Bourbaki

Theorem 1

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

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

Suppose that I is directed, that the sets Sα\mathfrak{S}_\alpha satisfy conditions (i) and (ii), and that the inverse system (Eα,fαβ)(\mathrm{E}_\alpha, f_{\alpha\beta}) has the following properties:

(iii) For each pair of indices α\alpha, β\beta such that α⩽β\alpha \leqslant \beta, and each xα∈Eαx_\alpha \in \mathrm{E}_\alpha, we have f−1αβ(xα)∈Sβ\overset{-1}{f}_{\alpha\beta}(x_\alpha) \in \mathfrak{S}_\beta.

(iv) For each pair of indices α\alpha, β\beta such that α⩽β\alpha \leqslant \beta, and each Mβ∈Sβ\mathrm{M}_\beta \in \mathfrak{S}_\beta, we have fαβ(Mβ)∈Sαf_{\alpha\beta}(\mathrm{M}_\beta) \in \mathfrak{S}_\alpha.

Let E=lim←⁡Eα\mathrm{E} = \varprojlim \mathrm{E}_\alpha and let fα:E→Eαf_\alpha : \mathrm{E} \to \mathrm{E}_\alpha be the canonical mapping for each α∈I\alpha \in \mathrm{I}. Then:

(a) For each α∈I\alpha \in \mathrm{I} we have

(19)fα(E)=⋂β⩾αfαβ(Eβ).(19) \qquad f_\alpha(\mathrm{E}) = \bigcap_{\beta \geqslant \alpha} f_{\alpha\beta}(\mathrm{E}_\beta).

(b) If Eα\mathrm{E}_\alpha is non-empty for each α∈I\alpha \in \mathrm{I}, then E is non-empty.

Let Σ\Sigma be the set of families A=(Aα)α∈I\mathfrak{A} = (\mathrm{A}_\alpha)_{\alpha \in \mathrm{I}} which satisfy the following conditions :

(20)Aα≠∅andAα∈Sαfor allα∈I;(20) \qquad \mathrm{A}_\alpha \neq \emptyset \quad \textit{and} \quad \mathrm{A}_\alpha \in \mathfrak{S}_\alpha \quad \textit{for all} \quad \alpha \in \mathrm{I};
(21)fαβ(Aβ)⊂Aαwheneverα⩽β.(21) \qquad f_{\alpha\beta}(\mathrm{A}_\beta) \subset \mathrm{A}_\alpha \quad \textit{whenever} \quad \alpha \leqslant \beta.

If A=(Aα)\mathfrak{A} = (A_\alpha) and A′=(Aα′)\mathfrak{A}' = (A'_\alpha) are any two elements of Σ\Sigma, let the relation A⩽A′\mathfrak{A} \leqslant \mathfrak{A}' mean that Aα⊃Aα′A_\alpha \supset A'_\alpha for all α\alpha. Clearly Σ\Sigma is ordered by this relation.

(1) Let us first show that the ordered set Σ\Sigma is inductive. Let L be a totally ordered set and let λ→Aλ=(Aαλ)α∈I\lambda \to \mathfrak{A}^\lambda = (A^\lambda_\alpha)_{\alpha \in I} be a strictly increasing mapping of L into Σ\Sigma. For each α∈I\alpha \in I, put Bα=⋂λ∈LAαλB_\alpha = \bigcap_{\lambda \in L} A^\lambda_\alpha. Then it is immediately seen that the family B=(Bα)α∈I\mathfrak{B} = (B_\alpha)_{\alpha \in I} satisfies (21); by reason of (i) and (ii), it also satisfies (20), hence belongs to Σ\Sigma; and it is clear that B\mathfrak{B} is an upper bound of the set of the Aλ\mathfrak{A}^\lambda.

(2) Let A=(Aα)\mathfrak{A} = (A_\alpha) be a maximal element of Σ\Sigma. We shall show that Aα=fαβ(Aβ)A_\alpha = f_{\alpha\beta}(A_\beta) whenever α⩽β\alpha \leqslant \beta. Let Aα′=⋂β⩾αfαβ(Aβ)A'_\alpha = \bigcap_{\beta \geqslant \alpha} f_{\alpha\beta}(A_\beta) for all α∈I\alpha \in I, and let us show that A′=(Aα′)\mathfrak{A}' = (A'_\alpha) belongs to Σ\Sigma. Note first that if α⩽β⩽γ\alpha \leqslant \beta \leqslant \gamma, we have fαγ(Aγ)=fαβ(fβγ(Aγ))⊂fαβ(Aβ)f_{\alpha\gamma}(A_\gamma) = f_{\alpha\beta}(f_{\beta\gamma}(A_\gamma)) \subset f_{\alpha\beta}(A_\beta) by (21); moreover, fαβ(Aβ)∈Sαf_{\alpha\beta}(A_\beta) \in \mathfrak{S}_\alpha by (iv), and fαβ(Aβ)≠∅f_{\alpha\beta}(A_\beta) \neq \emptyset by (20). Hence conditions (i) and (ii) show that A′\mathfrak{A}' satisfies (20). Also, if α⩽β\alpha \leqslant \beta, we have

fαβ(Aβ′)⊂⋂γ⩾βfαβ(fβγ(Aγ))=⋂γ⩾βfαγ(Aγ);f_{\alpha\beta}(A'_\beta) \subset \bigcap_{\gamma \geqslant \beta} f_{\alpha\beta}(f_{\beta\gamma}(A_\gamma)) = \bigcap_{\gamma \geqslant \beta} f_{\alpha\gamma}(A_\gamma);

and for each δ⩾α\delta \geqslant \alpha there exists γ∈I\gamma \in I such that γ⩾δ\gamma \geqslant \delta and γ⩾β\gamma \geqslant \beta, so that fαγ(Aγ)⊂fαδ(Aδ)f_{\alpha\gamma}(A_\gamma) \subset f_{\alpha\delta}(A_\delta) and consequently

⋂γ⩾βfαγ(Aγ)=⋂δ⩾αfαδ(Aδ)=Aα′.\bigcap_{\gamma \geqslant \beta} f_{\alpha\gamma}(A_\gamma) = \bigcap_{\delta \geqslant \alpha} f_{\alpha\delta}(A_\delta) = A'_\alpha.

Hence A′\mathfrak{A}' satisfies (21) and therefore belongs to Σ\Sigma. Since Aα′⊂AαA'_\alpha \subset A_\alpha for all α\alpha, the maximality of A\mathfrak{A} in Σ\Sigma implies A′=A\mathfrak{A}' = \mathfrak{A}, and our assertion is proved.

(3) We shall establish next that if A=(Aα)\mathfrak{A} = (A_\alpha) is a maximal element of Σ\Sigma, then each of the AαA_\alpha consists of a single element. Let xα∈Aαx_\alpha \in A_\alpha. For each β⩾α\beta \geqslant \alpha, put Bβ=Aβ∩f−1αβ(xα)B_\beta = A_\beta \cap \overset{-1}{f}_{\alpha\beta}(x_\alpha); if β\beta is not ⩾α\geqslant \alpha, put Bβ=AβB_\beta = A_\beta; then we shall see that B=(Bβ)\mathfrak{B} = (B_\beta) belongs to Σ\Sigma. If β\beta is not ⩾α\geqslant \alpha, the relation β⩽γ\beta \leqslant \gamma implies fβγ(Bγ)⊂fβγ(Aγ)⊂Aβ=Bβf_{\beta\gamma}(B_\gamma) \subset f_{\beta\gamma}(A_\gamma) \subset A_\beta = B_\beta. If, on the other hand, α⩽β⩽γ\alpha \leqslant \beta \leqslant \gamma, then since

f−1αγ(xα)=f−1βγ(f−1αβ(xα)),\overset{-1}{f}_{\alpha\gamma}(x_\alpha) = \overset{-1}{f}_{\beta\gamma}(\overset{-1}{f}_{\alpha\beta}(x_\alpha)),

we have

fβγ(f−1αγ(xα))⊂f−1αβ(xα),f_{\beta\gamma}(\overset{-1}{f}_{\alpha\gamma}(x_\alpha)) \subset \overset{-1}{f}_{\alpha\beta}(x_\alpha),

and since fβγ(Aγ)⊂Aβf_{\beta\gamma}(A_\gamma) \subset A_\beta, we again have fβγ(Bγ)⊂Bβf_{\beta\gamma}(B_\gamma) \subset B_\beta, so that the family B\mathfrak{B} satisfies (21). Since Aα=fαβ(Aβ)A_\alpha = f_{\alpha\beta}(A_\beta) whenever α⩽β\alpha \leqslant \beta by part (2) of the proof, it is clear that Bβ≠∅\mathrm{B}_\beta \neq \emptyset for all β∈I\beta \in \mathrm{I}. Finally, by virtue of (i) and (iii), we have Bβ∈Sβ\mathrm{B}_\beta \in \mathfrak{S}_\beta for all β∈I\beta \in \mathrm{I}, and hence B∈Σ\mathfrak{B} \in \Sigma. Since Bβ⊂Aβ\mathrm{B}_\beta \subset \mathrm{A}_\beta for all β∈I\beta \in \mathrm{I}, the maximality of A\mathfrak{A} implies that Bβ=Aβ\mathrm{B}_\beta = \mathrm{A}_\beta for all β\beta, and in particular Aα={xα}\mathrm{A}_\alpha = \{x_\alpha\}.

(4) We are now in a position to prove Theorem 1. Let us start with (a). We have

fα(E)⊂⋂β⩾αfαβ(Eβ).f_\alpha(\mathrm{E}) \subset \bigcap_{\beta \geqslant \alpha} f_{\alpha\beta}(\mathrm{E}_\beta).

Conversely, let

xα∈⋂β⩾αfαβ(Eβ),x_\alpha \in \bigcap_{\beta \geqslant \alpha} f_{\alpha\beta}(\mathrm{E}_\beta),

and put

Bβ=f−1αβ(xα)\mathrm{B}_\beta = \overset{-1}{f}_{\alpha\beta}(x_\alpha)

if β⩾α\beta \geqslant \alpha, and Bβ=Eβ\mathrm{B}_\beta = \mathrm{E}_\beta otherwise. By the definition of xαx_\alpha, the Bβ\mathrm{B}_\beta are non-empty, and we have Bβ∈Sβ\mathrm{B}_\beta \in \mathfrak{S}_\beta for all β∈I\beta \in \mathrm{I} by virtue of (iii) and (i); moreover, it is evident that fβγ(Bγ)⊂Bβf_{\beta\gamma}(\mathrm{B}_\gamma) \subset \mathrm{B}_\beta whenever β⩽γ\beta \leqslant \gamma. Hence B=(Bβ)∈Σ\mathfrak{B} = (\mathrm{B}_\beta) \in \Sigma. Let A=(Aα)\mathfrak{A} = (\mathrm{A}_\alpha) be a maximal element of Σ\Sigma such that A⩾B\mathfrak{A} \geqslant \mathfrak{B} (the existence of A\mathfrak{A} follows from (1) and §2, no. 4, Theorem 2, Corollary 1). Since, by (3), Aβ\mathrm{A}_\beta is of the form {yβ}\{y_\beta\} for all β∈I\beta \in \mathrm{I}, it follows that y=(yβ)y = (y_\beta) belongs to E, and fα(y)=yα=xαf_\alpha(y) = y_\alpha = x_\alpha by definition.

Finally, (a) implies (b). We may as well assume that I is not empty (otherwise there is nothing to prove). The hypothesis that the Eα\mathrm{E}_\alpha are non-empty implies that fαβ(Eβ)≠∅f_{\alpha\beta}(\mathrm{E}_\beta) \neq \emptyset for all β⩾α\beta \geqslant \alpha. Since the fαβ(Eβ)f_{\alpha\beta}(\mathrm{E}_\beta), for fixed α\alpha and variable β⩾α\beta \geqslant \alpha, form a left directed set of subsets of Eα\mathrm{E}_\alpha belonging to Sα\mathfrak{S}_\alpha, condition (ii)′' proves that

⋂β⩾αfαβ(Eβ)≠∅.\bigcap_{\beta \geqslant \alpha} f_{\alpha\beta}(\mathrm{E}_\beta) \neq \emptyset.

Hence fα(E)≠∅f_\alpha(\mathrm{E}) \neq \emptyset by (a), and a fortiori E≠∅\mathrm{E} \neq \emptyset. Q.E.D.

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

Cites

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