42 Groups Acting on Themselves by Left Multiplication - Cayley's Theorem
--23 ✅ 2025-02-22
2
Question
List the elements of as and label these with the integers respectively. Exhibit the image of each element of under the Left Regular Representation of into .
Proof
Here:
Label
Element in
1
1
2
(1 2)
3
(2 3)
4
(1 3)
5
(1 2 3)
6
(1 3 2)
Then applying is done by considering applied to each in our set.
Clearly . For example . What we can do is use the labels and see the group action on each element. Namely, let . Then we get the following table:
v First (arg) | Second ()>
Associated
1
2
3
4
5
6
2
1
5
6
3
4
3
6
1
5
4
2
4
5
6
1
2
3
5
...
6
...
... and while we could fill out the whole table, really we're looking for each associated with each element via this group action. Notice that because is a homomorphism them for example:
Thus, this implies we do an induction. If then which is the trivial subgroup (which itself is normal). If satisfies the proposition then for some where and (still) then which itself must be a normal subgroup using the inductive hypothesis (apply to be the subgroup of itself. Then must be normal by the hypothesis, as desired). This completes the induction.
For the deduction, let . Then we just need to construct a subgroup such that , and then we can apply our proposition with . Namely, to construct this, since is a group then it is non-empty. Furthermore, it must have a non-identity element (since ). Then such that . Then choose . Notice that since then . Furthermore, by Finite Groups, Order of x divides Order of G, then divides , thus as desired.
Let be non-abelian where . Notice first that since is non-abelian, then . Consider making the subgroup . Notice that by Finite Groups, Order of x divides Order of G, then must divide 6. Clearly the order cannot be 1 (since is a non-identity element and is in ) but it also cannot be 6 (since if it was then since then there is some power where , but that would contradict not commuting as ). By a similar argument, we've shown that , so then another subgroup can be made .
Therefore, either or and the same is for . One could easily find contradiction in to show that one has order 2 and the other had order 3 (if not, then that implies or as contradictions). Without loss of generality, suppose . That implies . By Paul's Theorem then . But by the orders equaling then consequently (by being an Isomorphism, namely ). I'm doing this to show that while there will be one subgroup that follows to be normal, the other one won't which we can always choose!
Similarly, notice that is non-normal because due to the lack of commutativity. Thus choosing as our subgroup will always work! If the other case where occurred, then but we could choose instead as our non-normal subgroup.
Lastly, let's make an injective homomorphism into via Cayley's Theorem. Namely, without loss of generality that and as described before. We note some equivalences:
Thus notice that this is similar to . If we could show that and then notice that the first relation is correct. We'd want to show that . Notice we can eliminate the candidates for :
If can't be since then right-cancellation gives which is incorrect.
If which is wrong as they don't commute.
It can't be as then contradicts the lack of commutativity.
Thus and thus we can treat as just ! This helps us fill the table as now is an equivalence we can use. Some other equivalences that are derived then:
Thus , but one can show that via the map defined by:
Since these are the same generators and have the same equivalence relations. Notice but for the last relation:
Thus this is a valid isomorphism!
This suggests that all:
Abelian Groups of order 6 are congruent to because there is either elements of order 2, 3, or 6 (from our argumentation before), but by our same reasoning one would show that and thus which implies is Cyclic Group and thus congruent.
Non-abelian groups are congruent to by our reasoning above.
classifying all groups of order 6!