Let $G$ be a finite group acting on a finite set $X$, and $\chi$ be the corresponding permutation character over the field $\mathbb{C}$.

1)If $G$ acts transitively, then $\chi=\mathbb{1}+\theta$, ($\mathbb{1}$ is trivial character of $G$); moreover the trivial character appears only once in $\chi$.

2)If $G$ acts doubly transitively, then the character $\theta$ is irreducible.

(Ref. Linear Representations of finite groups -Serre)

Question: What can be said about $\chi$ when the action is primitive, $3$-transitive, or in general $n$-trnasitive?


If the action is primitive, this means only that a point stabilizer is a maximal subgroup. I do not think that this puts much restriction on $\theta,$ but others may be able to say more than I can. In general, the character inner product $1 + \langle \theta, \theta \rangle$ is the rank of $G$ as a permutation group (when $G$ is transitive). This is the number of orbits that a point stabilizer $G_{\alpha}$ has on the points other than $\alpha.$ Higher transitivity puts further restrictions on how $\theta$ restricts to $G_{\alpha}$, but I at least do not know of many easily described properties in terms of $k$-transitivity for $k >2$ which can be attributed to $\theta$.


While, as Geoff has said in his answer, the only nice properties of $\chi$ that come to mind if the action is $3$-transitive, say, are about what happens if $\chi$ is restricted to a point stabiliser, $k$-transitive sets are nice for a different reasons. Namely, you can produce more irreducible characters than just $\chi$ out of them. E.g. you can consider the set of unordered pairs of elements of $X$ (or more generally, the set of $l$-element subsets for $l\lt k$). Of course, the action is not quite transitive, because e.g. the pair $(x,x)$ can only be sent to another pair with equal entries, but you can analyse this defect and obtain explicit descriptions of other irreducible characters.

For a guided solution with nice results in the case of symmetric groups, see exercise 7 on one of my representation theory exercise sheets.

  • $\begingroup$ Thanks! To compute the inner products, does one just compute the characters by counting fixed points, and then sum up? I computed the characters on cycles, but duty called before I could organize the rest, and I'm not sure I am skilled enough at binomial coefficients to simplify the inner product. $\endgroup$ – Jack Schmidt Aug 29 '11 at 16:30
  • 1
    $\begingroup$ @Jack Thanks, I am glad you like it! Note that $\langle \pi_k,\pi_l\rangle = \langle \pi,1\rangle$, where $\pi$ is the permutation character corresponding to the set $X_k\times X_l$. Now, this inner product is just the number of $G$-orbits on that latter set. It remains to count the orbits in a clever way. I will not spoil the rest of the fun for you. If you do it this way, no manipulation of binomial coefficients is necessary, all you need is a small but clever insight to count the orbits. Feel free to email me if you would like me to tell you the rest of the solution. $\endgroup$ – Alex B. Aug 30 '11 at 4:09

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.