CHAPTER VII

The Mind's Knowledge of Other Minds

Analysis of the Belief in Other Minds Natural Expression Conventional Expression The Logical Status of the Belief in Other Minds The Logical Connexion between Belief in Matter and Belief in Other Minds The Argument for Other Minds on the Assumption of Matter The Argument for Other Minds without the Assumption of the Existence of Matter Summary of Conclusions Do we "perceive" Other Minds?

      The proper analysis of our belief in the existence of other minds, and the question of how it can be justified, have been far less thoroughly discussed by philosophers than the corresponding questions about matter and our alleged knowledge of it. Many philosophers have wanted to deny the reality of material objects, and have felt that it was a feather in their caps when they succeeded in doing so to the satisfaction of themselves and their followers. But, seemingly, no one wants to be a Solipsist; and scarcely anyone has admitted himself to be one. It has been left to rival philosophers to tell him that, on his principles, he ought to be one; and this has generally been regarded as a charge to be repelled and not as a compliment to be thankfully acknowledged. We should be doing too much credit to human consistency if we ascribed this to the fact that all convinced Solipsists have kept silence and refused to waste their words on the empty air. It would seem then that we have a stronger belief in the existence of other minds than in the existence of material things. No one in his senses doubts either proposition in practice; but the philosopher can and does doubt the latter in his study, whilst, even in that chaste seclusion, he seems to be unable or unwilling to doubt the former. I do not think that this difference can be ascribed either to the fact that the evidence for the existence of other minds is more cogent than the evidence for the existence of matter, or to the fact that we have a stronger instinctive belief in the former than in the latter. I think that the real expianation is that certain strong emotions are bound up with the belief in other minds, and that no very strong emotions are bound up with the belief in matter. The position of a philosopher with no one but himself to lecture to, and no hope of an audience, would be so tragic that the human mind naturally shrinks from contemplating such a possibility. It is our business, however, to stifle our emotions for the present, and to follow the argument whithersoever it may lead.

Analysis of the Belief in Other Minds.

      I wish to begin, as usual, with propositions about which every one will agree. Now, I think it would be admitted by every one that the perception of a foreign body of a certain kind, which moves, alters its expression, makes noises, and so on, in certain characteristic ways, is a necessary part of the basis of our belief in the existence and activity of another mind. The only exception to this statement that I can think of is that some few people have claimed under exceptional circumstances to be in direct communion with God or with other spirits without perceiving a characteristic kind of body moving in characteristic ways or making characteristic sounds. But such claims are rare and hard to test. Setting aside such exceptional cases for the present, I think we may say that the above proposition would generally be admitted.

      I want at once to remove two possible misunderstandings.

1. I say only that the perception of such physical object and events is a necessary part of the basis of our belief in other minds. I do not say that it is sufficient. There may be other ingredients which are equally necessary.
2. I say only that the perception of such physical objects and events is a necessary part of the basis of the belief in other minds. I express no opinion at present about the nature of the connexion between this perception and this belief. In particular I must not be understood to be asserting that the perceptual judgment forms a premise from which the belief in other minds is inferred. The sensing and selecting of a certain sensum is a necessary part of the basis of our perceptual belief in the existence of a certain physical object or the happening of a certain physical event. But the two are not connected as premise and conclusion of an argument.

      We must next distinguish between our belief that a certain external body is animated by a mind other than our own, and the belief that at a certain moment a certain mental event which does not belong to my mental history is happening in intimate connexion with a certain external body. A mind, in the sense of an Empirical Self, consists of a oumber of simnitaneous and successive mental events united into a whole of a certain characteristic structure. Hence, whenever we believe that a certain external body is animated by a certain mind, we are no doubt equally justified in believing that there is a series of mental events of some kind intimately connected with this body. But it might well happen that we believed much more strongly that this body is animated by a mind than that a particular mental event of a certain specific kind was going on in this mind at a certain time and expressing itself by a certain panicular perceptible bodily change. I may be practically certain that the body of my friend is animated by a mind; and yet very doubtful, on a certain occasion when I see him frowning, whether he is angry, or thinking deeply, or in pain. Again, it is very far from certain that all mental events must occur as members of those sets of interrelated mental events which we call "Empirical Selves". We might then strongly believe that a certain movement of an external body is the outward expression of a certain mental event which does not belong to our mind; and yet we might be very doubtful whether this external body is animated by a mind at all. E.g., I might feel tolerably certain that, when an insect is injured and writhes about, there is a feeling of pain which is no part of my mind and is intimately connected with these writhing movements of the insect's body. And yet I might be very doubtful whether the insect's body is animated by anything that could reasonably be called a "mind". This feeling might be quite isolated. Or, if it be in fact a member of a group of interrelated simultaneous and successive mental events, this group might be so poor in content and so loose in structure as not to deserve the name of "mind". When we contemplate other human bodies and their behaviour (including in this their speech and writing) we do believe both that they are animated by minds and that certain specific mental events are going on in those minds when the bodies behave in certain specific ways. Most men believe that the bodies of cats and dogs are animated by minds, and also that certain specific events are going on in these minds when cats and dogs behave in certain ways. Even if we doubt this on philosophic reflexion, we find it very difficult not to act as if we believed it. But even here I think we are slightly more certain that there are specific experiences connected with certain specific bodily behaviour than that the body of the animal is animated by a mind. If one sees and hears an animal, such as a dog or a rabbit, with its leg caught in a trap, it is practically impossible at the time to doubt that there is a painful feeling which is being expressed by its struggles and cries. But one would feel a little less certain that the body of a dog or a rabbit is animated by anything that could fairly be called a "mind", unless one were already convinced that all mental events must belong to some mind. When we come to living beings which are very different from ourselves, such as insects, we feel rather doubtful about postulating mental events at all; more doubtful about the precise character of the mental event which accompanies a given movement; and extremely doubtful about the supposition that there is a mind which animates the body of the insect.

      Let us henceforth confine ourselves to the case of human bodies and their perceptible behaviour. I think it will be agreed that, when we see anything which has the characteristic shape, size, appearance, and movements of a human body, we treat it as if it were animated by a mind like our own. And, if it responds to this treatment in the way in which we expect it to do, we have no doubt whatever on this point. If it does not, we have much the same feeling of shock and surprise as we have when we lift something which looks like a heavy weight but is really made of painted cardboard. Our presumption that this body is animated by a mind is, as I have said, based on its general appearance in the first instance; but it is supported or refuted by the order and connexion (or the lack of these) which we afterwards find in its behaviour. A particular, and vitally important, case of this general principle is that of connected rational speech. If something which looks like a human being talks in a connected way, and makes appropriate answers to questions, it is not practically possible to doubt that it is animated by a mind. Since a mind is a whole of suitably interrelated mental events, it is natural enough that the basis for our belief in other minds should be suitably connected series of physical events rather than any particular isolated physical event.

      Besides this general belief that things which look like human bodies and perform certain chains of behaviour are animated by minds, we have more specific beliefs about what is going on in those minds on certain occasions. On observing certain facial expressions, such as frowning, we have a tendency to believe that there is a certain emotion, such as anger, in the mind which animates the body which we are observing. On hearing a human body emit certain sounds we have a tendency to believe that the mind which animates it is suffering a painful sensation, and so on. Here we must draw a sharp distinction between two different ways in which we may come to believe that a certain mind is having a certain experience.

      (i) The body which we are perceiving may make a series of noises or conventional gestures which form a coherent and intelligible "sentence". This sentence we understand, and we then tend to believe that the mind which animates this body is having the experience which we understand that it is describing to us. In that case we assign to the mind, not only the particular experience which it is describing, but also a certain kind of cognitive experience. When a body emits the series of noises: "I have toothache", we commonly believe that the mind which animates it has this peculiar feeling, unless we have some special reason to think that it is lying. But, in addition to this, we always believe that it is making a judgment of some sort. If it is trying to tell the truth, it is making a judgment about its present experience. And, if it is lying, it is still making a judgment about something. We may call this way of arousing belief in the existence of a certain mental event a "conventional expression".

      (ii) We ascribe on certain occasions certain experiences to a foreign mind even when it does not and cannot tell us of them. And, conversely, beings who cannot understand spoken or written words probably ascribe certain experiences to other minds on certain occasions. If I see a baby smiling or hear a dog snarling I ascribe a pleasant feeling to the baby and a feeling of anger to the dog, although they cannot describe their experiences to me by any conventional expression. And, on the other hand, it seems most likely that babies and dogs often know when those who surround them are angry or pleased, although such creatures could not understand what we were saying if we tried to describe our mental states to them. I will call this way of arousing belief in the existence of a certain experience in a foreign mind "natural expression". The experiences which most obviously have characteristic natural expressions are certain emotions, such as anger, and pleasant or painful bodily feelings. But I think that there are other experiences which have characteristic natural expressions. If I am doing something and another human body starts to struggle with me and try to hinder me I can hardly help believing that the mind which animates this body has a volition which is opposed to what I am doing. If, on the other hand, the other human body starts to co-operate with me I can hardly help believing that the mind which animates it has a volition which is in accord with what I am doing. Again, if I see another human body performing a chain of actions which seem well adapted to lead up to a certain end; if I see it avoiding obstacles or trying to modify them; I can hardly help believing that this body is animated by a mind which is not only desiring a certain end but is also thinking about the proper means to gain it. Lastly, whenever a body emits coherent and intelligible sentences it is always naturally expressing the kind of experience called "judgment" or "supposition" or "questioning", no matter what else it may be conventionallyexpressing. If it says "I have toothache", and I believe what the words mean, this set of sounds conventionally expresses to me the presence of a feeling of toothache; but the mere fact that the words form an intelligible sentence naturally expresses to me the presence of an act of judgment. When I say that I have a certain state of mind I deliberately and conventionally express only that state of mind; but I involuntarily and naturally express in addition the state of mind called "judging". It seems clear that natural expression is more primitive and fundamentaJ than conventional expression, so we will begin with it.

Natural Expression.

      The question to be considered in this subsection is: "What is the nature of the connexion between perceiving a certain facial expression or natural gesture in a foreign body and believing that there is a certain mental event going on at the time and expressing itself through this bodily event?"

      In the first place, I must point out that we are probably intellectualising the situation when we talk of a "belief in" or "judgment about" the mental event. As in the case of perception, it would be truer to say that we act as if we believed that such and such an event was happening, and are surprised if the results of our action give us the lie. So we had better talk of a "quasi-belief". A dog acts as it would be reasonable for him to act if he believed his master to be angry when his master shows the external signs of anger; but it is very doubtful whether the dog has the peculiar experience of judging or believing a proposition. And, in most of our intercourse with other human beings, we are in the same position as the dog in the example. The only difference is that we can reflect and afterwards rnake the judgment "in accordance with which" we have been acting; whilst a dog or a baby presumably cannot.

      I can now clear the ground by making certain negative statements.

      (1) It seems to me to be absolutely certain that the belief in other human minds, and the belief that a certain human mind is having a certain experience on a certain occasion, are not reached by inference, even if they can be afterwards justifed up to the hilt by inference. It is perfectly certain that I do not now make an inference when I see my friend frowning and believe that he is angry. And the notion that, as a baby, I began by looking in a mirror when I felt cross, noting my facial expression at tbe time, observing a similar expression from time to time on the face of my mother or nurse, and then arguing by analogy that these external bodies are probably animated by minds like my own, which are feeling cross, is too silly to need refutation. If the belief in other minds and other mental events were reached in this way, it might perhaps be entertained as a bold speculative opinion by a few exceptionally ingenious and observant persons at the ripe age of thirty-five. Its actual strength and its universal distribution would be utterly inexplicable on this hypothesis.

      (2) Next, I think it is equally clear that our ascription of minds to other human bodies, and our ascription of certain states to these minds on certain occasions, cannot be due to direct associations acquired in the course of our lives. A direct association between a certain facial expression and a certain emotion could arise only in the following way. We should need to have often seen a certahl expression (e.g., a frown) on our own faces when we felt a certain emotion (e.g., anger). If this could happen often enough, the visual appearance of this facial change might be associated with this emotion. When we saw a similar expression on another face we might automatically believe in the existence of a similar emotion, and we might locate it in the mind which animates this other body. But

1. the conditions which would be needed for the establishment of such an association are not and cannot in fact be fulfilled. We cannot see our own facial expressions in mirrors; and most of us pass through life with very little direct perceptual knowledge of what we "look like" when we feel angry, or pleased, or in pain. Of course, the same remarks do not apply to the natural expression of states of mind by interjectional noises. A baby who is in pain and howls can hear itself howling. An association might therefore be formed in its mind between the feeling of pain and the sound of howling. If it now heard a similar howl from another baby, it might automatically ascribe a painful feeling to it. Still, it is plain that this will not carry us very far; for many states of mind which have a natural expression are not naturally expressed by characteristic sounds, but by characteristic facial modifications.
2. Even if an association were somehow established between the visual appearance of a certain facial expression and the occurrence of a certain state of mind in ourselves, it is not obvious that this would suffice to explain our belief that this state is now happening in another mind when we see this expression on another face. The sight of this expression on another face might simply evoke a mild occurrence of the feeling in my own mind. But, even if it evokes in me the idea of that feeling, I might make various uses of this idea. I might just as well think of the past occurrence of this feeling in myself as believe in the present occurrence of this feeling in another. We may conclude then that the supposed association could not in fact be formed in the course of our lives; and that, even if it were, it would not suffice by itself to account for our belief that a certain kind of event is happening in another mind when we see a certain kind of expression on another face.

      There are two kinds of direct association which may be formed in the course of our lives, and must be carefully distinguished from the kind of association which I have just rejected.

1. When I frown, or have any other characteristic bodily modification, there is no doubt a characteristic bodily feeling. It is therefore very likely indeed that an association is quickly formed between certain of my mental states and the bodily feelings which are connected with their natural expression. But this kind of association evidently does not carry me beyond my own mind and its states.
2. It may be that a certain facial expression in another body which I can perceive has often been followed by overt action on the part of that body, and that this has often ended by producing some characteristic sensation and emotion in my mind. E.g, it might be that, as a rule, when I have seen people frowning, they have followed up their frowns by blows; and these may have caused pain and fear in my mind. An association may thus be formed between the perception of another's frown and the expectation of pain and fear in myself. But exactly the same kind of association may be formed in connexion with objects to which we do not ascribe minds or mental states, as when a " burnt child" learns to "dread the fire." Hence this kind of association cannot suffice to account for our belief in the existence of other minds, or of certain mental states which do not belong to our own minds.

      (3) It is now clear that we do not come to believe by a process of inference that other human bodies are animated by minds, and that we do not come to believe this through associatioos which have been formed in the course of our lives. And it is not in consequence of inference or of acquired associations that we ascribe certain states to other minds on seeing certain facial expressions on other bodies. Hence only two alternatives seem to be left. Either

1. there are certain cognitive situations which actually contain other minds or certain of their states as objective constituents; or
2. the visual appearance of certain bodily forms, movements, gestures, and modifications, has for us an unacquired meaning; so that, from the first, we pass from perceiving such things to believing that the perceived body is animated by a mind, and that this mind is owning such and such an experience.

I will now say something about these two alternatives.

      For reasons which have been repeated ad nauseam in the case of perception and introspection it is not possible that a mind, in the sense of an Empirical Self which may endure for years, can literally be a constituent of a cognitive situation which may last only for a few minutes. So at most we can suppose that the objective constituents of certain of our cognitive situations are mental events which in fact form parts of the history of other Empirical Selves. There is unfortunately no name, corresponding to "perception" and "introspection", for those situations in which we seem to be in direct cognitive contact with other minds and their states. But, as we shall often have to refer to them, we had better invent a name for them. For want of a better word, let us call them "extraspective situations". This name is not to imply any special theory about the right analysis of such situations.

      We must next notice that, even if mental events which are not owned by our minds be parts of the objective constituent of an extraspective situation, they are never the whole of its objective constituent. The perception of another body and of certain movements or modifications of it is essential to extraspection; and so one part of the objective constituent of any extraspective situation is the visual and other sensa by which the foreign body appears to us in perception. If there be cognitive situations in which a mental event belonging to another mind is the sole objective constituent, they must be classed separately. They might be called "telepathic" or "telegnostic", as distinct from "extraspective" situations. It will be remembered that we said that a perceptual situation always involves a sensational situation, and that it is perhaps doubtful whether purely sensational situations actually exist or are causally possible. On the alternative which we are at present considering, some if not all extraspective situations would involve a telegnostic situation; whilst all would involve a perceptual situation. It is of course certain that perceptual situations can and do exist apart from extraspective situations; but a person who accepted the present alternative about extraspective situations might quite legitimately doubt whether purely telegnostic situations exist or are causally possible.

      The next point to notice is this. It would not be necessary for an upholder of the present theory to assert that all extraspective situations contain a foreign mental event as part of their objective coostituent. Suppose that a considerable number of extraspective situations involve telegnostic situations; that they all involve perceptual situations with characteristic objective constituents; and that purely telegnostic situations rarely, if ever, arise. Then an association would be formed between such and such visual and auditory appearances and such and such foreign mental events. Suppose now that a purely perceptual situation were to arise, having for its objective constituents these characteristic visual and auditory sensa. Then this association would probably be excited, and we should automatically believe in the existence of such and such a foreign mental event, though it is not part of the objective constituent of the present cognitive situation. Beliefs reached in this way might often be true; but they might often be false. Thus the present theory is quite compatible with the existence of delusive extraspective situations.

      Now the kind of association which would certainly be formed if the present theory be true would presumably work both ways. Suppose then that a purely telegnostic situation were to arise. The association would now tend to call up images of a human being with the facial expression which corresponds to the mental event which is being telegnostically cognised, and it might lead to a perceptual belief that his body is now present to our senses. The result would be to produce an extraspective situation which is delusive on its perceptual side. If the reader will do what most philosophers are too proud and most scientists too prejudiced to do, and will study the evidence for "telepathy" in Phantasms of the Living, and the evidence which has accumulated since 1886 as marshalled by Mrs Sidgwick in the S.P.R. Proceedings for October 1922, he will see that a large proportion of the cases are of the kind just suggested. There is little evidence for pure telegnosis; but there is a great deal of excellent evidence for the existence of extraspective situations which are delusive on their perceptual side and veridical on their telegnostic side. By this I mean that the mental event which the experient claims to be apprehending really has happened at much the same time in a mind whose body is far away; but that the apprehension of this event has generally been accompanied by and bound up with an hallucinatory perceptual experience in which this body seems to be present to the experient.

      On the other alternative which we have to consider the extraspective situation never contains a foreign mental state as part of its objective constituent; i.e., it never involves a telegnostic situation. We must suppose that the innate constitution of human beings (and probably of other gregarious animals) is such that, when one sees any body which in fact resembles his own closely enough, he instinctively believes it to be animated by a mind like his own. And we must suppose that, whenever one sees in another a facial expression or hears a noise which in fact resembles closely enough the facial expression or noise which in him is the natural expression of a certain kind of experience, he instinctively ascribes a similar experience to the mind which he believes to animate the body of the other. If there be other creatures like ourselves, and if we be largely dependent on them, we could not have survived unless we either had instinctive beliefs of this kind, which are in the main true, or had telegnostic knowledge of some of their mental states.

      The next question is whether we can decide between these two alternatives. In the first place, we might accept the positive part of the second alternative, and deny or doubt the negative part. We might admit that there are such instinctive beliefs as the second theory assumes, and that there is also in certain cases genuine telegnosis. And it might be that those particular perceptual situations which call forth the instinctive beliefs are also those which are most favourable to the occurrence of a telegnostic situation. I think we may say at once that an analysis of extraspective situations which accepted telegnosis and denied the existence of these instinctive beliefswould be far less plausible than one which denied telegnosis and accepted the existence of these instinctive beliefs. For it is by no means certain that such instinctive beliefs, eked out by subsequent inference and association interpreted in terms of these beliefs, would not suffice to account for all the known facts about extraspection. After all, such a theory would leave our extraspective beliefs in foreign minds and mental events in no worse position than our perceptual beliefs in external physical events and things, if my analysis of sense-perception be admitted. In fact it would be in a slightly stronger position. For I have argued that we probably do have direct inspective knowledge of some mental events, whilst we probably do not have this kind of knowledge of any physical events. Thus we do not need any special postulate or category to provide us with the notion of a mental event, or to assure us that there are instances of such things; whilst, if I am right, we probably do need something of the kind in the case of physical events and things.

      If then we had nothing but ordinary extraspective situations to consider, I should be inclined to say that the assumption of a telegnostic factor is unnecessary and ungrounded, though it might still be true. But the actual position is somewhat different. From the study of abnormal phenomena it seems to me to be practically certain that there is such a mental power as telegnosis. It is therefore not a groundless assumption that it may be operative in some normal extraspective situations. And it is evident that, if it were present there mixed up with a perceptual situation and with instinctive beliefs about foreign mental events, it would be almost impossible to detect it. It therefore seems to me quite likely that there may be a telegnostic factor in many normal extraspective situations, i.e., that their objective constituents may include foreign mental events. I do not see that I can prove this, or that anyone else can disprove it; I merely say that I think there is a faint balance of probability in its favour in view of all the known facts.

Conventional Expression.

      A complete treatment of this subject would occupy volumes, and would be far beyond the powers of the present writer. I will therefore confine myself to a few remarks which seem specially relevant to our present purpose.

      As I have said, all intelligible sentences are natural expressions, on the occasion of which we believe that we are in presence of a foreign mind which is owning a process of thinking, judging, and so on. The vast majority of intelligible sentences are not about the | mental states of the person who utters them, and are therefore not conventional expressions of his state of mind. A sentence is a conventional expression of a state of mind only in so far as it asserts that the person who utters it is having this state. Even those sentences which do conventionally express a state of mind are very likely to be misleading. A man who says that he is having a certain emotion or volition may be intentionally trying to deceive us; or he may have introspected inaccurately, and be honestly mistaken; or he may be unable to find words which adequately express the results of a perfectly correct process of introspection; or we may be stupid and misunderstand the words which he uses. For all these reasons it is rash to believe that a certain man is having a certain experience at a certain time merely because he says that he is doing so. A prudent person checks such statements by noting the natural expressions of the speaker at the time and his subsequent actions and statements.

      In fact, for our present purpose, the natural and unintentional expression, which belongs to all intelligible sentences as such, is far more important than the conventional and deliberate expression, which belongs to a small minority of them. The hearing or reading of intelligent and intelligible discourse (whether we accept or reject what it asserts) is the occasion par excellence on which we feel perfectly certain of the presence of a foreign mind as distinct from the presence of mere mental events. To avoid an obvious criticism I must here make one important qualification. I say that intelligible discourse is the natural sign of the presence of a mind and of the presence of "thought", in a wide sense. But I am quite well aware that a great deal of intelligible and intelligent speaking and writing is accompanied by very little thought about its ostensible subject-matter. It would not be unfair to say that, while we are speaking and writing most, we are thinking least; and that, while we are thinking most, we are speaking and writing least. Anyone who prepares lectures knows that he was thinking about the subject, and not speaking, during his preparation; whilst he can largely let his mind "go on holiday" (to use an excellent phrase of Descartes) during the actual delivery of the lecture. We must not therefore say that the utterance of an intelligible discourse is a sign that the mind is now thinking ahout what the body is talking about. But we do feel perfectly sure that an intelligible discourse can be uttered only by a body which is animated by a mind that has thought and is capable of thinking again. Even here we must make a further qualification. There are many intelligible sentences, uttered in ordinary conversation, which are neither the expression of a present thought about their subject-matter nor the result of past thought about this subject-matter in the mind of the person who utters them. Many "expressions" of political and religious "opinion", which occur in the conversation of quite intelligent men, are of this nature. Nevertheless, these sentences would not have been spoken if someone else at some time in the past had not exercised his mind on these subjects.

      This brings us to a point where, as it seems to me, both Behaviourists and Bergsonians have gone wrong through failure to recognise an important distinction. Seeing how much of our alleged "thinking" is just the automatic reeling off of sentences or the mechanical manipulation of symbols, Behaviourists have tended to hold that all "thinking" reduces without residue to this. And Bergsonians have tended to contrast the merely mechanical processes of the "intellect" with a mysterious and superior faculty of "intuition", which is apparently supposed to be manifested in its purest form in the instinctive behaviour of animals. Now it seems to me that we must distinguish between what I call "fluid" and "crystallised" thinking. We must recognise that, whilst the greater part of any so-called process of "thinking" is of the latter kind, it must also contain short spells of the former. And we must recognise that the latter presupposes the previous occurrence of the former in the same mind or in some other mind. Anyone who considers what happens when he solves some problem for himself will recognise the difference. He would commonly he said to he "thinking" about the problem during the whole course of his work. Now, during the greater part of this period, he is certainly only manipulating symbols almost mechanically according to rules. But

1. at the beginning of the work, and at isolated intervals during the course of it, he must cease to do this and must contemplate face to face the actual ahstract objects with which he is concerned and their actual relations to each other. When he does this he is performing acts of "fluid" thinking; and no facility in manipulating symbols is any substitute for this. The power to perform acts of fluid thinking constitutes that difference between a man and a well-trained parrot which the Behaviourists (doubtless from excess of modesty) are so loath to admit.
2. I can now manipulate symbols blindly according to rules, and can feel confident that the result will accord with the real relations of things, only because I or my predecessors directly contemplated the things and their relations and made up a symbolism whose rules of operation were seen to accord with the relations of the things symbolised. Thus the symbolism is just the "crystallisation" of the past fluid thinking of myself or others; and, if it were not, there would not be the faintest reason to treat these operations with symbols according to rules as anything more than solemn trifling. It is simply unintelligible to me how this fact can escape the notice of any competent person, or why the tacit denial of it should be supposed to mark a wonderful advance in psychology.

      The position of the Bergsonians is less silly than that of the Behaviourists; for the former do at least recognise that mere crystallised thinking will not account for the facts. But why they should identify intellectual processes with that mechanical manipulation of words and symbols which I call "crystallised thinking", I cannot imagine. And I am equally at a loss to understand why they should suppose that the missing factor, which I call "fluid thinking", is specially manifested in instinctive actions. For these seem to be extremely like that mechanical reeling off of sentences which is supposed by them to be the special province of intellect, as opposed to intuition.

      To sum up. When we hear intelligible and intelligent discourse uttered we cannot help believing, either

1. that we are in the presence of a foreign mind which is thinking about the subject-matter of the discourse now, or has done so in the past; or
2. that at any rate there has been a foreign mind which did think about this subject-matter and is an essential condition of the possibility of the present utterance.

Which of these alternative beliefs we arrive at depends on the special circumstances in which the words are uttered.

The Logical Status of the Belief in Other Minds.

      So far I have confined myself to a purely descriptive discussion of extraspective situations. I have tried to show that they certainly do not involve inference; that our extraspective beliefs cannot be explained by direct associations which have arisen in the course of our lives; that they almost certainly depend on an innate and instinctive meaning which attaches from the first to certain perceived objects and events; and that it is not unlikely that some at least of them actually contain foreign mental events as part of their objective constituents. The purely logical question that remains is: "Granted that such beliefs are not in fact reached by inference, can they be supported by inference?"

      The Logical Connexion between Belief in Matter and Belief in Other Minds. I shall begin by considering a question which seems to be of very great interest and to have failed to receive the attention which it deserves. This is: "What logical connexion, if any, is there between the belief in Matter and the belief in other Minds?" This question first forced itself on my attention when reading the philosophy of Berkeley. It will be remembered that Berkeley denies the existence of matter, but is perfectly certain of the existence of himself and of God. He says very little about the existence of other finite spirits; but I think it is certain that he felt no doubt about the existence of other human minds. Now one can see that a Berkeleian has a right to be certain of the existence either of God or of other finite spirits. For he has certain sensations which are not due to his own volitions, and he holds that the only possible cause of anything is a volition of some mind. Hence he has a right to be sure of the existence of some mind other than his own, which has votitions. What seems more doubtful is whether he has a right to believe both in the existence of God and in that of other finite spirits. And this raises the general question whether a person who doubted or denied the existence of matter would have as good right to believe in the existence of other finite spirits as a person who accepted the existence of matter and held that we are in cognitive contact with it in perception. Corresponding to this would be the question whether a person who doubted the existence of other finite spirits would have as good a right to believe in the existence of matter as one who believed that there are other finite spirits and that we are in cognitive contact with them in extraspective situations. The second question is the easier of the two, and I will dismiss it before dealing with the first.

      If a man doubted or denied the existence of other spirits, it seems plain that he would be deprived of some of the grounds which ordinary men have for believing in matter. One ground which might be alleged for the view that my table is not a mere bundle of sensa, existentially dependent on myself, is that other people tell me that, when they are in my room and I am out of it, they are subject to perceptual situations with a very similar objective constituent to that of my perceptual situation when I am in the room. Another ground which might he alleged for believing in the independent existence of matter is that other people tell me that they know by bodily feeling that their bodies continue to exist when I cease to perceive them. Now suppose that I doubted or denied the existence of other minds. I should of course still hear and understand these utterances which apparently come out of the mouths of other human bodies. But, in so far as they asserted that a mind which animates these bodies has perceived or is perceiving something, I should have to doubt or deny the statement. If my gramophone said to me: "I saw your table all the time you were out of the room," I should not hold that this added any weight to the belief that my table existed in my absence except in so far as I believed that another human being who had been in the room had recorded this observation. Now, on the hypothesis under consideration, all statements uttered apparently by other human bodies will be in the position of statements uttered by gramophones, with the important difference that the "records" will not have been made by bodies which are animated by minds. And there would be no reason to attach any weight to these utterances. Of course the belief in matter is not reached by inference. It is, therefore, psychologically possible that a man might cease to believe in the existence of other minds and yet continue to believe just as strongly as before in the existence of matter. But, if he tried to defend his belief to himself by arguments, he would certainly be in a weaker position than a man who believed in the existence of other human minds.

      We now come to the other and harder question: "Would a man who doubted or disbelieved the existence of matter have as good a right to believe in the existence of other minds as one who accepted the existence of mauer and believed that he was in cognitive contact with it in perception?" in order to answer this question let us consider the sensa that we sense and the feelings that we feel, without regard to the question whether they are really appearances of our own and of other bodies. We can start by dividing them up into two great groups; viz.

1. those which we naturally regard as appearances of our own body; and
2. those which we naturally regard as appearances of foreign bodies.

The group (B) divides into two sub-groups; viz.

1. those which we naturally regard as appearances of other human bodies, and
2. those which we naturally regard as appearances of non-human bodies.

We will consider first the resemblances and differences between the contents of these groups.

      Group (A) consists mainly of bodily feelings; but it also contains certain characteristic visual, tactual, and auditory sensa. For I can "see" and "touch" parts of "my own body", and can "hear" "my own voice". Group (B) contains no bodily feelings in either of its sub-groups; but consists wholly of visual, tactual, and auditory sensa. The contents of sub-group (a) resemble that part of the contents of (A) which does not consist of bodily feelings. But it is much richer in content than the corresponding part of group (A); for I can "see" and "feel" much more of "other human bodies" than of "my own". The contents of sub-group (b) bear no special resemblance in detail to those of sub-group (a) or of group (A); for "human bodies" have a characteristic appearance, and "human voices" have a characteristic sound.

      Let us next consider the relations of my will to these various groups of sensa which I sense and of feelings which I feel. The great mass of feelings in group (A) is quite independent of my will. But, when I "will to move my body", I initiate certain changes in this mass of bodily feeling (viz., certain muscle- and joint-sensations, etc.). These changes are followed as a rule by certain characteristic changes in the visual, tactual, or auditory sensa of group (A); e.g., I may "hear myself speaking", "see my arm moving", and so on. These may be followed by characteristic changes in the sensa of group (B); e.g., I may "see a chair being moved by my hand" or may "see myself kicking another human body, and hear it cry out". The only way in which I can voluntarily affect the sensa in group (A) or in group (B) is by first initiating certain characteristic changes in the bodily feelings of group (A).

      These are the facts which are available for an argument by analogy when we confine ourselves strictly to what we can discover by inspection and introspection. The question which we have now to ask is whether the argumeot would be weaker or stronger according to whether we do or do not believe that these feelings and sensa are appearances of material things. The only way to test this is to consider in detail how the argument would run on each alternative assumption. I will begin by oonsidering the argument on the common-sense assumption that these sensa and feelings are appearances of material things.

The Argument for other Minds on the Assumption of Matter.

      On this view the sensa of group (A) will be appearances to me of that material thing which is my own body; those of sub-group (a) of group (B) will be appearances to me of that material object which is another human body. Now, on this supposition, the effect of my volitions is not directly to modify the feelings which I feel or the sensa which I sense. The direct result of my volitions is to produce internal changes in my body; and the changes which I observe in the feelings of group (A) are collateral results of these internal bodily changes. The physical consequence of these internal bodily changes is certain overt bodily movements of my limbs, tongue, etc. And the changes which I observe in the sensa of group (A) are merely collateral results of these overt bodily movements. Finally, the physical consequence of these overt bodily movements is, or may be, certain changes in external physical things; and the changes which I observe in the sensa of group (B) are merely collateral results of these external physical changes.

      On this view, another mind like mine would be one which animates a body like mine; wbich can directly produce changes within this body by willing; and can thus indirectly produce overt movements in its own body and changes in external bodies. If the other mind is like mine, its body appears to it as a group (A') of feelings which it feels and of certain characteristic sensa which it senses; and other bodies appear to it as a group (B) of sensa which it senses. This group (B') will divide into a sub-group (a') which is the appearance to it of other human bodies, and a sub-group (b') which is the appearance to it of external non-human bodies.

      On the present assumption, the argument by analogy for the existence of other human minds would run somewhat as follows. The resemblance of the sensa of sub-group (a) to the sensa of group (A), which are appearances to me of the outside of my own body, suggests that the sensa of sub-group (a) are appearances of an external body which outwardly resembles mine. Since it outwardly resembles mine, it is likely that it also resembles mine inwardly. Now the changes which I observe from time to time in the sensa of sub-group (a) resemble those which I observe from time to time in the sensa of group (A). The latter are appearances to me of overt movements of my own body, and the former are appearances to me of overt movements of an external body which outwardly resembles mine. So this external body resembles mine, not only in its outward form, but also in its overt movements. It is, therefore, likely that the internal changes which determine the overt movements of this external body resemble the internal changes which determine the overt movements of my own body. Now these internal changes in my own body are determined by my volitions, and appear to me as changes in the bodily feelings which I feel. It is, therefore, likely that the similar internal changes, which I assume on grounds of analogy to be taking place in the foreign body, are due to volitions. Now they are certainly not determined by any volition which I can introspect; and they are often contrary to volitions of mine which I can introspect. Hence it is probable that they are determined by volitions which do not belong to my mind. Now the order and connexion which I find among the changes of sensa in sub-group (a) resembles the order and connexion which I find among the sensa in group (A). So probably the overt movements of the external body have a similar order and connexion to that of the overt movements of my own body. But I know that this order and connexion in my own case is due to the fact that the successive volitions which determine the movements are not isolated mental events but are states of a more or less coherent and rational mind. I therefore infer that the postulated volitions are probably not merely isolated mental events, but belong to some mind other than my own, which is connected with the foreign body as mine is connected with my body. Now I know that my body appears to me through a mass of feelings which I feel and of sensa which I sense; I know that other bodies appear to me through sensa which I sense; and I know that, when I voluntarily produce internal changes in my body, these appear to me as changes in my bodily feelings. As I have postulated a mind with volitions like mine, connected with a body like mine in the same way in which my mind is connected with my body, I argue by analogy that probably this other body appears to this other mind by feeliogs which it feels and certain sensa which it senses; that probably other bodies appear to it through sensa which it senses; and that probably, when it produces internal changes voluntarily in its own body, these changes appear to it as changes in its bodily feelings.

      It is evident that such an argument as I have been describing has seme weight, if we grant the fundamental assumption that it makes about the connexion of sensa with material objects. And it is evident that this assumption forms an integral part of the basis of the argument. I argue to the existence of another mind like mine by way of the existence of another body which looks like mine and moves like it. And I believe in the existence of this other body because I be1ieve that the sensa of my (a)-sub-group are appearances of it, whilst the similar sensa of my (A)-group are appearances of my own body. The next question is whether I could legitimately argue to the existence of another mind like mine from the same facts without the assumption that the sensa which I sense and the feelings which I feel are appearances of material objects.

The Argument for Other Minds without the Assumption of the Existence of Matter.

      It is plain that, on the present alternative, the argument, if it be possible at all, must be very different in detail. We can no longer say that the immediate effect of my volition is to produce internal changes in my body, and that the changes in the feelings of group (A) are merely collateral effects of these internal bodily changes. We shall have to suppose that the immediate effect of my volitions is simply to produce changes in the feelings of group (A). Again, we can no longer say that the internal bodily changes produce directly overt bodily movements, and that the changes in the sensa of group (A) are simply collateral effects of these overt bodily movements. Instead we shall have to suppose that the voluntarily initiated changes in the feelings of group (A) directly produce changes in the sensa of group (A). Finally, we can no longer suppose that the overt bodily movements cause changes in external physical objects, and that the changes in the sensa of group (B) are merely collateral results of these external physical changes. We shall have to suppose that the changes in the sensa of group (A) in certain cases directly produce changes in the sensa of group (B).

      On this view another mind like mine would not be one which animates another body like mine. And it would not manifest itself to my mind by first directly affecting its own body and then indirectly affecting mine. For neither of us will have bodies. Another mind like mine will simply be one that feels a certain set of feelings and senses certain characteristic sensa, which together constitute an (A')-group. It will moreover sense another group of sensa (B'), and this will divide into sub-sets (a') and (b'). And it will be able directly to affect by its will some of the feelings in its (A')-group, and thence indirectly some of the sensa in its (A')-group, and thence at the secood remove some of the sensa in its (B')-group. But none of these feelings and sensa will be the appearances to it of material objects. Can any argument from analogy be founded on such a basis?

      So far as I can see it could only take the following form. The sensa of the (a)-sub-group of my (B)-group resemble the sensa of my (A)-group. And certain changes which I observe in the former resemble certain changes which I observe in the latter. Now these changes in the sensa of my (A)-group are immediately caused by changes in the feelings of my (A)-group. And these in turn are initiated by my volitions. The (a)-suh-group contains no feelings which I feel, and its changes are not correlated with any volition that I can introspect. Indeed they are often contrary to volitions which I can introspect. Now I might argue from the similarity of the (a)-sensa and their changes to the (A)-sensa and their changes that there is prohably a set of feelings which I do not feel, which are related to the (a)-sensa as I know the (A)-feelings to be related to the (A)-sensa. That is, I might argue that probably the (a)-sub-group is really part of a foreign (A')-group as well as being a part of my (B)-group. And I might argue that probably there are volitions that I cannot introspect, and that sometimes conflict with those which I can introspect, which directly produce changes in these hypothetical feelings and thus indirectly produce the changes which I from time to time observe in the sensa of my (a)-sub-group. Having reached this point, I might carry the analogy further. I sense and fee1 the contents of the (A)-group, and I own the volitions which directly affect the feelings, and thus indirectly affect the sensa, of this group. It is therefore probable that there is another mind which senses and feels the contents of this hypothetical (A)-group of which my (a)-sub-group is a part; and that this other mind owns the supposed volitions which directly affect the hypothetical feelings of this (A')-group and thus indirectly affect the sensa of this group. The analogy might then be concluded as follows. My mind senses a (B)-group beside sensing and feeling an (A)-group And this (B)-group splits into an (a)- and a (b)-sub-group. It is therefore probable that the supposed foreign mind senses a (B')-group beside sensing and feeling an (A')-group, and that this (B')-group splits into an (a')- and a (b')-sub-group.

      I do not know whether this argument from analogy will appear convincing to the reader. At any rate it seems to me to be the only one that could be used on the present supposition. It of course leads to a very different view of the interconnexion of minds from that which is held by common-sense. But this is natural enough, since common-sense believes that sensa and bodily feelings are appearances of material objects, whilst we have been explicitly rejecting this assumption in the present subsection. On the present supposition the group of sensa which I naturally take to be appearances to me of your body and the group of sensa which you naturally take to be appearances to you of your body partially overlap, so that some of them are sensed by both of us. You voluntarily produce certain changes in your feelings, which in turn produce certain characteristic changes in these common sensa. I notice these changes; remark their likeness to certain changes which I voluntarily produce in those sensa which I naturally take to be the appearance to me of my own body; find that they are not connected with changes in my feelings which I have voluntarily initiated, and that they often conflict with my volitions; and so I conclude that they are probably due to a foreign mind, which produces them by first voluntarily affecting certain bodily feelings which it feels and I do not.

      It will be noticed that this argument from analogy presupposes that certain sensa which are sensed by me are also capable of being sensed by another mind. Is this essential to the argument on the present supposition? I think that it is. So long as I confine myself to sensa and feelings, and make no assumption about their being appearances of material objects, the only sensa which I khow that I can affect voluntarily are sensa that I sense. That is, the only voluntary action on sensa with which I am acquainted will be immanent voluntary action. If I were to postulate another mind, which can voluntarily affect sensa which I sense and it does not, I should be postulating a mode of action for which I have no positive ground of analogy. For this would be transeunt voluntary action. I cannot be sure that the property of "being sensed by M" is not a necessary condition of the property of "being affected by M's volitions". It is of course perfectly possible that this is not so; it is perfectly possible that I do indirectly affect by my volitions sensa which I do not sense. I do not know that I cannot do this. But we cannot take a bare possibility as a ground for an argument from analogy. We must argue from what we know to be true, not from what we do not know to be false. Thus, although it is perfectly possible that there might be a plurality of minds, and that one might communicate with another by voluntarily affecting sensa which are sensed by the latter and not by the former, yet this would remain a bare possibility. I should have no positive ground of analogy for believing anything of the kind, if my only starting point is that I can voluntarily affect some of the sensa which I sense.

      Now it has been held by most philosophers that all sensa are essentially private, i.e., that if a sensum s be sensed by a mind M it cannot be sensed by any other mind. I now claim to have proved that, if we hold that sensa are private and also deny that they are appearances of material objects, it is impossible to produce a valid argument from analogy to the existence of other minds. If, however, we keep either of these assumptions and reject the other, it is possible to produce a valid argument for the existence of other minds from analogy. A fortiori, if we reject both of these assumptions, it is possible. The advantage for the present purpose of a belief in matter is this. If we believe that sensa are appearances of material objects, then, even if sensa themselves be essentially private, they are signs of the existence of something which is not private. The voluntary action which we observe in our own case is now transeunt from the beginning, for it immediately affects our bodies (which are public objects) and not the sensa which we sense (which may be private objects). So the arguments for the existence of other minds really are strengthened by the belief that sensa are appearances of material objects. For the analogy is valid, on this view, whether sensa be private or not; whilst, on the opposite view, it is valid only if some sensa at least be public.

      A final question remains to be raised. It might be said that it is sufficient for the present purpose that sensa should be assumed to be appearances of something public and neutral; it is not necessary that this something should be matter. This is of course perfectly true, so far as it goes. But, if we assume that sensa are appearances of something public and neutral, and that this something is not matter, what can it be? The only other plausible alternative is that the sensa which I sense are directly appearances of minds. Now, if we start with this assumption there is no need to use an argument from analogy to prove the existence of other minds. All sensa will be known from the outset to be appearances of some mind, and we shall merely have to seek for reasons for believing that some of the sensa which we sense are not appearances of our own minds. How we should set about doing this I do not know; but I do not see that any argument from analogy would be either necessary or useful.

Summary of Conclusions.

      Our belief in the existence of other minds is not reached by inference; and our belief in the existence of material objects is not reached by inference. Nevertheless, each of these beliefs can be rendered probable by certain inverse or analogical arguments, provided we admit that they have a finite antecedent probability. But the two beliefs are not logically independent of each other. For some, at any rate, of the arguments which support the belief in matter depend on our accepting the statements of other people about their perceptions; and the acceptance of such statements presupposes our belief in other minds. Again, arguments by analogy to support our belief in other minds presuppose either

1. that the feelings which we feel and the sensa which we sense are appearances to us of material objects, or
2. that some sensa are capable of being sensed by more than one mind.

Since the second condition is doubtful, whilst the first is sufficient even if the second be false, it follows that arguments by analogy in support of our belief in other minds are stronger if we believe that sensa are appearances of matter than if we do not.

Do we "perceive" Other Minds?

      I will end this chapter by trying to clear up a question which seems to me to be largely verbal. Some people are concerned to maintain that we "perceive" other minds; some people are concerned to deny it. And much heat is often engendered by this controversy. What we have to notice is that the question has three possible meanings. There are two senses of it in which the answer is certainly affirmative, and a third sense in which it is a fair matter of controversy.

1. If the question means: "Are there situations in which we believe in the present existence of certain mental states which do not belong to ourselves, and ascribe them to other minds without any process of inference?" the answer is "Yes".
2. If the question means: "Do the objective constituents of such situations have certain peculiar characteristics which distinguish them from the objective constituents of other situations?" the answer is again "Yes". For the sensa which are contained in the objective constituent of an extraspective situation always have those peculiar characteristics which lead us to take them as appearances of a human body or a human voice. And they change and succeed each other in perfectly characteristic ways.

If (1) and (2) be all that is meant by the question: "Do we perceive other minds?" the answer is that we certainly do perceive them. But

3. the question may mean: "Do extraspective situations contain a peculiar kind of objective constituent, which is not contained in other situations?" in that case the answer is doubtful. As I have said, I think that it is slightly more probable than not that some extraspective situations involve telegnostic situations, and therefore contain in their objective constituent the foreign mental event which we are said to be extraspecting. If this were certain, it would be certain that we "perceive" other mental events and other minds, even in this third sense. But it is not certain; and, therefore, if "perception" be taken in this very rigid sense, I can only say that it seems to me slightly more likely than not that we sometimes "perceive" other minds.

   ---

   Contents -- Go to Section C

   ---