Loading field map
Model theory is a subfield of mathematical logic that studies the relationship between formal languages and the mathematical structures that interpret them. It asks: given a set of axioms expressed in a formal language, what can we say about the structures that satisfy those axioms? And conversely, given a class of structures, what formal properties do they share? The field thus sits at the intersection of logic and algebra, using logical tools to understand mathematical structures and using algebraic insights to understand logical systems.
The core enterprise of model theory is the classification of mathematical structures by their logical properties. A language (or signature) specifies a collection of symbols for constants, functions, and relations. A structure for that language assigns actual mathematical objects and relations to those symbols. A theory is a set of sentences in the language; a model of the theory is a structure that makes all those sentences true. The fundamental question is: what can be said about the class of all models of a given theory?
This question has several dimensions. One is compactness: if every finite subset of a theory has a model, does the whole theory have a model? The compactness theorem, a landmark result, answers yes for first-order logic, and this fact has deep consequences—for example, it implies that any first-order theory with arbitrarily large finite models also has infinite models, and that there exist nonstandard models of arithmetic. Another dimension is categoricity: a theory is categorical if all its models of a given infinite cardinality are isomorphic. Morley's theorem (1965) showed that if a countable first-order theory is categorical in one uncountable cardinality, it is categorical in all uncountable cardinalities—a result that launched the modern classification program.
The stakes of model theory are both internal to mathematics and philosophical. Internally, model theory provides tools for analyzing algebraic structures (fields, groups, ordered sets) and for proving independence results—showing that certain statements cannot be proved or disproved from given axioms. Philosophically, model theory illuminates the nature of mathematical truth, the limits of formalization, and the relationship between syntax (formal sentences) and semantics (mathematical meaning).
Model theory emerged gradually from the confluence of several logical traditions. In the late 19th and early 20th centuries, mathematicians like Richard Dedekind, Giuseppe Peano, and David Hilbert worked on axiomatizing mathematical theories—arithmetic, geometry, set theory—and studying their models. The Löwenheim–Skolem theorem (1915, 1920) showed that any countable first-order theory with an infinite model has models of every infinite cardinality, revealing a fundamental limitation of first-order logic: it cannot uniquely characterize infinite structures.
The field crystallized in the 1930s and 1940s with the work of Alfred Tarski, who developed the formal concept of truth in a structure and initiated the systematic study of definable sets. Tarski and his students, particularly Abraham Robinson, established the basic theorems: compactness, the completeness theorem (due to Kurt Gödel), and the elementary chain theorem. Robinson's work on nonstandard analysis (1960s) applied model theory to analysis by constructing models of the real numbers that contain infinitesimals.
The modern period began in the 1960s and 1970s with Michael Morley's categoricity theorem and the subsequent development of stability theory by Saharon Shelah. Shelah's work transformed model theory from a collection of techniques into a systematic classification program, aiming to understand the structure of models of a theory by analyzing the complexity of definable sets.
Model theory is not divided into rival schools in the way some fields are, but it does contain distinct research programs that address different aspects of the subject. These programs coexist and interact, with many model theorists working across multiple approaches.
The classical tradition, rooted in Tarski and Robinson, focuses on the basic properties of first-order logic: compactness, completeness, elementary embeddings, and the construction of models. Its methods include the compactness theorem (used to build models with prescribed properties), the Löwenheim–Skolem theorem (used to control cardinalities), and ultraproducts (a construction that builds new models from families of existing ones). A central technique is the method of diagrams: given a structure, one expands the language with constants for its elements and writes down all true atomic sentences; a model of this diagram is an elementary extension of the original structure.
Classical model theory also studies elementary equivalence (two structures satisfy the same first-order sentences) and elementary embeddings (maps that preserve all first-order formulas). The back-and-forth method provides a way to characterize when two structures are isomorphic or elementarily equivalent, particularly for countable structures. This tradition remains foundational: every model theorist learns these techniques, and they are applied across the field.
Stability theory, initiated by Shelah in the 1970s, is the most influential research program in contemporary model theory. It addresses the problem of classifying theories by the complexity of their models. The central notion is stability: a theory is stable if there is a bound on the number of types that can be realized in a model of a given size. Unstable theories (like the theory of the natural numbers with addition and multiplication, or the theory of the real numbers as an ordered field) have many complex definable sets; stable theories (like the theory of algebraically closed fields) have a more tractable structure.
Shelah developed a hierarchy of stability classes: superstable, stable, unstable, and totally transcendental theories. For each class, he proved structural theorems about the models. The culminating result is the main gap theorem: for a countable first-order theory, either the theory has a "structure theory" (its models can be classified up to isomorphism by a reasonable set of invariants) or it has the maximum possible number of models in every uncountable cardinality. This theorem does not say that every theory falls neatly into one of these two categories, but it establishes a precise sense in which classification is possible for some theories and impossible for others.
Stability theory also introduced the concept of forking, a notion of independence that generalizes linear independence in vector spaces and algebraic independence in fields. Forking allows model theorists to define a notion of "dimension" for models of stable theories, leading to a geometric understanding of their structure.
Geometric model theory, sometimes called model-theoretic geometry, studies the interaction between model theory and algebraic geometry. It grew out of the observation that algebraically closed fields—a central example in stability theory—have a rich geometric structure. The key idea is to analyze definable sets in a structure using notions of dimension, independence, and closure that mimic those of algebraic geometry.
A landmark result is the Zilber trichotomy principle (proposed by Boris Zilber in the 1980s), which conjectures that in a strongly minimal set (a definable set where every definable subset is either finite or cofinite), the geometry of independence is either trivial (like a set with no structure), modular (like a vector space), or field-like (like an algebraically closed field). This principle was later refined and proved in many cases by Ehud Hrushovski and others, though it fails in general without additional assumptions. Hrushovski's work used geometric model theory to prove the Mordell–Lang conjecture for function fields (a deep result in number theory), demonstrating the power of model-theoretic methods outside logic.
Geometric model theory also studies o-minimal structures (ordered structures where every definable subset of the line is a finite union of intervals and points), which provide a tame setting for analysis and geometry. O-minimality has been used to prove results in real algebraic geometry, Diophantine geometry, and analysis.
Finite model theory is a distinct subfield that studies the relationship between logic and finite structures. It diverges from classical model theory because many of the central theorems (compactness, Löwenheim–Skolem) fail for finite structures. Instead, finite model theory focuses on descriptive complexity: which classes of finite structures can be defined by sentences in a given logic? This connects to computational complexity theory, because the expressive power of a logic often corresponds to the complexity of deciding whether a finite structure satisfies a sentence.
For example, the class of finite graphs that are connected cannot be defined in first-order logic, but can be defined in second-order logic or in first-order logic with a transitive closure operator. Finite model theory has developed its own techniques, including Ehrenfeucht–Fraïssé games (which characterize when two finite structures are elementarily equivalent) and 0-1 laws (which state that for certain logics, the probability that a random finite structure satisfies a sentence is either 0 or 1). While finite model theory is sometimes considered a separate field, it shares foundational concepts with classical model theory and many model theorists work in both areas.
Contemporary model theory is a mature but active field. Stability theory remains a central framework, but it has been extended in several directions. Simple theories (a class broader than stable theories) were introduced by Shelah and later developed by Byunghan Kim and others; they retain some of the good properties of stable theories while including examples like the theory of the random graph. NIP theories (theories that do not have the independence property) include both stable theories and o-minimal theories, and have been studied extensively for their combinatorial and geometric properties.
Model theory continues to find applications in other areas of mathematics. In number theory, Hrushovski's work on approximate subgroups and the model theory of difference fields has led to new results. In algebra, model-theoretic methods have been used to study groups, fields, and valued fields. In analysis, o-minimality provides a framework for studying tame geometric structures. The field also interacts with set theory, particularly through the study of large cardinals and forcing (which can be used to construct models of set theory with prescribed properties).
A notable recent development is the rise of abstract elementary classes (AECs), a framework that generalizes first-order model theory to contexts where the language may not be first-order. AECs were introduced by Shelah to study classes of structures that are not axiomatizable in first-order logic (such as the class of all models of a sentence in infinitary logic). This area has grown significantly, with results on categoricity, amalgamation, and the existence of models in various cardinalities.
The field remains unified by its core questions: how do logical languages constrain the structures that interpret them, and how can we classify those structures? Different approaches—stability, geometry, finite model theory, AECs—address different aspects of these questions, but they share a common toolkit of concepts (types, definability, saturation, indiscernibles) and a common commitment to understanding the interplay between syntax and semantics.