Take two sets $E_1$ and $E_2$, and assume $f$ is a function $E_1 \to E_2$. Take now a family of subsets of $E_2$ and call it $(O_i)_{i\in I}$, and consider the family $\left(f^{-1}(O_i)\right)_{i\in I}$ of subsets of $E_1$. Call $\mathcal B_1$ the $\sigma$-algebra on $E_1$ generated by $\left(f^{-1}(O_i)\right)_{i\in I}$ and $\mathcal B_2$ the $\sigma$-algebra on $E_2$ generated by $(O_i)_{i\in I}$

Is it true that $\mathcal B_1 = f^{-1}(\mathcal B_2)$ ? What I see is that at least $\mathcal B_1 \subset f^{-1}(\mathcal B_2)$.

If this is not true how would one prove that for a measurable function $f: E \to \overline{\mathbb{R}}$ the preimage of any borel set of $\overline{\mathbb{R}}$ is a measurable set of $E$, when measurability has been defined by: $f^{-1}(]c;+ \infty[)$ is a measurable set for any $c\in\mathbb{R}$


For any subcollection $\mathcal{C}$ of $\wp\left(E_{1}\right)$ or $\wp\left(E_{2}\right)$ let $\sigma\left(\mathcal{C}\right)$ denote the $\sigma$-algebra generated by $\mathcal{C}$.

Instead of $\left(O_{i}\right)_{i\in I}$ I will write $\mathcal{V}$ so that $\sigma\left(\mathcal{V}\right)$ denotes the $\sigma$-algebra generated by $\mathcal{V}$.

Defining $f^{-1}\left(\mathcal{C}\right)=\left\{ f^{-1}\left(C\right)\mid C\in\mathcal{C}\right\} $ for any subcollection $\mathcal{C}$ of $\wp\left(E_{2}\right)$ you are actually asking whether the following statement is true: $$f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)=\sigma\left(f^{-1}\left(\mathcal{V}\right)\right)$$ The answer to that question is: "yes".

Start by proving the following statements:

  • $f^{-1}\left(\mathcal{A}\right)$ is a $\sigma$-algebra whenever $\mathcal{A}\subseteq\wp\left(E_{1}\right)$ is a $\sigma$-algebra.
  • $\left\{ A\in\wp\left(E_{2}\right)\mid f^{-1}\left(A\right)\in\mathcal{B}\right\} $ is a $\sigma$-algebra whenever $\mathcal{B}\subseteq\wp\left(E_{1}\right)$ is a $\sigma$-algebra.

They are not too difficult to prove, especially because preimages are quite nice to work with.

The first statement tells us immediately that $f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)$ is a $\sigma$-algebra, and this of course with $f^{-1}\left(\mathcal{V}\right)\subseteq f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)$.

This allows the conclusion that $\sigma\left(f^{-1}\left(\mathcal{V}\right)\right)\subseteq f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)$.

The second statement tells us that $\left\{ A\in\wp\left(E_{2}\right)\mid f^{-1}\left(A\right)\in\sigma\left(f^{-1}\left(\mathcal{V}\right)\right)\right\} $ is a $\sigma$-algebra and this with $\mathcal{V}\subseteq\left\{ A\in\wp\left(E_{2}\right)\mid f^{-1}\left(A\right)\in f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)\right\} $.

This allows the conclusion that $\sigma\left(\mathcal{V}\right)\subseteq\left\{ A\in\wp\left(E_{2}\right)\mid f^{-1}\left(A\right)\in\sigma\left(f^{-1}\left(\mathcal{V}\right)\right)\right\} $ or equivalently $f^{-1}\left(\sigma\left(\mathcal{V}\right)\right)\subseteq\sigma\left(f^{-1}\left(\mathcal{V}\right)\right)$.

| cite | improve this answer | |
  • $\begingroup$ Thank you very much, this seems completely correct to me ! $\endgroup$ – incas Aug 29 '16 at 12:54
  • $\begingroup$ You are very welcome. $\endgroup$ – drhab Aug 29 '16 at 15:15

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.