MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X$ be a standard Borel space: that is, a topological space equivalent to a Borel subset of $\Bbb R$. It is known that for any probability measure $p$ on $X$ and any universally measurable set $A\subseteq X$ there exists a Borel set $B\subseteq X$ such that $p(A\Delta B) = 0$. Moreover, if $f:X\to\Bbb R$ is universally measurable, there exists a Borel-measurable $f'$ such that $f=f'$ $p$-a.e.

I wonder, whether the following result holds true:

If $Y$ is also a Borel space and $g:X\to Y$ is universally measurable, there exists a Borel map $g':X\to Y$ such that $g = g'$ $p$-a.e.

A possible proof that I have in my mind is the following: the countable case is trivial, so assume that $Y$ is uncountable, then there exists a Borel isomorphism $\phi:Y\to\Bbb R$. As a result, we can say that $$ g = \phi^{-1}\circ (\phi\circ g) $$ where $\phi\circ g:X\to\Bbb R$ is clearly universally measurable. Hence, there exists a Borel function $f':X\to\Bbb R$ such that $\phi\circ g = f'$ $p$-a.e. If we define $g':=\phi^{-1}\circ f'$, we obtain that $g':X\to Y$ is a Borel map and that $g=g'$ $p$-a.e.

Is the proof correct?

share|cite|improve this question
The proof is correct, but you might write standard Borel space instead of just Borel space. The latter is Mackey's term for a measurable space. – Michael Greinecker Jul 11 '13 at 13:07
@MichaelGreinecker: thanks, I just followed the terminology of the Bertsekas and Shreve's book. Have you seen such a statement as in OP somewhere? I haven't find it in BS. – Ilya Jul 11 '13 at 13:13
Lemma 1.2 in Crauel, Random Probability Measures on Polish Spaces, is slightly stronger. It says that if a function from a probbility space to a separable metric space is measurable with respect to the completion, it equal some measurable function almost surely. – Michael Greinecker Jul 11 '13 at 13:22
@MichaelGreinecker: nice, would you post this as an answer? – Ilya Jul 11 '13 at 13:24
up vote 3 down vote accepted

The proof is correct. The following result of a stronger assertion is taken from Hans Crauel's book Random Probability Measures on Polish Spaces, where it is Lemma 1.2.:

Proposition: Let $(\Omega,\Sigma,\mu)$ be a probability space, $Y$ a separable metric space and $f_0:\Omega\to Y$ be measurable with respect to the $\mu$-completion of $\Sigma$. Then there is a measurable function $f:\Omega\to Y$ such that the set $\{\omega:f_0(\omega)\neq f(\omega)\}$ has $\mu$-outer measure zero.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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