In set theory and mathematical logic, the Lévy hierarchy, introduced by Azriel Lévy in 1965, is a hierarchy of formulas in the formal language of the Zermelo–Fraenkel set theory, which is typically called just the language of set theory.
This is analogous to the arithmetical hierarchy, which provides a similar classification for sentences of the language of arithmetic.
In the language of set theory, atomic formulas are of the form x = y or x ∈ y, standing for equality and set membership predicates, respectively.
The first level of the Lévy hierarchy is defined as containing only formulas with no unbounded quantifiers and is denoted by
[1] The next levels are given by finding a formula in prenex normal form which is provably equivalent over ZFC, and counting the number of changes of quantifiers:[2]p. 184 A formula
is called:[1][3] As a formula might have several different equivalent formulas in prenex normal form, it might belong to several different levels of the hierarchy.
[citation needed] Lévy's original notation was
) due to the provable logical equivalence,[4] strictly speaking the above levels should be referred to as
) to specify the theory in which the equivalence is carried out, however it is usually clear from context.[5]pp.
[6] The Lévy hierarchy is sometimes defined for other theories S. In this case
by themselves refer only to formulas that start with a sequence of quantifiers with at most i−1 alternations,[citation needed] and
formulas in the language of the theory S. So strictly speaking the levels
of the Lévy hierarchy for ZFC defined above should be denoted by
The Lévy hierarchy has the following properties:[2]p. 184 Devlin p. 29