Bourbaki

Lemma 1

Tag 00WZ, permanent. Cite it as [Bourbaki, Tag 00WZ].

Lie Groups and Lie Algebras, chapter VII, § 5. Decomposable linear Lie algebras

Let n\mathfrak{n} be a Lie subalgebra of gl(V)\mathfrak{g}\mathfrak{l}(V) consisting of nilpotent endomorphisms, and N the subgroup expn\mathfrak{n} of GL(V)(\mathbf{G}\mathbf{L}(V) (§3, no. 1, Lemma 1).

(i) Let ρ\rho be a finite dimensional linear representation of n\mathfrak{n} on W, such that the elements of ρ(n)\rho (\mathfrak{n}) are nilpotent, W′W' a vector subspace of W stable under ρ,ρ1\rho ,\rho_1 and ρ2\rho_2 the subrepresentation and quotient representation of ρ\rho defined by W′,π,π1,π2W',\pi , \pi_1, \pi_2 the representations of N compatible with ρ,ρ1,ρ2(\rho , \rho_1, \rho_2(§3, no. 1). Then π1,π2\pi_1, \pi_2 are the subrepresentation and quotient representation of π\pi defined by W′W'.

(ii) Let ρ1,ρ2\rho_1, \rho_2 be finite dimensional linear representations of n\mathfrak{n} such that the elements of ρ1(n)\rho_1(\mathfrak{n}) and ρ2(n)\rho_2(\mathfrak{n}) are nilpotent, and π1,π2\pi_1, \pi_2 the representations of N compatible with ρ1,ρ2\rho_1, \rho_2. Then π1⊗π2\pi_1\otimes \pi_2 is the representation of N compatible with ρ1⊗ρ2\rho_1\otimes \rho_2.

(iii) Let ρ1,ρ2\rho_1, \rho_2 be finite dimensional linear representations of n\mathfrak{n} on vector spaces V1,V2V_1,V_2, such that the elements of ρ1(n)\rho_1(\mathfrak{n}) and ρ2(n)\rho_2(\mathfrak{n}) are nilpotent, ρ\rho the representation of n\mathfrak{n} on Hom(V1,V2_1,V_2) determined by ρ1,ρ2\rho_1, \rho_2. Let π1,π2\pi_1, \pi_2 be the representations of N compatible with ρ1,ρ2\rho_1, \rho_2, and π\pi the representation of N on Hom(V1,V2_1,V_2) determined by π1,π2\pi_1, \pi_2. Then π\pi is the representation of N compatible with ρ\rho.

Assertion (i) is clear. Let ρ1,ρ2,π1,π2\rho_1, \rho_2, \pi_1, \pi_2 be as in (ii). If x∈nx\in \mathfrak{n}, we have, since ρ1(x)⊗1\rho_1(x)\otimes 1 and 1⊗ρ2(x)1\otimes \rho_2(x) commute,

exp(ρ1(x)⊗1+1⊗ρ2(x)\rho_1(x)\otimes 1 + 1\otimes \rho_2(x)) == exp(ρ1(x)⊗1\rho_1(x)\otimes 1). exp(1 ⊗ρ2(x)\otimes \rho_2(x))

= (exp ρ1(x)\rho_1(x))⊗1.1⊗\otimes 1.1\otimes (expρ2(x)\rho_2(x))

= (exp ρ1(x)\rho_1(x))⊗\otimes (expρ2(x)\rho_2(x))

=π1=\pi_1(exp xx)⊗π2\otimes \pi_2(exp xx)

=(π1⊗π2= (\pi_1\otimes \pi_2)(exp xx),

hence (ii). Let ρ1,ρ2,ρ,π1,π2,π,V1,V2\rho_1, \rho_2, \rho , \pi_1, \pi_2, \pi ,V_1,V_2 be as in (iii). If v1∈v_1\in EndV1_1 and v2∈v_2\in EndV2_2, denote by Rv1R_{v_1} and Lv2L_{v_2} the maps u→uv1u \rightarrow uv_1 and u→v2uu \rightarrow v_2u from Hom(V1,V2_1,V_2) to itself; these maps commute and ρ(x)u=(Lρ2(x)−Rρ1(x))u\rho (x)u= (L_{\rho_2(x)}-R_{\rho_1(x)})u, so

expρ(x).u=\rho (x).u= exp Lρ2(x)L_{\rho_2(x)}. exp R−ρ1(x).uR_{-\rho_1(x)}.u

=Lexpρ2(x).Rexp(−ρ1(x)).u = L_{exp\rho_2(x)}.R_{exp(-\rho_1(x))}.u
=Lπ2(expx).Rπ1(exp(−x)).u = L_{\pi_2(expx)}.R_{\pi_1(exp(-x))}.u

=π=\pi(exp xx).u.u,

hence (iii).

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.