2
$\begingroup$

Let $\Gamma$ be a consistent set of $L$-sentences with infinite cardinality. If it has an infinite model, then there exists a model for $$ \Gamma' =\Gamma \cup \{\lnot c_a = c_b : a \neq b \}, $$ where $c_i$ is a constant symbol.

How would I go about proving this? I'm imagining that I can extend the model for $\Gamma$ to have a larger domain (to include more constant symbols) but I don't know how to show that it is actually a model for $\Gamma'$. Maybe with compactness?

$\endgroup$
5
  • 1
    $\begingroup$ Yes, compactness. $\endgroup$ – Asaf Karagila Dec 8 '18 at 20:43
  • $\begingroup$ Yes, with compactness. But it should be assumed that the constant symbols $c_i$ do not appear in $\Gamma$, otherwise the statement is not true. $\endgroup$ – Alex Kruckman Dec 8 '18 at 20:44
  • $\begingroup$ @AlexKruckman Why is that assumption necessary? $\endgroup$ – bof Dec 8 '18 at 20:45
  • $\begingroup$ For example, in the vocabulary $(c_a,c_b)$, $\Gamma$ could be the set $\{c_a = c_b\}$. $\endgroup$ – Alex Kruckman Dec 8 '18 at 20:48
  • $\begingroup$ @AlexKruckman Thanks, that makes sense. Now, compactness requires that every finite subset of $\Gamma'$ has a model but I'm not sure how I would go about proving that for every subset. Any hints? $\endgroup$ – bof Dec 8 '18 at 20:53
1
$\begingroup$

A finite subset of $\Gamma'$ is a finite subset of $\Gamma$, together with the assertions that finitely many $c_1,\dots,c_n$ are pairwise distinct.

Take any infinite model of $\Gamma$ (which exists by assumption), interpret $c_1,\dots,c_n$ as distinct elements, and interpret the rest of the constant symbols arbitrarily (we are free to do this, since the constant symbols are not mentioned in $\Gamma$). This is a model of our finite subset of $\Gamma'$. That's all there is to it.

This is the standard proof of the (upwards) Löwenheim-Skolem theorem.

$\endgroup$
4
  • $\begingroup$ Thank you. This was very helpful. $\endgroup$ – bof Dec 8 '18 at 21:05
  • $\begingroup$ I'm sorry but I have another question. Why can't we take any infinite model of $\Gamma$ and interpret $c_i$ as distinct elements and use it for the model of $\Gamma'$? What's special about finite subsets? $\endgroup$ – bof Dec 8 '18 at 21:19
  • 1
    $\begingroup$ Because there might be more constants then there are elements of the model. $\endgroup$ – Alex Kruckman Dec 8 '18 at 21:39
  • 1
    $\begingroup$ The point of the Löwenheim-Skolem theorem os to take a theory that has some infinite model (say a countably infinite one) and prove that it has models of all larger cardinities. $\endgroup$ – Alex Kruckman Dec 8 '18 at 21:41

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.