3
$\begingroup$

Consider the set $\mathcal{P}(\mathbb{R})$ of all subsets of $\mathbb{R}$, the set of real numbers. It has a natural partial order: $A \leq B$ iff $A \subseteq B$.

Can one extend this order to a total order?

(I was discussing this with a friend and we didn't know if this is possible. If we replace $\mathbb{R}$ by any finite set, this is possible. We were not sure even when we replace $\mathbb{R}$ by $\mathbb{N}$.)

$\endgroup$
1

1 Answer 1

10
$\begingroup$

Michael Greinecker's answer (in a comment) is correct in the context of the usual axioms of set theory, including the axiom of choice. If one works in set theory without the axiom of choice, then one cannot prove that $\mathcal P(\mathbb R)$ admits any total order at all (even without the requirement that it extend $\subseteq$). On the other hand, the $\subseteq$ ordering of $\mathcal P(\mathbb N)$ can be explicitly extended to a total ordering, given by lexicogaphically ordering the characteristic functions of subsets of $\mathbb N$.

$\endgroup$
4
  • 1
    $\begingroup$ Uhm...it's Michael, actually. :-) $\endgroup$ Jan 5, 2012 at 23:20
  • $\begingroup$ That's true... :) $\endgroup$
    – expmat
    Jan 6, 2012 at 0:40
  • $\begingroup$ See also mathoverflow.net/questions/26861/… where I explain why ${\mathcal P}({\mathbb R})$ lacks a linear ordering in some well studied models of set theory without choice. $\endgroup$ Jan 6, 2012 at 3:51
  • $\begingroup$ Apologies to Michael for my error, and thanks to Andres for correcting it. $\endgroup$ Jan 6, 2012 at 15:09

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.