C. D. Broad, Examination of McTaggart's Philosophy, 1933



It will be remembered that McTaggart has claimed to show to anyone who might doubt or deny it that there is at least one existent. We discussed this claim in Chap. II of this book. In Chap. V, § 59, of the Nature of Existence he professes to show that anything that was existent would necessarily have some other characteristic beside that of being an existent.

It seems to me that this is self-evidently true. Existence is a purely formal characteristic; in fact, in its primary sense, it seems to me to be equivalent to particularity, and all other senses of it are derived from this. Now it is surely evident that there could be nothing which was simply and solely a particular, and had no non-formal qualities and no non-formal relations.

McTaggart, however, prefers to use arguments to prove that the supposition that there might be something which had existence and no other characteristic is self-contradictory. His first argument is that anything that had existence and no other characteristic would be a "perfect and absolute blank", and that to say that this existed would be equivalent to saying that a non-entity existed. This seems to me to be false or circular. If it had any characteristic at all, it would not be a perfect and absolute blank; and, by hypothesis, it has the characteristic of being existent. The only ground for saying that it would be a "perfect and absolute blank" is the assumption that the absence of all other characteristics entails the absence of the characteristic of being an existent and so leaves no characteristics at all. But this is exactly what the argument set out to prove. So the argument is circular, since it can prove its conclusion only by assuming it as one of its premises.

The second argument is as follows. Take any characteristic you like, e.g., squareness. If there were an existent which lacked squareness, it would have, in addition to the characteristic of being an existent, the negative characteristic of non-squareness. Now any existent either has squareness or lacks it, and so every existent has, beside the characteristic of being existent, either that of being square or that of being non-square. There are two comments to be made on this argument. (i) It would leave it possible that all the characteristics of an existent except that of being an existent might be negative. (ii) It assumes that there are positive characteristics beside that of being existent. If there were no such characteristic as squareness, nothing would be square and nothing would be non-square. The Law of Excluded Middle applies to such characteristics as there are and to them only. Thus there is at this point a suppressed empirical premise, viz., that there is at least one other positive characteristic beside that of being an existent. No one is likely to refuse to grant this premise to McTaggart. But, since in this part of his book McTaggart is emulating the White Knight in his anxiety to provide against even such unlikely contingencies as mice on the charger's saddle, it is unfortunate that his mouse-trap should have so many holes in it.

In § 61 of the Nature of Existence McTaggart argues that it is certain, with regard to any existent, that there are characteristics which it does not possess. His ground is that there are incompatible qualities, such as red and blue or round and square. If x is red it is not blue, and so there is at least one quality which it lacks. If x is blue it is not red, and so there is at least one quality which it lacks. And, if x is neither red nor blue, there are at least two qualities which it lacks. But x must be either red or blue or neither red nor blue, and so in any case it will lack at least one quality. It seems clear that McTaggart's argument requires a new empirical premise, viz., that there are incompatible positive qualities. This again must be granted to him, but it ought to have been made explicit. It would have been denied by Leibniz and by many other philosophers before Kant.

It is evident that McTaggart himself feels some uneasiness at this point, for he professes to answer an objection of this kind in the footnote to § 62. But, although his answer to the objection which he there states is correct, the objection is not the fundamental one and his answer is irrelevant to the latter. He takes the objector to assert that our knowledge that red and blue are incompatible is empirical and not a priori, so that an empirical premise has been surreptitiously introduced at this point. To this he answers, quite rightly, that, although we need perception to make us acquainted with the qualities redness and blueness, yet, once we are acquainted with them, we can see that they are necessarily incompatible. This, however, does not answer my contention that a new empirical premise has been introduced at this point, and that our knowledge of the conclusion that every existent lacks certain qualities is therefore empirical and not a priori. The premise of McTaggart's argument is that there are incompatible qualities. No conclusion derived from this can be known a priori unless this premise itself is known a priori, i.e., unless we can see that it is necessary that there should be incompatible positive qualities. And surely we cannot see this. Our actual position is that we know empirically that there are certain qualities, e.g., redness and blueness, with regard to which we can see that they are necessarily incompatible. From this we can infer that as a matter of fact every existent lacks some positive quality or other. But we cannot infer from it that it is necessary that every existent should lack at least one positive quality.

In § 62 it is asserted that anything that exists will have as many qualities, positive and negative, as there are positive qualities. This, of course, follows at once from the Law of Excluded Middle. Again, if there be three mutually incompatible qualities, e.g., red, green, and blue, any particular will have at least two negative qualities. For it will be either non-red and non-green and non-blue; or red, and therefore non-green and non-blue; or green, and therefore non-blue and non-red; or blue, and therefore non-red and non-green. If we are to pass to the categorical conclusion that every particular has in fact at least two negative qualities, we shall need the premise that there are at least three mutually incompatible positive qualities. And our knowledge of this premise is empirical.

Must an existent have at least one positive quality beside that of being an existent? McTaggart says that it must, since it will have the quality of being many-qualitied, which differs from existence and is positive. This leaves it possible that an existent might have no positive original qualities except existence, and might stand in no positive original relations. I should have thought that one could see that this supposed state of affairs is impossible, and could assert as a self-evident truth that any existent must have some positive original quality beside existence or stand in some positive original relation. Some people would perhaps go further and say that it is self-evident that any existent must have some positive original quality beside existence. What precisely is the alternative which they wish to exclude by this assertion? Suppose, if possible, that x is a particular with no original quality except existence. With regard to any particular it seems logically possible that there should have been this particular even though there had been no other except this and its parts, if it has any. If then x were a particular with no original qualities except existence, it would be logically possible that there should have been no facts about x except the fact that it is an existent and facts about its relations to its own parts if it has any. Presumably those who assert that every particular must have some positive original quality beside existence mean to deny that it is logically possible for there to be a particular about which the only facts would be that it is an existent and that it stands in certain relations to its own parts. Now that the case has been clearly put to him the reader must be left to decide for himself whether this is logically possible or not.

1. Substance.

Although the name "substance" has not yet been used, we have in fact been talking about "substances" in McTaggart's sense of the word. The name is explicitly introduced in Chap. VI of the Nature of Existence, and that chapter is devoted to an explanation of his usage of the term. I will begin by expounding McTaggart's account of Substance, and will then make some further remarks of my own.

1.1. McTaggart's Notion of Substance. In § 67, p. 68, of the Nature of Existence, a "substance" is defined or described as something which is existent, which has qualities and is related, and is not itself either a quality or a relation. It will be noticed that the description contains negative characteristics. The description is plainly inadequate for McTaggart's purposes, since it would make facts substances; for facts have characteristics. For example, a fact is necessary or contingent, and one fact may entail another fact. And, on McTaggart's view, all facts would be existents, since they are indirectly or directly about existents. Now he certainly does not mean to include facts among substances, and so we must certainly add at the end of his description of "substance" the further alternative "or a fact".

McTaggart says that his is the traditional definition of "substance". It seems to me quite certain that he is mistaken on this point. As he remarks later, a sneeze or a flash of lightning, or a group whose members are a certain sneeze and a certain flash, would all count as substances on his definition. Now it is certainly extremely paradoxical to call such entities as these "substances"; so it seems most unlikely that people can have meant by "substance" what McTaggart describes, though they might find it hard to say precisely what else they did mean. I think that McTaggart's definition would be accepted as a satisfactory description of a "particular ", and I propose to substitute the latter word for McTaggart's word "substance" except when I give due warning to the contrary. We must remember that the description is not strictly a definition of "particularity"; the notion of being a particular is, no doubt, as indefinable as that of being a characteristic. The two are correlative, and each can be described only by reference to the other. We must also remember that there may be fundamentally different kinds of particular, e.g., "substances", in the ordinary sense, whatever that may be, and events.

In §§ 65-7 McTaggart tries to prove that there are particulars. It will be remembered that he has professed to show that there are existents. It therefore remains to show that not all existents could be characteristics or facts. The essential point in his argument is that the existence of qualities, of relations, and of facts is derivative. A quality has existence only by qualifying, either positively or negatively, some existent. A relation has existence only by relating terms which are existents. And, we might add, a fact is existent only by being a fact about some existent. Now it is impossible that the existence of anything should be derived, in this way, from the existence of something else whose existence in turn is derived in this way from that of something else, and so on without end. Therefore, if there be any existents, there must be some which are neither qualities nor relations nor facts. Now there are existents. Therefore there must be particulars. (McTaggart, of course, does not himself consider the case of facts. But it does seem quite clear that there must be some facts which are not themselves about facts, and so the omission is not serious.) This argument seems to me conclusive.

McTaggart then considers, and tries to refute, an alternative view, viz., that what is called a "substance" is really a highly complex quality. I am pretty certain that those who have professed to hold this and similar views have done so through a confusion between determinate qualities and occurrents, which are the manifestations of determinate qualities throughout periods of time, e.g., between a certain perfectly determinate quality of squeakiness, which may be manifested on many different occasions, and a particular squeaking, which is the manifestation of this squeakiness at a certain place and throughout a certain five minutes. What they really held may be most accurately expressed as follows. For any statement in which the name of a "thing" or "substance" in the ordinary sense occurs one can always substitute a statement in which no such name occurs. The substituted statement will be about occurrents and their relations, and it will express all that is true in what was expressed by the former. (Of course we must not expect that the grammatical predicate of the old statement will appear unmodified in the new one; if the subject is changed in the way suggested whilst the predicate is left as it was, we shall merely get nonsense, such as "I opened that series of actual and possible sensations which I call my umbrella".) Now occurrents are particulars, and therefore "substances" in McTaggart's sense, though not "things" or "substances" in the ordinary sense. And so the theory which these people were trying to state, unless I have been unduly charitable to them and they really meant the nonsense which they talked about "the characteristics of particulars being particular", does involve the existence of "substances" in McTaggart's sense. But they confused occurrents, which are particulars, with determinate qualities, which are universals, and so stated their theory in the plainly nonsensical form which McTaggart takes literally and tries to refute.

The theory, as McTaggart interprets it, is quite certainly false. But his refutation in § 66 is not really conclusive, and it could be met by a more careful statement of the position. The argument comes to this. Take the fact that Smith is happy and virtuous. If the theory under discussion be true, Smith is a highly complex quality of which happiness is one constituent and virtue another. What then does happiness qualify? It is equally nonsensical to say that happiness is happy, that virtue (or any other component in this complex quality) is happy, or that the complex quality taken as a unit is happy. Yet it is not nonsense to say that Smith is happy.

McTaggart is able to use this argument because he forgets, or refuses to acknowledge, that, when a new meaning is proposed for the subject of a sentence, the assertor cannot fairly be expected to retain precisely the same meaning as before for the copula and the predicate. It is as if a certain note in a tune had been altered in key, and one insisted that all the rest should be kept as before and then complained of a discord. The supporters of the theory under discussion could have answered McTaggart as follows: "Our theory is that the sentence Smith is happy can be replaced without loss or gain of meaning by the sentence There is a certain complex quality, which includes as a component the quality of evoking the use of the name 'Smith' in certain men, and this complex quality contains as a component the quality of happiness. You have no right to insist that we shall interpret the copula in a way which is appropriate only to your interpretation of the subject and shall combine the unmodified copula with our modification in the interpretation of the subject". No doubt the theory is false, even when thus stated; but McTaggart's objection in § 66 is not valid, for the reason just given.

In §§ 68-70 McTaggart considers, and tries to refute, certain objections to the notion of Particulars. The objections are, I think, all variations on Locke's theme of a substance being "a something, I know not what". It is said, quite truly, that no idea corresponds to the phrase "particular not characterised by any characteristics". It is concluded that no idea corresponds to the word "particular" itself. This plainly does not follow, and it is not in fact true. When I say of something that it is a particular I mean that it has the formal characteristics of having qualities, of standing in relations, of being an existent, and of not being a quality or a relation or a fact. All this is perfectly intelligible. And it is not rendered unintelligible by my knowledge that anything which had these formal characteristics would necessarily also have some non-formal characteristics. McTaggart rightly points out that a precisely similar argument might have been used to prove that no idea corresponds to the phrase "existent characteristic". For it could truly be said that no idea corresponds to the phrase "existent characteristic not qualifying directly or indirectly some particular". And from this anyone who accepts the first argument ought to conclude that no idea corresponds to the phrase "existent characteristic". A consistent user of the argument against particulars would therefore be left with the doctrine that the word "particular" and the phrase "existent characteristic" are equally meaningless. I am not clear that this would worry such a person so much as McTaggart evidently thought that it should. He might admit the conclusion, and propose some entirely different analysis of the facts which McTaggart analyses in terms of particulars and existent characteristics.

Much later in the book, in § 92, McTaggart suggests three causes which may have led people who never doubted that there are characteristics to feel doubtful whether there are particulars. (i) lt has often been thought that what we are acquainted with in sensation is sense-qualities, e.g., determinate shades of redness, determinate forms of squeakiness, and so on, and not particulars. And so sensation was not recognised to be indubitable evidence for there being particulars, whilst it was thought to be indubitable evidence for there being characteristics. Such a view of sensation is plainly mistaken. What I sense is not redness or squeakiness, but some particular which manifests redness or squeakiness to me. McTaggart thinks that the mistake arose because we are not much interested in sensa, as such, but only in their qualities as signs of the presence of such and such physical objects. I do not see that this can suffice to explain the origin of the mistake. For the very reason which is alleged to have made people concentrate on the characteristics of sensa and ignore their particularity would surely force them to attend to the particularly of the things of which the sensa are believed to be signs. (ii) Every particular has characteristics, but there are some characteristics which belong, not to any particular, but only to other characteristics. (iii)We know a great deal about certain characteristics, e.g., perfect straightness or circularity, without knowing whether any particular is characterised by them. But all that we can ever know about any particular is that it has such and such characteristics. The two facts last mentioned have made characteristics seem less dependent on particulars than particulars are on characteristics. This may have led to the belief or the hope that particulars might be dispensed with and that their work might be done by bundles of characteristics. Whether the belief arose in this way or not, it is certainly mistaken.

In § 69 McTaggart deals with an argument of Prof. Stout's, which now appears on pp. 255-6 of the latter's Studies in Philosophy and Psychology. "What", asks Prof. Stout, "is the subject itself as distinguished from the attributes? It would seem that its whole being must consist in being that to which its attributes belong. But how can the whole being of anything consist in its being related to something else? There must be an answer to the question: What is it that is so related?" Prof. Stout uses this rhetorical question to lead up to the doctrine that a subject is a complex whole composed of attributes interrelated in a peculiar way, a doctrine which he develops more fully in the essay on The Nature of Universals and Propositions which comes at the end of the Studies and was first published in 1921 after Vol. I of the Nature of Existence was completed.

McTaggart's answer is as follows. Suppose that Smith is happy and virtuous. Then it is true that there will be the relational fact that Smith is characterised by happiness, and there will be the relational fact that Smith is characterised by virtue. But these relational facts are not ultimate. They are derived respectively from the non-relational fact that Smith is happy and the non-relational fact that Smith is virtuous. If you insist that "Smith is happy" is just a loose phrase for what would be more accurately expressed by "Smith is characterised by happiness", you cannot consistently stop at that point. There is just as much or as little reason to insist that "Smith is charactetised by happiness" is a loose phrase for what would be more accurately expressed by "Smith is referent, and happiness is relatum, to the relation of being characterized by". And, once started on this course, there is no place at which one can consistently stop. We must then admit that there are non-relational facts as well as relational facts about Smith. So, when Stout asks us: "What is it that stands to happiness, virtue, etc., in the relation of being characterised by them?", we can answer: "A happy, virtuous, human particular".

I do not know whether Prof. Stout would accept this answer, for I cannot understand what precisely he means by his question. If the question means: "Under what category does that which has qualities and stands in relations fall?" the answer is that it falls under the category of Particularity. If the question means: "What kind of thing is that which has these qualities and stands in these relations?", the answer consists in mentioning certain of its fundamental properties, which either logically entail the rest or are, in the actual world, trustworthy signs of the presence of the rest. This is what we do, in answer to the question "What is that?", when we say "That is a circle" or "That is a bit of gold" or "That is a horse". If the question be interpreted in either of these two ways, there seems to be no difficulty in answering it. If it be interpreted in neither way, I must confess that I do not know what it means, and I do not believe that it could be answered in terms of any theory.

I think it is clear from Prof. Stout's essay on The Nature of Universals and Propositions that he and McTaggart are largely at cross purposes in this controversy. There are two sources of misunderstanding. In the first place, there is McTaggart's extremely wide use of the word "substance". As he remarks in § 72 of the Nature of Existence, on his definition, an event such as a sneeze will count as a substance. So will a whist party, or the group whose members are all red-haired archdeacons. He admits that most people would refuse to call these "substances"; but he claims that other people really mean by "substance" what he means by it and that they refuse the name to events and to groups of particulars only through inconsistency. It is very important to notice that McTaggart is really making a factual assumption here, which he never examines and which has far-reaching consequences for his system. He assumes without question that there is no fundamental distinction between what are called "occurrents" and what are called "continuants". Now most people believe, rightly or wrongly, that there is a fundamental distinction. They would be prepared to call them both "particulars" or "existent substantives", but they would confine the name "substances" to continuants. That is why I use the word "particular" where McTaggart uses the word "substance". His word "substance" covers particulars, whether they be occurrents or continuants, and groups or aggregates of particulars. It is a most extraordinary fact that McTaggart never discussed the common opinion that there are at least two fundamentally different kinds of particulars, viz., occurrents and continuants, and assumed, without any kind of argument, that it is a baseless prejudice.

Now Prof. Stout regards the distinction as fundamental. He would refuse to call an event, such as a sneeze, a "substance ". When he discusses the nature of "substance" he is discussing the nature of continuants. Stated accurately, his problem may be put as follows: "What is the right analysis of the facts which are expressed by sentences in which a continuant-name appears as subject, such as Smith is happy?" The analysis which he rejects, and which he assumes that McTaggart accepts, is that the continuant-name designates or describes a single particular existent of an unique kind which is qualified by a number of universals, such as happiness, virtue, humanity, etc.

Prof. Stout's own alternative is not easy to state clearly for the following reason. Just as McTaggart, believing there to be no fundamental distinction between occurrents and continuants, has provided no names for these two ostensibly different kinds of particular, so Prof. Stout, holding a peculiar theory of his own about universals, has not provided any means of expressing the ostensible distinction between an occurrent, such a red flash, and the perfectly determine quality -- a certain shade of redness -- of which this occurrent is a manifestation.

In terms of the distinctions for which we have now provided names we can state Prof. Stout's theory as follows: (i) There are occurrents, and each occurrent is a particular existent and therefore a "substance" in McTaggart's sense, though not in Stout's. (ii) In ordinary language each occurrent would be said to be a manifestation, throughout a period of time, of a single perfectly determinate quality. (iii) Such language suggests that every occurrent is a complex, consisting of a Hoc, which is a particular, and a perfectly determinate Quale, of which this Hoc is a manifestation and of which an indefinite number of other Haec might also be manifestations. This is a mistake. There is no such internal complexity in an occurrent, and there are no universal perfectly determinate Qualia in each of which a plurality of Haec can participate. The fundamental fact is that each occurrent has to certain other occurrents an ultimate relation of exact qualitative likeness; the mistake is to attempt to analyse this fact into the participation of a number of Haec in a single common Quale. This is a view which McTaggart, as we have seen, rejected with very little discussion. I do not know of any conclusive reason either for or against it. (iv) A substance, in Stout's sense, i.e., a continuant, is a group of dissimilar occurrents, i.e., of substances in McTaggart's sense, interrelated in an unique way. On the rejected theory a continuant is a single Hoc inhered in by many different Qualia, each of which is capable of inhering in many different Haec. On Stout's theory a continuant is a set of peculiarly interrelated Haec each of which is unlike all the other members of the set and is exactly like other Haec which are not members of the set.

The upshot of the discussion is that the essential differences between McTaggart and Stout reduce to the two following: (i) McTaggart accepted universal perfectly determinate qualities, and regarded exact likeness between two particulars as dependent on, or analysable into, the inherence in both of a common quality. Stout rejects this view. (ii) McTaggart saw no objection to one and the same particular being inhered in by a number of different qualities. Stout would apparently object to this even if he thought that there were universal determinate qualities. As regards the second point there are two remarks to be made. (a) Stout's theory assumes that there are universals, though it denies that there are any universal determinate qualities. For, presumably, it is one and the same relation of exact likeness which relates many different pairs of occurrents. And, presumably, the many different continuants are so many different sets, each consisting of a number of dissimilar occurrents interrelated by the same peculiar relation. Now it is difficult to see what objection there could be to a common quality inhering in each of a number of particulars which would not apply equally to a common relation interrelating the members of each of a number of different sets of particulars. (b) Even on Stout's theory there must be something analogous to the co-inherence of several qualities in a single particular. For consider the sensa, which are sensed when two flashes of lightning are seen. They may be exactly alike in colour and dissimilar in outline, or exactly alike in outline and dissimilar in colour, or exactly alike in both respects, or dissimilar in both respects. No one would suggest that, for this reason, each such sensum must be a complex whole composed of two peculiarly interrelated particulars, one of which is susceptible of shape-comparison but not of colour-comparison, and the other of colour-comparison but not of shape-comparison. We should say that there is a single sensum which is like a certain other in one respect and unlike this other in another respect. If it has to be granted in any case that one and the same particular can be the common term of different kinds of likeness and unlikeness to other particulars, it is difficult to see any obvious absurdity in the supposition that one and the same particular might be inhered in by several different qualities.

Before leaving this topic there is one other remark that seems worth making. As we shall see later, McTaggart held that every particular is divided into parts within parts without end. He therefore must have held that every particular is a set of interrelated particulars. We have seen that Stout holds that every continuant is a set of peculiarly interrelated occurrents, and that occurrents are particulars. Now, as we have said, McTaggart never discussed the division of particulars into occurrents and continuants. We therefore do not know what he supposed to be the ground on which this distinction is based. But it seems quite possible that, if he had considered the matter, he would have agreed with Stout that what is called a "continuant" is a group of particulars, none of which would be called "continuants", interrelated in a characteristic way. It is in fact difficult to see what other view he could have taken. Thus it is quite possible that, on this point at least, there is no essential difference of opinion between the two philosophers.

*1.2. Independent Discussion of the Notion of Substance. It will be evident from the above critical account of McTaggart's doctrine of Substance that its main defect is the complete lack of any attempt to discuss the common division of particulars into continuants or "things" and occurrents or "events" or "states". Even if one regards this distinction as not ultimately valid, it is so deeply rooted in our language and our thought that no one is justified in ignoring it. If it is mistaken, it is the business of any philosopher who makes great use of the notion of Substance to indicate the facts which led to the mistake being so commonly made. A secondary defect is that the status of Facts is left uncertain. I have the impression that the notion of Facts was an afterthought in McTaggart's mind, and that he never got quite clear about them or their position in his system. In Chap. IV of this work (p. 57) I have quoted passages from Chap. II of the Nature of Existence which show clearly that he sometimes gave, as examples of facts, entities which he elsewhere regards as events and therefore, in his sense, substances.

I propose therefore to say a little on this topic, though its difficulty is so great that I do not expect to say anything of much value. I shall begin by considering a certain distinction which we all do in fact make, whatever may be the right analysis of it.

*1.21. Processes and Things. There are certain predicate-phrases, such as "going-on", "taking place", "happening", etc., which it seems appropriate to conjoin with certain kinds of substantive names and phrases, and quite inappropriate to conjoin with certain others. It is sensible to say: "There is a noise going on" or "There is a movement taking place". It would be nonsensical to say: "There is a chair going on in my bedroom" or "There are several books taking place in my study ". There is then, prima facie, a distinction between two sorts of substantive, which we will call "Processes" and "Things" respectively. A noise or a movement seems a clear instance of a process, and a chair or a self seems a clear instance of a thing.

It is obvious that Things are substances in McTaggart's sense; and it is fairly easy to see that processes are so too. Processes have characteristics. A noise may be loud, continuous, "buzzy", and so on. A movement may be slow, jerky, rectilinear, and so on. Processes can have temporal relations to each other; e.g., a certain buzzing may partially overlap in time a certain hissing, and one buzzing may be louder than another, and so on. Also it seems quite plain that processes are not characteristics. Is it possible to hold that they are facts? It seems to me that this is not possible. There are certain predicates which can be applied to processes and cannot properly be applied to facts. We can say of a certain process that it is loud and "buzzy "; we can say of a certain other process that it is soft and "tinkly"; and we can say of the two that they go on simultaneously. But surely it would be nonsense to talk of a loud or a soft or a "buzzy" or a "tinkly" fact, or to speak of two facts as "going on simultaneously". Facts do not "go on" nor are they simultaneous, though there is the fact that processes go on and there is the fact that some processes go on simultaneously. It would appear then that Processes answer to McTaggart's description of "substances" and to our description of "particulars".

It will be noticed that our clear instances of Things have been either physical objects, like chairs and books, or minds. Our clear instances of Processes include both physical processes, like the movement of a golf-ball, and sensible processes, like a noise. Now there are particulars which are not clearly instances of either Things or Processes. What are we to say, for example, about a visual image or a visual sensum? We should hesitate to call it a thing, and we should hesitate to call it a process; though we should not hesitate to call a movement of a visual sensum in a visual field a process. For the present I am going to set aside these ambiguous particulars, and to concentrate attention on the clear instances of things and of processes.

We talk of processes "starting", "going on ", and " stopping". I have made the applicability of such predicates the distinguishing mark of processes. But do we never apply them in the same sense to things? We say: "A buzzing started, went on for some time, and then stopped". Now I can, of course, quite properly say of a train that it started, went on, and stopped. But here I plainly mean that a movement of the train started, that it went on for some time, and that it then stopped and was succeeded by a resting of the train which went on for some time. It is plain that here "starting", "going on", and "stopping", as applied to a thing, are derived from these notions as applied to a certain kind of process, viz., a movement.

There is, however, another sense in which we might apply these predicates to things. We might say that a chair "started" when its construction was completed, that it "went on" for some years, until finally it "stopped" when it was broken up. The more natural phrases to use here would be " started to exist", "went on existing ", and "stopped existing". It seems to me that a more accurate expression for these facts would be that a certain set of things, none of which were chairs, started to be a chair, went on being a chair, and eventually stopped being a chair. The set started to be a chair when its members in the course of their movements got into certain spatial and dynamical relations to each other; it went on being a chair so long as they went on resting in these relations; and it stopped being a chair when they started moving out of these mutual relations. Thus the starting, going on, and stopping of things which are recognised to be compounds seems to be analysable in terms of the starting, going on, and stopping of certain processes in other things, viz., their elements, which existed before, and will go on existing after, standing in those special mutual relations which are characteristic of the internal structure of the compound. It seems to me doubtful whether there is any other meaning that one can attach to the coming-to-be and the passing-away of Things. And so, if there be any things which are not compounds composed of other things suitably interrelated, I do not clearly understand what would be meant by saying that they come to be or that they pass away.

It will be remembered that Kant, in criticising the Scholastic argument from the simplicity of the soul to its immortality, said that it might cease to exist by "elanguescence", as a sound dies away without "coming to bits". Now this objection might be interpreted in two different ways. (i) It might mean "Even though the soul be a simple thing, yet it might nevertheless pass away by elanguescence as a sound does". Or (ii) it might mean "It is possible that the soul is not a thing but a very complex process, and in that case it might pass away by elanguescence as a sound does." If interpreted in the first way, Kant's argument seems to me to be plainly invalid; it is meaningless to apply the predicate "passing away by elanguescence" to a subject which is not a process but a thing. On the second interpretation, the argument may be valid; but, in so far as it is directed against the Scholastics, it is an ignoratio elenchi, since they would not have admitted for an instant that the soul might be a process and not a thing.

There is one further remark which may be worth making at this point. It has sometimes been said that, if there were anything in the Scholastic argument, it would prove the pre-existence as well as the post-existence or the soul. The truth of the matter seems to me to be the following. If a soul be a simple thing, then neither its coming to be (if it does come to be) nor its passing away (if it does pass away) is intelligible to us. But I may have good reason to believe that there is a fact corresponding to a certain statement S which is unintelligible to me, and I may have good reason to believe that there is no fact corresponding to another statement S' which is equally unintelligible to me and is of the same general form as S. There might, for example, be two formulae in a treatise on the Theory of Numbers, which were both equally unintelligible to Prof. Littlewood's bedmaker, and which never could be made intelligible to her. Yet, if Prof. Littlewood pointed to the first, and said "That is a true statement about certain numbers", and then pointed to the second and said "That is a false statement about the same numbers", his bedmaker would have very good reason to believe that there is a fact about numbers corresponding to the first and that there is no such fact corresponding to the second. Now let us substitute in this parable God for Prof. Littlewood; any human being, however intelligent, for his bedmaker; and the two statements "The soul comes to be at the time when the body is conceived" and "The soul ceases to be at the time when the body dies" for the two formulae. The Scholastics would say that God has told us, or that we can infer from other things which God has told us, that there is a fact which concords with the first unintelligible statement, and that there is a fact which discords with the second unintelligible statement. This is, so far as I can see, a perfectly consistent position, whether it be in fact true or not.

The next point to notice is that certain temporal phrases can be properly adjoined with names of things, but cannot properly be conjoined with names of processes; and conversely. We can say of a Thing that it has "age"; we cannot say that it has "temporal extension". On the other hand, we can predicate temporal extension of a Process, but cannot talk of its age. We talk of an "old" building, and we speak of it as "ageing" or "getting older". But we cannot talk of the history of a building as being old or getting older. We can, however, talk of the history of a building as being "long" and as "getting longer". We certainly could not speak of a building as "being long" or "getting longer" except in the obviously different spatial sense. We do indeed sometimes talk of "ancient history" as well as of "ancient buildings". But we mean by the former phrase processes which came to an end long ago, and we do not mean this by the second phrase. In the first phrase we could substitute "remote" for "ancient", but we certainly could not do so in the second. To put the distinction in general terms, we talk of Things as "enduring" or "persisting through" a period of time. We talk of Processes as "going on for" longer or shorter periods of time.

It is important to notice the above distinction. Yet it is equally important to notice the following linguistic fact. The two sentences "This thing is old" and "The history of this thing is long" seem to be two different ways of expressing precisely the same fact. Similarly the two sentences "This thing is getting older" and "The history of this thing is getting longer" seem to be two different ways of expressing a single fact.

Closely connected with the above distinction is a distinction with respect to temporal parts. Any process can properly be said to have successive temporal parts. These are shorter processes which together make up the longer process by adjunction, as shorter lines put end to end make up a longer line. Such temporal parts are called "successive total phases". We could say that the process of underpinning St Paul's Cathedral is a part of, or a phase in, the history of St Paul's. We could say that the dome is a part of, but not a phase in, St Paul's. We cannot say that the process of underpinning is a part of the cathedral; and we cannot say that the dome is a part of, or a phase in, the history of the cathedral; though the history of the dome is, no doubt, in some sense a part of the history of the cathedral.

Processes have certain characteristics which depend on the fact that they have temporal extension and are divisible into successive adjoined total phases. We can say of a process, e.g., a noise or a movement, that it is "steady" or that it is "fluctuating". If it is fluctuating, we can say that it varies "suddenly" or "continuously", that it varies "periodically" or "non-periodically", and so on. It is evident that these characteristics of a longer process depend on the characteristics of the successive shorter phases which together make it up. Such adjectives as these cannot be conjoined with the names of Things. Nevertheless, the two sentences "There is a periodically fluctuating movement going on" and "Something is moving periodically", seem to be just two different ways of expressing the same fact.

The next point to be considered is this. A man who is sitting in a room with me may say "I hear a buzzing noise". Then, after an interval, he may say "I don't hear that buzzing any longer". And then, after another interval, he may say "Now I hear that same buzzing noise again". He may also say "That chair which I see today is the same which I saw yesterday, but it is now in a different place". Now, as regards the first series of statements, I think that anyone would, on reflexion, accept the following as an equivalent and more accurate way of expressing his meaning. "There was a noise which I did hear a short time ago. That is no longer going on, and I am no longer hearing it. There is a noise which I am hearing. This was not going on, and I was not yet hearing it, when that was going on and I was hearing that. But that and this are specially closely related, either directly or indirectly, so that they may be regarded as successive, though not adjoined, phases in a single process." But no one, without a great deal of argument, would admit that any statement of this kind expresses what he meant by his statement about the chair. What any ordinary man believes is that in this case one and the same particular has persisted through a period of time, has been seen by him on two successive and separated occasions, and has been in one place on one occasion and in another place on the second occasion. There is for him no question of This and That in the present case. "This, which I saw yesterday and now see again, was there and is now here" expresses the plain man's view of the situation.

Closely connected with the distinction just mentioned there is another distinction between Things and Processes. There are certain kinds of adjective which may be called "dispositional adjectives". Obvious examples are words like "poisonous", "fusible", "massive", etc. These are properly conjoined with thing-names and not with process-names. We speak of a bit of arsenic as "poisonous", of a bit of wax as "fusible", of a bit of gold as "massive", and so on. Many adjectives are ambiguous in this respect. I should say of a pillar-box that it is "red". If I came to distinguish between the pillar-box and the visual sensibile which I sense when I look at a pillar-box, I should say that the sensibile is "red" too. As applied to the pillar-box the word "red" is a dispositional adjective; as applied to the visual sensibile it is non-dispositional. By saying that the pillar-box is red I mean at least that, if any normal observer were to look at it in daylight, it would look red to him. And I might mean no more than this. By saying that the visual sensibile is red I mean something which could not possibly be expressed by a conditional sentence. A man may believe that the pillar-box is red in the non-dispositional sense also. Most men do not explicitly distinguish between visual sensibilia and the surfaces of physical objects; and, since it is quite clear when one is looking at a pillar-box, one is acquainted with something that is red in the non-dispositional sense, it is natural to ascribe non-dispositional redness to the pillar-box itself. I think that, if we tried to express for them the view which plain men cannot express clearly for themselves, it would be somewhat as follows. "The pillar-box is red in the non-dispositional sense, and that is why it would look red to any normal observer who viewed it in daylight; i.e., its non-dispositional redness is the ground of its dispositional redness." However this may be, the facts are as follows. We must distinguish between the visual sensibile which I sense when I look at a pillar-box and the surface of the pillar-box. And, when the distinction has been made, we see that it is almost certain that the pillar-box is red in the dispositional sense, that it is highly doubtful whether it is red in the non-dispositional sense, that it is quite certain that the sensibile is red in the non-dispositional sense, and that it is nonsensical to say that the sensibile is red in the dispositional sense. Almost every adjective which we are justified in ascribing with any confidence to a physical thing is either explicitly dispositional or is an ambiguous adjective used in its dispositional sense.

Now whenever we conjoin a dispositional adjective to a substantive we are expressing in a categorical form a hypothetical proposition of the following kind. "If this were in a certain state, and were in certain relations to certain other things of certain specified kinds, then certain events of a specific kind would happen either in it or in one of these other things." Now such a statement implies at least that "This" is the name of something which may be in various states, and may stand in various relations to other things, at various times. It implies more even than this. It implies that this, which in fact was in a certain state and stood in certain relations to certain things at a certain time, might instead have been in a different state or stood in different relations to the same or different things at that time, and would then have behaved in a certain specifically different way from that in which it in fact did behave then. I admit the extreme difficulty of analysing the meaning of such statements as these, and of knowing whether any of them could possibly be true. But it is a fact that such sentences are constantly being spoken and written in daily life, and that their grammatical subjects are always thing-names and not process-names.

The contents of this sub-section may now be summed up as follows. In Indo-European languages, at any rate, there are at least two kinds of substantive-name, viz., thing-names and process-names. There are several different kinds of adjective-phrase which can be conjoined with thing-names to give intelligible sentences. If any of these be conjoined with process-names the result is nonsense. Similarly, there are several kinds of adjective-phrase which can be conjoined with process-names to give intelligible sentences. If any of these be conjoined with thing-names the result is nonsense. This linguistic fact may suggest that there are two fundamentally different, though no doubt closely interconnected, kinds of particulars. It certainly makes it incumbent on any philosopher, such as McTaggart, who thinks otherwise, to go very carefully into the question and to give very good reasons for regarding the prima facie distinction as mistaken. On the other hand, we must not take the linguistic distinctions in the only group of languages with which most of us happen to be familiar too seriously. We have seen that, in some cases at least, two sentences, one with a thing-name and the other with a process-phrase as grammatical subject, seem to be simply two different ways of expressing precisely the same fact. (Cf., e.g., "This is old and is getting older" and "The history of this is long and is getting longer".) It is conceivable then that we are concerned here with something analogous to the fact that precisely the same geometrical relation, e.g., that of collinearity, may be expressed either in Cartesian or in Polar Co-ordinates, and will look extremely different in the two modes of expression.

Perhaps I may use the above analogy as a defence against a critic who might say that it is futile to approach metaphysical subjects by way of language and grammar. One might just as well say that it is futile to approach geometrical and physical problems by way of co-ordinates. Some system of co-ordinates must be used. And the geometrical and physical facts which are independent of co-ordinates will emerge as a certain community of form between different expressions of the same fact in different systems of co-ordinates.

*1.22. Can either Things or Processes be dispensed with? In the first place, what exactly does this question mean? Suppose it were found that sentences which contain thing-names could all be replaced, without loss or gain of meaning, by sentences which contain process-names and do not contain thing-names. Then we could say that Things can be dispensed with in favour of Processes. Suppose, on the other hand, that the exact opposite of this were true. Then we could say that Processes can be dispensed with in favour of Things. A third possibility is the following. It might be that sentences containing thing-names and sentences containing process-names could both be replaced, without loss or gain of meaning, by sentences containing neither thing-names nor process-names but a certain other kind of name. We could then say that both Things and Processes can be dispensed with in favour of a certain third kind of entity.

Now many people have held that it is self-evident that any process, whether of "change" or of "quiescence", must be a "state of" or a "process in" a thing. If so, it would seem clear that things cannot be dispensed with in favour of processes. Some people have gone a step further. They have said that the statement that a certain process is going on in a certain thing can be replaced, without loss or gain of meaning, by the statement that this thing is the common subject of a certain set of facts of a peculiar kind. The facts are of the form: "x has the characteristic c1, at t1, x has the characteristic c2 at t2 ... x has the characteristic cn at tn". Here c1 ... cn are determinates under a certain determinable C, and each of these facts except the first and last contains as a constituent one and only one of the moments between t1 and tn. It may happen that c1 ... cn are all identical. We should then say that x has been in a state of quiescence with regard to C throughout the interval between t1 and tn. Otherwise we should say that x has been in a state of change with respect to C throughout this interval.

Is it really obvious that every process must be a state of, or a process in, some thing? Consider the two statements: "There is a noise going on" and "There is a movement taking place". There is one kind of question which we can reasonably ask about both, viz., "What kind of noise?" or "What kind of movement?". But, in the second case, it seems plain that we can raise the further question "What is moving?" The answer might be "This golf-ball " or "This wave" or "This shadow" or "This coloured patch in my visual field". (I have purposely introduced this particular selection of answers in order to show that, in some cases, what would be said to be "moving" is something that would not commonly be said to be a "thing".) Now can this kind of question be raised in connexion with every process, or only in connexion with some? In particular can it be raised in connexion with a buzzing or a hissing?

When I am told that there is a buzzing going on I can, of course, ask "What is buzzing?" This is verbally similar to the question "What is moving?" The answer that I should expect to get would be "A bee" or "A gnat", or something of that kind. Some physical object would be mentioned in which a physical process, such as the rapid rhythmic movement of wings, is supposed to be going on and causing the "buzzy" process. This shows that the question means "In what physical thing is the process going on which is responsible for the noise which I am hearing?" But that is not in the least what I mean when I ask "What is moving?", and am told "It is a certain coloured patch in a visual field" or "It is a golf-ball". Here I am not asking about what is responsible for the movement, but what is the subject of the movement. The question is whether I can reasonably raise the question "What is the subject of this buzzing?", when it is clearly understood that by "this buzzing" I am referring to an auditory process which is "buzzy" and not to a physical process of rhythmic movement. >p>
In order to put the case quite fairly we ought to compare statements of the same level of determinateness. We ought to compare "There is a noise going on" with "There is a movement going on", or to compare "There is a buzzing going on" with "There is a circling going on". Granted that the two statements about movements are to be interpreted to mean respectively "Something is moving" and "Something is moving in a circular way", what are we to say of the two statements about noises? Those who hold it to be self-evident that every process is a state of some thing must, presumably, be prepared to say that the two statements about noises are to be interpreted to mean respectively "Something is noising" and "Something is noising in a buzzy way". It will be noticed at once that this analysis is so far-fetched that I have had to invent the word "noising" in order to express it. Of course we have the word "sounding "; but this plainly would not have served our purpose, for it is bells, trumpets, and other sonorous physical objects which are said to "sound".

The next point to notice is this. When we ask "What is moving?" we know what kind of answer to expect. We shall be told "It is a golf-ball" or "It is a wave" or "It is a red circular patch in my visual sense-field", or something of that kind. A physical thing or process, or a visual sensibile, will be indicated or described to us in answer to our question. But what kind of answer would one expect to the question "What is noising?", if one could bring oneself to ask it? I have not the faintest idea.

No doubt many people believe, or think they believe, that noises are, in some sense, "mental". But even they, when told that a buzzing was going on, would hardly be prepared to say "Some mind is buzzing; or, to speak more accurately, some mind is noising buzzily". The mentalist's difficulty here forces on our attention a distinction which I have not yet explicitly drawn. I stated the theory, which we are at present examining, in the form that every process is a "state of" or a "process in" some thing. We have seen that, if "P is a state of T" means "T is the subject of the process P", it is far from obvious that noises are states of anything, and most difficult to conceive what they could be states of. It might be suggested at this point that P can be a process in T without being a state of T, though P cannot be a state of T without being a process in P. The mentalist might then say that a buzzing is necessarily a process in a mind, though a mind is never the subject of such a process. Either noises are not states of anything, or else they are states of things which are not minds but are certain parts of minds. It is not, perhaps, ridiculous to suggest that a mind might have certain parts which are subjects of processes such as buzzing, hissing, etc.

This expedient might help the mentalist, but we are not specially interested in him and his troubles. It does nothing to help the theory which we are discussing. For the question now becomes: "Is it obvious that every process is a state of some thing; i.e., that every process has some thing for its subject in the sense in which, for example, a certain movement has a certain golf-ball for its subject?" And we are no nearer than before to seeing what, if anything, is the subject of a noise-process. Even if we accepted mentalism we could only say that, if those processes which are noises have subjects which are things, these things, though they are certainly not minds, must be in some sense mental.

At this stage it seems worth while to raise the following question. Granted that a noise is a process and that a movement is a process, how close is the analogy between the two? In the case of movement we can distinguish between (a) a state of resting, and (b) a state of moving. I can literally "see" a thing resting, and I can literally "see" it moving. Then among states of movement we can distinguish between those which are constant in direction and uniform in speed and those which are not. Among the latter we can distinguish those which are constant in direction and non-uniform in speed (accelerated rectilinear movements), those which are constant in speed and non-uniform in direction (e.g., uniform circular movements), and those which vary in both respects, and so on.

Now it might seem plausible to hold that the speed of a movement is analogous to the intensity of a sound, and that the direction of a movement is analogous to the pitch of a sound. I do not know what would be analogous to the tone-quality of a sound unless it might be compared with the colour of a moving sense-object. If we accept these analogies, we must notice one profound difference between sounds and movements. We can literally "see" certain objects in a state of rest, i.e., of positional quiescence; they "look" characteristically different from similar objects in a state of motion, i.e., of positional change. When I "see something start to move" I see this object first in a state of positional quiescence and then in a state of positional change. Now there is no analogy to this in the case of noises. I do often have the kind of experience which I should describe by saying that "I hear a certain kind of noise start". But I certainly could not say that in such cases I first heard something in a state of auditory quiescence and then heard it in a state of auditory change. Rest seems to be a lower limit of motion, as speed decreases indefinitely, and both rest and motion can be "seen". The lower limit of noise, as intensity decreases indefinitely, is quiet; but, whilst noise can be "heard", quiet cannot be "heard".

Such facts as these seem to me to increase the difficulty of holding that, whenever a noise is going on, there is some thing which is "noising". We shall not only have to admit that we have no idea what sort of things are the subjects of this sort of process. We shall also have to admit that we have no idea of the characteristic in respect of which such things are changing when they become the subjects of such processes. The characteristic of spatial position is familiar to us because we see some things resting, others moving among them, and things which we have seen resting beginning to move. But, when a thing which is capable of "noising" is quiescent in respect of that characteristic which varies when such things "noise", it just ceases altogether to be an object of acquaintance to us.

The upshot of the discussion is that it is very far from clear that every process must have a subject which is a thing. There are some processes, e.g., movements, with regard to which this principle is highly plausible; but there are others, e.g., noises, with regard to which it is not plausible at all. We must therefore be prepared to admit the possibility of what I will call "Absolute Processes".

If it is doubtful whether all processes are states of things, it is still more doubtful whether processes could be dispensed with in favour of things and certain sets of facts about things. The latter view can, however, be attacked more directly. Undoubtedly, whenever there is the fact that a thing is moving there is also a set of facts of the form: "x is at s1, at t1, x is at s2 at t2, ... x is at sn at tn". And whenever there is the fact that a thing is resting there is also a set of facts of the form: "x is at s at t1, x is at s at t2, ... x is at s at tn". It is also true that for every different kind of movement, e.g., rectilinear or circular, uniform or accelerated, and so on, there will be a corresponding difference in such a set of facts about the occupation of positions at instants. But we must remember that a thing can quite literally be "seen" to rest and "seen" to move, just as it can be seen to be black or seen to be green. Again, a movement can quite literally be "seen" to be constant or variable in direction, in speed, and so on. Surely it is nonsensical to talk of "seeing" (in this quite literal sense) a set of facts, or of "seeing" such a set as having certain sensible characteristics. There is the movement of x, with its characteristic peculiarities, and there is the fact that x moved in a characteristic way. The latter can be "analysed", in one sense, into a conjunction of facts of the kind indicated above. The former can be "analysed", in another sense, into an adjunction of phases. Each kind of analysis is most intimately correlated with the other, but nothing that can be analysed in one of these ways can be analysed in the other of them. I conclude then that Processes cannot be dispensed with in favour of Things and Facts.

We can now pass to the opposite question. Can Things be dispensed with in favour of Processes? Let us begin with those processes which seem most unfavourable to the theory under discussion, viz., physical movements. Is it not obvious that here at least there is never a movement without some thing that moves? To this one might answer that plain men and scientists do constantly use sentences of the form "x is moving", where "x" is seen on reflexion not to be the name of a thing. For everyone, except Prof. Prichard and some of his pupils, talks quite shamelessly of waves and of shadows as "moving"; and no one regards a wave or a shadow as a thing. Now we know that anything that Prof. Prichard maintains is likely to be important and worth very serious consideration. And we do not want to quarrel about words. I propose therefore to substitute the words "transmission of states" and "translation of stuff" for the word "movement" as applied respectively to a wave and to a golf-ball. I think that Prof. Prichard's doctrine could then be fairly stated as follows. "No doubt there is transmission of state, and people do often use sentences, which are of the appropriate form for expressing translation of stuff, to express transmission of state. But (a) there is translation of stuff as well as transmission of state. And (b) whenever there is transmission of state, that which is transmitted is a state of some thing."

Let us now return to the question of waves and their "motion". When a person says that a wave is moving in a certain direction and with a certain velocity what he means is the following. There is a track in some thing, e.g., in a pool of water. Each particle of water in this track is being translated to and fro about a fixed mean position in a certain characteristic period. And, if a particle which occupies a mean position x in this track has reached any point y in its vibratory course at any moment t, then the particle which occupies the mean position x + xi in this track will reach the corresponding point in its vibratory course at a moment t + xi/v, where v is a constant which is characteristic of the process. Here it is quite obvious that transmission of state presupposes translation of stuff, for the essence of the business is that each particle is the subject of a to-and-fro movement of translation about a fixed mean position. Now there are certainly processes which appear, to be non-periodic translations of things, e.g., the flight of a golf-ball from a tee to the next green. It would be possible to hold that all such apparent instances of non-periodic translation of things are really instances of transmission of states, and that they are analogous to the "motion" of waves. But this would be of no philosophic interest if the states transmitted be all themselves instances of periodic translation of things. And, if one has to admit the occurrence of periodic translation of stuff, there is no reason why there should not be non-periodic translation of stuff, as there prime facie appears to be.

I should agree then with what I suppose that Prof. Prichard would assert, viz., that the existence and the success of "hydrodynamical " theories of the atom, such as Lord Kelvin's Vortex Theory, have no tendency to show that the notion of translation of stuff can be dispensed with. At most they would show that all those macroscopic processes which have commonly been regarded as translations of material particles are really transmissions of state. But the states transmitted would be states of periodic translation of the particles of some other kind of stuff, viz., the Ether, whatever that may be.

This, however, is by no means the end of the matter. In the first place, suppose one were to take the Newtonian theory of Substantival Absolute Space seriously, as I have gathered that Prof. Prichard does. Then, it seems to me, one could dispense with the Ether, and with translation of stuff, altogether. For the periodic translation of an ether-particle about a mean position one could substitute the periodic pervasion of a set of regions in Absolute Space by a certain determinate quality. And one could then deal with the apparent non-periodic translation of material particles on the same lines as before. There would be no "things", as distinct from regions of Absolute Space; and the latter would not of course "move" in any sense of that ambiguous word. Ultimately nothing could be said to "move" except qualities; and a quality would "move", in the sense that it pervaded now one and now another region of Absolute Space. It might be suggested then that one must admit either translation of stuff or Substantival Absolute Space, but that, if you are prepared to accept the latter, you can dispense with the former.

Is it possible to get beyond this point? The first thing to notice is that we do talk of what are admittedly Processes as "changing" in certain respects, or "remaining constant" in certain respects. Everyone would understand me if I were to say "I have been hearing a certain noise for some time now; it has remained of the same pitch, but it has got steadily louder". I can be said to "hear" the change in loudness and the constancy in pitch just as properly and literally as I can be said to "hear" the noise itself. It seems obvious that there is a fairly close analogy between what is expressed by the following two statements (a) "I have been 'hearing' a certain noise for some time, and I have been 'hearing' it altering in loudness", and (b) "I have been 'seeing' a coloured patch in my visual field for some time, and I have been 'seeing' it altering in position, i.e., moving". Now the statement about the noise seems to be capable of analysis. The noise is a long process, composed of temporally adjoined shorter phases, each of which is itself a noise. At each different moment within a certain period I have sensed a different short phase of this longer process. If two moments be near enough together, the phase which I sensed at the second of them partially overlaps the phase which I sensed at the first of them. There is a phase which I sensed at the first moment, and have ceased to sense at the second; there is a phase which I sense at the second, and was not yet sensing at the first; and there is a phase which I sense at both. The nearer the two moments are together the more nearly does that which is sensed at the second coincide with that which was sensed at the first. The identity, which enables me to talk of "this noise" and to say that I have been "hearing it" for some time, resolves itself into the peculiarly intimate way in which these successive and partially overlapping noises are interrelated in respect of their qualities. It is in virtue of this that they count as different phases in a single process. The diversity, which enables me to say that this noise "has been changing" and that I have "heard it changing" resolves itself into certain qualitative dissimilarities between these successive and closely interrelated phases.

We must now try to give a more accurate account of this vague notion of "qualitative dissimilarity" and "similarity" between successive phases of a process. It seems to me that we must introduce a conception which I will call that of "Quality-ranges". Suppose we take a certain noise which began suddenly, went on for a period with continually increasing loudness, and then suddenly stopped. It is plainly nonsensical to ascribe any determinate degree of loudness to the process as a whole or to any phase of it. If anything here could be said to have a determinate degree of loudness it would be an instantaneous cross-section of the process. This would not be a phase of it, and it has all the appearance of being a highly artificial intellectual construction. But we could say of the process that it has a certain "range of loudness", and we could say of every phase of the process that it has a certain range of loudness. Quality-ranges, in this sense, belong only to processes and their phases, and it is doubtful whether any other characteristic but quality-ranges can properly be predicated of processes and phases. Now we are accustomed to regard determinate loudnesses as primary, and to define the notion of a range of loudness in terms of them. I wish to suggest that, however convenient this may be, in practice, it is philosophically a reversal of the true order. I am going to take the notion of quality-range as fundamental, and to show that the notion of determinate quality can be derived from it.

I will now state some of the most important facts about quality-ranges. (i) They have magnitude, and the magnitude may, as a special case, be zero. Suppose that, in ordinary language, a noise started with a certain loudness and stopped with the same loudness, then, no matter what variations of loudness it may have undergone in the interval, the loudness-range of the process as a whole is zero. (ii) Some phases of a process may have a greater quality-range than the process as a whole has. In particular, the quality-range of a process may be zero, whilst that of every phase of the process is finite. Suppose, for example, that, in ordinary language, a noise started with a certain loudness, went up to a maximum, then dropped, and finally ceased suddenly with the same loudness with which it began. Then its range of loudness would be zero, but every phase of it would have a finite range of loudness. (iii) If, on the other hand, every phase of a process has zero range of loudness, then the process as a whole has zero range of loudness. In such cases, and in such only, we say that the noise has "remained constant in loudness". (iv) The following proposition is commonly assumed to be self-evident. Let P be any process, which has a quality-range of a certain kind, e.g., a range of loudness. Let epsilon be any degree of this quality-range, no matter how small. Then there is an integer nepsilon, such that any phase of P whose duration is less than 1/nepsilon of the duration of P will have this quality-range to a less degree than epsilon. This is the assumption which underlies the proposition that any noise could be analysed into an infinite series of successive "instantaneous events" each with a perfectly determinate degree of loudness. The cash-value of this statement is that, if you take any range of loudness, however small, then any noise will have a set of adjoined successive phases, such that every member of this set will have a lesser range of loudness than the assigned one. (v) Quality-ranges of the same kind, e.g., ranges of loudness, can differ in what may be called their "position on a scale" as well as in degree. Thus, two noises may have the same degree of loudness-range, whether finite or zero, and yet differ in their positions on the scale of loudness. There might, for example, be two noises which were constant in loudness, and therefore both had zero loudness-range, and one of them might be louder than the other. If two noises have finite ranges of loudness, the latter may be either co-terminous or not co-terminous on the scale. If co-terminous, they may either not overlap or one may wholly overlap the other. If they are not co-terminous, then either (a) they will be wholly separated on the scale, or (b) they will partially and only partially overlap, or (c) one will wholly overlap the other. The five possibilities are illustrated by straight lines in the following diagram:

It seems evident that two immediately successive noises, each of finite range of loudness, must have conterminous ranges if they are to be successive phases in a single noise. On one alternative the noise will continually increase in loudness; on the other it will go up to a maximum and then go down again.

The position which we have now reached is the following. (i) We have argued that some Processes, at any rate, seem to be absolute and not to have any Thing for their subject. We have given noises as a plausible instance of Absolute Processes. (ii) We have now seen that certain predicates, such as "changing in respect of a certain characteristic" and "remaining identical through change", can properly be applied to processes, such as noises, even though these should be absolute. For we have seen how sentences with such subjects and predicates are to be analysed, and we have seen that the equivalent sentences do not contain any word or phrase standing for an identical Thing or Subject which persists through a period of time. (iii) We also saw, in dealing with physical motion, that all processes which are commonly regarded as translations of macroscopic things can equally well be regarded as transmissions of states. The question remained whether these transmitted processes were not themselves states of microscopic things. (The words "macroscopic" and "microscopic" are here used in the technical sense in which physicists employ them. They mean, roughly, "in principle perceptible by the senses, if aided by theoretically perfect instruments" and "in principle imperceptible by the senses, however aided", respectively.) (iv) If, now, it be admitted that there may be absolute processes, and that there is a perfectly good sense in talking of "changes taking place in" such processes, it becomes possible to suggest that the microscopic processes are absolute, and that these macroscopic processes which are commonly regarded as translations of things are really transmissions of microscopic absolute processes.

This last suggestion now needs some further explanation and elaboration. Up to the present we have said only that there is a "fairly close analogy" between a noise persisting and changing in loudness, on the one hand, and a patch persisting and moving about in a visual field, on the other. It might be answered that coloured patches in visual fields are of the nature of Things, not of Processes; that changes in loudness or pitch are not movements in any sense whatever; and that therefore we have not refuted the contention that, whatever other kinds of movement there may be, there must be some movements which are literally translations of things, and thus some processes which are, in a quite ultimate sense, states of things. We are thus at last brought face to face with a question which we set aside at an earlier stage of this discussion, viz., "What is the nature of visual sensibilia, images, etc.?"

Before tackling this question I am going to make a further remark about noises, which, if it expresses a correct observation, is highly important in the present connexion. I am inclined to think, though I am by no means certain, that I have experiences which can properly be described by saying that I sometimes "hear a sound remaining stationary" and sometimes "hear a sound moving". When a car stops outside my window with the engine running I have the first kind of experience, and when it starts moving along the street I have the second. The doubt which attaches to the interpretation that I have tentatively put on these experiences is, of course, the following. I cannot feel quite sure that these are purely auditory experiences. Undoubtedly, when the car begins to move, I quite literally "hear" a certain noise changing in certain respects in which it was before constant, e.g., in intensity. And it is possible that I do not literally "hear" the noise moving, but merely believe, on the basis of these auditory changes and of past visual experiences, that the physical object which is "making the noise" is moving. There is no such doubt about the experience which I describe as "seeing a coloured patch moving in my visual field". Here it is certain that I quite literally "see" a peculiar kind of change, which we call "sensible movement".

Now, supposing for the moment that I do have experiences which can properly be described as "hearing a noise remaining stationary" and "hearing a noise moving", it seems plain that they would have to be analysed in the same general way as the experiences which I describe as "hearing a noise keeping constant in loudness" and "hearing a noise changing in loudness". We should have to suppose that there is a peculiar kind of quality-range, which we will call "Place-range". Every noise will have some place-range, finite or zero, just as every noise has some loudness-range. And every phase of a noise will have some place-range. Place-ranges, like loudness-ranges, will have position on a scale, though the scale in this case will probably not be one-dimensional as it is in the case of loudness-ranges. If a noise is such that every phase in it has zero place-range, we say that the noise remains stationary. If, on the other hand, successive adjoined phases of the noise have finite place-ranges, and the place-ranges of adjoined phases are co-terminous in their position on the place-range scale, we say that the noise is moving about. The notion of a perfectly determinate "instantaneous place" can be derived in the same way, and on the same assumption, as the notion of a perfectly determinate "instantaneous loudness".

The discussion of this hypothetical example at least shows that there is no difficulty of principle in dealing with rest and motion entirely in terms of absolute processes, their successive phases, and certain quality-ranges of such processes and their phases. It remain to consider whether the rest and motion of visual sensibilia and images can in fact be dealt with in this way. Now it seems antecedently unlikely that visual sensibilia and images should be fundamentally different in nature from auditory ones. Everyone admits that the latter are processes, and so there is an antecedent, probability that the former are too. It seems to me reasonable to suppose that the fundamental visual particulars are processes which, following Mr Wisdom's example, we might call "Colourings". A colouring might be a "redding" or a "greening", just as a noise might be a buzzing or a hissing.A colouring does not have a determinate colour, any more than a noise has a determinate loudness; but it has a colour-range, which may be zero or finite, just as a noise-process has a loudness-range. The phases of a colouring are themselves colourings and have colour-ranges, finite or zero. Now colourings, in addition to having colour-ranges, have shape-ranges, extension-ranges, and place-ranges. And all their phases have all these quality-ranges. It is, of course, nonsensical to ascribe a determinate shape, extension, or place to a process or to its phases; but it is possible to give a meaning to the statement that a certain instantaneous cross-section of a colouring has a certain determinate colour, shape, size, and place. We have seen how to do this in the simple case of noises and their loudness, and no difference of principle is involved.

Why are we so much inclined to regard visual sensibilia as different in nature from auditory ones? The explanation lies in the following purely contingent fact. A great many colourings go on for very considerable periods, and are such that every phase of them has zero or nearly zero range of colour, extension, place, and shape. Again, even when successive phases of a colouring have finite and co-terminous place-ranges, they often have zero or nearly zero colour-range, shape-range, and extension-range. In such cases we talk of a "patch" of constant size and colour moving visibly about. It may also happen that successive phases of a colouring have finite and co-terminous colour-ranges, but have zero or nearly zero place-range, shape-range, and extension-range. In such cases we talk of a "patch", of constant shape and size, resting and visibly undergoing changes of colour. Now noises seldom go on for very long. If they do, it is not very common for every phase of them to have zero or nearly zero loudness-range and pitch-range. Moreover, noises have far fewer kinds of quality-range than colourings have. In particular, they seem not to have shape-range, and it is doubtful whether they have extension-range. I have suggested that they do have place-range, but I have admitted that this might be questioned. These differences between visual and auditory sensibilia seem adequate to explain the fact that we hesitate to regard the former as processes whilst we unhesitatingly regard the latter as processes. It may be remarked that no one hesitates call a flash an "event" or "process", and that it is almost incredible that what we are acquainted with when we are said to "see a flash" should be fundamentally different in nature from what we are acquainted with when we are said to "see a coloured patch". My own view is that what we are acquainted with in each case is a short event. In the case of the flash this event is so short as to be wholly contained in a single specious present, and it is not a phase in a longer process of colouring. In the other case we are acquainted, at each one of a series of moments, with a different one of a series of short colourings which are successive phases in a long process of colouring.

We can now bring this long discussion to an end. Statements which grammatically predicate motion, or rest, or qualitative change, or qualitative quiescence, of Things, seem to be replaceable, without loss of meaning, by more complicated statements about Processes, their phases, and the quality-ranges of processes and their phases. On the other hand, there are Processes which cannot plausibly be regarded as states of Things. Thus there seems reason to think that the notion of Things could be dispensed with in favour of the notion of Absolute Processes. This does not mean that the notion of Things is invalid; but only that it is less ultimate than the notion of Processes. It must also be remarked that we have not dealt with dispositional properties, or with the conception that a thing might have been in different circumstances at a given moment from those in which it in fact was, and that it would have behaved in a characteristically different manner from that in which it in fact did. This question may be considered more conveniently at a slight later stage, in connexion with McTaggart's Principle of Extrinsic Determination, where it is highly relevant. Until there we cannot be sure whether the notion of Things can be dispensed with in favour of Processes.

Contents -- Chapter 8