Let $(X,\mathcal M)$ be a measurable space. Are these two equivalent?

  1. $f:(X,\mathcal M)\to (\mathbb R,\mathcal B(\mathbb R))$ is measurable.

  2. $f:(X,\mathcal M)\to (\mathbb R,\{(a,\infty):a\in\mathbb R\})$ is measurable.

Here, $\mathcal B(\mathbb R)$ is the collection of all borel sets in $\mathbb R$. To say that $f:(X,\mathcal M)\to (Y,\mathcal N)$ is measurable means $f:X\to Y$ and $f^{-1}(B)\in\mathcal M$ for all $B\in\mathcal N$. (I know that $\{(a,\infty):a\in\mathbb R\}$ is not a $\sigma$-algebra, but let's not worry about that.)

I'm curious about this because in the book Rudin. "Principles of Mathematical Analysis" 3/e. p. 310. (Chapter 11. The Lebesgue Theory), the author defines a measurable function to be a function that satisfies 2, while other books use 1 to define a measurable function.

It's easy to prove that 1 implies 2, since $(a,\infty)\in\mathcal B(\mathbb R)$ for all $a\in\mathbb R$. But I want to know if 2 implies 1.

  • $\begingroup$ 2 implies 1 because the smallest $\sigma$-algebra containing the sets $(a,\infty)$ contains the standard topology on $\mathbb R$, so the two generated $\sigma$-algebras are the same. $\endgroup$ – Ashwin Trisal Aug 27 '18 at 6:54

2 implies 1.

Let it be that for every $a\in\mathbb R$ we have $f^{-1}((a,\infty))\in\mathcal M$ or equivalently that $f^{-1}(\mathcal V)\subseteq\mathcal M$ where $\mathcal V:=\{(a,\infty)\mid a\in\mathbb R\}$.

Then: $$\sigma(f^{-1}(\mathcal V))\subseteq\mathcal M\tag1$$ and it can be shown that in general:$$\sigma(f^{-1}(\mathcal V))=f^{-1}(\sigma(\mathcal V))\tag2$$

For a proof of that see here.

So based on $(1)$ and $(2)$ we find:$$f^{-1}(\sigma(\mathcal V))\subseteq\mathcal M$$

The final step is proving that:$$\sigma(\mathcal V)=\mathcal B(\mathbb R)$$

It is evident that $\sigma(\mathcal V)\subseteq\mathcal B(\mathbb R)$ and for the other side it is enough to prove that:$$\tau\subseteq\sigma(\mathcal V)$$where $\tau$ denotes the standard topology on $\mathbb R$.

For this realize that every open set in $\mathbb R$ can be written as countable union of open intervals, secondly that $(a,b]=(a,\infty)\cap(b,\infty)^{\complement}$ and thirdly that $(a,b)=\bigcup_{q\in\mathbb Q\cap(a,b)}(a,q]$.


(2)$\implies $(1) is obvious. For the converse, consider $$\mathcal A=\{X\subset \mathbb R\mid f ^{-1}(X)\in \mathcal M\}.$$

It's a $\sigma -$algebra that contain open set (why ?) and thus $\mathcal B(\mathbb R)\subset \mathcal A$ what prove the claim.


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.