Show that any choice set for the rational equivalence relation on a set of positive outer measure must be uncountably infinite

I have been thinking about this problem on and off for a few days, but I haven't been able to come with much. Let $E$ be a set with positive outer measure, and $C_E$ some choice set of $E$. I know that $E$ must be uncountable, otherwise it will have a outer measure of zero. This means that $E$ must contain an uncountable number of irrational numbers. Initially I thought that the equivalence class of an irrational would be a singleton, but this isn't so since $(1-\pi) + \pi = 1$. I could use a hint.


Let $[x] = \{y \in E ~|~ x-y \in \Bbb{Q} \}$ denote the rational equivalence class of the on $E$ of the element $x \in E$. Then $[x] = \{x + q ~|~ \mbox{ for some } q \in \Bbb{Q}\}$, which suggests that the elements of $[x]$ can be put into a one-to-one correspondence with some subset $\Bbb{Q}$, making the cardinality of $[x]$ at most countably infinite. As I noted above, a set with positive outer measure must be uncountably infinite. Now if there were only a countably infinite number of equivalence classes, then these equivalence classes, which partition $E$, couldn't add up to make the uncountable set $E$ (i.e., it would be impossible for their union to be $E$). Hence, there must be an uncountable number of equivalence classes. Since $C_E$ is formed by taking a single element from each equivalence class, it's clear that $C_E$ must also be uncountably infinite.

How does that sound?

  • $\begingroup$ What if the equivalence classes are not disjoint? $\endgroup$ – Bijayan Ray May 15 at 10:21
  • 1
    $\begingroup$ @BijayanRay Equivalence classes are always disjoint, provided that the relation is in fact an equivalence relation (i.e., a relation that is reflexive, symmetric, and transitive). $\endgroup$ – user193319 May 15 at 10:27

Think about the following three questions:

  • What is the cardinality of a single rational equivalence class?

  • What is the cardinality of the union of countably many rational equivalence classes?

  • Finally, what do you know about the cardinality of a set of positive measure?

  • $\begingroup$ I came up with solution based upon your three questions and edited my post to contain it. Perhaps you could take a look when you get a chance. $\endgroup$ – user193319 Sep 21 '17 at 23:36
  • 1
    $\begingroup$ @user193319 You've got it. $\endgroup$ – Noah Schweber Sep 21 '17 at 23:36

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.