0
$\begingroup$

David Marker's Model Theory:
Theorem 4.2.11 Suppose that $T$ is a complete theory in a countable language and $A \subseteq M \models T$ is countable. If $|S_n^M(A)| < 2^{\aleph_{0}}$, then $|S_n^M(A)| < {\aleph_{0}}$.
Proof. Suppose that $|S_n^M(A)| > {\aleph_{0}}$. We claim that $|S_n^M(A)| = 2^{\aleph_{0}}$ . Because $|S_n^M(A)| > {\aleph_{0}}$ and there are only countably many $L_A$-formulas, there is a formula $\phi$ such that $|[\phi]| > {\aleph_{0}}$.
Claim If $|[\phi]| > {\aleph_{0}}$, there is an $L_A$-formula $\psi$ such that $|[\phi \wedge \psi]| > {\aleph_{0}}$ and $|[\phi \wedge \neg\psi]| > {\aleph_{0}}$. Suppose not. Let $p = \big\{\psi(\bar{v}):|[\phi \wedge \psi]| > {\aleph_{0}}\big\} $. Clearly, for each $\psi$ either $\psi \in p$ or $\neg \psi \in p$ but not both. We claim that $p$ is satisfiable. Suppose that $\psi_1,...,\psi_m \in p$. Either $\psi_1\wedge...\wedge\psi_m \in p$, in which case $\big\{\psi_1,...,\psi_m \big\} ∪ Th_A(M)$ is satisfiable, or $\neg\psi_1 ∨ . . . ∨ \neg\psi_m \in p$. Because $[\neg\psi_1 ∨ . . . ∨ \neg\psi_m]=[\neg\psi_1]∪...∪[\neg\psi_m]$, we must have $|[\neg\psi_i]| > {\aleph_{0}}$ for some $1\leq i\leq m$, a contradiction. Thus $p \in S_n^M(A)$.

My question is about this part of the proof which says if $\psi_1\wedge...\wedge\psi_m \in p$ then $\big\{\psi_1,...,\psi_m \big\} ∪ Th_A(M)$ is satisfiable!
I can't see why that's true. Would be thankful for your help.

$\endgroup$
  • $\begingroup$ Does "$[\chi]$" mean "the number of models of $T\cup\{\chi\}$"? If not, what does it mean? $\endgroup$ – Noah Schweber Dec 14 '16 at 16:37
  • $\begingroup$ @NoahSchweber It means the set of all types containing $\chi$. It's Marker's notation for the basic clopens in the Stone space. $\endgroup$ – Alex Kruckman Dec 14 '16 at 17:25
2
$\begingroup$

If $\psi_1\land \ldots\land \psi_m\in p$, then by definition $|[\psi_1\land \ldots\land \psi_m]|>\aleph_0$. In particular, this set is nonempty. So there's some type $q\in S^M_n(A)$ containing $\psi_1\land\ldots\land \psi_m$. Since types are consistent, $\{\psi_1,\dots,\psi_m\}\cup \mathrm{Th}_A(M)$ is satisfiable.

$\endgroup$
  • $\begingroup$ Thank you Alex, it was simply beautiful! $\endgroup$ – Aref Dec 14 '16 at 18:00

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.