CHAPTER I

THE CLUE TO THE DISCOVERY OF ALL PURE CONCEPTS OF THE UNDERSTANDING

That the understanding is “an absolute unity” is repeated. From this assertion, thus dogmatically made, without even an attempt at argument, Kant deduces the important conclusion that the pure concepts, originating from such a source, “must be connected with each other according to one concept or idea (Begriff oder Idee).” And he adds the equally unproved assertion:

“But such a connection supplies a rule by which we are enabled to assign its proper place to each pure concept of the understanding and by which we can determine in an a priori manner their systematic completeness. Otherwise we should be dependent in these matters on our own discretionary judgment or merely on chance.”

In the next section he sets himself to discover from an examination of analytic thinking what this rule or principle actually is, and in so doing he for the first time discloses, in any degree at all adequate, the real nature of the position which he is seeking to develop. He connects the required principle with the nature of the act of judging, considered as a function of unity.

=Section I. The Logical Use of the Understanding.=--This section, viewed as introductory to the metaphysical deduction of the categories, is extremely unsatisfactory. It directs attention to the wrong points, and conceals rather than defines Kant’s real position. Its argumentation is also contorted and confused, and only by the most patient analysis can it be straightened out. The commentator has presented to him a twofold task from which there is no escape. He must render the argument consistent by such modification as will harmonise it with Kant’s later and more deliberate positions, and he must explain why Kant has presented it in this misleading manner.

The title of the section would seem to imply that only the discursive activities of understanding are to be dealt with. That is, indeed, in the main true. Confusion results, however, from the clashing of this avowed intention with the ultimate purpose in view of which the argument is propounded. Kant is seeking to prove that we can derive from the more accessible procedure of the discursive understanding a clue sufficient for determining those pre-logical activities which have to be postulated in terms of his new Copernican hypothesis. But though that is the real intention of this section, it has, unfortunately, not been explicitly recognised, and can be divined by the reader only after he has mastered the later portions of the Analytic. Kant’s argument has also the further defect that no sufficient statement is given either of the nature of the discursive concept or of its relation to judgment. These lacunae we must fill out as best we can from his utterances elsewhere. I shall first state Kant’s view of the distinction between discursive and synthetic thinking, and then examine his treatment of the nature of the concept and of its relation to judgment.

As already noted, the distinction between transcendental and general logic marks for Kant all-important differences in the use of the understanding. In the one employment the understanding, by creative synthetic activities, generates from the given manifold the complex objects of sense-experience. In so doing it interprets and organises the manifold through concepts which originate from within itself. By the other it discriminates and compares, and thereby derives from the content of sense-experience the generic concepts of the traditional logic. Now Kant would seem to argue in this section that if the difference in the origin of the concepts in those two cases be left out of account, and if we attend only to the quite general character of their respective activities, they will be found to agree in one fundamental feature, namely, that they express functions of unity. Each is based on the spontaneity of thought--on the spontaneity of synthetic interpretation on the one hand, of discrimination and comparison on the other. This feature common to the two types of activity can be further defined as being the unity of the act whereby a multiplicity is comprehended under a single representation. In the judgment “every metal is a body” the variety of metals is reduced to unity through the concept body. In an analogous manner the synthetic understanding organises a manifold of intuition through some such form of unity as that of substance and attribute. That is the category which underlies the above proposition, and which renders possible the specific unity of the total judgment. To quote the sentence with which in a later section Kant introduces his table of categories:

“The same understanding, and by the same operations by which in concepts, by means of analytic unity, it has produced the logical form of a judgment, introduces, by means of the synthetic unity of the manifold in intuition in general, a transcendental element into its representations....”

Now Kant’s exposition is extremely misleading. As his later utterances show, his real argument is by no means that which is here given. We shall have occasion to observe that Kant is unable to prove, and does not ultimately profess to prove, that it is “the same understanding,” and still less that it is “the same operations,” which are exercised in discursive and in creative thinking. But this is a criticism which it would be premature to introduce at this stage. We must proceed to it by way of preliminary analysis of the above exposition. Kant’s argument does not rest upon any such analogy as that just drawn, between the concepts formed by consciously comparing contents and the concepts which originate from within the understanding itself. Both, it is true, are functions of unity, but otherwise there is, according to Kant’s own teaching, not the least resemblance between them. A generic or abstract concept expresses common qualities found in each of a number of complex contents. It is itself a content. A category, on the other hand, is always a function of unity whereby contents are interpreted. It is not a content, but a form for the organisation of content. It can gain expression only in the total act of judging, not in any one element such as the discursive concept. But though the analogy drawn by Kant thus breaks down, his argument is continued in a new and very different form. It is no longer made to rest on any supposed resemblance between discursive and creative thinking, regarded as co-ordinate and independent activities. It now consists in the proof that the former presupposes and is conditioned by the latter. Through study of the understanding in its more accessible discursive procedure, we may hope to discover the synthetic forms according to which it has proceeded in its pre-logical activities. When we determine the various forms of analytic judgment, the categories which are involved in synthetic thinking reveal themselves to consciousness.

Thus in spite of Kant’s insistence upon the conceptual predicate, and upon the unity to which it gives expression, immediately he proceeds to the deduction of the categories, the emphasis is shifted to the unity which underlies the judgment as a whole. What constitutes such propositions as “all bodies are divisible,” “every metal is a body,” a unique and separate type of judgment is not the character of the predicate, but the category of substance and attribute whereby the predicate is related to the subject. To that category they owe their specific form; and it is a function of unity for which the discursive understanding can never account. As Kant states in the Prolegomena, if genuine judgments, that is, judgments that are “objectively valid,” are analysed,

“...it will be found that they never consist of mere intuitions connected only (as is commonly believed) by comparison in a judgment. They would be impossible were not a pure concept of the understanding superadded to the concepts abstracted from intuition. The abstract concepts are subsumed under a pure concept, and in this manner only can they be connected in an objectively valid judgment.”

Thus the analogy between discursive and a priori concepts is no sooner drawn than it is set aside as irrelevant. Though generic concepts rest upon functions of unity, and though (as we shall see immediately) they exist only as factors in the total act of judging, there is otherwise not the least resemblance between them and the categories. The clue to the categories is not to be found in the inherent characteristics of analytic thinking, or of its specific products (namely, concepts), but solely in what, after all abstraction, it must still retain from the products which synthetic thinking creates. Each type of analytic judgment will be found on examination to involve some specific function whereby the conceptual factors are related to, and unified with, the other elements in the judgment. This function of unity is in each case an a priori category of the understanding. That is the thesis which underlies the concluding sentence of this section.

“The functions of the understanding [i.e. the a priori concepts of understanding] can be discovered in their completeness, if it is possible to state exhaustively the functions of unity [i.e. the forms of relation] in judgments.”

The adoption of such a position involves, it may be noted, the giving up of the assertion, which is so emphatically made in the passage above quoted, that it is by the same activities that the understanding discursively forms abstract concepts and creatively organises the manifold of sense. That is in no respect true. There is no real identity--there is not even analogy--between the processes of comparison and abstraction on the one hand and those of synthetic interpretation on the other. The former are merely reflective: the latter are genuinely creative. Discursive activities are conscious processes, and are under our control: the synthetic processes, are non-conscious; only their finished products appear within the conscious field. This, however, is to anticipate a conclusion which was among the last to be realised by Kant himself, namely that there is no proof that these two types of activity are ascribable to one and the same source. The synthetic activities--as he himself finally came to hold--are due to a faculty of imagination.

“Synthesis in general ... is the mere result of the power of imagination, a blind but indispensable function of the soul, without which we should have no knowledge whatsoever, but of which we are scarcely ever conscious.”

This sentence occurs in a passage which is undoubtedly a later interpolation. The “scarcely ever” (selten nur einmal) indicates Kant’s lingering reluctance to recognise this fundamental fact, destructive of so much in his earlier views, even though it completes and reinforces his chief ultimate conclusions. With this admission Kant also gives up his sole remaining ground for the contention that there must be a complete parallelism between discursive and creative thinking. If they arise from such different sources, we have no right to assume, without specific proof, that they must coincide in the forms of their activity. This is a point to which we shall return in discussing Kant’s formulation of the principle which is supposed to guarantee the completeness of the table of categories.

This unavowed change in point of view is the main cause of confusion in this section. Its other defects are chiefly those of omission. Kant fails to develop in sufficient detail his view of the nature of the discursive concept, or to make sufficiently clear the grounds for his assertion that conception as an activity of the understanding is identical with judgment. To take the former point first. Kant’s mode of viewing the discursive concept finds expression in the following passage in the Introduction to his Logic:

“Human knowledge is on the side of the understanding discursive; that is, it takes place by means of ideas which make what is common to many things the ground of knowledge: and hence by means of attributes as such. We therefore cognise things only by means of attributes. An attribute is that in a thing which constitutes part of our cognition of it; or, what is the same, a partial conception so far as it is considered as a ground of cognition of the whole conception. All our concepts, therefore, are attributes, and all thought is nothing but conception by means of attributes.”

The limitations of Kant’s view of the concept could hardly find more definite expression. The only type of judgment which receives recognition is the categorical, interpreted in the traditional manner.

“To compare something as a mark with a thing, is called ‘to judge.’ The thing itself is the subject, the mark [or attribute] is the predicate. The comparison is expressed by the word ‘is,’ ... which when used without qualification indicates that the predicate is a mark [or attribute] of the subject, but when combined with the sign of negation states that the predicate is a mark opposed to the subject.”

Kant’s view of analytic thinking is entirely dominated by the substance-attribute teaching of the traditional logic. A concept must, in its connotation, be an abstracted attribute, and in its denotation represent a class. Relational thinking, and the concepts of relation, are ignored. Thus, in the Aesthetic, as we have already noted, Kant maintains that since space and time are not generic class concepts they must be intuitions. This argument, honestly employed by Kant, shows how completely unconscious he was of the revolutionary consequences of his new standpoint. Even in the very act of insisting upon the relational character of the categories, he still continues to speak of the concept as if it must necessarily conform to the generic type. In this, as in so many other respects, transcendental logic is not, as he would profess, supplementary to general logic; it is its tacit recantation. Modern logic, as developed by Lotze, Sigwart, Bradley, and Bosanquet, is, in large part, the recasting of general logic in terms of the results reached by Kant’s transcendental enquiries. Meantime, sufficient has been said to indicate the strangely limited character of Kant’s doctrine of the logical concept.

But on one fundamental point Kant breaks entirely free from the traditional logic. The following passage occurs in the above-quoted pamphlet on The Mistaken Subtlety of the Four Syllogistic Figures:

“It is clear that in the ordinary treatment of logic there is a serious error in that distinct and complete concepts are treated before judgments and ratiocinations, although the former are only possible by means of the latter.” “I say, then, first, that a distinct concept is possible only by means of a judgment, a complete concept only by means of a ratiocination. In fact, in order that a concept should be distinct, I must clearly recognise something as an attribute of a thing, and this is a judgment. In order to have a distinct concept of body, I clearly represent to myself impenetrability as an attribute of it. Now this representation is nothing but the thought, ‘a body is impenetrable.’ Here it is to be observed that this judgment is not the distinct concept itself, but is the act by which it is realised; for the idea of the thing which arises after this act is distinct. It is easy to show that a complete concept is only possible by means of a ratiocination: for this it is sufficient to refer to the first section of this essay. We might say, therefore, that a distinct concept is one which is made clear by a judgment, and a complete concept one which is made distinct by a ratiocination. If the completeness is of the first degree, the ratiocination is simple; if of the second or third degree, it is only possible by means of a chain of reasoning which the understanding abridges in the manner of a sorites.... Secondly, as it is quite evident that the completeness of a concept and its distinctness do not require different faculties of the mind (since the same capacity which recognises something immediately as an attribute in a thing is also employed to recognise in this attribute another attribute, and thus to conceive the thing by means of a remote attribute), so also it is evident that understanding and reason, that is, the power of cognising distinctly and the power of forming ratiocinations, are not different faculties. Both consist in the power of judging, but when we judge mediately we reason.”

In the section before us this same standpoint is maintained, but is expressed in a much less satisfactory manner. Concepts are no longer spoken of as complete judgments. In the above passages Kant always speaks of the concept as the subject of the proposition; it is now treated only as a predicate. This difference is significant. The concept as subject can represent the judgment as a whole (or at least it does so from the traditional standpoint to which Kant holds); the concept as predicate is merely one element, even though it be a unifying element, in the total act of judging. This falling away from his own maturer standpoint would seem to be due to Kant’s lack of clearness as to the nature of the analogy which he is here drawing between analytic and synthetic thinking. It is connected with his mistaken, and merely temporary, comparison of a priori with discursive concepts. His position in 1762 alone harmonises with his essential teaching. Now, as then, he is prepared to view judgment as the sole ultimate activity of the understanding, and therefore to define understanding as the faculty of judging.

But the new Critical standpoint compels Kant to reinterpret this definition in a manner which involves a still more radical transformation of the traditional doctrine. The categories constitute a unique type of concept, and condition the processes of discursive thought. They are embodied in the complex contents from which analytic thinking starts; and however far the processes of discursive comparison and abstraction be carried, one or other of these categories must still persist, determining the form which the analytic judgment is to take. The categorical judgment can formulate itself only by means of the a priori concept of subject and attribute, the hypothetical only by means of the pure concept of ground and consequence, and so with the others. And there are in consequence just as many categories as there are forms of the analytic judgment. This is how the principle of the metaphysical deduction must be interpreted when the later and deeper results of the transcendental deduction are properly taken into account. In deducing the forms of the understanding from the modes of discursive judgment Kant is virtually maintaining that analytic judgment involves the same problems as does judgment of the synthetic type. The categories can be derived from the forms of discursive judgment only because they are the conditions in and through which it becomes possible.

But though Kant, both here and in the central portions of the Analytic, seems to be on the very brink of this conclusion, it is never explicitly drawn. As we shall see, it would have involved the further admission that there is no absolute guarantee of the completeness of the table of categories, and no satisfactory method of determining their interrelations. To the very last general logic is isolated from transcendental logic. The Critical enquiry is formulated as if it concerned only such judgments as are explicitly synthetic. The principle of the metaphysical deduction is not, therefore, stated by Kant himself in the above manner; and we have still to decide the difficult question as to what the principle employed by Kant in the deduction actually is.

Kant makes a twofold demand upon the principle. It must enable us to discover the categories, and it must also in so doing enable us to view them as together forming a systematic whole, and so as having their completeness guaranteed by other than merely empirical considerations. The principle is stated sometimes in a broader and sometimes in a more specific form; for on this point also Kant speaks with no very certain voice. The broader formulation of the principle is that all acts of understanding are judgments, and that therefore the possible ultimate a priori forms of understanding are identical with the possible ultimate forms of the judgment. The more specific and correct formulation is that to every form of analytic judgment there corresponds a pure concept of understanding. The first statement of the principle is obviously inadequate. It merely reformulates the problem as being a problem not of conception but of judgment. If a principle is required to guarantee the completeness of our list of a priori concepts, it will equally be required to guarantee the completeness of our list of judgments. Even if the above principle be more explicitly formulated, as in the Prolegomena, where judging is defined as the act of understanding which comprises all its other acts, it will not enable us to guarantee the completeness of any list of the forms of judgment or to determine their systematic interrelation. We are therefore thrown back upon the second view. This, however, only brings us face to face with the further question, what principle guarantees the completeness of the table of analytic judgments. And to that query Kant has absolutely no answer. The reader’s questionings break vainly upon his invincible belief in the adequacy and finality of the classification yielded by the traditional logic.

The fons et origo of all the confusions and obscurities of this section are thus traceable to Kant’s attitude towards formal logic. He might criticise it for ignoring the interdependence of conception, judgment, and reasoning; he might reject the second, third, and fourth syllogistic figures; and he might even admit that its classification of the forms of judgment is not as explicit as might be desired; but however many provisos he made and defects he acknowledged, they were to him merely minor matters, and he accepted its teaching as complete and final. This unwavering faith in the fundamental distinctions of the traditional logic was indeed, as we shall have constant occasion to observe, an ever present influence in determining alike the general framework and much of the detail of Kant’s Critical teaching. The defects of the traditional logic were very clearly indicated in his own transcendental logic. He showed that synthetic thinking is fundamental; that by its distinctions the forms and activities of analytic thought are predetermined; that judgment in its various forms can be understood only by a regress upon the synthetic concepts to which these forms are due; that notions are not merely of the generic type, but that there are also categories of relation. None the less, to the very last, Kant persisted in regarding general logic as a separate discipline, and as quite adequate in its current form. He continued to ignore the fact that the analytic judgment, no less than the synthetic judgment, demands a transcendental justification.

The resulting situation is strangely perverse. In the very act of revolutionising the traditional logic, Kant relies upon its prestige and upon the assumed finality of its results to make good the shortcomings of the logic which is to displace it. By Kant’s own admission transcendental logic is incapable of guaranteeing that completeness upon which, throughout the whole Critique, so great an emphasis is laid. General logic is allowed an independent status, sufficient to justify its authority being appealed to; and the principle which is supposed to guarantee the completeness of the table of categories is so formulated as to contain no suggestion of the dependence of discursive upon synthetic thinking. Formal logic, Kant would seem to hold, can supply a criterion for the classification of the ultimate forms of judgment just because its task is relatively simple, and is independent of all epistemological views as to the nature, scope, and conditions of the thought process. Since formal logic is a completed and perfectly a priori science, which has stood the test of 2000 years, and remains practically unchanged to the present day, its results can be accepted as final, and can be employed without question in all further enquiries. Analytic thinking is scientifically treated in general logic; the Critique is concerned only with the possibility and conditions of synthetic judgment. The table of analytic judgments therefore supplies a complete and absolutely guaranteed list of the possible categories of the understanding. But the perverseness of this whole procedure is shown by the manner in which, as we shall find, Kant recasts, extends, or alters, to suit his own purposes, the actual teaching of the traditional logic.

As noted above, the asserted parallelism of analytic and synthetic judgment rests upon the further assumption that discursive thinking and synthetic interpretation are the outcome of one and the same faculty of understanding. It is implied, in accordance with the attitude of the pre-Critical Dissertation, that understanding, viewed as the faculty to which all thought processes are due, has certain laws in accordance with which it necessarily acts in all its operations, and that these must therefore be discoverable from analytic no less than from synthetic thinking. The mingling of truth and falsity in this assumption has already been indicated. Such truth as it contains is due to the fact that analytic thinking is not co-ordinate with, but is dependent upon, and determined by, the forms of synthetic thinking. Its falsity consists in its ignoring of what thus gives it partial truth. The results of the transcendental deduction call for a complete recasting of the entire argument of the metaphysical deduction. And when this is done, there is no longer any ground for the contention that the number of the categories is determinable on a priori grounds. On Kant’s own fundamental doctrine of the synthetic, and therefore merely de facto, character of all a priori principles, the necessity of the categories is only demonstrable by reference to the contingent fact of actual experience. The possible conceptual forms are relative to actual and ultimate differences in the contingent sensuous material; and being thus relative, they cannot possibly be systematised on purely a priori grounds. This Kant has himself admitted in a passage added in the second edition, though apparently without full consciousness of the important consequences which must follow.

“This peculiarity of our understanding that it can produce a priori unity of apperception solely by means of the categories, and only by such and so many, is as little capable of further explanation as why we have just these and no other functions of judgment, or why space and time are the only forms of our possible intuition.”

STAGES IN THE DEVELOPMENT OF KANT’S METAPHYSICAL DEDUCTION

The character of the metaphysical deduction will be placed in a clearer light if we briefly trace the stages, so far as they can be reconstructed, through which it passed in Kant’s mind. We may start from the Dissertation of 1770. Kant there modifies his earlier Wolffian standpoint, developing it, probably under the direct influence of the recently published Nouveaux Essais, on more genuinely Leibnizian lines.

“The use of the intellect ... is twofold. By the one use concepts, both of things and of relations, are themselves given. This is the real use. By the other use concepts, whencesoever given, are merely subordinated to each other, the lower to the higher (the common attributes), and compared with one another according to the principle of contradiction. This is called the logical use.... Empirical concepts, therefore, do not become intellectual in the real sense by reduction to greater universality, and do not pass beyond the type of sensuous cognition. However high the abstraction be carried, they must always remain sensuous. But in dealing with things strictly intellectual, in regard to which the use of the intellect is real, intellectual concepts (of objects as well as of relations), are given by the very nature of the intellect. They are not abstracted from any use of the senses, and do not contain any form of sensuous knowledge as such. We must here note the extreme ambiguity of the word abstract.... An intellectual concept abstracts from everything sensuous; it is not abstracted from things sensuous. It would perhaps be more correctly named abstracting than abstract. It is therefore preferable to call the intellectual concepts pure ideas, and those which are given only empirically abstract ideas.” “I fear, however, that Wolff, by this distinction between the sensuous and the intellectual, which for him is merely logical, has checked, perhaps wholly (to the great detriment of philosophy), that noblest enterprise of antiquity, the investigation of the nature of phenomena and noumena, turning men’s minds from such enquiries to what are very frequently only logical subtleties. Philosophy, in so far as it contains the first principles of the use of the pure intellect, is metaphysics.... As empirical principles are not to be found in metaphysics, the concepts to be met with in it are not to be sought in the senses but in the very nature of the pure intellect. They are not connate concepts, but are abstracted from laws inherent in the mind (legibus menti insitis), and are therefore acquired. Such are the concepts of possibility, existence, necessity, substance, cause, etc. with their opposites or correlates. They never enter as parts into any sensuous representation, and therefore cannot in any fashion be abstracted from such representations.”

The etcetera, with which in that last passage Kant concludes his list of pure intellectual concepts, indicates a problem that must very soon have made itself felt. That it did so, appears from his letter to Herz (February 21, 1772). He there informs his correspondent, that, in developing his Transcendentalphilosophie (the first occurrence of that title in Kant’s writings), he has

“...sought to reduce all concepts of completely pure reason to a fixed number of categories [this term also appearing for the first time], not in the manner of Aristotle, who in his ten predicaments merely set them side by side in a sort of order, just as he might happen upon them, but as they distribute themselves of themselves according to some few principles of the understanding.”

Though in this same letter Kant professes to have solved his problems, and to be in a position to publish his Critique of Pure Reason (this title is already employed) “within some three months,” the phrase “some few principles” clearly shows that he has not yet developed the teaching embodied in the metaphysical deduction. For its keynote is insistence upon the necessity of a single principle, sufficient to reduce them not merely to classes but to system. The difficulty of discovering such a principle must have been one of the causes which delayed completion of the Critique. The only data at our disposal for reconstructing the various stages through which Kant’s views may have passed in the period between February 1772 and 1781 are the Reflexionen, but they are sufficiently ample to allow of our doing so with considerable definiteness.

In the Dissertation Kant had traced the concepts of space and time, no less than the concepts of understanding, to mental activities.

“Both concepts [space and time] are undoubtedly acquired. They are not, however, abstracted from the sensing of objects (for sensation gives the matter, not the form of human cognition). As immutable types they are intuitively apprehended from the activity whereby the mind co-ordinates its sensuous data in accordance with perpetual laws.”

Now the Dissertation is quite vague as to how the “mind” (animus), active in accordance with laws generative of the intuitions space and time, differs from “understanding” (intellectus), active in accordance with laws generative of pure concepts. Kant’s reasons, apart from the intuitive character of space and time, for contrasting the former with the latter, as the sensuous with the intellectual, were the existence of the antinomies and his belief that through pure concepts the absolutely real can be known. When, however, that belief was questioned by him, and he had come to regard the categories as no less subjective than the intuitional forms, the antinomies ceased to afford any ground for thus distinguishing between them. The intuitional nature of space and time, while certainly peculiar to them, is in itself no proof that they belong to the sensuous side of the mind.

A difficulty which immediately faced Kant, from the new Critical standpoint, was that of distinguishing between space and time, on the one hand, and the categories on the other. This is borne out by the Reflexionen and by the following passage in the Prolegomena.

“Only after long reflection, expended in the investigation of the pure non-empirical elements of human knowledge, did I at last succeed in distinguishing and separating with certainty the pure elementary concepts of sensibility (space and time) from those of the understanding.”

The first stage in the development of the metaphysical deduction would seem to have consisted in the attempt to view the categories as acquired by reflection upon the activities of the understanding in “comparing, combining, or separating”; and among the notiones rationales, notiones intellectus puri, thus gained, the idea of space is specially noted. The following list is also given:

“The concepts of existence (reality), possibility, necessity, ground, unity and plurality, parts, all, none, composite and simple, space, time, change, motion, substance and accident, power and action, and everything that belongs to ontology proper.”

In Reflexionen, ii. 507 and 509, the fundamental feature of such rational concepts is found in their relational character. They all agree in being concepts of form.

Quite early, however, Kant seems to have developed the view, which has created so many more difficulties than it resolves, that space and time are given to consciousness through outer and inner sense. Though still frequently spoken of as concepts, they are definitely referred to the receptive, non-spontaneous, side of the mind. This is at once a return to the Dissertation standpoint, and a decided modification of its teaching. It holds to the point of view of the Dissertation in so far as it regards them as sensuous, and departs from it in tracing them to receptivity.

The passage quoted from the letter of 1772 to Herz may perhaps be connected with the stage revealed in the Reflexionen already cited. “Comparing, combining, and separating” may be the “some few principles of the understanding” there referred to. That, however, is doubtful, for the next stage in the development likewise resulted in a threefold division. This second stage finds varied expression in Reflexionen, ii. 483, 522, 528, 556-63. These, in so far as they agree, distinguish three classes of categories--of thesis, of analysis, and of synthesis. The first covers the categories of quality and modality, the second those of quantity, the third those of relation.

Reflexionen, ii. 528 is as follows:

(c) Connection: co-ordination: whole and part [Synthesis = ] [anticipates the later category of reciprocity], simple and compound; subordination: (1) Subject and predicate. (2) Ground and consequence.

This, and the connected Reflexionen enumerated above, are of interest as proving that Kant’s table of categories was in all essentials complete before the idea had occurred to him of further systematising it or of guaranteeing its completeness by reference to the logical classification of the forms of judgment. They also justify us in the belief that when Kant set himself to discover such a unifying principle the above list of categories and the existing logical classifications must have mutually influenced one another, each undergoing such modification as seemed necessary to render the parallelism complete. This, as we shall find, is what actually happened. The logical table, for instance, induced Kant to distinguish the categories of quality from those of modality, while numerous changes were made in the logical table itself in order that it might yield the categories required.

But the most important alteration, the introduction of the threefold division of each sub-heading, is not thus explicable, as exclusively due to one or other of the two factors. The adoption of this threefold arrangement in place of the dichotomous divisions of the logical classification and of the haphazard enumerations of Kant’s own previous lists, seems to be due to the twofold circumstance that he had already distinguished three categories of synthesis or relation (always the most important for Kant), and that this sufficiently harmonised with the logical distinction between categorical, hypothetical, and disjunctive judgments. He then sought to modify the logical divisions by addition in each case of a third, and finding that this helped him to obtain the categories required, the threefold division became for him (as it remained for Hegel) an almost mystical dogma of transcendental philosophy. In so far as it involved recognition that the hard and fast opposites of the traditional logic (such as the universal and the particular, the affirmative and the negative) are really aspects inseparably involved in every judgment and in all existence, it constituted an advance in the direction both of a deeper rationalism and of a more genuine empiricism. But in so far as it was due to the desire to guarantee completeness on a priori grounds, and so was inspired by a persistent overestimate of our a priori powers, it has been decidedly harmful. Much of the useless “architectonic” of the Critique is due to this scholastic prejudice.

This fundamental alteration in the table of logical judgments is introduced with the naive assertion that “varieties of thought in judgments,” unimportant in general logic, “may be of importance in the field of its pure a priori knowledge.” In the Critique of Judgment we find the following passage:

“It has been made a difficulty that my divisions in pure philosophy have almost always been threefold. But this lies in the nature of the case. If an a priori division is to be made, it must be either analytic, according to the principle of contradiction, and then it is always twofold (quodlibet ens est aut A aut non A); or else synthetic. And if in this latter case it is derived from a priori concepts (not as in mathematics from the a priori intuition corresponding to the concept) the division must necessarily be a trichotomy. For according to what is requisite for synthetic unity in general, there must be (1) a condition, (2) a conditioned, and (3) the concept which arises from the union of these two.”

The last stage, as expressed in the Critique, was, as we have already noted, merely an application of his earlier position that all thinking is judging. This appreciation of the inseparable connection of the categories with the act of judging is sound in principle, and is pregnant with many of the most valuable results of the Critical teaching. But these fruitful consequences follow only upon the lines developed in the transcendental deduction. They are bound up with Kant’s fundamental Copernican discovery that the categories are forms of synthesis, and accordingly express functions or relations. The categories can no longer be viewed, in the manner of the Dissertation, as yielding concepts of objects. The view of the concept which we find in the Dissertation is, indeed, applied in the Critique to space and time--they are taken as in themselves intuitions, not as merely forms of intuition--but the categories are recognised as being of an altogether relational character. Though a priori, they are not, in and by themselves, complete objects of consciousness, and accordingly can reveal no object. They are functions, not contents. That, however, is to anticipate. We must first discharge, as briefly as possible, the ungrateful task of dwelling further upon the laboured, arbitrary, and self-contradictory character of the detailed working out of the metaphysical deduction. The deduction is given in Sections II. and III.

=Section II. The Logical Function of the Understanding in Judgment.=--Kant’s introductory statement may here be noted. If, he says, we leave out of consideration the content of any judgment, and attend only to the mere form, we “find” that the function of thought in a judgment “can” be brought under four heads, each with three subdivisions. But Kant himself, in this same section, recognises in the frankest and most explicit manner, that the necessary distinctions are only to be obtained by taking account of the matter as well as of the form of judgments. And even after this contradiction is discounted, the term “find” may be allowed as legitimate only if the word “can” is correspondingly emphasised. The distinctions were not derived from any existing logic. They were reached only by the freest possible handling of the classifications currently employed. Examination of the table of judgments, and comparison of it with the table of categories, supply conclusive evidence that the former has been rearranged, in highly artificial fashion, so as to yield a more or less predetermined list of required categories.

=1. Quantity.=--Kant here frankly departs from the classification of judgments followed in formal logic; and the reason which he gives for so doing is in direct contradiction to his demand that only the form of judgment must be taken into account. The “quantity of knowledge” here referred to is determinable, not from the form, but only from the content of the judgment. Also, the statement that the singular judgment stands to the universal as unity to infinity (Unendlichkeit) is decidedly open to question. The universal is itself a form of unity, as Kant virtually admits in deriving, as he does, the category of unity from the universal judgment.

=2. Quality.=--Kant makes a similar modification in the logical treatment of quality, by distinguishing between affirmative and infinite judgments. The proposition, A is not-B, is to be viewed as neither affirmative nor negative. As the content of the predicate includes the infinite number of things that are not-B, the judgment is infinite. Kant, in a very artificial and somewhat arbitrary manner, contrives to define it as limitative in character, and so as sharing simultaneously in the nature both of affirmation and of negation. The way is thus prepared for the “discovery” of the category of limitation.

=3. Relation.=--Wolff, Baumgarten, Meier, Baumeister, Reimarus, and Lambert, with very minor differences, agree in the following division:

{Simple = Categorical { Judgments{ {Copulative (i.e. categorical with more { {than one subject or more than one {Complex {predicate). {Hypothetical. {Disjunctive.

Kant omits the copulative judgment, and by ignoring the distinction between simple and complex judgments (which in Reimarus, and also less definitely in Wolff, is connected with the distinction between conditional and unconditional judgments) contrives to bring the remaining three types of judgment under the new heading of “relation.” They had never before been thus co-ordinated, and had never before been subsumed under this particular title. It is by no means clear why such distinctions as those between simple and complex, conditioned and unconditioned, should be ignored, and why the copulative judgment should not be recognised as well as the hypothetical. Kant’s criterion of importance and unimportance in the distinctions employed by the logicians of his day was wholly personal to himself; and, though hard to define, was certainly not dictated by any logic that is traceable to Aristotelian sources. His exposition is throughout controlled by foreknowledge of the particular categories which he desires to “discover.”

=4. Modality.=--Neither Wolff nor Reimarus gives any account of modality. Baumgarten classifies judgments as pure or modal (existing in four forms, necessity, contingency, possibility, impossibility). Baumeister and Thomasius also recognise four forms of modality. Meier distinguishes between pure judgment (judicium purum) and impure judgment (judicium modale, modificatum, complexum qua copula), but does not classify the forms of modality. Lambert alone classifies judgments as “possible, actual (wirklich), necessary, and their opposite.” But when Kant adopts this threefold division, the inclusion of actuality renders the general title “modality” inapplicable in its traditional sense. The expression of actuality in the assertoric judgment involves no adverbial modification of the predicate. Also, in its “affirmative” and “categorical” forms it has already been made to yield two other categories.

Kant speaks of the problematic, the assertoric, and the apodictic forms of judgment as representing the stages through which knowledge passes in the process of its development.

“These three functions of modality are so many momenta of thought in general.”

This statement has been eulogised by Caird, as being an anticipation of the Hegelian dialectic. As a matter of fact, Kant’s remark is irrelevant and misleading. The advance from consciousness of the problematic, through determination of it as actual to its explanation as necessary, represents only a psychological order in the mind of the individual. Logically, knowledge of the possible rests on and implies prior knowledge of the actual and of the necessities that constitute the actual.

=Section III. The Categories or Pure Concepts of the Understanding.=--The first three pages of this section, beginning “General logic abstracts,” and concluding with the word “rest on the understanding,” would seem to be a later interpolation. Embodying, as they do, some of the fundamental ideas of the transcendental deduction, they express Kant’s final method of distinguishing between general and transcendental logic. But they are none the less out of harmony with the other sections of the metaphysical deduction. They are of the nature of an after-thought, even though that afterthought represents a more mature and adequate standpoint. In A 55-7, where Kant defines the distinction between general and transcendental logic, the latter is formulated in entire independence of all reference to pure intuition. Kant, indeed, argues that just as there are both pure and empirical intuitions, so there are both pure and empirical concepts. But there is no indication that he has yet realised the close interdependence of the two types of a priori elements. Even when he proceeds in A 62 to remark that the empirical employment of pure concepts is conditioned by the fact that objects are given in intuition, no special reference is made to “the manifold of pure a priori intuition.” Now, however, Kant emphasises, as the fundamental characteristic of transcendental logic, its possession of a pure manifold through reference to which its pure concepts gain meaning. Thus not only does transcendental logic not abstract from the pure a priori concepts, it likewise possesses an a priori material. It is in this twofold manner that it is now regarded as differing from formal logic.

The accounts given of the metaphysical deduction by Cohen, Caird, Riehl, and Watson are vitiated by failure to remark that this latter standpoint is a late development, and is out of keeping with the rest of the deduction. Riehl’s exposition has, however, the merit of comparative consistency. He explicitly recognises the important consequence which at once follows from acceptance of this later view, namely, that it is by their implying space and time that the categories differ from the notions which determine the forms of judgment; in other words, that the categories are actualised only as schemata. The category of substance, for instance, differs from the merely logical notion of a propositional subject, in being the concept of that which is always a subject, and never a predicate; and such a conception has specific meaning for us only as the permanent in time. Logical subjects and predicates, quantitative relations apart, are interchangeable. The relation between them is the analytic relation of identity. The concept of subject, on the other hand, transcendentally viewed, that is, as a category, is the apprehension of what is permanent, in synthetic distinction from, and relation to, its changing attributes. In other words, the transcendental distinction between substance and accidents is substituted for that of subject and predicate. Similarly the logical relation of ground and consequence, conceived as expressive of logical identity, gives way to the synthetic temporal relation of cause and effect. And so with all the other pure forms. As categories, they are schemata. Kant has virtually recognised this by the names which he gives to the categories of relation. But the proper recognition of the necessary interdependence of the intuitional and conceptual forms came too late to prevent him from distinguishing between categories and schemata, and so from creating for himself the artificial difficulties of the section on schematism.

In A 82 Kant states that he intentionally omits definitions of the categories. He had good reason for so doing. The attempt would have landed him in manifold difficulties, since his views were not yet sufficiently ripe to allow of his perceiving the way of escape. In A 241 (omitted in second edition) Kant makes, however, the directly counter statement that definition of the categories is not possible, giving as his reason that, in isolation from the conditions of sensibility, they are merely logical functions, “without the slightest indication as to how they can possess meaning and objective validity.”

It cannot be too often repeated that the Critique is not a unitary work, but the patchwork record of twelve years of continuous development. Certain portions of the transcendental deduction, of which A 76-9 is one, represent the latest of all the many stages; and their teaching, when accepted, calls for a radical recasting of the metaphysical deduction. The bringing of the entire Critique into line with its maturest parts would have been an Herculean task; and it was one to which Kant, then fifty-seven years of age, was very rightly unwilling to sacrifice the time urgently needed for the writing of his other Critiques. The passage before us is one of the many interpolations by which Kant endeavoured to give an external unity to what, on close study, is found to be the plain record of successive and conflicting views. Meantime, in dealing with this passage, we are concerned only to note that if this later mode of defining transcendental logic be accepted, far-reaching modifications in Kant’s Critical teaching have to be made. The other points developed in A 76-9 we discuss below in their proper connection.

=The same Function, etc.=--This passage has already been sufficiently commented upon. Kant here expresses in quite inadequate fashion the standpoint of the transcendental deduction. The implication is that analytic and synthetic thinking are co-ordinate, one and the same faculty exercising, on these two levels, the same operations. The true Critical teaching is that synthetic thinking is alone fundamental, and that only by a regress upon it can judgments be adequately accounted for. This passage, like the preceding, may be of later origin than the main sections of the metaphysical deduction.

=Term “Categories” borrowed from Aristotle.=--Cf. below, p. 198.

=Table of Categories. Quantity.=--Kant derives the category of unity from the universal, and that of totality (Allheit) from the singular. These derivations are extremely artificial. In Reflexionen, ii. 563, Kant takes the more natural line of identifying totality with the universal, and unity with the singular. Probably the reason of Kant’s change of view is the necessity of obtaining totality by combining unity with multiplicity. That can only be done if universality is thus equated with unity. Watson’s explanation, that Kant has reversed the order of the categories, seems to be erroneous.

=Quality.=--Cf. above, p. 192.

=Relation.=--The correlation of the categorical judgment with the conception of substance and attribute is only possible owing to Kant’s neglect of the relational judgment and to the dominance in his logical teaching of the Aristotelian substance-attribute view of predication. The correlation is also open to question in that the relation of subject and predicate terms in a logical judgment is a reversible one. It is a long step from the merely grammatical subject to the conception of that which is always a subject and never a predicate.

Kant’s identification of the category of community or reciprocity with the disjunctive judgment, though at first sight the most arbitrary of all, is not more so than many of the others. Its essential correctness has been insisted upon in recent logic by Sigwart, Bradley, and Bosanquet. In Kant’s own personal view co-ordination in the form of co-existence is only possible through reciprocal interaction. The relation of whole and part (the parts in their relations of reciprocal exclusion exhausting and constituting a genuine whole) thus becomes, in its application to actual existences, that of reciprocal causation. The reverse likewise holds; interaction is only possible between existences which together constitute a unity. Kant returns to this point in Note 3, added in the second edition. The objection which Kant there considers has been very pointedly stated by Schopenhauer.

“What real analogy is there between the problematical determination of a concept by disjunctive predicates and the thought of reciprocity? The two are indeed absolutely opposed, for in the disjunctive judgment the actual affirmation of one of the two alternative propositions is also necessarily the negation of the other; if, on the other hand, we think of two things in the relation of reciprocity, the affirmation of one is also necessarily the affirmation of the other, and vice versa.”

The answer to this criticism is on the lines suggested by Kant. The various judgments which constitute a disjunction do not, when viewed as parts of the disjunction, merely negate one another; they mutually presuppose one another in the total complex. Schopenhauer also fails to observe that in locating the part of a real whole in one part of space, we exclude it from all the others.

=Modality.=--The existence of separate categories of modality seems highly doubtful. The concepts of the possible and of the probable may be viewed as derivative; the notion of existence does not seem to differ from that of reality; and necessity seems in ultimate analysis to reduce to the concept of ground and consequence. These are points which will be discussed later.

Aristotle’s ten categories are enumerated by Kant in Reflexionen, ii. 522, as: (1) substantia, accidens, (2) qualitas, (3) quantitas, (4) relatio, (5) actio, (6) passio, (7) quando, (8) ubi, (9) situs, (10) habitus; and the five post-predicaments as: oppositum, prius, simul, motus, habere. Eliminating quando, ubi, situs, prius, and simul as being modes of sensibility; actio and passio as being complex and derivative; and also omitting habitus (condition) and habere, as being too general and indefinite in meaning to constitute separate categories; we are then left with substantia, qualitas, quantitas, relatio, and oppositum. The most serious defect in this reduced list, from the Kantian point of view, is its omission of causality. It is, however, a curious coincidence that when substance is taken as a form of relatio, and oppositum as a form of quality, we are left with the three groups, quality, quantity, relation. Only modality is lacking to complete Kant’s own fourfold grouping. None the less, as the study of Kant’s Reflexionen sufficiently proves, it was by an entirely different route that Kant travelled to his metaphysical deduction. Watson does not seem to have any ground for his contention that the above modified list of Aristotle’s categories “gave Kant his starting-point.” It was there indeed, as the reference to Aristotle in his letter of 1772 to Herz shows, that he first looked for assistance, only, however, to be disappointed in his expectations.

=Derivative concepts.=--Cf. above, pp. 66, 71-2.

=I reserve this task for another occasion.=--Cf. A 204 = B 249; A 13; above, p. 66 ff., and below, pp. 379-80.

=Definitions of categories omitted.=--Cf. above, pp. 195-6, and A 241 there cited; also below, pp. 339-42, 404-5.

=Note 1.=--On this distinction between mathematical and dynamical categories cf. below, pp. 345-7, 510-11.

=Note 2.=--This remark is inserted to meet a criticism which had been made by Johann Schulze, and to which Kant in February 1784 had replied in terms almost identical with those of the present passage.

“The third category certainly springs from the connection of the first and second, not, indeed, from their mere combination, but from a connection the possibility of which constitutes a concept that is a special category. For this reason the third category may not be applicable in instances in which the other two apply: e.g. one year, many years of future time, are real concepts, but the totality of future years, that is, the collective unity of a future eternity, conceived as entire (so to say, as completed), is something that cannot be thought. But even in those cases in which the third category is applicable, it always contains something more than the first and the second taken separately and together, namely the derivation of the second from the first, a process which is not always practicable. Necessity, for example, is nothing else than existence, in so far as it can be inferred from possibility. Community is the reciprocal causality of substances in respect of their determinations. But that determinations of one substance can be produced by another substance, is something that we may not simply assume; it is one of those connections without which there could be no reciprocal relation of things in space, and therefore no outer experience. In a word, I find that just as the conclusion of a syllogism indicates, in addition to the operations of understanding and judgment in the premisses, a special operation peculiar to reason ..., so also the third category is a special, and in part original, concept. For instance, the concepts, quantum, compositum, totum, come under the categories unity, plurality, totality, but a quantum thought as compositum would not yield the concept of totality unless the concept of the quantum is thought as determinable through the composition, and in certain quanta, such as infinite space, that cannot be done.”

Kant’s assertion that in certain cases the third category is not applicable is misleading. His proof of the validity of the category of reciprocity in the third Analogy really consists in showing that it is necessary to the apprehension of spatial co-existence; and if, as Kant maintains, consciousness of space is necessary to consciousness of time, it is thereby proved to be involved in each and every act of consciousness. It is presupposed in the apprehension even of substantial existence and of causal sequence. His proof that it is a unique category, distinct from the mere combination of the categories of substance and causality, does not, therefore, assume what his words in the above letter would seem to imply, that it is only occasionally employed. The same remark holds in regard to totality; it is presupposed even in the apprehension of a single year. Kant’s references, both here and in other parts of the Critique, to totality in its bearing upon the conception of infinitude, reveal considerable lack of clearness as to the relation in which it stands to the Idea of the unconditioned. Sometimes, as in this letter, he would seem to be identifying them; elsewhere this confusion is avoided. In B 111 totality is defined as multiplicity regarded as unity, and in A 142-3 = B 182 its schema is defined as number. (The identification of totality with number has led Kant to say in B 111 that number is not applicable in the representation of the infinite, a much more questionable assertion than that of the letter above quoted.) The statement that necessity is existence in so far as it can be inferred from possibility, or that it is existence given through possibility, is similarly misleading. Kant’s true position is that all three are necessary to the conception of any one of the three.

Thus Kant’s reply to Schulze, alike in his letter and in Note 2, fails to indicate with any real adequacy the true bearing of Critical teaching in this matter; and consequently fails to reveal the full force of his position. Only in terms of totality can unity and plurality be apprehended; only through the reciprocal relations which determine co-existence can we acquire consciousness of either permanence or sequence; only in terms of necessity can either existence or possibility be defined. The third category is not derived from a prior knowledge of the subordinate categories. It represents in each case a higher complex within which alone the simpler relations defined by the simpler concepts can exist or have meaning.

=B 113-16, § 12.=--This section, of no intrinsic importance, is an example of Kant’s loving devotion to this “architectonic.” His reasoning is extremely artificial, especially in its attempt to connect “unity, truth, and perfection” with the three categories of quantity. The Reflexionen show how greatly Kant was preoccupied with these three concepts, seeking either to base a table of categories upon them (B. Erdmann’s interpretation), or to reduce them to categories (Adickes’ interpretation). For some time Kant himself ranked with those who “incautiously made these criteria of thought to be properties of the things in themselves.” In Reflexionen, ii. 903, we find the following statement: “Unity (connection, agreement), truth (quality), completeness (quantity).” In ii. 916 Kant makes trial to connect them, as conceptions of possibility, with the categories of relation. In ii. 911 and 912 the later view, that they are logical in character and function, appears, but leads to their being set in relation to the three faculties of understanding, judgment, and reason. This is conjectured by B. Erdmann to have been Kant’s view at the time of the first edition. ii. 915, 919, 920 present the view expounded in the section before us. Erdmann remarks that in this section Kant “is settling accounts with certain thoughts which in the ’seventies had yielded suggestions for the transformation of ontology into the transcendental analytic.”