Prime Models and Minimal Models

Core Idea

A prime model of a complete theory T is a model that embeds into every model of T. Prime models represent the minimal model generated by the theory itself with no extraneous structure. They exist for every countable complete theory and are unique up to isomorphism over the empty set.

How It's Best Learned

Construct prime models explicitly by omitting types and using the completeness theorem. Compare prime models with saturated models and understand when they coincide.

Explainer

From your study of complete first-order theories and the Löwenheim-Skolem theorems, you know that a complete theory T can have models of many different cardinalities — countably many, uncountably many, models of any infinite size. This proliferation raises a natural question: is there a *smallest* or *most basic* model of T, one that contains only what the theory forces and nothing extra? That is the idea behind a prime model.

A model P of a complete theory T is prime if it embeds elementarily into every other model of T. An elementary embedding is stronger than an isomorphism of substructures — it is an injection j: P → M such that for every formula φ(x₁,...,xₙ) and every tuple a₁,...,aₙ from P, P satisfies φ(a₁,...,aₙ) if and only if M satisfies φ(j(a₁),...,j(aₙ)). So P is not just a substructure of every model — it is an *elementary* substructure, meaning every first-order property holds in P if and only if it holds in the corresponding part of M. This is the precise sense in which P is minimal: you cannot take anything away, because every other model of T contains an isomorphic copy of P.

To understand what makes a model prime, it helps to think about types. A type is a set of formulas describing the possible first-order behavior of an element (or tuple) in a model. A complete type fully specifies the first-order description of an element relative to the theory. In a prime model, every element realizes only principal types — types that are isolated by a single formula. Intuitively, a principal type is one that the theory forces: if there is an element satisfying that one formula, it must also satisfy the entire type. A prime model is built from only such "forced" elements — nothing extra is thrown in for free. This is why prime models are also called "atomic models" — every element is described by a single formula that determines everything else about it.

Prime models are unique up to isomorphism over the empty set, assuming T is complete and countable. This uniqueness mirrors the uniqueness of other canonical constructions in mathematics — the rational numbers as the prime model of dense linear orders without endpoints, the algebraic closure of ℚ as a kind of prime model for algebraically closed fields of characteristic zero (though here the precise notion is slightly different). The contrast with saturated models is instructive: where a saturated model is as large as possible — realizing every type that is consistent — a prime model is as small as possible, realizing only the types it is forced to. These two extremes bracket the spectrum of models of a complete theory, and understanding both is essential for classifying the structure of first-order theories.