Consider the hypothesis that every quantified sentence in every natural language contains some expression or other whose extension constrains the domain of quantification. If the hypothesis is correct, quantificational domains are fixed in fundamentally different ways in natural languages and in standard artificial languages. For those of us who think sentences with different logical forms express different propositions, it would mean that no proposition expressed in a typical formal language is expressible in any natural language. The article begins by clarifying the hypothesis and by arguing that there is no easy way to defeat it—prima facie counterexamples are many, but none of them withstand serious scrutiny. Then the article presents a real counterexample—a hitherto largely neglected reading of quantified attitude-reports where the attitude verb splits the quantifier from its restrictor. The article shows that split readings can occur not only with attitude verbs but with modals, tense, and aspectual markers as well. The upshot of the article is that although there is a real threat to our comfortable assumption that the sort of bare quantification we routinely employ in logic, mathematics, and the sciences can be fully captured in English, there is also reason to think that the threat can be averted.

The text of this article is only available as a PDF.
You do not currently have access to this content.