• 沒有找到結果。

Exercises (§1.9, p.62)

N/A
N/A
Protected

Academic year: 2022

Share "Exercises (§1.9, p.62)"

Copied!
5
0
0

加載中.... (立即查看全文)

全文

(1)

Basic Algebra (Solutions)

by Huah Chu

Exercises (§1.9, p.62)

1. Let G = (Q, +, O), K = Z. Show that G/K ' the group of complex numbers of the form e2πiθ, θ ∈Q, under multiplication.

Proof. Define a homomorphism φ : G → {e2πiθ|θ ∈ Q} by θ → e2πiθ. Then ker φ = K

and φ is surjective. ¤

2. Show that a → a−1 is an automorphism of a group G if and only if G is abelian, and if G is abelian, then a → ak is an endomorphism for every kZ.

Proof. (1) φ : a → a−1 is an automorphism

⇔ For all a, b ∈ G, (ab)−1 = φ(ab) = φ(a)φ(b) = a−1b−1.

⇔ For all a, b ∈ G, ab = ba, that is, G is abelian.

(2) G is abelian. ¿From (ab)k = akbk, we have a → ak is an endomorphism. ¤

3. Determine Aut G for (i) G an infinite cyclic group, (ii) a cyclic group of order six, (iii) for any finite cyclic group.

Sol. (i) Let G = hai be an infinite cyclic group. The generators of G are a and a−1. Hence, for φ ∈ Aut G, φ(a) = a or a−1. Hence Aut G = {1G, φ : a → a−1} 'Z/2Z.

(ii) Let G = ha|a6 = 1i. The generators of G are a and a5 by exercise 4, §1.5. hence Aut G = {1G, φ : a → a5} 'Z/2Z.

(iii) Let G = hai by any finite cyclic group with |G| = n. Then all generators of G are ak, (k, n) = 1. Then Aut G is the set of all homomorphisms defined by φ : a → ak, (k, n) = 1.

Remark. In the case of (iii), Aut G is isomorphic to the group of units of the multi- plicative monoid (Z/nZ, ·). Its structure will be determined in Chap. 4, §11. (Thm.

4.19, 4.20).

4. Determine Aut S3.

Sol. We shall show that Aut S3 ' S3.

(2)

Step 1. The elements of S3 are 1, a = (123), a2 = (132), b = (12), ab = (13), (a2b = (23). Then we have the relation ba = a2b. Using this relation, the reader can verify that

(ambn)(apbq) = am+(n+1)pbn+q, m, p = 0, 1, 2; n, q = 0, 1 (∗) easily. Since an automorphism preserves the order of an element, hence, for φ ∈ Aut G, φ(a) = ai and φ(b) = ajb for some i = 1, 2, j = 0, 1, 2.

Step 2. Define the map φij : G → G by φij :

½ a → ai

b → ajb , i = 1, 2, j = 0, 1, 2. Then φij ∈ Aut G:

We have φij(am) = aim, φij(amb) = aim+jb by the definition of φij. Using these, it is easy to see that φij is bijective. Then we check that φij is a homomorphism in the following four cases:

(i) x = amb, y = anb. Then φ((amb)(anb)) = φ(am+2n) (by (∗)) = ai(m+2n). On the other hand, φ(amb)φ(anb) = aim+jbain+jb = aim+j+2(in+j) = aim+2in+3j = ai(m+2n). The other three cases: (ii) x = amb, y = an, (iii) x = am, y = anb, and (iv) x = am, y = an are left to the reader.

Step 3. It is easy to see that φ10 = 1, φ11 and φ12 have order 3. φ20, φ21 and φ22 have order 2. We define the mapping Φ : S3 → Aut S3 by ai 7→ φ1i, aib 7→ φ2i. The reader can verify that it is an isomorphism.

Remark. (1) Since S3 = ha, b|a3 = b2 = 1, ba = a2bi (See §1.11), to prove that φij is a homomorphism, it is enough to check that (φij(a))3 = (φij(b))2 = 1, φij(b)φij(a) = ij(a))2ij(b)).

(2) In fact, Aut Sn ' Sn for all n 6= 6, and Aut S6/S6 ' Z/2Z, (c.f. I. J. Rotman:

The theory of groups, p.132, or B. Huppert Endlich Gruppen I, p.173–177).

(3) For other remark, see the remark after exercise 5.

5. Let a ∈ G, a group, and define the inner automorphism (or conjugation) Ia to be the map x → axa−1 in G. Verify that Ia is an automorphism. Show that a → Ia is a homomorphism of G into Aut G with kernel the center C of G. Hence conclude that Inn G ≡ {Ia|a ∈ G} is a subgroup of Aut G with Inn G ' G/C. Verify that Inn G is a normal subgroup of Aut G. Aut G/ Inn G is called the group of outer automorphisms.

Proof. The last statement follows from φIaφ−1(b) = Iφ(a)(b). We leave all the verifica-

tions to the reader. ¤

Remark. A group G is complete in case C(G0) = 1 and Aut G ' G. Exercise 2 in

§1.4 and the remark in the above exercise show that Sn is complete for n 6= 2, 6.

(3)

It can be shown that if G is simple of composite order, then Aut(G) is complete.

6. Let G be a group, GL the set of left translations aL, a ∈ G. Show that GLAut G is a group of transformations of the set G and that this contains GR. GLAut G is called the holomorph of G and is denoted as Hol G. Show that if G is finite, then

| Hol G| = |G|| Aut G|.

Proof. (1) If gL∈ GL, φ ∈ Aut G, then φgLφ−1= φ(g)L. ¿From this fact, we can prove that GLAut G is a group.

(2) Since gL−1gR(x) = g−1xg = Ig−1 ∈ Aut G, hence gR = gLIg−1. And GR GLAut G.

(3) To prove | Hol G| = |G|| Aut G|, it suffices to show that GL∩Aut G = {1}. Since gL(1) = g 6= φ(1) for φ ∈ Aut G, g 6= 1, the result follows. ¤

7. Let G be a group such that Aut G = 1. Show that G is abelian and that every element of G satisfies the equation x2 = 1. Show that if G is finite then |G| = 1 or 2.

Proof. (1) let G be a group with Aut G = 1. Then G/C ' Inn G = 1 where C is the center of G (by exercise 5). Hence G is abelian. If G is abelian, a → a−1 is an automorphism (by exercise 2). The assumption Aut G = 1 implies that a = a−1 for all a, that is, a2 = 1.

(2) Suppose |G| is finite and G 6= 1.

Step 1. We prove that G contains elements a1, . . . , ar such that every element of G can be written in a unique way in the form a − 1k1· · · akrr, ki = 0, 1:

For this purpose, we show that, for all i, there exists a normal subgroup H = ha1, . . . , aii of G such that every element of H can be written as ak11· · · akrr, ki = 0, 1, uniquely. We prove this statement by induction on i. Note that any subgroup of G is normal since G is abelian.

Take any 1 6= a1 ∈ G, then ha1i is normal in G. Suppose we have H = ha1i × · · · × haii. Take any ai+1 ∈ G − H. Then H ∩ hai+1i = 1 since |hai+1i| = 2. Because G is abelian any element of hH, ai+1i can be written in the form hb with h ∈ H, b ∈ hai+1i.

Moreover, the expression is unique: If h1b1 = h2b2, then h−12 h1 = b2b−11 ∈ H ∩hai+1i = 1 and h1 = h2, b1 = b2. Hence the statement.

Step 2. Suppose n ≥ 2. Define the mapping α : G → G by ak11ak22· · · aknn = ak12ak21ak33· · · aknn. Obviously, α is a nontrivial automorphism. This contradicts to the hypothesis Aut G = 1. Thus n = 1 and |G| = 2.

Remarks. (1) We reprove Step 2 in the language of vector space. In Step 1, we have shown that G is abelian and x2 = 1 for all x. Regard G as an additive group, then G is a vector space over finite fieldZ/2Z (§4.13) and an automorphism is just a nonsingular linear transformation. Let {a1, . . . , an} be a basis of G. Suppose dim G ≥ 2, then G

(4)

has a nontrivial nonsingular linear transformation a1 7→ a2, a2 7→ a1, and ai → ai, i > 2. A contradiction.

(2) When G is an infinite abelian group with x2 = 1 for all x, we can still regard G as a vector space over Z/2Z. In this case, using Zorn’s lemma, we can find a base for G. Hence it is not difficult to construct a nontrivial nonsingular linear transformation on G.

8. Let α be the automorphism of a group G which fixes only the unit of G(α(a) = a ⇒ a = 1). Show that a → α(a)a−1 is injective. Hence show that if G is finite, then every element of G has the form α(a)a−1.

Proof. Let α be a fixed point free automorphism (α(a) = a ⇒ a = 1). Suppose α(a)a−1 = α(b)b−1. Then α(b−1a) = b−1a. Hence b−1a is fixed by α and b−1a = 1.

Thus a → α(a)a−1 is injective.

If |G| < ∞, by the pigeon hole principle, the mapping is surjective. ¤ 9. Let G and α be as in 8, G finite, and assume α2 = 1. Show that G is abelian of odd order.

Proof. (1) For any element g of G, g has the form α(a)a−1. α(g) = α(α(a)a−1) = α2(a)α(a−1) = aα(a)−1 = g−1. Thus G is abelian by exercise 2.

(2) Next we show that |G| is odd. Suppose to the contrary, there is a ∈ G with order 2 (exercise 13, §1.2). Then α(a) = a−1 = a, contradicts to the hypothesis about

α. ¤

Remark. An automorphism α of G is said to be fixed point free if it leaves only the unit fixed. This exercise shows that: if G admits a fixed point free automorphism of order 2, then G is abelian. Some further results are:

Suppose that G admits a fixed point free automorphism α of order n. (1) If n = 3, then G is nilpotent (for the definition, see Basic Algebra, I, p.243, exercise 6) and x commutes with α(x) for all x. (2) If n is a prime, then G is nilpotent (John G.

Thompson). (3) G is solvable in general (for the definition, see Basic Algebra, I, p.237).

For more details, we refer to D. Gorenstein: Finite groups, chap. 10, pp.333–357 and D. Gorenstein. Finite simple groups.

10. Let G be a finite group, α an automorphism of G, and set I = {g ∈ G|α(g) = g−1}.

Suppose |I| > 3|G|. Show that G is abelian. If |I| = 3|G|, show that G has an abelian

(5)

Proof. (1) Let I = {g ∈ G|α(g) = g−1} and |I| > 34|G|. For any h ∈ I, claim:

I ∩ h−1I ⊂ C(h). In fact, if x ∈ I ∩ h−1I, then x − h−1g with g, x ∈ I. Now α(h−1g) = (h−1g)−1 = g−1h; on the other hand α(h−1g) = α(h)−1α(g) = hg−1. Thus g−1 ∈ C(h). It follows that g ∈ C(h) and x = h−1g ∈ C(h) also.

Since |I| = |h−1I| > 34|G|, so |I ∩ h−1I| > 12|G|. Thus C(h) is a subgroup of order

> 12|G|. Then C(h) = G and h ∈ C(G), the center of G. Because this holds for any h ∈ I, so |C(G)| ≥ 34|G| and G = C(G), G is a abelian.

(2) Suppose |I| = 34|G|. Then G can not be abelian, otherwise, I is a subgroup of G.

Hence there exists h ∈ I − C(G). Let K = I ∩ h−1I, then K = C(h) and |K| = 12|G|, by the proof of (1). Since [G : K] = 2, K is normal. The only property remains to prove is that K is abelian.

For any k = h−1g ∈ K = C(h), then g ∈ C(h). Thus for k1 = h−1g1, k2 = h−1g2 K, g1g2 ∈ C(h) ⊂ I. Then (g1g2)−1 = φ(g1g2) = φ(g1)φ(g2) = g1−1g2−1and g1commutes

with g2. So k1 commutes with k2. ¤

Remark. The reader is urged to find a finite non-abelian group G and its automor- phism α such that |I| = 34|G|. In fact let G = {±1, ±i, ±j, ±k} be the quaternion group and α the inner automorphism determined by i. Then |{g ∈ G : α(g) = g−1}| = 6.

參考文獻

相關文件

Let {f n } be a sequence of real-valued continuous functions on [a, b] which converges uniformly to a continuous function f on

Since every point of S is an accumulation point of S, S has no isolated point... Is it true that if x is an accumulation point of A, then x is also an accumulation point

A scheme is integral if it is nonempty and for every nonempty affine open set U ⊂ X, the ring O X (U ) is an integral domain..

Remark 2: Indicate the properties of limit given below when you use it in the test.. Show that (a n ) is convergent and evaluate

Here is

Remark: All the sequences are sequence of real numbers.. Formula that might be useful: Let θ

(The argument is similar as those in the above exercises.) The only possibility is p ∈ A... The proof is given

(i) [5%] Give an example of a nondegenerate symmetric bilinear form of Witt index 1 on a two- dimensional real vector space.. (ii) [5%] Give an example of nonzero quadratic form on