Abstract Algebra - Fall 2024

Worksheet 8

Coset and Lagrange theorem

Equivalent relation and partition

Definition 1. A subset R of S × S is called a relation on a set S. If (a, b) ∈ R we denote

aRb

Definition 2. A relation R on S is called an equivalent relation if it satisfies the following

properties:

1. (Reflexive) xRx.

2. (Symmetric) If xRy then yRx.

3. (Transitive) If xRy and yRz then zRz.

Question 1. Let n ∈ Z+ and define a relation R on Z as xRy ⇐⇒ x = y (mod n).

Verify that R is an equivalent relation.

We have x = x (mod n) (reflexive).

If x = y (mod n) then y = x (mod n) (symmetric).

If x = y (mod n) and y = z (mod n), then y = z (mod n) (transitive).

Definition 3. A collection Si (i ∈ I) of subsets of S is called a partition of S if

[
Si ∩ Si = ∅ ∀i ̸= j and Si = S.
i∈I

1
Question 2. Define a collection {Si |i ∈ Zn } of subsets of Z as Si = {z ∈ Z|z = i(mod n)}.
Verify that {Si } is a partition of Z.

If z = i( modn), then z ̸= j( modn) with i ̸= j( mod n). So Si ∩ Sj = ∅ for i, j ∈ Zn and

i ̸= j.

For any x ∈ Z, there exists S

i ∈ Zn such that x = i( modn) as a remainder of the division of x
for n. So x ∈ Si . Therefore i∈Zn Si = Z

So {Si } is a partition of Z.
Question 3. Let ∼ be an equivalent relation on S. Define Sa = {x ∈ S|x ∼ a} to be the
equivalent class of a. Show that {Sa } form a partition of S by the following steps

a. For any x ∈ S ∃a ∈ S such that x ∈ Sa .

For any x ∈ S, we have x ∼ x, so x ∈ Sx .
b. If Sa ∩ Sb ̸= ∅ then Sa = Sb .

If Sa ∩ Sb ̸= ∅, then there exists m ∈ Sa ∪ Sb . Therefore, a ∼ a and m ∼ b. Using transitive,

a ∼ b. or Sa = Sb .

Coset

Let G be a group and H is a subgroup of G. Define a relation ∼L on G as

a ∼L b if and only if a−1 b ∈ H.

Denote aH = {ax | x ∈ H}, then we can write: a ∼L b if and only if b ∈ aH.

Question 4. Show that ∼l is an equivalent relation on G.

Suppose there exists a ∼L b, therefore there exists h ∈ H such that b = ah. Then, a = bh−1 .
Since b−1 ∈ H, a ∈ bH. So b ∼L a . Similarly, a ∼L a.

We also have if a ∼L b and a ∼L c, there exists h1 and h2 in H such that a = bh1 and a = bh2 .
Therefore, bh1 h−1 −1
2 = c. Since h1 h2 ∈ H, c ∈ bH or b ∼L c.

Question 5. Show that the equivalent class of a is aH, which means {aH|a ∈ G} is the
partition of G induced by the equivalent relation ∼L .

1. If a ∼l b, then b ∈ aH.

2
2. We want to prove if b′ ∈ aH, then a ∼l b’. We have there exists h1 ∈ H such that b′ = ah.
So a ∼l b′ So from the definition of coset, aH is the equivalent class of a.

Definition 4. Let H be a subgroup of a group G. The subset aH = {ax | x ∈ H} is called

the left coset of H containing a. We know that the collection of all left cosets of a subgroup
H forms a partition of G.

Question 6. Find all the left cosets of 3Z in Z.

1 + 3Z

2 + 3Z

0 + 3Z

Question 7. Find all the left cosets of ⟨3⟩ in Z12 .

1 + ⟨3⟩

2 + ⟨3⟩

⟨3⟩

Remark 1. From direct computation we see that all cosets have the same cardinality. We will
prove that this is correct for all groups.

Question 8. Let H ≤ G and a ∈ G. Find a bijection from H to aH and show that |H| = |aH|.

Consider the mapping from H to aH such that x is mapped to ax.

One to one : if a = b, then ax = bx.

Onto: If ax = bx, a = b (cancellation law).

So the mapping is bijection and cardinality is equal.

Lagrange theorem. Let H ≤ G, then the order of H divides the order of G.

Question 9. Prove Lagrange theorem.

Consider all classes formed by aH where a ∈ G, each classes have cardinality |H|. Since the
union of all those classes form G, we have |G| = k|H| or we proved the Langrage theorem.

Remark 2. The followings are easy corollaries of Lagrange theorem.

3
Question 10. Show that every group of prime order is cyclic.

Suppose there exist a group of prime K, the only subgroups of K has order 1 and |K|. The
subgroup with order 1 is the identity element. So beside it, K is the smallest subgroup since
the cyclic group generated by each elements are all the subgroups of K. That means cyclic
groups of other elements except identity one is K or K is a cyclic group.

Question 11. Show that the order of an element of a finite group divides the group’s order.

It follows by Langrage theorem since the order of a finite group is the order of a subgroup.

Definition 5. The number of left cosets is called the index of H in G and denoted (G : H).

Remark 3. We can define the right cosets of H using the relation ∼R defined by

a ∼R b if and only if ab−1 ∈ H.

Then ∼R is also an equivalent relation and the equivalent classes Ha are called the right cosets
of H. We can show that the right cosets also have the same cardinality as H. Just don’t assume
that left cosets are always the same as right cosets!

Question 12. Let H = ⟨µ⟩ be subgroup of D4 . Find all left cosets and right cosets of H.

Solution

The distinct right cosets are:

H, Hr, Hr2 , Hr3 , Hµr, Hµr2 , Hµr3

The distinct left cosets are:

H, rH, r2 H, r3 H, µrH, µr2 H, µr3 H

Remark 4. For abelian groups, left cosets and right cosets are the same. For non-abelian,
there are some special cases. When that happens, the cosets will form a group called the factor
group.

Question 13. Let ϕ : G → G′ be a group homomorphism. Show that for a, b ∈ G, if

a−1 b ∈ Ker(ϕ) then ab−1 ∈ Ker(ϕ), which implies that the left cosets of Ker(ϕ) are the same as
the right cosets.

Let Ker(ϕ) denote the kernel of the homomorphism ϕ, defined as:

Ker(ϕ) = {g ∈ G | ϕ(g) = e′ }

4
where e′ is the identity element in the group G′ .

Assume a−1 b ∈ Ker(ϕ). By the definition of the kernel, we have:

ϕ(a−1 b) = e′ .

Since ϕ is a homomorphism, we can express ϕ(a−1 b) as:

ϕ(a−1 b) = ϕ(a−1 )ϕ(b) = ϕ(a)−1 ϕ(b).
Thus, we have:
ϕ(a)−1 ϕ(b) = e′ .
This implies:
ϕ(a)−1 = ϕ(b) ⇒ ϕ(a) = ϕ(b).

Now, we need to show that ab−1 ∈ Ker(ϕ). We can calculate:

ϕ(ab−1 ) = ϕ(a)ϕ(b−1 ) = ϕ(a)ϕ(b)−1 .
Since we know ϕ(a) = ϕ(b), we can substitute:
ϕ(ab−1 ) = ϕ(b)ϕ(b)−1 = e′ .
Therefore, we conclude that:
ab−1 ∈ Ker(ϕ).

Next, we show that the left cosets of Ker(ϕ) are the same as the right cosets. The left coset of
Ker(ϕ) corresponding to an element a ∈ G is given by:
aKer(ϕ) = {ak | k ∈ Ker(ϕ)}.
Similarly, the right coset is:
Ker(ϕ)a = {ka | k ∈ Ker(ϕ)}.

To demonstrate that these cosets are the same, consider any b ∈ G. If a−1 b ∈ Ker(ϕ), we
already established that ab−1 ∈ Ker(ϕ).

Thus, we find:
aKer(ϕ) = a(a−1 b)Ker(ϕ) = bKer(ϕ).

This implies that for any a, b ∈ G, the left cosets and right cosets can be identified through
their respective representatives, establishing that they are indeed the same.

In conclusion, we have shown that if a−1 b ∈ Ker(ϕ), then ab−1 ∈ Ker(ϕ), which leads to the
fact that the left cosets of Ker(ϕ) are the same as the right cosets.