Group Theory and the Rubik’s Cube

by
Lindsey Daniels

A project submitted to the Department of

Mathematical Sciences in conformity with the requirements
for Math 4301 (Honours Seminar)

Lakehead University
Thunder Bay, Ontario, Canada
copyright (2014)
c Lindsey Daniels
 Abstract

The Rubik’s Cube is a well known puzzle that has remarkable group theory properties.
The objective of this project is to understand how the Rubik’s Cube operates as a group
and explicitly construct the Rubik’s Cube Group.

i
 Acknowledgements

I would like to thank my supervisors Dr. Adam Van Tuyl and Dr. Greg Lee for their
expertise and patience while preparing this project.

ii
 Contents

Abstract i
Acknowledgements ii
Chapter 1. Introduction 1
Chapter 2. Groups 3
1. Preliminaries 3
2. Types of Groups 4
3. Isomorphisms 6
Chapter 3. Constructing Groups 9
1. Direct Products 9
2. Semi-Direct Products 10
3. Wreath Products 10
Chapter 4. The Rubik’s Cube Group 12
1. Singmaster Notation 12
2. The Rubik’s Cube Group 12
3. Fundamental Theorems of Cube Theory 16
4. Applications of the Legal Rubik’s Cube Group 20
Chapter 5. Concluding Remarks 22
Chapter 6. Appendix 23
Bibliography 24

iii
 CHAPTER 1

Introduction

In 1974, Ernö Rubik invented the popular three dimensional combination puzzle known
as the Rubik’s Cube. The cube was first launched to the public in May of 1980 and quickly
gained popularity. Since its launch, 350 million cubes have been sold, becoming one of
the best selling puzzles [2]. By 1982, the cube had become part of the Oxford English
Dictionary and a household name [5]. In 1981, David Singmaster published Notes on
Rubik’s ‘Magic Cube’ which was the first analysis of the Rubik’s Cube, and provided
an algorithm for solving it. Singmaster also introduced ‘Singmaster Notation’ for the
different rotations of the cube [10]. Today, numerous methods for solving the cube exist.
When the cube was first introduced to the public, the focus was on solving the puzzle.
Today, the Rubik’s Cube is still popular; however, the focus has changed. Speed-cubing
competitions are held through the World Cube Association, where participants attempt
to solve the cube as fast as possible [2] (the current world record for solving the cube
is 5.55 seconds [7]). There is also interest in finding the maximum number of minimum
moves needed to put the cube into its solved state from any position. This number is
called God’s Number and in 2010 was determined to be 20 [9]. God’s Number, however,
does not say which twists and turns are needed to solve the cube, it merely states what
the maximum number of moves is. The challenge for the solver is to find the 20 moves
(or less) that are required [8].
Since its creation, the cube has been studied in a variety of fields such as computer
science, engineering and mathematics. In mathematics, the Rubik’s Cube can be described
by Group Theory. The different transformations and configurations of the cube form
a subgroup of a permutation group generated by the different horizontal and vertical
rotations of the puzzle [2]. The solution to the cube can also be described by Group
Theory [5]. Group Theory allows for the examination of how the cube functions and
how the twists and turns return the cube to its solved state. This project will explore the
construction of this permutation group, as well as the associated properties and theorems.
This project will follow the method of David Joyner’s Adventures in Group Theory:
Rubik’s Cube, Merlin’s Machine and Other Mathematical Toys to construct the Rubik’s
Cube Group. To begin, in Chapter 2, the preliminary properties of a group are reviewed.
The different types of groups needed to construct the Rubik’s Cube Group will be defined,
as well as the First Group Isomorphism Theorem. Chapter 3 presents the three different
products that are used in the Rubik’s Cube Group. Some of the related properties of
these products are also described. In Chapter 4, Singmaster Notation will be introduced

1
Chapter 1. Introduction 2
and the First and Second Fundamental Theorems of Cube Theory are presented. Here,
the Rubik’s Cube Group will be explicitly constructed. In Chapter 5 a short summary
is provided, along with some possible extensions of this paper. Finally, in Chapter 6 an
appendix of move sequences is provided.
 CHAPTER 2

Groups

In this chapter, the definition of a group and some of the associated properties are
reviewed. Several different types of groups are discussed, as well as the different isomor-
phism theorems.
The main sources for this chapter are [4] and [5].

1. Preliminaries

Before the Rubik’s Cube Group can be constructed, many definitions from group
theory will be needed. A review of the essential definitions from group theory are provided.
Definition 2.1. Let G be a set with a binary operation ∗ such that
∗:G×G→G
(g1 , g2 ) 7→ g1 ∗ g2 .
Then G is a group under this operation if the following three properties are satisfied:
(1) For every a, b and c in G, (a ∗ b) ∗ c = a ∗ (b ∗ c) (associativity).
(2) There exists an element e such that a ∗ e = e ∗ a = a for all a in G (identity
element).
(3) For every element a in G, there exists a−1 such that a ∗ a−1 = a−1 ∗ a = e
(inverses).
Example 2.2. Let G be the set of integers, G = Z, and x, y, z ∈ G under the
operation of addition.
• Since (x + y) + z = x + y + z = x + (y + z), G is associative.
• The identity element of G is 0 since x + 0 = 0 + x = x.
• For each x ∈ G, there exists −x ∈ G with x + (−x) = 0. So G contains
inverses.
So G is a group under addition.
Notice that if the operation on the integers is changed to multiplication, then G would
not be a group since the set would not contain inverses. For example, take the number
1 1 1
2 ∈ Z. The inverse of 2 would be since 2 ∗ = 1, but ∈ / Z.
2 2 2
Definition 2.3. The order of a group G, denoted |G|, is the number of elements in
G.
3
Chapter 2. Groups 4
Definition 2.4. Let H be a subset of a group G. If H is a group with the same
operation as G, then H is a subgroup of G.
Example 2.5. Let G be the set of integers modulo 6, G = Z6 = {0, 1, 2, 3, 4, 5}. Then
G is a group under addition mod 6. The order of G is |G| = 6. A subgroup of G would
be H = {0, 2, 4} under addition mod 6.
Definition 2.6. A group G is a finite group if |G| < ∞.
Example 2.7. For each positive integer n > 1, G = Zn is a finite group since |G| = n.
Definition 2.8. Let G be a group and H ⊂ G. The set aH = {ah | h ∈ H} for any
a ∈ G is a left coset of H in G. Likewise the set Ha = {ha | h ∈ H} for any a ∈ G is a
right coset of H in G.
Theorem 2.9 (Lagrange’s Theorem). If G is a finite group and H is a subgroup of
G, then |H| divides |G|. Furthermore, the number of distinct right (or left) cosets of H
in G is |G|/|H|.

A proof of Lagrange’s Theorem can be found in [4].

Definition 2.10. Let G and H be finite groups and H ⊂ G. The index of H in G
is [G:H] = |G|/|H|.
Example 2.11. If G = Z6 = {0, 1, 2, 3, 4, 5} and H = {0, 2, 4}, then [G : H] =
|G|/|H| = 6/3 = 2. So there are 2 distinct left cosets of H in G, and these two cosets are
{0, 2, 4} and {1, 3, 5}.
Definition 2.12. Let G and H be groups and H ⊂ G. The subgroup H is a normal
subgroup of G, denoted by H / G, if, for each a in G, a−1 Ha = H (or aH = Ha).
Example 2.13. Let G = Z6 = {0, 1, 2, 3, 4, 5} and H = {0, 2, 4}. For each g ∈ G,
g + H = H + g since in Z addition is commutative. So H is a normal subgroup of G and
denote by H / G.
Lemma 2.14. Let S1 , S2 , ..., Sn denote finite sets. Then
|S1 × S2 × ... × Sn | = |S1 | · |S2 | · ... · |Sn |.

Proof. Let S1 × S2 × ... × Sk = {(s1 , s2 , ..., sn )|si ∈ Si }. Now, there are |S1 | choices
for s1 , |S2 | choices for s2 , |S3 | choices for s3 , and so on. By the multiplication principle:
|S1 × S2 × ... × Sk | = |S1 | · |S2 | · ... · |Sk |.


2. Types of Groups

Many special types of groups can be constructed. In this section, the relevant groups
that will be needed to construct the Rubik’s Cube Group are outlined.
Chapter 2. Groups 5
Definition 2.15. A group G is a cyclic group if there is some element g in G such
that G = {g n |n ∈ Z}. The element g is a generator of the group G, denoted G = hgi.
The group Cn denotes the cyclic group of order n.
Example 2.16. If G = Z6 = {0, 1, 2, 3, 4, 5}, then a cyclic subgroup would be < 2 >=
{0, 2, 4}.
Definition 2.17. A permutation of a set G is a one-to-one and onto function from
G to itself.
Definition 2.18. A cycle is a permutation of the elements in a set X = {1, 2, 3, ..., n}
such that x1 7→ x2 7→ x3 7→ ... 7→ x1 where xi ∈ X.
Definition 2.19. Any permutation can be written as a product of its cycles. This
is called cycle notation. If in the permutation, an element is sent to itself, the cycle is
omitted from the cycle notation. Also, the identity permutation is denoted by (1).
Example 2.20. Take the set X = {1, 2, 3, 4} and the permutation σ : X → X where
σ(1) = 2, σ(2) = 4, σ(3) = 1 and σ(4) = 3. As a cycle, σ is 1 7→ 2 7→ 4 7→ 3 7→ 1 and the
cycle notation is (1243).
Definition 2.21. A cycle (x1 x2 ...xk ) is called a cycle of length k. Moreover, a
permutation that can be expressed as a cycle of length 2 is called a 2-cycle.
Definition 2.22. If a permutation can be expressed as an even number of 2-cycles,
then the permutation is even. If a permutation can be expressed as an odd number of
2-cycles, then the permutation is odd.
 
1 2 3 4 5
Example 2.23. Consider the permutation . In cycle notation, the
2 1 4 3 5
permutation would be (12)(34)(5) or more simply, (12)(34). Since the permutation can
be expressed by two 2-cycles, the permutation is even.
Definition 2.24. The permutation group of the set S is the set of all permutations
of S that form a group under composition.
Example 2.25. Let T = {1, 2, 3}. A permutation of T would be ρ : T → T where
ρ(1) = 2, ρ(2)
 = 3, and ρ(3) = 1. The permutation can be written completely as
1 2 3
ρ= or in cycle notation ρ = (123). The set of all permutations of T is
2 3 1
           
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
T1 = , , , , , .
1 2 3 1 3 2 2 1 3 2 3 1 3 1 2 3 2 1

 T1 ={(1), (23), (12), (123), (132), (13)}. The identity of T1 is the per-
In cycle notation,
1 2 3
mutation e = = (1).
1 2 3
Definition 2.26. The permutation group of n elements, denoted Sn is called the
symmetric group.
Chapter 2. Groups 6
Definition 2.27. The group of all even permutations, denoted An , is called the
alternating group.
Definition 2.28. Let G be a group and H / G. Then the factor group is the group
G/H={aH | a ∈ G} under the operation (aH)(bH) = abH for a, b ∈ G.
Example 2.29. If G = Z6 = {0, 1, 2, 3, 4, 5} and H = {0, 2, 4}, then the factor group
G/H = {a + H| a ∈ G}
= {(0 + {0, 2, 4}), (1 + {0, 2, 4}), (2 + {0, 2, 4}), (3 + {0, 2, 4}), (4 + {0, 2, 4}), (5 + {0, 2, 4})}
= {{0, 2, 4}, {1, 3, 5}, {2, 4, 0}, {3, 5, 1}, {4, 0, 2}, {5, 1, 3}}
= {{0, 2, 4}, {1, 3, 5}}.

3. Isomorphisms

One of the important concepts in group theory is understanding how to construct

isomorphisms.
Definition 2.30. A function φ from a group G to a group H is a homomorphism
if φ preserves the group operation; that is, if φ(ab)=φ(a)φ(b) for all a, b ∈ G.
Example 2.31. Take G = S4 and H= {1, −1} under the operation multiplication.
1 a even
Define the map σ : G → H with σ(a) = for every a ∈ G.
−1 a odd
To check that σ is a homomorphism, the 4 possible cases will be verified: a and b both
even, a odd and b even, a even and b odd, a and b both odd. Also, recall that the product
of two even functions is even, the product of two odd functions is even, and the product
of an even function with an odd function is odd.
If a and b are even, then φ(ab) = φ(a)φ(b) = (1)(1) = 1
If a is odd and b is even, then φ(ab) = φ(a)φ(b) = (−1)(1) = −1
If a is even and b is odd, then φ(ab) = φ(a)φ(b) = (1)(−1) = −1
If a and b are odd, then φ(ab) = φ(a)φ(b) = (−1)(−1) = 1

It is clear that even permutations are sent to 1 and odd permutations are sent to −1.
Thus σ is a homomorphism.
Definition 2.32. An isomorphism is a homomorphism φ : G → H that is a one-
to-one and onto. If such a function exists, then G is isomorphic to H and denote this
by G ∼= H.
Example 2.33. Let G = Z and H = 2Z both under the operation addition. Then
φ : Z → 2Z is an isomorphism where φ(a) = 2a for all a ∈ Z.
Definition 2.34. An automorphism is an isomorphism from a group G onto itself.
The set of automorphisms of a group G is denoted by Aut(G).
Chapter 2. Groups 7
Example 2.35. Take G to be any group under the operation of addition. Then
ρ : G → G is an automorphism where ρ(a) = a for all a ∈ G.
Definition 2.36. Let G and H be groups and let f : G → H be a homomorphism.
Then the kernel of G is the set ker(f ) = {g  G | f (g) = eH }, where eH is the identity
element of H.
Example 2.37. Take f : Z → Zn where a 7→ a(mod n). The identity of Zn is 0. So
ker(f ) = {a  Z|a = bn, b  Z} = nZ.
Lemma 2.38. Let f : G → H be a homomorphism for any two groups, G and H.
Then ker(f ) is a normal subgroup of G and G/ ker(f ) is a group.

Proof. First, note that ker(f ) 6= ∅ since eG 7→ eH by properties of homomorphisms.

Next, to show ker(f ) is a subgroup of G, it is enough to show that if a, b ∈ ker(f )
then ab−1 ∈ ker(f ). Let a, b ∈ ker(f ), then f (a) = eH and f (b) = eH . So f (ab−1 ) =
f (a)f (b−1 ) = f (a)(f (b))−1 = eH e−1
H = eH . Thus, ab
−1
∈ ker(f ).
Let g ∈ G and k ∈ ker(f ). Now, ker(f ) is a normal subgroup of G since:
f (gkg −1 ) = f (g)f (k)(f g −1 ) since f is a homomorphism
= f (g)eH f (g −1 ) by definition of kernel and since k ∈ ker(f )
= f (g)(f (g))−1
= eH .
So gkg −1 ∈ ker(f ). By the definition of a normal subgroup, ker(f ) is normal.
Finally, by definition of a factor group, G/ ker(f ) is a group. 
Theorem 2.39 (First Isomorphism Theorem of Groups). Let φ be a group homomor-
phism from G to H. Then the map from G/ ker(φ) to φ(G) defined by g ker(φ) → φ(g) is
an isomorphism. That is, G/ ker(φ) ∼
= φ(G).

Proof. Define a map

ρ : G/ ker(φ) → φ(G)
a ker(φ) 7→ φ(a).
For ρ to be an isomorphism, ρ must be well-defined, one-to-one, onto and a homomor-
phism.
First, ρ is well-defined. Suppose:
a ker(φ) = b ker(φ) ⇐⇒ ab−1 ∈ ker(φ) properties of cosets
⇐⇒ φ(ab−1 ) = eH where eH is the identity of H
⇐⇒ φ(a)φ(b−1 ) = eH since φ is a homomorphism
⇐⇒ φ(a)(φ(b))−1 = eH since f is a homomorphism
⇐⇒ φ(a) = φ(b).
Chapter 2. Groups 8
So ρ is well-defined.
Second, ρ is one-to-one. Suppose:
ρ(a ker φ) = ρ(b ker φ).
Then, φ(a) = φ(b) by mapping of ρ.
So, φ(ab−1 ) = eH .
=⇒ ab−1 ∈ ker φ
=⇒ a ker φ = b ker φ.
Thus ρ is one-to-one.
Next, ρ is onto since if b ∈ φ(G), then there exists some a ∈ G such that φ(a) = b
and ρ(a ker φ) = φ(a) = b.
Finally, ρ is a homomorphism.
ρ((a ker φ)(b ker φ)) = ρ(ab ker φ)
= φ(ab)
= φ(a)φ(b) since φ is a homomorphism
= ρ(a ker φ)ρ(b ker φ).
So ρ is a homomorphism and it follows that ρ is an isomorphism. 

0 if a even
Example 2.40. Take φ : S4 → Z2 where φ(a) = . Then Im φ = Z2 .
1 if a odd
Also note that the identity in Z2 is 0. Now, the kernel of φ is ker φ = {all even permutations} =
A4 . By the First Isomorphism Theorem of Groups, S4 /A4 ∼ = Z2 .
 CHAPTER 3

Constructing Groups

In this chapter, the construction of groups using direct products, semi-direct products,
and wreath products will be examined.
The main sources for this chapter are [4] and [3].

1. Direct Products

Given integers a and b, a new integer can be created by multiplying a and b. That is,
a · b = ab. The same concept can be applied to groups. New groups can be formed by
taking two existing groups, say G1 and G2 , and ‘multiplying’ them together.
Definition 3.1. Let G1 and G2 be groups. Then the direct product of G1 and G2
0 0 0 0 0
is the set G1 × G2 under the operation (g1 , g2 ) · (g1 , g2 ) = (g1 g1 , g2 g2 ) for g1 , g1 ,  G1 and
0
g2 , g2  G2 .
Example 3.2. Let G1 = Z2 and G2 = Z2 . Then
A = G1 × G2
= Z2 × Z2
= {(0, 0), (0, 1), (1, 0), (1, 1)}.

Example 3.3. R2 = R×R = {(x, y)|x, y ∈ R} under addition. That is, if a, b, c, d ∈ R,

then (a, b) + (c, d) = (a + c, b + d).
Definition 3.4. Let G be a group and X be a set. Define a map G × X → X. Then
the group G acts on X if the following happen:
• ex = x for all x ∈ X
• gx ∈ X for all g ∈ G and x ∈ X
• (mn)x = m(nx) for all m, n ∈ G and x ∈ X
Example 3.5. Let G = S4 and X = {1, 2, 3, 4}. Some examples of G acting on X
are:
• [(12)(34)]2 = 1.
• [(1234)]3 = 4.
• [(132)(12)]2 = 3.
9
Chapter 3. Constructing Groups 10
2. Semi-Direct Products

To construct the Rubik’s Cube Group, a more general product than the direct product
of two groups will be needed.
Definition 3.6. Let G1 and G2 be subgroups. Then A = G1 o G2 is a semi-direct
product if:
(1) A = G1 G2 .
(2) G1 ∩ G2 = eA where eA is the identity element of A.
(3) G1 / A.
Example 3.7. The group Sn can be written as a semi-direct product: Sn = An o
h(12)i. Note that An ∩ h(12)i = e. Now, for any a ∈ Sn and b ∈ An , aba−1 ∈ An since
sgn(b) = 1 and for any s ∈ {−1, 1}, sgn(aba−1 ) = sgn(a)sgn(b)sgn(a−1 ) = s2 = 1. But
this means that aba−1 is even, and thus aba−1 ∈ An . So An / Sn =⇒ Sn ∼ = An o h(12)i.
Example 3.8. The dihedral group (the group of reflections and rotations of a regular
polygon with n sides) Dn = {hr, si|rn = s2 = e, srs = r−1 } can be expressed as a semi-
direct product. Let G1 = hri where r are the rotations of order n and G2 = hsi where s
are the reflections of order 2. Now G1 ∩ G2 = e and G1 / Dn . So G1 o G2 ∼
= Dn .

3. Wreath Products

The product of two groups can be generalized from semi-direct products even further
to wreath products.
Definition 3.9. Let X be a finite set, G a group and H a group acting on X. Fix a
labelling of X, say {x1 ,x2 ,..., xt }, with |X| = t. Let Gt be the direct product of G with
itself t times. Then the wreath product of G and H is Gt o H = Gt o H where H acts
on Gt by its action on X.
Remark 3.10. Here, the action of H on G is by conjugation; that is, if g ∈ G, then
the action of H on Gt is (g1 , g2 , ..., gt )h = (g1h , g2h , ..., gth ).

The wreath product of two groups G and H is constructed by:

(1) write H as a permutation group on n items.
(2) make n copies of the group G.
(3) H acts on the copies of G by permuting the elements.
The wreath product of G by H is a semi-direct product of a direct products of n copies
of G by H.
Example 3.11. Let G = Zm , H = Sn and X = {1, 2, 3, ...n}. Then the wreath
product of G by H is Znm o Sn where ρ : Sn → Aut(Znm ) is defined by ρ(σ)(x1 , x2 , ..., xn ) =
(xσ(1) , xσ(2) , ..., xσ(n) ). The group Znm o Sn is called the generalized symmetric group.
Chapter 3. Constructing Groups 11
The wreath product just shuffles the elements of Znm
according to the action from Sn .
The result is a permutation of the original element. Hence, Znm oSn is called the generalized
symmetric group.
Example 3.12. Let G = Z2 , H = S3 and X = {1, 2, 3}. The wreath product of G by
H is Z32 o S3 . The elements of the wreath product Z32 o S3 are:
{(0, 0, 0)σ, (1, 0, 0)σ, (0, 1, 0)σ, (0, 0, 1)σ, (1, 1, 0)σ, (0, 1, 1)σ, (1, 0, 1)σ, (1, 1, 1)σ}
where σ ∈ S3 .

The wreath product permutes the factors of G according to the action h on X. So if

x ∈ G, then the wreath product would take the components of g and shuffle them around
according to the action h on the set X.
 CHAPTER 4

The Rubik’s Cube Group

The Rubik’s Cube is a 3 × 3 × 3 cube. The cube can be manipulated by rotating the
faces of the cube. There are six faces, with each face composed of nine facets. On each
face, the center facet is fixed, and is unmoveable. In total, there are 6 · 9 = 54 facets on
the cube. Each facet is also coloured, and solving the cube requires that each face be a
solid colour. That is, the nine facets of the side must all be the same colour.
In this chapter, the Rubik’s Cube Group will be defined. As well as some of the
associated theorems and applications of the group.
The primary sources for this chapter are [5] and [1].

1. Singmaster Notation

To solve the Rubik’s cube, a series of turns of the faces are needed. To describe these
turns, the notation introduced by David Singmaster [10] will be used. For this notation,
assume that the cube is sitting on a flat surface and each turn of the face will be a one
quarter turn (90 degrees) clockwise.
• Let U denote the upward (top) face.
• Let F denote the front face.
• Let L denote the left face.
• Let R denote the right face.
• Let B denote the back face.
• Let D denote the downward (bottom) face.
It is noted that the clockwise turns are done as if the solver is looking at that particular
face, and then turns the face in the clockwise direction. The inverse of each move would
be the 90 degree rotation of the face counter-clockwise and denoted Mi−1 , where Mi ∈
{U, F, L, R, B, D}.
Example 4.1. The combination F LU would result in the front face of the cube being
rotated 90 degrees, then the left face by 90 degrees and finally the upper face 90 degrees.
The inverse of F LU would be the move U −1 L−1 F −1 .

2. The Rubik’s Cube Group

On the Rubik’s Cube, there are 54 facets that can be arranged and rearranged through
twisting and turning the faces. Any position of the cube can be describe as a permutation
12
Chapter 4. The Rubik’s Cube Group 13
from the solved state. Thus, the Rubik’s Cube group is a subgroup of a permutation
group of 54 elements.
Definition 4.2. The permutation group G = hF, L, U, D, R, Bi ⊂ S54 is called the
Rubik’s Cube Group.

There are two different classifications of the Rubik’s Cube Group: the Legal Rubik’s
Cube Group and the Illegal Rubik’s Cube Group. The difference between the two being
that the Illegal Rubik’s Cube Group allows the solver to take the cube apart and rearrange
the facets. In neither case is the solver allowed to remove the stickers from each facet. As
expected, the Rubik’s Cube Group is a subset of the Illegal Rubik’s Cube group.
Now, not all of the permutations of S54 will be possible on the Rubik’s Cube. The
middle facet on each side of the cube is fixed and cannot be permutated to a different
position on the cube. Furthermore, any valid permutation on the cube will send corner
facets to corner positions and edge facets to edge positions. Any other permutations
will not be physically possible on the cube. Hence, G is only a subset of S54 and not
isomorphic to the full permutation group.
1
2

2
1

1
2

1
2
1

2
2 1 1 2
3 3
1 2 2 1
2 2
1 1

Figure 1. The different types of facets on a Rubik’s Cube: 1 denotes the

facets that make up corner cubes, 2 denotes facets that make up edge cubes
and 3 denotes the fixed center cubes [6].

2.1. Corner Cubes. As shown in Figure 1, each corner cube consists of three facets.
Now, there are a total of eight corner cubes on a Rubik’s Cube and each of the facets that
comprise the corner cube lie on three different sides of the cube.

As shown in Figure 2, facet A is on the upper face, facet B is on the left face, and
facet C is on the front face. Now, it is possible to reorient the facets of a center cube:
facet A is in the position where facet B is, facet B is moved to where facet C was, facet
C moved to the position of facet B; and facet A can be moved to the position of facet
C, facet C to the position of facet B and facet B to the position of facet A. In terms of
groups, this means that the facets of a corner cube belong to the cyclic group of three
elements C3 . Moreover, since there are eight corner cubes, the orientation of any facet of
Chapter 4. The Rubik’s Cube Group 14

A
B C

Figure 2. The 3 facets that make up a corner cube [6].

a corner cube can be described by the set C3 × C3 × C3 × C3 × C3 × C3 × C3 × C3 = C38 .

Now, the possible arrangements of the corner cubes can be described similarly. Again,
any of the eight corner cubes can occupy any of the corner cube positions of the Ru-
bik’s Cube. So, the possible arrangements of the corner cubes can be described by the
permutation group of eight elements, S8 .
Lemma 4.3. The position of all of the corner facets on the Rubik’s Cube can be de-
scribed by the group C38 o S8 .

Proof. This follows from the definition of wreath product and from the fact that any
corner cube position can be described by its position on the cube and the cycle orientation
of the three facets of the corner cube. 

2.2. Edge Cubes. Every edge cube in the Rubik’s Cube consists of two facets, as
shown in 1 and there are 12 edge cube on the Rubik’s Cube. Note that for every edge
cube, each of the two facets of an edge cube lie on different faces of the cube.
Y

Figure 3. The 2 facets that make up an edge cube [6].

As in figure 3, facet X is on the left face and facet Y is on the upper face. Like-
wise, it is also possible for facets X and Y to switch places. That is, facet X would be
repositioned to where facet Y is and facet Y would be moved to the position where facet
X is. In terms of groups, the facets of any edge cube belong to the cyclic group of two
elements C2 . In addition, there are 12 edge cubes on the Rubik’s Cube and any edge
cube can occupy an edge cube spot. Thus any facet of an edge cube will be in the set
C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 × C2 = C212 .
Likewise to describe the different arrangements of the edge cubes. There are 12 edge
Chapter 4. The Rubik’s Cube Group 15
cubes on the Rubik’s Cube and any edge cube can be in an edge cube spot. Thus, the
possible arrangements of the edge cubes of the Rubik’s Cube can be described by the
permutation group of 12 elements, S12 .

Lemma 4.4. The position of all of the edge facets on the Rubik’s Cube can be described
by the group C212 o S12 .
Proof. This follows from the definition of wreath product and from the fact that any
edge cube position can be described by its position on the cube and the cycle orientation
of the two facets of the corner cube. 

2.3. Cube Position. From Lemma 4.3 any corner cube position can be expressed
as a 8-tuple and from Lemma 4.4 any edge cube position can be expressed as a 12-tuple.
However, to determine the individual components of the tuples, a fixed numbering system
will be needed.

2 1 1
0 U 0
0 0

0 0

1
1 0 B 2
0

1 1
1 2 R
0 2 1 1 1 1 1 2 1 0
L 1 1
0 D 0

2 0 1 F 1 2
0 0

0 0
1
1 2 1
0

Figure 4. The fixed orientation markings, as denoted in [5], for the facets
of the Rubik’s Cube [6].

For any arbitrary facet, the position of the facet is assigned the corresponding number
above. Even though the facets will be moving around the cube, the numbering system
remains fixed.
Example 4.5. Consider the top edge cube on the front face of the Rubik’s Cube [6].
It begins with a number of 1. Now, by doing the move FR, the facet is moved to the
upper face on the right side. This position of the edge cube is assigned the number 0.
U

Move FR
−−−−−→
L F L F

Remark 4.6. With each turn, the edge cube’s orientation number is changed by either
0 mod 2 or 1 mod 2.
Example 4.7. Consider the cube below [6] and the upper, front and left face corner
cube.
Chapter 4. The Rubik’s Cube Group 16

U L F Move by F F R
Move by F
−−−−−−→ −−−−−−→
L F

D
U

U
Move by F Move by F
−−−−−−→ −−−−−−→
F R L F

Tracking the blue facet, it begins with the number 1, then has number 2, number 1,
number 2, and then back to number 1 to complete the cycle. Next, the purple facet
starts with number 0, then goes to number 1, number 0, number 1, and then back to 0
to complete the cycle. Finally, the orange facet starts with number 2, then number 0,
number 2, number 0, and then back to number 2.
Remark 4.8. With each turn of the R, L, F or B face, the corner facet orientation
number is changed by either 1 mod 3 or 2 mod 3. With each turn of the U or D face, the
numbering remains unchanged (0 mod 3).
Remark 4.9. The orientation number for any facet is determined by comparing the
position of the facet on the Rubik’s Cube to the fixed numbering shown in Figure 4.

2.4. The Illegal Rubik’s Cube Group. The Illegal Rubik’s Cube Group allows
the solver to take the cube apart and reassemble it in any orientation. Again, some of
the orientations are not physically possible on the cube. When all the possible positions
of the facets are combined as a whole, some of the arrangements will not be physically
possible on the cube.
Lemma 4.10. The Illegal Rubik’s Cube Group is I = (C212 o 12) × (C38 o S8 ).
Proof. This follows from Lemma 4.3, Lemma 4.4 and the definition of the direct
product. 

3. Fundamental Theorems of Cube Theory

To be able to distinguish between the legal and illegal Rubik’s Cube Group, the First
and Second Fundamental Theorems of Cube theory are needed.
The First Fundamental Theorem of Cube Theory gives the criteria for solvable ar-
rangements of the Rubik’s Cube. The illegal Rubik’s Cube group allows the solver to
take the cube apart and reassemble it. However, the cube may get reassembled in an
arrangement that is not solvable. For example, putting 19 of the cubes back in the solved
state and putting the last cube in upside down.
Chapter 4. The Rubik’s Cube Group 17
Theorem 4.11 (First Fundamental Theorem of Cube Theory). [1]
Let v ∈ C38 , r ∈ S8 , w ∈ C212 , and s ∈ S12 . The 4-tuple (v, r, w, s) corresponds to a possible
arrangement (position) of the cube if and only if:
(1) sgn(r) = sgn(s) (equal parity of permutations).
(2) v1 + v2 + v3 + ... + v8 = 0 (mod 3) (conservation of the total number of twists).
(3) w1 + w2 + w3 + ... + w12 = 0 (mod 2) (conservation of the total number of flips).

Proof. (⇒) Let v ∈ C38 , r ∈ S8 , w ∈ C212 , s ∈ S12 and g ∈ G where g is a move that
rearranges the cube from the solved state to a state (v, r, w, s). So g can be written as
g = M1 M2 ...Mn where Mi ∈ {F, L, U, B, R, D}.
(1) With each move a total of four edge cubes and four corner cubes are moved; that
is, the same number of corner cubes are moved and the same number of edge
cubes are moved. Note that each permutation is a 4-cycle, which is odd and has
Yn
sgn = −1. So for each g: sgn(r) = sgn(Mi ) = sgn(s)
k=1
(2) Note that if Mi is U or D, then v remains unchanged, since the all corner cubes
remain on the same face. If Mi is R, L, F , or B, then two corner cubes are
moved. One corner cube is moved down off the U face and one corner cube is
moved up onto the U face. So, the components of v are either decreased by
1(mod3) or increased by 1(mod3), respectively. But this means that for any R,
L, F , or B, v1 + v2 + v3 + ... + v8 = 1 (mod3) − 1 (mod3) = 0 (mod3). So
v1 + v2 + v3 + ... + v8 = 0 (mod3) for any move g.
(3) For each move g a total of four edge cubes will be reoriented. So w1 + w2 + w3 +
... + w12 = 4 (mod2).
(⇐) Let A = (v, r, w, s) and let A satisfy conditions (1), (2), and (3).
Condition (1) says that sgn(s) = sgn(r). So there is equal parity of permutations.
Thus the permutations of the corner cubes and edge cubes are either both even or both
odd. Assume that sgn(s) = sgn(r) = 1; that is, the permutations are even. If the per-
mutations are odd, simply apply one of the basic moves (B, F, L, U, R, D) and the new
position will satisfy sgn(s) = sgn(r) = 1.
Now, consider the move for a corner 3-cycle. Take M = RB −1 RF 2 R−1 BRF 2 R2 for
example. The move M cycles the upper-front-left, upper-front-right, and upper-back-
right corner cubes without changing the position of the other cubes. Denote the upper-
front-left cube as a1 , the upper-front-right cube as a2 , and the upper-back-right cube
as a3 and denote the remaining corner cubes as a4 , a5 , a6 , a7 and a8 . For every ai , from
a4 , a5 , a6 , a7 , a8 , there exists a move x from {B, F, L, U, R, D} of at most two moves (that
is, two of B, F, L, U, R, D) such that ai is moved to the position of a3 without changing
the position of a1 and a2 . Now, apply the transformation xM x−1 . This move creates the
3-cycle (a1 , a2 , ai ). This 3-cycle can be obtained for any of the ai ’s; that is, the 3-cycles
(a1 , a2 , a3 ), (a1 , a2 , a4 ), (a1 , a2 , a5 ), (a1 , a2 , a6 ), (a1 , a2 , a7 ) and (a1 , a2 , a8 ) can all be ob-
tained by the appropriate move x. But, this generates all the even permutations of the
Chapter 4. The Rubik’s Cube Group 18
corner cubes. Therefore, there exists an appropriate move x that will return all the corner
cubes to their home positions.
Now, consider the move for an edge 3-cycle. Take the move M ∗ = R2 U F B −1 R2 F −1 BU R2
(see appendix for diagram). The move M ∗ cycles the upper-front, upper-back and upper-
right edge cubes without changing the position of any of the other cubes. Denote the
upper-front cube as b1 , the upper-back cube as b2 and the upper-right cube as b3 and
denote the remaining edge cubes by b4 , b5 , b6 , b7 , b8 , b9 , b10 , b11 and b12 . Likewise with the
corner cubes, for any bi from b4 , b5 , b6 , b7 , b8 , b9 , b10 , b11 , b12 , there exists a move y from
{B, F, L, U, R, D} of at most 2 moves so that the edge cube bi is moved to the position of
b3 without changing the position of b1 and b2 . Now, apply the transformation yM ∗ y −1 .
This transformation creates the 3-cycle (b1 , b2 , bi ). But using an appropriate choice for y,
the 3-cycles (b1 , b2 , b3 ), (b1 , b2 , b4 ), (b1 , b2 , b5 ), (b1 , b2 , b6 ), (b1 , b2 , b7 ), (b1 , b2 , b8 ), (b1 , b2 , b9 ),
(b1 , b2 , b10 ), (b1 , b2 , b11 ), and (b1 , b2 , b12 ). These generate all the even permutations of the
edge cubes. Thus, there exists an appropriate move y that will return all of the edge
cubes into their home positions.
All that is left to do is to reorient the cubes so that the facets are colour matched.
Condition (2) says that there is a conservation of total twists; that is, the number
of clockwise twists is equal to the number of counterclockwise twists. This means there
exists a move which twists exactly 2 corner cubes and preserves the orientation and po-
sition of all the other cubes, namely the move M1 = (R−1 D2 RB −1 U 2 B)2 which twists
the upper-front-right corner cube by 120 degrees and twists the bottom-down-left cube
by −120 degrees. Note that the move M1 can be modified to obtain a similar result for
any 2 corner cubes. To begin to match the facets of the corner cubes, first twist any
clockwise and counterclockwise pairs into their solved orientations. The remaining corner
X8
cube orientations will occur in triples since the corner cubes obey vi = 0 mod3. So
i=1
they will occur in either 3 clockwise twists or 3 counterclockwise twists. Call these 3 cubes
c1 , c2 , and c3 . The remaining corner cubes can be solved by a sequence of corner twisting
moves, say M1∗ = L−1 D2 LBD2 B −1 U BD2 B −1 L−1 D2 LU −1 or a similar move for two of the
remaining corner cubes that need reorienting. Now, M1∗ will solve one of the remaining
corner cubes, say c1 , and reorient the other corner cube, say c2 , into the opposite position
to the untouched corner cube, c3 . That is, if c3 needs to be solved by a clockwise twist
then M1∗ will reorient c2 to a position that needs a counterclockwise twist to be solved
and vice versa. The remaining two cubes can be solved with the appropriate move M1 .
Thus all the corner cubes are in their solved states.
X12
Condition (3) says that there is a conservation of total flips. Since wi = 0 is done
i=1
mod 2, there is an even number of edge cubes that need to be flipped. But there exists
a move that flips exactly 2 edge cubes and preserves the orientation and position of the
remaining cubes. Take the move M2 = LF R−1 F −1 L−1 U 2 RU RU −1 R2 U 2 R. The move M2
flips the upper-front edge cube and the upper-right edges cube, while leaving the position
Chapter 4. The Rubik’s Cube Group 19
and orientation off all the other cubes untouched. The move M2 can be modified appro-
priately so that any 2 edges cubes can be flipped and the position and orientation of all
the other cubes will be preserved. Since there are an even number of edge cubes, all the
edge cubes are able to return to their solved orientations.
Thus A is a solvable position on the Rubik’s Cube. So (v, r, w, s) is a possible arrange-
ment of the Rubik’s Cube. 

The Second Fundamental Theorem of Cube Theory gives the criteria for legal moves
on the Rubik’s Cube.

Theorem 4.12 (Second Fundamental Theorem of Cube Theory). [1]

An operation of the cube is possible if and only if the following are satisfied:
(1) The total number of edge and corner cycles of even length is even.
(2) The number of corner cycles twisted right is equal to the number of corner cycles
twisted left (up to modulo 3).
(3) There is an even number of reorienting edge cycles.

Proof. (⇒) Let M be an operation on the cube that takes the cube from the solved
state to position g = (v, s, w, r), where v ∈ C38 , r ∈ S8 , w ∈ C212 , and s ∈ S12 .
(1) By (1) of Theorem 4.11, sgn(r) = sgn(s). But this means that the permutation
is even. So the length of the edge and corner cycles is even.
(2) For any move M , the corner cubes are moved either right, left or not at all. So
the cycle changes the sum of vi by 2, 1 or 0 (mod3) respectively. By Theorem
X8
4.11, vi = 0 the number of right twists is equal to the number of left twists.
i=1
(3) Note that an edge cycle only reorients if it is changed by an odd number; that
X12
is wj = 1 for some j = {1, 2, 3, ..., 12}. By Theorem 4.11, wi = 0. But
i=1
this means that if one edge cycle is reorienting, then another edge cycle must
be reorienting since the sum is zero. Thus, there must be an even number of
reorienting edge cycles.
(⇐) Suppose that (1), (2), and (3) hold. By Theorem 4.11, there exists move M that
takes the cube from the the solved state to the state g. There also exists move M −1 that
takes the cube from the state g to the solved state. Now, by assumption M and M −1
satisfy (1), (2), and (3). But M and M −1 are both valid operations on the Rubik’s Cube.
Thus, if (1), (2), and (3) hold, then the operation is valid. 

With the two fundamental theorems of cube theory, any possible position and oper-
ation on the Rubik’s Cube can be defined. Also, the theorems eliminate the physically
impossible arrangements and moves from the group.
Chapter 4. The Rubik’s Cube Group 20
4. Applications of the Legal Rubik’s Cube Group

Using the criteria of the First and Second Fundamental Theorems of Cube Theory,
the Illegal Rubik’s Cube Group can be reduced to the group G0 = {(v, r, w, s)|v ∈ C38 , r ∈
S8 , w ∈ C212 , and s ∈ S12 } where G0 has the properties of Theorem 4.11 and Theorem
4.12.
By Lemma 4.10, the Illegal Rubik’s Cube Group is defined to be I = (C212 o S12 ) ×
8
(C3 o S8 ). However, by the conditions of Theorem 4.11, the group is double counting
some positions of the facets. Condition (2) of Theorem 4.11 determines the position of
the corner cubes, but note that once 7 of the corner cubes have their arrangement, the
last cube’s position would automatically be determined by the given formula. Likewise,
condition (3) determines the orientation of the edge cubes. Once 11 edge cubes are given
a position, the final edge cube is automatically determined by the formula. Condition (2)
reduces the group by a factor of C3 and condition (3) reduces the group by a factor of
C2 . By reducing I, the group G0 is obtained.
Note that G0 ⊂ I; however, G0 is not quite the Rubik’s Cube Group. Some additional
reduction will be done to obtain the Rubik’s Cube Group G.
Theorem 4.13. There exists an isomorphism:
G0 ∼
= (C 7 o S8 ) × (C 11 o S12 )
3 2

and
|G0 | = |S8 ||S12 ||C211 ||C37 | = 8! · 12! · 211 · 37
.

Proof. By Theorem 4.11, the First Isomorphism Theorem of Groups and the def-
inition of semi-direct product G0 ∼ = (C37 o S8 ) × (C211 o S12 ). By Lemma 2.14 |G0 | =
|S8 ||S12 ||C211 ||C37 | = 8! · 12! · 211 · 37 . 

Now, to obtain the Rubik’s Cube Group G, G0 must be further reduced. Condition
(1) of Theorem 4.11 says that the number of even permutations is equal to the odd
permutations. So G0 must be further reduced by a factor of C2 .
Lemma 4.14. The Rubik’s Cube Group, G, can be expressed as G = (C37 o S8 ) × (C210 o
S12 ).

Proof. This follows from Lemma 4.10 and Theorem 4.11. 

Corollary 4.15. The Rubik’s Cube Group G is the kernel of the homomorphism
φ : G0 → {1, −1}
(v, r, w, s) 7→ sgn(r)sgn(s).
In particular, G ⊂ G0 is normal of index 2 and
|G| = 8! · 12! · 210 · 37 .
Chapter 4. The Rubik’s Cube Group 21
Proof. Let G0 = (C37 (C211
o S8 ) × o S12 ), H = {−1, 1} and φ : G0 → H where
~ s) 7→ sgn(r)sgn(s). Then ker(φ) = {(v, r, w, s)|φ(v, r, w, s) = eH }, where eH = 1.
(~v , r, w,
By Theorem 4.13 and the First Isomorphism Theorem of Groups G0 / ker(φ) ∼ = G, where
7 10 10 7
G = (C3 o S8 ) × (C2 o S12 ). Next, by Lemma 2.14, |G| = 8! · 12! · 2 · 3 and [G0 : G] =
(8!·12!·211 ·37 )
(8!·12!·210 ·37 )
= 2. 
 CHAPTER 5

Concluding Remarks

This paper explored some of the group theory applications to the Rubik’s cube and
constructed the Rubik’s Cube Group. The Rubik’s Cube Group was shown to be G =
hR, B, L, U, F, Di, which is a subgroup of S54 . The First and Second Fundamental Theo-
rems of Cube Theory were presented, which gave the criteria for all the possible arrange-
ments and moves allowed on the cube. The fundamental theorems redefined the Rubik’s
Cube group to G = (C37 o S8 ) × (C210 o S12 ). Furthermore the group G was shown to be the
kernel of the homomorphism of G0 = (C37 o S8 ) × (C211 o S12 ) → {−1, 1}.
The scope of this paper was restricted to the 3 × 3 × 3 Rubik’s Cube Group; however,
the method developed in this project can be extended to describe the group structure of
the 4 × 4 × 4 and 5 × 5 × 5 Rubik’s Cube. Moreover, the algorithm for solving any of the
3 cubes can be describe in terms of group operations.

22
 CHAPTER 6

Appendix

Move Sequence [1] Diagram [6]

U
L F

RB −1 RF 2 R−1 BRF 2 R2

U
L F

R2 U F B −1 R2 F −1 BU R2 U

L F

(R−1 D2 RB −1 U 2 B)2
U

L F

R(U 2 RF −1 D2 F R−1 )2 R−1

U

R F

LF R−1 F −1 L−1 U 2 RU RU −1 R2 U 2 R

23
 Bibliography

[1] Brandelow, Christoph. Inside the Rubik’s Cube and Beyond. Birkhäuser. (1982). 12, 17, 19, 23
[2] Demaine, Erik D.; Demain, Martin L; Eisenstat, Sarah; Lubiw, Anna; Winslow, Andrew. Algorithms
for Solving Rubik’s cubes. Lecture Notes in Computer Science, 6942 (2011) 689-700. 1
[3] Dummit, David S.; Foote, Richard M. Abstract Algebra. Prentice Hall. (1999). 9
[4] Gallian, Joseph A. Contemporary Abstract Algebra. Brooks/Cole, Cengage Learning. (2010). 3, 4, 9
[5] Joyner, David. Adventures in Group Theory: Rubik’s Cube, Merlin’s Machine, and other Mathemat-
ical Toys. John Hopkins University Press. (2008). 1, 3, 12, 15
[6] Kottwitz, Steven. Example: Sudoku 3D cube. texexample.net. (2008). Link at
http://www.texample.net/tikz/examples/sudoku-3d-cube/ 13, 14, 15, 23
[7] Reynolds, Tim. World cube Association Official Results. World Cube Organization. (2014). Link at
https://www.worldcubeassociation.org/results/regions.php 1
[8] Rokicki, Tomas. Twenty-Two moves suffice for Rubik’s Cube. Math Intelligencer. 32, No. 1 (2010)
33-40. 1
[9] Rokicki, Tomas; Kociemba, Herbert; Davidson, Morley; Dethridge, John. The diameter of the Ru-
bik’s Cube is twenty. SIAM J. Discrete Math, 27, No. 2 (2013) 1082-1105. 1
[10] Singmaster, David. Notes on Rubik’s ’Magic Cube’. Enslow Pub Inc. (1981). 1, 12

24