MARCUS AND THE NEW THEORY OF REFERENCE:

Published in: Synthese, Volume 104, No. 2, August 1995, pp. 217-244. Reprinted in (eds. James Fetzer and Paul Humphreys), The New Theory of Reference: Kripke, Marcus and Its Origins Kluwer Academic Publishers: Synthese Library Series, 1998, pp. 37-61.

We are engaged in an inquiry in the history of philosophy. Specifically, we are inquiring into the historical origins of the New Theory of Reference. Relevant to this inquiry are arguments about the meaning of texts by Marcus, Kripke and others. I do not believe it is relevant or helpful to adopt the sort of language that Soames uses in his reply, when he sees fit to call my work “careless, incompetent, grotesquely inaccurate, shameful, scandalous, and irresponsible”. Philosophical disagreements are not solved by the disputants labelling each other’s work with a variety of negative and emotive epithets; they are solved by presenting sound arguments, and I shall confine myself to presenting argu ments in my reply to Soames.

In my original paper I discussed seven ideas that Marcus introduced into the New Theory of Reference. Soames took issue with each of these seven claims, and I shall examine each of his seven criticisms in turn. (In my original paper, I talked of six ideas, since I included the modal and epistemic arguments for direct reference under one heading, but here I distinguish them for the sake of increased clarity.)

But first I should note that Soames does not quote Marcus’ original 1961 paper, which is the paper in dispute, but her revision of this paper, published in Modalities (1993), which differs at a number of places from the original 1961 article. I quoted only the 1961 article, but Soames, when he sometimes tries to show that my quotations are misleading or incomplete, does not in fact quote the 1961 article, but Marcus’ 1993 revision of this article. Since this is an historical inquiry into the relation of Marcus’ 1961 paper to Kripke’s works of 1971 and 1972, it is her 1961 paper that is the proper object of examination.

Before I examine Soames’ specific criticisms of my seven arguments, I should make an observation about a general claim that Soames made. He said that Marcus could not possibly be the originator of the main ideas of the New Theory of Reference, since she presented her ideas in too few number of pages. Soames writes about my claim about Marcus: “One clue that this claim cannot be correct comes from the observation that proper statements and defenses of the various doctrines of the so called ‘new theory’ would not fit into such a small space” (p. 193, my italics). But surely the following argument is invalid: if an idea takes up only three pages in Marcus, but is spread with numerous illustrations and repetitions over thirty pages in Kripke, that fact is sufficient to deprive Marcus of priority.

By comparison, F.P. Ramsey is generally credited with priority for the Dutch book argument for justifying the axioms of probability on the personalist or subjectivist interpretations of the axioms. But this accreditation is based on exactly two sentences on page 182 of his essay “Truth and Probability” (Foundations of Mathematics, Routledge and Kegan Paul, 1931). De Finetti developed this argument much more elaborately several years later (over 30 pages in Fundamenta Mathematica, Vol. 17, pp. 298 but this does not erase the fact of Ramsey’s priority. Indeed, de Finetti acknowledged that Ramsey originated the argument. Similarly, Marcus presented in a very concise fashion almost all the main ideas and arguments of the New Theory of Reference in her 1961 article; Kripke’s extensive elaboration upon her ideas, with the introduction of numerous illustrations and a new terminology, does not count against Marcus’ priority, as I shall now argue in detail.

I argued that the first idea Marcus introduced into the New Theory of Reference is that proper names are directly referential and not equivalent to definite descriptions. Soames admits that Marcus did present the idea that proper names are directly referential and not equivalent to definite descriptions, but claims that Marcus’ presentation of this idea was antedated by Fitch and Smullyan. Soames writes:

Although Marcus did regard names as directly referential, and did not take them to be disguised descriptions, she did not introduce the idea; nor did she claim to. As we have seen, Smullyan and Fitch both invoked that idea in earlier responses to Quine that Marcus took herself to be repeating and elaborating. (pp. 198-9).

The textual evidence does not support Soames’ claims about Fitch and Smullyan. In an article that Soames does not mention, Fitch’s article ‘Attribute and Class’, published in 1950 in Philosophical Thought in France and United States (edited by Farber, with a preface written in 1949), Fitch explicitly attributes the idea that names and descriptions behave differently in modal and other contexts to his doctoral student Ruth Barcan Marcus (her name was then Ruth Barcan). Fitch writes:

The system of modal logic developed by Ruth Barcan suggests that the simplest view is that no identities should be regarded as merely contingent and that identified entities should be everywhere intersubstitutable ... Furthermore, if entities X and Y have been identified with each other, it seems reasonable to suppose that the names of X and Y should also be everywhere intersubstitutable where they are being used as names. According to Church’s view, on the other hand [ is distinct from Ruth Barcan’s view] two names of the same thing might differ in sense and so not be intersubstitutable. (p. 252, edited Farber 1950, Buffalo.)

This passage confirms what Marcus told me in an interview, that she developed this view of names and descriptions in course of writing her thesis in 1943—45 and explained them in conversation to her dissertation adviser Fitch. This is in keeping with the passage from Fitch I just quoted, where Fitch mentioned these ideas and attributes them to Marcus. Notice that he is not advancing these ideas as his own ideas; rather, he is explaining Marcus’ ideas and contrasting them with Church’s ideas as two alternate views of the behavior of names in intensional contexts. Fitch refers again to his indebtedness to conversations with Marcus in his paper on the ‘Problem of Morning Star and Evening Star’, where he says “The writer wishes to acknowledge his debt to Miss Barcan for some helpful discussions concerning the ideas of this paper”. (Philosophy of Science, Vol. 16, 1949, p. 139, n. 6). Thus, Fitch received — and admitted receiving — the idea about the distinctive behavior of names from Marcus and it is false that Marcus repeated and elaborated on Fitch’s original idea.

Soames also claims that Smullyan earlier developed the idea that names are directly referential and are not disguised descriptions. Soames is also mistaken on this point. He refers to Smullyan’s 1948 article on ‘Modality and Descrip tion” in The Journal of Symbolic Logic and his 1947 review (in the same journal) of Quine’s article on modal logic. The reference to the 1948 article on ‘Modality and Description’ is a red herring, since Smullyan there discusses the issue of whether the modal paradoxes suggested by Quine can be resolved by treating the relevant expressions, not as names, but as descriptions or class abstracts. For example on page 35 he says that:

Our discussion of the argument, ABC, may seem to have depended, at certain crucial points, upon our decision, to interpret the expression, ‘the number of the planets’ as a definite description. What would have been the situation had the phrase in question been interpreted as a class abstract, as synonymous with Ж (x is equinumerous with the set of planets)’ for example?

Smullyan also goes on to show how the modal paradoxes can be avoided if the expression is interpreted as a class abstract. In this paper, Smullyan discusses only descriptions and class abstracts; names are not discussed at all, contrary to Soames’ contention.

Smullyan’s 1947 review of Quine’s paper, is, however, relevant, but even in this paper there is only one sentence about names — and even this one sentence Smullyan rejects as false. Furthermore, this sentence (even apart from the fact that Smullyan rejected it as false) does not even imply that names are directly referential and are not disguised descriptions. The sentence in question is on page 140. Smullyan writes:

if ‘Evening Star’ and ‘Morning Star’ proper-name the same individual they

(B) Evening Star is congruent with Evening Star and it is not necessary that Evening Star is congruent with Morning Star.

Soames’ gloss on this sentence is that “Smullyan adds to this claim that coreferential names are synonymous, which strongly suggests that genuine names have no meaning apart from their referents, another view later endorsed by Marcus” (p. 196).

I think Soames is misinterpreting Smullyan’s remark. If coreferential names are synonymous, that does not mean, imply or even suggest they are directly referential. And if coreferential names are synonymous in modal contexts, such as Smullyan’s statement B, that still does not mean, imply or suggest they are directly referential. For it is perfectly consistent with the claims that they are indirectly referential, that they express a world-indexed sense, such as Linsky and Plantinga claimed. “Evening Star” and “Morning Star” could both express the sense “whatever is actually the planet Venus”. Smullyan shows no awareness whatsoever of this distinction between direct reference and world- indexed senses, and it would be a mistake to read an awareness of this distinction, and an advocacy of direct reference rather than world-indexed sense, into his one sentence remark about synonymity. Further, Soames’ suggestion that Smullyan advocated this idea about the expressions being synonymous names is false, since Smullyan in the very next paragraph goes on to reject the idea that “Evening Star” and “Morning Star” should be regarded as synonymous names in modal contexts. He instead says that they should be regarded as descriptions. Smullyan writes: “On the other hand, if, more naturally, we view “Evening Star” and “Morning Star” as abbreviations of descriptive phases, we find that A expresses an evidently impossible proposition” (p. 140, my italics), where A is another modal proposition.

Thus, Soames’ claim that Marcus’ theory that ordinary names are directly referential and not disguised descriptions was first espoused by Fitch and Smullyan is an incorrect claim. The idea is original with Marcus. (Of course, we are talking about the contemporary New Theory of Reference; as everyone knows, Mill held a direct reference theory in the 19th century.) We should also point out that some others claimed in passing that ordinary proper names are directly referential, such as Paul Fitzgerald, but they did not support their claims with the arguments about the nonequivalence of names and descriptions in modal and epistemic contexts that have since become part of the New Theory of Reference.

My second claim was that Marcus introduced the idea that entities are singled out by definite descriptions, but that assigning the entities a name is semantically different than describing them. Marcus’ view is that singling out an entity by unique descriptions is a precondition of language, and that singling out entities this way is the only way to single them out (Marcus, 1961: 309). It follows that when we assign a name to a thing that we have singled out, we can assign the name only on the basis of a unique description that we are using to single out the thing. Soames claims that this does not imply that a specific description is linked as a matter of semantics to the name, and thereby that Marcus did not originally state the idea of reference-fixing descriptions later stated by Kripke. But Soames here appears to misunderstand Marcus and the concept of a reference-fixing description. The crucial premises in Soames’ argument are that (quoting from Soames’ paper):

…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. Similarly, 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. Finally, if one later finds that the description one associated with the name fails to pick out anything, or fails to pick out the individual one has in mind, then it will follow that the name failed to refer, or at least failed to refer to the individual one had in mind. None of this is remotely suggested in the quotation from Marcus (pp. 199—200).

None of this is remotely suggested in the quote from Marcus since these ideas are neither Marcus’, nor Kripke’s, nor ideas that are consistent with the concept of a reference-fixing description in the New Theory of Reference. Soames is assuming that if a name N has its referent semantically fixed by a description D, then certain conditions must be met that violate the epistemic argument for direct reference that Marcus first put forth and Kripke, Salmon and others later repeated. According to the New Theory of Reference, a description may serve to originally fix the direct referent of a name, but the description is not everafter associated with the name by competent speakers of the language. The a priori association need be present only on the original reference-fixing occasion.

Soames’ second claim (viz., that if one later finds that the description one associated with the name fails to pick out what one had in mind, then the name will fail to refer) is inconsistent with the fact that descriptions can be used in a referential as well as an attributive way, as Keith Donnellan first pointed out. Even if we reject Donnellan’s formulation of the notion of a referential definite description, as does Kripke (‘Naming and Necessity’, note 37, pp. 349), it is still the case that the reference of a name can be fixed by a description that fails to apply to the named object. Soames’ contrary claim is directly contradicted by Kripke in footnote 34 of ‘Naming and Necessity’ (p. 348), where Kripke writes:

“Following Donnellan’s remarks on definite descriptions, we should add that in some cases, an object may be identified, and the reference of the name fixed,

using a description which may turn out to be false of its object. The case where the reference of ‘Phosphorus’ is determined by ‘morning star’, which later turns out not to be a star, is an obvious example. In such cases, the description which fixes the reference clearly is in no sense known a priori to hold of the object ...“.

Thus, Soames’ two conditions for a reference-fixing description are incon sistent with this notion as it is understood by Kripke, Marcus and other proponents of the New Theory of Reference. Further, he ignores the argument in Marcus’ paper that names do have their reference fixed by descriptions. The argument is easily extracted from the theses advocated by Marcus. Marcus advances these theses:

(4) When we first assign a name to an entity, we do so by means of singling out the name’s referent by a unique description.

Thus, I conclude that Soames’ attempt to show that Marcus did not introduce the second idea into the New Theory of Reference is not successful.

Soames agrees with me that Kripke’s modal argument for direct reference was earlier stated by Marcus in her 1961 paper and that Marcus should have been credited for this. But Soames erroneously claims that Fitch and Smullyan offered the modal argument before Marcus. I have already quoted the passage from Fitch where he says this argument was suggested to him by his student Marcus. As for Smullyan, he did not present the modal argument that proper names are directly referential. The nearest he gets to this argument is to mention (but then reject) the claim that if “Evening Star” and “Morning Star” are synonymous, and name the same individual, then it is false that it is not necessary that the Evening Star is congruent with the Morning Star.

Several comments can be made about this sentence in Smullyan’s review of Quine’s article. First, Smullyan is not advancing the modal argument for direct reference. The modal argument is that names are co-referential in modal contexts and therefore are directly referential. But Smullyan instead says that names are co-referential in modal contexts and therefore are synonymous, which is a different argument altogether. Furthermore, as I have already pointed out, Smullyan does not endorse, but in fact rejects, the argument for synonymity he presents. After mentioning the idea that “Evening Star” and “Morning Star” are proper names and therefore synonymous in modal contexts, he goes on to reject this view. He says “On the other hand, if, more naturally, we view ‘Evening Star’ and ‘Morning Star’ as abbreviations of descriptive phrases ...“ (p. 140, my italics). In sum, Smullyan did not endorse or even mention the modal argument for direct reference. It was only Marcus who argued that proper names in ordinary language are directly referential, and who used their behavior in modal contexts to show this. This idea is original with Marcus. I quoted her argument in my original paper and I will not quote her again: in any case, Soames does not dispute the fact that Marcus presented this argument. (I should add the qualification made in my original paper ‘Marcus, Kripke, and the Origin of the New Theory of Reference’, namely, that the modal argument shows only that proper names do not express contingent descriptive senses; it does not show that proper names do not express modally stable senses. The epistemic argument, discussed below, is required to refute the Linsky-Plantinga “modally stable sense” theory of proper names).

My reply to Soames was written in response to a version of his paper which I received in mid-December, 1994, and Soames later added some new material to this version; the new version is the one he read at the APA colloquium and that is printed in this issue of Synthese. His new material consists of the long paragraph on page 202 beginning with “I should mention that, historically”, the four sentences beginning with “There are two problems here” on page 201, footnotes 26, 34 and 36, and several minor wording changes and connections elsewhere in the paper. Here I shall partially respond to footnote 34, which is on the modal argument; in Section 8 I will respond to other aspects of footnote 34.

In footnote 34, Soames argues that there “is an important feature of Kripke’s version of the modal argument that does not seem present in Marcus’ version. Both Kripke and Marcus use modal properties of names to distinguish them from descriptions” (p. 231, n. 34). Soames goes on to suggest that Kripke’s version of the modal argument goes like this:

(2b) Co-referring ordinary proper names are intersubstitutable in modal contexts.

Soames also suggests that Marcus’ version of the modal argument goes like this:

(5c) Co-referring ordinary proper names are intersubstitutable in modal contexts.

This added footnote indicates that Soames has not correctly identified the modal argument for direct reference. Neither argument (A) nor argument (B) are modal arguments for direct reference. The conclusion of the modal arguments for direct reference is proposition (4), which Soames makes a premise of argument (B) and which he omits altogether from argument (A). Soames’ decision to use the phrase “modal arguments” to refer to (A) and (B) is idiosyncratic and does not conform to how this phrase is used by New Theorists of Reference. For example, Nathan Salmon writes that:

A number of arguments have been advanced in favor of the central thesis of the theory of singular direct reference. Although the arguments are many and varied, most of them may be seen as falling under one of three main kinds: modal arguments, epistemological arguments, and semantic arguments ... The modal arguments are chiefly due to Kripke. [ Reference and Essence, Princeton University Press, 1981. PP. 23—24]

there is a weakness in the modal arguments; they show only that names and indexical singular terms are not descriptional in terms of the simple sorts of properties that come readily to mind, properties like the authorship of a certain work ... But faced with the modal arguments, some descriptional theorists, such as Linsky [ p. 84] and Plantinga [ have moved to fancier descriptions employing modally indexed properties ... Thus the modal arguments seem ineffective against the thesis that proper names are descriptional in the Linsky way. [ pp. 26—27]

By distinguishing argument (A) from argument (B), and claiming (A) is present only in Kripke’s writings and (B) present only in Marcus’ writings, Soames is in effect changing the subject, which is whether Marcus first stated the modal argument that proper names are not descriptional but directly referential. And that Marcus first stated this argument is evident from my quotations in ‘Marcus, Kripke, and the Origin of the New Theory of Reference’, pp. 182—3, for these quotations show that she argued that co-referring proper names are intersubstitutable in modal contexts and therefore are not equivalent to contingent descriptions. (Of course, we need the caveat that the modal argument for direct reference only shows that ordinary proper names are not equivalent to contingent descriptions, as Salmon, Plantinga, Linsky and others have recognized.)

The fourth argument I claimed Marcus introduced into the New Theory of Reference was the necessity of identity. Soames’ reply is that in Marcus’ 1946—47 papers, she proved that the necessity of identity involved only variables, and there is no direct route from this to the view that the theorem of the necessity of identity also applies to proper names. But Soames’ reply is unsuccessful in two respects.

First, he is incorrect in assuming that it is an obvious fact that there is no direct route from variables to proper names. There arguably is a formal equivalence between variables and proper names, an equivalence that has been recognized as early as 1939 with Quine’s article, ‘Designation and Existence’ (Journal of Philosophy, 1939), where Quine writes:

…we might equivalently omit express mention of existential generalization and describe names simply as those constant expressions which replace variables and are replaced by variables according to the usual logical laws of quantification. (Vide, Readings in Philosophical Analysis, editedby Feigl and Sellars, 1949, p. 50.)

Thus, Soames is making a dubious claim when he says that it is not clear that his (NI-i) (about variables) goes to his (NI-2) (about names) rather than to his (NI-3)) (about any singular descriptions).

But more importantly than this point about variables and proper names is the fact that Marcus herself applied the theorem of the necessity of identity to proper names in her 1961 paper. She takes great pains to make this application explicit in propositions (13) through (18) of her paper and in her section entitled “Semantic Constructions” (pp. 319—321). So it is simply false that this application was first made by Kripke in the 1970s.

One of the additions Soames subsequently made to the paper he sent me in mid-December is a paragraph on page 202, the paragraph beginning with “I should mention that ...“ and ending with”.., the truth of the others”. In this new paragraph, Soames criticizes Quine, Marcus, Smullyan, Fitch and others for assuming that “the intelligibility of objectual quantification depended upon universal instantiation and existential generalization being universally truth- preserving, where genuine singular terms are involved” (p. 202). Soames states that this assumption is false and that this assumption was taken by Marcus and others as a reason to believe that if (NI-i) is true, then (NI-2) and (NI-3) are true, where the relevant theses are

(NI-2) For all proper names a and b, if a = b is true then necessarily a = b is true.

(NI-3) For all singular terms a and b (including singular definite descriptions), if a = b is true then necessarily a = b is true.

I think Soames’ added paragraph about these three theses is misguided in several respects.

First, the motive for his discussion of the relations among (NI-i), (NI-2)) and (NI-3) in his claim that I am “just plain wrong to cite Marcus’ early systems of quantified modal logic as introducing the version — (NI-2) — of the necessity of identity that corresponds to the modal argument” (p. 202). However, I did not state that Marcus’ early articles on quantified modal logic introduced the thesis (NI-2). Rather, I stated that her early articles introduced (NI-i) to the NewTheory of Reference. (See pp. 183—4 of my ‘Marcus, Kripke, and the Origin of the New Theory of Reference’, where I stated that in her early articles Marcus presented a “formal proof of the necessity of identity in her extension of S4 {Barcan, 1946; 1947], which is a fourth component she introduced into the New Theory of Reference. She showed that

is a theorem of QS4 ...“). Marcus introduced (NI-2) to the New Theory of Reference in her 1961 article, in propositions (13)—(18) and in the section on “Semantic Constructions”.

Second, Marcus does not argue that (NI-1) entails (NI-2) and (NI-3), and Soames makes no attempt to provide textual evidence that Marcus presents such an argument. I also do not argue that there is such an entailment. Soames’ criticism of the belief that there is such an entailment seems to be a case of setting up a straw-woman (Marcus) and a straw-man (myself).

Third, in this section (section #4, “The Necessity of Identity”) of my reply to Soames, I was in part replying to a statement made in the paper I received from Soames in mid-December, 1994 (a statement that also appears in the version he read at the APA colloquium), viz, the statement that “the quantifica tional version (NI-i) of the necessity of identity, articulated and derived by Marcus, no more provides a direct route to the true thesis (NI-2) involving ordinary proper names than it provides a direct route to the false thesis (NI-3)” (p. 202). In his new paragraph, Soames takes the originally vague phrase “provides a direct route to” to have a precise meaning, namely, entails. If “provides a direct route to” is taken to mean entails, then of course the true thesis (NI-i) no more provides a direct route to the true thesis (NI-2) than it provides a direct route to the false thesis (NI-3).

In order for (NI-1) to entail (NI-2), additional premises are needed, such as the premise that proper names do not express contingent descriptive senses and the premise that variables and proper names are formally equivalent in some contexts. A point I was making in section #4 of this paper was that Soames’ originally vague claim that (NI-i) “no more provides a direct route to” (NI-2) than to (NI-3) is a dubious claim. It is dubious since it can be plausibly argued that there is a “direct route” from (NI-1) to (NI-2) in the sense that the true thesis (NI-2) can be logically derived from the true thesis (NI-i) in conjunction with some plausible supplementary premises about proper names and variables (e.g., that proper names are formally equivalent to variables in some contexts), whereas the false thesis (NI-3) cannot be logically derived from the true thesis (NI-i) in conjunction with supplementary premises.

Another crucial idea is the concept of rigid designation. Soames allows that Marcus implicitly treated proper names as rigid designators. But Soames claims that in order to have the concept of a rigid designator, one needs a semantical framework for modal systems that features the notion of a referent of a singular term at a possible world, and he denies that Marcus had this notion. Soames seems to make a twofold error here.

First, Marcus did present (in her 1961 paper) a semantical construction for a modal language that featured the notion of a referent of a singular term at a possible world. She presents a model-theoretic system that (to quote from her 1961 paper) “corresponds to the Leibnizian distinction between true in a possible world and true in all possible worlds” (p. 320). Specifically, she presents (again quoting from her 1961 paper):

a language (L), with truth functional connectives, a modal operator (à [diamond]) a finite number of individual constants, an infinite number of individual variables, one two-place predicate (R), quantification and the usual criteria for being well-formed. A domain (D) of individuals is then considered which are named by the constants of L. A model of L is defined as a class of ordered couples (possibly empty) of D. The members of a model are exactly those pairs between which R holds. To say therefore that the atomic sentence R(a₁a₂ of L holds or is true in M, is to say that the ordered couple (b₁, b₂) is a member of M, where a and a are the names in L of b and b ... If [ sentence] A is àB then A is true in M if and only if B is true in some model M (Marcus, p. 319).

Here the models M, M₁ etc. are possible worlds, the individuals in the domain D are members of the worlds, and these individuals are referents of the individual constants. Clearly, we do have here a semantics for modal logic that features the notion of a referent of a singular term at a possible world.

Kripke presented a largely similar semantics for modal logic in his 1963 paper ‘Semantical Considerations on Modal Logic’ with the main difference between Marcus’ and Kripke’s semantics being that Kripke (following Stig Kanger’s 1957 work Provability in Logic [Stockholm]) allowed that the domain of individuals is not constant across models, whereas Marcus held the domain of individuals is constant across models. Marcus’ models or worlds differ from one another in that her relation R holds between different individuals in different worlds. But this difference between Marcus and Kripke about the constancy of the domain of individuals is not relevant to the basic concept of rigid designation; rather, it is relevant to whether this concept has two species, weak rigid designation and strong rigid designation (as Kripke held) or whether this weak/strong distinction cannot be made. The relevant fact is that one can derive from both Marcus’ semantical system and Kripke’s semantical system the notion of a referent of a singular term at a possible world and the associated concept of a rigid designator.

But it should be added that Soames’ claim that we need a model-theoretic semantics with the notion of a possible world in order to have the notion of rigid designation is a dubious claim. The concept of rigid designation can be fully stated using only modal operators (“possibly”, “necessarily”) and sub- junctives (“might have been”, etc.), without need of a semantics for modal logic and without the notion of a referent of a singular term at a possible world. Indeed, this is how Kripke states it in ‘Naming and Necessity’. Kripke writes:

One of the intuitive theses I will maintain in these talks is that names are rigid designators. Certainly they seem to satisfy the intuitive test mentioned above; although someone other than the U.S. President in 1970 might have been the U.S. President in 1970 (e.g. Humphrey might have), no one other than Nixon might have been Nixon. ... For example, “the President of the U.S. in 1970” designates a certain man, Nixon; but someone else (e.g., Humphrey) might have been the President in 1970, and Nixon might not have; so this designator is not rigid. (Kripke, 1972, p. 270)

Here the notion of rigid designator is explained in terms of subjunctive expressions “might have been”, “might not have been”, and no mention of possible worlds is needed. This is how Marcus first introduced the notion in her 1961 article. Note that the distinctions made in the following passage from Marcus are logically equivalent to the distinctions made in the passage I just quoted from Kripke’s ‘Naming and Necessity’. Marcus writes:

Let us now return to (10) and (15). [ is “The evening star eq the morning star” and (15) is “Scott is the author of WAVERLY”]. If they express a true identity, then”Scott” ought to be anywhere intersubstitutable for “the author of WAVERLY’ in modal contexts, and similarly for “the morning star” and “the evening star”. If they are not so universally intersubstitutable — that is, if our decision is that they are not simply proper names for the same thing; that they express an equivalence which is possibly false, e.g., someone else might have written WAVERLY, the star first seen in the evening might have been different from the star first seen in the morning—then they do not express identities (Marcus, 1961, p. 311)

This passage conveys the distinction between rigid and nonrigid designators that is also conveyed in the passage quoted from Kripke. If we add to this the fact that Marcus had a semantics for modal logic that featured the notion of a referent of a singular term in a world, then it is quite clear that Marcus was in full possession of the concept of rigid designation in 1961. Soames writes:

“Smith’s characterization of Kripke as contributing only the name ‘rigid designation’ for a concept discovered by Marcus is a grotesquely inaccurate caricature” (p. 203). I suggest that so far from being a grotesquely inaccurate caricature, it is the literal truth.

A further idea that Marcus introduced into the New Theory of Reference is the epistemological argument for direct reference. Consider this passage from Marcus’ 1961 paper:

You may describe Venus as the evening star, and I may describe Venus as the morning star, and we may both be surprised that, as an empirical fact, the same thing is being described, but it is not an empirical fact that

Soames claims that if Marcus is making an epistemological point here, she cannot be using “empirical” to mean “contingent”. But Soames’ claim is not true. Even if we construe “empirical” as meaning “contingent”, the epistemic argument for direct reference is still logically implied by this passage. The argument goes

(1) If “Venus” expresses the sense of “the evening star”, then it is knowable a priori that Venus is the evening star.

(2) This is not knowable a priori, since we may be surprised to learn that, as a matter of contingent fact, Venus is the evening star.

(3) “Venus” does not express the sense of “the evening star” (or any other descriptive sense) but is directly referential.

Soames concentrates on the word “empirical” in the passage from Marcus, but the epistemic argument for direct reference is sufficiently conveyed by Marcus’ point that we may be surprised to discover that Venus is the evening star, which is something that we could not be surprised to discover if we knew this fact a priori by virtue of the meaning of the word “Venus”.

However, Marcus is in fact using “empirical” in its normal philosophical sense to mean a posteriori. Given this, it is even more clear that her passage implies the epistemic argument that proper names are directly referential. The argument goes:

5) If “Venus” has the sense of “the evening star”, it would not be an a posteriori fact that Venus is the evening star.

(6) “Venus” does not have the sense of “the evening star” (or any other descriptive sense, since this argument can be repeated for other descriptive senses).

Soames apparently does not see this argument in Marcus’ paper. He writes: “To get an epistemological argument against descriptivism one would have to argue directly, without appeal to Millian premises, that one cannot know a priori that N is D, for relevant names N and descriptions D. Marcus does not do this” (p. 207). However, I have just shown that she did do this.

At the risk of belaboring the point, I would note that one premise of her argument is that it is an empirical or a posteriori fact that Venus is the evening star. This is not a Millian premise, where a Millian premise is one that states that names are directly referential and do not express senses. The second premise is that “if ‘Venus’ has the sense of ‘the evening star’, it would not be an a posteriori fact that Venus is the evening star”. This is also not a Millian premise; it does not state that the word “Venus” is directly referential. The conclusion is that the word “Venus” does not have the sense of “the evening star”; this is a Millian thesis, but it is not a premise of the argument but the conclusion.

In my paper, I gave a different illustration of the epistemic argument in order to show how the epistemic argument could be used to refute Plantinga’s and Linsky’s theory that names express modally stable senses, i.e., senses that include the actuality operator. My example was of the name “Venus” purportedly expressing the modally stable sense “whatever is actually the morning star and the evening star”. Soames makes a puzzling comment about my example. He correctly states (on page 206) that the modally stable sense I mention, “whatever is actually the evening star and morning star”, and the example I give, is not a reconstruction of Marcus’ example, but is my own invention. But Soames seems to think that this statement counts as a criticism of my example, whereas it is clear from my paper that I make no claim to be analyzing Marcus’ different example. Rather, I introduced my own example in order to show how “the epistemic argument for direct reference” can be used to refute the Linsky-Plantinga theory that names express modally stable senses.[1]

I turn now to the idea of a posteriori necessities. Soames denies that the notion

You may describe Venus as the evening star, and I may describe Venus as the morning star, and we may both be surprised that, as an empirical fact, the same thing is being described. But it is not an empirical fact that

If, as Smith maintains, Marcus is trying to make an epistemological point here, then by “an empirical fact” she must not mean “a contingent fact”, for then there would be no epistemological point at issue. Rather, she must be taken to mean that these sentences do not express facts that are knowable only by empirical means. In other words they are knowable a priori. But if this is right, then true identity statements involving names are not characterized as examples of the necessary a posteriori, and Smith’s contention that Marcus introduced this notion collapses. (p. 204)

Soames is here making an object/metalevel confusion that prevents him from understanding both Marcus’ passage and the basic idea behind the necessary a posteriori. In the first sentence quoted, Marcus is saying that we may both be surprised that our two proper names of Venus, call them “a” and “b”, are co referring. She further states that it is an empirical fact that they are co-referring. In the first sentence, she is talking about metalevel facts about the reference of the two proper names. It is a contingent and a posteriori fact that the two proper names are co-referring. This is the case if the two proper names are homonyms (two uses of “Venus”), or whether one is “Hesperus” and the other “Phosphorus”, or whether we call them “a” and “b”. The reason this metalevel fact is contingent is that the two names might have had different referents. If your use of “Venus” is determined by the reference-fixing rule that “Venus” refers to the evening star, and my use of “Venus” is determined by the reference-fixing rule that “Venus” refers to the morning star, then it is a contingent fact that our two uses have the same referent, for it is not necessary that whatever has the property of being the evening star also has the property of being the morning star. Not only is this fact contingent, it is a posteriori, since it is only by observing the world that I can discover that one and the same body has these two properties and thereby is the referent of both uses of “Venus”.

Thus, the a posteriori feature of a posteriori necessary sentences pertains to the metalevel facts about the referents of the expressions in the sentence and thus to the metalevel facts about what is said by the sentence (i.e., what de re proposition is expressed by the sentence). Marcus uses the “what is said” terminology, and was the first to introduce this terminology (as Joseph Almog pointed out in his 1986 article ‘Naming Without Necessity” in The Journal of Philosophy, Vol. 83, pp. 220, n. 8), but we can also use the terminology of the “de re proposition” or “singular proposition” expressed by a sentence.

The necessary aspect of a posteriori necessary sentences pertains to the object level or, in other terms, to what is said by the sentences, the de re proposition expressed. In the first sentence in the quote, Marcus was talking about the metalevel, about our knowledge of the empirical fact that “the same thing is being described” by our two uses of the name “Venus”. But in the second sentence in the quote, Marcus is talking about the object level, about what is said by her sentences. Her second sentence reads: “But it is not an empirical fact that:

Note that Marcus is no longer talking about facts about what our words refer to; rather, she is talking about the identity of Venus with itself. This is what is said by the sentence, the de re proposition expressed. What is said by the sentence, that Venus is identical with Venus, is not an empirical or a posteriori fact. It is an a priori fact, since it is a mere tautology, an identity proposition.

Thus, what is empirical is that two names (be they two uses of “Venus” or “Hesperus” and “Phosphorus”) both refer to Venus. And what is not empirical is that Venus is identical with Venus. We have this correlation: the metalevel fact (mentioned by Marcus in her first sentence) is a posteriori and contingent, and the object level fact (mentioned by Marcus in her second sentence) is a priori and necessary.

Generally speaking, what is said by an a posteriori necessary sentence may be a tautology or analytic truth that is known a priori. This is not only true in Marcus’ theory of the a posteriori necessary, but also in Kripke’s and Putnam’s. For example, what is said by the sentence “Water is H₂O is the tautology H₂O is H₂O since the word “water” directly refers to H₂O This tautologous proposition is the object level fact. This tautological fact is known a priori; obviously, nobody learns a posteriori that H₂O is identical with H₂O The metalevel fact is the a posteriori fact; the metalevel fact is that the word “water” refers to H₂O This can only be known a posteriori, by observing the world to see what satisfies the reference-fixing description that governs the use of the word “water”. The description that fixes the reference of the word “water” is the description “whatever is causally responsible for the observable properties of a thirst-quenching, clear, tasteless liquid”. We learn a posteriori that H₂O rather than some other chemical structure, such as Putnam’s XYZ, is what satisfies the reference-fixing rule of use of “water”.

Now compare this with Soames’ remark on page 205: “Thus, according to Marcus true identity statements involving genuine names never express empiri cal facts, but rather are tautologically, or analytically true. Surely this is inconsistent with interpreting her as introducing the notion of the necessary a posteriori identities”. This remark betrays Soames’ object level/metalevel confusion. What is said or expressed by true identity sentences involving genuine names is tautologically true, and what is said is known a priori. But that is perfectly consistent with knowing a posteriori the metalevel fact that the sentences in question express this tautological proposition rather than some other proposition. In order to know that they express this de re proposition, and not some other de re proposition, we need to observe the world to determine what satisfies the reference-fixing rules of the relevant words in the sentence.

Soames’ account of the notion of a posteriori necessity implies that it is logically impossible that there be any a posteriori necessities. For Soames claims that if what is said by a sentence is a priori, then the sentence cannot be an a posteriori necessary sentence. On this construal, the sentence “water is H₂O would count as an a priori necessity, since what is said by this sentence, that H₂O is H₂O is known a priori.

A correct understanding of the a posteriori necessary would show that the identity sentence “water is H₂O is an a posteriori necessity because: (a) On the metalevel, it is an a posteriori fact that this sentence expresses the de re proposition H₂O is H₂O and not some other de re proposition, such as XYZ is H₂O (to borrow Putnam’s example). (b) On the object level, what is expressed by this sentence, the de re proposition that H₂O is H₂O is a necessary fact, indeed, a tautology, and is known a priori.

We may summarize this discussion by considering Soames’ sentence: “on the view articulated by Marcus, true identity statements involving names are knowable a priori” (p. 205). A correct phrasing of Marcus’ view would read instead “on the view articulated by Marcus, true identity statements involving names are knowable a priori on the object level but are knowable a posteriori on the metalevel”.

I believe a central reason why Soames did not recognize the argument for a posteriori necessary in Marcus’ work is that he uses her absence of the current Kripkean terminology (“a posteriori necessities”, etc.) as evidence that she did not have the concept of the a posteriori necessity. For example, Soames takes pains to emphasize that “It should be noted that Marcus nowhere in the article says that there are true identity statements which are knowable only a poster- ion” (p. 204). Certainly, Marcus never called any of her arguments “the argument for a posteriori necessity” or the like.

But the point, rather, is that these arguments are logically implied by what she does say, which can be put very succinctly. Soames admits that Marcus implicitly treated proper names as directly referential rigid designators. Given this admission, it is a short step to realize that this doctrine itself entails the doctrine of the a posteriori necessity. If “a” and “b” are directly and rigidly referential names of the same thing, it immediately follows (given Marcus’ theorem of the necessity of identity) that “alb” states a necessary truth, but the fact that “a” and “b” are coreferential (and that “alb” thereby states a necessary truth) is known only a posteriori. On the object level, “aIb” is necessary and a priori, but on the metalevel, it is contingent and a posteriori. We don’t even need to argue about the meaning of Marcus’ passage that I quoted, since the doctrine of the a posteriori necessity is logically entailed by the doctrines that Soames himself admits were espoused by Marcus.

One of the additions Soames made to the paper I received in mid-December is footnote 26, which analyses Marcus’ response to one of Kripke’s questions in the Discussion after her presentation of ‘Modalities and Intensional Languages’ on February 7, 1962 at the Boston Colloquium for the Philosophy of Science. Soames’ new footnote pertains to his quotation (on page 198) from Marcus’ Discussion in which she elaborates upon her doctrine of a posteriori necessities. In this quotation, Marcus links the epistemological status of co referring names to their modal and semantic status, in a way that implies the doctrine of the a posteriori necessary. Soames’ new footnote 26 contains the remark that this linkage marks “a very important contrast” between Marcus’ theory and Kripke’s.

Part of the “very important contrast” that Soames alleges is based on his incorrect attribution to Marcus of the view that names are mere tags entails or “is the ground for” the thesis that coreferring names are intersubstitutable in modal contexts; Marcus does not hold this view since Marcus’ theory is that further premises are needed for this entailment to go through, e.g., the premise that identities are necessary.

But more germane to our discussion of a posteriori necessity is that in this new footnote Soames seems to find suggestive evidence of a contrast between Marcus and Kripke in that “Marcus responds to a straightforward question about the necessity of identities involving names (and their intersubstitutivity in modal constructions) with remarks about epistemology and synonymy” (p. 212, note 26).

Soames’ belief that Marcus’ linkage of these issues marks a contrast between her and Kripke is indicative of his failure to grasp the fact that in her response she explains her argument for the a posteriori necessary, as I will now show in detail. In the 1962 Discussion, Kripke somewhat innocently or naively asks Marcus if her theory requires the thesis “that for any two names ‘A’ and ‘B’, of individuals, ‘A = B’ should be necessary, if true at all? (See “Discussion on the Paper of Ruth B. Marcus”, Synthese, Vol. XIV, no. 2/3, September, 1962, pp. 141—42. This Discussion is reprinted with minor changes in Marcus’ Modalities; see p. 33). Since it should have been obvious from Marcus’ talk that this thesis was one of the major points of her talk, Marcus chooses to elaborate on her epistemological theory rather than simply respond that of course her theory requires this thesis, as she repeatedly emphasized in her talk.

Marcus’ elaboration includes the epistemological point about having re course to a dictionary that Kripke later incorrectly represented in ‘Identity and Necessity’ and ‘Naming and Necessity’ — misrepresentations which played a major part (perhaps the major part) in preventing philosophers from recognizing that Marcus possessed the theory of the a posteriori necessary. Soames’ footnote 26 pertains to the following passage from the 1962 ‘Discussion’ in Synthese, p. 142 (Soames quotes this passage from Marcus’ Modalities, pp. 33—34, which contain a few minor stylistic changes that are irrelevant to its meaning). My emphases are added. Marcus says:

Presumably, if a single object had more than one tag, there would be a way of finding out such as having recourse to a dictionary or some analogous inquiry, which would resolve the question as to whether the two tags denote the same thing. If “Evening Star” and “Morning Star” are considered to be two proper names for Venus, then finding out that they name the same thing as “Venus” names is different from finding out what is Venus’ mass, or its orbit. It is perhaps admirably flexible, but also very confusing to obliterate the distinction between such linguistic and properly empirical procedures.

Note several things about this famous remark and Kripke’s even more famous criticisms of it in ‘Identity and Necessity’ (1971, pp. 142—43) and ‘Naming and Necessity’ (1972, P. 305). First, Marcus implies that having recourse to a dictionary is merely one example of finding out that the same object has more than one proper name. This is the significance of the “such as” that I italicized and the phrase “or some analogous inquiry”. An analogous inquiry, but different from having recourse to a book, would be to learn from conversation with other people that the same object has more than one name. Marcus is correct in making these claims. One way we do find out that the same object has more than one name is by finding more than one name listed for the object in a dictionary.

But notice that Kripke’s representation of Marcus’ remarks is misleading when he criticizes it in ‘Identity and Necessity’ (1971, pp. 142—43) and ‘Naming and Necessity’ (1972, p. 305). He writes in ‘Identity and Necessity’ (pp. 142— 43): “You have an object a and an object b with names ‘John’ and ‘Joe’. Then, according to Mrs. Marcus, a dictionary should be able to tell you whether or not ‘John’ and ‘Joe’ are names of the same object. Of course, I do not know what ideal dictionaries should do, but ordinary proper names do not seem to satisfy this requirement”. Kripke here represents Marcus’ claim — which is that having recourse to a dictionary is merely one example of a way of finding out how we may fluid out that the same object has more than one name — as the claim that having recourse to a dictionary is a necessary condition (a “require ment”) of finding out they have more than one name. This misrepresentation is compounded by Kripke deliberately choosing an example (“Joe” and “John”) where we do not learn from a dictionary whether they are co-referring. In such an example, we would find out they are co-referring by “some analogous inquiry”, such as asking somebody if Joe is also called “John”.

Kripke’s second major misrepresentation of Marcus’ theory appears in his next sentence, which is presented as a criticism of Marcus. Kripke writes: “You certainly can, in the case of ordinary proper names, make quite empirical discoveries that, let’s say, Hesperus is Phosphorus, though we thought other wise” (1971, p. 143). This sentence has misled countless readers into thinking that Marcus did not have the doctrine of the necessary a posteriori. Note that it mis-states Marcus’ view; she does not deny that we can “in the case of ordinary proper names, make quite empirical discoveries [such as that] as that Hesperus is Phosphorus”. What she implies in her remarks is that it is a different sort of empirical discovery than discovering that Venus has a certain mass.

Marcus correctly observes that ‘If Evening Star” and ‘Morning Star’ are considered to be two proper names for Venus, then finding out that they name the same thing as ‘Venus’ names is different from finding out what is Venus’ mass, or its orbit” (‘Discussion’ in Synthese 14 (2/3), 1962, p. 142). Certainly, looking through a telescope at Venus or bouncing electromagnetic rays off of Venus is a different sort of procedure from learning that it is a linguistic convention that “Hesperus” and “Venus” are both used to name the same object. As Marcus continues, “It is perhaps admirably flexible, but also very confusing to obliterate the distinction between such linguistic and properly empirical procedures” (my emphasis). It is clear from the context that by a “properly empirical procedure” Marcus means a scientific observation or an observation of a physical thing.

And surely she is correct that there is some sort of distinction between the procedure of observing the physical properties of physical things and the procedure of learning about the references of names. But this distinction is far from implying that we do not learn a posteriori (“empirically”, in a sense that is broader than the so-called “proper procedural” sense Marcus distinguishes in the “Discussion”) that “Venus” and “Hesperus” name the same thing. As Marcus says in her 1961 paper, “You may describe Venus as the evening star and I may describe Venus as the morning star, and we may both be surprised that as an empirical fact, the same thing is being described” [1961, p. 310, my emphasis].

Thus, it is Kripke who errs in representing Marcus’ theory, not Marcus who errs in stating in her theory, as Kripke alleges in the following passage from ‘Identity and Necessity’: “I thus agree with Quine, that ‘Hesperus is Phos phorus’ is (or can be) an empirical discovery; with Marcus, that it is necessary. Both Quine and Marcus, according to the present standpoint, err in identifying the epistemological and metaphysical issues” (‘Identity and Necessity’, p. 154, n. 13). But the error is Kripke’s, for “the present standpoint”, that” ‘Hesperus is Phosphorus’ is (or can be) an empirical discovery” and state a necessary truth, is Marcus’ standpoint. Soames writes in his paper that “Smith maintains that Saul Kripke learned these doctrines from her, initially misunderstood them, and, when he later straightened things out, mistakenly took the doctrines to be his own” (p. 191). I maintain the present instance is a case in point.

(In a forthcoming paper, ‘Direct, Rigid Designation and A Posteriori Necessity’, I argue there are two very different senses of “a posteriori necessity” that have been confused in the post-Marcus literature. One sense pertains to identifications and I discuss this sense in William Lane Craig and Quentin Smith, Theism, Atheism and Big Bang Cosmology [Clarendon Press, Oxford, 1993], pp. 184—185, and the other pertains to laws of nature and is discussed in Quentin Smith, Language and Time [ University Press, New York, 1993, pp. 203—204] and ‘The Metaphysical Necessity of Natural Laws’, forthcoming.)

I conclude that seven of the main ideas and arguments of the New Theory of Reference are indeed present in Marcus’ 1961 paper and that Kripke merely elaborated upon them and gave them a new terminology. Is it scandalous that this conclusion should have been given credence by the APA committee that selected the papers for this meeting, as Soames suggests? I suggest that it is not, but instead should be viewed the way the APA committee viewed it, namely, as an important issue about the history of recent analytic philosophy that deserves discussion. This contribution to the study of the history of philosophy is no more scandalous than is McCracken’s recent book Malebranche and British Philosophy, in which McCraken shows (on pages 257ff.) that many of the allegedly original ideas in Hume’s A Treatise of Human Nature were taken, without due acknowledgement, from Malebranche, and that this historical fact about Hume and Malebranche has gone unrecognized for centuries. This is simply one of the ways in which contributions to the study of the history of philosophy are made. Our concern here is not attacks or defenses of certain personalities or the doling out of credit and blame to our friends or colleagues, as Soames characterizes this discussion. Rather, our concern in the history of philosophy is the same as our concern in other areas of philosophy, a concern to find out the truth.

In the section on the modal argument, I added some material showing that Soames’ new footnote 34 misrepresents the modal argument for direct reference. Here I would like to make an additional response to this new footnote, that it reveals an internal inconsistency in Soames’ paper. Footnote 34 shows that Soames concedes my general thesis that Marcus had many of the ideas of the New Theory of Reference prior to Kripke, even though she is not credited with them. However, this concession is inconsistent with the conclusion Soames draws in his paper. Soames’ conclusion is worth comparing with his added footnote. His conclusion is that providing Marcus with proper credit:

does not result in a reassessment of the seminal role of Kripke and others as primary founders of contemporary nondescriptivist theories of reference. In fact, I think Smith does Kripke a grave injustice. Smith writes as if Kripke appropriated the major views expressed in Naming and Necessity from Marcus while denying her proper credit, and suggests that it is a scandal that the rest of the profession was thereby duped. We have not been duped; there was no misappropriation. Rather, what Smith has done is to mistakenly read many of Kripke’s arguments and doctrines back into Marcus, and then to insinuate that Kripke is guilty of theft. This, it seems to me, is shameful. If there is any scandal here it is that such a carelessly and incompetently made accusation should have been given such credence (pp. 208—9).

There are several problems with this passage, even apart from its “colorful” language and suggestion that the APA program committee behaved scandalously. One problem is that Soames mistakenly represents the conclusion of my paper as the claim that “Kripke is guilty of theft”. In fact, the conclusion of my paper was that Kripke unconsciously assimilated Marcus’ ideas (as opposed to consciously assimilating them and deliberately not crediting them to her, which is what a “guilty of theft” charge would imply). The “guilty of theft” interpretation of Kripke’s relation to Marcus is not supported by the available evidence, for if Kripke wanted to “steal and conceal” he would not have discussed or referred to her 1961 paper at all, whereas he discusses it in both ‘Identity and Necessity’ and ‘Naming and Necessity’. The conclusion of my paper was instead that:

a reasonable explanation of why Kripke did not attribute the central features of the “New Theory” to Marcus is that he originally misunderstood Marcus’ New Theory of Reference. When he eventually understood it, after a year or two, the insight that came made it seem that the ideas were new. I suspect that such instances occur fairly frequently in the history of thought and art. (‘Marcus, Kripke, and the Origin of the New Theory of Reference’, p. 188)

Although several philosophers have privately disagreed in print or conversa tion with my conclusion and have argued on the basis of the evidence I presented that Kripke was guilty of plagiarism (I will leave them to make their own case publicly), I believe my contrary conclusion, that Kripke unconsciously assimilated ideas that were originally misunderstood, is better supported by the evidence. In addition to the fact that Kripke discusses and refers to Marcus’ paper (which he would not do if he wanted to hide his main source), there is evidence he originally misunderstood many of Marcus’ ideas that he later espoused and elaborated upon in the early 1970s. I mentioned in my first paper Kripke’s 1962 misconstrual of Marcus’ theory that is revealed in his attribution to her of the view that “the tags are the ‘essential’ denoting phrases for individuals”, but empirical descriptions are not, and thus we look to statements containing ‘tags’, not descriptions, to ascertain the essential properties of individuals” (see page 142 of the ‘Discussion’ in Synthese and page 34 in Marcus’ Modalities).

(In footnote 46 of his paper, Soames denies that Kripke is here misunder standing Marcus and denies that Kripke is saying that tags are essentially denoting phrases of the individuals they denote; Soames puts a different and rather fantastic construction on Kripke’s remark, so fantastic that I will simply refer the reader to Soames’ footnote 46 and let the reader decide for herself if there is any plausibility to Soames’ claim that this construction is what Kripke meant by his remark.)

Other remarks made by Kripke also suggest he did not adequately under stand Marcus’ theory. For example, he has to ask Marcus: “Now, do you admit the notion of ‘identically the same’ at all?” (‘Discussion’, p. 136; Modalities, p. 28), whereas it should have been evident to anyone who understood Marcus’ talk that most of her talk consisted of her arguing for and explaining her notion of ‘identically the same’. Marcus has to respond to Kripke’s question that “I admit identity on the level of individuals certainly” (‘Discussion’, p. 136; Modalities, pp. 28 29). And later Kripke has to ask about her theory: “does it not require that for any two names, ‘A’ and ‘B’, of individuals, ‘A = B’ should be necessary, if true at all?” (‘Discussion’, pp. 141—142; Modalities, p. 33), whereas it would have been plainly obvious to someone who fully comprehended her talk that this was one of her main points.

Further evidence that Kripke did not adequately understand Marcus’ 1962 talk is that he himself says (in the 1980 Preface to Naming and Necessity, p. 5) that “Eventually I came to realize — this realization inaugurated the aforemen tioned work of 1963 — that the received presuppositions against the necessity of identities between ordinary names were incorrect”. If Kripke fully understood Marcus’ theory in 1962, why does he not say that he first realized at Marcus’ 1962 talk that the received presuppositions against the necessity of identities between ordinary names were incorrect? For these reasons, among others, it seems to me that the “plagiarism” explanation of Kripke’s relation to Marcus is false, and that the “unconscious assimilation of originally misunderstood ideas” explanation of the relation is correct.

But we have not yet touched on the main problem with the passage I quoted from Soames’ concluding remarks, namely, this his conclusion is inconsistent with his new footnote 34 and with the main body of his paper. Soames concludes that I “mistakenly read many of Kripke’s arguments and doctrines back into Marcus” and that my general thesis that “Kripke appropriated the major views expressed in Naming and Necessity from Marcus” is an incompe tently and carelessly made accusation. However, in footnote 34 and in the main body of his paper, Soames expressly concedes that many of the points made in my paper about Marcus’ priority to Kripke are true. It will be instructive to quote footnote 34 in its entirety and to insert some comments in brackets and italics:

This is an important feature of Kripke’s version of the modal argument that does not seem present in Marcus’ version. [Soames here concedes my point that Marcus did have the modal argument, in some version, prior to Kripke, even though he mis-states the modal argument in the way I described in section 3.] Both Kripke and Marcus use modal properties of names to distinguish them from descriptions. [Soames here concedes my point that Marcus had the distinction between names and descriptions, and an argument for this distinction, prior to Kripke.] However, the way they establish the modal properties of names is quite different. Kripke provides examples in which we evaluate simple sentences containing proper names (but no modal opeators) at alternative possible worlds (counterfactual circumstances). He uses these examples to show that the referent of a name at one world is the same as its referent at other worlds, and hence that names are rigid. It then follows that identities involving names are necessary if true, and that coreferential names are intersubstitutable in modal constructions. Marcus, on the other hand, seems to think that these modal properties of names somehow follow from the quantified version of the necessity of identity, either by itself or in conjunction with her controversial view that names are “mere tags”. [Soames here concede,s my four points that Marcus possessed before Kripke (1) the theory that coreferential names have the modal property of being intersubstitutable in modal constructions, (2) the theory that identities involving names are necessary f true, (3) the theory that names are rigid, (4) the theory that the referent of a name at one world is the same as its referent at another world.]

This new footnote is consonant with the main body of Soames’ paper, in which these and other of my points about Marcus’ priority are also conceded. For example, on page 197 Soames expressly concedes that in her 1961 paper Marcus makes the point “that (i) coreferential names are substitutable in modal contexts without change in truth value whereas descriptions are not, (ii) identity statements involving names are necessary if true, (iii) names are not equivalent to descriptions, (iv) descriptions are not genuine singular terms, (v) proper names have no meanings over and above their referents, and (vi) coreferential names have the same semantic content ...“.

Soames also concedes on page 196 that Marcus “can be credited with an implicit recognition that genuine names are rigid designators”. (But Soames oddly seems to think that my quotation from Marcus’ 1970 talk is meant to support my claim that Marcus had this notion before 1963—64, the years Kripke admitted he first came to understand this notion. See Soames, p. 203). Further, on page 200 Soames concedes both that Marcus had a “version of ‘the modal argument’ “ and that for this reason she (and Smullyan and Fitch) “deserve a degree of recognition that they are often not given”. Soames also allows that “Marcus’ quantified version of the necessity of identity anticipates the contemporary view that variables are rigid designators” (p. 208). How are all these concessions consistent with Soames’ conclusion that providing Marcus “with proper credit does NOTresult in a reassessment of the seminal role of Kripke and others as primary founders of contemporary non-descriptivist theories of reference” (p. 208, my emphases)?

If we eliminate the ideas that Soames concedes are present in Marcus’ early writings from the contemporary non-descriptivist theories of reference, we do not even have a non-descriptivist theory of reference, let alone several philosophers who deserve instead of Marcus the title of a “primary founder” of contemporary non-descriptivist theories of reference. If Soames is to draw a conclusion that is consistent with footnote 34 and the main body of his paper, he should have concluded that the evidence I presented established that Marcus was a primary founder of the New Theory of Reference, even though Soames does not agree with all of the points I made about her primacy. This conclusion would at least enable Soames’ paper to have the virtue of self-consistency, even if it would not thereby acquire the virtue of being entirely accurate.[2]

Fitch, F.: 1950, ‘Attribute and Class’, in Farber (ed), Philosophical Thought in France and Uniled State Buffalo.

Fitch, F.: 1949, The Problem of the Morning Star and the Evening Star’, Philosophy of Science, 16, 137—141.

Marcus, RB, et al.: 1962, ‘Discussion on the Paper of Ruth B. Marcus’, Synthese, 14, 132—143.

McCracken, C.J.: 1983, Malebranche and British Philosophy, Oxford University Press.

Ramsey, F.P.: 1931, Truth and Probability’, Foundations of Mathematics. Routledge and Kegan Paul.

Quine, W.V.O.: 1949, ‘Designation and Existence’, in FeigI and Sellars, W. (eds), Readings in Philosophical Analysis (New York: Appleton-Century-Crofts, Inc.), Reprinted from The Journal of Philosophy 36, 701—709.

Smith. Q.: 1995, ‘Marcus, Kripke and the Origin of the New Theory of Reference’, Synthese, 104, 179—189.

Smith, Q. and W.L. Craig.: 1993, Theism, Atheism and Big Bang Cosmology Clarendon Press, Oxford.

Soames, S.: 1995, ‘Revisionism about Reference: A Reply to Smith’. Synthese, 104, 191—216.

Smullyan. A.: 1947, Review of Quine’s ‘The Problem of Interpreting Modal Logic’, The Journal of Symbolic Logic, 12,139—141.

Smullyan. A.: 1948, ‘Modality and Description’, The Journal of Symbolic Logic, 13, 31—37,

[1] {May, 1995 addition: At the APA meeting, I crossed out my original version of the paragraph beginning “In my paper, I gave a different illustration ...“ and did not read this paragraph at the APA colloquium, since the second half of the original version of this paragraph needed revision. I have now omitted the problematic parts of this paragraph and added three new sentences (beginning with “He correctly states [ page 2061 that the modally stable sense …”.}

[2] {May, 1995 addition: I am very much indebted to Arthur Falk for helpful and extensive email correspondence between December 16 and December 25, 1994 about the origins of the New Theory of Reference. Arthur Falk’s insights enabled the reply to Soames I read at the APA meeting to be much better than it would otherwise have been. I am indebted to William Vallicella for the reference to the Malebranche/Hume relation and to Robert Fogelin for calling my attention to an instructive analogy between the Marcus/Kripke relation regarding the New Theory of Reference and the Fogelin/Kripke relation regarding the Skeptical Paradox Interpretation of Wittgenstein’s private language argument. See pages 241—246 of the second edition of Robert Fogelin’s Wittgenstein (Routledge and Kegan Paul, London, 1987).}