Our introductory study had a twofold purpose; to determine objectively the problem for contemporary philosophy and to discover how we must set about the task of solving it. The conclusion we have reached is that our problem is the form of the personal and that we may hope to resolve it only by starting from the primacy of the practical. For we have seen that it is the assumption of the primacy of the theoretical in our philosophical tradition which institutes a formal dualism which cannot be resolved; that the basic form of this dualism is the division of experience into theoretical and practical and that this dualism makes it impossible to think the unity of the Self and so to determine the form of a personal experience. We have therefore to begin by rejecting dualism through asserting the primacy of the practical.

What is here proposed is that we should substitute the ‘I do’ for the ‘I think’ as our starting-point and centre of reference; and do our thinking from the standpoint of action. Clearly we must ask at once what this can mean and what are its implications? It may assist us in our attempts to answer this question if we notice that we cannot ask in advance whether this is possible. That can only be discovered in the attempt. For any reasoned objection to its possibility would presuppose the primacy of the theoretical and would therefore be invalid. I emphasize this at the beginning because it provides the form of the answer to innumerable objections which must arise in our minds as we proceed. We must not underestimate the difficulty of the enterprise to which we are committed. We have to shift the centre of gravity in our philosophical tradition and to alter our established mode of thinking. To propose this is easy; to accomplish it is so difficult that complete success at the first attempt is inconceivable. We are largely creatures of habit; not least in our reflective activities. To change our standpoint is to transform our habits of thought. It is not to exchange one theory for another but to change the basis of all theory. To achieve this must it seems to me be a long co-operative process; a stumbling advance in country where there are no beaten paths to follow and where every step may lead us astray. No man however great his intellectual powers might be could trace all the ramifications of the process of readjustment or avoid the tendency to relapse unconsciously into the familiar standpoint which he is seeking to overcome. I certainly have no confidence in my own ability to succeed; and I should relinquish the attempt unless I were convinced that we have no option but to try; and to hope that our failures will prove stepping-stones to final success.

The proposal to start from the primacy of the practical does not mean that we should aim at a practical rather than a theoretical philosophy. It may indeed have the effect of concentrating our attention upon action as the primacy of the theoretical tends to concentrate attention upon the problems of knowledge. What it does mean is that we should think from the standpoint of action. Philosophy is necessarily theoretical and must aim at a theoretical strictness. It does not follow that we must theorize from the standpoint of theory. Kant one may recall compared his change of standpoint to the Copernican revolution in astronomy. The present proposal might use the same analogy. Copernicus proposed to conceive the planetary system from the standpoint of an observer on the sun. This might seem at first sight impossible since the astronomer is necessarily upon the earth. Yet the objection is unfounded. Are we as philosophers in any worse case? We are indeed tied to an activity of reflection; but why should we be unable to reflect from the standpoint of action? Granted that in reflective activity the Self is subject we need not conclude that we are debarred from thinking as though we were in action and so from the standpoint of the Self as agent. This would indeed be impossible if there were an unbridgeable dualism between theory and practice; and from the standpoint of the ‘Cogito’ there is. Since this is our habitual standpoint in philosophical reflection we are apt to think that the practical standpoint excludes the theoretical as the theoretical excludes the practical; or more naturally that we have a choice between a ‘mentalist’ and a ‘materialist’ systematic in theory; that we must either be ‘realists’ or ‘idealists’ and that which of the two we choose need make no difference in practice. But this reveals that we are still thinking from the theoretical standpoint. If we substitute for this the practical standpoint the dualism between theory and practice disappears. Why this should be so is the first issue that requires clarification.

The Self that reflects and the Self that acts is the same Self; action and thought are contrasted modes of its activity. But it does not follow that they have an equal status in the being of the Self. In thinking the mind alone is active. In acting the body indeed is active but also the mind. Action is not blind. When we turn from reflection to action we do not turn from consciousness to unconsciousness. When we act sense perception and judgement are in continuous activity along with physical movement. When we think we exclude overt bodily movement at least; what more we exclude depends upon the denotation we choose to give to the term ‘thought’ which in its usage is highly ambiguous. But perhaps we may say that the ‘purer’ our thought becomes the more it excludes not merely perception but all sensuous elements and moves in a shadowy world of abstract and general ideas. Action then is a full concrete activity of the self in which all our capacities are employed; while thought is constituted by the exclusion of some of our powers and a withdrawal into an activity which is less concrete and less complete. Indeed when we consider the contrast in this fashion it tends to present itself as an abstract duality; in which action and thought are the positive and negative poles of a personal experience which moves in its actuality between them. In a somewhat analogous fashion black and white present themselves as absolute contraries though in reality they are the ideal limits of a series of grey tones. In the case of visual experience too one of the limits is positive and the other negative. Both are ideal limits; but while pure white is the complete fulness of light black is its complete exclusion so that the darker the tone the nearer we are to an exclusion of light and so to the exclusion of the possibility of vision. A man born blind does not see only black. He does not see at all. It may indeed be said that this does not make white any more positive than black since pure white means the total exclusion of darkness. This is true in the abstract. Yet in fact the exclusion of darkness is merely a double negative. It means the exclusion of the exclusion of light.

‘Acting’ and ‘thinking’ then are in abstract conception exclusive contraries. In actuality they are ideal limits of personal experience; and ‘acting’ is the positive while ‘thinking’ is the negative limit. This is what is expressed by the assertion of the primacy of the practical or as Kant expressed it the primacy of practical reason. The inclusion of the term ‘reason’ adds nothing essential. For what we are considering is personal experience or the activity of the Self; and reason is the differentia of the personal. The concept of ‘action’ is inclusive. As an ideal limit of personal being it is the concept of an unlimited rational being in which all the capacities of the Self are in full and unrestricted employment. As limited and finite persons such a fulness of positive being lies beyond our range. This does not affect the concept but only its application to the particular case. The limitation marks the fact that we are never fully active without restriction or qualification in our experience as agents. The important issue which is not an empirical one is that ‘action’ without thought is a self-contradictory conception. We do indeed talk of ‘acting without thinking’; but this again refers to empirical instances. It means that we acted without considering the relevant issues in a particular situation that we acted without taking proper thought. It cannot mean that we acted with no thought at all; still less that we acted without ever having thought about anything. For only a thinking being could act without thinking. ‘Thought’ on the other hand is an exclusive concept and therefore negative. As an ideal limit—as ‘pure’ thought—it denotes an activity of the Self which is purely formal and completely without content. Now the purely formal is equivalent to nothing; for there cannot be a form which is not the form of something and a purely formal activity is therefore an activity which is no activity. This no doubt is why we tend to say that when we are thinking we are doing nothing. ‘Pure thought’ is not merely impossible for us it is impossible in the nature of things. It is not therefore as a concept self-contradictory but merely secondary derivative and negative; a concept formed by exclusion and therefore relative to what is excluded. It is a necessary concept since negation is necessary to thought. So in pure mathematics zero is a number though it represents nothing. But the analogy of vision may help us better. We can report our experience of a pitch-black night either by saying that we see an unlimited darkness or that we see nothing or that we cannot see; and these three forms of statement are equivalent. But the first two are relative to our capacity to see and to our experience of seeing. If we had been born blind only the third would be appropriate; but it would also be without meaning for us.

We should note here if only to avoid possible misunderstanding as we proceed that this implies a definition of ‘action’. The term is loosely used and can give rise to serious ambiguity. We talk of the actions of animals; we even refer to the action of an acid upon a metal. When we use the term in this way we are employing an anthropomorphic metaphor. There is no objection to this for such looseness is essential to the proper functioning of language. But we should be aware of what we are doing especially when we are engaged upon an abstract inquiry. In the strict sense of the term only a person can ‘act’ or in the proper sense ‘do’ anything. If this is not agreed I shall not argue the point. It is sufficient for the present purpose to say that in this inquiry we shall use the term ‘action’ and the other terms which are its derivatives and synonyms in this sense. We are concerned with personal activities with the agency of the Self. In this context action and thought both imply rationality. We may however use the term ‘activity’ without this implication as a generic term with a wider significance so that we can distinguish both thought and action as modes of rational activity. As a further aid to definition we may add that action is activity in terms of the distinction between ‘right’ and ‘wrong’ and that thought is activity in terms of the distinction between ‘true’ and ‘false’.

With these points clear we can go a step farther. If the concept of ‘pure thought’ is derived from the concept of action by exclusion then thought so far as it is actual falls within action and depends upon action. Action is primary and concrete thought is secondary abstract and derivative. This must mean that the distinction between ‘right’ and ‘wrong’ which is constitutive for action is the primary standard of validity; while the distinction between ‘true’ and ‘false’ is secondary. In some sense though not necessarily directly it must be possible to distinguish between ‘right’ and ‘wrong’ before distinguishing between ‘true’ and ‘false’ and so without reference to the truth or falsity of a judgement and to derive the latter from the former. In other words a theory of knowledge1 presupposes and must be derived from and included within a theory of action.

We may now formulate our starting-point more clearly. We have to substitute for the ‘I think’ as our centre of reference the ‘I do’. The ‘I think’ is not ultimate; it is the negative mode of the activity of the Self and presupposes the ‘I do’. For one possible answer to the question ‘What are you doing?’ is that I am thinking. We must therefore conclude that the ‘I do’ is the primary principle which is presupposed in all our experience; and that ‘acting’ and ‘thinking’ are opposite modes of ‘doing’ acting being the positive and thinking the negative mode. The Self then is not the thinker but the doer. In its positive doing it is agent; in its negative doing it is subject. We shall distinguish these two fundamental concepts by calling the one the concept of the Self-as-subject and the other the concept of the Self-as-agent.

Our next step must be to determine more definitely the relation between the Self-as-agent and the Self-as-subject. The formal relation of a positive to the negative which is derived from it by exclusion is still too abstract. But this abstract distinction falls within the unity of the Self. It is the same Self that is both agent and subject. Now the logical correlate of the Self is the Other and we shall have to use this distinction of Self and Other later on. But for the moment let us employ a more common term and speak of the Other as the world. We may then say that since the world is the correlate of the Self the world in which the Self as agent acts is the same world which as subject it knows.2

It might appear however that this assertion of the primacy of the Self as agent has been achieved through a mere verbal transposition. Must we not raise a prior question ‘How do we know that the self is agent?’ Clearly if this question is logically prior then the primacy of the practical cannot be maintained. But the question only seems to be a prior question from the standpoint of the Self as subject. The answer is simply that if when acting we did not know that we were acting we would not be acting. If any occurrence is to be an act of mine I must know that I perform it. This indeed is the meaning of the statement that action in distinction from thought is an inclusive concept. To do and to know that I do are two aspects of one and the same experience. This knowledge is absolute and necessary. It is not however knowledge of an object but what we may call ‘knowledge in action’.

Consider now the Self in relation to the world. When I act I modify the world. Action is causally effective even if it fails of the particular effect that is intended. This implies that the Self is part of the world in which it acts and in dynamic relation with the rest of the world. On the other hand as subject the Self stands ‘over against’ the world which is its object. The Self as subject then is not part of the world it knows but withdrawn from it and so in conception outside it or other than its object. But to be part of the world is to exist while to be excluded from the world is to be non-existent. It follows that the Self exists as agent but not as subject.

We can now understand how the antithesis of matter and mind originates. Matter like most of our fundamental concepts is a practical conception. It signifies originally like the corresponding terms in Greek and Latin ὕλη and materia the stuff of which something is made. The reference is to the technical activities of craftsmen. Clay is the matter or material from which the potter makes his pots. In general therefore matter is that which is acted upon or that which has form imposed upon it by an agent. Formed matter we call ‘body’. Now since nothing can be acted upon which offers no resistance that is to say which does not react upon the agent the agent must himself be material. As agent therefore the Self is the body. Conversely the Self as subject is the mind. For as subject the Self is non-agent withdrawn from action and therefore non-body. Since the Subject is the negative aspect of the Self all characterizations of the Self as subject participate in its negativity and must be defined by reference to practical experience. This is the reason why the terms which we use to denote or describe ‘mental’ or ‘subjective’ processes and activities were originally and often still are used with reference to practical experience and its objects. It is often said that they are used metaphorically when applied to the theoretical world. But if so they are not ordinary metaphors. They are practical expressions used as it were with a negative sign attached. The unicorn is a non-existent animal; and when we make a mental note of something we make a note which is not in fact made. Thus as an agent I am a body operative material and existent; as a subject I am a mind causally ineffective immaterial and non-existent.

These verbal paradoxes now that their significance has been explained will I hope no longer sound foolish or shocking. They are formal and conceptual. They are rooted in the fact that to think is to discriminate and the bare form of discrimination is the distinction between positive and negative. They do however set a logical problem which requires consideration. We must ask how the relation of the Self as agent to the Self as subject is to be logically construed. This is indeed the problem of the logical form of the personal. But before we can attempt an answer we must consider shortly what is meant by a logical form.

Logic itself is a highly problematical discipline. That in some sense it is formal no one doubts; but form is essentially relative and any form must be the form of something. A pure form in an absolute sense is a pure nothing. But when we ask ‘What is it that “has” the form which the logician seeks to determine?’ we face the problem of logic. For to this question different answers have been given; and in consequence different ‘logics’ have been constructed which are incompatible with one another. I have in mind in particular the difference between idealist or dialectical logic and realist or formal logic. The dialectical logic of the idealists—I should describe it myself as the logical form of the organic—is derived from Kant's transcendental logic by suppressing in the manner I have already described part of Kant's own logical doctrine which the developing organic concept considered to be the persistence of a pre-critical view which Kant had failed to overcome. For Kant maintained that the traditional formal logic was an a priori science like pure mathematics and considered therefore that his transcendental logic was quite compatible with it. The idealists considered them incompatible. May it not be that Kant was right in this instance as in others? If so it must mean that idealist and formal logicians are studying the forms of different things.

Consider first from the standpoint of the ‘I do’ the procedure of mathematics. If I have a practical problem which admits of an algebraic solution I have a double task before me. I must first formulate the relevant aspects of the concrete situation in a set of algebraic equations. When I have done this I must go on to solve the equations by a procedure which is dictated by rules for the manipulation of algebraic symbols. These two tasks though interrelated are distinct. The first demands a concrete grasp of an actual situation and consists in constructing a symbolic representation of the situation or of that part of it which is relevant to the problem. The second is a series of transformations of the symbolic representation in accordance with general rules and requires no understanding of the situation which is formally represented. It can indeed be done for me by someone else or by a machine constructed for the purpose.

Now any actual activity of thinking involves these two processes or at least their analogues. It involves firstly a constructive process of symbolization usually in words. This is much more than a mere description; for it involves relevant selection and combination. Secondly it involves the drawing of inferences for which the symbolic representation provides the premisses. Any actual process of inference presupposes a combination of premisses. Once the premisses have been combined in the requisite fashion the conclusion follows of necessity though it may not always be obvious what it is that is implied. This part of the thinking is a manipulation of symbols according to rules and with a knowledge of the rules and a suitable symbolic system it can be accomplished by a machine. But for the selection and combination of the premisses no rules can be given. The importance of this in the discovery of the unknown is apt to be overlooked if we limit our attention to the process of proof or demonstration. For when we prove or demonstrate a proposition the conclusion of the inference is already known and this dictates the selection of premisses from which it can be inferred. Proof is a secondary and derivative process. But a set of facts may be quite well known and yet their implications remain undiscovered. Some of them may be known to one person the others to another; or all may be known to the same individual. But nothing follows from them unless they are ‘thought together’ in the requisite form.

The logician can take either of these two formal aspects of thought as the central problem of his science. He may construct a logic which is concerned with the forms of inference and so with the rules for the manipulation of a linguistic symbolism. The problem of the form of representation he may consider to lie outside his province; and the simplest way of doing this is to take language as a presupposition and appeal to the common modes of speech to settle questions which concern the propriety of linguistic usage. On this side the business of logic will arise from the deficiencies of ordinary language—its ambiguities and looseness of structure and will consist in devising a symbolic system which will not be subject to such defects and which will lend itself like a mathematical symbolism to exact manipulation in accordance with unambiguous rules.

Now this aspect of logical theory is not problematic. It does not give rise to fundamental controversy between different types of logic. But if the logician concentrates his attention upon the other aspect—upon the adequacy of the forms of representation which the process of formal inference must presuppose—the case is different. The logic of implication determines the form of the relation between a conclusion and its premisses. It makes it possible to determine by a formal analysis whether a particular conclusion is or is not correctly inferred from the combination of certain premisses. But it has nothing to say about the truth of the conclusion unless the truth of the premisses is presupposed. This is not solely a material question. It has its formal aspect. For the premisses themselves are symbolic constructions. Propositions are not data; they are formulations of what is given in a concrete experience. This formulation involves the analysis of something that is given as a unity. The elements distinguished have to be represented by symbols and their unity must be represented by a pattern of relations which binds the symbols together into a whole. Such a unity-pattern as we may call it is synthetic theoretical and formal. It is therefore a logical form for the representation of the actual unity of the object to which our thought refers; that is to say the unity which is apprehended in practical experience which is not necessarily synthetic and which is necessarily neither theoretical nor formal. It was with this form of representation that Kant was concerned in his transcendental logic and since his day a unity-pattern of this kind has often been referred to as a system of categories.

Now a logic of the form of representation unlike the logic of implication is problematical. For it gives rise to the question whether there is only one unity-pattern—one system of categories—or more than one; and if so which of them is the most adequate for the representation of Reality: which means ultimately as we have seen for the representation of the unity of the Self. It is on this issue that the difference between the dialectical logic of modern idealists and the formal logic of the realists properly rests though this is apt to be obscured by the failure to distinguish clearly between the two aspects of logical form. Formal logic at least in its developed modern form assumes even when it does not assert the exclusive adequacy for all knowledge of the mathematical unity-pattern which represents the unity of the object as a summation of identical elements. The dialectical logic however rests upon the doctrine that this pattern is inadequate to represent the unity of a directed process such as is recognized in all living creatures. It offers instead an organic unity-pattern in which the elements are qualitatively different and functionally related by their complementary contributions to a unity of development.3

We may now return to the question which we formulated earlier. How is the relation of the Self as agent to the Self as subject to be construed logically? In other words How can we represent formally the unity of the Self in its two modes of activity? The crux of this problem lies as we have seen in this that formally the Self as subject is the negation of the Self as agent and since it is by its own activity that the Self withdraws from action into reflection its subjecthood is its self-negation. Thus the unity of the Self is a unity of self-affirmation and self-negation.

Such a unity can be represented neither in a mathematical nor in a dialectical form. For in the mathematical form positive and negative exclude one another. The rule of representation is that everything must be represented as always identical with itself. This is necessary because the pure form is a relation of identities and therefore all differences in the real must be represented by differences in the relations between elements while the elements themselves remain unaffected by any change in their relations to one another. Consequently the unity of a positive and its negative is unthinkable. The sum of +a and –a is always zero. Positive and negative cancel one another out. If then we attempt to represent the Self through the mathematical unity-pattern the result is necessarily a dualism of mind and body that is to say of the Agent-self and the Subject-self. The Self can be represented either as a physical system or as a mental system and these two systems exclude one another. Yet both are necessary. The unity of the two must be postulated but cannot be represented. It must be postulated because without a relation between the ‘mental’ and the ‘physical’ systems neither knowledge nor action is possible.

The organic form might seem to be in a better position; and indeed it has the advantage that it can represent the spontaneity of the Self and to this extent is more adequate. Moreover it can represent a unity of positive and negative elements which stand in a necessary dialectical relation as thesis and antithesis in the unity of a process of development. If then we conceive the Self not as substance but as organism and represent it through the organic unity-pattern it might appear that its positive and negative modes as agent and as subject could be represented as thesis and antithesis in a dialectical process of self-development. But this is an illusion. For thesis and antithesis represent successive phases in the development of a unitary system. If we represent action and thought as thesis and antithesis in a self-development we must represent them as successive phases in the development of the Self. But then it must be impossible to represent the same Self as at one and the same time both Agent and Subject. For when it is Agent it will not be Subject; and if this were actually the case then the Self could never know that it was Agent nor could it ever act with knowledge. The positive and negative phases would still exclude one another and no synthesis would be possible. This we may recall is precisely Kierkegaard's criticism of the Hegelian philosophy. The dialectic of the personal life he maintains is a dialectic without a synthesis.

The result of this inadequacy of the organic unity-pattern is that again it gives rise to a philosophical dualism. The Self may be represented either as a developing organic system of thought or as a developing organic system of action. Either a dialectical idealism or a dialectical materialism becomes possible but no unity of the two. When Karl Marx set out to make the Hegelian social philosophy practical declaring that philosophers hitherto have only explained the world while the task is to change it he could claim that he had merely inverted the dialectic. He found it standing on its head he asserted and turned it right side up. He did this by substituting the Self as ‘worker’ for the Self as ‘thinker’ without changing the organic unity-pattern. The result is a dialectic of the practical in place of a dialectic of the theoretical life.

Neither the form of the material nor the form of the organic is adequate to represent the unity of the Self. But by discovering the ground of their inadequacy we make it possible to recognize the structure which must be required in an adequate form of representation. The unity of the Self is neither a material nor an organic but a personal unity. The logical form of such a unity is one which represents a necessary unity of positive and negative modes. The Self is constituted by its capacity for self-negation. It must be represented as a positive which necessarily contains its own negative.

A formal conception of this sort is at once so abstract and so foreign to our traditional modes of thought that it will seem incomprehensible until it is referred to the types of experience which render it necessary. The necessary references will be given progressively in the sequel. But it is desirable to suggest at once some of the problems of personal experience which appear prima facie to possess such a form. The most familiar is no doubt the experience which we refer to as the moral struggle. From the time of Plato at least this has been both a commonplace of the moralists and a constant obstacle to the intellectual effort to represent the self as a unity. Constituent elements in our nature seem to be at war with one another. There is an active and dynamic contradiction between desire and reason between flesh and spirit between the ‘law in my members’ and ‘the law of my mind’. Different elements within us seem to be in competition with one another for the control of behaviour. Attempts to describe the relation between these elements in the Self tend to deny the unity which they seek to analyse by lending to the parts the characters which properly belong only to the whole. The one Self becomes two selves—a higher self and a lower self a controlling self and a self that is to be controlled. We have to learn to control ourselves yet the very conception of self-control is paradoxical for as Plato pointed out the man who is master of himself is thereby also slave of himself. The fact of experience that is expressed in all this is that the unity of the Self contains and indeed is constituted by a practical contradiction between its elements.

Another puzzling example is our capacity for self-deception. Again the fact is undeniable yet the effort to think it seems to lead to absurdity. To deceive myself can only mean to persuade myself to believe what I know to be untrue. This seems a plain impossibility yet we constantly find that we have accomplished it. We repress unacceptable desires and censor our own thoughts so that they are prevented from coming to consciousness. How is this possible unless we are conscious of them? To think this is to think a consciousness which contains the unconscious as a constituent element.

Consider finally a group of difficulties which more closely concern the business of philosophy. In the field of the personal the attempt at definition meets with peculiar difficulties. If with Aristotle we define Man as a rational organism we have to reckon with the fact that very few men behave or even think with any high degree of rationality. We find one another if not ourselves rather unreasonable creatures. Yet this fact does not vitiate the definition because only a rational being can behave irrationally. Philosophy is therefore driven to distinguish between the rational and the irrational elements in the Self and both are then constituents of its unity. The irrational element is necessary to the constitution of the rational whole.

This difficulty of definition faces us whenever we try to give an account of any mode of personal activity. In the Theaetetus for instance Plato pointed out that the difficulty in giving an account of knowledge is that it is a false account unless it allows for the possibility of error. How can we think falsely? For the capacity to think is the capacity to draw correct conclusions from given premisses. In that case if we draw incorrect conclusions we are not thinking. Yet a thinking that could not be incorrect could not be correct either; and thinking is an activity in terms of the distinction between true and false. To think is indeed to discriminate between what is incorrectly and what is correctly thought. If I could not think wrongly I could not think at all. Though I must define thinking as thinking truly mere can be no thinking which does not need its own negation to constitute it.

It is the same in ethics. Morality is doing the good. There is no problem here; the problem of morality is the problem of evil. An account of morality to be a true account must show how evil is possible. To do this it must show that evil is both real and necessary. In fact most ethical theories represent evil either as necessary but unreal; or as real but unnecessary and so they fail as theories. Morality is acting in terms of a distinction between good and evil. We can define it only in terms of doing good or acting rightly. But in its reality as an experience it is otherwise. The problem that faces us when we seek to do what is right is the impossibility of doing it without an admixture of what is wrong. In actual practice the good includes its negation in its own constitution. So Kant rightly pointed out that for a pure will duty could not exist.

These examples may perhaps suffice for the moment. We shall conclude our formal exposition by attempting to express the meaning of this logical form—of a positive which includes and is constituted by its own negative—in four propositions which though still formal are not purely logical but indicate the reference of the pure form to our actual experience. The first proposition is as follows:

1. The Self is agent and exists only as agent.

This proposition requires no further comment since the first part of the chapter was concerned to expound it. It is the positive assertion which defines the existential character of the Self. We may pass therefore to the second proposition which is concerned with the negative aspect which is excluded from the definition. It runs as follows:

2. The Self is subject but cannot exist as subject. It can be subject only because it is agent.

This proposition summarizes what we have already discussed in considering the relation between knowledge and action. There cannot be a pure subject since this is the pure negation of agency and a self which does not act cannot exist. The Self-as-subject is the Self conceived negatively as the Self in its non-existence. An activity of pure thought is a non-existent activity; that is to say it is a mere idea though a necessary idea and in no sense illusory. Illusion only enters if we posit it as an existent. When therefore we indicate the experience to which the idea refers we have to point to the fact that the Subject can exist only as an aspect of the Self as agent. It is the negative aspect of the existence of the Agent. In other words thought is at once the contrary of action and something that we do.

This brings us to the third proposition which with the fourth concerns the functional rather than the existential relation between the agency and the subjecthood of the Self. It is this:

3. The Self is subject in and for the Self as agent.

We have already indicated that the negative aspect of the personal is included in its existence that is in action. Now since action is deliberate activity which intends a change in the Other its negative aspect is included in this intention; and consequently it is not merely in action but also for action. This signifies that knowledge in its primary aspect at least arises in action; that is to say in an activity which does not aim at knowledge. If then we mean by ‘thinking’ an activity which aims at knowledge it is not true that knowledge in general is the result of thinking. On the contrary thinking presupposes knowledge. Our knowledge of the world is primarily an aspect of our action in the world. This indeed is reflected in the fact that our knowledge of the truth of a conclusion depends upon a prior knowledge of the truth of its premisses. We can only think about what we already know. This primary knowledge is the knowledge that arises in action apart from any theoretical intention. It is this knowledge to which we sometimes refer when we use—somewhat ambiguously—the term ‘experience’; as when we say that we know ‘by experience’ that Great Britain is an island. What we mean is that going to the continent from this country involves crossing the sea. It is worth noting that this piece of knowledge is theoretically an empirical generalization and therefore hypothetical. In reality it is absolutely certain.

The proposition means also that theoretical activities in which the intention is knowledge fall within action and have an essential reference to action. The reference may of course be indirect and in a sense must always be so. Sometimes this activity of reflection may be a subordinate activity falling within a practical intention as when I have to stop and think before I can proceed with what I am doing. Sometimes however we reflect simply in order to know; and our third proposition does not deny this. What it asserts is that if the result of reflection is knowledge if the theoretical conclusion can be significantly true or false then it can function as a determinant in action and modify the form of action whether in a particular case it does so or not; and this is not accidental but essential to the constitution of knowledge. In other words the question which a theoretical activity seeks to answer can only arise in practical experience directly or indirectly; and the answer can be true or false only through a reference to action. Thought cannot provide a criterion of truth but at most a criterion of the correctness of the process of thinking.

The fourth proposition is in a sense the converse and complement of this. We may formulate it as follows:

4. The Self can be agent only by being also subject.

We may expand this by noticing first that the existence of the Self is its agency. The existence of the Self therefore according to this fourth proposition depends upon its capacity to be subject and as such not agent. We have already expressed this by saying that the Self exists in virtue of its own self-negation. At the limit of abstraction there is an identity of action and reflection. To act and to know that I am acting are two aspects of one experience; since if I did not know that I was acting I should not be acting. And since to act is to do something I must know to some extent what I am doing if I am doing it. There cannot be action without knowledge. Yet action is logically prior to knowledge for there can be no knowledge without an actual activity which supports it; but there can be actual activity without knowledge. Such activity however is not action but only movement; or at most reaction to stimulus; not a deliberate effort to modify the Other.

• 1. I use the term ‘theory of knowledge’ in a very general sense; such that any attempt to determine the distinction between ‘true’ and ‘false’ implies a theory of knowledge.
• 2. Even Kant, in spite of his distinction between a phenomenal and a noumenal world, could not take exception to this, since the phenomenal world is the noumenal world as it appears to the Subject.
• 3. The characterizations of the unity patterns are approximate only, but sufficient for the present purpose. A further treatment of the subject may be found in an earlier work of mine, Interpreting the Universe (Faber and Faber, 1933).