Frederick Adams
Professor of Cognitive Science & Philosophy
University of Delaware

Abstract

This paper defends a direct reference view of empty names, saying that empty names literally have no meaning and cannot be used to express truths.  However, all names, including empty names, are associated with accompanying descriptions that are implicated in pragmatically imparted truths.  A sentence such as “Vulcan doesn’t exist” pragmatically imparts that there is no tenth planet.  This view is defended against objections.
 

1. Introduction
Theories of direct reference maintain that names contribute their bearers to propositions expressed by sentences or utterances on occasions of use.  On these theories, names contribute only their bearers and have no surplus meaning.  The sentence “Wittgenstein taught school in Kirchberg” expresses the proposition identified by the ordered pair <Wittgenstein, having taught school in Kirchberg>.  Whether theories of direct reference are correct is a matter of dispute.  Kripke’s (1972) arguments against descriptive theories of the meanings of names were persuasive, but descriptive theorists have persuasively fought back (Stanley, 1997).  In this paper, I will not attempt to settle the larger dispute between the two opposing camps, but will focus on the matter of what a direct reference theorist ought to say about names that lack bearers.  In order even to be a contender in the larger dispute between referentialists and descriptivists, theories of direct reference must have a plausible story to tell about the use of names that name nothing in utterances and sentences both positive and negative.
 Positive uses of empty names include “Santa Claus lives at the North Pole.”  What proposition is expressed by this sentence?  Does the sentence express a truth or a falsehood?  Negative uses of empty names include “Santa Claus does not exist.”  What proposition is expressed by this sentence?  Is this sentence true or false?  Until direct reference theorists have a plausible account of such sentences and the role of names in them, direct reference theories are not even in the game (so to speak).
 In what follows, I will sketch an account of the role of names that lack bearers and of their contribution to what is expressed on both their positive and negative use.  I will offer a theory that is consistent with theories of direct reference and a theory that claims that names make the same contribution on an occasion of use regardless of whether they actually name something or not.  I then will defend the view from recent criticisms.
2. The View
On the view that I like (Adams & Stecker, 1994), names that name nothing mean nothing and say nothing.  Lacking referents, their use in sentences or utterances, convey no bearer into the propositional structure that may be expressed.  This is true whether employed in positive or negative sentences.  So “a is F” expresses < __, being F>, and “a is not F” expresses <___, not being F> when “a” names nothing.  There are good reasons (Braun, 1993) to think that propositional structures are expressed by such sentences or utterances, and there are equally good reasons to think that neither sentential form expresses a truth (Adams & Stecker, 1994).  This becomes particularly troubling for sentences such as “Santa Claus does not exist” or “Vulcan does not exist” because these negative existential sentences surely seem to be true.  Yet, if the above account is correct, they express something of the form <__, non-existence> (or better, there is not an x such that x = ___, where one cannot fill the blank).  Thus, on the above account, negative existential sentences employing empty names do not express truths (or falsehoods).  This plainly seems wrong, (at least at first).
 To make the view more palatable, it must explain away the intuition that negative existential sentences employing empty names express truths.  It does so by finding something true that, although not literally semantically expressed by such sentences, is pragmatically imparted.  It does this by appealing to two mechanisms: association and pragmatic implication.  When we hear new names we cognitively place them into a mental file or dossier (Recanati, 1993).  We associate descriptions with those names, but the descriptions do not give the meanings of the names (as they are contingently associated with the names through a process of mental association—just as the empiricist, associationist philosophers and psychologists have long maintained).  If we acquire the name “Ludwig Wittgenstein” we may associate the descriptions “author of the Tractatus” or “famous student of Lord Russell.”  These descriptions help us cognitively track the individual named.  When we say “Wittgenstein was a philosopher” we may impart or imply that a famous student of Lord Russell was a philosopher, but this is not literally expressed by our utterance (what is literally expressed is <Wittgenstein, being a philosopher>).  Still it is information that is there to be conveyed and would be conveyed in a community of speakers who typically share the same cognitive associations of descriptions and names.  The mechanism of pragmatic conveyance is Gricean (Grice, 1989) and is well known.  Hence, we will help ourselves to the distinction between what a sentence literally semantically expresses and what its utterance can pragmatically convey.
 These associations take place whether names are empty or not.  When children acquire the name “George Washington” they typically acquire the associated descriptions “first president of the U.S.”, “man who had wooden teeth.”  When they acquire “Santa Claus” they associate with the name “jolly fat man who lives at the North Pole” or “man who brings presents on Christmas day.”  The negative existential “Santa Claus does not exist” expresses no literal truth.  Still it pragmatically imparts truths: truths such as that there is no jolly fat man who lives at the North Pole and wears a red suit or that there is no one who brings presents to the world’s children on Christmas Day.  These things, though not literally semantically expressed by the negative existential sentence, are true.  They include information that may be pragmatically conveyed by the utterance of the negative existential sentence.  It is these truths that we mistake for the literal truth of the sentence “Santa Claus does not exist” when we take that sentence to be true.  If this account of the role of names is correct, the sentence itself is strictly truth-valueless, despite our intuitions to the contrary.  The theory explains away contrary intuitions, in the ways just indicated (Adams, Stecker & Fuller, 1992), and can be extended to fictional names (Adams, Fuller & Stecker, 1997).  The account is unified, attributing to names the same kind of semantic contribution in all contexts whether the names are filled or empty.  Names contribute their bearers on occasions of use (if they have bearers) and otherwise make no semantic contribution at all.
3. Objections & Replies
In this section I will elaborate the theory by considering some recent objections and replies.  The first objection is that different sentences using different empty names seem to say different things, but on this account they would all say the same thing, so the theory must be false (Everett, manuscript a).  Consider “Santa Claus doesn’t exist,” “Father Xmas doesn’t exist,” and “Superman doesn’t exist.”  The first two may say similar things, but the latter does not.  However, all three should say that there is not an x such that x = ___ (where one cannot fill the blank).  That is, they should all say the same thing on our view.
 The reply is that they all do literally say the same incomplete negative existential propositional structure—one lacking a truth-value, but they do not all say it in the same way.  They express it using different names.  In that, the situation is not all that different from “Fa” and “Fb” where a = b.  They express the same proposition but do so using different names.  The main difference is that, in cases of empty names, no complete propositions are expressed.  Similarly, with filled names we may associate different descriptions.  With “Tony Curtis” we associate “famous film star and father of Jamie Lee.”  With “Bernie Schwartz” we may associate “gray-haired gentleman dressed in black.”  Tony Curtis = Bernie Schwartz.  Yet saying “Tony Curtis is here” may appear to say something different than “Bernie Schwartz is here” because of the different associations (in the minds of those not in the know).
 Similarly, we may associate the same descriptions with “Santa Claus” and with “Father Xmas,” but not with “Superman.”  The differences in descriptions are due to differences in the lore surrounding the religious and holiday figures versus the comic strip hero.  But any apparent differences of expression are due to the differences of associated descriptions and not due to difference of literal, semantic content expressed.  “Santa doesn’t exist” pragmatically imparts, but does not literally semantically express that no jolly fat man in a red suit who brings presents to the world’s children exists.  Perhaps “Father Xmas doesn’t exist” pragmatically imparts the same thing (if it has the same associations).  But “Superman doesn’t exist” pragmatically imparts that there is no man of steel.  It imparts other things because of different associated descriptions with “Superman” (though these do not give the meaning of the name).  That accounts for the apparent differences.  Therefore, our theory is not falsified by these apparent differences.
 Another, more serious objection is that our theory cannot explain the modal profile of sentences that contain empty names (Everett, manuscript a).  The sentence “Santa is identical to John Perry” seems not only to express a falsity, but a necessary falsity.  Yet on our view, it expresses the truth valueless propositional structure <____ = John Perry> (and it does so using the name ‘Santa’).  If we try to account for the merely apparent falsity of the sentence by relying on associated descriptions, we may pragmatically impart something such as that the jolly fat man in the red suit living at the North Pole who brings presents on Christmas day is John Perry.  This is not literally what the sentence above expresses, but it may be pragmatically imparted, as per our account.  However, this proposition, while false, is at best contingently false.  Thus, our view cannot explain the apparent necessity, as well as apparent falsity of the above type of sentence.
 In reply, our view can explain both the apparent falsity and necessity, but does so in different ways.  The apparent falsity is already explained nicely above.  The above sentence pragmatically imparts that John Perry is identical with the jolly fat man in the red suit living at the North Pole and who brings presents on Christmas day.  This is false, even though it is only pragmatically implied (not literally expressed).
It is true that it is only contingently false.  However, this proposition is not what explains the apparent necessity.  That is explained by our semantical intuitions.  When we see an expression of the form “a = b,” we know that it is either a necessary truth or a necessary falsehood (if it expresses a genuine proposition at all).  We know this because we know (Kripke, 1972) that all identities are necessary.  “Tony Curtis = Bernie Schwartz” expresses something necessary.  We know this even before we learn that it expresses a truth.  “Tony Curtis = John Perry” expresses a necessary falsehood, and we know of the necessity by the semantic structure of the identity sentence even prior to knowing that it is false.
 We know that “Santa” is not a name for John Perry.   So it is not like the first identity sentence above.  This will incline us to think it is like the second.  Unfortunately, “John Perry = Santa” does not literally express a truth or a falsehood, but if it did, it would express a necessary falsehood, and we know this via our semantical modal intuitions alone.  That is how the appearance of necessity is explained (away).  This is a different explanation than that of the appearance of falsity, but an explanation nonetheless.
 I will close by considering the view that empty names are descriptive names (Kripke, 1972, Evans, 1982).  Suppose we associate “the inventor of the zip” with the name “Julius.”  Then suppose there was no single inventor of the zip.  “Julius does not exist” might be taken to express the proposition that there is no inventor of the zip.  This would be literally true, not merely pragmatically imparted.  So it may appear to do a better job than our account of handling empty names.
 I think this view should be rejected.  First, if  “Julius” names nothing because the description “the inventor of the zip” fails to pick something out, then “Julius does not exist” still expresses the truth valueless propositional structure that there is not an x such that x = __ (where one picked out by “inventor of the zip” is to fill the blank).  Still there is no name to fill the blank, and there is no advantage over our view.  If one says the proposition expressed is that there is no inventor of the zip, then “inventor of the zip” would be giving the meaning of “Julius.”  Empty names would have descriptions as their meaning (contrary to filled names).  There then would be no uniform contribution that names make, whether filled or empty.  Second, suppose there were a single inventor of the zip, viz., Ken.  Then “Julius is smart” should express that Julius (aka Ken) is smart, but on this alternative view it would express that the inventor of the zip is smart (and not act like a name at all).  Finally, Recanati (1993) maintains that “Julius” is a name and once attached to its bearer (aka Ken), we could truly assert that Julius didn’t invent the zip.  But on the view where “inventor of the zip” gives the meaning of “Julius,” “Julius does not exist” would express that the inventor of the zip didn’t invent the zip.  Not good!

References
Adams, F. Stecker, R. & Fuller, G.  1992.  “The Semantics of Thought,” Pacific Philosophical Quarterly, 73, 375-389.
Adams, F. & Stecker, R. 1994.  “Vacuous Singular Terms,” Mind and Language, 9, 387-401.

Adams, F., Fuller, G. & Stecker, R. 1997.  “The Semantics of Fictional Names,” Pacific Philosophical Quarterly, 78, 128-148.

Braun, D. 1993.  “Empty Names,” Nous, 27, 449-469.

Evans, G. 1982.  The Varieties of Reference.  Oxford: Oxford University Press.

Everett, A. 2000.  “Referentialism and Empty Names.”  In Everett & Hofweber, 37-60.

Everett, A.  Manuscript a.  “Empty Names and ‘Gappy’ Propositions, “ (forthcoming in Philosophical Studies).

Everett, A. Manuscript b.  “Empty Names, Truth Conditions, and Propositions.”

Everett, A. & Hofweber, T. (eds.) Empty Names, Fiction, and the Puzzles of Non-Existence.  Stanford: CSLI Press, 2000.

Grice, H. 1989.  Studies in the Ways of Words.  Cambridge, MA: Harvard University Press.

Kripke, S. 1972.  “Naming and Necessity.”  In Davidson D. & Harman, G. (eds.), Semantics and Natural Language.  Dordrecht: Reidel.

Recanati, F. 1993.  Direct Reference: From Language to Thought.  Oxford: Blackwell.

Stanley, J. 1997.  “Names and Rigid Designation.” In Hale, B. & Wright, C. (eds.), Companion to the Philosophy of Language.  Oxford: Blackwell Press, 555-585.