2
$\begingroup$

Let $G$ be an algebraic group and $W$ denote the underlying affine Weyl group. I will label representations of the principal block of $G$ by their alcoves, which in turn I label by the corresponding element of the Weyl group $(W,S)$.

Given $w, y\in W$ Carter-Payne construct a homomorphism from $\Delta(w)$ to $\Delta(y)$ providing $w<y$ in the Bruhat order and $\ell(w)+1=\ell(y)=\ell$.

My question (and it should be silly and easy hopefully) is this: given a reduced expression $\underline{y}$ for $y$, how can I tell if a subexpression $\underline{w}$ is of length $\ell-1$? My suspicion is that the answer is as follows:

Fix $\underline{y}=s_1s_2\dots s_p$ with $s_i\in S$. A subexpression $\underline{w}=s_1s_2\dots s_{k-1}\widehat{s}_k s_{k+1}\dots s_\ell$ is reduced of length $\ell-1$ if and only if $s_k$ is the first or last occurrence of the generator $s_k$ in the expression $\underline{y}$.

I can prove this for type $A$, but I'm not really a dab hand at other Weyl groups. My feeling is that this should be true and well known, but I can't find a reference.

$\endgroup$
2
  • 1
    $\begingroup$ Your claim for type A is not true. Take $\underline{y} = s_1 s_2 s_3 s_1 s_4 s_2 s_1$, a reduced expression for $y = 43251$ and $\underline{w} = s_1 s_2 s_3 s_4 s_2 s_1$, a reduced expression for $w = 42351$. $\endgroup$ Commented Apr 8, 2020 at 0:48
  • $\begingroup$ Thanks! I guess that's the end of that then! :) $\endgroup$ Commented Apr 8, 2020 at 9:48

0

You must log in to answer this question.