Quantification and Modality

D. Terence Langendoen

Professor Emeritus, University of Arizona

Linguistics Colloquium

University of Arizona

13 March 2009

Abstract

The close relation between quantification and modality has been noted before. Melvin Fitting & Richard Mendelsohn, First Order Modal Logic, 1998, p. 108, state "The modal operators are like disguised quantifiers". Arnold Koslow, A Structuralist Theory of Logic, 1992, p. 312, states "It seems to be no accident that the quantifiers look very much like modals". In this presentation, I propose, following Koslow, that the first-order universal and existential quantifiers qualify as necessity and possibility modals respectively, but that the favor is not returned -- the standard modals are not quantifiers.

To show this, I first work through some simple examples of alethic (only analytic propositions are necessary) and epistemic (analytic and known synthetic propositions are necessary) modals. I then take up a simple example involving universal and existential quantifiers over one-place predicates that demonstrates why these quantifiers are modal, and how they differ from alethic and epistemic modals.

Next I turn to the question whether there are any other quantificational modals, and show that there are: In a domain with infinitely many entities, such as positive integers, to quantify over, 'for all but finitely many' is a necessity modal and 'for infinitely many' is a possibility modal.

However, other quantifiers, such as 'for all but at most one', 'for at least two', 'for most', 'for many', and many others, fail to qualify as modals, but come sufficiently close that they may be called 'quasimodals'; the ones that "count down" from the universal quantifier (like 'for all but at most one', 'for most') being quasinecessities and the ones that "count up" from the existential quantifier (like 'for at least two', 'for many') being quasipossibilities.

Finally, I ask whether there are any non-quantificational quasimodals, and conclude that there are, including the epistemic quasinecessity 'probably' and quasipossibility 'not improbably'.