Types and Saturated Models

Motivation: The "Portrait" of an Element in a Model

The type of an element is the aggregate of all properties possessed by that element in a given theory. Two models may have different types of elements. A "saturated" model "contains" all possible types—it is maximally rich. The theory of types is the key to understanding when models are isomorphic and to constructing "canonical" models of a theory.

Types

Partial n-type (over the empty set of parameters): A set of formulas $p(x_1,\ldots,x_n)$ such that $p \cup T$ is satisfiable (i.e., there exist elements in some model of $T$ which satisfy all $\varphi \in p$).

Complete n-type: For every formula $\varphi(x_1,\ldots,x_n)$: either $\varphi \in p$, or $\neg\varphi \in p$. The maximally informative type.

Realization of a type: Elements $(a_1,\ldots,a_n) \in M^n$ realize type $p$ if $M \models \varphi(a_1,\ldots,a_n)$ for all $\varphi \in p$.

Space of types $S_n(T)$: The set of all complete $n$-types of $T$. Stone topology: $[\varphi] = {p \in S_n(T) : \varphi \in p}$—the basic open sets. This is a compact topological space.

Saturated Models

$\kappa$-saturated model $M$: For every $A \subseteq M$, $|A| < \kappa$, and every type $p(x)$ over $A$ (satisfiable in an extension of $M$): $M$ realizes $p$.

Fully saturated: $M$ is $\kappa$-saturated with $\kappa = |M|$.

Theorem: Two saturated models of the same cardinality of the same complete theory are isomorphic. Saturated models are "canonical" representatives of the theory.

ω-Categoricity

$T$ is ω-categorical: $T$ has a unique (up to isomorphism) countable model.

Ryll-Nardzewski Theorem (1959): $T$ is ω-categorical $\iff S_n(T)$ is finite for every $n$.

Examples: $\operatorname{Th}(\mathbb{Q},<)$ is ω-categorical (DLO); $\operatorname{Th}(\infty$-atomic Boolean algebra) is ω-categorical; $\operatorname{Th}(\mathbb{Z},+)$ is not ω-categorical.

Numerical Example

Problem: Describe the complete 1-types in the theory DLO ($\operatorname{Th}(\mathbb{Q},<)$).

Step 1. DLO is a complete theory. Complete 1-types $p(x)$ are complete sets of formulas $\varphi(x) \in \operatorname{Th}(\mathbb{Q},<) \cup {\varphi(x)}$.

Step 2. Over the empty set of parameters: the unique complete 1-theory of DLO describes "an element in a dense linear order without endpoints." There are no additional parameters → all elements of $\mathbb{Q}$ have the same type (DLO is ω-categorical → one 1-type) ✓.

Step 3. Complete 2-types $p(x,y)$ over the empty set: three possibilities: $(x < y)$, $(x = y)$, $(x > y)$. Each of these defines a complete 2-type (in DLO the formula $x < y$, $x = y$, or $y < x$ is completely "extendable" to a complete type, since there are no other parameters). In total: 3 complete 2-types ✓.

Step 4. Ryll-Nardzewski check: $|S_1(\mathrm{DLO})| = 1$ (finite), $|S_2(\mathrm{DLO})| = 3$ (finite), $|S_3(\mathrm{DLO})| = \ldots$ orders of three elements: $3! = 6$ linear orders, plus coincidences → in DLO always one of $3!/$all permutations $= 6$ linear orders... wait, coincidences are also possible ($a = b < c$, $a < b = c$, etc.). A finite number ✓ → DLO is ω-categorical ✓.

Real Application

Database theory: types correspond to the "complete description" of a row in a given schema—all values of all attributes. Saturation—is an analogue of "completeness" of a representative sample. In data models: a saturated database contains "everything possible," which is important for substantiating the semantics of queries.

Additional Aspects

The partial type of an element $a$ in a model $M$ over a set of parameters $A$ is a consistent set of formulas $p(x)$ such that for a suitable $a'$ in an elementary extension, $p(a')$ is true. A complete type contains, for each formula, either it or its negation. The space of types $S(A)$ is a compact Stone space; the topological structure is connected to properties of the theory (stability, saturation). A saturated model realizes all types of small cardinality; such models are almost independent of choice and serve as "universal containers." At the intersection of model theory and algebra, saturation gives rise to theorems about quantifier elimination (quantifier-eliminable theories—algebraically closed fields of characteristic $0$, real closed fields; Tarski's theorem)—the foundation of automated geometry and computer algebra systems.

Connection with Other Branches of Mathematics

The notion of type and saturated model is organically connected with algebra through the works of Chevalley-MacLane, Robinson, and Los on elementary classes of algebraic structures. In field theory, the description of types over a subfield corresponds to the classification of polynomial roots and extensions: in algebraically closed fields, the types of one-dimensional elements over a subfield encode minimal polynomials and the arrangement of roots; saturated models of such theories coincide with "sufficiently large" universal fields used in Diophantine geometry (Lang, Deligne).

In differential equations, saturated models underlie the concept of Kolchin's differentially closed field and Morley's theory of categoricity: types describe solutions of differential equations with parameters, while saturated models play the role of the "general solution," containing all formally possible branches. In non-Archimedean analysis (Robinson, Lax) saturated extensions of the real numbers provide the existence of hyperreal numbers and allow one to formalize methods resembling numerical schemes with infinitely small steps.

Stone topology on the space of types is connected with functional analysis and probability theory through representations of Boolean algebras and ultrafilters: Stone's theorem on the representation of Boolean algebras and the construction of ultrafilter limits in probability employ the same apparatus. In nonstandard probability, Los and Kechris interpret saturated models as carriers of "rich" probability spaces, where types correspond to limit distributions. In numerical methods, the idea of saturation resonates with the construction of complete lattice grids and adaptive state spaces: each "local scenario" (type) should be represented in the model to guarantee algorithmic stability for all parameter configurations.

Historical Note and Development of the Idea

The origins of the notions of type and saturation appear in the works of Tarski in the 1940s on the elementary theory of fields, but the modern language formed in the 1950s–60s. Stone in 1936 introduced compact spaces, now called Stone spaces, which later turned out to be the natural topological shell for sets of types. Abraham Robinson, in his book "Model Theory and the Theory of Fields" (1959), systematized superstructures and saturated extensions, developing the ideas of Los's theorem about ultraproducts. The formal definition of types and saturated models in the familiar form is associated with the Tarski–Keisler school, and especially Morley and Shelah. Morley, in a 1965 article in the Journal of Symbolic Logic, proved the theorem about categoricity in uncountable cardinalities, where saturated models play a central role. Shelah, in the monumental series "Classification Theory" (beginning in the 1970s), developed the apparatus of stability, ranks, and saturation, turning work with types into a fundamental tool of theory classification.