Chapter 4

Binary Relations, Binary Operations

and Groups

In this final chapter we move on from studying the objects of mathematics, like numbers or
matrices, to looking at the ways we compare or combine them. We’ll study binary relations,
which are comparisons between elements of a set (such as < or =), and binary operations,
which are ways of combining elements of a set to produce another element of a set (such as
addition or multiplication of integers, or connectives ∧ and ∨ for statements). Some binary
operations on sets behave ‘nicely’ and the sets with these operations are called groups, for
example the set of integers under addition. Groups occur in many guises and group theory is
a very important part of mathematics. We’ll study groups in the second half of this chapter.

4.1 Binary Relations on a Set

Let A be a nonempty set. Informally a binary relation R on A can be thought of as a statement

about an ordered pair of elements of A. Examples of relations on the set Z of integers are <
and =, because the statements ‘a < b’ and a = b are either true or false. If the statement is
true for the pair (a, b), then we say a is related to b and write aRb. (In the case of <, we write
a < b.) If the statement is false we write aRb.\

More formally, a binary relation on A is a subset R of A × A. [The set R is the set of all pairs
for which the relation is true.] If (a, b) ∈ R, then we say that a and b are related and write aRb.

If A is finite then a relation R on A can be represented by its adjacency matrix. Given some
ordering a1 , a2 , . . . , an of the elements of A, then the adjacency matrix of R is an n × n matrix
(rij )n×n where

1 if ai Raj ;
rij =
\ j.
0 if ai Ra
88 Proof & Structure Chapter 4: Relations, Operations, Groups

Example 4.1.1 The relation > on the set A = {1, 2, 3, 4, 5, 6} taken in this order has the
following adjacency matrix.
 
0 0 0 0 0 0

 1 0 0 0 0 0 


 1 1 0 0 0 0 


 1 1 1 0 0 0 

 1 1 1 1 0 0 
1 1 1 1 1 0

Example 4.1.2 The following table gives a list of sets together with a relation on that set. In
this table E is a universal set and P (E) is its power set. S is the set of statements which can be
made from a collection p1 , p2 , . . . , pn of statements, using the connectives ¬, ∧, ∨ and ⇒, and
M(m, n) is the set of m × n matrices over R.

Set Relation
Z <
Z ≤
Z | (divides)
Z congruence modulo n
Z R0 , where mR0 n if and only if m + n is even
Z+ R1 , where mR1 n if and only if gcd(m, n) > 1
R R2 , where xR2 y if and only if x − y ∈ Q
Z × Z+ R3 , where (a, b)R3 (c, d) if and only if ad = bc
P (E) ⊆
S ⇐⇒ , where s1 ⇐⇒ s2 if and only if s1 and s2 are equivalent statements.
M(n, n) R4 , where AR4 B if and only if there is an
invertible matrix P such that P A = B
M(n, n) R5 , where AR5 B if and only if there is an
invertible matrix P such that P −1 AP = B
M(n, n) R6 , where AR6 B if and only det AB ≥ 0

Exercise 4.1 Give the adjacency matrix of the relations | , R0 and R1 when restricted to the
set A = {1, 2, 3, 4, 5, 6} (taking the elements in increasing order).

An alternative way of representing a relation R on a finite set A is a directed graph Γ in which

the vertex set is A and for a, b ∈ A, an edge is drawn from a to b if and only if aRb. We have
to put an arrow on each edge because aRb is not the same as bRa in general. A directed graph
is known as a digraph.
4.1 Binary Relations on a Set Proof & Structure 89

Example 4.1.3 The following are the digraphs associated with the relations >, |, R0 and R1
when restricted to the set A = {1, 2, 3, 4, 5, 6}.

�1
�
�5
6
�� ��2 ��4 3�

1 �
�� ��
��
5 3 �
� � ��
� 2 6
4
">" "|"

1 � 4
�� 2
� �� �

� � 1
�
� � �� �
3 5 ��6
2 4
� � 5
�
�
��3
��
6
" R0 " " R1 "

Types of Relation

A relation R on a set A is reflexive if for all a ∈ A, aRa. If a relation is reflexive, then there
is a loop at every vertex in the corresponding digraph, and the main diagonal of the adjacency
matrix is (1, 1, . . . , 1).

Example 4.1.4 We now consider some of the relations given in Example 4.1.2.
The relation < on Z is not reflexive, since no integer is less than itself, but ≤ is reflexive on Z.
For all integers m and n, n divides m − m = 0. Thus the relation congruence modulo n is
reflexive on Z.
\ 1 1 and so R1 is not reflexive on Z+ . Note that for all integers m > 1,
Since gcd(1, 1) = 1, 1R
gcd(m, m) = m > 1 and so mR1 m. However, a relation R is reflexive on a set if and only if
aRa for all a in the set.
The relation R5 on M(n, n) is reflexive, since for all A ∈ M(n, n), In−1 AIn = A. (Here, In is
the identity matrix.)

Exercise 4.2 Determine which, if any, of the remaining relations given in Example 4.1.2 are
reflexive.

A relation R on a set A is symmetric if for all a, b ∈ A, if aRb, then bRa. If a relation is

symmetric, then for every pair of vertices in the corresponding digraph, one of the following
occurs,
90 Proof & Structure Chapter 4: Relations, Operations, Groups

� ��
��
�

� �

and the adjacency matrix is symmetric.

Example 4.1.5 We now consider some of the relations given in Example 4.1.2.
The relations < and ≤ on Z are not symmetric. For example, 5 ≤ 9 but 9 ≤ 5.
For all integers l, m and n (with n ∈ Z+ ), if l ≡ m (mod n), then n divides l − m. Hence n
divides −(l − m) = m − l and so m ≡ l (mod n). Thus the relation of congruence modulo n
is symmetric on Z.
For all m, n ∈ Z+ , gcd(m, n) = gcd(n, m). Thus R1 is symmetric on Z+ .
The relation R5 on M(n, n) is symmetric since for all A, B ∈ M(n, n), if AR5 B, then there
is an invertible matrix P such that P −1 AP = B. Let Q = P −1 . Then Q−1 BQ = P BP −1 =
P P −1AP −1 P = A. Thus BR5 A.

Exercise 4.3 Determine which, if any, of the remaining relations given in Example 4.1.2 are
symmetric.

A relation R on a set A is transitive if, for all a, b, c ∈ A, if aRb and bRc, then aRc. If a
relation is transitive, then in the corresponding digraph, for every directed path u to v to w
there is an edge joining u to w. In particular, if there are two edges of a triangle, then the
whole triangle is part of the digraph.
v
�

u�
�
�w

Example 4.1.6 We now consider some of the relations given in Example 4.1.2.
The relations < and ≤ on Z are transitive.
For all integers a, b and c, and n ∈ Z+ , if a ≡ b (mod n) and b ≡ c (mod n), then n divides
a − b and b − c. Hence n divides (a − b) + (b − c) = a − c and so a ≡ c (mod n). Thus the
relation of congruence modulo n is transitive on Z.
The relation R1 is not transitive on Z+ . For example, 5R1 10 and 10R1 16 since gcd(5, 10) =
5 > 1 and gcd(10, 16) = 2 > 1. However, 5R\ 116, since gcd(5, 16) = 1.
The relation R5 on M(n, n) is transitive since for all A, B, C ∈ M(n, n), if AR5 B and BR5 C,
then there are invertible matrices P and Q such that P −1 AP = B and Q−1 BQ = C. Let
S = P Q. Then S −1 AS = Q−1 P −1 BP Q = Q−1 BQ = C. Thus AR5 C.

Exercise 4.4 Determine which, if any, of the remaining relations given in Example 4.1.2 are
transitive.
4.2 Equivalence Relations Proof & Structure 91

4.2 Equivalence Relations

A relation ∼ on a set A is an equivalence relation on A if ∼ is reflexive, symmetric and
transitive on A. That is, ∼ is an equivalence relation on A if for all a, b, c ∈ A the following
hold:
ER1 a ∼ a;
ER2 if a ∼ b, then b ∼ a;
ER3 if a ∼ b and b ∼ c, then a ∼ c.

Example 4.2.1 Congruence modulo n on Z is an equivalence relation, as is ⇐⇒ on S; ≤ on

Z is not an equivalence relation.

Exercise 4.5 Which, if any, of the relations in Example 4.1.2 are equivalence relations?

Let ∼ be an equivalence relation on a (nonempty) set A and let a ∈ A. The equivalence class
of a, denoted [a], is the set of elements of A which are related to a. That is

[a] = {b ∈ A : b ∼ a}.

Note that a ∈ [a] and so [a] = ∅.

Example 4.2.2 Consider the equivalence relation of congruence modulo 4. Then

[3] = {. . . , −9, −5, −1, 3, 7, 11, . . .} = {3 + 4m : m ∈ Z}.

More generally, for k ∈ Z, [k] = {k + 4m : m ∈ Z}.

Example 4.2.3 The relation R2 on R, where xR2 y if and only if x − y ∈ Q is an equivalence

relation. For r ∈ R, [r] = {r + q : q ∈ Q}.

Lemma 4.2.4 Let ∼ be an equivalence relation on a set A. Then for all a, b ∈ A,

a ∼ b if and only if [a] = [b].

Proof (⇐) Suppose that [a] = [b]. Since a ∼ a, a ∈ [a]. Thus a ∈ [b] and so a ∼ b by the
definition of [b].
(⇒) Suppose a ∼ b. Then by the symmetric property, b ∼ a. Let x ∈ [b]. Then x ∼ b. So
x ∼ b and b ∼ a. Thus, by transitivity, x ∼ a and hence x ∈ [a]. Therefore [b] ⊆ [a]. A similar
argument gives [a] ⊆ [b]. Hence [a] = [b].

Lemma 4.2.5 Let ∼ be an equivalence relation on a set A, and let a, b ∈ A. If [a] = [b], then
[a] ∩ [b] = ∅.
92 Proof & Structure Chapter 4: Relations, Operations, Groups

Proof (Contrapositive) Suppose [a] ∩ [b] = ∅. Then there exists x ∈ [a] ∩ [b]. Since x ∈ [a],
x ∼ a. Thus, by Lemma 4.2.4, [x] = [a]. Since x ∈ [b], x ∼ b. So by Lemma 4.2.4, [x] = [b].
Thus [a] = [x] = [b]. That is, [a] = [b].

Let ∼ be an equivalence relation on a set A. Since every element a ∈ A lies in exactly one
equivalence class, we can reconstruct the classes by giving a list of one representative of each
class. That is, we can choose a collection ai ∈ A, where i ∈ I for some indexing set I, such
that every element is in exactly one of the sets [ai ], i ∈ I. That is {[ai ] : i ∈ I} is a partition of
the elements of A. We call the ai equivalence class representatives.

Example 4.2.6 Consider the equivalence relation on Z of congruence modulo 4. Then [a] = [b]
if and only if a ≡ b (mod 4). There are four equivalence classes, and one set of equivalence
class representatives is {0, 1, 2, 3}. Another set is {8, 5, 2, 3}, as [8] = [0] and [5] = [1]. Any list
containing precisely one representative of each equivalence class is fine.

4.3 Binary Operations

Let A be a nonempty set. We can think of a binary operation ∗ on A as taking two elements
a, b of A (in that order) and producing a new element of A which we denote by a ∗ b. More
formally, a binary operation ∗ on A is a mapping from A × A to A; we denote the image of
(a, b) by a ∗ b. To show that an operation ∗ is a binary operation, we need to show that it is
defined for each pair a, b ∈ A, and that the product a ∗ b is an element of A.

Example 4.3.1 Let A be one of Z, Q, R or C. Then +, − and × are binary operations on

A. Note that ÷ is not a binary operation since, for all x ∈ A, x ÷ 0 is not defined. However,
on A∗ = A \ {0}, ÷ is a binary operation except when A = Z. In the case A = Z, although
m ÷ n is defined for all m, n ∈ Z∗ , the result is not usually an element of Z∗ . (For example,
/ Z∗ .)
12 ÷ 30 = 0.4 ∈

Example 4.3.2 Let A = Zn = {0, 1, . . . , n − 1}. Then ⊕ and ⊗ are binary operations on A.
[Recall that for a, b ∈ Zn , a ⊕ b and a ⊗ b are the elements of Zn congruent modulo n to a + b
and ab respectively.]

Example 4.3.3 Let P (E) be the power set of E. Then ∩, ∪ and \ are binary operations on
P (E).

Example 4.3.4 Let A be the set of all mappings from a set X to itself (so each f in A, is
a mapping f : X → X). Then ◦, the composition of mappings, is a binary operation on A,
defined as follows: if f, g ∈ A, then for all x ∈ X, f ◦ g(x) = f (g(x)).

Example 4.3.5 Matrix addition is a binary operation on M(m, n) and matrix multiplication
is a binary operation on M(n, n).

The previous examples are all familiar operations. However, it is possible to define all kinds of
‘unusual’ binary operations, as the following examples illustrate.

Example 4.3.6 Let A = Z+ . Then ∗ defined by m ∗ n = mn is a binary operation on A. For

example 3 ∗ 2 = 32 = 9.
4.3 Binary Operations Proof & Structure 93

Example 4.3.7 Let A = C. Then ⊙ defined by w ⊙z = wz is a binary operation on A. (Recall

that z is the complex conjugate of z, so if z = x + iy, then z = x − iy. This is sometimes known
as z ∗ .) For example

(3 + 2i) ⊙ (5 + 3i) = (3 + 2i)(5 − 3i) = 15 − 9i + 10i + 6 = 21 + i.

Example 4.3.8 Let A be any set and let c ∈ A. Then • and · defined by a • b = a and a · b = c
respectively are (rather dull) binary operations on A.

Multiplication Tables
Let A be a finite set. Then we can list the elements of A in some order a1 , a2 , . . . , an , say. If ∗
is a binary operation on A, then the multiplication table of A, with respect to this ordering,
is an n × n array in which the entry in the ith row and j th column (1 ≤ i, j ≤ n) is ai ∗ aj .

Example 4.3.9 The following are the multiplication tables for Z6 under ⊕ and ⊗ respectively.

⊕ 0 1 2 3 4 5 ⊗ 0 1 2 3 4 5
0 0 1 2 3 4 5 0 0 0 0 0 0 0
1 1 2 3 4 5 0 1 0 1 2 3 4 5
2 2 3 4 5 0 1 2 0 2 4 0 2 4
3 3 4 5 0 1 2 3 0 3 0 3 0 3
4 4 5 0 1 2 3 4 0 4 2 0 4 2
5 5 0 1 2 3 4 5 0 5 4 3 2 1

We can also use multiplication tables to define a binary operation on a finite set A.

Example 4.3.10 The following multiplication tables define three different binary operations
on the set A = {a, b, c, d}.

∗ a b c d ∗ a b c d ∗ a b c d
a a b d c a d a b c a c a d b
b b b d c b c d a b b a b c d
c d d a b c a b c d c d c b a
d c c b c d b c d a d b d a c

Special Properties of Binary Operations

We now consider some properties that certain binary operations satisfy. It should be empha-
sised that these are not random choices. These properties are shared by many familiar binary
operations, and so merit investigation in general.

A binary operation ∗ on a set A is commutative on A if for all a, b ∈ A, a ∗ b = b ∗ a. If A

is finite, and ∗ is commutative, then the multiplication table of A under ∗ is symmetric about
the main diagonal (NW – SE). We also say that A is commutative under ∗.

Example 4.3.11 R is commutative under +, but not under −. For example, 7−3 = 4 = 3−7.

Example 4.3.12 Both ⊕ and ⊗ are commutative on Zn .

94 Proof & Structure Chapter 4: Relations, Operations, Groups

Example 4.3.13 The binary operation ∗ on Z+ defined by m ∗ n = mn is not commutative.

For example 2 ∗ 3 = 23 = 8, but 3 ∗ 2 = 32 = 9.

Exercise 4.6 Which of the following binary operations are commutative on the given set?
Give examples to illustrate the cases where the binary operation is not commutative.

(a) C under ×;
(b) Q∗ under ÷;
(c) P (E) under ∩;
(d) M(2, 2) under matrix multiplication;
(e) C under ⊙, where w ⊙ z = wz;
(f) any set A under •, where a • b = a;
(g) any set A under ·, where a · b = c for some constant c ∈ A.
A binary operation ∗ on a set A is associative on A if for all a, b, c ∈ A, a ∗ (b ∗ c) = (a ∗ b) ∗ c.
Unfortunately there is no easy way to verify that a binary operation on A is associative from
its multiplication table.

Example 4.3.14 R under + is associative, but under − is not associative. For example
4 − (8 − 3) = 4 − 5 = −1, but (4 − 8) − 3 = −4 − 3 = −7.

Example 4.3.15 Both ⊕ and ⊗ are associative on Zn .

Example 4.3.16 The binary operation ∗ on Z+ defined by m ∗ n = mn is not associative. For

example, 2 ∗ (3 ∗ 2) = 2 ∗ 9 = 512, but (2 ∗ 3) ∗ 2 = 82 = 64.

Exercise 4.7 Which of the binary operations (a)–(g) given in Exercise 4.6 are associative
on the given set? Give examples to illustrate the cases where the binary operation is not
associative.

Exercise 4.8 The definitions of commutativity and associativity may seem very similar; both
are concerned with the order in which the elements are combined. But they are independent
properties. Establish this by giving examples of binary operations ∗ and ⋄ on given sets such
that ∗ is commutative but not associative, and ⋄ is associative but not commutative.

Note You may state without proof that the following binary operations on the given sets are
both commutative and associative (and remember that Z ⊆ Q ⊆ C).

Important: you must first check that the given operation ∗ is a binary operation on the given
set – that is, that the product of elements under ∗ is an element of the set.

Set Binary Operation

Any subset of C +
Any subset of C ×
Any subset of Zn ⊕
Any subset of Zn ⊗
Any set of subsets of E ∩
Any set of subsets of E ∪
Any subset of M(m, n) matrix addition
4.3 Binary Operations Proof & Structure 95

You can also state without proof that a set of mappings under composition of mappings and a
subset of M(n, n) under matrix multiplication are associative.

A binary operation ∗ on a set A has an identity element e ∈ A if for all a ∈ A, a∗e = a = e∗a.
If ∗ is a binary operation on a finite set of size n with an identity e, then in the multiplication
table, with respect to some ordering a1 , a2 , . . . , an of the elements
 of A, the row corresponding

a1
 a2 
 
to e is a1 a2 . . . an and the column corresponding to e is  .. . By convention we try to
 . 
an
order the elements of A so that e = a1 .
Example 4.3.17 R under + has an identity, namely 0. However R under − has no identity.
To see this, suppose e is an identity element. Then for all x ∈ R, x − e = x and e − x = x.
From x − e = x we get e = 0, and from x − e = x we get e = 2x. Choosing x = 0 gives 0 = x,
which is a contradiction. Thus R under − has no identity element.

Example 4.3.18 0 and 1 are identity elements of Zn under ⊕ and ⊗ respectively.

Example 4.3.19 The binary operation ∗ on Z+ defined by m ∗ n = mn has no identity. To see

this, suppose that e is an identity element. Then, for all m ∈ Z+ , e ∗ m =√m and m = m ∗ e.
That is em = m and m = me . Putting m = 2 gives e2 = 2 and so e = 2 ∈ / Z, which is a
contradiction.

Exercise 4.9 Which of the binary operations (a)–(g) given in Exercise 4.6 have an identity
element on the given set? In each case, either give an identity element or explain why there is
no identity element.

Let ∗ be a binary operation on a set A which has an identity element e ∈ A. An element a ∈ A

is invertible if there is an element b ∈ A such that

a ∗ b = e = b ∗ a.

If such an element b exists, then it is referred to as the inverse of a. Note that b must be an
element of A. For example, in Z+ under ×, n1 is not an inverse of n ∈ Z+ because n1 ∈ / Z+
(unless n = 1).

If ∗ is a binary operation on a finite set A which has an identity e, then a is invertible with
inverse b if the (a, b)th entry and the (b, a)th entry of the table are both e.
Example 4.3.20 R under + has identity 0. For all a ∈ R, if we set b = −a, then b ∈ R and
a + b = a + (−a) = 0 and b + a = (−a) + a = 0. Thus every element of a is invertible under +.

Example 4.3.21 0 and 1 are identity elements of Zn under ⊕ and ⊗ respectively. For m ∈ Zn ,
n − m ∈ Zn and
m ⊕ (n − m) = 0 = (n − m) ⊕ m.
Thus every element of Zn under ⊕ is invertible. Things are not so straightforward for ⊗;
whether an element m is invertible or not turns out to depend on both m and the modulus n.
For example 5 is invertible in Z6 under ⊗ (since 5 ⊗ 5 = 1) but 2 is not invertible. This will be
covered in more detail in the next section.
96 Proof & Structure Chapter 4: Relations, Operations, Groups

Example 4.3.22 The binary operation ∗ on Z+ defined by m ∗ n = mn has no identity and

so there are no invertible elements.
Exercise 4.10 Determine which of the elements of the following sets under the given binary
operations are invertible.
(a) C under ×;
(b) P (E) under ∩;
(c) M(2, 2) under matrix multiplication.
Let ∗ be a binary operation on a set A. An element a ∈ A is idempotent if a ∗ a = a.

It is customary to write a ∗ a as a2 . However, care should be taken not to generalise to higher

‘powers’ of an element, unless ∗ is associative on A. For example, if we write a3 , do we mean
(a ∗ a) ∗ a or a ∗ (a ∗ a)? This is one reason why associativity is such a useful property.
Example 4.3.23 For R under +, if x ∈ R is idempotent, then x + x = x and so x = 0. Thus
the identity element 0 is the only idempotent element.
Example 4.3.24 For the binary operation ∗ on Z+ defined by m ∗ n = mn , if m is idempotent,
then mm = m and so m(mm−1 − 1) = 0. Thus either m = 0 or m = 1. Since 0 ∈ / Z+ , m = 1 is
the only idempotent element.
Example 4.3.25 Let P = P (E) be the power set of some universal set E. Then, since A ∪ A =
A for all A ∈ P, every element of P under ∪ is idempotent.
Exercise 4.11 For each of the following binary operations on the given set, determine all of
the idempotent elements.
(a) Q∗ under ÷;
(b) P (E) under ∩;
(c) M(2, 2) under matrix multiplication;
(d) C under ⊙, where w ⊙ z = wz;
(e) any set A under •, where a • b = a;
(f) any set A under ·, where a · b = c for some constant c ∈ A.

4.4 Some Basic Results for Binary Operations

We now consider some results for binary operations on a set which satisfy one or more of the
properties given in the previous section.
It should be noted that in the examples and exercises in the previous section, if a set under a
binary operation has an identity element, then there is only one identity element. This is not
a requirement of the definition of an identity element, but is true in general as the following
result shows.
Lemma 4.4.1 Let ∗ be a binary operation on a set A. If A under ∗ has an identity element,
then this identity element is unique.
The proof of this is a typical uniqueness proof. We suppose e and ê are both identity elements.
Then we show that e = ê.
4.4 Some Basic Results for Binary Operations Proof & Structure 97

Proof Suppose e and ê are both identity elements of A under ∗. Then, for all a ∈ A

a∗e =a = e∗a
a ∗ ê = a = ê ∗ a.

Now e, ê ∈ A. Thus, setting a = ê in a ∗ e = a gives ê ∗ e = ê. Similarly, setting a = e in

ê ∗ a = a gives ê ∗ e = e. Combining the two expressions for ê ∗ e gives ê = e. Thus if A under
∗ has an identity element, then it is unique.

The next result considers the invertible elements of a set under a binary operation. We can
show that an invertible element has a unique inverse provided that the binary operation is
associative on the set.

Lemma 4.4.2 Let ∗ be an associative binary operation on a set A with an identity element e.
If a ∈ A under ∗ is invertible, then a has a unique inverse.

Note how the proof depends heavily on the associative property.

Proof Let a ∈ A be invertible and suppose b and b̂ are inverses of a. Then

a∗b =e= b∗a

a ∗ b̂ = e = b̂ ∗ a.

Consider b̂ ∗ (a ∗ b). Since a ∗ b = e we have

b̂ ∗ (a ∗ b) = b̂ ∗ e = b̂.

Also, using the associative property and b̂ ∗ a = e, we have

b̂ ∗ (a ∗ b) = (b̂ ∗ a) ∗ b = e ∗ b = b.

Combining the two expressions for b̂ ∗ (a ∗ b) gives b̂ = b. Thus a has a unique inverse.
Lemma 4.4.3 Let ∗ be an associative binary operation on a set A with an identity element e.
If a ∈ A under ∗ is invertible and idempotent, then a = e.

Proof Let a ∈ A be invertible and idempotent. Then there is an element b ∈ A such that
a ∗ b = e = b ∗ a and a2 = a. Consider a2 ∗ b. Since a is idempotent we have

a2 ∗ b = a ∗ b = e.

Also, using the associative property and a ∗ b = e, we have

a2 ∗ b = (a ∗ a) ∗ b = a ∗ (a ∗ b) = a ∗ e = a.

Combining the two expressions for a2 ∗ b gives a = e.

Note that if ∗ is an associative binary operation on a set A with an identity element e and all
elements are invertible, then Lemma 4.4.3 says that the only idempotent element is e.

Lemma 4.4.4 Let ∗ be an associative binary operation on a set A with an identity element e.
If x2 = e for all x ∈ A, then ∗ is commutative on A.
98 Proof & Structure Chapter 4: Relations, Operations, Groups

Proof Let a, b ∈ A. Now a2 = e and b2 = e. Let x = a∗ b. Then x2 = e. So (a∗ b) ∗ (a∗ b) = e.

Now multiply on the left by a and on the right by b:

a ∗ (a ∗ b) ∗ (a ∗ b) ∗ b = a ∗ e ∗ b = a ∗ b

By associativity, and the fact that a2 = e and b2 = e, the left hand side becomes

(a ∗ a) ∗ b ∗ a ∗ (b ∗ b) = a2 ∗ (b ∗ a) ∗ b2 = e ∗ b ∗ a ∗ e = b ∗ a.

Therefore b ∗ a = a ∗ b. This holds for all a, b ∈ A. That is, ∗ is commutative on A.

In some cases the preceding lemmas can be used to show that a binary operation on a set is
not associative, as the following examples illustrate.

Example 4.4.5 Consider the binary operation ∗ on the set {e, a, b, c} defined by the following
multiplication table.
∗ e a b c
e e a b c
a a c e e
b b e c b
c c e b c
Then ∗ on {e, a, b, c} has an identity, namely e. However, b and c are both inverses of a and so,
by Lemma 4.4.2, ∗ on {e, a, b, c} is not associative.

Example 4.4.6 Consider the binary operation ∗ on the set {e, a, b, c} defined by the following
multiplication table. ∗ e a b c
e e a b c
a a a e c
b b e b a
c c c a c
Then ∗ on {e, a, b, c} has an identity, namely e. However, a and b are both invertible and
idempotent and so, by Lemma 4.4.3, ∗ on {e, a, b, c} is not associative.

Example 4.4.7 Consider the binary operation ∗ on the set {e, a, b, c, d} defined by the follow-
ing multiplication table.
∗ e a b c d
e e a b c d
a a e c d b
b b d e a c
c c b d e a
d d c a b e
Then ∗ on {e, a, b, c, d} has an identity, namely e. Also x2 = e, for all x ∈ {e, a, b, c, d}. However,
a ∗ b = c and b ∗ a = d and so ∗ on {e, a, b, c, d} is not commutative. Thus, by Lemma 4.4.4, ∗
on {e, a, b, c, d} is not associative.

Exercise 4.12 Show that each of the binary operations defined in Examples 4.4.5, 4.4.6 and
4.4.7 is not associative on the given set by finding elements x, y and z in the set such that

x ∗ (y ∗ z) = (x ∗ y) ∗ z.
4.5 Groups Proof & Structure 99

4.5 Groups
There are many examples of binary operations in mathematics, so we ask ourselves what types
of properties do binary operations often have? Over time, three particular properties have
attracted much attention, leading to one of the most fruitful areas of mathematics today,
namely Group Theory. A group is just a set with an associative binary operation in which
there is an identity element and every element has an inverse. Groups are important because
they are so ubiquitous. For example, the set of integers (with addition) is a group, but so is
the set of symmetries of any shape, and there are countless other examples. In this section we
look at many examples of groups and establish some of their properties. Our first task is to
define formally what is meant by a group.

Definition A group is a non-empty set, G, together with a map ∗ defined on G × G, such

that the following axioms hold:

G0 For all g, h ∈ G, g ∗ h ∈ G. That is, G is closed under ∗. This is equivalent to saying that
∗ is a binary operation.
G1 For all g, h, k ∈ G, (g ∗ h) ∗ k = g ∗ (h ∗ k). That is, ∗ is associative on G.
G2 There is an element e ∈ G, such that for all g ∈ G, g ∗ e = e ∗ g = g. That is, G under ∗
has an identity element.
G3 For all g ∈ G, there is an element g −1 ∈ G, such that g ∗ g −1 = e = g −1 ∗ g. That is, every
element of G is invertible under ∗.

We denote a group as a pair (G, ∗), or just G if the binary operation is clear from the context.

Note Sometimes a group is defined as a non-empty set G together with a binary operation
∗, such that axioms G1, G2 and G3 hold. This is equivalent, because by definition, if ∗ is a
binary operation on G, then G is closed under ∗.

The ∗ is usually dropped and we write gh instead of g ∗ h, for the product of g and h. However,
this does not mean that the binary operation is necessarily multiplication, it is just that it gets
a little tedious always writing the ∗.

The identity element is sometimes denoted by 1, which is not necessarily the number 1. If we
are considering several different groups at the same time, then an identity element of G will,
for clarity, be denoted eG or 1G .

4.6 Examples of Groups

Numbers
Z, Q, R and C are all groups under addition. If G is one of these sets, and ∗ is addition, then
eG = 0 and for any g ∈ G, g −1 = −g.

If we consider multiplication, we see that the only possibility for the identity element is 1,
because if eg = g, then g(e − 1) = 0, so either e = 1 or g = 0. Since this equation must hold
for all g ∈ G, we must have e = 1. Now since gg −1 = 1, we get that g −1 = 1/g. This does not
100 Proof & Structure Chapter 4: Relations, Operations, Groups

work when g = 0. Therefore Z, Q, R and C are not groups under multiplication. We can rescue
the situation with the following definition:

Definition Let X be a subset of numbers. Then define X ∗ = X − {0}.

For x ∈ X ∗ , we can now define 1/x, and so Q∗ , R∗ and C∗ are all groups under multiplication.
Z∗ is not a group under multiplication because, for example, 1/2 ∈/ Z∗ , so 2 has no inverse.

Definition Let G be a group under ∗. Suppose a subset H of G is also a group under ∗.

Then H is called a subgroup of G.

Example 4.6.1 2Z = {2m : m ∈ Z} is a subgroup of Z under addition.

√ √
Example
√ 4.6.2 Q( 2) = {a + b 2 : a, b ∈ Q} is a subgroup of R under addition. Moreover,
Q( 2)∗ is a subgroup of R∗ under multiplication.

Exercise 4.13 Verify that the sets in the previous two examples are groups under the stated
binary operations.

Although the elements are not actually numbers, we observe that Zn is a group under ⊕, with
identity 0. We also note that in the group (Zn , ⊕) every element can be expressed in terms of
1, since, for all i ∈ Zn ,
i = 1 ⊕ 1 ⊕ · · · ⊕ 1.
i times

We say that (Zn , ⊕) is a cyclic group generated by 1. (More on cyclic groups later.)
Although (Zn , ⊕) is a group, (Zn , ⊗) is not a group. (Why?) However the set Un given by

Un = {m ∈ Zn : gcd(m, n) = 1}

is a group under ⊗. When n is prime, Un = {1, 2, . . . , n − 1} and in this case we sometimes use
the notation Z∗n for Un .

Example 4.6.3 The multiplication table for U5 = {1, 2, 3, 4} under ⊗ is

⊗ 1 2 3 4
1 1 2 3 4
2 2 4 1 3
3 3 1 4 2
4 4 3 2 1
The multiplication table for U8 = {1, 3, 5, 7} under ⊗ is
⊗ 1 3 5 7
1 1 3 5 7
3 3 1 7 5
5 5 7 1 3
7 7 5 3 1

The previous examples lead to the following obvious definition.

Definition A group (G, ∗) is finite if G is finite, and is infinite otherwise.

4.6 Examples of Groups Proof & Structure 101

Matrices
Let F be any one of Z, Q, R or C and let Mm,n (F ) denote the set of m × n matrices over F .
Then Mm,n (F ) is a group under matrix addition.

Clearly if Mm,n is to have any hope of being a group under matrix multiplication, then we
must have m = n, but this is not enough. However, if we consider only the invertible n × n
matrices over F , and restrict to the cases F = Q, R and C, then we get a group over matrix
multiplication. This group is denoted by GLn (F ) and is called the general linear group of
degree n over F .

Technical Note We had to exclude Z because the inverse of a matrix with integer entries is
1 2
not guaranteed to have integer entries. For example the matrix M = is an element
0 3
3 −2
of M2,2 (Z) but the inverse is M −1 = 3−1 . Now 3 is not invertible in Z, because
0 1
1
3
∈
/ Z. Hence M is not invertible in M2,2 (Z).

Matrices give us an instance of why the group axioms are useful: they guarantee that can-
cellation is possible. Given a binary operation ∗, it would be nice to be able to deduce that
a ∗ b = a ∗ c implies b = c. To do this we need all the group axioms. Firstly we note that a has
an inverse a−1 .
a∗b = a∗c
−1
a ∗ (a ∗ b) = a−1 ∗ (a ∗ c)
(a−1 ∗ a) ∗ b = (a−1 ∗ a) ∗ c) by associativity
e∗b = e∗c
b = c by the properties of the identity
Cancellation is certainly not an automatic property of binary operations, as the following
exercise shows.
2 3 1 7
Exercise 4.14 Consider the matrices A, B and C given by A = , B = ,
4 6 2 0
4 1
C= . Show that AB = AC, even though B = C. Which of A, B and C are elements
0 4
of GL2 (R)?
Again, many subsets of GLn (F ) are themselves groups under matrix multiplication, and the
following examples show that they may be either finite or infinite.
Example 4.6.4 SLn (F ) = {M ∈ GLn (F ) : det M = 1} is a group under matrix multiplica-
tion, called the special linear group of degree n over F .
Example 4.6.5 The following set Q8 is a group under matrix multiplication.
1 0 −1 0 i 0 −i 0
Q8 = , , , ,
0 1 0 −1 0 −i 0 i
0 1 0 −1 0 i 0 −i
, , ,
−1 0 1 0 i 0 −i 0
Exercise 4.15 Verify that SLn (F ) and Q8 are groups under matrix multiplication.
102 Proof & Structure Chapter 4: Relations, Operations, Groups

Geometry
Let Φ ⊆ Rn (we will only consider n = 2 or 3 here). Φ is sometimes called a figure. A
symmetry of Φ is a bijection α : Φ → Φ, which is distance preserving. That is, for any
x, y ∈ Φ,
d(x, y) = d(α(x), α(y)),
where d(x, y) denotes the distance from x to y.

The set of symmetries of a figure Φ, under composition of mappings, forms a group, denoted
by Sym(Φ). This is verified as follows:

Closure Let α, β ∈ Sym(Φ). It must be shown that β ◦ α ∈ Sym(Φ). That is,

d(β ◦ α(x), β ◦ α(y)) = d(x, y),

for all x, y ∈ Φ. Now

d(β ◦ α(x), β ◦ α(y)) = d(β(α(x)), β(α(y)))

= d(α(x), α(y))
= d(x, y).

Associativity Sym(Φ) is associative, since composition of mappings is an associative opera-

tion.

Identity element The identity mapping I : Φ → Φ, defined by I(x) = x, for all x ∈ Φ, is a

symmetry and an identity element of Sym(Φ).

Inverses Since any element α of Sym(Φ) is a bijection, the inverse mapping α−1 : Φ → Φ
exists. It remains to show that a−1 is a symmetry. Now,

d(x, y) = d(α ◦ α−1 (x), α ◦ α−1 (y))

= d(α(α−1 (x)), α(α−1(y)))
= d(α−1 (x), α−1 (y)).

Thus α−1 ∈ Sym(Φ).

Exercise 4.16 Describe the symmetries of the following figures in R2 .

Exercise 4.17 (HARD!) Determine the number of elements of Sym(C), where C is a cube.
4.6 Examples of Groups Proof & Structure 103

Example 4.6.6 The figure Φ does not have to be bounded in size. For example, suppose that
Φ(⊆ R2 ) is the infinite ‘brick’ pattern shown in the diagram.

In this case Sym(Φ) is an infinite group. This is because, in addition to the reflections and
rotations we have encountered already, there are two other types of symmetries.

The first of these is a translation, which moves every point the same distance in a specified
direction. Two such translations are shown on the diagram with a directed line segment. Each
of these translations can be composed with itself an arbitrary number of times, so we see that
there are infinitely many translations. Hence Sym(Φ) is infinite.

The second symmetry is a glide reflection, which consists of a reflection in a line l, followed
by a translation in the direction of l.

The symmetry groups of unbounded figures in R2 are called 2-dimensional crystallographic

or wallpaper groups. There are essentially 17 ‘different’ crystallographic groups.

It can be shown that every symmetry of a figure in R2 has to be either a rotation, a reflection,
a translation or a glide reflection.

Dihedral Groups Let Pn be a regular n-sided polygon (sometimes described as an ‘n-gon’).

Then Sym(Pn ) has 2n elements. There are n rotations α0 , α1 , . . . , αn−1 about the centre of Pn
through 0, 2π , 4π , . . . , 2π(n−1)
n n n
respectively (so αk is a rotation through 2πk
n
). There are also n
reflections where if n is odd these are in lines through a vertex and the midpoint of the opposite
side, and if n is even n/2 are in lines through opposite vertices, and n/2 are in lines bisecting
opposite sides of Pn .

Sym(Pn ) is a fairly important group and is called the dihedral group with 2n elements and
is denoted by Dih(2n).

You can check that if α = α1 and β is one of the n reflections, then αβ = βα. This leads to
the following definition.

Definition A group (G, ∗) is abelian or commutative if x ∗ y = y ∗ x for all x, y ∈ G. It is

nonabelian or noncommutative otherwise.

So, for example, (Z, +) is abelian and (Dih(2n), ◦) is nonabelian.

Exercise 4.18 Which of the groups given so far in this chapter are abelian?
104 Proof & Structure Chapter 4: Relations, Operations, Groups

Permutations
Recall that a permutation of a set X is a bijection from X to itself. If X is a finite set of size
n, then there are n! distinct permutations of X.

The set of permutations of X, denoted by SX , forms a group under composition of mappings.

Example 4.6.7 α : R → R, α(x) = 2x − 3 is a permutation.

β : R → R, β(x) = x2 is not a permutation since β(2) = 4 = β(−2), so β is not injective.

Example 4.6.8 Let X = {1, 2, 3, 4}. Then α : X → X, where α(1) = 1, α(2) = 4, α(3) = 2
and α(4) = 3, is a permutation of X.

Convention If X is finite of size n, then we usually take X to be the first n positive integers
(that is, X = {1, 2, . . . , n}). In this case SX is usually denoted Sn , or Sym(n).
Note that Sn is not abelian for n ≥ 3, because for instance a quick calculation shows that
(12)(23) = (123) but (23)(12) = (132).
Recall from Algebra 1 that we can write any permutation in Sn as a product of disjoint cycles,
and also as a product of transpositions. We say a permutation is even if it can be written as a
product of an even number of transpositions and odd otherwise.

Exercise 4.19 Let X = {1, 2, . . . , n} with n ≥ 2.

(a) Show that exactly half of the permutations of X are even permutations.
(b) Let An = { even permutations of X}. Show that (An , ◦) is a group with 12 n! elements.

An is called the alternating group of a set of size n.

4.7 Properties of Groups

Now that we have a large class of examples, we move on in this section to deduce, from the
axioms, some simple properties of groups. We will introduce some of the fundamental ideas
of group theory. Recall that if a binary operation on a set has an identity element, then
this element is unique. Thus any group has a unique identity element. Also, since the binary
operation for a group is associative, every element of a group has only one inverse. The following
result gives some other consequences of the group axioms. Note that we suppress the binary
operation ∗ and use multiplicative notation.

Lemma 4.7.1 Let (G, ) be a group with identity e. Then, for all a, b, c ∈ G,

(a) (a−1 )−1 = a;

(b) (ab)−1 = b−1 a−1 ;
(c) if ab = e, then b = a−1 ; and
(d) if ac = a, then c = e.
4.7 Properties of Groups Proof & Structure 105

Proof (a) a−1 a = e = aa−1 and so, by definition, a is an inverse of a−1 . Now (a−1 )−1 is also
an inverse of a−1 . Thus, by the uniqueness of inverses, (a−1 )−1 = a.
(b) (ab)(b−1 a−1 ) = a(bb−1 )a−1 = aea−1 = aa−1 = e. Similarly (b−1 a−1 )(ab) = b−1 (a−1 a)b =
b−1 eb = b−1 b = e. Thus, by the uniqueness of inverses, (ab)−1 = b−1 a−1 .
(c) if ab = e, then b = eb = a−1 ab = a−1 e = a−1 . That is, b = a−1 .
(d) if ac = a, then a−1 ac = a−1 a. Simplifying both sides gives the result.
Note that part (c) means that if we are checking whether, h = g −1 , for some group elements g
and h, we only need to check that gh = e.

The Order of an Element and Cyclic Groups

Let G be a group with identity e. For a ∈ G and n ∈ Z+ , let an = aa · · · a, so a1 = a, a2 = aa,
n times
a3 = aaa and so on. By convention a0 = e. And we can extend this idea for all integers n by
putting a−n = (a−1 )n = (an )−1 . With this notation, it is easy to show the following result.
Lemma 4.7.2 Let G be a group. Then for all a ∈ G and all m, n ∈ Z,
(a) am an = am+n
(b) (am )n = amn

Warning Do not be seduced by this notation into thinking that all the rules of powers of
numbers apply here. In particular, if a, b ∈ G and n ∈ Z, then in general (ab)n = an bn . This
rule only holds when a and b commute.

Note: if G is abelian, we often use the symbol + for the binary operation, and an is usually
written na. In this case Lemma 4.7.2 is written (a) ma + na = (m + n)a; (b) n(ma) = (mn)a.

Let G be a group with identity e and let a ∈ G. Suppose that there is a non-zero integer m
such that am = e. If m < 0, then a−m = (am )−1 = e−1 = e, so we can assume m > 0. Let
n = min{m ∈ Z+ : am = e}.
Then n is called the order of a and is denoted by o(a). If there is no non-zero integer m such
that am = e, then we say that a has infinite order. (This is equivalent to saying there is no
positive integer m such that am = e.)
Example 4.7.3 In (R∗ , ×), −1 has order 2, since (−1)2 = 1. But 3 has infinite order, since
3n = 1 for all n ∈ Z+ .
Example 4.7.4 α = (123)(45) ∈ S5 has order 6, since α2 = (123)(45)(123)(45) = (132),
α3 = (45), α4 = (123), α5 = (132)(45) and α6 = 1.
Example 4.7.5 In (Z6 , ⊕), o(0) = 1, o(1) = 6, o(2) = 3, o(3) = 2, o(4) = 3 and o(5) = 6.
If G is a group with identity e, then o(e) = 1. Note that the word ‘order’ is also used to refer
to the size of a group. So the order of (Zn , ⊕) is n, and (R, +) has infinite order. These two
uses are linked though, because the order of elements in a group is related to the order of the
group. One extreme example of this is the following result.
Proposition 4.7.6 Let G be a group with identity e and let a ∈ G. Then the set
A = {am : m ∈ Z}
is a group (under the same binary operation as G) and the order of A is o(a).
106 Proof & Structure Chapter 4: Relations, Operations, Groups

Proof We need to check the group axioms.

Closure This follows since, by Lemma 4.7.2, am1 am2 = am1 +m2 ∈ A.
Associativity Since A ⊆ G, the associativity of A follows automatically from the associativity
of G.
Identity e = a0 ∈ A.
Inverses a−m ∈ A and is the inverse of am .
Finally, to find the order of A, let am , al ∈ A, with m > l, and suppose am = al . Then
am−l = am a−l = al a−l = a0 = e. Therefore a has order at most m − l. In particular o(a) is
finite. Therefore, if a has infinite order, this means that whenever m = l, am = al . Hence A
has infinitely many elements, so A has infinite order.
Suppose that o(a) = n < ∞. Then for 0 ≤ l < m < n, am−l = e, because m − l < n. Therefore
am = al . So each of e, a, a2 , . . . , an−1 is distinct. We claim that A = {e, a, a2 , . . . , an−1 }. For any
integer m, there exists an integer q and an integer r with 0 ≤ r ≤ n − 1 such that m = qn + r.
So, using Lemma 4.7.2

am = aqn+r = anq ar = (an )q ar = eq ar = ar .

It follows that A = {e, a, a2 , . . . , an−1 }. Hence A has order n. Thus we have shown that A is a
group of order o(a).

It is possible that the group A defined in Proposition 4.7.6 is the whole group G. In this
case, every element of G can be written as an integer power of a. This leads to the following
condition.

Definition A group G is cyclic if there is an element a ∈ G such that G = {am : m ∈ Z}.

We say that a generates G and write G = a . Sometimes we denote a cyclic group with n
elements by Cn .

Example 4.7.7 (Z6 , ⊕) is cyclic, generated by 1 or 5.

Example 4.7.8 (Z, +) = 1 , so Z is an infinite cyclic group.

More generally, if G is a group and a1 , a2 , . . . , as are elements of G we say that G is generated

by a1 , a2 , . . . , as (written G = a1 , a2 , . . . , as ) if every element of G can be expressed as a
finite product of a1 , a2 , . . . , as and their inverses. The elements a1 , a2 , . . . , as of G are called
generators of G. Note that if G is not cyclic, then any generating set of G must contain at
least two elements.

Example 4.7.9 S3 = (12), (13), (23) . This is because every permutation can be written as
a product of transpositions, and (12), (13) and (23) are the only transpositions in S3 . Since S3
is not abelian, Proposition 4.7.10 will prove that S3 is not cyclic.

Exercise 4.20 Can you find any elements other than 1 that generate (Z, +)? Can you find
any two-element generating sets neither of whose elements generates (Z, +) on its own?

Proposition 4.7.10 Let G be a group. If G is cyclic, then G is abelian.

4.7 Properties of Groups Proof & Structure 107

Proof Let G = a . Then for all b, c ∈ G, there exist integers m and n such that b = am
and c = an . Consequently, by Lemma 4.7.2, we have bc = am an = am+n = an+m = an am = cb.
Hence G is abelian.

The converse of Proposition 4.7.10 is false. For example, (Q, +) is an infinite abelian group
which is not cyclic. A finite example is the subset V = {1, (12)(34), (13)(24), (14)(23)} of S4 .
V is called a Klein 4 group and is a finite abelian group which is not cyclic.

Exercise 4.21 Show that V and (Q, +) are not cyclic.

Multiplication Tables
If (G, ) is a finite group, then we can construct its multiplication table. The following are
examples of the multiplication tables of some finite groups.

Example 4.7.11 The table for (Z5 , ⊕) is as follows.

⊕ 0 1 2 3 4
0 0 1 2 3 4
1 1 2 3 4 0
2 2 3 4 0 1
3 3 4 0 1 2
4 4 0 1 2 3

Example 4.7.12 The dihedral group (Dih(6), ◦) (symmetries of an equilateral triangle).α1 is

rotation through 2π/3 anticlockwise about the centre of the triangle, α2 is rotation through
4π/3 anticlockwise. Let β1 be reflection in the axis through the top vertex, β2 reflection in the
axis through the bottom right-hand vertex and β3 reflection in the axis through the bottom
left-hand vertex. Then, for example, to find the entry marked with ∗ in the table, we need to
calculate α2 ◦ β2 , so the result of doing β2 followed by α2 . The diagram below shows this.

1 3 1
β2 α2
3 2 1 2 2 3

Hence α2 ◦ β2 = β1 . Check as many of the other table entries as you can stand!

◦ e α1 α2 β1 β2 β3
e e α1 α2 β1 β2 β3
α1 α1 α2 e β2 β3 β1
α2 α2 e α1 β3 β1∗ β2
β1 β1 β3 β2 e α2 α1
β2 β2 β1 β3 α1 e α2
β3 β3 β2 β1 α2 α1 e
108 Proof & Structure Chapter 4: Relations, Operations, Groups

Exercise 4.22 Find the multiplication

√ √tables of the following:
(a) the subset D = {1, − 21 + 23 i, − 12 − 23 i} of C under ×;
(b) (Z4 , ⊕);
(c) (U8 , ⊗);
1 0 −1 0 i 0 −i 0
(d) the subset , , , of GL2 (C), under matrix
0 1 0 −1 0 −i 0 i
multiplication.
You should convince yourself that the sets in (a) and (d) are groups.

With the exception of associativity each of the group axioms is easy to check from a multipli-
cation table. Let S = {s1 , s2 , . . . , sn }.

Closure S is closed if every entry in the table is an element of S.

Identity element e ∈ S is an identity element of S if the row corresponding to e is

s1 s2 . . . sn and the column corresponding to e is

s1
s2
..
.
sn

Inverses sj ∈ S is the inverse of si ∈ S if si sj = e and sj si = e, that is, if the (i, j)th entry
and the (j, i)th entry of the table are equal to e.

It is also easy to check if (S, ) is abelian. This occurs if for all 1 ≤ i, j ≤ n, si sj = sj si . In

other words, the (i, j)th entry and the (j, i)th entry of the table are equal. In other words, (S, )
is abelian if the table is ‘symmetrical’ about the main diagonal.

Exercise 4.23 Find a criterion to check, from the multiplication table of a finite group G,
whether or not G contains an element of order 2. Use this to decide which of the groups in
Examples 4.7.11, 4.7.12 and Exercise 4.22 have elements of order 2. Do you think there is a
link between the order of the group and whether or not it has an element of order 2?

4.8 Subgroups
We have already encountered the idea of a subgroup. Here is a reminder of the definition.

Definition Let (G, ∗) be a group with identity e. A non-empty subset H of G is a subgroup

of G if (H, ∗) is a group. We write H ≤ G.

Observe that G and {e} are both subgroups of G. In view of this we call H a proper subgroup
if H = G, and a non-trivial subgroup if H = {e}.
4.8 Subgroups Proof & Structure 109

Notes (i) Observe that the binary operation stays the same. So for example (R+ , ×) is not
a subgroup of (R, +). But it is a subgroup of (R∗ , ×).
(ii) By Proposition 4.7.6, if a ∈ G, then A = {am : m ∈ Z} is a subgroup of G.
(iii) If H is a subgroup of G, then the identity of (H, ∗) must be e. To see this, suppose f is
the identity of H. Then f ∗ f = f . Now f has an inverse in G, in other words an element g of
G for which f ∗ g = e(= g ∗ f ). So (f ∗ f ) ∗ g = f ∗ g = e. But f ∗ (f ∗ g) = f ∗ e = f . These
expressions are equal because ∗ is associative. Thus e = f .

Example 4.8.1 (Z, +) ≤ (Q, +) ≤ (R, +) ≤ (C, +) and (Q∗ , ×) ≤ (R∗ , ×) ≤ (C∗ , ×).

Example 4.8.2 The subgroups of (Z6 , ⊕) are {0}, {0, 3}, {0, 2, 4}, Z6 . To show this, we look
at candidate sets H, and use the fact that to be a subgroup, H must be closed under the group
operation, which here is ⊕. Also any subgroup must contain the identity element 0. This means
the only subgroup with one element is the trivial subgroup {0}. Next we move on to consider
sets H that contain at least one non-identity element. If 1 ∈ H, then H = Z6 , as Z6 = 1 . If
2 ∈ H, then the cyclic group generated by 2 must be contained in H. That is {0, 2, 4} ⊆ H.
Now {0, 2, 4} is a cyclic group. So it is a subgroup of H. Considering the next possibility, 3,
we find the subgroup {0, 3}. If 4 ∈ H, we get {0, 2, 4} again. Since Z6 = 5 , if 5 ∈ H, then
H = Z6 .
Next we note that if we add any more elements to any of the non-trivial proper subgroups so
far discovered, they are forced, by the closure property, to be the whole of Z6 . Therefore the
only subgroups of Z6 are {0}, {0, 3}, {0, 2, 4}, Z6 .

Exercise 4.24 Find all subgroups of (Z8 , ⊕).

Example 4.8.3 The dihedral group Dih(2n) has, among others, a subgroup of order n con-
sisting of all the rotations together with the identity, and n subgroups of order 2 consisting of
a reflection together with the identity.

Example 4.8.4 For n ∈ N, nZ = {nz : z ∈ Z} is a subgroup of (Z, +).

Proposition 4.8.5 (The subgroup criterion) Let G be a group and let H be a non-empty
subset of G. Then H is a subgroup of G if and only if H satisfies
(i) for all h, k ∈ H, hk ∈ H;
(ii) for all h ∈ H, h−1 ∈ H.

Proof Suppose that H is a subgroup of G. Then in particular, H is closed (hence (i) holds)
and each element h of H has an inverse in H which, by uniqueness of inverses, is equal to h−1 .
Hence (ii) holds.
For the converse, suppose that H satisfies (i) and (ii). Now the binary operation on H is asso-
ciative because G is a group. The closure and inverse properties are (i) and (ii). So all we need
to show is that the identity element e is contained in H. Since H is non-empty, there is some
element h ∈ H. By (ii), h−1 ∈ H. Then by (i), hh−1 = e ∈ H. Therefore H is a subgroup.

In the special case where H is a finite subset, the situation is even better, as the following
Proposition shows.

Proposition 4.8.6 (The finite subgroup criterion) Let G be a group and let H be a finite
non-empty subset of G. Then H is a subgroup of G if and only if H is closed.
110 Proof & Structure Chapter 4: Relations, Operations, Groups

Proof If H is a subgroup, then clearly H is closed. For the converse, suppose that H is finite,
non-empty and closed. By the subgroup criterion it is sufficient to show that for all h ∈ H,
h−1 ∈ H. So let h ∈ H. Since H is closed, hh = h2 ∈ H. Hence hh2 = h3 ∈ H, and in general
hn ∈ H for all n > 0. But H is finite, and therefore h must have finite order. That is, hn = e
for some n ∈ Z+ . Now hhn−1 = hn = e. Therefore hn−1 = h−1 , and so h−1 ∈ H, as required.
Hence H is a subgroup of G.

Exercise 4.25 Give an example of a group G, and a subset H such that H is closed but not
a subgroup of G. (Hint: In light of the last result, H will have to be infinite.)

The next theorem, Lagrange’s Theorem, is very important in group theory. To prove it we need
to define certain subsets of a group. Let G be a group, and H a subgroup. Then for g ∈ G,
the right H-coset of g in G is the set Hg = {hg : h ∈ H}.

Exercise 4.26 Let G = {1, −1, i, −i}. Show that (G, ×) is a group. Let H = {1, −1}. Show
that H is a subgroup of G, and that H1 = H, H(−1) = H, Hi = {i, −i} and H(−i) = {i, −i}.

We have the following lemma.

Lemma 4.8.7 Let G be a group, H ≤ G, and g1 , g2 ∈ G. Then |Hg1| = |H| and either
Hg1 ∩ Hg2 = ∅ or Hg1 = Hg2 .

Proof We have Hg1 = {hg1 : h ∈ H}. Then every element of this set is different. To see this,
suppose hg1 = h′ g1 . Then hg1 g1−1 = h′ g1 g1−1 , and hence h = h′ . Therefore |Hg1| = |H|.

For the second part, if Hg1 ∩ Hg2 = ∅, there is nothing to prove. So suppose that there exists
x ∈ Hg1 ∩ Hg2 . We must show that Hg1 = Hg2. Since x ∈ Hg1 , there exists h1 ∈ H such that
x = h1 g1 . And since x ∈ Hg2 , there exists h2 ∈ H such that x = h2 g2 . Thus h1 g1 = h2 g2 and
hence g1 = h−1
1 h2 g2 .
Now suppose y ∈ Hg1 . Then y = hg1 for some h ∈ H. So

y = hg1 = hh−1
1 h2 g2 ∈ Hg2 .

Hence Hg1 ⊆ Hg2.

Similarly if z = hg2 ∈ Hg2, then using g2 = h−1 −1
2 h1 g1 we get z = hh2 h1 g1 ∈ Hg1 . Thus
Hg2 ⊆ Hg1.
Combining the two set inclusions we find that Hg1 = Hg2 . So cosets are either disjoint or
equal.

Theorem 4.8.8 Lagrange’s Theorem If G is a finite group and H is a subgroup of G,

then |H| divides |G|.

Proof Suppose |H| = m. Let g ∈ G. Then g ∈ Hg, as g = eg and e ∈ H. Therefore every

element of G lies in an H-coset. But since, by Lemma 4.8.7 distinct cosets do not intersect, g
lies in only one coset. Therefore, every element of G lies in exactly one coset. Let the (distinct)
cosets of H be H1 , H2 , . . . Hk say. Then

G = H1 ∪ H 2 ∪ . . . ∪ Hk .
4.8 Subgroups Proof & Structure 111

Since these are all disjoint, we get

|G| = |H1 | + |H2 | + · · · + |Hk |.

But each coset, by Lemma 4.8.7, has size m = |H|. Therefore

|G| = m + m + · · · + m = km = k|H|.

Hence |G| is a multiple of |H|. That is, |H| divides |G|.

Exercise 4.27 If |G| = 20, list the possible orders of subgroups of G.

Corollary 4.8.9 If G is a group of prime order p, then G is cyclic and any non-identity
element of G generates G.

Proof If H ≤ G, then by Lagrange’s Theorem |H| divides p, and so |H| = 1 or p. Now, if

a ∈ G and a = e, then by Proposition 4.7.6, the set A = a is a subgroup of G and |A| > 1
because e and a are elements of A. Therefore |A| = p. Thus A = G and so G is generated by
a. That is G is cyclic. Moreover the only restriction on the element a was that a = e. Thus
any non-identity element generates G.

Corollary 4.8.10 Let G be a finite group and a ∈ G with o(a) = n. Then n divides |G|.

Proof By Proposition 4.7.6, a ≤ G and so by Lagrange’s Theorem n divides |G|.

Exercise 4.28 Analogously to the definition of a prime number, let’s call a finite group G
‘prime’ if G has exactly two subgroups, the trivial subgroup {eG }, and G itself. Prove that a
finite group G is prime if and only if |G| is prime. (Hint: If x ∈ G has composite order mn,
consider the subgroup generated by xm .)

Exercise 4.29 (hard – too hard for the exam!) Does it make sense to remove the word ‘finite’
from the definition given in Exercise 4.28? In other words, are there any infinite groups with
exactly two subgroups? If not, why not? And if so, can you give an example?

Exercise 4.30 (i) Suppose A is a group of order 2. Explain why A = {e, a} for some a with
o(a) = 2.
(ii) Suppose B is a group of order 3. Explain why B = {e, b, b2 } for some b with o(b) = 3.
(iii) Consider the group (U7 , ⊗). (Remember that U7 = {1, 2, 3, 4, 5, 6} and that the operation
⊗ is multiplication modulo 7.) Give the order of U7 and the possible orders of subgroups.
(iv) List a2 and a3 for all a ∈ U7 (recall that, for example a2 = a ⊗ a). Use this information to
list the order of each element. Now, using (i) and (ii), list all the subgroups of U7 . Is U7 cyclic?
Explain your answer.

The converse of Lagrange’s Theorem is false – a group G does not necessarily have a subgroup
of order n for each n dividing |G|. For an example of this, see Appendix 6.
 112 Proof & Structure Chapter 4: Relations, Operations, Groups

Exam-Style Questions
Section A
Exercise 4.31 Let R be a relation on the set {1, 2, 3, 4, 5, 6, 7} defined by mRn if and only if
mn is divisible by 4.
(a) Write down the adjacency matrix for R on {1, 2, 3, 4, 5, 6, 7}. [2]
(b) Determine whether R is (i) reflexive; (ii) symmetric; (iii) transitive. [3]
Exercise 4.32 The relation ∼ on R is defined by x ∼ y if and only if sin2 x + cos2 y = 1.
(a) Show that ∼ is an equivalence relation on R. [4]
(b) Describe the elements of the equivalence class [π]. [1]
Exercise 4.33 For x, y ∈ R let x ∗ y = xy − x.
(a) Find the idempotent elements of R under ∗. [2]
(b) Show that there is no identity element for ∗ on R. [3]
Exercise 4.34(a) Let G be a group with identity e. Show that if g 2 = e for every g ∈ G, then
G is abelian. [3]
(b) Neither of the following multiplication tables represents a group. In each case, explain why
not.
a b c d a b c d
a a b c d a a b c d
b b a d c b b a d c [2]
c d c a b c c b a d
d c d b a d d c b a

Section B
Exercise 4.35(a) Let V and W be sets of matrices defined as follows.
x y
V = : x, y, z ∈ R and xz = 0 ;
0 z
1 y
W = :y∈R .
0 1
(i) Show that V is a group under matrix multiplication. [4]
(ii) Give an example to show that V is not abelian. [2]
(iii) Find all the elements of order 2 in V . [3]
(iv) State the Subgroup Criterion. [2]
(v) Using the subgroup criterion, or otherwise, show that W is a subgroup of V . [3]
(b) (i) Let G be a finite group and let a ∈ G with o(a) = n. Show that the set A = {am : m ∈ Z}
is a subgroup of G and that |A| = n. [4]
(ii) State Lagrange’s Theorem for finite groups. [1]
(iii) Deduce from (i) and (ii) that the order of an element of a finite group divides the order of
the group. [1]
Note: Exam questions, particularly Section B questions, often combine elements from different
chapters.
4.8 Subgroups Proof & Structure 113

Historical Notes

Niels Henrik Abel (5 August 1802 – 6 April 1829)

Abel was born in Norway in 1802. His father died when Abel was a boy, leaving his mother
heavily in debt, at which point she became an alcoholic. Abel showed promise at mathematics
in school, although he enjoyed it rather more when the school appointed a new maths teacher in
1817. The previous one had got the sack for punishing a boy so violently that he actually died.
When Abel went to university he was so obviously poor that some of his professors supported
him financially through his student years. But his life was never easy. He submitted a brilliant
paper to the Paris Academy in 1826, which the referees (Cauchy and Legendre) did nothing
with. When Abel died in 1829 Cauchy had still not reported on the paper, which wasn’t
published in full until 1841. He couldn’t get a job because his work had not been published,
and he eventually succumbed to tuberculosis at only 26 years old. He died cold, hungry and
bitter. In a cruel twist of fate a job offer for him arrived in the post two days after his death.
Still, at least he got abelian groups named after him, eh?

Joseph-Louis Lagrange (January 25, 1736 – April 10, 1813)

Lagrange was certainly one of the finest mathematicians of his time. He is famous for his
works on celestial mechanics, but also made important contributions to pretty much every
area of mathematics, from analysis to algebra. Lagrange’s theorem as we have stated it would
not have been recognisable to him, because the language of abstract groups had yet to be
developed. Lagrange only proved his theorem in the context of permutations. Suppose you
have a polynomial with n variables, such as 3a − b + 3c + d. Lagrange asked how many different
polynomials can you get by permuting the variables in all possible ways. For example if you
interchange a and b you get 3b − a + 3c + d. He proved that the number of different polynomials
divides n!. How on earth is this related to the Lagrange’s theorem we proved? Well the set of
all possible permutations is just Sn , which has order n! (in our example S4 , where (12) means
‘a → b and b → a’, and so on). The set F of permutations that keep a given polynomial fixed
turns out to be a subgroup. In our example F = {1, (13)}. And every coset of F corresponds
to a different polynomial that can be produced by permuting 3b − a + 3c + d. So our version
of Lagrange’s theorem tells us that |F | × ( number of cosets = |Sn | = n!. In other words our
version of Lagrange’s theorem is a more general version of what Lagrange proved. Incidentally
Lagrange was so organised that he planned all his papers in his head first and then wrote them
out usually without needing to make a single correction.
114 Proof & Structure Chapter 4: Relations, Operations, Groups

Learning Outcomes

After studying this chapter, and the related exercises, you should be able to

• calculate adjacency matrices and directed graphs (digraphs) of binary relations on finite
sets;
• explain what it means for a binary relation on a set to be reflexive, symmetric, or tran-
sitive, and determine whether a given binary relation has these properties;
• define and work with the terms ‘equivalence relation’ and ‘equivalence class’;
• understand the notion of a binary operation, along with the properties of commutativity,
associativity, identity elements, inverses and idempotent elements;
• calculate the multiplication tables for binary operations on finite sets, and use them to
deduce properties of those operations;
• be aware of the basic results for binary operations, such as the uniqueness of identity
elements, where they exist;
• use basic results about binary operations to deduce from the multiplication table that a
given binary operation on a finite set is not associative;
• give the axioms for a group and show that a given set is a group;
• define the following sets and be able to show that they are groups: (Z, +), (Q, +), (R, +),
(C, +), (Zn , ⊕);
• be familiar with the notation S ∗ for a set, and show that the following are groups: (Q∗ , ×),
(R∗ , ×), (C∗ , ×), (Un , ⊗);
• be aware that sets of matrices are a good source of groups, and understand the notation
GLn (F ) and SLn (F ), where F is one of Zn , Q, R, or C;
• find the symmetry groups of simple figures in R2 and R3 , be familiar with the dihedral
group of order 2n, and give a generating set for Dih(2n);
• be aware that the set Sn of permutations is a group;
• understand what is meant by the terms abelian and cyclic and determine whether or not
a given group is abelian or cyclic;
• prove basic properties of groups from the group axioms;
• understand what is meant by the order of an element and of a group, and be able to
calculate these orders;
• show whether or not a subset of a group is a subgroup, and understand the terms proper
subgroup and non-trivial subgroup;
• state and prove the subgroup criterion;
• state and apply Lagrange’s Theorem.