A Crash Course In Group Theory (Version 1.

0)
Part I: Finite Groups
Sam Kennerly
June 2, 2010
with thanks to Prof. Jelena Maricic, Zechariah Thrailkill, Travis Hoppe,
Erica Caden, Prof. Robert Gilmore, and Prof. Mike Stein.

Contents
1 Notation

2 Set Theory

3 Groups
3.1 Definition of group
3.2 Isomorphic groups
3.3 Finite Groups . . .
3.4 Cyclic Groups . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

7
. 7
. 8
. 9
. 10

4 More Groups
4.1 Factor Groups . . . . . . . . . . .
4.2 Direct Products and Direct Sums
4.3 Noncommutative Groups . . . . .
4.4 Permutation Groups . . . . . . .

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

12
12
14
15
19

Why learn group theory? In short, the answer is: group theory is the
systematic study of symmetry. When a physical system or mathematical
structure possesses some kind of symmetry, its description can often be dramatically simplified by considering the consequences of that symmetry. Results from group theory can be very useful if (and only if) one understands
them well enough to look them up and use them.
The purpose of these notes is to provide readers with some basic insight
into group theory as quickly as possible. Prerequisites for this paper are
the standard undergraduate mathematics for scientists and engineers: vector
calculus, differential equations, and basic matrix algebra.
Simplicity and working knowledge are emphasized here over mathematical completeness. As a result, proofs are very often sketched or omitted in
favor of examples and discussion. Readers are warned that these notes are
not a substitute for a thorough study of modern algebra. Still, they may be
helpful to scientists, engineers, or mathematicians who are not specialists in
modern algebra. Readers who desire an in-depth study of the subject may
find this document useful as an outline and/or a quick-reference guide.
Like any good mathematical game, group theory is almost cartoonishly
simple at first but the most advanced results are nightmarishly complicated.
For example, a team of mathematicians recently found all irreducible unitary
representations of the exceptional Lie group E8 . The result was stored as
one 453,060 x 453,060 matrix of polynomials requiring 60GB of disk space.
These notes will not attempt any such task. Rather, the goal is to summarize the most important definitions and results while keeping an eye out
for possible uses in the physical sciences. Part I covers finite groups: groups
with only a finite number of elements. Part II will cover continuous groups,
but at the time of this writing, it is not finished yet.

Notation

Sections 1 and 2 are intended as background for those unfamiliar

with formal logic and mathematical notation. Impatient readers
are advised to skip directly to Section 3: Groups and refer back to
these sections later as necessary.
Notational conventions used in this document:
Bold words indicate new definitions.
Italic words indicate terms defined later in the text.
Quote marks identify informal terminology.
Small capital letters indicate informal definitions that are
unconventional and/or lacking in mathematical rigor.
Set theory notation:
Z
the set of all integers
N, Q, R, C
the sets of natural, rational, real, and complex numbers.
2
R
the set of all ordered pairs of real numbers, e.g. (-4.5, 7)
R3
the set of all ordered triplets of real numbers, e.g. (1, 2, 3)
n
R
the set of all ordered n-tuplets of real numbers
xS
x is an element of the set S
S1 S2
the set of elements in S1 or S2 or both
S1 S2
the set of elements in S1 and S2
S {x, y, z} the set of all elements in S except x, y, and z.
Formal logic notation:

for all
there exists
such that
implies
is logically equivalent to
is proportional to
is isomorphic to

A slash through a symbol means the negation of that symbol:

6 there does not exist
6

does not imply

3

Analysis notation:
|x|
absolute value of x: the positive square root of x2
[a, b] the real numbers between a and b including the endpoints: a x b
(a, b) the real numbers between a and b not including the endpoints: a < x < b
Linear algebra notation:

M
Mjk
T
M

M
|xi
hx|
hy|xi
||x||2
|xi
M
or

a matrix

the element on the jth row and kth column of M

(formed by replacing every Mjk with Mkj )
the transpose of M
(formed by complex conjugating every element of M
)
the conjugate of M

the matrix adjoint to M (formed by transposing M )

a vector (also represented as a column of components)
the dual vector to |xi (also represented as a row of components)
the inner product of hy| and |xi
the square magnitude of a vector x (also represented as hx|xi)
on the column vector |xi
the vector formed by operating the matrix M
direct sum or direct product of two groups

Some practice for readers unfamiliar with formal logic notation:

mathematical notation

English translation

6 free lunch

There is no such thing as a free lunch.

(x {things that glitter}) 6 (x = gold)

Not all that glitters is gold.

(U thinks) U

U thinks, therefore U exists.

x Z, (x > 0) (x N)

All positive integers are natural numbers.

(z Z : x 2 = z) (x 2Z)

x divided by 2 is an integer if and

only if x is even.

Set Theory

Set theory is an important and fascinating subject which, for the purposes
of this document, we will almost completely ignore. A sincere and thorough
investigation would lead us to deep problems in axiomatic mathematics
about which the author is more or less ignorant.
In the interest of speed, we will make use of the following informal
definitions. Readers interested in a rigorous discussion of these terms may
want to consult a book on axiomatic set theory.
Definition 1. A set is a collection of things or ideas.1 The individual
contents of a set S are called elements of S. Sets are often indicated by
writing their elements, or a description of their elements, in brackets.
Definition 2. The empty set has no elements. It is denoted . Note that
is not the same as {0}, the set containing only the element 0.
Definition 3. A subset of a set S is a set whose elements are all elements
of S. Note that is a subset of any set S and that any set S is a subset of
itself. S and are called improper subsets of S; any other subset of S is
proper. T is a proper subset of S is denoted T S.
Definition 4. A map : A B is a method for associating elements of A
to elements of B.
Definition 5. The image of a map is the set of all (a) where a A.
Definition 6. A map from A to B is onto if every element in B is
mapped to by at least one element of A: b B, a A such that
b = (a). is called a surjection.
Definition 7. A map from A to B is one-to-one if no two or more
elements in A are mapped to any one element in B: (a) = (b) a = b.
is called an injection.
Definition 8. A map from A to B is invertible if and only if every
element of A is mapped to exactly one element of B and every element of B
is mapped to by an element of A. In other words, an invertible map is
one-to-one and onto. Invertible maps are also called bijections.
1

Bertrand Russell used this definition in 1901 to construct the following paradox: Define
S as { all sets which are not elements of themselves }. Now (S is an element of itself)
(S is not an element of itself). Russells Paradox is a disaster for my definition of set.

If is invertible, then we can always define an inverse map 1 such that

(a) = b a = 1 (b) for all a A and b B.
Examples and counterexamples:
(x) = 2x 3 is an invertible map from R to itself. If we define
1 (y) = 21 (y + 3), then y = (x) x = 1 (y) for all real x and y .
(x) = x2 is not an invertible map from R to itself. 1 is ambiguous
because is not one-to-one: (2) and (2) are both 4, so 1 (4) could
be either. also fails to be onto: it misses all the negative numbers.
(x) = x2 is an invertible map from R+ to itself. (R+ is the set of all

positive real numbers.) Its inverse map is of course 1 (y) = y.

Groups

3.1

Definition of group

Definition 9. A binary operator from A to B maps an ordered pair of

elements of A to one element of B. (A trinary operator maps an ordered
triplet of elements of A to one element of B, and so on.)
Definition 10. A binary operator is said to be closed on A if and only if
(a, b) A for all a, b A. (Equivalently, the image of is a subset of A.)
You are probably familiar with many closed binary operators:
+ is a closed binary operator on Z: 2 + 3 = 5
is a closed binary operator on Q: 13 35 = 15
(the vector cross product) is a closed binary operator on R3 :
(a, b, c) (x, y, z) = (bz cy, cx az, ay bx)
Examples of binary operators that are not closed:
is not closed on Z: 5 2 = 52 , which is not an integer
is not closed on R2 : (a, b) (x, y) = ax + by, which is not an element of R2
Definition 11. A group (G, ?) consists of a set G and a binary operator ?
defined in such a way that the following four rules are true:
0)
1)
2)
3)

? is closed on G: if a, b G, then (a ? b) G
? is associative: if a, b, c G, then (a ? b) ? c = a ? (b ? c)
G contains the identity of ?: e G such that a G, (a ? e) = a
Inverses exist: a G, z G such that (a ? z) = e

Examples and counterexamples:

(Z, +) is a group:
1) (a + b) + c = a + (b + c) for any integers a, b, c
2) a + 0 = a for any integer a (0 is called the additive identity)
3) for every integer a, there is an integer a such that a + (a) = 0
(Z, ) is not a group:
(1 2) 3 6= 1 (2 3). Subtraction is not associative, so rule 1 fails.
(R3 , ) is not a group:
There is no cross product identity vector in R3 , so rule 2 fails.
(Q, ) is almost a group, but not quite:
7

1) (a b) c = a (b c) for all rationals a,b,c

2) a 1 = a for all rationals a (1 is called the multiplicative identity)
3) 0 does not have a multiplicative inverse in Q: 01 6 Q.
This looks silly, but its important! Every operation in a group can be
undone with an inverse operation. Multiply by zero is somehow too
destructive; no inverse operation exists. For practice, check that we can
form a group by removing 0 from Q: (Q {0}, ) is a group.
WARNING: It is very common to refer to the group (Z, +) as
the group Z. Remember that a set without an operation is not a
group! Sometimes its obvious what is meant; Z can form a group
under addition but not subtraction, multiplication, or division.

3.2

Isomorphic groups

Define the set 2Z = { all even integers }. Is (2Z, +) a group?

First check that + is closed on 2Z: a, b Z, 2a + 2b = 2(a + b) 2Z
1) + is still associative.
2) 0 2Z, so the identity element exists.
3) 2a + (2a) = 0, so additive inverses exist.
As a group, (2Z, +) behaves identically to (Z, +). All we did was relabel
the nonzero elements: rename 1 2 , 2 4 , etc. The elements still
behave the same way: 6 + 7 = 13 in (Z, +) and 12 + 14 = 26 = 13
renamed in (2Z, +).
In general, to show that two groups behave the same, we must show there is
a relabeling map from one to the other that does not change the group
structure. In this case, (a) = 2a is a relabeling map from Z to 2Z.
Definition 12. A group homomorphism is a map from G to H
compatible with the group operations of (G, ?) and (H, ) :
(g1 ) (g2 ) = (g1 ? g2 ) for all g1 , g2 G.
Definition 13. A group isomorphism is an invertible group
homomorphism. If is invertible, then 1 is an isomorphism from (H, )
to (G, ?). (To prove this, define h1 = (g1 ), etc. and use the definitions of
invertible map and group homomorphism.)
Definition 14. Two groups (G, ?) and (H, ) are isomorphic to each
other if a group isomorphism exists between them. Isomorphic groups in
this document are denoted (G, ?)
= (H, ) or sometimes G
= H.
8

In our example, was a map from Z to 2Z. Both groups used + as their
operation and (a) = 2a is compatible: 2a + 2b = 2(a + b) a, b Z. is
also invertible, so we conclude that the groups are isomorphic: Z
= 2Z.
Mathematicians view isomorphic groups as identical structures with
different labels on their elements and/or a different name for the operator.
We can operate two elements of G together, or we can operate their
relabeled counterparts in H, un-label the result back to G, and get the
same answer: g1 ? g2 = 1 ((g1 )(g2 )). Invertibility is important if we
want to say two groups are the same.

3.3

Finite Groups

Definition 15. If (G, ?) is a group, the order of this group is |G|, the
number of elements in G. If |G| N, G is called a finite group.
The groups (Z, +) and (Q {0}, ) are not finite. Lets find some that are.
Definition 16. Define modular addition for any n N: let a + b (mod
n) denote the remainder when a + b is divided by n. Define Zn as the set
of integers {0 ... (n 1)}.
Note that a + b (mod n) is a closed binary operation on Zn . Examples:
6 + 9 (mod 10) = 5 , 2 + 2 (mod 4) = 0 , 2100 + 1 (mod 2) = 1.
For any n N, ( Zn , + (mod n) ) is a finite group.
1) + (mod n) is associative. (a + b) + c = a + (b + c)
2) 0 is the identity. a + 0 = a
3) the inverse of a is n a. [a + (n a)] (mod n) = n (mod n)= 0.
4) |Zn | is a natural number by definition, so Zn is finite.
Example: Z4 = { 0, 1, 2, 3 } has the following structure table:2
How to read this table:
+ (mod 4)
0
1
2
3
2

0
0
1
2
3

1
1
2
3
0

2
2
3
0
1

3
3
0
1
2

To find 1 + 2, find 1 in the left column and 2 in

the top row. The result here is 3. Notice that
the table is symmetric about its diagonal - this
will not be true for all groups! Remember that in
(mod 4) addition, 2+2 = 0, 2+3 = 1, and 3+3 = 2.

Also known as a Cayley table. Arthur Cayley originally called them squares.

3.4

Cyclic Groups

The cyclic groups Zn are especially straightforward: for any n N, we can

generate the entire group Zn by operating 1 on itself repeatedly. Before
introducing them, we pause for some new notation:
Definition 17. For any element a in a group (G, ?), define the powers of
a: an = (a ? a? . . . ?a) where a appears in parentheses n times (and n N).
WARNING: The notation (a)(b) or ab is very common shorthand for a ? b in group theory. This will look completely ridiculous when the operator being represented is addition: (1)(3) means
(1 + 3), 25 means (2 + 2 + 2 + 2 + 2), and so on. We adopt power
notation for group operations now because it will be far more useful
for matrix multiplication and symmetry transformations later.
Definition 18. Let a be an element of a group (G, ?) and let d denote the
identity of (G, ?). The smallest natural number k such that ak = d is called
the order of element a and is denoted |a| = k.
Definition 19. If the set {a, a2 , . . . , ak } contains every element of G, then
a is said to generate (G, ?). If so, the order of the group G will be the
same as the order of a. |G| = |a| = k.
Not all groups can be generated by repeated operations of the same element
with itself. Groups that can be generated this way are called cyclic:
Definition 20. A finite group G of order n is called cyclic if it can be
written {a, a2 ,. . . , an } for some element a G.
Examples of cyclic groups:
The group Z4 can be written {1, 12 , 13 , 14 } = {1, 2, 3, 0}. (As expected,
power notation looks absurd when + is the operation were denoting.)
Cut a square out of cardboard and place it in front of you. Define  as the
operation rotate the square 90 counterclockwise. Define n to mean do
 n times. {, 2 , 3 , 4 } is a cyclic group. The group operation is rotation
composition and the identity is 4 . This group is isomorphic to Z4 .
If z C and z n = 1, z is called an nth root of unity. For any n N, the
nth roots of unity form a group under multiplication. For example, the
fourth roots of unity {1, , 1, } are generated by :
10

1 = , 2 = 1, 3 = , 4 = 1. This group is isomorphic to Z4 .

If youve ever heard someone refer to as a 90 rotation, this is what
he/she meant! Generators are important and will show up later in new
clothes as Lie algebras. For practice, find the 6th complex roots of unity.
(Hint: Draw 30 -60 -90 triangles in the complex plane.)
You may have already conjectured the following theorem:
Theorem 1. Every cyclic group of order n is isomorphic to Zn .
Proof : Construct an isomorphism explicitly. (G, ?) is a cyclic group of
order n G = {a, a2 , . . . , an } for some a G. Define (a) = 1,
(a2 ) = 2, . . . , (an ) = n. Then is compatible: for any j, k N,
(aj ? ak ) = (aj+k ) = j + k (mod n) = (aj ) + (ak ) (mod n) . Each
element of G is mapped to exactly one element of Zn and vice versa, so is
invertible. Therefore is an isomorphism.
Mathematicians refer to all three of the examples above as the cyclic group
of order 4. Each is isomorphic to the others regardless of what symbols are
used to denote the elements or the group operation. The notation
Zn = {a, a2 , . . . , an } is used here for any cyclic group of order n.
It is worth noting that there is another group of order 4 which is not cyclic.
Alternately called the Klein group, the Vierergruppe,3 or Z2 Z2 , it is
the smallest noncyclic group. All groups of order 4 are isomorphic to either
Z4 or the Klein group. Its structure table is:
?
1
i
j
k

1
1
i
j
k

i
i
1
k
j

j
j
k
1
i

k
k
j
i
1

Notice that every element is its own inverse and order

of operations does not matter. The Klein group is
isomorphic to the direct sum group Z2 Z2 , to be
defined in the next section.

Evariste
Galois showed that this group guarantees the solvability of quartic
(fourth-order) polynomials. Galois research into higher-order polynomials
is often considered the beginning of modern group theory.

German for four-ish group, or something like that. Klein is named for Felix Klein
and is coincidentally the German word for small.

11

More Groups

Notice what happens if we try to generate Z4 from the wrong element:

0 + 0 = 0 generates the set { 0 }
2 + 2 = 0, 0 + 2 = 2 generates the set { 0, 2 }
3 + 3 = 2, 2 + 3 = 1, 1 + 3 = 0 generates the set { 2, 1, 0, 3 } = Z4
Apparently 3 is a perfectly good generator! The other two have missed
some elements, but they have generated miniature groups of their own.
Definition 21. If (G, ?) is a group, H is some subset of G, and (H, ?) is a
group, then (H, ?) is called a subgroup of (G, ?). If H is a proper subset of
G, then (H, ?) is a proper subgroup of (G, ?).
To prove that a subset H forms a subgroup, use the Subgroup Test: show
that (a, b H) (a ? b1 H). Note that we already know ? is associative.
Be sure that H contains the identity of (G, ?) so that a ? a1 H.
Example: { 0, 2 } is a group under the operation + (mod 4). It is a proper
subgroup of Z4 and is isomorphic to Z2 .
Example: { 0 } is a group under + (mod 4). It is isomorphic to the group
Z1 (often called the trivial group). Any group with one element is
(trivially) cyclic and therefore isomorphic to Z1 .
Theorem 2. Every subgroup of a cyclic group is cyclic. Zj is isomorphic
to a subgroup of Zn if and only if j is a factor of n.
Theorem 3. If (G, ?) is finite and (H, ?) is a subgroup, then |H| is a
factor of |G|. The order of every element in G is a factor of |G|. [Szekeres]
We will see many more examples of subgroups in the next sections.

4.1

Factor Groups

Definition 22. Let (H, ?) be a subgroup of (G, ?). For any element a G,
define the left coset aH to be the set {a ? h0 , a ? h1 , a ? h2 , . . . } where h0 is
the identity and the other hj are all the other elements of H.
Example: H = {0, 2} is a subgroup of Z4 . Its left cosets are
0H = {0, 2}, 1H = {1, 3}, 2H = {2, 0}, 3H = {3, 1} .

12

In this example, 0H and 2H are different names for the set {0, 2} and 1H
and 3H are different names for the set {1, 3}. Also notice that the cosets of
H partition G; they split it into sets that do not overlap. Keep in mind
that each coset of H is a subset of G, but not necessarily a subgroup of
(G, ?). In our example, {1, 3} is not a subgroup of Z4 .
Definition 23. Define right cosets of H in G: Ha = {a, h1 ? a, h2 ? a, . . . }.
A subgroup H is normal if aH = Ha for all a G.
The left and right cosets of {0, 2} in Z4 are equal because the operation +
(mod n) is commutative: a + b = b + a always. Operations that do not
commute can sometimes cause abnormal subgroups.4
If a group (G, ?) has a normal subgroup (H, ?), we can try to factor out
the behavior of H to form a simpler group. This new group will have cosets
of H as its elements and a new operation  defined as follows:
Definition 24. Let (G, ?) have a normal subgroup (H, ?). Define the
coset operation  : aH  bH = (a ? b)H = {a ? b, a ? b ? h1 , a ? b ? h2 , . . . }
where a, b are any two elements of G and the hj are all the elements of H.
The cosets of H under the operation  form a group called the factor
group G/H, pronounced G modulo H or G slash H.
Example: The cosets of H = {0, 2} in Z4 are 0H and 1H from above.
{0H, 1H} under  forms the factor group Z4 /H. Its structure table is:

0H
1H

0H
0H
1H

1H
1H
0H

0H
0H
1H
1H

 0H
 1H
 0H
 1H

= (0 + 0)H
= (0 + 1)H
= (1 + 0)H
= (1 + 1)H

= 0H
= 1H
= 1H
= 2H = 0H

This structure table looks exactly like the table for Z2 , so Z4 /H is

isomorphic to Z2 . Earlier we showed that H itself is also isomorphic to Z2 .
Abusing our notation somewhat, we can write Z4 /Z2
= Z2 .
Definition 25. Every group (G, ?) contains itself and Z1 as factor groups.
If no other factor group can be defined, then (G, ?) is a simple group.
Simple is a dangerous term. A monster group has been discovered
which is simple and of order 808,017,424,794,512,875,886,459,904,961,710
757,005,754,368 billion. In retrospect, prime might have been a better
name for an unfactorable group.
4

A noncommutative group with no abnormal subgroups is called Hamiltonian after

Sir William Rowan Hamilton and his famous example, the quaternion group Q8 .

13

4.2

Direct Products and Direct Sums

Factor groups can be used to break a group apart into simpler groups.
Similarly, we can define direct product groups that combine two groups to
make a more complicated one:
Definition 26. Given two groups (G, ?) and (H, ), a new group called
the direct product group (G, ?) (H, ) can be constructed. The
elements are ordered pairs (g, h) where g G and h H and the operation
is defined (g1 , h2 ) (g2 , h2 ) = (g1 ? g2 , h1 h2 ).
Example: R+ , the set of all positive real numbers, forms a group under
multiplication. The set R+ R+ consists of all ordered pairs (x, y) where x
and y are positive real numbers. Define (x, y) (a, b) = (xa, yb) to form the
group (R+ R+ , ). (1, 1) is the identity and (x, y)1 = (x1 , y 1 ).
The direct product group G H will always contain a normal subgroup
isomorphic to G and another one isomorphic to H. To see this, let 1g and
1h denote the identities of G and H and consider the sets {(g, 1h )} and
{(1g , h)} where g is any element of G and h is any element of H. These two
sets form subgroups isomorphic to G and H respectively.
Definition 27. If the operations of G and H are denoted by addition
(rather than multiplication or some other symbol), the term direct sum is
used instead of direct product. Direct sums are denoted G H. It is
conventional to reserve the symbol + only for commutative operations.
Notice that Z2 Z2 consists of the elements {(0, 0), (1, 0), (0, 1), (1, 1)}. If
the elements are renamed {1, i, j, k}, their structure table is identical to
that of the Klein group from before and so Klein
= Z2 Z2 .
WARNING: Since we know that Z4 /Z2
= Z2 , it is very tempt
ing to write Z4 = Z2 Z2 . Dont do it! Not only are we using bad
notation by representing additive groups with , but Z2 Z2 is not
isomorphic to Z4 . Direct products do not undo factor groups.
In the next section we will see that D3 /Z3
= Z2 and Z6 /Z3
= Z2 despite
the fact that D3 and Z6 are not isomorphic. While there is a close analogy
between factoring groups and factoring natural numbers, be careful not to
take the analogy too far!

14

4.3

Noncommutative Groups

So far, every group operation we have seen obeys the commutative

property: a ? b = b ? a for any a, b G. Not all groups behave this way.
Definition 28. A group (G, ?) in which a ? b = b ? a for all a, b G is
called commutative or Abelian.5
Theorem 4. All cyclic groups are commutative.
Proof : Any cyclic group (G, ?) has a generator a such that every element
of G can be written as aj for some j N. Thus the result of any operation
can be written as aj ? ak = aj+k = ak+j = ak ? aj for some j, k N.
Essentially, a commutes with itself and therefore so do the elements it
generates. Noncyclic groups can also be commutative, but many are not.
Example: The set of all 3x3 real matrices with nonzero determinant form a
group under the operation of matrix multiplication. (Any square matrix
has a multiplicative inverse if and only if its determinant is not zero.) This
group is called GL(3, R), the General Linear 3-dimensional real group.
Some elements of GL(3, R) commute, but these two certainly dont:

0 1 0
0 0 1
0 1 0
1 0 0 0 1 0 = 0
0 1
0 0 1
1 0 0
1 0 0
Now multiply the first

0 0 1
0
0 1 0 1
1 0 0
0

two matrices in reverse

1 0
0 0

=
0 0
1 0
0 1
0 1

order:

1
0
0

Reversing the order of operations does not give the same result!
Experiment: The Special Orthogonal group SO(3) of all possible rotations
in a 3-dimensional vector space is noncommutative. Find an object that is
not perfectly symmetrical and choose three directions x,y, and z that are
perpendicular to each other. Rotate an object 90 counterclockwise as seen
by someone looking in the positive-y direction. Now rotate the object 90
counterclockwise as seen by someone looking in the positive-z direction.
5

Named after Norwegian mathematician Niels Henrik Abel, pronounced ah-bell.

15

Draw a sketch of the object. Return the object to its original position and
repeat, but this time do the z-axis rotation first. Is the object oriented the
same way as before? (Hint: No.)
SO(3) is extremely important in theoretical physics and chemistry, but for
now, we will set it aside and practice on finite groups. The symmetry
groups of regular 2D polygons are called the dihedral groups Dn and are
never commutative.6 The smallest noncommutative group (i.e. the one with
the fewest elements) is D3 , the symmetry group of an equilateral triangle.
Experiment: Construct a cardboard model of a triangle and pretend it is
absolutely perfect, with equal 60 angles and equal sides. Label the vertices
A,B, and C. Set the triangle on the desk in front of you like this:
A
B

T
 TT

If the triangle is left flat on the table, it can be rotated 120 , 240 , or 360 ,
or integer multiples of those. Name those transformations {r, r2 , r3 }. Since
r3 is equivalent to dont touch the triangle, lets rename it the identity
transformation and write r3 = 1.
If we allow operations in three dimensions, the triangle can also be flipped
in three different ways. Flip over the triangle so that the top point is
unmoved and call this transformation t (for top vertex). Flips around the
other two vertices will be named u and v as shown below:

B

T

T

T
 TT

T
 TT

T
 TT

T
 TT

TT

identity

C A

TT

B C

r2

A C

B B

A A

Now define (a . b) to mean do transformation b, then do transformation a

to the result. (The notation may look backwards, but it will be useful for
matrix transformations later.) The set of all six transformations
{1, r, r2 , t, u, v} forms a group under the operation ., called transformation
composition. The group is called D3 and its structure table is:
6

We could define commutative groups D1 and D2 , but no 1- or 2-sided polygons exist.

16

.
1
r
r2
t
u
v

1
1
r
r2
t
u
v

r
r
r2
1
u
v
t

r2
r2
1
r
v
t
u

t
t
v
u
1
r2
r

u
u
t
v
r
1
r2

v
v
u
t
r2
r
1

This structure table is not symmetric

because D3 is not a commutative group.
Be sure to read the table in the correct
order: the result of r . t is listed in row
r, column t and it is v. Check the accuracy of the table by drawing triangles
or using the cardboard model.

For practice, we will now find all subgroups and factor groups of D3 . Notice
that the product of two rotations is always another rotation: {1, r, r2 } looks
promising as a potential subgroup. Apply the Subgroup Test: for any two
rotations a and b, is a . b1 a rotation? Yes! The inverse of r is r2 , the
inverse of r2 is r, and the product of two rotations is another rotation.
Also, the identity is contained in {1, r, r2 }, so {1, r, r2 } is a subgroup. In
fact, its a cyclic group of order 3 because r generates {1, r, r2 }.
Stare at the structure table for D3 and look for other subgroups. {1, t} is a
subgroup: it includes 1 and it passes the Subgroup Test. Similarly, {1, u}
and {1, v} are each subgroups. Because there are only 6 elements in D3 , its
not hard to see that no other proper subgroups exist. ({1} is not proper!)
Are these subgroups normal? If so, we can use them to build factor groups.
Look at the rotation subgroup first: denote {1, r, r2 } as H and use the
structure table to find its left and right cosets aH and Ha for all a G:
1H = {1, r, r2 }

rH = {r, r2 , 1}

tH = {t, tr, tr2 } = {t, u, v}

vH = {v, vr, vr2 } = {v, t, u}

r2 H = {r2 , 1, r}

uH = {u, ur, ur2 } = {u, v, t}

The left cosets are apparently 1H = {1, r, r2 } and tH = {t, u, v}; the other
left cosets are each equivalent to one of these. We have partitioned D3 into
a rotation coset 1H and a flip coset tH. Check the right cosets H1 and
Ht to see if H is normal:
H1 = {1, r, r2 } = 1H (of course! 1 commutes with everything.)
Ht = {t, rt, r2 t} = {t, v, u} = tH.
H is a normal subgroup, so we can define the factor group D3 /H by using
the coset operation  as before: aH  bH = (ab)H = {ab1, abr, abr2 }.
17

Written out completely, we have:

1H  1H = 1H
1H  tH = tH
tH  1H = tH
2
2 2
2 2
tH  tH = (t )H = {t , t r, t r } = {1, r, r2 } = 1H
The structure table of D3 /H looks like either of these:

1H
tH

1H
1H
tH

tH
tH
1H


rotations
flips

or

rotations
flips
rotations
flips
flips
rotations

The factor group D3 /H has only two elements, so it is isomorphic to Z2 . H

is known to be isomorphic to Z3 , so we feel justified writing D3 /Z3
= Z2 .
Exercise: Show that these subgroups of D3 are not normal: {1, t}, {1, u},
{1, v}. These subgroups cannot be used to form factor groups of D3 .
Be careful not to assume that Z3 Z2
= D3 . The group Z6 also can be

factored Z6 /Z3 = Z2 . If groups factored the way natural numbers do, we

could conclude that D3
= Z6 , which is wrong! Z6 is cyclic and D3 is not;
D3 is not even commutative. Apparently (G/H
= K) 6 (G
= H K).
What does the real Z3 Z2 look like? First, lets use the notation
instead of because were using + as our operator. Elements of Z3 Z2
look like this: (0, 0), (0, 1), (1, 0), (1, 1), (2, 0), (2, 1)
The operation for Z3 Z2 is strange: the left elements are added (mod 3)
and the right elements are added (mod 2): (a, b) ? (x, y) = (a + x (mod 3) ,
b + y (mod 2) ). The structure table looks like:
?
(0,0)
(0,1)
(1,0)
(1,1)
(2,0)
(2,1)

(0,0)
(0,0)
(0,1)
(1,0)
(1,1)
(2,0)
(2,1)

(0,1)
(0,1)
(0,0)
(1,1)
(1,0)
(2,1)
(2,0)

(1,0)
(1,0)
(1,1)
(2,0)
(2,1)
(0,0)
(0,1)

(1,1)
(1,1)
(1,0)
(2,1)
(2,0)
(0,1)
(0,0)

(2,0)
(2,0)
(2,1)
(0,0)
(0,1)
(1,0)
(1,1)

(2,1)
(2,1)
(2,0)
(0,1)
(0,0)
(1,1)
(1,0)

This table is symmetric

about its diagonal, so
order of operations does
not matter. Z3 Z2 is
commutative.

Exercise: Show that Z3 Z2 is cyclic of order 6 and therefore Z3 Z2

= Z6 .
Construct an isomorphism explicitly. (Hint: Pick an element of Z3 Z2 and
18

operate it on itself repeatedly. Keep trying until you find an element that
generates the entire group. Map that element to a generator of Z6 .)

4.4

Permutation Groups

Definition 29. If a set S has a finite number of elements, then any

invertible map from S to itself is called a permutation of S.
Informal Definition: If some number of things are arranged in a
specific order, a permutation is a way of rearranging those things.
Instead of drawing triangles (or constructing a cardboard model) to study
D3 , we could have simply written down the letters {A, B, C} and every
possible permutation of them:
permutation (A,B,C) (C,A,B) (B,C,A) (A,C,B) (C,B,A) (B,A,C)
name
1
r
r2
t
u
v
These ordered triplets form a group under the operation of permutation
composition: a . b = do permutation b, then do permutation a to the
result. This group is isomorphic to D3 and an isomorphism is shown
explicitly in the table above. Groups formed by permutations are
important enough to have their own name:
Definition 30. For any n N, define the symmetric group Sn as the set
of all permutations of n distinct elements under the operation of
permutation composition.
WARNING: The name symmetric group does not imply that
the structure tables of Sn are symmetric! If n > 2, then Sn is not
commutative and thus its structure table is not symmetric.
S3 , the group of all permutations of 3 elements, is isomorphic to D3 , the
symmetry group of a regular triangle. Can we extrapolate and claim that
D4 is isomorphic to S4 ? No! Construct a cardboard square and label its
vertices {A, B, C, D}. Consider the permutation (A,B,C,D) (D,B,C,A):
A

This transformation would require an exceptionally flexible square! D4 is

not isomorphic to S4 , but D4 is isomorphic to a subgroup of S4 . All
19

symmetry transformations of a square can be represented by permutations

of 4 elements, but not all permutations of 4 elements represent symmetries
of a square. This conclusion is a special case of the following theorem:
Theorem 5. Every finite group is isomorphic to a subgroup of some Sn .
We are now ready, at last, to give an informal definition of finite group:
Informal Definition: A finite group is a list of things together with
several ways to rearrange those things and instructions about how to put
them back where they started.
There are 6 distinct permutations of 3 elements and 24 distinct
permutations of 4 elements. For any n N, it is well known that the order
of Sn is n factorial: n! = n (n 1) (n 2) 2 1. Thus, given a
finite group, we can know everything about it by choosing a large enough n,
writing out all n! elements of Sn , and looking for subgroups.
Faced with this task, we now declare ourselves experts at finite group
theory and abandon the subject. Here are more exercises to fill space:
Definition 31. A derangement of an ordered n-tuple (a1 , a2 , . . . , an ) is a
permutation that maps no element to itself: k N : k n, (ak ) 6= ak .
Example: The two non-identity rotations of our cardboard triangle
represent derangements of (A,B,C): (C,A,B) and (B,C,A). The non-identity
rotations of any regular n-sided polygon will always represent a subset of
the derangements of an ordered list of n elements.
Exercise: Write down all 9 derangements of (A,B,C,D). Label the vertices
of a square with the letters ABCD. Which derangements represent
symmetries of the square shown below? (Hint: Some of the derangements
are impossible transformations of the square. Here is one of them:)
A

impossible

(A,B,C,D)

(D,C,A,B)

Exercise: Rubiks group - the group of all operations on a Rubiks cube

that do not involve disassembling the cube - is a subgroup of S48 with
43,252,003,274,489,856,000 elements. Scramble a Rubiks cube and solve it.

20