1
$\begingroup$

Let $(\Omega, \mathcal{A})$ be a measurable space and $X:(\Omega, \mathcal{A})\rightarrow (\mathcal{X}, \mathcal{F})$ a measurable function and $f:(\mathcal{X}, \mathcal{F})\rightarrow (\mathcal{Z}, \mathcal{G})$ another measurable function.

I have stumbled across the following statement in a textbook $\sigma(f(X))\subseteq \sigma(X)$ and $\sigma(f(X))=\sigma(X)$ if and only if $f$ is one-to-one.

Now, my problem is that I dont understand if the author uses one-to-one in meaning an injective or a bijective mapping. Usually one-to-one means injection but here I have the feeling that we require $f$ to be a bijection. Can someone help?

EDIT: To add to my confusion:

$\sigma(T(X)) = \sigma(X) \iff T$ is bijective here it is being said that the statement is true if $f$ is bijective

When do we have $\sigma(X)= \sigma (f(X))$? here it is being said that the statement is true even if $f$ is only injective

But another uses suggests that the statement actually fails even when $f$ is bijective.

So what is correct now?

$\endgroup$
5
  • $\begingroup$ The equality can fail even if $f$ is bijective. $\endgroup$ Commented Oct 24 at 11:16
  • $\begingroup$ See: math.stackexchange.com/q/3781898/1104384 $\endgroup$ Commented Oct 24 at 11:17
  • $\begingroup$ @KaviRamaMurthy when would this statement then be correct? Doesnt the link by Bruno B suggest that the statement is true if $f$ is bijective? $\endgroup$ Commented Oct 24 at 11:23
  • 1
    $\begingroup$ Only for standard Borel spaces. You are having general measurable spaces, so even bijection is not enough, $\endgroup$ Commented Oct 24 at 11:28
  • $\begingroup$ @KaviRamaMurthy thanks a lot! So which of the spaces would need to be standard borel and would in that case then injection or bijection be enough? $\endgroup$ Commented Oct 24 at 11:30

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.