Derived contexts: A new argument for their usefulness

Abstract

This paper aims to vindicate the adequacy and usefulness of the notion of derived context (Stalnaker 1988), by showing how it helps to successfully account for phenomena that lay beyond the data that motivated Stalnaker to first introduce the notion. We will focus on phenomena related to Binding Theory (BT) principles (B) and (C) (Chomsky 1981, 1993). These principles account for an impressive array of data. They are subject, though, to well known counterexamples. As part of a more general attempt to defend standard BT, this paper will focus on one specific kind of counterexamples to the BT (some of which have not yet been considered in the literature), and will show how they can be successfully accounted for without any substantive departure from standard BT. All is needed is (1) to take the semantic restriction placed by the Binding Principles to be not that of co-reference (or co-valuation) but rather that of presupposed co-reference (or presupposed co-valuation) (as in Heim 1993), and (2) to appeal in the appropriate way to the existence of derived contexts.