Chapter 1

Groups and Subgroups

1.1 Section 5
Group Homomorphism

A group homomorphism from a group G to a group G′ is a function

ϕ : G → G′

such that for all a, b ∈ G,

ϕ(ab) = ϕ(a)ϕ(b).

That is, a homomorphism preserves the group operation.

If, in addition, ϕ is one-to-one and onto, then ϕ is called an isomorphism, and

the groups G and G′ are said to be isomorphic.

G∼
= G′ .

One-step Subgroup Test

A nonempty subset H of a group is a subgroup if and only if for all x, y ∈ H, we have
xy −1 ∈ H.

1
Theorem 1.1.1. A nonempty subset H of a group G is a subgroup of G if and only if

ab−1 ∈ H for all a, b ∈ H.

Proof. (⇒) Suppose H is a subgroup of G and let a, b ∈ H. Since b ∈ H and H

is closed under inverses, b−1 ∈ H. Because H is closed under the group operation,
ab−1 ∈ H. This proves the condition.
(⇐) Conversely, suppose H is a nonempty subset of G with the property that
ab−1 ∈ H for all a, b ∈ H. We show H is a subgroup using the one–step subgroup test.
Choose some h ∈ H (possible because H is nonempty). Put a = h and b = h in the
hypothesis to get hh−1 = eG ∈ H, so the identity of G lies in H.
Next, let x ∈ H. Using the hypothesis with a = eG and b = x we obtain

eG x−1 = x−1 ∈ H,

so H is closed under taking inverses.

Finally, let x, y ∈ H. Applying the hypothesis with these x, y gives xy −1 ∈ H. But
then multiplying on the right by y (which lies in H) and using closure under inverses
we obtain
(xy −1 )y = x(y −1 y) = xeG = x ∈ H,

which is just a tautology here; instead observe that closure under the product of arbi-
trary elements now follows from closure under inverses and the given condition: since
y −1 ∈ H, applying the hypothesis with a = x and b = y −1 yields

x(y −1 )−1 = xy ∈ H.

Thus H is closed under the group operation.

Having shown eG ∈ H, H is closed under inverses, and H is closed under the group
operation, we conclude that H is a subgroup of G. 2

2
 Question 1: If H is a subgroup of G, then

ϕ[H] = { ϕ(h) | h ∈ H }

is a subgroup of G′ . That is, an isomorphism carries subgroups into subgroups.

Solution:

Solution Let ϕ : G → G′ be an isomorphism and let H be a subgroup of G. We shall

show that ϕ[H] is a subgroup of G′ .

Since H ≤ G, the identity eG ∈ H. Hence, ϕ(eG ) = eG′ ∈ ϕ[H]; thus, ϕ[H] is

nonempty.

Let x1 , x2 ∈ ϕ[H]. Then there exist h1 , h2 ∈ H such that

x1 = ϕ(h1 ) and x2 = ϕ(h2 ).

Because ϕ is a homomorphism,

 −1
x1 x−1
2 = ϕ(h1 ) ϕ(h2 ) = ϕ(h1 )ϕ(h−1 −1
2 ) = ϕ(h1 h2 ).

(Here we have used the fact that ϕ(h2 )−1 = ϕ(h−1

2 ), since

ϕ(h2 )ϕ(h−1 −1
2 ) = ϕ(h2 h2 ) = ϕ(eG ) = eG′ .

Since H is a subgroup of G, we have h1 h−1 −1

2 ∈ H. Therefore, ϕ(h1 h2 ) ∈ ϕ[H], and

thus x1 x−1
2 ∈ ϕ[H].

By the one–step subgroup test, ϕ[H] is a subgroup of G′ .

Hence, an isomorphism carries subgroups into subgroups. 2

Theorem 1.1.2. Let ϕ : G → G′ be an isomorphism. If G is cyclic, then G′ is cyclic.

3
Proof. Suppose G = ⟨a⟩. Since ϕ is onto, every element of G′ has the form ϕ(g) for
some g ∈ G. But g = an for some n ∈ Z, so

ϕ(g) = ϕ(an ) = ϕ(a)n .

Thus every element of G′ lies in ⟨ϕ(a)⟩, and hence

G′ = ⟨ϕ(a)⟩.

Therefore G′ is cyclic, generated by ϕ(a). 2

Theorem 1.1.3. If H and K are subgroups of an abelian group G, then

HK = { hk | h ∈ H, k ∈ K }

is a subgroup of G.

Proof. Let G be abelian and H, K ≤ G. Define

HK = {hk | h ∈ H, k ∈ K}.

Since eG ∈ H and eG ∈ K, we have eG = eG eG ∈ HK. Hence HK is nonempty.

Let x, y ∈ HK. Then there exist h1 , h2 ∈ H and k1 , k2 ∈ K such that

x = h1 k1 and y = h2 k2 .

We compute:
xy −1 = h1 k1 (h2 k2 )−1 = h1 k1 k2−1 h−1
2 .

Since G is abelian, we may rearrange the terms:

xy −1 = h1 h−1 −1
2 k1 k2 .

4
Now h1 h−1 −1
2 ∈ H and k1 k2 ∈ K, because H and K are subgroups. Thus xy
−1 ∈ HK.

By the one–step subgroup test, HK is a subgroup of G. 2

Remark. In the nonabelian case, HK need not be a subgroup of G. A sufficient

condition is that HK = KH, for example, when either H or K is normal in G.

Theorem 1.1.4. A cyclic group that has only one generator has at most two elements.

Proof. Let G = ⟨g⟩ be cyclic.

Case 1: G is finite of order n. The generators of G are precisely the powers g k

with gcd(k, n) = 1. If n ≥ 3 then 1 and n − 1 are distinct integers relatively prime to n,
so both g = g 1 and g n−1 are generators of G. Hence when n ≥ 3 there are at least two
distinct generators. Therefore, if G has only one generator we must have n ≤ 2. Thus
the finite possibilities are |G| = 1 or |G| = 2.

Case 2: G is infinite. If G = ⟨g⟩ is infinite then g and g −1 are distinct and each
generates G. Thus an infinite cyclic group has at least two generators.

Combining the two cases, a cyclic group with only one generator can only be of
order 1 or 2, so it has at most two elements. 2

Theorem 1.1.5. Let G be an abelian group (written multiplicatively) with identity

e. Then the set
H = {x ∈ G | x2 = e}

of all elements of order dividing 2 is a subgroup of G.

Proof. We apply the one–step subgroup test. First H is nonempty since e2 = e, so

e ∈ H.
Let a, b ∈ H. Then a2 = e and b2 = e. Using commutativity of G we compute

(ab)2 = abab = aabb = a2 b2 = e · e = e,

so ab ∈ H. Thus H is closed under the group operation.

5
 Finally, if a ∈ H then a2 = e, hence a−1 = a and so a−1 ∈ H. (Equivalently, (a−1 )2 =
(a2 )−1 = e−1 = e.)
By the one–step subgroup test (nonempty and closed under xy −1 , or equivalently
closed under products and inverses) H is a subgroup of G. 2

Theorem 1.1.6. Let G be a group and let a be a fixed element of G. Then

Ha = {x ∈ G | xa = ax}

is a subgroup of G.

Proof. We use the one-step subgroup test.

(1) Non-emptiness: Since ea = ae = a, the identity element e ∈ Ha . Hence Ha is
non-empty.
(2) Closure under products and inverses: Let x, y ∈ Ha . Then by definition,
xa = ax and ya = ay. We must show that xy −1 ∈ Ha , i.e., that (xy −1 )a = a(xy −1 ).
Compute:
(xy −1 )a = x(y −1 a).

Since y ∈ Ha , we have y −1 a = ay −1 (because multiplying ya = ay on the left by y −1

gives a = y −1 ay, and multiplying again by y −1 on the right gives y −1 a = ay −1 ).
Thus,
(xy −1 )a = x(ay −1 ) = (xa)y −1 .

Since x ∈ Ha , xa = ax, so

(xa)y −1 = (ax)y −1 = a(xy −1 ).

Hence, (xy −1 )a = a(xy −1 ), which means xy −1 ∈ Ha .

Since Ha is nonempty and closed under the operation xy −1 , it follows that Ha is a
subgroup of G. 2

6
Theorem 1.1.7. Let G be a group and let S be a nonempty subset of G. Then

HS = {x ∈ G | xs = sx for all s ∈ S}

is a subgroup of G.

Proof. We apply the one-step subgroup test.

(1) Non-emptiness: For each s ∈ S, we have es = se = s. Hence, e ∈ HS and so
HS is nonempty.
(2) Closure under xy −1 : Let x, y ∈ HS . Then for each s ∈ S, we have xs = sx
and ys = sy. We must show that xy −1 ∈ HS , i.e., (xy −1 )s = s(xy −1 ) for all s ∈ S.
We begin by showing that if y ∈ HS , then y −1 ∈ HS . Since ys = sy, multiplying on
the left by y −1 gives s = y −1 sy. Multiplying on the right by y −1 yields y −1 s = sy −1 ,
so y −1 ∈ HS .
Now for x, y ∈ HS and for all s ∈ S,

(xy −1 )s = x(y −1 s) = x(sy −1 ) = (xs)y −1 = (sx)y −1 = s(xy −1 ).

Thus, (xy −1 )s = s(xy −1 ) for all s ∈ S, so xy −1 ∈ HS .

Since HS is nonempty and closed under xy −1 , it follows that HS is a subgroup of
G. 2

Remark. The subgroup HS is called the centralizer of S in G, and is denoted by

CG (S). When S = {a}, it reduces to CG (a) = {x ∈ G | xa = ax}.

Definition 1.1.8. Let G be a group. The center of G, denoted by Z(G), is defined as

Z(G) = {x ∈ G | xg = gx for all g ∈ G}.

The center Z(G) is a subgroup of G.

7
Theorem 1.1.9. Let G be a group and let

HG = {x ∈ G | xg = gx for all g ∈ G}

be the center of G. Then HG is an abelian subgroup of G.

Proof. We have already observed that HG is a subgroup of G. To prove it is abelian,

let a, b ∈ HG . By definition of the center, a commutes with every element of G, hence
with b. Therefore
ab = ba.

Thus any two elements of HG commute, and so HG is abelian. 2

Assignment

Theorem 1.1.10. Let G be a group and let H and K be subgroups of G. Then

H ∩ K = {x ∈ G | x ∈ H and x ∈ K}

is a subgroup of G.

Theorem 1.1.11. Every cyclic group is abelian.

Proof. Let G be cyclic, so G = ⟨a⟩ for some a ∈ G. Choose arbitrary elements x, y ∈ G.

Then there exist integers m, n such that

x = am and y = an .

Therefore
xy = am an = am+n = an+m = an am = yx.

Thus any two elements of G commute, so G is abelian. 2

Theorem 1.1.12. If a group G has no proper nontrivial subgroups, then G is cyclic.

8
Proof. If G = {e}, then G = ⟨e⟩ is cyclic. So suppose G is nontrivial and choose
a ∈ G with a ̸= e. The subgroup ⟨a⟩ generated by a is nontrivial (it contains a), and
by hypothesis it cannot be proper. Hence ⟨a⟩ = G, so G is cyclic. 2

Theorem 1.1.13. Let G be a cyclic group with n elements and generated by a. Let
b ∈ G and let b = as . Then b generates a cyclic subgroup H of G containing n/d
elements, where d is the greatest common divisor of n and s. Also, ⟨as ⟩ = ⟨at ⟩ if and
only if gcd(s, n) = gcd(t, n).

Proof. That b generates a cyclic subgroup H of G is known from Theorem 5.17. We

need to show only that H has n/d elements. Following the argument of Case II of
Theorem 6.10, we see that H has as many elements as the smallest positive power m
of b that gives the identity. Now b = as , and bm = e if and only if (as )m = e, or if and
only if n divides ms. What is the smallest positive integer m such that n divides ms?
Let d be the greatest common divisor of n and s. Then there exist integers u and v
such that
d = un + vs.

Since d divides both n and s, we may write

1 = u(n/d) + v(s/d),

where both n/d and s/d are integers. This last equation shows that n/d and s/d are
relatively prime, for any integer dividing both of them must also divide 1. We wish to
find the smallest positive m such that

ms m(s/d)
=
n (n/d)

is an integer.
From the division property, we conclude that n/d must divide m, so the smallest
such m is n/d. Thus the order of H is n/d.

9
 Observe that Zn as a model for a cyclic group of order n, we see that if d is a divisor
of n, then the cyclic subgroup ⟨d⟩ of Zn has n/d elements, and contains all the positive
integers m less than n such that gcd(m, n) = d. Thus there is only one subgroup of
Zn of order n/d. From above, this shows at once that if a is a generator of the cyclic
group G, then ⟨as ⟩ = ⟨at ⟩ if and only if gcd(s, n) = gcd(t, n). 2

Theorem 1.1.14. Let r and s be positive integers. The set

H = { nr + ms | n, m ∈ Z }

is a subgroup of (Z, +).

Proof. View Z as an additive group. We apply the one–step subgroup test (for
additive groups: nonempty and closed under subtraction).
First, 0 = 0 · r + 0 · s ∈ H, so H is nonempty.
Next, let x, y ∈ H. Then there exist n1 , m1 , n2 , m2 ∈ Z with

x = n1 r + m1 s, y = n2 r + m2 s.

Consider the difference

x − y = (n1 − n2 )r + (m1 − m2 )s,

and note (n1 − n2 ), (m1 − m2 ) ∈ Z. Hence x − y ∈ H. By the one–step test, H is a

subgroup of Z. 2

Problem: A subgroup of a cyclic group is cyclic. Give counter example for the
converse statement

Theorem 1.1.15. The converse of above statement is false. There exists a group G
such that every proper subgroup of G is cyclic, yet G itself is not cyclic.

10
Proof. Take G = V4 = {e, a, b, c}, the Klein four group, with multiplication given by

a2 = b2 = c2 = e, ab = c, bc = a, ca = b.

The proper subgroups of V4 are exactly

{e, a}, {e, b}, {e, c},

each of which is cyclic of order 2 (generated respectively by a, b, c). Thus every proper
subgroup of V4 is cyclic.
However, V4 is not cyclic: every nonidentity element has order 2, so no single
element generates all four elements of V4 . Hence V4 is not cyclic even though all its
proper subgroups are cyclic. This provides the required counterexample. 2

Theorem 1.1.16. Let p and q be distinct primes. The number of generators of the
cyclic group Zpq is
ϕ(pq) = (p − 1)(q − 1),

where ϕ denotes Euler’s totient function.

Proof. View Zpq as the additive cyclic group of integers modulo pq. An element
k ∈ Zpq has order
pq
ord(k) = .
gcd(k, pq)
Thus k is a generator (i.e. has order pq) if and only if gcd(k, pq) = 1. Hence the
generators of Zpq are exactly the residue classes represented by integers relatively prime
to pq. The number of such residues in {0, 1, . . . , pq − 1} is Euler’s totient ϕ(pq).
Since p and q are distinct primes, an integer 1 ≤ k ≤ pq − 1 fails to be coprime to
pq precisely when it is divisible by p or by q. Counting,

{1 ≤ k ≤ pq : gcd(k, pq) = 1} = pq −(multiples of p)−(multiples of q)+(multiples of pq).

There are q multiples of p in 1, . . . , pq, and p multiples of q, and exactly 1 multiple of

11
pq. Thus
ϕ(pq) = pq − q − p + 1 = (p − 1)(q − 1).

Therefore the number of generators of Zpq is (p − 1)(q − 1). 2

Theorem 1.1.17. Let p be a prime and let r ≥ 1 be an integer. The number of

generators of the cyclic group Zpr is

ϕ(pr ) = pr − p r−1 = p r−1 (p − 1),

where ϕ denotes Euler’s totient function.

Proof. An element k ∈ Zpr (represented by k ∈ {0, 1, . . . , pr − 1}) is a generator of Zpr

precisely when gcd(k, pr ) = 1, because

pr
ord(k) = ,
gcd(k, pr )

so ord(k) = pr iff gcd(k, pr ) = 1.

Hence the number of generators equals the number of integers 1 ≤ k ≤ pr that are
coprime to pr , i.e. ϕ(pr ). For a prime power pr the integers not coprime to pr are
exactly the multiples of p. There are p r−1 such multiples among 1, 2, . . . , pr . Therefore

ϕ(pr ) = pr − p r−1 = p r−1 (p − 1).

This is the desired count of generators. 2

Theorem 1.1.18. Let p be a prime. The group Zp has no proper nontrivial subgroups.

Proof. Since Zp is a finite group of order p, Lagrange’s theorem implies that the
order of any subgroup divides p. Because p is prime the only positive divisors of p are
1 and p. Thus any subgroup of Zp has order either 1 or p. A subgroup of order 1 is
the trivial subgroup {0}, and a subgroup of order p is Zp itself. Therefore there are no
proper nontrivial subgroups of Zp . 2

12