21.1. STRONG AND WEAK MEASURABILITY 647

⇐Next suppose x is strongly measurable. Then there exists a sequence of simple func-tions xn which converges to x pointwise. Hence for all f ∈ X ′, f ◦ xn is measurable andf ◦ xn→ f ◦ x pointwise. Thus x is weakly measurable.

The same method of proof yields the following interesting corollary.

Corollary 21.1.8 Let X be a separable Banach space and let B (X) denote the σ algebraof Borel sets. Let H be a dense subset of X ′. Then B (X) = σ (H) ≡F , the smallest σ

algebra of subsets of X which has the property that every function, x∗ ∈ H is measurable.

Proof: First I need to show F contains open balls because then F will contain theopen sets and hence the Borel sets. As noted above, it suffices to show F contains closedballs. Let D′ be those functionals in B′ defined in Lemma 21.1.6 contained in H. Then

{x : ∥x−a∥ ≤ r} =

{x : sup

x∗∈D′|x∗ (x−a)| ≤ r

}= ∩x∗∈D′ {x : |x∗ (x−a)| ≤ r}= ∩x∗∈D′ {x : |x∗ (x)− x∗ (a)| ≤ r}

= ∩x∗∈D′x∗−1(

B(x∗ (a) ,r))∈ σ (H)

which is measurable because this is a countable intersection of measurable sets. Thus Fcontains open sets so σ (H)≡F ⊇B (X) .

To show the other direction for the inclusion, note that each x∗ is B (X) measurablebecause x∗−1 (open set) = open set. Therefore, B (X)⊇ σ (H) .

It is important to verify the limit of strongly measurable functions is itself stronglymeasurable. This happens under very general conditions.

Lemma 21.1.9 Let X be a metric space and suppose V is an open set in V . Then thereexists open sets Vm such that

· · ·Vm ⊆V m ⊆Vm+1 ⊆ ·· · , V =∞⋃

m=1

Vm. (21.1.1)

Proof: Recall that if S is a nonempty set, x→ dist(x,S) is a continuous map from X toR. First assume V ̸= X . Let

Vm ≡{

x ∈V : dist(x,VC)> 1

m

}Then for large enough m, this set is nonempty and contained in V. Furthermore, if x ∈ Vthen it is at a positive distance to the closed set VC so eventually, x ∈Vm. Now

Vm ⊆Vm ⊆{

x ∈V : dist(x,VC)≥ 1

m

}⊆V

Indeed, if p is a limit point of Vm, then there are xn ∈Vm with xn→ p. Thus dist(xn,VC

)→

dist(

p,VC)

and so p is in the set on the right. In case X = V, let Vm ≡ B(ξ ,m) . ThenVm ⊆Vm ⊆

{x ∈V : dist

(x,VC

)≥ 1

m

}and the union of these Vm equals V .

What of limits of measurable functions? The next theorem says that the usual theoremabout limits of measurable functions being measurable holds.

21.1. STRONG AND WEAK MEASURABILITY 647<Next suppose x is strongly measurable. Then there exists a sequence of simple func-tions x, which converges to x pointwise. Hence for all f € X’, fox, is measurable andf OX — fox pointwise. Thus x is weakly measurable. JJThe same method of proof yields the following interesting corollary.Corollary 21.1.8 Let X be a separable Banach space and let &(X) denote the o algebraof Borel sets. Let H be a dense subset of X'. Then B(X) = 0(H) = F, the smallest oalgebra of subsets of X which has the property that every function, x* € H is measurable.Proof: First I need to show ¥ contains open balls because then ¥ will contain theopen sets and hence the Borel sets. As noted above, it suffices to show ¥ contains closedballs. Let D’ be those functionals in B’ defined in Lemma 21.1.6 contained in H. Thenfr: |-all <7 = {¥ sup eal srtx*ED!yep {x: |x* (x—a)| <r}= NAep {x: |x" (x) —x* (a)| <r}= Neepx! (BO) €o(H)which is measurable because this is a countable intersection of measurable sets. Thuscontains open sets so 0 (H) = ¥ D BX).To show the other direction for the inclusion, note that each x* is A(X) measurablebecause x*~! (open set) = open set. Therefore, @(X)Do(H). WlIt is important to verify the limit of strongly measurable functions is itself stronglymeasurable. This happens under very general conditions.Lemma 21.1.9 Let X be a metric space and suppose V is an open set in V. Then thereexists open sets V,, such thatVin Vim © Ving Co V= (Vn (21.1.1)Proof: Recall that if S is a nonempty set, x > dist (x,S) is a continuous map from X toR. First assume V 4 X. Let. 1Vin = EV: dist (x,V°) > }mThen for large enough m, this set is nonempty and contained in V. Furthermore, if x € Vthen it is at a positive distance to the closed set V© so eventually, x € Vin. NowVin CVn © {¥eV dist (V9) > ~| CVIndeed, if p is a limit point of V,,, then there are x, € Vi, with x, — p. Thus dist (x,,V°) >dist (p,V©) and so p is in the set on the right. In case X = V, let Vi, = B(€,m). ThenVin CVn C {x € V : dist (x,V°) > +} and the union of these V,, equals V.What of limits of measurable functions? The next theorem says that the usual theoremabout limits of measurable functions being measurable holds.