0% found this document useful (0 votes)
50 views4 pages

2 4

Uploaded by

23Square
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
50 views4 pages

2 4

Uploaded by

23Square
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Hungerford: Algebra

II.4. The Action of a Group on a Set

1. Let G be a group and A a normal subgroup of G.


(a) Show that G/A operates on A by conjugation.
(b) Let CG (A) denote the centralizer of A in G. Show that G/CG (A) is isomorphic to a subgroup
of Aut(A).
(c) Suppose that A is abelian. Obtain a homomorphism from G/A 7→ Aut(A).
Proof: (a) Let Ω = A and G acts on Ω by conjugation: ∀g ∈ G and ∀x ∈ Ω, g(x) = gxg −1 . Since
Ω = A  G, g(x) ∈ Ω. As eG (x) = x and g2 g1 (x) = g2 (g1 xg1−1 ) = g2 (g1 xg1−1 )g −1 = (g2 g1 )(x), this
defines A as a G-set.
(b) Moreover, each g : A 7→ A defined by g(x) = gxg −1 is a bijection and a homomorphism (see
(c) below). Therefore, we have a map φ : G 7→ Aut(A) by φ(g)(x) = g(x), for any x ∈ Ω = A. As
A is a G-set, this map φ is a group homomorphism. Note that Ker(φ) = CG (A), and so by the first
isomorphism theorem, G/CG (A) is isomorphic to a subgroup of Aut(A).
(c) If A is abelian, then A ⊆ CG (A). By the 3rd isomorphism theorem, G/A = ∼ (G/CG (A))/(CG (A)/A).
This suggests the following map G/A 7→ Aut(A): ∀g ∈ G/A, and ∀x ∈ A, f (g)(x) = gxg −1 . To see
that this map is well defined, we pick g, g 0 ∈ g ∈ G/A. Then, there is an a ∈ A such that g 0 = ga.
As A is abelian, axa−1 = x. it follows that

f (g 0 )(x) = (g 0 )x(g 0 )−1 = (ga)x(a−1 g −1 ) = gxg −1 = f (g)(x),

and so this map f is well defined. As ∀x, x0 ∈ A, and ∀g ∈ G/A, f (g)(xx0 ) = g(xx0 )g −1 =
gxg −1 gx0 g −1 = f (g)(x)f (g)(x0 ), f (g) is a homomorphism from A into A. To see that f (g is an
epimorphism, we pick x ∈ A. Then x0 = g −1 (x) ∈ A. Thus x = f (e)(x) = f (g)(x0 ), and so f (g) is
onto. The same argument also indicates that f (g) is one to one. Hence f (g) is in aut(A).
2. If H, K are subgroups of G such that H  K, show that K ≤ NG (H).
Proof: Note that NG (H) = {g ∈ G : gH = Hg}. Since H  K, ∀k ∈ K, kH = Hk, and so
k ∈ NG (H).

3. If a group G contains an element a having exactly two conjugates, then G has a proper normal
subgroup N 6= {e}.
Proof: Let G acts on Ω = G by conjugation. Then the orbit aG has exactly two elements. Therefore
[G : Ga ] = |aG | = 2, and so N = Ga  G. If N = {e}, then |G| = 2. It follows that G ∼ = Z2 , and
so G is abelian. But if G is abelian, every conjugate class has exactly one element, contrary to the
assumption that G has an element a with exactly two conjugates.

4. Let H be a subgroup of G. The centralizer of H in G is

CG (H) = {g ∈ G : ∀h ∈ H, gh = hg}.

Show that CG (H) ≤ NG (H).

1
Proof: Recall that NG (H) = {g ∈ G : gH = Hg}. Thus ∀c ∈ CG (H), and ∀h ∈ H, ch = hc, and so
cH = Hc. Thus c ∈ NG (H).

5. Let H be a subgroup of G. The factor group NG (H)/CG (N ) is isomorphic to a subgroup of


Aut(H).
Proof: Let NG (H) act on H by conjugation. As shown in the proof of Exercise 1 in this section,
this action induces a homomorphism from NG (H) 7→ Aut(H) with kernel CG (H). By the first
isomorphism theorem, NG (H)/CG (N ) is isomorphic to a subgroup of Aut(H).

7. Let G be a group, and let In(G) be the set of all inner automorphisms of G. Show that
In(G)  Aut(G).
Proof: Let φ ∈ Aut(G). For each ginG, let σg ∈ In(G) be such that ∀x ∈ G, σg (x) = gxg −1 . Then
∀x ∈ G,
φ(σg )φ−1 (x) = φ(gφ−1 (x)g −1 ) = φ(g)x(φ(g))−1 = σφ(g) (x).

Thus φσφ−1 ∈ In(G).

8. Exhibit an automorphism of Z6 that is not an inner automorphism.


Proof: As Z6 is abelian, and as if G is abelian, then In(G) = {e}, it suffices to find a φ ∈
Aut(Z6 ) − {e}. Write Z6 = {0, 1, 2, 3, 4, 5}. Define φ(i) = 5i. Then φ ∈ Aut(Z6 ) − {e}.

9. If G/C(G) is cyclic, then G is abelian.


Proof: It suffices to show that G = C(G). Suppose that G − C(G) 6= ∅. Since G/C(G) is cyclic,
there exists an element g ∈ G − C(G) such that G/C(G) = hgi and |g| > 1. We shall show that
g ∈ C(G) also, contrary to the assumption that g ∈ G − C(G).
Arbitrarily pick g 0 ∈ G − C(G). Then g 0 ∈ G/C(G). As G/C(G) = hgi, there exists an integer k
such that g 0 = g k . Therefore, for some element c ∈ C(G), we have g 0 = cg k . It follows by c ∈ C(G)
that gg 0 = gcg k = cg k g = g 0 g, and so g ∈ C(G).

11. Let G be a group containing an element a not of order 1 or 2. Show that G has a nonidentity
automorphism.
Proof: Consider the automorphism φ(x) = x−1 . We can verify that this is an automorphism
(Exercise I-2.2). As a is not of order 1 or 2, φ(a) 6= a, and so φ is not the identity map.

12. Any finite group is isomorphic to a subgroup of An .


Proof: Let G be a group with |G| = m < ∞. By Cayley’s Theorem, G is isomorphic to a subgroup
of Sm . Therefore, it suffices to show that Sm is isomorphic to a subgroup of An , where n = 2m.
This is the same to construct a monomorphism from Sm into An .
To construct a monomorphism from Sm into An , we assume that Sm is the symmetric group
on {1, 2, ..., m}. Let {10 , 20 , ..., m0 } be a set disjoint from {1, 2, ..., m}, and Sm
0
be the symmetric
0 0 0 0 0
group on the set {1 , 2 , ..., m }, and let σ 7→ σ denote the isomorphism from Sm onto Sm . Define
0 0 0
Φ : Sm 7→ An as follows: for any σ ∈ Sm , Φ(σ) = σσ (For example, Φ(1, 2) = (1, 2)(1 , 2 )).

2
For any τ1 , τ2 ∈ Sm , Φ(τ1 τ2 ) = τ1 τ2 τ10 τ20 . Since {1, 2, ..., m} and {10 , 20 , ..., m0 } are disjoint, τ2 and
τ10 are disjoint and so Φ(τ1 τ2 ) = τ1 τ2 τ10 τ20 = τ1 τ10 τ2 τ20 = Φ(τ1 )Φ(τ2 ). Hence Φ is a homomorphism.
Let e, e0 denote the identity of Sm and An , respectively. It remains to show that Ker(Φ) = {e}.
Suppose that for some σ ∈ Sm , Φ(σ) = e0 . As Φ(σ) = σσ 0 , and as σ and σ 0 are on disjoint sets of
elements, we must have σ = e, which proves that Φ is a monomorphism.

13. If H < G and H ≤ G such that [G : H] < ∞, then G has a normal subgroup N with
[G : N ] < ∞.
Proof: Since [G : NG (H)][(NG (H) : H] = [G : H] < ∞, [G : NG (H)] = m < ∞. Thus there are m
conjugates of H: H1 , · · · , Hm . Use the fact that

if A < G and B < G, then [A : A ∩ B] ≤ [G : B]

to get
[G : ∩m m m
i=1 Hi ] ≤ Πi=1 [G : Hi ] = [G : H] < ∞.

−1
Note that N = ∩m i=1 Hi = ∩x∈G xHx  G.
Remark: One can also obtain N by letting G act on the set of all left cosets of H in G. Then the
kernel of the action if N .

14. If |G| = pn, with p > n, p prime, and H is a subgroup of order p, then H  G.
Proof: Let Ω be the set of left cosets of H in G. Let G acts on Ω by left multiplication, and let
N = Ker(G/Ω) be the kernel of the action. Then N  G.
Recall that N = {n ∈ G : ∀aH ∈ Ω, (ga)H = aH}. Therefore, ∀n ∈ N , nH = n(eH) = eH = H,
and so N ≤ H. If N = H, then H = N  G, and the proof is completed.
Therefore, we assume that N 6= H, and so |N | < |H|. By Lagrange Theorem, |N | is a factor
of |H| = p. Since p is a prime, we must have N = {e}. Therefore, G is isomorphic to a subgroup
of Sn0 , where n0 = |Ω| = [G : H] = (np)/p = n. By Lagrange Theorem again, np = |G| is a factor
of |Sn | = n!. But p > n is a prime, and so (pn)|n! is impossible. This contradiction shows that we
must have H = N .

Supplemental Problems

16. Show that if |G| = p2 with p prime, then G ∼


= Zp2 or G ∼
= Zp × Zp .
Proof: If G has a element a with |a| = p2 , then G = ∼ Zp2 . Thus we assume that ∀a ∈ G − {e},
|a| = p.
Claim 1 G is abelian. Since |G| = p2 , C(G) 6= {e}. If C(G) = G, then G is abelian and the claim
holds. Suppose that |C(G)| = p. Then |G/C(G)| = p, and so G/C(G) is cyclic. By Exercise 9 in
this section, G = C(G), a contradiction. Hence G must be abelian.
Pick x ∈ G − {e}, and let X = hxi. Then X ∼ = Zp , as |x| = p. Since |G/X| = p, we must have

G/X = Zp , and so we may assume that for G/X = hyi. By Exercise II-2.3, there exists y ∈ y such
that |y| = p. Let Y = hyi.

3
Claim 2 G = XY . ∀g ∈ G. Then gX = y i X, for some integer i, as G/X = hyi. It follows by Claim
1, that for some integer j, g = y i xj = xj y i ∈ XY . This proves G = XY .
Claim 3 X  G and Y  G. This follows from the fact that G is abelian.
Claim 4 X ∩ Y = {e}. Suppose that g = xj = y i ∈ X ∩ Y , for some integers 0 ≤ i, j ≤ p − 1. Then
y i = X, and so p|i. Since 0 ≤ i ≤ p − 1, we must have i = 0, and so g = e.
By Claims 1-4, G ∼=X ×Y ∼ = Zp × Zp .

You might also like