Introduction
Lagrange’s theorem is a beautiful result in group theory that has many neat applications. The notion of cosets is fundamental to the proof of Lagrange’s theorem and, independent of their utility in proving Lagrange’s theorem, they’re a very neat construction.

Cosets and equivalence relations
Let G and H be groups with H \le G. We say that a is congruent to b modulo H and write a \equiv b \pmod{H} if a^{-1} b \in H (note that we use -a+b if the operation of our groups is addition). The important thing to notice is that congruence modulo H is in fact an equivalence relation on G. In case the notion of equivalence relation is unfamiliar, I’ll define it; an equivalence relation is a binary relation on some set S that satisfies the following criteria (we’ll denote the operation by \rho):

    \[a \rho a, \forall a \in S  \quad (\text{reflexivity})\]

    \[a \rho b \implies b \rho a,  \forall a, b \in S \quad (\text{symmetry})\]

    \[a \rho b \text{ and } b \rho c \implies a \rho c,  \forall a, b, c \in S \quad (\text{transitivity})\]

If x \in G, then its equivalence class with respect to congruence modulo H is the set \{xh : h \in H\} = xH. We refer to this set as the left coset of x with respect to H. As you can probably guess, we define the right coset of x with respect to H as \{hx : h \in H\} = Hx. For groups G and H with H \leq G, we define \{gH:g \in G\} as the set of left cosets of H in G. The set of right cosets of H in G is defined analogously.

Proposition: Congruence between two numbers modulo H is an equivalence relation

Proof: Reflexivity: If a \in G, then a^{-1}a=e \in H and therefore a \equiv a \pmod{H}
Symmetry: If a,b \in G and a \equiv b \pmod{H}, then a^{-1}b \in H, and therefore (a^{-1}b)^{-1}=b^{-1}a \in H. But this means that b \equiv a \pmod{H}
Transitivity: Suppose a,b,c \in G, where a \equiv b \pmod{H} and b \equiv c \pmod{H}. Then a^{-1}b, b^{-1}c \in H. But in this case, H contains their product, a^{-1}bb^{-1}c=a^{-1}c. Thus, a \equiv c \pmod{H} \quad \blacksquare

The following proposition about equivalence classes (sets of elements that are ‘equivalent’ to each other under the equivalence relation and are denoted [a]) is very useful.

Proposition: The equivalence classes of a set partition it

Proof: Let’s say we’re working with some nonempty set S. We want to show that A = \{[a] : a \in S\} is a partition of S. Due to the reflexivity of equivalence relations, we know that a \rho a \implies a \in [a] \implies [a] \neq \emptyset. Hence each equivalence class is nonempty. Suppose that x \in [a] and x \in [b], which implies x \rho a and x \rho b. By symmetry, we have that a \rho x and then by transitivity we know that a \rho b, which shows that [a] = [b], showing that two equivalence classes only have common elements when they are the same equivalence class and demonstrating that equivalence classes partition a set. \quad \blacksquare

This of course means that, since cosets are an equivalence class, the left cosets H in G partition G. The index of H in G, denoted [G:H], is the number of left cosets of H in G

Lagrange’s theorem
Having spoken a little about cosets, we can now tackle Lagrange’s theorem.

Lagrange’s Theorem: Let G and H be groups. H \le G \implies |H| divides |G|

Proof: To prove this, we just have to use what we’ve learned about cosets. We’ve known that G is partitioned into left cosets, which means |G| is simply the sum of the sizes of the left cosets. If ah_1 = ah_2 then h_1 = h_2 by the cancellation law and hence aH consists of exactly |H| elements. This means that |G| =|H| \times [G:H]

\therefore |H| divides |G| \quad \blacksquare

Corollaries
I’ll now show two useful corollaries of Lagrange’s Theorem.

Corollary 1: Let G be a finite group and a \in G. Then the order of a divides the order of G.

Proof: The order of the element a is also the order of the cyclic subgroup generated by a, and that must divide the order of G \space  \quad \blacksquare

Corollary 2: Every group of prime order is cylic

Proof: Take e \neq a \in G where \lvert G \rvert is prime. As \lvert a \rvert divides \lvert G \rvert and \lvert a \rvert \neq 1, we must have that \lvert a \rvert = \lvert G \rvert. But then \lvert G \rvert = \lvert \langle a \rangle \rvert, and therefore G=\langle a \rangle \quad \blacksquare

Trending