1 ‘Quasi-historical’, because they are at best intended to capture philosophical lore, not historical fact.

2 Mackie, however, has argued that the modem view also captures a feature of Locke's theory, namely its assumption that real natures exist, and that they might be referred to even though they could never be known. Mackie, John ‘Locke's Anticipation of Kripke,’ Analysis 34 (1974) 177-78.CrossRefGoogle Scholar But seen. 13 below.

3 See Kripke, S. ‘Naming and Necessity,’ in Davidson, D. and Harman, G. eds., Semantics of Natural Language, (Dordrecht, Netherlands: D. Reidel 1972) 253–355;CrossRefGoogle Scholar Putnam, H. ‘The Meaning of “Meaning”,’ in Mind, Language, and Reali ty: Philosophical Papers, vol. 2 (Cambridge, UK: Cambridge University Press 1975).CrossRefGoogle Scholar A good general account of the Modern View is provided by Schwartz, Stephen P. ‘Natural Kind Terms,’ Cognition, 7 (1979) 301-15.CrossRefGoogle Scholar For a far more detailed account, see Salmon, Nathan U. Reference and Essence (Princeton, NJ: Princeton University Press 1981.)Google Scholar

4 Cf. Salmon, 86, fn.

5 See Browning, Douglas ‘Presidential Address: Believing in Natural Kinds; South West Journal of Philosophy, 9 (1978) 135-48CrossRefGoogle Scholar: ‘Natural kinds, if there are any, are natural because each thing of that kind is of just that kind and not another kind by virtue of what that thing is on its own and apart from whatever decisions may be made by a sorter … its qualifications for occupancy [of the natural kind] would not easily include reference to another of that kind or to anything outside that kind, including our own interest as investigators.’ (136-7).

6 Cf. Brody, B. Identity and Essence (Princeton, NJ: Princeton University Press 1980),CrossRefGoogle Scholar discussed by Linsky, B. in ‘Is Transmutation Possible?', Philosophical Studies, 41 (1982) 367-81.CrossRefGoogle Scholar

7 See Ayers, Michael ‘Locke versus Aristotle on Natural Kinds,’ Journal of Philosophy, 78 (1981) 247-72,CrossRefGoogle Scholar esp. 256.

8 See Schwartz, Stephen P. ‘Formal Semantics and Natural Kind Terms,’ Philosophical Studies, 38 (1980) 189-98,CrossRefGoogle Scholar esp. 193, for references to supporters of Stability and arguments against it.

9 Arguments for the latter view and against Brody are to be found in Linsky (ibid.). Price, Marjorie in ‘Identity through Time,’ Journal of Philosophy, 74 (1977) 201-17,CrossRefGoogle Scholar claims equipollence in the sense that no natural kind is stable.

10 Actually it amounts to much the same problem as what Plato called Meno's Paradox (Plato, Meno, 80 ff.).

11 See Ayers. My ‘quasi-historical’ sketch above owes all but its crudity to that excellent article. Ayers argues that Locke's view was plausible at a time when biology and chemistry were as much of a mess as sociology is today, but that he was wrong about biology because ‘[c]ontrary to his principle (cf III, vi, 50; IV, vi, 4) reality can supply the boundary to the denotation of a word by supplying a rough boundary’ (Ayers 269). Locke was wrong about chemistry, because ‘the discovery of elements and chemical combination … has confirmed the ordinary or primitive view of natural kinds with independent sharp edges, refuting Locke's alternative picture of a chemical world with no “Chasms or gaps“’ (ibid.).

12 See also Derjaguin, Boris V. ‘Superdense Water.’ Scientific American, 223 (1970) 52–71,CrossRefGoogle Scholar on the hypothesis that water contains another kind, ‘water II.’ with different polimerization properties.

13 Evan Fales has argued in ‘Natural Kinds and Freaks of Nature.’ Philosophy of Science, 49 (1982) 67-90, that only fundamental entities (and not composite or ‘derivative’ entities such as biological taxa) can truly be said to form natural kinds. His argument rests partly on the existence of biological ‘freaks of nature.'

14 Bernard Linsky (ibid.) has argued that transmutation of a member of one natural kind into a member of another is sometimes physically possible and sometimes not. Thus the transformation of Galatea into a living being is physically impossible, but that of lead into gold is not. But surely the differences between these impossibilities is merely one of degree: the line between technological fiction and science fiction is not clear.

15 But what if quarks are the ultimate particle, and come in just four types? Then perhaps there are indeed just four natural kinds, satisfying most of the requirements (i)-(vi) if not (vii). But this possibility is not of great comfort to the Modern View. For the drive to unity in science is likely to leave us with the conviction that if there are just four kinds of quarks, then that needs to be explained, in terms of something of which there is only one. So it seems we won't be satisfied until requirement (ii) is violated.

16 Paul Churchland has recently argued that Thinker’ may indeed be a natural kind, but his argument concerns thinkers as biological organisms. His argument is based precisely on the rejection of a functional characterization of any proper natural kind. See Paul Churchland, ‘Is Thinker a Natural Kind?,’ Dialogue, 21 (1982) 223-38.

17 Philosophers disagree as to whether biological species should be counted in. See e.g., Dupre, John ‘Natural Kinds and Biological Taxa,’ Philosophical Review, 90 (1981) 66–90,CrossRefGoogle Scholar who argues that the theory of natural kind terms proposed by Putnam and Kripke is not applicable to species, because biological terms are not at the right level of theory. Others take it for granted that species are natural kinds. See e.g., Michael Ruse, ‘Species: Natural Kinds, Individuals, or What?’ (unpublished paper presented at the Simon Fraser Conference on Natural Kinds, February 1983).

18 Putman, D. ‘Natural Kinds and Human Artifacts,’ Mind, 91 (1982) 418-19.CrossRefGoogle Scholar

19 See Putnam, H. ‘Minds and Machines,’ in Anderson, A. ed., Minds and Machines (Englewood-Cliffs, NJ: Prentice-Hall l964) 72–97.Google Scholar

20 This might not be true, as Ayers observes, if the fundamental properties of matter were observable (if extension for example remained the fundamental property of the natural kind ‘Matter’). But the perspicuity of fundamental properties quickly became lost, since Locke, in the complexities of sub-microscopic theory. ‘[N]either the fundamental properties now ascribed to primary particles by present-day physics nor, therefore, the particles themselves could be either empirically identified or conceived independently of the general law governing the behavior of those particles.’ (Ayers, 255).

21 It may, of course, function so on the appropriate de re reading. But the standard doctrine is that it does not so function when it is read as referring to an open class.

22 Cf. Goodman, Nelson. Fact, Fiction, and Forecast, 2nd ed. (New York: Bobbs Merril Co. 1965) 20Google Scholar: ‘the first is accepted as true while many cases of it remain to be determined, the further, unexamined cases being predicted to conform with it. The second …., on the contrary, is accepted as a description of contingent fact after the determination of all cases, no prediction of any of its instances being based upon it.'

23 And what of the rate of explosion in the original Big Bang? John Leslie has recently argued that had this parameter been different in minuscule proportion the whole course of the universe would have been different enough to preclude the possibility of life, for example. From this it would follow that in the possible world where the Big Bang took a fraction of a second less time - surely a contingent matter - there would have been no biological laws at all - unless there can be laws that never have instances. Whatever (if anything) explains the rate of the Big Bang, then, must be deeper than any biological (and at least some chemical?) laws. See John Leslie, The Need to Explain Life’ (paper read to the Canadian Philosophical Association, June 1982). See also Gale, George The Anthropic Principle,’ Scientific American 245 (1981) 154-65.CrossRefGoogle Scholar

24 This is related to what Goodman calls ‘cotenability’ (Goodman, 15).

25 That is, since Goodman himself: ‘projectibility … may give us a way of distinguishing ‘genuine’ from merely ‘artificial’ kinds, or more genuine from less genuine kinds, and thus enable us to interpret ordinary statements affirming that certain things are or are not of the same kind, or are more akin than certain other things (Goodman, 121).

26 An earlier version of this paper was presented at a Conference on Natural Kinds at Simon Fraser University, February 1983. I am grateful for comments from Mohan Matthen, delivered at that conference. Thanks also to Kathryn Morgan and Bernard Katz.