It's well-known that any abelian transitive subgroup A of a symmetric group $S_{n}$ has order $n$. Moreover, does anyone know how many abelian transitive subgroups of $S_{n}$ and what do they look like?

  • $\begingroup$ My immediate gut reaction is that they should all be cyclic, but I don't really know how to prove it, or whether it's really true. And just counting the number of different transitive cyclic subgroups isn't entirely trivial either. $\endgroup$ – Arthur Jun 6 at 9:34
  • 3
    $\begingroup$ @Arthur The Klein-4 group is an example of an abelian, non-cyclic transitive subgroup of $S_4$. They do not need to be cyclic. $\endgroup$ – Jack Crawford Jun 6 at 9:39
  • $\begingroup$ @Ling I believe you meant to say that $n$ divides its order, not that it has order $n$, right? If it has order exactly $n$, I can't believe I haven't heard of this until now! My galois theory exam was this morning! $\endgroup$ – Jack Crawford Jun 6 at 9:53
  • 2
    $\begingroup$ @JackCrawford The order is precisely n because I restrict to abelian subgroup cases, see (math.stackexchange.com/questions/128098/…) $\endgroup$ – Ling Jun 6 at 11:35

Every abelian group $A$ of order $n$ acts transitively on itself by left translation (aka Cayley action). Therefore we can view any such $A$ as a transitive subgroup of $S_n$. Furthermore, given that this kind of an action necessarily has trivial stabilizers, such an action has to be isomorphic to the translation action. In other words, a given abelian group $A$ of order $n$ has an essentially unique (up to conjugation by an element of $S_n$) transitive action of this type.

The problem of listing abelian groups of a given order $n$ (up to isomorphism) is relatively easy, but does depend on having the full factorization $n=\prod_{i=1}^kp_i^{a_i}$ with $p_i$ ranging over the prime factors of $n$. You need to list all the possible $p$-parts for $p=p_i,i=1,2,\ldots,k$. This amounts to partitioning $a_i$ in all possible ways, and the group is then a direct sum of its Sylow $p_i$-subgroups. The number of different (up to isomorphism) abelian groups of order $n$ is thus $$ \prod_{i=1}^kp(a_i), $$ where $p$ is the partition function.

  • 1
    $\begingroup$ Yes that's a nice "formula" for the number of conjugacy classes of abelian subgroups of $S_n$. The number of groups in the conjugacy class of such a group $A$ is equal to $(n-1)!/|{\rm Aut}(A)|$, so the exact number abelian subgroups of $S_n$ is more difficult to express concisely. $\endgroup$ – Derek Holt Jun 7 at 7:10

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.