Let (Xι)ι∈I be a family of sets whose index set I is the union of a family (Jλ)λ∈L of sets. Then
ι∈I⋃Xι=λ∈L⋃(ι∈Jλ⋃Xι),
and (if L=∅ and Jλ=∅ for each λ∈L)
ι∈I⋂Xι=λ∈L⋂(ι∈Jλ⋂Xι)
("associativity" of union and intersection).
Let x be an element of ⋃ι∈IXι. There exists an index ι∈I such that x∈Xι. Since I is the union of the family (Jλ)λ∈L, there exists an index λ∈L such that ι∈Jλ, whence x∈⋃ι∈JλXι, and consequently
x∈λ∈L⋃(ι∈Jλ⋃Xι).
Conversely, let x be an element of this set. There exists an index λ∈L such that x∈⋃ι∈JλXι, hence there exists an index ι∈Jλ (and therefore ι∈I) such that x∈Xι; it follows that
x∈ι∈I⋃Xι.
Now suppose that L=∅ and Jλ=∅ for each λ∈L. Then I=∅. Let x be an element of ⋂ι∈IXι. If λ∈L, we have x∈Xι for each ι∈Jλ (since Jλ⊂I), whence x∈⋂ι∈JλXι. Since the last inclusion is true for all λ∈L, x belongs to ⋂λ∈L(⋂ι∈JλXι). Conversely, let x be an element of this latter set, and let ι be any element of I. There exists λ∈L such that ι∈Jλ; since x∈⋂ι∈JλXι, we have x∈Xι, which is true for all ι∈I. Hence x∈⋂ι∈IXι. This completes the proof.
For families of subsets of a given set the second part of Proposition 2 remains valid without restriction on L and Jλ.
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.
Nothing.