A Few Notes on Cayley’s Theorem

David Meyer

[email protected]

Last update: November 13, 2022

1 Introduction
Group Theory is the study of symmetry. Cayley’s Theorem is a fundamental theorem in Group Theory, and
the topic of these notes.

Before diving into Cayley’s Theorem, a couple of notes:

• The Symmetric Group Sym(G) or sometimes Sn , where n = |G| (G is finite), is the set of all bijections
from G to itself with function composition as the group operation. That is, Sym(G) = Sn = S|G| =
{f : G → G | f is an bijection}.
• We use the symbol ' (or sometimes ∼
=) to mean that the groups G and H are isomorphic. That is,
G ' H =⇒ ∃f | f : G → H where f is a bijection and a homomorphism. See Equation (1).

• To show that f is one-to-one, show that f (x) = f (y) =⇒ x = y.

• To show that f is onto, pick an arbitrary h ∈ H and show that ∃g ∈ G | f (g) = h.

Recall also that if we have two groups (G, ∗) and (H, ·) we say that (G, ∗) is isomorphic to (H, ·) if there
exists a bijection f : G → H which satisfies the homomorphism property:

f (x ∗ y) = f (x) · f (y) ∀x, y ∈ G (1)

That is, f is a bijection (one-to-one and onto) and f is also a homomorphism.

Any bijective function f which satisfies Equation (1) is called a group isomorphism from G to H. The basic
idea of (G, ∗) being isomorphic to (H, ·) is that (G, ∗) and (H, ·) are ”algebraically equivalent”. That is, there
is a one-to-one correspondence between elements of G and elements of H where the outcomes of operations
on elements of G are matched with the outcomes of the corresponding operations on the corresponding
elements of H.

2 Cayley’s Theorem
Theorem 2.1. Cayley’s Theorem: If G is a group then there exists a subgroup H of Sym(G) such that
G is isomorphic to H.
Proof: Suppose that G is a group. Then to prove Cayley’s Theorem we need to find a subgroup H of
Sym(G) and a bijective homomorphism f : G → H. My roadmap for the proof looks like

1
 1. Define φa : G → G for each a ∈ G and show that φa is a bijection
2. Define H = {φa | a ∈ G} and show that H is a subgroup of Sym(G)
3. Define f : G → H and show that f is both a bijection and a homomorphism

BTW, a nice thing about the proof of Cayley’s theorem is that it is a constructive proof: the statement of
the theorem is that a certain group H exists. In the course of the proof of the theorem one can actually
show not only that such an H exists but also how to actually find it. We’ll see an example of this below
(Section 3.2).

2.1 Define φa : G → G for each a ∈ G and show that φa is a bijection

To start, for each fixed element a ∈ G define φa : G → G by the map x 7→ ax. That is

φa (x) = ax ∀x ∈ G (2)

Luckily it turns out that each φa is a bijection. To see this we need to show that φa is one-to-one and onto.
First, consider that φa is one-to-one since

φa (x) = φa (y) # to show φa is 1-to-1 show φa (x) = φa (y) ⇒ x = y

⇒ ax = ay # definition of φa (x) (Equation (2))
⇒ a−1 (ax) = a−1 (ay) # multiply by a−1 ; a ∈ G & G a group ⇒ a−1 ∈ G (3)
⇒ (a−1 a)x = (a−1 a)y # multiplication is associative
⇒ x=y # a−1 a = 1

So φa is one-to-one.

Aside on cancellation laws: Note that in (3) we used the fact that a ∈ G and that G is a group so
a−1 ∈ G. Here we have a−1 a = 1, which essentially gives us a cancellation law 1 ; in (3) this allows us to
”cancel” the a on both sides. Now, what if we don’t have access to multiplicative inverses? We might be
faced with this situation if we have a ring, where we don’t in general have multiplicative inverses2 . So if we
don’t have multiplicative inverses how do we go about showing that something is one-to-one?

One approach is to factor out a and note that by assumption, a 6= 0 so something else must be. For example

φa (x) = φa (y) # to show φa is 1-to-1 show that φa (x) = φa (y) ⇒ x = y

⇒ ax = ay # definition of φa (x) (Equation (2))
⇒ ax − ay = 0 # subtract ay from both sides
⇒ a(x − y) = 0 # factor out a
⇒ x−y =0 # a 6= 0 by assumption so x − y = 0
⇒ x=y # so φa is one-to-one

Getting back to showing that φa is a bijection, we next need to show that φa is onto. To do this pick an
arbitrary y ∈ G (here G is the range). Then a−1 y ∈ G (here G is the domain) and so φa (a−1 y) = a(a−1 y).
Since multiplication is associative we have φa (a−1 y) = a(a−1 y) = (aa−1 )y = y. So φa is onto and hence φa
is a bijection.
1 Note that having a cancellation law is equivalent to saying there are no zero divisors.
2A ring with multiplicative inverses is called a division ring (or skew field). Example: the quaternions.

2
2.2 Define H = {φa | a ∈ G} and show that H is a subgroup of Sym(G)
Now we can define H = {φa | a ∈ G}. Since each element of H is a bijection from G to G and since Sym(G)
is the set of all bijections from G to G we know that H ⊆ G. To show that H is a subgroup of Sym(G) we
also need to show that H is closed under function composition and inversion.

To show closure under function composition we need to show that α, β ∈ H ⇒ α ◦ β ∈ H. To see this
consider α, β ∈ H. Then there exists a ∈ G such that α = φa . Similarly there exists b ∈ G such that β = φb .
So we know that

α ◦ β = φa ◦ φb (4)

and so for any x ∈ G we have

(α ◦ β)(x) = (φa ◦ φb )(x) # α ◦ β = φa ◦ φb (Equation (4))

= φa (φb (x)) # definition of function composition
= φa (bx) # φb (x) = bx (definition of φb )
(5)
= a(bx) # φa (x) = ax (definition of φa )
= (ab)x # multiplication is associative
= φab (x) # φg (x) = gx where g = ab

So α ◦ β = φa ◦ φb = φab . Since ab ∈ G (G is closed under multiplication) we know that φab ∈ H. Now

we have φab ∈ H and α ◦ β = φab which together imply that α ◦ β ∈ H. So H is closed under function
composition.

To show that H is closed under inversion we need to show that α ∈ H ⇒ α−1 ∈ H. To see this consider
α ∈ H. Then there exists an a ∈ G such that α = φa . Since a ∈ G and since G is a group, a−1 ∈ G and so
φa−1 ∈ H. Note further that for any x ∈ G

(φa−1 ◦ φa )(x) = φa−1 (φa (x)) # definition: (f ◦ g)(x) = f (g(x))

= φa−1 (ax) # φa (x) = ax
= a−1 (ax) # definition: φa−1 (x) = a−1 (x) (6)
= (a−1 a)x # multiplication is associative
= x # a−1 a = 1

and

(φa ◦ φa−1 )(x) = φa (φa−1 (x)) # definition: (f ◦ g)(x) = f (g(x))

= φa (a−1 x) # φa−1 (x) = a−1 x
= a(a−1 x) # definition: φa (x) = ax (7)
= (aa−1 )x # multiplication is associative
= x # aa−1 = 1

Recall that if a function f is a bijection we know (f −1 ◦ f )(x) = (f ◦ f −1 )(x) = x. From (6) and (7) we see
that φa−1 is the inverse of φa . More specifically φa−1 = φ−1
a . Since α = φa , α
−1
= φ−1
a = φa−1 ∈ H. So H
is closed under inversion.

3
2.3 Define f : G → H and show that f is a homomorphic bijection
We still need to show a homomorphic bijection f from G to H. One way to do this is to define f (g) = φg
for all g ∈ G. Then to show that f is a bijection we need to show that f is both one-to-one and onto.

To see that f is one-to-one consider

f (a) = f (b) # to show f is 1-to-1 show that f (a) = f (b) ⇒ a = b

⇒ φa (x) = φb (x) # definition of f (g) : f (g) = φg for all g ∈ G
⇒ φa (a) = φb (a) # evaluate at a ∈ G
⇒ aa = ba # definition of φg : φg (x) = gx for all g ∈ G
⇒ (aa)a−1 = (ba)a−1 # multiply by a−1 ; a ∈ G and G a group ⇒ a−1 ∈ G
⇒ a(aa−1 ) = b(aa−1 ) # multiplication is associative
⇒ a=b # aa−1 = 1

So f is one-to-one.

To show that f is onto, choose a α ∈ H. Then there exists an a ∈ G such that α = φa . However we know
that f (a) = φa and φa = α so we know that f (a) = α. So f in onto and since we saw that f is one-to-one,
f is a bijection.

Finally, to show that f is also a homomorphism we want to show that f (ab) = f (a) ◦ f (b). To see this
consider that for any a, b ∈ G we have

f (ab) = φab # definition of f

= φa ◦ φb # Equation (5)
= f (a) ◦ f (b) # definition of f

So f is a homomorphism.

This completes the proof of Cayley’s Theorem.

3 Examples
3.1 (Z4 , +) → (G, ·)
Let (Z4 , +) be the set Z4 = {0, 1, 2, 3} with addition modulo 4 and let (G, ·) be the set G = {1, −1, i, −i}
(the fourth roots of unity) with the usual multiplication on C. Then (Z4 , +) ' (G, ·). To see that Z4 is
isomorphic to G, let f : Z4 → G be the bijection

0 −→ 1
1 −→ i
2 −→ −1
3 −→ −i

Here are the Cayley tables for Z4 and G:

To show that f is an isomorphism we need to show that f is a homomorphism, that is, that f (x + y) =
f (x) · f (y). Since there are only n2 = 42 = 16 values for f (x + y) we can just enumerate them:

4
 + 0 1 2 3 · 1 i −1 −i
0 0 1 2 3 1 1 i −1 −i
1 1 2 3 0 i i −1 −i 1
2 2 3 0 1 −1 −1 −i 1 i
3 3 0 1 2 −i −i 1 i −1

Table 1: Z4 Table 2: G

f (0 + 0) = f (0) =1 =1·1 = f (0) · f (0)

f (0 + 1) = f (1) =i =1·i = f (0) · f (1)
f (0 + 2) = f (2) = −1 = 1 · −1 = f (0) · f (2)
f (0 + 3) = f (3) = −i = 1 · −i = f (0) · f (3)
f (1 + 0) = f (1) =i =i·1 = f (1) · f (0)
f (1 + 1) = f (2) = −1 = i · i = f (1) · f (1)
f (1 + 2) = f (3) = −i = i · −1 = f (1) · f (2)
f (1 + 3) = f (0) =1 = i · −i = f (1) · f (3)
f (2 + 0) = f (2) = −1 = −1 · 1 = f (2) · f (0)
..
.
f (3 + 3) = f (2) = −1 = −i · −i = f (3) · f (3)

So the bijection f : Z4 → G above is a homomorphism and hence f is a group isomorphism.

3.2 The Klein 4-group

The Klein 4-group is the group K = {e, a, b, c} where e is the identity element and the group operation is
defined by the Cayley table below (Table 3).

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

Table 3: The Klein 4-group Operation

Here K is not isomorphic to Z4 . To see this notice that there are 24 bijections from Z4 and K: |K| = |Z4 | = 4
so there are n! = 4! = 24 possible bijections from Z4 to K. Since we need f (0) = e that leaves 3! = 6 bijections
that could be homomorphisms. For example, consider the bijection

0 −→ e
1 −→ a
2 −→ c
3 −→ b

This bijection is not a homomorphism since f (1 + 3) = f (4) = f (0) = e while f (1) · f (3) = ab = c, so
f (1 + 3) 6= f (1) · f (3).

One way to see that Z4 is not isomorphic to K is to recognize that every element of K satisfies the equation
x · x = e (a key property of the Klein 4-group). However not every element of Z4 satisfies the equation
x + x = 0.

5
This gives a clue as to how to prove, by contradiction, that Z4 is not isomorphic to K. Specifically, suppose
that Z4 is isomorphic to K. Then there exists a bijection f : Z4 → K such that f (x + y) = f (x) · f (y) for all
x, y ∈ Z4 . Well, we know by definition that f (0) = e and since f is one-to-one we also know that f (1) 6= e.
Since f is a homomorphism we also know that

f (1 + 1) = f (1) · f (1)

However, since f (1) ∈ K and all elements of K satisfy x · x = e we can conclude that f (1) · f (1) = e, so
f (1 + 1) = f (2) = e. Now we have f (0) = e and f (2) = e which is a contradiction since we assumed that f
was one-to-one. So the original assumption that Z4 is isomorphic to K is false.

Ok, but Cayley’s Theorem says there is a subgroup H of S4 which is isomorphic to K. How to find H? Since
as noted above Cayley’s Theorem is constructive, we should be able to follow the approach used in the proof
to find H. Here we let H = {φe , φa , φb , φc } where, for all x ∈ K

φe (x) = ex # φe (x) = x
φa (x) = ax
φb (x) = bx
φc (x) = cx

Now we can rewrite the Cayley table for the Klein 4-group (Table 3) as

 
e a b c
φe =
e a b c
e a b c
φa =
 a e c b
e a b c
φb =
b c e a
e a b c
φc =
c b a e

Now, if we relabel K by the bijection

e a 
 b c

   
y y y y
1 2 3 4

we can represent K in cyclic notation:

 
1 2 3 4
φe = = (1)
1 2 3 4
1 2 3 4
φa = = (12)(34)
2 1 4 3
1 2 3 4
φb = = (13)(24)
3 4 1 2
1 2 3 4
φc = = (14)(23)
4 3 2 1

Now we can see that K ' H where H = {(1), (12)(34), (13)(24), (14)(23)}. That is, f : K → H is the
bijection

6
 f: 1
 2
 3
 4

   
y y y y
(1) (12)(34) (13)(24) (14)(23)

The Klein 4-group K has many other interesting properties, including

• K is the smallest non-cyclic group

• K is the underlying group of the four-element field

• K is the symmetry group of a non-square rectangle
• K is the group of bitwise exclusive or operations on two-bit binary values

• K = Z2 × Z2 , the direct product of two copies of the cyclic group of order 2

4 Conclusions
5 Acknowledgements
LATEX Source
https://www.overleaf.com/read/nbfyqkwsfmyc

References