Let $\mathscr{F}$ be a $\sigma$-field on $\mathbb{R}^d$ such that every continuous function $f:\mathbb{R}^d\to\mathbb{R}$ that vanishes outside a bounded interval is $\mathscr{F}$-measurable. Prove that $\mathscr{B}(\mathbb{R}^d)\subseteq\mathscr{F}.$

I interpereted "vanishes outside a bounded interval" as the following:

Let $g$ be a continuous function such that $g:\mathbb{R}^d\to\mathbb{R}$ and $f=g1_{A}$ where $A\subseteq\mathbb{R}^d$ bounded. Then since $f$ is measurable, take any open bounded set $(a,b)\in\mathbb{R}$ and we have $f^{-1}(a,b)\in\mathscr{F}.$ Intuitively, since $\mathscr{B}(\mathbb{R}^d)$ is a $\sigma$-field every collection $\mathscr{A}\subseteq\mathscr{B}(\mathbb{R}^d)$ consists of subsets of $\mathbb{R}^d$.

Let $\mathscr{A}=\{A_1,A_2,A_3...\}$ where each $A_i\subseteq\mathbb{R}^d$. Is it true that $f(\mathscr{A})=\{(a_1,b_1),(a_2,b_2),...\}$ where each $(a_i,b_i)\subseteq\mathbb{R}$? If it's the case that $f$ maps each $A_i\in\mathbb{R}^d$ to a $(a_i,b_i)\in\mathbb{R}$, then isn't it clear that $\mathscr{A}\in\mathscr{F}$ from the argument in the above paragraph? What am I missing here?


Note that $\mathscr B(R^d)$ is generated by the set $S:=\{ \prod _{i=1} ^ d I_i : I_i\ are\ bounded\ open\ intervals\ of\ R\} $ i.e. $\mathscr B(R^d)$ is the smallest $\sigma-algebra$ contining $S$. Now for each element $J$ of $S$ you can construct a continuous map $f:R^d\rightarrow R$ which vanishes outside $J$ and $f(J)$ does not contains $0$.Notcice we can do when $d=1$ i.e. for $-\infty<a_i<b_i<\infty$ we have a continuous function $f_i$ such that $0 \notin f_i((a_i,b_i))$ and $f_i(R-(a_i,b_i))=\{0\}$ for $i=1,2,..,d$. So whenever $ J=\prod _{i=1} ^ d (a_i,b_i)$ your $f$ will be $\prod _{i=1} ^ d f_i$.

Then by hypothesis $f$ is measurable,so that $f^{-1}(R-\{0\})=J$ is in $\mathscr F$ , each element of $S$ is in $\mathscr F$ but $\mathscr B(R^d)$ is the smallest $\sigma-$algebra containing $S$, hence $\mathscr B(R^d) \subseteq \mathscr F$

  • $\begingroup$ Thank you! Why is it important that $f(J)$ does not contain 0? $\endgroup$ – John Simpleton Oct 9 '18 at 15:55
  • $\begingroup$ Note the equality $f^{-1}(R-\{0\})=J$ $\endgroup$ – S.D. Oct 9 '18 at 15:57
  • $\begingroup$ Is it necessary to provide an example of a continuous function that exhibits this behavior or is the existence inherent from the problem statement? $\endgroup$ – John Simpleton Oct 9 '18 at 16:10
  • 1
    $\begingroup$ Construction of $f_i$ is followed form graph of $f_i$ : join $(a_i,0)\in R^2$ with $(\frac{a_i+b_i}{2},1)\in R^2$ by line segment and join $(\frac{a_i+b_i}{2},1)\in R^2$ with $(b_i,0)\in R^2$ by another line segment and finally consider the two ray $(-\infty,a_i)$, $(b_i,\infty)$ as parts of graph. $\endgroup$ – S.D. Oct 9 '18 at 16:19
  • $\begingroup$ Makes perfect sense. Can't thank you enough for your help. $\endgroup$ – John Simpleton Oct 9 '18 at 16:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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