I Introduction

Angle Grinder Man removes wheel locks from cars in London.¹ He is something of a folk hero, saving drivers from enormous parking and towing fines, and has succeeded thus far in eluding the authorities. In spite of his cape and lamé tights, he is no fiction; he's a real person. By contrast, Pegasus, Zeus and the like are fictions. None of them is real. In fact, not only is each of them different from the others, all differ from Angle Grinder Man. After all, Zeus throws thunderbolts but doesn't remove boots from cars; unlike Superman, Angle Grinder Man couldn't leap over a parked Mini, and all sightings suggest that he is a human being, not a horse.

According to the charmingly austere theory of Direct Reference, a proper name's meaning is simply its referent.² Two proper names with [End Page 419] the same referent are synonymous. Assuming compositionality³ , and that, for instance, the proper names 'Hesperus' and 'Phosphorus' are coreferential, S believes that Hesperus is shining iff S believes that Phosphorus is shining; likewise for any agent S and attitude or speech act expression other than 'believes'. Further, a proper name lacking a referent is without meaning.⁴ Whereas 'Angle Grinder Man' is meaningful, 'Superman' and 'Zeus' lack meaning, whence the sentences 'Superman flies' and 'Zeus flies' do not differ in meaning. Some proponents of Direct [End Page 420] Reference hold in addition that the sentences 'Superman flies' and 'Zeus flies' express one and the same "gappy proposition", <___ , flies>. (Braun 1993; Braun 2005) Further, according to at least one articulation of Direct Reference, 'Zeus does not exist' expresses only a gappy proposition and is neither true nor false.⁵

Some of the above consequences may appear to be knock-down refutations of Direct Reference. However, proponents of Direct Reference have attempted to explain away the counterintuitive appearance of these consequences in either of two ways. One I shall call the Ways of Believing Defense, and the other I shall call an Implicature Defense. On the Ways of Believing Defense, different proper names are associated with different ways of believing propositions. Further, one can believe a single proposition in one way while disbelieving it in another with no lapse in rationality. By contrast, on the Implicature Defense, our firm conviction that 'S believes that Hesperus is shining' can differ in truth conditions from 'S believes that Phosphorus is shining,' is due to our conflation of pragmatic implicature with semantic content.⁶ Likewise, our firm intuition that 'Superman flies' can be true even though 'Superman' is empty is due to our conflation of what is pragmatically conveyed by uses of 'Superman' in a speech act, with that name's literal meaning.

The Implicature Defense as it applies to coreferential proper names in attitude and speech act contexts has been challenged in, for instance, Green 1998, and that challenge need not be duplicated here. My focus will instead be on the Implicature Defense as it applies to the behavior of names lacking bearers. It has been claimed (for instance in Adams and Stecker 1994; Adams, Fuller and Stecker 1997) that our firm convictions about the behavior of bearerless proper names can be accounted for in pragmatic terms. This position has been criticized in Everett 2000, 2003, Taylor 2000, and Reimer 2001. Adams and Dietrich 2004 persuasively rebuts the criticisms of Everett, unconvincingly (as we explain below) replies to Taylor, and does not reply to Reimer.

This essay replies to Taylor 2000 on behalf of the Implicature Defense of the [End Page 421] Direct Reference position on empty names, and shows Reimer's criticism of the Implicature Defense of that position to be unpersuasive. I also present a new criticism of the Implicature Defense of the Direct Reference position on empty names in the form of a trilemma: Either (i) it appeals to conversational...