meaning of a local name is “known” only to the people in the locality, assuming like Locke that we can only know it through acquaintance with the bearer. Such knowledge will be unavailable to the greater part of the community that uses the language. There are a few proper names that are understood by the whole community (he mentions “the Sun” as an example, but presumably “Caesar” and “Moses” would do), but that is because the bearers are known to the whole community.29
These explanations are inadequate. In what sense are the very individuals Caesar and Moses “known” to the whole community that understands their names? How exactly are we acquainted with them? How is it we understand names like “Frodo” and “Sherlock Holmes” that have no bearers, so that there is no possibility of knowing or being acquainted with the bearers? I shall argue that our knowledge begins and ends with the text that introduces those names. If the texts (The Gallic wars, The Lord of the Rings, etc.) are available to the whole community, the whole community is able to understand the names. Acquaintance with the text is both sufficient and necessary. No acquaintance with the individual is necessary, nor is it even sufficient. If I am confronted with Caesar, I do not know that this individual conquered Gaul, unless I am know that that this person is Caesar.
Hence, if our knowledge begins and ends with the text that introduces names such as “Caesar,” “Moses” and “God,” we cannot understand the proper name “God” without acquaintance with the Hebrew Bible, or some text that co-refers with it, and so cannot have a singular conception of God.
Organization of the Book
In the second chapter, I defend the co-reference thesis, arguing that there is a natural reading of a text to which all authors will try to conform, not always successfully, and this natural reading determines the co-reference of the singular terms occurring in the text. There is evidence for systematic rules or heuristics known to both author and reader, and which in principle suggests the possibility of mechanical systems or algorithms for determining co-reference. A common reference is an objective property of the text. No acquaintance or knowledge of a bearer is necessary. I claim that, in principle, a computer could analyze the text of The Lord of the Rings to determine the set of terms, which co-refer with “Frodo Baggins,” or with “Gandalf,” and so forth.
In the third chapter, I introduce the dependency and the reference theses through the puzzling phenomenon (sometimes called the puzzle of unbound anaphora) of proper names introduced by indefinite description, such as “a man named Moses” and “a woman named Martha.” It can be shown that these proper names co-refer in some sense with their indefinite antecedent, yet the antecedent itself does not refer. I claim this is because the meaning of a singular term is non-transportable:30 it co-refers with some indefinite antecedent, but the whole purpose of an indefinite term is its indifference to what came before in the order of reading, hence no singular term that occurs before the antecedent can refer forward to it in the way that a subsequent anaphor can refer back. No term prior to the antecedent can have the meaning of an anaphor term subsequent to it.31 This, and not Locke’s conception of acquaintance, is what distinguishes general terms, which are transportable, from singular terms, which are not. Singular anaphor terms are semantically dependent on the chain introduced by some indefinite antecedent, and so dependent on the text itself. They have no general dictionary-defined meaning, which we can understand as part of learning the language. Our understanding begins and ends with some text. This is the dependency thesis.
The reference thesis is that we can make a reference statement in a text where the referring term of that statement has guaranteed co-reference with the mentioned term, so that the term must itself have a non-transportable meaning. We cannot specify the reference of a term used in a narrative without that specification itself being a part of the narrative.
In chapter 4, I address the question of co-reference between texts that are not independent, that is, when a text such as the New Testament mentions or cites the Hebrew Bible, and I distinguish between citation and reference.
Chapter 5 concerns implied or imputed co-reference between independent texts, where it is not signified that the same person is a character in both texts, but where we can reasonably infer that it is the same person, or where we could write a history using information from both texts, where the co-reference was signified. For example, the identity we assume between the man called “Moses” in Exodus and the man of the same name in Numbers or Deuteronomy is signified. If the men were different, the author would have said so, therefore to understand the text at all, we have to understand the identity. But the identity we assume between the man called “Pilate” in Philo’s De Legatione ad Gaium, and the man of the same name in the Gospel account, while fairly certain, is merely probable. The identity is not part of the meaning of the separate accounts, since they were independently written by different authors. It is not as though Philo would tell us whether the man he calls “Pilate” is or is not the same as the man of the same name in the Gospels. He was almost certainly unaware of the Gospels. Likewise, the author of the Gospels does not mention Philo’s text.
I argue that history involves taking separate accounts such as this—perhaps very short accounts contained in baptismal records and other kinds of register—and imputing co-reference by converting the probable identity into semantic identity. Thus, in writing “Pilate was ordered by Tiberius to remove the shields from the palace in Jerusalem” and later “he handed Christ over to be crucified,” the historian converts a probable identity between the man who Philo said removed the shields and the man who the Gospel said handed Jesus over, into a semantic one, via the pronoun “he.” I take up the philosophical implications of such probable (i.e., contingent) identity in the following chapter.
Chapters 6 and 7 concern two corollaries of the co-reference thesis that conflict with standard semantics. Chapter 6 considers, and rejects, two contemporary philosophical theories of reference. The first is strong Millianism or “direct reference,” which holds that the proposition, the bearer of truth and falsity, is not a linguistic item, a sentence capable of truth and falsity, but a Russellian proposition: an extralinguistic item expressed by an assertoric sentence. According to this theory, the semantic value of a proper name is the bearer itself: the proposition expressed by “Peter preached in Galilee” contains both Peter and Galilee. I argue that the semantic value of a proper name consists solely in its (intralinguistic) co-reference with its anaphoric antecedents, and that because its meaning is not transportable, it can have no semantic connection to its bearer. Thus, identity statements flanked by proper names can be informative when the proper names do not co-refer. Such an identity is not like “catsup is ketchup,” where the transportable meaning of the two terms is identical.
The second theory, following from the first, is that identity statements involving proper names, if true, are necessarily true. It was first stated by Ruth Barcan Marcus in 1947, becoming widely known and accepted after Saul Kripke’s defense of it in Naming and Necessity. I reject the theory by rejecting the principle of substitution on which it depends. Even if “Cicero” refers to Cicero, and Cicero is the same person as Marcus, it need not follow that “Cicero” refers to Marcus.
Chapter 7 is about existence. The co-reference thesis implies that proper names are meaningful even when empty. A proper name acquires its meaning through guaranteed co-reference alone, and so does not need a bearer in order to have a meaning. Hence, the fideist and the atheist can meaningfully disagree, given that “God does not exist” denies a referent, but not a meaning for the name “God.” However, the thesis involves two logical problems. First, “God does not exist” implies “something does not exist” in standard logic, and “there is no such thing as God” implies “there is no such thing as something.” Second, it follows from the reference thesis that “‘God’ refers to God” is true, even if “God” has no referent, that is, even if God does not exist. Yet standard logic requires that “a R b” implies “for some x, a R x.”
I
