
be defiued How many ditfereut conmutative binary compositions
a +b + = (ä+ 6) + c for all . b, E Z: cn be defined on S?
6S HIGHER ALGEBRA GROUPS 69

2.2. Groupoid. = = = a a in G.
eoa aoe =a and foa aof for all
Then
Let G a non-empty set on which a binary composition o is
be defined We have eof = f, by the property e; of
Some is G by the composition o and and also eof = e, by the property of f.
algebraic strucure imposed on (G,
becomes an algebraic system. The algcbraic system (G, o) is said to be a Therefore e = f.
9Toupoid. The groupoid (G.o) is comprised of two entities, the set G and
the composition o on G. The same set G may form different groupoids Theorem 2.2.2. If a groupoid (G, o) contains a left identity as well as
with respect to different binary compositions on it. a right identity then they are equal and the equal element is the identity
element in the groupoid.
Examples Proof. Let e be a left identity and f be a right identity in (G, o).
1. (Z, +) and (Z,-) are both groupoids. They are different Then eoa = a for all a in G, aof = a for all a in G.
systems although the underlying set is Z in each case.
algebraic
We have = the
eof f, by property of e;
2. are
and also eof = e, by the property of f.
(Q.+).(R. +). (Q..), (R..) groupoids.
Therefore e = f. This proves that e is an identity element in the
3.
and e is the
(Z+). (Z ) are groupoids. only identity element in the
groupoid by the Theorem 2.3.1,
4. (M2(R), +) is a groupoid, where + is the matrix addition.
(M2(R). groupoid.
.
is a groupoid, where is the matrix multiplication.
Definition.
Definitions. Let (G, o) be a groupoid containing the identity element e. An element
A groupoid (G, o) is said to be a commutative
groupoid if the binary a in G is said to be invertible if there exists an element a' in G such that
composition o is commutative. d'oa = aoa = e. a' is said to be an inverse of a in the groupoid.
An element e in G is said to be an element in the
identity groupoid An element a in G is said to be left invertible if there exists an element
(G,) if aoe = eoa = a for all a in G.
bin G such that boa = e. b is said to be a left inverse of a in the groupoid.
Example 4. (Z, +) is a commutative groupoid but (Z, -) is not a com An a is said to be right invertible if there exists an
mutative groupoid. 0 is an identity element in (Z,+). There is no
element in G
elementc in G such that aoc = e. c is said to be a right inverse ofa in
element in (Z, -).
identity the groupoid.
Definitions. Examples (continued).
An element e in G is said to be a right identity in the 7. 1 is the identity element in the groupoid (Z..). -1 in Z is invertible
if aoe = a for all a in G.
groupoid (G,o)
because z.(-1) = =
(-1).z 1 holds in Z for c =-1. 2 in Z has no
An element e in G is said to be a
left inverse in the groupoid because there is no element c in Z such that
left identity in the groupoid (G, o
if eoa = a for all a in G. .2 1. Also 2 has no right inverse in the groupoid because there is no
element y in Z such that 2.y = 1.
Examples (continued). 8. 1 is the identity element in the groupoid (Q,.). 2 in Q is invertible
5. In the groupoid
=
(Z, +), 0 is a left identity as well as a right identity because there exists an element in Q such that 2.2 2. = 1. 0 in Q
In the groupoid (Z. .),1 is a left identity as well a right identity. is not invertible.
6. In the groupoid (Z, -), there is no
left identity, but 0 is a right identity Definition. If e be just a left identity in the
groupoid (G,o), then an
Theorem 2.2.1. If a groupoid (G, o) contains an identity element, element a in G is said to be left e-invertible if there
that element is unique. the in G such that boa =
exi_ts an element b
e and a is said to be right e-invertible if
there exists
an élement c in G suck that aoc =
e. b is said to be a
Proof. Let there be two identity elements e and f in left e-inverse of a
(G. o). and c is said be a right e-inverse of a.
 70 HIGHER ALGEBRA
GROUPSs
71
When e is just a a
right identity, then left e-inverse and a right e 2.4. Monoid.
inverse of an element can be detined in a similar manner.
A
semigroup (G,o) containing the is said to be a
Examples (continued). monoid. Therefore an algebraic system identityis element
(G, o) said to be a
= monoid if
9. In the groupoid (i) ao(boc) (aob)oc for all a, b, c E G; and
(Z,-),0 is a right identity. An element a in Z has
left 0-inverse as well as a a (ü) there exists an element e in G such that eoa =
aoe =
right 0-inverse in the groupoid. inG. a for all a
10. In the
* =
a,
groupoid (Z, ») where is defined by a*b a+ 2b, b E
is a right identity. 3 in Z is Z,0 A monoid (G, o) is said to be a
left 0-invertible but not commutative monoid ifo be
in Z is left 0-invertible as well right 0-invertible. 4 tative. commu-
as right 0-invertible.

2.3. Semigroup. Examples

1. (Z, +) is a monoid, 0
being the identity element.
A groupoid (G, o) is said to be a if o is
semigroup associative. 2. (Z..) is a 1
A monoid, being the identity element.
semigroup (G, o) is said to be a commutative semigroup if o 3. Let E be the set of all even
commutative. is is a
not a monoid. integers. Then (E,.) semigroup but
Examples. 4. (Zn) is a monoid, I being the identity element.
1. 5.
(Z,+) is a semigroup. (Q. +). (R, +) are semigroups. (M2(R), .) is a
monoid, the identity matrix I2 the
2. (Z,.) is a
semigroup. (Q,.). (R, .) are semigroups. element. It is not a comnutative monoid. being identity
3. (Z,-) is not a
semigroup. Theorem 2.4.1. In a monoid
(M, o) if an element a
a
it has be invertible then
.
is a
2) semigroup. It is a commutative semigroup. unique inverse.
Proof. Since a is
=
aoa= a'oa
invertible, there exists an element a'
Let (G, o) be a e, e in M such that
semigroup and a E G. Then G. aoa e of a.
being the identity element. a' is said to be an
=
ao(aoa) (aoa)oa, since o is associative. inverse
each of them is written Dropping the parantheses, Let there be two inverses a'
,
as aoaoa. a" of a in the
Then aoa a'oa =
=
monoid.
E =
=
Thus aoaoa E G, e and aoa"
G,... aoaoaoa element. a"oa e,e being the
Parantheses may, however, be inserted identity
=
in any manner for the We have
of calculation. purpose a'o(aoa")
But =
a'o(aoa")
(a'oa)oa", since o is associative.
a'oe a and =
=
=
The positive integral a E Therefore a' a". This =
(a'oa)oa" eoa" a".
powers of G are defined as follows.
= =
a, a =
proves the uniqueness of the inverse
a aoa, a" aoaoa, .... a*l =
of a.
a"oa for all n E N. Theorem 2.4.2. In a monoid
as well as if an element a
be a
Theorem 2.3.1. Let (S, o) right invertible then a (M,o) be left invertible
semigroup and a E S. Then am*n is invertible.
a oa for all m. n E N. Let e
Proof. be the identity element
=
and b be a left
Proof. a"*7 aoao = inverse, c be a
oa =
inverse of a. Then boa e, aoc
(m+n times ) e. right
=
=
a oa (aoao .oa)o(aoaooa) We have bo(aoc)
n times
times But =
(boa)oc, since o is associative.
bo(aoc) boe =b and = =
aoao oa Therefore b =
(boa)oc eoc c. =
m +n times ), since o is c and boa
Thercfore a"*" =
associative. aob= e. This shows that a is
a" oa". invertible.
Definition. In a monoid an invertible eloment is said to be a unit.
 HIGHER ALGEBRA4 GROUPS 73
72

2.5. QuasigroupP they are y = a,y=C.

two solutions and
The column of b contains a. Therefore the equation yob = a has a
A groupoid (G, ) is said to be a guasigroup if for any two elements
a , b E G each of the equations aoz = b and yoa = bhas a unique solution solution G. y b is = olution. The column of b does not contain c.
=
c has no solution in G.
in G. Therefore the equation yob

Examples. Worked Examples.

1. (Z,+) is a groupoid. Let a,b E Z. The equation a + = b has the o aob = ZZ.
on Z by a+b- ab for a, b E
1..Define a binary composition
solutionz = b- a in Z and the equation y + a = b has the solution Show that (Z,o) is a monoid.
y b - a in 2. Therefore (Z, +) is a quasigroup. Let a E Z,b E Z. Then aob = a +b ab E Z. Therefore Z is closed
2. (Z,.) is a groupoid. 2 e Z, 3 e Z. The equation 2.c = 3 has no under o.
solution in Z. Therefore (Z,.) is not a quasigroup. Let a, b, cEZ.
3. (Z,-) is a quasigroup but not a semigroup. ao(boc) = ao(b + c- bc) = a+(b+c bc) -a(b +c-bc)
= ab- bc-ca +abc.
4 . (M(R),.) is a groupoid. The equation A.X = B has no solution at+b +c
(aob)oc = (a +b ab)oc (a+b-ab) +c- (a +b- ab)c
M2, if A be a singular matrix. Therefore (M2,.) is not a quasigroup a+b +c ab- be- ca + abc.

As o is
Let G bea finite set and a, b, c, de G. The solvability of the equation ao(boc) =(aob)oc, associative.
Let us examine if there exists an element e in Z such that eoa = aoe
ao b can be read from the entries along the row of a in the compo
sition table. If the row of a contains b in the column of the element a for alla in Z. = a
=
eoa=a e +a-ea
then aoc b and therefore c is a solution of the equation aoT b. If e(1-a) =0
appears only once in the row of a, the solution of the equation aor6 e =0
is unique. and aoe = a a+e ae a
=

Similarly, the solvability of the equation yoa b can be read from

If
e(1 a ) =0
the entries along the columnn of a in the composition table.and the column e 0.
then doa =b therefore d
of a contains b in the row of the element d
=
This shows that 0 is the identity element.
of the equation yoa b. Ifb appears only once in thecolumn
is a solution
of a then the solution of the equation yoa = b is unique. Therefore (Z, o) is a monoid.
For example, lct us examine the o-composition table for the set G 2. Let (S. ) be a finite semigroup and a E S. Prove that there exist
fa.b, c. positive integers m, n such that a * = a". Deduce that a is an
idempotent clement in the semigroup.
In a scemigroup (S, o), an element z is said to be an idempotent ele-
ment if zot=
t.
Since (S, o) is a semigroup and a e S,a,a.a all belong to S.
The row b twice. Therefore the equation bor = b
of b contains has Since S is a finite set, there exist positive integers m and p (where
two solutions and they aret =a, T=C. p m ) such that aP = a
= a has a 1. Then = for some
Therefore the equation bor solution Let p m +n,n 2 amt a" positive integers
The row of b contains a.
solution. The row of b does not contain c. Therefore the
inG. t =bis a m, n.
= a+ =am+2n
a =a*n a" oa" a oa
eqution bot c has no solution in G. =
am+2n oan am+2 =am+3n
aTn+ a+2 +a+noa
The column of b contains b twice. Thercfore the equation yob= bihas
74
HIGHER ALGEBRA

GROUPS
Hence a=a"m*= amn =..= am+mn 75
=
Again a3 aoa= aoatmn =a2mtmn, 2.6. Groups.
Similarly, a A non-empty set G is
a
=a3m+mn, am=atm+mn..,anm =anm+mn to form
Therefore an
a " is composition o, if
said group with respect to a
idempotent element. binary
Note. In a finite (i) G is closed under the
semigroup (5,o) for each a E S, a" is an
element for some neN. composition o,
idempotent (i) o is associative,
3. Let (S, o) be a
semigroup with the identity element e. If each (ii) there exists an element e in G such that
in S be right invertible element eoa = aoe all a
prove that each clement in S is also left in G, =a for
Let a ES. Let a be a
right inverse
invertible. (iv) for each element a in
and let a" be a right G, there exists an element a' in
inverse = G such that
of a'. Then aoa' = e and a'oa" =e. of a
=
a' oa= aoa e.
=
=
a'oa =(a'oa)oe (a'oa)o(a'oa") a'o(aoa')oa" a'oeoa" =
a'oa" The group is denoted by the
This shows that a is
left invertible. =e. symbol (G, o).
The element e is said to be an
identity element in the group. We shall
prove that there is only one such element in the
group and therefore e
Exercises 7 will be said to be the identity element.
1. Define a o
The element a' is said to be
binary composition on Q by aob = an inverse of a. We
2a+b, a,b E Q. Show that a has
shall prove that each
Q, o) is a quasigroup but not a semigroup. element only one inverse and therefore a will be
Does there exist (i) a left identity in (Q.o) ? inverse of a. said to be the
(ü) a right identity in (Q, o) ?
2. Let R =
R-{0}. Define a binary o on = Definition. A group (G,o) is said to be
R by aob lab|, a, be a commutative
but not a
R. Show that (R,o) is a semigroup composition
quasigroup. abelian group the name of (after group or an
is commutative.
Norwegian mathematician N. Abel) if o
3. Let M be the set of all real matrices
):+b#o.
i) (M,o) is a semigroup where o is matrix multiplication,
1Prove that Theorem 2.6.1. A group (G, o) contains
only one identity eloment.
Proof. Let e, f be two identity elements in the group.
Then eoa aoe = = = =
(i) there is no left identity in the semigroup, a and foa
aof a for all a in G.
(iäi) =
o 1 0 i a right identity. We have eof f, by the property of e;
and also eof =
e, by the property of f.
4.
=
Define a binary composition * on the set Zx Z by
(a, b) (c, d) Therefore e f and this proves
(a +c, a +d),(a, b), (c, d) E Zx 2. Prove that (Z x Z,+) is a uniqueness of identity element.
but not a monoid.
semigroup Note. The identity element in (G, o) is denoted by ec.
5. Prove that (Zx Z,) is a commutative monoid, where is defined by Theorem 2.6.2.
(a,b) (c, d)= (ac,bd), (a, 6), (e, d) e ZxZ. In a group (G.o) each element has only one inverse.
) Find the units (invertible elements) in the monoid. Proof.Let a E G and a', a" be two inverses of a.
=
Then a'oa =aoa=e
and a"oa =aoa"
(i) Find the idempotent elements in the monoid.
e, e being the identity element.
We have a'o(aoa") =(aoa)oa, since o is associative.
6. Find the units and the idempotent elements in the monoid But d'o(aoa") a'oe = =
a' and (a'oa)oa" =eoa"
(i) (Zs,), (i) (Ze, ), (üi) (Zs, ). =a".
Therefore a' =
a" and this proves that the inverse of a is
7. In a finite monoid (M, o) if the identity element e is the
only idempotent
uniquc.
element, prove that each element of the monoid is invertible. Note. The inverse of a is denoted by a . Therefore
for
aoa= aoa = e holds. each a in G,
 276 HIGHER ALGEBRA
RINGS AND FIELDS
2Tt
Therefore o is not a
homomorphism. (ii) Since a is a unit in R, R is a ring with unity.
4. Let R =
=
(C,+, .)
and o: R^ R be defined
by o(2) z, 2 ¬ Since a is a unit in R, a" E R and a.al = a . a = I holds in I
z is the conjugate of the
complex number z. Cand R,
being the unity in R.
Let 21,22 E C. Then z1 + 22 E
C, z1.22 E C. = =
=
=
Since o is a homomorphism, o(a).p(a-1) ¢(a-1).ø(a) ¢(T).
o(z1) , d(z2) 2, ó(z1+22) 21 =
+22, o(z1.2a) =21 Since is onto, d(I) is the unity in R and it follows that p{«) is a
= unit in R' and [o(a)] = ¢(a-1).
We have b(z1 +2)
=
=22+22 ~i+2 ó(z1) +¢(22) and
oz1.22) =Z1.22 2.2= o(z1).¢(z2). This completes the proof.
Therefore is a homomorphismn.
o is a bijection. Hence o is an Note. If o: R R be an onto homomorphism and R' be a ring
witha
isomorphism. unity, then R may not be a ring with unity.
Theorem 3.14.1. Let R and R be two rings and o : R let R be the ring of 2 x 2
homomorphism. Then R be a For example, matrices Z of the form over
(i) p{0) 0', where=
0 is the zero element in R and 0' is
( R =Z and ¢:R- R be defined by ó a=a. Then
that in R o is a ring epimorphis1m. Here R' is a ring with unity, but R is not.
(ii) o(-a) = -o(a) for all a E R.
Proof. (i) 0 = 0 + 0 in R. Definition. Let R and R' be two rings and o: R - R' be a homomor-
Therefore o(0) = o(0 +0) in R' phism. Then the set d(R) = {d(a) : a ¬ R} is a subset of R. It is said
to be the homomorphic image of R.
=
o(0) + ¢(0), since ó is a
homomorphism.
Since 0' is the zero =
: R a
element in R', d(0) +0 o(0) + ¢(0). Theorem 3.14.3. Let R and R' be two rings and o R' be
Therefore 0' =
o(0), by left cancellation law for addition. homomorphism. Then d(R) is a subring of R.
(ii) Let a E R. Then a +(-a) = (-a) +a = 0 in R. Proof. o(R) is a non-empty subset of R, since 0'(= ¢(0)) ¬
We have p{0) = o(a +(-a)) = o(a) + d(-a) and also
$(R).
Let a E ®(R),b E Then there exist elements
o(R). a, b in R such
= that d(a) = a , d(6) =b.
d(0) $((-a) +a) =ô(-a)+ dla).
Therefore d(a) +o(-a) = o(-a) + ola) = 0'. Now a ' - b = #(a) -d(b) = dla - 6) ¬ d(R) and
=
This shows that a'b = ola)o(b) o(ab) E o{R).
d(-a) is the additive inverse of é(a) in R.
That =
Thus o' e E d(R)
is, do(-a) -p(a). d(R),& implies a'- be (R) and a'd E dR).
This completes the proof. This proves that o(R) is a subring of R'.

Theorem 3.14.4. Let R and R' be two rings and o: R - R be

Theorem 3.14.2. Let R and R be two and R » R' an onto
rings o: be an homomorphism. If R be a commutative ring, then R' is commutative:
onto homomorphism. Then
6) if R has a unity I, then R has o(1) as unity; d(R), E P(R). Then there exist elements a,b in R such
that o(a)
Let a' E=#', p(6) =b.
(ii) if a be a unit in R, then o(a) is a unit in R and [o(a)]=l = o l a )
a'= bla).o(b) = ö(a.b), since d is a homomorphism
Proof. (i) Let a E R. Since o is onto, there is an element a in R such ob.a), since R is commutative
that o(a) = a'. =since d is a homomorphism
We have a.I = I.a = a in R. = b.a.
olb).p(a),
=
Since is a homomorphism, dla).o(1) o{I).ó(a) This proves that o(R) is commytative.
=o(a).
That = This holds for all a' in
is, a'.o(T) d(T).a' =d in R. R. Note. The converse of the theorem is not true. This can be established
This shows that o(I) is the unity in R. by considering the example given in the Note of the Theorem 3.12.2.
 RINGS AND FIELDS 279
HIGHER ALGEBRRA
278

and u'r ¬ ¢(U).

R' Thus u' E o(U),r' E R' > r u ' E d{U)
R and R be two rings and ó: R be A
Theorem 3.14.5. Let an ideal of R'.
homomorphism. Then This proves that o(U) is
is a subring of R'; Proof left to the reader.
(i) if S be a subring of R, then d(S) (i)
and U be ideal of R an
¬ R: d(r) = ¬ S'} is a Note. If d: R ^ R be a ring homomorphism
(ii) if S' be a subring of R', then d-1(S') {r but may not be ideal of R'. an
subring of R. then o(U) is an ideal of o(R)
»
=
=
Z x Z and o: R R' be defined by
is a non-empty subset of R, since 0'(=o(0)) E o(S). For example, let R Z, R'
Proof. (i) d{$) :
= Then p is a homomorphism (not onto).
Let a' E E ¢(S). Then there exist elements a, b in S suchE o(n) (rn, n) n E Z. :
o(S), b =
= The ring 22 is an ideal of R but o(2Z) {(27n, 2n) n ¬ Z} is not
that o(a) = a', o(b) b. =

of since (2,2) E d(2Z), (1,3) ER

but (2,2)(1,3) (2,6) £
a E S,b ¬ S a - b E S and ab e S. an ideal R',
Since S is a subring of R,
Therefore dla- b) E ¢(S) and d(ab) e ¢(S). d(22).
But o(a-b) = d(a) -d(6) = a'- 8 and dlab) = ola)b(6) ==a'b.
Definition.
Therefore a' - b E p(S) and a'b ¬ o(S). Let R and R' be two rings and
let o: R ^ R be a homomorphism.
e d(S). =
Thus a' E E do(S) » d - b E d{S) and a'b
o(S),b is a subset of R given by ker d
The kernel of o, denoted by ker o,
a of R®. That is, ker ¢
This proves that is subring = where 0' is the zero element of R'.
o(S) {a E R: o(a) 0', zero element of
= elements of R that are mapped to the
Since S' is a subring of R', the zero element 0 E
S'. Since d(0) 0, is the set of those
(ii)
0 E(S") and therefore o(S') is a non-empty
subset of R. R'.
E S'. R ^ R' be a
Let a E 1(S'),b ¬ d-1(S"). Then o(a) ¬ S',$(6) Theorem 3.14.7. Let R and R'
be two rings and o:
-

E S' > Then ker o is an ideal of R.

Since S' is a subring of R', dla) E S', ¢{6) ¢(a) o[b) ES homomorphism.
=
where 0' is the zero element in R'.
and o(a)d(6) E S". Proof. ker o= {a E R: ¢(a) 0'},
-
= 0 E ker p.
- = and o(ab). ker o is a non-empty subset of R, because
But o(a) d(b) dla b) ola)¢(b)
Therefore a - b e ¢(S') and ab E o(S'). Let a,b E ker d. Then o(a) =0, d{b) ==0.
-

- = = + o(-b) = ola) ¢(6) = 0'.

is a subring of R. dla 6) o(a +(-6)) d(a)
-
This proves that d(S') bE ker d 4)
This shows that a
and o: R » R' be an =
Theorem 3.14.6. Let R and R' be two rings = = 0'
Let r E R. Then d(a.r) ola).d(r) 0'.¢(r)
onto homomorphism. Then and o(r.a) = d(r).$(a) = o(r).0' = 0'.
ideal of ideal of R'; Thus a E ker o and r E R* a.r E ker d and r.a E ker o . . (i)
(i) if U be an R, then o(U) is an
= R: o(z) ¬ U'} is an and it follows that ker o is an ideal of R.
(ii) if U' be an ideal of R', then o(U') {z e From (i) (ii)
ideal of R. This completes the proof.
U s a subring of R.
Proof. (1) Since U is an ideal of R, Examples (continued).
is a subring of R'.
By the previous theorem, d(U) of ker = R.
u in U, r 5. Forthe homomorphism d Ex.1, o
Let a' E d(U), r E R'. Then there exist elements in R
=
6:For the homomorphism o of Ex.4, ker d= {0}.
such that o(u) u, ¢(r) = r . ru ¬ U ur EU.
an u E U,r ¬ R> and
Since U is ideal of R, » R' be an
o(ru) E d(U) and o(ur) E ¢(U) Theorem 3.14.8. Let R and R' be two rings and d : R are
ker o,
But o(ru) = o(r)¢(u) and o(ur) = ¢(u)$(r). oto homomorphism. Then the ideals of R, each containing
Therefore r'u E d{U) and u'r ¬ d{U). in one-to-one correspondence with the ideals of R'.
 280 HIGHER ALGEBRA RINGS AND FIELDS
281
Proof. Let S be the set of all ideals of R, each o(a.b) = o(a).o(b)
containing ker d and S =
be the set of all ideals of R'.
o(b).dla), since R' is commutative
=
Let UE S. Then U is an ideal
of R and ker o p{U) is an cU. o{b.a), since o is a
of R', i.e., d{U) E S'..
(i) ideal Since homomorphism.
o is an isomorphism, ¢ is one-to-one and
Let U' E =
S'. Then U' is an ideal of R. o(b.a) > a.b = b.a. therefore o(a.b)
¢-(U') is an ideal of R.
Let a E ker o. Then This proves that R is a commutative ring.
d(a) =0, the zero element of R and
pla) e U. This implies a E ¢-(U'). Therefore ker o C o(U").
therefore (ii) Let R be a ring with unity I. Since o is also a
Hence o(U") E S...
(ii) follows from the Theorem 3.14.2.(i) that R' is a ring epimorphism, it
From (i) and (ii) it follows that to ring with unity o(I).
an element of
each element of S there
S', and conversely. corresponds Conversely, let R be a ring with unity I'. Let I be the pre-image of I'
under . Then o(I) = I'.
This proves the theorem.
Let a e R. Then = =
d(a.I) ola).¢(1) ¢(a).I' ¢la) and =
=
Theorem 3.14.9. Let R and R' be two o(I.a) $(T).¢{a) =!.o(a) =d(a).
rings and R^ R Since is o
be'an onto one-to-one, it follows that a.I I.a = =
a in R.
Thus a.I = I.a = a
homomorphism. Then o is an isomorphism if and o: only if ker o {0} holds for all a ¬ R. This
Proof. Let o be an isomorphism. Then o is one-to-one. unity in R. proves that I is the
We have o(0) =
0 and therefore 0 is a of 0. Since p
one-to-one, 0 is the only pre-image of 0'. So pre-image is =
(ii) Let a',b be non-zero elements of R'. Then there exist
unique ele =
ker d {0} ments a,b in R such that =
=
=
p(a) a'. o(6) b. Since o is an isomorphism,
= = =
-
Conversely, let ker d {0}. Let a,b ¬ R such that o(a) o(b).
-
a#0,b 0, because a 0 p(a) =0,b 0 p{b) 0.
= = -
Then o(a Since R contains no divisor
b) ó(a) ¢{b) 0'. This implies a b E ker d. a
= =
Since of zero, #0,b#0 ab # 0.
ker o= {0},a- b 0, i.e., a b.
Thus o(a) = =
Since is an ab
ø({6) a b. This proves that o is one-to-one. isomorphism, #0> dlab) # 0. Therefore o(a)o(b) #
0, i.e., a'V # 0. Thus a' # 0, b #0
This that R'
a'b #0 in R'.
So o is an one-to-one and onto proves contains no divisor of zero.
homomorphism. That is, d is ang
isomorphism. The proof of the converse
This the part left to the reader.
completes proof. Let a
(iv) be a unit in the ring R. Since
d is also a ring
Theorem 3.14.10. Let R and R' be two follows from the Theorem it
rings and o: R R be an 3.14.2.(ii) that o(a) is a unit inepimorphism,
the ring R'.
isomorphism. Then The proof of the converse
part left to the reader.
(i) if R be a commutative ring, then R' is a commutative ring, and Corollary. The isomorphic image of an integral domain is an
conversely; domain, the isomorphic image of a division integral
be a isomorphic image of a field is a field. ring is a division ring, the
(ii) if R ring with unity, then R' is a ring with unity, and con
versely;
an
Theorem 3.14.11. Let U be ideal of a ring R. Then the
(iii) if R be a ring without divisors of zero, then is a : R =
mapping
R ring without R/U defined by 6(s) t+U, z ¬ R is an onto
divisors of zero, and conversely; .
with kernel U. homomorphism
if a be a unit in R, then is a unit in Proof. =
=
(iv) o(a) R, and Let æ,y E R. Then
conversely. =
4() r+U, O(Y) y+U.
=
Let R be a e(r+y) (r+y) +U =
Proof. (i) commutative ring. Since o is an
isomorphism, = =
(« +U)+ (y+U) 6(r) +0(y);
o (xy) *y +U (z +U)(y +U) 0(r)0(y). =
is also an epimorphism and therefore R' is commutative, by
3.14.4.
Theorem This proves that 6 is a
homomorphism.
Converseky, Let R be a commutative ring and let a,b E R. The zero element in the quotient ring R/U is UU.
 RINGS AND FIELDS 283
HIGHER ALGEBRA
282

= is the trivial homomorphism.

= = U *rEU. When ker o F, o
=
ker 0{r e R:0 (z) U}. 0(T) U t+U of a field F is either an isomor
Therefore ker 0 =U. It follows that any homomorphism
This completes the proof. phism o r the trivia homomorphism.
be the natural homomorphism
Note. This homomorphism 6 is said to
from the ring R to the quotient ring R/U. Worked Examples.
and o: R» R be a ring homomorphism.
R the quotient ring 1. Let R be a ring with unity 1
The theorem says that for every ideal U of a ring = that ker = R.
the ring R, the kernel of the homomor If o(1) 0 prove
R/U is a homomorphic image of =
= 0. Therefore 0
phism being U. Let r E R. Then d(r) = ¢(r.1)
= o(r).o(1) o(r)
for all r E R. This proves that ker d= R.
Theorem 3.14.12. (Fundamental theorem of ring homomor
a
phism) 2. Let R be a ring with unity 1 and d: R + R' be ring homomorphism.
with kernel U If R contains no divisor of zero and ker ó # R, prove that o(1) is the
Let o be a homomorphism of a ring R onto a ring R'
Then R' is isomorphic to the quotient ring R/U. identity element of R'.
=
=
Let us define the map : R/U R' by b(a +U) = ola),a+ First we
prove that o(1) # 0. If o(1) 0, then for anyr¬R, ¢(r)
Proof. =
is well defined in the sense that if 0 and this =
implies ker o R, a contradiction.
UE R/U. First w e show that or.1)= ¢(r)$(1)
= = =
a + U = a + U then v(a'+0) = p(a). Therefore 0. Now
[$(1)]2 ¢(1).¢(1) ¢(1.1) o(1) # 0.
d(1) #
a'+U = a+U a ' - a EUd{#- a) = 0, since ker d =U - = - = -

¢l«') = p(a) > v(a' +U) = ola) Let E R'. Then ro(1) r"o(1) r'(o(1)]2 r p(1) r'o(1)
=
o in R'. Since R' contains no divisor of zero and o(1) # 0, it
r'o(1)
This shows that b is well defined. = r'.
follows that r'o(1)
-
i s a homomorphism, because the product r1, it follows that
= definition Similarly, by considering o(1)\o(1)}r
bf(a+U)+(b+U)] =v{(« +6) + U] ola+b), by o(1)r = r.
= =
o(a) + o(b) p(a +U) +b{b +U) and
+ =
Therefore r'o(1) = o(1)r' = r' for all r' E R'. This proves that ø(1)
b[a+U)(B+U)] = y[(ab) U] ¢(ab), by definition
= éla)d(b) = v(a+ U)V{6+ U). is the identity element of RR.

h is one-to-one, because 3. Prove that the only homomorphisms from the ring Z to Z are the
=
= > dla) =0 p(a-b) trivial homomorphism and the identity homomorphism.
bCa+U) %(b+U) d(a) d(6) - d{6)
=
0a b eU» (a +U) (b+U). Let o : Z Z be a homomorphism. We consider two cases.
form o(a) for
Finally, b is onto, because each element of R' is of the Case 1. o(1) = 0. Then for any z E Z, d(r) = d(r.1) = o(T)¢(1) = 0
some a E R and since the pre-image of o(a) is a+U in R/U.
to R. and this implies ker
=Z. Therefore d is the
Thus is an isomorphism and R/U is isomorphic trivial homonorphism.
Case 2. o(1) #0. Then since the codomain ring is Z and it contains no
This completes the proof.
divisor $(1) =1 by the previous Example.
of zero,
of a ring R is a quotient ring R/U for
Note. Every homomorphicofimage Let e
n Z and n is a positive integer.
some choice of the ideal U R. Then o(n) =o(1+1+ +1) =$(1) +¢(1) + +o(1) ==n... (1)
null ideal (0) and
Corollary. Since the only ideals of a field F a r e the Since is
F itself, for any homomorphism o of a field F, eiher ker o = {OF or a homomorphism, ¢(0) =0 (is)
non-zero a = = -
For any in Z, 0 = and
ker o= F. o(0) ¢la a) d(a) + ¢(-a)
When ker o= {0}, the homomorphic image of F is F, i.e., i s therefore p(-a) = -¢(a).
an epimorphism and and since ker ó = {0}, it is a monomorphism Let n e Z and n is a n =
negative integer. Let -m, where m isa
is isomorphism in this case.
Consequently, o an' positive integer.
 284 HIGHER ALGEBRA
RINGS AND FIELDS
285
= = =
Then o(n)
ó(-m) -d(m) - m = n . . (ii) 3. (i) Prove that the rings Z and 2Z are not
From (i), (i) and (ii) it follows that o(a) = a for isomorphic.
all a E Z and (i) Prove that the rings Z and Z x Z are not
therefore ó is the identity homomorphism. isomorphic.
(Hint.(i) Z is a ring without divisors
a
Note. The only isomorphism from the of zero, but the ring Zx Z is
ring Z onto Z is the identity divisors of zero.] ring with
mapping
4. Let R be a ring with unity 1 and
4. Prove that an o: Z+ R be defined by =
ideal U of a ring R is maximal if and only if the
quotient n EZ. o(n) nl for all
ring R/U is simple.
Show that o is a ring homomorphism. Find ker o if char R =
There exists a ring epimorphism o: R - char R =0. (i) 6, (ii)
=
R/U such that ker d =U.
Since o(R) R/U, the ideals of R/U are in one-to-one correspondence
with the ideals of R, each containing ker o. 5. Let o : ZxZ Z be defined by
is an onto homomorphism. Determine d(m, n) = m, (m, n) E ZxZ. Show that o
Let U be maximal in R. Then the only ideals of R
containing U are
ker o. Prove that (Zx
Z)/ker Z.
= is
Deduce that S = {(0, a) : a E
= Z} a prime ideal but not a maximal ideal
U and R. Now o(U)
in Z x Z.
U, the trivial ideal of R/U and o(R) R/U.
Therefore the only ideals of R/U are R/U and the null ideal.
This =
proves that the quotient ring R/U is simple. (Hint. Here ker d S. Z x Z is a commutative ring with unity and the quotient
ring(Zx Z)/S is an integral domain.
Conversely, let the quotient ring R/U be a simple ring. Then the only 6. Let R be the ring of all real
ideals of R/U are R/U and the null ideal {60}. valued continuous functions on closed
interval
Since U is an ideal of R, there exists a ring epimorphism d: R-R/UT [0, 1. A mapping d: R R is defined by d{f) = fG), f Ethe
that o is an onto R. Show
=
such that ker p U. Since o is an epimorphism, the ideals of
homomorphism. Determine ker p. Prove that R/ker o R .
in
R/U are Deduce that S = {f ER:
one-to-one correspondence with the ideals of R, each containing U f() = 0} is a maximal ideal in R.
=
-(R/U) = R and o-({®}) = ker o = UCR. So the only ideals (Hint. Here ker o S. R is a commutative ring with
unity and the quotient ring
R/S is a field.]
of R containing U are U and R. lence U is maximal in R.
T. Let o : Z+ Zs be
defined by o(r) = the remainder
Show that p is an onto of z (mod 5) for each
a EZ. homomorphism. Prove that Z/ker ó s.
Exercises 24 Deduce that 5Z is a maximal ideal in Z.
1. Which of the following are ring homomorphisms ? 8. Let R be the ring
=
:a,be Zand o: R -Z is defined by
(i) Let R= and d: R - R is defined by d(«) l=], r e R.
(R, +,),
(i) Let R= Ma(R), R = (R, +,.) and o: R R is defined by =a-b. Show that o is a ring
a-.| homomorphism. Determine ker d.
Show that the ring R/ker ó is isomorphic to Z.
Ma(R). Examine if ker o is (i) a
(:2])-s* [: 2]-[: *]- prime ideal of R, (ii) a maximal ideal in R.
(iü) Let R =Z(V ) and o: R Ris defned by ola+bv2) =a-bv2, a
b2e Z(V2). 3.15. Field of
quotients.
(iv) Let R= Ma(R) and o:R-- R is defned by dlA) = A', A E Ma(R Definition. A ring R is said to be embedded in a
ring S if there exists
a
2. Let d: R R be a ring homomorphism. Prove that monomorphism o of the ring R onto a subring of S. In this case the
= E R and all n e Z; ring S is also said to be an extènsion of the ring R.
(i) o(na) nøla) for all a R is
= allneN.
isomorphic with o(R), a subring of S. In this case we identify R
(i) ó(a") [¢(a)]* for all a R and with o(R) so that R may be considered
as a subring of S.
286 HIGHER ALGEBRA RINGS AND FIELDS 287

Examples. ker d= {a e R: ¢(a) = 0} = {0r} and this implies ó is a monomor-

1. The ring Z can be embedded in the ring Q by the mapping d: Z
0 phism.
defined by d(n)= i n e Z. Here d is a monomorphism and d(Z) is, Therefore the ring R is embedded in the ring 5, a ring with unity and
subring of the ring Q. char S =n.
= n then R can be embedded
a
2. The ring Z be embedded in the ring Z x Z by the mapping ¢: Z Note. If R be ring with unity and char R
is a char S = n.
Zx Z defined by d(n) = (n, 0), n E Z. HHere
monomorphism and a
in ring S with
unity and
o(Z) is a subring of the ring Z x Z.
Remark. A ring R with divisors of zero cannot be embedded in an
3. The ring R can be embedded in the ring C by the mapping o : R C
defined by o(r)= T+0i, r E R. Here d is a monomorphism and integral domain, since a subring of an integral domain is a ring with no
o(R) is divisor of zero.
a subring of the ring C. Our next attempt is to search whether we can have a similar extension
in case of an integral domain. The following theorem comes up.
Theorem 3.15.1. A ring R can be embedded in
ring S with unity a

Proof. Let S=ZxR. Let + and be defined on S by (m,a) + (7,b) Theorem 3.15.3. An integral domain D can be embedded in a field.
(m+Tn,a +6), (m, a).(n, b) = (mn, mb + ra+a.b), where ma = a +a+
Proof. Let T= {(«,v):ae D,y ED and y #03.
.+a(m times). Let a binary relation p be defined on T by (a, b) p (c, d) if and only if
Then (S,+,-) is a ring with unity, (1,0) being the unity. ad = bc.

Let : R» S be defined by = for a E R. Then For any (a, b) e T, (a, 6) p (a, b), since ab = ba. Therefore p is
d(a) (0, a) gd is a
homomorphism, since reflexive.
= = and
la+b) (0,a +b) (0,a) + (0,6) = ola) + ¢{b) Let (a, ) , (c, d) e T and (a, b) p (c, d). Then ad = bc. This implies
o(a.b) = (0, a.b) = (0, a).(0, b) = ola).o{b). cb = da and therefore (c, d) p (a, 6). Hence p is symmetric.
= = a
ker d= {a e R: ola) 0} {0r} and this implies d is monomo Let (a, b), (c, d), (p, a) ET and (a, b) p (c, d), (c, d) p (p, q) both hold.
phism. Then ad = be and cq = dp.
This implies adq = beq and bcq = bdp
Therefore the ring R is embedded in the ring S, a ring with unity.
adq = bdp
Note. If R be a ring with unity then R can be embedded in a ring S aq = bp, since d #0.
with unity. Therefore (a, 6) p (p, g). Hence p is transitive.
Theorem 3.15.2. A ring R of characteristic rn can be embedded in a Consequently, p is an equivalence relation.
=
ring S with unity and char:S n. Let Q(D) be the set of all distinct p-equivalence classes. Let [a, b)
Proof. Let S = Zn x R. Let + and be defined on S by (m,a) denote the class containing (a, b).
(n,6)= (m +n,a + 6), (m, a).(Tn, b) = (mn, mb + na + a.b), where Let us define + and. on the set Q(D) by
ma = a + a + . + a (m times). a, b+ [e,d = [ad +be, bd], [a, b].[e, d]= [ac, bd.
Then (5, +,.) is a ring with unity, (1,0) being the unity. As b0,d #0 and D is an integral domain, bd # 0. So the set Q(D)
Now n(1,0) = ( r , 0 ) = (0,0), since char Zn = n and therefore is closed under + and
First we show that the compositions + and . are well defined in the
char S n . , =

Let o: R S be defined by o(a) = (0, a) for a e R. Then o isa sehnse that if (a', &) E [a,b}, (¢, d') e [e, d then [a',] + [ad + [¢,d']
be, bd and [a', b].[¢,d] = [ac, bd).
homomorphism, since E ab' a'b > = ...

= = and (a',8) la,b] (a, 6) p (a', b') (i)

o(a+b)= (0,a +b) (0, a)+ (0, 6) dla) + ¢(b)
d(a.6) = (0, a.b) = (0, a).(0, 6) = d(a).¢{8).
(C,d) E [c, d > ( , d) p (¢,d') cd' =cd .. (ii)
 288 HIGHER ALGEBRA
RINGS AND FIELDS
289
Multiplying (i) by dd', (ii) by bb and then adding, we have Note. For any element in the field =
=
[a, 6] of quotients, [a, b] [a, 1].[1,b] =
ab' dd+ cd'bb a'bdd' + =
c dbb and this gives (ad + bc)bd (a'd a, 1].6, 11= la, 1]/b, 1] = #la)/d(6).
b'c)bd. Therefore (ad+be, bd) p (a'd+bc, bd') and hence [ad+be, bd This shows that every element of the field of
la'd +be, bd']. quotients of the integral
domain D can be expressed as a.b, where a E D,b e
Again from (i) and (ii) ab'cd' a'bc'd. Therefore = D,b# 0..
(ac, bd) p (a'e,ba'
and hence [ac, bd] =
[a'c,&d']. Theorem 3.15.4. The field of quotients F of an integral domain D is
=
= the smallest field containing D.
Thus [a',b]+[e,#'] [ad +be, bd] and [a', b],[e,d'] [ae, bd).
Next we show that Proof. Let K be any field containing D.
Q(D) is a field under + and.
Let 2 E F. Then r =
(i) +is commutativ, since for [a, b], [c, d E Q(D), [a,b] + a.b1, for some a,b in D with b 4 0.
=
ad +be, bd] [cb + da, db) le, d + la, }. =
[c, d Since D CK, a ¬ D,b E D,b #0 implies a E K,b E K,b
Since K is a field, a E K,b¬ K,b
#0.
i) It is a matter of simple verification that + is associative. #0 a.b-l e K, i.e., sE K.
Thus E F » z ¬ K and therefore F C K.
(ii) [0, 1] E Q(D) and it is the additive identity element. Since K is any
containing D, it follows that F is the smallest field containing D. field
-a, b] is the additive inverse of the element [a, b] E Q(D), since
= =
Note. Whenever an integral domain D is a
a, b+ [-a, b] [0, b] and [0, 6] [0, 14, since (0, b2) p (0,1).
not be the field of quotients of D. For
subring of a field F, F may
(v) i s commutative as well as associative on Q(D). is a subring of each of the fields
example, the integral domain Z
Q and R. As Q is the smallest field
(vi) [1, 1] E Q(D) and it is the multiplicative identity. containing Z, the field Q is the field of quotients of Z, but the field R is
not.
(vii) Let [a, b] be a non-zero element in Q(D). Then a 0 and there-
fore [b, a] E Q(D). = =
b, a].[a, ö) [ba, ab) [1,1), since (ba, ab) P (1,1). Worked Examples.
Therefore [b, a] is the inverse (multiplicative) of la, b].
1. Find the field of
(vii) The distributive law holds in Q(D). quotients of the integral domain Z.
Let F be the field of quotients. An element of F
Thus Q(D) is a field. where m E Z,n E Z,n #0, i.e.,
is of the form m/n,
every element of F is a rational number.
We now show that the domain D is isomorphic with some subdomain Therefore F cQ, where Q is the field
of Q(D).
of all rational numbers.
Let a E Q. Thenr =
p/q, where P, are integers and q 4 0. This
Let us define a D - =
mapping ó: Q(D) by øla) a E shows that a ¬ F and therefore Q C F. q
la,1] for D.
To prove that is a homomorphism, let a e D,b e D. Consequently, F = Q.
Then a +bE D, a.b e D.
2. Find the field of the
oa +b)= la +b,1] = la, 1] + b, 1) = o(a) + ¢(b), dla.b) = [ab, 1] = of quotients integral domain Z[i].
a, 1].[, 1] = d(a).d(U). Let F be the field of
where m E
quotients. An element of F is of the form
m/n,
So is Z[i), n e Z[i]. n # 0.
a homomorphism.
i s one-to-one, because o(a) = o(6) » [a, 1] = [b, 1} and this implies m = =
a+ bi, n c-+ di, where a,
(a, 1) p (b, 1) and therefore a = b. b, c, de Z, (¢,d) # (0,0).
=
- + are rational numbers.
MD) is a subdomain of Q(D) andD is isomorphic with é(D) under i=p+gi, where P, q
P+qi is an element
the mapping o. of the field Q[i]. Therefore FcQij.
Consequently, the integral domain D is embedded in the field Q(D) Let +yi e Qij. Then z E Qy ¬ Q and a = =
y $, where
Q(D) is called the field of quotients of the donain D. T, 3,u,v E Z, s # 0,v #0.

HA-19
 HIGHER ALGEBRA RINGS AND FIELDS 291
290

= +. and =
C F and therefore r.s,u.vl E F Definition. Two polynomials f(r) ao +a1T +a2 g(z)
T, 3,u,vE Z>r, s, u, v E Z) are said to be denoted by =
f(r) g(z),
since s #0,v#0. bot+b1 +b2at2+. in R[«] equal,
=
Hence E F, i.e., F. So Q[j] c F. if ai bi for each i.
+i z +yi e have the same de-
It follows from definition that two equal polynomials
Consequently, F =Qj.
gree.
We define 'addition' on the set R{T.
3.16. Polynomial Rings. Let f(r)= ao + a1t +a2 + and g(z) = bo + b1t +b2r +
the sum f(«) +9(2) is defned to be the
Let R be a ring and » be a formal symbol. A polynomial in a over be elements in RT]. Then
R.
R, usually denoted by f(r), glr), P(r),.. is an infinite formal sum polynomial co +C1T+C2T + , where Ci = ai +b; E
ag+a1+ag +.+@nt"+, It is a matter of simple verification that +) is a commutative
(R[#],
=
the zero polynomial.
where a E R
and a; 0 for all but a finite number of values of i. ai are group. The identity element is
called coefficients of the polynomial. We define 'multiplication' on the set R[z].
Iffor some n > 0, an #0 and a; = 0 for all i> n then an is said to and
Let fr) = ao t ar + a22+ gl«) =bo + b1z + b2a2 +..
be the leading coeficient of the polynomial; n is said to be the degree of be elements in R[æ]. Then the product f(z).9(z) is defined to be the
the polynomial f(r) and we write deg(f) = n or deg(f(r)) = n.
polynomial co +C1* +C2r + , where ck= aobk +a1bk-1+:+akbo.
If no such n exists, then the polynomial is of the form ao +Oz +0a It can be verificd that (R{r], +,.) is a ring.
I t is said to be a constant polynomial and the polynomial is said to
be of degree 0. The ring (R[z], +,.) is said to be the polymomial ring over the ring R.
= = to be
If R be a ring with unity 1, then the ring (R[r], +,.) is also a ring
If 0 for all i .,
0,1,2,.. the polynomial is said
a and no
the zero with unity, the unity being the constant polynomial 1.
polynomial degree is assigned to it.
a commutative ring, then the is also a com-
We agree to write the polynomial ao f R be ring (R[z],+, .)
+a1+ a2 +..
+*
with leading mutative ring.
coefficient an, as the finite formal sum ao +a1r +a2* +apT*
Here the degree of the polynomial is n. Theorem 3.16.1. If R be a ring with no divisor of zero, then the ring
Thus a a R can as (Rr], + ) is a ring with no divisor of zero.
polynomialof degree n over ring be written do t
+ +anT", where ao, a1, . ,a,n E R and an #0. Proof. Let f(z) = ao +a1t +a2 + +am and g(r) = bo +b17 -+
a1T+a2T
Note. The formal
bat+.+bn*" be non-zero elemnents in R[: with leading coefficients
symbol r cán be thought of as an elerment in an over
ring of R and it is assumed to commute with every element of R t13 Gmbn respectively.
also called an indeterminate over R. Let
the
us consider the isproduct f(r).g(7). The coefficient of amtn in
product f(z)-9(«) ambn and it is non-zero, since R contains no
= we this coefficient
If for some i, ai 1, the unity in R, drop from the divisors of zero. Therefore
formal sum and we write, for example, the f(c).9(a) is a non-zero polynomial in R[r
polynomial 2+3r +1r* +1 This proves that the ring (R[:],
+,.) contains no divisor of zero.
over the ring Z as the polynomial 2 +37 + +*.
=
some 0 some
Corollary. If D be an integral domain, then the polynomial ring
If for i, aj and aj # 0 for j > i, then we drop the Dz],+,.) is an integral domain.
term a;t* from the formal sum and we write, for example, the polynomia In particular, if F be a field, then also the
2+Or + 1r +3 over the ring Z as the polynomial 2+T" + 3r°. is an integral domain.
polynomial ring (F[T]), +, )
Thus 0, 2, z, are all polynomials with coefficients in the
2+* ring Z.
Theorem 3.16.2. If R be a ring and be in
The set of all polynomials over a ring R is denoted by R[c]. f(r), g(z) polynomials R:)
then deg(f(r)g(*)) <deg(f(«))+deg(9(z)).
 292 HIGHER ALGEBRA RINGS AND FIELDS
293

In particular, if R be an then
integral domain, deg(f(T)g(z)) Corollary. If D be an integral domain, then the units ofD are the
deg(f())+deg(g(*). units in the domain D[z]. only
Let =
Proof. In particular, if F be a
f(z) ao +a1t +a2* + = field, then the non-zero elements of F are the
+am and g(s) bo +b1* -+
bat+ +bnt" be polynomials in R[z] with leading coeficients only units in the domain F[s].
respectively. Then deg(f(z)) = m, deg(g(z)) = n. am, b
=
f(T)g(r) aobo +(agb1 +a1bo)z + We have seen that if R be a
If amba= 0, then deg(f(c)g(c)) < m+n.
+ambnt* ring, then R[r], the set of all polynomials
over R, is also a ring. If R be an
integral domain, then the polynomial
= ring R{z] is also an integral domain.
If ambn #0, then
deg(f(«))g(T)) m
Therefore deg(f(«)g(«)) £ deg(f(«))++.
deg(9(«). If some further condition be
imposed on the ring R, a natural
Second part. If R be an enquiry
integral domain, then 0 = comes
about the nature of the polynomial ring
0, since R contains no divisor of zero.
am # 0, bn # ambn R[z] up.
Therefore =
= deg(f())+deg(g(r). deg(f(:)g(T)) m+n We state here an important theorem in this connection, without proof.
Theorem 3.16.4. If D be a unique factorisation domain, then D{z] is
Examples. a UFD.
1. Let =
f(r) 32 +2x +4, g(r) 2z"+ +3 E Zsl=] =

=
Note 1. It follows from the theorem that if D be a
Then =
PID, then D[z] is a
3.
f(æ)g(r) r+4r3+r in Zs[T]. Here des(f(c)) 2, deg(g(z))= UFD. But D[«] may not be a PID. This
can be established
deg(f(«)g(«)) =4< deg(f(r))+deg(g(z)). example 3, page 260. by the worked
=
=
2. Let f (r) r2
-+2ax +4,9(r) 2x3 +*2+3 E Zslr]. Note 2. It follows from the theorem that if D be a
=
=
Then f(r)g(r) 2a5+5r* + 4r3 + in Zs{r]. Here then Da] is a UFD. Euclidean domain,
deg(9()) =3. deg(f(T)g(=)) =5 deg(f(r))+deg(o(«).deg(f(r)) 2, =

Since the integral

=
domain Z is a Euclidean
3. Let f(r) = domain, it follows from
3a2+2x +4, g(r) 2a +2+3 e Zs[a]. the same example [Ex.3,
=
page 260] that if D be an Euclidean
Then f(r)g(r) s +24 +3a2+a +2 in Zs[z]. Here DT] may not be a domain,
=
deg(f(«)) =2 = =
PID and consequently,
D[s] may not be a Euclidean
deg(g(r) 3. deg(f(r)g(r)) 5 degs(f(r))+deg(9(r)). domain.
Note 3. It
the
Theorem 3.16.3. If R be a ring with unity, then the units in follows from the theorem that if F be a
R are the only units of the ring is a UFD. In field, then F[*]
ring R[z]. subsequent theorems it will be established that FT is a
Proof. Let f(r) be a unit in R[z]. Then f(z) is a non-zero principal ideal domain and also a Euclidean domain.
R\T| and there exists some non-zero polynomial g(a) in R\c|polynomial in Note 4. The polynomial
= =
such th8t ring F[z] over a field F is not a field.
f(T)g(c) 9(r)f(r) 1, 1 being the unity in R[T]. The identity element in the
=
ring F[z] is the constant 1. If
=
Therefore deg(f(T)g(«)) deg 1 or deg(f(«))-+ deg(g(«)) no
0. Since P(T) be a non-constant polynomial in F{T] then there ispolynomial
both f(«) and g(c) are non-zero
polynomials in Rl7), deg(f(c)) 2 0 g() in F[z] such that p(r)4(z) = 1. polynomial
deg(g(«)) 2 0. For example, the
=
polynomial z in F[z] has no multiplicative inverse
in Flz]. Therefore is not a
Therefore the equality deg(f(:))+ deg(g(z)) 0 can hold only
= =
when Flz] field.
deg(f(r)) 0 and deg(g(c)) 0, i.e., only when f(z), g(«) are
non-zero
constant polynomials in R[T]. The non-zero constant Theorem 3.16.5.
are the non-zero elements of R.
polynomials in R[c (Division algorithm)
Let F be a field and
Thus a
f(z), g(7) ¬ F{z] with g(r) # 0. Then there
non-zero
element a in R is a unit in R{T] if there exists exist unique polynomials
non-zero element b in R such that ab
a = = =
q(r) and r(s) in F[s] such that
ba 1, i.e., if a is a unit in R. fl) g(z)a(«) + r(r),
where either = 0 or
In other words, the units in R are the only units in
R[T]. r(z) deg(r(«)) <
deg(9(z)).
 RINGS AND FIELDS 295
HIGHER ALGEBRA
294

Theorem 3.16.6. An element a in a field F is a zero of the polynomial

= <
Proof. If f(r) 0 or if f(a) # 0 but deg(f(«)) deg(g(r)), then in if z -
a is a factor of f(r).
So we may assume f(r) F(¢]
f()= Og(T) + f(z) and the theorem is satisfied.
that deg(f(=)) 20 and deg(f(«)) 2 deg(o(«). Proof left to the reader.
= 0.
Case 1. degf(=) Since deg(9(=) deg(f(), deg(ol"))0 Irreducible polynomial. Let F be a field. A nonconstant polynomial
this case. Let f(7) = a,g(r) = b for some a,bEF. Then a =b{b-"a)+0
and the theorem is satisfied. f(E) in F|c] is said to be an irTeducible polynomial in F[a] if f(z) cannot
be expressed as the product g(T)h(«) of two polynomials g(z) and h(r)
Case 2. 0. We use the second principle of induction. in Fll both of lower degree than deg(f(r).
deg(f(=)) >
Let us assume that the of degree 1) a 0 in F(#) is
theorem holds for all polynomials flr) of de
gree less than k and al non-zero polynomials g(r) such that deg(f(#))D Every linear polynomial (i.e., a +b,
clearly irreducible.
deg(g(-). distinction
= Let
k and Since F is a feld, F[r| is a UFD and therefore there is no
Let deg(f(#)) = n n deg(9(-)) such that k. a b between a prime element and an irreducible element in Fz].
be the leading coefficients of f(r) and g(«) respectively. Then f ( c )
ab*-"g(r) is a
polynomial in F[z] of degree k. An irreducible polynomial f(r) in Fel is also said to be irreducible
By induction hypothesis, there exist polynomials q(«) and r(*) in over the field F.
-

Fll such that f() ablak-g(7) =g(r)g(r)+r(), where either A polynomial f(r) in F{r] may be irreducible over the field F, but
r(z) = 0 or deg(r(z))< deg(g(T)). it may not be irreducible over a larger field K containing P as a subfield.
or, f(a) =g(«)[4%ba*-7 - q(z)] 4r(c).
For example, the polynomial 2 -2 is irroducible over the fieldQ but
Let s(r) = [asb'ak- - q(«)] E F{T]. is not so over the field R, because a2-2 = (z - V2)(z+ V2) in R.
= + where s(7),r(z) e F[lz] and either
Then f() g(7)s(z) r(z),
= 0 or
deg(r(z)) < deg(g(z)) and this shows that the theorem
holds Irreducibility test for quadratic and cubic polynomials can be easily
r(r)
for all polynomials f(z) of degree k and all non-zero polynomials g() done by the following theorem.
such that deg(f(«)) 2 deg(o(T).
Theorem 3.16.7. Let S(r) be a polynomial in Flz] of degree 2 or 3.
theorem holds for all polý
By the second principle of induction, the Then f() is reducible over Fif and only if it has a zero in F.
nomials f(z), g(r) ¬ F[z] with g(r) # 0.
Proof. Let f(r) be reducible over F. Then f(z) = g(z)h(r), where
=
let f(r) g{«)71(«) + ri(z) and f(r)
-
To prove uniqueness, -
deg(g()) < deg(f(z)) and deg(h(c)) < deg(f(«). Since deg(f(«)) is
g(T)g2(7) + r2(7). Then 9(r)lg1(z) g2(«)]= r2(7) ri(z). either 2 or 3, one of g(c) and h(7) must be of degree 1. I, say, g(a) is
Since the degree of r2(«) - ri(«) is less than the degree of g(«), this
=
of degree 1, then g(s) = ar+6, where a, b E F and a z# 0. Clearly, g(z)
- =
0, i.e., if q1(r) q2(7). Then we have
has a zero, -a-b, in F and consequently, - a b is a zero of f(c).
can hold only if q1(c) g2(r)
ri()= T2(7).
This completes the proof. Conversely, let a E F be a zero of the polynomial f(c) in F(z]. Then
is called the remainder in the divi - a is a factor of f(s) and this proves that f(z) is reducible over F.
( ) is called the quotient and r(z)
sion of f(c) by g(T). f(z) is said to be divisible by g(r) if the remainder This completes the proof.
be zero. In this case, is said to be a factor of f(c).
r(T) g(z)
Example. The polynomial + r +1 is irreducible over Z2, since the
Definition. An element a in a field F is said to be a zero (or a root) of is of degree 3 and none of the elements of the field Z2 is a
a in if = 0 in F.
e olynomial
zero of the polynomial.
polynomial f(«) F[c] f(a)
For example, 1 in Zs is a zero of the polynomial f(r) = z*+2r2+T+1
in Zs|c], since f(1) = 0; 1 is a zero of the polynomial f(z) = a + a 2 T h e r e are other elegant theorems for testing irreducibility of polyno-
mials over a field F.
in R[z), since f(1) = 0; the polynomial a* +2 in R[z] has no zero in R
296 HIGHER ALGEBRA
RINGS AND FIELDS
297
We state here without proof one such theorem due to Eisenstein.
Proof. Let U be an ideal in =
If U then U =
Theorem 3.16.8.(Eisenstein) Plz]. {O} (0), a
ideal. principal
= Let
Let p be a prime integer. Let
f(r) a,t"+an-1"-+. +ao E U# {0} and let g(r) be a
and an É0 (mod p), a; = 0 =
Zel non-zero polynomial in U of minimal
(mod p) for i 0,1,..., n-1 with ao 0 degree.
(mod p*). Then f(«) is over
irreducible Q. ¬
If the degree of g(r) be 0, then
g(») Fl:] and it is a unit. So
U= (1)= Flz] and it is
Examples. principal.
If the degree of g(r) be 2 1, let
1. =
=
Taking p 2, the polynomial - 2 is
f(c) be any polynomial in U. Then
irreducible over f() 9(=)a(«) +r(«) where
=
2.
q(z), r(z) ¬ F[E] and ither r(r) 0 or
Takingp 3, 275 -3a4 +92+3 is irreducible over Q
the degree of r(z) < the
degree of g(rs).
3. =
Since U is
Takingp 5, 2x0- 25r an ideal, g(r) e U,4(r) e F[»] g(«)a(z) e U and
+10r+5x +20 is irreducible over Q. f(T) ¬ U, g(r)g(r) ¬ U >r(«) ¬U.
-
4. For every the
prime p, polynomial " p is irreducible over
Since g(z) is a polynomial of minimal
Q. =
in Fa], it follows that
=
and therefore U =
r(a) 0. Consequently, f(r) g(7)q(r) degree
Worked Examples. principal ideal. (9(«)), a
1. Show that the Thus every ideal of
polynomial
a* +z+1 is irreducible over
Z5. Fl«} is principal and F[z] is a PID.
If
3+x +1 be factored as g(«)h(x), where both g(z) and This completes the proof.
are h(z)
polynomials of lower degree, then one of them,
= -
g(z) must be of Theorem 3.16.10. IfF be a field, then an
=
degree 1 and then g(z) r a for some a E Z5. say,
=
Therefore f(«) 0.
=
maximal if and only if p(r) is irreducible overideal (p(r)) # (0) in Fle] is
But f(0) =1, f(1) 3, f(2) 1, f(3) 1, f(4) 1, showing = =
F.
+c+1 has no zero
in Z5. that Proof. Let the ideal (p(z)) be maximal in Then (p(«)) # F[:]. So
Hence z +c+1 is irreducible over p() is not a unit in F[2| and therefore Flr].is a
Z5. =
p(s) polynomial of
2 1. Let p(c)
2. Find all irreducible 2 over
f(c)g(r) be a factorisation in F[E]. Then f(z)degree is a
polynomials of degree Zg. factor of P(c) and therefore
(p(«)) c (f(c)).
over
If (p(z)) be a
A polynomial of degree 2 Za is of the form ard+ br +C, proper subset of (S(T)), then =
a, b, c E Z3 and a #0. where (f(«)) F{z] because of
maximality of (p(r). In this case f(r) is a unit.
If (p(z)) =
Therefore a takes up 2 non-zero elements, b takes up 3 (f(«)), then p(«) and f(r) are associates in
elements,
takes up 3 elements of Zg independent of one another. The total c case g() is a unit. F[r]. In this
of such polynomials is 18. number Thus p(-) =
f(r)g(r) implies is a unit or g(«) is a unit.
in
This proves that p() is irreducibleeither f(z)
They are s*, ** +1, c* + 2, c +a, * + +1, r* +r +2, r+ F[a].
2r, + 2ax +1, **+2c+2, 2x, 27+1, 27 + 2, 2 +z, 2:x* +z+ Conversely, let p(«) be irreducible in F[z]. Let U be an ideal in
1, 22+ c +2, 2ax* +2x, 2ar+2x +1, 2ax+ 2r +2. such that (p(z)) C U C Flr]
F[z). Since F{T] is a principal ideal domain,
U =(f(z)) for some
=
The reducible polynomials are , f(z) ¬ F|z).
=
z* +2 (z +2)(c + 1), z*+r =
=
P()) c (f(7)) >p(r) f(«)g(z) for some g(z) ¬
w(r +1), r*+c+1= (r +2)(T +2), *i+2x z(z +2 ), r* +2c+1=E is irreducible. F{T]. But p(r)
is a
=
Therefore either f(z)
= unit.
(c+1)(r+1), 2r2, 2a +1 (r+1)(2r +1), 2c+a «(2x+1), 2r+ unit or g(r) is a
t+2 =
=
f(a) is a unit =
2(c+1)(c+1), 2:x+22 2x(T+1), 2+2+2 2(x+2)(c+2). implies (f(«)) Flz], i.e., U F[»]. =

9) is a unit implies p(z) and

The irreducible polynomials are +1,r* =
f(z) are associates and therefore
+r+2, *+2z+2,22+ (p(r)) U =

2,2x+a+1,2+2x +1. (f(T)), i.e., (p(«).

Thus p(z) cUC F(z] implies either U Fla] or
Theorem 3.16.9. If F be a field, then is a implies that the ideal U =(p(z).This
F[c] principal ideal domain
This completes the proof.
(p(r)) is maximal.
 RINGS AND FIELDS 299
HIGHER ALGEBRA
298
(z) is not a maximal
is Since the quotient ring i s not a field, the ideal
Theorem 3.16.11. If F be. field, then F[»]
a Euclidean domain.
ideal in Z[e]
Proof. Since F is a field, F is an integral domain and therefore Fll is
define a mapping v from the set of non-zero in Zal7]. Hence prove
an integral domain. Let us 4. Show that r* +1 is an irreducible poynomial
that the quotient ring is a field of 9 elements.
elements of F[z] to the set of non-negative integers by
=
vf()) deg(f(=)) for all non-zero polynomials f(c) E Flc] Let f(z) =2+1. If f(r) be reducible in Za[z], then f(z) = g(z)h(z),
non-zero then = T -a
If f(x), g(r) be
4) polynomials in F[a], where both g(r) and k(:) are polynomials of degree 1. Let g(T)
= = 2,
v[f(a)g(r)] = deg(f(r)g(=)) for some a E Z3. Then f(a) 0. But f(0) 1, f(1)= 2, f(2)=
=deg(f(z))+ deg(g(«)), since F[a] is an integral domain So +1 is irreducible over Z3.
showing that r* 4+1 has no z e r o in Z3. s
2 deg(f(c)) = v(f(z)). Since Za is a field, the ideal (r+1) is a maximal ideal in Zaz].
since Zalc] is a ring
If
f(r), g(*) be n o n - z e r o polynomials in F[z], then by division Consequently, the quotient ring i s a field,
(ii) with unity.
algorithm, there exist unique polynomials q(r) and r(z) in F[æ] such
that f(z) = q(z)g(r) + r(r), where either r(r) = 0 or deg(r(«))<
Any element of the quotient ring is of the form g(z) + (r2 +1), where
= 0 or <
deg(g()), i.e., either r(«) v(r(7)) v{g(T). g(T) is a polynomial in Za[z].
= =
Thus both the conditions for a Euclidean valuation are satisfied by By division algorithm, s(r) (r2+-1)h(r)+r(z), where either r(»)
=
1.
and hence is a Euclidean domain. 0 or deg(r(z))
F[z] In either case, r(r) = ar +b, where a,b are arbitrary elements in Z3.
This completes the proof.
Thus g(z) + ( r + 1) = a z + 6 + ( * + 1 ) , a E Z3,b E Z3.
a field, then is a principal ideal domain.
Corollary. If F is F[r] Since each of a and b can be chosen independently in 3 ways, the total
mmber of elements of the quotient ring (i.e., of the field) is 9.
Worked Examples (continued).
3. Show that the mapping o: Z[z] » Z defind by o(f(7)) = f(0), for
Exercises 25
f(r) E is a homomorphism.. Find ker p.
Z[T] but not a maximal
Deduce that the ideal (r) is a prime ideal in Z[z],
1. Find the feld
ideal in Z[(s). of quotients ofthe integral domain Z[V={a+bv2:a, be Z}.
a
every polynomial in Z[7] is mapped to its constant an example of ring R and two polynomials f(), g(=) in R{rl such
By the mapping d .thatGive
deg(f(c)g(=)) < deg(f(«)+ deg(o(«).
term.
= 3. Give an example of a ring R and two polynomials f(«), 9(z) in R[a] such
= bo + b17 +b2 F
Let f(c) ag + a1t +a2r + . +an",g(r)
that =
+ b n z " E Z{c. Then f(c) + 9(7) = (ao +bo) + (1 +b1)r + deg(f(:)s(«)) deg(f(*))+ deg(9(*))
f(r)g(r) = (a0bo) + (a0bi + a1bo)x +*' 4. Examine if the polynomial
= = and (i) 2+4z - 2 is ireducible over Q,
+ g(=))
b(f(z) ag +bo d(f(=)) + o(o(«))
S(T)g(=)) = aobo = ¢(f(a))olo(=). (ii) 8 +6 +9+24 is irreducible over Q,
This proves that ó is a homomorphism.
(iii) +2ac+2 is irreducible over R.
ker o = {f E Z[:f(0) = 0}. Therefore z E ker ô and ker o 5 Examine if the polynomial

p(r) for i.e., ker d= (r). (i) **+t+r+ 1 is irreducible over Z2,
all p(7) e Z[7], (ii) 2a +3 +2az +3 is irreduçible over Z7,
Z Z . Therefore
By the homomorphism theorem, e r
Since Z[r] is a commutative ring with unity and the quotient ring (ii) 2a75+5+2a+1 is irreducible over Zs.
is an integral domain, the ideal (z) is a prime ideal in Z[« Hint. (ii) Use Fermat's Theorem.
300 HIGHER ALGEBRA

=
6. Show that J {f(«) ¬ Z[«] : 6|5 (0)} is an ideal but not a prime ideal in

Hint. The quotient ring Zel/J is a ring of 6 elements of

having divisors zero]
7. Show that J = E
{f(c) Z[¢]: 5|f(0)} is a maximal ideal in Zle].
8. Show that the ideals (r) and (3a) in are
Z[z] different ideals but they are
the same ideal in QT]
Let I = =
[Hint. (=), J (3) be ideals in Let f(z) E I for
Qz]. some f(z) ¬ Q=)
Then =
zf(a) 3z.f(e) ¬ J, since f(c) Qc). ¬
Therefore I C J. Let 32f(«) E J:
for some f(e) E =
Q=]. Then 3f(z) z.3f(z) E I, since 3f(z) E Q=]. Therefore
JcI. Hence I =J]
9. Find all irreducible polynomials of 3 in
degree Z2lz).
10. Show that the polynomial z +z+1 is irreducible in Z2[z]. Show that
the quotient ring is a field of 4 elements.

11. Show that the + +1 is irreducible in

polynomial Z2[r]. Show that
the quotient ring is a field of 8 elenents.