# Thread: Formula with parameters

1. ## Formula with parameters

Hi everybody. Can someone help me finding a solution for this problem:

"Let M be a model and phi(x) be a formula with parameters in M, x being a single variable. Prove that if phi(M) contains the parameters of phi(x), then phi(M) is not an elementary substructure of M."

I'm not even fully understanding the problem, since I can't figure it out what the "parameters of phi(x)" are.

Thank you so much
bye!

2. Hi, and welcome to the forum.

Let $M=\langle D,\Sigma,I\rangle$ where $D$ is a domain, $\Sigma$ is a signature, and $I$ is an interpretation of $\Sigma$ on $D$.

I can't figure it out what the "parameters of phi(x)" are.
This probably means elements of D that occur in phi.

Prove that if phi(M) contains the parameters of phi(x), then phi(M) is not an elementary substructure of M.
What is phi(M)? I'll assume it is $\langle D',\Sigma,I'\rangle$ where $D'=\{a\in D\mid M\models\varphi(a)\}$, and $I'$ is the restriction of $I$ on $D'$. However, it is not clear that this is a structure, namely, that $D'$ is closed with respect to functions. (It may happen that for a functional symbol $f\in\Sigma$ and $a\in D'$, $I(f)(a)\notin D'$.) So, I'll assume that there are no functional symbols in $\Sigma$.

Next, what happens if $D'=D$, e.g., when $\varphi(x)$ is $x=x$? Then phi(M) = M and so phi(M) is an elementary substructure of M. So, I'll assume that $D'\subset D$, but $D'\ne D$, i.e., there exists an $a\in D$ such that $M\not\models\varphi(a)$.

In this case, what about $\exists x\,\neg\varphi(x)$?

Maybe I made too many assumptions here...

3. Hi and thank you for the quick reply! You made the right assumptions, I had a reply from a classmate of mine that clarified me the meaning of "parameters"... Apparently (I can't follow the lectures) our professor has a slightly uncommon approach to the subject, so the "parameters" are elements of a set that in this case, but not necessarily, coincide with D.

However, this:

Originally Posted by emakarov
Next, what happens if $D'=D$, e.g., when $\varphi(x)$ is $x=x$? Then phi(M) = M and so phi(M) is an elementary substructure of M.
is the exact same thought I had, that made me submit the problem to the community, since the definition of elementary substructure the professor gave to us implies $D'\subseteq D$.

Only a bit of a correction to your statement... in the hypotesis of the exercise it says "a formula with parameters" and $x=x$ doesn't have any. But the core of the reasoning is correct, I think it is sufficient putting $D=\{0,1,2\}, \Sigma=\{<\}$ with its normal interpretation, and $\varphi(x)=x<2 \vee x=2$. Am I right?

4. I believe so.

I tend to understand "a formula with parameters" as "a formula that is allowed to contain parameters". But to change my example, you could take phi to be x = x /\ a = a for some a in D, or x = a -> x = a.

Maybe your instructor wanted to claim that phi(M) is not a proper elementary substructure of M.

5. Originally Posted by emakarov
Maybe your instructor wanted to claim that phi(M) is not a proper elementary substructure of M.
This could be, I'll inspect this possibility!

Thank you!!