Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let AC denote the Axiom of Choice. Let PP denote the so-called "Partition Principle" which states that "If S is a non-empty set and T is a non-empty set of pairwise disjoint subsets of S, then S can be mapped onto T". It is well known that "AC implies PP" is provable in ZF, but the question of whether "PP implies AC" is provable in ZF has long been an open problem. What is the present status of this problem? Has any progress been made on it? Or-on the other hand-have any models of ZF been constructed in which pp is true and some consequence of AC -such as the existence of non-measurable sets of real numbers-is false? If "pp implies AC" could be proved in ZF, this would seem to provide a powerful philosophical argument for accepting AC. In my opinion, any set theory in which pp can be disproved yields a really counter-intuitive picture of the "Set-theoretical Universe".

share|improve this question
PP+DC implies the existence of a nonmeasurable set - mathoverflow.net/questions/22927/… –  François G. Dorais Jul 6 '13 at 18:47
I think I'm missing something. Fix some $t\in T$; why isn't the following map a surjection? $f: x\mapsto s\in T$ if $x\in s$, $x\mapsto t$ if $\forall s\in T(x\not\in s)$. Since the sets in $T$ are pairwise disjoint, and $T$ is nonempty, this is well defined, and can be proved to exist by ZF, and since each element of $T$ is a subset of $S$ this map is clearly onto. (Unless $T$ happens to contain $\emptyset$, but then just take $t=\emptyset$.) What am I missing? –  Noah S Jul 6 '13 at 18:51
Garabed, PP states that if $A\leq^\ast B$ then $A\leq B$, where $\leq$ means there is an injection and $\leq^\ast$ means there is a surjection. –  Asaf Karagila Jul 6 '13 at 18:54
Noah is right. The conclusion of PP should not be that S can be mapped onto T (which is trivial) but that T can be mapped one-to-one into S. –  Andreas Blass Jul 6 '13 at 19:03
Related. –  Andres Caicedo Jul 6 '13 at 20:17
show 11 more comments

1 Answer

up vote 11 down vote accepted

To my best knowledge, the Banaschewski-Moore paper from some twenty years ago is pretty much the last recorded progress on the topic.

The two main papers on the subject are the Banaschewski-Moore paper and a paper by Higasikawa, both are from more or less twenty years ago.

  1. Bernhard Banaschewski, Gregory H. Moore, The dual Cantor-Bernstein theorem and the partition principle, Notre Dame J. Formal Logic 31 (3), (1990), 375–381.

  2. Masasi Higasikawa, Partition principles and infinite sums of cardinal numbers. Notre Dame J. Formal Logic 36 (1995), no. 3, 425–434.

You can find a nice diagram of implications in Gregory Moore's book "Zermelo's Axiom of Choice" (which, oddly enough, is the second time I refer to on this site today).

I have some master plan on how to prove its independence from the axiom of choice, but it's a wild and vague dream at the moment which doesn't worth much mentioning except for the fact that I believe, at the moment, that PP does not imply the axiom of choice. For whatever that is worth.

One interesting fact on $\sf PP$ is that it implies the existence of non-measurable sets of real numbers all by itself. $\sf PP$ implies $\sf DC$, as well $\aleph_1\leq2^{\aleph_0}$ and therefore implies the existence of a non-measurable set. But in fact even much weaker versions of $\sf PP$ imply the existence of non-measurable, for example $\sf WPP$ which asserts $A\leq^\ast B\rightarrow B\nless A$, or in other words: if $B$ can be mapped onto $A$ then it cannot have a strictly smaller cardinality.

The reason that $\sf WPP$ implies the existence of a non-measurable set is that $\Bbb R$ can always be mapped onto $[\Bbb R]^\omega$, the set of countably infinite sets of real numbers, and of course can be mapped into that set injectively.

Since in $\sf ZF$ we have $\Bbb R\leq[\Bbb R]^\omega\leq^\ast\Bbb R$, in $\sf ZF+WPP$ we have that $\Bbb R$ is equipotent with $[\Bbb R]^\omega$. Sierpinski proved from this assumption that there exists a non-measurable set.

Sierpinski, W. "L’axiome de M. Zermelo et son rôle dans la théorie des ensembles et l’analyse." Bulletin de l’Académie des Sciences de Cracovie, Classe des Sciences Math., Sér. A (1918), 97-152.

share|improve this answer
To Noah and Asaf: I stated a version of pp which simply –  Garabed Gulbenkian Jul 7 '13 at 17:46
@Garabed: You might want to make this comment on the question, rather than my answer (and you should ping Noah using @ in front of his user name, I already know there is a comment coming up...). –  Asaf Karagila Jul 7 '13 at 17:47
@Noah: I stated a version of PP which simply stipulates –  Garabed Gulbenkian Jul 7 '13 at 17:51
I stated a version of PP which simply stipulates that the –  Garabed Gulbenkian Jul 7 '13 at 17:53
@Garabed: This answer is not the question. You should make this comment on your own question, this way Noah is notified as well. Be advised that pressing Enter submits the comment now, but you can edit your comment (for the first five minutes) using the "edit" link which appears not far from your user name after the comment. –  Asaf Karagila Jul 7 '13 at 17:58
show 2 more comments

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.