Understanding the Irreps of a Particular Representation of Symmetric Group Product

39 Views Asked by At

I'm reading this paper on the element distinctness problem, and I'm having some trouble parsing Claim 2. I've recently been going through The Symmetric Group by Sagan (Chapters 1 and 2 so far).

Here's the premise. For an alphabet $\Sigma$ of size $n$, consider the list of distinct elements $z=(z_1,\dots, z_n)$. The index permutations $\pi\in S_n$ just change the order of lists: $$z_\pi:=\pi z=(z_{\pi^{-1}(1)},\dots, z_{\pi^{-1}(n)}).$$ The alphabet permutations $\tau\in S_\Sigma$ relabel elements in the list: $$z^\tau:=\tau z=(\tau(z_1),\dots, \tau(z_n)).$$

Now we consider the matrices $V^\tau_\pi$ which perform the map $V^\tau_\pi=z^\tau_\pi$. Here's where my questions start, as the the author states "$V$ is isomorphic to the regular representation of $S_n$ and $S_\Sigma$.'' So, I know that the regular representation of $S_n$ is just consists of the algebra $\mathbb C[S_n]$, meaning I need an $S_n$-isomorphism between $S_n$ and $V$, yet I'm unclear on what $V$ is. I'd think it'd be $V=\text{span}\{V^\tau_\pi\mid \pi\in S_n, \tau\in S_\Sigma\}$, but the basis $\{V^\tau_\pi\}$ obviously has more than $n!$ elements, so that led me to believe they're referring to $\{V_\pi^\varepsilon\}$ and $\{V^\tau_\varepsilon\}$, is this correct?

On the other hand, Claim 2 says that $$V\cong \bigoplus_{\lambda\vdash n}\lambda\times \lambda,$$ which led me to the following question: When they say $V\cong \bigoplus_{\lambda\vdash n}\lambda\times \lambda$, is it convention for $\lambda$ to refer to the Specht module $S^\lambda$? Or is that even what they mean? I think it should be more clear why the claim is true when I know the answers to these two questions.

Thanks!