Let E be a graded bigebra (resp. skew graded bigebra). For every element f of degree 1 in E∗gr, the left and right inner products are derivations (resp. antiderivations) of the algebra E.
Let x∈Ep,y∈Eq (p≥1,q≥1). By Proposition 5 of no. 3, we can write
c(x)=x⊗1+1≤j≤p−1∑(i∑xij′⊗xij,p−j′′)+1⊗x c(y)=y⊗1+1≤k≤q−1∑(i∑yi,k′⊗yi,q−k′′)+1⊗y
where xij′ and xij′′ belong to Ej, yik′ and yik′′ to Ek. If E is a graded bigebra, the component of c(xy)=c(x)c(y) belonging to E1⊗E, is equal to
i∑xi,1′⊗xi,p−1′′y+i∑yi,1′⊗xyi,q−1′′ and hence by definition
(xy)←f=i∑f(xi,1′)xi,p−1′′y+i∑f(yi,1′)xyi,q−1′′ =(x←f)y+x(y←f) and the right inner product by
f is a
derivation. If on the other hand
E is a skew graded bigebra, the component of
c(xy) belonging to
E1⊗E is equal to
i∑xi,1′⊗xi,p−1′′y+(−1)pi∑yi,1′⊗xyi,q−1′′ and this time we obtain
(xy)←f=(x←f)y+(−1)px(y←f) which shows that
i(f) is then an
antiderivation. The argument is similar for the left inner product by
f.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.