I think that this statement is very well-known but I am a bit unclear on some of the reasoning.

I am aware that $ V_{ \kappa }$ models ZFC when $ \kappa $ is an inaccessible cardinal. Therefore by Downward Lowenheim Skolem theorem we know that there is some countable elementary submodel $M$ such that $ M \preccurlyeq V_\kappa$, which since ZFC is a collection of sentences , must also model ZFC( right?). Now I read in Jech:

Thus there is $ E \subset \omega \times \omega $ such that $ \mathfrak{A} = ( \omega, E)$ is a model of ZFC.

One is then to verify that $ V_\kappa \models $ ZFC.

If I am not mistaken I think that we wan to take advantage of the fact that $ \Delta_0$ formulas are upward absolute which requires that our model be transitive . I think this is why we want such a $ \mathfrak{A}$ model as above. However, I do not know why we can assert the existence of such a model. Is it due to the Mostowski collapsing function, $ \pi$?

Thank you for any hints/help.

  • 1
    $\begingroup$ Related: math.stackexchange.com/questions/802565/… $\endgroup$
    – Asaf Karagila
    Oct 19, 2015 at 22:35
  • 3
    $\begingroup$ Look: If there is a model of set theory at all ($V_\kappa$ or whatever), there is a countable one, because this shows that $\mathsf{ZFC}$ is a countable theory that is consistent. Once you have a countable model, you have that, for any countably infinite set $A$, there is a countable model with universe $A$. $\endgroup$ Oct 19, 2015 at 23:09
  • $\begingroup$ Ah, I didn't see that post. I guess I should have searched under a different title. Thanks @AsafKaragila. $\endgroup$
    – Jmaff
    Oct 19, 2015 at 23:13

1 Answer 1


Once you have a model $(M,{\in})$ where $M$ is countable, you also have a bijection $f:\omega\to M$ -- because that's what it means for $M$ to be countable. Now let $$ E = \{ (a,b)\in\omega\times\omega \mid f(a)\in f(b) \} $$

This makes $(\omega,E)$ isomorphic to $(M,{\in})$. Since $M$ is known to satisfy ZFC, so does $(\omega,E)$.

  • 1
    $\begingroup$ You need a little more, though, to show that $V_{\kappa} \models (\mathfrak{A} \text{ is a countable model of ZFC})$. It suffices to show that the bijection $f \in V_{\kappa}$: then $\mathfrak{A} \in V_{\kappa}$, and confirming the rest is routine. But clearly, $\sup_{n<\omega} (rank(f(n)) < \kappa$, so $f \in V_{\kappa}$. $\endgroup$
    – BrianO
    Oct 19, 2015 at 20:41

You must log in to answer this question.

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