next up previous
Next: Zadrozny on Cardinalities Up: Some Interesting Attempts Previous: Some Interesting Attempts

Perlis's Commonsense Set Theory

Perlis's approach was to develop a series of theories towards a complete commonsense set theory. He first proposed an axiom scheme of set formation for a naive set theory which he named (Perlis 1987):

Here is any formula and Ind is a predicate symbol with the intended extension ``individuals.'' This theory lacks further axioms, like an axiom of extensionality, which can be easily added. However, Ind can sometimes be critically rich, i.e., if is the same with Ind itself, then y may be too large to be an individual. (This is the case of Cantor's Axiom of Abstraction.) Therefore, a theory for a hierarchical extension for Ind is required. To support the cumulative hierarchy, Perlis extended this theory to a new one called using Ackermann's Scheme (Ackermann 1956) which is a formal principle of this hierarchy:

where can be interpreted as ``x can be built up as a collection from previously obtained entities.'' is consistent with respect to ZF. Unfortunately, it is hierarchical and hence not able to deal with self-referring sets.

Perlis finally proposed which is a synthesis of the universal reflection theory of Gilmore-Kripke (Gilmore 1974), which forms entities regardless of their origins and self-referential aspects, and the hierarchical theory of Ackermann (1956). GK set theory has the following axiom scheme where each well-formed formula has a reification (name) with variables free as in and distinguished variable x

where y does not appear in (cf. Note 16). There is also a definitional equivalence (denoted by ) axiom:

GK is consistent with respect to ZF (Perlis 1985). Perlis then proposed the following axioms to augment GK:

These axioms provide extensional constructions, i.e., collections determined only by their members. Thus, while GK provides the representation of circularity, these axioms support the cumulative construction mechanism. This theory can deal with problems like non-profit organization membership described earlier (Perlis 1988). But Perlis could not prove the consistency of yet because this requires linking two notions of membership of the two theories.



next up previous
Next: Zadrozny on Cardinalities Up: Some Interesting Attempts Previous: Some Interesting Attempts



Varol Akman
Sat Jan 13 15:54:04 EET 1996