*In which we think about the link between equivalence relations and partitions, and meet cosets.*

- Definition of an
*equivalence class*. - Definition of a
*partition*of a set. - Theorem 27:
*Let be an equivalence relation on a set . The equivalence classes of partition .*To prove this, we checked that each equivalence class is non-empty, that between them they cover the set (their union is the set), and they are pairwise disjoint, using each of reflexivity, symmetry and transitivity along the way. - Theorem 28:
*Let be a partition of a set . For , write for the unique part in with . Define a relation on by if and only if . Then is an equivalence relation.*We checked reflexivity, symmetry and transitivity. - Corollary 29:
*There is a bijection between equivalence relations on a set and partitions of the same set $latex $.*This was a consequence of Theorems 27 and 28. - Definition of , and of binary operations and on .
- Lemma 30:
*The operations and on are well defined.*Once we’d worked out what we needed to worry about, it was straightforward to check that in fact we didn’t need to worry. - Proposition 31:
*is an Abelian group. Moreover, it is cyclic and isomorphic to . Furthermore, is associative and commutative on , and is distributive over .*The proof of this is an exercise. - Proposition 32:
*Take . Then*has a multiplicative inverse in if and only if .*If is prime, then is a field.**Let has a multiplicative inverse be the set of*units*in . Then is a group under multiplication.*The proof is an exercise on Sheet 4.

- Definition of a
*left coset*of a subgroup in a group . Definition of the notation for the set of cosets. Definition of the*index*of a subgroup in a group. Definition of*right cosets*.

### Understanding today’s lecture

Can you prove Propositions 31 and 32?

What are the left cosets of in ? What is the index of in ?

Pick a group and a subgroup. What are the left cosets? For example, you could pick a small dihedral group like (the symmetries of a square) and explore left cosets of subgroups of .

### Further reading

In Lemma 30, we showed that two functions we’d just written down were well defined. What does well defined mean? Here are some useful thoughts from Tim Gowers.

I’ve written before about the result that the non-zero integers modulo a prime form a group under multiplication, which is closely related to (for prime) being a field.

And here’s something a bit lighter, in case you’re feeling old as the end of term approaches.

### Preparation for Lecture 8

Can you say anything interesting about the index of a subgroup in a finite group , in terms of the orders of and ?

Take a finite group , and a subgroup . Can you show that the left cosets of partition ? What is the corresponding equivalence relation on the set ?

Can you show that all the left cosets of in have the same size?

April 17, 2017 at 2:53 pm

[…] come across (left) cosets of a subgroup in a group . They’re the translates where . Intriguing question: can we […]

April 24, 2017 at 10:10 am

[…] that are self-inverse — ) or 2 (). By counting the number of each, and remembering that equivalence classes partition the set, we saw that the number of classes of size 1 is even. Since it’s also at least 1, there must be […]

May 3, 2017 at 10:07 am

[…] Proposition 56: The orbits of an action partition the set. We defined an equivalence relation whose equivalence classes are precisely the orbits, and were then done by Theorem 27. […]