Bourbaki

Remark

Tag 03OO, permanent. Cite it as [Bourbaki, Tag 03OO].

Theory of Sets, chapter III, § 7. Inverse limits and direct limits

Suppose that, for each α∈I\alpha \in \mathrm{I}, uα:Eα→E′u_\alpha : \mathrm{E}_\alpha \to \mathrm{E}' is a mapping such that the family (uα)(u_\alpha) satisfies (23). Consider the direct system (Eα,iβα)(\mathrm{E}_\alpha, i_{\beta\alpha}) relative to I, where Eα′=E′\mathrm{E}'_\alpha = \mathrm{E}' for all α∈I\alpha \in \mathrm{I}, and iβαi_{\beta\alpha} is the identity mapping of E′\mathrm{E}'. Then (no. 5, Example 2) E′\mathrm{E}' can be identified canonically with lim→⁡Eα′\varinjlim \mathrm{E}'_\alpha. If uαu_\alpha is considered as a mapping of Eα\mathrm{E}_\alpha into Eα′\mathrm{E}'_\alpha, then (uα)(u_\alpha) is a direct system of mappings, and the mapping u:E→E′u : \mathrm{E} \to \mathrm{E}' defined by (24) is identified with the direct limit of this system of mappings. Hence, by abuse of language, we write u=lim→⁡uαu = \varinjlim u_\alpha.

If J is a subset of I which is directed (with respect to the induced pre-ordering), it is clear that the pair consisting of the subfamily (Eα)α∈J(\mathrm{E}_\alpha)_{\alpha \in \mathrm{J}} and the family (fβα)(f_{\beta\alpha}), where α⩽β\alpha \leqslant \beta and α∈J\alpha \in \mathrm{J} and β∈J\beta \in \mathrm{J}, is a direct system of sets relative to J; it is said to be obtained by restricting the index set to J. Let E, E′ respectively denote the direct limits of the families (Eα)α∈I(\mathrm{E}_\alpha)_{\alpha \in \mathrm{I}} and (Eα)α∈J(\mathrm{E}_\alpha)_{\alpha \in \mathrm{J}}, and for each α∈I\alpha \in \mathrm{I} let fα:Eα→Ef_\alpha : \mathrm{E}_\alpha \to \mathrm{E} denote the canonical mapping. Then (fα)α∈J(f_\alpha)_{\alpha \in \mathrm{J}} is a direct system of mappings, and consequently g=lim→⁡fαg = \varinjlim f_\alpha is a mapping of E′ into E, called canonical. Moreover, if J′ is a directed subset of J, if E″ is the direct limit of the family (Eα)α∈J′(\mathrm{E}_\alpha)_{\alpha \in \mathrm{J}'}, and if

g′:E′′→E′andg′′:E′′→Eg' : \quad \mathrm{E}'' \to \mathrm{E}' \qquad \text{and} \qquad g'' : \mathrm{E}'' \to \mathrm{E}

are the canonical mappings, then it follows immediately from Proposition 6 that

(28)

g′′=g∘g′.g'' = g \circ g'.

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.

Cited by

Nothing.

Cites

Transcribed from the printed edition, 2004, Springer. Not translated.