Inmathematical logic, atheory iscategorical if it has exactly onemodel (up to isomorphism).[a] Such a theory can be viewed asdefining its model, uniquely characterizing the model's structure.
Infirst-order logic, only theories with afinite model can be categorical.Higher-order logic contains categorical theories with aninfinite model. For example, the second-orderPeano axioms are categorical, having a unique model whose domain is theset of natural numbers
Inmodel theory, the notion of a categorical theory is refined with respect tocardinality. A theory isκ-categorical (orcategorical inκ) if it has exactly one model of cardinalityκ up to isomorphism.Morley's categoricity theorem is a theorem ofMichael D. Morley (1965) stating that if afirst-order theory in acountable language is categorical in someuncountablecardinality, then it is categorical in all uncountable cardinalities.
Saharon Shelah (1974) extended Morley's theorem to uncountable languages: if the language has cardinalityκ and a theory is categorical in some uncountable cardinal greater than or equal toκ then it is categorical in all cardinalities greater than κ.
Oswald Veblen in 1904 defined a theory to becategorical if all of its models are isomorphic. It follows from the definition above and theLöwenheim–Skolem theorem that anyfirst-order theory with a model of infinitecardinality cannot be categorical. One is then immediately led to the more subtle notion ofκ-categoricity, which asks: for which cardinalsκ is there exactly one model of cardinalityκ of the given theoryT up to isomorphism? This is a deep question and significant progress was only made in 1954 whenJerzy Łoś noticed that, at least forcomplete theoriesT over countablelanguages with at least one infinite model, he could only find three ways forT to beκ-categorical at some κ:
In other words, he observed that, in all the cases he could think of,κ-categoricity at any one uncountable cardinal impliedκ-categoricity at all other uncountable cardinals. This observation spurred a great amount of research into the 1960s, eventually culminating inMichael Morley's famous result that these are in fact the only possibilities. The theory was subsequently extended and refined bySaharon Shelah in the 1970s and beyond, leading tostability theory and Shelah's more general programme ofclassification theory.
There are not many natural examples of theories that are categorical in some uncountable cardinal. The known examples include:
There are also examples of theories that are categorical inω but not categorical in uncountable cardinals. The simplest example is the theory of anequivalence relation with exactly twoequivalence classes, both of which are infinite. Another example is the theory ofdenselinear orders with no endpoints;Cantor proved that any such countable linear order is isomorphic to the rational numbers: seeCantor's isomorphism theorem.
Every categorical theory iscomplete.[1] However, the converse does not hold.[2]
Any theoryT categorical in some infinite cardinalκ is very close to being complete. More precisely, theŁoś–Vaught test states that if a satisfiable theory has no finite models and is categorical in some infinite cardinalκ at least equal to the cardinality of its language, then the theory is complete. The reason is that all infinite models are first-order equivalent to some model of cardinalκ by theLöwenheim–Skolem theorem, and so are all equivalent as the theory is categorical inκ. Therefore, the theory is complete as all models are equivalent. The assumption that the theory have no finite models is necessary.[3]