Main Index Foundation of mathematics Set theory
  Subject Index
comment on the page

Zermelo-Fraenkel Set Theory

In 1908 Ernst Zermelo proposed the axiomatic set theory now called Zermelo set theory. One of its drawback was that it did not allow to construct arbitrary ordinal numbers.  In 1922  Abraham Fraenkel and Thoralf Skolem independently gave impetus to formulation of the Axiom of Replacement. This axiom together with Zermelo set theory gives what we call the Zermelo-Fraenkel set theory, abbreviated as ZF [1], p. 31-96.

Similarly as the naïve set theory also ZF depends on a single primitive non-defined ontological notion, that of set, and a single ontological assumption, that all mathematical objects are sets. The second underlying item is a single primitive dyadic relation, the set membership: typeset structure means that typeset structure is an element of typeset structure 1 .  In other words, we have two primitive (i.e. undefined) concept in axiomatization of set theory:

Note that set equality and the concept of belonging typeset structure are two different things. Set equality is transitive, as follows from the axioms below. Unlike set equality, belonging is not transitive: point is an element of a straight line, typeset structure, line typeset structure is an element of a bundle of lines typeset structure, but typeset structure is not element of typeset structure.

The axioms then govern how sets behave and interact. The axiom of the Zermelo-Fraenkel set theory  (ZF) are:

(Z1) Axiom of Existence: typeset structure

Axiom Z1 ensures the existence of at least one set.

(Z2) Axiom of Extensionality:   typeset structure
Note that this axiom of extension does not discribe only a logically necessary property of equality of two sets but touches in a non-trivial way also the concept of  belonging. For instance, if typeset structure and typeset structure are integers, write typeset structure whenever typeset structure divides typeset structure. The axiom says that if typeset structure and typeset structure have the same divisors then typeset structure and typeset structure are equal.   But this is not true, take typeset structure and typeset structure. Thsu even it is not for the first moment apparent, the axiom of extensionality provides a non-trivial pronouncement about belonging.

It defines that two sets are equal if and only if they contain the same elements. Thus typeset structure.

Define the notion of a subset as follows: if  typeset structure and typeset structure are two sets, then typeset structure is a subset of typeset structure,  write  typeset structure, if every element of typeset structure  is also an element of typeset structure. Then the Axiom of Extensionability can be expressed also in the following way:

(typeset structure) Axiom of Extensionality: Two sets typeset structure are equal if and only if  typeset structure and typeset structure.

Let typeset structure denote a formula in first-order theory that contains typeset structure as a free variable:
(Z3) Axiom of Separation (Subset Axiom or Axiom of Comprehension): typeset structure
That is, if typeset structure is a set and typeset structure is a condition on sets, there exists a set, in the common set-builder noatation denoted as typeset structure, whose elements are precisely the elements for typeset structure is true.  

(Z4) Unordered Pair Axiom: typeset structure
There exists a set containing given two elements (sets).

(Z5) Union Axiom: typeset structure
If typeset structure is a set, then there exists a set containing every element of each typeset structure.  This set is called the union and denoted typeset structure.

(Z6) Powerset Axiom: typeset structure
If  typeset structure is a set, then there exists a set typeset structure, usually denoted as typeset structure with the property that typeset structure if and only if any element typeset structure is also an element of typeset structure.

Axioms Z1 - Z6 are also called axioms of transformations and with their help  all of finite mathematics can be constructed.

The next axiom ensures the existence of an infinite set. In some axiom systems it is replaced by a weaker axiom Z1.  In what follows typeset structure, where typeset structure is some fixed set.

(Z7) Axiom of Infinity: typeset structure
Another formulation: There exists a non-empty set typeset structure with the property that, for any typeset structure, there is some typeset structure such that  typeset structure but typeset structure.

Axiom of infinity guarantees the existence of a set with elements typeset structure, typeset structure, typeset structure, ...

Here typeset structure denotes a function which takes any object as argument and yields as its value the singleton set with that object as its only member . Thus the set typeset structure which existence is ensured by the Axiom of Infinity is the set typeset structure.

By the way, the set of non-negative integers can be modeled using this set typeset structure when typeset structure is interpreted as typeset structure, typeset structure as typeset structure, typeset structure as typeset structure, etc. In this case typeset structure is the successor function which maps any non-negative integer to its successor value.

Another possibility for typeset structure is typeset structure.  

Let typeset structure be a predicate in 2 variables not depending on typeset structure. The next axiom of replacement scheme describes how new sets can be defined from exiting sets using relationship  typeset structure that defines typeset structure as a function fo typeset structure
(ZF8) Axiom of Replacement: typeset structure,
or in words: given any set typeset structure such that for any typeset structure, there is a unique typeset structure (denoted as typeset structure) such that typeset structure holds for typeset structure and typeset structure, there is a set B which we get by replacing every element of typeset structure with the object for which typeset structure holds.

Notation typeset structure is an abbreviation for typeset structure.

Axiom of replacement the construction that if  typeset structure is a set, and there is a way to assign each set typeset structure to some set typeset structure, then the image of  typeset structure under this assignment is a set.

(Z9) Axiom of Regularity (Axiom of Foundation): typeset structure.

The Axiom of Regularity, in combination with the Unordered Pair Axiom rules out every set that contains itself as a member. To see this remember that the Unordered Pair Axiom says that for every element of a set, there must be the singleton that contains it and nothing else. But the Axiom of Regularity rules this out, i.e. there is no set that is a member of itself.

(Z10) Axiom of Choice: If typeset structure is a function2 with non-empty domain typeset structure and for each typeset structure, typeset structure is a non-empty set then there is a function typeset structure also with domain typeset structure such that for each typeset structure we have typeset structure.

If we define  typeset structure, then the Axiom of choice can be formally define as

typeset structure

Axiom of Choice can be formulated in many ways.  .  The general idea behind this axiom is that for any collection of sets, you can choose one member from each of them. Here the values of function typeset structure give the initial collection of sets, and the function typeset structure is the selector function that chooses a single element from each of the sets that are the values of typeset structure.

The above axioms without axiom Z8 form the so-called Zermelo set theory, abbreviated as Z. In this system it is possible to develop the almost all parts of the classical mathematics, as arithmetic, analysis, functional analysis, etc. However it is impossible to develop within Z the theory of cardinals greater than typeset structure.  From this reason Fraenkel proposed to add a new axiom Z8.  The axioms(Z1)-(Z9)  form a foundation for what is now called Zermelo-Fraenkel set theory, or ZF. This set of axioms along with the Axiom of Choice is often denoted ZFC.  Since Axiom of Choice is non-constructive, it is customary to state explicitly when a mathematical result is proved using it.

In 1957, Richard Montague proved  [2]  that ZF (and consequently also ZFC) cannot be be finitely axiomatized, that is its construction requires at least one axiom schema.

ZFC is believed to be consistent 3. A.Abian and S.Lamacchia proved  [3] that if in a theory of sets every set is a power set and if the Powerset Axiom is valid then the Axioms of Extensionality, Union and of Choice are valid. They also proved [4]

Abian previously proved [5] that  the Axioms of Extensionality, Powerset, Union and of Choice form a consistent and independent system of axioms.

As mentioned above ZFC requires an infinite number of first-order axioms. Therefore the question whether it is possible to develop a set theory with a finite number of axioms is very natural. The question was answered in the affirmative by J. von Neumann in 1925.  However von Neumann used functions rather than sets as his primitive notion. The first-order version which is widely known now is due to the reworking in the late 1930's by Bernays with a contribution of Gödel. It is called von Neumann-Bernays-Gödel set theory, abbreviated to NBG set theory.

For the origin of notation used in set theory visit .

Notes

1 The symbol typeset structure has its origin in a notation introduced by Peano. He derived it from Greek typeset structure. He wrote in 1889: The sign typeset structure signifies is.

2 Note that the notion of a function can be correctly defined within the first-order theory.

3 Consistency of an axiom system means that is impossible to deduce within the system that a statement and its negative are both true.

References

[1]  Fraenkel, A. A., & Bar-Hillel, Y. (1958). Foundation of Set theory. Amsterdam: North-Holland Publishing Co..

[2]  Montague, R. (1961). Semantic closure and non-finite axiomatizability I. In . <Last> (Ed.), Proceedings of the Symposium on Foundations of Mathematics, (Warsaw, 2-9 September 1959). (pp. 45-69). Oxford: Pergamon.

[3]  Abian, A., & Lamacchia, S. (1965). Some consequences of the axiom of power-set. J. Symb. Log., 30, 293-294.

[4]  Abian, A., & Lamacchia, S. (1978). Some the consistency and independence of some set-theoretical axioms. Notre Dame Journal of Formal  Logic, 19, 155-158.

[5]  Abian, A. (1969). On the independence of set theoretical axioms.. Amer. Math. Monthly, 76, 787-790.

Cite this web-page as:

Štefan Porubský: Zermelo-Fraenkel Set Theory.

Page created  .