If there is a ''set'' W in V which is a standard model of ZF, and the ordinal κ is the set of ordinals which occur in W, then Lκ is the L of W. If there is a set which is a standard model of ZF, then the smallest such set is such a Lκ. This set is called the minimal model of ZFC. Using the downward Löwenheim–Skolem theorem, one can show that the minimal model (if it exists) is a countable set.

Of course, any consistent theory must have a model, so even within the minimal model of set theory there are sets which are models of ZF (assuming ZF is consistent). However, those set models are non-standard. In particular, they do not use the normal element relation and they are not well founded. (http://en.wikipedia.org/wiki/Constructible_universe)

I don't get it. OK, so in this case, one assumes that there exists some standard model that is not a class, but a set. Then one constructs a constructible universe inside such set which is also a set. Then one uses Lowenheim-Skolem to show that the smallest such model is countable.

My question is, it seems that the model is externally not standard (viewed from a bigger model) - so why call it standard first place? Or am I being confused, and is the model really standard? And the quote in the later part seems to say that the model is not standard, which confuses me much.

And I do not get the second paragraph - why would any consistent theory having at least one model relates to the minimal model containing models of ZF? I am not sure how Godel's completeness theorem is being matched here.


If you have a standard model $M$, and $N$ is a countable submodel (not even elementary, just a substructure which is a model of $\sf ZFC$) then $N$ is well-founded, because its $\in$ is a subset of the real $\in$ of the universe of sets (which is well-founded by the axiom of foundation).

We can therefore collapse $N$ to a countable transitive model, which is then a standard model as wanted.

Within the minimal model, however, we can't do this sort of trick anymore. There is no set which is a standard model that we can begin with. However the standard integers cannot encode a proof of $\lnot\text{Con}(\sf ZFC)$, so the minimal standard model also satisfies $\text{Con}(\sf ZFC)$ and therefore has a model. But this model cannot be standard, it has to be ill-founded.

  • $\begingroup$ Asaf: can you possibly add a few words on how that ill-founding 'hides' from the internal model (presumably in some fashion akin to how a model's countability 'hides' internally in that the bijections which show countability are not part of the model)? I can't speak for OP but that would certainly help me; I'm having a hard time quite grokking how ill- (or well-)foundedness aren't absolute. $\endgroup$ – Steven Stadnicki Jun 26 '13 at 15:38
  • $\begingroup$ @Steven: I'm not 100% sure about your request (but I'll be happy to add more once I am sure about it). Models of set theory may not be aware that there is a decreasing sequence $f_{n+1}\mathrel{E} f_n$ such that $f_n\in M$ and $E=\in^M$. So even if every $f_n$ is in the model, the set $\{f_n\mid n\in\omega\}\notin M$ -- so the model does not know about this violation of the axiom of foundation. $\endgroup$ – Asaf Karagila Jun 26 '13 at 15:41
  • $\begingroup$ I think that's just what I was after: while there is such a sequence (and so 'externally' the model is ill-founded) the internal model itself isn't aware of it because it can't 'collect' all the $f_i$. That makes sense to me at least. $\endgroup$ – Steven Stadnicki Jun 26 '13 at 15:48
  • $\begingroup$ And thinking about it more: this doesn't violate $\omega$-consistency either, I presume, because it's not a question of having some non-exhibited example, but simply of not having the collection of the $f$ - which really makes it almost exactly identical to the situation of internal-vs.-external countability in countable models. $\endgroup$ – Steven Stadnicki Jun 26 '13 at 17:00
  • $\begingroup$ Thanks for all your comments, but I am not sure if I really got this issue correctly. What Godel's completeness theorem says is that for any countable first-order theory, there is a model, and a statement is a theorem of a theory only if a statement is true in all models. So we have a standard model M, we create a submodel N, by Lowenheim-Skolem, the smallest possible elementary (well, it's not necessary, but anyway.) model is countable, as the created countable model uses the exact set relation, it is standard by Mostowski collapse lemma. $\endgroup$ – sets Jun 27 '13 at 2:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.