If (G , +) is a finitely generated abelian group, then there is a finite

subset S = {a1 , a2 , · · · , an } ⊆ G such that G = hSi. Then any

Xn
g ∈ G can be written as g = mi ai where mi ∈ Z.
i=1
Example
Let G = { 2mn : m, n ∈ Z} ≤ (Q, +). Suppose that G is finitely
generated. Then
m1 m2 mr
G =h , , · · · , nr i
2n1 2n2 2
for some m1 , . . . , mr , n1 , . . . , nr . We may assume that 2 6 | mi and
n1 ≥ n2 ≥ · · · ≥ nr .
Example
Let G = { 2mn : m, n ∈ Z} ≤ (Q, +). Suppose that G is finitely
generated. Then
m1 m2 mr
G =h , , · · · , nr i
2n1 2n2 2
for some m1 , . . . , mr , n1 , . . . , nr . We may assume that 2 6 | mi and
n1 ≥ n2 ≥ · · · ≥ nr . As G + G = G , we have G = 2n1 G , and
m1 m2 mr
G = 2n1 G = 2n1 h n
, n , · · · , nr i
2 1 2 2 2
= hm1 , m2 2n1 −n2 , · · · , mr 2n1 −nr i ⊆ Z,
which is impossible. Thus G is not finitely generated.
Example
Z × Z, Z × Zn , Zm × Zn are finitely generated abelian groups, but
(Q, +) is not.
Example
Z × Z, Z × Zn , Zm × Zn are finitely generated abelian groups, but
(Q, +) is not.
Proof. The first part of this question is obvious. Let us prove that
the abelian group (Q, +) is not finitely generated.
Example
Z × Z, Z × Zn , Zm × Zn are finitely generated abelian groups, but
(Q, +) is not.
Proof. The first part of this question is obvious. Let us prove that
the abelian group (Q, +) is not finitely generated.
To the contrary, suppose that Q is finitely generated, i.e.,
m1 m2 mr
Q=h , ,··· , i,
n1 n2 nr
where mi , ni ∈ Z \ {0} and n1 ≥ n2 ≥ · · · ≥ nr > 0.
Example
Z × Z, Z × Zn , Zm × Zn are finitely generated abelian groups, but
(Q, +) is not.
Proof. The first part of this question is obvious. Let us prove that
the abelian group (Q, +) is not finitely generated.
To the contrary, suppose that Q is finitely generated, i.e.,
m1 m2 mr
Q=h , ,··· , i,
n1 n2 nr
where mi , ni ∈ Z \ {0} and n1 ≥ n2 ≥ · · · ≥ nr > 0.
Then
m1 m2 mr
Q = n1 n2 · · · nr Q = n1 n2 · · · nr h , ,··· , i
n1 n2 nr
= hm1 n2 · · · nr , n1 m2 n3 · · · nr , · · · , n1 n2 · · · nr −1 mr i ⊆ Z,
which is a contradiction. 
Theorem (Fundamental Theorem of Finitely Generated
Abelian Groups)
Every nontrivial finitely generated abelian group G is isomorphic to
a direct product of nontrivial cyclic groups

Zm1 × Zm2 × · · · × Zmr × Zn ,

where r , n ∈ Z+ , m1 , · · · , mr ∈ 1 + N with mi |mi+1 for

i = 1, 2, · · ·, r − 1. The parameters r , n, m1 , · · · , mr are unique.
Equivalently we can restate Fundamental Theorem of Finitely
Generated Abelian Groups as follows.
Equivalently we can restate Fundamental Theorem of Finitely
Generated Abelian Groups as follows.
Theorem
Every finitely generated abelian group G is isomorphic to a direct
product of cyclic groups

Zp1 r1 × · · · × Zps rs × Zn

where pi ’s are primes and s, n are nonnegative integers. Moreover,

piri , n are uniquely determined by G .
A special case of the above theorem is the following corollary.
Corollary
If G is a finite abelian group, then

G∼
= Zp1 r1 × · · · × Zps rs ,

where pi ’s are primes and ri ’s are positive integers.

A special case of the above theorem is the following corollary.
Corollary
If G is a finite abelian group, then

G∼
= Zp1 r1 × · · · × Zps rs ,

where pi ’s are primes and ri ’s are positive integers.

Corollary
If m is square-free, then every abelian group of order m is cyclic.
Corollary
If m is square-free, then every abelian group of order m is cyclic.
Proof. We know that m = p1 · · · ps , where pi ’s are distinct primes.
Therefore, G ∼
= Zp1 × · · · × Zps ∼
= Zp1 ···ps . 
Example
Find all non-isomorphic abelian groups of order 1089.
Example
Find all non-isomorphic abelian groups of order 1089.
Solution. Note that
1089 = 32 × 112 = 3 × 3 × 112 = 32 × 11 × 11 = 3 × 3 × 11 × 11.
Therefore, all abelian groups of order 1089 are the following:
Example
Find all non-isomorphic abelian groups of order 1089.
Solution. Note that
1089 = 32 × 112 = 3 × 3 × 112 = 32 × 11 × 11 = 3 × 3 × 11 × 11.
Therefore, all abelian groups of order 1089 are the following:
(1). Z 2 × Z 2 ∼
3 11 = Z1089 .
(2). Z3 × Z3 × Z112 ∼
= Z3 × Z3×112 ;
(3). Z32 × Z11 × Z11 ∼
= Z11 × Z32 ×11 ,
(4). Z3 × Z3 × Z11 × Z11 ∼
= Z33 × Z33 .

The partition function p(n) is the number of different partitions
of n. By convention p(0) = 1, p(n) = 0 for n negative. For
example, p(1) = 1.
The partition function p(n) is the number of different partitions
of n. By convention p(0) = 1, p(n) = 0 for n negative. For
example, p(1) = 1. Since 

 1+1+1+1
 
1 + 1 + 1 1 + 1 + 2
( 

 1+1 
3= 1+2 2= and 4 = 2 + 2
 2 
3, 1+3
 




4,

The partition function p(n) is the number of different partitions
of n. By convention p(0) = 1, p(n) = 0 for n negative. For
example, p(1) = 1. Since 

 1+1+1+1
 
1 + 1 + 1 1 + 1 + 2
( 

 1+1 
3= 1+2 2= and 4 = 2 + 2
 2 
3, 1+3
 




4,

we have p(2) = 2, p(3) = 3, p(4) = 5, p(5) = 7, p(6) = 11.
The first few values of the partition function, starting with
p(0) = 1, are:

1, 1, 2, 3, 5, 7, 11, 15, 22, 30, 42, 56, 77, 101, 135, 176, 231, 297, 385, 490,
627, 792, 1002, 1255, 1575, 1958, 2436, 3010, 3718, 4565, 5604, · · ·
The first few values of the partition function, starting with
p(0) = 1, are:

1, 1, 2, 3, 5, 7, 11, 15, 22, 30, 42, 56, 77, 101, 135, 176, 231, 297, 385, 490,
627, 792, 1002, 1255, 1575, 1958, 2436, 3010, 3718, 4565, 5604, · · ·

The exact value of p(n) for larger values of n is very large, for
example
p(100) = 190, 569, 292,
p(1000) = 24, 061, 467, 864, 032, 622, 473, 692, 149, 727, 991.
The first few values of the partition function, starting with
p(0) = 1, are:

1, 1, 2, 3, 5, 7, 11, 15, 22, 30, 42, 56, 77, 101, 135, 176, 231, 297, 385, 490,
627, 792, 1002, 1255, 1575, 1958, 2436, 3010, 3718, 4565, 5604, · · ·

The exact value of p(n) for larger values of n is very large, for
example
p(100) = 190, 569, 292,
p(1000) = 24, 061, 467, 864, 032, 622, 473, 692, 149, 727, 991.

As of September 2017, the largest known prime number that

counts a number of partitions is p(1289844341), with 40000
decimal digits. We have the limit:
√
4n 3p(n)
lim √ = 1.
n→∞
e π 2n/3
Theorem
Let m = p1k1 p2k2 · · · prkr where pi , 1 ≤ i ≤ r are distinct primes and
ki are positive integers. Then there are p(k1 )p(k2 ) · · · p(kr )
non-isomorphic abelian groups of order m, where p(k) is the
partition function.
Example
Consider the expression 24 · 32 . All abelian groups of order 24 are


 Z2 × Z2 × Z2 × Z2

Z2 × Z2 × Z4



Z4 × Z4

Z2 × Z8




Z16 .


All abelian groups of order 32 are

(
Z3 × Z3
Z9 .

So there are 10 abelian groups of order 24 · 32 .

The number of abelian groups of order 24 35 52 7 is

p(4)p(5)p(2)p(1) = 5 · 7 · 2 · 1 = 70.
Properties of Homomorphisms

Recall: Let G , G 0 be groups. A map φ : G → G 0 is a

homomorphism if φ(ab) = φ(a)φ(b), ∀a, b ∈ G . The following are
special types of homomorphisms:
(1) φ is called an epimorphism if φ is onto;
(2) φ is called a monomorphism if φ is 1-1;
(3) φ is called an isomorphism if φ is bijective (i.e., both 1-1 and
onto).
Example
Let G1 , G2 , · · · , Gn be groups. Let G be the direct product
G = G1 × G2 × · · · × Gn . Then πi : G → Gi ,

(g1 , g2 , · · · , gi , · · · , gn ) 7→ gi

is a homomorphism.
Example
Let G1 , G2 , · · · , Gn be groups. Let G be the direct product
G = G1 × G2 × · · · × Gn . Then πi : G → Gi ,

(g1 , g2 , · · · , gi , · · · , gn ) 7→ gi

is a homomorphism.

Example
Let φ : Z → Zm , a 7→ a. Then φ is a an epimorphism.
Example
Let G be a group and h, k ∈ G . Let ϕ : Z × Z → G , where
(m, n) 7→ hm k n . Give a necessary and sufficient condition such
that ϕ is a homomorphism.
Example
Let G be a group and h, k ∈ G . Let ϕ : Z × Z → G , where
(m, n) 7→ hm k n . Give a necessary and sufficient condition such
that ϕ is a homomorphism.
Solution. We need to show that
ϕ((m1 , n1 ) + (m2 , n2 )) = ϕ((m1 , n1 ))ϕ((m2 , n2 )). So, take
(m1 , n1 ) = (0, 1), and (m2 , n2 ) = (1, 0). Then the product

hk = ϕ((1, 1)) = ϕ((0, 1))ϕ((1, 0)) = kh.

Example
Let G be a group and h, k ∈ G . Let ϕ : Z × Z → G , where
(m, n) 7→ hm k n . Give a necessary and sufficient condition such
that ϕ is a homomorphism.
Solution. We need to show that
ϕ((m1 , n1 ) + (m2 , n2 )) = ϕ((m1 , n1 ))ϕ((m2 , n2 )). So, take
(m1 , n1 ) = (0, 1), and (m2 , n2 ) = (1, 0). Then the product

hk = ϕ((1, 1)) = ϕ((0, 1))ϕ((1, 0)) = kh.

Conversely, if hk = kh, then we must have

ϕ((m1 , n1 ) + (m2 , n2 )) = ϕ((m1 + m2 , n1 + n2 ))

= hm1 +m2 k n1 +n2
= hm1 hm2 k n1 k n2
= hm1 k n1 hm2 k n2
= ϕ((m1 , n1 ))ϕ((m2 , n2 )).
Therefore, the necessary and sufficient condition for ϕ to be a
homomorphism is that hk = kh. Note that

ϕ : Z × Z → hh, ki ≤ G .


Therefore, the necessary and sufficient condition for ϕ to be a
homomorphism is that hk = kh. Note that

ϕ : Z × Z → hh, ki ≤ G .


Example
Let φ : G → G 0 be an epimorphism. If G is abelian, then so is G 0 .
Therefore, the necessary and sufficient condition for ϕ to be a
homomorphism is that hk = kh. Note that

ϕ : Z × Z → hh, ki ≤ G .


Example
Let φ : G → G 0 be an epimorphism. If G is abelian, then so is G 0 .
Proof. Since φ is onto, then for all a0 , b 0 ∈ G 0 , there exist a, b ∈ G
such that φ(a) = a0 , φ(b) = b 0 .
Therefore, the necessary and sufficient condition for ϕ to be a
homomorphism is that hk = kh. Note that

ϕ : Z × Z → hh, ki ≤ G .


Example
Let φ : G → G 0 be an epimorphism. If G is abelian, then so is G 0 .
Proof. Since φ is onto, then for all a0 , b 0 ∈ G 0 , there exist a, b ∈ G
such that φ(a) = a0 , φ(b) = b 0 . Therefore,
a0 b 0 = φ(a)φ(b) = φ(ab) = φ(ba) = φ(b)φ(a) = b 0 a0 . So G 0 is
abelian. 
Theorem
Let φ : G → G 0 be a group homomorphism. Then
(1) φ(e) = e 0 .
(2) φ(a−1 ) = φ(a)−1 , ∀a ∈ G .
(3) If H ≤ G , then φ(H) ≤ G 0 .
(4) If K 0 ≤ G 0 , then φ−1 (K 0 ) ≤ G .
Theorem
Let φ : G → G 0 be a group homomorphism. Then
(1) φ(e) = e 0 .
(2) φ(a−1 ) = φ(a)−1 , ∀a ∈ G .
(3) If H ≤ G , then φ(H) ≤ G 0 .
(4) If K 0 ≤ G 0 , then φ−1 (K 0 ) ≤ G .

Proof. (1). Since φ(e)e 0 = φ(ee) = φ(e)φ(e), we see that

φ(e) = e 0 .
Theorem
Let φ : G → G 0 be a group homomorphism. Then
(1) φ(e) = e 0 .
(2) φ(a−1 ) = φ(a)−1 , ∀a ∈ G .
(3) If H ≤ G , then φ(H) ≤ G 0 .
(4) If K 0 ≤ G 0 , then φ−1 (K 0 ) ≤ G .

Proof. (1). Since φ(e)e 0 = φ(ee) = φ(e)φ(e), we see that

φ(e) = e 0 .
(2). We know that

φ(a)φ(a)−1 = φ(e) = φ(aa−1 ) = φ(a)φ(a−1 ).

Thus φ(a−1 ) = φ(a)−1 , ∀a ∈ G .

Theorem
Let φ : G → G 0 be a group homomorphism. Then
(1) φ(e) = e 0 .
(2) φ(a−1 ) = φ(a)−1 , ∀a ∈ G .
(3) If H ≤ G , then φ(H) ≤ G 0 .
(4) If K 0 ≤ G 0 , then φ−1 (K 0 ) ≤ G .

Proof. (1). Since φ(e)e 0 = φ(ee) = φ(e)φ(e), we see that

φ(e) = e 0 .
(2). We know that

φ(a)φ(a)−1 = φ(e) = φ(aa−1 ) = φ(a)φ(a−1 ).

Thus φ(a−1 ) = φ(a)−1 , ∀a ∈ G .

(3), (4) exercises 
Definition
Let φ : G → G 0 be a group homomorphism. We define the kernel
of φ as

ker(φ) = φ−1 (e 0 ) = x ∈ G |φ(x) = e 0 ≤ G .


Definition
Let φ : G → G 0 be a group homomorphism. We define the kernel
of φ as

ker(φ) = φ−1 (e 0 ) = x ∈ G |φ(x) = e 0 ≤ G .



Theorem
Let φ : G → G 0 be a homomorphism. Then φ is 1-1 if and only if
ker(φ) = {e}.
Definition
Let φ : G → G 0 be a group homomorphism. We define the kernel
of φ as

ker(φ) = φ−1 (e 0 ) = x ∈ G |φ(x) = e 0 ≤ G .



Theorem
Let φ : G → G 0 be a homomorphism. Then φ is 1-1 if and only if
ker(φ) = {e}.
Proof.
φ is 1 − 1 ⇐⇒ “φ(x) = φ(y ) =⇒ x = y ”
⇐⇒ “φ(x)φ(y )−1 = e 0 =⇒ x = y ”
⇐⇒ “φ(xy −1 ) = e 0 =⇒ x = y ”
⇐⇒ “φ(x) = e 0 =⇒ x = e ”
⇐⇒ ker(φ) = {e} .
Theorem
Let φ : G → G 0 be a homomorphism and let H = ker φ.
Theorem
Let φ : G → G 0 be a homomorphism and let H = ker φ.
(a) For any a ∈ G we have

φ−1 ({φ(a)}) = {x ∈ G |φ(x) = φ(a)} = aH = Ha.

Theorem
Let φ : G → G 0 be a homomorphism and let H = ker φ.
(a) For any a ∈ G we have

φ−1 ({φ(a)}) = {x ∈ G |φ(x) = φ(a)} = aH = Ha.

(b) For any K ≤ G , φ−1 ({φ(K )}) = KH = HK .

Proof. (a) To show that both sets are the same, let’s show they’re
both subsets of each other.
x ∈ φ−1 ({φ(a)}) ⇐⇒ φ(x) = φ(a)
⇐⇒ φ(a)−1 φ(x) = e 0
⇐⇒ φ(a−1 )φ(x) = e 0
⇐⇒ φ(a−1 x) = e 0
⇐⇒ a−1 x ∈ H
⇐⇒ x ∈ aH.
Proof. (a) To show that both sets are the same, let’s show they’re
both subsets of each other.
x ∈ φ−1 ({φ(a)}) ⇐⇒ φ(x) = φ(a)
⇐⇒ φ(a)−1 φ(x) = e 0
⇐⇒ φ(a−1 )φ(x) = e 0
⇐⇒ φ(a−1 x) = e 0
⇐⇒ a−1 x ∈ H
⇐⇒ x ∈ aH.

So φ−1 ({φ(a)}) = aH.

Proof. (a) To show that both sets are the same, let’s show they’re
both subsets of each other.
x ∈ φ−1 ({φ(a)}) ⇐⇒ φ(x) = φ(a)
⇐⇒ φ(a)−1 φ(x) = e 0
⇐⇒ φ(a−1 )φ(x) = e 0
⇐⇒ φ(a−1 x) = e 0
⇐⇒ a−1 x ∈ H
⇐⇒ x ∈ aH.

So φ−1 ({φ(a)}) = aH.

Similarly φ−1 ({φ(a)}) = Ha.
(b). φ−1 ({φ(K )}) = ∪k∈K φ−1 ({φ(k)}) = ∪k∈K Hk = HK .
(b). φ−1 ({φ(K )}) = ∪k∈K φ−1 ({φ(k)}) = ∪k∈K Hk = HK .
Similarly φ−1 ({φ(K )}) = KH. 
(b). φ−1 ({φ(K )}) = ∪k∈K φ−1 ({φ(k)}) = ∪k∈K Hk = HK .
Similarly φ−1 ({φ(K )}) = KH. 
(b). φ−1 ({φ(K )}) = ∪k∈K φ−1 ({φ(k)}) = ∪k∈K Hk = HK .
Similarly φ−1 ({φ(K )}) = KH. 

Corollary
Let φ : G → G 0 be a homomorphism. Then ker φ  G .
Remark
To show that φ : G → G 0 is an isomorphism, we need to show that
(1) φ is homormophism;
(2) φ is onto;
(3) ker(φ) = {e}.
Theorem
Let (G , ·) be a group and H  G . Then τ : G → G /H defined by
g 7→ gH is a group homomorphism with kernel H.
Theorem
Let (G , ·) be a group and H  G . Then τ : G → G /H defined by
g 7→ gH is a group homomorphism with kernel H.
Proof. It is easy to see that τ is a homomorphism:

τ (g1 g2 ) = g1 g2 H = g1 H · g2 H = τ (g1 )τ (g2 ), ∀g1 , g2 ∈ G .

Theorem
Let (G , ·) be a group and H  G . Then τ : G → G /H defined by
g 7→ gH is a group homomorphism with kernel H.
Proof. It is easy to see that τ is a homomorphism:

τ (g1 g2 ) = g1 g2 H = g1 H · g2 H = τ (g1 )τ (g2 ), ∀g1 , g2 ∈ G .

Then for g ∈ G with τ (g ) = H, we see that

gH = H,

so it follows that g ∈ H. Then ker(τ ) ⊆ H.

Theorem
Let (G , ·) be a group and H  G . Then τ : G → G /H defined by
g 7→ gH is a group homomorphism with kernel H.
Proof. It is easy to see that τ is a homomorphism:

τ (g1 g2 ) = g1 g2 H = g1 H · g2 H = τ (g1 )τ (g2 ), ∀g1 , g2 ∈ G .

Then for g ∈ G with τ (g ) = H, we see that

gH = H,

so it follows that g ∈ H. Then ker(τ ) ⊆ H. Since τ (H) = H, it

follows that H ⊆ ker(τ ). So ker(τ ) = H. 
Theorem (The Fundamental Homomorphism Theorem)
Let G , G 0 be groups, and ϕ : G → G 0 be a group homomorphism
with kernel H.
(a) ϕ(G ) ≤ G 0 ;
(b) µ : G /H → ϕ(G ) defined by µ(gH) = ϕ(g ) is an isomorphism;
(c) ϕ = µ ◦ τ.

用于研究群的结构
Proof. (a) was proved before.
Proof. (a) was proved before.
(b). First show that µ is well-defined: for any g1 H = g2 H ∈ G /H,
we have

µ(g1 H) = ϕ(g1 ) = ϕ(g1 )ϕ(H) = ϕ(g1 H) = ϕ(g2 H) = ϕ(g2 ) = µ(g2 H).

Proof. (a) was proved before.
(b). First show that µ is well-defined: for any g1 H = g2 H ∈ G /H,
we have

µ(g1 H) = ϕ(g1 ) = ϕ(g1 )ϕ(H) = ϕ(g1 H) = ϕ(g2 H) = ϕ(g2 ) = µ(g2 H).

Next, we verify that

µ((g1 H)(g2 H)) = µ(g1 g2 H) = ϕ(g1 g2 ) = ϕ(g1 )ϕ(g2 ) = µ(g1 H)µ(g2 H)

so µ is a homomorphism.
Proof. (a) was proved before.
(b). First show that µ is well-defined: for any g1 H = g2 H ∈ G /H,
we have

µ(g1 H) = ϕ(g1 ) = ϕ(g1 )ϕ(H) = ϕ(g1 H) = ϕ(g2 H) = ϕ(g2 ) = µ(g2 H).

Next, we verify that

µ((g1 H)(g2 H)) = µ(g1 g2 H) = ϕ(g1 g2 ) = ϕ(g1 )ϕ(g2 ) = µ(g1 H)µ(g2 H)

so µ is a homomorphism. For the kernel of µ, suppose that

gH ∈ ker(µ), i.e.,
µ(gH) = ϕ(g ) = e.
So g ∈ H, i.e., gH = H. Then ker(µ) = {H}. Thus µ is injective.
Proof. (a) was proved before.
(b). First show that µ is well-defined: for any g1 H = g2 H ∈ G /H,
we have

µ(g1 H) = ϕ(g1 ) = ϕ(g1 )ϕ(H) = ϕ(g1 H) = ϕ(g2 H) = ϕ(g2 ) = µ(g2 H).

Next, we verify that

µ((g1 H)(g2 H)) = µ(g1 g2 H) = ϕ(g1 g2 ) = ϕ(g1 )ϕ(g2 ) = µ(g1 H)µ(g2 H)

so µ is a homomorphism. For the kernel of µ, suppose that

gH ∈ ker(µ), i.e.,
µ(gH) = ϕ(g ) = e.
So g ∈ H, i.e., gH = H. Then ker(µ) = {H}. Thus µ is injective.
Clearly µ is onto. Thus µ is an isomorphism.
Proof. (a) was proved before.
(b). First show that µ is well-defined: for any g1 H = g2 H ∈ G /H,
we have

µ(g1 H) = ϕ(g1 ) = ϕ(g1 )ϕ(H) = ϕ(g1 H) = ϕ(g2 H) = ϕ(g2 ) = µ(g2 H).

Next, we verify that

µ((g1 H)(g2 H)) = µ(g1 g2 H) = ϕ(g1 g2 ) = ϕ(g1 )ϕ(g2 ) = µ(g1 H)µ(g2 H)

so µ is a homomorphism. For the kernel of µ, suppose that

gH ∈ ker(µ), i.e.,
µ(gH) = ϕ(g ) = e.
So g ∈ H, i.e., gH = H. Then ker(µ) = {H}. Thus µ is injective.
Clearly µ is onto. Thus µ is an isomorphism.
(c). We have (µ ◦ γ)(g ) = µ(γ(g )) = µ(gH) = ϕ(g ). Thus
ϕ = µ ◦ γ. 
Theorem (Second Isomorphism Theorem)
Let (G , ·, e) be a group, H ≤ G and N  G . Then H ∩ N  H and
H/(H ∩ N) ∼ = NH/N.
Proof. It is easy to verify that NH ≤ G . Consider the natural
homomorphism
π : G → G /N, a 7→ Na.
Let π|H be the restriction to H, i.e.,

π|H : H → G /N, a 7→ Na, ∀x ∈ H.

Note that Na = aN. Clearly π|H is a homomorphism with

ker π|H = H ∩ ker π = H ∩ N

and

imπ|H = {Na|a ∈ H} = {Nna|a ∈ H, n ∈ N} = NH/N.

Hence, H ∩ N  H, NH/N  G /N and H/(H ∩ N) ∼

= NH/N. 
Theorem (Third Isomorphism Theorem)
Let (G , ·, e) be a group and let H, K  G with K ≤ H ≤ G . Then
H/K  G /K and G /H ∼ G /K
= H/K
Proof. Define the map

θ : G /K → G /H,
aK → aH, ∀a ∈ G .

If aK = bK , then a−1 b ∈ K ≤ H, i.e., aH = bH. So θ is a

well-defined map. Since

θ((aK )(bK )) = θ(abK ) = abH = (aH)(bH) = θ(aK )·θ(bK ), ∀a, b ∈ G ,

then θ is a homomorphism. Clearly imθ = G /H, and

ker θ = {aK |a ∈ H} = H/K .

So, we have H/K  G /K and G /H ∼

=
G /K
H/K . 
Example
Define θ : (R, +, 0) → (C∗ , ·, 1) with
r 7→ exp(2πir ) = cos(2πr ) + i sin(2πr ). This is a group
homomorphism since

θ(r + s) = exp(2πi(r + s)) = exp(2πir ) exp(2πis) = θ(r )θ(s).

We know that the kernel is Z
R. Clearly the image is the unit

circle (U, ·). We conclude that

R/Z ∼
= (U, ·)
Proposition (Direct product theorem)
Let H1 , H2 ≤ G . Suppose the following are true:
(1) H1 ∩ H2 = {e};
(2) (∀ai ∈ Hi ) a1 a2 = a2 a1 ;
(3) G = H1 H2 .
Then G ∼
= H1 × H2 .
Proof. Define f : H1 × H2 → G by f ((a1 , a2 )) = a1 a2 . Then it is a
homomorphism since

f ((a1 , a2 ) ∗ (b1 , b2 )) = f (a1 b1 , a2 b2 )

= a1 b1 a2 b2
= a1 a2 b1 b2
= f ((a1 , a2 ))f ((b1 , b2 )).
Proof. Define f : H1 × H2 → G by f ((a1 , a2 )) = a1 a2 . Then it is a
homomorphism since

f ((a1 , a2 ) ∗ (b1 , b2 )) = f (a1 b1 , a2 b2 )

= a1 b1 a2 b2
= a1 a2 b1 b2
= f ((a1 , a2 ))f ((b1 , b2 )).

Surjectivity follows from (3). We will show that the kernel is {e}.
Proof. Define f : H1 × H2 → G by f ((a1 , a2 )) = a1 a2 . Then it is a
homomorphism since

f ((a1 , a2 ) ∗ (b1 , b2 )) = f (a1 b1 , a2 b2 )

= a1 b1 a2 b2
= a1 a2 b1 b2
= f ((a1 , a2 ))f ((b1 , b2 )).

Surjectivity follows from (3). We will show that the kernel is {e}.
If f ((a1 , a2 )) = e then a1 a2 = e and a1 = a2−1 . Since a1 ∈ H1 and
a2−1 ∈ H2 , we have a1 = a2−1 ∈ H1 ∩ H2 = {e}.
Proof. Define f : H1 × H2 → G by f ((a1 , a2 )) = a1 a2 . Then it is a
homomorphism since

f ((a1 , a2 ) ∗ (b1 , b2 )) = f (a1 b1 , a2 b2 )

= a1 b1 a2 b2
= a1 a2 b1 b2
= f ((a1 , a2 ))f ((b1 , b2 )).

Surjectivity follows from (3). We will show that the kernel is {e}.
If f ((a1 , a2 )) = e then a1 a2 = e and a1 = a2−1 . Since a1 ∈ H1 and
a2−1 ∈ H2 , we have a1 = a2−1 ∈ H1 ∩ H2 = {e}. Thus a1 = a2 = e
and ker f = {e}. Therefore G ∼ = H 1 × H2 . 
Corollary
If H1 , H2  G such that H1 ∩ H2 = {e} and G = H1 H2 , then
G∼ = H1 × H2 .
Corollary
If H1 , H2  G such that H1 ∩ H2 = {e} and G = H1 H2 , then
G∼ = H1 × H2 .

Example
(1) SLn (R)  GLn (R).
(2) H = {A ∈ GLn (R)| det(A) = ±1}  GLn (R).
Proof. (1). For A ∈ SLn (R), B ∈ GLn (R), we simply need to
show that
BAB −1 ∈ SLn (R).
Proof. (1). For A ∈ SLn (R), B ∈ GLn (R), we simply need to
show that
BAB −1 ∈ SLn (R).
From the definitions of the groups, we know that
det(A) = 1, det(B) 6= 0. Recall the formula

det(BAB −1 ) = det(B) det(A)(det(B))−1 = det(A) = 1.

Proof. (1). For A ∈ SLn (R), B ∈ GLn (R), we simply need to
show that
BAB −1 ∈ SLn (R).
From the definitions of the groups, we know that
det(A) = 1, det(B) 6= 0. Recall the formula

det(BAB −1 ) = det(B) det(A)(det(B))−1 = det(A) = 1.

It follows that SLn (R)  GLn (R) since we know previously that
SLn (R) ≤ GLn (R).
(2). Define a map

ϕ : GLn (R) → (R+ , ·), A 7→ | det(A)|.

So ϕ is an onto homomorphism.
(2). Define a map

ϕ : GLn (R) → (R+ , ·), A 7→ | det(A)|.

So ϕ is an onto homomorphism. We know that

H = ker(ϕ) = ϕ−1 ({1})  GLn (R).


Definition
Let (G , ·, e) be a group. The subgroup

G 0 = hxyx −1 y −1 , ∀x, y ∈ G i

is the commutator subgroup of G , also denoted [G , G ].

Definition
Let (G , ·, e) be a group. The subgroup

G 0 = hxyx −1 y −1 , ∀x, y ∈ G i

is the commutator subgroup of G , also denoted [G , G ].

Theorem
Let G be a group. Then the following statements hold:
(1). G 0  G .
(2). If N  G , then G /N is abelian ⇐⇒ G 0 ⊆ N.
Proof. (1). Denote aba−1 b −1 := [a, b]. Then we must have
(aba−1 b −1 )−1 = [b, a]. Then for x ∈ G , [a, b] ∈ G 0 ,

x[a, b]x −1 = x[a, b]x −1 [a, b]−1 [a, b]

= [x, [a, b]] [a, b] ∈ G 0 .

Thus G 0  G .
Proof. (1). Denote aba−1 b −1 := [a, b]. Then we must have
(aba−1 b −1 )−1 = [b, a]. Then for x ∈ G , [a, b] ∈ G 0 ,

x[a, b]x −1 = x[a, b]x −1 [a, b]−1 [a, b]

= [x, [a, b]] [a, b] ∈ G 0 .

Thus G 0  G .
(2). We have

G /N is abelian ⇐⇒ aNbN = bNaN

⇐⇒ aNbNa−1 Nb −1 N = N
⇐⇒ aba−1 b −1 N = N
⇐⇒ [a, b] ∈ N, ∀a, b ∈ G
⇐⇒ G 0 ⊆ N.
Proof. (1). Denote aba−1 b −1 := [a, b]. Then we must have
(aba−1 b −1 )−1 = [b, a]. Then for x ∈ G , [a, b] ∈ G 0 ,

x[a, b]x −1 = x[a, b]x −1 [a, b]−1 [a, b]

= [x, [a, b]] [a, b] ∈ G 0 .

Thus G 0  G .
(2). We have

G /N is abelian ⇐⇒ aNbN = bNaN

⇐⇒ aNbNa−1 Nb −1 N = N
⇐⇒ aba−1 b −1 N = N
⇐⇒ [a, b] ∈ N, ∀a, b ∈ G
⇐⇒ G 0 ⊆ N.

Corollary
Let G be a group. Then G is abelian if and only if G 0 = {e}.