Colloquium for the Philosophy of Science where it was read and discussed. According to Smith, if Kripke and those following him had properly read, understood, and cited Marcus, they would have found a picture with all of the following theses of NTR:
[T1] Proper names are directly referential, rather than being strongly equatable with, or disguised, definite descriptions.
[T2] Just because a definite description has been used to fix the referent of a name, this does not mean that the name goes on to carry the meaning of that description.
[T3] The modal argument refutes descriptivist theories of proper names.
[T4] Statements of identity express necessary propositions (i.e., the necessity of identity).
[T5] Proper names are rigid designators.
[T6] There exist necessary a posteriori truths. (Smith 1995a, 182–86)
From here, we should also survey Soames’ and Burgess’ responses in Soames (1995) and Burgess (1998). Soames’ (1995) primary issues with Smith (1995a) can be broken down into general worries and specific concerns with each of T1–T6. On the more general side, he says that “[Marcus’] entire discussion of the meaning and reference of names, as well as their relations to descriptions, covers only five or six pages [. . .] and defenses of the various doctrines of the so-called ‘new theory’ would not fit into such a small space” (Soames 1995, 193). With respect to T1–T6, individually, Soames says something different about each that he believes should make us question each one of Smith’s attributions to Marcus. In particular, insofar as T1 and T3 are to be found in work by Marcus, they are also to be found in works by Smullyan (1947, 1948) and Fitch (1949). So, all three of them provided “significant anticipations of some of the central theses of contemporary non-Fregean theories of reference,” but nothing that would single out Marcus as a founder of such theories (Soames 1995, 208).
Something similar can be said for Soames’ response to Smith’s claims about T4 and T5—there is a sense in which they contain a kernel of truth, but ultimately need to be qualified significantly. In particular, Marcus can be attributed with an implicit step along the way toward T4 and T5, but she does not anticipate the full-blown theses that Kripke argued for in the arena. To see this, lets first look at T4—the necessity of identity. Here, Soames rightly points out that this is an ambiguous phrase which can mean that:
(1) statements of identity involving variables express necessary propositions (i.e., (∀x) (∀y) (x = y→ x = y)),
(2) statements of identity involving proper names express necessary propositions, or
(3) statements of identity involving singular terms express necessary propositions.
This is important because the necessity of identity with respect to variables and proper names—(1) and (2)—seem correct, while the necessity of identity with respect to singular terms (n.b. contingent definite descriptions) is false (Soames 1995, 202). Furthermore, while Marcus (1947) proved that the necessity of identity with respect to variables is true, she does not discuss the necessity of identity with respect to proper names—the view that Kripke (1980) got to. Given this, we can no more attribute to her the true view—(2)—that there is a necessity of identity with respect to proper names than we can the false view—(3)—that there is a necessity of identity with respect to singular terms, generally. Similarly, with respect to T5, Soames argues that “[Marcus] can be seen, following Smullyan and Fitch, as implicitly treating proper names as rigid” (Soames 1995, 203). That said, without a possible-worlds semantics, Marcus’ early work does not have the notion of a referent at a world and, thus, cannot be said to have any explicit views concerning rigid designation—reference to just one object across all possible worlds.
Finally, Soames’ responses to T2 and T6 are much less congenial. With respect to T2—the idea that just because a definite description has been used to fix the referent of a name, this does not mean that the name goes on to carry the meaning of that description—Soames argues that Smith misunderstands Marcus’ (& Kripke’s) work and cites it problematically. To make his case for this, he begins by pointing out that understanding exactly what is behind T2 is somewhat difficult. Because of this, he spends some time trying to spell out exactly what T2 asserts, ultimately claiming that T2 requires holding that:
(1) “sometimes the referent of a name is semantically fixed by an associated definite description.”
(2) “even though the description is not synonymous with the name.”
(3) “if the referent of a name is semantically fixed by a certain description, then being a competent speaker who understands the name will involve knowing that if the name has a referent at all, it must be one that satisfies the description.”
(4) “if a name N has its referent semantically fixed by a description D, then one who understands the sentence ‘If N exists, then N is D’ will know, without empirical investigation, that it expresses a truth.” (Soames 1995, 199–200)
Unfortunately, Soames claims nothing quite to the level of any of (1)–(4) can be found in Marcus (1961) because “What she was really saying was that despite the fact that recognizing something as a thing presupposes a readiness to apply descriptions to it, nevertheless we have a linguistic device, the proper name, which allows us to refer to a thing without describing it” (Soames 1995, 200).
As for T6, Soames again argues that Smith has failed to understand Marcus’ and Kripke’s views on the matters at hand. In particular, Soames points out that Marcus could not have introduced the existence of the necessary a posteriori because Marcus fails to use “a posteriori” anywhere in her paper and identifies the necessary with the tautologous, with the analytic. Furthermore, Marcus holds that co-referential proper names are intersubstitutable in all contexts. So, if the following is true—“Krista knows a priori that Hesperus = Hesperus”—so too should the next sentence be true—“Krista knows a priori that Hesperus = Phosphorus.” Given this, Soames says that “it is as clear as anything can be that Marcus did not in her 1961 article embrace the notion of necessary a posteriori identities” (Soames 1995, 205).
With Soames’ (1995) detailed responses to the particular theses in Smith’s original paper aside, it will also be instructive to look at Burgess’ (1998) responses. This is because Burgess takes a step back to argue that the paper does not just suffer from problems of detail, but also has major problems in its underlying methodology. More specifically, Burgess “begin[s] by listing ten reasons why Smith’s original paper should never have been taken seriously as a contribution to historical scholarship, ten ways in which the paper conspicuously violates elementary rules of historiographical methodology” (Burgess 1998, 125). Here, I will focus on just the first four of these reasons. This is because Burgess is just straightforwardly correct about the last six. That said, admitting this does not undermine the primary goal of Smith (1995a). Rather, it shows that it needs significant work to clean it up. Furthermore, Burgess himself admits that there is an important difference between the first four and the last six when he says “[a]s a result of the four methodological errors just enumerated, Smith approaches the texts asking the wrong question,” whereas the last six simply cause problems of interpretation in answering those questions (Burgess 1998, 127).
With these preliminaries aside, we move to the list of reasons that Burgess provides for saying that Smith (1995a), the beginning of this saga, never should have happened in the first place. In the order he considers them, the original paper should never have been allowed to be presented because:
[R1] Smith reduces a complex philosophical theory to a list of a half-dozen
