Toward ‘perfect collections of properties’: Locke on the constitution of substantial sorts

Canadian Journal of Philosophy

Volumn 29, Issue 4, 1999, Pages 551-593

  Shapiro, Lionel ^a  

^a UNIVERSITY OF PITTSBURGH   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33744909006     PISSN: 00455091     EISSN: 19110820     Source Type: Journal    
DOI: 10.1080/00455091.1999.10715991     Document Type: Article

References (55)

1. Locke, John. 1975. An Essay Concerning Human Understanding Edited by: Nidditch, P. H., Oxford: Clarendon Press. book IV, chapter xii, §14: 648; also IV.xii.6–7: 642–3. The expression in my title appears in the marginal summary of III.vi.19–20:449. All italics in citations from the Essay are Locke's; boldface indicates my emphasis.
2. Joseph, Joseph. 1996. ‘Locke's Semantics and the New Theory of Reference to Natural Kinds,’. Locke Newsletter, 27: 41–64. Cf. at 44–6. Note Locke's semantic vocabulary: words ‘belong truly’ to things, which in turn ‘have the right to’ names. By contrast, it has long been recognized that neither ‘signifying’ nor ‘standing for’ is a reference relation (e.g. Norman Kretzmann, ‘The Main Thesis of Locke's Semantic Theory,’ Philosophical Review771968 175–96); the same holds for the relation between a word and its ‘meaning,’ defined as the ‘Idea it stands for’ (III.iii.10: 413). We will soon see that a substance-name is properly ‘referred to’ neither the idea it ‘signifies’ nor the class of those things to which it ‘belongs truly.’ Finally, a principal task for this paper will be to explain the relation between a Lockean natural kind and the name that is ‘used for’ it, the name by which it is ‘called.’
3. Essay,’ Philosophical Review, 488 For a lone dissent, see Martha Brandt Bolton, ‘Substances, Substrata, and Names of Substances in Locke's 85 (1976, 513. On Bolton's reading, the ‘confused Idea of Substance’ included as a component in each idea of a substantial sort (II.xii.6: 165, II.xxiii.3: 297, III.vi.21: 450) plays the role of an actuality operator. As an example, assume that the qualities rational, animal exhaust the constituent qualities in the abstract idea of man. Bolton denies that being a rational animal need be a sufficient condition for conforming to this idea. Instead, roughly, something in any possible world will conform to the idea provided it possesses all properties flowing from the explanatory ‘internal constitution’ shared by all actual rational animals (worries about ill-definedness are addressed, in my view superficially, at 507–8n45). It should soon become apparent that I share much of the motivation behind this reading. However, I will argue that we can embrace a strong version of Bolton's key insight (cf. 506–8) that the kind represented by a substance-idea is to some degree independent of the idea's descriptive content without insisting, in the absence of specific evidence, that Locke treats conformity to a nominal essence as modally rigid.
4. Essay,’ History of Philosophy Quarterly, 73 This habit, together with Locke's acknowledged use of ‘idea’ where we would expect ‘quality’ (II.viii.8: 134), may reflect more than an innocent ambiguity: see Jonathan Bennett, ‘Ideas and Qualities in Locke's 13 (1996, 88
5. 1991. New Essays on Human Understanding Vol. 2, 76Cambridge: Cambridge University Press. They were acutely perceived by Leibniz in his comment on III.xi.24 ([1703–5], trans, P. Remnant and J. Bennett (1981), 354). Discussing the same passage in Locke, Michael Ayers warns not to ‘import into his arguments contradictions and tensions that do not exist there’ (Locke: Epistemology and Ontology (London: Routledge vol.). I hope to vindicate Leibniz.
6. I follow Bolton in posing this question as a challenge to conventionalist readings (‘Substances, Substrata, and Names,’ 495–9).
7. 1991. 105–18. Here I endorse the customary view that Locke employs a notion of the ‘internal constitution’ or ‘real essence’ of a sort of substance, understood as an aspect of the internal constitution of each individual of the sort. While this assumption should garner support throughout my paper, David Owen's thesis that real essences and internal constitutions are possessed only by particulars merits response (‘Locke on Real Essence,’ History of Philosophy Quarterly8). Where Locke identifies a sense of ‘real essence’ that ‘relates to a Sort, and supposes a Species’ (III.vi.6: 442), Owen construes him as referring to the internal constitution of a particular, considered as the source of those qualities rendered essential by subsumption of this particular under a sort determined by a nominal essence (113–14). Other passages are resilient to Owen's strategy, though. Consider the parallelism of ‘the nominal Essence of Gold’ and ‘the real Essence’ (III.vi.2: 439), or Locke's claim that knowledge of the ‘Properties of Gold’ and other substances could be obtained if we had ‘specifick Ideas of their real Essences in our own Minds’ (IV.vi.11: 585; cf. IV.xii.9: 645). For ‘internal’ or ‘real constitutions’ of substantial sorts, see II.xxxi.10: 382, III.vi.3: 440, and IV.vi.15: 590. Locke is plainest in reply to Stillingfleet: while the particular ‘internal constitutions’ of finite things are alterable, the abstract ‘internal constitution or real essence of a species’ is Unchangeable’ (First Letter [1696–7], in Works (London, 1823), vol. 4 90–1). Still, he makes do wherever possible with a notion of real essence less liable to suggest a substantial form: the internal constitution of a particular. Indeed, I believe the entirety of chapter III.iii on ‘General Terms’ should be understood in this manner, as Owen reveals it can be. Owen's scrupulously nominalist reading of this chapter is certainly supported by its entirely different alignment of the perishable/permanent distinction with that between real and nominal essences (III.iii.19: 419).
8. Bolton's explication of Locke's semantics (summarized in note 3) does allow an object in a possible but non-actual world to ‘conform to our description of gold, but lack some of the properties typical of gold, and so fail to be gold’ (‘Substances, Substrata, and Names,’ 507–8).
9. Owen. ‘Locke on Real Essence,’ 111
10. A second passage in Book IV might appear to support the innocuous reading by implying that whatever secondary qualities ‘result from’ the ‘Constitution of the insensible parts’ of a substance ‘consequently must always co-exist with that complex Idea we have of it’ (IV.iii.11: 545). Though the point can't be fully argued here, I believe this conclusion would be unwarranted. In section IV.vi.7, Locke retrospectively summarizes two ‘Reasons mentioned, Chap. 3,’ clearly corresponding to the theses of sections IV.iii.11 and 12. The wording of this summary strongly suggests that when Locke refers in IV.iii.11 to additional qualities resulting from the ‘same Constitution of the insensible parts of Gold’ on which the qualities in our nominal essence depend, he does not have in mind qualities resulting from gold's real essence. Rather, he is referring to those secondary qualities (should there be any) necessarily possessed by anything that shares each of the various real constitutions underlying the qualities in the nominal essence. The ‘reason’ knowledge of the necessary coexistence of qualities was found to be unattainable is not that we lack knowledge of real essences, but that ‘we know not the real Constitutions of Substances, on which each secondary Quality particularly depends’ (IV.vi.7: 582).
11. Ideas Context establishes the ‘same subject’ as a sort, rather than a particular. A directly prior mention of ‘these simple that…are united in the same Subject’ refers back to ‘these [simple ideas], as united in the several Sorts of Things.’ Two sentence later, Locke supports his conclusion that people have ‘different Ideas of the same Substance’ by observing that ‘the Properties of any sort of Bodies [are] not easy to be collected.’ It is only in the next section that he raises the additional point that ‘any particular thing existing’ can be classified into various sorts.
12. Species The passage under discussion occurs as part of an enumeration of conditions whose obtaining would be ‘necessary’ for us to be able to ‘distinguish substantial Beings into, according to the usual supposition, that there are certain precise Essences or Forms of Things, whereby all the Individuals existing, are, by Nature, distinguished into Species’ (III.vi.14: 448). Hence advocates of the innocuous reading might dismiss Locke's complaint about how ‘hard it is’ to express the imperfection of nominal essences as merely drawing a consequence of that mistaken ‘usual supposition’ (cf. LaPorte, ‘Locke's Semantics,’ 50–5). However, it would be farfetched not to regard the complaint as pronounced in propria voce. After all, it isn't the expressive predicament that Locke cites as a difficulty for his opponent. Rather, it is the thesis he is trying to express (‘Our nominal Essences of Substances, not perfect Collections of Properties’), one whose obtaining is not predicated on his opponent's mistaken supposition.
13. The maneuver is futile as long as the semantic ascent appealed to is the familiar variety entailing the identity of gold with the substance called ‘gold’. If I am right, though, Locke is best viewed as implicitly trying to employ the phrase ‘the substance called “gold”’ in a sense that won't license this ‘disquotational’ identity. In Section IV.2, I sketch an account of what that sense could be.
14. For further examples of Locke's nonconstitutive use of ‘determine that Species’, see III.xi.20: 519 and IV.xii.14: 648. The subject of the above passage must be the sort, not the particular ring: nothing is ‘inseparable’ from an individual considered as such, nor does it possess ‘properties’ (III.vi.6: 442).
15. Troyer. ‘Locke on the Names of Substances,’ 28
16. Mackie, J. L., 1993. Problems from Locke 24–5. (Oxford: Clarendon Press 1976), 88. See also Hilary Kornblith's chapter on ‘Locke and Natural Kinds’ in Inductive Inference and its Natural Ground: An Essay in Naturalistic Epistemology (Cambridge, MA: MIT Press
17. Mackie. Problems 97
18. Ayers. 1994. Locke Cambridge: Cambridge University Press. 2: 68. See also §IV of Paul Guyer, ‘Locke's Philosophy of Language,’ in The Cambridge Companion to Locke, V. Chappell, ed. (
19. 1975. Mind, Language and Reality: Philosophical Papers, Volume 2 239Cambridge: Cambridge University Press. See e.g. ‘The Meaning of “Meaning”,’ in
20. Kornblith does address Locke's argument regarding specific differences in nature. But where Locke ‘appears to be offering a conceptual argument’ that specific differences are unintelligible apart from our abstract ideas, Kornblith seeks to reinterpret this as the ‘empirical argument’ that our beliefs about such differences can only be ‘explained’ by appeal to ‘our own conceptual activity’ (Inductive Inference, 45–7). Despite advancing this argument that the postulation of ‘real kinds in nature’ is explanatorily idle, Kornblith adds, Locke elsewhere embraces such kinds (16–17, 23–5). What Locke denies, on this reading, is that real kinds can serve to regulate taxonomic practice. By contrast, I will argue that to the extent that Locke does embrace mind-independent kinds, it is precisely in virtue of his explicit recognition of their regulatory role.
21. III.iii.17: 418, III.vi.14–18 [summary]: 448, IV.iv.13: 569; see also III.VI.27: 454, III.x.21: 502.
22. Locke My account of Locke's reasoning owes much to Ayers 2: 71–3); where I will depart from him is in stressing this argument's compatibility with a recognition of naturally privileged kinds. It is in terms of Locke's skepticism about the utility of an absolute accident/property distinction that we must understand two passages seemingly implying that individuals bear properties, contrary to Locke's express denial. ‘If things were distinguished into Species, according to their real Essences,’ Locke argues a mere page after that denial, ‘it would be as impossible to find different Properties in any two individual Substances of the same Species, as it is to find different Properties in two Circles, or two equilateral Triangles’ (III.vi.8: 443, also III.iii.17: 418). Considered in the context of the reductio he is offering, his intent should be clear. Supposing quantities of vitriol are found to ‘betray Qualities so different from one another, as to frustrate the Expectation and Labour of very wary Chymists,’ Locke sees no reason why the chemists may not recognize a narrower species possessing as a property one or more of the qualities at issue (to be regarded either as a subspecies of vitriol or perhaps as true vitriol). Were the extension of the name they use assumed to constitute a Scholastic infima species, this move would be arbitrarily precluded.
23. Locke Newsletter, 45 Two attempts at doing justice to Locke's anti-conventionalist strand without presupposing essentialist metaphysics are Bolton's reading discussed in note 3 and a reply by Ruth Mattern (‘Locke on Natural Kinds as the “Workmanship of the Understanding”,’ 17 (1986, 92, esp. 68). Mattern offers a clear statement of the seeming tension in Locke, but I am not convinced that her proposed resolution adequately respects the anti-conventionalism.
24. Locke. 1990. Drafts Vol. 4, 27–55. Draft B, §75: 183, §75: 181n, and §84: 191. See also n.12. Pauline Phemister points to the traditional nature of Draft B's discussion of natural kinds in ‘Real Essences in Particular,’ Locke Newsletter21 at 28–9. Passages like these may explain Locke's remark to Molyneux about the ‘difficulty I often found my self under when I was writing of that subject [of species], where I was very apt to suppose distinct species I could talk of without names’ (letter of 20. January 1693, #1592 in The Correspondence of John Locke, E.S. de Beer, ed. (Oxford: Clarendon 1976-), vol.626).
25. Locke. Drafts Draft B, §72: 176–7
26. Bolton. See ‘Substances, Substrata, and Names,’ 500–1, 503–5. In restricting his claim to the species of ‘natural’ substances, Locke acknowledges that the nominal and real essences of a sort of artifact can coincide, e.g. in the mind of the artificer (see III.vi.40: 464–5). The fact that a sort's possession of its properties rests on a foundation other than the combination of qualities in its nominal essence may not actually suffice to establish a divergence between nominal and real essences. For this, Locke also seems to require that we intend to hold the nominal essence responsible to something it only ‘inadequately’ represents (either the foundation itself or an unknown number of properties flowing from it). In the case of modes, neither of these two conditions is satisfied: ‘all the properties of the Species’ depend on the nominal essence, whence there is nothing else we ‘would have express'd’ by the name (III.V.14: 436). Secondary qualities might present a case in which the first condition for divergent essences is satisfied (cf. Mackie, Problems, 90), but here the second condition clearly is not (II.xxxi.12: 383).
27. Kitcher, P., and Salmon, W., eds. 1989. Scientific Explanation 424–5. Minneapolis: University of Minnesota Press. See e.g. Philip Kitcher, ‘Explanatory Unification and the Causal Structure of the World,’ in
28. Mancosu, Paolo. 1996. Philosophy of Mathematics and Mathematical Practice in the Seventeenth Century New York: Oxford University Press. See ch. 1. While the above analysis of Locke's example accords with the semantically innocuous reading, the same analysis is consistent with its rejection. Leibniz, for instance, invokes naturally privileged (‘physical’) species whose boundaries are imperfectly approximated by our ‘provisional’ nominal definitions (New Essays, 312, 325, 400ff). Nonetheless, he presents a nearly identical example, explaining that one can define the parabola in terms of an ‘external feature’ and subsequently devise a ‘more perfect idea’ by adding a further such property, all the while unaware of ‘the figure's inner essence’ that serves as a ‘key to further knowledge’ of its properties (346, 402) and explains why they obtain (cf. 295). Given his use of this example as an objection to Locke's characterization of the substance/mode distinction, it is ironic that Leibniz fails to recognize Locke's own use of the same example to concede the contingency of the fact that our modal ideas serve as archetypes and as such ‘cannot but be adequate Ideas’ (II.xxxi.3: 376). Cf. Roger Woolhouse, ‘Locke's Theory of Knowledge,’ in The Cambridge Companion, 160.
29. Ayers. Locke 2: 67
30. Ayers. Locke 2: 75; see also Guyer, ‘Locke's Philosophy of Language,’ 143. For Locke's distinction between ‘civil and ‘philosophical Discourse,’ see III.ix.3: 476 and III.ix.15: 484.
31. Ayers never claims it does: it isn't clear he sees any connection between Locke's advocacy of natural history and his thesis that our substance-ideas are ‘imperfect and inadequate’ (see esp. Locke, 2: 76).
32. Ayers. 1984. “ 13 ”. In Locke 273 2: 67–8 (my italics), 38, 70; see also Phemister, ‘Real Essences,’ 35–6, 45–6; Peter Alexander, Ideas, Qualities and Corpuscles: Locke and Boyle on the External World (Cambridge: Cambridge University Press chapter esp. R.S. Woolhouse, Locke's Philosophy of Science and Knowledge (New York: Barnes & Noble 1971), 118–19.
33. 1991. Locke 127–48. Ayers does note this qualification: ‘In general there is the implicit suggestion that repeated observation of coexisting qualities and powers is requisite to justify the presumption of a recurrent underlying cause of their union, and so to justify the formation of a complex idea’ 2: 79–80). See also Woolhouse, who makes the further claim that correlations are at issue (Locke's Philosophy of Science, 117–18, 132). In neither case does the restriction constitute a general abandonment of the traditional picture: neither author connects the failure of a nominal essence to be determined by a real essence with its ‘imperfection’ or ‘inadequacy,’ a defect both appear to account for via the semantically innocuous reading (Woolhouse, Locke's Philosophy of Science, 120; Ayers, Locke, 1: 102, 2: 76). I know of only one reader of the Essay who has both recognized the role of correlations and suggested a connection between microstructural uniformity and ‘adequacy’: Richard Boyd, ‘Realism, Anti-Foundationalism, and the Enthusiasm for Natural Kinds,’ Philosophical Studies61 at 130–3.
34. Drafts Compare the slightly weaker condition in Draft B: An Englishman's ‘complex Idea of a Swan is a kinde of affirmation that where such a kinde of shape colour bignesse with such a necke & legs doth exist there also whole feet are joynd with them or such a kinde of voice as that of a swan is. i.e. where a great number of those simple Ideas doe exist togeather that the rest are also’ (Locke, Draft B, §63: 166–7; see also Draft A, §1: 4–5).
35. Problems Mackie allows that Locke ‘hints vaguely at this sort of progress’ 98n), and Kornblith sketches a similar account of Locke's implicit understanding of ‘chemical method’ (Inductive Inference, 26–8), one he says ‘flatly contradicts’ Locke's ‘official position’ that ‘real kinds’ can exercise no constraint on taxonomic practice (17). Besides my denial that Locke espouses that position, the main discrepancy between our readings lies in Kornblith's insistence, based on a failure to distinguish Locke's ‘properties’ from ‘qualities,’ that any difference in qualities implies two thing can't be members of the same ‘real kind’ (26–8, 36–7). This may explain his attribution to Locke of the view that ‘[w]ere we in a position to observe [corpuscular structure], we would see directly how it is that nature divides the world into kinds’ (17).
36. Locke's Philosophy of Science I don't wish to imply that the characterization of substance-ideas as those intended to represent kinds with causally explanatory real essences fully accounts for Locke's substance/mode distinction. Following Woolhouse and others, I suspect the distinction involves a separate and potentially orthogonal ontological strand chs. 4 and 7; Mackie, Problems, 99–100; Ayers, Locke, 2: ch. 8). Whether and how each of these strands is associated with the idea of ‘substance in general’ are vexed questions beyond this paper's scope.
37. I have adapted this term, as well as the idea of a semantic ‘revision procedure,’ from Anil Gupta's writings on truth.
38. In particular, I see little reason why Locke wouldn't allow the expansion of nominal essence by a number of qualities simultaneously, should no single quality be discoverable.
39. Drafts Locke gives a more detailed explanation of the ‘mistake & error‘, in Draft B, §93i: 205–6 (cf. also III.vi.29: 456).
40. The present reading thus lends support to Boyd's conjecture that Locke's solution to the ‘problem of the “inadequacy” of kinds of substances’ is intended to apply to the ‘problem of inductive categories’ as well (‘Enthusiasm for Natural Kinds,’ 131–2).
41. Locke See e.g. Ayers, 2: 67–8; Guyer, ‘Locke's Philosophy of Language,’ 131; Wool-house, Locke's Philosophy of Science, 101.
42. Mackie. Problems 87
43. Ayers. Locke 2: 71
44. Locke Ayers rejects any such reading 2: 76), whence it is unclear how he thinks the imperfection of our substance-ideas could even be relevant to Locke's claim that our species-boundaries wouldn't match supposed natural ones. I suspect the reason Ayers doesn't notice this difficulty lies in his neglecting to distinguish between the ‘imperfection’ of substance-ideas at issue here (their ‘inadequacy’ or failure to perfectly represent their archetypes) and the ‘imperfection’ of substance-names (the uncertainty of their signification).
45. Species It may appear that Locke is not contrasting two imperfections that are always present, but rather our normatively imperfect ideas of the ‘lowest’ with our designedly imperfect ideas of the ‘more comprehensive Classes’ (III.vi.32: 459). Clearly, Locke doesn't restrict designed imperfection to higher genera: just as the ‘the Genus, or more comprehensive, is but a partial Conception of what is in the Species,’ so also the ‘Species [is] but a partial Idea of what is to be found in each individual’ (460). But might he not countenance a privileged category of ‘specific’ nominal essences, conceived as the only ones to which the normative notion is applicable, perhaps those intended to represent naturally lowest species? (This is Liebniz's view of definition in the New Essays, 401–2.) The passage I am about to display confirms the conclusions of Section II.4 in denying ‘just Authority’ to our privileging of any level of generality: Locke's ‘specific’ ideas are distinguished only relative to a stage of inquiry. If Locke does mean to restrict his inadequacy thesis to just these substance-ideas, it would have to be for a pragmatic reason, e.g., to ensure that idea-perfection preserves our current hierarchy's subsumption-relations.
46. Moreover, it may not be easy to correct this ‘oversight.’ The presumption that Locke's revision procedure will converge on some natural kind would stand in need of justification, and may conflict with belief in the plenitude of such kinds.
47. Chappell, V., ed. 1998. “ ‘The Relevance of Locke's Theory of Ideas to his Doctrine of Nominal Essence and Anti-Essentialist Semantic Theory’ (1992), reprinted in ”. In Locke 214–25. Oxford: Oxford University Press. See Martha Brandt Bolton
48. This identification may appear to fly in the face of Locke's contrast between the claim that ‘what has the real Essence of Gold is malleable’ and the mere claim that ‘what I call Gold is malleable’ (III.x.17: 499). The latter, however, is not intended by Locke as an exploitation of semantic ascent. Parallel passages reveal it to be the claim that malleableness ‘is part of the Definition, part of the nominal Essence the Word Gold stands for’ (III.vi.50: 470, IV.viii.5: 612–3, IV.vi.9: 583).
49. Bolton. Cf. ‘Substances, Substrata, and Names,’ 509.
50. Donnellan. 1983. “ ‘Kripke and Putnam on Natural Kinds,’ in ”. In Knowledge and Mind Edited by: Ginet, C., and Shoemaker, S., 102Oxford: Oxford University Press. Wilson, ‘Predicate Meets Property,’ Philosophical Review91 (1982) 549–89, passim.
51. Would Locke abjure criticism if the Jamaican instead started calling ice ‘glass’? Here the response ‘But ice and glass have different internal constitutions!’ would have greater bite: it isn't clear there exists a Lockean natural kind toward which the man's word might be interpreted as pointing. Interestingly, Locke's example traces to the early drafts, where he uses it to strikingly different effect: he illustrates people's liability to ‘mistake the meaneing of words by their unacquaintednesse with the things themselves’ by imagining a Jamaican who mistakenly calls frozen water ‘glasse or chrystall or stone’ (Locke, Drafts, Draft A, §2:9, Draft B, §84: 192)! If Locke's positive aim in III.vi.13 were to explain that the proper assignment of meanings is unconstrained by the natural kinds of things, there would have been no need to change the example—in fact, the revised scenario would be highly tendentious.
52. Donnellan. 1999. ‘Kripke and Putnam,’ 103–4; Wilson, ‘Predicate Meets Property,’ 553. For a bullet-biting defense of a ‘diachronic division of linguistic labor,’ cf. Henry Jackman, ‘We Live Forwards but Understand Backwards: Linguistic Practices and Future Behavior,’. Pacific Philosophical Quarterly, 80: 157–77.
53. Wilson. ‘Predicate Meets Property,’ 582, 578
54. 1993. The Advancement of Science: Science without Legends, Objectivity without Illusions 171–3. New York: Oxford University Press. Responses of a similar shape have not been uncommon in the philosophical literature since Locke. One current proponent points to ‘a line of thought that runs from Kant through Peirce to recent writers such as Sellare and Putnam’ (Philip Kitcher,).
55. I have profited from discussions with and comments by Joseph L. Camp and Elizabeth L. Jockusch, as well as from suggestions by anonymous referees. An ancestor of this paper was written while I was supported under a National Science Foundation Graduate Fellowship.