Dedekind complete models

Baizhanov B.S.


Throughout in the paper we assume that $M$ is an totally
ordered structure.

{\bf Definition 1.}
A partition $(A,B)$ of a model $M$ is said to be {\it cut} if $A <
B$. Here, $A < B \iff \forall a \in A \; \forall b \in B \; (a < b).$
A cut is said to be {\it rational} if $A$ has maximal element or $B$
has minimal one, or one of them is empty.
We say that cut is {\it quasirational} if $A$ and consequently $B$
are definable.
A non-quasirational cut is said to be {\it irrational}.
A model $M$ is said to be {\it Dedekind complete} if any its cut is
rational.
We say that model $M$ is {\it quasi Dedekind complete} if
any its cut is quasirational.
Let $M\prec N$ We say that the cut $(A,B)$ in $M$ is realized in $N$
if there exists $\alpha\in N\setminus M$ such that $A < \alpha
< B.$\\
{\bf Definition 2.} [L. van den Dries, Pillay, Steinhorn] A totally
ordered structure $M$ is
said to be {\it o-minimal,} if every definable (with parameters)
subset of $M$ is a finite union of points in $M$ and intervals
$(a,b),\; a \in M \bigcup \{ -\infty \},\; b \in M
\bigcup\{\infty\}$. \\
{\bf Definition 3.}[Marker-Steinhorn] Let $M \prec N \models T$, $T$ be
an o-minimal theory. A model $M$ is to be said {\it Dedekind
complete in $N$} if there is no non-rational cut in $M$ realized in $N$.\\
{\bf Definition 4.}[Dickmann, Marker-Macpherson-Steinhorn] A totally
ordered structure $M$ is said to be {\it weakly o-minimal,} if every
definable (with parameters) subset of $M$ is finite union of convex
sets.

Theory $T$ is {\it weakly o-minimal} if every model of $T$ is weakly
o-minimal.

{\bf Definition 5.}
Let $M \prec N \models T$, $T$ be a weakly o-minimal
(w.o.m.) theory. We say that $M$ is {\it quasi Dedekind complete in $N$} if there is no
irrational cut in $M$ realized in $N$.\\
Lou van den Dries had studied definable types over real closed fields
and proved the following result:\\
{\bf Theorem 1.} Every type over $(R,+,\cdot,0,1)$ is
definable.\\
D.Marker and Ch.Steinhorn had strengthened this result
expanding it for o-minimal theories:\\
{\bf Theorem 2.} Let $T$ be an o-minimal
theory, $M \models T$. Then the following holds:
\begin{enumerate}
\item $M$ is Dedekind complete $\iff$ every $n$--type of
$\bigcup_{n<\omega} S_n(M)$ is definable.
\item Let $M \prec N$.
Then $(\forall \bar \alpha \in N \setminus M)\; tp(\bar \alpha /
M) $ is definable $\iff M $ is Dedekind complete in $N$.
\end{enumerate}
We prove the following theorem:\\
{\bf Theorem 3.}
Let $M$ be a model of a w.o.m. theory $T$. Then $M$ is quasi Dedekind
complete iff any $p(\bar v) \in \bigcup_{n<\omega} S_n(M)$ is
definable.\\
And we also have the following:\\
{\bf Proposition 1.}
There is a w.o.m. theory with two models $M$, $N$ such that $M\prec N$
and $\exists \bar\alpha\in N\setminus M$:
$$tp(\bar\alpha/M) \mbox{ is non-definable } \mbox{ and }
M \mbox{ is quasi Dedekind complete in }N.$$