0

I would like to study Sylow 2-subgroups of $S_{2^n}$. I studied Sylow 2-subgroups of $D_n$, but I couldn't find reference for $S_{2^n}$. Could anyone explain about Sylow 2-subgroups of $S_{2^n}$?

Thank you.

Shaun
  • 47,747
RV269
  • 29
  • 1
    You want to consider the graph automorphisms for a perfect $k$-level binary rooted tree for the $2$-Sylow subgroup of $S_{2^k}$. My answer below generalizes to all Sylows of all finite symmetric groups. – anon Oct 27 '21 at 03:54
  • Do you know what a permutation group or a group action is? I ask as for me (at the time I learned about it) runway44's answer would have been easier to digest if phrased in these terms. – j.p. Oct 27 '21 at 05:58

1 Answers1

5

I will explain the $p$-Sylow subgroup of any $S_n$. This requires understanding wreath powers.

Consider a perfect $k$-level $p$-ary rooted tree. That is, each node has $p$ child nodes, starting with one node at level $0$ and a total of $p^k$ nodes at level $k$. Below we have a $2$-level $3$-ary rooted tree depicted.

tree

Label the nodes at the bottom level $1$ thru $p^k$. Any automorphism of the tree corresponds to a permutation of the labels $\{1,\cdots,p^k\}$. Consider all automorphisms that come from cycling, in order, all leaves emanating from a given node (rather than an arbitrary permutation of them).

In the picture above, we get three separate cyclic groups generated by $(123)$, $(456)$, $(789)$ (which then form an internal direct product within $S_9$), but cycling the nodes directly below the root (and dragging their children with them) yields a cyclic group generated by $(147)(258)(369)$. The direct product of the first three cyclic groups and the latter cyclic group form a semidirect product, $(C_3\times C_3\times C_3)\rtimes C_3$, within $S_9$. This is called a wreath product, denoted $C_3\wr C_3$.

If a group $H$ acts on a set $\Omega$ used to index copies of $G$ in a direct product $\prod_\Omega G$, then the wreath product $G\wr_\Omega H$ is a semidirect product of $\prod_\Omega G$ and $H$, where conjugating $\prod_\Omega G$ by $H$ amounts to $H$ permuting the coordinates of $\prod_\Omega G$ according to how $H$ acts on $\Omega$. Thus, it has cardinality $|G\wr_\Omega H|=|G|^{|\Omega|}|H|$. If the set $\Omega$ is understood from context it can be dropped from the notation to write $G\wr H$.

In general th cycling automorphisms generate a subgroup which is a wreath power

$$ C_p^{\wr k} : = \underbrace{C_p\wr C_p\wr C_p\cdots\wr C_p}_k $$

(wreath products are associative up to isomorphism, so I didn't bother parenthesizing). By induction and the geometric sum formula, this has cardinality $|C_p^{\wr k}|=p^{p^{k-1}+\cdots+p+1}=p^{(p^k-1)/(p-1)}$.

The $p$-Sylow subgroup of $S_n$ can be determined from $n$'s base-$p$ expansion:

$$ n = \sum a_k p^k \\ P = \prod a_k C_p^{\wr k} $$

Above, we adopt the notation $mG:=\overbrace{G\times\cdots G}^m$ for notational suggestiveness. To see how $P$ is a subgroup of $S_n$, consider drawing $a_k$-many perfect $k$-level $p$-ary rooted trees then labelling all of their lowest level nodes $1$ thru $n$. To see it's a Sylow subgroup, check it's the correct size using Legendre's formula.

anon
  • 155,259
  • Thank you for your comment. But is there a more elementary approach? I haven't learned wreath product nor Legendre's formula. I will try to understand your detailed answer but I am just curious if there is easy way to understand without getting into p-ary roots and advanced concepts. Thanks a lot! – RV269 Oct 27 '21 at 03:56
  • 2
    @RV269 The case of the $2$-Sylow of $S_{2^k}$ is a bit simpler (see my comment under your question), but it still more or less requires thinking about these things. You can get away with staring at a binary tree without hearing about the word "wreath," but you still effectively have to work with the tree structure. Any other approach will just be this approach but without the visual aid. – anon Oct 27 '21 at 03:58
  • @runaway44 - very nicely explained! – Nicky Hekster Oct 27 '21 at 07:47
  • @RV269 If it helps, here's an example in concrete terms. Say $n=8$. Then a Sylow $2$-group is generated by: $4$ transpositions $(12),(34),(56),(78)$ – $2$ permutations that act like transpositions of two-groups $(13)(24), (57)(68)$ – and one permutation that acts like transposing groups of four elements $(15)(26)(37)(48)$. You can think about $(13)(24)$ for example as a "pointwise transposition of sets" $([12][34])$ (if my ad hoc notation makes sense to you. Likewise $(15)(26)(37)(48) = ([1234][5678])$. – Milten Oct 27 '21 at 12:06
  • @RV269 As for avoiding Legendre's formula, you can derive fairly easily that the Sylow $2$-groups of $S_{2^k}$ are of order $2^{2^k-1}$. The hard part then is to verify the order of the subgroup generated by the permutations I gave above. You don't necessarily need to know about wreath products, but I'd say you do need to know about (internal) direct and semidirect products to work it out. – Milten Oct 27 '21 at 12:12