Bourbaki

Proposition 10

Tag 095B, permanent. Cite it as [Bourbaki, Tag 095B].

Integration, V, p. 71, § 6. Images of a measure

With the above notations, suppose that π\pi is ν\nu-measurable. Let gg be the function t↦λt∗(1)t \mapsto \lambda_t^*(1) on TT. For π\pi to be ν\nu-proper, it is necessary and sufficient that gg be locally μ\mu-integrable, in which case

π(ν)=g⋅μ. \pi(\nu) = g \cdot \mu.

We begin by arguing on the assumption that gg is finite locally μ\mu-almost everywhere; we will rid ourselves of this auxiliary hypothesis at the end of the proof. Since π\pi is by hypothesis ν\nu-measurable, to say that π\pi is ν\nu-proper is equivalent to saying that ν∗(f∘π)<+∞\nu^*(f \circ \pi) < +\infty for every function f∈K+(T)f \in \mathscr{K}_+(T); gg being finite locally almost everywhere, we are under the conditions for applying assertion c) of Prop. 5 of §3, No. 2. Therefore

∫∙(f∘π) dν=∫∙dμ(t)∫∙(f∘π) dλt=∫∙f(t)g(t) dμ(t), \int^\bullet (f \circ \pi)\, d\nu = \int^\bullet d\mu(t) \int^\bullet (f \circ \pi)\, d\lambda_t = \int^\bullet f(t)g(t)\, d\mu(t),

from the fact that λt\lambda_t is concentrated on π‾−1(t)\overline{\pi}^{-1}(t). We know that gg is μ\mu-measurable, since Λ\Lambda is μ\mu-adequate (§ 3, No. 1, Def. 1). To say that the first member is finite for every f∈K+(T)f \in \mathscr{K}_+(T) is therefore equivalent to saying that gg is locally μ\mu-integrable (§ 5, Prop. 1), and in this case (14) follows at once from the above relations.

It therefore only remains to eliminate the auxiliary hypothesis. If gg is locally μ\mu-integrable, then gg is finite locally μ\mu-almost everywhere, and the hypothesis is indeed satisfied. Let us assume that π\pi is ν\nu-proper, and let us show that gg is finite locally almost everywhere. Let K\mathfrak{K} be the μ\mu-dense set of compact sets KK such that Λ∣K\Lambda|K is vaguely continuous; since gg is measurable, we are reduced to showing that every compact set K∈KK \in \mathfrak{K} such that g∣K=+∞g|K = +\infty is μ\mu-negligible. Now, let H\mathscr{H} be the set of functions h∈K+(X)h \in \mathscr{K}_+(X) such that h≤1h \leq 1; set gh(t)=λt(h)g_h(t) = \lambda_t(h), denote by Λh\Lambda_h the μ\mu-adequate mapping t↦h⋅λtt \mapsto h \cdot \lambda_t, by νh\nu_h the integral of Λh\Lambda_h, and by ff an element of K+(T)\mathscr{K}_+(T) such that f≥φKf \geq \varphi_K. Applying formula (14) to Λh\Lambda_h, which does satisfy the auxiliary hypothesis, we obtain:

∫(f∘π) dν≥∫(f∘π) dνh=∫fgh dμ. \int (f \circ \pi)\, d\nu \geq \int (f \circ \pi)\, d\nu_h = \int fg_h\, d\mu.

But the functions fgh∣Kfg_h|K form an increasing directed set of continuous functions on KK, whose upper envelope has the value +∞+\infty; by Dini’s theorem (GT, X, §4, No. 1, Th. 1), one can choose hh so that fgh∣Kfg_h|K is greater than or equal to an arbitrary positive number nn, and it follows that ∫(f∘π) dν≥n μ(K)\int (f \circ \pi)\, d\nu \geq n\, \mu(K). Since the first member is finite because π\pi is ν\nu-proper, it then follows that μ(K)=0\mu(K) = 0.

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Cites

Transcribed from the printed volume. Not translated.