36

Assuming ZFC has a model, is there a model of ZFC such that in that model, ZFC has no model?

Also, assuming ZFC has a model, is there a model of ZFC such that in that model, ZFC is inconsistent?

Asaf Karagila
  • 393,674
dankness
  • 1,141

2 Answers2

50

Yes.

Recall that by the Completeness Theorem, having a model and being consistent are the same thing. Also, by Incompleteness, ZFC doesn't prove its own consistency. Finally, ZFC can prove the Soundness Theorem - that an inconsistent theory has no models!

So - assuming ZFC has a model - ZFC is consistent. If ZFC is consistent, then ZFC can't prove "ZFC is consistent." By completeness, this means there's a model of ZFC satisfying "ZFC is inconsistent." Since ZFC proves the Soundness Theorem, this model must think that ZFC has no model!

Note that there's a subtlety to the claim "There is a model $M$ of ZFC not containing a model of ZFC." What that means is that there is an $M$ which satisfies ZFC+"There is no model of ZFC." However, since ZFC is not finitely axiomatizable, what $M$ thinks ZFC is will be different from what ZFC actually is! And in fact every model $M$ of ZFC contains a structure $N$ which classically is a model of ZFC, but within $M$ appears to not be a model of ZFC! Neat, huh? See Joel David Hamkins' answer to https://mathoverflow.net/questions/51754/clearing-misconceptions-defining-is-a-model-of-zfc-in-zfc.

Community
  • 1
Noah Schweber
  • 245,398
  • 4
    The post by Joel is probably my favorite use of the Reflection theorem. – Asaf Karagila Jun 14 '16 at 22:32
  • 1
    @AsafKaragila Yup, it's pretty sweet. :D – Noah Schweber Jun 14 '16 at 22:32
  • 2
    Neat! Now I'm curious: Which axiom of ZFC could $N$ appear to violate within $M$? – dankness Jun 15 '16 at 09:11
  • @dankness: Axioms which are coded by non-standard integers; and it is possible that the proof of the violation does not have a standard length either. Namely, $M$ thinks that the axioms prove a contradiction, but to prove this contradiction we need to appeal to integers which are not really integers, just $M$ mistakes them to be such objects. – Asaf Karagila Jun 15 '16 at 12:50
  • @AsafKaragila: Now I'm confused: As far as I understand the concept of a model, a model will assign a truth value to every proposition, so I don't see how my question has anything to do with proofs. More specifically my question was for which ZFC axiom $\phi$, $M$ assigns "True" to the proposition "$N$ assigns 'False' to the proposition $\phi$". Your answer seems to be about why $M$ assigns "True" to the proposition "there is a proof of a contradiction in ZFC". Am I misunderstanding something? – dankness Jun 15 '16 at 13:15
  • 1
    @dankness: Yes, but the object in $M$ which $M$ "thinks" is the theory we denote by ZFC is not the actual theory ZFC from $V$. It contains more axioms, whose length is not an actual natural number, but rather an object that $M$ thinks of as a natural number. Similarly, the inconsistency, or proof of contradiction, need not be a finite proof. It can be of a non-standard length, or appeal to inference rules which are themselves not actual inference rules. – Asaf Karagila Jun 15 '16 at 13:17
  • @AsafKaragila Ohh, thank you! So the axioms that $M$ "thinks" $N$ violates are some nonstandard instantiations of the axiom schemas of specification and replacement then? – dankness Jun 15 '16 at 13:27
  • 1
    @dankness: Yes, $N$ is a model of an $M$-finite fragment of $M$-ZFC; it just happened that as this fragment is taken to be a non-standard fragment, it includes all the standard axioms, so from the point of view of the universe, $N$ is a model of ZFC. – Asaf Karagila Jun 15 '16 at 13:33
  • 1
    This answer is a bit confusing. :D – Buffer Over Read Nov 13 '16 at 21:35
  • @TheBitByte In what way? If you ask a specific question, I might be able to address it. – Noah Schweber Nov 13 '16 at 21:37
  • @NoahSchweber Let me inquire about something slightly similar but not exactly related: $con(\sf ZFC)$ cannot be proved in or using $\sf ZFC$ itself, but only in some theory "greater" than $\sf ZFC$ in some way, right? – Buffer Over Read Nov 13 '16 at 21:48
  • 1
    @TheBitByte Yes, that's correct (assuming ZFC is consistent). For example, $ZFC+Con(ZFC)$ can do the job :P. – Noah Schweber Nov 13 '16 at 21:58
  • @NoahSchweber Wait, what do you mean by $ZFC + con(ZFC)$? – Buffer Over Read Nov 13 '16 at 22:02
  • 1
    @TheBitByte This is ZFC + "ZFC is consistent" (although on the face of it this is not a statement in the language of set theory, it can be so coded - this is the real meat of Goedel's incompleteness theorem!). Note that although ZFC+Con(ZFC) can prove that ZFC is consistent, it can't prove that it itself is consistent! Weird though that may seem, there are consistent theories which prove that they are inconsistent; see e.g. this question. – Noah Schweber Nov 13 '16 at 22:05
  • @NoahSchweber This is getting confusing... :D – Buffer Over Read Nov 13 '16 at 22:09
  • @TheBitByte I suggest you ask a specific question about it, then, and not here on the comments thread. – Noah Schweber Nov 13 '16 at 22:13
  • How does this change if you use NBG instead of ZFC? NBG is in fact finitely axiomatizable. Would there be room for "fake" axioms in a model of NBG+"there is no model of NBG"? – Zemyla May 09 '18 at 07:43
14

Noah gave the correct answer. It is true, that if $\sf ZFC$ has any models, then there are models of $\sf ZFC$ in which there are no models of $\sf ZFC$.

But let me point out an important piece of information here.

We formulate statements about consistency in terms of natural numbers. This is because the language of set theory is finite, so we can encode formulas as natural numbers in a very robust way. How robust? Well, like a computer-program robust. It's a mechanical algorithm to move on thing into the other.

Why am I bringing this up? Because it serves the point that natural numbers are important to the statement that something is consistent or inconsistent.

More specifically, we work in a universe $V$. In $V$, it is true that $\sf ZFC$ is consistent. Now, it is important to point out that by this we mean "what $V$ interprets as the predicate which defines '$\sf ZFC$' in its own natural numbers". But we can show that if $M$ is a model of $\sf ZFC$ inside $V$, and $M$ and $V$ agree on the notion of "natural number", then both have the same first-order theory for the natural numbers.

In particular, the predicate '$\sf ZFC$' and the statement of its consistency are absolute between $V$ and $M$. And since $V$ thinks that $\sf ZFC$ is consistent, so must $M$.

This means that if $M$ is a model of $\sf ZFC$ which thinks that $\sf ZFC$ is inconsistent two things must be true:

  1. $M$ has non-standard integers (compared to $V$).
  2. The interpretation of $\sf ZFC$ inside $M$ is different the one in $V$.

Namely, '$\sf ZFC$' can be seen as a predicate in the language of arithmetic, and since in $M$ we have more natural numbers, this predicate is interpreted to have more axioms than it should really have; and first-order logic have more inference rules; and proofs can suddenly be "too long" as far as $V$ is concerned. All these things allow $M$ to find out a proof of contradiction from $\sf ZFC$.

This is the reason why we can find such models. And models whose natural numbers agree with that of the universe are called $\omega$-models, and in such models it is necessarily the case that $\sf ZFC$ is consistent, simply because the natural numbers agree with the universe, so the interpretation of '$\sf ZFC$' and the rules of first-order logic, are the same as in $V$. Since in $V$ we know that $\sf ZFC$ is consistent, it must be consistent in any $\omega$-model.

So while $\sf ZFC$ proves that if there are models of $\sf ZFC$, then there are models of $\sf ZFC+\lnot\operatorname{Con}(ZFC)$; it is not the case that $\sf ZFC$ proves that if there are models of $\sf ZFC$, then there are $\omega$-models of $\sf ZFC$, as this would be a violation of the incompleteness theorem.

Asaf Karagila
  • 393,674
  • 1
    Maybe I'm being a bit thick, but I don't quite see how your last sentence is true; can you elaborate? – Noah Schweber Jun 14 '16 at 22:39
  • You mean about $\omega$-models? – Asaf Karagila Jun 14 '16 at 22:40
  • Ah, I thought you were going to point out that any model $M $ of ZFC that thinks ZFC is inconsistent actually has a (code for) a model $N $ of ZFC, by reflection. Of course, $M $ does not recognize this property of $N $. – Andrés E. Caicedo Jun 14 '16 at 22:40
  • @Andrés: Noah already pointed that out (in the link to Joel's post on MO)... – Asaf Karagila Jun 14 '16 at 22:41
  • 2
    @AsafKaragila Yes. I don't immediately see why ZFC proving "if there's a model of ZFC, then there's an $\omega$-model of ZFC" violates Incompleteness - what if ZFC proves $\neg Con(ZFC)$, in which case it trivially proves that statement? – Noah Schweber Jun 14 '16 at 22:41
  • (Of course, we hope ZFC doesn't prove $\neg Con($ZFC$)$, but . . . :P) – Noah Schweber Jun 14 '16 at 22:42
  • @Noah: Because if $M$ is an $\omega$-model, then it is a model of ZFC+Con(ZFC). If you can prove that Con(ZFC) implies Con(ZFC+Con(ZFC)), you are breaking the incompleteness theorem. – Asaf Karagila Jun 14 '16 at 22:42
  • 1
    @AsafKaragila Let me rephrase my question. Can you prove "If ZFC is consistent, then ZFC doesn't prove "if ZFC has a model, it has an $\omega$-model"" without a correctness assumption on ZFC? – Noah Schweber Jun 14 '16 at 22:43
  • @Noah: What do you mean by a correctness assumption? – Asaf Karagila Jun 14 '16 at 22:44
  • 1
    @AsafKaragila E.g. $\Sigma^0_1$ correctness; if you prefer, a soundness assumption. My point is that your argument can't be completely without an extra assumption, since T=ZFC+Con(ZFC)+"ZFC proves $\neg Con$(ZFC)" proves that (ZFC is consistent, but that ZFC proves that (if ZFC has a model, then ZFC proves that ZFC has an $\omega$-model)). – Noah Schweber Jun 14 '16 at 22:44
  • @Noah: I still don't think I follow. We know that Q is correct; since ZFC interprets Q, it is either inconsistent or it is correct as well. – Asaf Karagila Jun 14 '16 at 22:45
  • 1
    Hehe. I hadn't yet read Noah's answer. – Andrés E. Caicedo Jun 14 '16 at 22:47
  • @Andrés: The only valid excuse is "I was too busy watching the videos from MM70" ;-) – Asaf Karagila Jun 14 '16 at 22:49
  • Is it possible that ZFC$\vdash$ X but not ZFC$\vdash$(ZFC$\vdash$ X)?what about vice versa? – Vivaan Daga Jul 27 '22 at 10:57
  • @Shinrin-Yoku: No, if ZFC proves something, this can be coded as a standard integer (relative to the meta-theory), so ZFC must prove that it proves that. On the other hand, if we can have a meta-theory that allows us to prove that ZFC proves that ZFC is inconsistent, in which case ZFC proves "ZFC proves a contradiction", while in reality ZFC does not prove a contradiction (note that any such contradiction must be coded via a non-standard integer). – Asaf Karagila Jul 27 '22 at 13:56