There exists a unique involutive automorphism X→X∗ of the ring R(a) such that [E]∗=[E∗] for every finite dimensional a-module E.
The uniqueness is clear. By Lemma 3, there exists a homomorphism X→X∗ from the additive group R(a) to itself such that [E]∗=[E∗] for every finite dimensional a-module E. We have (X∗)∗=X, so this homomorphism is involutive. It is an automorphism of the ring R(a) since (E⊗F)∗ is isomorphic to E∗⊗F∗ for all finite dimensional a-modules E and F. Q.E.D.
Let U(a) be the enveloping algebra of a,U(a)∗ the vector space dual of U(a). Recall (Chap. II, §1, no. 5) that the coalgebra structure of U(a) defines on U(a)∗ a commutative, associative algebra structure with unit element. For any finite dimensional a-module E, the map u→ Tr(uE) from U(a) to k is an element τE of U(a)∗. If 0⟶E′⟶E⟶E′′⟶0 is an exact sequence of finite dimensional a-modules, then τE=τE′+τE′′. Hence, by Lemma 3 there exists a unique homomorphism, which we denote by Tr, from the additive group R(a) to the group U(a)∗ such that Tr[E] =τE for every finite dimensional a-module E. If k denotes the trivial a-module of dimension 1, it is easy to check that Tr[k] is the unit element of U(a)∗. Finally, let E and F be finite dimensional a-modules. Let u∈U(a) and let c be the coproduct of U(a). By definition of the U-module E⊗F (Chap. I, §3, no. 2), if c(u)=∑iui⊗ui′,
This means that τEτF=τE⊗F. Thus, Tr:R(a)→U(a)∗ is a homomorphism of rings.
Let a1 and a2 be Lie algebras, f a homomorphism from a1 to a2. Every finite dimensional a2-module E defines by means of f an a1-module, hence elements of R(a2) and R(a1) that we denote provisionally by [E]2 and [E]1. By Lemma 3, there exists a unique homomorphism, denoted by R(f), from the group R(a2) to the group R(a1) such that R(f)[E]2=[E]1 for every finite dimensional a2-module E. Moreover, R(f) is a homomorphism of rings. If U(f) is the homomorphism from U(a1) to U(a2) extending f, the following diagram is commutative
In what follows we take for a the splittable semi-simple Lie algebra g. The ring R(g) is called the representation ring of g. For all λ∈P++, we denote by [λ] the class of the simple g-module E(λ) of highest weight λ.
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.