4
$\begingroup$

What are the places where character table values of $S_n$ occurs naturally? one such an example is when we write power sum symmetric function of order n in terms of Schur function of order n the coefficient of each monomial will be the character value of the respective conjugacy class and respective irreducible representation in the character table of $S_n$.

$\endgroup$
  • 3
    $\begingroup$ One is tempted to say: In the representation theory of the symmetric groups! ;-) $\endgroup$ – Johannes Hahn Apr 12 '13 at 17:54
  • 3
    $\begingroup$ It is very weird that you accepted an answer so soon: you asked for what is essentially a list... $\endgroup$ – Mariano Suárez-Álvarez Apr 12 '13 at 18:21
  • 1
    $\begingroup$ Should be CW it would seem. $\endgroup$ – Benjamin Steinberg Apr 12 '13 at 18:32
  • $\begingroup$ I am new to mathoverflow. cant I accept more than one answer? I found that answer interesting. $\endgroup$ – GA316 Apr 12 '13 at 18:35
3
$\begingroup$

My favourite place where the character table of $S_n$ occurs naturally is in the enumeration of branched covers of Riemann surfaces (Hurwitz numbers).

In more detail, let $\Sigma$ be a closed Riemann surface of $g$ with $k$ marked points $x_1, \ldots, x_k$, and fix an integer $n\geq 1$. For each marked point $x_i$ pick a partition $\mu^i$ of $n$.

We define $\mathcal H_n(g,\mu_1 , \ldots ,\mu_k)$ to be the sum of $1/|Aut(f)|$ for each isomorphism class of branched covers $f: \widetilde{\Sigma} \to \Sigma$ which are ramified precisely at $x_i$ with ramification type $\mu^i$.

Then there is a formula due to Hurwitz (or perhaps Frobenius, or Burnside, depending on which version...?):

$\mathcal H_n(g,k) = \sum_{\nu \in P(n)} \left(\prod_{i=1} ^k|ccl(\mu^i)| \frac{\chi_\nu(\mu^i)}{\chi_\nu(1)}\right) \left( \frac{\chi_\nu(1)}{n!} \right)^{2-2g}$.

Here $\nu \in P(n)$ indexes the irreducible representations $\chi_\nu$ of $S_n$, and the $\mu^i$ are identifed with conjugacy classes in $S_n$.

$\endgroup$
  • $\begingroup$ Of course, this isn't really specific to the symmetric group... we have the same formula for every finite group $G$ (enumerating $G$-Galois covers). But the symmetric group case has historical significance. $\endgroup$ – Sam Gunningham Apr 12 '13 at 18:09
  • $\begingroup$ what is ccl in the expression for $Hn$? $\endgroup$ – GA316 Apr 12 '13 at 18:14
  • $\begingroup$ That means the size of the conjugacy class in $S_n$ corresponding to the partition $\mu ^i$. There is a formula for this, I just didn't want to write it down... $\endgroup$ – Sam Gunningham Apr 12 '13 at 18:20
  • $\begingroup$ @SamGunningham I don't think your first comment is quite correct. You count degree $n$ covers unbranched away from $p_1,...,p_k$ by counting homomorphisms $\pi_1(\Sigma\setminus\{p_1,...,p_k\})\to S_n$, and those with ramification data $\mu_1,...,\mu_k$ by forcing the images of loops around $p_i$ to lie in the right conjugacy class. For $\Sigma=\mathbf{P}^1$ a result of Frobenius then gives your formula. Maps $\pi_1(\Sigma\setminus\{p_1,...,p_k\})\to G$ do not correspond to $G$-Galois covers (otherwise all degree $n$ covers would be $S_n$ Galois), but those with $\text{Aut}(f)\subseteq G$. $\endgroup$ – Meow Mar 20 at 14:56
  • $\begingroup$ @Meow Apologies, you are correct. To make sense of what I say, you should replace the phrase "$G$-Galois cover" with "principle $G$-bundle" (I am not assuming the total space is connected). Note that there is an equivalence of categories between principal $S_n$-bundles and $n$-fold covers (taking the associated bundle with respect to the standard $S_n$-action on the set $\{1,2, \ldots, n\}$). I apparently have been using the terminology wrong... $\endgroup$ – Sam Gunningham Mar 20 at 15:27
1
$\begingroup$

As a somewhat trivial example, the sign of a permutation is a character of $S_n$, and the sign function appears in many contexts.

For instance, the sign function is used when calculating the determinant of a matrix by summing over permutations in $S_n$. Replacing the sign character in this formula with another character of $S_n$ gives the definition of the immanant of a matrix.

$\endgroup$
  • $\begingroup$ thanks. but I want little more natural answer. $\endgroup$ – GA316 Apr 12 '13 at 18:51

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.