Let(Eα,fβα)be a direct system of magmas and letEbe its direct limitsfα:Eα→Ethe canonical homomorphisms. Suppose a magmaFand homomorphismsuα:Eα→Fare given such thatuα=uβ∘fβαforα≤β. There exists one and only one homomorphismu:E→Fsuch thatuα=u∘fαfor allα∈I. If the magmasEαandFand the homomorphismsfβαanduαare unital, the homomorphismuis unital.
We know (Set Theory, III, § 7, no. 6, Proposition 6) that there exists one and only one mapping u:E→F such that uα=u∘fα for all α∈I. We verify that u is a homomorphism: let x,y be in E, α in I and xα,yα in Eα such that x=fα(xα) and y=fα(yα). Then xy=fα(xαyα), whence
We consider now the unital case and let eα denote the identity element of Eα, e that of E and e′ that of F. Choose α∈I, then e=fα(eα), whence
u(e)=u(fα(eα))=uα(eα)=e′
for uα is unital. Hence u is unital.
By analogy with the notion of direct system of magmas, that of a direct system of monoids or groups can be formulated. Proposition 1 shows that the magma E which is the limit of a direct system of monoids (Eα,fβα)α,β∈I is a monoid. We show that E is a group if the Eα are groups: let x∈E,α∈I and xα∈Eα be such that x=fα(xα); the element y=fα(xα−1) of E is the inverse of x (§ 2, no. 3). The universal property of Proposition 2 goes over immediately in the case of a direct system of monoids or groups.
The reader is left to define a direct system of rings. Let (Aα,fβα) be such a direct system; let A=limAα and fα:Aα→A the canonical homomorphisms. There exists (Proposition 2) on A one and only one addition and multiplication characterized by x+y=fα(xα+yα),xy=fα(xαyα) for α∈I,xα,yα in Aα and x=fα(xα),y=fα(yα). Under addition A is a commutative group and multiplication is associative and unital. Finally, for x,y,z in A, choose α in I and xα,yα and zα in Aα with
and the relation x(y+z)=xy+xz is proved analogously. In other words, A has the structure of a ring, characterized by the fact that fα is a ring homomorphism for all α∈I.
The ring A is called the direct limit of the rings Aα. Proposition 2 extends immediately to the case of rings.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.