Kant's Theory of Knowledge

by H. A. Prichard

GENERAL CRITICISM OF THE TRANSCENDENTAL DEDUCTION OF THE CATEGORIES

The preceding account of Kant's vindication of the categories has included much criticism. But the criticism has been as far as possible restricted to details, and has dealt with matters of principle only so far as has been necessary in order to follow Kant's thought. We must now consider the position as a whole, even though this may involve some repetition.[1] The general difficulties of the position may be divided into two kinds, (1) difficulties involved in the working out of the theory, even if its main principles are not questioned, and (2) difficulties involved in accepting its main principles at all.

The initial difficulty of the first kind, which naturally strikes the reader, concerns the possibility of performing the synthesis. The mind has certain general ways of combining the manifold, viz. the categories. But on general grounds we should expect the mind to possess only one mode of combining the manifold. For the character of the manifold to be combined cannot affect the mind's power of combination, and, if the power of the mind consists in combining, the combining should always be of the same kind. Thus, suppose the manifold given to the mind to be combined consisted of musical notes, we could think of the mind's power of combination as exercised in combining the notes by way of succession, provided that this be regarded as the only mode of combination. But if the mind were thought also capable of combining notes by way of simultaneity, we should at once be confronted with the insoluble problem of determining why the one mode of combination was exercised in any given case rather than the other. If, several kinds of synthesis being allowed, this difficulty be avoided by the supposition that, not being incompatible, they are all exercised together, we have the alternative task of explaining how the same manifold can be combined in each of these ways. As a matter of fact, Kant thinks of manifolds of different kinds as combined or related in different ways; thus events are related causally and quantities quantitatively. But since, on Kant's view, the manifold as given is unrelated and all combination comes from the mind, the mind should not be held capable of combining manifolds of different kinds differently. Otherwise the manifold would in its own nature imply the need of a particular kind of synthesis, and would therefore not be unrelated.

Suppose, however, we waive the difficulty involved in the plurality of the categories. There remains the equally fundamental difficulty that any single principle of synthesis contains in itself no ground for the different ways of its application.[2] Suppose it to be conceded that in the apprehension of definite shapes we combine the manifold in accordance with the conception of figure, and, for the purpose of the argument, that the conception of figure can be treated as equivalent to the category of quantity. It is plain that we apprehend different shapes, e. g. lines[3] and triangles[4], of which, if we take into account differences of relative length of sides, there is an infinite variety, and houses,[5] which may also have an infinite variety of shape. But there is nothing in the mind's capacity of relating the manifold by way of figure to determine it to combine a given manifold into a figure of one kind rather than into a figure of any other kind; for to combine the manifold into a particular shape, there is needed not merely the thought of a figure in general, but the thought of a definite figure. No 'cue' can be furnished by the manifold itself, for any such cue would involve the conception of a particular figure, and would therefore imply that the particular synthesis was implicit in the manifold itself, in which case it would not be true that all synthesis comes from the mind.

This difficulty takes a somewhat different form in the case of the categories of relation. To take the case of cause and effect, the conception of which, according to Kant, is involved in our apprehension of a succession, Kant's view seems to be that we become aware of two elements of the manifold A B as a succession of events in the world of nature by combining them as necessarily successive in a causal order, in which the state of affairs which precedes B and which contains A contains something upon which B must follow (i. e. a cause of B), which therefore makes it necessary that B must follow A.[6] But if we are to do this, we must in some way succeed in selecting or picking out from among the elements of the manifold that element A which is to be thus combined with B. We therefore need something more than the category. It is not enough that we should think that B has a cause; we must think of something in particular as the cause of B, and we must think of it either as coexistent with, or as identical with, A.

Kant fails to notice this second difficulty,[7] and up to a certain point avoids it owing to his distinction between the imagination and the understanding. For he thinks of the understanding as the source of general principles of synthesis, viz. the categories, and attributes individual syntheses to the imagination. Hence the individual syntheses, which involve particular principles, are already effected before the understanding comes into play. But to throw the work of effecting individual syntheses upon the imagination is only to evade the difficulty. For in the end, as has been pointed out,[8] the imagination must be the understanding working unreflectively, and, whether this is so or not, some account must be given of the way in which the imagination furnishes the particular principles of synthesis required.

The third and last main difficulty of the first kind concerns the relation of the elements of the manifold and the kinds of synthesis by which they are combined. This involves the distinction between relating in general and terms to be related. For to perform a synthesis is in general to relate, and the elements to be combined are the terms to be related.[9] Now it is only necessary to take instances to realize that the possibility of relating terms in certain ways involves two presuppositions, which concern respectively the general and the special nature of the terms to be related.

In the first place, it is clear that the general nature of the terms must correspond with or be adapted to the general nature of the relationship to be effected. Thus if two terms are to be related as more or less loud, they must be sounds, since the relation in question is one in respect of sound and not, e. g., of time or colour or space. Similarly, terms to be related as right and left must be bodies in space, right and left being a spatial relation. Again, only human beings can be related as parent and child. Kant's doctrine, however, does not conform to this presupposition. For the manifold to be related consists solely of sensations, and of individual spaces, and perhaps individual times, as elements of pure perception; and such a manifold is not of the kind required. Possibly individual spaces may be regarded as adequate terms to be related or combined into geometrical figures, e. g. into lines or triangles. But a house as a synthesis of a manifold cannot be a synthesis of spaces, or of times, or of sensations. Its parts are bodies, which, whatever they may be, are neither sensations nor spaces nor times, nor combinations of them. In reality they are substances of a special kind. Again, the relation of cause and effect is not a relation of sensations or spaces or times, but of successive states of physical things or substances, the relation consisting in the necessity of their succession.

In the second place, it is clear that the special nature of the relation to be effected presupposes a special nature on the part of the terms to be related. If one sound is to be related to another by way of the octave, that other must be its octave. If one quantity is to be related to another as the double of it, that quantity must be twice as large as the other. In the same way, proceeding to Kant's instances, we see that if we are to combine or relate a manifold into a triangle, and therefore into a triangle of a particular size and shape, the elements of the manifold must be lines, and lines of a particular size. If we are to combine a manifold into a house, and therefore into a house of a certain shape and size, the manifold must consist of bodies of a suitable shape and size. If we are to relate a manifold by way of necessary succession, the manifold must be such that it can be so related; in other words, if we are to relate an element X of the manifold with some other Y as the necessary antecedent of X, there must be some definite element Y which is connected with, and always occurs along with, X. To put the matter generally, we may say that the manifold must be adapted to or 'fit' the categories not only, as has been pointed out, in the sense that it must be of the right kind, but also in the sense that its individual elements must have that orderly character which enables them to be related according to the categories.

Now it is plain from Kant's vindication of what he calls the affinity of phenomena,[10] that he recognizes the existence of this presupposition. But the question arises whether this vindication can be successful. For since the manifold is originated by the thing in itself, it seems prima facie impossible to prove that the elements of the manifold must have affinity, and so be capable of being related according to the categories. Before, however, we consider the chief passage in which Kant tries to make good his position, we may notice a defence which might naturally be offered on his behalf. It might be said that he establishes the conformity of the manifold to the categories at least hypothetically, i. e. upon the supposition that the manifold is capable of entering into knowledge, and also upon the supposition that we are capable of being conscious of our identity with respect to it; for upon either supposition any element of the manifold must be capable of being combined with all the rest into one world of nature. Moreover, it might be added that these suppositions are justified, for our experience is not a mere dream, but is throughout the consciousness of a world, and we are self-conscious throughout our experience; and therefore it is clear that the manifold does in fact 'fit' the categories. But the retort is obvious. Any actual conformity of the manifold to the categories would upon this view be at best but an empirical fact, and, although, if the conformity ceased, we should cease to be aware of a world and of ourselves, no reason has been or can be given why the conformity should not cease.

The passage in which Kant vindicates the affinity of phenomena in the greatest detail is the following:

"We will now try to exhibit the necessary connexion of the understanding with phenomena by means of the categories, by beginning from below, i. e. from the empirical end. The first that is given us is a phenomenon, which if connected with consciousness is called perception[11].... But because every phenomenon contains a manifold, and consequently different perceptions are found in the mind scattered and single, a connexion of them is necessary, which they cannot have in mere sense. There is, therefore, in us an active power of synthesis of this manifold, which we call imagination, and the action of which, when exercised immediately upon perceptions, I call apprehension. The business of the imagination, that is to say, is to bring the manifold of intuition[12] into an image; it must, therefore, first receive the impressions into its activity, i. e. apprehend them."

"But it is clear that even this apprehension of the manifold would not by itself produce an image and a connexion of the impressions, unless there were a subjective ground in virtue of which one perception, from which the mind has passed to another, is summoned to join that which follows, and thus whole series of perceptions are presented, i. e. a reproductive power of imagination, which power, however, is also only empirical."

"But if representations reproduced one another at haphazard just as they happened to meet together, once more no determinate connexion would arise, but merely chaotic heaps of them, and consequently no knowledge would arise; therefore the reproduction of them must have a rule, according to which a representation enters into connexion with this rather than with another in the imagination. This subjective and empirical ground of reproduction according to rules is called the association of representations."

"But now, if this unity of association had not also an objective ground, so that it was impossible that phenomena should be apprehended by the imagination otherwise than under the condition of a possible synthetic unity of this apprehension, it would also be a pure accident that phenomena were adapted to a connected system of human knowledge. For although we should have the power of associating perceptions, it would still remain wholly undetermined and accidental whether they were associable; and in the event of their not being so, a multitude of perceptions and even perhaps a whole sensibility would be possible, in which much empirical consciousness would be met with in my mind, but divided and without belonging to one consciousness of myself, which however is impossible. For only in that I ascribe all perceptions to one consciousness (the original apperception) can I say of all of them that I am conscious of them. There must therefore be an objective ground, i. e. a ground to be recognized a priori before all empirical laws of the imagination, on which rests the possibility, nay even the necessity, of a law which extends throughout all phenomena, according to which we regard them without exception as such data of the senses, as are in themselves associable and subjected to universal rules of a thorough-going connexion in reproduction. This objective ground of all association of phenomena I call the affinity of phenomena. But we can meet this nowhere else than in the principle of the unity of apperception as regards all cognitions which are to belong to me. According to it, all phenomena without exception must so enter into the mind or be apprehended as to agree with the unity of apperception, which agreement would be impossible without synthetical unity in their connexion, which therefore is also objectively necessary."

"The objective unity of all (empirical) consciousness in one consciousness (the original apperception) is therefore the necessary condition even of all possible perception, and the affinity of all phenomena (near or remote) is a necessary consequence of a synthesis in the imagination, which is a priori founded upon rules."

"The imagination is therefore also a power of a priori synthesis, for which reason we give it the name of the productive imagination; and so far as it, in relation to all the manifold of the phenomenon, has no further aim than the necessary unity in the synthesis of the phenomenon, it can be called the transcendental function of the imagination. It is therefore strange indeed, but nevertheless clear from the preceding, that only by means of this transcendental function of the imagination does even the affinity of phenomena, and with it their association and, through this, lastly their reproduction according to laws, and consequently experience itself become possible, because without it no conceptions of objects would ever come together into one experience."[13]

If it were not for the last two paragraphs[14], we should understand this difficult passage to be substantially identical in meaning with the defence of the affinity of phenomena just given.[15] We should understand Kant to be saying (1) that the synthesis which knowledge requires presupposes not merely a faculty of association on our part by which we reproduce elements of the manifold according to rules, but also an affinity on the part of the manifold to be apprehended, which enables our faculty of association to get to work, and (2) that this affinity can be vindicated as a presupposition at once of knowledge and of self-consciousness.

In view, however, of the fact that, according to the last two paragraphs, the affinity is due to the imagination,[16] it seems necessary to interpret the passage thus:

'Since the given manifold of sense consists of isolated elements, this manifold, in order to enter into knowledge, must be combined into an image. This combination is effected by the imagination, which however must first apprehend the elements one by one.'

'But this apprehension of the manifold by the imagination could produce no image, unless the imagination also possessed the power of reproducing past elements of the manifold, and, if knowledge is to arise, of reproducing them according to rules. This faculty of reproduction by which, on perceiving the element A, we are led to think of or reproduce a past element B—B being reproduced according to some rule—rather than C or D is called the faculty of association; and since the rules according to which it works depend on empirical conditions, and therefore cannot be anticipated a priori, it may be called the subjective ground of reproduction.'

'But if the image produced by association is to play a part in knowledge, the empirical faculty of reproduction is not a sufficient condition or ground of it. A further condition is implied, which may be called objective in the sense that it is a priori and prior to all empirical laws of imagination. This condition is that the act by which the data of sense enter the mind or are apprehended, i. e. the act by which the imagination apprehends and combines the data of sense into a sensuous image, must make the elements such that they have affinity, and therefore such that they can subsequently be recognized as parts of a necessarily related whole.[17] Unless this condition is satisfied, even if we possessed the faculty of association, our experience would be a chaos of disconnected elements, and we could not be self-conscious, which is impossible. Starting, therefore, with the principle that we must be capable of being self-conscious with respect to all the elements of the manifold, we can lay down a priori that this condition is a fact.'

'It follows, then, that the affinity or connectedness of the data of sense presupposed by the reproduction which is presupposed in knowledge, is actually produced by the productive faculty of imagination, which, in combining the data into a sensuous image, gives them the unity required.'

If, as it seems necessary to believe, this be the correct interpretation of the passage,[18] Kant is here trying to carry out to the full his doctrine that all unity or connectedness comes from the mind's activity. He is maintaining that the imagination, acting productively on the data of sense and thereby combining them into an image, gives the data a connectedness which the understanding can subsequently recognize. But to maintain this is, of course, only to throw the problem one stage further back. If reproduction, in order to enter into knowledge, implies a manifold which has such connexion that it is capable of being reproduced according to rules, so the production of sense-elements into a coherent image in turn implies sense-elements capable of being so combined. The act of combination cannot confer upon them or introduce into them a unity which they do not already possess.

The fact is that this step in Kant's argument exhibits the final breakdown of his view that all unity or connectedness or relatedness is conferred upon the data of sense by the activity of the mind. Consequently, this forms a convenient point at which to consider what seems to be the fundamental mistake of this view. The mistake stated in its most general form appears to be that, misled by his theory of perception, he regards 'terms' as given by things in themselves acting on the sensibility, and 'relations' as introduced by the understanding,[19] whereas the fact is that in the sense in which terms can be said to be given, relations can and must also be said to be given.

To realize that this is the case, we need only consider Kant's favourite instance of knowledge, the apprehension of a straight line. According to him, this presupposes that there is given to us a manifold, which—whether he admits it or not—must really be parts of the line, and that we combine this manifold on a principle involved in the nature of straightness. Now suppose that the manifold given is the parts AB, BC, CD, DE of the line AE. It is clearly only possible to recognize AB and BC as contiguous parts of a straight line, if we immediately apprehend that AB and BC form one line of which these parts are identical in direction. Otherwise, we might just as well join AB and BC at a right angle, and in fact at any angle; we need not even make AB and BC contiguous.[20] Similarly, the relation of BC to CD and of CD to DE must be just as immediately apprehended as the parts themselves. Is there, however, any relation of which it could be said that it is not given, and to which therefore Kant's doctrine might seem to apply? There is. Suppose AB, BC, CD to be of such a size that, though we can see AB and BC, or BC and CD, together, we cannot see AB and CD together. It is clear that in this case we can only learn that AB and CD are parts of the same straight line through an inference. We have to infer that, because each is in the same straight line with BC, the one is in the same straight line with the other. Here the fact that AB and CD are in the same straight line is not immediately apprehended. This relation, therefore, may be said not to be given; and, from Kant's point of view, we could say that we introduce this relation into the manifold through our activity of thinking, which combines AB and CD together in accordance with the principle that two straight lines which are in the same line with a third are in line with one another. Nevertheless, this case is no exception to the general principle that relations must be given equally with terms; for we only become aware of the relation between AB and CD, which is not given, because we are already aware of other relations, viz. those between AB and BC, and BC and CD, which are given. Relations then, or, in Kant's language, particular syntheses must be said to be given, in the sense in which the elements to be combined can be said to be given.

Further, we can better see the nature of Kant's mistake in this respect, if we bear in mind that Kant originally and rightly introduced the distinction between the sensibility and the understanding as that between the passive faculty by which an individual is given or presented to us and the active faculty by which we bring an individual under, or recognize it as an instance of a universal.[21] For we then see that Kant in the Transcendental Deduction, by treating what is given by the sensibility as terms and what is contributed by the understanding as relations, is really confusing the distinction between a relation and its terms with that between universal and individual; in other words, he says of terms what ought to be said of individuals, and of relations what ought to be said of universals. That the confusion is a confusion, and not a legitimate identification, it is easy to see. For, on the one hand, a relation between terms is as much an individual as either of the terms. That a body A is to the right of a body B is as much an individual fact as either A or B.[22] And if terms, as being individuals, belong to perception and are given, in the sense that they are in an immediate relation to us, relations, as being individuals, equally belong to perception and are given. On the other hand, individual terms just as much as individual relations imply corresponding universals. An individual body implies 'bodiness', just as much as the fact that a body A is to the right of a body B implies the relationship of 'being to the right of something'. And if, as is the case, thinking or conceiving in distinction from perceiving, is that activity by which we recognize an individual, given in perception, as one of a kind, conceiving is involved as much in the apprehension of a term as in the apprehension of a relation. The apprehension of 'this red body' as much involves the recognition of an individual as an instance of a kind, i. e. as much involves an act of the understanding, as does the apprehension of the fact that it is brighter than some other body.

Kant has failed to notice this confusion for two reasons. In the first place, beginning in the Analytic with the thought that the thing in itself, by acting on our sensibility, produces isolated sense data, he is led to adopt a different view of the understanding from that which he originally gave, and to conceive its business as consisting in relating these data. In the second place, by distinguishing the imagination from the understanding, he is able to confine the understanding to being the source of universals or principles of relation in distinction from individual relations.[23] Since, however, as has been pointed out, and as Kant himself sees at times, the imagination is the understanding working unreflectively, this limitation cannot be successful.

There remain for consideration the difficulties of the second kind, i. e. the difficulties involved in accepting its main principles at all. These are of course the most important. Throughout the deduction Kant is attempting to formulate the nature of knowledge. According to him, it consists in an activity of the mind by which it combines the manifold of sense on certain principles and is to some extent aware that it does so, and by which it thereby gives the manifold relation to an object. Now the fundamental and final objection to this account is that what it describes is not knowledge at all. The justice of this objection may be seen by considering the two leading thoughts underlying the view, which, though closely connected, may be treated separately. These are the thought of knowledge as a process by which representations acquire relation to an object, and the thought of knowledge as a process of synthesis.

It is in reality meaningless to speak of 'a process by which representations or ideas acquire relation to an object'.[24] The phrase must mean a process by which a mere apprehension, which, as such, is not the apprehension of an object, becomes the apprehension of an object. Apprehension, however, is essentially and from the very beginning the apprehension of an object, i. e. of a reality apprehended. If there is no object which the apprehension is 'of', there is no apprehension. It is therefore wholly meaningless to speak of a process by which an apprehension becomes the apprehension of an object. If when we reflected we were not aware of an object, i. e. a reality apprehended, we could not be aware of our apprehension; for our apprehension is the apprehension of it, and is itself only apprehended in relation to, though in distinction from, it. It is therefore impossible to suppose a condition of mind in which, knowing what 'apprehension' means, we proceed to ask, 'What is meant by an object of it?' and 'How does an apprehension become related to an object?'; for both questions involve the thought of a mere representation, i. e. of an apprehension which as yet is not the apprehension of anything.

These questions, when their real nature is exhibited, are plainly absurd. Kant's special theory, however, enables him to evade the real absurdity involved. For, according to his view, a representation is the representation or apprehension of something only from the point of view of the thing in itself. As an appearance or perhaps more strictly speaking as a sensation, it has also a being of its own which is not relative[25]; and from this point of view it is possible to speak of 'mere' representations and to raise questions which presuppose their reality.[26]

But this remedy, if remedy it can be called, is at least as bad as the disease. For, in the first place, the change of standpoint is necessarily illegitimate. An appearance or sensation is not from any point of view a representation in the proper sense, i. e. a representation or apprehension of something. It is simply a reality to be apprehended, of the special kind called mental. If it be called a representation, the word must have a new meaning; it must mean something represented, or presented,[27] i. e. object of apprehension, with the implication that what is presented, or is object of apprehension, is mental or a modification of the mind. Kant therefore only avoids the original absurdity by an illegitimate change of standpoint, the change being concealed by a tacit transition in the meaning of representation. In the second place, the change of standpoint only saves the main problem from being absurd by rendering it insoluble. For if a representation be taken to be an appearance or a sensation, the main problem becomes that of explaining how it is that, beginning with the apprehension of mere appearances or sensations, we come to apprehend an object, in the sense of an object in nature, which, as such, is not an appearance or sensation but a part of the physical world. But if the immediate object of apprehension were in this way confined to appearances, which are, to use Kant's phrase, determinations of our mind, our apprehension would be limited to these appearances, and any apprehension of an object in nature would be impossible.[28] In fact, it is just the view that the immediate object of apprehension consists in a determination of the mind which forms the basis of the solipsist position. Kant's own solution involves an absurdity at least as great as that involved in the thought of a mere representation, in the proper sense of representation. For the solution is that appearances or sensations become related to an object, in the sense of an object in nature, by being combined on certain principles. Yet it is plainly impossible to combine appearances or sensations into an object in nature. If a triangle, or a house, or 'a freezing of water'[29] is the result of any process of combination, the elements combined must be respectively lines, and bricks, and physical events; these are objects in the sense in which the whole produced by the combination is an object, and are certainly not appearances or sensations. Kant conceals the difficulty from himself by the use of language to which he is not entitled. For while his instances of objects are always of the kind indicated, he persists in calling the manifold combined 'representations', i. e. presented mental modifications. This procedure is of course facilitated for him by his view that nature is a phenomenon or appearance, but the difficulty which it presents to the reader culminates when he speaks of the very same representations as having both a subjective and an objective relation, i. e. as being both modifications of the mind and parts of nature.[30]

We may now turn to Kant's thought of knowledge as a process of synthesis. When Kant speaks of synthesis, the kind of synthesis of which he usually is thinking is that of spatial elements into a spatial whole; and although he refers to other kinds, e. g. of units into numbers, and of events into a temporal series, nevertheless it is the thought of spatial synthesis which guides his view. Now we must in the end admit that the spatial synthesis of which he is thinking is really the construction or making of spatial objects in the literal sense. It would be rightly illustrated by making figures out of matches or spelicans, or by drawing a circle with compasses, or by building a house out of bricks. Further, if we extend this view of the process of which Kant is thinking, we have to allow that the process of synthesis in which, according to Kant, knowledge consists is that of making or constructing parts of the physical world, and in fact the physical world itself, out of elements given in perception.[31] The deduction throughout presupposes that the synthesis is really manufacture, and Kant is at pains to emphasize the fact. "The order and conformity to law in the phenomena which we call nature we ourselves introduce, and we could not find it there, if we or the nature of our mind had not originally placed it there."[32] He naturally rejoices in the manufacture, because it is just this which makes the categories valid. If knowing is really making, the principles of synthesis must apply to the reality known, because it is by these very principles that the reality is made. Moreover, recognition of this fact enables us to understand certain features of his view which would otherwise be inexplicable. For if the synthesis consists in literal construction, we are able to understand why Kant should think (1) that in the process of knowledge the mind introduces order into the manifold, (2) that the mind is limited in its activity of synthesis by having to conform to certain principles of construction which constitute the nature of the understanding, and (3) that the manifold of phenomena must possess affinity. If, for example, we build a house, it can be said (1) that we introduce into the materials a plan or principle of arrangement which they do not possess in themselves, (2) that the particular plan is limited by, and must conform to, the laws of spatial relation and to the general presuppositions of physics, such as the uniformity of nature, and (3) that only such materials are capable of the particular combination as possess a nature suitable to it. Moreover, if, for Kant, knowing is really making, we are able to understand two other prominent features of his view. We can understand why Kant should lay so much stress upon the 'recognition' of the synthesis, and upon the self-consciousness involved in knowledge. For if the synthesis of the manifold is really the making of an object, it results merely in the existence of the object; knowledge of it is still to be effected. Consequently, knowledge of the object only finds a place in Kant's view by the recognition (on the necessity of which he insists) of the manifold as combined on a principle. This recognition, which Kant considers only an element in knowledge, is really the knowledge itself. Again, since the reality to be known is a whole of parts which we construct on a principle, we know that it is such a whole, and therefore that 'the manifold is related to one object', because, and only because, we know that we have combined the elements on a principle. Self-consciousness therefore must be inseparable from consciousness of an object.

The fundamental objection to this account of knowledge seems so obvious as to be hardly worth stating; it is of course that knowing and making are not the same. The very nature of knowing presupposes that the thing known is already made, or, to speak more accurately, already exists.[33] In other words, knowing is essentially the discovery of what already is. Even if the reality known happens to be something which we make, e. g. a house, the knowing it is distinct from the making it, and, so far from being identical with the making, presupposes that the reality in question is already made. Music and poetry are, no doubt, realities which in some sense are 'made' or 'composed', but the apprehension of them is distinct from and presupposes the process by which they are composed.

How difficult it is to resolve knowing into making may be seen by consideration of a difficulty in the interpretation of Kant's phrase 'relation of the manifold to an object', to which no allusion has yet been made. When it is said that a certain manifold is related to, or stands[34] in relation to, an object, does the relatedness referred to consist in the fact that the manifold is combined into a whole, or in the fact that we are conscious of the combination, or in both? If we accept the first alternative we must allow that, while relatedness to an object implies a process of synthesis, yet the relatedness, and therefore the synthesis, have nothing to do with knowledge. For the relatedness of the manifold to an object will be the combination of the elements of the manifold as parts of an object constructed, and the process of synthesis involved will be that by which the object is constructed. This process of synthesis will have nothing to do with knowledge; for since it is merely the process by which the object is constructed, knowledge so far is not effected at all, and no clue is given to the way in which it comes about. If, however, we accept the second alternative, we have to allow that while relatedness to an object has to do with knowledge, yet it in no way implies a process of synthesis. For since in that case it consists in the fact that we are conscious of the manifold as together forming an object, it in no way implies that the object has been produced by a process of synthesis. Kant, of course, would accept the third alternative. For, firstly, since it is knowledge which he is describing, the phrase 'relatedness to an object' cannot refer simply to the existence of a combination of the manifold, and of a process by which it has been produced; its meaning must include consciousness of the combination. In the second place, it is definitely his view that we cannot represent anything as combined in the object without having previously combined it ourselves.[35] Moreover, it is just with respect to this connexion between the synthesis and the consciousness of the synthesis that his reduction of knowing to making helps him; for to make an object, e. g. a house, is to make it consciously, i. e. to combine materials on a principle of which we are aware. Since, then, the combining of which he speaks is really making, it seems to him impossible to combine a manifold without being aware of the nature of the act of combination, and therefore of the nature of the whole thereby produced.[36] But though this is clearly Kant's view, it is not justified. In the first place, 'relatedness of the manifold to an object' ought not to refer both to its combination in a whole and to our consciousness of the combination; and in strictness it should refer to the former only. For as referring to the former it indicates a relation of the manifold to the object, as being the parts of the object, and as referring to the latter it indicates a relation of the manifold to us, as being apprehended by us as the parts of the object. But two relations which, though they are of one and the same thing, are nevertheless relations of it to two different things, should not be referred to by the same phrase. Moreover, since the relatedness is referred to as relatedness to an object, the phrase properly indicates the relation of the manifold to an object, and not to us as apprehending it. Again, in the second place, Kant cannot successfully maintain that the phrase is primarily a loose expression for our consciousness of the manifold as related to an object, and that since this implies a process of synthesis, the phrase may fairly include in its meaning the thought of the combination of the manifold by us into a whole. For although Kant asserts—and with some plausibility—that we can only apprehend as combined what we have ourselves combined, yet when we consider this assertion seriously we see it to be in no sense true.

The general conclusion, therefore, to be drawn is that the process of synthesis by which the manifold is said to become related to an object is a process not of knowledge but of construction in the literal sense, and that it leaves knowledge of the thing constructed still to be effected. But if knowing is obviously different from making, why should Kant have apparently felt no difficulty in resolving knowing into making? Three reasons may be given.

In the first place, the very question, 'What does the process of knowing consist in?' at least suggests that knowing can be resolved into and stated in terms of something else. In this respect it resembles the modern phrase 'theory of knowledge'. Moreover, since it is plain that in knowing we are active, the question is apt to assume the form, 'What do we do when we know or think?' and since one of the commonest forms of doing something is to perform a physical operation on physical things, whereby we effect a recombination of them on some plan, it is natural to try to resolve knowing into this kind of doing, i. e. into making in a wide sense of the word.

In the second place, Kant never relaxed his hold upon the thing in itself. Consequently, there always remained for him a reality which existed in itself and was not made by us. This was to him the fundamental reality, and the proper object of knowledge, although unfortunately inaccessible to our faculties of knowing. Hence to Kant it did not seriously matter that an inferior reality, viz. the phenomenal world, was made by us in the process of knowing.

In the third place, it is difficult, if not impossible, to read the Deduction without realizing that Kant failed to distinguish knowing from that formation of mental imagery which accompanies knowing. The process of synthesis, if it is even to seem to constitute knowledge and to involve the validity of the categories, must really be a process by which we construct, and recognize our construction of, an individual reality in nature out of certain physical data. Nevertheless, it is plain that what Kant normally describes as the process of synthesis is really the process by which we construct an imaginary picture of a reality in nature not present to perception, i. e. by which we imagine to ourselves what it would look like if we were present to perceive it. This is implied by his continued use of the terms 'reproduction' and 'imagination' in describing the synthesis. To be aware of an object of past perception, it is necessary, according to him, that the object should be reproduced. It is thereby implied that the object of our present awareness is not the object of past perception, but a mental image which copies or reproduces it. The same implication is conveyed by his use of the term 'imagination' to describe the faculty by which the synthesis is effected; for 'imagination' normally means the power of making a mental image of something not present to perception, and this interpretation is confirmed by Kant's own description of the imagination as 'the faculty of representing an object even without its presence in perception'.[37] Further, that Kant really fails to distinguish the construction of mental imagery from literal construction is shown by the fact that, although he insists that the formation of an image and reproduction are both necessary for knowledge, he does not consistently adhere to this. For his general view is that the elements combined and recognized as combined are the original data of sense, and not reproductions of them which together form an image, and his instances imply that the elements retained in thought, i. e. the elements of which we are aware subsequently to perception, are the elements originally perceived, e. g. the parts of a line or the units counted.[38] Moreover, in one passage Kant definitely describes certain objects of perception taken together as an image of that 'kind' of which, when taken together, they are an instance. "If I place five points one after another, . . . . . this is an image of the number five."[39] Now, if it be granted that Kant has in mind normally the process of imagining, we can see why he found no difficulty in the thought of knowledge as construction. For while we cannot reasonably speak of making an object of knowledge, we can reasonably speak of making a mental image through our own activity, and also of making it in accordance with the categories and the empirical laws which presuppose them. Moreover, the ease with which it is possible to take the imagining which accompanies knowing for knowing[40]—the image formed being taken to be the object known and the forming it being taken to be the knowing it—renders it easy to transfer the thought of construction to the knowledge itself. The only defect, however, under which the view labours is the important one that, whatever be the extent to which imagination must accompany knowledge, it is distinct from knowledge. To realize the difference we have only to notice that the process by which we present to ourselves in imagination realities not present to perception presupposes, and is throughout guided by, the knowledge of them. It should be noted, however, that, although the process of which Kant is normally thinking is doubtless that of constructing mental imagery, his real view must be that knowledge consists in constructing a world out of the data of sense, or, more accurately, as his instances show, out of the objects of isolated perceptions, e. g. parts of a line or units to be counted. Otherwise the final act of recognition would be an apprehension not of the world of nature, but of an image of it.

'This criticism,' it may be said, 'is too sweeping. It may be true that the process which Kant describes is really making in the literal sense and not knowing, but Kant's mistake may have been merely that of thinking of the wrong kind of synthesis. For both ordinary language and that of philosophical discussion imply that synthesis plays some part in knowledge. Thus we find in ordinary language the phrases 'putting 2 and 2 together' and '2 and 2 make 4'. Even in philosophical discussions we find it said that a complex conception, e. g. gold, is a synthesis of simple conceptions, e. g. yellowness, weight, &c.; that in judgement we relate or refer the predicate to the subject; and that in inference we construct reality, though only mentally or ideally. Further, in any case it is by thinking or knowing that the world comes to be for us; the more we think, the more of reality there is for us. Hence at least the world for us or our world is due to our activity of knowing, and so is in some sense made by us, i. e. by our relating activity.'

This position, however, seems in reality to be based on a simple but illegitimate transition, viz. the transition to the assertion that in knowing we relate, or combine, or construct from the assertion that in knowing we recognize as related, or combined, or constructed—the last two terms being retained to preserve the parallelism.[41] While the latter assertion may be said to be true, although the terms 'combined' and 'constructed' should be rejected as misleading, the former assertion must be admitted to be wholly false, i. e. true in no sense whatever. Moreover, the considerations adduced in favour of the position should, it seems, be met by a flat denial of their truth or, if not, of their relevance. For when it is said that our world, or the world for us, is due to our activity of thinking, and so is in some sense made by us, all that should be meant is that our apprehending the world as whatever we apprehend it to be presupposes activity on our part. But since the activity is after all only the activity itself of apprehending or knowing, this assertion is only a way of saying that apprehending or knowing is not a condition of mind which can be produced in us ab extra, but is something which we have to do for ourselves. Nothing is implied to be made. If anything is to be said to be made, it must be not our world but our activity of apprehending the world; but even we and our activity of apprehending the world are not related as maker and thing made. Again, to speak of a complex conception, e. g. gold, and to say that it involves a synthesis of simple conceptions by the mind is mere 'conceptualism'. If, as we ought to do, we replace the term 'conception' by 'universal', and speak of gold as a synthesis of universals, any suggestion that the mind performs the synthesis will vanish, for a 'synthesis of universals' will mean simply a connexion of universals. All that is mental is our apprehension of their connexion. Again, in judgement we cannot be said to relate predicate to subject. Such an assertion would mean either that we relate a conception to a conception, or a conception to a reality[42], or a reality to a reality; and, on any of these interpretations, it is plainly false. To retain the language of 'relation' or of 'combination' at all, we must say that in judgement we recognize real elements as related or combined. Again, when we infer, we do not construct, ideally or otherwise. 'Ideal construction'[43] is a contradiction in terms, unless it refers solely to mental imagining, in which case it is not inference. Construction which is not 'ideal', i. e. literal construction, plainly cannot constitute the nature of inference; for inference would cease to be inference, if by it we made, and did not apprehend, a necessity of connexion. Again, the phrase '2 and 2 make 4' does not justify the view that in some sense we 'make' reality. It of course suggests that 2 and 2 are not 4 until they are added, i. e. that the addition makes them 4.[44] But the language is only appropriate when we are literally making a group of 4 by physically placing 2 pairs of bodies in one group. Where we are counting, we should say merely that 2 and 2 are 4. Lastly, it must be allowed that the use of the phrase 'putting two and two together', to describe an inference from facts not quite obviously connected, is loose and inexact. If we meet a dog with a blood-stained mouth and shortly afterwards see a dead fowl, we may be said to put two and two together and to conclude thereby that the dog killed the fowl. But, strictly speaking, in drawing the inference we do not put anything together. We certainly do not put together the facts that the mouth of the dog is blood-stained and that the fowl has just been killed. We do not even put the premises together, i. e. our apprehensions of these facts. What takes place should be described by saying simply that seeing that the fowl is killed, we also remember that the dog's mouth was stained, and then apprehend a connexion between these facts.

The fact seems to be that the thought of synthesis in no way helps to elucidate the nature of knowing, and that the mistake in principle which underlies Kant's view lies in the implicit supposition that it is possible to elucidate the nature of knowledge by means of something other than itself. Knowledge is sui generis and therefore a 'theory' of it is impossible. Knowledge is simply knowledge, and any attempt to state it in terms of something else must end in describing something which is not knowledge.[45]