2
$\begingroup$

For a sentence $\varphi$, we'll define $Spec(\varphi)$ to be the set of all $n\in\mathbb{N}$ for which there is a model $M$ with $|D^M|=n$, such that $M\models\varphi$.
Let $\Sigma=\{P(\cdot),R(\cdot)\}$, where $P,\,R\,$ are relations.
$\boldsymbol{Disprove:} $ There is a sentence $\psi$ above $\Sigma$ such that $Spec(\psi)=\{n\,|\,n\in\mathbb{N}_{odd} \}$.

I have a disproof, but I wonder if part of it is unnecessary.
This is the disproof I have which got all the points (the question taken from an exam):
Assume by contradiction that such $\psi$ exists. Because $1$ is an odd number, there is a model $M$, with $|D^M|=1$ such that $M\models\psi$. Without loss of generality let $D^M=\{a\}$.
Define a new model, $M'$, such that: $D^{M'}=\{a,b\}$ and the next holds:
$\{a\}\in R^M \iff \{a,b\}\in R^{M'}$
$\{a\}\in P^M \iff \{a,b\}\in P^{M'}$
$M'$ is an expansion of $M$. We will define an equivalence relation, $\boldsymbol{\sim}$, with respect to the relations $P,R$. Namely, $\boldsymbol{\sim}$ will have four equivalance classes: as the number of options to be in $P$ or $R$.
By the way $M$ defines $P$ and $R$, there is only one equivalence class, which is: $\{a,b\}$. Let $[\bar{a}]$ be it's representor.
Now we will define an additional model, $N$, as follow: $D^N$ will contain the representatives of the equivalence classes of $\boldsymbol{\sim}$ with respect to $M$. $N$ will define the realtions $P$ and $R$ as follows:
$\{a,b\}\in R^{M'}\iff \{\bar{a}\}\in R^N$
$\{a,b\}\in P^{M'}\iff \{\bar{a}\}\in P^N$
So we actually have:
$M'\models\psi \iff N\models\psi$
From transitivity we have:
$\{a\}\in R^M \iff \{\bar{a}\}\in R^N$
$\{a\}\in R^M \iff \{\bar{a}\}\in P^N$
Namley $N$ is isomorphic to $M$, and becuase $M\models\psi$ we have due to isomorphism $N\models\psi$ which is contradiction, $|D^N|$ is an even number (2).

My question is: wasn't it sufficient to show that $M\models\psi \iff M'\models\psi\,\,$?
Or the whole reason for defining $N$ was because we can't show the above? I can't really understand what $N$ was needed for.

Would appreciate help!

$\endgroup$

1 Answer 1

2
$\begingroup$

You are right that $N$ is used in this way. If you want a proof without $N$, you will have to prove in some other way (for example, by sentence induction) that $M'\models\psi$.

I understand you that it is very intuitive why $M'\models\psi$, but sadly formalism doesn't care about intuition, so we need a rigorous proof.

$\endgroup$
1
  • $\begingroup$ Got it, thank you! :) $\endgroup$
    – Ella
    Commented Jun 17, 2021 at 5:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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