Group Theory

THEORY

ONE

INTRODUCTION

1.1Preliminaries

We shall begin by reviewing some notions from intuitive set theory.

Throughout the book, equality will mean identity. A set is determined by its members; that is, if S and T are sets, then S = T iff (*) x ∈ S iff x ∈ T. The notation {1, 2, 3}, for example, means the set whose members are 1, 2, and 3. The notation {x | P) means the set of x such that P of powers of primes, the set (later group or ring) J of rationals. Other permanent notation will be introduced from time to time (see the index of notation).

The number of elements of a set S will be called the order of S and denoted by o(S). A set S is a singleton iff o(S) = 1, and an n-ton iff o(S) = n . A set S is a subset of a set T, written S ⊂ T,* iff (*) if x ∈ S, then x ∈ T. If S is a set and Ti is a set for each i ∈ S, then

If S = {1, 2}, then T1 ∪ T2 is usually used instead of ∪ {Тi | i ∈ S) and T1 ∩ T2 instead of ∩ {Ti | i ∈ S}, although there is a conceptual difference. The notation ∪ {Ti | i ∈ S} will be used for ∪ {Ti | i ∈ S} if Ti,·∩ Tj = ø whenever i ≠ j, and will be called the disjoint union of the family {Ti | i ∈ S}. If each element of S is a set, then ∪ {i | i ∈ S} will sometimes be written simply ∪ SS and ∩ S have similar meanings. If S and T are sets, then S\T will denote the set {x | x ∈ S and x ∉ T}.

Two ordered pairs (a, b) and (c, d) are equal iff a = c and b = d. A relation R is a set* of ordered pairs. Its inverse

The relation R is symmetric iff R = R−1. The domain of R is the set

(that is, the domain of R is the set of first coordinates of elements of R). The range of R is given by Rng(R) = Dom(R−1) (that is, the range of R is the set of second coordinates of elements of R). A relation R is said to be from A into В iff Dom(R) = A and Rng(R) ⊂ B, and onto B iff В = Rng(R). If S ⊂ Dom(R), then the restriction of R to S is the set

If S ⊂ Dom(R), then we set SR = Rng(R | S). If R and R′ are relations, then their product is the relation

A relation R is transitive iff RR ⊂ R. The product of relations is associative, that is

1.1.1. If R1 R2, and R3 are relations, then R1(R2R3) = (R1R2)R3.

Proof. Let (a, d) ∈ R1(R2R3). Then ∃ b such that (a, b) ∈ R1 and (b, d) ∈ R2R3. Hence ∃ c such that (b, c) ∈ R2 and (c, d) ∈ R3. Therefore (a, c) ∈ R1R2 (because of the properties of b). But then (a, d) ∈ (R1R2)R3 (because of the properties of c). Hence R1(R2R3) ⊂ (R1R2)R3. Similarly, one shows that (R1R2)R3 ⊂ R1(R2R3), and the equality follows. ||

The Cartesian product of two sets S and T is the set

A relation R is on S iff R ⊂ S × S. A relation R on S is reflexive iff {(a, a) | a ∈ S} ⊂ R. An equivalence relation on S is a relation on S which is reflexive, symmetric, and transitive. If R is an equivalence relation on S, then a subset В of S is an equivalence class with respect to R iff ∃ a ∈ S such that В = {b | (a, b) ∈ R}. It turns out (Exercise 1.1.6) that S is then the disjoint union of its equivalence classes.

While the preceding concepts are useful, the notion which will be used most frequently is that of function. A function is a relation ƒ such that if (a, b) ∈ ƒ and (a, c) ∈ f then b = c. If f is a function and (a, b) ∈ f then we shall often write af = b or f (a) = b (usually the former). It is clear that two functions ƒ and G are equal iff Dom(f) = Dom(g), and af = ag for all a ∈ Dom(f). A function ƒ is 1−1 iff the relation f−1 is a function. If S is a set, the identity function on S is Is = {(s, s) | s ∈ S}. Is will often be denoted by I. It is 1−1 from S onto S, and such that sIs = s for all s ∈ S.

1.1.2. The product of functions is a function.

Proof. Let ƒ and G be functions. Let (a, d) ∈ fg and (a, e) ∈ fg. Then ∃ b such that (a, b) ∈ ƒ and (b, d) ∈ g, and ∃ c such that (a, c) ∈ ƒ and (c, e) ∈ g. Since ƒ is a function, b = c. Therefore, since G is a function, d = e. Hence fg is a function. ||

A binary operation on S is a function ∘ from S × S into S. The fact that (a, b) ∘ = c will be written a ∘ b = c. A binary operation ∘ on S is commutative iff

(i) if a ∈ S and b ∈ S, then a ∘ b = b ∘ a.

A binary operation ∘ on S is associative iff

(ii) if a ∈ S, b ∈ S, and c ∈ S, then (a ∘ b) ∘ c = a ∘ (b ∘ c).

If ∘ is a binary operation on S, T ⊂ S, and U ⊂ S, then T ∘ U is defined to be the set {x ∘ y | x ∈ T and y ∈ U}.

It will now be shown that associativity implies general associativity. Toward this end, let ∘ be a binary operation on a set S. Let f1 be the function with domain S such that f1(a1) = {a1} for all a1 ∈ S, define fn to be the function whose domain is the set of n-tuples (a1, …, an) with ai ∈ S, such that

It is inductively evident that fn(al,. ., an) is not empty. Moreover, the operation ∘ is associative iff f3(al, a2, a3) is a singleton for all a1 ∈ S, a2 ∈ S, and a3 ∈ S.

1.1.3. (General associative law.) If ∘ is an associative binary operation on S and a1 ∈ S, …, an ∈ S, then fn(a1, …, an) is a singleton for all .

Proof. This is obvious for n = 1 and n = 2. Now use induction, and let n > 2. Let z ∈ fn(a1, …, an) and z′ ∈ fn(a1, …, an). Then ∃ x, у, x′, y′, and natural numbers r and t such that

If r = t, then by the induction hypothesis, x = x′ and y = y′, so z = x ∘ у = x′ ∘ y′ = z′. If r < t, say, then by the induction hypothesis, ft−r(ar+l, …, at) is a singleton {v}. Hence, again by the induction hypothesis, x′ = x ∘ v and у = v ∘ y′. Therefore, by associativity,

Hence, by induction, fn(a1, …, an) is a singleton. ||

The sole element of fn(a1, …, an) is denoted by a1 ∘ … ∘ an or by π {ai i n}. It is intuitively clear that Theorem 1.1.3 states that all legal insertions of parentheses into the product a1 ∘ a2 … ∘ an which make the result computable lead to the same result (see Exercise 1.1.19).

In a similar manner, it may be shown that if ∘ is a commutative and associative binary operation on S, then the product of any n elements of S is defined and is independent of the order of factors and positions of parentheses. The proof of this fact will be omitted.

EXERCISES

1.1.4.Show that ∪ and ∩ are commutative and associative (as binary operations).

1.1.5.If A and В are sets, then A = В iff A ⊂ В and В ⊂ A.

1.1.6.Let R be an equivalence relation on a set S. Prove that S is the disjoint union of its equivalence classes with respect to R.

1.1.7.If R is a symmetric and transitive relation on S, then R ∪ {(a, a) | a ∈ S} is an equivalence relation on S.

1.1.8.Show how a plane may be considered as the Cartesian product of the set of real numbers with itself.

1.1.9.In the sense of Exercise 1.1.8, find the geometric condition under which a relation in the plane is symmetric; under which it is reflexive; under which it is transitive.

1.1.10.If o(S) = n is finite, then

(a)o(S × S) = n²,

(b)there are 2n² relations on S,

(c)there are 2n²−n reflexive relations on S,

(d)there are 2(n²+n)/2 symmetric relations on S,

(e)there are 2(n²−n)/2 reflexive and symmetric relations on S.

1.1.11.Let R be a relation on S, and let

such that all (xi, xi+1) ∈ R}. Prove that

(a)R* is the smallest transitive relation on S which contains R;

(b)If R is symmetric, so is R*;

(c)If R is reflexive and symmetric, then R* is an equivalence relation.

1.1.12.Let S be a set, T the set of relations on S, and let

Then ∘ is an associative binary operation on T. Hence the product of relations satisfies the general associative law.

1.1.13.Let S be a set and let F be the set of relations on S which are functions. Prove that if ∘ is the product as in Exercise 1.1.12, then ∘ | F is an (associative binary) operation on F.

1.1.14.Prove that if R1 and R2 are relations, then

1.1.15.Which of the four processes of arithmetic—addition, subtraction, multiplication, and division—are binary operations on J? Which of those that are operations are commutative? Which are associative?

1.1.16.If S and T are sets, then o(S) = o(T) iff ∃ a 1−1 function from S onto T. Prove that if R is a relation, then o(R) = o(R−1).

1.1.17.If o(S) = n is finite, then

(a)there are nn² binary operations on S;

(b)there are n(n²+n)/2 commutative binary operations on S.

1.1.18.Define unary operation, ternary operation, and n-arу operation on S. How many of each are there if o(S) = m?

1.1.19.(a)Show that the general associative law implies that if ∘ is associative, then

(b)Give a similar interpretation of Theorem 1.1.3 for the case of a product of five elements.

1.1.20.(a)Show that if ∘ is a binary operation on a set S, then, in the notation of 1.1.3, fn(al, …, an) is a set containing at most g(n) elements, where g(1) = 1, g(2) = 1, and recursively,

(b) Compute g(4) and compare with Exercise 1.1.19.(a).

1.1.21.Show that the bound given in Exercise 1.1.20.(a) is the best possible, i.e., that for each natural number n there are a set S, a binary operation ∘ on S, and a1 ∈ S, …, an ∈ S such that fп(a1, …, an) is a g(n)-ton.

1.1.22.Prove that if f and G are functions, x ∈ Dom(f), and xf ∈ Dom(g), then x(fg) = (xf)g.

1.1.23.If f is a function, then

1.1.24.Prove that if S × T = U × V, then S = U and T = V. In particular, if S × S = T × T, then S = T.

1.1.25.Prove that if f is a binary operation on S and also a binary operation on T, then S = T.

1.2Definitions and first properties

A group is an ordered pair (G, ∘) such that G is a set, ∘ is an associative binary operation on G, and ∃ e ∈ G such that

(i) if a ∈ G, then a ∘ e = a,

(ii) if a ∈ G, then ∃ a−1 ∈ G such that a ∘ a−1 = e.

When there is no danger of misunderstanding, the group (G, ∘) will be denoted by G, and the element a ∘ b by ab (see Exercises 1.2.11 and 1.2.16). Sometimes the operation ∘ will be denoted by + and called addition. Since general associativity holds (Theorem 1.1.3), parentheses will usually be omitted from products of several factors. The element e is called the identity of G. (See Theorem 1.2.5 for justification of the article the.) The symbol e will always denote the identity of whatever group G is under consideration (eG will be used if necessary for the sake of clarity). The element a−1 is called the inverse of a (see Theorem 1.2.6).

1.2.1. a−la = e.

Proof. Compute a−1aa−1(a−1)−1. (This means that one expands

in two directions to obtain the proof:

1.2.2. ea = a.

Proof. Compute aa−1a.

1.2.3. If a ∈ G and b ∈ G, then ∃ | x ∈ G such that ax = b.

Proof. If ax = b, then x = ex = a−1ax = a−1b by 1.2.2 and 1.2.1. Conversely, if x = a−1b, then ax = aa−1b = eb = b.

1.2.4. If a ∈ G and b ∈ G, then ∃ | x ∈ G such that xa = b.

1.2.5. The element e is unique.

Proof. In Theorem 1.2.3 let b = a.

1.2.6. If ab = e, then b = a−1.

Proof. In Theorem 1.2.3 let b = e.

.

1.2.8. (a−1)−1 = a.

Proof. By 1.2.1. ||

If a ∈ G, define a⁰ = e, define an to be the product of n a’s, and a−n as (a−1)n.

1.2.9. (Exponents.) If a ∈ G and r and s are integers, then (i) aras = ar+s and (ii) (ar)s = ars.

Outline of Proof. If either r or s is 0, the theorem is obvious. If r and s are positive, then (i) follows from the definition of the sum of finite cardinals (and the general associative law) and (ii) from the definition of product of finite cardinals. If one of r and s is negative and the other positive, then (i) follows from successive cancellations. In particular, it follows that a−r = (ar)−1. As to (ii), letting both r > 0 and s > 0,

Again

by an earlier remark and the last sentence. Finally, if r > 0 and s > 0, then

and

by an earlier remark and earlier cases.

EXERCISES

1.2.10.State the theorems of this section in the additive notation.

1.2.11.Let G = {a, b}. Prove that there are exactly two binary operations ∘ on G such that (G, ∘) is a group. (This proves that the convention of denoting a group (G, ∘) by G is actually incorrect. However, we shall continue to use this abbreviation.)

1.2.12.State (ii) in the definition of group in terms of a unary operation on G.

1.2.13.If (G, ∘) is a group, then G is not empty.

1.2.14.Let S = {a, b}.

(a)Define a binary operation on S which is commutative but not associative.

(b)Define a binary operation on S which is associative but not commutative.

1.2.15.Generalize Exercise 1.2.14 to any set S with more than one element.

1.2.16.Prove that if (G, ∘) and (H, ∘) are groups, then G = H. This means that the set G of a group is determined by the operation ∘ of the group. This fact permits one to define a group as an operation ∘ with certain properties. Make this definition.*

1.3Permutation groups

Examples of groups will be given in the next section. In this section, a special type of group, permutation groups, will be introduced, both to facilitate the construction of examples and for its own sake.

A permutation of a set M is a 1−1 function from M onto M.

1.3.1. If M is a set and Sym(M) is the set of permutations of M, then Sym(M) is a group.

Proof. If f ∈ Sym(M) and G ∈ Sym(M), then by Theorem 1.1.2, fg is a function from M into M. If x ∈ M, then, since G is onto M, ∃ y ∈ M such that yg = x. Since ƒ is onto M, ∃ z ∈ M such that zf = y. Hence z(fg) = x. Therefore fg is onto M. By Exercise 1.1.14, (fg)−1 = g−1f−1. Since f and G are 1−1, f−1 and g−1 are functions. By Theorem 1.1.2, (fg)−1 is a function. Hence fg is 1−1. This proves that fg is a permutation of M, that is that fg ∈ Sym(M).

By Exercise 1.1.13, the associative law is satisfied in Sym(M). Clearly, I ∈ Sym(M). If x ∈ M and G ∈ Sym(M), then x(gI) = (xg)I = xg. It follows that gI = g.

If f ∈ Sym(M), then f−1 is a function since ƒ is 1−1. Since ƒ is onto M, f−1 is from M onto M. To prove that f−1 is 1−1, suppose that x ∈ M and y ∈ M are such that xf−1 = yf−1. Then ∃ a ∈ M such that (x, a) ∈ f−1 and (y, a) ∈ f−1, hence (a, x) ∈ ƒ and (a, у) ∈ F. Since ƒ is a function, x = y. This proves that f−1 is 1−1 and therefore that f−1 ∈ Sym(M). It remains to be shown that the inverse function f−1 is an inverse with respect to the operation. If x ∈ M and f ∈ Sym(M), then (x, xf) ∈ f so (xf, x) ∈ f−1, and x(ff−1) = (xf)f−1 = x = xI. Hence ff−1 = I. Therefore Sym(M) is a group. ||

Definition. The group Sym(M) is called the symmetric group on the set M.

Definition. A permutation group is an ordered pair (M, G), where M is a set and G is a group of permutations of M (and where the operation in G is the usual multiplication of functions described earlier). The degree of (M, G) is o(M). The elements of M are called letters.

If V and U are subsets of M and G, respectively, then

If V or U is a singleton, say U = {u}, the notation is simplified to Vu (instead of V{u}).

If ai ∈ M for i = 1, …, and ai ≠ aj. if i ≠ j, then (a1, …, an) means the permutation G ∈ Sym(M) such that aig = ai+1 for i = 1, …, n−1, ang = a1, and bg = b for all other b ∈ M. Such a permutation is called an n-cycle. Similarly, (…, a−1, a0, a1, …) means the G ∈ Sym(M) such that aig = ai+1 for all integers i, and bg = b for all other b ∈ M. This latter permutation is called an ∞-cycle. Any 1-cycle equals e, of course. Two cycles with no letters in common are called disjoint.

Note that

Let (M, G) be a permutation group of finite degree, and let G ∈ G. There are two standard ways of writing g, a single example of which will suffice for purposes of illustration. If M = {1, 2, 3, 4, 5, 6}, 1g = 3, 2g = 5, 3g = 4, 4g = 1, 5g = 2, and 6g = 6, then G may be written in the forms:

The first form will be called a two-row form of g. Either the second or the third form is called a cyclic decomposition of G (or a one-row form).

1.3.2. Iƒ (M, G) is a permutation group of finite degree and G ∈ G, then G is the product of pairwise disjoint cycles. This cyclic decomposition is unique except for order of the cycles and inclusion or omission of 1-cycles.

Proof. The theorem is obvious if o(M) = 1. Induct on o(M). Let G ∈ G and x1 ∈ Mand distinct letters x1, …, xi of M such that

by finiteness of the set M and the fact that G is 1−1. Now G | M\{x1, …, xi} is a permutation (if М\{x1, …, xi} is nonempty), hence by the induction hypothesis G | М\{x1, …, xi} is the product c2 … cm of pairwise disjoint cycles, and G = (x1, …, xi)c2 … cm. Therefore a cyclic decomposition always exists. In any cyclic decomposition of g, the cycle containing x1 must be (x1, …, xi) so that the uniqueness follows readily. ||

If M is infinite and G ∈ Sym(M), then G need not be the product of a finite number of cycles. Nevertheless, Theorem 1.3.2 is still true in a certain sense.

If S ⊂ Sym(M) and for each a ∈ M, there is at most one y ∈ S such that ay ≠ a, then we define a function π {y | y ∈ S} by, for a ∈ M,

We call π {y | y ∈ S} the formal product of the members of S, and will abbreviate it πy when no confusion is possible.

1.3.3. If M is a set and π {y | y ∈ S} is a formal product of elements of Sym(M), then π {y | y ∈ S} ∈ Sym(M).

Proof. It is clear from the definition of formal product and a(πу) that π {у | y ∈ S} is a function from M into M.

Suppose that a(πy) = b(πy) with a ≠ b. If b(πу) = b, then ∃ x ∈ S such that ax = b, while by = b for all у ∈ S. Therefore ax = b = bx and x is not 1−1, a contradiction. Hence WLOG,* a(πy) = b(πу) = c, c ≠ a, c ≠ b. Thus ∃ x ∈ S and y ∈ S such that ax = c = by. Since a ≠ b, x ≠ y. But then cx ≠ c and cу ≠ c, contradicting the definition of formal product. It follows that if a ≠ b, then a(πy) ≠ b(πy), so that πу is 1−1.

Let b ∈ M. If by = b for all y ∈ S, then b(πу) = b. Suppose ∃ x ∈ S such that bx ≠ b. Since x ∈ Sym(M), ∃ a ∈ M such that ax = b. Then a ≠ b, so that a(πу) = b. Therefore πy maps M onto M, so that πу ∈ Sym(M).

1.3.4. If G ≠ e is a permutation, then G is a formal product of pairwise disjoint cycles, G = π {c | c ∈ S}, with no c ∈ S a 1-cycle. If G = π {c | c ∈ S′} is a second such product, then S = S′.

Proof. Let G ∈ Sym(M), a ∈ M, and ag ≠ a. Let

It is obvious that c defined in (1) is a cycle and ag = ac. It follows that if G = π {c | c ∈ S} is a formal product of cycles of length at least 2, then S is the set of cycles of the form (1).

Conversely, let S be the set of cycles of the form (1). Suppose b ∈ M, c1 ∈ S, c2 ∈ S, bc1 ≠ b, and bc2 ≠ b. Then, if

then

while if

then

In either case, it follows that

Therefore π {c | c ∈ S} is a formal product. Moreover, from (1), ag = a(πc) in all cases. Hence G = πc. In a formal product of a set S of cycles of length at least 2, the cycles are automatically disjoint. ||

If a ∈ M, x ∈ Sym(M), and y ∈ Sym(M), then a(x(x−1уx)) = a(yx). From this simple fact, there results the following useful rule.

1.3.5. If x and y are permutations of M, then a cyclic decomposition of x−1yx is formed from one of y by replacing each letter by the letter standing below it in a two-row form of x.

Proof. In a cyclic decomposition of y, the letter a is followed (cyclically) by the letter ay. The letters standing under a and ay in a two-row form of x are ax and ayx, respectively. By the remark preceding the theorem, ayx is the letter following ax in a cyclic decomposition of x−1yx. ||

Example. Let

Then x−1yx = (5, 4, 3)(l, 6)(2).

1.3.6. If x and y are permutations of M, then y and x−1уx have the same number of cycles of each length in their cyclic decompositions.

Proof. By Theorem 1.3.5.

EXERCISES

1.3.7.(1, 2, …, n) = (1, 2)(1, 3)… (1, n).

1.3.8.If g is a permutation of a finite set, then g is a finite product of (not necessarily disjoint) 2-cycles.

1.3.9.The character of g ∈ Sym(M) is the number Ch(g) of letters in M fixed by g. Thus, if M = {1, 2, 3, 4} and g = (1, 2, 3), then Ch(g) = 1. Prove the following statements :

(a)Сh(e) = o(M);

(b)Ch(x−1yx) = Ch(y).

1.3.10.Compute the following products:

(a)(1, 2, 3)(2, 3, 5)(1, 4, 5);

(b)

(c)x−1yx if x = (1, 2, 3) and y =(1, 3)

1.3.11.Prove the following converse of Theorem 1.3.6. If y and z are permutations of M or n = ∞ then y and z have the same number of n-cycles and Ch(y) = Ch(z), then ∃ x ∈ Sym(M) such that z = x−1уx. (M may be infinite.)

1.3.12.Show that (1, 2)(1, 3) = (1, 2, 3), whereas (1, 3)(1, 2) = (1, 3, 2). Hence if o(M3, then Sym(M) has noncommutative operation.

1.4Examples of groups

Although little machinery of group theory has been set up, it seems worth while to give some examples of groups at this stage.

Definition. A group G is Abelian iff (*) if a ∈ G and b ∈ G, then ab = ba.

Example 1. If M is a set, then G = Sym(M) is a group (Theorem 1.3.1). If o(M) < 3, then G is Abelian; if o(M3, then G is non-Abelian (see Exercise 1.3.12). If M is finite, then o(Sym(M)) = (o(M))! If M is infinite, then o(Sym(M)) = 2∘〈M〉.

Example 2. Let M = {1, 2, 3, 4} and let

It is readily verified that (M, G) is a permutation group with identity element (1). Each element is its own inverse, and the product of any two of the three nonidentity elements is the third one. This group is often called the four-group. It is Abelian.

Example 3. The integers J form an Abelian group under the operation of addition.

Example 4. The form an Abelian group under addition.

The following permanent notation will be used occasionally. If G is a group with identity e, let G# = G\{e}.

Example 5. # is an Abelian group under multiplication.

Example 6. Let P be a plane and Q ∈ P. Let G be the set of rotations of P about Q. Then (P, G) is an Abelian permutation group.

Example 7. Let , and let G be the set of rotations of a plane P about a point Q ∈ P through 2πk/n radians, k = 0, 1, …, n−1. Then (P, G) is an Abelian permutation group, and o(G) = n.

Definition. An isometry of Euclidean n-space En is a function T from En into En such that if x ∈ En and у ∈ En, then d(x, y) = d(xT, yT), where d(x, y) is the distance from x to у.

Example 8. If G is the set of all isometries of En, then (En, G) is a permutation group.

Proof. It is clear that the product of two isometries is an isometry and that IEn is an isometry. Associativity follows from the fact that the operation of multiplying functions is associative. If T ∈ G and x and y are distinct points of En, then xT ≠ yT. Hence T is 1−1. If T is onto M, then T−1 is a function from En onto En which preserves distance, and is therefore an isometry. Hence, it only remains to be shown that an isometry T from En into En is onto. A proof of this fact will now be sketched.

If x, y, and z are points such that

then

It follows readily that if L is a line in En, then LT is also a line. An affine subspace of En is a nonempty subset S such that if a and b are distinct points of S then the line through a and b is in S. It follows from the preceding remarks that, if S is an affine subspace, so is ST. If S0, S1, …, Sn = En is a strictly increasing sequence of affine subspaces (which clearly exists), so is S0T, S1T, …, SnT. Therefore, Dim(Si+1T) > Dim(SiT) (where Dim denotes dimension). Hence Dim(Snn. Therefore, SnT = En, and T is onto En. ||

Example 9. Let S be a nonempty subset of En. The set G of isometries T of En such that ST = S is a group.

Example 10. In Example 9, let n = 2, and let S be a regular polygon with k sides. The resulting group G of isometries is called a dihedral group. If k 3, then G is non-Abelian. It has 2k elements: k rotations about the center of S, and k products of a rotation and a fixed reflection about a line through the center and one vertex. ||

Let {Ms | s ∈ S} be a family of sets. Then the Cartesian product × {Ms | s ∈ S} is the set of functions ƒ from S such that sf ∈ Ms for all S ∈ S. (In case o(S) = 2, this differs slightly from the previous definition in that the present product is unordered, while the previous one was ordered.)

Example 11. Let {(GS, ∘s) | s ∈ S} be a family of groups. Then the direct product π {(GS, ∘s) | s ∈ S} has × {Gs | s ∈ S} as set, and addition defined by the rule

(that is, addition is made componentwise). The direct product of groups is again a group. If S is finite, say S = {1, …, n}, it is more customary to call

the direct product. However, there is a formal distinction between the two concepts in this case (see Exercise 2.1.32 and earlier remarks). The latter notation will also be used.

Example 12. The external direct sum ΣΕ {Gs | s ∈ S} of a family {Gs | s ∈ S} of groups is the set of f ∈ π {Gs | s ∈ S} such that sf is the identity of Gs for all but a finite number (perhaps 0) of s ∈ S. The external direct sum is a group under the operation given in Example 11.

Example 13. As a special case of Example 11, let S be a set and G a group. Then the sum of two functions ƒ and G from S into G is given by

The set of these functions forms a group π {G | s ∈ S} under addition. The identity elements of this group will be denoted by OSG or O, even in situations where the group itself is not under consideration. Thus

The inverse of T ∈ π {G | s ∈ S} will be denoted by −T. Thus

EXERCISES

1.4.1.Prove that the external direct sum and direct product of Abelian groups is Abelian.

1.4.2.In Example 9, let n = 3 and let S be a cube. Show that o(G) = 48.

1.4.3.In Example 9, if S is an n-dimensional cube, what is o(G)? Describe the elements of G in case S is the unit cube.

1.5Operations with subsets

Definition. If А1, …, An are subsets of a group G, then the product A1Α2 … An is the set of elements of G of the form a1a2 … an, where ai ∈ Ai. The following convention will be used: if a factor A of a product is a singleton {x}, then x will be used instead of {x}. For example, if x ∈ G, then SxT means S{x}T.

1.5.1. S(TU) = (ST)U = STU for all subsets S, T, and U of G.

1.5.2. If G is Abelian, then AB = ВA for all subsets A and В of G.

1.5.3. A(∪{Bi | i ∈ S}) = ∪ {ABi | i ∈ S},

1.5.4. A(∩ {Bi | i ∈ S}) ⊂ ∩{ABi | i ∈ S}.

EXERCISES