5
$\begingroup$

Let $G$ be a nontrivial group, denote $\text{Aut}(G)$ the group of all its automorphisms of $G$ and denote $S(G)$ the symmetric group on $G$, e.g. the set of all bijections $f:G\rightarrow G$. I would be interested, whether $\text{Aut}(G)$ and $S(G)$ can ever be isomorphic as a groups.

For a finite case, we can make the observation, that any permutation $f\in S(G)$ not fixing the identity cannot be automorphism, so in the finite case $S(G)$ and $\text{Aut}(G)$ aren't even equinumerous.

Could someone provide a rigorous argument for the infinite case? It seems like it is in fact impossible to have these two isomorphic as a groups, but can they atleast have the same cardinality?

We know that if $|G|=\kappa$ then $|S(G)|=2^\kappa$.

E: Adding some more ideas, someone could use:

Denote $S_{fix}(G)$ the set of permutations of $G$ that fix the identity element $e\in G$. Certainly we get $$ \text{Aut}(G)\preceq S_{fix}(G)\preceq S(G) $$ If we can show that $S_{fix}(G)$ and $S(G)$ aren't equinumerous, then we are done proving that cardinalities cannot ever be equal.

So, the cardinality problem is solved, now can we show that $\text{Aut}(G)$ and $S(G)$ cannot be isomorphic in the infinite case (or construct a counterexample?).

$\endgroup$
  • 1
    $\begingroup$ Note that $S(G)$ and $\operatorname{Aut}(G)$ are equinumerous if $|G|=1$. $\endgroup$ – Servaes Jul 14 '19 at 11:00
  • $\begingroup$ Yeah, that's a trivial case, I added that I require $G$ to be nontrivial. $\endgroup$ – Michal Dvořák Jul 14 '19 at 11:01
2
$\begingroup$

Suppose $C_\infty^\infty = \langle a_1 \rangle_\infty \times \langle a_2 \rangle_\infty \times \langle a_3 \rangle_\infty \times ...$ is the direct product of countably many isomorphic copies of an infinite cyclic groups. Then $|C_\infty^\infty| = \aleph_0$. And we have $S(\{a_1, a_2, a_3, ...\}) \leq Aut(C_\infty^\infty) \leq S(C_\infty^\infty)$. Thus $2^{\aleph_0} = |S(\{a_1, a_2, a_3, ...\})| \leq |Aut(C_\infty^\infty)| \leq |S(C_\infty^\infty)| = 2^{\aleph_0}$, which results in $|Aut(C_\infty^\infty)| = |S(C_\infty^\infty)|$

$\endgroup$
  • $\begingroup$ Could you please more elaborate why does $S(a_1,\ldots)$ inject into $Aut(C^\infty_\infty)$ ? $\endgroup$ – Michal Dvořák Jul 14 '19 at 18:15
  • $\begingroup$ @MichalDvořák, this is a free abelian group of countable rank. Any permutation of its free generators induces an automorphism. You can understand elements of $C_\infty^\infty$ as finitary integer sequences (integer sequences with finitely many non-zero entries) equipped of an operation of entry-wise addition. If you change the order of the entries, you will: 1) still get finitely sequences, 2) this operation will "respect" addition, 3) this operation is invertible. Thus we can conclude, that it is an automorphism. $\endgroup$ – Yanior Weg Jul 14 '19 at 18:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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