On Problems with Descriptivism: Psychological Assumptions and Empirical Evidence

Abstract

We offer an empirical assessment of description theories of proper names. We examine empirical evidence on lexical and cognitive development, memory, and aphasia, to see whether it supports Descriptivism. We show that description theories demand much more, in terms of psychological assumptions, than what the data suggest; hence, they lack empirical support. We argue that this problem undermines their success as philosophical theories for proper names in natural languages. We conclude by presenting and defending a preliminary alternative account of reference from a developmental perspective.

A Cognitive Theory of Empty Names

“…The semantic assumption is based on empirical research from cognitive psychology and psycholinguistics according to which proper names do not have descriptive semantic content (see Garcia-Ramirez and Shatz 2011). I refrain from appealing to non-descriptive semantic content beyond that of the referent (see Sainsbury 2005).…”

Rigidity and Contingency

“…Some cases involve material constitution as part of the relation between the identity (non-identity) pair, others do not. Because of space limitations I will only present four cases owed to Gibbard (1975), Lewis (1986), and Karmo (1983), but there are plenty more (see Noonan, 1991, and1993;Perry, 1972;Myro, 1986a and1986b; for a general survey see Gallois, 1998;andSchwartz, 2013). Gibbard's [1975:191] example is perhaps the most well known case of contingent identity.…”

“…Yet, this widespread agreement has not precluded philosophers from noticing that the theory has problems with contingent identity (see Gibbard, 1975;Lewis, 1986;Gallois, 1986 and1998;and Stalnaker, 1997). Some have tried to offer an account that fits Kripke's model.…”

“…As I said above, Gibbard (1975), Lewis (1986), Gallois (1986), andStalnaker (1997) have already noticed the threat posed by E 4 and, thus, offer (or suggest) a way to fix the semantic theory. The literature includes only three alternative approaches, Gibbard's (1975) sortal rigidity; Lewis's (1986) quasi-rigidity also suggested by Stalnaker (1997); and Gallois's (1998) restricted rigidity. For reasons that I will present, none of them has gained supporters.…”

“…Let me now consider the last alternative available in the literature. To achieve the needed referential flexibility exhibited by names in statements (4) to (8), Gallois (1986) and(1998) proposes that we relativize all predicates/properties to a world. Thus, the property of being identical to a, for any a, is itself relativized to a world.…”