1 Kripke, Saul A. “Naming and Necessity,” in Harman, and Davidson, eds., Semantics of Natural Language (Dordrecht, 1972), 253–355CrossRefGoogle Scholar, and 763-769. References in brackets in the text are to page numbers of that book. The transcript of another lecture covering some of the same material is published under the title “Identity and Necessity” in Munitz, Milton K. ed., Identity and Individuation(New York, 1971)Google Scholar (Henceforth ‘Munitz’).

2 Partly, no doubt, through Kripke's own influence. He mentions the following as having independently expressed similar views on one point or another: Albritton, Chastain, Donnellan, Putnam, Slote, Stroud. Other points of contact are indicated below.

3 Quine's opposition to the notions of modal logic is well known. It may seem curious that in “Two Dogmas of Empiricism” his ground for rejecting them is that they are all interdefinable but not definable in observation or non-modal terms. One might think that ineliminability is an argument in favour of a certain mode of discourse, since if the terms of modal logic were definable in other terms, they would be theoretically superfluous, like singular terms (cf. Word and Object, § 38). This contrary strategy is basically Chisholm's when he argues for the uniqueness of mentalist discourse (Perceiving: Ithaca, 1967; ch. 11). More reasonably, however, Quine's strictures may be viewed as a demand that modal terms be given some explanation that someone of his philosophical temper will find intelligible. And this condition is, perhaps, satisfied by model-theoretic semantics. Nevertheless one may still be troubled about how modal logic interpreted by model theory is itself to be interpreted in the real world. Perhaps we carry models in our heads, like Leibniz's God. But if that is the answer, Quine's philosophical attitude has been vindicated in essential respects. This is not a point which I can argue here.

4 Except in a Pickwick-Frege sense. We might say: the sense of a proper name is just that it is assigned to such and such. But that gives no guidance for interpretation in oblique contexts (277 and 346).

5 For an evocation of the troubles such a quest incurs, d. Chisholm, “Identity through Possible Worlds: Some Questions,” in Nous I (1967).

6 As counselled by Lewis, David “Counterpart Theory and Quantified Modal Logic,” journal of Philosophy, 65 (1968).CrossRefGoogle Scholar

7 Cf. Ishiguro, H. Leibniz's Philosophy of Logic and Language,(London, 1972), p. 123Google Scholar:

Leibniz's account, though formally identical thus far with that of Lewis, is less bizarre. For possible worlds exist in God's mind, and so there seems to be a place for real possibility attaching to an actual individual: namely God's capacity to create other worlds. This means that Leibniz has not analyzed out actual possibilities completely. If he had, as Lewis does, then ‘God could have created another world’ would mean ‘There is a world where a counterpart of God creates a different world’. This I find (if there are gradation of oddity among theological statements) the more bizarre, in being the more thorough-going, version of counterpart theory.

8 Hintikka, J. Knowledge and Belief,(Ithaca, 1962), pp. 141-2.Google Scholar See below, pp. 454 ff.

9 One might think the case where ϕ is ‘is named ‘N”provides a trivial counterexample. But it does not. (283-6).

10 This does not follow if the difference between de re and de dicto is defined purely in formal terms, as holding between expressions beginning ‘(Quantifier)(Modal Operator)’ and those beginning ‘(Modal Operator)(Quantifier)’. This is the procedure adopted by Plantinga in “De Re and De Dicto,” Nous Ill (1969), in which he proposes reducing de re modalities to de dicto ones involving proper names. But the point of preferring de dicto modalities (for those, like Quine, who do) is that such modalities have an unmysterious source in language: “Necessity has its source in the way we talk about things, not in the things we talk about.” (Quine, Ways of Paradox, p. 174, quoted by Plantinga p. 247.) Once we give up descriptivism, which assigns sense to names, names have reference without sense. So any necessity that attaches to ‘ ϕ N’, even if it is expressed formally as de dicto, must be due to the referent not the name. Hence it must be de re. Perhaps Plantinga implies this, for he is concerned to show that no gain in clarity results from reducing de re to de dicto.

11 Marcus, Ruth Barcan “Essential Attributor,” journal of Philosophy 68 (1971), p. 191.CrossRefGoogle Scholar Aristotle may have thought that his notion of Form could account for both kinds of essence. Cf. Albritton, R. and Sellars, W. “Substance and Form in Aristotle” (symposium) journal of Philosophy 54 (1957).CrossRefGoogle Scholar

12 As argued by Rorty, Amelie Oksenberg in “Essential Properties in the Actual World,” Review of Metaphysics XXV (1972).Google Scholar

13 For the causal theory of knowledge, vid. Goldman, Alvin “A Causal Theory of Knowledge,” journal of Philosophy, 64 (1967).CrossRefGoogle Scholar

14 For a·defense of this perspective, vid. Sellars, W. “Phenomenalism,” in Science, Perception, and Reality,(London & New York, 1963), esp. pp. 87–91.Google Scholar

15 Hintikka, V. op. cit., 141-2, and cf. Sleigh, Robert C. Jr. “Restricted Range in Epistemic logic,” journal of Philosophy 69 (1972), p. 67.Google Scholar

16 Word and Object, § 35. For Russell's, distinction of scope, vid. “On Denoting,” in Logic and Knowledge,ed. Marsh, (London, 1956), p. 52;Google Scholar and cf. also Smullyan, A. F. “Modality and Description,” in Linsky, L. ed., Reference and Modality,(Oxford, 1971).Google Scholar

17 Cf. also Munitz, where his reasons are stated more fully, and from which the argument in the text below is taken.

18 Cf. 333, where a similar point is made in relation to the question whether this (wooden) table might have been made of ice:

19 The same problem arises in connection with mythical species. The sight of partially submerged rhinoceri may have led people to believe that there were animals like horses but with a horn, which they named ‘unicorn’. So we would have a baptism and a causal chain from it to us. But in the mean time various legends sprang up, which are in fact false beliefs about unsubmerged rhinoceri. Nevertheless, unicorns are clearly not rhinoceri. (This point was suggested by a conversation with Peter Geach).

20 For careful proofs of this, cf. Wiggins, D. Identity and Spatio-Temporal Continuity (Blackwell, 1967), p. 5,Google Scholar and R. Cartwright, “Identity and Substitutivity,” in Munitz.

21 Cf. n. 10 to p. 451 above.

22 Given this motivation, the need for some such strategy is formally proved by Quine's demonstration that unrestricted substitutivity both of co-referential terms and of logical equivalents ensures truth-functionality (Quine, Ways of Paradox, pp. 161-2) and so are incompatible with modalities operating on sentences de dicto. If necessity de re is to be reduced to de dicto, therefore, some restriction on substitutivity must be imposed. The strategy has many variants. In Hintikka, the privileged terms are those that satisfy the schema: ‘(Ex)Ka(-=x)’, trickily read as ‘a knows who -is’. (Op. cit., 144). For Kaplan, the priviledged terms are those that “represent” the thing or are “standard names” of it. (“Quantifying in,” in Davidson, & -Hintikka, eds., Words and Objections,(Dordrecht, 1969)Google Scholar.) Sosa, Ernie requires that his terms be “distinguished,” which turns out to be a context-relative matter (“Propositional Attitudes de Dicto and de Re,” journal of Philosophy 67 (1970).CrossRefGoogle ScholarPlantinga gives a lucid account of this strategy and its motives, and requires merely that his privileged terms be proper names. As I pointed out above, his resulting reduction is a reductio, no doubt by design. (vid. fn. 10 top. 451 above). Cf. also the paper by Sleigh cited in fn. 15.

23 Another strategy with which we can assume Kripke would have equally little sympathy consists in confining quantified modal logic to forms that are not committed to essentialism in the full sense. Parsons, (“Essentialism and Quantified Modal Logic,” Philosophical Review 78 (1969)CrossRefGoogle Scholar) argues that we can be content to think of the necessity of arithmetical or analytic truths as de dicto only, confining de re necessities to trivial ones. While Parson's aim in this article and in “Grades of Essentialism in Quantified Modal Logic” (Nous I (1967)) appears to be to refute Quine, he endorses Quine's central philosophical assumptions without reservation.

24 Cf. Quine, “Natural Kinds,” in Essays in Honor of Hempel,ed. Rescher, et al. (Dordrecht, 1970)Google Scholar:

25 Cf. Quine, “Two Dogmas of Empiricism,” in from a logical point of view, Harvard, 1953, and Goodman, N. face Fiction, and Forecast, Harvard, 1955.Google Scholar

26 There is no need to quibble about whether any arbitrary limitations are entailed by the term ‘physical’ here. For as Chomsky once remarked, we call any phenomenon ‘physical’ just as soon as we understand it.

27 Contrast Goodman, N. “The Way the World Is,” in The Review of Metaphysics, XIV (1960)Google Scholar:

26 Such a doubt is expressed by Quine in “Natural Kinds” (21-2). We can take his point without sharing his feeling that natural kinds and their fellow-travellers (cause, subjunctive conditionals, dispositions, etc.) are scientifically “disputable.”

29 Note a misprint at a crucial stage here. The argument just sketched concludes: “the apparent contingency … thus can be explained by some sort of qualitative analogue as obtained in the case of heat.”(339). Obviously this should read: ‘ … thus cannot be explained’.

30 Quoting Wittgenstein, Philosophical Investigations, §50.

31 The point at issue between the advocates of type-type identities and the advocates of particular identities (cf. n. 74, 354). The latter deny what the former assert, that pain is a natural kind.

32 Though it is clear from this argument that this is not the only strategy open to the advocate of the identity thesis. He could also hope that science might show that it is of the essential nature of certain brain states to be painful, thus allowing him to reject 1(b) while still granting 1(a).

Note moreover that 2(a) may be accepted as describing a priori knowledge about the referent of ‘pain’ but not about its meaning. This is what is required by taking ‘pain’ as a name for a natural kind (vid. 346, n. 22, and cf. above, p. 449 with n. 4, pp. 450-1 and p. 463).