1. a. H = {1,
a6,
b,
ba6} is a subgroup because
It is not normal in D4 because Ha = {a, a7, ba, ba7} while aH = {a, a7, ba11, ba5}.
c. This is closed because a2b = ba10, a4b = ba8, a6b = ba6; (bak)2 = 1 for each k. Hence H is a subgroup. It is normal, being of index 2.
3. A4 has no subgroup of order 6 [Exercise 34 §2.6] so the only subgroups (normal or not) have order 1, 2, 3 or 4. If
, it contains no 3-cycle and so equals
K. Thus
K A4 being unique of its order. If
, let
H =
(1, 2, 3)
without loss of generality. But (1 4)
−1(1 2 3)(1 4) = (2 3 4) shows
H is not normal. Finally
without loss of generality. But (1 3)
−1(1 2)(1 3) = (2 3) shows
H is not normal.
5. First
aKa−1 is a subgroup, by Theorem 5 §2.3, and
aKa−1 ⊆
aHa−1 ⊆
H because
H G. If
h H, we must show
h(
aKa−1)
h−1 ⊆
aKa−1. We have
h−1Kh =
K because
K H, so
because K G.
7. Let
H = {1,
h}. Given
g G, we have
g−1hg H because
H G. But
g−1hg ≠ 1 because
h ≠ 1, so
g−1hg =
h. Thus
hg =
gh for all
g G; that is
h Z(
G). If
H = {1,
a,
a2} ⊆
D3 = {1,
a,
a2,
b,
ba,
ba2} then
H D3 but
Z(
D3) = {1}.
9. (1, 1)
D, (
g,
g)
−1 = (
g−1,
g−1)
D; (
g,
g)(
g1,
g1) = (
gg1,
gg1)
D. So
D is always a subgroup of
G ×
G. If
G is abelian, then
G ×
G is abelian so every subgroup is normal. Conversely, if
D is normal, let
g,
a G. Then (
a,
g)
−1(
g,
g)(
a,
g)
D, that is (
a−1ga,
g)
D. This means
a−1ga =
g, so
ga =
ag. Hence
G is abelian.
11. Let
H and
K be subgroups of
G with
and
. Then
H ∩
K = {1} by Lagrange's theorem. Moreover
H G because it is unique of its order, and similarly
K G. Hence
G ≅
H ×
K by Corollary 2 of Theorem 6. Since
p and
q are primes,
H and
K are cyclic of relatively prime orders. Hence
H ×
K is cyclic by Exercise 25 §2.4.
13. a. Let
g G,
h H. Write
,
xi X,
. Then
so it suffices to show
xkhx−k H for all
x X,
. But then
xkhx−k = (
xhx−1)
k H because
xhx−1 H by hypothesis. The converse is clear.
15. a. Since
a ∉
K,
Ka ≠
K, so
Ka =
Gn
K (because
). Similarly
b−1 ∉
K gives
Kb−1 =
G −
K. Thus
Ka =
Kb−1, so
ab K.
17. a. If
H =
ad ,
d n, let
n =
md. Since
bak is selfinverse for all
k, we have
2.
19. Assume that
HK ⊆
KH. If
g KH, say
g =
kh, then
g−1 =
h−1k−1 HK ⊆
KH. If
g−1 =
k1h1, then
proving that
KH ⊆
HK. Hence
HK =
KH and Lemma 2 applies.
21. always holds because
is a subgroup containing
H and
K. But
HK is itself a subgroup (Lemma 2) and contains
H and
K, so
23. a. Let
Dn = {1,
a,
a2, . . .,
an−1,
b,
ba,
ba2, . . .,
ban−1} where
o(
a) =
n,
o(
b) = 2 and
aba =
b. If
n = 2
m, let
H = {1,
am}. We have
amb =
ban−m =
bam , so
H ⊆
Z(
Dn). Thus
H Dn. Let
K = {1,
a2,
a4, . . .,
an−2,
b,
ba2, . . .,
ban−2}. We have
o(
a2) =
m,
o(
b) = 2, and
a2ba2 =
a(
aba)
a =
aba =
b. Thus
K ≅
Dm. Moreover,
K Dn because it is of index 2. But then we have
G ≅
H ×
K ≅
C2 ×
Dm by Corollary 2 of Theorem 6 since
24.
a. If H is characteristic in G then aHa−1 = σa(H) ⊆ H where σa is the inner automorphism.
c. Let
G =
a ×
a where
o(
a) = 2, and define
σ :
G →
G by
σ(
x,
y) = (
y,
x). It is easy to verify that
σ is an automorphism of
G and, if
H =
(
a, 1)
= {(1, 1), (
a, 1)}, then
H G, but
σ(
H)
H, so
H is not characteristic in
G.
3. If
a G, let
σa :
G →
G denote the inner automorphism. Then
σa(
K) =
K because
K G, so
σa :
K →
K is an automorphism of
K. Hence
σa(
H) =
H because
H is characteristic in
K, that is
aHa−1 =
H.
g. If
G =
a and
H =
am , let
σ aut G and let
σ(
a) =
an. If
h =
amk H, then
σ(
h) = [
σ(
a)]
mk =
anmk = (
ank)
m H. Thus
σ(
H) ⊆
H. If
G is finite it is easier:
σ(
H) is a subgroup of
G of the same order as
H so (Theorem 7 §2.4)
σ(
H) =
H. It fails if
G is abelian by (c).
i. Clearly
K ⊆ 1
G(
H) =
H. If
τ aut G, then
K ⊆
τ−1σ(
H) for all
σ aut G by the definition of
K, so
τ(
K) ⊆
σ(
H) for all
σ. Thus
τ(
K) ⊆
K. Similarly,
τ−1(
K) ⊆
K, so
K ⊆
τ(
K). Thus
K =
τ(
K) and
K is characteristic. Finally, let
C ⊆
H,
C a characteristic subgroup of
G. If
σ aut G, then
C =
σ(
C) ⊆
σ(
H), so
C ⊆
K by the definition of
K.
25.
a. We have 1
N(
X) because 1
X1
−1 =
X. If
a,
b N(
X), then we have
aXa−1 =
X =
bXb−1. Thus
a−1 N(
X) because
and
ab N(
X) because (
ab)
X(
ab)
−1 =
a[
bXb−1]
a−1 =
a[
X]
a−1 =
X. Hence
N(
X) is a subgroup of
G.
c. Suppose
K is a subgroup of
G and
H K. If
k K, then
k−1Hk =
H by Theorem 3, so
k N(
H). Thus
K ⊆
N(
H).
26.
a. Write
K =
core H = ∩
aaHa−1. Clearly 1
K. If
g,
g1 K, then
gg1 aHa−1 for all
a, so
gg1 K. Also,
g−1 a−1H(
a−1)
−1 for all
a, so
g aHa−1 and
g K. Hence
K is a subgroup. If
g G and
k K then
gkg−1 g[(
g−1a)
H(
g−1a)
−1]
g−1 =
aHa−1 for all
a, as required.
c. core H ∩
K ⊆
H ∩
K ⊆
H and
core H ∩
K G gives
Similarly, core (H ∩ K) ⊆ core K, so core (H ∩ K) ⊆ core H ∩ core K. On the other hand, core H ⊆ H and core K ⊆ K gives core H ∩ core K ⊆ core H ∩ K. Since core H ∩ core K G, this gives core H ∩ core K ⊆ core (H ∩ K).
27.
a. First,
is a subgroup of
G. If
and
X ⊆
N G, then 1
H and the fact that
g,
g1 N implies that
g−1 N and
gg1 N. Thus
, so
is a subgroup. But if
a G and
then we have
aga−1 aNa−1 =
N for all
N G,
X ⊆
N, so
. Thus
for all
a G, so
. Clearly
.
c. so
. If
G =
S3,
H = {
ε,
σ,
σ2} and
K = {
ε,
τ}, then
,
, so
.
28. a. If
c,
c1 C(
X), then (
cc1)
x =
c(
c1x) =
c(
xc1) = (
cx)
c1 =
xcc1, for all
x X, so
cc1 C(
X). Since
cx =
xc for all
x X, it follows that
xc−1 =
c−1x; that is
c−1 C(
X). Finally, 1
C(
X) is clear.