MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I feel sort of silly asking this question. Unless I'm very much mistaken the paper I'm reading assumes the following statement:

Let $G$ be a finite group. We may embed it via the Cayley embedding into an ambient permutation group $G \leq S_{|G|}$. Then any automorphism of $G$ comes from conjugation by an element in $N_{S_{|G|}}(G)$.

Is this statement true?

share|cite|improve this question
Note that left and right Cayley embeddings may be different (although this doesn't affect the answer to your question). – S. Carnahan Jun 11 '10 at 19:29
The normalizer you mention is called the holomorph. It is the semi-direct product of G and Aut(G) and is very often used as a containing group where automorphisms become group elements. I think you'll find this in most group theory texts, though the permutation description is clearest in Burnside. Remember an automorphism is just a permutation of the elements of a group. – Jack Schmidt Jun 12 '10 at 15:16
This result reminds me of the Skolem-Noether theorem on automorphisms of simple subalgebras of central simple algebras, or at least the special case where the CSA is $M_n(K)$. Is there any connection here? – Pete L. Clark Jun 12 '10 at 18:13
up vote 15 down vote accepted

The statement is true. Let $g \in G$ and $\pi \in Aut(G)$. Let $\lambda_{g}$ be the corresponding left translation by $g$. Regard $\pi$ and $\lambda_{g}$ as elements of $Sym(G)$. Then for all $x \in G$,

$(\pi \lambda_{g} \pi^{-1})(x) = (\pi \lambda_{g}) ( \pi^{-1}(x)) = \pi( g \pi^{-1}(x)) = \pi(g) x = \lambda_{\pi(g)}(x)$.

Thus $\pi \lambda_{g} \pi^{-1} = \lambda_{\pi(g)}$.

share|cite|improve this answer
You would think this would be a more famous fact. I was always under the impression that outer automorphisms are much more mysterious than inner automorphisms. – James D. Taylor Jun 11 '10 at 19:27
This observation is the basis of a famous construction of Hall. If you iteratively form the embeddings: $G_{0} \to G_{1} \to \cdots \to G_{n} \to \cdots$ where $G_{0} = Sym(5)$ and $G_{n} \to G_{n+1} = Sym(G_{n})$, then the union is a simple locally finite group $G$ such that: (a) every finite group embeds in $G$; and (b) every isomorphism between finite subgroups is induced by an inner automorphism of $G$. – Simon Thomas Jun 11 '10 at 19:34
That sounds interesting. Do you have a reference? – James D. Taylor Jun 11 '10 at 19:35
Outer automorphisms are much more mysterious than inner ones: look at the symmetric group $G=S_6$, for instance. You get no practical help in studying its outer automorphisms just by knowing the general fact you quoted. Here the left or right Cayley embedding involves an enormous symmetric group. – Jim Humphreys Jun 11 '10 at 19:36

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.