Bourbaki

Proposition 2

Tag 0163, permanent. Cite it as [Bourbaki, Tag 0163].

Lie Groups and Lie Algebras, chapter VIII, § 9. The formula of Hermann Weyl

Let λ,μ∈P++\lambda , \mu\in P_{++}. In R(g)\mathscr{R}(\mathfrak{g}), we have

[λ].[μ]=∑ν∈P++m(λ,μ,ν)[ν] [\lambda ].[\mu] =\sum_{\nu\in P_{++}}m(\lambda , \mu, \nu )[\nu ]

with

m(λ,μ,ν)=∑w,w′∈Wε(ww′)P(w(λ+ρ)+w′(μ+ρ)−(ν+2ρ)) m(\lambda , \mu, \nu ) =\sum_{w,w'\in W}\varepsilon (ww')\mathfrak{P}(w(\lambda +\rho ) +w'(\mu+\rho )-(\nu + 2\rho ))

Let E,FE,F be finite dimensional simple g\mathfrak{g}-modules of highest weights λ,μ\lambda , \mu. Let lνl_{\nu} be the length of the isotypical component of E⊗FE\otimes F of highest weight ν\nu. It suffices to show that

lν=∑w,w′∈Wε(ww′)P(w(λ+ρ)+w′(μ+ρ)−(ν+2ρ))(2) l_{\nu}=\sum_{w,w'\in W}\varepsilon (ww')\mathfrak{P}(w(\lambda +\rho ) +w'(\mu+\rho )-(\nu + 2\rho )) \tag{2}

Put c1=c_1= ch(E) =∑σ∈Pmσeσ,c2==\sum_{\sigma\in P}m_{\sigma}e^{\sigma},c_2= ch(F), and d=J(eρ)d= J(e^{\rho}), where J is defined as in no. 2. We have ∑ξ∈P++lξ\sum_{\xi\in P_{++}}l_{\xi}ch[ξ]=\xi ] = ch(E ⊗F\otimes F) =c1c2=c_1c_2

so, after multiplying by dd and using Th. 1,

∑ξ∈P++lξJ(eξ+ρ)=c1J(eμ+ρ)=(∑σ∈Pmσeσ)((∑w∈Wε(w)ew(μ+ρ))(3) \sum_{\xi\in P_{++}}l_{\xi}J(e^{\xi+\rho}) =c_1J(e^{\mu+\rho}) =(\sum_{\sigma\in P}m_{\sigma}e^{\sigma})((\sum_{w\in W}\varepsilon (w)e^{w(\mu+\rho)}) \tag{3}
=∑τ∈P(∑w∈Wε(w)mτ+ρ−w(μ+ρ))eτ+ρ =\sum_{\tau\in P}(\sum_{w\in W}\varepsilon (w)m_{\tau+\rho-w(\mu+\rho)})e^{\tau+\rho}

Now, if ξ∈P++,ξ+ρ\xi \in P_{++},\xi +\rho belongs to the chamber defined by B (Chap. VI, §1, no. 10); thus, for all w∈Ww\in W distinct from 1, we have w(ξ+ρ)∉P++w(\xi +\rho )\notin P_{++}. Consequently, the coefficient of eν+ρe^{\nu+\rho} in ∑ξ∈P++lξJ(eξ+ρ)\sum_{\xi\in P_{++}}l_{\xi}J(e^{\xi+\rho}) is equal to lνl_{\nu}. In view of (3), we obtain

lν=∑w∈Wε(w)mν+ρ−w(μ+ρ) l_{\nu}=\sum_{w\in W}\varepsilon (w)m_{\nu+\rho-w(\mu+\rho)}

that is, by Prop. 1,

lν=∑w,w′∈Wε(w)ε(w′)P(w′(λ+ρ)−(ν+ρ−w(μ+ρ)+ρ)) l_{\nu}=\sum_{w,w'\in W}\varepsilon (w)\varepsilon (w')\mathfrak{P}(w'(\lambda +\rho )-(\nu +\rho -w(\mu+\rho ) +\rho ))

which proves (2).

Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.

Cited by

Cites

Transcribed from the printed volume. Not translated.