§ 1

At this point in the argument a critic may be tempted to observe that while I have talked lengthily about perception, I have made little use of the labours which generations of philosophers have expended upon the subject. Have they, it will be asked, done nothing to clear away the difficulties with which it seems to be beset? Their labours, I should reply, have assuredly not been without valuable results; but it cannot be pretended that they are yet in a position to supply us with an agreed theory. On the contrary, I hazard the conjecture that since the days when Berkeley first turned the attention of the philosophic world to certain aspects of the problem, many controversies have been started, and few solutions have been generally accepted.

Consider the notorious differences which, within my own recollection, have divided thinkers of the first rank in the English-speaking world. Psychological idealists or mentalists (as, after Sidgwick, I prefer to name them) would say that my perception of a red bull in a green field was a complex of sensations; and that sensations, like feelings and desires, are mental states to which no independent material objects correspond. “Cosmothetic idealists” as Hamilton called them, or “Transfigured realists” as Spencer called them, would agree that they are mental states; while holding that they are states to which material or non-mental objects do correspond. The new realists say that they are, at one and the same time, both mental states1 and non-mental entities—thus differing from mentalists in asserting that there is an independent reality to be known, and from transfigured realists in asserting that our knowledge of it is direct.

§ 2

Now, so far as I can see, these three philosophic theories have nothing in common except the attribute of complete incompatibility with the scientific theory of perception, which is the one which we are concerned to establish. This, as we know, maintains that the perception of the red patch on the green patch is a mental result of cerebral changes set going through the optic nerve by light rays of various frequencies selectively reflected from certain illuminated surfaces! Can we force any such doctrine into conformity with mentalism?

Evidently not. If all the entities of which we know, or can know, anything are mental and only mental, if they are mental as ideas are mental, the material universe required by science dissolves into an unsubstantial dream. The enduring framework of the external world must be expunged from our picture of reality, and what remains will be processions of mental “movies,” with or without a “Self” to which they are presented.

These extreme conclusions, which could logically be pressed still further into solipsism, are consistently entertained by no one; and various are the expedients employed to mitigate their rigour. None of these expedients, so far as I know, have met with much favour. Berkeley's device was theological. He brought in God to provide an ideal world more stable than any compounded from the fleeting feelings of the human mind. Hume's device was practical. He lived, it seems, a double life—in his library, a mentalist; a realist elsewhere. Mill's device was verbal. He postulated what he called “permanent possibilities of sensation”; apparently under the impression that by devising a formula which contained the word “permanent” he had done all that either science or common sense could reasonably demand.

It is unnecessary, I think, to dwell further upon the insufficiency of this doctrine in any of its forms to perform the one task which, at the moment, we are asking it to undertake. In itself it has much to recommend it. But as a foundation for common sense, either in its primitive or in its scientific form, it seems almost ludicrously insufficient. “If idealism be true,” says Herbert Spencer, “evolution is a dream”; and I feel bound to agree with him.2

§ 3

“Cosmothetic idealism,” “Transfigured realism,” and other doctrines of representative knowledge are in little better case. Framed to combine all that seems true in mentalism with all that seems necessary for scientific realism, they have in large measure the defects of both and the merits of neither. They hold that there is an external reality—so far agreeing with science. They deny that we can ever perceive it, or indeed anything but our own sensations-so far agreeing with mentalism.

But they are for ever debarred from welding the two halves of their theory into a coherent whole, because they are quite unable to find any logical bridge from the sensations of which we are immediately conscious to the external world in which we are resolved to believe. Confined, each one of us, to the contemplation of his own mental states, we may assume, if we will, the existence of an outside reality, but it can never be more than an unproved and unprovable conjecture.3

Such speculations, however, are not merely vain; they are also at the present moment held in small esteem. I turn therefore to the last of the philosophic theories I have mentioned, which is vigorously advocated by an able and self-confident school on both sides of the Atlantic. The New Realists are in strong and conscious opposition to every form of philosophic idealism—psychological, transcendental, “cosmothetic,” what you will. Indeed, from my point of view, they are almost too deeply absorbed in their anti-idealist campaign. Their polemic is conducted with a vigour which somewhat upsets the balance of their exposition; and the problems with which we are here specially concerned receive the less attention at their hands.

§ 4

I make no attempt (need I say it?) to summarise their philosophy; contenting myself with a commentary on the particular doctrines which seem most relevant to our main line of argument.4

The first of these refers to the “status” of the object perceived. The New Realists agree (as I have already said) with the mentalists in thinking that this always belongs to the content of consciousness. But while the mentalists regard this proposition as throwing a blaze of light on the essential character of experienced reality, the New Realists treat it as a barren tautology, self-evident indeed but purely verbal. According to them it merely states that what we are conscious of is in consciousness; that what we experience is experienced;—propositions without doubt veracious, but not adding anything of importance to the sum of human knowledge.

The point may be well worth making. Yet I cannot but suspect that mentalists have some more solid justification for their opinions than a bare tautology. Many philosophers, belonging to different generations of thinkers and different schools of thought, not deficient in natural acumen, and by no means anxious to agree with each other, have held the view that the presentations of sense do to some extent share the essential nature of ideas; like them are psychical; like them are transitory; like them have no existence except as experienced. Nor did it ever occur to realists like Reid, or to critics like Kant, that an effective campaign could be conducted against Berkeley's mentalism by treating it as no more than a question of words. However this may be, I need not labour the point. Let us suppose, for the sake of argument, with the New Realists, that while the object perceived, in so far as it is perceived, is mental, it may nevertheless exist quite apart from mind, and though partially revealed in perception, may nevertheless be wholly independent of it.

§ 5

Now what relation has this theory to the common-sense view of the external world in the naïve form discussed in the last chapter? In two most important respects the resemblance is evidently very close. Common sense assumes (1) that things exist as we perceive them, and (2) that their character and duration are quite independent of our perception. We never, in our ordinary moments, suppose them to be constituted by thought or sophisticated through the operation of psychological laws; neither do we imagine that they only exist so long as they happen to be experienced. If I am right in supposing that to both these doctrines the New Realism subscribes, we have only to ask whether it has developed or explained them in such a way as to avoid the perplexities by which the common-sense statement of them is certainly beset.

It will be remembered that these perplexities were due in part to the divergence between the external world as we perceive it and the external world as modern science assures us that it is. Common sense does not solve these difficulties, but, as we have seen, it has its own method of eluding them. Fundamentally this method largely consists in the free use of unacknowledged transitions between the direct and the indirect method of gaining knowledge through perception. The reality of an independent world of things, situated in space, and capable of producing experiences by acting on organisms which are themselves to be reckoned among things situated in space, is assumed whenever the assumption is convenient; and it always is convenient when awkward questions are asked about the different appearances presented to the same observer by the same “thing” in different perspectives or under altered circumstances. We then for the moment cease to regard experience as directly introducing us to external fact. We suddenly treat it as due to a combination of causes, among which the “thing” experienced is but one—so that we are free to attribute differences in the appearance of the same external object, not to changes in the object itself, but to changes in the conditions under which we happen to perceive it.

This method, as has been seen, is open to criticism; and, if I rightly understand them, is utterly rejected by the New Realists. They hold a theory very like that of naïve commonsense, but purged of the incoherences required to make common sense tolerable in practice. As a result they present us with a conception of the material universe which, to me at least, seems equally repellent to common sense in its primitive and in its scientific form—a conception most difficult to realise and quite impossible to accept.

§ 6

Let us consider for a moment what this conception is. There is an interesting passage in Hume's “Treatise on Human Nature,” to which Professor Montague and Professor Perry5 have called attention, where the great doubter plays with the supposition that perceptions and objects are identical—both possessing “continued and uninterrupted being,” both being “sometimes present to the mind, sometimes absent from it,” both remaining in either event essentially unchanged. This I understand to be the opinion also of the New Realists. For them the world consists of elements which are mental when experienced, but in themselves are “neutral”—determined (that is to say) neither by mental nor by physical relations—sometimes imperceptible, and always independent of perception. Not everything therefore is perceived, but everything which is perceived exists independently as it is perceived. In other words, every “percept” has a separate and enduring existence out of all necessary relation to any process of perception whatever. Its status as a “percept ” is transitory and accidental. It has but a passing place in somebody's mental experience, suffering no essential change when it either makes or breaks this temporary connection. What it seems to be, that, in very fact, it is; what it is while under observation, that it will be when observed no more. I conclude that on this theory there are no such things as “mere appearances”; and that what in ordinary parlance we call the “aspects ” of any object are, in fact, persistent and self-sufficing entities, enduring occupants of an external world.

The speculative difficulties of such a theory seem to me serious, but my main concern is not so much with them as with the bearing of the theory itself upon accepted science. To me, I own, the two appear quite incompatible. If I am right in believing that science always assumes, and always must assume, that perception is the end product of a temporal process which begins with physical happenings in the world of space, time, and matter, we are dealing with a doctrine difficult enough to harmonise with any philosophic theory, but quite impossible to harmonise with neo-realism as I have described it. For in neo-realism thus understood there seems no room left for this causal view of perception. The planet Mars, to take Professor Perry's example, has an independent and persisting existence as a small red star. A small red star it was when unperceived, a small red star it will remain when it is perceived no more. The only difference which perception makes is that the planet while perceived becomes somehow part of some one's mental experiences, without being itself in any way affected. This no doubt is excellent realism. But it is not the realism of Science; nor do I understand how it can be made consistent with any scientific theory of perception whatever. It does not explain how the planetary object seen by the naked eye can be the same as the very different-looking planetary object observed through an astronomical telescope, nor how, speaking generally, numerical identity can ever be associated with diversity of appearance. But if that which seems different must always be different, then surely there is an end of scientific realism.

§ 7

It may perhaps be urged that the theory I have been commenting on, whether tenable or untenable, is not the theory of the New Realism; and I am by no means sure that it is. For the neo-realists, judged by many authoritative passages in their writings, seem also to hold a very different view. They are acutely aware of the shortcomings of “naïve realism.” They know well that there is “an elaborate mechanism underlying sense perception,” that this brings in its train “temporal and spatial aberrations,” so that “what we perceive will depend not only upon the nature of the object but on the nature of the medium through which its energies have passed on their way to our organism, [and] also upon the condition of our sense organs and brain.” Furthermore they recognise that it is urgently incumbent upon them to “amend the realism of common sense so as to make it compatible with relativity”; that is, as I understand the use here made of this ambiguous word, to find a method of eliminating the adulterations introduced into the object perceived by the causal process of perception.

But I cannot discover that they have made any contribution to this most desirable end beyond the invention or discovery of the “neutral entity”—the persistent and independent object, which can at the same time be part of a mental experience and of a material world, while always remaining, in its essential nature, quite independent of both. Now this discovery may play a most useful part in the polemic directed by the neo-realists against “subjectivism”; a controversy in which they are specially interested. But for us subjectivism, or “mentalism,” has already been weighed and found wanting. It has been rejected as an utterly inadequate basis for science; and what we are now concerned to discover is not whether mentalism is defective, but whether the New Realism, in its second or more developed form, can give us anything better.

By the New Realism in its more developed form I mean the doctrine which introduces into Hume's tentative theory (already discussed) considerations based on the causal process of perception. Hume's suggestion (as you will remember) was to the effect that our perceptions are not merely our perceptions, but are also independent and persisting entities. Now, if I am right in supposing that the modified doctrine of Neo-realism asserts or admits that these entities—independent though they be—are qualified in experience by the “personal equation,” by “temporal and spatial aberrations,” by the “nature of the medium” through which they are perceived, and by the “elaborate mechanism underlying sense perception,” does it not plunge us into perplexities akin to those which have so effectually discredited theories of representative perception? These theories assert that we have no direct acquaintance with external reality, but only with mental experiences from which by some undiscovered process the character of external reality may be inferred. Neo-realism in its first form does not go so far. It always claims that our acquaintance with what it calls “neutral” entities is unqualified and direct. But in its second form it apparently regards that acquaintance as inevitably modified by the “mechanism of perception.” In order, therefore, to correspond with reality immediate experience requires correction. But how is any correction to be applied? If we admit that the “neutral entities” suffer “aberrations” whenever they become elements in an experience, by what standard can we estimate either the magnitude of the aberration or its character? We can no more get outside our own experience into a position whence we can compare some “neutral entity” as it is “in itself” with the same entity after it has been dealt with by the “elaborate mechanism underlying sense perception,” than the believer in representative perception can compare his experiences with the realities he thinks they represent. Errors by which every possible observation is infected, no possible observation can either measure or amend.

§ 8

I cannot, therefore, believe that neo-realism really gives us much assistance in our particular line of enquiry. We are asking it (a) to indicate the cognitive process by which we may reasonably argue from our immediate experiences to the character of the material world, and (b) to bring this into harmony with the causal process by which the material world reveals itself in our immediate experiences. I cannot see that it performs either of these tasks. Perhaps it does not desire to perform them. In so far as it adopts Hume's half-forgotten speculation, it seems to conceive the external world in a manner quite incompatible with ordinary scientific realism. For, by treating every percept that ever has been, or that ever can be, as an independent denizen of the “neutral” limbo, it chokes up all the familiar avenues of physical causation, and makes it impossible to form any tolerable conception of this overcrowded universe, or of its relation to those who suppose themselves to perceive it. If, on the other hand, it adopts the scientific version of the causal process of perception without any attempt to combine it with a revised version of the cognitive process, it leaves us where it found us, face to face with the unresolved discord between the causes of belief and its reasons—and this in the region where we least expect to find it—the region of everyday experience, of plain matter of fact, of the accepted sciences, and of ordinary common sense.

§ 9

It may be worth adding a brief postscript to the preceding argument. I have already observed that one of the difficulties which everyone feels about mentalism is that it seems to supply no basis for our beliefs in the enduring world required by science. Does realism fare so very much better? Every experience of an object is for the realist, as for the Berkeleyan mentalist, a transitory event. It is true that the mentalist regards the experience as exhausting the fact. He identifies esse with percipi. For him an object which is not being perceived does not exist. It is also true that for ordinary common sense, and perhaps even more completely for the neo-realist, an object that was once perceived but is perceived no longer may nevertheless continue to exist as it did before. But surely also it may not. The mentalist argues that the percept from its very nature cannot endure. The realist denies this; but can he give any very satisfactory reason for thinking that, from its very nature, the percept will endure? So far as I can remember, he never makes the attempt.

I admit, of course, that his view on this point seems to be the simplest, the easiest, and (as we say) the most natural. And so it obviously is, if we assume what methodological doubt requires to be proved—namely, a world which is independent and enduring. But because mentalism is false, must this assumption be true? May they not both be false? The point deserves consideration, and plainly we cannot settle it by immediate intuition. We have not, and cannot have, any direct experience of such attributes as independence and endurance, for in this connection independence means independence of experience, and endurance means endurance when not experienced. Either, then, their reality must be established by some process of inference which is to me unknown; or else our assumption that they qualify the universe of material objects must be counted among the beliefs, inevitable but unproved, whose number and importance become more and more apparent as our investigations proceed.

• 1.

  Or at least belong to the content of consciousness.

• 2.

  I have developed this argument in ch. ix of “Philosophic Doubt.”

• 3.

  I have criticized H. Spencer's form of this doctrine at some length in “ Philosophic Doubt, ” ch. xi. The difficulty here referred to is in substance identical with that already dealt with in connection with attempts to base science entirely on the indirect argument from effect to cause.

• 4.

  See Professor Perry's “Present Philosophic Tendencies, ” ch. xiii.

• 5.

  “Present Philosophic Tendencies, ” p. 306.