1. In this lecture I want to try to justify the rather large claims for the importance of the judgment-theory of cognition made at the close of my last lecture. To this end, I shall invite your attention to some examples of its impact upon controversies prominent in the recent history of philosophy. My concluding example will relate directly to the problem of the self, and this will naturally receive very much fuller discussion.

Let us begin with the problem of the External World.

The problem of how, if at all, we can philosophically justify belief in an External World has taken historically two main forms, which we may distinguish as the seventeenth century form and the twentieth century form respectively. The earlier calls in question the legitimacy of any transcendence by the mind of its own ‘ideas’; the later, the legitimacy of any transcendence by the mind of ‘sensa’ or ‘sense-data’. The problem in either form loses all meaning, I think, if one accepts the analysis of cognition involved in the judgment-theory.

According to the seventeenth century formulation, I directly apprehend in cognition my own ideas, and I directly apprehend nothing else. These ideas, although both ‘mental’ and ‘mine’, are not indeed, or not necessarily, states or processes of my mind, like, e.g. desiring and perceiving. (And it is worth noting, in parenthesis, that critics of the Representationalist epistemology have sometimes made things much too easy for themselves by forgetting this, as though the Representationalist were committed to the absurd doctrine that all cognition is a kind of introspection.) The ideas apprehended are mental ‘objects’, mental ‘contents’, not mental states or processes. It remains true, however, that for this view they are mental, and they are mine; and that is sufficient to raise an acute problem as to how, if at all, I can legitimately pass beyond them to the affirmation of an external world, a world existing independently of my mind and its ideas.

Not many words are needed to show that, for the judgment-theory of cognition which I have been advocating, the problem has no existence. For on this view, cognition does not start with apprehension of ideas or ideal contents. It is from the beginning, and it continues to be, a process of ideally characterising an independent objective reality. We have distinguished on the judgment-theory, it is true, something that we called ‘ideal content’. But this ideal content is not, as it is for the Representationalist, an ‘object’ of cognition. It is just the meaning or complex of meanings which we predicate of, and by which we seek to characterise, the objective reality. Accordingly, if all cognition involves judgment, an objective reality ‘beyond’ or independent of our ideas is presupposed in all cognition. And it clearly does not make sense to treat as problematic, and in need of independent justification, something which in all our cognitions we presuppose.

The problem takes a somewhat different shape, but is resolvable in similar fashion, within the more modern ‘sense-datum’ frame-work. According to the typical sense-datum view, sensing gives us direct acquaintance with particular existents like ‘this red’ or ‘this sweet’; and we directly apprehend nothing else. But, unlike the Representationalist, the believer in sense-data is not committed to holding that these direct objects are either mental or private in character. They may be or they may not be. That is one of the problems that most actively engage his attention. Hence the question that typically arises for him is not exactly that of the reality of an external world; since sensa, if interpreted (with whatever difficulty) as non-mental and public entities, may intelligibly enough be said themselves to constitute an ‘external world’. His typical problem is rather that of the existence of the kind of external world believed in by ‘common-sense’; i.e. an external world which is something more than just sensa in various relationships and groupings, an external world which consists of ‘things’ that are in some way other than the sensory qualia which we normally take to ‘belong’ to them. Are we entitled to affirm the existence of ‘things’ which are not reducible to ‘families’ of actual and possible sensa, or is the so-called ‘physical thing’ merely a logical construction out of sensa and exhaustively definable in terms of sensa? Such in rather broad terms, seems to me to be the problem about the ‘external world’ that is most significant from the standpoint of sense-datum theory; the problem to which, e.g. (until its defects became too palpable) Phenomenalism was widely offered as the best solution.

But for us this problem too is just a pseudo-problem which arises out of, and gets all its point from, a false analysis of the fundamental nature of cognition. If our judgment-theory be valid, then the ‘sensum’, understood as a direct object of cognition, is sheer myth. There is no distinguishable cognitive activity of sensing of which the sensum is the direct object. The so-called ‘sensum’, like the ‘idea’ of the subjectivist, in so far as it is anything for, or enters in any way into, the life of cognition, is not an object, but just an ideal content or meaning by which we characterise objective reality in judgment. The perceived ‘green’, if it really is cognised as green, is a determinate shade of greenness affirmed as characterising a certain spatial expanse in the objective reality. It is an element intellectually distinguishable within the unity of the judgment, not an entity per se. No question can legitimately arise, therefore, as to whether there is or is not an external reality other than the so-called sensa in their several groupings. Such a question rests upon the hypostatisation of an abstraction: upon the imputing of an independent existential status to that which is, for cognition, always and only a character affirmed of objective reality in judgment. The acceptance of a real world ‘beyond’ so-called sensa is thus, on the judgment-theory of cognition, implied in all cognition, and cannot intelligibly be called in question.

What is perhaps most commonly regarded in philosophy today as the basic problem of perception—the problem of how we ought to understand the relationship between sensory qualia and the physical thing to which we normally take them to ‘belong’—is not, of course, the same as the problem we have just been discussing. Nothing that has been said above is in any way intended to deny that genuine problems of perception remain. It is intended to deny only the propriety of formulating and discussing these problems in certain very familiar terms. So long as the problems are stated in terms of the relation to physical things of sensa which are supposed to be themselves distinct existences, their discussion is, in my view, hopelessly bedevilled from the start. It cannot be a ‘real’ problem to discover how an existent X is related to some other existent, if in fact X is not an existent at all, but is from the beginning and throughout apprehended as, and only as, characterising an existent. Hence any problem so posed is at bottom a pseudo-problem. Discussion of problems of perception in these terms may yield—indeed it clearly has yielded in the pages of philosophers like H. H. Price and A. J. Ayer—many by-products of an illuminating kind. But that is the most that can be expected. For the problems in the form in which they are stated just do not exist, and therefore cannot be ‘solved’.

2. It will be appropriate at this juncture to say a word or two about the impact of the judgment-theory upon the Verifiability theory of meaning. For I think that this theory, though it may not strictly depend upon, nevertheless owes a good deal of its appeal to, the belief that sensa, as directly apprehended in a cognitive process of sensing, have a special place of privilege and priority in the realm of knowledge.

The Verifiability Principle, as is well known, has undergone very substantial modifications since its first crude expression in the startling dictum that ‘the meaning of a statement is the method of its verification’. And I am not going to flog dead horses by recapitulating the criticisms to which the earlier formulations were subjected. It can hardly, however, without losing its distinctive character altogether, retreat much beyond the point it has now reached. In so far as it is still a power in the land (and it is not altogether easy to judge how far that is the case) it can, I think, be formulated in some such terms as these—‘a statement has (cognitive) meaning only if it is possible, at least in principle if not in fact, to point to sensory observations which are relevant to the determination of its truth or falsehood.’ In this later formulation the principle is offered merely as a criterion, and not also as a definition, of meaning, and is at least not now (as might be argued of its earliest formulation) logically indefensible.

Even in so comparatively moderate a formulation, however, the Verifiability Principle must be uncompromisingly rejected if our judgment theory of cognition is valid. If cognition involves judgment, it involves, as we have seen, reference to an objective reality existing independently of the sensory qualia by which in judgment we attempt to characterise it. Now it can hardly be contended that there is ‘no meaning’ in the statement that there is an objective reality independent of these sensory qualia, if in fact every cognition presupposes that there is! Yet this statement is surely just the kind of ‘metaphysical’ statement about the supra-sensible which we are commanded in the name of the Verifiability Principle to abjure as meaningless. For no sensory observation can be pointed to which will be evidence towards either its confirmation or its rejection. On the judgment-theory of cognition the evidence in confirmation of this ‘metaphysical’ statement is simply the fact that we cannot think at all without accepting what the statement asserts. Can there be better ‘evidence’ for anything? I suggest that so far from its being the case that metaphysical statements are pseudo-statements, the real state of affairs is that the much canvassed ‘problem’ whether metaphysical statements can be meaningful is a pseudo-problem. It arises from the failure to notice that cognition involves elements of metaphysical meaning from the beginning.

One ought perhaps to remark in passing that although no further notable modification has occurred, or indeed is easily conceivable, in the content of the Verifiability Principle, there has occurred a modification which amounts almost to a revolution in the status which, on the whole, its champions now claim for it. It is now pretty generally agreed among those who still think the principle of some value that it is not a necessary truth, to which the philosopher is bound to submit, but should rather be regarded as a useful methodological postulate by which all philosophers will do well to be guided if they wish to philosophise with profit. But the reply to this, in the light of the criticisms we have just passed, is obvious. It can hardly be useful for philosophy—however useful it may be for other pursuits not concerned with ultimate truth—to proceed upon a principle whose falsity is implied in every cognition.

3. Another vexatious problem from which we are happily delivered by the judgment-theory is the so-called ‘Problem of Error’; which may be taken here as including the problem of ‘unreal objects’. This problem has been a thorn in the very vitals of all Realist epistemologies. Modern Realism arose primarily as a reaction against the vast claims made by Idealism for the importance of ‘mind’ in the scheme of things, and in particular against what was believed to be the Idealist insistence upon the mind-dependent character of the ‘object’ of cognition. It is the very first article of the Realist creed that the object of cognition exists independently of the cognising mind. As the Realist fully appreciated, however, this epistemological premise at once raises troublesome questions about the object of erroneous cognition. If we believe that Charles I died in his bed, what we apprehend—our ‘object’—cannot be said without most violent paradox to have real existence in any ordinary sense; and yet, according to the Realist premise, it must apparently have some kind of objective reality assigned to it. What kind? So too with cognised unicorns, centaurs, mirage lakes, the present King of France, and the like. As objects of cognition they have to be assigned to objective reality; but as unreal objects they cannot be assigned to objective reality in any ordinary sense. The puzzle is to find an extraordinary sense in which the curious notion of ‘unreal reality’ has intelligible meaning.

In a manner these matters are now vieux jeux; and I shall not weary you by describing the truly desperate intellectual contortions of the Realist philosophers, in the U.S.A. more particularly, as they struggled to invent some plausible account of error while still holding fast to their epistemological premise. More than any other factor it was the difficulties over error that brought about the early disintegration of the very vigorous and able school of American philosophers who assumed for themselves the title of ‘New Realists’. All that I want to do here is to draw attention to, and comment briefly upon, two points. The first is that on the judgment-theory of cognition no epistemological problem of error arises. The second is that the prima facie plausibility of the Realist's premise concerning the independence of the object cognised arises from a confusion between two senses of the term ‘object’; there is, as we shall see, a sense of ‘object’ in which the Realist premise does hold good; but in that sense our judgment-theory fully acknowledges its validity.

As regards the first point. What gives rise to the so-called problem, we saw, was the Realist assumption that unicorns, the proposition that Charles I died in his bed, and the rest, in so far as, and just because, they are objects of cognition, must have some sort of independent existence. But according to our judgment-analysis of cognition, the cognitive situation which the Realist postulates—of an apprehending mind on the one side and an object apprehended on the other—is a misleading simplification of what is in fact a much more complex state of affairs. Take the false proposition ‘Charles I died in his bed’. On the judgment-theory, to cognise this proposition is to assert (mentally) that the ideal complex of terms-in-relation which is what we mean by ‘Charles I having died in his bed’ characterises the objective reality. Thus the proposition, as cognised, is not an object over against the mind. It is just the ideal content or meaning which the judging mind predicates of reality. As such it has, no more than ‘ideas’ or ‘sensa’, an existence independent of the cognising mind—which was the Realist postulate that gave rise to the whole problem. Considered by itself, the ‘proposition’ is a mere abstraction: and the only being which an abstraction has is as a constituent of that from which it has been abstracted—in this case the judgment.

Mutatis mutandis, the same is in principle true, I think, of all types of ‘unreal object’. There is no way of cognising them except in judgment, and the so-called ‘object’ turns out, on analysis, to be an ideal content, or complex of meanings, by which the judging mind attempts to characterise objective reality. It is not an independent entity, and there is no problem, accordingly, about the sort of ‘unreal reality’ to be assigned to it.

I pass on to my second point. I strongly suspect that most of the troubles that afflict the Realist epistemology flow from an initial confusion between two senses of the term ‘object’. In insisting upon the independent reality of the object of cognition, the Realist is insisting upon what is, in one sense of the term ‘object’, a perfectly genuine postulate of all cognition. The cognitive process would be self-stultifying if the character of the object which we seek to know were affected by the process of knowing it. But everything turns here upon what we understand by the ‘object’ that is thus independent of the cognising mind. It is crucial to recognise the distinction between the ‘object’ as that which we are seeking to know—the object ‘as it really is’—and the object in its character as actually cognised by us. The former we do, and must, think of in all cognition as having its being independently of the cognising mind. And our judgment-theory of cognition takes due account of this through its contention that what we aim at characterising by our ideal content or meaning in judgment is always the independent objective reality. But the ‘object’ in the latter sense, the object in its character as actually cognised, is a totally different matter. The object as actually cognised is the object in the character we ascribe to it in the judgment. Why on earth should we suppose that the object as thus characterised by us is independent of the cognising mind? To suppose that the mind-independence of the ‘object’ in this sense is a postulate of cognition seems really to be nonsense.

Once this confusion between the two senses of ‘object’ is resolved, and we recognise that it is not the object in its character as cognised whose independence of the cognising mind is presupposed in the cognitive situation, the Realist's difficulties over ‘false objects’ disappear. For these objects are merely ‘objects-as-cognised’, and thus do not have the independent existence which is for the Realist the cause of all the trouble.

4. Perhaps I may be permitted to linger a moment or two longer over this confusion between the two senses of the ‘object’ of cognition; for it seems to me to be very fundamental, and to have had consequences as unhappy as they have been extensive. I have very little doubt that the main attraction of the twentieth-century Realist revolt against the Idealist epistemology lay for most people in the belief that the Realist, in insisting upon the mind-independence of the object of cognition, was introducing a much-needed rectification of a radical blunder committed by the Idealist. And this, I respectfully submit, is a delusion. The Idealist fully acknowledges the object's independence in the only sense in which that is a postulate of cognition. When I cognise this ink-pot, certainly I postulate in my cognising that there is something there to be cognised independently of my cognising it. That is every bit as true for Idealist as for Realist epistemology. But for the former what is postulated as existing independently is the object in its character as I seek to cognise it, not in its character as I actually do cognise it. The Realist differs from the Idealist in that he apparently wants to hold it to be a postulate of cognition that the ink-pot in its character as we do cognise it is independent of the cognising mind. For the acceptance of that postulate I can see no justification whatever. On the other hand it has, I think, a fairly obvious psychological explanation in the confusion between two different senses of the term ‘object’.

It would not be difficult to supplement these samples of much-debated philosophical problems which lose all point if the judgment-theory of cognition should happen to be true. I venture to hope, however, that enough has been said to make good my claim that this theory has philosophical implications of the most fundamental kind. If this be granted, I think it ought also to be granted that there is an obligation upon philosophers to give some study to the judgment-theory before deciding to dismiss it.

5. I turn now to a philosophical implication of the judgment-theory of a rather different kind, one which bears much more directly upon the central problem of the present course. I want to consider for the remainder of the lecture the implication of the judgment-theory in respect of the cognising subject.

The judgment-theory of cognition finds one of its sharpest contrasts with rival theories in the emphasis it places upon the activity aspect of cognition; and in thus emphasising the activity aspect it at the same time draws attention to something which is apt to become obscured in Empiricist accounts of cognition, viz. that there can be no cognition apart from a cognising subject.

On Empiricist theories of cognition, for which the dominant rôle is played by sense-experience, the cognising subject tends, at any rate in the earlier phases of cognition, to be no more than a passive spectator of what is ‘given’ in sense. Now it is difficult in any field of operations to overlook the presence of an active player; but not of a passive spectator. And the spectator is the more readily overlooked where, as is the case when we perceive simple sensory qualia like ‘red’, the ‘spectating’ aspect is by no means easy to distinguish and hold clearly before the mind. Hence a transition is often unconsciously made from treating the sensory cognition as the passive reception of a sensum by a subject mind, to treating it as just the occurrence of a sensum. The recipient, the subject mind, tends to drop out of the picture altogether.

Even on Empiricist premises, however, a subject mind is much less easily dispensed with when one passes from the more elementary to the more advanced phases of cognition. Here what at least looks like an activity aspect forces itself upon one's attention; e.g. in such functions as comparing, relating and inferring. But, though difficult, the task of eliminating the ‘thinker’ has not even here proved beyond the Empiricist's ingenuity. It has to be remembered that the Empiricist subscribes to a basic principle of methodology which strongly predisposes him to seek for an account of cognition free from any reference to a subject mind; the principle, namely, that nothing should be accepted as an authentic reality which is not an object of experience. Since the experiencing subject cannot, as such, be an object of experience, the Empiricist thinker starts with a powerful initial bias in its disfavour.

The classic example of an Empiricism which keeps stubbornly to its aim of excluding from its theory of cognition all such ‘unverifiable’ entities as subject minds is to be found, of course, in the philosophy of David Hume. For Locke, and still more for Berkeley, the activity aspect of cognition in one shape or another seemed inexpugnable, and neither of these philosophers, in consequence, shows any inclination to deny the existence of a subject. But to Hume, in at least one prominent strand in his thinking, it appeared that the Law of Association of Ideas could plausibly be called upon to perform the duties for which less rigorous Empiricists supposed that they must invoke a subject mind. The attempt is boldly made to depict the whole cognitive life of man in terms of particular mental events—impressions and ideas—which enter into relations with one another in accordance with this impersonal, external ‘Law of Association’—the ‘gentle force of attraction’ which was the counterpart in Hume's ‘statics and dynamics of the mind’ of Newton's ‘Law of Gravitation’ in the sphere of matter.¹ The individual mind becomes for Hume the mere theatre in which a set of particular mental events happen. And the spatial metaphor ‘theatre’ is not, of course, meant by Hume to be taken too seriously, as though it denoted some reality additional to what goes on ‘in’ it. It signifies no more, I think, than Hume's recognition that the happenings in what, in common parlance, is called a ‘mind’ are inter-related in some way which constitutes them a distinctive kind of unity. When we get behind the metaphor, the reality of the individual mind consists in nothing more than the reality of particular mental events and groups of mental events in their distinctive unity.

The judgment-theory, on the other hand, construes cognition as essentially an active process, because judgment is essentially an active process. This activity-aspect of judgment can perhaps best be brought out by contrasting the connection of ideas as we experience it in a case of mere association, with the connection of ideas as we experience it in a judgment. Consider, e.g. our experience when the perception of a frozen pond excites in our mind the image of one's self skating on its surface, as contrasted with our experience when the same perception is the occasion of our framing the judgment that the pond is ‘bearing’. In the former, the connection is experienced as something that happens in or to our mind. We do nothing about it. The image of one's self skating just ‘rises up’ in our consciousness. But in the latter, in the judgment, we are aware, however dimly, that we are doing something about it. The idea represented by the term ‘bearing’ is not experienced as something that just ‘happens’ in our mind following upon our perception of the frozen pond. In a real sense it is we who make it happen in our mind, by our active cognitive interest in, and consequent attention to, the character of the pond.

6. But at this point it would be wise, I think, to indulge in a partial digression. The emphasis I have been laying upon the active aspect of cognition, coupled with my earlier refusal to allow existence to so-called ‘sense-data’ or ‘sensa’, may have created a suspicion that I want to deny to ‘mere sensation’ any effective part in the cognitive life. I must make it clear that such is not in fact my view. On the contrary, I take sensation to be, though not itself a mode of cognition, a vital determining factor in cognition. Let me elaborate this point a little.

Although the mind is never, I think, in any judgment, a mere recipient of ideas; although it is always, in some degree, active in initiating the changes which characterise its cognitive field when a judgment is made; nevertheless, it is not to be denied that in some judgments there is an aspect of passivity which is a great deal more striking than the aspect of activity. I have in mind especially those elementary sensory judgments in which we affirm that some simple sensory quale—a colour, a sound, or a smell—characterises a part of our physical environment. On these occasions, while it is doubtless true that no judgment would take place but for the initiating activity of the mind, it is also true that, if the mind does judge at all, it feels compelled to judge in a certain determinate way. And this compulsion is felt as coming from a source external to the judging mind. There is, of course, a sense in which in all judgment the mind is conscious of a compulsion to judge in one way rather than another; for there is an obligation to observe logical consistency acknowledged, implicitly or explicitly, in all our judging. But this ‘logical’ compulsiveness is, as it were, internal to the mind, and is in that respect quite different from the special kind of compulsiveness of which we are conscious in simple sensory judgments. If a coloured patch is presented to me and I am asked what colour I see, I normally feel a compulsion which is not in the least like a logical compulsion to judge that the patch has this and not that determinate colour. And if I am asked what is the source of this compulsion, I do not see what answer I can well offer except just ‘the given in sensation’ or some similar expression. It seems to me, therefore, that not only must we recognise an element of passivity, as well as of activity, in many judgments, but that this element is most naturally identified with the function in experience of ‘sensation’.

But we must, I think, pursue the matter a stage or two further. For on this question of the rôle of sensation in cognition the school of epistemologists with which I am in general well content to associate myself, the idealist school which insists upon the centrality of the judgment in cognition, does seem to me to have laid itself open to criticism. They have, in point of fact, very little of a direct character to say on the question—which is in itself a ground for complaint, since the question is obviously important. But the language they tend to use about sensory cognition implies the ascription to the intellect of functions which, in the case of basic sensory cognitions at any rate, can hardly be defended. They are ready enough to grant, as a rule, that in sensory cognition the activity from the side of the mind is not the whole story. There must be, they acknowledge, some kind of a ‘datum’. But this datum, apparently, is for them something upon which thought ‘gets to work’ in the way of interpretation. So much seems implied by their frequent assertion that there is always ‘interpretation’ as well as ‘datum’, and that the object of sensory cognition is always an ‘interpreted datum’.

But this formula, ‘an interpreted datum’, surely signifies a gross over-estimate of the part played by the intellect in basic sensory cognitions? Without doubt, in the higher reaches of sense-perception much goes on that can very properly be called ‘interpreting data’. Presented with a small, brown, smooth, oval shape, we interpret these data as ‘egg’. On the basis of our past experience we can say to ourselves, ‘these signify “egg.”’ But where is the interpretation, and what is it that is interpreted, in a basic sensory cognition where a blue patch, say, is first cognised as blue? Here there is no passage, such as ‘interpretation’ involves, from that which is initially apprehended to the interpretation of it. There are no initially apprehended characteristics to be interpreted. What is present, I have suggested, is a felt extra-logical compulsion to affirm blueness of the patch: but this is as different as well can be from an act of ‘interpreting’ the patch as blue.

Our basic sensory cognition, then, though a judgment, is very misleadingly described when it is said to involve ‘interpreting a datum’. The determinate content the judgment affirms is due not to interpretation by the intellect but to the extra-logical compulsion exerted by pure sensation or pure sensing.

Can we say any more by way of characterising this elusive ‘sensation’, which is not a mode of cognition, and is yet a basis for and determinant of cognition? I think we can say a little, but not very much. I think we must at least say that it is something which falls within the general field of ‘experience’. It is not a physical process in the brain (or elsewhere)—though we are able, on what appear very good grounds, to correlate it with such processes. When we suppose ourselves compelled by ‘sensation’ to judge in certain ways, we do not suppose that it is some brain process that so compels us, and we may not even be aware that a brain process is going on. But while it seems proper to regard ‘pure’ sensing as, in its intrinsic nature, an experience, it is an experience for which the distinction of subject from object does not exist; an experience in which sensing and sensed are in indivisible unity. It thus falls into the general category of immediate experience, or feeling. I am inclined to believe that this is the least misleading way to regard ‘pure’ sensation—viz. as a specific sort of immediate experience—and I doubt if we can profitably characterise it much further.

7. In any event, it is rather more than time that we took up again the main thread of our argument. I had been pointing out how the judgment theory of cognition differs from other theories in its emphasis upon the activity aspect. Now the implication of that emphasis is that the cognising subject is thrown into sharp relief. Subscription to the judgment theory relieves us of any temptation to look upon the individual mind as just a sort of mise en scène for particular cognitive events. ‘Activity’ implies a subject that is active. And—though some philosophers have at least talked as though they thought otherwise—that which is active in activity cannot possibly be the activity itself. To suppose that activities can themselves be active is surely to commit an outsize example of what Professor Ryle would call a ‘category mistake’, and of what Professor Moore a generation back would have called, more picturesquely if less precisely, a plain ‘howler’.

The impossibility of dispensing with a cognitive subject distinguishable from, though not of course separable from, particular cognitions, seems to me, then, to be one of the plain implications of the judgment-theory. At the same time I should not wish to leave the impression that, in my belief, the need for positing a cognitive subject vanishes if one does not happen to be persuaded that the judgment-theory of cognition is the true one. That theory, by the stress it lays on the activity aspect, merely high-lights the absurdity of trying to get along without a cognitive subject. But the absurdity remains, I should contend, though it may be less conspicuous, on any theory of cognition; always provided that, unlike Behaviouristic and quasi-Behaviouristic theories, it really is a theory of cognition. There is in fact a simple enough argument (already hinted at) which seems to me conclusive against all attempts to exorcise from cognition a subject distinct from the actual cognising. It is as follows. Let X be the psychical operation of cognising or apprehending, and Y be the object that is cognised or apprehended. Now presumably the apprehended implies an apprehends of some sort. What then is the apprehends of Y? According to those who deny a distinguishable subject, it can only be X, the actual operation of apprehending. But this is surely absurd. An apprehending cannot be that which apprehends. What is ‘known’ cannot be known to the operation of knowing. It can be known only to a subject which, while engaged in the knowing, is not itself identical with the knowing. As Professor Bowman observes in concluding his own admirable statement of the argument, ‘To represent a mental process as at once a psychical performance and a performer is out of the question.’²

It is hardly surprising, therefore, that even those theories for which cognition is never more than a mere passive awareness are not really able to proceed far in their epistemology without having recourse to language which tacitly presupposes the cognising subject which their official conclusions repudiate. One need only think of the intolerable paradoxes in, e.g. Hume's explanation of how the ‘bundle of perceptions’ that is the individual mind comes to frame the ‘illusion’ that it is something other than a mere bundle. It is odd how little Hume's modern disciples seem to have learnt from their master's own revealing confession of failure in the well-known Appendix to the Treatise, where he comes next-door to acknowledging the total shipwreck of his atomistic interpretation of the human mind. An atomic mind, he now strongly suspects, just could not perform those reflections upon its own experiences which (as he earlier supposed) delude it into supposing that it is not atomic. Dr. F. R. Tennant has well said that ‘no one ever has really dispensed with the subject of consciousness, whatever terms he may have used to hush up its existence.’³

But it is still a far cry, of course, from the vindication of the ‘cognitive subject’ to the vindication of something that might reasonably claim to be called a ‘self’. In the next Lecture I hope to make more appreciable headway towards this goal by examining the nature of the cognitive subject from the standpoint of its own experience of itself. Our topic will, in short, be what is commonly known as ‘self-consciousness’.