2 Cf., e.g., Feyerabend, “Problems of Empiricism,” (pp. 145–260 in Colodny, R. (ed.), Beyond the Edge of Certainty. Englewood Cliffs: Prentice Hall, 1965)Google Scholar; and Kuhn, T. Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1962, 1970)Google Scholar, Ch. 5 et passim. For a convenient summary and critical assessment of their views on the matter, cf. my “The Search For Philosophic Understanding of Scientific Theories” (in F. Suppe (ed.), The Structure of Scientific Theories. Urbana: University of Illinois Press, 1973) Sections V-B-1-b, d, and V-B-2.

3 For samples of these various sorts of criticisms, cf., Shapere, D. “The Structure of Scientific Revolutions,” Philosophical Review, 73 (1964): 383-394,CrossRefGoogle Scholar and “Meaning and Scientific Change” (pp. 41–85 in Colodny, R. (ed.), Mind and Cosmos, Pittsburgh: University of Pittsburgh Press, 1966)Google Scholar; Scheffler, I. Science and Subjectivity (Indianapolis: Bobbs Merrill, 1967)Google Scholar; and my “The Search For Philosophic Understanding of Scientific Theories,” op. cit., which critically assesses such objections.

4 For argument that Scheffler's arguments (op. cit.) so require the presentation of such an analysis of facts, cf. my op cit., Section V-B-2.

5 P. 38 of Strawson, P. “Truth,” Proceedings of the Aristotelian Society, Supplementary Volume 24 (1950)Google Scholar; reprinted in Pitcher, George (ed.), Truth (Englewood Cliffs: Prentice Hall, 1964), pp. 32–53;Google Scholar all page references are to this reprinting.

6 I am using ‘proposition’ the way Geach, P. does in Reference and Generality (Cornell: Cornell University Press, 1962)Google Scholar: “A proposition is something which we propound or put forward—it may or may not be asserted.” (p. 25) propositions thus are linguistic entities. By formulating my analysis in terms of propositions rather than statements, it applies to a wider class of cases.

7 Some authors (e.g. Herbst, P. “The Nature of Facts,” reprinted on pp. 134–136 of Flew, A. Essays in Corceptual Analysis (London: Macmillan, 1960)Google Scholar) restrict facts to being what is stated or expressed by a posteriori propositions (or a subset thereof), whereas other authors (e.g., White, A. Truth (Garden City: Anchor Books, 1970)Google Scholar, ch. 4) maintain the every true proposition (including a priori ones) expresses a fact. I do not want to enter that dispute here, my present concern being restricted to analyzing what it is for a posteriori propositions to express facts. Accordingly, in this paper I use ‘fact’ as being synonymous with ‘empirical fact’.

8 Tractatus Logico-Philosophicus, tr. Pears, D. F. and McGuinness, B. F. (London: Routledge and Kegan Paul, 1961)Google Scholar, propositions 1-1.12. In my use of Wittgenstein I make no claim that I am being faithful to his doctrines or intentions. It is especially important to note that my simple particulars need not be the same sorts of things as Wittgenstein's objects. Cf. Copi, I. M. “Objects, Properties, and Relations in the Tractatus,” Mind, Vo . LXVII N. S., No. 226, April, 1958, pp. 145–165CrossRefGoogle Scholar, for an account of Wittgenstein's notion of objects.

9 Cf. Strawson, op. cit.; Herbst, op. cit.; and White, op. cit.

10 Thus I am denying, for the moment, that particulars and their intrinsic characteristics are intentional; later I consider this possibility. By a particular I mean “particular, individual, existing things or substances, mental or physical, or both (or perhaps neither), each of which has many characteristics, that is qualities, properties, or relations” (p. 14 of Baylis, C. “Introduction” to his Methaphysics (New York: Macmillan, 1965)Google Scholar). Different ontologies will give varying analysis of what these particulars are, and what sorts of intrinsic properties and relationships there are; our analysis allows any choice of ontology so long as the particulars in the world have intrinsic characteristics and enter into intrinsic relationships.

11 The above characterizations are by no means exhaustive of the possible compound complex particulars; whenever any logical operation on the intrinsic relationships characterizing complex particulars results in an intrinsic relationship, a compound complex particular of the sort exists. Examples of such logical operations include those of existential and constructive definition; cf. Smullyan, R. Theory of Formal Systems rev. ed. (Annals of Mathematical Studies, No. 47, Princeton: Princeton University Press, 1961), pp. 21–22CrossRefGoogle Scholar, 29-30, for characterizations of these forms of definition.

12 Notice I am not claiming that there are characteristics of characteristics, etc. I am only allowing the possibility so that my account may be as neutral as possible on this issue.

13 Cf. also Herbst, ibid., for a similar claim. I deny, however, the acceptability of tying facts to varifiability or direct confirmation in the way Herbst does.

14 Cf. “On Sense and Reference,” pp. 56–78 in Geach, P. and Black, M. (eds.), Translations from the Philosophical Writing of Gottlob Frege(Oxford: Basil Blackwell, 1960)Google Scholar; and “On Concept and Object,” ibid., pp. 42–55.

15 Against this claim it has been suggested to me that examples such as the following one show that referential equivalence (given the same logical form) is not enough: “The detective informed the reporters, ‘We have uncovered two important facts about the case: (1) Jones left the house at 5:30 and (2) the murderer left the house at 5:30’.” For assuming Jones is the murderer, on my account does it not follow that (1) and (2) assert the same fact, whereas it clearly is the case that they state different facts? But this objection misfires. For in the circumstances ‘the murderer’ clearly is being used in (2) as an attributive definite description (in the sense of K. Donnellan, “Reference and Definite Descriptions,” Philosophical Review, 75 (1966): 231-304); so (2) predicates both being a murderer and leaving the house at 5:30 of Jones, whereas (1) only predicates leaving the house at 5:30 of Jones. Thus, despite the fact that (1) and (2) have the same grammatical form, they do not have the same logical form (cf. Geach, op. cit., and my “Misidentification, Truth, and Knowing That,” Philosophical Studies, 23 (1972): 128-139Google Scholar, for further discussion of the difference between grammatical and logical forms of propositions); so (1) and (2) do not express the same fact by my criterion. More generally (in terms of developments which follow), they will not express the same fact since they do not have the same extensions.

16 I owe this suggestion to Wilfrid Sellars. In order to not have properties change with the demise of particulars, it is necessary to have names for all particulars. And since particulars may change, these names will have to be temporal names-i.e., will have to be names which apply to individuals only for particular times. Incidentally, we will say that an intrinsic characteristic is intensional if and only if it can only be specified in an intensional language.

17 This idea is Fregean; cf. Geach and Black, op. cit., e.g., p. 58. ‘S’ is taken as the logical subject, not the grammatical subject, of the proposition; cf. Geach, op. cit., pp. 27–29.

18 This definition can be construed as a variation on and extension of the definition of empirical truth for scientific theories I have presented in “On Partial Interpretation,” J. Philosophy, 68 (1972): 57-76, and also in “What's Wrong with the Received View on the Structure of Scientific Theories?” Philosophy of Science, 39 (1972): 1-19. A discussion how this is so is found in Section V of my “Theories, Their Formulations, and the Operational Imperative,” Synthese, 11 (1972): 101-136, where ‘factual truth’ is used in the sense of ‘empirical truth’ here.

19 For evidence that this is so, cf. my “Theories, Their Formulations, and the Operational Imperative,” op. cit., “Some Philosophical Problems in Biological Speciation and Taxonomy” (forthcoming in the proceedings of the First Ottawa Conference on the Conceptual Bases of Classification, held at University of Ottawa, October, 1971), and “Theories and Phenomena,” forthcoming in Theory and Decision.

20 For a proof that it does (albeit a controversial one), see Wiggins, D. Identity and Spatio-Temporal Continuity (Oxford: Blackwell's, 1967)Google Scholar, Section 1.2.

21 Note that their plausibility stems from the Individuation of Complex Particulars principle. They are not plausible in the alternative world just considered.

22 Thus the question which world to accept turns centrally on how one resolves the sort of issues raised by David Wiggins, op. cit. The analysis presented here has obvious applications to those issues.

23 For a characterization of intentional particulars, Cf. Sellars, W. and Chisholm, R. “Intentionality and the Mental,” pp. 597–639 in Feigl, H. et. al., Minnesota Studies in the Philosophy of Science, Vol. II (Minneapolis: University of Minnesota Press, 1958)Google Scholar.

24 Many sense data theories of perception apparently do, as do many idealisms. In Seeing and Knowing(Chicago: University of Chicago Press, 1969), Dretske, Fred commits himself to the existence of real hallucinatory rats (pp. 46–49Google Scholar)-which presumably are intentional particulars. At places Feyerabend, op. cit., and Kuhn, op. cit., apparently commit themselves to intentional particulars.

25 It is worth noting that, in its basic version, my analysis embodies a treatment of universals very much like that which Bambrough, Renford attributes to Wittgenstein in his “Universals and Family Resemblances,” Proceeding of the Aristotelean Society, Vol. LXI (1960-61): 207-222Google Scholar; reprinted in Pitcher, G. (ed.), Wittgenstein: The Philosophical Investigations (Garden City: Anchor Books, 1966)CrossRefGoogle Scholar. The relationships between Bambrough's analysis and mine are discussed in my “Some Philosophical Problems in Biological Speciation and Taxonomy,” op. cit.

26 Although Hanson often is interpreted as holding the same view as do Kuhn, and Feyerabend, in his Patterns of Discovery(Cambridge: Cambridge University Press, 1958), Chapter IIGoogle Scholar, he in fact holds a position embodying essentially the position I have just taken.

27 Despite possible problems with the substitutivity of truth-functional equivalents in epistemic contexts, proponents of the KK thesis typically allow the sort of move I have just made. E.g., cf. Section Ill of Hilpinen, R. “Knowing That One Knows and The Classical Definition of Knowledge,” Synthese, 21 (1970): 109-132CrossRefGoogle Scholar, and Ackermann, R. Belief and Knowledge (Garden City: Anchor, 1972)CrossRefGoogle Scholar, ch. 6.

28 Op. cit., Chapter 6, and p. 128.

29 Note that most advocates of the KK thesis (cf., e.g., the works cited in n. 27) maintain that I know that I know that ϕ if and only if I know that ϕ, I know that I believe that ϕ and I know that I am justified in believing that ϕ. This is, in effect, to assume that ‘ϕ’ and ‘ ‘ϕ’ is true’ state the same fact; and their arguments for the KK thesis essentially involve this supposition, The reason why the supposition is false on our analysis (given White's claims) is that two truth-functionally equivalent propositions do not always assert the same fact. These considerations therefore show that it is false that knowing and knowing that one knows are the same state, as many advocates of the KK thesis have suggested.

30 It is not clear to me whether this form of argument can be extended to show that the KK thesis fails for knowledge which is not factual knowledge of the world. Whether it can depends on at least the following issues: (a) Is White correct that all true propositions assert facts (cf. note 7)? (b) Is assumption (1) plausible for non-empirical knowledge? (c) How are facts expressed by non-empirical propositions individuated? The answers to (b) and (c) are obviously interconnected: e.g., if I know that even integers are divisible by two without remainder, it plausibly might be maintained that I thereby know that 2 is an even integer, that 4 is, that 6 is, etc., and hence if I know the first fact I know an infinite number of facts; and therefore that the answer to (b) is negative. But whether this is so depends, inter alia, on the answer given to (c).