It is now time to gather together the various threads of the earlier chapters, and to see whether we can come to any conclusions about the probable position and probable prospects of Mind in the Universe. It appears to me that seventeen different types of metaphysical theory are possible theoretically on the relation between Mind and Matter. I will first proceed to justify this very startling statement, and to enumerate, classify, and name the theories. Afterwards I shall consider the strong and weak points of each, and see whether we can come to any tentative decision between them.

The Seventeen Types of Theory.

    In order to understand the discussion that follows the reader should refer back to the section on Pluralism and Monism in Chapter 1, where I defined the notion of "Differentiating Attributes" and distinguished them from other kinds of attribute. He should also refer to Chapter II, where I distinguished between those non-differentiating attributes which are "Emergent" and those which are not. I propose here to call non-differentiating attributes which actually apply to certain things in the world, but are not emergent, "Reducible Attributes". It will be necessary to introduce one further distinction which we have not so far made use of. Some attributes have application, i.e., there are things in the Universe which have these attributes in some determinate form. Other attributes have no application. The characteristic of being a fire-breathing serpent, or of being mistress of the Duke of Bletchley, applies to nothing in the world. Now it is held by many people that there are characteristics which do not in fact apply to anything but which seem to some or all men to apply to something. E.g., if Dr M'Taggart be right, it can be proved that the characteristic of being extended cannot apply to anything. But it certainly seems to all men as if there were extended things. I propose to call a characteristic which seems to apply to certain things, but does not in fact apply to anything, a "Delusive Characteristic". I am going to use words in such a way that Differentiating Attributes, Emergent Qualities, and Reducible Qualities, are to be understood to have application and therefore not to be delusive. With these preliminary explanations we can pass to our classification of theoretically possible types of metaphysical theory about Mind and Matter.

    We have to consider the two attributes of "mentality" and "materiality". We at once find three great divisions of possible theories.

1. Both mentality and materiality might be differentiating attributes.
2. One might be a differentiating attribute and the other not. Or
3. it might be that neither is a differentiating attribute.

We now proceed to divide up these three types of theory in turn.

    (1, 1) Both mentality and materiality may be capable of belonging to the same substance; or
(1, 2) it may be that no substance can have both these differentiating attributes.

    (2, 1) Mentality might be a differentiating attribute and materiality not; or
(2, 2) materiality might be a differentiating attribute and mentality not.
We now further subdivide these alternatives as follows.
(2, 11) Materiality, though not a differentiating attribute, might still have application; or
(2, 12) materiality might be a delusive characteristic. Similarly
(2, 21) mentality, though not a differentiating attribute, might still have application; or
(2, 22) mentality might be a delusive characteristic.
The alternatives (2, 12) and (2, 22) need no further subdivision ; but the alternatives (2, 11) and (2, 21) both need further subdivision. Let us begin with (2, 11). It might be that materiality is
(2, 111) an emergent characteristic; or
(2, 112) that it is a reducible characteristic.
Similarly, it might be
(2, 211) that mentality is emergent ; or
(2, 212) that it is reducible.
This completes the subdivisions of alternative (2).

    We pass now to the subdivisions of alternative (3). Granted that neither mentality nor materiality is a differentiating attribute, there are three alternatives open.
(3, 1) Both attributes might have application; or
(3, 2) one might have application and the other be delusive; or
(3, 3) both might be delusive.
The last alternative needs no further subdivision; the first two require to be further subdivided. We will begin with (3, 1). If mentality and materiality both have application, they may
(3, 11) both be emergent; or
(3, 12) one maybe emergent and the other reducible; or
(3, 13) they may both be reducible.
The first and third of these alternatives need no further subdivision, but the second divides into two. It may be
(3, 121) that mentality is emergent and materiality reducible; or
(3, 122) that materiality is emergent and mentality reducible.
It remains to subdivide (3, 2). If one of the attributes has application and the other is delusive, it may be
(3, 21) that mentality has application and materiality is delusive; or
(3, 22) that materiality has application and mentality is delusive.
Each of these latter alternatives subdivides into two viz.,
(3, 211) that mentality is emergent; or
(3, 212) that mentality is reducible; or
(3, 221) that materiality is emergent; or
(3, 222) that materiality is reducible.

    We have now got our seventeen alternative theories, which I will recapitulate and name.

    (1, 1) Mentality and materiality are both differentiating attributes which can belong to the same substance. This I will call "Dualism of Compatibles".

    (1, 2) Mentality and materiality are both differentiating attributes, but they cannot both belong to the same substance. This I will call "Dualism of Incompatibles".

    (2, 12) Mentality is a differentiating attribute, but materiality is delusive. This I will call "Pure Mentalism".

    (2,22) Materiality is a differentiating attribute, but mentality is delusive. This I will call "Pure Materialism".

    (2, 111) Mentality is a differentiating attribute, and materiality is an emergent characteristic. This I will call "Emergent Mentalism".

    (2, 112) Mentality is a differentiating attribute, and materiality is a reducible characteristic. This I will call "Reductive Mentalism".

    (2, 211) Materiality is a differentiating attribute, and mentality is an emergent characteristic. This I will call "Emergent Materialism".

    (2, 212) Materiality is a differentiating attribute, and mentality is a reducible characteristic. This I will call "Reductive Materialism".

    (3, 11) Neither mentality nor materiality is a differentiating attribute, but both are emergent characteristics. This I will call "Emergent Neutralism".

    (3, 13) Neither mentality nor materiality is a differentiating attribute, but both are reducible characteristics. This I will call "Reductive Neutralism".

    (3, 121) Neither mentality nor materiality is a differentiating attribute, but mentality is an emergent characteristic and materiality is a reducible characteristic.

    (3, 122) Neither mentality nor materiality is a differentiating attribute, but mentality is a reducible characteristic and materiality is an emergent characteristic. I class these two alternatives together under the name of "Mixed Neutralism".

    (3, 211) Neither mentality nor materiality is a differentiating attribute, but mentality is emergent and materiality is delusive.

    (3, 212) Neither mentality nor materiality is a differentiating attribute, but mentality is a reducible characteristic and materiality is delusive. I class these two alternatives together under the name of "Mentalistic Neutralism".

    (3, 221) Neither mentality nor materiality is a differentiating attribute, but mentality is delusive and materiality is an emergent characteristic.

    (3, 222) Neither mentality nor materiality is a differentiating attribute, but mentality is delusive and materiality is a reducible characteristic. I class these two alternatives together under the name of "Materialistic Neutralism".

    (3, 3) Neither mentality nor materiality is a differentiating attribute, and both of them are delusive. This I call "Pure Neutralism".

    We have now got our seventeen alternative possible theories about Mind and Matter definitely stated. I propose now to take them in order, to explain more fully what each of them means, and to consider the strong and weak points (if any) in each of them. It may then be possible to make a tentative decision between them.

Discussion of the Seventeen Types of Theory.

    It will save time and simplify the discussion if we begin by eliminating those alternatives which are quite plainly impossible. It is easy to see that any theory which makes mentality a delusive characteristic is self-contradictory. For to say that mentality is a delusive characteristic is to say that it in fact belongs to nothing, but that it is misperceived or misjudged to belong to something. But, if there be misperceptions or misjudgments, there are perceptions or judgments; and, if there be perceptions or judgments, there are events to which the characteristic of mentality applies. This enables us at once to eliminate (2, 22) Pure Materialism; (3, 221) and (3, 222) the two forms of Materialistic Neutralism; and (3, 3) Pure Neutralism. We have thus reduced our alternatives to thirteen.

    There are two other types of theory which, I believe, can be positively refuted. These are (2, 112) Reductive Mentalism, and (2, 212) Reductive Materialism. So far as I am aware, Reductive Mentalism has never been held; but Reductive Materialism flourishes to-day under the name of "Behaviourism". I will therefore take the latter theory first, and try to prove that it is absurd.

    Reductive Materialism or "Behaviourism". This theory holds that there really are material objects, and that materiality is a differentiating attribute. And it also holds that the characteristic of being a mind or being a mental process reduces to the fact that a certain kind of body is making certain overt movements or is undergoing certain internal physical changes. Of course many writers who call themselves "Behaviourists" are really Epiphenomenalists, and Epiphenomenalism is an entirely different doctrine; but there is a residue of quite genuine Behaviourists, and it is with their views which we are now concerned.

    Behaviourism in psychology may be compared to mechanism in biology. But there is a very important difference between the problem of life and that of mind, which makes Behaviourism in psychology much less plausible than mechanism in biology. The one and only kind of evidence that we ever have for believing that a thing is alive is that it behaves in certain characteristic ways. E.g., it moves spontaneously, eats, drinks, digests, grows, reproduces, and so on. Now all these are just actions of one body on other bodies. There seems to be no reason whatever to suppose that "being alive" means any more than exhibiting these various forms of bodily behaviour. That is why Substantial Vitalism, which is the biological analogue of Cartesian Dualism in psychology, is a dead issue; and why the whole controversy about life is really between Emergence and Mechanism. But the position about consciousness, certainly seems to be very different. It is perfectly true that an essential part of our evidence for believing that anything but ourselves has a mind and is having such and such experiences is that it performs certain characteristic bodily movements in certain situations. E.g., we observe it avoiding obstacles, repeating some series of movements again and again with suitable variations until a certain end is gained, giving appropriate answers to questions, and so on. When external bodies behave in these ways, we are inclined to associate minds and mental processes with them; and, when they do not, we are inclined to deny these to them. Now, if this were the only evidence that we ever had in any case for the existence of minds and mental processes, it may be admitted that the latter could, at most, be regarded as purely hypothetical causes of certain kinds of bodily behaviour. And it might then be plausible, though it would certainly not be logically necessary, to suggest that "having a mind" simply means "behaving in such and such ways". But it is plain that our observation of the behaviour of external bodies is not our only or our primary ground for asserting the existence of minds and mental processes. And it seems to me equally plain that by "having a mind" we do not mean simply "behaving in such and such ways". These points can be made clear as follows.

    (a) We certainly ascribe mental processes to ourselves as well as to others, and it is perfectly certain that here our ground for saying that we are having such and such an experience is not the fact that we have observed our bodies to be behaving in such and such ways. When I say that I am seeing a chair or hearing a bell I am asserting the occurrence of an experience. Now it is possible that, whenever I have the first kind of experience, my body is behaving in one characteristic way; and that, whenever I have the second kind of experience, my body is behaving in a characteristically different way. But, even if this be in fact true, it is perfectly certain that this is not my ground for saying that I see a chair or hear a bell. I often know without the least doubt that I am having the experience called "seeing a chair" when I am altogether uncertain whether my body is acting in any characteristic way. And again I distinguish with perfect ease between the experience called "seeing a chair" and the experience called "hearing a bell" when I am quite doubtful whether my bodily behaviour, if any, on the two occasions has been alike or different. If then the Behaviourist argues that mental processes, in so far as they differ from bodily behaviour, are purely hypothetical causes of such behaviour; and that we shall keep nearer to the observable facts by dropping these hypothetical entities altogether; the answer is to deny his premise. If we confine ourselves to bodily behaviour it is perfectly certain that we are leaving out something of whose existence we are immediately aware in favourable cases.

    (b) However completely the behaviour of an external body answers to the behaviouristic tests for intelligence, it always remains a perfectly sensible question to ask "Has it really got a mind, or is it merely an automaton" It is quite true that we have no available means of answering such questions conclusively. It is also true that, the more nearly a body answers to the behaviouristic tests for intelligence, the harder it is for us in practice to contemplate the possibility of its having no mind. Still, the question: "Has it a mind?" is never silly in the sense that it is meaningless. At worst it is silly only in the sense that it does not generally express a real doubt, and that we have no means of answering it. It may be like asking whether the moon may not be made of green cheese; but it is not like asking whether a rich man may have no wealth. Now, on the behaviouristic theory, to have a mind just means to behave in certain ways; and to ask whether a thing which admittedly does behave in these ways has a mind would be like asking whether Jones, who is admittedly a rich man, has much wealth. Since the question can be raised, and is evidently not tautologous or self-contradictory, it is clear that when we ascribe a mind or a mental process to an external body we do not mean simply that it behaves in certain characteristic ways. If the Behaviourist answers that, whatever we do mean, this is all that we ought to mean, I have two comments to make.

1. We have a right to mean more, because we know that in our own case there is more.
2. I would invite the Behaviourist to explain how, on his own theory, we can ever have come to make the mistake which he says that we do make.

If in fact we can observe nothing but bodily behaviour in ourselves and others, how did we ever come to entertain the hypothesis that there is something more than these; and how did we come to suppose that there are better grounds for assuming the presence of this extra factor in some cases than in others? On the ordinary view this fact is easily explicable. We know that there is something more than bodily behaviour in our own case, because we can directly observe it. We find that certain kinds of experience in ourselves are accompanied by certain types of bodily behaviour. If we find external bodies which resemble our own behaving in the way in which ours behave when we have a certain kind of experience, we assume that there is a similar kind of experience associated with these bodies. And we feel more confidence in this conclusion the more closely the external body and its behaviour resemble our own. Of course such an inference may be wrong in any particular case; it is even possible theoretically that it is wrong in all cases. But it is at least intelligible, on the ordinary view, how we come to make this hypothesis, and why we feel more certain of it in some cases than in others. All this would be completely inexplicable, it seems to me, if Behaviourism were the whole truth. If Behaviourism be true we all make a mistake which it would be impossible for us even to think of unless Behaviourism were false.

    I propose now to go rather more into detail about the behaviouristic analysis of certain special kinds of mental process. But, before I do this, I must clear up a certain ambiguity about the meaning of "bodily behaviour". When the Behaviourist says that all mental processes reduce without residue to the fact that the body is behaving in a certain specific way he to does not mean to confine himself to gross overt actions, like shrieking or kicking. His attempts to reduce all mental processes to bodily behaviour would have no plausibility at all if he were restricted to this narrow sense of "behaviour". He always includes also at least such bodily movements as changes of blood pressure, incipient movements in the tongue and throat, convergence and accommodation of the eyes, and so on. This, I think, is quite legitimate; for there is no essential difference between movements which are difficult to observe simply because they go on inside the body and those which are overt and easily observable without special instruments of precision. I will lump together all such changes under the name of "molar behaviour", as contrasted with "molecular behaviour"; and I will call a Behaviourist who thinks that all mental processes can be reduced without residue to molar behaviour a "molar Behaviourist".

    But it is very difficult to get the Behaviourist to stop at this point. When overt behaviour, supplemented by changes of blood-pressure, incipient movements in the throat, etc., seems inadequate to make the behaviouristic analysis of some mental process seem plausible, the Behaviourist is very liable to appeal to hypothetical molecular movements in the brain and nervous system. If you say to him that two obviously different mental processes, A and B are accompanied by indistinguishable molar behaviour, or that qualitatively indistinguishable mental processes are accompanied on different occasions by obviously different kinds of molar behaviour, he is liable to say: "Well, at any rate, the correlated molecular changes in the brain and nervous system must have been different in the one case and exactly alike in the other." As no one knows anything about these, no one can deny that this may be true. The Behaviourist then proceeds to identify the mental process with the supposed molecular changes. This I will call "molecular Behaviourism". In this form of course there is nothing new about Behaviourism; it is just old-fashioned materialism which has crossed the Atlantic under an alias. It is true that all Behaviourism is a form of Reductive Materialism; but it does at least claim to be a new form. And this claim can be upheld only if it be interpreted to mean molar Behaviourism. Of course, what happens is that a man starts as a Molar Behaviourist and is then pushed back by criticism into Molecular Behaviourism, at which stage his theory has lost most of its interest.

    I am now going to consider the behaviouristic account of perception, because perception is the mental process to which Behaviourism can most plausibly be applied. If it fails to give an adequate account of perception, it is incredible that it should give an adequate account of memory, imagination, or abstract thinking. I will first explain why I hold this. If a certain kind of mental process is to be reduced to a certain kind of bodily behaviour it is evidently a necessary, though by no means a sufficient, condition that there shall be a one to one correlation between the two. That is, it must be quite certain that this mental process never happens without this bodily behaviour, and that this bodily behaviour never happens without this mental process. If either ever happens without the other it is certain that the two cannot be identical; though of course they might not be identical even if one never did happen without the other. There is a one to one correlation between the events in the life of Augustus and the events in the life of the second Roman emperor, because the name "Augustus" and the phrase "the second Roman emperor" denote the same person. There is also a one to one correlation between the movements of the needles of two connected telegraphic instruments. But these are nevertheless two different sets of movements, and not a single set with two different names. The Behaviourist has to show that mental and bodily events are connected as incidents in the life of Augustus and incidents in the life of the second Roman emperor are connected. He will have done nothing relevant if he shows only that they are connected as the movements of the two telegraph-needles are connected.

    Now I do not see the least reason to believe that there is any kind of molar behaviour which always goes on when I am thinking and never goes on at any other time; and the same remarks apply to remembering and imagining. I see still less reason to believe that there is one kind of molar behaviour which always happens and only happens when I am thinking of Cleopatra's Needle, and another kind which always happens and only happens when I am thinking of the Binomial Theorem. If you say that there may be molecular differences in these cases I cheerfully admit it, since neither you nor I can possibly know anything about the matter. It seems to me then that the irreducible minimum of conditions necessary for applying Behaviourism to thinking, remembering, and imagining, are plainly lacking in the present state of our knowledge. On the other hand, it is more or less plausible to hold that there is a certain type of bodily behaviour which always happens and only happens when I am perceiving something. And it is more or less plausible to hold that this behaviour differs in a characteristic way according to whether I am perceiving A or B. What I propose to do is to show (a) that there are cases where we should be said to be perceiving a certain thing, and where it is by no means certain that there is any characteristic molar behaviour which might not have taken place when we were not perceiving this thing; and (b) that, even where there is a one to one correlation between my perception of a certain object and certain molar behaviour, there is something more involved in the perception than the mere occurrence of this molar behaviour.

    (a) Whenever we perceive we perceive some definite object, e.g., a chair on one occasion and a cat on another occasion. A complete account of the act of perceiving in behaviouristic terms must therefore do two things.

1. It must mention some special kind of bodily behaviour which is always present when we are perceiving and is never present when we are merely imagining or thinking or remembering. And
2. this kind of behaviour must be such that a meaning can be given to the statement that one bit of behaviour of the perceptual kind refers specially to a certain chair and that another bit of behaviour of the perceptual kind refers specially to a certain cat.

    If I am already moving about I shall no doubt as a rule avoid those obstacles which I perceive, and I shall stumble into objects which lie in my way and which I do not perceive. In such cases there is no doubt certain bodily behaviour which has a specific relation to those objects which I am said to perceive. And there is no bodily behaviour having this specific relation to objects which I am said not to perceive. It is in such cases as these that the behaviouristic analysis of perception has most plausibility. For here we may admit that at any rate the irreducible minimum of conditions for the possibility of such an analysis seems to be fulfilled. But even these examples are subject to the following criticism. We are not dealing here merely with the perception of an object, but with the perception of an object which is an obstacle to some already intended and initiated course of movement. Surely, as I walk about a room (and, still more so, as I stand still), there are many objects which I perceive and in reference to which there is no specific molar bodily behaviour of the kind mentioned. There are plenty of objects which I neither avoid nor stumble over, because they are not in my way or because I am not moving about. And yet I may be perceiving them.

    Perhaps the most plausible kind of bodily behaviour to take as present in all cases of perception, and as specially correlated with the object which is said to be perceived, would be the convergence and accommodation of the eyes on to the place where the object is. But, in the first place, this applies only to visual perception. Secondly, since it is admitted that I can see two things in different places at the same time, though not perhaps with equal clearness, it is evident that my perception of one at least of them cannot be accompanied by the convergence of my eyes on to the place which it occupies. Lastly, even when we find a man blundering into an obstacle, it is not safe to assume that he did not perceive it unless we know what his intentions and wishes are at the time and that his body is under the control of his will. No doubt it is generally safe to assume that a person does not want to be dashed to pieces. And no doubt it is generally safe to assume that, if a person did not want to fall over a precipice, he could stop himself from doing so when it stares him in the face. Subject to these conditions it would no doubt be reasonable to conclude that a man who walked over a precipice had not perceived it. But it is perfectly notorious that people sometimes do walk over precipices which they perceive, because they want to be dashed to pieces. And it is probable that some people walk over precipices which they perceive, although they do not wish to be dashed to pieces, because the height exercises a fascination over them which paralyses their wills. Such happenings are relatively uncommon because the wish to be dashed to pieces is much rarer than the wish not to be, and because contra-voluntary ideo-motor actions on this scale are very rare indeed. But they do happen. It is thus very doubtful whether we can find any kind of molar bodily behaviour which always takes place when a person would be said to perceive a certain object and which never takes place when a person would be said not to perceive this object. And, unless this can be found, the attempt to reduce perception to some kind of molar bodily behaviour which has some special reference to the perceived object fails in limine.

    (b) I propose now to waive this objection, and to assume for the sake of argument that careful enough investigation of a man's body would disclose some specific kind of molar behaviour which always takes place when he perceives A and never takes place when he does not perceive A. I shall now show that, even if this be so, there is always something involved in the statement that this man perceives A over and above the fact that his body is behaving in this specific way. It is quite certain that, whenever it is true to say that I see something, it is true that I have a sensation of colour; that, whenever it is true that I hear a bell, it is true that I have a sensation of a noise; and so on. In fact every perception involves a sensation as an essential factor, although it involves something else as well. Perception, therefore, cannot be reduced to the fact that my body is behaving in a certain way towards a certain external object unless the sensational element in it can be reduced to bodily behaviour. This quite obvious fact may be illustrated as follows. There is a perfectly specific relation between the movements of a compass-needle and those of a magnet held near it. On a purely behaviouristic analysis of perception there can be no possible reason to doubt that the compass-needle perceives the magnet. Yet, as a matter of fact, nearly every one would deny that the needle perceives the magnet; and the few people who would suggest that it does would admit that this is a paradoxical proposition which needs to be recommended by an elaborate set of arguments such as those used by Spinoza or by Schopenhauer in establishing their peculiar metaphysical systems of hylozoism. This is because we know that a sensational element is an essential factor in what we understand by a "perception", and because we are very doubtful whether there is anything of the kind in the case of the needle and the magnet.

    Now can statements of the form: "I am aware of a red patch" or "I am aware of a tinkling noise" be reduced to statements of the form: "This body, or some part of it, is behaving in such and such a way"? If not, behaviourism has manifestly failed even in the cases which are antecedently most favourable to it.

    Let us suppose, for the sake of argument, that whenever it is true to say that I have a sensation of a red patch it is also true to say that a molecular movement of a certain specific kind is going on in a certain part of my brain. There is one sense in which it is plainly nonsensical to attempt to reduce the one to the other. There is a something which has the characteristic of being my awareness of a red patch. There is a something which has the characteristic of being a molecular movement. It should surely be obvious even to the most "advanced thinker" who ever worked in a psychological laboratory that, whether these "somethings" be the same or different, there are two different characteristics. The alternative is that the two phrases are just two names for a single characteristic, as are the two words "rich" and "wealthy"; and it is surely obvious that they are not. If this be not evident at first sight, it is very easy to make it so by the following considerations. There are some questions which can be raised about the characteristic of being a molecular movement, which it is nonsensical to raise about the characteristic of being an awareness of a red patch; and conversely. About a molecular movement it is perfectly reasonable to raise the question: "Is it swift or slow, straight or circular, and so on?" About the awareness of a red patch it is nonsensical to ask whether it is a swift or a slow awareness, a straight or a circular awareness, and so on. Conversely, it is reasonable to ask about an awareness of a red patch whether it is a clear or a confused awareness; but it is nonsense to ask of a molecular movement whether it is a clear or a confused movement. Thus the attempt to argue that "being a sensation of so and so" and "being a bit of bodily behaviour of such and such a kind" are just two names for the same characteristic is evidently hopeless. And this is what the Behaviourist has really got to do.

    Of course, when a man says that all mental states are reducible to bodily behaviour, he may not mean anything so radical as this. He may admit, e.g., that to be a sensation of red is one characteristic and to be a molecular movement of a certain kind is another characteristic. He may merely wish to deny that the two characteristics belong to different events or substances. He may wish to maintain only that there is one event which has the two characteristics of being an awareness of a red patch and of being a molecular movement of a certain kind; that there are events which have only material characteristics, and none which have only mental characteristics; and that the mental properties of those events which do have mental properties are completely determined by the material properties which these events also have. Such a doctrine, whether true or false, cannot be dismissed at once as plainly absurd. But it is certainly not Behaviourism, and is not a form of Reductive Materialism; it is a form of the theory (2, 211), i.e., of Emergent Materialism.

    It seems to me then that Reductive Materialism in general, and strict Behaviourism in particular, may be rejected. They are instances of the numerous class of theories which are so preposterously silly that only very learned men could have thought of them. I may be accused of breaking a butterfly on a wheel in this discussion of Behaviourism. But it is important to remember that a theory which is in fact absurd may he accepted by the simple-minded because it is put forward in highly technical terms by learned persons who are themselves too confused to know exactly what they mean. When this happens, as it has happened with Behaviourism, the philosopher is not altogether wasting time by analysing the theory and pointing out its implications.

    Reductive Mentalism (2, 112). Reductive mentalism would be the counterpart of Behaviourism. It would consist in holding that the material characteristics of being extended and public, of having position, motion, etc., are reducible to combinations of purely mental characteristics. And there is precisely the same reason to deny this as to deny the opposite doctrine of Reductive Materialism. So far as I know, the present theory has never been held. All mentalists with whose works I am acquainted have held that material characteristics are delusive appearances of certain mental characteristics. This is obvious in the case of Leibniz, Hegel, Ward, Bradley, and M'Taggart. Berkeley's theory, on the face of it, is somewhat different. He holds that sensa really do have some material characteristics. They really are extended, coloured, hot, etc., and they really do move about in sense-fields. But

1. they are also mental events. And
2. they do not have all the characteristics of matter. For they are private, fleeting, and incapable of interacting with each other.
3. The remaining characteristics of matter are ascribed to God's habits of volition by Berkeley. These are permanent, neutral, and capable of causal action. But they are not extended or movable; and they are mental.

Thus, in the end, materiality is a delusive characteristic for Berkeley as for other mentalists. There is nothing which has all the characteristics of materiality; though there are some things which have some of these characteristics, and other things which have the rest of them. For Berkeley materiality is a delusive characteristic, in the sense in which the characteristic of being a mermaid is delusive; i.e., it is a compound characteristic which applies as a whole to nothing, though it can be analysed into factors each of which does apply to something. For M'Taggart or Hegel materiality is delusive in a still more radical sense. It is a compound characteristic some of whose factors apply to nothing. E.g., nothing, on their view, is really extended.

    Are Mentality and Materiality compatible Characteristics? We have now reduced our original seventeen candidates to the more wieldy number of eleven. Can we make a further reduction? The most promising question to raise at this point is the following. Granted that mentality and materiality are distinct and mutually irreducible characteristics, could any substance or event possess both of them? If we could answer this in the negative, if we could show that it is as absurd to suppose that the same event or substance could be both mental and material as to suppose that the same material substance could be at once red and blue all over, we could make a considerable clearance. We should certainly get rid of (1, 1) Dualism of Compatibles; (2, 111) Emergent Mentalism, and (2, 211) Emergent Materialism. The effect on the various Neutralistic Theories would be less marked; it would, I think, still be possible to keep them by a suitable statement. We should simply have to suppose that mental and material characteristics both emerge from certain arrangements of the same Neutral Stuff; but that they never both emerge from the same arrangements of this Neutral Stuff.

    Opposite answers have been given to the question at issue by different philosophers of eminence. Mr Locke saw no reason why God should not have endowed a material substance with the power to think. Descartes and Dr M'Taggart held that it is impossible that the same event or substance should be both mental and material. The argument against the compatibility of the two characteristics is this. Suppose a certain event were both mental and material. Since it is material it must have a certain shape and size. But is it not plainly nonsense to talk of a circular thought of an inch in diameter? Again, suppose that at a certain moment I am wanting my tea and thinking of the square root of minus-one. If this volition and this thought be also material events they must have spatial positions and stand in spatial relations. But is it not plainly nonsense to talk of a volition being two inches to the north-west of a thought?

    I should like to believe that these arguments were conclusive, because it would greatly simplify our problem if they were. But I cannot honestly say that they seem to me to be conclusive. I admit, of course, that such statements as have been made above sound very odd, and that no one ever thinks of making them. But these facts seem capable of explanation. Suppose for the moment that it were true to say of a certain event both that it occupies a circular region in a brain and that it is someone's desire for his tea. It would have to be admitted that the two characteristics of this one event are known in quite different ways. There is one and only one person to whom its mental characteristics are directly manifested, viz., the person whose desire it is. Its material characteristics never manifest themselves to this person; and, strictly speaking, they do not manifest themselves to anyone. For no one perceives the position, shape, and other physical characteristics of this event. If the latter characteristics be known at all, they are known only to physiologists by an elaborate and precarious process of hypothetical and analogical reasoning. Now, since no one is acquainted with the material characteristics of any mental event, even if such events do have material characteristics, and since every one is acquainted with the mental characteristics of some mental events, it is not surprising that it should sound odd to ascribe determinate spatial qualities and relations to those events which are thoughts, volitions, etc. We may admit

1. that the presence of mentality does not entail the presence of materiality in the same object, as, e.g., the presence of colour entails that of extension. So far as one can see, an event could have been mental without being extended or material.
2. That nothing which manifests mentality also manifests materiality.

There is thus no direct empirical evidence that what has mentality ever has also materiality, as there is direct empirical evidence that what is red may also be hot. But the real question is: "Does the presence of mentality in an object entail the absence of materiality from it?" We may admit (a) and (b) without admitting that this question must be answered in the affirmative.

    Now I cannot see by direct inspection that what is material cannot also be mental. Is there any indirect way of proving this proposition? It might be said that there is a fundamental difference between mentality and all those qualities which admittedly can belong to extended objects. Every non-spatial quality which admittedly can belong to an extended object is an extensible quality; i.e., it is such that any object which possesses it must be extended. It is obvious that colour, temperature, etc., are extensible qualities in this sense. Now it is certain that mentality is not an extensible quality, in this sense. For, as we have seen, it is plainly logically possible that an event might have mentality without being extended. Now it might be suggested that it is a self-evident proposition that every non-spatial quality of an extended object must be an extensible quality. If this be accepted, it follows that mentality cannot be a quality of any extended object, and therefore that mentality and materiality are incompatible characteristics. But I do not find the suggested proposition self-evident on careful inspection. I do indeed find it self-evident that every extended object must have some non-spatial extensible quality; but this is quite a different proposition from the suggested principle that every non-spatial quality of any extended object must be an extensible quality. Hence I see no impossibility in the supposition that one and the same thing or event may have both mental and material characteristics. And I can see what causes may have made men think that the two characteristics are incompatible even if they be really compatible.

    It might perhaps be admitted that, so long as we confine ourselves to isolated mental events, we cannot see why they might not also be material. But it might be said that, when we consider that mental events are states of mind whilst material events are states of body, and that the characteristic interrelations of mental events within a mind are utterly unlike the characteristic interrelations of material events within a body, we see that it is necessary to assume two sets of events and not just a single set of events with two different characteristics. This again is not obvious to me. If two events have each two different determinable characteristics A and B in the determinate forms a₁ b₁ and a₂ b₂ respectively, they can obviously stand at the same time in two very different determinable relations to each other, one in virtue of the determinable A and the other in virtue of the determinable B. E.g., two musical notes may be identical in temporal position, whilst one is an octave lower than the other in pitch. It therefore seems perfectly possible that a set of events, each of which had both material and mental characteristics, might form a whole of the material kind in virtue of the relations which depend on the material characteristics of the events, and might also form a whole of the mental kind in virtue of the relations which depend on the mental characteristics of the events. Thus a series of notes is at once a tune, in virtue of the relations of pitch which depend on the auditory characteristics of the notes; and a series of events each of which lasts so long and is separated by such and such a time-gap from its neighbours, in virtue of the temporal characteristics of the notes.

    I should like to point out that the doctrine that every event which is mental is also material would in no way entail the view that all causation is physical causation. To say that the event e₂ is determined by purely physical causation would mean that the necessary and sufficient condition of the occurrence of e₂ is the occurrence of an event e₁ having a certain determinate material characteristic theta₁. This event might also have a determinate mental characteristic psi₁; but, if the causation be purely physical, the possession of psi₁ by e₁ will be causally irrelevant to e₂. Now it is obvious that this is only one of three possible alternatives. In the first place, the possession of theta₁ by e₁ might be necessary but not sufficient to determine e₂. It might be that the possession of psi₁ by e₁ was also necessary, and that the complete cause of e₂ is the occurrence of an event e₁ with the material characteristic theta₁ and the mental characteristic psi₁. Secondly, it might even be the case that the possession by e₁ of theta₁ was causally irrelevant to e₂ and that the occurrence of an event e₁ having the mental characteristic psi₁ was necessary and sufficient to determine e₂. This might be true even if in fact there are no events which have mental characteristics without having material characteristics. We can call these two alternatives which we have just been mentioning "mixed causation" and "purely mental causation" respectively. I have thus shown that the types of theory at present under discussion do not preclude mixed causation and purely mental causation, as might perhaps be thought by some.

    On the whole, then, I can see no conclusive objection to the possibility that one and the same event should have both mental and material characteristics or that one and the same substance should be both a mind and a body. Hence I cannot reject off-hand the three types of theory which imply that this possibility is realised. So we are still left with the eleven alternatives with which we started this subsection.

    Theories which make Materiality delusive. We have seen that no theory which makes mentality delusive can be accepted. We cannot reject on the same grounds theories which make materiality delusive. Nevertheless we might be able to reject such theories on other grounds. Let us consider this question next. We are left with three types of theory which make materiality delusive, viz. (2, 12) Pure Mentalism, and (3, 211) and (3, 212) the two forms of Mentalistic Neutralism.

    Of these theories I believe that Pure Mentalism, both in its less radical Berkeleian form and in the more radical form in which it is held by Leibniz, Hegel and M'Taggart, may be rejected. The theory has a negative and a positive side. The negative side is that materiality is a delusive characteristic. The positive side is that things which have nothing but mental qualities and relations are misperceived to have material qualities and relations. Now I see no reason to believe the negative proposition, and strong reasons to doubt the positive proposition. Materiality is a complex characteristic, which I have analysed in Chapter IV.

1. All the arguments to prove that some of the constituent characteristics of materiality (e.g., extension) are delusive seem to me to be plainly fallacious (like Bradley's) or to depend on premises which I see no ground for accepting (like M'Taggart's).
2. The arguments to prove that, whilst none of the constituent characteristics of materiality are delusive, materiality as a whole is delusive seem to me to prove something important, but not this proposition. I therefore see no ground to believe that materiality is delusive either in the more radical sense of M'Taggart and Leibniz or in the less radical sense of Berkeley.
3. It seems to me most unlikely that things which had nothing analogous to spatial qualities and relations and did not form a quasi-spatial order of at least three dimensions could present that particular system of interconnected appearances which external objects do present.

And I do not see that minds or mental events, connected by the only mental relations with which we are acquainted, would fulfil these conditions. If you say that there may be many mental relations with which we are unacquainted I of course agree. But why call these particular relations "mental" if they do not resemble any mental relation that we are acquainted with, and do resemble spatial relations?

    My detailed reasons for making these assertions will be found in Chapter IV, and especially in the subsection headed: In what Sense can we accept Physical Objects? On the whole then I think we may reject Pure Mentalism, and thus reduce the number of theories which are worth serious consideration to ten.

    Would the arguments which I have used against Pure Mentalism apply to Mentalistic Neutralism? They would apply to any form of Neutralism which refuses to allow quasi-spatial qualities and relations to the neutral stuff itself or to certain combinations of this neutral stuff. But they would not be fatal to such a form of Mentalistic Neutralism as Mr Russell puts forward in his Analysis of Mind. For Mr Russell mentality is not delusive, since it does belong to certain groups of suitably interrelated sensa. And materiality is delusive, in the less radical sense in which it is so for Berkeley, though not in the more radical sense in which it is so for M'Taggart. For Mr Russell's neutral stuff is sensa; and these really are extended and spatially related to each other, though they lack the remaining characteristics which are essential to materiality. These other characteristics really do belong to certain groups of interrelated sensa; but these groups are not literally extended. Hence every characteristic involved in materiality has application, though materiality itself has no application. Mr Russell's theory is therefore a form of Mentalistic Neutralism. And it is a form to which my arguments do not apply, since his neutral stuff really does have spatial qualities and relations. It is therefore possible for Mr Russell to "save the appearances", whilst, as far as I can see, it is not possible for a Pure Mentalist to do so. So we cannot rule out the two forms of Mentalistic Neutralism at this stage, provided that they are suitably stated.

    General Remarks on Neutralistic Theories. It will be noticed that six of the ten theories which we still have on our hands are forms of Neutralism. It will therefore be wise to consider now the general conditions which any neutralistic theory must fulfil, in the hope that these may exclude some of the suggested forms of Neutralism. I understand by Neutralism the doctrine that neither mentality nor materiality is a differentiating attribute, so that the fundamental stuff of which the existent world is made consists of one or more substances which are neither mental nor material This fundamental stuff must have some differentiating attributes, and by hypothesis these are neither mentality nor materiality. Now of course it might be suggested that the differentiating attributes are utterly unknown to us. In that case we could hardly deny the possibility of any of the six remaining forms of Neutralism. But this purely agnostic Neutralism is not worth serious consideration; for it is useless to trouble about a theory which, from the nature of the case, could explain nothing. I shall assume then that the differentiating attributes of the fundamental substance or substances are attributes which we are acquainted with.

    In that case the possible suppositions about the nature of the fundamental stuff are very restricted. We are directly acquainted at most with two kinds of existent, viz., sensa or images and sense-fields, on the one hand, and mental states and minds, on the other. The only empirical attributes with which we can claim to be acquainted are the qualities of each of these two kinds of existent, the relations of sensa or images to each other in sense-fields, the relations of mental states to each other in minds, and the relations of mental states to sensa or images. In addition to these empirical characteristics, which we become acquainted with by abstraction from instances which manifest them to us, there are, I think, certain categorial characteristics, such as the relation of substance and state, cause and effect, etc. We become acquainted with these on reflection when suitable material is presented to us by our senses or by introspection; but we do not reach our knowledge of them simply by abstraction from instances which manifest them to us. Now these empirical and categorial characteristics are the only materials that any human being has or can have for constructing a theory of the Universe. It follows that the Neutralist who is not content to be merely agnostic about his neutral stuff can ascribe to it nothing but a selection from these characteristics. The neutral stuff must be supposed either

1. to have some of the factors included in materiality and none of those included in mentality; or
2. to have some of the factors included in mentality and none of those included in materiality; or
3. to combine some of the factors of mentality with some of the factors of materiality.

Let us now consider the effects of these three hypotheses on the six forms of Neutralism which still remain.

    Before we can do this a little preliminary explanation is needed. I have admitted that materiality is a complex characteristic. The fundamental factor involved in it is extension. This, if I am right, carries with it some extensible quality, but not any particular extensible quality. The other characteristics are publicity, persistence, and existential independence of any observing mind. It is therefore easy to understand what is meant by the supposition that the fundamental stuff has some but not all the characteristics involved in materiality. It presumably means that extension, at any rate, is ascribed to it, and existential independence of any observing mind. But the corresponding supposition about mentality needs some further explanation.

    So far I have neither asserted nor denied that mentality is a complex characteristic. Some people would no doubt hold that it is simple. If this be so, the hypotheses (b) and (c) are ruled out, and the Neutralist is left with hypothesis (a). But I think that it is arguable that mentality is a complex characteristic, and that it may be analysed somewhat as follows.
(i) The irreducible minimum involved in mentality would seem to be the fact which we express by the phrase "feeling somehow", e.g., feeling cross or tired or hungry. It seems to me to be logically possible that this characteristic, which we might call "sentience", could belong to a thing or event which had no other mental characteristic. But this possibility depends partly on the view that we take about the proper analysis of "feeling somehow"; and I can discuss the question better when I have mentioned the other factors involved in mentality.

    (ii) There is plainly a difference between the fact that something exists and has such and such qualities and relations and the fact that something manifests its existence and manifests certain qualities and relations. Now the converse of manifestation is acquaintance, such as we have in sensing and in imaging. I think that some people would claim to reduce sentience to acquaintance with certain peculiar existents and their qualities. It might be held, e.g., that there are certain peculiar qualities, called is "tiredness", "hungriness", "crossness", etc., and that these qualities characterise certain things from time to time. When this happens something is tired or hungry or cross. Now a thing which is tired or hungry or cross may manifest itself to itself, or to something else which is uniquely connected with itself, as having these qualities. We then say that "tiredness is felt" or that "there is a feeling of tiredness". I introduce the two alternatives of "manifesting itself to itself" and "manifesting itself to something which is uniquely connected with itself", because there seem to be two alternative forms of the theory under discussion.
(alpha) It might be held that tiredness, crossness, etc., are qualities which belong to minds, and that to feel tired is to be acquainted with one's own mind as having the quality of tiredness at the moment. Or
(beta) it might be held (and it apparently is held by Professor Laird and Professor Alexander) that tiredness, crossness, etc., are qualities which belong, not to minds, but to living organisms. In that case to feel tired is to be acquainted with one's own organism or with some part of it as having the quality of tiredness at the moment. Of course a complete account of all feelings might need to combine both these alternatives.

    If we accepted either of these alternatives we could take acquaintance as the fundamental characteristic involved in mentality. My tentative statement that there might be things which had sentience and no other mental characteristic would then have to be modified as follows. We should have to say that there might be things which had no mental characteristic except acquaintance with themselves or their organisms as having certain peculiar qualities, such as tiredness, crossness, etc.

    (iii) Whether sentience be a mental characteristic distinct from acquaintance, or whether it just be acquaintance with certain special objects as having certain special qualities, it is plain that mentality as we know it in ourselves involves a further characteristic. This may be called "referential cognition". We believe in the existence of things and events which we are not at the moment acquainted with, and we believe them to have certain qualities and relations which they are not manifesting to us at the moment. I have tried to show that even perception is referential cognition. It seems clear that there could be no referential cognition without acquaintance; and it seems to me logically possible that there might be things which have sentience and acquaintance without referential cognition.

    We must next notice that we find in ourselves two different kinds of referential cognition, which I will call "intuitive" and "discursive". When I perceive a chair or a pink rat, my perception is an instance of intuitive referential cognition; when I merely think of a chair or a pink rat, my thought is an instance of discursive referential cognition. I can have both intuitive and discursive referential cognition of certain objects; but there are many objects, such as Julius Caesar or a Hydrogen atom, of which I can have only discursive cognition. Now I think it is impossible for there to be referential cognition of the discursive kind in a being which has not referential cognition of the intuitive kind; but it seems logically possible that there should be things which have intuitive referential cognition without discursive referential cognition.

    (iv) Finally, we find in ourselves what may be called "affective attitudes". Conations and the various kinds of emotions are examples of these. An affective attitude consists in "feeling somehow towards something". Now, if I am to be able to take up an affective attitude towards something, this something must fulfil one of the following conditions.

• (alpha) it may be a feeling which is felt by me. E.g., I may dislike my present feeling of hunger.
• (beta) it may be something which I am acquainted with, and which is not a feeling of mine. E.g., I may be pleased with the brightly coloured visual sensa which I am acquainted with when I see a firework display.
• (gamma) It may be the epistemological object of an intuitive referential situation of which I am subject. E.g., a drunkard may be frightened at the pink rats which he sees. And
• (delta) it may be the epistemological object of a discursive referential situation of which I am subject. E.g., I may desire the dinner which I am now thinking of.

It is therefore logically impossible that any affective attitude should exist in a thing that did not possess any other mental characteristic; but it does seem possible that any or all the three mental characteristics previously enumerated should exist in a thing which did not have any affective attitudes whatever. Again, there are certain affective attitudes, such as volition, which could occur only in a being which has discursive cognition; there are others, such as anger, which presuppose at least intuitive, but not necessarily discursive referential cognition; and there are others, such as mere liking and disliking, which presuppose no more than acquaintance or mere sentience.

    The upshot of this discussion is as follows. If "mentality" means the peculiar characteristic of human minds, we must admit that it is complex. Its factors may be divided first into Affective Attitudes and Other Factors. The relation between them is that it is logically possible for the Other Factors to occur without any of the Affective Attitudes, whilst it is not logically possible for any of the Affective Attitudes to occur without at least one of the Other Factors. Secondly, we can arrange the Other Factors in a hierarchical order, such that the earlier could occur without the later but the later could not occur without all the earlier ones. This order is Sentience, Acquaintance, Intuitive Referential Cognition, and Discursive Referential Cognition. We can now see exactly what would be meant by ascribing some but not all of the factors of mentality to the supposed neutral stuff. In the first place, it would mean that the neutral stuff was supposed to have the earlier but not the later factors (of this hierarchy). And, secondly, the Neutral Monist might ascribe or refuse to ascribe affective attitudes to his neutral stuff. (There are some affective attitudes, such as volition, which he must refuse to ascribe to it if he refuses to ascribe the higher members of the hierarchy of Other Factors to it.)

    We are now in a position to consider the three hypotheses, and to note the effects of each on the six remaining forms of Neutralism. I will first try to show that no form of Neutralism which makes mentality a reducible characteristic is compatible with any of the three hypotheses. It is immediately obvious that no such form of Neutralism is compatible with hypothesis (a), which ascribes to the neutral stuff no characteristics except some of the factors of materiality. For we have already argued that mentality cannot be reduced to materiality as a whole; and, if this be so, a fortiori, it cannot be reduced to a part of materiality.

    It is easy to show that such forms of Neutralism are also incompatible with hypothesis (b), which ascribes to the neutral stuff no characteristics except some of the factors of mentality. If mentality is to be a reducible characteristic on this hypothesis we must suppose that the higher terms in the hierarchy of mental factors can be reduced to the lower terms of this hierarchy. Now it seems to me that this is plainly impossible when we clearly understand what is required. Let us take an example. There is a certain event which has the characteristic of being a perception of a pink rat. Let us make the most favourable assumption possible for the reductive type of theory. Let us suppose that this perception has no existent constituents except events which are feelings and events which are acquaintances with sensa and images. We are to suppose then that this perception consists of such events interrelated in certain characteristic ways, and of nothing else. It seems to me that it would still be impossible to deduce from the fact that it has this structure and is composed of these constituents, and from laws which are entirely about feelings and sensations, that this event will be the perception of an epistemological object and that this epistemological object will be a pink rat. Unless we had actually met with events which were perceptions and had epistemological objects I do not see that we could possibly have suspected that a whole composed of feelings and sensations interrelated in certain ways would have the property of being the perception of a certain epistemological object. Thus the characteristic of being a perception is not a reducible characteristic, like the behaviour of a clock, but is at best an emergent characteristic, like the behaviour of silver-chloride. Similarly, I do not see the least reason to believe that the characteristic of being a discursive cognition could be reduced to characteristics which come lower in the hierarchy of mental factors. If such reductions can be effected it is quite certain that no one has made even a plausible beginning of performing the reduction. I conclude then that all forms of Neutralism which make mentality a reducible characteristic are incompatible with hypothesis (b) as well as with hypothesis (a).

    What about hypothesis (c), which ascribes to the neutral stuff some of the factors of mentality and some of the factors of materiality? The forms of Neutralism at present under discussion could be consistent with (c) only on the supposition that the higher factors of mentality, though not reducible to the lower factors alone, are reducible to these eked out with some of the factors of materiality. And I cannot see the least reason to believe that the addition of these factors of materiality would help the proposed reduction. I am therefore inclined to reject all forms of Neutralism which make mentality a reducible characteristic, on the ground that they are inconsistent with all the intelligible hypotheses that we can make about the neutral stuff. We thus get rid of (3, 13) Reductive Neutralism, (3, 122) the second form of Mixed Neutralism, and (3, 212) the second form of Mentalistic Neutralism.

    We are thus left with three forms of Neutralism, viz., (3, 11) Emergent Neutralism, (3, 121) the first form of Mixed Neutralism, and (3, 211) the first form of Mentalistic Neutralism. We must now see how the three hypotheses affect these three alternatives. We find that hypothesis (a) is consistent with all of them. The same is true of hypothesis (c). But hypothesis (b) would exclude all but the first of them, as I will now show. In the first place, if no characteristic be ascribed to the neutral stuff except some of the factors of mentality, it is plain that no form of Neutralism which makes materiality a reducible characteristic can be accepted. For we have already argued that materiality cannot be reduced to mentality as a whole; and, if this be so, a fortiori, it cannot be reduced to a part of mentality. This removes (3, 121) the first form of mixed Neutralism. Moreover, this hypothesis is incompatible with any form of Neutralism which makes materiality a delusive characteristic. For we have argued that it is almost incredible that what has nothing but mental qualities and relations should appear to have spatial qualities, motion, and spatial relations. If this be true, it is, a fortiori, incredible that what has only some of the factors of mentality should appear to have spatial qualities, to move, and to stand in spatial relations. This removes (3, 211) the first form of Mentalistic Neutralism. Thus we may finally classify the three surviving forms of Neutralism as follows. (I) Theories compatible with all the alternative hypotheses: -- (3, 11) Emergent Neutralism. (II) Theories compatible with (a) and (c) but not with (b): -- (3, 121) the first form of Mixed Neutralism, and (3, 211) the first form of Mentalistic Neutralism.

    The Seven remaining Types of Theory. We have now reduced our original seventeen types of theory to a modest seven. I shall now take these seven survivors in order, and mention what seem to me to be the strong and weak points to each.

    (1) We will begin with the two forms of Dualism. If I am right in holding that materiality and mentality are both complex characteristics analysable into several factors, it cannot strictly be said that either is a differentiating attributes. For it is part of the definition of a differentiating attribute that it shall be simple and unanalysable. What then must we understand the two types of Dualism to mean when they say that mentality and materiality are "differentiating attributes"? I think it is fairly easy to see what they mean. Dualism means to assert of materiality the following propositions.

1. There is something which possesses all the factors of materiality, so that materiality is not in any sense a delusive characteristic.
2. None of the factors of materiality are reducible to or emergent from the other factors of materiality, or mentality, or a combination of both.

This is of course quite consistent with the belief that there may also be some things which have some of the factors of materiality and not others. It merely insists that, if there be such things, the remaining factors of materiality are neither reducible nor emergent qualities of certain complex wholes composed of these things. E.g., sensa would have some but not all the factors of materiality, and Dualism is not compelled to deny the existence of sensa. But it is compelled to assert that there are also material things and that the characteristics which these have and sensa lack are not emergent or reducible characteristics of certain groups of interrelated sensa.

    I take it that Dualism asserts a similar pair of propositions about mentality. It asserts

1. that there is something which possesses all the factors of mentality. And
2. that the higher factors of mentality are neither reducible to nor emergent from the lower factors of mentality, or materiality, or a combination of both.

This is quite consistent with admitting that there may be things that have only some of the lower factors of mentality. E.g., the minds of oysters might have nothing but sentience. But Dualism is compelled to assert that there are some things which have all the factors of mentality. And it is compelled to assert that the characteristics which human minds have and the minds of oysters lack are not emergent or reducible characteristics of certain groups of interrelated things which have nothing but sentience.

    We now understand what is asserted in common by theories which make mentality and materiality both "differentiating attributes". It is evident that, if we accept Dualism, five alternative views are possible about the relations between mentality and materiality.

1. They may be incompatible with each other.
2. The possession of materiality may entail that of mentality, but not conversely.
3. The possession of mentality may entail that of materiality, but not conversely.
4. The possession of either may entail that of the other. And
5. the possession of one may entail neither the possession nor the absence of the other.

Now I have argued that the theory that the two are incompatible, though plausible at first sight, is really quite groundless. I therefore reject (1, 2) the Dualism of Incompatibles, which was Descartes' theory. We can therefore confine ourselves to (1, 1) the Dualism of Compatibles, which includes the remaining four alternatives. I see no reason for, and strong empirical reasons against, both (b) and (d). On the face of it there are plenty of things which are material and have none of the factors mentality. I do not know that (b) has ever been maintained. But I understand Spinoza to have asserted (d); if so, he appears to me to have produced no reasons good or bad for his belief. We are thus left with (c) and (e). Let us begin with (c).

    It is certain that we have no empirical evidence in normal experience for the existence of anything which possesses any factor of mentality without being also material. Thus the empirical facts are all in accordance with the view that the possession of mentality entails that of materiality. But, in the first place, they do not require this view. It might be that mentality and materiality are logically indifferent to each other, and yet that, in the actual world or in the only part of it that comes under our observation in this life, it is a fact that mentality is always accompanied by materiality whilst materiality is sometimes unaccompanied by mentality. In the second place, the empirical facts would suggest causal rather than logical dependence of mentality on materiality. For the only things which exhibit signs of mentality are organisms, i.e., certain very special and complex material structures. Now this would make mentality, not a "differentiating attribute", even in the wider sense in which we are at present using the phrase, but an emergent characteristic. Finally, when I reflect on mentality and materiality as carefully as I can, I cannot see that the presence of the former logically entails that of the latter, I do not find the least difficulty in conceiving of a being which had all the factors of mentality and none of the factors of materiality. Some philosophers (perhaps St Thomas) seem to have taken an intermediate view. They seem to have held that some of the lower factors of mentality (e.g., sentience) entail materiality, whilst some of the higher factors (e.g., discursive cognition) do not entail it. And some philosophers seem to have gone further, and to have maintained that the higher factors of mentality exclude materiality. I see no ground for holding either alternative. I cannot see any a priori reason why an immaterial being should not have sentience and sensation; or why a material being should not have discursive cognition.

    We are thus left with alternative (e) viz., that mentality and materiality are logically indifferent to each other. This supposition divides into five factual alternatives analogous to the five logical alternatives which we have already discussed.

• (alpha) Nothing in fact has both mentality and materiality.
• (beta) Everything which has materiality has in fact mentality, but there are some things which have mentality without materiality.
• (gamma) Everything which has mentality has in fact also materiality, but there are some things which have materiality without mentality.
• (delta) Everything that has mentality has in fact materiality, and everything that has materiality has in fact mentality. And
• (epsilon) some things have mentality without materiality, some things have both characteristics, and some things have materiality without mentality.

Now we must admit that all direct empirical evidence of the normal kind favours (gamma). We are not acquainted with anything that is certainly mental and certainly immaterial. And we are acquainted with many things which seem to be material and do not show the least sign of being mental. Of course it remains possible that there may be things which are mental and immaterial (e.g., angels), but that they cannot or do not manifest themselves to us. But, in that case, there can be no direct empirical evidence for their existence. Again, it remains possible that those minds which do manifest themselves to us are immaterial, but that they stand in specially intimate relations with organisms and have to manifest themselves by means of these organisms. But, unless we hold that it is logically impossible for what is mental to be also material, there seems to be nothing in the normal phenomena to suggest this. And I have argued in Chapter XII (i) that there are serious difficulties in squaring this Instrumental Theory of the relation of body and mind with the known facts, and (ii) that the abnormal facts dealt with by Psychical Research do not on the whole support it and do in certain respects conflict with it. The existence of immaterial mental substances therefore remains a mere possibility for which there appears to be no evidence whatever, normal or abnormal, a priori or empirical.

    Again, there is plainly no direct empirical evidence for the view that everything which is material is in fact also mental. If we accept it at all we must accept it as a hypothesis which goes beyond and appears to conflict with the observable facts. We must therefore ask (i) whether this hypothesis has any appreciable antecedent probability, and (ii) whether it explains anything that could not have been explained equally well without it. (i) I do not think that anyone has had the hardihood to ascribe all the factors of mentality to every material substance. At most they have regarded some of the lower mental factors, such as sentience, as differentiating attributes which belong to all bits of matter. But even this seems rash to the last degree. It is no doubt true that the evidence for sentience fades gradually away as we go lower in the scale of organisms. It is therefore quite possible that sentience extends below the point at which direct evidence for it ceases. But everything which we have the least ground for believing to be sentient is a living organism; i.e., a highly complicated material structure consisting of millions of molecules. Moreover, every organism is composed of a comparatively few chemical elements, viz., Carbon, Oxygen, Hydrogen, Nitrogen, Sulphur and Phosphorus. It is an enormous extrapolation to ascribe any kind of sentience to inorganic matter which does not consist of these elements. Thus the antecedent probability of the hypothesis of universal sentience seems to me to be vanishingly small.

    (ii) I think that the only empirical ground that anyone has ever had for ascribing some of the factors of mentality to all material objects is indirect. It looks as if things which obviously have mentality had developed in the course of time from things which obviously had materiality and seem not to have had any factor of mentality. And it is thought that this alleged fact could be more easily explained if we assumed that nonliving matter really has some of the factors of mentality in spite of all appearances to the contrary. We can deal with this contention very shortly. Any theory which regards mentality as a "differentiating attribute", even in the looser sense defined in this subsection, ipso facto, renounces all hope of "explaining" its occurrence. For, by definition, it holds that none of the higher factors of mentality are reducible to or emergent from the lower factors of mentality alone or in combination with materiality. Hence it is not of the least advantage to a Dualistic Theory to ascribe some of the factors of mentality to all matter in the hope that it will thereby explain the occurrence of the complete characteristic of mentality in certain material objects.

    I therefore conclude that the only form of Dualism for which there is the least evidence, either direct or indirect, is (gamma) which asserts that everything which has mentality has also materiality, whilst some things which have materiality have no factor of mentality. Is there any good reason to accept Dualism in this form? I do not think that there is. All the empirical facts which make it unreasonable for the Dualist to accept any but this particular form of Dualism make it unreasonable for him to accept Dualism at all. We find that nothing exhibits mentality except living organisms, and that all living organisms of a certain kind and degree of complexity do exhibit a certain number of the factors of mentality. As we pass to more and more complex organisms we find that higher factors of mentality are exhibited by them. This suggests most forcibly that all factors of mentality are emergent or reducible; or, if not, that only the lowest factor of mentality is a differentiating attribute, whilst the higher factors are reducible to or emergent from it alone or in combination with materiality. The normal facts seem to me to be altogether against Dualism, and the abnormal facts seem to me not to require a Dualism of mentality and materiality. We may therefore pass on to the five types of theory which still remain.

    (2) We can at once dismiss (2, 111) Emergent Mentalism. For this makes mentality a differentiating attribute; and I have just argued that this is most improbable, even when it is interpreted in the looser sense in which we had to interpret it in order to make it possible at all. Hence the only type of theory which remains for discussion under (2) is (2, 211) Emergent Materialism. This asserts

1. that materiality is a differentiating attribute, and
2. that mentality is an emergent characteristic.

    I have already explained what a person must be understood to mean by calling materiality a "differentiating attribute" in spite of its being a complex characteristic. He must mean that all the factors of materiality are differentiating attributes, and that there are in fact some things which have them all. Now I have argued in Chapter IV that this is possible and even highly probable. I have therefore no objection to the first part of the Theory of Emergent Materialism. The second part needs a little further explanatlon. It might be held in a more or a less radical form. The more radical form would assert that even the lowest factor of mentality, viz., sentience, is an emergent characteristic of certain kinds of material complex. The higher factors of mentality would of course be emergent characteristics of complexes composed of complexes which have the lower factors. The less radical form would make sentience a differentiating attribute, which in fact belongs only to material things and perhaps only to some of them. The higher factors of mentality would, as before, be emergent characteristics of complexes composed of complexes which have the lower factors.

    I see no reason to prefer the less radical form of Emergent Materialism to the more radical form. If the higher factors of mentality could be reduced to the lowest factor there might be something to be said for making sentience a differentiating attribute which belongs to some or all material things and to nothing else. But actually no factor in the mental hierarchy can be reduced to any or all of the lower factors, whether taken by themselves or combined with material characteristics. And, since we therefore must postulate emergence at every stage but the first, there seems to be no advantage in refusing to postulate it at the first stage too.

    If there were no facts to be considered except the normal ones, and we rejected all the alleged abnormal facts dealt with by Psychical Research, I should regard Emergent Materialism as on the whole the most reasonable view to take of the status and relations of matter and mind in Nature. The only question would be whether one of the forms of Neutralism might not be preferable. I shall ignore the alleged abnormal facts for the present, and shall discuss the three forms of Neutralism which still remain.

    (3) Of the three remaining forms of Neutralism the only one against which I have a positive objection to offer is (3, 121) the first form of Mixed Neutralism. This type of theory makes materiality a reducible characteristic. We have agreed that it could not be reduced to any or all the factors of mentality. Hence the only question that remains is whether there are some factors of materiality to which the remaining factors of materiality could be reduced. If this were so it would be possible to ascribe only the fundamental factors of materiality to the neutral stuff; and to hold that materiality as a whole is reducible, in the sense that its remaining factors could be proved to belong to certain complex wholes composed of elements which have only these fundamental factors. So far as I know, this cannot be done. At any rate no one has done it, or, to the best of my knowledge attempted to do it. I shall therefore reject (3, 121) the first form of Mixed Neutralism.

    The two remaining forms are (3, 11) Emergent Neutralism, and (3, 211) the first form of Mentalistic Neutralism. Both these types of theory have actually been held by distinguished philosophers. Professor Alexander's theory in Space, Time, and Deity is a form of Emergent Neutralism. Mr Russell's theory in the Analysis of Mind is a form of the first kind of Mentalistic Neutralism. Now Emergent Neutralism takes materiality to be an emergent characteristic, and the first form of Mentalistic Neutralism takes materiality to be a delusive characteristic. Either view appears to me to be possible, but I do not see any good reason to believe either. I have argued that it is possible and even probable that materiality is a "differentiating attribute", in the sense that there are some things which have all the factors of materiality and that none of these factors are either reducible or emergent. There are not the same strong empirical reasons for refusing to take materiality as a differentiating attribute, in this sense, as there are for refusing to take mentality as a differentiating attribute in a similar sense. So far as I can see, the only merit of these two remaining forms of Neutralism is that they introduce rather more unity into the world as a whole than we could admit on other theories. But this seems to me to be a very minor virtue, and quite insufficient to justify any strong preference for these theories.

    It will be noticed that Professor Alexander's theory may be regarded as a more radical form of Emergent Materialism. Emergent Materialism regards materiality as a "differentiating attribute" in the looser sense, and it regards mentality as an emergent characteristic of certain material aggregates. Professor Alexander, so far as I can understand him, ascribes to his neutral stuff only some of the factors of materiality, viz, spatio temporal characteristics, and none of the factors of mentality. Now, if this be the right interpretation, I have a positive objection to Professor Alexander's form of Emergent Neutralism. For it seems to me to be impossible for anything to have only spatio-temporal characteristics; it seems to me that anything that has spatio-temporal characteristics must also have some extensible quality. This, however, is a minor point for the present purpose. If Professor Alexander admitted this objection it would be quite easy for him to ascribe some non-spatio-temporal extensible quality to his neutral stuff and then carry on as before.

    In calling Mr Russell's theory a form of the first kind of Mentalistic Neutralism I am aware that I am being perhaps too charitable. It is indeed quite certain that he regards materiality as a delusive characteristic, in the milder sense in which Berkeley did so. That is, he holds that nothing has all the characteristics of matter, though he admits that each characteristic of matter belongs to something. What I am not certain about is whether he regards mentality as an emergent characteristic or as a reducible characteristic. I am sure that he would like to hold the latter view, because it would shock more intensely more of the people whom he likes to shock. If he really does hold this view his theory is a form of the second kind of Mentalistic Neutralism, and I have rejected this long ago.

    The upshot of the discussion is that, if we confine ourselves to normal phenomena, Emergent Materialism, Emergent Neutralism, and the first form of Mentalistic Neutralism are all possible; and that there is nothing to suggest any theory which would give to Mind a more important and self-subsistent status in Nature. As between these three theories I prefer Emergent Materialism to any form of Emergent Neutralism or Mentalistic Neutralism with which I am acquainted. This is partly because I think it possible and probable that materiality is a differentiating attribute, in the sense there are some things which have all the factors of materiality and that none of these factors are emergent or reducible. And it is partly because I think that there are difficulties in the only two forms of Neutralism with which I am well acquainted, viz., Professor Alexander's and Mr Russell's. But it is quite possible that these difficulties will some day be removed by these learned men themselves, or by successors who will enjoy the advantage of their pioneer work. In the meanwhile, if I were forced to choose between the two, I think I should give a slight preference to Mr Russell's form of Neutralism. If I am not to regard materiality as a "differentiating attribute", in the looser sense, I think it is more profitable to regard it as a delusive characteristic, in the milder sense of Berkeley and Russell, than as an emergent characteristic. The former view seems to me to fit better into the facts of perception than the latter.

    Final Considerations. It only remains to consider what effect the abnormal and supernormal facts, which I have so far ignored in this Chapter, will have on the above tentative conclusions. For this purpose I must refer the reader back to the latter part of Chapter XII. I suggested there that, in the present state of Psychical Research, there is some evidence for persistence after bodily death, but hardly any that justifies a belief in survival. And I suggested that the facts which are at present reasonably well established are best explained by a peculiar form of the view that mentality is an emergent characteristic. This theory I called the "Compound Theory". The essential point of it is that mentality is an emergent characteristic of a compound composed of a living brain and nervous system and of something else which is capable of persisting for some time after the death of the body and of entering into temporary combination with the brain and nervous system of certain peculiarly constituted human beings called "mediums". This something else I called a "Psychic Factor".

    About the nature of this Psychic Factor there is very little that can at present be said with certainty. We can say positively of it that it must be capable of carrying traces of experiences which happened to the mind of which it was formerly a constituent. And negatively we can say that there is at present no reason to believe and strong reason to doubt that it has the higher factors of mentality. It remains possible that it may have some of the lower factors of mentality, such as sentience; but I do not see anything in the facts to require or to suggest this hypothesis. Could it have materiality as a whole, or any of the factors of materiality? It certainly has the factor of persistence; but this is common to all substances and is not peculiar to matter. If we say that it is material we must admit that it is an unusual kind of matter. It is not destroyed by the breaking up of the body with which it was connected; it does not manifest. itself to sense-perception; and it does not produce ordinary physical and chemical effects. On the other hand, there is no reason why it should not be material in spite of these peculiarities. If the Psychic Factor be material, and if it have none of the factors of mentality, the Compound Theory is a form of Emergent Materialism. Thus a slightly modified form of Emergent Materialism is compatible with all the well-established supernormal facts, so far as I can see. A fortiori, these facts will be compatible with slightly modified forms of the two kinds of Neutralism which I have admitted to be possible.

    It is perhaps just worth while to mention that, if the alleged phenomena of materialisation and telekinesis were well established, we might be able to see our way to a more definitely materialistic view of the Psychic Factor. For the so-called "ectoplasm". which is alleged to be involved in these phenomena, would be a peculiar kind of matter associated with the ordinary matter of the living human body. It would then be plausible to suggest that the ectoplasm, which is involved in the physical phenomena of Psychical Research, is identical with the Psychic Factor, which is required to explain some of the mental phenomena of Psychical Research. At present, however, the physical phenomena do not stand on the same evidential level as the mental phenomena. It is much harder to rule out all possibility of fraud in the former than in the latter. And it is harder to make fraud impossible in the case of alleged materialisation than in the case of alleged telekinesis. I therefore do not wish this suggestion to be regarded as any more than a pure speculation.

    The upshot of the discussion is that such abnormal phenomena as are at present reasonably well established do not require us to accept anything more than a slightly modified form of one of the three types of theory which we accepted as possible in the last subsection. They do not give us any reason to ascribe to Mind a more important and self-subsistent status in Nature than we were prepared to give it when we confined ourselves to the normal phenomena. It is of course possible that in future we might get empirical evidence for survival as distinct from mere persistence. Some people, who have much more extensive practical experience of Psychical Research than I have, would say that we have such evidence already; but, as I have tried to show in Chapter XII, this is doubtful. In that case we might be forced to look with more favour on Dualistic theories which make mentality a differentiating attribute. But we must estimate probabilities on the evidence that we have, and not on the evidence which our successors may have. And, on all the evidence which is available to me, which I have tried to state as fairly as I can to the reader in the course of this book, I judge the most likely view to be some form of the Compound Theory which is compatible with Emergent Materialism.

Prospects of Mind in Nature.

    I shall end this book by saying something about the probable prospects of Mind in Nature. This is a question which could not be profitably asked until we had formed some opinion of the status of Mind in Nature. But it does not follow that we can make any very definite answer to it even now.

    The first point that I want to make is that there is nothing in Mentalism as such to justify an optimistic view of the prospects of Mind, and nothing in Materialism as such to justify a pessimistic view of the prospects of Mind. Let us consider, e.g., the Leibnitian form of Mentalism. If this be true there are certainly many minds which are so stupid and confused that aggregates of them appear to us as material objects. And, although we are not so confused as this, we are confused enough to misperceive such aggregates as material objects though they are really mental. Now it is quite consistent with Mentalism as such that all the minds in the Universe should be getting steadily more and more stupid and confused. It is quite possible that the minds which are now so confused that aggregates of them appear to us as material objects were once minds like our own. And it is quite possible that our minds were once clear enough not to misperceive aggregates of very confused minds as material objects. Mentalism seems to be an optimistic theory only because it is confused with Idealism.

    By "Idealism" I understand the doctrine that the nature of the Universe is such that those characteristics which are "highest" and most valuable must either be manifested eternally or must be manifested in greater and greater intensity and in wider and wider extent as time goes on. It so happens that most idealists have been Mentalists; but, as I have just shown, Mentalism is no guarantee of Idealism. Leibniz and Berkeley were both Mentalists; but their optimism was based on the fact that they were Theists, and not on the fact that they were Mentalists. And it is perfectly possible to be a Theist without being a Mentalist, or a Mentalist without being a Theist

    I will now show that, just as Mentalism as such does not entail Idealism, so Idealism is not incompatible with Materialism as such. Suppose that mentality is an emergent characteristic of certain complicated material aggregates. It remains quite possible that the actual configuration of matter in the Universe and the actual laws of matter are such that aggregates of this kind must grow more and more complex, and that a larger and larger proportion of matter must be aggregated in this way as time goes on. In that case we should have an idealistic view of the Universe combined with a Materialistic view of the status and nature of Mind. Materialism is supposed to be incompatible with Idealism only because it happens to be associated with a particular view about the actual laws of matter and the actual configuration of matter in the Universe. And, as we have seen, the former theory about the status of Mind does not entail any particular view about the laws and configurations of Matter.

    It is perhaps worth while to remark before going further that Mentalism would make it probable that pain and frustration are much more widely distributed in the Universe than there is any reason to suppose on a materialistic or dualistic theory. We all treat what we regard as bits of inorganic matter wholly as means. Now, if the Leibnitian form of Mentalism be true, what we take to be lumps of coal and pokers are really colonies of spirits of low intelligence. These spirits must presumably have at least sentience; for this is the lowest factor of mentality. Yet, when I burn a bit of coal in my fire and poke it with my poker, I certainly treat all these confused but sentient spirits merely as means. And it seems quite possible that, under these circumstances, their sentience will be tinged with pain. This pain would be absent if Materialism or Dualism were true.

    I claim now to have shown that there is no special connexion between Mentalism as such and a cheerful view of the prospects of Mind, and no special connexion between Materialism as such and a depressing view of the prospects of Mind. The question that remains is whether we know anything about the laws and configuration of Matter which would throw any light on the probable prospects of Mind, supposing that some form of Emergent Materialism is true. But, before we discuss this final question, there is a more fundamental issue to be faced. Some philosophers, e.g., Mr Bradley and the Dean of St Paul's, have held that the whole notion of perpetual progress is logically impossible. If this be so, it is obviously needless to discuss the question whether it is causally possible. And it would be still more futile to discuss the question whether it is likely or unlikely to be a fact.

    Is Perpetual Progress logically possible? We must begin by drawing some distinctions which are commonly overlooked. (i) I distinguish between perpetual and uniform progress. The latter implies the former, but the former does not imply the latter. To say that s uniformly progresses means that every later state of s is better than every earlier state of s. To say that s perpetually progresses is to assert the following two propositions. (a) if x be any state of s there is a state of s which succeeds x and is better than x itself and all x's predecessors. And (b) if x be any state of s there is no state of s which succeeds x and is worse than x itself and all x's predecessors. This of course leaves it quite open that some of the successors of any state of s are worse than this state or than some of its predecessors. The definition is meant to allow of fluctuations of value, provided that their maxima increase and their minima do not as time goes on. I need hardly say that perpetual progress, thus defined, is perfectly compatible with the view that the value of no state of s will surpass a certain finite magnitude. For, although the successive maxima always increase, they may increase at a diminishing rate as time goes on. It is also consistent with the definition that the successive minima should continually decrease, and that they should approach the same limit as the successive maxima. In that case s, though perpetually progressing, would perpetually approach (though it would never exactly reach) a permanent condition of constant finite value. Now I think that most philosophers who have objected to the notion of perpetual progress have done so because they supposed that this would entail (a) a uniform, and (b) an unlimited increase in the value of the thing which is said to be "perpetually progressing." We now see that this is just a gross mistake, which these persons might have avoided by half an hour's study of any decent introduction to mathematics even if they were not acute enough to avoid it for themselves. I may add that personally I see no logical objection even to the notion of uniform and unlimited progress; but it is not necessary to defend this view, since it is not entailed by the notion of perpetual progress.

    (ii) There is another important point about progress which we have so far ignored. We have spoken as if the whole value of a thing resided in its successive states. And we are even liable to talk as if the whole value of a thing were concentrated in its final state, if it has one. This is of course a mistake. The value of a thing does not reside in any one of its states; nor is the value of a thing the sum of the values of its successive states. It would obviously be absurd to identify the value of the Byzantine Empire with the value of that slice of its history which occupied the last second before Constantinople was taken by the Turks. And it would be equally absurd to identify the value of the Byzantine Empire with the sum of the values of all the successive slices of its history from the foundation of Constantinople by Constantine till its capture by Mahomet II. For it is plain that one important factor in the value of the Empire is the fact that after a time it went from better to worse and not from worse to better. The value of a persistent thing is a quality which inheres in its whole history from its beginning, if it had one, to its end, if it has one. If it be still existing its value is a quality which inheres in its history up to and including its present state. No doubt the value of a thing does depend in some complicated way on the values of its successive states; but it does not depend on this alone. It is plain that the temporal order in which the states happen, and other relations between them, make a great difference to the value of the persistent thing of which they are successive states. It is necessary to modify our definition of "perpetual progress" in the following way to meet these facts. The more accurate definition will run as follows. To say that s perpetually progresses is to assert two propositions; viz., (a) if x be any state of s then there is a state of s which succeeds x and is such that the value of the whole history of s up to and including y is greater than the value of the whole history of s up to and including x. And (b) if x be any state of s then there is no state y of s which succeeds x and is such that the value of the whole history of s up to and including y is less than the value of the whole history of s up to and including x.

    This would of course leave it possible that some later patches of the history of s are worse than some earlier patches of its history. I think it would even leave it possible that some later patches of the history of s are worse than all the earlier patches of its history. Suppose that Ax is the history of s up to and including x. And suppose that xB is a patch of the history of s which immediately succeeds Ax. Suppose further that xB is worse than Ax and than every slice of Ax. It is nevertheless theoretically possible that the whole AB, composed of Ax and xB, should be no worse than Ax. For the value of a whole is not in general the algebraic sum of the values of its parts. E.g., Ax might be an early life of continual sin, and xB might be the appropriate amount of regret in the mind of the sinner for this particular kind and amount of sin. Then I should say that AB is better than Ax. And yet xB might be worse than Ax. The pain without the sin might have a greater disvalue than the sin without the pain. It is certainly arguable that sin without sorrow may be less bad than that state of sorrow for purely imaginary sins which occurs in religious melancholia.

    Now it is sometimes objected by philosophers to the notion of perpetual progress that it places the whole value of a thing in a last state which, by hypothesis, can never be reached. If the notion involved this it would no doubt be self-stultifying. But we have just seen that it involves nothing of the kind. Hence this objection is quite irrelevant. And I think we may safely conclude that there is no purely logical objection to the notion of perpetual progress.

    Is Perpetual Progress causally possible? Granted that perpetual progress is logically possible, it only remains to ask whether it is causally possible on that view of the status of Mind which has seemed most probable in the light of all the facts which are at present available.

    I will first remind the reader that the view about the status of Mind which I have asserted to be most probable an the available evidence is that mentality is an emergent quality of a compound composed of a living brain and nervous system and another constituent which is not always at once destroyed when the brain and nervous system are broken up. As we obviously know extremely little about this other constituent, I will say what I have to say about it at once, and clear it out of the way. It may have some of the lower factors of mentality, though there is no need to suppose that it has. And, whether it has any of the factors of mentality or not, it may be matter of a peculiar kind. All that we positively know about this constituent is that it is capable of carrying traces of past experiences and of certain personal peculiarities. We do not know how persistent it may be, and we do not know what conditions, if any, are capable of destroying it. But we do know that it is not immediately destroyed by those processes which destroy brains and nervous systems.

    It is therefore possible that, even if a cosmic disaster were to destroy all living organisms (and therefore, on our view, all minds) in the Universe, the other constituents of these minds might persist indefinitely. We might imagine them blowing about the Universe for millions of years, like seeds or spores or uncombined chemical elements, waiting for suitable material conditions. Eventually the necessary and sufficient conditions for the existence of living organisms might once more be fulfilled in some part of the Universe, and some of these constituents might unite with them to form those compounds which have mentality as an emergent characteristic. And the new minds, thus formed, might derive certain advantages from the traces left by the experiences of the old minds. The new minds might then develop for millions of years, thus adding fresh traces to the persistent constituents of the wholes of which they are emergent characteristics. Another cosmic disaster might eventually happen, and the process described above might be repeated. Such a series of events as I have been imagining would be quite consistent with perpetual, though not with uniform, progress.

    I have spoken as if these persistent constituents merely waited passively for the development of living organisms, and then combined with them. This may be true. But it is also possible that they play a more active part. It is possible that the development of living organisms out of inorganic matter depends on the agency of such persistent constituents as well as on the fulfilment of certain conditions in ordinary matter. We never find highly developed organisms without minds, any more than we find minds without organisms. It therefore seems not unlikely that the persistent constituents of minds act as cause factors in the original production of living organisms from inorganic matter.

    It remains to consider brains and nervous systems, the other and more familiar constituents of those compounds which have the emergent characteristic of mentality. On our view they are just as necessary for the existence of minds as are the persistent constituents which we have been discussing. Even if the persistent constituents persist indefinitely, a time will come after which there will be no more minds, if the laws and configurations of ordinary matter be such that after a time there can be no more living organisms. And, unless the laws and configurations of ordinary matter allow of perpetual progress in the complexity of living organisms, there is no likelihood of perpetual progress in the mental realm. (I say "no likelihood" and not "no possibility" here, because it must be admitted that a great deal of mental progress can be secured merely by improved social organisation and recording of experiences, without any fundamental change in the individual human organism or its mind.)

    Now, so far as I can see, the situation, according to our present knowledge about matter, is roughly as follows.
(1) There is one and only one alleged law of physics which would seem to make it very probable that, after a certain time, there will be no more living organisms, or only organisms of uniformly decreasing complexity. This is of course the law which asserts a perpetual decrease in the amount of available energy in the Universe. Now there are two remarks to be made I about this alleged law. In the first place, it is a statistical law. All alleged laws of nature are only probable; but this law is peculiar in that it is a statement about probability, which most laws are not. It says that a process by which the available energy of the Universe should be increased is not causally impossible but is extremely unlikely to happen at any given moment. It does not follow that it is extremely unlikely to happen at some moment or other within a sufficiently long period of time. It is extremely unlikely that a person will be involved in a railway accident on any assigned day, but it is by no means unlikely that he will be involved in one some day or other if he travels by train every day for a hundred years. It seems not improbable that the Universe as a whole may pass through successive phases of "running down" and "winding up"; or that, while one part of it is "running down", another part of it may be "winding up". And perhaps the fact that there still is available energy supports the view that such processes of "winding up" do happen from time to time or from place to place. A full discussion of this last point would take us into difficult questions about finite and infinite duration, and about (what is quite a different thing) having a beginning or having no be ginning in time. I do not propose to enter into this subject at the end of a long book which is mainly concerned with quite different problems.

    Secondly, even if the amount of available energy in the Universe continually diminishes, it does not seem to me to be impossible that there should always be organisms and that they should perpetually grow more complex. It is very easy to commit a fallacy here. We can see that, if any existing organism were put into conditions very different from those in which it has lived and in which its immediate ancestors have lived, it would die or degenerate. And we are liable to conclude from this that no organism could live or be highly developed in these supposed conditions. This argument forgets that, if the conditions change slowly enough, the organisms may have time to adapt themselves to the new conditions, and that this adaptation need not take the form of degeneration. It is certain that no existing organism, or only very simple kinds of existing organism, could live if suddenly placed in conditions in which there is much less available energy than there is at present. But it does not follow from this that, if the available energy slowly diminished, the descendants of complex organisms might not be able to live and flourish under the new conditions. It is quite certain that there is no very close correlation between the mere size of the brain and nervous system, or the mere strength of the organism, and high intellectual development. If anyone denied this he might be advised to "go to the ant, to consider her ways, and be wise". For here we have very high intellectual development accompanied by a very small brain and very feeble bodily strength. On the whole then I do not think it is by any means certain that the law of the degradation of energy precludes the possibility of perpetual mental progress, in the sense defined by us.

    (2) Granted that the laws of physics oppose no insuperable bar to the perpetual progress of Mind, it might still be true that the special configurations of matter in the actual Universe do so. About this possibility I have the following remarks to make. We have to take the ultimate laws of matter as unalterable data; but the configurations of matter are certainly to some extent under our control. And they certainly come more and more under our control the better we understand the laws of matter. In particular we are able to make material complexes which have new emergent qualities: this happens whenever we synthesise a new chemical compound. I have already pointed out that we cannot set limits to the automatic power of adaptation possessed by organisms, provided their environment changes slowly enough. I wish now to point out that the existence of minds which understand the laws of matter makes a great difference to the environment itself. There are, e.g., thousands of chemical compounds now existing which probably never existed before in the history of the Universe, and which almost certainly would not have existed if there had not been minds which came to understand the laws of chemical combination and the properties of the chemical elements. Now the greater part of the mind's knowledge and control over inorganic matter is quite new. It dates from the time of Galileo and Newton. Such as it is, it has been gained under the most unfavourable circumstances by the work of a comparatively small group of men, surrounded, influenced, and often opposed by a majority whose minds are warped by the superstitions and heated by the emotions of patriotism and religion. If so much control over inorganic nature has been gained in so short a time and under such unfavourable conditions, it would be rash to set limits to the possible developments of this control in the future.

    The next point to notice is that the human mind has not as yet gained any comparable degree of knowledge of and control over living organisms. When we remember how long it was before it understood the fundamental structure of the inorganic world, and how much this understanding depended on the insight of a few men of genius, like Galileo, Newton, Dalton, and Maxwell, we shall not be surprised at this. And, when we remember how quick and cumulative has been the growth of our control over inorganic matter since Newton's time, we need not despair of a similar growth of control over living organisms. We must remember, however, that, even if we had enough knowledge of biology, physiology, and genetics, to produce healthier and healthier bodies and better and better brains, there might be insuperable psychological difficulties in applying it. For here we come against a solid mass of primitive emotions and superstitions, many of which are crystallised in theological dogmas and supported by the authority of vast ecclesiastical organisations like the Church of Rome. This leads me to the last remark which I wish to make

    The beginnings of a genuine science of organisms exist, and progress in this science might at any moment become rapid. Supposing that Europe does not relapse into barbarism before America has emerged from it, it is quite possible that the next two hundred years may witness as great an advance in our knowledge of living matter as the last two hundred years witnessed in our knowledge of inorganic matter. But, so far as I can see, there are not even the beginnings of a scientific psychology of the individual or of communities. And, unless this defect can be remedied, there seems to be no hope either of devising a stable yet progressive social system or of making the vast alterations in men's minds which would be necessary before they could work such a system and live happily in it.

    Now undoubtedly the greatest immediate threat to the further progress of the human mind is the unequal development of these three branches of knowledge; i.e., the relatively high degree of our control over inorganic nature, combined with our still very rudimentary knowledge of biology and genetics, and with the complete absence of a scientific psychology and sociology. The first and least obvious danger of this state of affairs is that our environment and mode of life are changed deliberately, profoundly, and very quickly by the application of physical and chemical knowledge. The human organism has had no time to adapt itself spontaneously to these changes; for the spontaneous evolutionary adaptation of organisms is an extremely slow process. It therefore seems not unlikely that there is a great and growing disharmony between human organisms and their environment; and that, unless this l can be corrected, the physical and mental qualities of the human race may degenerate. Now it cannot be corrected except by a deliberate modification of human organisms, which shall proceed as fast as the deliberate modification of their environment now proceeds. And this is possible only if we have a scientific knowledge of biology, physiology, and genetics comparable in extent and accuracy to our knowledge of physics and chemistry.

    The more obvious danger of this unequal development of our knowledge lies in the fact that human control over inorganic nature provides men with means of destroying life and property on a vast scale; whilst the present emotional make-up of men, and their extraordinarily crude and inept forms of social organisation, make it only too likely that these means will be used. This danger, so far as I can see, could be averted only by deliberately altering the emotional constitution of mankind, and deliberately constructing more sensible forms of social organisation. And it is quite useless to attempt the latter without the former. In order to do this a vast development of scientific psychology would be needed for two different reasons. In the first place, it would obviously be needed in order to know how to alter the emotional make-up of the individual. But this would not be enough. We might know how to do these things, and yet it might be quite impossible to get people to submit to having these things done to them. For this purpose we should need an enormous development of what Kant calls "the wholesome art of persuasion"; and this could arise only on the basis of a profound theoretical knowledge of the factors which produce, modify, and remove non-rational beliefs.

    Conclusion. The conclusion of the whole matter seems to be that perpetual mental progress is certainly not logically impossible, and certainly not causally inevitable, in the sense of being bound to happen whatever we may do. On the other hand, there seems to be no positive reason to believe that it is causally impossible, in the sense that it is bound not to happen whatever we may do. So far as we are concerned, the possibility depends on our getting an adequate knowledge and control of life and mind before the combination of ignorance on these subjects with knowledge of physics and chemistry wrecks the whole social system. Which of the runners in this very interesting race will win, it is impossible to foretell. But physics and death have a long start over psychology and life.

---

Contents

---