Can a $\Delta_0$ formula $\varphi$ not be independent of ZFC?

187 Views Asked by At

I proved that a $\Delta_0$ formula $\varphi$ cannot be independent of ZFC, i.e.,

(*) If $\varphi$ is a $\Delta_0$ formula and $V$ is a model of ZFC, then one of the followings must hold:

(i) $V\models\varphi$, or

(ii) $V\models\neg\varphi$

I proved (*) by a Boolean valued model $V^B$:

Lemma

(i) If $\varphi(x,\ldots)$ is a $\Delta_0$ formula then $\lVert\varphi(\check{x},\ldots)\rVert\in\{0,1\}$.

(ii) If $\varphi$ is a $\Delta_0$ formula, then $\varphi(x,\ldots)$ if and only if $\lVert\varphi(\check{x},\ldots)\rVert=1$.

But I'm still not sure whether the proposition (*) is true since it seems powerful. If it is true, is there a proof without utilizing the Boolean valued model?

1

There are 1 best solutions below

0
On BEST ANSWER

There are several issues here.

The first is that formulas are generally with free variables. So what do you mean when you say that a formula is independent of a theory? Do you mean its universal closure (i.e. add $\forall$ on all the free variables?)

Or do you mean that if two models of set theory agree on elementary diagram of some object(s) $x$, then they also agree on their $\Delta_0$ properties?

The second is that you only proved that logic is Boolean. If $V$ is a structure of a language $\cal L$ (in our case, the $\{\in\}$), then for every sentence, $\varphi$, either $V\models\varphi$ or $V\models\lnot\varphi$. This is not about $\Delta_0$ or anything else. Just about how truth is defined—relative to a given structure—for every sentence.


But since you use Boolean valued models, I'm guessing that you're trying to get something out of forcing.

If $V$ is a model of $\sf ZF$, then in any generic extension $V[G]$, we have that $V$ is a transitive class of $V[G]$ (in the broad sense, at least, since it might not be definable in the absence of choice in $V[G]$). And since $\Delta_0$ formulas are absolute between transitive classes that contain all the relevant sets (i.e. the parameters), you immediately get that every $\Delta_0$ formula is true in $V$ if and only if it is true in $V[G]$.

The same can be said about the construction of inner models, e.g. $L$ or $\rm HOD$.

In fact, more can be said: if $W$ is an end-extension of $V$, then they also agree on $\Delta_0$ truths with respect to sets in $V$. (Here an end-extension means that we add new ordinals to $V$, but there is no $x\in W\setminus V$ whose rank is an ordinal in $V$.)


So why does [presumably] Jech even want to prove that sort of lemma? For starters, it's a good exercise in understanding how forcing works. Secondly, it is useful for later proofs. Thirdly, at that point, we haven't talked about the generic extension, where $V$ is a transitive subclass, so we can't quite appeal to that sort of absoluteness.