8
$\begingroup$

Let $m>1$ be an odd natural number, $x$ a $m$-cycle in $A_m$, the alternating group in $m$ letters, $C$ the conjugacy class of $x$ in $A_m$.

Question: How can I describe the elements in the set $\{ j \mid x^j \in C\}$ in terms of $m$?

For instance, if $C^\prime$ is the conjugacy class of $x$ in $S_m$, the symmetric group in $m$ letters, then $$ \{ j \mid x^j \in C\} = \{ j \mid (j,m)=1 \}, $$ where $(j,m)$ is the greatest common divisor of $j$ and $m$. But in $A_m$, $C^\prime$ splits in two conjugacy classes of $A_m$ of the same size: $C$ and the conjugacy class of $(1 2)\times(1 2)$ in $A_m$.

Thank you in advance. Fernando.

$\endgroup$
1
  • $\begingroup$ Retagged to add nt, co, and finite-groups based on tags of similar questions. $\endgroup$ Commented Feb 26, 2010 at 6:23

1 Answer 1

6
$\begingroup$

The set is the quadratic residues when $m$ is prime, but usually not when $m$ is composite. For example, $(0,1,2,3,4,5,6,7,8)$ is conjugate to $(0,2,4,6,8,1,3,5,7)$ in $A_9$ even though $2$ is not a square mod $9$, so there is no additional condition beyond $(j,9)=1$.

For $m$ odd, the sign of the permutation on $\mathbb Z/ m\mathbb Z$ of multiplication by $j$ is the Jacobi symbol $\big(\frac jm\big)$. (This perspective on the Jacobi symbol is natural from one of Gauss's proofs of quadratic reciprocity, but it's also theorem 1 here. Also see Zolotarev's lemma.) Since there are two conjugacy classes of $m$-cycles in $A_m$, $\big(\frac jm\big)=+1$ iff $x$ is conjugate to $x^j$ in $A_m$.

$\endgroup$
3
  • $\begingroup$ The link to math.us.edu.pl seems to be broken. I'm also unable to find any snapshot saved on the Wayback Machine. $\endgroup$ Commented Sep 13, 2023 at 15:44
  • $\begingroup$ @TheAmplitwist, re, based on the URL, it might be Szyjewski - Zolotarev's proof of Gauss reciprocity and Jacobi symbols, whose Theorem 1 does indeed seem to state the desired claim. $\endgroup$
    – LSpice
    Commented Sep 13, 2023 at 20:12
  • 1
    $\begingroup$ @LSpice Thanks! You're probably right, that must be the article that was meant. $\endgroup$ Commented Sep 13, 2023 at 20:15

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .