0
$\begingroup$

Let $\mathfrak h$ a Cartan subalgebra of the complex semisimple Lie algebra $\mathfrak g$ and let $V$ be a representation of $\mathfrak g$. Let $v \in V$ be a weight vector of weight $\lambda$, that is, $h v = \lambda(h) v$ for all $h \in \mathfrak h$ and let $n_\alpha \in \mathfrak g_\alpha$ be a root vector for the root $\alpha \in \mathfrak h^\ast$, that is, $[h,n_\alpha] = \alpha(h) n_\alpha$ for all $h \in \mathfrak h$. It is easy to calculate that $n_\alpha v$ (action of $n_\alpha \in U(\mathfrak g)$ on $v$) then is a weight vector of weight $\lambda + \alpha$. Now if $n_\alpha \in \mathfrak g_\alpha, n_\beta \in \mathfrak g_\beta$ are considered as elements in the universal enveloping algebra $U(\mathfrak g)$, then I want to deduce that $n_\beta n_\alpha v$ is a weight vector of weight $\lambda + \alpha + \beta$. How is this calculation done? (note: $n_\beta n_\alpha$ denotes multiplication in $U(\mathfrak g)$) Do I have to invoke the PBW-theorem and how?

$\endgroup$
4
  • $\begingroup$ $ n_\beta n_\alpha v= n_\beta(n_\alpha v)$! $\endgroup$ Sep 12, 2015 at 19:18
  • $\begingroup$ i tried $h(n_\beta n_\alpha v) = (h n_\beta) (n_\alpha v) = (n_\beta h - 2 n_\beta) (n_\alpha v)$ but I dont see how to proceed. Inserting what I already know would give $h(n_\beta n_\alpha v) = n_\beta (\lambda(h)+\alpha(h))n_\alpha v - 2 n_\beta n_\alpha v$, but I dont see how to proceed. (I assume relations $[h,n_\alpha] = -2 n_\alpha$, same for $\beta$ and $[n_\alpha,n_\beta] = h$) $\endgroup$
    – Mekanik
    Sep 12, 2015 at 19:45
  • $\begingroup$ You just need to use the result you already have twice. This is simpler than you are making it. $\endgroup$ Sep 12, 2015 at 19:53
  • $\begingroup$ ok I understand my mistake now, i assumed $[h,n_\alpha] = -2 n_\alpha$ for some reasons instead of simply $[h,n_\alpha] = \alpha(h) n_\alpha$... $\endgroup$
    – Mekanik
    Sep 13, 2015 at 13:59

1 Answer 1

0
$\begingroup$

You have a result that says for all weights $\mu$ and elements $u \in V_\mu$ (the $\mu$-weight space of $V$) and all roots $\gamma$ and all $n_\gamma \in \mathfrak{g}_\gamma$ we have $n_\gamma u \in V_{\mu+\gamma}$.

Apply this once with $u=v, \mu = \lambda, \gamma = \alpha$: you get $n_\alpha v \in V_{\lambda + \alpha}$. Apply it again with $u=n_\alpha v, \mu = \lambda + \alpha, \gamma = \beta$ and you get $n_\beta (n_\alpha v) \in V_{\lambda + \alpha + \beta}$. Isn't this what you want?

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .