Model Theory

Model theory is the branch of mathematical logic that studies the relationship between formal languages and the mathematical structures those languages can describe.

9 frameworks, from 1950 to today · Mathematical Logic · Mathematics

still a working view ended · each leads to its entry below

Overview

Model theory is the branch of mathematical logic that studies the relationship between formal languages and the mathematical structures those languages can describe12. A formal language here is a precisely specified vocabulary—symbols for relations, functions, and constants—together with rules for building sentences1. A structure is a set equipped with actual relations, functions, and constants interpreting those symbols1. A structure is a model of a theory (a set of sentences) when every sentence in the theory is true in that structure1. Model theorists ask: given a theory, what do its models look like? Given a structure, what can a language say about it? These questions turn out to connect logic to algebra, geometry, number theory, and combinatorics, and the answers have reshaped how mathematicians understand whole areas of their subject.

From Löwenheim and Skolem to the Compactness Theorem

Model theory grew out of the late-nineteenth- and early-twentieth-century effort to put mathematics on rigorous logical foundations. Leopold Löwenheim proved in 1915 that if a first-order sentence has a model at all, it has a countable one; Thoralf Skolem strengthened and clarified this in the 1920s. The result, now called the Löwenheim–Skolem theorem, says that first-order logic cannot pin down the size of an infinite structure: any first-order theory with an infinite model has models of every infinite cardinality13. This was philosophically startling—a first-order description of the real numbers must also be satisfied by countable structures—but it became a foundational tool rather than a paradox.

The decisive result was the compactness theorem, proved by Kurt Gödel in 1930 as a corollary of his completeness theorem for first-order logic14. Compactness says that a set of first-order sentences has a model if and only if every finite subset does14. In practice this lets mathematicians build structures with prescribed properties by ensuring finite consistency. Abraham Robinson and others recognized in the 1950s that compactness could be used constructively, not just as a metatheorem, and this recognition marks the birth of model theory as an independent subject rather than a branch of proof theory.

Elementary Equivalence and the Birth of Classical Model Theory

The central notion of classical model theory is elementary equivalence: two structures are elementarily equivalent when they satisfy exactly the same first-order sentences5. A theory is complete when all its models are elementarily equivalent. Alfred Tarski's work on the decidability of elementary algebra and geometry in the late 1940s and early 1950s gave the first rich examples: the theory of real closed fields is complete and decidable, meaning there is an algorithm deciding whether any first-order sentence holds in the real numbers. Tarski's quantifier-elimination method—showing every formula is equivalent to one without quantifiers—became a template for later work1.

In the 1950s and 1960s, a generation of logicians built the apparatus of the field. Types—sets of formulas consistent with a theory, describing how an element or tuple could behave—became the basic objects of study. The omitting types theorem, developed by Leon Henkin and others, says a countable theory can omit any type that is not forced to be realized. Saturation and homogeneity gave model theorists a way to work inside very rich models where every consistent type is realized1. These tools let model theorists compare structures by their first-order content rather than by their algebraic presentation.

Nonstandard Analysis: Infinitesimals Made Rigorous

Abraham Robinson's 1960 monograph Non-standard Analysis showed that compactness could resurrect infinitesimals. By taking a nonstandard model of the theory of the real numbers—one containing elements larger than every standard natural number—Robinson obtained a rigorous framework in which infinitesimals exist and calculus can be done as Leibniz and Newton imagined. The transfer principle, a consequence of elementary equivalence, lets statements about standard reals be translated to nonstandard ones. Robinson's approach gave new proofs of classical theorems and, in the hands of later researchers, new results in analysis, probability, and mathematical economics. It also demonstrated that model-theoretic methods could illuminate mainstream mathematics, not just logic.

Classification Theory: Counting and Categorizing Models

The most ambitious program of classical model theory was classification theory, developed by Saharon Shelah from the late 1960s onward. The problem: given a complete first-order theory, how many models does it have in each infinite cardinality? Shelah introduced a hierarchy of stability properties—stability, superstability, and related notions—that measure how complicated a theory's types can be16. The stability spectrum of a theory records the number of models in each cardinality, and Shelah proved a sweeping dichotomy: unstable theories have the maximum possible number of models, while stable theories have a structured, analyzable spectrum. The main gap theorem (Shelah, 1970s–80s) showed that for countable theories, the number of models is either small and controlled by a well-understood invariant or maximal.

Classification theory introduced a vocabulary—forking, independence, canonical bases, regular types—that turned model theory into a kind of abstract geometry. The aim was to find, inside any stable theory, an analogue of algebraic or linear independence, and to classify models by the dimension-like invariants this independence yields.

Geometric Stability Theory and the Model Theory of Valued Fields

In the 1980s and 1990s, geometric stability theory pushed classification ideas toward geometry. Boris Zilber, Ehud Hrushovski, and others studied strongly minimal sets—definable sets with no nontrivial definable structure—and asked what geometries they can carry. Zilber's trichotomy conjecture proposed that every strongly minimal set is either trivial, or linear (like a vector space), or field-like (like an algebraically closed field). Hrushovski refuted the conjecture in general in the late 1980s by constructing new strongly minimal sets, but the geometric framework proved extraordinarily productive.

The model theory of valued fields became a showcase. The theory of algebraically closed valued fields (ACVF) was analyzed by Haskell, Hrushovski, and Macpherson in the 2000s, revealing a rich structure of definable sets and connecting to rigid analytic geometry. The p-adic numbers and their model theory, developed by Ax, Kochen, and later Denef and others, yielded applications to number theory and arithmetic geometry. Hrushovski's 1996 proof of the Mordell–Lang conjecture in characteristic zero for function fields—using model-theoretic methods—was a landmark demonstration that the subject could settle questions in algebraic geometry.

O-Minimality: Ordering and Geometry

A parallel development was o-minimality, introduced by Anand Pillay and Charles Steinhorn in 1984. A structure on a dense linear order is o-minimal if every definable subset of the line is a finite union of points and intervals17. This simple condition turns out to force strong geometric regularity: definable sets in higher dimensions are finite unions of cells, and definable functions are piecewise monotone and differentiable almost everywhere. Wilkie's 1996 theorem that the real exponential field is o-minimal opened the door to applications in real analytic geometry, and o-minimality became a central tool in the study of tame topology and in Diophantine geometry (notably through the Pila–Wilkie theorem on rational points).

Finite Model Theory: When the Domain Is Finite

Finite model theory asks the same questions as classical model theory but restricts attention to finite structures2. Here the compactness theorem fails, and the subject connects directly to computational complexity. A landmark result is Fagin's theorem (1974): the class of properties expressible in existential second-order logic is exactly the class NP28. This and related results—the descriptive complexity program—characterize complexity classes by the logical languages that capture them8. Finite model theory also studies 0-1 laws (the probability that a random finite structure satisfies a given sentence tends to 0 or 1) and the expressive power of fixed-point logics, which are used in database theory and verification.

Neo-Stability Theory and Continuous Model Theory

In the 2000s and 2010s, neo-stability theory extended classification ideas beyond the stable setting. Work by Shelah, Artem Chernikov, and others developed NIP (dependent) theories, NTP₂ theories, and rosy theories, each defined by combinatorial conditions on definable sets. These classes include many theories of interest in algebra and geometry—valued fields, differentially closed fields, and the p-adics—that are unstable but still tame in specific ways. The program seeks to find the right notions of independence and dimension in these broader contexts.

Continuous model theory, developed by Itaï Ben Yaacov, Alexander Berenstein, C. Ward Henson, and Alexander Usvyatsov in the 2000s, adapts model theory to structures whose truth values lie in a metric space rather than {true, false}. This is the natural setting for Banach spaces, operator algebras, and other analytic structures. The continuous analogue of compactness and the development of metric structures have made model theory applicable to functional analysis, and the subject now has a flourishing interaction with operator algebra theory.

The Present

Model theory today is a broad field with several active centers. Classical model theory and classification theory continue through neo-stability and the study of NIP theories1. Geometric stability theory and o-minimality remain central, with ongoing work on the structure of definable sets in valued fields, differential fields, and analytic settings. Finite model theory is closely tied to complexity theory and database theory12. Continuous model theory has opened analysis to model-theoretic methods. The unifying theme is still the original one: understanding what formal languages can and cannot say about mathematical structures, and using that understanding to prove theorems about the structures themselves1. Open questions include the full classification of NIP theories, the structure of definable sets in enriched o-minimal settings, and the precise relationship between model-theoretic tameness and algorithmic tractability in finite structures.

References

  1. Model theory (introduction, Basic model-theoretic concepts, Compactness and the Löwenheim–Skolem theorem, History, Eliminating quantifiers, Selected applications, Structures and types, Stability theory, The stability hierarchy, Minimality, Geometric stability theory, Finite model theory), Wikipedia
  2. Finite model theory (introduction, History), Wikipedia
  3. Löwenheim–Skolem theorem, Wikipedia
  4. Compactness theorem (History, introduction), Wikipedia
  5. Elementary equivalence, Wikipedia
  6. Stability spectrum, Wikipedia
  7. O-minimal theory, Wikipedia
  8. Descriptive complexity theory, Wikipedia

The frameworks of Model Theory

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1950-Present

    Classical Model Theory

    What does a first-order theory say about its models?

    About Classical Model Theory
    Classical model theory, dominant from roughly 1950 to 1970, addressed a foundational question: how can we classify the infinite mathematical structures that satisfy a given first-order theory? Its central insight was that the expressive power of first-order logic, while limited, is exactly strong enough to admit powerful metatheorems—compactness, the Löwenheim–Skolem theorems, and the ultraproduct construction—that allow us to transfer properties between structures and to build new models from old ones. This framework turned scattered results from the 1910s–1930s into a coherent research program, culminating in Michael Morley's categoricity theorem (1965), which showed that a countable first-order theory categorical in one uncountable cardinality is categorical in all uncountable cardinalities. Classical model theory provided the vocabulary and the core questions that later frameworks—stability theory, o-minimality, and model theory of fields—would refine and specialize.
    Core Curriculum
    A first course in classical model theory typically proceeds through these fundamental concepts, each building on the last:
    1. First-order languages, structures, and satisfaction – The basic syntax and semantics: a language LL consists of constant, function, and relation symbols; an LL-structure interprets these symbols in a domain; and Tarski's definition of satisfaction gives a rigorous meaning to truth in a structure.
    1. Compactness theorem – A theory has a model if and only if every finite subset does. This corollary of Gödel's completeness theorem (1929) is the workhorse for constructing nonstandard models and for transferring properties between structures.
    1. Löwenheim–Skolem theorems – If a first-order theory has an infinite model, it has models of every infinite cardinality (upward and downward versions). These theorems reveal the limits of first-order logic: no theory can pin down a unique infinite structure up to isomorphism.
    1. Elementary equivalence and elementary embeddings – Two structures are elementarily equivalent if they satisfy the same first-order sentences. An elementary embedding is a map between structures that preserves all first-order formulas. These notions allow us to compare structures without requiring isomorphism.
    1. Ultraproducts and the Łoś theorem – Given a family of structures and an ultrafilter on the index set, the ultraproduct is a quotient of the product. Łoś's theorem (1955) states that a first-order sentence holds in the ultraproduct if and only if it holds in a large (in the ultrafilter sense) set of the factors. Ultraproducts give a clean proof of compactness and enable transfer of properties.
    1. Types and saturation – A type is a set of formulas consistent with a theory; it describes the possible behavior of an element (or tuple) in a model. A model is κ\kappa-saturated if it realizes all types over subsets of size less than κ\kappa. Saturation is a key tool for constructing universal models and for understanding categoricity.
    1. Categoricity and Morley's theorem – A theory is categorical in a cardinal κ\kappa if it has exactly one model of size κ\kappa up to isomorphism. Morley's theorem (1965) proved that if a countable first-order theory is categorical in some uncountable cardinal, then it is categorical in every uncountable cardinal. This result launched the classification program in model theory.
    1. Model completeness and quantifier elimination – A theory is model-complete if every embedding between models is elementary. Quantifier elimination means every formula is equivalent to one without quantifiers. Abraham Robinson developed model completeness in the 1950s as a method for proving decidability and for applications to algebra (e.g., the theory of algebraically closed fields).
    These concepts form the teaching spine of classical model theory. They are not merely a list of theorems but a coherent toolkit for analyzing the relationship between syntax and semantics.
    Intellectual History
    The roots of classical model theory lie in the pre-history of logic. Leopold Löwenheim (1915) and Thoralf Skolem (1920) proved the first results about the existence of models of different cardinalities. Kurt Gödel's completeness theorem (1929) and compactness theorem (1930) provided the logical foundation. Alfred Tarski gave the subject its semantic basis with his definition of truth (1933) and, with his students, developed the concepts of elementary equivalence and elementary embeddings in the 1930s–1940s. However, these results were not yet part of a unified framework.
    The framework of classical model theory crystallized in the 1950s. Jerzy Łoś introduced the ultraproduct construction in 1955, giving a powerful algebraic method for building models and proving compactness. Abraham Robinson, in his 1956 book Complete Theories, developed model completeness and applied it to algebra, showing that the theory of algebraically closed fields is model-complete and hence decidable. Robert Vaught and others refined the notions of types and saturation. The landmark achievement came in 1965 when Michael Morley proved his categoricity theorem, which showed that the number of uncountable models of a countable theory is either one or infinite—a result that opened the door to stability theory.
    Key figures of the period include:
    • Alfred Tarski (1901–1983): provided the semantic foundation and, with his school, established the basic concepts of elementary equivalence and elementary embeddings.
    • Jerzy Łoś (1920–1998): introduced ultraproducts and proved the fundamental theorem bearing his name.
    • Abraham Robinson (1918–1974): developed model completeness and applied model theory to analysis (nonstandard analysis).
    • Michael Morley (1930–2020): proved the categoricity theorem, initiating the classification of theories.
    • Robert Vaught (1926–2002): contributed to the theory of types and saturation, and proved the Vaught two-cardinal theorem.
    Seminal works from the local period include:
    • Łoś (1955): “Quelques remarques, théorèmes et problèmes sur les classes définissables d'algèbres” – introduced ultraproducts.
    • Robinson (1956): Complete Theories – established model completeness.
    • Morley (1965): “Categoricity in Power” – proved the categoricity theorem.
    Classical model theory was not replaced but rather supplemented and specialized after 1970. Stability theory, beginning with Shelah's work in the 1970s, deepened the classification program. Model theory of fields and valued fields, o-minimality, and finite model theory each carved out a domain where the classical tools were refined or extended. The legacy of classical model theory is the enduring set of questions—about categoricity, definability, and the spectrum of models—that continue to drive research today.
    The 13 concepts its tower teaches
    • Categoricity
    • Compactness theorem
    • Elementary embeddings
    • Elementary equivalence
    • First-order languages, structures, and satisfaction
    • Łoś theorem
    • Löwenheim–Skolem theorems
    • Model completeness
    • Morley's theorem
    • Quantifier elimination
    • Saturation
    • Types
    • Ultraproducts

    Fly to its tower in Nousfarer

  2. 1960-Present

    Nonstandard Analysis

    Can infinitesimals be made rigorous?

    Fly to its tower in Nousfarer

  3. 1965-Present

    Model theory of valued fields

    What can logic say about valued fields?

    Fly to its tower in Nousfarer

  4. 1969-Present

    Classification Theory

    How many models does a theory have?

    Fly to its tower in Nousfarer

  5. 1974-Present

    Finite Model Theory

    What can logic express on finite structures?

    Fly to its tower in Nousfarer

  6. 1980-Present

    Geometric Stability Theory

    What geometries can definable sets carry?

    Fly to its tower in Nousfarer

  7. 1984-Present

    O-minimal theory

    When are definable sets geometrically tame?

    Fly to its tower in Nousfarer

  8. 2000-Present

    Continuous Model Theory

    How does logic work for metric structures?

    Fly to its tower in Nousfarer

  9. 2000-Present

    Neo-Stability Theory

    How far does stability generalize?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Classical Model TheoryNonstandard AnalysisModel theory of valued fieldsClassification TheoryFinite Model TheoryGeometric Stability TheoryO-minimal theoryContinuous Model TheoryNeo-Stability Theory

Beyond the field

Where Model Theory's ideas came from in other fields, and what they led to there: the strongest 12 of 26 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Model Theory is a planet in Nousfarer

Its city has a tower for each of these 9 frameworks. Land, walk in, and climb one concept at a time.

Land on Model Theory

Related fields

Fields its ideas are linked to

More in Mathematical Logic

See where Model Theory sits on the Map of Human Knowledge