Think about the world around you for a second. You know, from like the giant swirls of a galaxy way out there down to the tiny, perfect structure of a crystal , or even just the patterns you see in buildings. Yeah, there's this amazing regularity, isn't there? Exactly. A kind of order, a structure. Maybe you call it symmetry. It reminds me of that quote from D'Arcy Thompson, the biologist: "The beauty of a snow crystal depends on its mathematical regularity and symmetry." (Read about D'Arcy Wentworth Thompson's theory on how physical and mathematical principles shape the growth and structure of living organisms. Instead of focusing solely on Darwinian natural selection, Thompson emphasized the role of physical laws and geometry in biological form and development.)
See, this is going to be in informal way, but that doesn't mean you do not read it carefully. Please read it very carefully it will help you in building a good foundation on Group Theory.
Let's start with a concrete example: A Regular Tetrahedron (pyramid shape with four identical triangle faces – equilateral triangles, all edges the same length .
Now, before talking about it's symmetry in detail, let me briefly define symmetry: Imagine you have an object, like a perfectly square tile or a beautiful snowflake. A symmetry is a transformation (like a flip, a turn, or even just leaving it alone) that you can do to the object, and after you're done, it looks exactly the same as when you started. It fits perfectly back into its original outline or space. It's like if you pick it up, spin it around somehow, put it back down, I wouldn't know you'd done anything unless I'd put some mark on one corner.
Pick a tetrahedron by one of its vertices (the corners), and imagine an axis going straight through one corner (vertex) and piercing the middle of the opposite triangular face. You now spin the shape around that axis.
Here’s what you will get:
A rotation of 120 degrees (that’s 2π/3 radians) keeps the shape looking identical.
Do another spin, same amount—240 degrees or 4π/3 radians—still looks the same.
Do it a third time (360 degrees) and, well, you’re back where you started.
And because there are 4 vertices, that gives us 4 axes, each allowing two non-trivial rotations (120° and 240°), plus the identity (doing nothing at all). So far, that makes the total: 4 × 2 = 8 distinct rotations
(We’re not double-counting the 360° spins because those are basically just the “do-nothing” move.)
Now take two edges that don’t touch each other (non-adjacent)—there are exactly three such pairs in a tetrahedron. Connect their midpoints with another imaginary axis. These axes run through the middle of the tetrahedron, and they let you spin it 180 degrees (aka π radians). And guess what? That spin also brings you back at something that looks exactly like the original tetrahedron. So you get:
3 more rotations, one for each edge-edge axis.
Now we count all the rotation, counting the do nothing operation once, we get total: 8+3+1 = 12 rotational symmetry.
Now here comes an interesting part, Suppose you take one of these 12 rotations and then, apply another one. what would be the result?
It’s always going to be from one of those 12 rotations we have already discussed. You cannot somehow create a 13th way! And this is what is called closure. The set is locked in—it’s a closed system under composition (formal word for doing an operation after some other operation).
https://www.geogebra.org/m/nph3q5rv
(In this not all rotation symmetry is shown, but it is sufficient for understanding purpose - Play with this for a while till you get a hang of it)
Let's again look at a familiar example and after this we will define a group formally.
We will now look at the equilateral triangle - three equal sides, three equal angles (60° each).Let’s label the corners: A, B, and C — so that we can keep track of what moves where.
There are exactly 6 such symmetries:
Identity (e): Do nothing. A stays A, B stays B, C stays C.
Rotation R120: Rotate 120° counterclockwise around the center. (A → B, B → C, C → A)
Rotation R240: Rotate 240° counterclockwise around the center. (A → C, B → A, C → B). This is like doing R120 twice.
Reflection FA: Flip the triangle about the line passing through vertex A and the midpoint of the opposite side BC. (A stays, B ↔ C)
Reflection FB: Flip about the line through B and the midpoint of AC. (B stays, A ↔ C)
Reflection FC: Flip about the line through C and the midpoint of AB. (C stays, A ↔ B)
I would like to make some observations here:
If you perform one symmetry on an object (say, rotate the triangle 120 degrees) and then you perform another symmetry on it (say, reflection ), the final position of the triangle is still one of its possible symmetric states. The combined/composite operation is itself a symmetry of the square.
For any object, there's always one symmetry that involves "doing nothing" – just leaving the object as it is. If you do this "do nothing" action and then another symmetry, you just get that other symmetry.
If you perform a symmetry action, you can always perform another symmetry action that brings the object right back to its starting position. For example, if you rotate a triangle 120 degrees clockwise, rotating it 120 degrees counter-clockwise (or 240 degrees clockwise) undoes the first rotation. This "undoing" action is also a symmetry, and when combined with the original action, it results in the "do nothing" state (our identity).
If you have three symmetry operations (let's call them A, B, and C) and you apply them one after another, it doesn't matter if you think of it as (doing A then B) and then doing C, OR doing A and then (doing B then C). The final state of the object will be exactly the same. This is known as the associativity property.
Now we will define what is a group in a more formal way. But what is the need, actually once we define a group in generality we will be in a position to identify groups easily. There may be other collection of objects that have the structure of the group and they maybe not even geometric shapes. To classify them formally we need a general definition.
Okay, so a group is fundamentally two things:
A set of elements (G). These could be rotations, numbers, matrices, whatever.
An operation ('◦') to combine any two elements from G to get another element, which must also be in G (closure!). - Binary Operation
This set and binary operation combo has to follow just three specific rules (axioms):
Associativity: If you're combining three elements (a,b,c) in order, it doesn't matter how you group them: (a◦b)◦c=a◦(b◦c). The parentheses can move, but the order can't.
Identity Element: There must be one special element (e) in G that doesn't change anything: a◦e=e◦a=a. For our tetrahedron, it was the "do-nothing" rotation.
Inverse Element: For every element x in G, there must be another element (y, often written x−1) in G that "undoes" it: x◦y=y◦x=e. For a rotation, it's rotating back.
So, you see, the very nature of symmetries—how they combine, how they can be undone, the existence of a "do nothing" option—lines up perfectly with the formal axioms of a group! In a way, the abstract definition of a group is inspired by these concrete properties of symmetries.
Okay so all the examples that I have discussed so far had geometric structure and some of their symmetries were easily visualised. Now I would like to discuss a very important symmetry group, which is very general in nature and also there is famous theorem that shows its generality.
Let’s talk about one of the most general and powerful types of symmetry: permutations—which just means “rearrangements.”
Think of a set of distinct items. Like 3 colored balls:
🔴 Red
🟢 Green
🔵 Blue
Each ball sits in its own spot. A permutation is just a way to shuffle these items around. You’re not adding or removing anything—you’re just reordering them. (more formally a permutation is a bijective map from a set to itself)
Let’s imagine this simple setup: A collection of 6 distinct points, sitting at fixed positions on a plane.
We’re not talking about a shape like a triangle or hexagon with straight sides—just 6 separate, labeled points, like pins on a board. Their positions are fixed, but what we’ll play with is which label (or identity) is where.
Now imagine we relabel or swap the points—shuffle them around without changing where the actual positions are. What you’re doing is:
Reassigning labels to the fixed spots
But you’re not changing the overall structure—all 6 spots are still filled, just their labels have changed.
Since all the points are distinct, every possible way of rearranging their labels is allowed, as long as each point ends up with a new unique label (or the same, if nothing changes). And each way to rearrange them is not keeping the structure, therefore all these possible rearrangements form a symmetry of this configuration. (Note: We’re thinking of the entire set as the object, not how it's drawn or how the points are connected, and this is the reason it is very fundamental in nature and once we get a good grasp of it you’re learning about the core idea of symmetry itself. And thanks to Cayley’s Theorem, you’re also secretly learning about every possible group, even the ones you haven't come across yet.
Keeping all this aside first let us find how many elements are there in this group and what is the group operation. To answer the first question we need to answer, in how many ways can we rearrange 6 distinct labels?
That’s just the number of permutations of 6 elements = 6! = 6 × 5 × 4 × 3 × 2 × 1 = 720
So there are 720 possible symmetries of this labeled set of 6 points.
This collection of 720 permutations forms a group, under the operation of composition ( do one permutation, then another and the result is still a valid permutation). It is easy to verify that the set of these permutations along with the composition operation satisfies all the group properties and this group is called the Symmetry group Sn
There's a super important idea in group theory called Cayley's Theorem. It basically says that every single group, no matter how abstract or weird it seems, is structurally the same as (isomorphic to) a subgroup (a part) of some Symmetric group Sn.
It's not just shapes! Let's take the set of all integers (Z) and the operation of addition (+). For now look at these numbers not just as quantities, but as actions or operations, specifically "adding that number." For example:
The integer 7 represents the action "Add 7.
"The integer -2 represents the action "Add -2" (which is just subtracting 2).
The integer 0 represents the action "Add 0."
This collection of "adding actions," with the operation being "one addition followed by another" (which is just regular addition), forms a group:
Closure: If you add any two integers, the result is always another integer. (e.g., "Add 5" then "Add -3" is "Add 2". And 2 is an integer).
Identity Element: The action "Add 0" is our identity. If you "Add 0" to any number, it doesn't change that number (x+0=x).
Inverse Element: For any action "Add a", its inverse is "Add −a". When you combine them, you get "Add 0" (our identity). (e.g., "Add 7" combined with "Add -7" results in "Add 0", since 7+(−7)=0).
Associativity: Adding integers is associative. For any integers a,b,c: (a+b)+c=a+(b+c). For instance, (2+3)+4=5+4=9, and 2+(3+4)=2+7=9.
Since the order of addition doesn't matter (e.g., "Add 2" then "Add 3" is the same as "Add 3" then "Add 2", because 2+3=3+2), this group (ℤ, +) is an commutative group ( also known as Abelian group named after Niels Henrik Abel) .
The way in which I have introduced the group (Z,+) is only because the core of this article is to relate groups with symmetry. But it is very tedious to always think in this way specially when the subject gets more complicated. Therefore it is better we familiarize ourselves with the abstract definition of '+' , the one we usually use from our childhood (it is a part of learning mathematics at each step we have to familiarize ourselves to move one step ahead)
So, (Z,+) is a group! and so are (Q,+) (rationals) and (R,+) (reals).
What about multiplication (*)? Is (Z,∗) a group?
Associative? Yes, (2×3)×4=2×(3×4).
Identity? Yes, it's 1. x×1=x.
Inverses? Noo. The inverse of 2 is 1/2. Is 1/2 an integer? No.
So, (Z,∗) is not a group because it fails the inverse axiom for most elements.
But! If you take the set of non-zero rational numbers Q*, or non-zero real numbers R*, or non-zero complex numbers C*, they do form groups under multiplication. We remove the zero because it has no multiplicative inverse, but everyone else does (1/x), and it stays within the set.
You might have noticed that for addition and multiplication of numbers, a+b=b+a and a×b=b×a. The order doesn't matter. This property is called commutativity. A group where the operation is commutative (x◦y=y◦x for all x,y) is called an Abelian group. All those number groups we just saw ((Z,+), (Q*,∗), etc.) are Abelian. But not all groups are! Sometimes, the order does matter. We will now see an interesting example:
Let's go back to shapes, but this time we will also include reflections. The dihedral group Dn is the group of all symmetries (rotations and reflections) of a regular n-sided polygon. Consider the hexagon (D6). It has:
6 Rotations: 0°, 60°, 120°, 180°, 240°, 300° (R0, R60, … etc.)
6 Reflections: 3 through opposite corners, 3 through midpoints of opposite sides (let's call them S1, S2 , … etc.)
That's a total of 12 symmetries. (Hey, same number as the tetrahedron rotations!!).
Let R be the smallest rotation (60°) and S be one specific reflection. We know R6
=e (rotate 6 times = do nothing) and S2
=e (flip twice = do nothing). But what happens when you mix them?
It turns out there's a crucial relationship: S◦R=R−1
◦S. This means: "Flip then rotate" is the same as "Rotate backwards, then flip".
Crucially, this means S◦R
=R◦S (unless R is 180° or 0°). If you flip the hexagon and then rotate it, it lands in a different position than if you rotate it first and then flip it. The order matters!
Let R be the smallest rotation (60°) and S be one specific reflection, we know that
But what happens when you mix them?
It turns out there's an important relationship between R and S:
Another interesting and useful example of non abelian group is The General Linear Group GL(n,R)
We started with the rotational symmetries of the 3D tetrahedron (12 elements) and the full symmetries of the 2D hexagon (D6 , also 12 elements). They have the same size.
Two different physical objects, yet the algebraic structure describing their symmetries has the same number of elements. What does that mean? Are these groups somehow related, or even the same in some deep mathematical sense, even though one describes 3D rotations and the other 2D spins-and-flips?
To answer this we need to study about isomorphism - which is a very crucial concept in entire mathematics and we also use this concept almost everyday in our lives. After studying the concept of isomorphism we will be ready to tell when two groups have same fundamental structure even when their elements are totally different.
The principles of group theory provide a powerful lens to see
underlying structures in the world around us and in mathematics itself. It's a cornerstone of modern algebra and has applications in physics, chemistry, computer science, and beyond!