a) The mapping ϱ:(g,h)↦ϱ(g,h) is a unitary representation of G×H in E ;
b) For every orthonormal basis (ei)i∈I of E1, the mapping from the Hilbert sum ⨁i∈IE2 into E defined by (yi)i∈I↦∑i∈Iei⊗yi is an isometric H-isomorphism of the Hilbert sum ⨁i∈Iϱ2 onto Res{e}×HG×H(ϱ);
c) For every orthonormal basis (fj)j∈J of E2, the mapping from the Hilbert sum ⨁j∈JE1 into E defined by (xj)j∈J↦∑j∈Jxj⊗fj is an isometric G-isomorphism of the Hilbert sum ⨁j∈Jϱ1 onto ResG×{e}G×H(ϱ).
The mapping (g,h)↦ϱ(g,h) is a homomorphism of G into GL(E)(cf. EVT, V, p. 28, no2). Let (ei)i∈I be an orthonormal basis of E1. By EVT, V, p. 29, prop. 3 and cor. 2, the mapping
u:(yi)↦i∈I∑ei⊗yi
is an isometry of the Hilbert direct sum F2=⨁i∈IE2 onto E. By means of this isometry, the representation ϱH:h↦ϱ(e,h) of H in E is identified with the direct-sum representation of the representations (ϱ2)i∈I in F2. In particular, it is a unitary representation of H in E (lemma 7). Analogously, the homomorphism ϱG:g↦ϱ(g,e) is a unitary representation of G in E.
Let (g,h)∈G×H. We have ϱ(g,h)=ϱG(g)∘ϱH(h); therefore ϱ(g,h) is unitary; moreover, ϱG(g) and ϱH(h) commute, hence lemma 1 of V, p. 377 implies that ϱ is a unitary representation of G×H.
Finally, the assertions concerning the restriction of ϱ to the subgroups {e}×H and G×{e} were obtained in the course of the preceding argument.
The representation (g,h)↦ϱ1(g)⊗ϱ2(h) of G×H is called the external tensor product of the unitary representations ϱ1 and ϱ2, and is denoted by ϱ1⊠ϱ2.
Let n∈N and let (Gi)1⩽i⩽n be a finite family of topological groups. Let ϱi be a unitary representation of Gi in a Hilbert space Ei for 1⩽i⩽n. One defines analogously a representation
ϱ1⊠⋯⊠ϱn
of G1×⋯×Gn in the Hilbert space E=E1⊗2⋯⊗2En (EVT, V, p. 27).
Suppose that Gi=G for 1⩽i⩽n. Let Δn:G→Gn be the homomorphism defined by g↦(g,...,g) for every g∈G. One denotes by ϱ1⊗⋯⊗ϱn the unitary representation (ϱ1⊠⋯⊠ϱn)∘Δn of G. One says that it is the tensor product of the unitary representations ϱi.
For every permutation σ of {1,...,n}, the canonical isomorphism
E1⊗2⋯⊗2En→Eσ(1)⊗2⋯⊗2Eσ(n)
(EVT, V, p. 28) is an isometric isomorphism
ϱ1⊗⋯⊗ϱn→ϱσ(1)⊗⋯⊗ϱσ(n)
of representations of G.
If ϱi=ϱ for 1⩽i⩽n, where ϱ is a unitary representation of G, one also denotes by ϱ⊗n the tensor-product representation of the ϱi, and one says that it is the ne tensor power of ϱ.
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.