0% found this document useful (0 votes)
30 views

Unit 9

This document introduces the topic of permutation groups. It discusses how permutation groups will be studied in detail, including the symmetric group Sn and its subgroups. The key topics that will be covered are: properties of permutations in Sn; the alternating group; and Cayley's theorem, which states that every group is isomorphic to a permutation group.

Uploaded by

Sweta Akhori
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
30 views

Unit 9

This document introduces the topic of permutation groups. It discusses how permutation groups will be studied in detail, including the symmetric group Sn and its subgroups. The key topics that will be covered are: properties of permutations in Sn; the alternating group; and Cayley's theorem, which states that every group is isomorphic to a permutation group.

Uploaded by

Sweta Akhori
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 30

Unit 9 Permutation Groups

UNIT 9

PERMUTATION GROUPS

Structure Page Nos.

9.1 Introduction 255


Objectives
9.2 Preliminaries 256
9.3 Properties of Permutations in Sn 259
9.4 Alternating Groups 267
9.5 Cayley’s Theorem 275
9.6 Summary 277
9.7 Solutions / Answers 278

9.1 INTRODUCTION
In this unit, we will discuss, in detail, a group that you studied briefly in
Sec.1.5, Unit 1, and Sec.2.4.2, Unit 2. This is the symmetric group. As you
have often seen in previous units, the symmetric group Sn , as well as its
subgroups, have provided us with many examples. As you know, the
symmetric groups and their subgroups are permutation groups. Historically, it
was the study of permutation groups and groups of transformations that gave
the foundation to group theory. A lot of work in the study of permutations was
undertaken by several European mathematicians in the 18th century.
However, the basis of this theory, and the notation that you will study, are
mainly due to the French mathematician, Augustin-Louis Cauchy, whose Fig.1: Cauchy
partial theorem you also studied in Unit 7. (1789-1857)

In Sec.9.2, we will help you recapitulate what you have studied about
permutations and permutation groups in Units 1 and 2.

In Sec.9.3, we shall look at several properties of elements of Sn . In particular,


you will see why every element of Sn is a cycle or a product of disjoint cycles.
Then you will study why 2-cycles generate Sn , for n ≥ 2.

In Sec.9.4, the focus will be on certain permutations, called even


permutations. You will see why the set of even permutations in Sn is a normal
subgroup of Sn , called the alternating group. This group has several
interesting properties, which you will also study.
255
Block 2 Normal Subgroups and Group Homomorphisms
Finally, in Sec.9.5, you will study a result by the famous mathematician,
Cayley, which says that every group is isomorphic to a permutation group.
Thus, every isomorphism class of groups can be represented by a permutation
group. This result is what makes permutation groups so important.

Please study this unit carefully, because it gives you a solid basis for studying
and understanding the theory of groups. We also suggest that you go through
Sec.1.5 and Sec.2.4.2 again, before beginning work on this unit.

Objectives
After studying this unit, you should be able to:
• express any permutation in S n as a product of disjoint cycles, and as a
product of transpositions;
• find out whether an element of S n is odd or even;

• prove that the alternating group of degree n , A n , is normal in Sn and is


n!
of order ;
2
• prove, and apply, the result that A n is simple ∀ n ≥ 5;

• state, and prove, Cayley’s theorem.

9.2 PRELIMINARIES
From Sec.1.5 (Unit 1) and Sec.2.4.2 (Unit 2), you know that a permutation on
a non-empty set X is a bijective function from X onto X. We denote the set
of all permutations on X by S(X).

Let us now gather some facts that you studied in Sec.2.4.2.


Suppose X is a finite set having n elements. For simplicity, we symbolise
these elements by 1, 2, K , n. Then, you know that the set of all permutations
on these n symbols is denoted by Sn .
You also know that we represent any f ∈ S n in a 2-line form as
⎛ 1 2 K n ⎞
f = ⎜⎜ ⎟⎟. …(1)
⎝ f (1) f (2) K f (n ) ⎠
How many elements do you think Sn has? To count them, look at f as in (1)
above. Now, there are n possibilities for f (1), namely, 1, 2, K , n. Once f (1)
has been specified, there are (n − 1) possibilities for f (2), namely,
{1, 2, K, n} \ {f (1)}, since f is 1-1 and onto. Thus, there are n (n − 1) choices
for f (1) and f (2). Continuing in this manner, you can see that there are n !
different possibilities for f ∈ Sn . Therefore, S n has n ! elements.

Now, let us look at the algebraic structure of S(X), for any set X. From the
course Calculus, you know that the composition of bijections from X to X is
a bijection from X to X. Hence, if f , g ∈ S(X), then f o g ∈ S(X). So,
composition is a binary operation on S(X). To help you regain practice in
computing the composition of permutations, consider an example.

Example 1: Find f o g, where f = (1 2 4 3) and g = (1 4 2) in S4 .


256
Unit 9 Permutation Groups
⎛ 1 2 3 4⎞
Solution: From Unit 2, you know that f = (1 2 4 3) = ⎜⎜ ⎟⎟, and
⎝ 2 4 1 3⎠
⎛ 1 2 3 4⎞
g = (1 4 2) = ⎜⎜ ⎟⎟.
⎝ 4 1 3 2⎠
Then, to get f o g, we first apply g and then apply f .
∴ f o g(1) = f (g (1)) = f (4) = 3,
f o g (2) = f (g (2)) = f (1) = 2,
f o g (3) = f (g(3)) = f (3) = 1,
f o g (4) = f (g (4)) = f (2) = 4.
⎛1 2 3 4⎞
∴ f o g = ⎜⎜ ⎟⎟.
⎝3 2 1 4⎠
We show this process diagrammatically in Fig. 2.

f og
g f
1 1
1
2
2
2
3 3
3

4 4
4
Fig. 2: (1 2 4 3) o (1 4 2) in S 4 .

***

Now, let us go back to S(X), for any set X. You have studied the proof of the
following result in Sec.2.4.2.

Theorem 1: Let X be a non-empty set. Then (S(X), o) is a group, called the


permutation group on X.

Thus, Sn is a group of order n ! . Recall, from Unit 2, that S n is called the


symmetric group of degree n.
Now, from Unit 2, you also know that if f ∈ Sn , then
⎛ f (1) f (2) K f (n ) ⎞
f −1 = ⎜⎜ ⎟.
⎝ 1 2 K n ⎟⎠

With the recap above, and the experience that you have gained in previous
units, you should now solve the following exercises.

E1) Show that (S n , o) is a non-commutative group for n ≥ 3.

E2) Show that Sm ≤ Sn if m ≤ n.

E3) Let G be a group and let g ∈ G. Show that f : G → G : f ( x ) = gx is in


S(G ). (f is called the left regular representation by g of G.)

257
Block 2 Normal Subgroups and Group Homomorphisms
At this point we would like to make a remark about the terminology and
notation.

Remark 1: In line with Remark 5 of Unit 2, from now on we will refer to the
composition of permutations as multiplication of permutations. We will also
drop the composition sign. Thus, we will write f o g as fg, unless we want to
stress the operation involved.

The two-line notation that we have used for a permutation of a finite set is
rather cumbersome. Let us see if there is a shorter notation. In case the
permutation is a cycle, you know that we can denote it in one line only, as in
Example 1.

Let us first recall how a cycle is written in one line from a 2-line format.
⎛1 2 3 4⎞
Consider the permutation f = ⎜⎜ ⎟ in S4 . Choose any one of the
⎝3 1 4 2 ⎟⎠
symbols, say 1 .
Now, we write down a left hand bracket followed by 1 : (1
Since f maps 1 to 3 , we write 3 after 1 : (1 3
Since f maps 3 to 4 , we write 4 after 3 : (1 3 4
Since f maps 4 to 2 , we write 2 after 4 : (1 3 4 2
Since f maps 2 to 1 (the symbol we started with), we
close the brackets after the symbol 2 : (1 3 4 2)
Now, since no more symbols are left in f , we write f = (1 3 4 2).

This means that a cycle maps each symbol to the symbol on its right, except
for the final symbol in the brackets, which is mapped to the first symbol.
1 3
If we had chosen 3 as our starting symbol, we would have got f = (3 4 2 1).
Note that this cycle is exactly the same as (1 3 4 2), because both cycles
show the same value for f (i), i = 1, 2, 3, 4. Hence, they both denote the
permutation which we have represented diagrammatically in Fig.3. This is an
2 4 example of a 4-cycle, or a cycle of length 4. Fig.3 may give you some
Fig. 3: (1 3 4 2). indication about why we call this function a cycle.

⎛1 2 3 4⎞
Now consider g = ⎜⎜ ⎟⎟. You should check that g = (1 2 4). But what
⎝2 4 3 1⎠
happened to 3 ? Since g (3) = 3, i.e., g fixes 3, we don’t include it in the cycle
representation of g.

More generally, we have the following definition.

Definition: A permutation f ∈ S n is called an r-cycle (or cycle of length r),


If f (s) = s, then we for r ≤ n , if there are r distinct integers x1 , x 2 , K , x r lying between 1 and n
say f fixes s. such that f ( x i ) = x i +1 ∀ i = 1, K , r − 1, f ( x r ) = x1 , and
If f (s ) ≠ s, we say that
f (k ) = k ∀ k ∈ {1, 2, K , n} \ {x1 , x 2 , K , x r }.
f moves s.
Then, we write f = (x1 x 2 K x r ).

In particular, a 2-cycle is called a transposition.


258
Unit 9 Permutation Groups
For example, the permutation f = (2 3) ∈ S3 is a transposition. Here
f (1) = 1, f (2) = 3 and f (3) = 2.
Note that, if we have f = (2 3) ∈ S7 , then f (2) = 3, f (3) = 2, and f (k ) = k for
k = 1, 4, 5, 6, 7.

In the next section you will see that transpositions play a very important role in
the theory of permutations.

Consider the following important observation about 1-cycles.

Remark 2: Consider any 1-cycle, say (3), in S4 . (3) maps 3 to itself, and
⎛1 2 3 4 ⎞
maps 1, 2 and 4 to 1, 2 and 4, respectively. Thus, (3) = ⎜⎜ ⎟⎟ = I,
⎝1 2 3 4 ⎠
the identity permutation. Thus, any 1-cycle (i) in S n is the identity
⎛1 2 K n ⎞
permutation I = ⎜⎜ ⎟⎟, since it maps i to i and the other (n − 1)
⎝1 2 K n ⎠
symbols to themselves.

You are already familiar with cycles in S3 . You know that there are two
3-cycles, (1 2 3) and (1 3 2). There are also three transpositions in S3 ,
namely, (1 2), (1 3) and (2 3). You have worked with these cycles in several
examples and exercises of previous units. Now you can work with other cycles
while solving the following exercises.

E4) Write down 2 distinct transpositions, 2 distinct 3-cycles and 2 distinct


7-cycles in S7 . Justify your choices.

⎛1 2 3 4 5⎞
E5) Write ⎜⎜ ⎟⎟ as a cycle.
⎝5 2 3 1 4⎠

E6) i) Show that (1 2) −1 = (1 2), (1 2 3) −1 = (3 2 1) and


(2 5 6 8) −1 = (8 6 5 2), in S8 .

ii) If f = (i1 i 2 K i r ) ∈ Sn , n ≥ r , then show that f −1 = (i r i r −1 Ki 2 i1 ).

E7) Give two distinct elements of [G, G ], where G = S4 .

Now that we have done a quick review of the basic concepts regarding
permutations, let us discuss important properties of elements of Sn .

9.3 PROPERTIES OF PERMUTATIONS


From what you have studied in the previous section, you may think that we
can express any permutation as a cycle. However, consider the following
example from S5 .
⎛1 2 3 4 5 ⎞
Let g = ⎜⎜ ⎟⎟.
⎝3 5 4 1 2⎠
259
Block 2 Normal Subgroups and Group Homomorphisms
Let us start with the symbol 1, and apply the procedure for obtaining a cycle to
g. We obtain (1 3 4) after three steps, because g maps 4 to 1, and hence,
we close the brackets, even though we have not yet written down all the
symbols in g. So, are the leftover symbols, 2 and 5, fixed by g ? No. We see
that g (2) ≠ 2 and g (5) ≠ 5.

So, as the next step, we simply choose any symbol that has not appeared so
far, say 2, and start the process of writing a cycle again. Thus, we obtain
another cycle (2 5). Now, all the symbols in {1, 2, 3, 4, 5} are exhausted. So,
what is g in terms of the cycles we have got? Let’s see.

If we write (1 3 4) (2 5) in the two-line format, what do we get?


⎛ 1 2 3 4 5 ⎞ ⎛1 2 3 4 5 ⎞ ⎛ 1 3 4 2 5 ⎞
⎜⎜ ⎟⎟ o ⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟,
⎝ 3 2 4 1 5 ⎠ ⎝1 5 3 4 2 ⎠ ⎝ 3 4 1 5 2 ⎠
which is the same as g.
∴ g = (1 3 4)(2 5).
We call this expression for g, as a product of two cycles that move different
symbols, the cycle decomposition of g. In Fig.4 we represent this by a
diagram which shows the decomposition clearly.

2 5

1 4

Fig. 4: g = (1 3 4) (2 5) .

Notice that the symbols in the two cycles that make up g form disjoint sets,
{1, 3, 4} and {2, 5}. Further, because of the arbitrary choice of the symbol at
the beginning of each cycle, there are many ways of expressing g. For
example,
g = (4 1 3)(2 5) = (2 5)(1 3 4) = (5 2)(3 4 1).
However, within each cycle, the same order has to be maintained. For
instance, we cannot replace (4 1 3) by (4 3 1), as h = (4 3 1)(2 5) is a
different function from g. Why? Note that g (1) = 3, but h (1) = 4.

So, we can write the product of the separate cycles with disjoint symbols in
any order. The choice of the starting element within each cycle is arbitrary,
ensuring that each cycle represents the same function.

So, you see that in this case g can’t be written as a cycle, but as a product of
cycles of the kind we now define.

Definition: Two cycles are called disjoint if they have no symbol in common.

Thus, disjoint cycles of length 2 or more move disjoint sets of symbols.


So, for example, the cycles (1 2) and (3 4) in S4 are disjoint. But (1 2) and
(1 4) are not disjoint, since they both move 1.
260
Unit 9 Permutation Groups
Now let us consider one more example.

⎛1 2 3 4 5 ⎞
Example 2: Write h = ⎜⎜ ⎟ ∈ S5 as a cycle, or a product of
⎝1 3 2 5 4 ⎟⎠
disjoint cycles.

Solution: As we did for g above, we start with the symbol 1. However, 1 is


fixed by h. So, we ignore 1, and move to 2. We get the cycle (2 3).
Now consider 4, and we get the cycle (4 5). Now all the symbols in h have
been exhausted.
So h = (2 3)(4 5).
Note that, by convention, we don’t include the 1-cycle in the expression for h ,
unless we wish to emphasise it, since it is just the identity permutation.
Thus, we simply write h = (2 3)(4 5), or h = (4 5)(2 3), ignoring (1).
***

Let us now generalise what we have noted above about the disjoint cycles.

Theorem 2: If σ1 and σ 2 ∈ Sn are disjoint cycles, then σ1σ 2 = σ 2 σ1.

Proof: Let σ1 = (a1 a 2 K a r ) and σ 2 = (b1 b 2 K bs ), with r + s ≤ n and


{a1 ,K, a r } ∩ {b1 ,K, bs } = «.
Let {1, 2, K , n} = {a 1 , K , a r , b1 , K , b s , c1 , c 2 , K , c k } with k = n − (r + s) ≥ 0, and
σ1 (ci ) = ci = σ 2 (c i ) ∀ i = 1, K , k.
Now σ1σ 2 (a i ) = σ1 (a i ), since σ 2 (a i ) = a i ∀ i = 1,K, r.
= a i+1 ∀ i = 1, K , r, and putting a r +1 = a 1.
Also σ 2 σ1 (a i ) = σ 2 (a i+1 ) = a i+1 ∀ i = 1, K , r.

Similarly, you can show that σ1σ 2 ( bi ) = σ 2σ1 ( bi ) ∀ i = 1,K, s, taking bs +1 = b1.

Finally, σ1σ 2 (ci ) = σ1 (ci ) = ci and σ 2 σ1 (ci ) = c i ∀ i = 1, K , k.

Thus, σ1σ 2 ( x ) = σ 2σ1 ( x ) ∀ x ∈ {1,K, n}.


Hence, σ1σ 2 = σ 2σ1.

Theorem 2 is why we had noted earlier that g = (1 3 4)(2 5) = (2 5)(1 3 4).

The process we have used earlier, to write g (and h in Example 2) as a


product of disjoint cycles, can be used to write any permutation that moves a
finite set of symbols in the same way.

Theorem 3 (Cycle decomposition): Every permutation of a finite set is either


a cycle or a product of disjoint cycles.

Proof: Note that the identity permutation can be trivially seen as a 1-cycle, or
a product of 1-cycles.
Now, S1 = {I}, S2 = {I, (1 2)}, and you have also seen that every element of
S3 is a cycle.
So, let’s assume n ≥ 4 and σ ∈ Sn is not a cycle. This means σ ≠ I.
261
Block 2 Normal Subgroups and Group Homomorphisms
Let x1 ∈ {1, 2, K , n} s.t. σ moves x1 , and let x 2 = σ( x1 ).
Then take x 3 = σ( x 2 ) = σ(σ( x1 )) = σ 2 ( x1 ), x 4 = σ( x 3 ) = σ3 ( x1 ), and so on.
Since {1, 2, K , n} is finite, by this process symbols will start repeating at some
point, say σi ( x1 ) = σ j ( x1 ) for some i < j.
If r = j − i, then x1 = σ r ( x1 ).
Let m be the least positive integer s.t. σ m ( x1 ) = x1 .
Then σ1 = ( x1 x 2 K x m ) is a cycle, and σ( x i ) = σ1 ( x i ) ∀ i = 1, K , m.
Now take y1 ∈ {1, 2,K, n} \ {x1 , x 2 ,K, x m }, where σ moves y1. Such a y1
exists since σ is not a cycle.
Then, using the same process as above, we get
σ 2 = ( y1 y 2 K ys ), for some s ≥ 2, and where yi = σi −1 ( y1 ) for i = 1,K, s.
Are σ1 and σ 2 disjoint? Suppose they are not. Then, for some i and j, we
get x i = y j , i.e., σi −1 ( x1 ) = σ j−1 ( y1 ), i.e., σi− j ( x1 ) = y1 , i.e., y1 = x i− j+1 , a
contradiction to the way y1 was chosen.
So σ1 and σ 2 are disjoint.

We can continue the process by which we got σ1 and σ 2 till all the symbols
moved by σ are exhausted. Note that for each i ∈ {1, 2, K, n} s.t. σ(i) = i, the
1-cycle (i) is I. Hence, we do not include this in the decomposition.
So, we finally get σ = σ1σ 2 K σ t as a product of t disjoint cycles of lengths
greater than 1.

Because of Theorem 3, any permutation in Sn , written in the 2-line format, can


be more conveniently expressed as a cycle decomposition. Do you agree?
Also, because of Theorem 2, the order in which the cycles in a decomposition
are written doesn’t matter.

If you have understood the discussion so far, you will be able to solve the
following exercises.

E8) Express each of the following permutations as products of disjoint


cycles.

⎛1 2 3 4 5⎞
i) ⎜⎜ ⎟⎟,
⎝ 5 4 2 1 3⎠

⎛1 2 3 4 5 6 7 8⎞
ii) ⎜⎜ ⎟⎟,
⎝8 4 7 2 1 3 6 5⎠

⎛1 2 3 4 5⎞
iii) ⎜⎜ ⎟⎟.
⎝ 4 5 3 1 2⎠

E9) Write (1 4)(2 3) ∈ S4 , (6 5 2 4)(3 1) ∈ S8 and (3 1 2)(4 6)(5 7) ∈ S8 in the


two-line format.

E10) Do the cycles (1 3) and (1 5 4) in S6 commute? Give reasons for your


answer.
262
Unit 9 Permutation Groups
E11) If f is an r-cycle, then show that o(f ) = r , i.e., f = I and f ≠ I , if s < r.
r s

(Hint: If f = (i1 i 2 K i r ), then f (i1 ) = i 2 , f 2 (i1 ) = i3 , K, f r −1 (i1 ) = i r .)

From E11, you know what the order of a cycle is. So, using this, and the cycle
decomposition of an element of Sn , can we obtain the order of an element of
Sn easily? Consider an example.

Example 3: Find o(g ), where g = (1 3 4)(2 5) in S5 .

Solution: Let σ1 = (1 3 4) and σ 2 = ( 2 5).


Then o(σ1 ) = o((1 3 4)) = 3, and o(σ 2 ) = o((2 5)) = 2.
Now, as seen in Theorem 2, σ1σ 2 = σ 2 σ1 .
So g 2 = (σ1σ 2 ) 2 = σ12 σ 22 = σ12 ≠ I, since o(σ1 ) = 3 and o(σ 2 ) = 2.
Also, g 3 = σ12 g = σ13σ 2 = σ 2 since o(σ1 ) = 3.
≠ I, since o(σ2 ) = 2.
Similarly, you should verify that
g 4 ≠ I, g 5 ≠ I, g 6 = σ16σ62 = (σ13 ) 2 (σ 22 )3 = I.
Thus, o(g ) = 6 = l.c.m(o(σ1 ), o(σ 2 ))
= l.c.m of the lengths of σ1 and σ 2 .
***

What you have found for g, in the example above, is true in general. This is
what the following theorem tells us, which was proved by the Italian
mathematician, Paolo Ruffini. Fig.5: Paolo Ruffini
(1765-1822)
Theorem 4: Let σ ∈ Sn , for n ≥ 3. Let σ = σ1σ 2 K σ r , as a product of disjoint
cycles. Then o(σ) is the least common multiple of the lengths of
σ1 , σ 2 , K , σ r .

Proof: In E11 you have proved that the order of a cycle of length p is p.
Now, consider σ = σ1σ 2 , where σ1 and σ 2 are disjoint cycles of lengths r
and s, respectively. Thus, o(σ1 ) = r and o(σ 2 ) = s. Let o(σ) = t and
l.c.m(r, s) = m.
As r m and s m, σ1m = I = σ m2 .
∴ σ m = σ1m σm2 , since σ1σ2 = σ 2σ1.
= I.
Since t = o(σ), we see that t m. …(2)
−t −1 t
Also I = σ = σ σ , so that σ = σ
t t
1
t
2 = (σ ) .
t
1 2 2

Since σ1 and σ 2 are disjoint, so are σ1 and σ −2 1 , and hence, σ1t and σ −2 t are
disjoint. Thus, they can only be equal if σ1t = I = σ 2− t .
So, r t , s t.
Thus, m t. …(3)
From (2) and (3), we conclude that m = t.

Let us now apply the strong principle of induction on the number of disjoint
cycles in the product.
263
Block 2 Normal Subgroups and Group Homomorphisms
If σ is a cycle, or a product of 2 cycles, you have seen that the theorem is
true.
Assume that the theorem is true for all permutations written as a product of
(r − 1) cycles.
Now, let σ = σ1σ 2 K σ r be a product of r disjoint cycles.
Then σ = ρσ r , where ρ = σ1σ 2 K σ r −1.
Hence, o(ρ) = m = l.c.m of the lengths l 1 , l 2 , K , l r −1 of σ1 , σ 2 , K , σ r −1 ,
respectively.
Now, let l r be the length of σ r and let o(σ) = t.
Then, as in the case for r = 2 above, you can show that
t = l.c.m(m, l r ) = l.c.m(l 1 , l 2 , K , l r ).
Hence, the result is true for a product of r disjoint cycles.
Thus, by the principle of induction, it is true in general.

Using Theorem 4, you can easily find the order of elements in Sn . Let us
consider an example.

⎛ 1 2 3 4 5 6 7 8 9 10 ⎞
Example 4: Find the order of σ = ⎜⎜ ⎟⎟ in S10 .
⎝ 7 8 5 6 3 4 2 1 9 10 ⎠

Solution: Note that σ = (1 7 2 8)(3 5)(4 6).


Thus, σ = σ1σ2σ3 , l(σ1 ) = 4, l(σ2 ) = 2 = l(σ3 ), where l(σ) = length (σ).
Hence, o(σ) = l.c.m(4, 2, 2) = 4.
***

You can see how the cycle decomposition eases the process of finding the
order of an element of Sn . Of course, this representation is also an elegant
representation of a permutation in Sn !

Now consider an important property of a cycle, or a product of disjoint cycles.

Example 5: From Unit 6, you know that α, β ∈ Sn are conjugates if ∃ σ ∈ Sn


Example 5(i) shows that
the conjugate of an r- s.t. α = σ β σ −1. For σ ∈ Sn , show that
cycle is an r-cycle.
Example 5(ii) tells us i) if β = ( x1 x 2 K x r ) in Sn , then σβσ −1 = (σ( x1 ) σ( x 2 ) K σ( x r )).
that conjugation
preserves the cycle ii) if ρ = σ1σ 2 K σs as a product of disjoint cycles in Sn , then
structure of a
permutation. σρσ−1 = α1α 2 K α s , a product of disjoint cycles in Sn with
length (α i ) = length (σi ) ∀ i = 1,K, s.

Solution: i) Note that σβσ −1 (σ( x i )) = σβ( x i ) = σ( x i +1 ) ∀ i = 1, K , r, taking


x r +1 = x1.
Also, let y ∈ {1,K , n} \ {σ( x1 ),K , σ( x r )}.
Then y = σ(z) for some z ≠ x i ∀ i = 1,K , r , as σ is a 1-to-1 map of
{1, K , n}.
So σβσ −1 ( y) = σβσ −1 (σ(z)) = σ(z), as β(z) = z.
∴ σβσ −1 = (σ( x1 ) σ( x 2 ) K σ( x r )), a cycle of length r.

ii) Let ρ = σ1σ 2 K σ s , as a product of disjoint cycles. Then by Example 29,


Unit 8, you know that conjugation by σ is the inner automorphism f σ .
264
Unit 9 Permutation Groups
−1 −1 −1 −1
∴ σρσ = (σσ1σ )(σσ 2 σ ) K (σσs σ ) …(4)

Also, by (i), σσi σ −1 only moves what σ i moves.


Hence, (4) is a representation of σρσ −1 as a product of disjoint cycles in
Sn , with length (σσi σ −1 ) = length (σ i ).
***

What is proved in Example 5 is very useful, and will be used several times in
this unit.

Try solving some related exercises now.

⎛ 1 2 3 4 5 6 7 8 9 10 ⎞
E12) Find o(σ), where σ = ⎜⎜ ⎟⎟ ∈ S10 .
⎝ 3 5 6 4 7 9 2 10 1 8 ⎠

E13) Give an example of cycles σ1 and σ 2 ∈ Sn , where


o(σ1σ 2 ) ≠ l.c.m(l(σ1 ), l(σ 2 )). Does Theorem 4 fail here? Why, or why
not?

E14) If σ ∈ Sn , must we have o(σ) ≤ n ? Must o(σ) n ? Give reasons for your
answers.

Now, let us discuss transpositions. Consider the cycle (1 5 3) in S5 . You


should verify that this is the same as the product (1 3)(1 5). You should also
verify that (1 5 3) = (1 5)(5 3). Note that the transpositions in either product are
not disjoint.

The same process can be used to show that any r-cycle


(i 1 i 2 K i r ) = (i 1 i r ) (i 1 i r −1 )K (i 1 i 2 ), a product of transpositions.
Also (i 1 i 2 K i r ) = (i 1 i 2 )(i 2 i 3 )K(i r −1 i r ), again a product of transpositions.
Note that, since the transpositions aren’t disjoint, they do not commute (see
E16). Further, as you have seen above, the expression of a cycle as a
product of transpositions is not unique.

Before discussing the importance of transpositions, try solving the following


exercises.

E15) Express the following cycles in S5 as products of transpositions:


i) (1 3 4) , ii) (4 3 1), iii) (2 4 5 3).

E16) Show that (i j)( j k ) ≠ ( j k )(i j) in Sn , n ≥ 3, for any three distinct


symbols i, j, k.

Let us now use the cycle decomposition of a permutation to prove a result


which shows why transpositions are so important in the theory of
permutations.

Theorem 5: Every permutation in Sn , n ≥ 2, can be written as a product of


transpositions.
265
Block 2 Normal Subgroups and Group Homomorphisms
Proof: By Theorem 3, you know that every permutation is a product of one or
more disjoint cycles. Also, you have just seen how every cycle is a product of
transpositions. Hence, every permutation is a product of one or more
transpositions.
Note that I = (1 2)(1 2). Thus, I is also a product of transpositions.

Let us see how Theorem 5 works in practice, through an example.

⎛1 2 3 4 5 6 7 ⎞
Example 6: Write σ = ⎜⎜ ⎟⎟ as a product of transpositions.
⎝ 3 6 4 1 2 5 7 ⎠

Solution: σ = (1 3 4)(2 6 5), as a product of disjoint cycles


= (1 4)(1 3)(2 5)(2 6).
***

Consider the following important comment here.

Remark 3: Note that if σ = τ1τ2 K τt as a product of transpositions, you


cannot conclude that o(σ) = 2. Theorem 4 cannot be applied here. It holds
only for a disjoint cycle decomposition, and here the τi are not disjoint.

Now, in Theorem 5 you have seen that the set of transpositions in S n


generates S n . In fact, there is a smaller set that generates Sn . Recall, from
Unit 4, that a set S generates a group G if every element of G is of the form
s1r1 s r22 K s rnn , s i ∈ S and ri ∈ Z.

Theorem 6: {(1 2), (1 3),K, (1 n )} generates Sn .

Proof: From Theorem 5, you know that any σ ∈ Sn is a product of


transpositions. Now take any transposition (i j) ∈ Sn . Then
(i j) = (1 i)(1 j)(1 i).
Hence, any σ ∈ Sn is a product of transpositions of the form (1 r ), for some
r ∈ {1,K, n}.
Thus, {(1 2), (1 3),K, (1 n )} generates Sn .

In fact, an even smaller set than the one given in Theorem 6 generates
Sn , n ≥ 3, namely, {(1 2), (1 2 K n )}. However, we will not prove this here.

Now you can try your hand at solving some exercises.

E17) Write the permutation in E8(ii) as a product of transpositions.

E18) Write (1 3 4)(5 7)(2 6 8) ∈ S8 as a product of transpositions of the form


(1 i), for some i ∈ {1,K,8}.

E19) Show that (1 2)(2 3) K (9 10) = (1 10)(1 9) K (1 2).

E20) Is the set of all transpositions in Sn a subgroup of Sn ? Why, or why not?

266
Unit 9 Permutation Groups
The decomposition given in Theorem 5 leads us to focus on a certain
important subgroup of Sn that we will now discuss.

9.4 ALTERNATING GROUPS


You have seen that any permutation in Sn can be written as a product of
transpositions. You have also seen that the factors in the product are not
uniquely determined. Not only this, even the number of factors in the product
can vary. For example, in S4 we have (1 2)(3 4)(1 4) = (1 2)(3 4)(2 3)(1 4)(3 2).
Here the LHS has 3 transpositions, and the RHS has 5 transpositions. Can
the RHS have 4, or 6, transpositions? Try to find any such representation.

Actually, all representations as a product of transpositions have one thing in


common – if a permutation is the product of an odd number of transpositions
in one such representation, then it will be a product of an odd number of
transpositions in any such representation. Similarly, if f ∈ Sn is a product of an
even number of transpositions in one representation, then f will be a product
of an even number of transpositions in any such representation. To see this
fact, we first need to define a concept new to you.

Definition: The signature of f ∈ Sn ( n ≥ 2) is defined to be


n
f ( j) − f (i)
sign f = ∏ . n
∏ α i = α1α 2 K α n
i , j =1 j−i i =1
i< j

For example, for f = (1 2 3) ∈ S3 ,


f (2) − f (1) f (3) − f (1) f (3) − f (2)
sign f = ⋅ ⋅
2 −1 3 −1 3− 2
⎛ 3 − 2 ⎞ ⎛1− 2 ⎞ ⎛1− 3 ⎞
=⎜ ⎟⎜ ⎟⎜ ⎟ = 1.
⎝ 1 ⎠⎝ 2 ⎠⎝ 1 ⎠
Similarly, if f = (1 2) ∈ S3 , then we still need to include the factors involving
f (3) for obtaining the signature of f .
f (2) − f (1) f (3) − f (1) f (3) − f (2)
So, sign f = ⋅ ⋅ (Note that f (3) = 3 here.)
2 −1 3 −1 3− 2
⎛ 1− 2 ⎞ ⎛ 3 − 2 ⎞ ⎛ 3 −1⎞
=⎜ ⎟⎜ ⎟⎜ ⎟ = −1.
⎝ 1 ⎠⎝ 2 ⎠⎝ 1 ⎠

Try some simple exercises now to get used to the signature of a permutation.

E21) What is the signature of I ∈ Sn ?

⎛ 1 2 3 4 5⎞
E22) Find the signature of σ = ⎜⎜ ⎟⎟ ∈ S5 .
⎝ 4 3 2 1 5⎠

From the examples you have seen so far, you may have concluded that the
signature is a function from Sn to Z. You will see that it is, in fact, a
homomorphism from Sn to ({1, − 1}, ⋅). Let us first show that sign preserves
the operation.
267
Block 2 Normal Subgroups and Group Homomorphisms

Theorem 7: sign : Sn → Q is a well-defined homomorphism, where n ≥ 2.

Proof: By definition, if f , g ∈ Sn s.t. f = g, then f (i) = g(i) ∀ i = 1,K, n.


∴ sign f = sign g in Q∗.
Thus, sign is well-defined.
n
f (g ( j)) − f (g (i))
Next, sign (f o g) = ∏
i , j=1 j−i
i< j

f (g( j)) − f (g(i)) g( j) − g(i)


=∏ ⋅∏ . …(5)
i, j g( j) − g(i) i, j j −1
i< j i< j

Now, as i and j take all possible pairs of distinct values from 1 to n , so do


g(i) and g( j), since g is a bijection. So f can be thought of as a permutation
of {g (1),K, g (n )}.
f (g ( j)) − f (g (i))
∴∏ = sign f .
i< j g ( j) − g (i)
∴ (5) tells us that sign (f o g ) = (sign f )(sign g ).

By Theorem 7, we know that, for instance,


sign (1 2 3 4) = sign ((1 2)(2 3)(3 4)) = sign (1 2) ⋅ sign (2 3) ⋅ sign (3 4).
You may think that this way of finding sign (1 2 3 4) seems to be much longer
than finding it directly. However, the following theorem, with Theorem 7, gives
us properties of the signature function that shorten the process. Of course, let
us not forget the crucial role of Theorem 5 in all this!

Theorem 8: Consider sign : Sn → Q∗ , n ≥ 2.

i) If t ∈ S n is a transposition, then sign t = −1.

ii) Im (sign ) = {1, − 1}.

Proof: i) Let t = (p q), where 1 ≤ p < q ≤ n. [To help you understand what is
going on, you may like to work with (1 2) ∈ Sn , as you go through the
proof.]
Now, only one factor of sign t involves both p and q, namely,
t (q ) − t ( p ) p − q
= = −1. …(6)
q−p q−p
Every factor of sign t that doesn’t contain p or q equals 1, since
t (i) − t ( j) i − j
= = 1, if i, j ≠ p, q. …(7)
i− j i− j
The remaining factors contain either p or q but not both. These can be
paired together to form one of the following products.

t (i) − t (p) t (i) − t (q) i − q i − p


= = 1, if i > q,
i−p i−q i−p i−q
t (i) − t (p) t (q) − t (i) i − q p − i
= = 1, if q > i > p, …(8)
i−p q−i i−p q −i
t (p) − t (i) t (q) − t (i) q − i p − i
= = 1, if i < p.
p−i q−i p−i q −i
268
Unit 9 Permutation Groups
Thus, taking the values of all the factors of sign t , from (6), (7) and (8)
you can see that sign t = −1.

ii) Let f ∈ S n . By Theorem 5, you know that f = t 1 t 2 K t r for some


transpositions t 1 , K, t r in Sn .
∴ sign f = sign ( t 1 t 2 K t r )
= (sign t 1 ) (sign t 2 ) K (sign t r ), by Theorem 7.
= (−1) r , by (i) above.
∴ sign f = 1 or − 1, depending on whether r is even or odd.
Hence, Im (sign ) ⊆ {1, − 1}.
You also know that sign t = −1, for any transposition t , and sign I = 1.
∴{1, − 1} ⊆ Im (sign ).
∴ Im (sign ) = {1, − 1}.

So Theorems 7 and 8 tell us that sign : S n → {1, − 1} is an epimorphism,


n ≥ 2.

Now, we are in a position to prove what we said at the beginning of this


section.

Theorem 9: Let f ∈ S n and let f = t1t 2 K t r = t1′ t ′2 K t ′s be two factorisations


of f into a product of transpositions. Then either both r and s are even
integers, or both are odd integers.

Proof: Let us apply the function sign : Sn → {1, − 1} to f = t1t 2 K t r .


By Theorem 8, you know that
sign f = (sign t1 ) (sign t 2 ) K (sign t r ) = (−1) r . …(9)
Also f = t1′ t ′2 K t ′s .
So sign f = (−1) s . …(10)
From (9) and (10) we get (−1) = (−1) .
r s

This can only happen if both s and r are even, or both are odd.
Thus, the number of factors occurring in any factorisation of f into
transpositions is always even, or always odd.

The theorem above leads us to the following definition.

Definition: A permutation f ∈ S n is called even if it can be written as a


product of an even number of transpositions. f is called odd if it can be
decomposed as a product of an odd number of transpositions.

For example, (1 2) ∈ S3 is an odd permutation. In fact, any transposition is an


odd permutation. On the other hand, any 3-cycle is an even permutation,
since (i j k ) = (i k )(i j). So sign (i j k ) = (−1)(−1) = 1.

Consider the following remark in this context.

Remark 4: What Theorems 8 and 9 tell us is that f ∈ Sn is odd iff


sign f = (−1). Thus, f ∈ Sn is even iff sign f = 1.

Now, here’s your chance to work with some odd and even permutations.
269
Block 2 Normal Subgroups and Group Homomorphisms

E23) Which of the permutations in E15 and E18 are odd?

E24) If f , g ∈ S n are odd, then is f o g odd too? Why?

E25) Is the identity permutation odd or even? Why?

Now we will consider an important subset of Sn , namely, A n = {f ∈ Sn f is


even}.
You will see that A n Sn .

Theorem 10: The set A n , of even permutations in Sn , forms a normal


n!
subgroup of S n of order .
2

Proof: You have already seen that the signature function,


sign : Sn → {1, − 1} is an epimorphism.
Now Ker (sign ) = {f ∈ Sn sign f = 1}
= {f ∈ Sn f is even}
= An .
Thus, A n Sn .
Further, by the Fundamental Theorem of Homomorphism,
(Sn A n ) ~ {1, − 1}.
o(Sn )
∴ o(Sn A n ) = 2, that is, = 2.
o( A n )
o(Sn ) n !
∴ o( A n ) = = .
2 2

Theorem 10 leads us to make the following important comments.

Remark 5: Note that Theorem 10 tells us that


i) the number of even permutations in S n equals the number of odd
permutations in S n , and

ii) A n has only 2 cosets in Sn , A n and σA n , where σ is any odd


permutation in Sn . Thus, Sn = A n ∪ (1 2)A n .

Theorem 10 leads us to the following definition.

Definition: A n , the group of even permutations in Sn , is called the


alternating group of degree n.
3!
Let us consider what A 3 looks like. Theorem 10 says that o(A 3 ) = = 3.
2
Since (1 2 3) = (1 3)(1 2), (1 2 3) ∈ A 3 . Similarly, (1 3 2) ∈ A 3 . Of course,
I ∈ A 3 . Also (1 2) ∉ A 3 . Similarly, (2 3) and (1 3) are not in A 3 .
∴ A 3 = {I, (1 2 3), (1 3 2)}.
You have already been working with this subgroup of S3 in several examples
of the earlier units.
270
Unit 9 Permutation Groups
Do the following exercises now.

E26) Prove that an r-cycle is odd if r is even, and even if r is odd.

E27) Write down all the elements of A 4 . Is A 4 abelian? Why?

E28) Give the two elements of S5 A 5 .

E29) Check whether or not all the odd permutations in Sn form a subgroup of
S n , n ≥ 3.

Now, for a moment, let us go back to Unit 5 and Lagrange’s theorem. This
theorem says that the order of the subgroup of a finite group divides the order
of the group. However, we did not give you an example there to show you why
the converse is not true. Now that you know what A 4 looks like, we are in a
position to disprove the converse.

Example 7: Show that 6 o(A 4 ), but A 4 has no subgroup of order 6. The converse of Lagrange’s
Theorem for finite groups is
Solution: Suppose, to the contrary, A 4 has a subgroup H of order 6. Then not true.

o(H) = 6, o(A 4 ) = 12. ∴ A 4 : H = 2. ∴ H A 4 (see Theorem 2, Unit 6).


Thus, A 4 H is a quotient group of order 2, and hence, is cyclic. Let
A 4 H = < Hg > .
Then (Hg) 2 = H ∀ g ∈ A 4 . (Remember H is the identity of A 4 H.)
∴ g 2 ∈ H ∀ g ∈A 4 .
As (1 2 3) ∈ A 4 , (1 2 3) 2 = (1 3 2) ∈ H.
Similarly, (1 3 2) 2 = (1 2 3) ∈ H.
By the same reasoning all the 3-cycles in A n are in H.
Thus, (1 4 2), (1 2 4), (1 4 3), (1 3 4), (2 3 4), (2 4 3) are also distinct elements
of H. Of course, I ∈ H.
Thus, H contains at least 9 elements.
∴ o(H) ≥ 9. This contradicts our assumption that o(H) = 6.
Therefore, A 4 has no subgroup of order 6.
***

We will use A 4 to provide another counterexample too. (See how useful A 4


is!) In Unit 6 you studied that if H N and N G, then H need not be
normal in G. Well, here’s an example as an exercise for you (se E30).

E30) Consider the subset V4 = {I, (1 2)(3 4), (1 4)(2 3), (1 3)(2 4)} of A 4 . ‘ ’ is not a transitive
relation.
Show that V4 A 4 . Also show that H = {I, (1 2)(3 4)} is a normal
subgroup of V4 , but H A 4 . (Hence, H V4 , V4 A 4 but H A 4 . )

E31) How many elements does A 5 have of order


i) 2? ii) 3? iii) 5? iv) 15?
Give reasons for your answers.
271
Block 2 Normal Subgroups and Group Homomorphisms
Now, let us find a generating set for A n . You have already seen that
S = {(1 i) 2 ≤ i ≤ n} generates Sn . Since no transposition lies in
A n , S ∩ A n = «. So, what could a generating set for A n be, apart from the set
A n itself (of course!)?

Theorem 11: The set of 3-cycles in Sn generates A n , n ≥ 3.

Proof: Let S be the set of all 3-cycles in Sn . Since any 3-cycle is an even
permutation, < S > ⊆ A n .
Now, in Theorem 6, you have seen that {(1 i) 1 < i ≤ n} generates Sn .
So, let σ ∈ A n . Then σ ∈ Sn .
So σ = (1 i1 )(1 i 2 ) K (1 i r ), where the i js are not necessarily distinct.
However, r = 2m for some m ∈ N, since σ ∈ A n .
Also, if i j = i j+1 , then (1 i j )(1 i j+1 ) = I.
So let us assume i j ≠ i j+1 ∀ j = 1, K, r − 1.
Now, (1 i1 )(1 i 2 ) = (1 i 2 i1 ).
Similarly, (1 i 3 )(1 i 4 ) = (1 i 4 i 3 ).
In this way, taking pairs of adjacent transpositions, we get σ as the product of
m 3-cycles of the form (1 i j), i, j = 2, K, n , i ≠ j.
Thus, σ ∈ < S > .
∴ An = < S > .

Let us consider an application of Theorem 11.

Example 8: Let H A n , n ≥ 3, and let (1 2 3) ∈ H. Show that (1 3 2) ∈ H.


Further, show that (1 i 3)(1 2 3)(1 3 i) ∈ H for i ∈ {1, 2, K, n} \ {1, 3}.

Solution: Since (1 3 2) = (1 2 3) −1 and H ≤ A n , (1 3 2) ∈ H.


Next, (1 i 3) ∈ A n and H A n . Hence,
−1
(1 i 3)(1 2 3)(1 i 3) = (1 i 3)(1 2 3)(1 3 i) ∈ H.
***

Now try solving an exercise which we shall require for proving a very important
corollary of Theorem 11.

E32) Let H A n , n ≥ 3, and let (1 2 3) ∈ H. Calculate the following:


i) (1 i 3)(1 2 3)(1 3 i),
ii) ( j 1 i)(1 i 2)( j i 1),
iii) (1 2 k )( j 2 i)(1 k 2),
where i, j, k ∈ {1, 2, K , n} are distinct, and are such that all the 3-cycles
in the products above are defined.
Show that all the elements you have just calculated lie in H. Hence
show that H = A n .

272
Unit 9 Permutation Groups
Now let us use E32 to prove a corollary of Theorem 11.

Corollary 1: Let H A n , n ≥ 3. If H contains a 3-cycle, then H = A n .

Proof: From Example 5, you know that every conjugate of a 3-cycle is a


3-cycle, and that conjugation preserves the operation in A n .
Now, let (i j k )∈ H, and let (r s t ) be any 3-cycle.
As in E32, taking σ1 = (i r k ), σ 2 = (s i r ) and σ3 = ( j t i), you should verify
that σ3σ 2 σ1 (i j k )(σ3σ 2 σ1 ) −1 = (r s t ).
Hence, (r s t ) ∈ H.
Thus, all the 3-cycles are in H. Hence, by Theorem 11, H = A n .

Consider the impact of the result above. It says that if any normal subgroup of
A n has one 3-cycle, then it must be all of A n . You will see how this is applied
as we discuss the simplicity of A n .

For i = 1, 2, 3, A i is simple. Why? Well, A1 = {I}; A 2 = {I}; A 3 = < (1 2 3) >,


which is of order 3, and hence simple (see Unit 6).
Also, in E30 you have seen that A 4 is not simple. What about A 5 ? The
following theorem answers this. This theorem was known to the young
mathematician, Galois. However, the first formal proof was given by the
Fig.6: Camille Jordan
French analyst and algebraist, Camille Jordan, in 1870. (1838-1922)

Theorem 12: A n is simple for n ≥ 5.

Proof: Let n ≥ 5. Suppose A n has a non-trivial normal subgroup H.


Since H ≠ {I}, ∃ σ′ ∈ H, σ′ ≠ I. Out of all such σ′, pick σ ∈ H such that σ
moves the least number of integers from 1 to n , say σ moves k integers.
Firstly, k ≠ 1, since if σ moves one element, it must move at least 2, being a
non-identity bijection.
However, k ≠ 2, since σ ∈ A n , and hence, σ cannot be a transposition.
If k = 3, then σ must be a 3-cycle. Then, by Corollary 1, H = A n .

Next, let k ≥ 4, and suppose σ = ( x1 x 2 K x r )ρ is a disjoint product, where


3 ≤ r < k.
Now, let α = ( x1 x 2 x 3 ), and β = σ −1ασα −1.
Since H A n , ασα −1 ∈ H. Hence, β ∈ H.
Now, if ρ moves an element, then it must move more than one element at
least.
Assume ρ( x r +1 ) = y1 ≠ x r +1 and ρ( x r + 2 ) = y 2 ≠ x r +2 , where x r +1 , x r +2 , y1 , y 2
are not any of the x i s, i = 1, K , r.
Then σ( x r +1 ) = y1 , and σ( x r +2 ) = y 2 .
Now β( x r +1 ) = σ −1ασα −1 ( x r +1 ) = σ −1ασ( x r +1 ), since α fixes x r +1.
= σ −1α( y1 ) = σ −1 ( y1 ), since α fixes y1.
= x r +1.
Also β( x1 ) = σ −1ασ( x 3 ) = σ −1α( x 4 ) = σ −1 ( x 4 ) = x 3 ≠ x1.
So β ≠ I.
273
Block 2 Normal Subgroups and Group Homomorphisms
Further, for any t fixed by σ, t ∉ {x1 , x 2 , K , x r }. So t is also fixed by β.
So β ∈ H, with β moving one less element than σ (since β( x r +1 ) = x r +1 ). This
is a contradiction to the way we chose σ. Hence, our assumption that σ can
be written as a disjoint product with one factor being a cycle of length ≥ 3 is
not possible.

So, now we are left with the possibility of k ≥ 4 and σ is a disjoint product of
transpositions, say,
σ = ( x1 x 2 )( x 3 x 4 )ρ,
where ρ is a product of an even number of disjoint transpositions.
Take γ = ( x1 x 2 x 3 ) and δ = σ −1γ σ γ −1. Then, you should show why δ ∈ H.
Further, you should check that
δ = (σ −1 ( x1 ) σ −1 ( x 2 ) σ −1 ( x 3 )) ( x 3 x 2 x1 ) = ( x 2 x1 x 4 )( x 3 x 2 x1 ) = ( x 3 x1 )( x 2 x 4 ).
So ( x 3 x1 )( x 2 x 4 ) ∈ H.
Now, since n ≥ 5, ∃ i ≠ x1 , x 2 , x 3 , x 4 in {1, 2, K , n}.
Let μ = ( x1 x 3 i). Then μ −1 = (i x 3 x 1 ).
Again, you should check that μ −1δμδ ∈ H.
Now μ −1δμδ = (μ −1 ( x 3 ) μ −1 ( x1 ))(μ −1 ( x 2 ) μ −1 ( x 4 ))(x 3 x1 )( x 2 x 4 )
= ( x1 i)( x 2 x 4 )( x 3 x1 )( x 2 x 4 )
= ( x1 i)( x1 x 3 ) = ( x1 x 3 i)
= μ.
Thus, μ ∈ H.
Hence, as in the case k = 3, H = A n .

Hence, in all the cases if H A n , H = {I} or H = A n .


Thus, A n is simple for n ≥ 5.

Though the proof above may seem a bit involved, please study it carefully,
doing each step yourself.

Using Theorem 12, we can now show what the non-trivial normal subgroups of
Sn look like. You have already seen that S1 = {I}, and that S2 = {I, (1 2)} is
simple. You also know that the only non-trivial normal subgroup of S3 is A 3 .
You have seen that S4 has at least two normal subgroups, V4 and A 4 .
What about Sn for n ≥ 5 ? You know that A n Sn ∀ n ∈ N. Are there any other
normal subgroups, as in S4 ? Consider the following result.

Theorem 13: A n is the only non-trivial proper normal subgroup of Sn ∀ n ≥ 5.

Proof: Let H be a non-trivial normal subgroup of Sn , n ≥ 5.


Then H ∩ A n An .
Hence, by Theorem 12, H ∩ A n = {I} or H ∩ A n = A n .
If H ∩ A n = A n , A n ⊆ H ⊆ Sn .
So (H A n ) (Sn A n ) , and o(Sn A n ) = 2.
Hence, (H A n ) = {I} or (H A n ) = (Sn A n ).
Thus, H = A n or H = Sn .

If H ∩ A n = {I}, then H has no even permutation, apart from I.


274
Unit 9 Permutation Groups
Hence, if σ ∈ H, σ ≠ I, σ is an odd permutation.
Now, suppose H has two distinct non-trivial elements σ1 , σ 2 . Then both are
odd. So σ1σ 2 is even. Also σ1σ 2 ∈ H, since H ≤ S n .
Hence, σ1σ 2 ∈ H ∩ A n = {I}.
Thus, σ 2 = σ1−1. Hence, H = < σ1 >, where o(σ1 ) = 2. Say, σ1 = (i j).
Since n ≥ 5, ∃ k ∈ {1, K, n} \ {i, j}.
Then, since H Sn , (i k )(i j)(i k ) −1 ∈ H, i.e., ( j k ) ∈ H, a contradiction, since
H = {I, σ1}. Hence H ∩ A n ≠ {I}, i.e., H ∩ A n = A n .

Thus, the only proper normal subgroups of Sn are {I} and A n .

In the proofs of both Theorems 12 and 13, we have left several steps for you
to check. So please do every step yourself.

Let us consider an example of how Theorems 12 and 13 are of help.

Example 9: Find all possible group homomorphisms from S6 to Z 7 .

Solution: Let f : S6 → Z 7 be a homomorphism.


Then Ker f S6 . Hence, Ker f = {I}, or Ker f = A 6 , or Ker f = S6 .
If Ker f = {I}, then S6 ~− Im f ≤ Z 7 . But o(S6 ) > 7 = o(Z 7 ).
Hence, we reach a contradiction. ∴ Ker f ≠ {I}.
If Ker f = A 6 , then S6 A 6 ~
− Im f ≤ Z7 . Here o(Im f ) = o(S6 A 6 ) = 2. But
2 o(Z 7 ). Hence, we reach a contradiction. ∴ Ker f ≠ A 6 .

Thus, the only possibility is Ker f = S6 , i.e., f (σ) = 0 ∀ σ ∈ S6 , i.e., f is the


zero map.
***

Try solving the following exercises now.

E33) Let f : A 7 → G be a group homomorphism. Show that o(G ) ≥ 2520 or


f (x) = e ∀ x ∈ A7 .

E34) What are the possible group homomorphisms from S5 to U5 , and why?
(Hint: Analyse the possibilities of Ker f , for each such f .)

And now let us see why permutation groups are so important in group theory.

9.5 CAYLEY’S THEOREM


In this course you have studied all kinds of groups – finite, infinite, abelian,
non-abelian, cyclic, non-cyclic. You have studied subgroups of C and C∗,
subgroups of M m× n (C), those of D 2 n , Sn , Un , and of so many other groups.
You have also seen that there are infinitely many isomorphism classes of
groups. Yet, it turns out, amazingly, that each of these isomorphism classes
has a permutation group in it. This was noticed, and proved, by Arthur Cayley,
the English mathematician by whose name we also call the group operation
275
Block 2 Normal Subgroups and Group Homomorphisms
tables that you have used again and again. Let us see what Cayley’s theorem
precisely says.

Theorem 14 (Cayley): Any group G is isomorphic to a subgroup of the


Recall that S(G ) is the group
permutation group S(G ). Thus, G can be viewed as a permutation group.
of all bijections from G to G.

Proof: For a ∈ G, we define the left regular representation


f a : G → G : f a ( x ) = ax.
In E3, you have shown that f a ∈ S(G ) ∀ a ∈ G.
Now let us define a function f : G → S(G ) : f (a ) = f a .

f is well-defined: If a = b in G, then ax = bx ∀ x ∈ G. So f a = f b in S(G ).

f is a homomorphism: To prove this, we note that


(f a o f b ) ( x ) = f a (bx ) = abx = f ab( x ) ∀ a , b ∈ G.
∴ f (ab) = f ab = f a o f b = f (a ) o f (b) ∀ a , b ∈ G.

f is 1-1: To prove this, consider


Ker f = {a ∈ G f a = I G }
= {a ∈ G f a( x ) = x ∀ x ∈ G}
= {a ∈ G ax = x ∀ x ∈ G}
= {e}, by left cancellation.
Thus, by the Fundamental Theorem of Homomorphism,
G Ker f ~− Im f ≤ S(G ),
i.e., G − Im f ≤ S(G ), since G ~
~ − (G {e}).
i.e., G is isomorphic to a subgroup of S(G ).

The importance of Theorem 14 needs to be stressed. Consider the following


comment in this regard.

Remark 6: If we put Theorem 14 above, together with Theorem 9, Unit 8, what


do you see? We find that there are as many non-isomorphic groups as the
number of non-isomorphic permutation groups, i.e., infinite. Each isomorphism
class is [H], where H ≤ S(X) for some X.

Let us consider an example of representing a group using Cayley’s theorem.

Example 10: Find a subgroup of S4 to which the Klein 4-group K 4 is


isomorphic.

Solution: Consider the multiplication table for K 4 :

• e a b c
e e a b c
a a e c b
b b c e a
c c b a e

Now using the left representation function, let us see what Im f (in Theorem
14) gives us. Using this table, you can see that f e = I, since
276
Unit 9 Permutation Groups
f e( x ) = x ∀ x ∈ G.
Next, looking at the second row of the table, we see that
f a (e) = a , f a (a ) = e, f a (b) = c, f a (c) = b.
So f a = (e a )(b c).
Similarly, you should show that f b = (e b)(a c) and f c = (e c)(a b).
Hence, K 4 ~ {I, (e a )(b c), (e b)(a c), (e c)(a b)}.
Now, just replace the symbols e, a , b, c by 1, 2, 3, 4, and you’ll get Im f = V4
(in E30), sitting in S4 .
∴ K 4 ~ V4 .
***

The example above leads us to make the following observation.

Remark 7: Note that a subgroup of S(G ) need not have all the algebraic
properties of S(G ). So, for example, if G is abelian of order 6, S(G ) would
be non-abelian, but the subgroup of S(G )(= S6 here) to which G is isomorphic
− V4 , which
must be abelian. Thus, in Example 10, S4 is non-abelian, but K 4 ~
is abelian.

Try solving the following exercises now.

E35) Obtain the subgroup of S4 to which Z 4 is isomorphic. Is Z 4 ~ A 4 ?


Why, or why not?

E36) If G is a finite group of order n , then show that G is isomorphic to a


subgroup of Sn .

E37) For each of the following groups, find a subgroup of S8 to which it is


isomorphic:
i) D8 , ii) Z 8 , iii) U8 , iv) Q 8 .

With this we wind up our discussion on permutation groups. We also close our
discussion on group theory. In the next block you will start studying ring
theory. Of course, you will keep using what you have learnt in the first two
blocks, because every ring is a group also, as you will see.

So, let us see what you have studied in this unit.

9.6 SUMMARY
In this unit, we have discussed the following points.

1. A brief recap about permutations, in general, with the focus on the group
(Sn , o), in particular. Sn is a finite non-abelian group of order n!, for
n ≥ 3.

2. The definition, and some properties, of cycles and transpositions.

3. Any non-identity permutation in Sn can be expressed as a disjoint


product of cycles.
277
Block 2 Normal Subgroups and Group Homomorphisms
4. The set of all transpositions generates Sn , n ≥ 2. Also,
{(1 2), (1 3), K , (1, n )} generates Sn , n ≥ 2.

n
σ( j) − σ(i)
5. For n ≥ 2, the function sign : Sn → {1, − 1} : sign(σ) = ∏
i , j =1 ( j − i)
i< j

is an epimorphism.

6. Odd and even permutations.

7. A n , the set of even permutations in Sn , is a normal subgroup of Sn of


n!
order , for n ≥ 2.
2

8. A n is generated by the set of 3-cycles in Sn , for n ≥ 3.

9. A n is simple for n ≥ 5.

10. The only non-trivial proper normal subgroup of Sn (n ≥ 5) is A n .

11. Cayley’s Theorem: Each group is isomorphic to some permutation


group.

9.7 SOLUTIONS / ANSWERS


E1) In Unit 2, you have seen that (Sn , o) is a group ∀ n ∈ N.
Since S1 = {I}, and o(S2 ) = 2, these groups are abelian. Now let us look
at S3 .
⎛ 1 2 3 ⎞ ⎛ 1 2 3 ⎞ ⎛1 2 3 ⎞
Since ⎜⎜ ⎟⎟ o ⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ and
⎝ 2 3 1 ⎠ ⎝ 3 2 1 ⎠ ⎝1 3 2 ⎠
⎛ 1 2 3 ⎞ ⎛ 1 2 3 ⎞ ⎛ 1 2 3 ⎞ ⎛1 2 3 ⎞
⎜⎜ ⎟⎟ o ⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ ≠ ⎜⎜ ⎟⎟, these two
⎝ 3 2 1 ⎠ ⎝ 2 3 1 ⎠ ⎝ 2 1 3 ⎠ ⎝1 3 2 ⎠
permutations don’t commute.
∴ S3 is non-abelian.
Now, consider any Sn , n ≥ 3. Since (1 2 3) and (1 3) are in Sn , the
same argument above shows that Sn is non-abelian ∀ n ≥ 3.

E2) See Example 17, Unit 8.

E3) If x = y in G , then gx = gy, i.e., f ( x ) = f ( y). So f is well-defined.


Next, f is 1-1 because f ( x ) = f ( y) ⇒ gx = gy ⇒ x = y, by the left
cancellation law.
Also, for any x ∈ G , f (g −1x ) = x. So f is surjective.
Hence, f ∈ S(G ).

E4) There can be several answers.


Our answer is (1 2) and (2 4), (1 3 5) and (1 2 3), (1 2 3 4 5 6 7) and
(2 1 3 4 5 6 7).
278
Unit 9 Permutation Groups
Note that (1 2) ≠ (2 4), since (1 2) takes 2 to 1 and (2 4) takes 2 to
4.
Similarly, explain why the permutations in your answer are distinct.

E5) Here 1 → 5, 5 → 4, 4 → 1, 2 → 2, 3 → 3. Thus, this is the cycle (1 5 4).

⎛ 1 2 3 4 K 8⎞ ⎛ 2 1 3 4 K 8⎞
E6) i) (1 2) = ⎜⎜ ⎟⎟, (1 2) −1 = ⎜⎜ ⎟⎟,
⎝ 2 1 3 4 K 8⎠ ⎝ 1 2 3 4 K 8⎠
i.e., (1 2) −1 = (2 1) = (1 2).
⎛1 2 3 4 5 K 8⎞ ⎛1 2 3 4 5 K 8⎞
(1 2 3) o (3 2 1) = ⎜⎜ ⎟o⎜ ⎟
⎝2 3 1 4 5 K 8 ⎟⎠ ⎜⎝ 3 1 2 4 5 K 8 ⎟⎠
⎛1 2 3 K 8⎞
= ⎜⎜ ⎟ = I.
⎝1 2 3 K 8 ⎟⎠
Hence, (1 2 3) −1 = (3 2 1).

ii) f (i k ) = i k +1 and f ( x ) = x ∀ x ≠ i k , where k = 1, K , r, putting


i r +1 = i1.
So f −1 (i k +1 ) = i k and f −1 ( x ) = x ∀ x ≠ i k , where k = 1, K , r.
Thus, f −1 = (i r i r −1 K i 2 i1 ).

E7) Consider g −1h −1gh , where g = (1 2), h = (1 3) and g = (2 3), h = (2 4).


Then you should verify that (1 2)(1 3)(1 2)(1 3) = (3 1 2), and
(2 3)(2 4)(2 3)(2 4) = (4 2 3).
Note that (3 1 2) ≠ (4 2 3), since, for example, (3 1 2) takes 4 to 4,
and (4 2 3) takes 4 to 2.

E8) i) (1 5 3 2 4).

ii) Here 1 → 8 → 5 → 1. All the symbols haven’t been covered as yet,


for instance, 2 is left. So we look at 2. We get 2 → 4 → 2.
Still all the symbols haven’t been covered, one of them being 3.
So we look at 3. We get 3 → 7 → 6 → 3.
Now all the symbols are covered. Hence, the permutation is the
product (1 8 5)(2 4)(3 7 6).

iii) (1 4)(2 5).

⎛1 4 2 3⎞ ⎛ 6 5 2 4 3 1 7 8⎞
E9) (1 4)(2 3) = ⎜⎜ ⎟⎟, (6 5 2 4)(3 1) = ⎜⎜ ⎟⎟,
⎝ 4 1 3 2⎠ ⎝ 5 2 4 6 1 3 7 8⎠
⎛ 1 2 3 4 6 5 7 8⎞
(3 1 2)(4 6)(5 7) = ⎜⎜ ⎟⎟.
⎝ 2 3 1 6 4 7 5 8⎠

⎛1 2 3 4 5 6⎞
E10) No, because (1 3)(1 5 4) = ⎜⎜ ⎟⎟ and
⎝5 2 1 3 4 6⎠
⎛1 2 3 4 5 6⎞
(1 5 4)(1 3) = ⎜⎜ ⎟⎟.
⎝3 2 5 1 4 6⎠
279
Block 2 Normal Subgroups and Group Homomorphisms
Since, for example, (1 3)(1 5 4) takes 3 to 1 and (1 5 4)(1 3) takes 3 to
5, the two are different.

E11) Let f = (i1 i 2 K i r ). Then, as you have seen in the proof of Theorem 3,
f (i1 ) = i 2 , f (i 2 ) = f 2(i1 ) = i 3 ,K, f r −1(i1 ) = i r , f r(i1 ) = f (i r ) = i1.
Similarly, f r (i k ) = i k ∀ k = 2,K, r.
∴ f r = I.
Also, for s < r, f s (i1 ) = i s +1 ≠ i1. ∴ f s ≠ I.
∴ o(f ) = r.

E12) σ = (1 3 6 9)(2 5 7)(8 10).


∴ o(σ) = l.c.m(4, 3, 2)
= 12.

E13) Consider σ1 = (1 2) = σ 2 , in S3 . Then σ1σ 2 = I.


So o(σ1σ 2 ) = 1.
But l.c.m(2, 2) = 2.
Theorem 4 doesn’t fail here because it doesn’t apply here. It only applies
to a product of disjoint cycles. Here σ1 and σ 2 are not disjoint.

E14) No, e.g., see E12 for both questions.

E15) i) (1 4)(1 3).

ii) (4 1)(4 3)(= (1 4)(3 4)).

iii) (2 3)(2 5)(2 4).

E16) (i j)( j k ) = (i j k ) and ( j k )(i j) = (i k j) ≠ (i j k ).

E17) (1 5)(1 8)(2 4)(3 6)(3 7).

E18) (1 3 4) = (1 4)(1 3), (5 7) = (1 5)(1 7)(1 5),


(2 6 8) = (2 8)(2 6) = (1 2)(1 8)(1 2)(1 2)(1 6)(1 2) = (1 2)(1 8)(1 6)(1 2),
since (1 2)(1 2) = I.
∴ (1 3 4)(5 7)(2 6 8) = (1 4)(1 3)(1 5)(1 7)(1 5)(1 2)(1 8)(1 6)(1 2).

E19) For any three symbols i, j and k ,


(i j)( j k ) = (i j k ).
Then, if m is yet another symbol,
(i j k )(k m) = (i j k m), and so on.
∴ (1 2)(2 3) K (9 10)
= (1 2 3)(3 4) K (9 10)
= (1 2 3 4) K (9 10)
= (1 2 3K10)
= (1 10)(1 9) K (1 2).

E20) No. Since I is not a transposition, it doesn’t lie in this set.


280
Unit 9 Permutation Groups
n
I( j) − I(i) n
j−i
E21) sign I = ∏
i , j=1 j −1
=∏
i , j=1 j − i
= 1.
i< j i< j

σ(2) − σ(1) σ(3) − σ(1) σ(3) − σ(2) σ(4) − σ(3) σ(4) − σ(2)
E22) sign σ = ⋅ ⋅ ⋅ ⋅
2 −1 3 −1 3− 2 4−3 4−2
σ(4) − σ(1) σ(5) − σ(4) σ(5) − σ(3) σ(5) − σ(2) σ(5) − σ(1)
⋅ ⋅ ⋅ ⋅ ⋅
4 −1 5−4 5−3 5−2 5 −1
3 − 4 2 − 4 2 − 3 1− 2 1− 3 1− 4 5 −1 5 − 2 5 − 3 5 − 4
= ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ = 1.
1 2 1 1 2 3 1 2 3 4

E23) The permutation in E15(iii) is odd, because it is a product of 3


transpositions.
Similarly, explain why the others are odd or not odd.

E24) sign (f ) = sign (g) = −1.


∴ sign (f o g) = (−1)(−1) = 1.
∴ f o g is even.

E25) sign I = 1. ∴ I is even.


You could also have argued that I = (1 2)(1 2), and hence I is even.

E26) You have seen that σ = (i1 i 2 K i r ) = (i1 i 2 )(i 2 i 3 ) K (i r −1 i r ), a product of


(r − 1) transpositions.
Thus, σ is odd if r is even, and σ is even if r is odd.

4!
E27) You know that o(A 4 ) = = 12.
2
Now I ∈ A 4 . Then all the 3-cycles are in A 4 .
So (1 2 3), (1 3 2), (1 2 4), (1 4 2), (1 3 4), (1 4 3), (2 3 4), (2 4 3) are in
A4.
Then all the possible disjoint products of two transpositions lie in A 4 .
They are (1 2)(3 4), (1 3)(4 2), (1 4)(2 3).
So we have obtained all the 12 elements of A 4 .

Regarding commutativity, note that


(1 2 3)(1 2 4) = (1 3)(2 4) and (1 2 4)(1 2 3) = (1 4)(2 3).
Thus, (1 2 3) and (1 2 4) don’t commute. Hence, A 4 is not abelian.

E28) o(S5 A 5 ) = 2. One element is the coset A 5 , another is (1 2)A 5 , since


(1 2) ∉ A 5 .
Note that (1 2)A 5 = σA 5 for any odd σ in Sn , since σ −1 (1 2) ∈A 5 .

E29) (1 2) and (1 3) are odd permutations in Sn , but (1 2)(1 3) is an even


permutation.
Hence, this set is not a subgroup of Sn .

You could also have argued this by noting that I is even, and hence
doesn’t lie in this set. Hence, this set is not a subgroup of Sn .
281
Block 2 Normal Subgroups and Group Homomorphisms
E30) From Example 5, you know that σνσ −1 has the same disjoint cycle
decomposition structure as that of ν ∀ ν ∈ V4 and σ ∈ A 4 .
Also V4 contains all elements that are products of two disjoint
transpositions in S4 .
Hence, σνσ −1 ∈ V4 ∀ σ ∈ A 4 , ν ∈ V4 .
Thus, V4 A4.
Since V4 : H = 2, H V4 .
Now (1 2 3)(1 2)(3 4)(1 2 3) −1 = (1 4)(2 3) ∉ H. Hence, H A4.

E31) We will use a counting argument to find this. The elements in A 5 are
3-cycles, 5-cycles and products of 2 disjoint transpositions.

i) The only even permutations of order 2 in S5 are those of the form


(1 2)(3 4), using 4 distinct symbols.
Also (1 2) = (2 1) and (1 2)(3 4) = (3 4)(2 1).
Hence, the total number of such elements in A 5 is
1 ⎡5× 4 3× 2⎤
= 15.
2 ⎢⎣ 2 2 ⎥⎦

ii) The only even permutations of order 3 are 3-cycles.


Also (1 2 3) = (2 3 1) = (3 1 2).
1
Hence, the number of distinct 3-cycles in S5 is (5 × 4 × 3) = 20.
3

iii) This is the number of distinct 5-cycles in S5 , i.e.,


1
(5 × 4 × 3 × 2 × 1) = 24.
5

iv) Since we have already found 59 elements of A 5 , and o(A 5 ) = 60,


these are all the non-trivial elements of A 5 . Hence, A 5 has no
element of order 15.

E32) We will use Example 5 to calculate the elements.

i) Let i ≠ 2. If σ1 = (1 i 3), then the given element is


σ1 (1 2 3)σ1−1 = (σ(1) σ(2) σ(3)), from Example 5.
= (i 2 1) ∈ H, since H A n , and σ1 ∈ A n .
If i = 2, then σ1 = (1 2 3). Hence, σ1 (1 2 3)σ1−1 = (1 2 3) ∈ H.

ii) Here σ 2 = ( j 1 i), and the given element is


σ 2 (1 i 2)σ 2−1 = (i j 2) ∈ H, using (i) above, and since σ 2 ∈ A n .
Here we have assumed j ≠ 1, 2.
As in (i), if j = 2, then σ 2 = (1 i 2), so that
σ 2 (1 i 2)σ −21 = (1 i 2) ∈ H.

iii) Here σ3 = (1 2 k ), and the given element is


σ3 (i j 2)σ3−1 = (i j k ) ∈ H, using (ii), and since σ3 ∈ A n .
282
Unit 9 Permutation Groups
Here too i, j, k ≠ 1, 2.

Hence, for any (i j k ) ∈ A n , (i), (ii), (iii) tell us


(i j k ) = (σ3σ 2 σ1 )(1 2 3)(σ3σ 2 σ1 ) −1 ∈ H.
Thus, every 3-cycle is in H. Hence, by Theorem 11, H = A n .

E33) Ker f A 7 . Hence, Ker f = {I} or Ker f = A 7 .


If Ker f = {I}, then f is 1-1, and f (A 7 ) ≤ G , where
o(f (A 7 )) = o(A 7 ) = 2520.
∴ o(G ) ≥ 2520.
If Ker f = A 7 , then f ( x ) = I ∀ x ∈ A 7 .

E34) Let f : S5 → U5 be a group homomorphism.


Then Ker f S5 . So Ker f = {I}, or Ker f = A 5 , or Ker f = S5 .
If Ker f = {I}, then S5 ~
− Im f ≤ U5 .
But then o(Im f ) = 5! and o( U5 ) = 5.
So we reach a contradiction.
Hence, Ker f ≠ {I}.

If Ker f = A 5 , then Im f ~
− (S5 A 5 ). Hence, Im f ≤ U5 of order 2. But
2 o( U5 ). So, by Lagrange’s theorem, this case is not possible.

Hence, the only possibility is Ker f = S5 , i.e., f : S5 → U5 : f (σ) = 1.

E35) You know that Z 4 = < 1 > and o( 1 ) = 4. Therefore, the subgroup of S4
isomorphic to Z 4 must be cyclic of order 4.
Further, it is generated by the permutation f 1 .
Now f 1 ( x ) = 1 + x ∀ x ∈ Z 4 .
∴ f 1 = ( 1 2 3 0 ), which is the same as the cycle (1 2 3 4).
∴ Z4 ~ − < (1 2 3 4) >, which is certainly not isomorphic to A 4 , as
(1 2 3 4) ∉ A 4 . Also note that A 4 is not cyclic. (Why?)

E36) Let G = {g1 , g 2 , K , g n }. Then S(G ) is the set of permutations on n


symbols. Hence, S(G ) = Sn .
− Im f ≤ Sn , where f is as in the proof of Cayley’s theorem.
Thus, G ~

E37) i) D8 = < {r, R r 2 = I, R4 = I, rR = R−1r} >.


Hence, D 8 must be isomorphic to a subgroup generated by
{σ1 , σ 2 σ12 = I, σ 42 = I, σ1σ 2 = σ 2−1σ1}.
Recall also, from Unit 2, that if we take r and R to be the
reflection and rotation shown in Fig.3 there, then σ1 = (2 4) and
σ 2 = (1 2 3 4).
You should also check that all the required conditions are satisfied.
− < (2 4), (1 2 3 4) > ≤ S8 .
So D8 ~

ii) − < (1 2 3 4 5 6 7 8) > ≤ S8 .


As in E35, Z 8 ~
283
Block 2 Normal Subgroups and Group Homomorphisms
iii) − Z8 , U8 ~
Since U8 ~ − < (1 2K8) > ≤ S8 .

iv) Verify that the Cayley table of Q8 is as below.

• I −I A −A B −B C −C
I I −I A −A B −B C −C
−I −I I −A A −B B −C C
A A −A −I I C −C −B B
−A −A A I −I −C C B −B
B B −B −C C −I I A −A
−B −B B C −C I −I −A A
C C −C B −B −A A −I I
−C −C C −B B A −A I −I

So, by looking at each row of the table, you can see that f I = IG ;
f − I ( x ) = − x ∀ x ∈ Q8 . So, f − I = (I − I)(A − A)(B − B)(C − C).
If we replace I, − I, A, − A, B, − B, C, − C by 1, K ,8, respectively,
in f − I , we get f − I = (1 2)(3 4)(5 6)(7 8).
Similarly, you should check that
⎛1 2 3 4 5 6 7 8⎞
fA = ⎜⎜ ⎟⎟ = (1 3 2 4)(5 7 6 8) ,
⎝ 3 4 2 1 7 8 6 5⎠
f − A = (1 4 2 3)(5 8 6 7),
f B = (1 5 2 6)(3 8 4 7),
f − B = (1 6 2 5)(3 7 4 8),
f C = (1 7 2 8)(3 5 4 6),
f − C = (1 8 2 7)(3 6 4 5).
∴ Q8 ~ − {f x x ∈ Q8 } ≤ S8 .

284

You might also like