The ideal a of Definition 1 is graded. For a,b in Lib(X), let B(a,b)=a.b+b.a. The formulae
(10)B(a,b)=Q(a+b)−Q(a)−Q(b) (11)Q(λ1.w1+⋯+λn.wn)=i∑λi2Q(wi)+i<j∑λiλjB(wi,wj) for
w1,…,wn in
M(X) and
λ1,…,λn in
K, show that the families
(Q(a))a∈Lib(X) and
(Q(w),B(w,w′))w,w′∈M(X) generate the same submodule of
Lib(X). As J is trilinear the ideal
a is generated by the homogeneous elements
Q(w),B(w,w′) and
J(w,w′,w′′), where
w,w′,w′′ are in
M(X), and hence is graded (Algebra, Chapter III, § 3, no. 3, Proposition 1).
Let the Lie algebra L(X)=Lib(X)/a be given the quotient graduation. The homogeneous component of L(X) of degree δ is denoted by Lδ(X); it is the submodule of L(X) generated by the images of the elements w∈M(X) such that ϕ(w)=δ.
We shall make special use of the following two cases: (a) Total graduation: we take Δ=N and ϕ0(x)=1 for all x∈X, whence ϕ(w)=l(w) for w in M(X). The K-module Ln(X) is generated by the images of the elements of length n in M(X), which we shall call alternants of degree n. We shall see later that the module Ln(X) is free and admits a basis consisting of alternants of degree n (no. 11, Theorem 1). Then L(X)=⨁n≥1Ln(X) and L1(X) admits X as basis (no. 2, Corollary 1 to Proposition 1). By the construction of M(X),
(12)Ln(X)=p=1∑n−1[Lp(X),Ln−p(X)] and in particular
(13)[Lm(X),Ln(X)]⊂Lm+n(X).
(b) Multigraduation: we take Δ to be the free commutative monoid N(X) constructed on X. The mapping ϕ0 of X into Δ is defined by (ϕ0(x))(x′)=δxx′, where δxx′ is the Kronecker symbol. For w∈M(X) and x∈X, the integer (ϕ(w))(x) is "the number of occurrences of the letter x in w". For α in N(X), we write ∣α∣=∑x∈Xα(x), whence ∣ϕ(w)∣=l(w) for all w in M(X). It follows that
(14)Ln(X)=∣α∣=n⨁Lα(X); obviously
(15)[Lα(X),Lβ(X)]⊂Lα+β(X)for α,β in N(X).
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.