Groups, Homomorphisms, and Group Actions
This section corresponds to Groups, Homomorphisms, and Group Actions — 2:44:50 in the video Proof of Fermat Last Theorem FROM SCRATCH. Noah Jang is the original creator of the exposition and examples referenced here. The interactive implementations, animations, additional explanations, and contextual material are my own work.
📌 Groups
The concept of groups plays a pivotal role in both physics and mathematics. It is deeply connected with the notion of symmetry.
Why do we need this?
From the website fermatically.com, we quote:
For Fermat’s Last Theorem, groups play a key role through Galois theory: the symmetries of the roots of a polynomial form a group – the Galois group. This group controls the algebraic structure of the associated field extension.
In Wiles’ proof, groups appear in several guises:
- The absolute Galois group \(G_\mathbb{Q}=\text{Gal}(\bar{\mathbb{Q}}/\mathbb{Q})\) — the central symmetry object of algebraic number theory. It acts on the division points of elliptic curves.
- Matrix groups \(GL_2(\mathbb{F}_p)\) and \(GL_2(\mathbb{Z}_p)\) — as target groups of the Galois representations that link elliptic curves with linear algebra.
- Hecke algebras — symmetries in the space of modular forms that generate algebraic structures on the Fourier coefficients.
Without groups, Wiles’ proof would not even be formulable.
Introduction
A square can be rotated (by 90°, 180°, 270°) and reflected (across four axes) – that makes a total of eight symmetries. Each individual symmetry maps the square onto itself, and two symmetries performed in succession yield another symmetry.
This observation can be generalised: wherever there are symmetries – in geometry, in physics, in number theory – there is a mathematical structure behind them that always obeys the same rules. This structure is called a group. — fermatically.com
Before defining a group, we first need to clarify one basic concept first.
🏷️ Binary operation
A binary operation \(*\) on a set \(S\) is a function that assigns to each ordered pair \((a,b)\) of elements of \(S\) some element of \(S\).
- Associativity: \((a*b)*c = a * (b* c)\)
- Commutativity: \(a*b = b*a\)
The most familiar binary operations we already know would definitely be addition and multiplication. For example, addition can be viewed as an operation that assigns the value 7 to the pair \((3, 4)\). Likewise, multiplication can be seen as an operation that assigns the value of 20 to the pair \((4, 5)\).
However, we may want to define more unusual operations as well.
🏷️ Group
A group \((G, *)\) is a set \(Gf\) with a binary operation \(*\) satisfying four axioms:
- Closure: For all \(a, b\in G,\ a * b \in G\).
- Associativity: The operation \(*\) is associative: \(\forall a,b,c \in G:\ (a*b)*c=a*(b*c)\)
- Identity element: There is an element \(e\in G\) such that \(e*x = a * e = a\) for all \(a \in G\).
- Inverses: For each \(a\in G\), there is an element \(a^{-1}\in G\), such that \(a*a^{-1}=a^{-1}* a=e\). This is called an inverse element.
👉 A group is abelian (or commutative) if its operation is also commutative — named after Niels Henrik Abel.
👉 The number of elements in a group \(G\) is called its order, denoted by \(|G|\).
👉 The order of an element \(g\) in a group \(G\) is the smallest positive integer \(n\) such that \(g^n = e\), where \(e\) is the identity element of the group.
✍️ Examples
The set of integers with addition is a group:
- The addition operation is clearly an operation on \(\mathbb{Z}\), as it always returns another integer ꪜ
- Addition is associative ꪜ
- The identity element is zero ꪜ
- Each element \(a\in\mathbb{Z}\) has an inverse, namely \(-a\) ꪜ
It is also an abelian group, since addition is commutative.
Similarly, \((\mathbb{Q}, +)\), \((\mathbb{R}, +)\), and \((\mathbb{C}, +)\) are also groups.
✍️ Counter example: \((\mathbb{Q}, *)\)
This is not a group, as there is no inverse element for zero: there is no such element for which we can write \(0*b=b*0=1\) ✗
So the same set can behave very differently, depending on what operations we put on it!
✍️ Example: \((\mathbb{Z}_n, +)\)
The group \((\mathbb{Z}_n, +)\). Remember that \(\mathbb{Z}_n\) is a subset \(\mathbb{Z}_n=\{0,1,2\cdots, n-1\}\).
Note that the addition operation is defined differently here: \(a + b\ (\text{mod}\ n)\)
Take for example \((\mathbb{Z}_3, +)\)
\[ \begin{eqnarray} 1 + 0 &=& 1 \\ 1 + 1 &=& 2 \\ 1 + 2 &=& 0 \end{eqnarray} \]
Let’s verify that \((\mathbb{Z}_5, +)\) is a group:
- The \(+\)-operation is associative ꪜ
- There is a zero element: \(0 + a = a + 0 =a\) ꪜ
- Every element has an inverse ꪜ
\[ \begin{eqnarray} 1 + 4 = 4 + 1 &=& 0\ (\text{mod}\ 5) \\ 2 + 3 = 3 + 2 &=& 0\ (\text{mod}\ 5) \\ \vdots \end{eqnarray} \]
Similarly, \((\mathbb{Z}_n, +)\) is a group.
✍️ Example: \((\mathbb{Z}_n \times \mathbb{Z}_m, +)\)
The addition operation is defined as
\[ (a_1, b_1) + (a_2, b_2) := ((a_1 +b_1)\ \text{mod}\ n, (a_2 + b_2) \ \text{mod}\ m) \]
With this operation, it forms a group. Let’s check this for \((\mathbb{Z}_3 \times \mathbb{Z}_4, +)\):
- With the above addition operation, we always get another element in the group ꪜ \[ \begin{eqnarray} 1 + 2 = 1 + 3 &=& (2, 1) \in \mathbb{Z}_3 \times \mathbb{Z}_4 \\ 2 + 3 = 3 + 2 &=& (1, 1) \in \mathbb{Z}_3 \times \mathbb{Z}_4 \end{eqnarray} \]
- The zero element is \((0, 0)\) ꪜ
- For every element, there is an inverse, for example \((2, 2) + (1, 2) = (0, 0)\) ꪜ
So indeed, \((\mathbb{Z}_n \times \mathbb{Z}_m, +)\) is a group.
❗ More generally, we have that
\[ (\mathbb{Z}_{n_1} \times \mathbb{Z}_{n_2} \times \mathbb{Z}_{n_3}\cdots, +) \]
forms a group.
✍️ Example: \((GL_n(\mathbb{R}), *)\)
The collection of invertible matrices \(GL_n(\mathbb{R})\) with matrix multiplication froms a group:
- \(\forall A, B\in GL_n(\mathbb{R}): A*B \in GL_n(\mathbb{R})\) as \(\det(AB)=\det(A)\det(B)\neq0\) ꪜ
- Matrix multiplication is associative ꪜ
- Obvisouly, \(GL_n(\mathbb{R})\) contains a zero element ꪜ
- As the determinant is non-zero by definition, every matrix has an inverse matrix ꪜ
✍️ Example: \(n\)-th roots of unity
If we define \(\mu_n = \{z\in\mathbb{C}\mid z^n=1 \}\) with multiplication, \((\mu_n, *)\) forms a group.
- \(\forall z_1, z_2 \in \mu_n: (z_1z_2)^n = z_1^nz_2^n =1\) ꪜ
- Multiplication is associative ꪜ
- There is one identiy element, namely \(1^n = 1\) ꪜ
- Every element has an inverse, as \(1 / z^n\) is also in the group ꪜ
🚨 When the operation is obvious, we usually just state the set without explicitly mentioning its operation!
🏷️ Subgroup
A subset of \(H\) of a group \(G\) is a subgroup (\(H\leq G\)) if \(H\) is itself a group under the operations of \(G\).
It is a concept that is very similar to a subspace or subset, so it is basically a smaller group sitting inside a larger group.
✍️ Examples
\[ (\mathbb{Z}, +) \leq (\mathbb{Q}_n, +) \]
Also,
\[ (SL_n(\mathbb{R}), *) \leq (GL_n(\mathbb{R}), *) \]
Another example would be \((\mu_2, *) \leq (\mu_4, *)\).
🏷️ Cyclic groups
A group \(G\) is cyclic if it can be generated by a single element. We write \(G=\langle a \rangle = \{a^n\mid n \in \mathbb{Z}\}\). The element \(a\) is a generator of \(G\).
So the whole group can be built by applying the generator multiple times on itself:
\[ \begin{array}{c} a \\ a * a \\ a * a * a \\ a * a * a * a \\ \vdots \end{array} \]
✍️ Examples
\[ \mathbb{Z} = \langle 1 \rangle = \langle -1 \rangle \]
Also
\[ \mathbb{Z}_n = \langle 1 \rangle = \langle -1 \rangle \]
Note that we also have
\[ \mathbb{Z}_7 = \langle 1 \rangle = \langle -1 \rangle = \langle 3 \rangle \]
as by adding 3, we get
\[ \{3, 6, 2, 5, 1, 4, 0\} = \mathbb{Z}_7 \]
Another example of a cyclic group is
\[ \mu_n=\langle e^{\frac{2\pi i}{n}}\rangle \]
🏷️ Isomorphic groups
So, to understand the group, we need to understand how the elements interact under that operation. One way to do that is to make an operation table.
For example, let’s look at the operation tables of both \(\mathbb{Z}_3\)
\[\begin{array}{|c|c|c|c|} + & 0 & 1 & 2\\ \hline 0 & 0 & 1 & 2\\ \hline 1 & 1 & 2 & 0\\ \hline 2 & 2 & 0 & 1 \end{array}\]and \(\mu_3\):
\[ \begin{array}{|c|c|c|c|} + & 1 & e^{\frac{2\pi i}{3}} & e^{\frac{4\pi i}{3}}\\ \hline 1 & 1 & e^{\frac{2\pi i}{3}} & e^{\frac{4\pi i}{3}}\\ \hline e^{\frac{2\pi i}{3}} & e^{\frac{2\pi i}{3}} & e^{\frac{4\pi i}{3}} & 1\\ \hline e^{\frac{4\pi i}{3}} & e^{\frac{4\pi i}{3}} & 1 & e^{\frac{2\pi i}{3}} \end{array} \]
Upon closer inspection, we notice that these operation tables have exactly the same structure, as there turns out to be an obvious one-to-one correspondence between the elements of each group:
\[ \begin{array}{l} 0 \longleftrightarrow 1 \\ 1 \longleftrightarrow e^{\frac{2\pi i}{3}} \\ 2 \longleftrightarrow e^{\frac{4\pi i}{3}} \\ \end{array} \]
👉 Two groups \((G, *)\) and \((G', *')\) are isomorphic if their elements can be paired up in a way that preserves the group’s structural properties. We write \(G\cong G'\).
❗ The two groups have very little in common. One group consists of integers, the other contains exponents with even imaginary numbers. Also, the first group is a group under addition, whereas the second group is a group under multiplication. However, although the two groups are very different, they have exactly the same structure❗
🧠 What matters is how the elements interact with each other. If two groups have the same operation structure after relabelling the elements, they are called isomorphic.
👉 \(\mathbb{Z}_3 \cong \mu_3\)
🏷️ Isomorphism: the map
The function(s) that can be used to check whether two groups are isomorphic or not is called an isomorphism. So what conditions should this function generally satisfy?
An Isomorphism \(\phi\) from a group \((G, *)\) to a group \((G', *')\) is a function \(\phi: G\rightarrow G'\) that satisfies two conditions:
- \(\phi\) is a bijection (one-to-one and onto).
- \(\phi\) preserves the operation: For all \(a,b\in G\), \[ \phi(A*B)=\phi(a)*'\phi(b) \]
An automorphism is an isomorphism from a group \(G\) to itself.
So the function, the isomorphism does not just match elements, it mathes them in a way that respects the operations of both groups.
❗ An isomorphism is a structure-preserving bijection between groups. If such an isomorphism exists between two groups, we say the two groups are isomorphic❗
✍️ Example 1: \(\mathbb{C}\cong\mathbb{R}^2\)
With \(\phi(a + bi) := (a, b)\) we can check that:
- The function clearly is a bijection ꪜ
- Preserves the operations:
\[ \begin{array}{ll} \phi\left((a_1+b_1i)+(a_2+b_2i)\right) &=& \phi\left((a_1+a_2)+(b_1i+b_2i)\right) \\ &=& (a_1 + a_2, b_1+b_2) \\ &=& (a_1, b1) + (a_2, b_2) \\ &=& \phi(a_1, b_1) + \phi(a_2, b_2)\quad ꪜ \end{array} \]
✍️ Example 2: \((\mathbb{R}, +)\cong(\mathbb{R}_+, *)\)
With \(\phi(x) := e^x\) we can check that:
- The function clearly is a bijection ꪜ
- Preserves the operations:
\[ \phi(x_1+x_2) = e^{x_1+x_2} = e^{x_1}e^{x_2} = \phi(x_1)\phi(x_2)\quad ꪜ \]
✍️ Example 3
\[\mathbb{Z}_4\cong\mu_4\quad\text{with}\ \phi(k) = i^k \]
Generally we have:
\[ \mathbb{Z}_n\cong\mu_n. \]
With the concept of isomorphism, we can start thinking about groups in a way that puts focus on their structure: when two groups are isomorphic, they are structurally the same, independent of how their symbols, elements, and binary operations are defined.
As a consequence, the following question make sense: “can we classify groups using isomorphisms”? Stated differently, can we make a list of possible group structures where we don’t distinguish between groups that are esentially the same.
Although this turns out to be extremely difficult, except for finite Abelian groups. This is called the fundamental theorem of finitely generated abelian groups.
🏷️ The fundamental theorem of finitely generated abelian groups
Every finitely generated abelian group \(G\) is isomorphic to a direct product of cyclic groups of the form
\[ \mathbb{Z}_{p_1} r_1 \times \mathbb{Z}_{p_2} r_2 \times \cdots \times \mathbb{Z}_{p_n} r_n \times \mathbb{Z}\times \cdots\times \mathbb{Z} \]
where the \(p_i\) are primes (not necessarily distinct).
For example, we suppose we know nothing about a group \(G\) except that it is abelinan and the order is \(|G|=12\). The theorem now states that
\[ G \cong \mathbb{Z}_4 \times \mathbb{Z}_3,\ \text{or}\ \mathbb{Z} \times \mathbb{Z}_2 \times \mathbb{Z}_2 \]
🏷️ Cosets
Let \(H\) be a subgroup of \(G\), and let \(a\in G\).
- The left coset of \(H\) containing \(a\) is the set \(aH=\{ah\mid h\in H\}\).
- The set of all left cosets of \(H\) forms a partition of the group \(G\).
- All cosets of \(H\) have the same size as \(H\).
✍️ Example 1: \(5\mathbb{Z} \leq \mathbb{Z}\)
So what are some cosets of \(5\mathbb{Z}\) (all multiples of 5)?
\[ \begin{eqnarray} 1 + 5\mathbb{Z} &=& \{1, 6, 11, 16, \ldots\}\\ 2 + 5\mathbb{Z} &=& \{2, 7, 12, 17, \ldots\}\\ 3 + 5\mathbb{Z} &=& \{-2, 3, 8, 13, \ldots\}\\ 4 + 5\mathbb{Z} &=& \{-1, 4, 9, 14, \ldots\}\\ 11 + 5\mathbb{Z} &=& \{-4, 1, 6, 11, 16, \ldots\} \end{eqnarray} \]
We now see that each coset is not unique (and nor is it a subgroup):
\[ 1 + 5\mathbb{Z} = 6 + 5\mathbb{Z} = 21 + 5\mathbb{Z} = \dots \]
A coset does partition a group nicely into disjoint pieces/partitions:
\[ \mathbb{Z} = 5\mathbb{Z} \cup (1 + 5\mathbb{Z}) \cup (2 +5\mathbb{Z}) \cup (3+ 5\mathbb{Z}) \cup (4 + 5\mathbb{Z}) \]
✍️ Example 2: \(H=\{0, 4\} \leq \mathbb{Z}_8\)
\(H\) is a subgroup of \(\mathbb{Z}_8\). The follwing four cosets partition the group \(\mathbb{Z}_8\):
\[ \begin{eqnarray} H &=& \{0, 4\}\\ 1 + H &=& \{1, 5\}\\ 2 + H &=& \{2, 6\}\\ 3 + H &=& \{3, 7\} \end{eqnarray} \]
✍️ Example 3: \(H=\{0, 4, 8\} \leq \mathbb{Z}_12\)
\(H\) is a subgroup of \(\mathbb{Z}_12\). The follwing four cosets partition the group \(\mathbb{Z}_12\):
\[ \begin{eqnarray} H &=& \{0, 4, 8\}\\ 1 + H &=& \{1, 5, 9\}\\ 2 + H &=& \{2, 6, 10\}\\ 3 + H &=& \{3, 7, 11\} \end{eqnarray} \]
🏷️ Factor groups (quotient groups)
If \(N\) is a normal subgroup of \(G\), the set of cosets, denoted by \(G/N\), forms the factor group. The operation is defined as \((aN)(bN)=(ab)N\). This operations is well-defined only if \(N\) is normal.
We have just seen that
\[ \mathbb{Z} / 5\mathbb{Z} = \{5\mathbb{Z}, 5\mathbb{Z}+1,5\mathbb{Z}+2, 5\mathbb{Z}+3, 5\mathbb{Z} + 4\} \]
👉 All the integers with the same remainder modulo 5 are grouped together into the same basket.
As such, this factor group closely resembles (i.e. is isomorphic to) \(\mathbb{Z}_5=\{0, 1, 2, 3, 4\}\). The operation between the cosets is then given by the multiplication of the representatives (\(a\) and \(b\)) of each coset:
\[ (5\mathbb{Z} + 2) + (5\mathbb{Z} + 4) := (5\mathbb{Z}) + (2 + 4) = (5\mathbb{Z} + 1) \]
Concluding, we have
\[ \mathbb{Z} / 5\mathbb{Z} \cong \mathbb{Z}_5 \]
The notion of a normal group falls outside the scope of this chapter, as going forward most groups will be normal anyhow.
📌 Homomorphisms
🏷️ Homomorphism
A homomorphism \(\phi\) from a group \((G, *)\) to a group \((G', *')\) is a map \(\phi:G\rightarrow G\) that preserves the group operation.
This means that for all \(a,b\in G\):
\[ \phi(a*b) = \phi(a) *'\phi(b) \]
A homomorphism is a map that respects the structure of the group.
For some context, let’s recall our two requirements of an isomorphism: bijection and preservation of the operation of the groups. If we remove the requirement of the bijection, we get a homomorphism. As a consequence, all isomorphism are homomorphisms.
✍️ Example 1
\[ \phi: \mathbb{Z} \rightarrow \mathbb{Z},\quad\phi(n)=2n \]
This is not a bijection, but it preserves the group action:
\[ \phi(n_1+n_2) = 2(n_1 + n_2) = 2n_1 + 2n_2 = \phi(n_1) + \phi(n_2)\ ꪜ \]
✍️ Example 2
\[ \phi: GL_n(\mathbb{R}) \rightarrow \mathbb{R}-\{0\},\quad\phi(A)=\det(A) \]
This is a homomorphism as
\[ \phi(AB) = \det(AB) = \det(A)\det(B) = \phi(A)\phi(B)\ ꪜ \]
🏷️ Kernel and image
The kernel of a homomorphism \(\phi: G\rightarrow G'\) is the set of elements in \(G\) that map to the identity in \(G'\).
\[ \ker(\phi) = \{g\in G\mid \phi(g)=e'\} \]
The image of \(\phi\) is the set of all elements in \(G'\) that are mapped to by some element of \(G\).
\[ \text{im}(\phi) = \{\phi(g)\mid g\in G\} \]
This is defined similarly as for the kernel and image of linear transformations.
✍️ Example 1
\[ \phi: \mathbb{Z} \rightarrow \mathbb{Z},\quad\phi(n)=3n \Rightarrow\ \begin{cases}\ker(\phi) &=& \{0\} \\ \text{im}(\phi) &=& 3\mathbb{Z}\end{cases} \]
✍️ Example 2
\[ \phi: \mathbb{R}^2 \rightarrow \mathbb{R},\quad\phi\begin{pmatrix} x \\ y \end{pmatrix}=x+y \Rightarrow\ \begin{cases}\ker(\phi) &=& \left\{a\begin{pmatrix} 1 \\ -1 \end{pmatrix}\mid a\in\mathbb{R} \right\} \\ \text{im}(\phi) &=& \mathbb{R}\end{cases} \]
📌 Group actions
Thus far we have looked at groups as objects of study by themselves. Very often though, the elements of a group represent a transformation that acts on something else, for example the group of rotations acting on vectors.
Why do we need this?
This becomes very important when we are going to look at Galois theories and modular, as the study of many abstract groups becomes way easier when we see how they act on other elements.
🏷️ Group actions and G-sets
A set \(X\) is called a G-set if there is a map (called a group action) \(*:G\times X\rightarrow X\) that satisfies the following conditions:
- \(e*x = x\quad \forall x\in X\)
- \((g_1g_2)*x = g_1*(g_2 * x)\quad \forall g_1, g_2 \in G\ \text{and}\ x\in X\)
✍️ Example 1: translations in the plane \(\mathbb{R}^2\)
\[ G=\mathbb{Z},\quad X=\mathbb{R}^2,\quad n*x=n+x \]
Does \(X\) suffice the conditions of a G-set?
- $ 0 * x =x$ ꪜ
- \((n_1+n_2) *x = n_1 * (n_2 * x)\) holds because of the associativity of addition ꪜ
✍️ Example 2: rotations of two-dimensiona vectors around the origin
\[ G=\mathbb{R},\quad X=\mathbb{R^2},\quad \theta(x, y) = \begin{pmatrix} \cos\theta & -\sin\theta \\ \sin\theta & \cos\theta\end{pmatrix}\begin{pmatrix}x \\ y \end{pmatrix} \]
Does \(X\) suffice the conditions of a G-set?
- \(\theta = 0 = \begin{pmatrix} \cos 0 & -\sin 0 \\ \sin 0 & \cos 0\end{pmatrix}\begin{pmatrix}x \\ y \end{pmatrix} = \begin{pmatrix}x \\ y \end{pmatrix}\) ꪜ
- \(\begin{pmatrix} \cos\theta_1 & -\sin\theta_1 \\ \sin\theta_1 & \cos\theta_1\end{pmatrix}\begin{pmatrix} \cos\theta_2 & -\sin\theta_2 \\ \sin\theta_2 & \cos\theta_2\end{pmatrix}\begin{pmatrix}x \\ y \end{pmatrix}\) holds because of the associativity of matrix multiplication ꪜ
The concept of orbit now follows quite naturally, as we often may want to know and study all outcomes resulting from group actions \(g_n*x\) (\(g_n \in G\)) on an element \(x\in X\).
🏷️ Orbit
Let \(X\) be a G-set. For an element \(x\in X\): the orbit of \(x\) is the set of all elements in \(X\) to which \(x\) can be moved by elements of \(G\):
\[ O_x = \{ g*x\mid g\in G\} \subseteq X \]
Stated differently, let all group actions work on \(x\) and collect the resulting elements. This collection is called an orbit. Generally, this is a subset of \(X\).
👨💻 Orbits demo
Experiment to see how each norm \(\|x\|\) determines exactly one orbit, and hence how the plane \(\mathbb{R}^2\) can be partitioned into orbits of the G-set of rotations.