CHAPTER IX

The Alleged Evidence for Unconscious Mental Events and Processes

Arguments for "Unconscious Sensations" Arguments for Undiscriminable Sensa Arguments for "Unconscious Perceptions" Arguments for "Unconscious Emotions" Arguments for "Unconscious Mental Processes" The "Unconscious" Factors in Conscious Mental Processes

      In the last chapter I pointed out a number of non-literal senses of the phrase "unconscious mental events", and I tried to give a sufficient description of literally unconscious mental events. Now there are a number of well-known facts, and of familiar arguments based on these facts, which profess to prove the existence of "unconscious mental events". Owing to the ambiguities of this phrase it is uncertain in what sense, if any, these arguments do prove the existence of "unconscious mental events". In particular it is uncertain whether they have any tendency to prove the existence of literal unconscious mental events, as described by us in the last Chapter. In the present Chapter I propose to consider some of the most familiar of these arguments, and to discuss what precisely, if anything, they prove.

      Before I begin I want to point out the difference between a sufficient description of an entity and a test for the existence of such an entity. We give a sufficient description of anything when we state a set of properties which together characterise it and do not all characterise anything else. But it may be that some of the properties mentioned in a sufficient description are such that it is difficult or impossible to tell by direct inspection whether they be or be not present in a particular case. Now it may be that there are other properties whose presence or absence can be readily detected by direct inspection. And it may be that their presence or absence involves the presence or absence respectively of those features in the sufficient description which cannot be directly inspected. If so, these properties constitute a test for the entity to which the description applies. Now it is evidently difficult or impossible to know by direct inspection or memory whether we formerly had an experience which we could not at the time have introspectively discriminated. The only test that we can well apply is certain facts about our present experiences, which are supposed to involve the past existence of an experience which we could not at the time have introspectively discriminated. Such tests always assume the truth of some proposition connecting the diagnostic properties with those mentioned in the description; and any doubt about these propositions throws doubt on the test.

  

Arguments for "Unconscious Sensations."

       There are certain facts, and certain arguments from them, which have been used since Leibniz's time to prove the existence of "unconscious sensations". Arguments about the roaring of the sea, the stopping of clocks, and so on, are what I have in mind; and they will be familiar to every one. Now, there are certain general remarks which apply to all such arguments, and throw grave doubt on their relevance to prove the existence of unconscious mental events.

       It will be admitted, I think, that the people who have used these arguments have not as a rule drawn any clear distinction between sensations and sensa. If they have done this, they have, nevertheless, generally assumed that sensa are themselves mental events and experiences of the person who senses them. And, even if they have distinguished between sensations and sensa, and have denied that sensa are mental events owned by the person who senses them, they have generally held that sensa are existentially dependent on being sensed by a mind. I will remind the reader that by a "sensum" I mean such an event as a coloured patch or flash in a visual field, or a noise. And by a "sensation" I mean such a situation as would be expressed by saying: "This noise is being sensed". Now sensa are nearly always outstanding parts or differentiations of spatially larger wholes, which I call "sense-fields". These in turn are generally temporal parts of longer strands of history which I call "special sense-histories". E.g., the sensum which is a certain coloured flash is generally an outstanding feature of a visual sense-field which is a sensibly coloured continuum of coexisting visual sensa. And this visual sense-field is generally a short slice of a longer whole which is a visual sense-history. The reader will find a more elaborate discussion of these conceptions in Part II of my Scientific Thought. Now, what the arguments under discussion prove, if they prove anything, is the existence of undiscriminated or undiscriminable sensa in a sense-field. For instance, the argument about the roaring of the sea tries to show that some auditory fields which seem quite homogeneous must yet be highly differentiated into distinct but undiscriminable noises. Now the relevance of such a conclusion to the existence of unconscious mental events depends very much on the view that we take about the nature of sensa.

       (1) If we hold that sensa are themselves mental events, an undiscriminable sensum will be, ipso facto, an undiscriminable mental event, i.e., an unconscious mental event as described above. But, personally, I see no ground whatever for thinking that a noise or a coloured patch in a visual field is itself a mental event, though of course the total situation whth contains such an object as its objective constituent is a mental event.

       (2) It might be held that noises and coloured patches, though not themselves mental events, can exist only as objective constituents of mental situations. This is a more plausible view, though I do not think that there are any conclusive arguments for it. Now it might seem that, if this were true, the existence of undiscriminable sensa would involve the existence of undiscriminable, and therefore unconscious, mental events. This, however, need be true only in a very trivial sense. Even if no sensum can exist except as objective constituent of a situation which also contains a non-objective constituent suitably related to the sensum, it does not follow that each different sensum needs a different non-objective factor and a different particular instance of this relation. Sensa are simply differentiated parts of a larver whole of the samc kind as themselves, viz, a sense-field. If this exists, all its differentiations exist ipso facto. If this whole be sensed, all its differentiations are thereby sensed; just as the fact that Cambridge is north of London involves the fact that Trinity College is north of London. Thus, even if all sensa have to be sensed in order to exist, and even if there be undiscriminable sensa, there is no need to assume any independent mental event beside the sensation of the whole field. And this is a literally conscious experience, since it would be perfectly easy to discriminate it introspectively if we wanted to.

       (3) Lastly, if, as I think most likely, sensa be neither themselves mental events nor dependent for their existence on being objective constituents of mental events, it is plain that the existence of undiscriminable sensa would have no tendency to prove the existence of undiscriminable experiences. We can say at once then that the arguments under consideration are wholly irrelevant to the existence of literally unconscious experiences, unless we accept the extreme view that noises, coloured patches, etc., are themselves mental events. It may, however, be of interest to glance at the arguments in order to see whether they prove the existence even of undiscriminable sensa.

Arguments for Undiscriminable Sensa.

       (1) Let us begin with the familiar argumcnt about the stopping clock. It is a fact that we may be sitting in a room in which a clock is ticking; and that, if we be suddenly asked by ourselves or by another whether the clock has been going, we cannot at once answer. But, if the clock should stop, we are liable to look up and say: "Why, the clock has stopped!" This is alleged to prove that we must have been "sub-consciously hearing" the ticking of the clock. Whst do such facts really suggest?

       (i) The negative part, the fact that we could not say off-hand whether the clock has been going or not, suggests that we did not in fact discriminate the ticking noise if it did in fact form part of our earlier auditory fields. If we had been discriminating it, we should almost certainly be able to remember it. (ii) The positive part, the fact that, when the clock does stop, our attention is arrested, does suggest that there is some difference or other between our total sense-field just before and just after the stoppage. Now, what the argument would need to prove is the following two propositions.

1. That our earlier fields did contain ticking noises; and
2. that these ticking noises were not merely undiscriminated but were undiscriminable.

Now it seems to me that the argument fails on both counts. If there were ticking noises in our earlier fields, there seems to be no reason to suppose that we could not have discriminated them if we had tried. All that we can say is that we did not in fact discriminate them, because we were attending to other things such as an interesting book. But it does not seem to me that the argument proves that there were ticking noises at all in our earlier auditory fields. The fact that we notice a difference when the clock stops needs for its explanation no other assumption except that, while the clock is going, it produces some modification somewhere in our total sense-object, which ceases when the clock stops. This modification might not be auditory at all; it might be simply a vague toning of our general bodily feeling. And, even if it be auditory, it need not take the form of a ticking noise. It might be simply a vague toning of our auditory field as a whole.

       It might be objected that, if this be all, why do we at once associate the change with the clock and say that it has stopped? To this I answer in the first place that it is doubtful whether we do this as a rule. It seems more accurate to say that we notice a difference, wonder what it is, look about, and finally fix the responsibility on the clock. But, even if we do sometimes immediately attribute the change to the stopping of the clock, this can be explained easily without assuming that there have been undiscriminated ticking noises in our earlier auditory fields. Suppose (to take the least favourable case) that the ticking of the clock simply produced a vague modification of our general bodily feeling. This modification must often in the past have been accompanied by auditory fields in which ticking existed and was discriminated and ascribed to the clock. Hence this kind of modification will have come to suggest by association ticking and clocks. Its cessation would therefore tend to make us look at the clock and ascribe the change to it. Thus, while I have no a priori objection to the possibility of undiscriminable sensa in sense-fields which we sense as wholes, I do not think that this argument has any strong tendency to prove that they exist.

      (2) I will now consider the argument about the waves of the sea. I will put it in a perfectly general form. When a number of similar physical stimuli are acting together we may be aware of a perfectly noticeable and characteristic sensum. When one of them acts separately, or only a few of them together, we may be unable to notice any sensum of the kind. It is argued that each of them must produce its own special sensum, and that each of these sensa must be undiscriminable.

      This is an atrocious argument. It assumes that the effect of a complex cause must be the sum of the effects which each of its parts would produce if acting separately. As a general proposition about causation there is nothing to be said for this. A very simple illustration of its folly is the following. The hearing of a certain note is physically conditioned by a series of compressions and rarefactions in the air which follow each other with a certain frequency. Now each separate compression or rarefaction is physically a part of this total cause. Will it be said that the note heard is the sum of the undiscriminable "notes" due to each separate compression and rarefaction? And, if so, how will you account for the fact that differences of note depend on the characteristic intervals between successive compressions or rarefactions?

       Even if we accepted this general principle about causation, it would be difficult to apply it to the case of sounds. Sounds have intensive magnitude, and it is not easy to attach a meaning to the statement that loud sounds are the sum of a number of coexisting soft sounds. It might be said, however, that a meaning can be given to this statement by considering the case of an orchestra; and that this example favours the original argument. Let us consider this point. It is true that, in listening to an orchestra, I do not as a rule discriminate the sounds due to the various instruments. But I can do so, if I choose to attend. Here of course each of the instruments would produce a noticeable and characteristic sound if played by itself. Now it might be said:

"The sound of the orchestra is the sum of the sounds of the separate instruments; and these separate sounds can be discriminated, although as a rule they are not. When we say that the noise produced by a sufficient number of separately inaudible stimuli is the sum of the undiscriminable noises due to each of them, we mean that the total noise is related to these undiscriminable noises as the sound of an orchestra is related to the discriminable sounds made by the separate instruments. And this is intelligible; because, in the case of the orchestra, you can both hear the whole and discriminate the parts."

I think that it would be better to drop the word "sum" and to substitute the word "pattern". The sound of the orchestra might be called a "pattern woven from" the sounds of the separate instruments; and, in this case, we can attend either to the pattern as a whole or to the elements out of which it is woven. The argument would then ask us to believe that a loud discriminable noise, like the roaring of the sea, is a pattern woven out of the many soft and separately indiscriminable noises due to the separate waves. In this case, it would be said, you can attend to the pattern as a whole but not to the elements out of which it is woven. Put in this way the argument is at least intelligible. An intermediate case between that contemplated by the argument and the example of the orchestra would be the following. We might have a number of precisely similar stimuli, eg., fog-horns, each of which separately would give rise to a noticeable sound. If they were to blow together we should be aware of a louder but qualitatively similar sound. Now, can we regard the noise which we hear when all these fog-horns are blowing together as a pattern woven out of the sounds which each of the fog-horns would make if it were blowing separately, just as we regard a symphony as a pattern woven out of the qualitatively dissimilar sounds of the separate instruments? If so, we might fairly argue by analogy that the sound of the sea is a pattern woven from a number of qualitatively similar but separately inaudible sounds due to each wave.

       The difficulty in the way of this argument is the following. It seems to me very doubtful whether the noises due to the separate fog-horns do exist within the noise made by all the fog-horns together. I do not think that we could discriminate these supposed separate noises, as we can discriminate the noises of the various instruments in the symphony. And this is certainly not because of their intrinsic faintness, since any one of them could be heard with perfect ease if it happened alone or against a qualitatively different background of sound. Thus it seems uncertain whether we have here an instance of a loud noise being a pattern woven out of a number of softer noises; for it is doubtful whether the softer noises which would be made by each foghorn if it were blowing separately exist literally witbin the louder noise made by all the fog-horns blowing together. It seems just as reasonable to hold that we have here a homogenrous auditory field, within which no sensa are discriminable because there are none to be discriminated. If this be so, the argument by analogy to prove that the noise of the sea is a pattern woven out of the faint undiscriminable sounds due to each wave, breaks down. Once again I have no a priori objection to the conclusion; I maintain only that the argument fails to prove it.

       (3) Before ending the present subsection I wish to point out the complete irrelevance of Stumpf's Argument for the present purpose. Stumpf's argument is perfectly valid; and it is quite true that it does not, as some of its critics have alleged, depend upon doubtful assumptions about the connexion between sensations and physiological stimuli. Stumpf might (and, for all I know, does) use his argument to the angels in Heaven, who have no bodies. It proves with almost complete certainty that some pairs of sensa, which seem to us to be exactly alike in quality or intensity, must really differ in these respects. But, even if sensa be experiences, no one has ever supposed that the qualities or intensities of sensa or the relations of identity or difference between their qualities or intensities are experiences. The unlikeness of two auditory sensa is not a third auditory sensum or any kind of sensum; it is a relation. Stumpf's argument merely establishes the fact that we can make mistakes about the relations of qualitative or intensive likeness between sensa which we sense; just as we may make mistakes about the likeness or unlikeness of anything else. This has no tendency to prove the existence of unconscious mental events in any sense whatever; except on two assumptions, one of which is probably false and the other of which is absurd. The first of these is that sensa are experiences owned by the mind which senses them. The second is that, in some sense, we must, ipso facto, know all the qualities and relations of our experiences merely because they are ours. Stumpf shows that I often judge sensum A to be exactly like sensum B, when in fact they cannot be exactly alike. On the two assumptions just stated I must in some sense know that A and B are not exactly alike if they be not so in fact. Thus, with these two assumptions, Stumpf's argument would prove the existence of unconscious knowledge about the relations of sensa, which conflicts with my conscious beliefs about the same subject. It would not prove the existence of unconscious sensations or of undiscriminated sensa, even on these assumptions. And, whatever one may think of the first assumption, the second is too silly to merit a moment's consideration.

  

Arguments for "Unconscious Perceptions".

       We can sometimes be pretty sure, on reflexion that, if we perceived a certain object on a certain occasion, this perception must have been unconscious, at least relatively to us. Now the circumstances may have been such that it would be surprising if no perception had existed at the time. Finally, we may afterwards have dreams, or say and do things in ordinary life or under hypnosis or psychoanalysis, which would be most naturally explained by the supposition that a perception did exist at the time. It will be noticed that, in order to establish the existence of an unconscious perception of a certain thing at a certain time, we must prove two propositions, one positive and the other negative:

1. We must prove that a perception of this thing did exist at this time; and
2. we must prove that the person who might have been expected to own this perception could not have detected it by introspection at the time, even if he had tried.

To prove the first point there are two main lines of argument. The first is that the general situation at the time was such that a perception of this thing might reasonably have been expected to happen then. The second is that some of the later experiences and acts of the person are such as would be likely to follow if he had perceived this thing, and not otherwise. To prove the second point we have again two main lines of argument. The first is that the person, not only does not remember percerving the thing, but also remembers not perceiving it. (I need scarcely say that these are two very different experiences.) The second is that some of the later experiences and acts of the person are such as would be likely to follow if he had not perceived the thing, and would be unlikely to follow if he had perceived it.

       A simple example would be one that we have already used, viz., looking for something, failing to find it, and yet discovering afterwards that it had been staring one in the face in the very drawer in which we have been looking. The argument would then run as follows:

"If I had recognised at the time that I was perceiving the object, I should certainly have found it. As I did not find it, it seems reasonable to suppose either that I was not perceiving it at all or that, if I was, this perception was not noticed by me. Now, if it existed, it is hardly likely to have escaped my notice by mere inadvertence; for this was the very experience which I was wanting and expecting at the time. Hence it seems probable that, if there was a perception of this object, and if it was owned by me, it was for some reason incapable of being introspectively discriminated by me at the time. Therefore, we must say either that there was no perception of the object at all; or that it was not owned by me; or that it was owned by me but could not have been introspectively discriminated by me at the time. If it was an experience of mine at all it must, therefore, have been a literally unconscious experience."

The argument might then continue as follows:

"The object was in such a position that light from it must have affected the central part of my retina; and, therefore, it is very unlikely that it did not produce a perceptual experience at all."

Lastly, it might be that in some cases we could add to this presumption something further of the following kind:

"Last night I dreamed of the object in a certain place in the drawer; and when I went this morning and looked again there it was."

Or again:

"I was hypnotised afterwards and told the hypnotist where this object was, and he found it there."

We should then have a pretty strong case, superficially at any rate, for the view that I had had a literally unconscious perception of this object when I was looking for it in the drawer.

       Let us now consider whether this case be really so strong as it seems at first sight. I think we may admit that such arguments make it highly probable that, if a perception existed at the time, it was literally unconscious relatively to the mind which was then controlling the body at any rate. So we may confine ourselves to the question:

"Is there any reason to believe that a perception of this object existed at all."

The evidence which is given for an affirmative answer to this question is derived, as I have said, from the later experiences or actions of the observer and from the general nature of the situation in which he was placed at the time. That is, we know that most people in the situation in which this man was placed would have had a certain kind of experience, and so we think it likely that he too must have had experience of this kind. And, again, we know that people can generally remember and be affected only by wbat they have already perceived. So when the man remembers a certain past event under hypnotism, or when he acts in ordinary life as if he had perceived this event, we assume that he must have perceived it. Now these are of course simply arguments from analogy; and, if the facts can be explained equally well in other ways, they do not carry much weight. Can the facts be explained equally well in other ways?

       In the first place, we must notice that the occurrence of an experience in the past is never a sufficient, even if it be a necessary, condition for the occurrence of a memory in the present. No memory will arise unless a trace, whatever that may be, has been left. Now, if a trace exists, it seems reasonable to suppose that it could give rise to mnemic phenomena even if no experience had accompanied its formation. Whether you choose to call any of these mnemic phenomena "memories" or not will be simply a matter of definition. 1, therefore, suggest the following hypothesis as a possible explanation of the facts with which we are at present dealing. When stimuli act on our nerves they usually give rise to two results, viz., an experience E and a trace T. But, under certain circumstances, only one of these results may happen; e.g., a trace may be formed, but no mental event may accompany its formation. If this trace be afterwards excited, the resulting experience will be exactly or very much like a memory of the experience E which normally accompanies the formation of the trace T. And, again, the resulting behaviour will be very much like that which normally follows from an experience such as E. Since the stimulus may thus have two alternative effects, it is reasonable to suppose that the total cause is of the form Sxy, where S is the stimulus. Here Sx produces an experience, and Sy produces a trace. We must suppose that the two factors x and y generally co-operate with each other, but that under certain circumstances x is inhibited. We then get a trace formed, but no mental event accompanies it. It is of course theoretically possible that, under other circumstances, y is inhibited. We should then get a mental event, but no trace would be left. Such mental events, if they exist, could never be remembered by any means, and could produce no mnemic effects. It seems to me that this simple and quite plausible hypothesis accounts for the facts perfectly well without the assumption of unconscious perceptions.

       I will now go a little more into detail. We must remember that the perception of a certain physical object involves sensation together with at least two other factors. In addition to sensing a certain field and its contents we must select a certain sensum from the rest of its sense-field; and in addition to this we must recognise this sensum as an appearance of a certain physical object. We might, therefore, fail to find a certain object which we were looking for, and which was staring us in the face, for mle of three reasons.

1. Because it was producing no sensible appearance in our sense-field; or
2. because the sensible appearance which it was producing was not selected and discriminated from the rest of the field; or
3. because this particular sensum was not recognised by us at the time as an appearance of the physical object for which we were looking.

I think it likely that the second and third possibilities are so closely connected as to form in practice only a single alternative. What I mean is this. I think that we select and discriminate certain sensa from the rest of the field largely because they represent certain physical objects to us. If a stimulus does excite certain traces of past experiences we both discriminate the sensum which it produces and take this sensum as an appearance of a certain physical object. If it fails to excite these traces we do not as a rule select and discriminate this sensum from the rest of the field; and a fortiori do not take it as an appearance of any particular physical object.

       It seems to me most likely that in the cases under discussion there really was a sensum in the visual field, which was in fact an appearance of the object that we were seeking; but that for some reason the traces which would usually be excited under such conditions were not excited, or, if they were, failed to produce their normal effect. This sensum was therefore not selected and discriminated from the rest of the field, and was not recognised as an appearance of the object for which we were looking. If this be so, there was no unconscious perception, for there was no perception at all. But there was an undiscriminated, and at the time probably undiscriminable, sensum in my visual field. This leaves a trace; and afterwards in dreams or under hypnosis the excitement of this trace excites those which failed to operate on the former occasion. We may then get what is, for all practical purposes, a memory of the physical object which we failed to perceive when we were searching for it. Or other mnemic results may follow which usually follow from actual perception. The psycho-analysts have many things of great interest to tell us about the emotional and conative factors which sometimes prevent sensations from developing into perceptions by preventing the usual traces from being excited or by inhibiting the usual results of such excitement.

       The conclusion of the matter is that, while I have no a priori objection to the existence of literally unconscious perceptions, I do not think that the facts which have been brought forward to prove their existence are adequate for the purpose. These facts can be explained quite as well by another hypothesis which is at least as plausible as the hypothesis of literally unconscious perceptions.

  

Arguments for "Unconscious Emotions."

       As an example of the facts which have been brought forward to prove the existence of literally unconscious emotions I will take the case of Livingstone and the lion. Livingstone tells us that he remembers that he felt neither fear nor pain while he was in the jaws of a lion and had given himself up for lost. There are other cases where men have been in extremely dangerous situations, and remember that they felt no fear at the time. Yet afterwards they dream of the incident, and the dream is accompanied by fear amounting to terror. It is reasonable to suppose that in such cases, if the emotion of fear existed at all while the incident was taking place, it must have been literally unconscious relatively to the agent at any rate. If Livingstone's memory be correct he looked at the time for the feeling of fear, which he expected to find, and was surprised to notice that it was absent. It is, therefore, reasonable to suppose that, if this emotion existed at all at the time it was not merely unnoticed by Livingstone but was unnoticeable by him at the time. This argument would not apply so strongly to men who have saved themselves in dangerous situations by their own efforts and resource. They must have been attending intently to external situations with which they had to deal promptly and effectively. They can, therefore, hardly have been attending at the time to their own emotions. When they say afterwards that they "remember not feeling fear" it would probably be much more accurate to say that they "do not remember feeling fear", which is a very different matter. In such cases the emotion, if it existed at all, was no doubt undiscriminated; but there seems no good reason to think that it was undiscriminable. But this explanation will not cover a case like Livingstone's; for he had ceased to struggle and had given himself up for lost.

      The evidence for supposing that an emotion of fear existed is twofold. In the first place, the situation was such that a man might be expected to feel fear in it. Secondly, the dreams and other later experiences of the agent are such as might be expected to be consequent on a past feeling of fear. Now this argument can be dealt with in precisely the same way as the argument for unconscious perceptions. I do not think that there is the least need to suppose that there was an emotional experience at all in Livingstone's case, or even that a trace was produced of the kind which an emotional experience generally leaves. All that we need to suppose is that the perception of the situation left an ordinary cognitive trace, and that the emotional experience which would normally accompany such a perception was inhibited. I take it that, although the external situation was such as normally produces pain and fear, the internal state of Livingstone's body at the time was abnormal. The result was that a perception arose as usual, but the emotion which would have accompanied it if the body had been in a normal state did not arise. This perception left a trace, as we know from the fact that Livingstone was able to remember perceiving the situation. When the patient, at some later date, goes to sleep and dreams this trace is excited and produces a dream of the original situation. The abnormal bodily state which inhibited pain and fear on the first occasion has now ceased. Hence the dream is accompanied by the sort of emotion which would normally have accompanied the original perception. There is therefore not the least need to suppose either that there was an emotion when the dangerous situation was being lived through, or that an "emotional trace" was left though no emotion existed. The cognitive trace, which was certainly left, is quite adequate to explain all the subsequent phenomena, both cognitive and emotional.

Arguments for "Unconscious Mental Processes."

      We have so far considered the evidence which has been brought forward to prove directly the existence of literally unconscious mental events, such as sensations, perceptions, and emotions. We have found this evidence to be quite inconclusive. But we must now introduce a distinction between mental events and mental processes. I do not pretend that this is an absolutely sharp distinction; but I think that a few examples will make clear what I have in mind by it. When I deliberately consider a number of possible alternative courses of action, weigh up the pros and cons of each, and finally decide on one of them to the exclusion of the rest, the whole long event may be called a conscious "mental process". Similarly, when I follow, or make up for myself, a long chain of reasoning the whole long event may be called a conscious "mental process." On the other hand, seeing a dog and feeling frightened of it would hardly be called a mental "process", though it is a mental "event". And feeling a twinge of toothache would hardly be called a "mental process", though it is a mental "event". We might roughly define a "mental process" as a long mental event which has a set of successive parts each of which is a mental event of a different kind from the whole. And the mental process is characterised by the nature of these parts and by the characteristic relations between them within the whole long event. For instance, a process of reasoning is a mental event which is divisible into three successive parts, none of which is itself a process of reasoning; viz.

1. contemplating and accepting the premises,
2. noticing that they logically entail a certain conclusion, and
3. passing from the state of merely contemplating this conclusion to believing it with a feeling of being justified in doing so.

(Very likely the analysis could be carried further, but this is enough to illustrate my meaning.) Now it might be that there is good direct evidence for the existence of unconscious mental processes, though there is no good direct evidence for the existence of unconscious mental events which are not, in this sense, processes. If this were so, there would be good indirect evidence for the existence of certain unconscious mental events which are not processes. For a mental process is a mental event having a set of successive parts which are mental events and not mental processes. We must therefore consider the alleged evidence for the existence of unconscious mental processes. We shall find that this is a somewhat complicated business, which cannot be completed in this Chapter. It will lead us up to the question of the real nature of traces and dispositions, which I reserve for the next Chapter.

The "Unconscious" Factor in Conscious Mental Processes.

      I think that it is best not to attack the question of the evidence for unconscious mental processes directly, but to make a detour by way of the unconscious factors in admittedly conscious mental processes. Suppose we consider an ordinary process of conscious desire which a man follows up to a successful result. When we analyse this we profess to find a process in which the man sets a certain end before himself, feels attracted by the prospect of it, chooses such and such means in order to get it, and so on. But how little of this actually is introspected or is introspectible! While the man is busy carrying out his purpose it is doubtful whether he has any literally conscious idea of the end or any literally conscious feeling of attraction towards it except for short spells separated by long intervals. Most of the time he is thinking of some minute detail in the means which he is using, and is not thinking about the end at all. Yet we say that idea of the end is constantly present, guiding his choice of means; and we say that the attraction which he feels for this end is a factor which is constantly present, making the agent persevere and surmount the obstacles which from time to time arise. Moreover, a conative process is often laid aside for a time, and then taken up again at the point where it was left; and this may go on for years. The writing of this book has been a conscious conative process which has, in a sense been going on for the last two years. But in every day of this period there have been long intervals during which introspection would discover nothing relevant to this conative process. Now here we are dealing with nothing odd or abnormal. I think that such examples show that we cannot identify the most normal conscious conation with those scrappy and separated bits of experience of which we have simultaneous undiscriminating awareness, and which we could no doubt introspectively discriminate if we tried. The fullest list of all these fragments with their observable interrelations is no more what we mean by a "conscious conation" than the set of interrelated sensa by which my table from time to time appears to me is what I mean by "my table". We may say that, just as the most ordinary statement about a physical object which I perceive goes beyond the sensa which I sense, so the most ordinary statement about a conscious "mental process" of mine goes beyond the scrappy and jumbled bits of experience which I could introspectively discriminate.

       It seems to me then that there is much more analogy between my perceptual knowledge of external physical things and processes and my introspective knowledge of my own mind and my conscious mental processes than is commonly admitted. Poor Locke and Kant have been abused like pickpockets for talking about "internal perception" and comparing it to "external perception". I believe that they were substantially correct; and that their main mistake was that they failed to draw certain very necessary distinctions which really make the analogy stronger than they tbought. In ordinary sense-perception we have to distinguish between perception itself and sensation on which it is based. My perception of a chair is based on my sensations of certain interrelated sensa which are appearances of it; and I should not say that I perceived the chair unless I did from time to time sense such sensa. But I mean something more by the "chair" than all these sensa; and I mean something more by "perceiving the chair" than just sensing these sensa. The notion of the chair contains as an essential factor the notion of a persistent something which joins up the various isolated "chair-sensa" which I sense, and which shows itself to me partially and imperfectly through them. Now it seems to me that we must draw a similar distinction in dealing with our introspective knowledge of our own conscious mental processes. In sense-perception we distinguished between

1. sensing
2. selecting and
3. using a sensed and selected sensum for perceiving.

I think that we must draw a similar set of distinctions in the case of our knowledge of our own mental processes. We must distinguish

1. undiscriminating simultaneous awareness of mental events
2. introspective discrimination of certain particular mental events, and
3. what I will call "introspective perception" of a conscious mental process by means of certain mental events which I have introspectively discriminated.

My introspective perception of an ordinary conscious mental process, such as a conation, is based on certain interconnected experiences of which I do have simultaneous undiscriminating awareness and some of which I may introspectively discriminate. I should not say that I "introspectively perceived" this conation of mine unless I were introspectively aware of these interconnected bits of experience. Nevertheless, I mean something more by my conation than these bits of experience; and I mean something more by "introspectively perceiving" it than just having simultaneous undiscriminating awareness of these bits of experience or just introspectively discriminating them. The notion of a conation involves as an essential factor the notion of a persistent something which joins up these relatively isolated bits of experience and reveals itself partially and imperfectly to me through them.

      I prefer not to call my "perception" of my own mental processes "internal percention", because this name seems to me to apply better to the perceptual knowledge which I have of certain physiological processes in my own body by means of bodily sensations, such as headache or toothache. I prefer to divide Perceptien into

1. Introspective, and
2. Sensuous; and then to divide Sensuous Perception into
   1. External and
   2. Internal.

I shall say that the introspectible experiences on which I base my introspective perception of a conscious mental process are "appearances" of the latter to me. This word is to have no implication of delusiveness. At present I do not wish to deny the possibility that the introspectible appearances of a conscious mental process may be literally slices of it; that is a question for subsequent discussion.

      If the view which I have been suggesting be accepted it will be seen that there is a close analogy between the problem of the nature of traces and dispositions and their relation to introspectible experiences, on the one hand; and the problem of the nature of physical obiects and processes and their relation to sensa, on the other. Now anout the latter problem there are, I think, five possible types of theory; and corresponding to each of these will be a possible type of theory about the nature of traces and dispositions.

1. Phenonnenatism. This is the theory that, strictly speaking, there are no physical objects or processes. There are just the sensa which we sense and the observable relations between them which put certain of them into certain groups. Corresponding to this would be Mr Russell's tentative suggestion of a special kind of Mnemic Causation which directly connects introspectible experienceS and does away with the need for assumine traces and dispositions.
2. Naive Realism. This is the theory that the sensa which are appearances of a physical object are literally spatio-temporal parts of that object, and that the spatio-temporal parts of it which are not manifested in sensation are of precisely the same nature as those which are so manifested. Corresponding to this would be the theory that the introspectible appearances of a conscious mental process are literally temporal slices of the whole process, and that the non-introspectible slices are of precisely the same nature as those which we can introspect.
3. Critical Realism. This is the theory that sensa are not literally spatio-temporal parts of the physical objects of which they are appearances; that there are certain characteristics which belong only to sensa and not to physical objects; and that there are other characteristics which belong to both, though not necessarily in the same determinate form. In the next Chapter I will try to explain the theory about traces and dispositions which would correspond to this.
4. Agnosticism. This is the theory that we have no means of telling what are the characteristics of those relatively permanent things and processes which manifest themselves partially to us by the interrelated sensa which we from time to time sense. Corresponding to this would be the view that we cannot tell whether what manifests itself partially to us by interrelated experiences is mental or physical or both or neither.
5. Mentalism. This is the theory that the relatively permanent conditions of interrelated sets of sensa are minds or states of mind, e.g., colonies of spirits of a low order of intelligence (Leibniz) or standing volitions in the mind of God (Berkeley). Corresponding to this would be the view that the bits of experience which we can introspect are appearances of purely physical or physiological processes, and that the intervals between these introspectible experiences are filled in with something which is physical or physiological and not in any sense mental.

I leave the discussion of these alternatives to the next Chapter.

Wholly "Unconscious Mental Processes".

       We are now in a position to consider the arguments for "unconscious mental processes" which are not merely factors in "conscious mental processes." A completely unconscious mental process would be a process of the same kind as those which appear to us or to others as a series of characteristically interrelated experiences, but which does not in fact appear to anyone in that way. It may be compared with a completely unperceived physical object or process, such as an unperceived fire. This is a process of the same kind as those which appear to ourselves or to others as a series of characteristically interrelated sensa, but which does not in fact manifest itself to anyone in this way. Naturally, a belief in a completely unconscious mental process, like a belief in a completely unperceived physical object or process, must rest on inference; and this inference will take the form of an argument from causation and analogy. To illustrate this I will consider the arguments which have led psycho-analysts and others to postulate wholly unconscious conations.

       These arguments are logically of the same type as those which led Adams and Leverrier to postulate the existence of the hitherto unperceived planet Neptune. In both cases they are arguments by analogy to explain certain observed perturbations and irregularities. We are quite familiar with the conflict between two conscious wishes. Each wish, in the absence of the other, would appear as a certain typical series of conscious experiences, though it would be more than the sum-total of these. When both coexist we can observe that the two series of conscious experiences do not go on side by side, but they are as a rule replaced by a modified series which is in some sense a compromise between the two. This modified series may resemble either of the unmodified series in various degrees, thus showing the relative strength of the two conscious wishes. And the conflict is accompanied by a characteristic conscious experience of strain and uneasiness. All these facts are of course quite familiar and open to introspective observation. Now we may sometimes find one and only one conscious wish; and it may manifest itself to superficial observation as a certain series of conscious experiences which would be the normal manifestation of a single wish. And yet, if we look more closely into our minds or study our actions more carefully, we may find isolated bits of conscious experience or isolated actions which are not normal appearances of this conscious wish but are more like the normal appearances of a certain conflicting wish. Or we may find, instead of these isolated and anomalous conscious experiences and actions, another peculiarity. We may find that our actual series of conscious experiences is more like the compromise series which occurs when we have two conflicting conscious wishes than like the pure series which occurs when we have a single conscious wish. Lastly, there will sometimes be a feeling of strain and uneasiness which resembles the feeling that we have when two conscious wishes conflict, although here we can introspectively perceive only one wish. Under these circumstances it is reasonable to suppose that there really are two processes going on, and that they are both such that we should call them "conscious wishes" if they manifested themselves as series of conscious experiences. And it is reasonable to suppose that they have to each other that kind of relation which often manifests itself as a conflict between two conscious wishes. But in this case, for some reason, one of these processes cannot manifest itself in a series of conscious experiences characteristic of a wish. It can manifest itself only by a few isolated and anomalous conscious experiences and actions, or by perturbing and compromising the series of conscious experiences which is the manifestation of one of these wishes. This hypothesis will of course be strengthened if we do have a vague feeling of strain and uneasiness which cannot be explained by any observable conflict between our conscious wishes. So far the argument of the psychoanalysts is exactly like the arguments by which Adams and Leverrier were led to suspect the existence of the planet Neptune, and to predict certain facts about it.

       The next stage was for astronomers to look for the hypothetical planet; and, as we know, they were eventually able to perceive it. The psycho-analysts would say that they had followed a similar course with similar results. The astronomers by their special technical methods succeeded in making the planet Neptune manifest itself by certain sensa characteristic of a planet such as Adams and Leverrier had suspected to exist. The psycho-analysts profess that, by using certain technical methods, they have in many cases made the previously unconscious wish manifest itself as a series of conscious experiences characteristic of such a wish as they had suspected to exist. If we accept this statement (and I see no reason why we should not) the analogy is complete.

      When we are considering the arguments for unconscious mental processes used by psycho-analysts and others, I think we must distinguish three questions.

1. Are the arguments logically of the same form as arguments which we should admit to be valid in other spheres? Here the answer is certainly in the affirmative. As I have just shown, they are of precisely the same form as those used by Adams and Leverrier in astronomy. True, they deal with a subject-matter which is much more complex and about which we know much less. And they cannot be put into mathematical form or tested by making exact measurements. This no doubt reduces the probability of their conclusions, but of their general formal validity there can be no doubt.
2. Do they make it probable that there are certain processes which do not give rise to characteristic sets of interrelated conscious experiences, but which are nevertheless of the same nature and capable of the same mutual relations as processes which do manifest themselves by such sets of conscious experiences? Here again I think there is little doubt that the answer is in the affirmative.
3. Have we a right to call such processes "mental"? This question is ambiguous. If it merely means: "Are they of the same nature as those processes which do manifest themselves as conscious experiences?", the answer is that they most probably are. If it means: "Are they composed of mental events which are of the same nature as the conscious experiences by which other processes of a similar kind manifest themselves?", the answer, if one can be given at all, must be left to the next Chapter. And the answer to it will depend largely on what view we take about the nature of the "unconscious parts" of "conscious mental processes". If we have to regard the "unconscious parts" of a "conscious mental process" as non-introspectible mental events, and the "conscious parts" as differing from them only in being introspectible, we shall have to answer the question affirmatively. Otherwise, we shall have to answer it with a definite negative; or with a confession of ignorance; or by analysing it into several questions, some of which can be answered affirmatively, others negatively, and others perhaps not at all.

         I have so far confined my attention to "unconscious wishes" and to the arguments which have been brought forward by psycho-analysts in favour of their existence. There is another set of facts which seem strongly to favour the hypothesis of "unconscious mental processes". These are the cases in which a patient under hypnosis is told to perform a certain act at a certain number of minutes (which may run into hundreds or thousands) after he has been awakened. It is found that certain patients perform the suggested action automatically at or very near to the suggested time. This seems to imply that the time which was given in minutes has been reduced to days and hours by some process of mental arithmetic, and that a watch has been kept for the arrival of the calculated moment. Yet the patient cannot discover this process of calculation and of watching by introspection. Here it certainly seems reasonable to suppose that a process has been going on similar to the processes which manifest themselves as a rule by the series of conscious experiences which we call "making a calculation". This process, for some reason, does not manifest itself in this way to the mind which normally controls this body. Whether the process can strictly be called "mental" depends upon considerations which must be discussed in the next Chapter.

          There is, however, one remark which must be made before ending this Chapter. Suppose that, on again hypnotising the patient, he says that he remembers making the calculation and watching for the calculated moment. If we accept this statement we shall have to suppose that a genuinely mental process has been going on, that it is literally unconscious relatively to the mind which normally controls this patient's body, but that it is literally conscious relatively to the mind which controls this body when it is hypnotised. And of course there may be good grounds for regarding these two "minds" as parts of a single mind which animates the body at both times. Now, supposing that such statements are made under hypnosis, I think that it is certain that some weight must be attached to them. We must remember, however, that a person under hypnosis is in an extremely suggestible state, and is very liable to answer questions in the way in which he supposes the qucstioner to want them answered. Thus I do not think that it would be reasonable to attach as much weight to the statements of a hypnotised person about his own experiences as to the statements of a normal waking person. We should at least need to be very certain that the questions were not put in such a way as to suggest even faintly to the patient that the questioner was wanting a certain kind of answer. I have not enough first-hand knowledge of the facts to feel sure whether these conditions have been fulfilled by the experimenters on this subject.

         Somewhat similar remarks apply to the statements made by alternating personalities, who claim to be co-conscious, like Sally Beauchamp. Such a personality, when in control, may tell us that it owned a certain continuous series of conscious experiences while another personality was in control. And this series may be such as to fill the gaps between certain fragmentary and isolated bits of conscious experience belonging to the latter personality. Here we have evidence which, taken at its face-value, would suggest that the "unconscious parts" of a conscious mental process in one personality are literally conscious experiences in another co-conscious personality. And, if we accepted this, we might be inclined to generalise this to normal cases, and to suggest that the difference here is that the coconscious personality never gets control of the body and therefore never has a chance of telling us about its experiences. Now I think that some weight must be given to the statements of personalities like Sally about their own experiences. Moreover, it cannot be said that Sally was particularly suggestible; she seemed to have a strong will of her own and to be quite capable of resisting unpalatable suggestions. But was the suggestion that she was co-conscious an unpalatable one? One gets the impression that she was very anxious to make herself out to be as important and mysterious as possible. (It will be remembered that at certain stages of the proceedings she claimed to be a "spirit".) There is obviously a fairly close connection between multiple personality and ordinary hysteria. Now hysterical persons can be extremely obstinate in many respects, as Sally was. But they show an embarrassing readiness to provide evidence for any theory about hysteria which they believe to be held by the doctor who is treating them. And probably they would specially welcome any theory which enables them to feel themselves to be mysterious beings who are creating a revolution in current medical and psychological concepts. Thus, although we cannot afford to neglect their statements about themselves and their experiences, we ought to view them with very great suspicion.

   ---

   Contents -- Go to Chapter X

   ---