Let $ (X,\Sigma,\mu) $ be a $ \sigma $-finite measure space and $ B $ a Banach space. A function $ f: X \to B $ is said to be strongly $ \mu $-measurable iff it is the almost-everywhere pointwise limit of a sequence $ (s_{n}: X \to B)_{n \in \mathbb{N}} $ of integrable simple functions, where an integrable simple function $ s: X \to B $ has the form $$ s = \sum_{(E,b) \in I} \chi_{E} \cdot b $$ for some finite subset $ I $ of $ \{ E \in \Sigma \mid \mu(E) < \infty \} \times B $.

Let $ G $ be a second-countable, locally compact Hausdorff group and $ \mu_{G} $ a fixed Haar measure on the Borel $ \sigma $-algebra $ \mathscr{B}(G) $ of $ G $. The second-countability condition implies that $ G $ is $ \sigma $-compact, which ensures that $ (G,\mathscr{B}(G),\mu_{G}) $ is a $ \sigma $-finite measure space. Let $ B $ be a separable Banach space and $ {L^{\infty}}(G,B) $ the set of all (equivalence classes of) $ B $-norm essentially bounded strongly $ \mu_{G} $-measurable functions from $ G $ to $ B $.

Note: $ {L^{\infty}}(G,B) $ is a Banach space, and except in trivial cases, it is always non-separable.

Question. If $ F: G \times G \to B $ is a strongly $ \mu_{G \times G} $-measurable function where $ F(x,\bullet) \in {L^{\infty}}(G,B) $ for all $ x \in G $, then is it true that the mapping \begin{align} G & \to {L^{\infty}}(G,B); \\ x & \mapsto F(x,\bullet) \end{align} is strongly $ \mu $-measurable?

Thank you all very much for your help!


1 Answer 1


I think the answer is no, even in the case of $G:=\mathbb{R}$ with the Lebesgue measure and $B: =\mathbb{R}$ as a Banach space.

Let $F: \mathbb{R}\times \mathbb{R} \to \mathbb{R} $ be the characteristic function of the half-plane above the diagonal: $F(x,y):=\chi_\mathbb{{R}_+}(y-x)$. So $F\in L^\infty(\mathbb{R}\times \mathbb{R} )$; however the corresponding map $\tilde F: \mathbb{R} \to L^\infty(\mathbb{R} )$ such that $x\mapsto F(x,\cdot)=\chi_{[x,\infty)}$ is not strongly measurable, because its range is not separable, as it is the discrete uncountable set $\{\chi_{[x,\infty)} : x\in \mathbb{R}\}$ (not even if we modify $\tilde F$ on some Lebesgue null set).

Rmk 1. A Banach space valued strongly measurable map $f$ on a measure space $X$ is necessarily essentially separable , that is, up to removing a null set $N\subset X$, its image $f(X\setminus N)$ is a separable subset of $B$ (indeed it is included in the closure of the countable union of the images of the simple functions that converge pointwise to $f$; and a simple function is exactly a measurable function with finite range).

Rmk 2. A theorem of Pettis states that a Banach space valued map $f$ on a measure space $X$ is strongly measurable if and only if: (i) it is essentially separable; (ii) it is weakly measurable (that is, measurable w.r.to the $\sigma$ algebra generated by the weak topology; (iii) its support is $\sigma$-finite. (See e.g. the Kōsaku Yosida's Functional Analysis. The proof is easy. ).

  • 2
    $\begingroup$ I think Yosida quotes Pettis' theorem with a reference for the proof. Anyway note that "strongly measurable" easily implies (i), (ii) and (iii); for the other implication, one can assume $X$ sigma-finite and B separable. Note also that in a separable Banach space any closed ball is a countable intersection of open half-spaces, therefore it is weakly measurable, so that the weak and the strong Borel algebras coincide. $\endgroup$ Oct 17, 2014 at 9:01
  • $\begingroup$ Thank you very much, Pietro. I need some time to look over your solution. One quick question: If I replace $ L^{\infty} $ by $ L^{2} $, do I get an affirmative answer? $\endgroup$ Oct 17, 2014 at 9:51
  • 2
    $\begingroup$ I think so. For measure spaces $(X,\mathcal{A},\mu)$ and $(Y,\mathcal{B},\nu)$, for a Banach space $B$, and for $1\le p <\infty$ the correspondence gives a normed space isomorphism between the Bochner spaces $$L^p(X\times Y,B)\to L^p(X,L^p(Y,B)).$$ This is essentially Fubini-Tonelli; the key point is that a dense subset of simple functions in $L^p(X\times Y,B)$ is obtained by using characteristic functions of rectangles of finite measure in $X\times Y$; and these are mapped on a dense set of simple functions in $L^p(X,L^p(Y,B))$. I think no further assumptions are needed on $X$, $Y$, $B$. $\endgroup$ Oct 17, 2014 at 11:18
  • $\begingroup$ Hi Pietro. Actually, I’m not sure about your last comment, as I’m only assuming that $ F: X \times Y \to B $ is strongly measurable and not that it belongs to $ {L^{p}}(X \times Y,B) $ for any $ p \geq 1 $. $\endgroup$ Oct 17, 2014 at 16:47
  • 1
    $\begingroup$ For $1\le p<\infty$, your hypotheses should ensure that $L^p(G,B)$ be separable ; so you only need to verify condition (ii) of Pettis'theorem... $\endgroup$ Oct 17, 2014 at 17:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.