1. It would be pretty generally agreed I think that the wave of hostility towards introspection in the early decades of this century manifested most strikingly in the self-denying and strangely paradoxical ordinance of the Behaviourist psychologists who proclaimed that the only effective way of studying the mind was to avert one's gaze from it and attend exclusively to the body has now even among psychologists in large measure spent itself. As Professor and Mrs. Knight have written in their recent Modern Introduction to Psychology: ‘… the opposition between introspectionists and behaviourists is no longer a live issue. The pure introspectionist view has long been abandoned and extreme behaviourism now has few adherents.’1 The controversy has however in many respects been not unfruitful. On the one side the difficulties dangers and limitations of the introspective method have been more precisely determined and have achieved a wider recognition. On the other side it has become increasingly clear that however necessary it be to exercise caution in the practice of this method and to supplement it and check it wherever practicable by various extro-spective techniques it is nevertheless not only an important but a strictly indispensable instrument for the acquirement of knowledge of mental states and processes.

This rapprochement is very welcome and relieves me of the tedious duty of traversing again much well-trodden ground. It will be sufficient I think if I make clear my position on one or two matters of importance that are stilt the subject of some controversy.

2. I take it that most people if asked to define what it is that introspection studies would reply in effect ‘psychical states and processes’. Yet Samuel Alexander was surely right in maintaining that as ordinarily used the term covers a good deal more than this.2 It covers the study not only of psychical states and processes but also of the ‘objects’ towards which these are directed in so far as such objects are presented through the said states or processes. Suppose e.g. I say to someone ‘I believe X’ and he replies ‘Are you sure you believe X?’ His question may mean either of two things which it means being generally made apparent by the relative stress laid on the words ‘believe’ and ‘X’. He may mean ‘Is your state of mind towards X one to which the term “belief” is really appropriate?’ Or he may mean ‘Is it really X towards which you are in a believing state and not perhaps XY or XYZ?’ Now in either case it is natural to say that we seek to answer the question by ‘looking into our minds’ to see what was really ‘there’ when we said ‘I believe X’. And it is also natural to call this ‘introspection’. But it seems quite clear that whereas in the former case we are indeed concerned to study a psychical state in the latter we are not. There we seek to ascertain not whether our psychical state was one of believing—that question is not being raised—but whether the proposition believed is correctly described as X.

Now whether Alexander is also right in insisting that it is only to the former case—the examination of psychical states and processes—that the term ‘introspection’ can legitimately be applied does not here concern us. Not that this is by any means a merely verbal question. Anyone who accepts Alexander's particular brand of realist epistemology must I think admit the propriety of Alexander's restriction upon the usage of the term. But in the present context our interest is only to justify introspection in the sense in which Alexander agrees that the term is appropriate. How far can we get by introspection valid knowledge of the self's psychical states and processes of the experiences of a self as in Alexander's terminology actually ‘enjoyed’?

It has sometimes been suggested that whatever knowledge introspection may yield it cannot be knowledge of the self proper. For in introspection the self in the nature of the case is turned into an experienced object which is just what the self qua self is not. It is of the very essence of the self qua self to be an experiencing subject. We do not therefore in introspection ever apprehend the self as it really is.

It does not seem to me however that there is very much in this objection. No doubt the self in introspection is in some sense an ‘object’. But it need not be so in any sense which precludes it from being also the self as subject the self in its actual subjective functioning. We may agree that the introspected object can never be the subject in its subjective functioning at the moment of introspecting. We do not directly introspect our own introspecting. But why should it not be the subjective functioning of the introspecting subject at some previous moment? Reluctance to admit this seems to arise from some misunderstanding about what we actually do when we introspect. It is insufficiently appreciated that when we want to introspect we begin (and indeed continue) by re-living through memory the experience in which we are interested. In introspection of course we re-live it ‘with a difference’. In the past experience there was as in all conscious experience some degree of self-awareness: but it was likely to be a very low degree—for in ordinary life attention is normally concentrated upon the ‘object’ experienced rather than upon the experience itself—and it was certainly not self-awareness controlled by any ‘scientific interest’. In introspection the past experience is lived through again but now with the difference that the self-awareness is both raised to a high degree and is also scientifically oriented. Alexander is surely near to the heart of the matter when he defines introspection (in so far as used for psychological purposes) as ‘enjoyment lived through with a scientific interest.’3

3. Is it only the self in its past subjective functioning that can be introspected? Is all introspection really retrospection? I have hinted at my own agreement with this view but the question merits much closer consideration. For if all introspection is retrospection certain disabilities must attach to it and they may be of a kind to prejudice seriously its claim to give satisfactory knowledge of the self.

Since Auguste Comte's classic criticism more than a century ago it has been widely acknowledged that at least in the case of the mind's cognitive processes introspection at the time of occurrence is not practicable. As Comte puts it ‘The thinker cannot divide himself into two of whom one reasons and the other observes him reason. The organ observed and the organ observing being in this case identical how could observation take place?’4 To this objection I know of no effective answer. I think we must grant that the cognitive processes at any rate cannot be introspected at the moment of their occurrence but only afterwards as revived in memory; and we must frankly accept whatever reduction of the certainty of introspection in this field is entailed by the need to put our trust in memory.

But do the objections to ‘introspection of the present moment’ hold of cognitive processes only? This has sometimes been claimed to be the case (and Comte has himself lent some countenance to the claim). It is held that we can retain unabated our confidence in the direct observations of feelings and conations since here ‘the organ observed and the organ observing’ are not one and the same. To take an obvious instance—the instance presumably most favourable to the possibility of ‘introspection of the present moment’—that of a persistent physical pain. The experience of the pain it is contended does not involve a cognitive process of the mind and there is accordingly no reason why the cognitive process of observing the experience of the pain should not be carried on simultaneously with the occurrence of the experience itself. And as a matter of plain fact is not something of the sort happening constantly? The doctor asks his patient to tell him just what kind of a pain he is suffering from and the patient does not require to rely upon his memory in order to give the information. He concentrates his attention upon the pain as it is now occurring and reports that it is dull or acute continuous or intermittent in his heart or in his head and so on.

I suggest nevertheless that this view is plausible only at first sight. It seems on fuller consideration that even an experience like that of pain must to some extent undergo change by being reflected upon by the victim of it. We can readily agree that the experienced pain does not involve what would ordinarily be called a ‘cognitive process’. But it does involve the subject's attention to it; involuntary attention no doubt but still attention. It is a common-place that one of the most efficacious ways of relieving a patient's pain is to distract his attention from it. There is even good evidence that a man can train himself to become virtually impervious to pain by prolonged practice in the art of concentrating upon extraneous matters ‘at will’. If then as happens when he sets out to introspect a man's attention is directed away from the pain it seems inevitable that the experience of the pain will be modified in some degree however slight. Introspection of a present pain does not involve Comte's ‘thinker’ dividing himself into two in the technical sense of the term ‘thinker’. But it does involve an attending subject dividing himself into two; and this feat seems no easier of accomplishment.

I find myself therefore forced to take the view that introspection of the present moment must fail of its proper object even where the experience to be introspected is one so remote from the cognitive activity as that of pain. I think there is in fact no mental state or process that can remain unaffected by being reflected upon ‘with a scientific interest’. We can of course in a sense introspect a present pain. But the point is that this ‘present pain’ is not the pain we set out to introspect. That pain has undergone modifications through the influence of the introspective activity directed upon it. The present pain is what it is at least in part through such influence.

If such be the case it becomes very necessary to enquire just how serious are the disabilities entailed in our having to rely upon memory in introspection generally.

4. On the whole I incline to the opinion that these disabilities have been much exaggerated. That memory is in principle fallible everyone admits. But there is a great deal of past experience in respect to which no one believes that memory does in fact deceive us. Where the experience to be remembered belongs to the very recent past and also includes items of which the subject was vividly conscious at the time accurate introspection (or retrospection) of such items seems well within human powers. I can be virtually certain that my experience of a moment ago was one of pain and that the pain was severe in intensity was of a throbbing character and was located in my big toe. Introspection runs into difficulties only where the past experience to be introspected was either remote in time or where though recent it included features some of which were only faintly apprehended. Thus I should find difficulty in introspecting the characteristics of the pain I suffered from an attack of lumbago fifteen years ago even though these characteristics were doubtless vividly enough focused in my attention during the original experience. Thus too if only a few moments ago I made a troublesome practical decision which required me to take into account a great variety of factors I should find great difficulty in introspecting the more dimly apprehended ingredients in the complex experience of my motivation despite the recency of their occurrence.

But granting that it is hard to be confident of the accuracy of our introspection either in the case of experiences long past or in the case of the finer details in even the most recent experiences it is apparent I think that at least one of these difficulties—remoteness in time—can in practice be largely discounted. For almost all of the mental states and processes which one is likely to want to introspect with a scientific interest either occur frequently in the natural course of events or if they do not are capable of being produced readily and as often as required by an act of constructive imagination. Introspection immediately after their occurrence is therefore almost always possible. Introspection compares very favourably indeed with extrospection of material things in regard to the easy accessibility of its objects and we may well agree with Stout that ‘this may be set down as the grand advantage of introspection which compensates in a high degree for its drawbacks.’5

The other difficulty however remains: and few who have seriously tried their hands at introspecting the more faintly experienced constituents of even the most recent complex mental states and processes are likely to under-rate it. There is no good reason indeed to suppose the difficulty insurmountable. Very much can be done in the way of correcting initial errors and oversights by frequent repetitions of the experience to be introspected and by careful comparison of one's findings with the findings of other workers in the same field. Still one must admit that even the most skilful and experienced practitioners of introspection would seldom venture to claim that their efforts however pertinacious have yielded a completely comprehensive and impeccably accurate picture of all that there is in any really complex mental experience.

It is a matter for satisfaction therefore that this very real difficulty has after all not much bearing upon the particular sort of empirical self-knowledge which is our concern in these lectures. It is sufficient for our purposes if the possibility of accurate introspection of the broad basic modes of experience is granted; and the possibility of this does not seem to be effectively impugned by any of the criticisms of introspection we have considered. One may fairly regard it as significant that there is virtually no disagreement about the classification of these basic modes by psychologists who do not deny themselves the use of introspection. The traditional psychological distinctions under the general rubrics of feeling conation and cognition have been arrived at by introspection and I see no good reason for doubting that introspection tells us substantially the truth about them. Moreover if I have been right in dismissing as unsound the contention that the states and processes which are the objects of introspection cannot because objects be states and process of the self qua subject then the truths which introspection yields can be taken to be truths about the self ‘proper’. The basic modes of experience disclosed by introspection will be real manifestations of the real self and as such will furnish evidence for valid conclusions about that self's ‘nature’.

Before we finally take leave of introspection there is just one further point to which I should like to draw attention. I have been talking almost exclusively about the alleged defects of introspection with only a casual reminder of the ‘grand advantage’ of introspection in regard to the ready accessibility of its objects of study. But there is another ‘grand advantage’ of introspection over extrospection of physical things which is at least as important; an advantage which might even be said to establish knowledge of the inner life as on an altogether superior plane to knowledge of the external world. The ‘objects’ we are trying to know in introspection are mental states and processes as they are for the experience of them. The ‘objects’ we are trying to know in observation of the external world on the other hand are physical states and processes not merely as they are for the experience of them but as they are in themselves. In the case of the external world we may be virtually certain that the object as experienced by us has the character X; but this certainly does not carry with it any certainty that X is a character of the physical object itself. Possibly X is only a subjectively conditioned ‘appearance’ of the real object. But in the case of internal observation or introspection the situation is quite otherwise. Here the distinction between the ‘object’ as it is for the experience of it and the ‘object’ as it is in itself does not arise. There is no meaning in asking what a mental state or process might be as not experienced. As not experienced it is not strictly a mental state or process at all. If e.g. we are sure on the basis of introspection that the mental state we are concerned to know was for our experience a state of pain then we are sure that the mental state just was a state of pain. If on the other hand we are interested in observing a physical object—say a tomato—we may be perfectly sure that the object as experienced was red and yet properly regard it as open to legitimate doubt whether the object in itself was red.

Our general conclusion about introspection then is that while expertise in it certainly requires much practice and some natural talent it is a far more competent instrument for the acquirement of empirical self-knowledge than is usually allowed and that there is no reason in principle why it should fail us in the broad use we make of it in these lectures.

5. I remarked a few moments ago that the mental states and processes revealed to us by introspection furnish valid evidence for conclusions about the self's nature. And indeed there is no other way of acquiring knowledge about the different aspects of the self's nature than by inference from observed facts of this sort. As Lotze has said ‘We have to conceive its [the self's] nature as it must be in order that it should pass through what we know in ourselves as its states and perform what we find in ourselves as its actions’. And he proceeds ‘Hence we must start from a comparison between mental phenomena; putting together the like and separating the unlike we shall sort the heterogeneous multitude into groups each of which includes all that have one common stamp and excludes whatever is of a divergent kind. Mental phenomena differ sufficiently among themselves to make it probable that this comparison if made steadily from one point of view will end in discovering several separate groups for whose peculiar distinctions no common expression can be found.… For the whole of each department of phenomena we must attribute to the soul a peculiar faculty to energise in that manner which predominates uniformly throughout all its component parts. Accordingly we must suppose the soul to possess as many separate faculties as there are groups of phenomena left unresolvable by observation; but we shall at the same time be left with the conviction that they are not imprinted in its nature as an unconnected assemblage of faculties but that there is between them an affinity by which as various manifestations of one and the same being they are harmonised into the whole of its rational development.’6

Despite (or as some may think because of) its somewhat old-fashioned flavour I should myself take little serious exception to this way of stating the situation; not even to the use of the term ‘faculty’—which after all means simply ‘power’ or ‘capacity’. Philosophers (who sometimes betray their common humanity paradoxically enough by curiously ovine propensities) have for a long time suffered from a sort of group-or flock-phobia about the use of this term. No doubt they are fearful lest they be identified with those earlier students of the mind who supposed that they had offered an explanation of some specific type of mental process when they had postulated in the mind a specific ‘faculty’ for its performance. But these earlier psychologists were not obviously wrong—indeed I should say they were obviously right—in holding that there are faculties; for after all whatever the mind does (as Locke said) there must be in it a ‘power’ of doing.

What is wrong—or rather what is dangerous—about ‘faculty’ talk is of course that one can so easily be beguiled by it into supposing that when one has ascribed a process to a ‘faculty’ one has ‘explained’ it. It is not so much false as just a pointless tautology to say of a man who adds 2 and 2 and gets 4 for the answer that he has a ‘faculty’ of addition. Of course he has in so far as ‘faculty’ means ‘power’. But the ‘explanation’ of the additive or any other cognitive process only begins when we are able to exhibit the power in question as a specific differentiation in a determinate set of circumstances of some more general cognitive power. In other words the ‘faculty’ explanation stops before explanation proper begins.

Yet it does not follow from this that there is no sense at all in talking about ‘faculties’. For ‘explanation’ has always certain limits; and it is in order to mark the limit of explanation that ascription to a faculty is in point. When in the course of explaining the less general by the more general we arrive at a mental process which appears incapable of being subsumed under any power more general than its own we may fitly signalise this ne plus ultra by saying that the process is the expression of an ultimate power or faculty: intending thereby of course no implication that we are ‘explaining’ anything. But just because ascription to faculties is proper and valuable only as marking the terminus ad quern of explanation it is best I think to confine the use of the word ‘faculty’ exclusively to ultimate and irreducible powers of the mind. Used in this way—and it is I believe Lotze's way—the reference of mental processes to faculties e.g. to faculties of cognition conation and feeling seems unexceptionable. And in point of fact a good many philosophers and psychologists who shudder at the very mention of the word and broadly hint that to believe in faculties is as intellectually disreputable as to believe in fairies show that they do not really take exception to the thing but only to the name by their own free use of terms like ‘power’ and ‘capacity’ which are fundamentally synonymous with the term they repudiate.

6. Now it is not to our purpose here to attempt any general inventory of the mind's ultimate powers or faculties. Broadly speaking there is virtual unanimity among psychologists about this: and certainly I have no original suggestions to offer. There is however one such power—a power which the observed facts apparently force us to postulate—with which we must I fear deal at some length. For much controversy has centred upon it. Moreover this power has a function in human experience almost as central as the basic powers of cognition conation and feeling to the effective exercise of which indeed it is quite indispensable. I refer to the mind's power of retaining within it in some form its past experiences and utilising them on receipt of appropriate stimuli in the course of its future experience.

The paramount importance for human life of this capacity is evident from the simple fact that in its absence there could be no such thing as ‘learning from experience’. Our world would be completely new for us at every moment. The development in the individual person of a body of knowledge of a character of sentiments could not even begin if cognitions conations and feelings passed out of existence as soon as they had occurred.

But note that mere ‘retention’ of past experiences is not enough. Past experiences might conceivably be retained in unconscious form in the mind but without their doing anything—in which case presumably we should not guess that they were retained. In fact we do know that they are retained (or most plausibly infer it) because we find we can give no intelligible account of the actual growth of an individual's experience without postulating that they are constantly at work exercising a determining influence upon the course of his future mental life in all its phases. The past experiences it would seem not merely survive but survive as tendencies or abilities of the mind to react on receipt of stimuli similar in character to the stimuli of the past experiences in certain ways which bear the unmistakable impress of these past experiences. The meaning that a perceived object has for us in adult life for instance we read into it in the light of past experiences of this and cognate objects even although these past objects have not persisted in the conscious mind during the interval since their occurrence and indeed are probably not even revived consciously in the mind now.

It will be apparent that what we are here about to deal with are these highly elusive ‘somethings’ called ‘dispositions’. Or to speak more accurately acquired dispositions; tendencies or abilities that are built up in the individual mind out of its past experiences. For these must be distinguished from congenital dispositions such as musical ability which there is good reason to suppose belong to the mind independently of its actual experience; though experience is of course necessary for their activation and development. To speak more accurately still we are concerned here with acquired mental dispositions; abilities and tendencies which prima facie at any rate arise out of mental experiences are excited into action by mental experiences and issue in mental experiences. That there are also congenital and acquired physiological dispositions bodily tendencies to specific patterns of bodily behaviour no one I think doubts.

In the little I have so far said about dispositions I have to some slight extent been mixing together fact and interpretation; e.g. in assuming the common-sense view that past experiences do ‘survive’ in some form. I hope that this may be excused in a preliminary exegesis whose object is no more than to indicate in general terms what it is that we are to discuss. But it is time now to disentangle fact and theory; for the true interpretation of the facts relating to so-called ‘dispositions’ has come in recent times to be a matter of lively debate.

7. The bare facts may for present purposes be outlined in a very few words.

1. Most and perhaps all conscious experiences do somehow exercise causal influence of one sort or another upon later experiences of the ‘same’ mind.

2. During the interval which may be very long indeed between the occurrence of a past experience and the occurrence of a later experience causally affected by it observation yields no direct evidence that the past experience is still ‘there’ in any form to exercise its influence. If there are ‘traces’ of some sort left in the mind by past experiences these traces are not apparently of the nature of conscious states; for the most patient introspection will not reveal them.

This is where the question of theory comes in. Do the past experiences perhaps persist in some form of unconscious traces filling the temporal gap between the past and the present experience and thus enabling us to interpret the situation without breach of the commonly accepted principle that the cause must be temporally continuous with its effect? Or is it perhaps the case that traces are not empirically observed because there just aren't any; in which case presumably the past experience achieves its effect by some kind of ‘action at a distance’—the ‘distance’ here being not spatial but temporal?

There would seem broadly speaking to have been three main theories put forward to account for the phenomena; one which does not assume traces and two which do. The one which does not we have just alluded to. It prefers the assumption of causal action at a (temporal) distance. Lord Russell calls it the Mnemic Causation theory. The two theories which assume that traces persist in the mind differ sharply according to their interpretation of the phrase ‘in the mind’. If we find it not unreasonable to believe that the mind has a relatively enduring structure we may conceive that traces are modifications of that structure brought about by past experiences and that through these traces which have thus become ingredient in the mind's structure the past experiences of a mind affect its future functioning in respect of relevant experiential situations. I propose to call this the Structure-Trace theory. On the other hand if we are determined to avoid any traffic with ‘substances’ we may hold that the traces persist not as modifications of the hypothetical structure of a hypothetical mind but as ingredients of actual mental events being passed on from the initial mental state to its successor in the series and thence to its successor and so on without intermission throughout the whole course of an individual's mental life. This I shall call the Event-Trace theory.

We have then (1) the Mnemic Causation theory; (2) the Structure-Trace theory; and (3) the Event-Trace theory.

Now so far as (1) and (3) are concerned a rather odd thing about them is that no one not even their sponsors really likes them very much. Formidable objections to them are freely admitted. Such appeal as they have they may be said to owe to an extrinsic consideration; the consideration namely that they would seem to be virtually the only theories open to us if we suppose ourselves obliged to reject the view that the mind is a substantival entity a relatively permanent being of which mental states are manifestations and adopt instead the ‘serial’ view according to which ‘mind’ is just a name for a succession of particular mental events inter-related in certain ways.

If on the other hand one happens to be satisfied with the substantival view of the self then the second of the three theories the theory that the mind's relatively permanent structure bears the traces left by past experiences seems obvious and natural. Once we accept the thesis that the mind is a relatively abiding entity not reducible to particular experiences there is no manifest objection in principle to conceiving this entity as having a structure which undergoes continual modification from its experiences. Nor would one expect modifications of structure to lie open to introspective observation as one might expect of traces supposed to be ingredient in actual mental states. And although admittedly one does not know how it seems by no means incredible that these modifications of the mind's structure should be such that in its future experiences the mind functions in a manner which manifests the traces left by its past experiences.

Since I myself regard some form of the substantival view as inescapable I do not find it easy to get greatly excited about theories of how we might interpret the influence of past upon later experiences on the assumption that the substantival view has to be repudiated. I should have been rather glad nevertheless had time permitted a closer examination of theories (1) and (3) for they seem to me to provide instructive illustration of the truly desperate straits to which those philosophers are reduced who conceive it their duty to try to account for the facts of human experience in terms of the Serial view. In the time at our disposal in these lectures however criticism is best reserved for doctrines which though in my view mistaken do have some plausibility and I shall make only the briefest of comments on the Mnemic Causation and Event-Trace theories.

8. It is common ground to all theories that if I now have a memory X of a childhood experience Y Y is part-cause of X. The peculiarity of the Mnemic Causation theory is to hold that Y does not exercise its causality through any chain of intermediaries (whether in the form of mental traces in or structural changes of my mind) which bridge the temporal gap between Y and X. Y has ceased to exist perhaps 50 years ago. Yet according to the theory this past Y somehow functions now as part-cause of my present memory X.

It is hardly surprising that this theory seems to have been found attractive by no philosopher except for a spell by Lord Russell. Perhaps the chief objection to it is that it requires us to accept a totally new kind of causality with no analogues elsewhere in human experience. Everywhere else in cases where the cause does not immediately precede its effect we find intermediaries bridging the temporal gap. ‘Mnemic Causation’ is a purely ad hoc invention which has no support from other regions of experience; and the only support it has from this the mental region of experience is negative—the alleged absurdity of alternative accounts which posit intermediary causes which are not empirically observable.

I should heartily agree with Russell however that one of the alternative accounts viz. the Event-Trace Theory is absurd. This is the theory that the initial state leaves traces of itself which are passed on successively from mental state to mental state through all the intermediate states between it and the specific state perhaps fifty years later which manifests its causal influence.

The merit of this theory is of course that it preserves temporal continuity between cause and effect. But at what a cost! It is hard to see how such a theory can be seriously held by anyone who has faced up to its implications. On either of the ‘Trace’ theories one has to suppose that all or almost all a man's conscious experiences leave traces of themselves and that these traces persist throughout the whole course of his life. On the Event-Trace theory it therefore follows that at any moment in the experience of an adult the particular experience he then enjoys is somehow the bearer of literally millions upon millions of these ‘traces’. Suppose my razor slips and I cut myself. All that I can find in my momentary experience is a feeling of sharp pain located in some part of my face. But according to the Event-Trace theory there must also be ingredient in that momentary experience in unconscious form the uncountable multitude of traces left in my mind by past experiences from my infancy onwards. It seems to me that if one can believe this one can believe anything. For notice it is not an experiencing subject but the momentary experience itself in which these traces must be supposed to reside. On the theory before us there is no enduring ‘subject’ to carry the traces. One can understand what is meant by the teeming subconscious depths of the Freudian Ego or even of the Leibnizian Monad. For these are substantival entities. But the teeming subconscious depths of a particular momentary mental state is another matter. Indeed I cannot help suspecting that a good many of those who profess some sympathy with the Event-Trace theory do really despite themselves think of the unconscious traces not as constituents of the particular mental events themselves but of an enduring subject in which these occur.

9. Let us now leave these highly artificial and barely credible theories about mental dispositions begotten by misplaced fidelity to the Serial view of the self and let us return for the few minutes that remain to the common sense doctrine which interprets dispositions in terms of the relatively permanent mental structure of the self. I can readily imagine objection being taken to the use of the term ‘structure’ in this context and I ought to say a word or two in its defence.

It must be freely admitted that ‘structure’ as applied to the mind is an analogical term derived from our experience of material things. We think of a material thing as having a structure if it is a whole of relatively permanent parts in relatively permanent relations to one another. Of the spatial relations of such parts we have generally very clear ideas. But there are of course other somewhat less clear types of physical relation e.g. dynamical relations which we also accept as contributing to the structure of the material thing. The structure of the atom could not be adequately described by indicating merely the spatial relationships of its parts; and much less could the structure of an organism. Now in applying the notion of ‘structure’ to the mind we frankly recognise that none of the physical relationships constitutive of ‘material’ structure obtain. But the application will be justified analogically if there is good reason to believe that the mind has relatively permanent ‘parts’ which stand in relatively permanent relations to one another.

And this it does seem reasonable to believe. We cannot to be sure perceive directly either the ‘parts’ of a mind or the interrelations of these parts. We perceive directly only the mind's manifestations. Knowledge of the mind's parts and relations can be got only by inference from the observed manifestations. But the inference seems fair enough once it has been accepted that there is an enduring subject. If from observation of these manifestations we find certain distinct and irreducible kinds of experiencing constantly recurrent e.g. conation cognition and feeling we may reasonably infer that there are different relatively permanent ‘parts’ or ‘organs’ of the mind in virtue of which the self strives cognises and feels. Again if we find that these basic experiences in which the self manifests itself stand in certain relatively permanent relationships to one another—as we surely do find else our experiences would be chaotic which they are not—we may reasonably infer that the parts of the self in virtue of which it so manifests itself stand also in orderly relationship to one another. No doubt the knowledge we so gain of the mind's parts and relationships is both indirect and highly indefinite. We can only say that they are what they must be in order that the mind's manifestations should be what we observe them to be. It is not to be denied that of much of the ultimate structure of the human mind we have vague ideas at best. But that the mind has a structure in the sense of being a whole of relatively permanent parts in relatively permanent relationships and that we can speak accordingly of ‘modifications of the mind's structure’ in the course of its experience seems a legitimate inference from direct observation of the mind's actual manifestations.

It will be clear then that in my view a disposition is a potentiality based on an actuality—the ‘actuality’ being the existent mental structure of the self as modified in a specific manner by the relevant past experiences of that self. This is I think in complete accord with what is ordinarily understood by the term ‘disposition’. When we say that a man has an ability to play bridge it is not the case that our meaning is exhausted by a set of hypothetical propositions stating how he is likely to respond if he is confronted with certain more or less determinate situations. Rightly or wrongly we mean also that as a result of his past experience his mind is now so constituted that if such situations confront him he will so respond.

What is it that is supposed to be wrong with this common-sense interpretation? Why is it rejected as naïve by so many present-day philosophers? The rock of offence is of course that it implies the substantival view of the self. It implies a mind which is a relatively permanent entity with a constitution of its own. The common-sense interpretation is unintelligible on the Serial view of the self which in one form or another is held to be forced upon us by a rigorously empirical approach to human experience.

I believe myself that a ‘rigorously empirical’ approach which if it is to be properly so described must take full account of the self's experience of itself in self-consciousness forces us to a very different conclusion. But concerning this enough has been said already. Rather I should like now (though it must be very briefly) to raise a question about the right of anyone who denies the substantival character of the mind to talk about ‘mental dispositions’ at all. I take it as agreed that a disposition is or involves some kind of potentiality—certainly it is very hard to talk about it intelligibly without using this or some similar term. What is it I want to ask that on the Serial view is supposed to ‘have’ this potentiality?

For surely something must ‘have’ it? A potentiality it seems clear cannot be itself merely potential. If it does not actually exist as a potentiality it is just another word for nothing. Again while actually existing as a potentiality it cannot exist in the void. A potentiality is a sheer abstraction unless thought of as characterising an actual someone or something. So much seems to myself beyond reasonable dispute. But in the case of mental dispositions what is the actual someone or something that has the potentiality? There is no problem here on the substantival theory of the self; but if we accept the Serial theory I can see no answer that does not make plain nonsense. Smith we say is ‘able’ to understand German. What is it that has this ability or potentiality? What conceivable meaning is there in speaking of a series of particular mental events inter-related as one will having the ability to understand German?—or for that matter having any other ability? If the series of particular mental events is all that there is to Smith's mind then that is all there is to it and there is no room for predicating of it an ability or potentiality.

Unless of course a potentiality is nothing actual but is itself merely potential. Is it unjust I wonder to suggest that the readiness to dispense with a mental substance to ‘carry’ mental dispositions is not unconnected with the fallacious belief that because the actualisation of a potentiality belongs to the future a potentiality itself has no actual existence now? If that belief were not a fallacy but valid there would be something to be said for an analysis of mental dispositions in terms solely of hypothetical propositions about future events. I must insist however that a potentiality that is not itself something actual is nothing at ail-not even a potentiality; that as an actual characteristic an actual location must be found for it; and that the Serialist being debarred from finding any actual location for it ought to abandon the notion altogether. And when the notion of potentiality goes so too as I see it must go any talk of ‘dispositions’. It is fortunate therefore that this disagreeable situation imposes itself only upon those who choose to deny the thesis the truth of which seems to me to be independently established that there are relatively permanent selves with relatively permanent mental structures.

10. One final point. The dispositions with which we have been concerned are dispositions which so far as the evidence of observation goes arise from mental experiences are excited into activity by mental experiences and issue in mental experiences. They are called therefore very properly ‘mental’ or ‘psychical’ dispositions and I think they must be accounted for primarily at least in terms of modifications of the mind's structure. But although it seems to me impossible to account for them as some have tried to do in terms of modifications of bodily structure alone nevertheless I by no means wish to rule out the hypothesis that our psychical dispositions (so-called) involve modifications of bodily structure also. It is I think highly probable that the experiences which give rise to modifications of mental structure invariably give rise also to corresponding modifications on the brain. After all it is never to be forgotten that in the empirical self (i.e. in the self qua manifested in human experience) mind and body are so intermixed ‘as to compose a certain unity’. If any one should care to contend that there is no such thing as a purely psychical disposition that all so-called psychical dispositions are at bottom psycho-physical I have no quarrel with him. On the contrary I suspect that he is right.

• 1. P. 7.
• 2. Space, Time, and Deity, Vol. I, p. 18.
• 3. Space, Time, and Deity, Vol. II, p. 89.
• 4. Quoted by William James, Principles of Psychology, Vol. I, p. 188.
• 5. Manual of Psychology, p. 45 (3rd edn.).
• 6. Microcosmus, Book II, Chap. II, pp. 168–9. (English translation by E. Hamilton and E. E. C. Jones.)