5
$\begingroup$

Existence of non-measurable sets requires the use of Axiom of choice. And one can construct models eg. Solovay model without using the Axiom of choice where there are no non-measurable sets. So, Can one start with the the hypothesis that non-measurable sets exist and then 'prove' the Axiom of choice?

$\endgroup$
1
  • 2
    $\begingroup$ You probably should read this question: mathoverflow.net/questions/42215/… In particular, note that ZF+HB (where HB is a form of the Hahn-Banach theorem) implies the existence of nonmeasurable sets, and that HB is strictly weaker than AC. $\endgroup$ Feb 27, 2011 at 8:05

1 Answer 1

8
$\begingroup$

The answer to your question is "no".

Using the technique of forcing one can easily violate the axiom of choice in a variety of ways (essentially, by focusing on "large enough sets") without affecting the fact that there are non-measurable sets of reals. This includes leaving a bit of choice to make sense of the construction of Lebesgue measure and development of its basic properties (the principle of "dependent choice for relations on reals" seems more than enough for this).

As far as I know, there are no known "choice or choice-like principles" even at the level of the reals known to be equivalent or to reasonably follow from the assumption that there are non-measurable sets.

(By the way, this is form 93 in "Consequences of the axiom of choice" by Howard and Rubin. The companion website does not reveal any known implications.)

$\endgroup$
2
  • $\begingroup$ Thanks for your answer. In the case where we violate AC while still having non-measurable sets, can you please elaborate on which properties of Lebesgue measure would require some choice ? $\endgroup$
    – Dushyant
    Feb 27, 2011 at 7:41
  • 1
    $\begingroup$ For example, you want: 1) Non-triviality (${\mathbb R}$ must have non-zero measure), 2) Countable additivity, and 3) Singletons have measure zero. But it is consistent with the failure of choice that the reals are a countable union of countable sets, which makes these three conditions impossible to satisfy simultaneously. $\endgroup$ Feb 27, 2011 at 8:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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