4
$\begingroup$

This is a repost of the same question on math.SE, which received several comments but no answers/comments on the first question.


Suppose $\kappa$ is a cardinal preserved in the generic extension $V[G]$. Let $Y \subseteq \kappa$ be an unbounded set in $V[G]$. Does there always exist an $X \in V$ such that $X \subseteq Y$ and $X$ is unbounded?

This question comes from Section 18 of James Cummings' Singular Cardinal Arithmetic. In this context, he forced with the Levy collapses $\operatorname{Col}(\omega,\delta^{+\omega}) \times \operatorname{Col}(\delta^{+\omega+2},<\kappa)$, and fixed "$X \subseteq \gamma$ unbounded of order type $\delta_V^{+\omega+1}$...". In the next line, he then says that "However, it is easy to see that there is $Y \in V$ with $Y \subseteq X$ unbounded".

This observation is certainly not easy for me, so my questions are:

  1. Why do such a $Y$ exist?
  2. Is the existence of such a $Y$ limited to forcing with Levy collapse?

Any help is appreciated.

Note: Andreas Blass provided a counterexample to the second question, which is a Mathias real.

$\endgroup$
3
  • 2
    $\begingroup$ Another example of a forcing adding an unbounded $Y \subseteq \kappa$ with no unbounded ground model set, somewhat similar to a Matthias real, is the Prikry forcing to add an $\omega$ cofinal sequence to a measurable cardinal $\kappa$ while preserving $\kappa$. The generic $\omega$-sequence you add to the measurable can't contain an infinite ground model set since $\kappa$ is regular in $V$. $\endgroup$ Commented Aug 23, 2021 at 9:11
  • $\begingroup$ If you force to add a subset $G$ of an infinite cardinal $\kappa$ using finite conditions (which is equivalent to adding $\kappa$-many Cohen reals), then $G$ is unbounded in $\kappa$ but contains no infinite set from the ground model. $\endgroup$ Commented Aug 23, 2021 at 14:22
  • 1
    $\begingroup$ I remember this from the introduction to Magidor's paper on the dual Jensen covering lemma, which shows that if a set of ordinals has some weak closure properties (primitive recursive closure) and $0^\#$ doesn't exist, then the set is the union of countably many constructible sets. This shows that under anti-large cardinal principles, with some reasonable constraints on the set in the extension, the answer to your question becomes yes. $\endgroup$ Commented Aug 23, 2021 at 14:30

1 Answer 1

4
$\begingroup$

Let $\mathbb{P} = \textsf{Col}(\delta^{+ \omega + 2}, < \kappa)$ and $\mathbb{Q} = \textsf{Col}(\omega, \delta^{+ \omega})$. Then $\mathbb{P}$ is $< \delta^{+ \omega + 2}$-closed so it doesn't add any new set of ordinals of order type $\leq \delta^{+ \omega + 1}$. Put $V_1 = V^{\mathbb{P}}$ and note that all cardinals $\leq \delta^{+ \omega + 2}$ are preserved in $V_1$. Suppose $V_1^{\mathbb{Q}} \models \mathring{Y}$ has order type $\delta^{+ \omega + 1}$. Since $|\mathbb{Q}| < \delta^{+ \omega + 1}$, we can find a condition $p \in \mathbb{Q}$ such that $p$ forces $ \delta^{+ \omega + 1}$ many ordinals into $\mathring{Y}$. Put $X = \{\alpha: p \Vdash_{\mathbb{Q}} \alpha \in \mathring{Y}\}$. Then $X \in V_1$ and therefore by the previous observation, $X \in V$.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .