The Online Library of Liberty

A project of Liberty Fund, Inc.

• Dedication
• Preface 1
• Translator’s Preface
• Translator’s Preface to Second Edition
• Introduction [p. 1]
• I: The Idea of Transcendental Philosophy
• II: Division of Transcendental Philosophy
• I: The Elements of Transcendentalism
• Part I: Transcendental Æsthetic [p. 19]
• Section I: Of Space
• Section II: Of Time
• Part II: Transcendental Logic [p. 50]
• Introduction the Idea of a Transcendental Logic
• I: Of Logic In General
• II: Of Transcendental Logic
• III: Of the Division of General Logic Into Analytic and Dialectic
• IV: Of the Division of Transcendental Logic Into Transcendental Analytic and Dialectic
• Division I: Transcendental Analytic In Two Books, With Their Chapters and Sections
• Book I: Analytic of Concepts
• Chapter I: Method of Discovering All Pure Concepts of the Understanding
• Section I: Of the Logical Use of the Understanding In General
• Section II: Of the Logical Function of the Understanding In Judgments [p. 70]
• Section III: Of the Pure Concepts of the Understanding, Or of the Categories
• Chapter II: Of the Deduction of the Pure Concepts of the Understanding [p. 84]
• Section I: Of the Principles of a Transcendental Deduction In General
• Section II: Of the a Priori Grounds For the Possibility of Experience
• I: Of the Synthesis of Apprehension In Intuition
• II: Of the Synthesis of Reproduction In Imagination
• III: Of the Synthesis of Recognition In Concepts [p. 103]
• IV: Preliminary Explanation of the Possibility of the Categories As Knowledge a Priori
• Section III: Of the Relation of the Understanding to Objects In General, and the Possibility of Knowing Them a Priori [p. 115]
• Book II: Analytic of Principles
• Introduction of the Transcendental Faculty of Judgment In General
• Chapter I: Of the Schematism of the Pure Concepts of the Understanding [p. 137]
• Chapter II: System of All Principles of the Pure Understanding [p. 148]
• Section I: Of the Highest Principle of All Analytical Judgments
• Section II: Of the Highest Principle of All Synthetical Judgments [p. 154]
• Section III: Systematical Representation of All Synthetical Principles of the Understanding
• I: [of the Axioms of Intuition 1 Principle of the Pure Understanding
• II: [ Anticipations of Perception
• III: [ the Analogies of Experience
• IV: The Postulates of Empirical Thought In General
• Chapter III: On the Ground of Distinction of All Subjects Into Phenomena and Noumena
• I.: Identity and Difference
• II.: Agreement and Opposition
• III.: The Internal and the External
• IV.: Matter and Form
• Note On the Amphiboly of Reflective Concepts
• Division II: Transcendental Dialectic In Two Books, With Their Chapters and Sections [p. 293]
• Introduction: 1.
• Book I: Of the Concepts of Pure Reason [p. 310]
• First Section of Ideas In General
• Second Section of Transcendental Ideas [p. 321]
• Third Section System of Transcendental Ideas [p. 333]
• Book II: Of the Dialectical Conclusions of Pure Reason
• Chapter I: Of the Paralogisms of Pure Reason [p. 341]
• Chapter II: The Antinomy of Pure Reason
• Section I: System of Cosmological Ideas
• Section II: Antithetic of Pure Reason
• Section III: Of the Interest of Reason In These Conflicts [p. 462]
• Section IV: Of the Transcendental Problems of Pure Reason, and the Absolute Necessity of Their Solution
• Section V: Sceptical Representation of the Cosmological Questions In the Four Transcendental Ideas [p. 485]
• Section VI: Transcendental Idealism As the Key to the Solution of Cosmological Dialectic
• Section VII: Critical Decision of the Cosmological Conflict of Reason With Itself
• Section VIII: The Regulative Principle of Pure Reason With Regard to the Cosmological Ideas [p. 508]
• Section IX: Of the Empirical Use of the Regulative Principle of Reason With Regard to All Cosmological Ideas
• Chapter III: The Ideal of Pure Reason
• Section I: Of the Ideal In General
• Section II: Of the Transcendental Ideal ( Prototypon Transcendentale )
• Section III: Of the Arguments of Speculative Reason In Proof of the Existence of a Supreme Being
• Section IV: Of the Impossibility of an Ontological Proof of the Existence of God [p. 592]
• Section V: Of the Impossibility of a Cosmological Proof of the Existence of God [p. 603]
• Section VI: Of the Impossibility of the Physico-theological Proof
• Section VII: Criticism of All Theology Based On Speculative Principles of Reason [p. 631]
• Chapter I: The Discipline of Pure Reason
• Section I: The Discipline of Pure Reason In Its Dogmatical Use
• Section II: The Discipline of Pure Reason In Its Polemical Use
• Section III: The Discipline of Pure Reason With Regard to Hypotheses
• Section IV: The Discipline of Pure Reason With Regard to Its Proofs
• Chapter II: The Canon of Pure Reason [p. 795]
• First Section: of the Ultimate Aim of the Pure Use of Our Reason
• Section II: Of the Ideal of the Summum Bonum As Determining the Ultimate Aim of Pure Reason
• Section III: Of Trowing, Knowing, and Believing [p. 820]
• Chapter III: The Architectonic of Pure Reason [p. 832]
• Chapter IV: The History of Pure Reason [p. 852]
• Supplement I: Motto to Second Edition
• Supplement II: Preface to the Second Edition. 1787. [p. Vii]
• Supplement III: Supplement Iv
• Supplement V
• Supplement Vi
• Supplement V
• Supplement Vi
• Supplement Vii
• Supplement Viii
• Supplement Ix
• Supplement X
• Supplement Xi
• Supplement Xii
• Supplement Xiii
• Supplement Xiv
• Supplement Xv
• Supplement Xvi a
• Supplement Xvi B
• Supplement Xvii
• Supplement Xviii
• Supplement Xix
• Supplement Xx
• Supplement Xxi
• Supplement Xxii
• Supplement Xxiii
• Supplement Xxiv
• Supplement Xxv
• Supplement Xxvi
• Supplement Xxvii
• Supplement Xxviii

BOOK I

ANALYTIC OF CONCEPTS

By Analytic of concepts I do not understand their analysis, or the ordinary process in philosophical disquisitions of dissecting any given concepts according to their contents, and thus rendering them more distinct; but a hitherto seldom attempted dissection of the faculty of the understanding itself, with the sole object of discovering the possibility of concepts a priori, by looking for them nowhere but in the understanding itself [p. 66] as their birthplace, and analysing the pure use of the understanding. This is the proper task of a transcendental philosophy, all the rest is mere logical treatment of concepts. We shall therefore follow up the pure concepts to their first germs and beginnings in the human understanding, in which they lie prepared, till at last, on the occasion of experience, they become developed, and are represented by the same understanding in their full purity, freed from all inherent empirical conditions.

CHAPTER I

METHOD OF DISCOVERING ALL PURE CONCEPTS OF THE UNDERSTANDING

When we watch any faculty of knowledge, different concepts, characteristic of that faculty, manifest themselves according to different circumstances, which, as the observation has been carried on for a longer or shorter time, or with more or less accuracy, may be gathered up into a more or less complete collection. Where this collection will be complete, it is impossible to say beforehand, when we follow this almost mechanical process. Concepts thus discovered fortuitously only, possess neither order nor systematic unity, but [p. 67] are paired in the end according to similarities, and, according to their contents, arranged as more or less complex in various series, which are nothing less than systematical, though to a certain extent put together methodically.

Transcendental philosophy has the advantage, but also the duty of discovering its concepts according to a fixed principle. As they spring pure and unmixed from the understanding as an absolute unity, they must be connected with each other, according to one concept or idea. This connection supplies us at the same time with a rule, according to which the place of each pure concept of the understanding and the systematical completeness of all of them can be determined a priori, instead of being dependent on arbitrary choice or chance.

Section I

Of the Logical Use of the Understanding in General

We have before defined the understanding negatively only, as a non-sensuous faculty of knowledge. As without sensibility we cannot have any intuition, [p. 68] it is clear that the understanding is not a faculty of intuition. Besides intuition, however, there is no other kind of knowledge except by means of concepts. The knowledge therefore of every understanding, or at least of the human understanding, must be by means of concepts, not intuitive, but discursive. All intuitions, being sensuous, depend on affections, concepts on functions. By this function I mean the unity of the act of arranging different representations under one common representation. Concepts are based therefore on the spontaneity of thought, sensuous intuitions on the receptivity of impressions. The only use which the understanding can make of these concepts is to form judgments by them. As no representation, except the intuitional, refers immediately to an object, no concept is ever referred to an object immediately, but to some other representation of it, whether it be an intuition, or itself a concept. A judgment is therefore a mediate knowledge of an object, or a representation of a representation of it. In every judgment we find a concept applying to many, and comprehending among the many one single representation, which is referred immediately to the object. Thus in the judgment that all bodies are divisible,1 the concept of divisible applies to various other concepts, but is here applied in particular to the concept of body, and this concept of body to certain phenomena of our experience. [p. 69] These objects therefore are represented mediately by the concept of divisibility. All judgments therefore are functions of unity among our representations, the knowledge of an object being brought about, not by an immediate representation, but by a higher one, comprehending this and several others, so that many possible cognitions are collected into one. As all acts of the understanding can be reduced to judgments, the understanding may be defined as the faculty of judging. For we saw before that the understanding is the faculty of thinking, and thinking is knowledge by means of concepts, while concepts, as predicates of possible judgments, refer to some representation of an object yet undetermined. Thus the concept of body means something, for instance, metal, which can be known by that concept. It is only a concept, because it comprehends other representations, by means of which it can be referred to objects. It is therefore the predicate of a possible judgment, such as, that every metal is a body. Thus the functions of the understanding can be discovered in their completeness, if it is possible to represent the functions of unity in judgments. That this is possible will be seen in the following section.

Section II

Of the Logical Function of the Understanding in Judgments [p. 70]

If we leave out of consideration the contents of any judgment and fix our attention on the mere form of the understanding, we find that the function of thought in a judgment can be brought under four heads, each of them with three subdivisions. They may be represented in the following table:—

As this classification may seem to differ in some, though not very essential points, from the usual technicalities of logicians, the following reservations against any [p. 71] possible misunderstanding will not be out of place.

1. Logicians are quite right in saying that in using judgments in syllogisms, singular judgments may be treated like universal ones. For as they have no extent at all, the predicate cannot refer to part only of that which is contained in the concept of the subject, and be excluded from the rest. The predicate is valid therefore of that concept, without any exception, as if it were a general concept, having an extent to the whole of which the predicate applies. But if we compare a singular with a general judgment, looking only at the quantity of knowledge conveyed by it, the singular judgment stands to the universal judgment as unity to infinity, and is therefore essentially different from it. It is therefore, when we consider a singular judgment (judicium singulare), not only according to its own validity, but according to the quantity of knowledge which it conveys, as compared with other kinds of knowledge, that we see how different it is from general judgments (judicia communia), and how well it deserves a separate place in a complete table of the varieties of thought in general, though not in a logic limited to the use of judgments in reference to each other.

2. In like manner infinite judgments must, in transcendental logic, be distinguished from affirmative ones, though in general logic they are properly classed together, and do not constitute a separate part in [p. 72] the classification. General logic takes no account of the contents of the predicate (though it be negative), it only asks whether the predicate be affirmed or denied. Transcendental logic, on the contrary, considers a judgment according to the value also or the contents of a logical affirmation by means of a purely negative predicate, and asks how much is gained by that affirmation, with reference to the sum total of knowledge. If I had said of the soul, that it is not mortal, I should, by means of a negative judgment, have at least warded off an error. Now it is true that, so far as the logical form is concerned, I have really affirmed by saying that the soul is non-mortal, because I thus place the soul in the unlimited sphere of non-mortal beings. As the mortal forms one part of the whole sphere of possible beings, the non-mortal the other, I have said no more by my proposition than that the soul is one of the infinite number of things which remain, when I take away all that is mortal. But by this the infinite sphere of all that is possible becomes limited only in so far that all that is mortal is excluded from it, and that afterwards the soul is placed in the remaining part of its original extent. This part, however, even after its limitation, still remains infinite, and several more parts of it may be taken away without extending thereby in the least the concept of the soul, or affirmatively determining [p. 73] it. These judgments, therefore, though infinite in respect to their logical extent, are, with respect to their contents, limitative only, and cannot therefore be passed over in a transcendental table of all varieties of thought in judgments, it being quite possible that the function of the understanding exercised in them may become of great importance in the field of its pure a priori knowledge.

3. The following are all the relations of thought in judgments:—

a. Relation of the predicate to the subject.

b. Relation of the cause to its effect.

c. Relation of subdivided knowledge, and of the collected members of the subdivision to each other.

In the first class of judgments we consider two concepts, in the second two judgments, in the third several judgments in their relation to each other. The hypothetical proposition, if perfect justice exists, the obstinately wicked is punished, contains really the relation of two propositions, namely, there is a perfect justice, and the obstinately wicked is punished. Whether both these propositions are true remains unsettled. It is only the consequence which is laid down by this judgment.

The disjunctive judgment contains the relation of two or more propositions to each other, but not as a consequence, but in the form of a logical opposition, the sphere of the one excluding the sphere of the other, and at the same time in the form of community, all the propositions together filling the whole sphere of the intended knowledge. The disjunctive judgment contains therefore [p. 74] a relation of the parts of the whole sphere of a given knowledge, in which the sphere of each part forms the complement of the sphere of the other, all being contained within the whole sphere of the subdivided knowledge. We may say, for instance, the world exists either by blind chance, or by internal necessity, or by an external cause. Each of these sentences occupies a part of the sphere of all possible knowledge with regard to the existence of the world, while all together occupy the whole sphere. To take away the knowledge from one of these spheres is the same as to place it into one of the other spheres, and to place it in one sphere is the same as to take it away from the others. There exists therefore in disjunctive judgments a certain community of the different divisions of knowledge, so that they mutually exclude each other, and yet thereby determine in their totality the true knowledge, because, if taken together, they constitute the whole contents of one given knowledge. This is all I have to observe here for the sake of what is to follow hereafter.

4. The modality of judgments is a very peculiar function, for it contributes nothing to the contents of a judgment (because, besides quantity, quality, and relation, there is nothing else that could constitute the contents of a judgment), but refers only to the nature of the copula in relation to thought in general. Problematical judgments are those in which affirmation or negation are taken as possible (optional) only, while in assertory judgments affirmation or negation is taken as real (true), in apodictic as necessary.1 Thus the two judgments, [p. 75] the relation of which constitutes the hypothetical judgment (antecedens et consequens) and likewise the judgments the reciprocal relation of which forms the disjunctive judgment (members of subdivision), are always problematical only. In the example given above, the proposition, there exists a perfect justice, is not made as an assertory, but only as an optional judgment, which may be accepted or not, the consequence only being assertory. It is clear therefore that some of these judgments may be wrong, and may yet, if taken problematically, contain the conditions of the knowledge of truth. Thus, in our disjunctive judgment, one of its component judgments, namely, the world exists by blind chance, has a problematical meaning only, on the supposition that some one might for one moment take such a view, but serves, at the same time, like the indication of a false road among all the roads that might be taken, to find out the true one. The problematical proposition is therefore that which expresses logical (not objective) possibility only, that is, a free choice of admitting such a proposition, and a purely optional admission of it into the understanding. The assertory proposition implies logical reality or truth. Thus, for instance, in a hypothetical syllogism the antecedens in the major is problematical, in the [p. 76] minor assertory, showing that the proposition conforms to the understanding according to its laws. The apodictic proposition represents the assertory as determined by these very laws of the understanding, and therefore as asserting a priori, and thus expresses logical necessity. As in this way everything is arranged step by step in the understanding, inasmuch as we begin with judging problematically, then proceed to an assertory acceptation, and finally maintain our proposition as inseparably united with the understanding, that is as necessary and apodictic, we may be allowed to call these three functions of modality so many varieties or momenta of thought.

Section III

Of the Pure Concepts of the Understanding, or of the Categories

General logic, as we have often said, takes no account of the contents of our knowledge, but expects that representations will come from elsewhere in order to be turned into concepts by an analytical process. Transcendental logic, on the contrary, has before it the manifold contents of sensibility a priori, supplied by transcendental [p. 77] æsthetic as the material for the concepts of the pure understanding, without which those concepts would be without any contents, therefore entirely empty. It is true that space and time contain what is manifold in the pure intuition a priori, but they belong also to the conditions of the receptivity of our mind under which alone it can receive representations of objects, and which therefore must affect the concepts of them also. The spontaneity of our thought requires that what is manifold in the pure intuition should first be in a certain way examined, received, and connected, in order to produce a knowledge of it. This act I call synthesis.

In its most general sense, I understand by synthesis the act of arranging different representations together, and of comprehending what is manifold in them under one form of knowledge. Such a synthesis is pure, if the manifold is not given empirically, but a priori (as in time and space). Before we can proceed to an analysis of our representations, these must first be given, and, as far as their contents are concerned, no concepts can arise analytically. Knowledge is first produced by the synthesis of what is manifold (whether given empirically or a priori). That knowledge may at first be crude and confused and in need of analysis, but it is synthesis which really collects the elements of knowledge, and unites them to a certain extent. It is therefore the first thing which we [p. 78] have to consider, if we want to form an opinion on the first origin of our knowledge.

We shall see hereafter that synthesis in general is the mere result of what I call the faculty of imagination, a blind but indispensable function of the soul, without which we should have no knowledge whatsoever, but of the existence of which we are scarcely conscious. But to reduce this synthesis to concepts is a function that belongs to the understanding, and by which the understanding supplies us for the first time with knowledge properly so called.

Pure synthesis in its most general meaning gives us the pure concept of the understanding. By this pure synthesis I mean that which rests on the foundation of what I call synthetical unity a priori. Thus our counting (as we best perceive when dealing with higher numbers) is a synthesis according to concepts, because resting on a common ground of unity, as for instance, the decade. The unity of the synthesis of the manifold becomes necessary under this concept.

By means of analysis different representations are brought under one concept, a task treated of in general logic; but how to bring, not the representations, but the pure synthesis of representations, under concepts, that is what transcendental logic means to teach. The first that must be given us a priori for the sake of knowledge of all objects, is the manifold in pure intuition. The second is, the synthesis of the manifold by means of [p. 79] imagination. But this does not yet produce true knowledge. The concepts which impart unity to this pure synthesis and consist entirely in the representation of this necessary synthetical unity, add the third contribution towards the knowledge of an object, and rest on the understanding.

The same function which imparts unity to various representations in one judgment imparts unity likewise to the mere synthesis of various representations in one intuition, which in a general way may be called the pure concept of the understanding. The same understanding, and by the same operations by which in concepts it achieves through analytical unity the logical form of a judgment, introduces also, through the synthetical unity of the manifold in intuition, a transcendental element into its representations. They are therefore called pure concepts of the understanding, and they refer a priori to objects, which would be quite impossible in general logic.

In this manner there arise exactly so many pure concepts of the understanding which refer a priori to objects of intuition in general, as there were in our table logical functions in all possible judgments, because those functions completely exhaust the understanding, and comprehend every one of its faculties. Borrowing a term of Aristotle, we shall call these concepts categories, [p. 80] our intention being originally the same as his, though widely diverging from it in its practical application.

This then is a list of all original pure concepts of synthesis, which belong to the understanding a priori, and for which alone it is called pure understanding; for it is by them alone that it can understand something in the manifold of intuition, that is, think an object in it. The classification is systematical, and founded on a common principle, namely, the faculty of judging (which is the same as the faculty of thinking). It is not the [p. 81] result of a search after pure concepts undertaken at haphazard, the completeness of which, as based on induction only, could never be guaranteed. Nor could we otherwise understand why these concepts only, and no others, abide in the pure understanding. It was an enterprise worthy of an acute thinker like Aristotle to try to discover these fundamental concepts; but as he had no guiding principle he merely picked them up as they occurred to him, and at first gathered up ten of them, which he called categories or predicaments. Afterwards he thought he had discovered five more of them, which he added under the name of post-predicaments. But his table remained imperfect for all that, not to mention that we find in it some modes of pure sensibility (quando, ubi, situs, also prius, simul), also an empirical concept (motus), none of which can belong to this genealogical register of the understanding. Besides, there are some derivative concepts, counted among the fundamental concepts (actio, passio), while some of the latter are entirely wanting.

With regard to these, it should be remarked that the categories, as the true fundamental concepts of the pure understanding, have also their pure derivative concepts. These could not be passed over in a complete system of transcendental philosophy, but in a merely critical [p. 82] essay the mention of the fact may suffice.

I should like to be allowed to call these pure but derivative concepts of the understanding the predicabilia, in opposition to the predicamenta of the pure understanding. If we are once in possession of the fundamental and primitive concepts, it is easy to add the derivative and secondary, and thus to give a complete image of the genealogical tree of the pure understanding. As at present I am concerned not with the completeness, but only with the principles of a system, I leave this supplementary work for a future occasion. In order to carry it out, one need only consult any of the ontological manuals, and place, for instance, under the category of causality the predicabilia of force, of action, and of passion; under the category of community the predicabilia of presence and resistance; under the predicaments of modality the predicabilia of origin, extinction, change, etc. If we associate the categories among themselves or with the modes of pure sensibility, they yield us a large number of derivative concepts a priori, which it would be useful and interesting to mark and, if possible, to bring to a certain completeness, though this is not essential for our present purpose.

I intentionally omit here the definitions of these categories, though I may be in possession of them.1 In the sequel I shall dissect these concepts so far as is [p. 83] sufficient for the purpose of the method which I am preparing. In a complete system of pure reason they might be justly demanded, but at present they would only make us lose sight of the principal object of our investigation, by rousing doubts and objections which, without injury to our essential object, may well be relegated to another time. The little I have said ought to be sufficient to show clearly that a complete dictionary of these concepts with all requisite explanations is not only possible, but easy. The compartments exist; they have only to be filled, and with a systematic topic like the present the proper place to which each concept belongs cannot easily be missed, nor compartments be passed over which are still empty.1

CHAPTER II

OF THE DEDUCTION OF THE PURE CONCEPTS OF THE UNDERSTANDING [p. 84]

Section I

Of the Principles of a Transcendental Deduction in General

Jurists, when speaking of rights and claims, distinguish in every lawsuit the question of right (quid juris) from the question of fact (quid facti), and in demanding proof of both they call the former, which is to show the right or, it may be, the claim, the deduction. We, not being jurists, make use of a number of empirical concepts, without opposition from anybody, and consider ourselves justified, without any deduction, in attaching to them a sense or imaginary meaning, because we can always appeal to experience to prove their objective reality. There exist however illegitimate concepts also, such as, for instance, chance, or fate, which through an almost general indulgence are allowed to be current, but are yet from time to time challenged by the question quid juris. In that case we are greatly embarrassed in looking for their deduction, there being no clear legal title, whether from experience or from reason, on which their [p. 85] claim to employment could be clearly established.

Among the many concepts, however, which enter into the complicated code of human knowledge, there are some which are destined for pure use a priori, independent of all experience, and such a claim requires at all times a deduction,1 because proofs from experience would not be sufficient to establish the legitimacy of such a use, though it is necessary to know how much concepts can refer to objects which they do not find in experience. I call the explanation of the manner how such concepts can a priori refer to objects their transcendental deduction, and distinguish it from the empirical deduction which shows the manner how a concept may be gained by experience and by reflection on experience; this does not touch the legitimacy, but only the fact whence the possession of the concept arose.

We have already become acquainted with two totally distinct classes of concepts, which nevertheless agree in this, that they both refer a priori to objects, namely, the concepts of space and time as forms of sensibility, and the categories as concepts of the understanding. It would be labour lost to attempt an empirical deduction of them, because their distinguishing characteristic is that they refer to objects without having borrowed anything from experience for their representation. [p. 86] If therefore a deduction of them is necessary, it can only be transcendental.

It is possible, however, with regard to these concepts, as with regard to all knowledge, to try to discover in experience, if not the principle of their possibility, yet the contingent causes of their production. And here we see that the impressions of the senses give the first impulse to the whole faculty of knowledge with respect to them, and thus produce experience which consists of two very heterogeneous elements, namely, matter for knowledge, derived from the senses, and a certain form according to which it is arranged, derived from the internal source of pure intuition and pure thought, first brought into action by the former, and then producing concepts. Such an investigation of the first efforts of our faculty of knowledge, beginning with single perceptions and rising to general concepts, is no doubt very useful, and we have to thank the famous Locke for having been the first to open the way to it. A deduction of the pure concepts a priori, however, is quite impossible in that way. It lies in a different direction, because, with reference to their future use, which is to be entirely independent of experience, a very different certificate of birth will be required from that of mere descent from experience. We may call this attempted physiological derivation (which cannot properly be called deduction, [p. 87] because it refers to a quaestio facti), the explanation of the possession of pure knowledge. It is clear therefore that of these pure concepts a priori a transcendental deduction only is possible, and that to attempt an empirical deduction of them is mere waste of time, which no one would think of except those who have never understood the very peculiar nature of that kind of knowledge.

But though it may be admitted that the only possible deduction of pure knowledge a priori must be transcendental, it has not yet been proved that such a deduction is absolutely necessary. We have before, by means of a transcendental deduction, followed up the concepts of space and time to their very sources, and explained and defined their objective validity a priori. Geometry, however, moves along with a steady step, through every kind of knowledge a priori, without having to ask for a certificate from philosophy as to the pure legitimate descent of its fundamental concept of space. But it should be remarked that in geometry this concept is used with reference to the outer world of sense only, of which space is the pure form of intuition, and where geometrical knowledge, being based on a priori intuition, possesses immediate evidence, the objects being given, so far as their form is concerned, through their very knowledge a priori in intuition. When we come, however, [p. 88] to the pure concepts of the understanding, it becomes absolutely necessary to look for a transcendental deduction, not only for them, but for space also, because they, not being founded on experience, apply to objects generally, without any of the conditions of sensibility; and, speaking of objects, not through predicates of intuition and sensibility, but of pure thought a priori, are not able to produce in intuition a priori any object on which, previous to all experience, their synthesis was founded. These concepts of pure understanding, therefore, not only excite suspicion with regard to the objective validity and the limits of their own application, but render even the concept of space equivocal, because of an inclination to apply it beyond the conditions of sensuous intuition, which was the very reason that made a transcendental deduction of it, such as we gave before, necessary. Before the reader has made a single step in the field of pure reason, he must be convinced of the inevitable necessity of such a transcendental deduction, otherwise he would walk on blindly and, after having strayed in every direction, he would only return to the same ignorance from which he started. He must at the same time perceive the inevitable difficulty of such a deduction, so that he may not complain about obscurity where the object itself is obscure, or weary too soon with our removal of obstacles, the fact being that we have [p. 89] either to surrender altogether all claims to the knowledge of pure reason — the most favourite field of all philosophers, because extending beyond the limits of all possible experience — or to bring this critical investigation to perfection.

It was easy to show before, when treating of the concepts of space and time, how these, though being knowledge a priori, refer necessarily to objects, and how they make a synthetical knowledge of them possible, which is independent of all experience. For, as no object can appear to us, that is, become an object of empirical intuition, except through such pure forms of sensibility, space and time are pure intuitions which contain a priori the conditions of the possibility of objects as phenomena, and the synthesis in these intuitions possesses objective validity.

The categories of the understanding, on the contrary, are not conditions under which objects can be given in intuition, and it is quite possible therefore that objects should appear to us without any necessary reference to the functions of the understanding, thus showing that the understanding contains by no means any of their conditions a priori. There arises therefore here a difficulty, which we did not meet with in the field of sensibility, namely, how subjective conditions of thought can have objective validity, that is, become conditions of the possibility of the knowledge of objects. It cannot be [p. 90] denied that phenomena may be given in intuition without the functions of the understanding. For if we take, for instance, the concept of cause, which implies a peculiar kind of synthesis, consisting in placing according to a rule after something called A something totally different from it, B, we cannot say that it is a priori clear why phenomena should contain something of this kind. We cannot appeal for it to experience, because what has to be proved is the objective validity of this concept a priori. It would remain therefore a priori doubtful whether such a concept be not altogether empty, and without any corresponding object among phenomena. It is different with objects of sensuous intuition. They must conform to the formal conditions of sensibility existing a priori in the mind, because otherwise they could in no way be objects to us. But why besides this they should conform to the conditions which the understanding requires for the synthetical unity of thought, does not seem to follow quite so easily. For we could quite well imagine that phenomena might possibly be such that the understanding should not find them conforming to the conditions of its synthetical unity, and all might be in such confusion that nothing should appear in the succession of phenomena which could supply a rule of synthesis, and correspond, for instance, to the concept of cause and effect, so that this concept would thus be quite empty, null, and meaningless. With all this phenomena would offer objects to our intuition, because intuition by itself does not require the functions [p. 91] of thought.

It might be imagined that we could escape from the trouble of these investigations by saying that experience offers continually examples of such regularity of phenomena as to induce us to abstract from it the concept of cause, and it might be attempted to prove thereby the objective validity of such a concept. But it ought to be seen that in this way the concept of cause cannot possibly arise, and that such a concept ought either to be founded a priori in the understanding or be surrendered altogether as a mere hallucination. For this concept requires strictly that something, A, should be of such a nature that something else, B, follows from it necessarily and according to an absolutely universal rule. Phenomena no doubt supply us with cases from which a rule becomes possible according to which something happens usually, but never so that the result should be necessary. There is a dignity in the synthesis of cause and effect which cannot be expressed empirically, for it implies that the effect is not only an accessory to the cause, but given by it and springing from it. Nor is the absolute universality of the rule a quality inherent in empirical rules, which by means of induction cannot receive any but a relative universality, that [p. 92] is, a more or less extended applicability. If we were to treat the pure concepts of the understanding as merely empirical products, we should completely change their character and their use.

Transition to a Transcendental Deduction of the Categories

Two ways only are possible in which synthetical representations and their objects can agree, can refer to each other with necessity, and so to say meet each other. Either it is the object alone that makes the representation possible, or it is the representation alone that makes the object possible. In the former case their relation is empirical only, and the representation therefore never possible a priori. This applies to phenomena with reference to whatever in them belongs to sensation. In the latter case, though representation by itself (for we do not speak here of its1 causality by means of the will) cannot produce its object so far as its existence is concerned, nevertheless the representation determines the object a priori, if through it alone it is possible to know anything as an object. To know a thing as an object is possible only under two conditions. First, there must be intuition by which the object is given us, though as a phenomenon only, secondly, there must be a concept by which [p. 93] an object is thought as corresponding to that intution. From what we have said before it is clear that the first condition, namely, that under which alone objects can be seen, exists, so far as the form of intuition is concerned, in the soul a priori. All phenomena therefore must conform to that formal condition of sensibility, because it is through it alone that they appear, that is, that they are given and empirically seen.

Now the question arises whether there are not also antecedent concepts. a priori, forming conditions under which alone something can be, if not seen, yet thought as an object in general; for in that case all empirical knowledge of objects would necessarily conform to such concepts, it being impossible that anything should become an object of experience without them. All experience contains, besides the intuition of the senses by which something is given, a concept also of the object, which is given in intuition as a phenomenon. Such concepts of objects in general therefore must form conditions a priori of all knowledge produced by experience, and the objective validity of the categories, as being such concepts a priori, rests on this very fact that by them alone, so far as the form of thought is concerned, experience becomes possible. If by them only it is possible to think any object of experience, it follows that they refer by necessity and a priori to all objects of experience.

There is therefore a principle for the transcendental [p. 94] deduction of all concepts a priori which must guide the whole of our investigation, namely, that all must be recognized as conditions a priori of the possibility of experience, whether of intuition, which is found in it, or of thought. Concepts which supply the objective ground of the possibility of experience are for that very reason necessary. An analysis of the experience in which they are found would not be a deduction, but a mere illustration, because they would there have an accidental character only. Nay, without their original relation to all possible experience in which objects of knowledge occur, their relation to any single object would be quite incomprehensible.

[There are three original sources, or call them faculties or powers of the soul, which contain the conditions of the possibility of all experience, and which themselves cannot be derived from any other faculty, namely, sense, imagination, and apperception. On them is founded —

1. The synopsis of the manifold a priori through the senses.

2. The synthesis of this manifold through the imagination.

3. The unity of that synthesis by means of original apperception.

Besides their empirical use all these faculties have a transcendental use also, referring to the form only and possible a priori. With regard to the senses we have discussed that transcendental use in the first part, [p. 95] and we shall now proceed to an investigation of the remaining two, according to their true nature.1 ]

Section II

Of the a priori Grounds for the Possibility of Experience

[That a concept should be produced entirely a priori and yet refer to an object, though itself neither belonging to the sphere of possible experience, nor consisting of the elements of such an experience, is self-contradictory and impossible. It would have no contents, because no intuition corresponds to it, and intuitions by which objects are given to us constitute the whole field or the complete object of possible experience. An a priori concept therefore not referring to experience would be the logical form only of a concept, but not the concept itself by which something is thought.

If therefore there exist any pure concepts a priori, though they cannot contain anything empirical, they must nevertheless all be conditions a priori of a possible experience, on which alone their objective reality depends.

If therefore we wish to know how pure concepts of the understanding are possible, we must try to find out what are the conditions a priori on which the possibility [p. 96] of experience depends, nay, on which it is founded, apart from all that is empirical in phenomena. A concept expressing this formal and objective condition of experience with sufficient generality might properly be called a pure concept of the understanding. If we once have these pure concepts of the understanding, we may also imagine objects which are either impossible, or, if not impossible in themselves, yet can never be given in any experience. We have only in the connection of those concepts to leave out something which necessarily belongs to the conditions of a possible experience (concept of a spirit), or to extend pure concepts of the understanding beyond what can be reached by experience (concept of God). But the elements of all knowledge a priori, even of gratuitous and preposterous fancies, though not borrowed from experience (for in that case they would not be knowledge a priori) must nevertheless contain the pure conditions a priori of a possible experience and its object, otherwise not only would nothing be thought by them, but they themselves, being without data, could never arise in our mind.

Such concepts, then, which comprehend the pure thinking a priori involved in every experience, are discovered in the categories, and it is really a sufficient deduction of them and a justification of their objective validity, if we succeed in proving that by them alone an object [p. 97] can be thought. But as in such a process of thinking more is at work than the faculty of thinking only, namely, the understanding, and as the understanding, as a faculty of knowledge which is meant to refer to objects, requires quite as much an explanation as to the possibility of such a reference, it is necessary for us to consider the subjective sources which form the foundation a priori for the possibility of experience, not according to their empirical, but according to their transcendental character.

If every single representation stood by itself, as if isolated and separated from the others, nothing like what we call knowledge could ever arise, because knowledge forms a whole of representations connected and compared with each other. If therefore I ascribe to the senses a synopsis, because in their intuition they contain something manifold, there corresponds to it always a synthesis, and receptivity can make knowledge possible only when joined with spontaneity. This spontaneity, now, appears as a threefold synthesis which must necessarily take place in every kind of knowledge, namely, first, that of the apprehension of representations as modifications of the soul in intuition, secondly, of the reproduction of them in the imagination, and, thirdly, that of their recognition in concepts. This leads us to three subjective sources of knowledge which render possible the understanding, and through it all experience as an empirical product of the understanding. [p. 98]

Preliminary Remark

The deduction of the categories is beset with so many difficulties and obliges us to enter so deeply into the first grounds of the possibility of our knowledge in general, that I thought it more expedient, in order to avoid the lengthiness of a complete theory, and yet to omit nothing in so essential an investigation, to add the following four paragraphs with a view of preparing rather than instructing the reader. After that only I shall in the third section proceed to a systematical discussion of these elements of the understanding. Till then the reader must not allow himself to be frightened by a certain amount of obscurity which at first is inevitable on a road never trodden before, but which, when we come to that section, will give way, I hope, to a complete comprehension.

I

Of the Synthesis of Apprehension in Intuition

Whatever the origin of our representations may be, whether they be due to the influence of external things or to internal causes, whether they have arisen a priori or empirically as phenomena, as modifications of the mind they must always belong to the internal [p. 99] sense, and all our knowledge must therefore finally be subject to the formal condition of that internal sense, namely, time, in which they are all arranged, joined, and brought into certain relations to each other. This is a general remark which must never be forgotten in all that follows.

Every representation contains something manifold, which could not be represented as such, unless the mind distinguished the time in the succession of one impression after another; for as contained in one moment, each representation can never be anything but absolute unity. In order to change this manifold into a unity of intuition (as, for instance, in the representation of space), it is necessary first to run through the manifold and then to hold it together. It is this act which I call the synthesis of apprehension, because it refers directly to intuition which no doubt offers something manifold, but which, without a synthesis, can never make it such, as it is contained in one representation.

This synthesis of apprehension must itself be carried out a priori also, that is, with reference to representations which are not empirical. For without it we should never be able to have the representations either of space or time a priori, because these cannot be produced except [p. 100] by a synthesis of the manifold which the senses offer in their original receptivity. It follows therefore that we have a pure synthesis of apprehension.

II

Of the Synthesis of Reproduction in Imagination

It is no doubt nothing but an empirical law according to which representations which have often followed or accompanied one another, become associated in the end and so closely united that, even without the presence of the object, one of these representations will, according to an invariable law, produce a transition of the mind to the other. This law of reproduction, however, presupposes that the phenomena themselves are really subject to such a rule, and that there is in the variety of these representations a sequence and concomitancy subject to certain rules; for without this the faculty of empirical imagination would never find anything to do that it is able to do, and remain therefore buried within our mind as a dead faculty, unknown to ourselves. If cinnabar were sometimes red and sometimes black, sometimes light and sometimes heavy, if a man could be changed now into this, now into another animal shape, if on the longest day the fields were sometimes covered with fruit, [p. 101] sometimes with ice and snow, the faculty of my empirical imagination would never be in a position, when representing red colour, to think of heavy cinnabar. Nor, if a certain name could be given sometimes to this, sometimes to that object, or if that the same object could sometimes be called by one, and sometimes by another name, without any rule to which representations are subject by themselves, would it be possible that any empirical synthesis of reproduction should ever take place.

There must therefore be something to make this reproduction of phenomena possible by being itself the foundation a priori of a necessary synthetical unity of them. This becomes clear if we only remember that all phenomena are not things by themselves, but only the play of our representations, all of which are in the end determinations only of the internal sense. If therefore we could prove that even our purest intuitions a priori give us no knowledge, unless they contain such a combination of the manifold as to render a constant synthesis of reproduction possible, it would follow that this synthesis of the imagination is, before all experience, founded on principles a priori, and that we must admit a pure transcendental synthesis of imagination which forms even the foundation of the possibility of all experience, such experience being impossible without the reproductibility of phenomena. [p. 102] Now, when I draw a line in thought, or if I think the time from one noon to another, or if I only represent to myself a certain number, it is clear that I must first necessarily apprehend one of these manifold representations after another. If I were to lose from my thoughts what precedes, whether the first parts of a line or the antecedent portions of time, or the numerical unities representing one after the other, and if, while I proceed to what follows, I were unable to reproduce what came before, there would never be a complete representation, and none of the before-mentioned thoughts, not even the first and purest representations of space and time, could ever arise within us.

The synthesis of apprehension is therefore inseparably connected with the synthesis of reproduction, and as the former constitutes the transcendental ground of the possibility of all knowledge in general (not only of empirical, but also of pure a priori knowledge), it follows that a reproductive synthesis of imagination belongs to the transcendental acts of the soul. We may therefore call this faculty the transcendental faculty of imagination.

III

Of the Synthesis of Recognition in Concepts [p. 103]

Without our being conscious that what we are thinking now is the same as what we thought a moment before, all reproduction in the series of representations would be vain. Each representation would, in its present state, be a new one, and in no wise belonging to the act by which it was to be produced by degrees, and the manifold in it would never form a whole, because deprived of that unity which consciousness alone can impart to it. If in counting I forget that the unities which now present themselves to my mind have been added gradually one to the other, I should not know the production of the quantity by the successive addition of one to one, nor should I know consequently the number, produced by the counting, this number being a concept consisting entirely in the consciousness of that unity of synthesis.

The very word of concept (Begriff) could have suggested this remark, for it is the one consciousness which unites the manifold that has been perceived successively, and afterwards reproduced into one representation. This consciousness may often be very faint, and we may connect it with the effect only, and not with the act itself, i.e. with the production of a representation. But in [p. 104] spite of this, that consciousness, though deficient in pointed clearness, must always be there, and without it, concepts, and with them, knowledge of objects are perfectly impossible.

And here we must needs arrive at a clear understanding of what we mean by an object of representations. We said before that phenomena are nothing but sensuous representations, which therefore by themselves must not be taken for objects outside our faculty of representation. What then do we mean if we speak of an object corresponding to, and therefore also different from our knowledge? It is easy to see that such an object can only be conceived as something in general = x: because, beside our knowledge, we have absolutely nothing which we could put down as corresponding to that knowledge.

Now we find that our conception of the relation of all knowledge to its object contains something of necessity, the object being looked upon as that which prevents our knowledge from being determined at haphazard, and causes it to be determined a priori in a certain way, because, as they are all to refer to an object, they must necessarily, with regard to that object, agree with each other, that is to say, possess that unity which [p. 105] constitutes the concept of an object.

It is clear also that, as we can only deal with the manifold in our representations, and as the x corresponding to them (the object), since it is to be something different from all our representations, is really nothing to us, it is clear, I say, that the unity, necessitated by the object, cannot be anything but the formal unity of our consciousness in the synthesis of the manifold in our representations. Then and then only do we say that we know an object, if we have produced synthetical unity in the manifold of intuition. Such unity is impossible, if the intuition could not be produced, according to a rule, by such a function of synthesis as makes the reproduction of the manifold a priori necessary, and a concept in which that manifold is united, possible. Thus we conceive a triangle as an object, if we are conscious of the combination of three straight lines, according to a rule, which renders such an intuition possible at all times. This unity of rule determines the manifold and limits it to conditions which render the unity of apperception possible, and the concept of that unity is really the representation of the object = x, which I think, by means of the predicates of a triangle.

No knowledge is possible without a concept, [p. 106] however obscure or imperfect it may be, and a concept is always, with regard to its form, something general, something that can serve as a rule. Thus the concept of body serves as a rule to our knowledge of external phenomena, according to the unity of the manifold which is thought by it. It can only be such a rule of intuitions because representing, in any given phenomena, the necessary reproduction of their manifold elements, or the synthetical unity in our consciousness of them. Thus the concept of body, whenever we perceive something outside us, necessitates the representation of extension, and, with it, those of impermeability, shape, etc.

Necessity is always founded on transcendental conditions. There must be therefore a transcendental ground of the unity of our consciousness in the synthesis of the manifold of all our intuitions, and therefore also a transcendental ground of all concepts of objects in general, and therefore again of all objects of experience, without which it would be impossible to add to our intuitions the thought of an object, for the object is no more than that something of which the concept predicates such a necessity of synthesis.

That original and transcendental condition is nothing else but what I call transcendental apperception. [p. 107] The consciousness of oneself, according to the determinations of our state, is, with all our internal perceptions, empirical only, and always transient. There can be no fixed or permanent self in that stream of internal phenomena. It is generally called the internal sense, or the empirical apperception. What is necessarily to be represented as numerically identical with itself, cannot be thought as such by means of empirical data only. It must be a condition which precedes all experience, and in fact renders it possible, for thus only could such a transcendental supposition acquire validity.

No knowledge can take place in us, no conjunction or unity of one kind of knowledge with another, without that unity of consciousness which precedes all data of intuition, and without reference to which no representation of objects is possible. This pure, original, and unchangeable consciousness I shall call transcendental apperception. That it deserves such a name may be seen from the fact that even the purest objective unity, namely, that of the concepts a priori (space and time), is possible only by a reference of all intuitions to it. The numerical unity of that apperception therefore forms the a priori condition of all concepts, as does the manifoldness of space and time of the intuitions of the senses.

The same transcendental unity of apperception [p. 108] constitutes, in all possible phenomena which may come together in our experience, a connection of all these representations according to laws. For that unity of consciousness would be impossible, if the mind, in the knowledge of the manifold, could not become conscious of the identity of function, by which it unites the manifold synthetically in one knowledge. Therefore the original and necessary consciousness of the identity of oneself is at the same time a consciousness of an equally necessary unity of the synthesis of all phenomena according to concepts, that is, according to rules, which render them not only necessarily reproducible, but assign also to their intuition an object, that is, a concept of something in which they are necessarily united. The mind could never conceive the identity of itself in the manifoldness of its representations (and this a priori) if it did not clearly perceive the identity of its action, by which it subjects all synthesis of apprehension (which is empirical) to a transcendental unity, and thus renders its regular coherence a priori possible. When we have clearly perceived this, we shall be able to determine more accurately our concept of an object in general. All representations have, as representations, their object, and can themselves in turn become objects of other representations. The only objects which can be given to us immediately are phenomena, and whatever in them refers immediately to the object is [p. 109] called intuition. These phenomena, however, are not things in themselves, but representations only which have their object, but an object that can no longer be seen by us, and may therefore be called the not-empirical, that is, the transcendental object, = x.

The pure concept of such a transcendental object (which in reality in all our knowledge is always the same = x) is that which alone can give to all our empirical concepts a relation to an object or objective reality. That concept cannot contain any definite intuition, and can therefore refer to that unity only, which must be found in the manifold of our knowledge, so far as it stands in relation to an object. That relation is nothing else but a necessary unity of consciousness, and therefore also of the synthesis of the manifold, by a common function of the mind, which unites it in one representation. As that unity must be considered as a priori necessary (because, without it, our knowledge would be without an object), we may conclude that the relation to a transcendental object, that is, the objective reality of our empirical knowledge, rests on a transcendental law, that all phenomena, if they are to give us objects, must be subject to rules [p. 110] a priori of a synthetical unity of these objects, by which rules alone their mutual relation in an empirical intuition becomes possible: that is, they must be subject, in experience, to the conditions of the necessary unity of apperception quite as much as, in mere intuition, to the formal conditions of space and time. Without this no knowledge is possible.

IV

Preliminary Explanation of the Possibility of the Categories as Knowledge a priori

There is but one experience in which all perceptions are represented as in permanent and regular connection, as there is but one space and one time in which all forms of phenomena and all relations of being or not being take place. If we speak of different experiences, we only mean different perceptions so far as they belong to one and the same general experience. It is the permanent and synthetical unity of perceptions that constitutes the form of experience, and experience is nothing but the synthetical unity of phenomena according to concepts.

Unity of synthesis, according to empirical concepts, would be purely accidental, nay, unless these [p. 111] were founded on a transcendental ground of unity, a whole crowd of phenomena might rush into our soul, without ever forming real experience. All relation between our knowledge and its objects would be lost at the same time, because that knowledge would no longer be held together by general and necessary laws; it would therefore become thoughtless intuition, never knowledge, and would be to us the same as nothing.

The conditions a priori of any possible experience in general are at the same time conditions of the possibility of any objects of our experience. Now I maintain that the categories of which we are speaking are nothing but the conditions of thought which make experience possible, as much as space and time contain the conditions of that intuition which forms experience. These categories therefore are also fundamental concepts by which we think objects in general for the phenomena, and have therefore a priori objective validity. This is exactly what we wish to prove.

The possibility, nay the necessity of these categories rests on the relation between our whole sensibility, and therefore all possible phenomena, and that original apperception in which everything must be necessarily subject to the conditions of the permanent unity of self-consciousness, that is, must submit to the general functions [p. 112] of that synthesis which we call synthesis according to concepts, by which alone our apperception can prove its permanent and necessary identity a priori. Thus the concept of cause is nothing but a synthesis of that which follows in temporal succession, with other phenomena, but a synthesis according to concepts: and without such a unity which rests on a rule a priori, and subjects all phenomena to itself, no permanent and general, and therefore necessary unity of consciousness would be formed in the manifold of our perceptions. Such perceptions would then belong to no experience at all, they would be without an object, a blind play of representations, — less even than a dream.

All attempts therefore at deriving those pure concepts of the understanding from experience, and ascribing to them a purely empirical origin, are perfectly vain and useless. I shall not dwell here on the fact that a concept of cause, for instance, contains an element of necessity, which no experience can ever supply, because experience, though it teaches us that after one phenomenon something else follows habitually, can never teach us that it follows necessarily, nor that we could a priori, and without any limitation, derive from it, as a condition, any conclusion as to what must follow. And thus I ask with reference to that empirical rule of association, which must always be admitted if we say that everything in the succession of events is so entirely subject to rules that nothing [p. 113] ever happens without something preceding it on which it always follows, — What does it rest on, if it is a law of nature, nay, how is that very association possible? You call the ground for the possibility of the association of the manifold, so far as it is contained in the objects themselves, the affinity of the manifold. I ask, therefore, how do you make that permanent affinity by which phenomena stand, nay, must stand, under permanent laws, conceivable to yourselves?

According to my principles it is easily conceivable. All possible phenomena belong, as representations, to the whole of our possible self-consciousness. From this, as a transcendental representation, numerical identity is inseparable and a priori certain, because nothing can become knowledge except by means of that original apperception. As this identity must necessarily enter into the synthesis of the whole of the manifold of phenomena, if that synthesis is to become empirical knowledge, it follows that the phenomena are subject to conditions a priori to which their synthesis (in apprehension) must always conform. The representation of a general condition according to which something manifold can be arranged (with uniformity) is called a rule, if it must be so arranged, a law. All phenomena therefore stand in a permanent connection according to necessary laws, and thus possess [p. 114] that transcendental affinity of which the empirical is a mere consequence.

It sounds no doubt very strange and absurd that nature should have to conform to our subjective ground of apperception, nay, be dependent on it, with respect to her laws. But if we consider that what we call nature is nothing but a whole of phenomena, not a thing by itself, but a number of representations in our soul, we shall no longer be surprised that we only see her through the fundamental faculty of all our knowledge, namely, the transcendental apperception, and in that unity without which it could not be called the object (or the whole) of all possible experience, that is, nature. We shall thus also understand why we can recognise this unity a priori, and therefore as necessary, which would be perfectly impossible if it were given by itself and independent of the first sources of our own thinking. In that case I could not tell whence we should take the synthetical propositions of such general unity of nature. They would have to be taken from the objects of nature themselves, and as this could be done empirically only, we could derive from it none but an accidental unity, which is very different from that necessary connection which we mean when speaking of nature.

Section III

Of the Relation of the Understanding to Objects in General, and the Possibility of Knowing Them a priori [p. 115]

What in the preceding section we have discussed singly and separately we shall now try to treat in connection with each other and as a whole. We saw that there are three subjective sources of knowledge on which the possibility of all experience and of the knowledge of its objects depends, namely, sense, imagination, and apperception. Each of them may be considered as empirical in its application to given phenomena; all, however, are also elements or grounds a priori which render their empirical application possible. Sense represents phenomena empirically in perception, imagination in association (and reproduction), apperception in the empirical consciousness of the identity of these reproductive representations with the phenomena by which they were given; therefore in recognition.

The whole of our perception rests a priori on pure intuition (if the perception is regarded as representation, then on time, as the form of our internal intuition), the association of it (the whole) on the pure syn- [p. 116] thesis of imagination, and our empirical consciousness of it on pure apperception, that is, on the permanent identity of oneself in the midst of all possible representations.

If we wish to follow up the internal ground of this connection of representations to that point towards which they must all converge, and where they receive for the first time that unity of knowledge which is requisite for every possible experience, we must begin with pure apperception. Intuitions are nothing to us, and do not concern us in the least, if they cannot be received into our consciousness, into which they may enter either directly or indirectly. Knowledge is impossible in any other way. We are conscious a priori of our own permanent identity with regard to all representations that can ever belong to our knowledge, as forming a necessary condition of the possibility of all representations (because these could not represent anything in me, unless they belonged with everything else to one consciousness and could at least be connected within it). This principle stands firm a priori, and may be called the transcendental principle of the unity of all the manifold of our representations (therefore also of intuition). This unity of the manifold in one subject is synthetical; the pure apperception therefore supplies us with a principle of the synthetical unity of [p. 117] the manifold in all possible intuitions.1

This synthetical unity, however, presupposes [p. 118] or involves a synthesis, and if that unity is necessary a priori, the synthesis also must be a priori. The transcendental unity of apperception therefore refers to the pure synthesis of imagination as a condition a priori of the possibility of the manifold being united in one knowledge. Now there can take place a priori the productive synthesis of imagination only, because the reproductive rests on conditions of experience. The principle therefore of the necessary unity of the pure (productive) synthesis of imagination, before all apperception, constitutes the ground of the possibility of all knowledge, nay, of all experience.

The synthesis of the manifold in imagination is called transcendental, if, without reference to the difference of intuitions, it affects only the a priori conjunction of the manifold; and the unity of that synthesis is called transcendental if, with reference to the original unity of apperception, it is represented as a priori necessary. As the possibility of all knowledge depends on the unity of that apperception, it follows that the transcendental unity of the synthesis of imagination is the pure form of all possible knowledge through which therefore all objects of possible experience must be represented a priori.

This unity of apperception with reference to [p. 119] the synthesis of imagination is the understanding, and the same unity with reference to the transcendental synthesis of the imagination, the pure understanding. It must be admitted therefore that there exist in the understanding pure forms of knowledge a priori, which contain the necessary unity of the pure synthesis of the imagination in reference to all possible phenomena. These are the categories, that is, the pure concepts of the understanding. The empirical faculty of knowledge of man contains therefore by necessity an understanding which refers to all objects of the senses, though by intuition only, and by its synthesis through imagination, and all phenomena, as data of a possible experience, must conform to that understanding. As this relation of phenomena to a possible experience is likewise necessary, (because, without it, we should receive no knowledge through them, and they would not in the least concern us), it follows that the pure understanding constitutes by the means of the categories a formal and synthetical principle of all experience, and that phenomena have thus a necessary relation to the understanding.

We shall now try to place the necessary connection of the understanding with the phenomena by means of the categories more clearly before the reader, by beginning with the beginning, namely, with the empirical.

The first that is given us is the phenomenon, [p. 120] which, if connected with consciousness, is called perception. (Without its relation to an at least possible consciousness, the phenomenon could never become to us an object of knowledge. It would therefore be nothing to us; and because it has no objective reality in itself, but exists only in being known, it would be nothing altogether.) As every phenomenon contains a manifold, and different perceptions are found in the mind singly and scattered, a connection of them is necessary, such as they cannot have in the senses by themselves. There exists therefore in us an active power for the synthesis of the manifold which we call imagination, and the function of which, as applied to perceptions, I call apprehension.1 This imagination is meant to change the manifold of intuition into an image, it must therefore first receive the impressions into its activity, which I call to apprehend.

It must be clear, however, that even this apprehension [p. 121] of the manifold could not alone produce a coherence of impressions or an image, without some subjective power of calling one perception from which the mind has gone over to another back to that which follows, and thus forming whole series of perceptions. This is the reproductive faculty of imagination which is and can be empirical only.

If representations, as they happen to meet with one another, could reproduce each other at haphazard, they would have no definite coherence, but would form irregular agglomerations only, and never produce knowledge. It is necessary therefore that their reproduction should be subject to a rule by which one representation connects itself in imagination with a second and not with a third. It is this subjective and empirical ground of reproduction according to rules, which is called the association of representations.

If this unity of association did not possess an objective foundation also, which makes it impossible that phenomena should be apprehended by imagination in any other way but under the condition of a possible synthetical unity of that apprehension, it would be a mere accident that phenomena lend themselves to a certain connection in human knowledge. Though we might have the power of associating perceptions, it would still be a matter of [p. 122] uncertainty and chance whether they themselves are associable; and, in case they should not be so, a number of perceptions; nay, the whole of our sensibility, might possibly contain a great deal of empirical consciousness, but in a separate state, nay, without belonging to the one consciousness of myself, which, however, is impossible. Only by ascribing all perceptions to one consciousness (the original apperception) can I say of all of them that I am conscious of them. It must be therefore an objective ground, that is, one that can be understood as existing a priori, and before all empirical laws of imagination, on which alone the possibility, nay, even the necessity of a law can rest, which pervades all phenomena, and which makes us look upon them all, without exception, as data of the senses, associable by themselves, and subject to general rules of a permanent connection in their reproduction. This objective ground of all association of phenomena I call their affinity, and this can nowhere be found except in the principle of the unity of apperception applied to all knowledge which is 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. This, without a synthetical unity in their connection, which is therefore necessary objectively also, would be impossible.

We have thus seen that the objective unity [p. 123] of all (empirical) consciousness in one consciousness (that of the original apperception) is the necessary condition even of all possible perception, while the affinity of all phenomena (near or remote) is a necessary consequence of a synthesis in imagination which is a priori founded on rules.

Imagination is therefore likewise the power of a synthesis a priori which is the reason why we called it productive imagination, and so far as this aims at nothing but the necessary unity in the synthesis of all the manifold in phenomena, it may be called the transcendental function of imagination. However strange therefore it may appear at first, it must nevertheless have become clear by this time that the affinity of phenomena and with it their association, and through that, lastly, their reproduction also according to laws, that is, the whole of our experience, becomes possible only by means of that transcendental function of imagination, without which no concepts of objects could ever come together in one experience.

It is the permanent and unchanging Ego (or pure apperception) which forms the correlative of all our representations, if we are to become conscious of them, and all consciousness belongs quite as much to such an all-embracing pure apperception as all sensuous intuitions belongs, as a representation, to a pure internal [p. 124] intuition, namely, time. This apperception it is which must be added to pure imagination, in order to render its function intellectual. For by itself, the synthesis of imagination, though carried out a priori, is always sensuous, and only connects the manifold as it appears in intuition, for instance, the shape of a triangle. But when the manifold is brought into relation with the unity of apperception, concepts which belong to the understanding become possible, but only as related to sensuous intuition through imagination.

We have therefore a pure imagination as one of the fundamental faculties of the human soul, on which all knowledge a priori depends. Through it we bring the manifold of intuition on one side in connection with the condition of the necessary unity of pure apperception on the other. These two extreme ends, sense and understanding, must be brought into contact with each other by means of the transcendental function of imagination, because, without it, the senses might give us phenomena, but no objects of empirical knowledge, therefore no experience. Real experience, which is made up of apprehension, association (reproduction), and lastly recognition of phenomena, contains in this last and highest [p. 125] (among the purely empirical elements of experience) concepts, which render possible the formal unity of experience, and with it, all objective validity (truth) of empirical knowledge. These grounds for the recognition of the manifold, so far as they concern the form only of experience in general, are our categories. On them is founded the whole formal unity in the synthesis of imagination and, through it, of1 the whole empirical use of them (in recognition, reproduction, association, and apprehension) down to the very phenomena, because it is only by means of those elements of knowledge that the phenomena can belong to our consciousness and therefore to ourselves.

It is we therefore who carry into the phenomena which we call nature, order and regularity, nay, we should never find them in nature, if we ourselves, or the nature of our mind, had not originally placed them there. For the unity of nature is meant to be a necessary and a priori certain unity in the connection of all phenomena. And how should we a priori have arrived at such a synthetical unity, if the subjective grounds of such unity were not contained a priori in the original sources of our knowledge, and if those subjective conditions did not at the same time possess objective validity, as being the grounds on which alone an object becomes possible in [p. 126] our experience?

We have before given various definitions of the understanding, by calling it the spontaneity of knowledge (as opposed to the receptivity of the senses), or the faculty of thinking, or the faculty of concepts or of judgments; all of these explanations, if more closely examined, coming to the same. We may now characterise it as the faculty of rules. This characteristic is more significant, and approaches nearer to the essence of the understanding. The senses give us forms (of intuition), the understanding rules, being always busy to examine phenomena, in order to discover in them some kind of rule. Rules, so far as they are objective (therefore necessarily inherent in our knowledge of an object), are called laws. Although experience teaches us many laws, yet these are only particular determinations of higher laws, the highest of them, to which all others are subject, springing a priori from the understanding; not being derived from experience, but, on the contrary, imparting to the phenomena their regularity, and thus making experience possible. The understanding therefore is not only a power of making rules by a comparison of phenomena, it is itself the lawgiver of nature, and without the understanding nature, that is, a synthetical unity of the manifold of phenomena, [p. 127] according to rules, would be nowhere to be found, because phenomena, as such, cannot exist without us, but exist in our sensibility only. This sensibility, as an object of our knowledge in any experience, with everything it may contain, is possible only in the unity of apperception, which unity of apperception is transcendental ground of the necessary order of all phenomena in an experience. The same unity of apperception with reference to the manifold of representations (so as to determine it out of one)1 forms what we call the rule, and the faculty of these rules I call the understanding. As possible experience therefore, all phenomena depend in the same way a priori on the understanding, and receive their formal possibility from it as, when looked upon as mere intuitions, they depend on sensibility, and become possible through it, so far as their form is concerned.

However exaggerated therefore and absurd it may sound, that the understanding is itself the source of the laws of nature, and of its formal unity, such a statement is nevertheless correct and in accordance with experience. It is quite true, no doubt, that empirical laws, as such, cannot derive their origin from the pure understanding, as little as the infinite manifoldness of phenomena could be sufficiently comprehended through the pure form of sensuous intuition. But all empirical laws are only particular determinations of the pure laws of the [p. 128] understanding, under which and according to which the former become possible, and phenomena assume a regular form, quite as much as all phenomena, in spite of the variety of their empirical form, must always submit to the conditions of the pure form of sensibility.

The pure understanding is therefore in the categories the law of the synthetical unity of all phenomena, and thus makes experience, so far as its form is concerned, for the first time possible. This, and no more than this, we were called upon to prove in the transcendental deduction of the categories, namely, to make the relation of the understanding to our sensibility, and through it to all objects of experience, that is the objective validity of the pure concepts a priori of the understanding, conceivable, and thus to establish their origin and their truth.

SUMMARY REPRESENTATION

OF THE CORRECTNESS AND OF THE ONLY POSSIBILITY OF THIS DEDUCTION OF THE PURE CONCEPTS OF THE UNDERSTANDING

If the objects with which our knowledge has to deal were things by themselves, we could have no concepts a priori of them. For where should we take them? If we took them from the object (without asking even the question, how that object could be known to us) our [p. 129] concepts would be empirical only, not concepts a priori. If we took them from within ourselves, then that which is within us only, could not determine the nature of an object different from our representations, that is, supply a ground why there should be a thing to which something like what we have in our thoughts really belongs, and why all this representation should not rather be altogether empty. But if, on the contrary, we have to deal with phenomena only, then it becomes not only possible, but necessary, that certain concepts a priori should precede our empirical knowledge of objects. For being phenomena, they form an object that is within us only, because a mere modification of our sensibility can never exist outside us. The very idea that all these phenomena, and therefore all objects with which we have to deal, are altogether within me, or determinations of my own identical self, implies by itself the necessity of a permanent unity of them in one and the same apperception. In that unity of a possible consciousness consists also the form of all knowledge of objects, by which the manifold is thought as belonging to one object. The manner therefore in which the manifold of sensuous representation (intuition) belongs to our consciousness, precedes all knowledge of an object, as its intellectual form, and constitutes a kind of formal a priori knowledge of all objects in general, if they are to be thought (categories). Their synthesis [p. 130] by means of pure imagination, and the unity of all representations with reference to the original apperception, precede all empirical knowledge. Pure concepts of the understanding are therefore a priori possible, nay, with regard to experience, necessary, for this simple reason because our knowledge has to deal with nothing but phenomena, the possibility of which depends on ourselves, and the connection and unity of which (in the representation of an object) can be found in ourselves only, as antecedent to all experience, nay, as first rendering all experience possible, so far as its form is concerned. On this ground, as the only possible one, our deduction of the categories has been carried out.]

[1 ]Veränderlich in the First Edition is rightly corrected into theilbar in later editions, though in the Second it is still veränderlich.

[1 ]As if in the first, thought were a function of the understanding, in the second, of the faculty of judgment, in the third, of reason; a remark which will receive its elucidation in the sequel.

[1 ]See, however, Karl’s remarks on p. 210 (p. 241 of First Edition).

[1 ]Here follows in the Second Edition, Supplement XII.

[1 ]That is a transcendental deduction.

[1 ]Read deren instead of dessen.

[1 ]The last paragraph is omitted in the Second Edition. There is instead a criticism of Locke and Hume, Supplement XIII. The Deduction of the Categories is much changed, as seen in Supplement XIV.

[1 ]This point is of great importance and should be carefully considered. All representations have a necessary relation to some possible empirical consciousness, for if they did not possess that relation, and if it were entirely impossible to become conscious of them, this would be the same as if they did not exist. All empirical consciousness has a necessary relation to a transcendental consciousness, which precedes all single experiences, namely, the consciousness of my own self as the original apperception. It is absolutely necessary therefore that in my knowledge all consciousness should belong to one consciousness of my own self. Here we have a synthetical unity of the manifold (consciousness) which can be known a priori, and which may thus supply a foundation for synthetical propositions a priori concerning pure thinking in the same way as space and time supply a foundation for synthetical propositions which concern the form of mere intuition.

The synthetical proposition that the different kinds of empirical consciousness must be connected in one self-consciousness, is the very first and synthetical foundation of all our thinking. It should be remembered that the mere representation of the Ego in reference to all other representations (the collective unity of which would be impossible without it) constitutes our transcendental consciousness. It does not matter whether that representation is clear (empirical consciousness) or confused, not even whether it is real; but the possibility of the logical form of all knowledge rests necessarily on the relation to this apperception as a faculty.

[1 ]It has hardly struck any psychologist that this imagination is a necessary ingredient of perception. This was partly owing to their confining this faculty to reproduction, partly to our belief that the senses do not only give us impressions, but compound them also for us, thus producing pictures of objects. This, however, beyond our receptivity of impressions, requires something more, namely, a function for their synthesis.

[1 ]Of may be omitted, if we read aller empirischer Gebrauch.

[1 ]That is, out of one, or out of the unity of apperception.