Introduction to Groups

Hong-Jian Lai

August 2000

1. Basic Concepts and Facts

(1.1) A semigroup is an ordered pair (G, ?) where G is a nonempty set and ? is a binary
operation on G satisfying:
(G1) a ? (b ? c) = (a ? b) ? c, ∀a, b, c ∈ G.
A semigroup G is a monoid if it also satisfies:
(G2) G has an element e (sometimes denoted by 1G , called the identity of G), such
that e ? a = a ? e = a, ∀a ∈ G.
A monoid is a group if (G3) below is satisfied.
(G3) For each a ∈ G, there is an a−1 ∈ G such that a ? a−1 = a−1 ? a = e. (a−1 is called
the inverse of a).

(1.1a) Z, Q, R, Zn and vector spaces with addition, and the corresponding multiplicative
groups. Groups obtained from taking products.

(1.1b) Permutations, Sn , symmetric group on n letters, and Dn , the dihedral group. (Dis-
play n = 3, 4).
(1.1c) Let F be a field. For each integer n ≥ 1, let

GLn (F ) = {A| A is an n × n matrix over F with det(A) 6= 0}.

Then GLn (F ) with matrix multiplication forms a group, called the general linear group
of degree n, whose identity is the identity matrix.

(1.2) (Thm 1.2) If G is a monoid, then identity of G is unique. If G is a group, then

(i) c ∈ G and cc = c =⇒ c = e
(ii) ∀a, b, c ∈ G, ab = ac =⇒ b = c and ba = ca =⇒ b = c. (Cancellation Laws).
(iii) ∀a ∈ G, a−1 is unique.
(iv) (a−1 )−1 = a.

1
 (v) (a ? b)−1 = b−1 ? a−1 .
(vi) ∀a, b ∈ G, ax = b and ya = b have unique solutions in G : x = a−1 b and y = ba−1 .
(vii) (Generalized Associative Law, Thm 1.6) For any a1 , a2 , · · · , an ∈ G, the value
of a1 a2 · · · an is independent of how the expression is bracketed.
Proof (1) - (vi) by definitions. For (vii), use induction on n to show every such expression
is equal to a1 (a2 (· · · an ) · · ·). (G1) implies the case when n=3.

(1.3) Let G be a semigroup. The G is a group iff each holds:

(i) ∃e ∈ G such that ∀a ∈ G, ea = a (left identity).
(ii) For each a ∈ G, ∃a−1 ∈ G such that a−1 a = e (left inverse).

(1.3A) Let G be a semigroup. The G is a group for all a, b ∈ G, the equations ax = b

and ya = b have solutions.
PF: Apply (1.2) in (1.3), and apply (1.3) to (1.3A).

(1.4) (Thm 1.5) Let R be an equivalence relation on a monoid G such that a1 Ra2 and
b1 Rb2 imply (a1 b1 )R(a2 b2 ). Then the set G/R of all equivalence classes of G under the
binary operation āb̄ = ab is also a monoid.
(1.4a) Zn = Z/nZ.
(1.4b) Q/Z.

(1.4c) For a group G, an element x ∈ G has order n if xn = 1 and n is the smallest

such positive integer, (n can be finite or infinite). The order of x is |x|.
(1.4d) In Zn , |1̄| = n; in S3 , |(12)| = 2. For any n ∈ Z − {0}, |n| = ∞.

(1.5) A group G is abelian if for any pair a, b ∈ G, ab = ba. (Commutative Law

holds). For abelian groups, we usually use addition to denote the binary operation.

(1.5a) ((Thm 1.7) Generalized Commutative Law) If G is a commutative semigroup,

and if a1 , ..., an ∈ G, then for any permutation i1 , ..., in of 1, 2, ..., n, ai1 ai2 · · · ain = a1 a2 · · · an .
PF: Induction on n ≥ 2.

(1.6) Notation:
(1.6a) (multiplication notation) a1 = a and an = a · an−1 , for n ≥ 2, and a0 = 1G . If g ∈ G
and H ⊆ G, then
gH = {gh|h ∈ H} and Hg = {hg|h ∈ H}.

2
If H, K ⊆ G, then
HK = {hk|h ∈ H ∧ k ∈ K}.

(1.6a) (addition notation) 1 · a = a and na = a(n − 1)a, for n ≥ 2, and 0 · a = 0G . If g ∈ G

and H ⊆ G, then
g + H = {g + h|h ∈ H}.

If H, K ⊆ G, then
H + K = {h + k|h ∈ H ∧ k ∈ K}.

(1.7) (Thm 1.9) Let G be a group. For any a ∈ G and m, n ∈ Z, (an )m = anm and
am an = am+n . (Same result for addition notation).

(1.8) (Direct Products) Let G and H be two groups. Let

G × H = {(g, h) : g ∈ G and h ∈ H}.

Then the operation (g1 , h1 )(g2 , h2 ) = (g1 g2 , h1 h2 ) is a binary operation on G × H. One

can verify that the set G × H with this binary operation will form a group. (When both
operations in G and in H are additive, we often use G ⊕ H for G × H).
(1.8a) If both G and H are abilian, then G × H is also abelian.

2. Homomorphisms and Subgroups

(2.1) Let G and H be groups. A map f : G → H is a homomorphism if for all x, y ∈ G,
f (xy) = f (x)f (y). f is an isomorphism if it is bijective. If f : G 7→ G is a homomorphism
(isomorphism, resp.), then f is also called an endomorphism (automorphism), resp.) of
G.
(2.1a) Linear transformations of vector spaces are examples of homomorphisms; Z2 × Z3 ∼
=
Z6 ; GL2 (Z2 ) ∼
= S3 .
(2.1b) Let f : G 7→ H is a group homomorphism. The kernel of f is ker(f ) = {a ∈ G :
f (a) = e in H}. For A ⊆ G, f (A) = {f (a) : a ∈ A} is the image of A, and we denote
Im(f ) = f (G), called the image of f . If B ⊆ H, then f −1 (B) = {a ∈ G : f (a) ∈ B} is the
inverse image of B.
(2.1c) (Thm 2.3) Let f : G 7→ H be a group homomorphism, and let eG and eH denote the
identities of G and H, respectively. Let 1G and 1H denote the identity maps in G and in
H, respectively. Then
(i) f (eG ) = eH .
(ii) For any a ∈ G, f (a−1 ) = [(f (a)]−1 .

3
 (iii) f is injective (called a monomorphism) iff ker(f ) = {e}.
(iv) f is onto (called an epimorphism) iff Im(f ) = H.
(v) f is an isomorphism iff there exists a homomorphism f −1 : H 7→ G such that
f f −1 = 1H and f −1 f = 1G .
(2.1d) Let G be a group, and let Aut(G) denote the set of all automorphisms of G. Then
Aut(G) with the map composition forms a group itself, called the automorphism group
of G.
(2.1e) AutZ ∼
= Z2 ∼
= Aut(Z6 ).

(2.2) Let G be a group and let H ⊆ G. If H is also a group, then H is a subgroup

of G, denoted by H < G.

(2.3) Let G be a group and let H ⊆ G be a nonempty subset of G. Then TFAE:

(i) H is a subgroup of G.
(ii) ∀a, b ∈ H, ab−1 ∈ H.
(iii) ∀a, b ∈ H, a−1 ∈ H and ab ∈ H.
Proof (i) =⇒ (ii) =⇒ (iii) =⇒ (i). (Show 1 ∈ H first).

(2.4) If H1 , H2 are subgroups of G, then H1 ∩ H2 is also a subgroup of G.

Proof Use (2.3).

(2.5) Let X ⊆ G. Denote

\
< X >= H.
X⊆H∧H≤G

Then < X >≤ G. (< X > is the smallest subgroup of G with X ⊆< X >). Call < X >
the subgroup generated by X, and elements in X are the generators of < X >.
Proof Use (2.3).
(2.5a) Let A ⊆ G, and let

Ā = {a11 a22 · · · ann |n ∈ Z, n ≥ 0, i = ±1 for each i}.

Then < A >= Ā.

Proof: Let H be a subgroup of G such that A ⊂ H. Since elements in a subgroup are
closed under multiplication, Ā ⊆ H. In particular, Ā ⊆< A >. On the other hand, it can
be checked by using (2.3) that Ā itself is a subgroup containing A, and so < A >⊆ Ā.
(2.5b) Let A ⊂ G such that ∀a1 , a2 ∈ A, a1 a2 = a2 a1 . Then < A > is abelian.
(2.5c) Let H ≤ G. If there is an a ∈ H such that H =< a >, then H is a cyclic (sub)group

4
and a is a generator of H.
(2.5d) (Z, +) =< 1 >, (Zn , +) =< 1̄ >. Q and R are not cyclic.

(2.6) Let σ = (123 · · · n) and


 (1n)(2(n − 1)) · · · ( n ( n + 1)) if n is even,
2 2
τ=
 (1n)(2(n − 1)) · · · ( n−1 n+3 )( n+1 ) if n is odd.
2 2 2

Let X = {σ, τ } ⊂ Sn . The D2n =< X > is called the dihedral group of order 2n. The
presentation of D2n is

D2n = {σ, τ |σ n = τ 2 = 1, στ = τ σ −1 }.

(2.6a) Group of rigid motions in R2 and in R3 .

(2.6b) < σ > is a cyclic subgroup of D2n . So is < τ >.

(2.7) Let G be a group and A ⊆ G. The centralizer of A is

CG (A) = {g ∈ G|gag −1 = a, ∀a ∈ A}.

The normalizer of A is
NG (A) = {g ∈ G|gAg −1 = A},

where gAg −1 = {gag −1 |a ∈ A}. The center of G is

Z(G) = CG (G).

(2.8) CG (A) ≤ G and NG (A) ≤ G.

Proof Use (2.3).
(2.8a) ND8 (< σ >) = D8 , CD8 (< σ >) =< σ >, and Z(D8 ) =< τ >.

3. Cyclic Groups
(3.1) Recall that the order of x is |x|. If |x| = n and if xm = e, then n|m.
Proof by Long Division, m = qn + r, where 0 ≤ r < n. (xn = 1) ∧ (xm = e) =⇒ xr = e.
Hence r = 0.

(3.2) (Thm 3.1) Let H be a subgroup of the additive group Z.

(i) Either H =< 0 >, or

5
 (ii) for some m ∈ Z − {0}, H =< m >, and |H| = ∞.

(3.3) (Thm 3.2) Let H =< x > be a cyclic group.

(i) If |H| = n < ∞, then H = {xi |0 ≤ i ≤ n} and the order of x is n. Moreover,
H∼
= Zn .
(ii) If |H| = ∞, then H = {xi |i ∈ Z} and no element of H − {1} has a finite order.
Moreover, H ∼= Z.
Proof Use definition of order.

(3.3A) Any two cyclic group with the same order are isomorphic. (finite or infinite)
Proof They are iso to either Zn of or to Z.

(3.4) Let x ∈ G, and let n 6= 0 be an integer.

(i) If |x| = ∞, then |xn | = ∞.
m
(ii) If |x| = m < ∞, then |xn | = = l.
(m, n)

(iii) If |x| = m < ∞ and d|m, then |xd | = m/d.

Proof: (i) follows by (ii) of (3.2).
(ii) Let y = xn and d = |y|. First y l = e and so by (3.1), d|l. Since 1 = (xn )d = xdn , by
(3.1), m|(dn) =⇒ m/(n, m)|dn/(n, m). Since (m/(n, m), n/(n, m)) = 1, l|d.
(iii) follows by (ii).

(3.5) Let H =< x >.

(i) Assume |x| = ∞. Then H =< xm >⇐⇒ m = ±1.
(ii) Assume |x| = n < ∞. Then H =< xm >⇐⇒ (n, m) = 1.
Proof Use (3.3) and then (3.4).

(3.6) (Structure of Subgroups of a Cyclic Group) Let H =< x >.

(i) If K ≤ H, then either K = {e}, or K =< xd >, where d is the smallest positive
integer such that xd ∈ K.
(ii) If |H| = ∞, then for any distinct nonnegative integer n and m, < xn >6=< xm >.
Furthermore, ∀m ∈ Z, < xm >=< x|m| >. (Thus the number of distinct subgroups of H is
the same as the cardinality of Z.)
(iii) If |H| = n < ∞, then for each positive integer m|n, there is a unique subgroup
< xd >≤ H such that | < xd > | = m, where d = n/m. Furthermore, < xm >=< x(n,m) >.
Proof (i) Assume K 6= {e}. Let P = {(n ∈ Z) ∧ (n > 0)|xn ∈ K}. Let d = min P. Then

6
xd ∈ K. Use long division to show K ∈< xd >.
(ii) < xn >=< xm >, then n|m and m|n and so n = m.
(iii) By (3.4)(3.4)(iii), | < xd > | = n/d = m. Let K ≤ H be such that |K| = m. By (3.6)(i),
K =< xl >, where l is the smallest non negative integer such that xl ∈ K. To prove the
uniqueness, write n = ql + r, with 0 ≤ r < l. As xr = (xn )(xql )−1 e(xql )−1 = (xql )−1 ∈ K,
and by the minimality of l, we have r = 0 and so l|n.
By (3.4)(ii),
n n n
= = |xl | = |K| = m = ,
l n, l d
and so l = [d and K =< xd >.

(3.7) More examples of groups:

(3.7a) (Define direct product G × H) V2 = Z2 × Z2 , a group each of whose proper subgroups
are cyclic, but the group is not cyclic.
(3.7b) Q8 , the quaternion group, is defined by

Q8 = {1, −1, i, −i, j, −j, k, −k},

whose identity is 1 and whose multiplication is defined as follows:

(−1)2 = 1, (−1)a = −a, ∀a ∈ Q8 , b2 = −1, ∀b ∈ Q8 − {1, −1},

and
ij = k, jk = i, ki = j, ji = −k, kj = −i, ik = −j.

Each of the proper subgroup of Q8 is cyclic, but Q8 is not abelian.

4. Alternating Groups (4.1) Recall permutations, cycles, and transpositions (2-cycles).

Every φ ∈ Sn is a product of transpositions.
Proof Every cycle is a product of transpositions:

(i1 i2 · · · ik ) = (i1 ik )(i1 ik−1 ) · · · (i1 i3 )(i1 i2 ).

Every φ ∈ Sn is a product of cycles (called the cycle decomposition of φ).

(4.2) No φ ∈ Sn (n ≥ 2) can be expressed both as a product of an even number of transpo-

sitions and as a product of an odd number of transpositions.
Proof We use e to denote the identity of Sn .
(Step 1) (4.2) holds for φ = e.

7
Suppose that Sn is the set of all permutations on the set {1, 2, · · · , n}, and that e = τk · · · τ1 ,
where each τi is a transposition. Let X = {x : 1 ≤ x ≤ n and x is involved in some of the
τ ’s }, and s = |X|.
Argue by induction on s. If s = 2, then we may assume that the involved letters are 1
and 2, and e = τk · · · τ1 , where each τi = (1, 2). Since e = (1, 2)(1, 2), k must be even.
Assume that s ≥ 3 and that (Step 1) holds for smaller values of s. Suppose that e =
τk · · · τ1 , where each τi is a transposition, and where the involved letters are in {1, 2, . . . , s}.
We further argue by induction on k. (Step 1) holds trivially if k = 2, and so we assume
futher that (Step 1) holds for smaller values of k.
Pick m ∈ X. Let τj be the 1st transposition (from R to L) that contains m. Then
τj+1 τj must be one in the left side of

(x, m)(x, m) = e
(m, y)(m, x) = (m, x)(x, y)
(y, z)(m, x) = (m, x)(y, z)
(x, y)(x, m) = (m, y)(x, y)

Hence the substitution of the left by the right either reduces the number of transpositions
by 2; whence by induction on k, (Step 1) holds; or moves the 1st transposition containing m
to the left by one step. Repeat this process (assuming that k remains unchanged) until the
first τj containing m is τk−1 . Then τt auk−1 must be one of the four cases listed above. In this
case, only the case τk = τk−1 = (x, m) will occur, as otherwise, after the process of pushing
m to the left, the right most transposition of a factoring of e is the only transpotition in
the factorization of e contains the element m, and so m must be moved, contrary to the
fact that e ∈ Sn is the identity permutation.
Therefore, such a process can eliminate the element m, without introducing any new
elements involved in the factorization, and without changing the parity of k. Now X
becomes |X| − 1, and so by induction on |X|, (Step 1) holds also for all values of k.
(Step 2) General Case: Suppose φ ∈ Sn has two factorizations:

φ = τ1 τ2 · · · τr = τ10 τ20 · · · τt0 ,

where τi ’s and τj0 ’s are transpositions. Then φ−1 = τr−1 · · · τ1−1 and so e = φφ−1 =
τr−1 · · · τ1−1 τ10 τ20 · · · τt0 . Hence r + t must be even, and so r and t must have the same parity.
(4.3) (Even and Odd Permutations) A permutation in Sn is even (or odd) if it can be
expressed as a product of an even (or odd) number of transpositions. The set of all even
permutations in Sn is denoted by An . An is a subgroup of Sn , called the Alternating

8
Group of degree n.
Proof Use (2.3) to show An ≤ Sn .

9
5. Cosets and Counting

(5.1) (Thm 4.2) Let G be agroup and let H < G. For any a, b ∈ G, define a ≡l b (mod H)
iff a−1 b ∈ H (a ≡r b (mod H) iff ab−1 ∈ H, resp.). Then both ≡l and ≡r are equivalence
relations.

(5.2) (Them 4.2) Each equivalence class of ≡l has the form gH, where g ∈ G, and is
called a left coset of H in G. Each equivalence class of ≡r has the form Hg, where g ∈ G,
and is called a right coset of H in G. Any element in a coset if a representative of the
coset. (Every statement below about left cosets can also have a right coset version.)
PF: Show that a and b are in the same class iff aH = bH.
(5.3) (Thm 4.2) ∀g ∈ G, |gH| = |H| = |Hg|.
PF: define a bijection.

(5.4) Let H < G. The index of H in Gdenoted [G : H], is the cardinal number of the set
of distinct left cosets of H in G.

(5.5) If K < H < G, then [G : K] = [G : H][H : K].

PF: Use (5.2). Show that K has [G : H][H : K] cosets in G.

(5.6) (Cor. 4.6: Lagrange) If H < G, then |G| = [G : H]|H|.

(5.7) Let H and K be finite subgroup of G, then |HK| = |H||K|/|H ∩ K|.

10
6. Normality, Quotients and Homomorphisms

(6.1) Let φ : G −→ H be a group homomorphism.

(i) φ(1G ) = 1H .
(ii) φ(g −1 ) = (φ(g))−1 , ∀g ∈ G.
(iii) φ(g n ) = (φ(g))n , ∀n ∈ Z.
(iv) The kernel of φ, kerφ = {g ∈ G|φ(g) = 1H } ≤ G.
(v) The image of G under φ, imφ = {h ∈ H|φ(g) = h, for some g ∈ G} ≤ H.
Proof (i): Use φ(1G 1G ) and Cancellation Law.
(ii): Use uniqueness of inverse.
(iii): Induction on n for n > 0, and use (ii) for negative n’s.
(iv) and (v): (Check ab−1 ∈ kerφ).

(6.2) For a map φ : X −→ Y and for each y ∈ Y , the subset φ−1 (y) = {x ∈ X|φ(x) = y}
is called a fiber of φ. Given a group homomorphism φ : G −→ H with K = kerφ, G/K
denotes the set of all fibers of φ. Define a binary operation ? on G/K by

φ−1 (a) ? φ−1 (b) = φ−1 (ab).

Then
(i) ? is well defined. (φ−1 (ab) is independent of the choices of a and b).
(ii) (G/K, ?) is a group, called the quotient group of factor group. The identity of
G/K is K and the inverse of gK is g −1 K.
(iii) φ−1 (a) = aK = {ak|k ∈ K} = Ka = {ka|k ∈ K}.
Proof: (i) Suppose that a0 ∈ φ−1 (a) and b0 ∈ φ−1 (b). The φ(a0 ) = φ(a) and φ(b0 ) = φ(b).
Thus φ(a0 b0 ) = φ(a0 )φ(b0 ) = φ(a)φ(b) = φ(ab).
(ii) Verify the group axioms. The inverse and the identity conclusions follow from the defi-
nition of the binary operation and the uniqueness of identity and inverse.
(iii) Since φ(ak) = a, aK ⊆ φ−1 (a). ∀x ∈ φ−1 (a), we can write x = ay (y = a−1 x). Thus
φ(a) = φ(x) = φ(a)φ(y) and so y ∈ K.

(6.3) For any N ≤ G and g ∈ G, gN and N g are called the left coset and the right
coset of N in G. Any element in a coset if a representative of the coset. (Every theorem
below about left cosets can also have a right coset version.) If G is finite, then
(i) ∀g ∈ G, |gN | = |N |, and
(ii) G is the disjoint union of distinct left (or right) cosets of N . (Valid even when
|G| = ∞.)
(iii) If φ : G → H is a homomorphism with ker(φ) = K, then every fiber of φ has the

11
same cardinality.
(iv) A homomorphism φ is injective iff ker(φ) = {1}.
Proof (i) It suffices to show that if n1 6= n2 and n1 , n2 ∈ N , then gn1 6= gn2 , which is
assured by Cancellation Laws.
(ii) Since G = {g ∈ G} ⊆ ∪g∈G gN , it suffices to show that if g1 N 6= g2 N , then g1 N ∩ g2 N =
∅. In fact, if g1 n1 = g2 n2 for some n1 , n2 ∈ N , then g1 = g2 n2 n−1
1 ∈ g2 N , and so g1 N ⊆ g2 N .
Similarly, g2 N ⊆ g1 N .
(iii) follows from (i) and (iv) follows from (iii).

(6.4) (6.1) can be restated in terms of left and right cosets.

Let G be a group and let K ≤ G be the kernel of some homomorphism from G. Then
the set of all left (or all right) cosets of K with the operation defined by

uK ◦ vK = (uv)K (or Ku ◦ Kv = K(uv))

is a group, denoted by G/K. The operation ◦ is well defined (independent of the choices of
the representatives).
(6.4a) Examples: φ : Z → nZ, for any fixed n ≥ 1 and n ∈ Z. Projections in R2 .
φ : S3 → Z3 .

(6.5) (This is another way to state (5.2)) Let N ≤ G. Then uN = vN ⇐⇒ u−1 v ∈ N .

Proof uN = vN =⇒ u ∈ vN =⇒ u−1 v ∈ N =⇒ v ∈ uN =⇒ uN = vN .

(6.6) (Thm 5.1 and Thm 5.5) Let N ≤ G. TFAE:

(i) The operation ◦ on the left cosets of N by

uN ◦ vN = (uv)N

is well defined.
(ii) ∀g ∈ G, and ∀n ∈ N , gng −1 ∈ N .
(iii) ∀g ∈ G, gN g −1 ⊆ N .
(iv) ∀g ∈ G, gN = N g.
(v) NG (N ) = G or equivalently ∀g ∈ G, gN g −1 = N .
(vi) N is the kernel of some homomorphism from G.
Proof (i) =⇒ (ii). Suppose that ◦ is well defined. ∀g ∈ G and ∀n ∈ N , (1g −1 )N = (ng −1 )N ,
and so by (6.5), gng −1 ∈ N .
(ii) =⇒ (i). Suppose that u0 ∈ uN and v 0 ∈ vN . Want to show (u0 v 0 )N = (uv)N . Since
u0 = un and v 0 = vn0 , for some n, n0 ∈ N , u0 v 0 = unvn0 = uv(v −1 nv)n0 = uvn00 ∈ (uv)N ,

12
where n00 = (v −1 nv)n0 .
(ii) ⇐⇒ (iii). Definition.
(iii) =⇒ (iv). By (iii), we have gN g −1 ⊆ N , and so gN ⊆ N g. Replace g by g −1 to get
N g ⊆ gN .
(iv) ⇐⇒ (v). NG (N ) = {g ∈ G|gN g −1 = N } = G.
(vi) =⇒ (i). (i) of (6.2).
(v) =⇒ (vi). Let P denote all the left cosets of N in G. By (i), (P, ◦) is a group. Define a
map π : G → P by π(g) = gN, ∀g ∈ G. Then π(gg 0 ) = (gg 0 )N = g(g 0 N g 0−1 )gN = gN g 0 N =
π(g)π(g 0 ), and so a homomorphism. The kernel of π, by (ii) of (6.2), is

ker(π) = {g ∈ G|φ(g) = N } = {g ∈ G|gN = 1N } = by (6.5) {g ∈ G|g ∈ N } = N.

(6.7) A subgroup N satisfying any one properties of (4.6) is called a normal subgroup of
G. Denote this fact by N  G. The homomorphism π in (vi) of (6.6) is called the natural
projection of G onto G/N .

(6.8) Let φ : G → H be a homomorphism.

(i) if H 0 ≤ H, then φ−1 (H 0 ) ≤ G.
(ii) If G0 ≤ G, then φ(G0 ) ≤ H.
Proof (i) Use (2.3) and (ii) of (4.1). (ii) of (4.8) = (v) of (4.1).

(6.9) For any H̄ ≤ G/N , N ≤ π −1 (H̄) ≤ G.

Proof It suffices to show that π −1 (H̄) ≤ G. Use (i) of (6.8).

(6.10) (Thm 5.6) Let f : G 7→ H be a group homomorphism and N  G such that

N < ker(f ). Then there exists a unique homomorphism f¯ : G/N 7→ H such that
(i) f¯(aN ) = f (a), ∀a ∈ G.
(ii) Im(f ) = Im(f¯) and ker(f¯) = ker(f )/N .
Moreover, (First Isomorphisn Theorem) f¯ is an isomorphism iff f is an epimorphism
and N = ker(f ).

(6.11) (Second Isomorphisn Theorem) If K  G and N  G, then K/(N ∩ K) ∼

= N K/N .
PF: Justfy N  HK. Find a homomorphism f : K 7→ HK/N with ker(f ) = (N ∩ K).

13
(6.12) (Third Isomorphisn Theorem) If KG and N G and if K < H then H/KG/K
and (G/K)/(H/K) ∼ = G/H.
PF: Justfy H/K  G/K. Find a homomorphism f : G/K 7→ G/H with ker(f ) = H/K.

(6.13) (Thm 5.11) Let f : G 7→ H be an epimorphism of groups. Then the assignement

K 7→ f (K) is
(i) a bijection between the set Sf (G) = {K ≤ G : ker(f ) ≤ K ≤ G} and the set
S(H) = {N ≤ H}; and
(ii) a bijection between the set Sf (G) = {K  G : ker(f ) ≤ K  G} and the set
S(H) = {N  H}.

(6.14) Z(G)  G.

(6.14a) If G/Z(G) is cyclic, then G is abelian.

(6.15) A group G is simple if |G| > 1 and if H ∈ {< 1 >, G} whenever H  G.

(6.15a) The only simple abelian groups are Zp , for prime p’s.
Proof (5.1).

(6.16) An example: being normal is not a transitive relation. Let G = D8 =<

r, s|r4 = 1, s2 = 1, rs = sr−1 >. Let H = {1, r2 , s, sr2 }, and K = {1, s}. Then since
|G : H| = 2, and |H : K| = 2, both K  H and H  G. However, rsr−1 = r2 s 6∈ K and so
K 6 G.
7. Some Counting Lemmas
(7.1) If |G| < ∞, then ∀g ∈ G, |g| divides |G|.
Proof H =< g >.

(7.2) If |G| = p is a prime, then G ∼

= Zp .
Proof Pick g ∈ G − {1} and set H =< g >.

(7.3) If |G : H| = 2, then H  G.
Proof By (5.6), G = H ∪ gH = H ∪ Hg, for any g 6∈ H. Therefore, ∀g ∈ G − H, g −1 ∈ G.
Hence gH = Hg and so by (iv) of (6.6), H  G.

14
(7.4) If H ≤ G and K ≤ G with max{|H|, |K|} < ∞, then

|H||K|
|HK| = .
|H ∩ K|

Proof Note that

HK = ∪h∈H hK = ∪m
i=1 hi K,

where h1 K, h2 K, · · · , hm K are the distinct cosets of the form hK with h ∈ H. (How many
distinct left cosets?) Note that for h, h0 ∈ H, by (4.5)

hK = h0 K ⇐⇒ h−1 h0 ∈ H ∩ K ⇐⇒ h(H ∩ K) = h0 (H ∩ K).

Hence m = number of distinct left cosets of (H ∩K) in H, and so by (5.6), m = |H|/|H ∩K|.
Then use (i) of (4.3).

(7.5) Suppose K ≤ G and H ≤ G. Then HK ≤ G ⇐⇒ HK = KH.

Proof Assume HK ≤ G. Then H, K ≤ HK and so KH ⊆ HK. To see the reverse
containment, by HK ≤ G, (hk)−1 = h0 k 0 ∈ HK, and so

hk = (h0 k 0 )−1 = k 0−1 h0−1 ∈ KH =⇒ HK ⊆ KH.

Conversely, assume HK = KH. Use (2.3).

(7.6) If H ≤ G, K ≤ G, and if H ≤ NG (K), then HK ≤ G.

Proof H ≤ NG (K) =⇒ HK = KH. Use (7.5).
(7.7) If H ≤ G, K ≤ G, and if K  G, then HK ≤ G.
Proof K  G implies that H ≤ NG (K).

15