2

Let $H,X,Y$ be the standard triple for $\mathfrak{sl}(2,\mathbb C)$. Let $\alpha$ be the root of $i\mathbb R H$ determined by $\alpha(H)=2$. Let $(\pi, V)$ be a finite dimensional continous representation of $SU(2)$. This gives $(\pi_*,V)$ a representation of $\mathfrak{sl}(2,\mathbb C)$. Let $V_\lambda$ denote the weight space of the weight $\lambda$ ie $V_\lambda=\{v\in V \text{ such that } Hv=\lambda v\}$. Therefore $$V_0=\{v\in V \text{ such that } Hv=0 v\}$$ and $$V_{\alpha/2}=\{v\in V \text{ such that } Hv= \alpha v/2 \}$$.

Show that $\pi_*$ is a direct sum of $\dim V_0+\dim V_{\alpha/2}$ irreducibles.

I know for compact groups we can decompose a representaiton by $\pi \sim \bigoplus_{\rho\in \hat G}\langle\pi,\rho \rangle \rho $. But $\mathfrak{sl}(2,\mathbb C)$ is not compact so we cannot use this.

My guess is there should be a way to decompose the representation into the different weights but I don't know how one would do this

Update:

I now that if $\rho$ is irreducible then $0$ and $\alpha/2$ cannot both be weights. If $\pi$ decomposes into a sum of $\rho_j$ does $\dim_{\pi}(V_k)$ in $\pi$ equal the $\sum_j \dim_{\rho_j}(V_k)$? I am also not sure on what I mean by $\dim_{\pi}(V_k)$ versus $\dim_{\rho_j}(V_k)$

Emily
  • 335
  • What is $V_0$ and $V_{\alpha/2}$? – Pratyush Sarkar Nov 26 '19 at 20:33
  • @PratyushSarkar They are the weight spaces for the weights $0$ and $\alpha/2$. I have added a description for $\alpha$ in the body but I don't fully understand roots. – Emily Nov 26 '19 at 20:43
  • @PratyushSarkar does this make sense now? – Emily Nov 26 '19 at 21:39
  • Look at this question and its answer, they should answer your question: https://math.stackexchange.com/q/3344288/96384. – Torsten Schoeneberg Nov 27 '19 at 07:43
  • @Emily Makes sense. But I'm not completely sure how to do this either. But you said that you know this is true for irreps. So now suppose you have an arbitrary rep $(\pi, V)$. Then this breaks into irreps $\bigoplus_j V_j$. Then those break into weight spaces $\bigoplus_{j, \eta} V_{j, \eta}$ where $V_{j, \eta}$ are weight spaces of $V_j$. Then it should be easy to see that weight spaces of $V$ are $V_\eta = \bigoplus_j V_{j, \eta}$. So then you can use the claim for irreps and just add it up. – Pratyush Sarkar Nov 27 '19 at 15:41

0 Answers0