Bertrand Russell


Russell Society Home Page

About Bertrand Russell

About the Russell Society

The BRS Library

Society Publications

Russell Texts Online

Russell Resources

JOIN the Russell Society!

Officers and Organization

Contact Us



Review of Russell’s Problems of Philosophy (1912)*

By Bernard Bosanquet

A GREAT deal may be said in 250 small pages by a man who knows what he wants to say. And this little book is a work of great interest and importance. Critics of Mr. Russell may regret – I sympathise with the feeling – that views which they hold to be fallacious should be so admirably and authoritatively presented to popular readers. But we may note that the books which Mr. Russell recommends for further study are first-rate philosophical classics, and the list, judiciously brief, shows no one-sidedness. And for the rest, the only remedy is that Mr. Russell’s critics should write books as good as his, if they can. After all, Wallace and Caird have had their chance with the general reader, and Mr. Russell has a right to his.

Mr. Russell explains in the Preface that he has not dealt equally with the whole field of philosophy, but has treated rather of the Theory of Knowledge, where a positive contribution seemed more readily made, than of Metaphysic, where results (so I understand him) might be more negative. Still the book presents a completer survey of his views, though of course less detailed in its departments, than he has published before. It is admirably clear both in statement and in arrangement, and charged at every point with novel and interesting matter. My attitude to it, unluckily, must be hostile on the whole. But my hostility assumes a peculiar form. For I believe that the author's fundamental demands are just, and that his error lies in holding that they can be- satisfied within the framework of his own logical conceptions, and cannot be satisfied within that which others have adopted.

I will first briefly survey the elements of the world as presented in Mr. Russell’s work, and then say something of the theory of knowledge involved in the representation.

First come the data of sense; not acts of sensation, and not mental states, but what is given in and through acts of sensation.

These, however, cannot be the objects of the external world. And this incapacity of theirs infects also what one would call in current language the perceived object, such as a wooden table. The relation of perceptive judgment to sense-data is discussed quite late in the book. I do not think the first three chapters recognise a distinction between the sense-data and what one calls the wooden table. In any case, the doctrine is, that neither the sense-data nor the wooden table (whether simply one with them or not) constitute the real external table.

This, known by inference from the sense-data as their cause, is a physical object, that is to say, something in space and capable of motion, but endowed with no sense-data of its own; i.e. I presume intrinsically and as unconditioned. For as conditioned it surely owns the original sense-data from which it is inferred. Physical objects are public and neutral, and are in physical space, which is also public and neutral; while sense-data are private and in private spaces, some of which also differ in kind (visual, tactual, etc.). The spatiality and motion of physical objects are therefore not such as we are acquainted with, but only possess relations corresponding to those we perceive. In physical space objects have only their true shapes, not the varied appearances which these shapes present to our sense-perception. (In spite of the author’s strongly held principle that objects of mind need not be mental states, with which I am quite at one, I am not sure that his conviction of the privacy of sense-data and sensuous space does not rest ultimately on the notion that they are mental states. A bodily change, e.g., in my eye, is not in principle private to me. Why is light so? See the treatment of light, pp. 45 and 65.)

Physical objects, then, cannot be directly known. “The actual thing which is the table is not strictly speaking known to us at all” (75).

Other people’s minds, like physical objects, are not known to us directly. They can be known only by introspection, and therefore each only by its owner. Thus, e.g., Bismarck alone can make the judgment of which he himself is a constituent, for to us the actual Bismarck is unknown. “When we make a statement about Julius Caesar, it is plain that Julius Caesar himself is not before our minds.” We are however justified in making judgments of this type, as “knowledge by description,” to which I will recur.

We also know past sense-data, inner or outer, directly by memory.

There are moreover universals, defined as anything which may be shared by many particulars. They are, but in a world which is neither mental nor physical. Relations, doubted or neglected by philosophers – surely not by Green, who talks of hardly anything else1 – are a principal class of these.

Whatever is given in sensation or is of the same nature as what is given in sensation, is particular. Proper names stand for particulars; other substantives with the remaining parts of speech, for universals.

Through universals we have a priori knowledge, mathematical, logical, and ethical. All a priori knowledge deals exclusively with the relations of universals, and depends on the fact, discovered by reflexion, that sometimes we can see these relations as self-evident. Truths which merely state what is given in sense are also self-evident. The truths immediately known or self-evident make up our intuitive knowledge of truths, which is the source of all our other knowledge of truths. The basis of Induction is such a self-evident truth; and the theory of Induction leads from the knowledge of things to the discussion of general truths, which in one form or another occupies more than the last half of the book. But “knowledge concerning the universe as a whole is not to be obtained by metaphysics” (221). Taken quite literally, this statement seems to me precisely wrong. Truths about the universe as a whole are just, I should have thought, what philosophy can establish. Particular facts beyond scientific experience, such as a future life, or the existence of a personal God, I strongly agree with the author that it cannot establish. And I believe that post-Kantian idealists would mostly take this view. I do not mean that light is not thrown on many particular problems by our general conception of the universe.

Even from this slight outline of the system before us, is it not obvious that we have here a very strange realism, in net result closely following Locke’s doctrine? Thus there is the real physical object, Locke's real essence, spatial and capable of motion, but not known to us as it is, hidden away from common perception by one veil at least, if not by two, according as the sense-data are reckoned as one with or other than the object of common perception. And this latter – or the sense-data which make it up – corresponds to Locke’s nominal essence composed of qualities relative to sense, and unlike their causes in the real essence. The physical object is indeed, as I gather, meant to be defended against the further analysis into sense-data which was fatal to Locke's primary qualities, by being deprived of all intrinsic sense-data altogether. But even if this defence has a meaning (for sense-data are, surely, ex hypothesi conditional, so that to demand that they should be intrinsic is a contradiction) it resolves itself into a further step towards agnosticism. We are to accept a physical object apart from conditions of perception. The abstraction, that is, grows thinner and thinner. Mr. Russell is indeed eager to prove, as he thinks, against “philosophers” that a thing may be known to exist although we do not know anything of it. I do not think that this has been disputed. But the very differently worded formula on the opposite page (p. 68) which seems to be taken as equivalent to denying the above, beginning, “if matter were essentially something with which we could not become acquainted” – i.e. we cannot know to exist that of which we essentially cannot know anything – this has been maintained and I think must be upheld. An old carpenter used to amuse children by saying, “Now I will show you something no one ever saw before,” and cut a bit of wood in two. Certainly we could have known that its inside existed, even if it had been impossible to cut it open. But something which essentially we could not know (Mr. Russell says, “become acquainted with,” but the meaning of the doctrine is deeper) is a miracle, that is, something the alleged conditions of whose existence contradict the conditions of knowledge, so that in as far as we prove it to exist, we ipso facto prove it not to be what it was said to be. Spencer’s Unknowable is the typical case in philosophy.

Nor do I see that so obvious a principle need be denied in defence of Mr Russell's physical objects or matter. They seem to be fairly well-connected with experience through the obvious sense-data which they cause. But if they are to be treated qua isolated unconditioned intrinsic existences, then I think they must be miracles, and cannot be known to exist. And I think this isolation is what Mr. Russell desires to insist on. He hints at another way of discovering their nature; but the qualification “intrinsic” still adheres to it (p. 54) and would cause the same difficulty.

However, there are physical objects, unknown except by inference and analogy, behind a single or double veil of perceived data and objects. Besides them we have minds, directly known only to their owners, and that by introspection, and universals, placed in a world neither physical nor mental. And there are self-evident truths of mathematics, logic, and morals, which remind us of those accepted by Locke in the same subjects. The statement that empirical philosophers strenuously denied the a priori character of mathematics can hardly be intended to include him.

Must not a scheme of realism which leaves standing such poor fragments of our things and truths, and those so arbitrarily selected, go the way which Locke’s has gone? No doubt it is better guarded against mentalism than his doctrine was; but immediacy and detachment, the heirs of mentalism, must surely do here the same destructive work which it did upon Locke.

What we who have been trained in another school miss alike in Locke and in Mr. Russell, is the point of view of the whole. Now this, I well understand, Mr. Russell intentionally rejects. It is only possible therefore to argue by pointing out what we take to be the disasters to which the rejection leads him.

He is obsessed, we should say, for example, by the conception of Idealism,2 according to which what appears as matter consists either of minds or of mental states. Against this he repeatedly advances the doctrine that the object of mind need not itself be mental, in this sense. Now I imagine that to post-Kantian idealists in general this position will seem obvious; but they will be inclined to hold Mr. Russell himself unfaithful to it in the spirit if not in the letter.

For why does he accept sense-data as primary factors of knowledge, and yet reject them as constituents of a real external object, unless, in spite of his own insistence that they are not acts of sensation, he retains some prejudice that as mental states they must belong to knowledge and yet cannot belong to things? I observed above upon the doubtful treatment of light. True, they are criticised as being changeable, inconsistent, conditional. But for anyone who recognises the conception of the whole, which is plainly involved from the first in the perception of an object, and apart from which the sense-data are not given at all, these data are not a primary basis, independent of the object, from which it is inferred, nor are they in the least inconsistent with each other, but on the contrary, form a system as the obviously necessary outcome of varying conditions. If we ask, which is the real colour, the real shape, etc., of the object, the answer surely is, “All of them, according to their conditions, each to each.” There is only a difficulty if we ask, What is the object as intrinsic or unconditioned? This is trying to get behind the whole by abstraction, and stultifies the inquiry into reality in a world like ours. Why should the object as apprehended not be an external object, as the whole system made known to us through sense-data, unless you think that sense-data are states of mind? And if science tells us that the wooden table, being further considered, must be held to be also “a collection of electric charges in violent motion,” that does not make it any the less a real wooden table. You may consider a tree as an aggregate of cells and fibres; and so it is; but it remains a real tree in space all the same. I certainly have no doubt whatever that the wooden table, an object constituted by all its conditioned qualities and reactions as a whole system, is a perceived physical object in a perceived space, one with my private space. And, again, I should have supposed that most post-Kantian idealists would agree with me.

Of course space and spatial objects as we perceive them are not for us ultimately satisfactory before metaphysical criticism. But this would tell us more about them; it would not go behind them or deny their existence.3 Here is one of my grounds for believing that if you want genuine hard Realism you can only get it from an idealist. Every one else picks and chooses according to some arbitrary standard, and ends by erecting a reality which leaves out nine-tenths of the facts; because he does not believe that the real is the whole.

A similar comment suggests itself on Mr. Russell’s doctrine of a priori knowledge. His criticism on Kant’s view (p. 132) is almost word for word the same as Green’s.4 But Green lives up to it; the question is if Mr. Russell does so. Both disown the idea of truth specially contributed by our own nature in the way of thought as opposed to the object. But Green infers from this the falsity in principle of the distinction between contingent and necessary truth, and the dependence of all truth on the whole system of experience – in short the coherence theory. Mr. Russell believes in a type and source of self-evidence dependent on isolated intuitions of relations between universals; in the “correspondence” theory of truth, which, along with his theory of the physical object as known by inference, must mean that you test one isolated inference (your belief) by another (the "fact ") without criticising the two by any common standard; and in a doctrine of Induction which seem to me to combine both errors, by resting an uncritical method dependent on cumulative instances, upon an uncritical a priori principle affirming an increase of probability to be generated by such accumulation. This dualism of fact and a priori truth is not of course grounded in Kant’s fashion; but it rests on a prejudice like his.

And all that is wanted for Induction is so simple. Postulate here what is postulated in the account of knowledge by description (to be discussed below), that there is a true proposition about the phenomenon in question; and further that some judgment can be formed as to the relative approximation of our knowledge to it, and the whole difficulty of uniformity vanishes. For suppose the laws of motion were going to cease to operate at twelve tonight, a true proposition about any phenomenon in which they are concerned would of course tell you so. You cannot interrogate nature except on the hypothesis that there is a truth about it, i.e., that it is a whole; and Mr. Russell’s view leaves the interrogation of nature entirely unaccounted for. Series of similar conjunctions are just what it does not rest upon.5 Or again, every association indicates a law. And you cannot establish the true form of a law by accumulating unanalysed cases under the first indication of it.

In short, the Inductive principle assumed a priori has no connexion with the general basis of logic, and will not account for the conception apart from which actual Inductive practice would be impossible. The idea of the whole is needed for all knowledge and influences it ab initio, and is not an assumption but a truth revealed by analysis. And if it were an assumption or a priori principle, it would be better to accept it as one which does explain Inductive practice and all cognition, than Mr. Russell’s principle, which does not explain even the former, and does not profess to explain the latter.

This notice is already too long, but there are still some important points which must be mentioned. The chapter on Knowledge by Acquaintance and Knowledge by Description6 is fundamental for Mr. Russell's position. It has the merit of raising the question what is the place of immediate experience in knowledge. And the answer given is that it forms a separate and primary type of knowledge, distinguished as knowledge of things by acquaintance from knowledge of truths, than which it is essentially simpler.

We have knowledge by acquaintance of sense-data outward and inward, of the past through memory, and of certain universals. We have not acquaintance with physical objects proper, nor with other people’s minds. Acquaintance is the foundation of all our knowledge, both of things and truths. Knowledge by description is, if I grasp it, the way in which we use knowledge by acquaintance as a substitute, so to speak, for the direct knowledge of things which we have never experienced. We describe, in terms with which we are acquainted, the true proposition which we know there is,7 and which we should like to make, but with the actual object concerned in which we are not acquainted. I cited above the instance of Julius Caesar. We cannot speak about him because we have not introspective acquaintance with his mind; and even in those who saw his living body had not direct knowledge of it, for it was a “physical object,” still less of him.

This, I said, is fundamental. No doubt our private and peculiar experience is somehow used in our knowledge of what is beyond it. Others know the same things through other private experiences. (I do not mean to admit that what we experience, e.g., in sense, is not continuous with what others experience.) This is clear. But the question is how to state the connexion. What Mr. Russell has done is to separate out the supposed data from the wholes and truths in which alone we apprehend them, and to erect them into a first line of knowledge, the foundation of all other.

Now this first line of knowledge, if it is knowledge at all, is of course extraordinarily meagre. And our knowledge of the second line, which contains practically the whole of our substantive cognition, is not admitted to apply to actual objects at all. Thus between the two our awareness of truths about things falls to the ground. The first line of knowledge is of “things,” but tells us no truths; the second tells us truths, but not about actual things.

Surely the error is in treating pure acquaintance, apart from judgment about wholes, as knowledge at all. It is admittedly not expressed in judgments, and I feel sure that the attempt to elicit judgments from it which preserve its alleged self-evidence must fail. (See the discussion on self-evidence of memory and perception Ch. 11. I read the result differently from Mr. Russell.) It cannot be got at as separate knowledge. It is rather a factor in all knowledge, by which thought is specified to certain objects.8 And if our knowledge is true, it is of the actual objects, systems which cannot be cut down to the terms of our acquaintance with them. It seems to me, I confess, absurd to say that only Bismarck could make a judgment of which he himself was a constituent. The self is a complex system, like other objects, and may be known from many points of view with equal grasp of its real being.

I have already referred to Mr. Russell’s well-known theory of truth and falsehood. It has a strong point in recognising that there is no objective falsehood, but that a false belief is a belief that real objects are related otherwise than as they are. It is not out of jealousy of priority, but to defend my bona fide belief in the kinship of robust idealism and thorough realism, that I point out what seems to me the same doctrine in Green’s argument that there is no unreal world, but only mistaken judgments as to the relations of real things.9 “A confusion of relations” was, as I remember from lectures, Green’s short working definition of error.

On the other hand, the point that truth is prior to coherence does not impress me greatly. What the Law of Contradiction does is just to say that truth is not self-contained in any proposition, but depends on the character of not being denied, all things considered. The appeal to non-negation is an appeal to coherence. The idea of the whole thus governs the conception of truth ab initio.

The criticism of Hegel, therefore, seems to me not to meet the point. It is really based on taking the nature of a thing to be bounded by what can be known of it through acquaintance. But the implied distinction, we have seen, is vicious, and in asserting truths you are asserting about the actual thing. You may be wrong; indeed the whole process is one of correction. That is inevitable. You find yourself in a contradiction; you must say something, and it must not be contradictory; therefore you must go farther, and contrive a reconciling truth; that is all. Plato describes the process quite clearly, and shows how it leads up to the fullest and most living concrete. Mr. Russell’s interpretation of the ‘Forms’ I hold to be on the whole a misrepresentation, though not without support in Plato. And his limitation of universals to abstractions, excluding individuals, I take to be one of the arbitrary distinctions by which he truncates experience.

This logical error, as I must hold it to be, affects profoundly his conception of the contrasted worlds of universals and of existences (p. 156). It is a dualism which divorces the being and logic of his universe from its life and love. I cannot say how deeply I regret that such a doctrine, absolutely fallacious, as I hold, in logic, and in its general bearing a mere formulation of popular prejudices,10 should go out to the world with Mr. Russell’s great authority. The typical universal is surely Plato’s all pervading pulse at once of thought and of desire. What Mr. Russell calls universals seem to me to be just the barest outlines of the substructure of the world, and to have a comparatively slight claim to the character of wholeness and pervasiveness which marks the true universal.

I referred above to Mr. Russell's view of the limits of philosophical knowledge. I strongly agree with it as against many theological philosophers; and I also accept in a very large measure his estimate of the value of philosophy, which seems to me very finely expressed. Only I am not sure what application he has in mind when he censures philosophies which recognise in the universe nothing alien to the Self. The ideas, say, of Plato, Spinoza, Hegel, Bradley, might equally well be described either in this language, or, in the phrase which carries Mr. Russell’s approval, “the union of Self and not-Self.” It would in my judgment be a very serious error to censure philosophies like these as treating the self in a way which makes it an obstacle to freedom of thought, and I am inclined to believe, as also I hope, that Mr. Russell has not committed it.


*  Bernard Bosanquet, “Review of Bertrand Russell, The Problems of Philosophy,” Mind, n.s. 21, no. 84 (Oct 1912), 556-64

1  Verbs and prepositions so far from being neglected have been a main source of categories from Aristotle downward. See Wallace, Introduction to Hegel’s Logic, p. 371

2 For Green’s characterisation of what amounts to this view of idealism, when advanced by Herbert Spencer, see Works, i., 386

3 The argument that there can be no contradiction in space, because logic has proved all sorts of spaces possible, amounts to very little. Possibility is a matter of point of view; from a very abstract point of view all sorts of impossibilities are possible. On given infinites see my Principle of Individuality, p. 394

4  Works, ii., 5

5  See for the true view, Green, Works, ii., 6

6 Locke, Essay iii., xi., 21

7 This suggestion, as I said above, seems to me most important. We only have to extend it to all objects, and the system of truth rises up single and systematic against the chaos of data and physical objects and self-evident principles

8 I am here following Prof. Stout

9 Prolegomena, sect. 23

10 Cf. James’ student in my Principle of Individuality, p. 10; James’ Pragmatism, p. 21