5
$\begingroup$

$\newcommand{\P}{\mathbb{P}}$ Recall the following formulation of the Laver property (Page 459, Combinatorial Set Theory: With a Gentle Introduction to Forcing (Second Edition) by Halbeisen):

Laver Property. Let $\mathcal{F}$ be the set of all functions $S : \omega \to \operatorname{fin}(\omega)$ such that for every $n < \omega$, $|S(n)| \leq \color{red}{2^n}$. A forcing notion $\P$ has the Laver property iff for all $f \in {}^\omega\omega \cap V$ in the ground model and every $\P$-name $\dot{g}$ for a function in ${}^\omega\omega$ such that $1 \Vdash_\P \forall n < \omega(\dot{g}(n) \leq f(n))$, we have $1 \Vdash_\P \exists S \in \mathcal{F} \cap V \, \forall n < \omega(g(n) \in S(n))$.

Is it possible to modify the bound $\color{red}{2^n}$ above? For instance, is the Laver property equivalent to the statement above with the bound replaced by $r^n$ for some $r < \omega$, or even by $h(n)$ where $h \in {}^\omega\omega \cap V$?

$\endgroup$

1 Answer 1

3
$\begingroup$

You can replace $2^n$ by any function $h: \omega \to \omega \setminus \{0\}$ in $V$ so that $h(n)$ diverges to infinity. For instance if you want to get the Laver property for $h_0$ from the Laver property for $h_1$, let $i_n$, for each $n$, be the least $m$ so that $$h_0(m') \geq h_1(n),$$ for every $m' \geq m$.

Instead of $\dot g$ simply consider a name for $$g'(n) = g \restriction [i_n, i_{n+1}).$$

The Laver property for $h_1$ gives $h_1(n)$ guesses for the values of $g$ on the interval $[i_n, i_{n+1})$ which is less than the number of guesses asked by $h_0$ on that interval (and in fact ever after $i_n$). The only additional guess we have to make is on $[0, i_0)$, but that's of course fine.

$\endgroup$
3
  • $\begingroup$ Thanks Jonathan. Do you know if there's any known property that's equivalent to the Laver property, with the bound replaced by some constant (i.e. $h$ may be chosen to be uniformly bounded)? $\endgroup$ Commented Aug 12 at 5:10
  • 1
    $\begingroup$ @ClementYung Yes, that would be equivalent to not adding any new real. For any $x \in 2^\omega$, let $f(n) = x \restriction n$ (which clearly gives rise to a bounded fundtion). If you could guess $c$ values for each $f(n)$, where $c$ is a constant, then also in the end there can only be at most $c$ possibilities for $x$ and you can compute these possibilities in $V$. $\endgroup$ Commented Aug 12 at 21:26
  • 1
    $\begingroup$ @ClementYung cnt'd... More specifically, if $S \in V$ is your guessing function consider $\{ y \in 2^\omega : \forall n (y \restriction n \in S(n)) \}$. If it had $c+1$ many elements, then there must be a point from which all of their initial segments are pairwise different, which is not possible. This set is a finite Borel set coded in $V$, so it is contained in $V$. $\endgroup$ Commented Aug 12 at 21:29

You must log in to answer this question.

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