A formula $\varphi$ such that no "small" transitive $M$ will satisfy $M\preccurlyeq _\varphi V$.

165 Views Asked by At

This is exercise II.5.6 from Kunen's Set Theory:

Assume Axiom of Choice. Find a formula $\varphi$ such that every transitive $M$ satisfying $M\preccurlyeq _\varphi V$ is of the form $R(\gamma)$ for some $\gamma = \beth _\gamma$.

(These $R(\gamma)$ are the "approximations of the universe" sometimes noted $V_\gamma$.)

It includes a hint:

$\varphi$ can be $\varphi_{0}(x)$$\land$$\varphi_{1}$ where $\varphi_0(x)$ says that $x$ is the set of the form $R(\alpha)$, and $\varphi_1$ is a sentence and a theorem of $ZFC$.

I imagine Kunen meant to write "where $\varphi_0(x)$ says that $x$ is $\textbf{a}$ set of the form $R(\alpha)$" but I copy it here verbatim just in case.

I don't get this hint, it seems to say $\varphi_1$ can be any theorem of $ZFC$ but this doesn't make sense as taking $\varphi_1$ to be $\exists z (z=z)$ we get $R(\omega)\preccurlyeq_\varphi V$ as for every $x\in R(\omega)$ we have $\varphi_0(x) ^{R(\omega)} \iff \varphi_0(x) ^{V}$. But $\omega \neq \beth_\omega$.

Am I misinterpreting the hint? I need help solving this exercise. Thank you for reading.

P.S. This exercise is probably related to the Reflection Theorem as it is given in the chapter dedicated to it.

1

There are 1 best solutions below

7
On BEST ANSWER

The point is that if $\varphi=\varphi_0(x)\wedge\varphi_1$ for $\varphi_1$ a theorem of ZFC, then $M\preccurlyeq_\varphi V$ iff

  • $M\preccurlyeq_{\varphi_0}V$, and

  • $M\models \varphi_1$.

So we're really getting to constrain $M$ in two ways: $M$ will be correct about $\varphi_0$, and $M$ will satisfy $\varphi_1$.

For example, take $\varphi_0(x)=$"$x=\langle y, z\rangle$ for some $y, z$ such that $y=\mathcal{P}(z)$" and $\varphi_1$ to be a large finite fragment of ZFC containing the Powerset axiom. Then $M\preccurlyeq_{\varphi}V$ implies that $M$ is closed under powersets: the $\varphi_1$-part says that $M$ contains "internal powersets" - that is, for each $z\in M$ there is some $y\in M$ such that $M$ thinks $y$ is the powerset of $z$ - and the $\varphi_0$-part implies that the "internal powersets" are in fact actual powersets.

Does this help?