Let d1,d2 be two ε-derivations of (A,A′,A′′) of respective degrees δ1,δ2. Then the ε-bracket
[d1,d2]ε=d1∘d2−εδ1,δ2d2∘d1 is an ε-derivation of degree δ1+δ2. Moreover, if d is an ε-derivation of (A,A′,A′′) of degree δ and if εδ,δ=−1, then d2=d∘d is a derivation.
Suppose x∈A is homogeneous of degree ξ; for all y∈A′,
d1(d2(xy))=((d1d2)(x))y+εδ1,δ2+ξ(d2x)(d1y)+εδ2,ξ(d1x)(d2y)+εδ1+δ2,ξx((d1d2)(y)) taking account of formulae (1) of no. 1. Exchanging the roles of
d1 and
d2, we obtain, after simplifications again using (1) (no. 1),
(d1d2)(xy)−εδ1,δ2(d2d1)(xy)=((d1d2)(x))y−εδ1,δ2((d2d1)(x))y+εδ1+δ2,ξx((d1d2)(y))−εδ1,δ2εδ1+δ2,ξx((d2d1)(y)) that is, writing
d=[d1,d2]ε and
δ=δ1+δ2,
d(xy)=(dx)y+εδ,ξx(dy) which proves that
d is an
ε-derivation.
On the other hand, if, in (11), we let d1=d2=d, δ1=δ2=δ and εδ,δ=−1, we obtain, since then εδ,δ+ξ=−εδ,ξ by (1),
d2(xy)=(d2x)y+ε2δ,ξx(d2y) and as
ε2δ,ξ=1 it is seen that
d2 is a derivation.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.