Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

More automorphisms

  1. Nov 20, 2004 #1
    Okay, so i'm having trouble understanding the following question:
    Determine the group of automorphisms of [tex]S_3[/tex].

    I understand that the automorphisms must match orders of the same element, and since there are three permutations of order 2 and two of order 3, there are 6 "possible" permutations. But I don't know where to go from here. I'm pretty sure there's a better way than to tediously go through all six possible automorphisms, and explicitly check whether each work or not. Am i missing something here? :confused:
    (I mean, how do you know when you've determined the group?)
  2. jcsd
  3. Nov 21, 2004 #2

    matt grime

    User Avatar
    Science Advisor
    Homework Helper

    You only need to look at the elements of order two since you know where (123) goes once you know where (12) and (23) go - that is the point of automorphisms - and this doesn't involve much work.

    Of course, you could just prove it has no outer automorphisms...

    In this case, every permutation of the 3 elements of order two is an automorphism, obviously, and this determines every automorphism, equally clearly, and hence the group Aut(S_3) = S_3.
  4. Nov 22, 2004 #3

    matt grime

    User Avatar
    Science Advisor
    Homework Helper

    I was thinking that this may well need more explanation.

    Every relabelling of the numbers 1,2,3 gives an automorphism of S_3 (and a relabelling of 1,..,n would give an automorphism of S_n). These correspond to the inner automorphisms - the ones where the group acts by conjugation. If you've done linear algebra it's a lot like a change of basis.

    This tells us that there is a copy of S_3 inside Aut(S_3). Now all we need to do is show that Aut(S_3) has at most 6 elements and we are done.

    Because S_3 has exactly 3 elements of order 2, and they generate S_3, then any permutation of them *might* be an automorphism, and any two automorphisms permute them in distinct ways, so there are at most 6 possible automorphisms, as we needed to show.

    Hence S_3 <= Aut(S_3) <=S_3

    so they are equal.

    The reason I thought I needed to clarify this was that I suspected it wasn't clear why this didn't show that the Aut(S_n) was something that it wasn't.

    It's important that n=3 here, so that all the elements of order 2 are of the same cycle type and generate the group.

    In general there is always a map from G to Aut(G) given by [tex] x \to f_x[/tex] where [tex]f_x(y)=x^{-1}yx[/tex] is how the aut f_x acts on G.

    1. Show this map from G to Aut(G) is a homomorphism
    2. What is the kernel?
    3. Hence conclude that when G=S_n it is an injection unless n=2

    FACT it is an isomorphism on S_n except for n=2 and n=6.
    Proof n=2 is easy since S_2 is abelian. n=6 is too hard for me to recall but essentially seems to be because 2*3=2+3+1. Have a look for a proof somewhere - there are some nice geometric ones available.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook