Definition:Divide (Model Theory)

From ProofWiki
Jump to navigation Jump to search


Let $T$ be a complete $\LL$-theory.

Let $\mathfrak C$ be a monster model for $T$.

Let $A$ be a subset of the universe of $\mathfrak C$.

Let $\map \phi {\bar x, \bar b}$ be an $\LL$-formula with free variables $\bar x$ and parameters $\bar b$ from the universe of $\mathfrak C$.

Let $\map {\operatorname{tp} } {\bar b / A}$ denote the type of $\bar b$ over $A$.

Then $\map \phi {\bar x, \bar b}$ $k$-divides in $\mathfrak C$ over $A$ if and only if:

there exists a sequence $\sequence {\bar b_i}_{i \mathop \in \N}$ such that $\map {\operatorname{tp} } {\bar b_i / A} = \map {\operatorname{tp} } {\bar b / A}$ for each $i \in \N$


for any distinct $k$-many terms $\bar b_{i_1}, \ldots, \bar b_{i_k}$ of the sequence, the set $\set {\map \phi {\bar x, \bar b_{i_1} }, \ldots, \map \phi {\bar x, \bar b_{i_k} } }$ is not satisfiable in $\mathfrak C$.

Let $\map \pi {\bar x, \bar b}$ be a set of formulas with parameters $\bar b$.

$\map \pi {\bar x, \bar b}$ $k$-divides over $A$ if it implies a formula $\map \phi {x, \bar c}$ which $k$-divides over $A$.

Formulas and sets of formulas are said to divide if they $k$-divide for some $k$.


To help understand the definition, consider the following informal discussion.

We can think of $\map \phi {x, y}$ as a description of $x$ in terms of $y$.

If $b$ and $b'$ have the same type over $A$, then as far as $A$ can tell, $b$ and $b'$ are the same thing, in the sense that they both satisfy the same formulas with parameters from $A$.

So $\map \phi {x, b}$ $k$-dividing over $A$ means that we can find an infinite sequence of parameters $b_i$ that $A$ can't tell apart, but nevertheless, no $x$ can satisfy any $k$ of the descriptions $\map \phi {x, b_i}$ at the same time.


Consider the language $\LL = \set \sim$ with one binary relation symbol $\sim$.

Let $\MM = \struct {\R^2, P}$ be the $\LL$-structure where $P$ is the equivalence relation defined by $\tuple {x_1, y_1} \mathrel P \tuple {x_2, y_2}$ if and only if $x_1 = x_2$.

That is, $P$ identifies points in the plane which project vertically to the same value on the $x$-axis.

Let $\mathfrak C$ be some monster model for $\MM$.

The formula $x \sim \tuple {0, 0}$ $2$-divides over $\O$, which can be seen by considering the sequence $\tuple {\tuple {0, 0}, \tuple {1, 0}, \tuple {2, 0}, \tuple {3, 0}, \ldots}$.

The language is not very complicated; every $\tuple {n, m}$ has the same type over $\O$, so in particular each term in this sequence has the same type as $\tuple {0, 0}$.

However, $x \sim \tuple {n, 0}$ and $x \sim \tuple {m, 0}$ cannot be simultaneously satisfied for $n \ne m$, since this would imply $\tuple {n, 0} \mathrel P \tuple {m, 0}$, which is impossible.

The formula $x \sim \tuple {0, 0}$ does not $k$-divide over $A = \set {\tuple {0, y}: y \in \R}$ for any $k$ however, since only points of the form $\tuple {0, y}$ are of the same type as $\tuple {0, 0}$ over $A$.