5
$\begingroup$

What is the simplest non-constructible set of integers (say, in the analytical hierarchy) that is compatible with the nonexistence of $0^\sharp$? In particular, can there still be a non-constructible $\Delta^1_3$ set of integers if one assumes the nonexistence of $0^\sharp$?

$\endgroup$

1 Answer 1

8
$\begingroup$

It is consistent, relative to just $ZFC$, that there exists a non-constructible $\Delta_3^1$ set of integers. Such a result was first produced in the paper ``Some allpications of almost disjoint sets'' by Jensen-Solovay, using almost disjoint forcing.

In fact we can prove the following stronger result which is due to Jensen ``definable sets of minimal degree'':

It is consistent that $V=L[R],$ where $R$ is minimal (in the degree of constructibility) and it is the unique solution of a $\Pi_2^1$ predicate (hence it is $\Delta_3^1$).

$\endgroup$
6
  • $\begingroup$ By "compatible with the nonexistence of $0^\sharp$" I mean consistent with ZFC+"$0^\sharp$ does not exist." Even if "there is a non-constructible $\Delta^1_3$ set of integers" is consistent with ZFC, I'm wondering if it is also consistent with ZFC+"$0^\sharp$ does not exist." $\endgroup$ Nov 29, 2014 at 6:27
  • $\begingroup$ It is certainly not compatible with ZF+$V = L$, and the thought is that ZFC+"$0^\sharp$ does not exist" is perhaps not too far away from ZF+$V = L$. $\endgroup$ Nov 29, 2014 at 6:34
  • 1
    $\begingroup$ As the results are equiconsistent with $ZFC$, you just need to start from L, and force the resulting model in the extension. In the extension $0^\sharp$ does not exist. $\endgroup$ Nov 29, 2014 at 6:36
  • 1
    $\begingroup$ The fact that in the absence of $0^\sharp,$ $V$ is close to $L$ is a consequence of Jensen's covering lemma, which says if $0^\sharp$ does not exist, then any uncountable set of ordinals is covered by a constructible set of ordinals of the same size. $\endgroup$ Nov 29, 2014 at 6:43
  • $\begingroup$ I see. If the existence of a non-constructible $\Delta^1_3$ set of integers implied $0^\sharp$ exists, then its consistency strength would be greater than that of ZFC, which it is not. I misread your meaning of "consistent, relative to just ZFC." I haven't been able to locate the two papers yet but I'll take your word for it and accept your answer! $\endgroup$ Nov 29, 2014 at 7:50

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.