Definition: Let be a group and a subgroup. Get be an element of . We call the set
a left coset of in , and the set
a right coset of in

  • In an Abelian group, there is no difference between left and right cosets
  • if is an additive group, we write cosets additively, e.g.
    • As additive groups are always Abelian,

Lemma: Let be a group and a subgroup. Let . Then if and only if
Proof: Suppose . Note that as is a subgroup of . But , and so , so for some . Hence as required.
Conversely, suppose . We want to show that . Let . Then for some . But , so , so . Therefore every belongs to . By a similar argument, every belongs to . Thus,

Definition: Let be a group and be a subgroup. Define the left index of in , denoted by , to be the number of left cosets of in . Likewise, define the right index of in to be the number of right cosets of in .

Lemma: Let be a group and a finite subgroup. If then and have the same number of elements as .
Proof: Let . We want to show that and have the same number of elements. Both and are finite, and the best way to show that two finite sets have the same number of elements is to set up a bijection between them.
Let
From the definition of it is clear that is in the coset whenever is in the subgroup . So the map makes sense.
To check this map is bijective we must check it is both injective and surjective:

  • Injectivity: Suppose two elements map to the same element in (). This means that . If we left multiply both sides by , we obtain , and thus
  • Surjectivity: Suppose , we want to show that is of the form for some . By definition, , so for some

Lemma: Let be a group and be a subgroup. Let . Then the cosets and are either equal or disjoint.
Proof: Suppose and are not disjoint. We want to show they are necessarily equal. As and are not disjoint, they must have a common element which we can denote . Then and . Thus Then By lemma we have

Definition: Let be a set and be subsets of . We say that is a partition of if the following conditions hold:

  • If then (All subsets are pairwise disjoint)
  • for

Lemma: Let be a finite group and a subgroup. The left cosets of in form a partition of .
Proof: Let be the distinct left cosets of . As they are distinct, we know by lemma that they are disjoint. Suppose now that . Then must equal one of the . But as . Hence, the cosets are not only disjoint, but every element of belongs to one of them, and so their union is equal to . Finally, so , so all cosets are non-empty, and so the cosets form a partition of .

Theorem (Lagrange’s Theorem): Let be a finite group and a subgroup. Then
Proof: Let be the distinct left cosets of . By lemma, these form a partition of , and so Now, by lemma Hence where is the number of left cosets of in , which we defined to be the index of in , so

Corollary: Let be a group and a subgroup. Then
Proof: This follows from Lagrange’s Theorem as the index is an integer.

Corollary: Let be a finite group. Let have order . Then
Proof: Let the cyclic group generated by . We know from lemma that . By corollary we have

Let be a subgroup of . A conjugate of has the form for some

We say that a subgroup of is normal if and only if for all . That is, is normal if and only if it is equal to all its conjugates. We write to denote that is a normal subgroup of .

Lemma: Let be a subgroup of . Then the following are equivalent.

  1. is normal in
  2. for all
  3. for all
  4. for all
  5. for all
    Proof: It is easy to see that and also that . Let’s do . Suppose for all . Then, since we have . Left multiplying by and right multiplying by gives As and we have

Lemma: Let be a finite group and let be a subgroup of of index 2. Then is normal in .
Proof: We want to show that for all . By lemma, if and only if .

  • Suppose first that . Then and , and so
  • Suppose instead that . Then . But has index 2 in and so has exactly 2 left cosets, which must be and . Therefore, and , since cosets form a partition. Thus, . Similarly, . Hence, .

Let be a group and a subgroup. We write for the set of left cosets of in :
If is normal subgroup of , we define multiplication on by
Lemma: Suppose is normal in . Then is well-defined.
Proof: Suppose and . For the definition to be sensible we want . By lemma we know that and . We want to prove that . However,
We know that as is normal and , we have . Therefore, .

Lemma: Let be a group and a normal subgroup. Then with is a group with identity element and inverses given by . Moreover, if is finite then We call the quotient group of over .
Proof: The fact that is a group easily follows from the fact is a group. For the last part, note that is the set of left cosets of , so by definition of the index. By Lagrange’s Theorem,

Let be a field. Let and be -vector spaces. A linear transformation is a map satisfying:

  • for all
  • for all and
    We define the kernel and image of the linear transformation by
    We recall that is a subspace of and is a subspace of .

Definition: Let , be groups and let be a map. We say that is a homomorphism of groups if for all

Definition: Let and be groups. A map is an isomorphism if it is a bijective homomorphism. If and are isomorphic we write .

Associated to any homomorphism are its kernel and image:

Theorem: Let be a homomorphism of groups. Then

  • is a normal subgroup of
  • is a subgroup of

Lemma: Let be a homomorphism of groups. Then is injective if and only if
Proof: Suppose . Let and suppose . Then . Thus and so and hence . Therefore is injective.
Conversely, suppose is injective. Let . Thus . As is injective, . Hence

Theorem (The First Isomorphism Theorem): Let be a homomorphism of groups. Let Then is a well-defined group isomorphism.
Proof: Lets show first that is well-defined. Let and suppose . sends to and to . For to be well-defined we want . tells us that and so by definition of kernel, . As is a homomorphism, we have , so . Thus, is well-defined.
To show that is a homomorphism note that

To show that is injective we apply lemma. Let . Then . Thus . Therefore, . It is clear from the definition of that it is surjective, and so is an isomorphism.

Let . An important homomorphism to know is the sign of a permutation. Recall that a permutation is even if it can be written as a product of an even number of transpositions, and odd if it can be written as a product of an odd number of permutations. A permutation cannot be both even and odd.

We define the sign of a permutation as follows: We obtain a homomorphism How do we know that sign is a homomorphism? We need to check that . Write as products of transpositions. Then
However, which is a product of transpositions. Thus Hence sign is a homomorphism

Theorem: Let . Then is a normal subgroup of . Moreover, Proof: Now that we’ve checked that sign is a homomorphism, we can observe that the kernel is defined to be So we know that is a normal subgroup of .
It is easy to see that sign is surjective, since and . So
Applying the first isomorphism theorem to gives an isomorphism
Thus . Therefore

Lemma: Let and be cyclic groups of order . Then and are isomorphic.
Proof: Let and , where and have order . Define
Is this well-defined? Suppose there are two integers and such that . The map sends to and to . For the definition to be sensible, must be equal to . tells me that and since has order we have . But also has order , so and hence . It follows that is well-defined.
We now need to check that is a homomorphism. We want which is the same as , which is true. Finally, we need to show that is a bijection. Let Then is well-defined as before and are mutual inverses. Since has an inverse it must be a bijection.

Lemma: Let be a prime. Any group of order is isomorphic to
Proof:

Definition: Let and be groups. We define the direct product of and to be i.e. is the set of ordered pairs where and . The binary operation on is
Lemma: is a group with identity element and inverse given by . If and are finite then Proof:

Theorem: The only groups of order 4 are and
Proof: What this theorem states is that any group of order 4 is isomorphic to either or . Lets write where has order 2. Then Note that has order 1 and all other elements have order 2. In particular, and are non-isomorphic as the latter has no elements of order 4.
Let be a group of order 4. We want to show that or . If is cyclic, then . Thus we may suppose that is not cyclic, and so has no element of order 4. Write . The elements have order 2 by Lagrange’s Theorem.

  • What is ? It must be one of the 4 elements of . If then giving a contradiction. If then giving a contradiction. Likewise gives a contradiction. Therefore, . We find that whenever we multiply two distinct elements among , we obtain the third one.
    Now, let and we just check (with the help of a multiplication table) that is an isomorphism. Thus, .

One way of thinking about isomorphisms of groups is that is a map that sends the multiplication table of to the multiplication table of .

Theorem (Fundamental Theorem of Finite Abelian Groups): Let be a non-trivial finite abelian group. Then there are integers , such that
Here, is the cyclic group of order . The integers are called the invariants of the finite abelian group . Note that they are positive integers, that each divides the next one, and that .

A regular -gon has symmetries (consisting of rotations and reflections). The set of these symmetries forms a group which we will denote , which is called the dihedral group of order .

  • Note that some sources will denote this , however has been chosen to be consistent with the resources from Algebra 1 (MA151) and Algebra 3 (MA268)

Lemma: Let . Suppose fixes vertices 1 and 2. Then ^512
Proof: The proof is geometric. Suppose fixes vertices 1 and 2. Then fixes the whole line segment joining these two vertices. But also fixes (the centre of the -gon). Therefore, fixes the whole -gon and so

Lemma: Let . Suppose and . Then ^513
Proof: Here we consider and as permutations of . Suppose and where and are vertex numbers. Let . Then By lemma we get , so

Theorem: Let . Then In particular, and is a normal subgroup of index 2.
Proof: Let . Let . Note that must map the vertices adjacent to the vertex 1 (2 and ) to the vertices adjacent to the vertex ( and . Therefore, we have 2 cases, either:

\begin{cases} n \mapsto k −1 & \\ 1 \mapsto k & \\ 2 \mapsto k+1 & \end{cases}$$ or $$a: \begin{cases} n & \mapsto k + 1 \\ 1 & \mapsto k \\ 2 &\mapsto k—1 \end{cases}$$ In the first case, let $b = r^{k-1}$. Note that $$a(1) = k = b(1), \quad a(2) = k + 1 = b(2)$$ so $a = b = r^k$ by [[#ae07c0|lemma]]. Now we take the second case. Note that $$s : \begin{cases} n & \mapsto 2 \\ 1 & \mapsto 1 \\ 2 & \mapsto n \end{cases}$$ So $$as : \begin{cases} n & \mapsto k—1 \\ 1 & \mapsto k \\ 2 & \mapsto k + 1 \end{cases}$$ Hence $as = r^{k-1}$ so $a = r^{k-1}s^{-1} = r^{k-1}s$. We conclude that any $a \in D{2n}$ belongs to either $R$ or $Rs$. Thus $D{2n} = R \cup Rs$. Note that any two cosets $R$ and $Rs$ are distinct as $s \not \in R$. Therefore $[D{2n} : R] = 2$ and so $R$ is normal in $D{2n}$ by [[#6f1afc|lemma]]. Therefore $Rs = sR$ and so $D_{2n} = R \cup RS$ as required. Lemma: With $r, s$ as above, $$r^n = \text{id}, \quad s^2 = \text{id}, \quad srs = r^{-1}$$ *Proof:* The first two relations are trivial. For the third, note that $s^{-1} = s$. Therefore, $srs = srs^{-1}$. As $R$ is normal in $D_{2n}$ and $r \in R$ we have $srs \in R$. So $srs = r^k$ for some $k$. We compute $$(srs)(1) = (sr)(s(1)) = (sr)(1) = s(r(1)) = s(2) = n$$ Hence $srs = r^{n-1} = r^{-1}$ ## Generators ## Group Presentations ## Homomorphisms From Groups Defined By Presentations ## The Quaternion Group $Q_8$ ## A Presentation for $D_{2n}$ # Classification of Groups ## Groups with Exponent 2 ## Groups of Order 6 ## Groups of Order 8 # Group Actions ## Group Actions ## Cauchy’s Theorem ## Conjugation ## Conjugacy Classes in $S_n$