NATURALISM is not science, and the mechanical theory of Nature, the theory which serves as its foundation, is not science either. There are still, happily, many scientific men of eminence who reject the naturalistic philosophy—for philosophy it is—emphatically and entirely. And as to the mechanical theory in particular, what I ventured to say of that six years ago is truer than ever to-day:—“The mechanical theory as a professed explanation of the world, receives its death-blow from the progress of mechanical physics itself” (above).

Nevertheless, though Naturalism and the natural sciences, the Mechanical Theory of the Universe and mechanics as a science are logically distinct, yet the two are at first sight very similar and historically are very closely connected. Between the natural sciences and philosophies of the idealist (or spiritualist) type there is indeed no danger of confusion, for all such philosophies necessarily involve criticism of the epistemological assumptions which science unconsciously makes. Not so with Naturalism, which is as innocent of any theory of knowledge as science itself. In fact Naturalism, like Materialism, is only physics treated as metaphysics—a mistaken identification, which has tainted our so-called Natural Philosophy at least since the days of Descartes. Naturalism is less dogmatic than Materialism, no doubt; owing to its agnostic reservation as to the nature of ultimate reality; but it insists emphatically on the priority of the material aspect of its Unknowable. When the essentially philosophical question, how best to systematise experience as a whole, arises, the naturalist—as we have seen—contends that we must begin from the physical side. Then only are the facts precise, determinate, and rigorously concatenated: every thought that ever stirred the human heart, not less than every breeze that ever rippled the bosom of the deep can, it holds, be traced to a perfectly definite redistribution of matter and motion. To the mechanical principles of this redistribution an ultimate analysis, it is said, brings us down; and from these principles—aided by the nebular hypothesis and the theory of natural selection—all subsequent synthesis is to be explained. Life and mind take throughout a secondary place: cosmical mechanism not only precedes them but determines them, and they are powerless to modify it. The mental becomes the ‘epiphenomenal,’ a merely incidental phosphorescence that regularly accompanies material processes of a certain type and complexity. That propositions of such philosophic generality and scope are legitimate deductions from physical science, few, if any, of our modern physicists are bold enough directly to maintain. But many of them consider that their science itself is attacked by those who seek to lay bare the latent metaphysics, the physical realism, on which the Mechanical Theory of the Universe rests.

The criticism of this theory in the preceding lectures has been so regarded. It has been described as an “attempt to prove that the science of mechanics is no true science at all”; and again as making the “exactest of sciences impossible”; and finally as exhibiting “a dislike, a contempt, a hatred, a loathing of everything connected with science!” In point of fact my criticism rests throughout on the expositions of a school of physicists—if one might call them so—steadily increasing in number and influence, who reject entirely the almost mediæval realism imparted by Descartes to modern physics. This realism has remained so long unquestioned, that to challenge it now seems to many to spell scientific anarchy. And yet it surely verges on extravagance to suppose that men like Kirchhoff or Poincaré—to mention only two out of many distinguished names—who do challenge it, are seeking “to invalidate the methods of science” or to prove that “mechanics is no true science at all.” To distinguish them from the old school, whom we may fairly term physical realists, we might call the new school physical symbolists. The term is not very happy, but it may at least serve to emphasise the one difference between the two which now specially concerns us. The question at issue is very simple. Both schools start, of course, from the same perceptual experiences; both employ an abstract conceptual system, differing in detail but essentially the same; both resort to the same methods of verification. But the one believes that it is getting nearer to the ultimate reality and leaving mere appearances behind it: the other believes that it is only substituting a generalised descriptive scheme that is intellectually manageable, for the complexity of concrete facts which altogether overtask our comprehension. On either view the value of physics as systematic knowledge about things is unaffected: its possibilities of future extension and of practical application are in either case the same. But the speculative difference between the two is immense, and in this respect the question which is right becomes important.

Which then is right? This question may be argued from two distinct standpoints: from the general point of view of epistemology or from the more special one of the logical structure and method of the science of physics itself. In the fourth part of these lectures I have attempted to show that epistemologically the realistic interpretation of physics is untenable, and in this first part my contention is that the symbolic character of physics is completely borne out by what we may call the internal evidence of the science itself as well as by its past history and recent progress. I should assuredly never have dreamt of daring to meddle with physics as a positive science, still less of attempting to invalidate its methods or belittle its splendid achievements. There is a striking passage in Mr. Bradley's Appearance and Reality, which I have had throughout before my eyes:—“As a working point of view, directed and confined to the ascertainment of some special branch of truth, Phenomenalism is of course useful and indeed quite necessary. And the metaphysician, who attacks it when following its own business, is likely to fare badly.1 But,” he continues in words that I have already quoted (above), “when Phenomenalism loses its head and, becoming blatant, steps forward as a theory of first principles, then it is really not respectable. The best that can be said of its pretensions is that they are ridiculous.”2 This blunder I believe that physical realism has perpetrated so far as it has advanced or defended the mechanical theory of nature. And it was solely against these ‘pretensions,’ and the realistic interpretation of physical conceptions on which they rest, that my strictures were aimed.

Sir A. W. Rücker, as President of the British Association in 1901, devoted his Inaugural Address to a defence of physical realism against the symbolic interpretations recently advocated by Professors Poincaré3 and Poynting4 and by myself. Principal Rücker, is more judicious than many of his predecessors and confines his defence within very moderate bounds: the mechanical theory of Nature he seems to reject altogether, as “repugnant to common sense.” He is satisfied with maintaining a via media between this and “the opposing assertion that atoms and the ether are mere figments of the scientific imagination.” “It is a mistake,” he says, “to treat physical theories in general, and the atomic theory in particular, as though they were parts of a scheme which has failed if it leave anything unexplained.” To treat physical science in this fashion would certainly be an unjustifiable mistake, for that is usually ‘respectable’ enough to acknowledge its limits and to mind its own business. But it is otherwise with Naturalism, which does “step forward as a theory of first principles” and claims to be a philosophy. For Naturalism, then, to admit that it leaves anything unexplained or that it is repugnant to common sense amounts pro tanto to a capitulation, as I have already urged.5 Nevertheless, in spite of his reservations, Principal Rücker's claims on behalf of atoms and the ether seem to exceed the bounds of scientific propriety. Let us endeavour to examine his argument.

He maintains that we can argue back “from the phenomena displayed by matter to the constitution of matter itself, which we cannot directly perceive”—and therefore cannot perceive at all, for indirect perception is, strictly speaking, no longer perception. This argument he conceives admits of development through successive steps in a “series of propositions, the proof of each of which is based upon a few crucial phenomena.” At the same time he frankly acknowledges that the development can never be complete, “that the ultimate nature of things is, and must remain, unknown.” But is not even this abated confidence untimely? It is true indeed that the progress of the physical sciences of late has been unprecedented. The recent vast extension of their experimental methods, and the many strange and unexpected discoveries to which these have led, compel our admiration. But what is the result? Whilst Principal Rücker was thus confident of getting nearer to “a copy and not a mere diagram of the truth,” we find another distinguished physicist, Professor Boltzmann, addressing a similar assembly, drawing the following very different picture of the prospect:—“The majority of the novel phenomena here described have been investigated as yet only superficially. … But theory has been hard put to it by the new facts. The intellectual tranquillity into which she lapsed from her belief that she had comprehended everything has been rudely shaken, and no attempt has yet been successful in bringing the new phenomena under so successful a point of view as the old. In fact, everything is still in a state of vacillation and fermentation.” Then, after referring to the “philosophical criticisms of the foundations of mechanics formulated by Kirchhoff” and “pushed to their last consequences” by Hertz, to the revolutionary ideas awakened by the advances of electrodynamics and thermodynamics, and to the consequent growth of ‘Energetics,’ he continues: “To-day the battle of opinion rages tempestuously. … What will the outcome be? Will the old mechanics with the old forces, stripped of its metaphysical garb, continue to persist in its main features, or is it to exist henceforward merely in the pages of history?… Are the essential constituents of the present molecular theory…to endure for all time, … or … is the conception of a pure continuum as the most adequate representation of nature to prove victorious? Will mechanical models in any case persist, or will new, non-mechanical models prove better adapted, and the component factors of energy control absolutely the domain?…Is it possible that the conviction will ever arise that certain representations are per se exempt from displacement by simpler and more comprehensive ones, that they are true? Or is it perhaps the best conception of the future to imagine something of which one has absolutely no conception?”6

And now to return to Sir Arthur Rücker. Of the successive steps “back from the direct impressions of our senses to the constitution of matter itself,” he develops only two. In the first step—which the ancient atomists had already taken—he infers the coarse-grainedness of matter from the phenomena of diffusion, expansion, and heat. But at the outset this inference is wholly analogical:7 what we see in a cloud of dust or a swarm of flies we imagine extended indefinitely. But Principal Rücker is not content. with this: any other interpretation of his crucial phenomena he regards as “absolutely unintelligible.” Such a position, however, is epistemologically indefensible: it would require us to regard every unanalysable fact, everything ultimate or sui generis as ipso facto absurd.8 The phenomena appealed to suggest indeed that matter, so far as they are concerned, is resolvable into discrete parts; but they do not show this to be either logically necessary or perceptually a fact. The most we can say is what Leibniz said long ago: “C'est ce qui remplit le mieux l'imagination.” If we make this assumption we can form a mental picture of processes that we might otherwise have to regard simply as facts. But there is no absurdity in simple facts.9 “Up to this point,” says Principal Rücker, no question arises as to whether the separate parts are, like grains of sand, mere fragments of matter; or whether, though they are the bricks of which matter is built, they have, as individuals, properties different from those of masses of matter large enough to be directly perceived.” Apparently it is only the second step that decides for the latter of these alternatives. In that case the first step need not further detain us. “That matter is grained in structure,” said Professor Poynting in the address already referred to, “is hardly more than an expression of the fact that in very thin layers it ceases to behave as in thicker layers. But when we pass on from this general statement and give definite forms to the granules, or assign definite qualities to the intergranular cement, we are dealing with pure hypotheses.” In other words, the first step can scarcely be said to carry us at at all beyond the phenomena displayed by matter towards the constitution of matter itself.

It is far otherwise with the second step when that at length begins. This is not so much a step as a leap; for in the exposition of it we find ourselves confronted at once by such terms as ‘physical basis of matter,’ ‘ultra-physical entities,’ ‘quasimaterial substances,’ and ‘concealed causes of physical phenomena.’

Such language at once rouses the suspicion that if in this step we are taken beyond the phenomena displayed by matter, it is only because, as Professor Poynting has said, “we are dealing with pure hypotheses”—perhaps metaphysical hypotheses. According to Principal Rücker, however, we are leaving “matter as it seems to be,” and “dealing with something…which has properties different from those of matter in bulk.” “To show this,” he says, “is easy; for if the basis of matter had the same constitution as matter,” then “in the case in which a hot body is prevented from losing heat to surrounding objects, its sensible heat should spontaneously decay by a process of self-cooling.” He adds: “No such phenomenon is known.” If this meant that the case supposed is a purely hypothetical one incapable of rigorous verification, that would be true. But what is meant, no doubt, is that, in spite of the “uncertainty of experiment,” this ideal case would actually be found to hold good if the conditions imposed could be realised. Let us grant this; but what then? This ideal case implies another with which it is implicitly compared, that, namely, of a collection of bodies of sensible mass in irregular motion, and cut off from interaction with all bodies outside. In both cases the total mass and the total energy will, it is assumed, remain unchanged; but in the second case the kinetic energy of the sensible masses will be “frittered away,” till at length we have one mass of uniform temperature, and afterwards no further change. In other words, the second case will eventually become identical with the first, the irregular sensible motions having been transformed into their equivalent of sensible heat—that is, into irregular, insensible motions of the particles reached in the first step. But now suppose that, instead of remaining constant, this sensible heat in turn were “frittered away” by a “process of self-cooling.” How should we interpret this result? We should assume that the irregular motions of these particles, the sensible heat, had been transformed into irregular motions of smaller particles out of which these were built up, i.e. into insensible heat: we should infer, that is to say, that there existed a heat of a higher order. As this does not happen, it may be concluded that, so far as the particular ‘crucial phenomenon’ is concerned, the said particles are ultimate. But that is no reason for describing them as the physical basis of matter, or even for attributing to them properties different from those of matter in bulk. So far they only differ from that as a single grain differs from a heap of grains; or—it may be—as a packet of grains differs from a heap of similar packets. But now, instead of isolating the body, the chemist claims to show, perhaps by heating it still more, that sometimes the one, sometimes the other of these alternatives holds. In the one case, since the supposed particles have been decomposed, he maintains that they must really have been compound; but in the other, where such decomposition cannot be effected, he maintains that the particle must at least so far be regarded as simple.10 And so we reach the notion of the eighty odd elementary substances in which for the present the chemical analysis of matter ends. Even with this analysis, however, we are not really beyond the first step. We have assorted matter in bulk into eighty heaps of grains of as many different kinds, but each grain still differs from its heap only in being a grain and not a heap. The phenomena displayed by these heaps lead us to imagine the grains, but we have yet to assign differentiating qualities to the grains themselves.

It is only with this task that Principal Rücker's second step properly begins. “The idea that entities exist possessing properties different from those of matter in bulk is,” he says, “forced upon us at the very threshold of our study of nature.” But so far as I can see, this very safe and general statement, which even Berkeley might have made, is not followed up by the mention of any crucial phenomena showing that the particles of the physicist or the elements of the chemist are such entities. In fact the whole tenor of his address gives one the impression that Principal Rücker only claims to have made his first step secure. Thus, to cite but one passage out of many: “The cogency of the proof that matter is coarse-grained is,” he says, “in no way affected by the fact that we may have grave doubts as to the nature of the granules.” With the attempt to determine this nature, the second step then, as I have said, professes to begin. Leaving the phenomenal behind us we are now to be confronted with the ‘ultra-physical entities,’ ‘quasi-material substances,’ etc., which I have already mentioned. At the outset Principal Rücker distinguishes two theories, or rather classes of theories, between which he does not venture to decide: (1) that the granules are “distinct in kind from the medium which surrounds them”; (2) “that they are parts of that medium existing in a special state.” These two theories, he thinks, “are not by any means mutually exclusive,” though how even an ultra-physical entity can be distinct in kind from that of which it is a part or a state is not clear. If we suppose, however, that the granules are not entities at all, but that on the one view they are stages in the analysis, or simplification of gross matter, regarded as indefinitely divisible, and on the other as stages in the synthesis or complication of ether regarded as absolutely continuous, then indeed the two theories cease to be mutually exclusive. They become complementary aspects of the same thing. The phenomenal atom is conceived as really ether. Such a view removes the mechanical stumbling-block of action at a distance, and provides a “physical basis of matter” that is at least worthy of the name. In principle, that is to say, it has these merits, but so soon as the attempt is made to fill in the outline, such merits seem to be purely formal. As Maxwell has said, “The properties of a body supposed to be a uniform plenum may be affirmed dogmatically but cannot be explained mathematically.” Hence Descartes' philosophical conception of matter, which was of this sort, succumbed to the Newtonian doctrine of attracting particles; and though Newton “sought for the mechanism of gravitation in the properties of an æthereal medium diffused over the universe,”11 his efforts, as is well known, were fruitless, and he was content merely to insist that some such medium must exist.

As regards the ether the situation in the present day can hardly be said to have fundamentally changed. Like Newton, Principal Rücker postulates a medium, to avoid the supposed absurdity that “matter can act where it is not”; and yet he advances no crucial phenomena that disclose the nature of “this simpler machinery immediately below the complexities of superficial phenomena.” As with Newton, so still; the medium is wanted to make the material mechanism clear, and yet the medium itself cannot be got mechanically to work as it is wanted. “It must be a medium which can be effective for transmitting all the types of physical action known to us: it would be worse than no solution to have one medium to transmit gravitation, another to transmit electrical effects, another to transmit light, and so on.”12 The crucial phenomena that will uniquely determine such a medium are, I take it, still far to seek, even if we leave aside the still more complex problem of its differentiation into material atoms.

As regards this atomic constitution of matter the only prospect seems to be—as I have already urged (cf. above)—that of an indefinite regress with no true atom as a resting-place till the purely dynamical notion of ‘centres of force’ is reached; and that can never happen. Still, great strides seem to have been made in this direction of recent years. What the periodic law of Meyer and Mendelejeff (cf. above) and the spectroscopic investigations of Sir Norman Lockyer suggested as possible, Professor J. J. Thomson appears actually to have accomplished. If his interpretation of his interesting experiments be correct, the chemical atom has been broken. It must be henceforth conceived as made up, according to its ‘atomic weight,’ of from 1000 to 240,000 ‘corpuscles,’ as he has aptly termed them because of their resemblance to the famous light corpuscles of Newton's Opticks. The chemical atom in fact becomes a system geometrically and kinematically far exceeding in complexity the solar system as known to the astronomer. The whole organon of mechanical conceptions—kinetic and potential energy, attractions, repulsions, impacts, elasticity, vibrations, rotations, and so on—are again involved. What reason is there, then, for confidence that the application of this conceptual apparatus to these corpuscles will not still entail the old distinction of matter in bulk and constituent granules? The pursuit of the ultimate atom is thus like the fabled pursuit of the mirage: the goal appears always within reach, yet is always receding.

One thing at least seems certain: this pursuit has so far led to the discovery of nothing in the nature of the granules that entitles them to the designation of ‘ultra-physical’ or ‘quasimaterial.’ Science ever since the time of Descartes has regarded only the so-called primary qualities of matter as objective or real, and these—configuration, motion, and force—are present alike in sensible masses and in material particles; “even an atom,” says Maxwell, “when we consider it as capable of rotation, must be regarded as consisting of many material particles” (cf. above). If, setting continuity at defiance, we were to overleap the interval between Professor J. J. Thomson's ‘bodies smaller than atoms,’ and betake ourselves to Boscovich's dimensionless centres of force, we might indeed call these ultra-physical entities and quasi-material substances. But most people would prefer to call them analytical abstractions, ‘convenient fictions’ which it would be unwarrantable to regard as substances or entities of any sort. Similar remarks are applicable to the conception of a primordial medium or ether. The science of hydrodynamics works with the conception of a perfect fluid and the theory of elasticity with the conception of homogeneous bodies, albeit the evidence is all against the existence of such fluids or bodies. When the ether is regarded as at once a perfect fluid and perfectly elastic, shall we say that it is no longer merely a working hypothesis, but that these are actual properties of an ultra-physical entity or quasi-material substance? And if the ether itself is a hypothesis, how can its differentiation into the vortex-atoms of Lord Kelvin, or the strain-atoms of Professor Larmor, or the like, be other than hypothetical?

It would be a sufficient triumph for science if every such hypothesis proved adequate to embrace all the known facts. The beautiful conception of Lord Kelvin, for example, has already failed to stand this test, and the attention of physicists is now challenged by another so entirely revolutionary that it is actually entitled An Inversion of Ideas as to the Structure of the Universe. According to its author,13 “the probability that there should be another structure for the universe which would satisfy the same evidence must be indefinitely small!” We may fairly confront such overweening confidence with the concluding words of Maxwell's admirable primer: “The investigations of molecular science have proceeded for the, most part by the method of hypothesis, and comparison of the results of the hypothesis with the observed facts. The success of this method depends on the generality of the hypothesis we begin with. If our hypothesis is the extremely general one that the phenomena to be investigated depend on the configuration and motion of a material system, then if we are able to deduce any available results from such an hypothesis, we may safely apply them to the phenomena before us. If, on the other hand, we frame the hypothesis that the configuration, motion, or action of the material system is of a certain definite kind, and if the results of this hypothesis agree with the phenomena, then, unless we can prove that no other hypothesis would account for the phenomena, we must still admit the possibility of our hypothesis being the wrong one.”14 Substantially the same position is put still more strongly by Professor Poincaré. Unless the principles of the conservation of energy and of least action are satisfied, no mechanical explanation is possible, and when they are satisfied there is not only one possible explanation but an infinity of such.15

The contention of Principal Rücker's Address—that Nature is really and truly a mechanism of atoms and ether, or else is unintelligible—is, then, we conclude, logically unsound. His second step, we find, adds nothing essentially new to his first. The conceptions of configuration and motion of masses cannot be made to carry us further from physical phenomena and nearer to ultra-physical reality by diminishing the scale. The conceptions of perfect rigidity, perfect elasticity, or perfect fluidity again, for which there is no empirical justification, are surely none the more entitled to be regarded as properties ‘of a substance other than ordinary matter’ because as ideals they help us to form a possible model of its working. But how far does Principal Rücker really intend to go? He is constantly talking of ‘mental pictures,’ while constantly protesting that atoms and ether must be more than these. Such procedure practically amounts to saying: In this case I can form no other picture, and therefore the reality must be like it. And yet Principal Rücker's confidence does not carry him thus far. He is fair enough to allow the abstract possibility of a different mental picture. Atoms and ether, then, cannot be either presented realities or necessities of thought. Nay, he allows “the tentative nature of some of our theories”; he admits “many outstanding difficulties.” After all, then, he is only defending a working hypothesis, and one, moreover, that has lost greatly in prestige in the last half century. But if the atomic and other theories of the constitution of matter are but working hypotheses, and hypotheses strictly confined to physical phenomena, there is no justification for a theory which maintains that mechanism is fundamental everywhere and reduces the facts of life and mind to epiphenomena—makes them, that is to say, a degree more phenomenal, a degree less real than matter and motion. Such is the mechanical theory of the universe. Save as he seems unwittingly to countenance that, we have then no quarrel with Sir Arthur Rücker.

• 1.

  Mr. Bradley was thinking perhaps of Hegel.

• 2.

  F. H. Bradley, Appearance and Reality, 2nd ed., 1897, p. 126.

• 3.

  Address to the International Congress of Physicists at Paris, 1900.

• 4.

  Presidential Address to the Mathematical and Physical Section of the British Association, 1899.

• 5.

  Cf. Lecture 8 above.

• 6.

  The Recent Development of Method in Theoretical Physics, translated from the Proceedings of the Gesellschaft deutscher Naturforscher und Aerzte, in the Monist, 1901, pp. 240, 255 f.

• 7.

  On this point Epicurus, the earliest atomist of whom we have precise records, appears to have insisted. Cf. Lange, Geschichte des Materialismus, i. p. 79.

• 8.

  Cf. article by Professor Poynting, Physical Law and Life, in the Hibbert Journal, July 1903.

• 9.

  Moreover, the energists, as distinct from the atomists, contrive to deal with these processes by other and fewer assumptions. Whether their treatment is equally complete is disputable, but at least they have made tremendous headway within the last few years.

• 10.

  Not all chemists, however, are thus confident; for certainly it must be said that here again we have neither logical necessity nor perceived fact. Accordingly, some chemists of the new school seek not only to dispense with the help of the atomic theory, but even deny that elementary substances still exist as such in their so-called compounds. Cf. on this point Professor Liveing's Crystallisation, a Royal Institution Lecture; Nature, xliv. p. 150; and especially Duhem, Le Mixte et la Combinaison chimique, 1902.

• 11.

  Maxwell, Encyclopœdia Britannica, s.v. Attraction.

• 12.

  Prof. Larmor, article Æther, Encyclopœdia Britannica, 10th Ed.

• 13.

  Professor Osborne Reynolds, Rede Lecture, 1902.

• 14.

  Matter and Motion, p. 124.

• 15.

  La Science et l'Hypothèse, pp. 256 f. Principal Rücker chooses to regard this as applicable only to “explanations of isolated phenomena,” but I can find no warrant for any such restriction, and Professor Poincaré himself explicitly maintains the contrary (cf. p. 197).