Generic Ontology of Linguistic Classification

Abstract

We introduce observational logic as a framework for linguistic classification. Observational theories consist of universally quantified monadic conditionals, where antecedent and consequent are built by finite conjunction and disjunction from primitive monadic predicates plus two predicates expressing existence and non-existence.

Adopting a Quinean conception of ontology, we require any model representing the “generic ontology” of an observational theory to satisfy the condition that indiscernibles are identical. Moreover, such a model should be “as general as possible” with respect to the theory, that is, if two predicates have identical extensions in the model, they should be equivalent in the theory. We investigate whether the latter condition enforces the model to be “as large as possible” provided that indiscernibles are identical. In general, the answer is negative. It turns out that the most problematic elements of the universe are those which neither are finitely specifiable nor can be approximated by finitely specifiable ones. The concluding discussion indicates how generic ontology can help to clarify the nature of linguistic entities. We briefly address issues such as partiality and cognitive adequacy.