Explanation - Oxford Scholarship Jump to ContentJump to Main Navigation
John Locke and Natural Philosophy$

Peter R. Anstey

Print publication date: 2011

Print ISBN-13: 9780199589777

Published to Oxford Scholarship Online: May 2011

DOI: 10.1093/acprof:oso/9780199589777.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 27 June 2016



(p.153) 8 Explanation
John Locke and Natural Philosophy

Peter R. Anstey (Contributor Webpage)

Oxford University Press

Abstract and Keywords

This chapter argues that, on the whole, Locke adheres to the principles of explanation and constraints on explanation that were typical of proponents of the mechanical philosophy and the corpuscular theory of matter. The chapter concludes by showing that there is little place for laws of nature in Locke's views on explanation in natural philosophy and that he did not incorporate Newton's notion of principles of natural philosophy into the Essay.

Keywords:   Boyle, contact criterion, Descartes, Familiarity Condition, intelligibility, laws of nature, Newton, reduction, simple ideas, superaddition

It is usuall in things we doe not comprehend as a cover of our ignorance for the explaining what we doe not understand to substitute something as unintelligible & hard to be explaind with this pretence that without such a supposition it can not be explained. whereby we are made much vainer in an opinion of our Knowledg though not one jot the more Knowing.

John Locke1

In the previous two chapters we examined Locke's vision of a corpuscular metric of the unobservable entities that make up everyday material bodies, and we saw just how Locke accommodated demonstrative reasoning in his method for natural philosophy. No doubt Locke came to believe that the corpuscular metric would ideally produce a kind of analogue to what Newton had achieved in celestial physics. It would include reasoning according ‘Lockean demonstration’ from natural philosophical principles derived from observation. In this chapter we turn to Locke's views on the nature of the explanation of the phenomena we experience in the external world. The issue is important, because an analysis of his views on empirical explanation is the final piece in the exposition of Locke's views on the nature and prospects of the science of natural philosophy. We will proceed by examining some principles of explanation to which Locke is committed and some constraints on the application of those principles which seem to be implied in his discussions of natural philosophical reasoning. The principles and their constraints, when taken together, enable us to construct a coherent picture of Locke's view of natural philosophy.

Lockean principles of empirical explanation

Locke does not have a theory of empirical explanation. Rather, what we find in the Essay and elsewhere is a wide range of claims about the nature of properties, matter, (p.154) and the phenomena of nature, all of which reveal that Locke is committed, implicitly at least, to two principles of explanation. The first and most explicit explanatory principle is the Contact Criterion:

Contact Criterion—all change in the material world, so far as we can conceive of it, occurs by contact of bodies in motion (apart from gravitational attraction).

Locke states this explicitly in the Essay. In the first three editions (E 1–3 II. viii. 11), Locke says:

The next thing to be consider'd, is, how Bodies operate one upon another, and that is manifestly by impulse, and nothing else. It being impossible to conceive, that Body should operate on what it does not touch, (which is all one as to imagine it can operate where it is not) or when it does touch, operate any other way than by Motion.2

In his subsequent correspondence with Stillingfleet, Locke acknowledged that Newton's claims about gravity in the Principia had prompted him to change his view:

It is true, I say, ‘that bodies operate by impulse, and nothing else’. And so I thought when I writ it, and can yet conceive no other way of their operation. But I am since convinced by the judicious Mr. Newton's incomparable book, that it is too bold a presumption to limit God's power, in this point, by my narrow conceptions. The gravitation of matter towards matter, by ways inconceivable to me, is not only a demonstration that God can, if he pleases, put into bodies powers and ways of operation above what can be derived from our idea of body, or can be explained by what we know of matter, but also an unquestionable and every where visible instance, that he has done so. And therefore in the next edition of my book I shall take care to have that passage rectified.3

He did rectify the passage by removing the clause ‘and nothing else’ and by changing it from a claim about body–body interaction to a claim about how bodies produce ideas in us.4 But he left other statements of the contact criterion untouched, as we see for example just six sections later, where he claims that the pain and sickness caused by manna

are confessedly nothing, but the effects of its operations on the Stomach and Guts, by the size, motion, and figure of its insensible parts; (for by nothing else can a Body operate, as has been proved:).5

In fact, Locke even made a countervailing change to his concession to Stillingfleet in the fourth edition, at E IV. x. 19. There he changed ‘[w]e cannot conceive how Thought (or any thing but motion in Body) can move Body’ to ‘we cannot conceive how anything but impulse of Body can move Body’ (taken also in the fifth edition).6 Thus it is not entirely clear just how far Locke really modified his views in the light of Newton's discoveries about gravity.

(p.155) What is clear is that he was committed to the Contact Criterion in a fairly strong form. This criterion is a central tenet of the mechanical philosophy, as espoused by the likes of Descartes and Boyle.7 It is, in effect, a criterion of causal explanation which privileges efficient causation over other forms of causation, though Locke does not present it as such.

A second principle of explanation to which Locke is committed can be stated as follows:

The Reduction Principle—all observable natural phenomena can, in principle, be explained by the small number of qualities possessed by the imperceptible parts of matter.8

Locke nowhere in the Essay explicitly states that this is an explanatory principle, but there is ample evidence that this principle is implicit in his articulation both of the primary/secondary quality distinction and of the nominal/real essence distinction. In Book II, chapter 8 (and elsewhere), Locke speaks of secondary qualities as powers which ‘depend on’, ‘flow from’, ‘spring from’, and are ‘produced’ by the primary qualities of the imperceptible parts of matter that make up the relevant object.9 Yet the clearest statement of this Reductive Principle is found in the last sentence of Locke's Elements of Natural Philosophy, composed in the late 1690s:

By the figure, bulk, texture, and motion, of these small and insensible corpuscles, all the Phenomena of bodies may be explained.10

The principle is reductive in so far as a large range of qualities and phenomena are claimed to have their ontological ground in a small set of underlying qualities, and this ontological reduction is explanatory.

It is well known that these two principles are central tenets of the mechanical philosophy of Descartes and Boyle. However, it is unlikely that they are sufficient to mark a commitment to the mechanical philosophy. An additional explanatory principle is required; and that is the claim that all explanations of natural phenomena are to be made by analogy with the functioning of machines. With respect to the machine analogy, Locke does use some common mechanical tropes, in the Essay and elsewhere;11 but, unlike Boyle, he does not seem to employ the machine analogy as an explanatory device.

Nevertheless, it is clear that he was well disposed to the corpuscular hypothesis (as distinct from the mechanical philosophy). Here the work of Lisa Downing has, (p.156) I believe, gone a long way to establishing that, while Locke was not an unequivocal adherent of the corpuscular matter theory, he did regard it as the most intelligible hypothesis in terms of the clarity of its primary concepts and of its explanatory potential.12 Wherein does its explanatory potential lie? Why does Locke give the example of ‘the corpuscularian Hypothesis, as that which is thought to go furthest in an intelligible Explication of the Qualities of Bodies’?13 As Downing has pointed out,14 the answer lies in Locke's theory of material qualities.

The centrepiece of Locke's theory of qualities, the primary and secondary quality distinction, maps nicely onto a similar distinction, adumbrated by Boyle, between the mechanical affections and the non‐mechanical or secondary qualities of bodies. Now Boyle's theory of material qualities was central to his corpuscularian hypothesis and, in particular, it was tied to his account of the explanatory superiority of that hypothesis. For Boyle as for Descartes, the new theory of qualities required an inversion of the explanatory direction of the traditional theory of qualities. Instead of the primae qualitates of bodies (hot, cold, wet, and dry), which were used to explain the elemental make‐up of bodies and their peculiar accidents, the new mechanical affections of shape, size, motion, and texture became the explanans. Thus, for the corpuscularians, the properties of hot, cold, wet, and dry were to be explained in terms of the mechanical affections of shape, size, and motion.15 So by adopting, and even by developing, the corpuscularian theory of properties in the drafts and then in the Essay, Locke implicitly committed himself to a new approach to explanation.

There is, however, another point that needs to be made with regard to the Reduction Principle, and this has to do with one particular property of the micro‐structure of the unobservable parts of bodies. Numerous comments that Locke makes throughout the Essay and elsewhere imply that he is committed to a particular form of the Reduction Principle, a form that has been dubbed by Ernan McMullin ‘structural explanation’.16 Locke's talk of the possibility of God superadding thought to matter ‘fitly disposed’ is perhaps the most notorious of the many passages in which he reveals a commitment to the view that the powers of material bodies derive from the structural arrangement of their constituent parts.17 The technical term for this property is ‘texture’, and Locke does include texture in a number of his lists of primary qualities.18 Moreover, structural explanations were championed by the mechanical philosophers and were a natural mode of explanation for those who emphasized the efficacy and comprehensiveness of explaining natural phenomena by analogy with the functioning of machines. We have seen that Locke does not fully appropriate the machine metaphor as an explanatory principle; yet his commitment to structural explanation (p.157) was such that he used the underlying structure of functional systems as a criterion of the preservation of identity in living things over time.19

Lockean explanatory constraints

In addition to the Contact Criterion and to the Reduction Principle, Locke appears to be committed to two very important explanatory constraints on empirical explanation. The first is what I have elsewhere dubbed the Familiarity Condition:20

Familiarity Condition—all explanations of the unobserved must be made in terms of properties and causes with which we are familiar.

Before we examine the textual warrant for the claim that Locke implicitly adhered to this condition, it is worth drawing attention to just how important this thesis was to some of Locke's closest natural philosophical mentors. When Robert Boyle discusses the nature of natural philosophical explanation in The Origine of Forms and Qualities, (According to the Corpuscular Philosophy) (hereafter Forms and Qualities) he speaks in the following terms:

to explicate a Phænomenon, being to deduce it from something else in Nature more known to Us, then the thing to be explain'd by It.21

Again, in his Defence of the Doctrine Touching the Spring of the Air, he claims: ‘to explain a thing is to deduce it from something or other in Nature more known’.22 Boyle's younger contemporary Isaac Newton agreed. His third rule of philosophizing, given in the second edition of the Principia, says,

Those qualities of bodies that cannot be intended and remitted and that belong to all bodies on which experiments can be made, should be taken as qualities of all bodies universally.23

Amongst the claims which Newton is making here and in his extensive elaboration of this rule is the claim that the properties of the unobservable parts of matter are to be inferred by analogy with those of the observable parts of matter. A manuscript comment on this rule is revealing: ‘This seems to be the foundation of all Philosophy. For otherwise one could not derive the qualities of imperceptible bodies from the qualities of perceptible (bodies)’.24

What is more interesting still is that in Locke's copy of the Principia, which was given to him by Newton, there is an annotation on p. 402 ‘correcting’ Hypothesis III. This is an ancestor of what was to become Rule III in the second edition of 1713. It says: (p.158)

Hypoth. III. The qualities of bodies that cannot be intended and remitted, and that belong to all bodies in which one can set up experiments, are the qualities of bodies universally.25

Unlike Rule III, this is a strongly metaphysical claim. Instead of the injunction that universal qualities of observable bodies ‘should be taken as’ qualities of those bodies that we cannot observe, Newton asserts that they are the universal qualities of bodies. This, as far as we know, is the only version of Rule III that Locke ever saw and, while it post‐dates the publication of the first edition of the Essay, it certainly reflects the confidence with which the Familiarity Condition was held by those within Locke's ambit.

It is clear, then, that Boyle and Newton are both committed to the Familiarity Condition.26 Part of the motivation for adopting this condition on natural philosophical explanation derives from the polemic against the speculative excesses of scholastic natural philosophy, which was more than ready to postulate the existence of a host of occult or inexplicable qualities in order to explain various natural phenomena. One need only allude to the hackneyed example of the dormitive virtue of opium to make the point.27

Turning to Locke's discussion of analogical reasoning in E IV. xvi. 12, we have already encountered the following claim about reasoning from the observed to the unobserved micro‐structure of material objects:

Analogy in these matters is the only help we have, and 'tis from that alone we draw all our grounds of probability.

Locke gives the example of rubbing two sticks together, producing heat, and inferring from this that ‘what we call Heat and Fire, consists in a violent agitation of the imperceptible minute parts of the burning matter’.28 He clearly regards analogical reasoning from the observed to the unobserved micro‐structure of material objects to be our only legitimate method of reasoning about the unobserved parts of bodies. The implication of the words ‘only’ and ‘alone’ is that Locke was committed to the Familiarity Condition of two of his Master‐Builders.29 A crucial instance of this, as we have seen, is the manner in which Locke accepted Newton's inductive generalization that all bodies have the power of gravitational attraction; and this is a direct application of analogical reasoning.

As stated, the Familiarity Condition is an epistemic constraint on natural philosophical explanation; we cannot infer the existence of any new properties or laws at the sub‐microscopic level. However, the condition rides closely alongside the stronger metaphysical thesis that Newton toyed with, namely that there are in fact no properties (p.159) at the unobservable level that are not also at the observable level. It is important that we do not saddle Locke with this thesis, for he did conceive that spiritual beings might have additional senses to our own, and he implies that they might be able to perceive qualities of bodies to which humans do not have epistemic access.30 We must therefore stress that Locke is committed only to the epistemic cousin of this stronger metaphysical thesis.

The Familiarity Condition also needs to be considered alongside a common vice in the conduct of the understanding: taking words for things that really exist in the world. Locke regards this as a particularly pernicious error in human reasoning, and one that he himself attempts to guard against: ‘I endeavour, as much as I can, to deliver my self from those Fallacies, which we are apt to put upon our selves, by taking Words for Things’ (E II. xiii. 18, p. 174). He has a sustained discussion of this error in E III. x. 14–16, and returns to it in the section on words in the Conduct (§29). Locke gives numerous examples of this kind of error, from various speculative systems of philosophy, and even from the new philosophy. The causes of the vice are various, ranging from uncritical acceptance of authority to prepossession by a speculative system or to lack of clarity and distinctness among ideas. Furthermore, while Locke does not make the connection, the problem is compounded by the tendency to take certain qualities, such as cold, to be positive when they are merely privative (Draft B, §58, E II. viii. 1–6). At E III. xi Locke prescribes a raft of techniques in the use of language, all designed to combat this and other abuses of words. And these techniques can be seen to complement the Familiarity Condition in so far as they will militate against the postulation of the existence of theoretical entities that are mere posits of theory and not real existents.

Now, while the advantage of a commitment to the Familiarity Condition is the proscribing of appeals to bizarre and ad hoc properties, there is also a severe cost to it. The history of science since Locke has revealed that nature is replete with qualities at the sub‐microscopic level (as measured in Locke's day) that bear no resemblance to properties with which Locke and his contemporaries were familiar.31

A second condition for natural philosophical explanation to which Locke is committed has to do with the way of ideas. Let us call it the Simple Ideas Condition.

Simple Ideas Condition—if an idea is in the mind, then it arrived there as a simple idea (or simple ideas) via sensation or reflection.

The Simple Ideas Condition is a further epistemic constraint on empirical explanation for Locke. It lies at the heart of his theory of knowledge and ideas, and it has the (p.160) important consequence that we cannot conceive of new properties for which we lack simple ideas. Locke uses the example of the man born blind, who lacks ideas of colour and who, barring a miracle, has no prospect of receiving such ideas.32 (Of course, Locke denies that the mind is equipped with any innate principles or ideas.) This constraint is closely allied to the Familiarity Condition insofar as it limits the range of possible inferences humans can make about the unobserved. Were we even to believe that there are unfamiliar properties at the sub‐microscopic level, we could have no conception of what they might be like. Therefore the route to such properties is doubly blocked: ex hypothesi we cannot perceive them, and we cannot reason to them analogically because they are unfamiliar.

Furthermore, the Simple Ideas Condition and the Familiarity Condition, together with the primary and secondary quality distinction, appear to entail the Contact Criterion. Thus, while the Contact Criterion and the Reduction Principle might seem at first sight to be important and independent explanatory principles from which one might develop a theory of empirical explanation, it is really Locke's theory of ideas and the Familiarity Condition that are at the heart of Locke's views on how we can explain the phenomena of the material world.

The argument from intelligibility

But on what grounds could Locke claim that this new theory of qualities was explanatorily superior to its rivals? ‘Intelligibility’ was a buzz word amongst the mechanical philosophers. Conversely, ‘unintelligible’ was a term of abuse frequently levelled against Aristotelian natural philosophical categories, against the Galenic theory of disease, and often against the principles of the chymists. Boyle uses the term and its cognates scores of times, with reference to the new theory of qualities and to the corpuscularian hypothesis in general. Such uses are easy to spot in other mechanical philosophers as well. The term and its cognates occur eighteen times in the Essay alone.

Now there was a generic intelligibility argument, propounded by the adherents of the mechanical philosophy, which was normally stated in a fairly imprecise manner. For example, Descartes famously claims at Principles IV, §198:

Now we understand very well how the different size, shape and motion of the particles of one body can produce various local motions in another body. But there is no way of understanding how these same attributes (size, shape and motion) can produce something else whose nature is quite different from their own—like the substantial forms and real qualities which many {philosophers} suppose to inhere in things; and we cannot understand how these qualities or forms could have the power subsequently to produce local motions in other bodies. Not only is all this unintelligible, but we know that the nature of our soul is such that different local motions are quite sufficient to produce all the sensations in the soul. What is more, we actually experience (p.161) the various sensations as they are produced in the soul, and we do not find that anything reaches the brain from the external sense organs except for motions of this kind. In view of all this we have every reason to conclude that the properties in external objects to which we apply the terms light, colour, smell, taste, sound, heat and cold—as well as the other tactile qualities and even what are called ‘substantial forms’—are, so far as we can see, simply various dispositions in those objects which make them able to set up various kinds of motions in our nerves {which are required to produce all the various sensations in our soul}.33

Boyle produces an analogous argument in Forms and Qualities:

I do not remember, that either Aristotle himself, (who perhaps scarce ever attempted it,) or any of his Followers, has given a solid and intelligible solution of any one Phænomenon of Nature by the help of substantial Forms; which you need not think it strange I should say, since the greatest Patrons of Forms acknowledg their Nature to be unknown to Us, to explain any Effect by a substantial Form, must be to declare (as they speak) ignotum per ignotius, or at least per æquè ignotum. And indeed to explicate a Phænomenon, being to deduce it from something else in Nature more known to Us, then the thing to be explain'd by It, how can the imploying of Incomprehensible (or at least Uncomprehended) substantial Forms help Us to explain intelligibly This or That particular Phænomenon? For to say, that such an Effect proceeds not from this or that Quality of the Agent, but from its substantial Form, is to take an easie way to resolve all difficulties in general, without rightly resolving any one in particular; and would make a rare Philosophy, if it were not far more easie then satisfactory.34

Analyses of the various versions of these arguments suggest that the notion of intelligibility was closely tied with conceivability, with familiarity, with parsimony, and with observation. Normally the first step of the implicit argument is of the form:

  1. 1. If x is conceivable (or familiar or parsimonious) then x is intelligible

  2. 2. x is conceivable (or familiar or parsimonious)

  3. 3. Therefore, x is intelligible.

This is followed by an argument by elimination or by an argument from comparative plausibility. The argument from comparative plausibility suffers from the threat of circularity, for it is tempting to cash out plausibility in terms of intelligibility.

Locke gives a rather vague eliminative version of the intelligibility argument in Draft B, appealing to conceivability and to the Simple Ideas Condition. He claims that material effects in nature are ‘noething else but modifications of motion’. He continues:

I thinke we cannot conceive [material effects] to be any other […] for what ever sort of action besides [this] produces any effect I confesse my self to have noe notion nor Idea of & soe are as far from my thoughts apprehension & knowledge & as much in the darke to me as the Ideas of colours to a blinde man.35

(p.162) We should not make too much of this intelligibility argument as an argument for the Contact Criterion, for Locke's concern here in Draft B is to elucidate the notion of power, and the passage is recycled in the chapter on mixed modes in the Essay (II. xxii. 11), where the focus is on the idea of causation. Moreover, Locke uses a similar argument form and the very same terminology at E II. xxiii. 28 (p. 311), when arguing against the intelligibility of the transfer of motion from one body to another by impulse: ‘the passing of Motion out of one Body into another; […] I think, is as obscure and unconceivable, as how our Minds move or stop our Bodies by Thought’. However, of interest here is the argument form, together with Locke's early appeal to the Simple Ideas Condition.

Locke's claim, at E IV. iii. 16, p. 547, about ‘the corpuscularian Hypothesis, as that which is thought to go farthest in an intelligible Explication of the Qualities of Bodies’ suggests a comparative plausibility form of the intelligibility argument. But it must be said that Locke nowhere goes to any pains to spell out an intelligibility argument for the corpuscularian philosophy. Rather, informal intelligibility arguments are deployed by him on the one hand for the new theory of qualities, and on the other hand against the intelligibility of the fundamental notions of corpuscular theory. In Locke's discussions of the new philosophy, intelligibility arguments cut both ways.

Nomological explanation

Rather surprising, however, is the fact that Locke has so little to say about nomological explanation. As we have seen, he is certainly aware of the recent discovery of determinable laws of nature in optics and mechanics; and there are at least fifteen specific references to laws of nature in his writings.36 However, unlike Boyle, who infers by transdiction that ‘the Laws of Motion take place […] in the smallest Fragments of Matter,37 Locke never offers nomological explanations of natural phenomena, and he makes almost no mention of laws of nature in the Essay. There are only two places in his published writings where he mentions their explanatory power. The first occurs in the opening paragraph of his review of Newton's Principia in 1688. There he says:

[T]he philosophers and principally the moderns imagine that God has prescribed the same laws for the formation and the conservation of his works and they have tried to explicate by (p.163) them divers effects of nature. Mr Newton sets himself the same aim and takes the same way in this treatise.38

This is merely a paraphrase of Newton's description of the project of the Principia in his ‘Preface to the Reader’.39 The second reference to the explanatory role of laws is found in Locke's Second Reply to Stillingfleet, and there Locke's comments are laced with irony:

Your second argument against accommodating mathematics to the nature of material things is, ‘that mathematicians cannot be certain of the manner and degrees of force given to bodies so far distant as the fixed stars; nor of the laws of motion in other systems’. A very good argument why they should not proceed demonstratively in this our system upon laws of motion, observed to be established here: a reason that may persuade us to put out our eyes, for fear they should mislead us in what we do see, because there be things out of our sight.40

The implication of Locke's ironic reply to Stillingfleet is that we should proceed demonstratively from the laws of motion (just as Newton has done). But there is no hint of this in the Essay, where laws of nature are only mentioned three times, and then only in passing.41

I claimed in the previous chapter that in the 1690s Locke came to acquiesce in Newton's notion of principles of natural philosophy and that, while this notion does not appear in the Essay, changes that Locke made to his discussion of principles and maxims in Book IV indicate the impact of Newton's demonstrative reasoning from principles. It is natural to ask, therefore, whether Locke believed that these principles are laws of nature.

It is beyond doubt that Newton regarded the principles of the Principia as laws. He says as much in the Introduction to Book III: ‘[t]hese principles are the laws and conditions of motions and of forces, which especially relate to [natural] philosophy’.42 Later, after Locke's death, Newton reiterated the point when writing to Roger Cotes about ‘the first Principles or Axiomes [of experimental philosophy] which I call the laws of motion’ and by claiming that the principles ‘are deduced from Phænomena & made general by Induction’ and are therefore as certain as we can expect in natural philosophy.43 But in the draft preface to the Opticks, written in 1703, Newton is working with a broader conception of principles of natural philosophy, one which encompasses impenetrability, the particulate nature of matter, and the existence of an infinite, eternal spirit.44 These latter three principles are hardly laws of nature.

(p.164) I contend that this broader conception has primacy in Locke's own use of the notion of principles of natural philosophy. For Locke nowhere makes an explicit connection between principles and laws. He does intimate the identity between principles and laws, at least in relation to gravity, in the Elements of Natural Philosophy. I quote in extenso:

It appears, as far as human observation reaches, to be a settled Law of Nature, that all Bodies have a Tendency, Attraction, or Gravitation towards one another.

The same force applied to two different bodies, produces always the same quantity of Motion in each of them. For instance, let a Boat, which with its loading is one tun, be tied at a distance, to another Vessel, which with its lading [cargo] is twenty six tuns: if the rope that ties them together be pulled, either in the less or bigger of these Vessels; the less of the two, in their approach one to another, will move twenty six foot, while the other moves but one foot.

Wherefore the quantity of matter in the Earth being twenty six times more, than in the Moon; the motion in the Moon towards the Earth, by the common force of attraction by which they are impell'd towards one another, will be twenty six times as fast as in the Earth; that is, the Moon will move twenty six miles towards the Earth, for every mile the Earth moves towards the Moon.

Hence it is, that in this natural tendency of Bodys towards one another, that in the lesser is consider'd as Gravitation; and that in the bigger as Attraction: because the motion of the lesser body (by reason of its much greater swiftness) is alone taken notice of.

This Attraction is the strongest, the nearer the attracting bodies are to each other: and in different distances of the same bodys, is reciprocally in the duplicate proportion of those distances. For instance, if two bodys, at a given distance, attract each other with a certain force, at half the distance, they will attract each other with four times that force; at one third of the distance, with nine times that force: and so on.

Two bodys, at a distance; will put one another into motion by the force of attraction: which is unexplicable by us, tho’ made evident to us by experience, and so to be taken as a Principle in Natural Philosophy.45

But he nowhere seems to make the general claim that laws are identical to, or a subset of, the principles of natural philosophy. Evidence of this is found in a telling passage of the Essay that hardly changed from the first to the fourth editions. At E IV. iii. 29 (p. 560), Locke speaks thus:

The Things that, as far as our Observation reaches, we constantly find to proceed regularly, we may conclude, do act by a Law set them; but yet by a Law, that we know not: whereby, though Causes work steadily, and Effects constantly flow from them, yet their Connexions and Dependancies being not discoverable in our Ideas, we can but have an experimental Knowledge of them.

Had he clearly identified some of his natural philosophical principles with laws, he might have modified this passage, for he would have seen that we can, and we do, have experimental knowledge of some laws of nature, and that these laws in some sense explain the necessary connections or constant regularities in nature. Such a change to (p.165) the Essay, however, would have ramified through Locke's theory of knowledge and would have required additional changes elsewhere in the text. For, according to Locke's theory, sensitive knowledge—that is, experimental knowledge—is only of particulars (E IV. iii. 5), and all general propositions concerning substances—for instance ‘All bodies gravitate towards one another’—while keenly sought after, are not knowledge at all (E IV. v. 10; IV. vi. 16). Yet his principles of natural philosophy are general propositions that are ‘teeming truths’, ‘fundamental truths’ from which new truths can be generated. Little wonder, then, that the principles that matters of fact justify do not appear in the later editions of the Essay. There remains, therefore, a discontinuity between Locke's mature thoughts on the method of natural philosophy and what he presented in the Essay, even in its final form of the fourth edition.46

The Newtonian doctrine of natural philosophical principles also bears on another issue in Locke's account of natural philosophy given in the Essay and in the correspondence with Stillingfleet: the evident tension between Locke's ideal of a demonstrative natural philosophy and the passages in which he posits that God might superadd certain qualities to matter. The various attempts to resolve this tension have polarized Locke's interpreters. Some, like Michael Ayers, claim that Locke maintained that a demonstrative science of nature is possible, though ‘out of our reach’ (E IV. iii. 26, p. 556).47 Others, such as Matthew Stuart, claim that the superaddition passages reveal that a demonstrative science of nature is ruled out in principle and that Locke ultimately succumbs to natural philosophical scepticism.48

In my view, the most honest assessment,—and here I am in agreement with Margaret Wilson49—is that Locke has two ostensibly inconsistent lines of reasoning on the nature of our knowledge of material substances. On the one hand, he had long held to the ideal of a demonstrative natural philosophy which is based upon a corpuscular metric once we gain epistemic access to the determinate qualities of corpuscles. On the other hand, he could genuinely see no way in which, say, the quality of gravity would derive from what we know about matter.50 The principles of natural philosophy, however, provide a means of resolving this tension. For example, the principle of gravitational attraction, the prime example of a superadded quality, just is a principle which matters of fact justify: the general proposition ‘All bodies gravitate to one another’ is a teeming truth, the foundation of all natural philosophy, and therefore a principle upon which a demonstrative natural philosophy can be based. The key move here is to see that the superadded quality, gravity, is identical to the quality that features in the foundational principle.

(p.166) Now the unresolved tension in Locke's view of gravity arises from the fact that we cannot demonstrate that gravity derives or flows from the nature of matter. But, if the principle of gravity is the foundation of natural philosophy, then there is no need to derive it from anything more fundamental. It should be stressed that the contingency of the laws of nature is not relevant here—and not because, as we have seen, Locke had little time for nomological explanation. The fact that God might be able to impose laws (or qualities) on bodies independently of their inner natures (and of the natures of bodies with which they causally interact) does not entail that we cannot, in principle, have a demonstrative science of nature. Should God change the laws, we would have a different science of nature, but a science of nature nonetheless. Thus, had Locke identified principles with laws, such ‘nomological principles’ would have formed the foundation from which a demonstrative science of nature would have developed. It is even in keeping with Locke's views that there may be different sciences of nature at different times, or in different worlds. In my view, then, Locke did have the resources at hand to resolve the tension between his ideal of a demonstrative natural philosophy and the doctrine of superaddition. Locke was not a natural philosophical sceptic. Unfortunately, however, the doctrine of the principles of natural philosophy never found its way into the Essay.


Where does this leave Locke? What sort of natural philosophy can one reasonably construct, given these theses about empirical explanation? Given the current limits of observation in Locke's day, which enabled microscopic magnification up to x500 (though this was extremely rare) and telescopic magnification up to x100 and above (though the field of view was very limited), the prospects of actually getting epistemic access to the corpuscular realm were extremely remote.

First, let us consider the knowledge of celestial bodies. Here the situation is markedly more encouraging. In the case of forces, we have a highly developed celestial dynamics with (determinable) laws of motion and gravitational attraction; we have epistemic access to determinate properties of celestial bodies; we have causal explanations of macro‐terrestrial and celestial phenomena (movement of the tides, planetary and cometary motions, and the like); we have predictive power. This represents a significant advance, and in Some Thoughts concerning Education Locke rightly praises the Newtonian achievement, seeing it as a model for approaching other parts of natural philosophy:

Though the Systems of Physicks, that I have met with, afford little encouragement to look for Certainty or Science in any Treatise, which shall pretend to give us a body of Natural Philosophy from the first Principles of Bodies in general; yet the incomparable Mr. Newton, has shewn, how far Mathematicks, applied to some Parts of Nature, may, upon Principles that Matter of Fact justifie, carry us in the knowledge of some, as I may so call them, particular Provinces of the (p.167) Incomprehensible Universe. And if others could give us so good and clear an account of other parts of Nature, as he has of this our Planetary World, and the most considerable Phænomena observable in it, in his admirable Book Philosophiæ naturalis principia Mathematica, we might in time hope to be furnished with more true and certain Knowledge in several parts of this stupendious Machin, than hitherto we could have expected.51

However, when Locke turns to transdictive inferences, that is, inferences from the observed to the sub‐microscopic, the situation is very different. In the case of forces, we cannot posit, say, repulsive forces between corpuscles by the Contact Criterion; nor can we posit, by the Contact Criterion, any attractive forces apart from gravity; and we cannot give an explanation of cohesion, in spite of the Reduction Principle. In the case of qualities of sub‐microscopic bodies, we cannot posit any new qualities by the Familiarity Condition and by the Simple Ideas Condition; we can infer the existence of familiar properties through analogy, by the Familiarity Condition, but we cannot give any determinate corpuscular explanations, in spite of the Reduction Principle.

Not surprisingly, therefore, Locke is far less sanguine about the prospects of this aspect of experimental natural philosophy than he is about the progress that has been made in celestial mechanics:

I deny not, but a Man accustomed to rational and regular Experiments shall be able to see farther into the Nature of Bodies, and guess righter at their yet unknown Properties, than one, that is a Stranger to them: But yet, as I have said, this is but Judgment and Opinion, not Knowledge and Certainty. This way of getting, and improving our Knowledge in Substances only by Experience and History, which is all that the weakness of our Faculties in this state of Mediocrity which we are in in this World, can attain to, makes me suspect, that natural Philosophy is not capable of being made a Science.52

As for speculative natural philosophy, as we have already seen, Locke sees here little hope of any advance. We have already encountered his comment, in Some Thoughts concerning Education, that ‘Natural Philosophy, as a speculative Science, I imagin we have none, and perhaps, I may think I have reason to say, we never shall be able to make a Science of it’.53 One important consequence is that Locke does not regard the corpuscular hypothesis or the mechanical philosophy as having any special heuristic value.54 He does not view them as setting desiderata for an experimental programme, but rather as the outcome of speculative reflection on the collection of observations and experiments performed by chymists and others on material bodies. On this point (p.168) then, Locke had a more realistic grasp of the yawning gap between theory and the current science than Boyle did.

A second point to note is the extent to which Locke's approach to empirical explanation is still wedded to the Aristotelian ideal of a demonstrative natural philosophy. To be sure, he has his own objections to the syllogistic, and the process of demonstrative reasoning is cashed out by using the Lockean theory of ideas. But there is nothing in his approach to explanation that resembles the hypothetico‐deductive method, and this is what makes it so different from a modern conception of empirical explanation.

Finally, it is now clear, with hindsight, that Locke was unduly pessimistic about the prospects of natural philosophy and that, within one hundred years of the publication of his Essay, significant advances were to be made in chemistry that would deliver quantitative results. Therefore Locke should not be regarded as a modern when it comes to empirical explanation: he should not be regarded as pioneering, providing, or even adumbrating a promising new approach to the acquisition of natural philosophical knowledge.


(1) Extract from ‘Ignorantia’, by Locke, Bodl. MS Locke c. 33, fol. 27v.

(2) For minor textual variations between editions, see the critical apparatus of the Essay, p. 135.

(3) Second Reply, pp. 467–8.

(4) For further discussion, see Downing 1997.

(5) E II. viii. 18, p. 138.

(6) See the critical apparatus in Locke 1975, p. 629.

(7) See for example Boyle's ‘About the Excellency and Grounds of the Mechanical Hypothesis’, Boyle Works, Vol. 8, p. 105.

(8) Gaukroger 2008 downplays the significance of reductive explanations in Locke and emphasizes the autonomy and explanatory adequacy of phenomenal explanations vis‐à‐vis mirco‐corpuscular explanations.

(9) Locke's secondary qualities are said at E II. viii. 14, 23, 26, II. xxiii. 9, IV. iii. 12, etc. to depend on primaries; at E II. xxxi. 13, to flow from primary qualities; at E IV. iii. 11, to spring from primary qualities; at E IV. iii. 13, to be produced by primary qualities.

(10) Elements of Natural Philosophy, Locke Works, Vol. 3, p. 330.

(11) See for example E IV. vi. 11. For other uses of ‘machine’ to refer to the world, see also STCE, §194, p. 248 (quoted below), and Locke's Journal entry for 8 February 1677, in Locke 1936, pp. 86 and 89. Locke denies the Cartesian doctrine of the bête machine at a number of places in the Essay: E II. xi. 11 (section entitled ‘Brutes abstract not’), E II. xxvii. 5.

(12) Downing 1998.

(13) E IV. iii. 16, p. 547, underlining added.

(14) Downing 1998, pp. 356–63.

(15) For further discussion, see Anstey 2000, pp. 24–8.

(16) McMullin 1978.

(17) E IV. iii. 6. See also E II. viii. 19.

(18) ‘Texture’ appears in lists of primary qualities at E II. viii. 10, 14, 18, 23 and 24, E II. xxi. 3, 73, E II. xxiii. 8.

(19) E II. xxvii. 4–5.

(20) Anstey 2000, pp. 54–8.

(21) Boyle Works, Vol. 5, pp. 351–2.

(22) Boyle Works, Vol. 3, p. 42.

(23) Newton 1999, p. 795. ‘Intended and remitted’ means roughly increase or decrease in magnitude.

(24) Quoted from Cohen 1966, p. 176.

(25) Quoting Cohen's translation in Cohen 1971, p. 24. For a facsimile of this annotation, see ibid., p. 48.

(26) For Descartes and the Familiarity Condition, see his Principles, Part 4, §201 (Descartes 1985, p. 287), and his The World, ch. 6 (ibid., p. 91). For Cordemoy, see Cordemoy 1679, p. 73 (Mihnea Dobre alerted me to this passage).

(27) See Hutchison 1982. For Locke's references to the properties of opium, see E II. xxiii. 8 and IV. iii. 25.

(28) E IV. xvi. 12, pp. 665–6.

(29) See the related discussion in R. A. Wilson 2002, pp. 207–8.

(30) E II. xxiii. 13, p. 304: ‘though we cannot but allow, that the infinite Power and Wisdom of God, may frame Creatures with a thousand other Faculties, and ways of perceiving things without them, than what we have: Yet our Thoughts can go no farther than our own, so impossible it is for us to enlarge our very Guesses, beyond the Ideas received from our own Sensation and Reflection’. In Draft A, §2, p. 11, Locke suggests that spiritual beings might be distinguished by qualities ‘whereof we have noe notion’. See also Draft B, §20, pp. 131–2.

(31) See Friedman 1974.

(32) E II. ii. 2, III. xxii. 11, III iv. 11–13, IV. iii. 23, IV. xvii. 9. See also Draft A, §43, Draft B, §20, pp. 75 and 131.

(33) Descartes 1985, p. 285.

(34) Boyle Works, Vol. 5, pp. 351–2, underlining added.

(35) Draft B, §150, p. 262. For the use of the term ‘unintelligible’ against the School philosophers, see Draft B, §§72 and 88, pp. 176–7 and 195.

(36) For references to laws of nature in Locke's writings, see his Essays on Law of Nature (= Locke 1954), p. 109; his review of Newton's Principia in Locke 1688, pp. 436–7; STCE, §192, p. 246; Examination of Malebranche, Locke Works, Vol. 9, p. 217 (motion of animal spirits, rules of refraction and dioptrics); Second Reply, p. 427; Elements of Natural Philosophy, Locke Works, Vol. 3, p. 304; Discourse of Miracles, ibid., Vol. 9, pp. 256 and 264; and Conduct, p. 282. There is a note made by Locke in 1690 upon reading Cudworth's True Intellectual System of the Universe (LL 896), which refers to phenomena that cross ‘the laws of Mechanisme’: see Bodl. MS Locke c. 33, fol. 28r. This note was copied in 1700 into Bodl. MS Locke d. 11, fol. 51v.

(37) ‘About the Excellency of the Mechanical Hypothesis’, Boyle Works, Vol. 8, p. 107.

(38) Locke 1688b, pp. 436–7.

(39) Newton 1999, p. 382.

(40) Second Reply, p. 427.

(41) E IV. iii. 13 and 29, twice. Locke mentions the ‘Rules of Geometry’ in the context of corpuscular explanations in his review of Boyle's Specific Medicines (Dunton 1692, p. 184), and again in his review of Newton's Principia, Locke 1688b, p. 437.

(42) Newton 1999, p. 793.

(43) Newton to Cotes, 28 March 1713, in Newton 1959–77, Vol. 5, p. 397. He is even more explicit in a draft of this letter that was never sent: ibid., p. 399.

(44) McGuire 1970, pp. 183–4. See also p. 181 for the claim that Newton's principles are ‘broader in scope’ than laws are.

(45) Locke 1720, pp. 181–3 = Locke Works, 3, pp. 304–5, underlining added. For another comment that hints that the principle of gravity is a law, see also STCE, §192, p. 246.

(46) On this point I am in broad agreement with John Yolton, who claims that Locke ‘was unwilling or unable to adjust his theory of knowledge and his account of the science of nature to take account of this new understanding [the method of Newton's Principia]’: Yolton 1970, p. 89.

(47) Ayers 1981a and 1991, Vol. 2, pp. 142–53.

(48) See Stuart 1996.

(49) See M. D. Wilson 1979.

(50) See Second Reply, pp. 467–8. Gravity is not mentioned in the Essay.

(51) STCE, §194, pp. 248–9.

(52) E IV. xii. 10, p. 645. See also E IV. iii. 26, pp. 556–7: ‘And therefore I am apt to doubt that, how far soever humane Industry may advance useful and experimental Philosophy in physical Things, scientifical will still be out of our reach: because we want perfect and adequate ideas of those very bodies which are nearest to us, and most under our command’.

(53) STCE, §190, pp. 244–5.

(54) For Boyle's view of the heuristic value of his speculative corpuscular philosophy, see Anstey 2002d.