Tableaux rules clarify definite descriptions in free logic theories

Tableaux Rules for Theories of Definite Descriptions in Positive Free Logic with Some Philosophical Remarks

Logic in Computer Science

Summary

Definite descriptions are phrases like 'the F' that talk about unique things. While these work well in classical logic, they are tricky in free logic where some things may not exist. The authors study different logical theories that try to handle these phrases, focusing on intermediate versions between very weak and very strong theories. They show that some promising middle-ground theories don't work as hoped and discuss the philosophical reasons why. This helps clarify the limits of defining unique descriptions in these logical systems.

What this means in practice

  • For logic system designers: Avoid using certain intermediate definite description theories in free logic that are shown to be equivalent to strong or problematic ones.
  • For automated reasoning developers: Use clarified tableau rules for definite descriptions to improve proof procedures in free logic-based reasoning systems.

A theory result. No direct application yet.

Authors

Nils Kürbis

Abstract

While Russell's theory of definite descriptions, expressions of the form 'the F', is almost incontrovertible in classical logic, in free logic matters stand differently. Several alternatives have been proposed. In Lambert's hierarchy, the weakest theory is his minimal free theory of definite descriptions (MFD), which only says that anything is the F iff it is the unique existing F, the strongest one is a chosen object theory in which all improper definite descriptions refer to the same object (FD2). Intermediate is Lambert's FD1, which adds ix(x = t) = t as an axiom to MFD. MFD is very weak. FD2 is very strong. One might hope to find intermediate systems in which ix(x = t) = t is provable from more fundamental principles. The present paper considers one avenue where one might hope to find them and closes it. Two theories, FDA and FDH, will be considered that are motivated by considerations from proof-theoretic semantics. The former turns out to be equivalent to FD2, the latter to have unreasonable consequences. A third theory, similarly motivated, is also not viable. The philosophical upshot of this is briefly discussed.