4
$\begingroup$

This might be a classic question, but since I am new to representation theory of the symmetric group, I am asking it here.

Suppose that $n=k+l+r$ where $k\geq l\geq r\geq 0$. Let $G$ be the symmetric group $S_n$ and $H$ be its Young subgroup $S_k \times S_l \times S_1^r$. What can be said about the induced representation $\mbox{Ind}_H^G 1$? What are its components? Can we say something about their multiplicities?

$\endgroup$
1
  • 3
    $\begingroup$ This is called the permutation module associated to the Young subgroup. The decomposition into irreducibles is given by the Kostka numbers - they have a nice combinatorial description. You can read a nice exposition in Sagan's book "The symmetric group". $\endgroup$
    – Noah White
    Aug 8, 2017 at 17:45

1 Answer 1

5
$\begingroup$

The answer is a special case of Young's rule. In my book, I give a very simple method for the slightly easier case where $r=0$. In that case we have: $$ \mathrm{Ind}_{S_k\times S_l}^{S_n} = \bigoplus_{s=0}^l V_{(n-s, s)}, $$ a multiplicity-free decomposition, where $V_{(n-s, s)}$ is an irreducible representation of dimension $\binom ns - \binom n{s-1}$.

In your more general case, the irreducible constituents will all be Specht modules of the form $V_{(s, t, u)}$, where $s\geq k$, $s+t\geq k+l$, and (of course) $s+t+u=k+l+r$. The multiplicities are given by Kostka numbers, which can be greater than one. For example, if $(k, l, r) = (3, 2, 1)$, then $V_{(5, 1)}$ occurs with multiplicity $2$.

The multiplicity of $V_{(s, t, u)}$ with $(s, t, u)=(5, 1, 0)$ can be calculated in sage as, for example:

sage: sage.libs.symmetrica.all.kostka_number([5, 1], [3, 2, 1])
2

More generally, replace [5, 1] with [s, t, u].

$\endgroup$
3
  • $\begingroup$ Can you please point out in which page does the case with $r=0$ appear in your book? Thank you! $\endgroup$
    – Jose Brox
    Jan 4 at 11:02
  • 1
    $\begingroup$ In Section 2.5, the representation $V_k$ is $V_{(n-k,k)}$ when $k\leq n-k$. $\endgroup$ Jan 7 at 0:11
  • $\begingroup$ Perhaps you could look into this related question of mine? math.stackexchange.com/questions/4844295 Thank you! $\endgroup$
    – Jose Brox
    Jan 15 at 22:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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