Tarski-Vaught test

From Wikipedia, the free encyclopedia

The Tarski-Vaught test (sometimes called Tarski's criterion) is a result in model theory which characterizes the elementary substructures of a given structure using definable sets. It is often used to determine whether a substructure of a structure is elementary, and is particularly useful in the construction of elementary substructures. Formally,

For a given first-order language \mathcal{L}, let \mathcal{N} be a structure with domain N, and \mathcal{M} be a substructure of \mathcal{N} with domain M (written \mathcal{M}\subseteq\mathcal{N}). Then \mathcal{M} is an elementary substructure of \mathcal{N} (written \mathcal{M}<\mathcal{N}) if and only if for every nonempty A\subseteq N definable in \mathcal{N} with parameters from M, A contains an element of M.


[edit] Proof

First suppose \mathcal{M}<\mathcal{N}. Let A\subseteq N be nonempty and definable in \mathcal{N} using \varphi(x,x_1,\ldots,x_n) with parameters b_1,\ldots,b_n\in M. We must prove A\cap M\ne\emptyset. Since A\ne\emptyset, we have

\mathcal{N}\models\exists x\varphi[b_1,\ldots,b_n]

(Note that the bracket notation here is used to indicate the semantic evaluation of the free variables of the formula.) But then, by definition of an elementary substructure,

\mathcal{M}\models\exists x\varphi[b_1,\ldots,b_n]

Thus there exists an a\in M such that \mathcal{M}\models\varphi[a,b_1,\ldots,b_n]. Again by definition of elementary substructure, \mathcal{N}\models\varphi[a,b_1,\ldots,b_n], so a\in A. Thus A\cap M\ne\emptyset as desired.

Conversely, suppose every nonempty subset of N definable in \mathcal{N} with parameters from M contains an element of M. We prove \mathcal{M}<\mathcal{N} by induction on formulas. The atomic and boolean cases are immediate from definitions since \mathcal{M}\subseteq\mathcal{N}. Suppose the result holds for \psi(x,x_1,\ldots,x_n) and consider \exists x\psi. Since M\subseteq N, it is immediate that for any a_1,\ldots,a_n\in M, if \mathcal{M}\models\exists x\psi[a_1,\ldots,a_n], then \mathcal{N}\models\exists x\psi[a_1,\ldots,a_n]. Suppose now \mathcal{M}\not\models\exists x\psi[a_1,\ldots,a_n] for some a_1,\ldots,a_n\in M. Thus for all a\in M,

\mathcal{M}\not\models\psi[a,a_1,\ldots,a_n]

so by the induction hypothesis, for all a\in M,

(*) \mathcal{N}\not\models\psi[a,a_1,\ldots,a_n]

Now consider the set A\subseteq N defined in \mathcal{N} by \psi(x,x_1,\ldots,x_n) using parameters a_1,\ldots,a_n. If there exists b\in N such that \mathcal{N}\models\psi[b,a_1,\ldots,a_n], then A\ne\emptyset. By our hypothesis then, there exists an a\in A\cap M. But this contradicts (*). Thus \mathcal{N}\not\models\exists x\psi[a_1,\ldots,a_n]. It follows that \mathcal{M}<\mathcal{N} as desired. Q.E.D.

[edit] Example

Consider the structure \mathcal{N}=(\mathbb{N},<) consisting of the naturals with the usual ordering. Let \mathcal{M}=(P,<_P) be the substructure consisting of the prime naturals with the induced ordering. Is \mathcal{M} an elementary substructure of \mathcal{N}? Tarski's criterion provides an immediate negative answer since, for example, the number 4 is definable in \mathcal{N} without parameters by the formula

\exists x_1,x_2,x_3((x_1<x_2\,\land\,x_2<x_3\,\land\,x_3<x)\land \forall x_4(x_4<x\rightarrow (x_4=x_1\,\lor\,x_4=x_2\,\lor\,x_4=x_3)))

which states (in \mathcal{N}) 'there exist exactly 3 things less than me'.

[edit] References