Let B be a basis of R, R+ (resp. R−) the set of positive (resp. negative) roots of (g,h) relative to B,n+=∑α∈R+gα,n−=∑α∈R−gα, S(h) the symmetric algebra of h, and J the ideal of S(g) generated by n+∪n−.
(i) S(g)=S(h)⊕J.
(ii) Let j be the homomorphism from the algebra S(g) to the algebra S(h) defined by the preceding decomposition of S(g). Let I(g) be the set of invariant elements of S(g). Let S(h)W be the set of elements of S(h) invariant under the operation of W. Then j∣I(g) is an isomorphism from I(g) to S(h)W.
Assertion (i) is clear. The Killing form defines an isomorphism from the vector space g∗ to the vector space g, which extends to an isomorphism ξ from the g-module S(g∗) to the g-module S(g). We have ξ(I(g∗))=I(g). The orthogonal complement of h with respect to the Killing form is n++n−(§2, no. 2, Prop. 1). If we identify h∗ with the orthogonal complement of n++n− in g∗, then ξ(h∗)=h, so ξ(S(h∗))=S(h) and ξ(S(h∗)W)=S(h)W. Finally, ξ−1(J) is the set of polynomial functions on g that vanish on h. This proves that ξ transforms the homomorphism i of Th. 1 into the homomorphism j of Cor. 2. Thus assertion (ii) follows from Th. 1 (i).
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.