A sentential operator that takes a formula, binds some set of variables within its scope, and is evaluated in a model on the basis of the elements of its domain satisfying the formula that it modifies. The most frequently encountered quantifiers are the existential and universal quantifiers and , although more expressive quantifiers can be encountered if generalized quantifiers are considered.
Semantically, there are multiple ways to interpret a standard quantifier. On the account due to philosopher Gottlob Frege (1848–1925), quantifiers are second-order functions that take propositional functions as arguments and return truth values (or other propositional functions). For example, in the case of classical logic, a model can associate a propositional function with any formula with free variable so that maps every element of the domain to a truth value according to whether satisfies the formula or not. On the Fregean interpretation, a quantifier is interpreted by a function that maps every propositional function to a truth value, determined by the scheme:
In other words, is true in a model if the propositional function that associates with maps all elements of the domain to truth, i.e., is satisfied by all elements.
A second approach due to logician Walter Carnielli (1952– ) first defines a distribution of a formula in a model as the set of truth values to which the propositional function maps elements of the domain. This approach interprets quantifiers as functions mapping such distributions to truth values, i.e., when is a set of truth values. On the distribution account, the above conditions can be rephrased:
where is the distribution of a formula .
Many generalizations of quantification have been introduced, including logics that permit quantification over formulae (cf. propositional quantifiers), quantification over groups of individuals (cf. plural quantification), and accounts of quantification that eliminate dependence relations between variables (cf. branching quantifier). Even more variations are made possible by representing natural language quantifiers as generalized quantifiers.