The Main Transcendental Problem, Part 3

§ 40.

§ 40.

PURE mathematics and pure science of nature had no occasion for such a deduction, as we have made of both, for their own safety and certainty. For the former rests upon its own evidence; and the latter (though sprung from pure sources of the understanding) upon experience and its thorough confirmation. Physics cannot altogether refuse and dispense with the testimony of the latter; because with all its certainty, it can never, as philosophy, rival mathematics. Both sciences therefore stood in need of this inquiry, not for themselves, but for the sake of another science, metaphysics.

Pure mathematics and pure natural science did not need the kind of deduction we have provided for each. They did not need it to establish their own security and certainty. Mathematics rests on its own evidence. Natural science, though it comes from pure sources of the understanding, rests on experience and thorough confirmation by experience. Physics cannot entirely reject or do without the testimony of experience. For all its certainty, physics can never rival mathematics as a philosophical science. Both sciences therefore needed this inquiry, not for their own sake, but for the sake of another science, metaphysics.

Metaphysics has to do not only with concepts of nature, which always find their application in experience, but also with pure rational concepts, which never can be given in any possible experience. Consequently the objective reality of these concepts (viz., that they are not mere chimeras), and the truth or falsity of metaphysical assertions, cannot be discovered or confirmed by any experience. This part of metaphysics however is precisely what constitutes its essential end, to which the rest is only a means, and thus this science is in need of such a deduction for its own sake. The third question now proposed relates therefore as it were to the root and essential difference of metaphysics, i.e., the occupation of Reason with itself, and the supposed knowledge of objects arising immediately from this incubation of its own concepts, without requiring, or indeed being able to reach that knowledge through, experience.²⁸

Metaphysics deals not only with concepts of nature, which always apply to experience. It also deals with pure concepts of reason, which can never be given in any possible experience. No experience can therefore establish or confirm the objective reality of these concepts (that is, show that they are not mere fantasies). Nor can experience establish or confirm the truth or falsity of metaphysical claims. Yet this part of metaphysics is precisely its essential aim. The rest is only a means to that end. Metaphysics therefore needs such a deduction for its own sake. The third question now proposed thus concerns, as it were, the root and defining feature of metaphysics. This is reason’s occupation with itself and the supposed knowledge of objects that arises directly from reason’s working over its own concepts. This knowledge does not require experience and indeed cannot be reached through experience.²⁸

Without solving this problem reason never is justified. The empirical use to which reason limits the pure understanding, does not fully satisfy the proper destination of the latter. Every single experience is only a part of the whole sphere of its domain, but the absolute totality of all possible experience is itself not experience. Yet it is a necessary [concrete] problem for reason, the mere representation of which requires concepts quite different from the categories, whose use is only immanent, or refers to experience, so far as it can be given. Whereas the concepts of reason aim at the completeness, i.e., the collective unity of all possible experience, and thereby transcend every given experience. Thus they become transcendent.

Reason can never be justified without solving this problem. Reason limits the pure understanding to use in experience. But this limit does not fully satisfy the understanding’s proper purpose. Each individual experience is only part of the whole domain of possible experience. The absolute totality of all possible experience is not itself an experience. Yet it is a necessary [concrete] problem for reason. Even representing this totality requires concepts quite different from the categories. The categories are used only immanently, that is, in relation to experience as it can be given. The concepts of reason, by contrast, aim at completeness, meaning the collective unity of all possible experience. They therefore go beyond every given experience and become transcendent.

As the understanding stands in need of categories for experience, reason contains in itself the source of ideas, by which I mean necessary concepts, whose object cannot be given in any experience. The latter are inherent in the nature of reason, as the former are in that of the understanding. While the former carry with them an illusion likely to mislead, the illusion of the latter is inevitable, though it certainly can be kept from misleading us.

Just as understanding needs categories for experience, reason contains the source of ideas. By ideas, I mean necessary concepts whose objects cannot be given in any experience. Ideas are inherent in the nature of reason, just as categories are inherent in the nature of understanding. Categories carry an illusion that is likely to mislead us. The illusion of ideas, however, is inevitable, though it can certainly be kept from misleading us.

Since all illusion consists in holding the subjective ground of our judgments to be objective, a self-knowledge of pure reason in its transcendent (exaggerated) use is the sole preservative from the aberrations into which reason falls when it mistakes its destination, and refers that to the object transcendently, which only regards its own subject and its guidance in all immanent use.

All illusion consists in mistaking the subjective basis of our judgments for an objective one. So pure reason must understand its own transcendent (exaggerated) use to guard against the errors it falls into when it mistakes its purpose. In such cases, reason applies to an object in a transcendent way what concerns only itself and its guidance in all immanent use.

§ 41. The distinction of ideas, that is, of pure concepts of reason, from categories, or pure concepts of the understanding, as cognitions of a quite distinct species, origin and use, is so important a point in founding a science which is to contain the system of all these a priori cognitions, that without this distinction metaphysics is absolutely impossible, or is at best a random, bungling attempt to build a castle in the air without a knowledge of the materials or of their fitness for any purpose. Had the Critique of Pure Reason done nothing but first point out this distinction, it had thereby contributed more to clear up our conception of, and to guide our inquiry in, the field of metaphysics, than all the vain efforts which have hitherto been made to satisfy the transcendent problems of pure reason, without ever surmising that we were in quite another field than that of the understanding, and hence classing concepts of the understanding and those of reason together, as if they were of the same kind.

§ 41. Ideas are pure concepts of reason, while categories are pure concepts of the understanding. These are entirely different kinds of knowledge, with different origins and uses. The distinction is crucial to establishing a science that contains the system of all this a priori knowledge. Without it, metaphysics is absolutely impossible. At best, it is a haphazard, clumsy attempt to build a castle in the air, without knowing the materials or whether they are fit for any purpose. Suppose the Critique of Pure Reason had done nothing more than first point out this distinction. Even so, it would have done more to clarify our conception of metaphysics and guide our inquiry in that field than all previous efforts. Those efforts have tried in vain to resolve the transcendent problems of pure reason. Yet they never suspected that they were in a field entirely different from that of the understanding. As a result, they grouped concepts of the understanding with concepts of reason, as if they were the same kind.

§ 42. All pure cognitions of the understanding have this feature, that their concepts present themselves in experience, and their principles can be confirmed by it; whereas the transcendent cognitions of reason cannot, either as ideas, appear in experience, or as propositions ever be confirmed or refuted by it. Hence whatever errors may slip in unawares, can only be discovered by pure reason itself—a discovery of much difficulty, because this very reason naturally becomes dialectical by means of its ideas, and this unavoidable illusion cannot be limited by any objective and dogmatical researches into things, but by a subjective investigation of reason itself as a source of ideas.

§ 42. All pure cognitions of the understanding share a feature: their concepts appear in experience, and their principles can be confirmed by experience. Transcendent cognitions of reason, however, cannot do either. Their ideas cannot appear in experience, and their propositions can never be confirmed or refuted by experience. So only pure reason itself can discover whatever errors may slip in unnoticed. This is very difficult, because reason itself naturally becomes dialectical through its ideas. This unavoidable illusion cannot be limited by objective, dogmatic inquiry into things. It can be limited only by a subjective investigation of reason itself as a source of ideas.

§ 43. In the Critique of Pure Reason it was always my greatest care to endeavor not only carefully to distinguish the several species of cognition, but to derive concepts belonging to each one of them from their common source. I did this in order that by knowing whence they originated, I might determine their use with safety, and also have the unanticipated but invaluable advantage of knowing the completeness of my enumeration, classification and specification of concepts a priori, and therefore according to principles. Without this, metaphysics is mere rhapsody, in which no one knows whether he has enough, or whether and where something is still wanting. We can indeed have this advantage only in pure philosophy, but of this philosophy it constitutes the very essence.

§ 43. In the Critique of Pure Reason, I always took the greatest care to distinguish the different kinds of knowledge. I also sought to trace the concepts belonging to each kind back to their common source. By knowing where these concepts came from, I could determine with confidence how to use them. This also gave me an unexpected but invaluable advantage. I could know that my list, classification, and description of a priori concepts were complete and based on principles. Without this assurance, metaphysics is a mere jumble. No one knows whether it contains enough, or whether anything is still missing and where. We can have this advantage only in pure philosophy, but it is essential to pure philosophy.

As I had found the origin of the categories in the four logical functions of all the judgments of the understanding, it was quite natural to seek the origin of the ideas in the three functions of the syllogisms of reason. For as soon as these pure concepts of reason (the transcendental ideas) are given, they could hardly, except they be held innate, be found anywhere else, than in the same activity of reason, which, so far as it regards mere form, constitutes the logical element of the syllogisms of reason; but, so far as it represents judgments of the understanding with respect to the one or to the other form a priori, constitutes transcendental concepts of pure reason.

I had found the origin of the categories in the four logical functions of all judgments of the understanding. So it was natural to seek the origin of the ideas in the three functions of syllogisms of reason. For once these pure concepts of reason, or transcendental ideas, are given, they could hardly originate anywhere else, unless they are innate. They must originate in the activity of reason itself. Considered in terms of mere form, this activity constitutes the logical element of syllogisms of reason. But insofar as it represents judgments of the understanding in relation to one form or another a priori, it constitutes transcendental concepts of pure reason.

The formal distinction of syllogisms renders their division into categorical, hypothetical, and disjunctive necessary. The concepts of reason founded on them contained therefore, first, the idea of the complete subject (the substantial); secondly, the idea of the complete series of conditions; thirdly, the determination of all concepts in the idea of a complete complex of that which is possible.²⁹ The first idea is psychological, the second cosmological, the third theological, and, as all three give occasion to Dialectics, yet each in its own way, the division of the whole Dialects of pure reason into its Paralogism, its Antinomy, and its Ideal, was arranged accordingly. Through this deduction we may feel assured that all the claims of pure reason are completely represented, and that none can be wanting; because the faculty of reason itself, whence they all take their origin, is thereby completely surveyed.

Formal differences make it necessary to divide syllogisms into categorical, hypothetical, and disjunctive types. The concepts of reason based on them therefore contained three ideas. The first was the idea of the complete subject (the substantial). The second was the idea of the complete series of conditions. The third was the determination of all concepts in the idea of a complete complex of what is possible.²⁹ The first idea is psychological, the second cosmological, and the third theological. All three give rise to Dialectics, each in its own way. Accordingly, the whole Dialectic of pure reason was divided into its Paralogism, its Antinomy, and its Ideal. Through this deduction, we may be assured that all the claims of pure reason are represented and none are missing. This is because the deduction completely surveys the faculty of reason itself, from which all those claims originate.

§ 44. In these general considerations it is also remarkable that the ideas of reason are unlike the categories, of no service to the use of our understanding in experience, but quite dispensable, and become even an impediment to the maxims of a rational cognition of nature. Yet in another aspect still to be determined they are necessary. Whether the soul is or is not a simple substance, is of no consequence to us in the explanation of its phenomena. For we cannot render the notion of a simple being intelligible by any possible experience that is sensuous or concrete. The notion is therefore quite void as regards all hoped-for insight into the cause of phenomena, and cannot at all serve as a principle of the explanation of that which internal or external experience supplies. So the cosmological ideas of the beginning of the world or of its eternity (a parte ante) cannot be of any greater service to us for the explanation of any event in the world itself. And finally we must, according to a right maxim of the philosophy of nature, refrain from all explanations of the design of nature, drawn from the will of a Supreme Being; because this would not be natural philosophy, but an acknowledgment that we have come to the end of it. The use of these ideas, therefore, is quite different from that of those categories by which (and by the principles built upon which) experience itself first becomes possible. But our laborious analytics of the understanding would be superfluous if we had nothing else in view than the mere cognition of nature as it can be given in experience; for reason does its work, both in mathematics and in the science of nature, quite safely and well without any of this subtle deduction. Therefore our Critique of the Understanding combines with the ideas of pure reason for a purpose which lies beyond the empirical use of the understanding; but this we have above declared to be in this aspect totally inadmissible, and without any object or meaning. Yet there must be a harmony between that of the nature of reason and that of the understanding, and the former must contribute to the perfection of the latter, and cannot possibly upset it.

§ 44. Another remarkable point emerges from these general considerations. Unlike the categories, the ideas of reason are of no use to the understanding in experience. They are entirely dispensable and even obstruct the principles of rational knowledge of nature. Yet they are necessary in another respect, which remains to be determined. Whether the soul is a simple substance or not makes no difference to our explanation of its phenomena. No possible sensory or concrete experience can make the concept of a simple being intelligible to us. The concept is therefore entirely empty as a source of insight into the cause of phenomena. It cannot serve at all as a principle for explaining what inner or outer experience provides. Likewise, the cosmological ideas of the beginning of the world or its eternity in the past are no more useful for explaining any event in the world itself. Finally, a sound principle of natural philosophy requires us to avoid all explanations of nature’s design based on the will of a Supreme Being. Such explanations would not be natural philosophy. They would be an admission that we have reached its limits. The use of these ideas is therefore quite different from the use of the categories. The categories, and the principles based on them, first make experience itself possible. But our painstaking analysis of the understanding would be unnecessary if we sought only knowledge of nature as experience can give it to us. In mathematics and natural science, reason works quite safely and well without any of this subtle deduction. Our Critique of the Understanding therefore joins with the ideas of pure reason for a purpose beyond the empirical use of the understanding. Yet we have already declared that purpose, in this respect, entirely inadmissible and without any object or meaning. Still, the nature of reason and the nature of the understanding must be in harmony. Reason must contribute to perfecting the understanding, and it cannot possibly upset it.

The solution of this question is as follows: Pure reason does not in its ideas point to particular objects, which lie beyond the field of experience, but only requires completeness of the use of the understanding in the system of experience. But this completeness can be a completeness of principles only, not of intuitions (i.e., concrete atsights or Anschauungen) and of objects. In order however to represent the ideas definitely, reason conceives them after the fashion of the cognition of an object. The cognition is as far as these rules are concerned completely determined, but the object is only an idea invented for the purpose of bringing the cognition of the understanding as near as possible to the completeness represented by that idea.

The answer to this question is as follows. In its ideas, pure reason does not point to particular objects beyond the field of experience. It requires only the complete use of the understanding within the system of experience. But this completeness can apply only to principles, not to intuitions (concrete perceptions, or Anschauungen) or objects. To give these ideas a definite form, however, reason conceives them as though they were knowledge of an object. The knowledge is completely determined by these rules. But the object is only an idea, invented to bring the knowledge gained by the understanding as close as possible to the completeness represented by that idea.

Prefatory Remark to the Dialectics of Pure Reason.

Introductory Note to the Dialectic of Pure Reason.

§ 45. We have above shown in §§ 33 and 34 that the purity of the categories from all admixture of sensuous determinations may mislead reason into extending their use, quite beyond all experience, to things in themselves; though as these categories themselves find no intuition which can give them meaning or sense in concreto, they, as mere logical functions, can represent a thing in general, but not give by themselves alone a determinate concept of anything. Such hyperbolical objects are distinguised by the appellation of Noümena, or pure beings of the understanding (or better, beings of thought), such as, for example, "substance," but conceived without permanence in time, or "cause," but not acting in time, etc. Here predicates, that only serve to make the conformity-to-law of experience possible, are applied to these concepts, and yet they are deprived of all the conditions of intuition, on which alone experience is possible, and so these concepts lose all significance.

§ 45. In §§ 33 and 34, we showed that the categories are free from any mixture of sensory elements. This purity may mislead reason into applying them to things in themselves, far beyond all experience. Yet the categories themselves find no intuition that can give them concrete meaning or sense. As mere logical functions, they can represent a thing in general. But they cannot by themselves provide a definite concept of anything. Such far-reaching objects are called Noümena, or pure beings of the understanding (or better, beings of thought). Examples include "substance" conceived without permanence in time, or "cause" that does not act in time, and so on. Here, predicates that serve only to make the lawfulness of experience possible are applied to these concepts. Yet the concepts are stripped of all the conditions of intuition, which alone make experience possible. As a result, the concepts lose all meaning.

There is no danger, however, of the understanding spontaneously making an excursion so very wantonly beyond its own bounds into the field of the mere creatures of thought, without being impelled by foreign laws. But when reason, which cannot be fully satisfied with any empirical use of the rules of the understanding, as being always conditioned, requires a completion of this chain of conditions, then the understanding is forced out of its sphere. And then it partly represents objects of experience in a series so extended that no experience can grasp, partly even (with a view to complete the series) it seeks entirely beyond it noumena, to which it can attach that chain, and so, having at last escaped from the conditions of experience, make its attitude as it were final. These are then the transcendental ideas, which, though according to the true but hidden ends of the natural determination of our reason, they may aim not at extravagant concepts, but at an unbounded extension of their empirical use, yet seduce the understanding by an unavoidable illusion to a transcendent use, which, though deceitful, cannot be restrained within the bounds of experience by any resolution, but only by scientific instruction and with much difficulty.

There is no danger, however, that the understanding will stray so recklessly beyond its own bounds on its own. It will not venture into the realm of mere creations of thought unless it is driven by laws external to it. But reason cannot be fully satisfied with any empirical use of the rules of the understanding, because such use is always conditioned. When reason demands a completion of this chain of conditions, the understanding is forced beyond its own sphere. Then, on the one hand, the understanding represents objects of experience in a series so long that no experience can grasp it. On the other hand, in order to complete the series, the understanding seeks noumena entirely beyond experience. It seeks noumena to which it can attach the chain, so that, having escaped the conditions of experience, it can reach a position that is, as it were, final. These are the transcendental ideas. According to the true but hidden ends of the natural purpose of our reason, these ideas may aim not at extravagant concepts, but at an unlimited extension of their empirical use. Yet they tempt the understanding, through an unavoidable illusion, into a transcendent use. This use is deceptive, but no decision alone can keep it within the bounds of experience. Only scientific instruction can do so, and even then with great difficulty.

I. The Psychological Idea.³⁰

I. The Psychological Idea.³⁰

§ 46. People have long since observed, that in all substances the proper subject, that which remains after all the accidents (as predicates) are abstracted, consequently that which forms the substance of things remains unknown, and various complaints have been made concerning these limits to our knowledge. But it will be well to consider that the human understanding is not to be blamed for its inability to know the substance of things, that is, to determine it by itself, but rather for requiring to cognise it which is a mere idea definitely as though it were a given object. Pure reason requires us to seek for every predicate of a thing its proper subject, and for this subject, which is itself necessarily nothing but a predicate, its subject, and so on indefinitely (or as far as we can reach). But hence it follows, that we must not hold anything, at which we can arrive, to be an ultimate subject, and that substance itself never can be thought by our understanding, however deep we may penetrate, even if all nature were unveiled to us. For the specific nature of our understanding consists in thinking everything discursively, that is, representing it by concepts, and so by mere predicates, to which therefore the absolute subject must always be wanting. Hence all the real properties, by which we cognise bodies, are mere accidents, not excepting impenetrability, which we can only represent to ourselves as the effect of a power of which the subject is unknown to us.

§ 46. People have long observed that the true subject of any substance remains unknown. This subject is what remains after we set aside all its accidents, or predicates. It is therefore what constitutes the substance of things. People have often complained about these limits to our knowledge. But we should consider that the human understanding is not at fault because it cannot know substance itself. The fault lies rather in demanding definite knowledge of substance, a mere idea, as if it were a given object. Pure reason requires us to seek the true subject of every predicate of a thing. It also requires us to seek the subject of that subject, which is itself necessarily only a predicate. We must continue this search indefinitely, or as far as we can. It follows that we must not regard anything we reach as an ultimate subject. However deeply we penetrate, our understanding can never think substance itself, even if all nature were revealed to us. For our understanding is discursive by nature. It thinks of everything through concepts, and therefore through mere predicates. The absolute subject must always be missing from these predicates. Hence all the real properties through which we know bodies are mere accidents. This includes impenetrability. We can represent impenetrability only as the effect of a power whose subject is unknown to us.

Now we appear to have this substance in the consciousness of ourselves (in the thinking subject), and indeed in an immediate intuition; for all the predicates of an internal sense refer to the ego, as a subject, and I cannot conceive myself as the predicate of any other subject. Hence completeness in the reference of the given concepts as predicates to a subject—not merely an idea, but an object—that is, the absolute subject itself, seems to be given in experience. But this expectation is disappointed. For the ego is not a concept,³¹ but only the indication of the object of the internal sense, so far as we cognise it by no further predicate. Consequently it cannot be in itself a predicate of any other thing; but just as little can it be a determinate concept of an absolute subject, but is, as in all other cases, only the reference of the internal phenomena to their unknown subject. Yet this idea (which serves very well, as a regulative principle, totally to destroy all materialistic explanations of the internal phenomena of the soul) occasions by a very natural misunderstanding a very specious argument, which, from this supposed cognition of the substance of our thinking being, infers its nature, so far as the knowledge of it falls quite without the complex of experience.

Now we seem to find this substance in our consciousness of ourselves, as thinking subjects. We even seem to find it in immediate intuition. For all predicates of inner sense refer to the ego as their subject. And I cannot conceive of myself as the predicate of any other subject. So experience seems to give us a complete reference of the given concepts as predicates to a subject. This subject seems to be not merely an idea, but an object, the absolute subject itself. But this expectation is disappointed. For the ego is not a concept,³¹ but only an indication of the object of inner sense. It indicates that object insofar as we know it through no further predicate. Consequently, the ego cannot itself be a predicate of anything else. But neither can it be a determinate concept of an absolute subject. As in all other cases, it only refers inner phenomena to their unknown subject. Yet this idea serves very well as a regulative principle for completely refuting all materialistic explanations of the soul’s inner phenomena. By a very natural misunderstanding, however, it gives rise to a very plausible argument. From this supposed knowledge of the substance of our thinking being, the argument infers its nature insofar as knowledge of it lies entirely outside the whole of experience.

§ 47. But though we may call this thinking self (the soul) substance, as being the ultimate subject of thinking which cannot be further represented as the predicate of another thing; it remains quite empty and without significance, if permanence—the quality which renders the concept of substances in experience fruitful—cannot be proved of it.

§ 47. But we may call the thinking self (the soul) a substance. This is because it is the ultimate subject of thought and cannot itself be represented as a predicate of something else. Even so, calling it a substance remains empty and meaningless if we cannot prove that the soul is permanent. Permanence is the quality that makes the concept of substance useful in experience.

But permanence can never be proved of the concept of a substance, as a thing in itself, but for the purposes of experience only. This is sufficiently shown by the first Analogy of Experience,³² and who ever will not yield to this proof may try for himself whether he can succeed in proving, from the concept of a subject which does not exist itself as the predicate of another thing, that its existence is thoroughly permanent, and that it cannot either in itself or by any natural cause originate or be annihilated. These synthetical a priori propositions can never be proved in themselves, but only in reference to things as objects of possible experience.

But permanence can never be proved from the concept of a substance as a thing in itself. It can be proved only for the purposes of experience. The first Analogy of Experience³² shows this sufficiently. Anyone who will not accept this proof can try to prove it for themselves. They can start with the concept of a subject that does not itself exist as a predicate of another thing. From that concept, they can try to prove that the subject’s existence is completely permanent. They can also try to prove that it cannot come into existence or be destroyed, either on its own or through any natural cause. These synthetic a priori propositions can never be proved in themselves. They can be proved only in relation to things as objects of possible experience.

§ 48. If therefore from the concept of the soul as a substance, we would infer its permanence, this can hold good as regards possible experience only, not [of the soul] as a thing in itself and beyond all possible experience. But life is the subjective condition of all our possible experience, consequently we can only infer the permanence of the soul in life; for the death of man is the end of all experience which concerns the soul as an object of experience, except the contrary be proved, which is the very question in hand. The permanence of the soul can therefore only be proved (and no one cares for that) during the life of man, but not, as we desire to do, after death; and for this general reason, that the concept of substance, so far as it is to be considered necessarily combined with the concept of permanence, can be so combined only according to the principles of possible experience, and therefore for the purposes of experience only.³³

§ 48. Suppose we infer that the soul is permanent from the concept of the soul as a substance. This inference can hold only for possible experience. It cannot hold for the soul as a thing in itself, beyond all possible experience. But life is the subjective condition of all our possible experience. So we can infer only that the soul is permanent during life. For a person's death ends all experience of the soul as an object of experience, unless the contrary is proved. But whether the contrary can be proved is the very question at issue. Therefore, the permanence of the soul can be proved only during a person's life. No one cares about that. It cannot be proved after death, as we want to do. The general reason is that the concept of substance can be necessarily linked with the concept of permanence only through the principles of possible experience, and therefore only for the purposes of experience.³³

§ 49. That there is something real without us which not only corresponds, but must correspond, to our external perceptions, can likewise be proved to be not a connexion of things in themselves, but for the sake of experience. This means that there is something empirical, i.e., some phenomenon in space without us, that admits of a satisfactory proof, for we have nothing to do with other objects than those which belong to possible experience; because objects which cannot be given us in any experience, do not exist for us. Empirically without me is that which appears in space, and space, together with all the phenomena which it contains, belongs to the representations, whose connexion according to laws of experience proves their objective truth, just as the connexion of the phenomena of the internal sense proves the actuality of my soul (as an object of the internal sense). By means of external experience I am conscious of the actuality of bodies, as external phenomena in space, in the same manner as by means of the internal experience I am conscious of the existence of my soul in time, but this soul is only cognised as an object of the internal sense by phenomena that constitute an internal state, and of which the essence in itself, which forms the basis of these phenomena, is unknown. Cartesian idealism therefore does nothing but distinguish external experience from dreaming; and the conformity to law (as a criterion of its truth) of the former, from the irregularity and the false illusion of the latter. In both it presupposes space and time as conditions of the existence of objects, and it only inquires whether the objects of the external senses, which we when awake put in space, are as actually to be found in it, as the object of the internal sense, the soul, is in time; that is, whether experience carries with it sure criteria to distinguish it from imagination. This doubt, however, may be easily disposed of, and we always do so in common life by investigating the connexion of phenomena in both space and time according to universal laws of experience, and we cannot doubt, when the representation of external things throughout agrees therewith, that they constitute truthful experience. Material idealism, in which phenomena are considered as such only according to their connexion in experience, may accordingly be very easily refuted; and it is just as sure an experience, that bodies exist without us (in space), as that I myself exist according to the representation of the internal sense (in time): for the notion without us, only signifies existence in space. However as the Ego in the proposition, "I am," means not only the object of internal intuition (in time), but the subject of consciousness, just as body means not only external intuition (in space), but the thing-in-itself, which is the basis of this phenomenon; [as this is the case] the question, whether bodies (as phenomena of the external sense) exist as bodies apart from my thoughts, may without any hesitation be denied in nature. But the question, whether I myself as a phenomenon of the internal sense (the soul according to empirical psychology) exist apart from my faculty of representation in time, is an exactly similar inquiry, and must likewise be answered in the negative. And in this manner everything, when it is reduced to its true meaning, is decided and certain. The formal (which I have also called transcendental) actually abolishes the material, or Cartesian, idealism. For if space be nothing but a form of my sensibility, it is as a representation in me just as actual as I myself am, and nothing but the empirical truth of the representations in it remains for consideration. But, if this is not the case, if space and the phenomena in it are something existing without us, then all the criteria of experience beyond our perception can never prove the actuality of these objects without us.

§ 49. We can also prove that something real outside us not only corresponds to our external perceptions, but must correspond to them. This proof concerns experience, not a connection between things in themselves. What can be satisfactorily proved is that something empirical exists, namely, some appearance in space outside us. For we are concerned only with objects belonging to possible experience. Objects that cannot be given to us in any experience do not exist for us. What is empirically outside me is what appears in space. Space, together with all the appearances it contains, belongs to our representations. Their connection according to laws of experience proves their objective truth. In the same way, the connection between appearances of the inner sense proves the reality of my soul as an object of that sense. Through outer experience, I am conscious of the reality of bodies as external appearances in space. In the same way, through inner experience, I am conscious of the existence of my soul in time. But I know the soul as an object of the inner sense only through appearances that make up an inner state. The essence in itself underlying these appearances is unknown. Cartesian idealism therefore does no more than distinguish outer experience from dreaming. It distinguishes the lawfulness of outer experience, as a test of its truth, from the irregularity and false illusion of dreams. In both cases, it assumes space and time as conditions for the existence of objects. It asks only whether the objects of the outer senses that we place in space when awake are really found there. Are they found there just as the object of the inner sense, the soul, is found in time? In other words, does experience provide reliable tests for distinguishing it from imagination? This doubt, however, can easily be resolved. We always resolve it in everyday life by examining how appearances connect in both space and time according to universal laws of experience. When our representation of external things agrees with those laws throughout, we cannot doubt that it constitutes truthful experience. Material idealism, which considers appearances as such only in terms of their connection in experience, can therefore be refuted very easily. It is just as certain an experience that bodies exist outside us in space as that I myself exist according to the representation of the inner sense in time. For “outside us” means only existence in space. However, the Ego in the proposition "I am" means more than the object of inner intuition in time. It also means the subject of consciousness. Similarly, body means more than outer intuition in space. It also means the thing-in-itself underlying that appearance. This being so, we can deny without hesitation that bodies, as appearances of the outer sense, exist as bodies apart from my thoughts. But an exactly similar question can be asked about myself as an appearance of the inner sense, that is, the soul as understood by empirical psychology. Do I exist in time apart from my faculty of representation? This question must likewise be answered in the negative. In this way, once everything is reduced to its true meaning, everything is settled and certain. Formal idealism, which I have also called transcendental idealism, actually abolishes material, or Cartesian, idealism. For if space is nothing but a form of my sensibility, then space, as a representation in me, is just as real as I myself am. All that remains to be considered is the empirical truth of the representations in space. But suppose this is not the case. If space and the appearances in it are something that exists outside us, then all the criteria of experience beyond our perception can never prove the reality of these objects outside us.

II. The Cosmological Idea.³⁴

II. The Cosmological Idea.³⁴

§ 50. This product of pure reason in its transcendent use is its most remarkable curiosity. It serves as a very powerful agent to rouse philosophy from its dogmatic slumber, and to stimulate it to the arduous task of undertaking a Critique of Reason itself.

§ 50. This product of pure reason in its transcendent use is its most remarkable curiosity. It is a powerful force for waking philosophy from its dogmatic slumber. It also spurs philosophy to undertake the difficult task of a Critique of Reason itself.

I term this idea cosmological, because it always takes its object only from the sensible world, and does not use any other than those whose object is given to sense, consequently it remains in this respect in its native home, it does not become transcendent, and is therefore so far not mere idea; whereas, to conceive the soul as a simple substance, already means to conceive such an object (the simple) as cannot be presented to the senses. Yet the cosmological idea extends the connexion of the conditioned with its condition (whether the connexion is mathematical or dynamical) so far, that experience never can keep up with it. It is therefore with regard to this point always an idea, whose object never can be adequately given in any experience.

I call this idea cosmological because its object always comes from the world of the senses. It uses only objects given to the senses. In this respect, it remains on familiar ground. It does not go beyond the world of the senses and is therefore not merely an idea in this respect. By contrast, to conceive of the soul as a simple substance is already to conceive of an object (the simple) that cannot be presented to the senses. Yet the cosmological idea extends the connection between the conditioned and its condition beyond anything experience can follow. This is true whether the connection is mathematical or dynamical. In this respect, it is therefore always an idea whose object can never be adequately given in any experience.

§ 51. In the first place, the use of a system of categories becomes here so obvious and unmistakable, that even if there were not several other proofs of it, this alone would sufficiently prove it indispensable in the system of pure reason. There are only four such transcendent ideas, as there are so many classes of categories; in each of which, however, they refer only to the absolute completeness of the series of the conditions for a given conditioned. In analogy to these cosmological ideas there are only four kinds of dialectical assertions of pure reason, which, as they are dialectical, thereby prove, that to each of them, on equally specious principles of pure reason, a contradictory assertion stands opposed. As all the metaphysical art of the most subtile distinction cannot prevent this opposition, it compels the philosopher to recur to the first sources of pure reason itself. This Antinomy, not arbitrarily invented, but founded in the nature of human reason, and hence unavoidable and never ceasing, contains the following four theses together with their antitheses:

§ 51. First, the usefulness of a system of categories is so clear and unmistakable here that this alone would prove it indispensable to the system of pure reason, even if there were no other proofs. There are only four such transcendent ideas, corresponding to the four classes of categories. In each class, however, these ideas concern only the absolute completeness of the series of conditions for something given that is conditioned. Corresponding to these cosmological ideas, there are only four kinds of dialectical assertions of pure reason. Because these assertions are dialectical, each has a contradictory assertion opposed to it, based on equally plausible principles of pure reason. All the metaphysical skill in making the finest distinctions cannot prevent this opposition. It therefore forces the philosopher to return to the original sources of pure reason itself. This Antinomy is not invented arbitrarily, but is founded in the nature of human reason. It is therefore unavoidable and never ceases. It contains the following four theses together with their antitheses:

Thesis.

Thesis.

The World has, as to Time and Space, a Beginning (limit).

The world has a beginning in time and a limit in space.

Antithesis.

Antithesis.

The World is, as to Time and Space, infinite.

The world is infinite in time and space.

Thesis.

Thesis.

Everything in the World consists of [elements that are] simple.

Everything in the world consists of simple elements.

Antithesis.

Antithesis.

There is nothing simple, but everything is composite.

Nothing is simple. Everything is made up of parts.

Thesis.

Thesis.

There are in the World Causes through Freedom.

There are causes in the world that arise from freedom.

Antithesis.

Antithesis.

There is no Liberty, but all is Nature.

There is no freedom, but everything is governed by nature.

Thesis.

Thesis.

In the Series of the World-Causes there is some necessary Being.

In the chain of causes of the world, there is a necessary being.

Antithesis.

Antithesis.

There is Nothing necessary in the World, but in this Series All is incidental.

Nothing in the world is necessary, but everything in this series could have been otherwise.

§ 52.a. Here is the most singular phenomenon of human reason, no other instance of which can be shown in any other use. If we, as is commonly done, represent to ourselves the appearances of the sensible world as things in themselves, if we assume the principles of their combination as principles universally valid of things in themselves and not merely of experience, as is usually, nay without our Critique, unavoidably done, there arises an unexpected conflict, which never can be removed in the common dogmatical way; because the thesis, as well as the antithesis, can be shown by equally clear, evident, and irresistible proofs—for I pledge myself as to the correctness of all these proofs—and reason therefore perceives that it is divided with itself, a state at which the sceptic rejoices, but which must make the critical philosopher pause and feel ill at ease.

§ 52.a. Here is the most remarkable phenomenon of human reason. Nothing like it can be found in any other use of reason. We commonly regard appearances in the sensible world as things in themselves. We also usually assume that the principles governing how these appearances combine apply universally to things in themselves, not merely to experience. Indeed, without our Critique, we cannot avoid making this assumption. If we do so, an unexpected conflict arises. The usual dogmatic approach can never resolve this conflict. For both the thesis and the antithesis can be supported by proofs that are equally clear, evident, and irresistible. I pledge myself to the correctness of all these proofs. Reason therefore sees that it is divided against itself. The sceptic rejoices at this state of affairs, but it must make the critical philosopher pause and feel uneasy.

§ 52.b. We may blunder in various ways in metaphysics without any fear of being detected in falsehood. For we never can be refuted by experience if we but avoid self-contradiction, which in synthetical, though purely fictitious propositions, may be done whenever the concepts, which we connect, are mere ideas, that cannot be given (in their whole content) in experience. For how can we make out by experience, whether the world is from eternity or had a beginning, whether matter is infinitely divisible or consists of simple parts? Such concept cannot be given in any experience, be it ever so extensive, and consequently the falsehood either of the positive or the negative proposition cannot be discovered by this touch-stone.

§ 52.b. We can make mistakes in metaphysics in various ways without fear of being found wrong. Experience can never refute us if we avoid contradicting ourselves. We can avoid contradiction even in synthetic propositions that are purely fictional. We can do so whenever the concepts we connect are mere ideas whose full content cannot be given in experience. For how can experience tell us whether the world has existed forever or had a beginning? How can experience tell us whether matter is infinitely divisible or consists of simple parts? These concepts cannot be given in any experience, however extensive. Consequently, experience cannot reveal whether either the affirmative or the negative proposition is false.

The only possible way in which reason could have revealed unintentionally its secret Dialectics, falsely announced as Dogmatics, would be when it were made to ground an assertion upon a universally admitted principle, and to deduce the exact contrary with the greatest accuracy of inference from another which is equally granted. This is actually here the case with regard to four natural ideas of reason, whence four assertions on the one side, and as many counter-assertions on the other arise, each consistently following from universally-acknowledged principles. Thus they reveal by the use of these principles the dialectical illusion of pure reason which would otherwise forever remain concealed.

There was only one way reason could unintentionally reveal its hidden dialectics, falsely presented as dogmatics. Reason would have to base an assertion on a universally accepted principle. It would then have to deduce the exact opposite with perfect accuracy from another equally accepted principle. This is precisely what happens with four natural ideas of reason. From these arise four assertions on one side and four opposing assertions on the other. Each follows consistently from universally accepted principles. In this way, the use of these principles reveals the dialectical illusion of pure reason. Otherwise, this illusion would remain hidden forever.

This is therefore a decisive experiment, which must necessarily expose any error lying hidden in the assumptions of reason.³⁵ Contradictory propositions cannot both be false, except the concept, which is the subject of both, is self-contradictory; for example, the propositions, "a square circle is round, and a square circle is not round," are both false. For, as to the former it is false, that the circle is round, because it is quadrangular; and it is likewise false, that it is not round, that is, angular, because it is a circle. For the logical criterion of the impossibility of a concept consists in this, that if we presuppose it, two contradictory propositions both become false; consequently, as no middle between them is conceivable, nothing at all is thought by that concept.

This is therefore a decisive test, which must expose any error hidden in reason’s assumptions.³⁵ Two contradictory statements cannot both be false unless the concept they both concern contradicts itself. For example, the statements "a square circle is round, and a square circle is not round," are both false. The first is false because the circle has four sides. The second is also false because the figure is a circle, so it cannot be not round, or angular. The logical test for an impossible concept is that, if we assume the concept, two contradictory statements both become false. Since no middle ground between them is conceivable, the concept expresses no thought at all.

§ 52.c. The first two antinomies, which I call mathematical, because they are concerned with the addition or division of the homogeneous, are founded on such a self-contradictory concept; and hence I explain how it happens, that both the Thesis and Antithesis of the two are false.

§ 52.c. I call the first two antinomies mathematical because they concern adding or dividing things of the same kind. These antinomies are based on a self-contradictory concept. This explains how both the Thesis and the Antithesis of each antinomy can be false.

When I speak of objects in time and in space, it is not of things in themselves, of which I know nothing, but of things in appearance, that is, of experience, as the particular way of cognising objects which is afforded to man. I must not say of what I think in time or in space, that in itself, and independent of these my thoughts, it exists in space and in time; for in that case I should contradict myself; because space and time, together with the appearances in them, are nothing existing in themselves and outside of my representations, but are themselves only modes of representation, and it is palpably contradictory to say, that a mere mode of representation exists without our representation. Objects of the senses therefore exist only in experience; whereas to give them a self-subsisting existence apart from experience or before it, is merely to represent to ourselves that experience actually exists apart from experience or before it.

When I speak of objects in time and space, I do not mean things as they are in themselves. I know nothing about those. I mean things as they appear, that is, as we experience them. Experience is the particular way human beings can know objects. I must not say that something I think of in time or space exists in time and space in itself, independently of my thoughts. That would contradict what I have said. For space and time, along with the appearances in them, do not exist in themselves or outside my representations. They are only ways in which we represent things. And it is plainly contradictory to say that a mere way of representing things exists without our representing them. Objects of the senses therefore exist only in experience. To give them an independent existence apart from experience or before it is merely to imagine that experience itself exists apart from experience or before it.

Now if I inquire after the quantity of the world, as to space and time, it is equally impossible, as regards all my notions, to declare it infinite or to declare it finite. For neither assertion can be contained in experience, because experience either of an infinite space, or of an infinite time elapsed, or again, of the boundary of the world by a void space, or by an antecedent void time, is impossible; these are mere ideas. This quantity of the world, which is determined in either way, should therefore exist in the world itself apart from all experience. This contradicts the notion of a world of sense, which is merely a complex of the appearances whose existence and connexion occur only in our representations, that is, in experience, since this latter is not an object in itself, but a mere mode of representation. Hence it follows, that as the concept of an absolutely existing world of sense is self-contradictory, the solution of the problem concerning its quantity, whether attempted affirmatively or negatively, is always false.

Now, if I ask about the size of the world in space and time, I cannot declare it infinite or finite. Neither claim can be established by experience. We cannot experience infinite space or an infinite time that has already passed. Nor can we experience the world as bounded by empty space or by an earlier empty time. These are mere ideas. The world's size, whether determined as infinite or finite, would therefore have to exist in the world itself, apart from all experience. But this contradicts the concept of a world of sense. Such a world is merely a collection of appearances whose existence and connection occur only in our representations, that is, in experience. Experience is not an object in itself, but merely a way of representing things. It follows that the concept of an absolutely existing world of sense is self-contradictory. Therefore, any answer to the question of its size is false, whether we say it is finite or infinite.

The same holds good of the second antinomy, which relates to the division of phenomena. For these are mere representations, and the parts exist merely in their representation, consequently in the division, or in a possible experience where they are given, and the division reaches only as far as this latter reaches. To assume that an appearance, e.g., that of body, contains in itself before all experience all the parts, which any possible experience can ever reach, is to impute to a mere appearance, which can exist only in experience, an existence previous to experience. In other words, it would mean that mere representations exist before they can be found in our faculty of representation. Such an assertion is self-contradictory, as also every solution of our misunderstood problem, whether we maintain, that bodies in themselves consist of an infinite number of parts, or of a finite number of simple parts.

The same is true of the second antinomy, which concerns the division of appearances. Appearances are mere representations. Their parts exist only as they are represented, and therefore only in the division or in a possible experience where they are given. The division reaches only as far as that possible experience reaches. Take the appearance of a body, for example. Suppose we assume that it already contains, before any experience, all the parts that any possible experience could ever reach. We would then be giving an existence prior to experience to a mere appearance, which can exist only in experience. In other words, we would be saying that mere representations exist before they can be found in our faculty of representation. That claim contradicts itself. So does every solution to our misunderstood problem, whether we claim that bodies in themselves consist of an infinite number of parts or of a finite number of simple parts.

§ 53. In the first (the mathematical) class of antinomies the falsehood of the assumption consists in representing in one concept something self-contradictory as if it were compatible (i.e., an appearance as an object in itself). But, as to the second (the dynamical) class of antinomies, the falsehood of the representation consists in representing as contradictory what is compatible; so that, as in the former case, the opposed assertions are both false, in this case, on the other hand, where they are opposed to one another by mere misunderstanding, they may both be true.

§ 53. In the first, mathematical class of antinomies, the assumption is false because it treats incompatible things as compatible within one concept. It treats an appearance as an object in itself. But in the second, dynamical class of antinomies, the representation is false because it treats compatible things as contradictory. In the first class, both opposing assertions are false. In the second class, however, the assertions oppose each other only because of a misunderstanding, so both may be true.

Any mathematical connexion necessarily presupposes homogeneity of what is connected (in the concept of magnitude), while the dynamical one by no means requires the same. When we have to deal with extended magnitudes, all the parts must be homogeneous with one another and with the whole; whereas, in the connexion of cause and effect, homogeneity may indeed likewise be found, but is not necessary; for the concept of causality (by means of which something is posited through something else quite different from it), at all events, does not require it.

Any mathematical connection necessarily assumes that the things connected are of the same kind, in the concept of magnitude. A dynamical connection, however, does not require this. In extended magnitudes, all the parts must be of the same kind as one another and as the whole. In a connection between cause and effect, the two may also be of the same kind, but this is not necessary. The concept of causality does not require it. Through causality, something is posited by means of something else quite different from it.

If the objects of the world of sense are taken for things in themselves, and the above laws of nature for the laws of things in themselves, the contradiction would be unavoidable. So also, if the subject of freedom were, like other objects, represented as mere appearance, the contradiction would be just as unavoidable, for the same predicate would at once be affirmed and denied of the same kind of object in the same sense. But if natural necessity is referred merely to appearances, and freedom merely to things in themselves, no contradiction arises, if we at once assume, or admit both kinds of causality, however difficult or impossible it may be to make the latter kind conceivable.

If we take the objects of the sensory world as things in themselves, and the laws of nature described above as laws of things in themselves, a contradiction is unavoidable. Likewise, a contradiction is unavoidable if we represent the subject of freedom as a mere appearance, like other objects. For we would be both affirming and denying the same property of the same kind of object in the same sense. But suppose we attribute natural necessity only to appearances and freedom only to things in themselves. Then no contradiction arises if we assume or accept both kinds of causality, however difficult or impossible it may be to understand how the second kind is possible.

As appearance every effect is an event, or something that happens in time; it must, according to the universal law of nature, be preceded by a determination of the causality of its cause (a state), which follows according to a constant law. But this determination of the cause as causality must likewise be something that takes place or happens; the cause must have begun to act, otherwise no succession between it and the effect could be conceived. Otherwise the effect, as well as the causality of the cause, would have always existed. Therefore the determination of the cause to act must also have originated among appearances, and must consequently, as well as its effect, be an event, which must again have its cause, and so on; hence natural necessity must be the condition, on which effective causes are determined. Whereas if freedom is to be a property of certain causes of appearances, it must, as regards these, which are events, be a faculty of starting them spontaneously, that is, without the causality of the cause itself, and hence without requiring any other ground to determine its start. But then the cause, as to its causality, must not rank under time-determinations of its state, that is, it cannot be an appearance, and must be considered a thing in itself, while its effects would be only appearances.³⁶ If without contradiction we can think of the beings of understanding [Verstandeswesen] as exercising such an influence on appearances, then natural necessity will attach to all connexions of cause and effect in the sensuous world, though on the other hand, freedom can be granted to such cause, as is itself not an appearance (but the foundation of appearance). Nature therefore and freedom can without contradiction be attributed to the very same thing, but in different relations—on one side as a phenomenon, on the other as a thing in itself.

As an appearance, every effect is an event, or something that happens in time. According to the universal law of nature, every effect must be preceded by a state that determines its cause to act. This state follows according to a constant law. But this determination of the cause to act must also be something that takes place or happens. The cause must have begun to act. Otherwise, we could not conceive of the cause acting before the effect. The effect, like the causality of the cause, would instead have always existed. Therefore, the determination of the cause to act must also have originated among appearances. Like its effect, it must be an event, which must in turn have its own cause, and so on. Natural necessity must therefore be the condition under which effective causes are determined. If freedom is to be a property of certain causes of appearances, however, it must be a power to start events spontaneously. This means starting events without the causality of the cause itself requiring any other ground to determine its start. But then the causality of the cause must not depend on states determined in time. In other words, the cause cannot be an appearance and must be considered a thing in itself. Its effects, however, would be only appearances.³⁶ If we can think, without contradiction, of beings of understanding [Verstandeswesen] as exercising such an influence on appearances, then natural necessity will apply to all connections of cause and effect in the sensuous world. At the same time, freedom can be granted to a cause that is not itself an appearance but the foundation of appearance. Nature and freedom can therefore be attributed to the very same thing without contradiction, but in different respects: on one side as a phenomenon, on the other as a thing in itself.

We have in us a faculty, which not only stands in connexion with its subjective determining grounds that are the natural causes of its actions, and is so far the faculty of a being that itself belongs to appearances, but is also referred to objective grounds, that are only ideas, so far as they can determine this faculty, a connexion which is expressed by the word ought. This faculty is called reason, and, so far as we consider a being (man) entirely according to this objectively determinable reason, he cannot be considered as a being of sense, but this property is that of a thing in itself, of which we cannot comprehend the possibility—I mean how the ought (which however has never yet taken place) should determine its activity, and can become the cause of actions, whose effect is an appearance in the sensible world. Yet the causality of reason would be freedom with regard to the effects in the sensuous world, so far as we can consider objective grounds, which are themselves ideas, as their determinants. For its action in that case would not depend upon subjective conditions, consequently not upon those of time, and of course not upon the law of nature, which serves to determine them, because grounds of reason give to actions the rule universally, according to principles, without the influence of the circumstances of either time or place.

We have a faculty connected to the subjective grounds that determine it. These grounds are the natural causes of its actions. In this respect, the faculty belongs to a being that is itself an appearance. But the faculty is also connected to objective grounds, which are only ideas, insofar as they can determine it. We express this connection with the word ought. This faculty is called reason. When we consider a human being entirely in terms of reason as objectively determinable, we cannot regard that person as a being of sense. Instead, this capacity belongs to a thing in itself, whose possibility we cannot understand. I mean that we cannot understand how an ought can determine its activity and become the cause of actions whose effects appear in the world of sense. Yet the ought has never actually taken place. Nevertheless, the causality of reason would be freedom with respect to its effects in the world of sense, insofar as we can regard objective grounds, which are themselves ideas, as determining those effects. For in that case, the action of reason would not depend on subjective conditions. It would therefore not depend on conditions of time, or on the law of nature that determines them. This is because the grounds of reason give actions a universal rule according to principles, unaffected by circumstances of time or place.

What I adduce here is merely meant as an example to make the thing intelligible, and does not necessarily belong to our problem, which must be decided from mere concepts, independently of the properties which we meet in the actual world.

What I offer here is only an example to make the matter clear. It does not necessarily belong to our problem. The problem must be decided from concepts alone, independently of the properties we encounter in the actual world.

Now I may say without contradiction: that all the actions of rational beings, so far as they are appearances (occurring in any experience), are subject to the necessity of nature; but the same actions, as regards merely the rational subject and its faculty of acting according to mere reason, are free. For what is required for the necessity of nature? Nothing more than the determinability of every event in the world of sense according to constant laws, that is, a reference to cause in the appearance; in this process the thing in itself at its foundation and its causality remain unknown. But I say, that the law of nature remains, whether the rational being is the cause of the effects in the sensuous world from reason, that is, through freedom, or whether it does not determine them on grounds of reason. For, if the former is the case, the action is performed according to maxims, the effect of which as appearance is always conform able to constant laws; if the latter is the case, and the action not performed on principles of reason, it is subjected to the empirical laws of the sensibility, and in both cases the effects are connected according to constant laws; more than this we do not require or know concerning natural necessity. But in the former case reason is the cause of these laws of nature, and therefore free; in the latter the effects follow according to mere natural laws of sensibility, because reason does not influence it; but reason itself is not determined on that account by the sensibility, and is therefore free in this case too. Freedom is therefore no hindrance to natural law in appearance, neither does this law abrogate the freedom of the practical use of reason, which is connected with things in themselves, as determining grounds.

I can now say, without contradiction, that all actions of rational beings are subject to natural necessity as appearances in experience. Yet those same actions are free when considered only in relation to the rational subject and its capacity to act by reason alone. What does natural necessity require? Only that every event in the world of sense can be determined according to constant laws. This means that each event must be related to a cause in appearance. The thing in itself underlying the appearance, and its causality, remain unknown. But I say that natural law holds in either case. It holds whether a rational being causes effects in the world of sense through reason, and thus through freedom, or does not determine them on grounds of reason. In the first case, the action is performed according to maxims. Its effect, as an appearance, always conforms to constant laws. In the second case, the action is not performed on principles of reason. It is subject to the empirical laws of sensibility. In both cases, the effects are connected according to constant laws. We require and know nothing more than this about natural necessity. In the first case, reason is the cause of these laws of nature and is therefore free. In the second case, the effects follow from the natural laws of sensibility alone, because reason does not influence them. But reason itself is not thereby determined by sensibility. It is therefore free in this case too. Freedom is therefore no obstacle to natural law in appearance. Nor does natural law cancel the freedom of reason in its practical use. That use of reason is connected with things in themselves as grounds of determination.

Thus practical freedom, viz., the freedom in which reason possesses causality according to objectively determining grounds, is rescued and yet natural necessity is not in the least curtailed with regard to the very same effects, as appearances. The same remarks will serve to explain what we had to say concerning transcendental freedom and its compatibility with natural necessity (in the same subject, but not taken in the same reference). For, as to this, every beginning of the action of a being from objective causes regarded as determining grounds, is always a first start, though the same action is in the series of appearances only a subordinate start, which must be preceded by a state of the cause, which determines it, and is itself determined in the same manner by another immediately preceding. Thus we are able, in rational beings, or in beings generally, so far as their causality is determined in them as things in themselves, to imagine a faculty of beginning from itself a series of states, without falling into contradiction with the laws of nature. For the relation of the action to objective grounds of reason is not a time-relation; in this case that which determines the causality does not precede in time the action, because such determining grounds represent not a reference to objects of sense, e.g., to causes in the appearances, but to determining causes, as things in themselves, which do not rank under conditions of time. And in this way the action, with regard to the causality of reason, can be considered as a first start in respect to the series of appearances, and yet also as a merely subordinate beginning. We may therefore without contradiction consider it in the former aspect as free, but in the latter (in so far as it is merely appearance) as subject to natural necessity.

Practical freedom is therefore preserved. This is the freedom in which reason has causal power through objectively determining grounds. Yet natural necessity is not diminished in the least with respect to the same effects, considered as appearances. The same remarks explain what we said about transcendental freedom and its compatibility with natural necessity. Both apply to the same subject, but not in the same respect. Every action a being begins from objective causes as determining grounds is always a first beginning. Yet the same action, in the series of appearances, is only a subordinate beginning. It must be preceded by a state of its cause that determines it. That state is itself determined in the same way by another state immediately preceding it. Thus, we can imagine a capacity to begin a series of states on its own, without contradicting the laws of nature. We can imagine this capacity in rational beings, or in beings generally, insofar as their causality is determined in them as things in themselves. For an action’s relation to objective grounds of reason is not a relation in time. In this case, what determines the causality does not precede the action in time. These determining grounds do not refer to objects of sense, such as causes among appearances. They refer instead to determining causes as things in themselves, which are not subject to the conditions of time. In this way, an action can be considered a first beginning, with respect to the causality of reason. Yet it can also be considered a merely subordinate beginning in the series of appearances. We may therefore consider the action free in the first respect without contradiction. In the second respect, insofar as it is merely an appearance, we may consider it subject to natural necessity.

As to the fourth Antinomy, it is solved in the same way as the conflict of reason with itself in the third. For, provided the cause in the appearance is distinguished from the cause of the appearance (so far as it can be thought as a thing in itself), both propositions are perfectly reconcilable: the one, that there is nowhere in the sensuous world a cause (according to similar laws of causality), whose existence is absolutely necessary; the other, that this world is nevertheless connected with a Necessary Being as its cause (but of another kind and according to another law). The incompatibility of these propositions entirely rests upon the mistake of extending what is valid merely of appearances to things in themselves, and in general confusing both in one concept.

The fourth Antinomy is resolved in the same way as reason's conflict with itself in the third. We must distinguish a cause within an appearance from the cause of the appearance itself, insofar as the latter can be thought of as a thing in itself. If we do this, both propositions can be fully reconciled. The first is that nowhere in the sensory world is there a cause whose existence is absolutely necessary, according to the same laws of causality. The second is that this world is nevertheless connected to a Necessary Being as its cause. This cause, however, is of another kind and acts according to another law. The apparent conflict between these propositions rests entirely on a mistake. We extend what is valid only for appearances to things in themselves, and generally confuse both under a single concept.

§ 54. This then is the proposition and this the solution of the whole antinomy, in which reason finds itself involved in the application of its principles to the sensible world. The former alone (the mere proposition) would be a considerable service in the cause of our knowledge of human reason, even though the solution might fail to fully satisfy the reader, who has here to combat a natural illusion, which has been but recently exposed to him, and which he had hitherto always regarded as genuine. For one result at least is unavoidable. As it is quite impossible to prevent this conflict of reason with itself—so long as the objects of the sensible world are taken for things in themselves, and not for mere appearances, which they are in fact—the reader is thereby compelled to examine over again the deduction of all our a priori cognition and the proof which I have given of my deduction in order to come to a decision on the question. This is all I require at present; for when in this occupation he shall have thought himself deep enough into the nature of pure reason, those concepts by which alone the solution of the conflict of reason is possible, will become sufficiently familiar to him. Without this preparation I cannot expect an unreserved assent even from the most attentive reader.

§ 54. This, then, is the statement and the solution of the whole antinomy in which reason becomes entangled when applying its principles to the sensible world. The statement alone would be a considerable contribution to our knowledge of human reason, even if the solution does not fully satisfy the reader. For the reader must struggle against a natural illusion that has only recently been exposed and that he had always taken to be genuine. For at least one result is unavoidable. Reason cannot be prevented from conflicting with itself as long as objects in the sensible world are taken for things in themselves, rather than for mere appearances, as they actually are. The reader is therefore compelled to reconsider the deduction of all our a priori cognition and the proof I have given for it, so that he can decide the question. That is all I require for now. For, in doing so, the reader will think deeply enough about the nature of pure reason to become sufficiently familiar with the concepts that alone make a solution to reason’s conflict possible. Without this preparation, I cannot expect even the most attentive reader to agree without reservation.

III. The Theological Idea.³⁷

III. The Theological Idea.³⁷

§ 55. The third transcendental Idea, which affords matter for the most important, but, if pursued only speculatively, transcendent and thereby dialectical use of reason, is the ideal of pure reason. Reason in this case does not, as with the psychological and the cosmological Ideas, begin from experience, and err by exaggerating its grounds, in striving to attain, if possible, the absolute completeness of their series. It rather totally breaks with experience, and from mere concepts of what constitutes the absolute completeness of a thing in general, consequently by means of the idea of a most perfect primal Being, it proceeds to determine the possibility and therefore the actuality of all other things. And so the mere presupposition of a Being, who is conceived not in the series of experience, yet for the purposes of experience—for the sake of comprehending its connexion, order, and unity—i.e., the idea [the notion of it], is more easily distinguished from the concept of the understanding here, than in the former cases. Hence we can easily expose the dialectical illusion which arises from our making the subjective conditions of our thinking objective conditions of objects themselves, and an hypothesis necessary for the satisfaction of our reason, a dogma. As the observations of the Critique on the pretensions of transcendental theology are intelligible, clear, and decisive, I have nothing more to add on the subject.

§ 55. The third transcendental Idea is the ideal of pure reason. It provides material for reason’s most important use. But if pursued only speculatively, this use is transcendent and therefore dialectical. Here, reason does not begin with experience, as it does with the psychological and cosmological Ideas. In those cases, reason goes astray by exaggerating its grounds as it strives, if possible, to reach the absolute completeness of their series. Instead, reason breaks entirely with experience. It begins with mere concepts of what constitutes the absolute completeness of a thing in general. From these, it forms the idea of a most perfect primal Being. By means of this idea, it proceeds to determine the possibility, and therefore the actuality, of all other things. This Being is not conceived as part of the series of experience. It is nevertheless presupposed for the purposes of experience, so that we can understand experience’s connection, order, and unity. So here the mere presupposition of such a Being, that is, the idea [the notion of it], is more easily distinguished from a concept of the understanding than in the previous cases. We can therefore easily expose the dialectical illusion. It arises when we take the subjective conditions of our thinking to be objective conditions of objects themselves. We also turn a hypothesis necessary to satisfy our reason into a dogma. The observations of the Critique on the claims of transcendental theology are intelligible, clear, and decisive. I therefore have nothing more to add on the subject.

General Remark on the Transcendental Ideas.

General Note on Transcendental Ideas.

§ 56. The objects, which are given us by experience, are in many respects incomprehensible, and many questions, to which the law of nature leads us, when carried beyond a certain point (though quite conformably to the laws of nature), admit of no answer; as for example the question: why substances attract one another? But if we entirely quit nature, or in pursuing its combinations, exceed all possible experience, and so enter the realm of mere ideas, we cannot then say that the object is incomprehensible, and that the nature of things proposes to us insoluble problems. For we are not then concerned with nature or in general with given objects, but with concepts, which have their origin merely in our reason, and with mere creations of thought; and all the problems that arise from our notions of them must be solved, because of course reason can and must give a full account of its own procedure.³⁸ As the psychological, cosmological, and theological Ideas are nothing but pure concepts of reason, which cannot be given in any experience, the questions which reason asks us about them are put to us not by the objects, but by mere maxims of our reason for the sake of its own satisfaction. They must all be capable of satisfactory answers, which is done by showing that they are principles which bring our use of the understanding into thorough agreement, completeness, and synthetical unity, and that they so far hold good of experience only, but of experience as a whole.

§ 56. The objects given to us in experience are incomprehensible in many respects. When we follow natural laws beyond a certain point, even while remaining entirely consistent with them, they lead us to questions that cannot be answered. For example, why do substances attract one another? But suppose we leave nature entirely, or trace its connections beyond all possible experience. We then enter the realm of mere ideas. There we cannot say that the object is incomprehensible or that the nature of things presents us with unsolvable problems. For we are no longer concerned with nature or with given objects of any kind. Instead, we are concerned with concepts that originate solely in our reason and with mere creations of thought. All the problems arising from our notions of these concepts must be solved. For reason can and must give a full account of its own procedure.³⁸ The psychological, cosmological, and theological Ideas are nothing but pure concepts of reason. They cannot be given in any experience. So the questions reason asks us about them are not posed by objects. They come from mere maxims of our reason, for its own satisfaction. All these questions must be capable of satisfactory answers. We can answer them by showing that these maxims are principles that bring our use of the understanding into thorough agreement, completeness, and synthetic unity. We must also show that these principles apply only to experience, but to experience as a whole.

Although an absolute whole of experience is impossible, the idea of a whole of cognition according to principles must impart to our knowledge a peculiar kind of unity, that of a system, without which it is nothing but piecework, and cannot be used for proving the existence of a highest purpose (which can only be the general system of all purposes), I do not here refer only to the practical, but also to the highest purpose of the speculative use of reason.

Although a complete whole of experience is impossible, the idea of a whole of knowledge organized by principles must give our knowledge a special unity. This is the unity of a system. Without it, our knowledge is merely piecemeal and cannot be used to prove the existence of a highest purpose. Such a purpose can only be the general system of all purposes. By a highest purpose, I mean not only the practical purpose of reason, but also the highest purpose of its speculative use.

The transcendental Ideas therefore express the peculiar application of reason as a principle of systematic unity in the use of the understanding. Yet if we assume this unity of the mode of cognition to be attached to the object of cognition, if we regard that which is merely regulative to be constitutive, and if we persuade ourselves that we can by means of these Ideas enlarge our cognition transcendently, or far beyond all possible experience, while it only serves to render experience within itself as nearly complete as possible, i.e., to limit its progress by nothing that cannot belong to experience: we suffer from a mere misunderstanding in our estimate of the proper application of our reason and of its principles, and from a Dialectic, which both confuses the empirical use of reason, and also sets reason at variance with itself.

The transcendental Ideas therefore express the distinctive use of reason as a principle for giving systematic unity to the understanding. But suppose we assume that this unity in our way of knowing belongs to the object we know. Suppose we treat what merely guides our inquiry as if it established facts about the object. And suppose we convince ourselves that these Ideas can extend our knowledge beyond all possible experience. In fact, their only purpose is to make experience itself as complete as possible. That is, they let nothing limit the progress of experience except what can belong to experience. If we make these assumptions, we merely misunderstand the proper use of reason and its principles. We also fall into a Dialectic that confuses the use of reason in experience and sets reason against itself.

Conclusion.

Conclusion.

On the Determination of the Bounds of Pure Reason.

On Defining the Limits of Pure Reason.

§ 57. Having adduced the clearest arguments, it would be absurd for us to hope that we can know more of any object, than belongs to the possible experience of it, or lay claim to the least atom of knowledge about anything not assumed to be an object of possible experience, which would determine it according to the constitution it has in itself. For how could we determine anything in this way, since time, space, and the categories, and still more all the concepts formed by empirical experience or perception in the sensible world (Anschauung), have and can have no other use, than to make experience possible. And if this condition is omitted from the pure concepts of the understanding, they do not determine any object, and have no meaning whatever.

§ 57. We have given the clearest arguments. It would therefore be absurd to hope that we can know more about an object than its possible experience allows. It would be equally absurd to claim even the smallest bit of knowledge about something not assumed to be an object of possible experience. Such knowledge would determine what that thing is like in itself. How could we determine anything in this way? Time, space, and the categories have no use other than to make experience possible. The same is even more true of all concepts formed through empirical experience or perception in the sensible world (Anschauung). They can have no other use. If this condition is removed from the pure concepts of the understanding, they determine no object and have no meaning at all.

But it would be on the other hand a still greater absurdity if we conceded no things in themselves, or set up our experience for the only possible mode of knowing things, our way of beholding (Anschauung) them in space and in time for the only possible way, and our discursive understanding for the archetype of every possible understanding; in fact if we wished to have the principles of the possibility of experience considered universal conditions of things in themselves.

But it would be even more absurd to deny that things in themselves exist. It would also be more absurd to claim that our experience is the only possible way to know things. Or that our way of perceiving things in space and time (Anschauung) is the only possible way. Or that our discursive understanding is the model for every possible understanding. In short, it would be more absurd to treat the principles that make experience possible as universal conditions of things in themselves.

Our principles, which limit the use of reason to possible experience, might in this way become transcendent, and the limits of our reason be set up as limits of the possibility of things in themselves (as Hume's dialogues may illustrate), if a careful critique did not guard the bounds of our reason with respect to its empirical use, and set a limit to its pretensions. Scepticism originally arose from metaphysics and its licentious dialectics. At first it might, merely to favor the empirical use of reason, announce everything that transcends this use as worthless and deceitful; but by and by, when it was perceived that the very same principles that are used in experience, insensibly, and apparently with the same right, led still further than experience extends, then men began to doubt even the propositions of experience. But here there is no danger; for common sense will doubtless always assert its rights. A certain confusion, however, arose in science which cannot determine how far reason is to be trusted, and why only so far and no further, and this confusion can only be cleared up and all future relapses obviated by a formal determination, on principle, of the boundary of the use of our reason.

Our principles limit the use of reason to possible experience. But they might themselves go beyond experience if a careful critique did not guard the bounds of reason’s empirical use and limit its claims. We might then treat the limits of our reason as limits on what things in themselves can be. Hume’s dialogues may illustrate this danger. Scepticism originally arose from metaphysics and its unrestrained dialectics. At first, scepticism might have declared everything beyond experience worthless and deceptive, merely to favor the empirical use of reason. But people gradually noticed something else. The very principles used in experience led them, almost without their noticing, beyond the reach of experience, apparently with equal justification. Then people began to doubt even the propositions of experience. But there is no danger here, because common sense will doubtless always assert its rights. Science, however, fell into some confusion. It cannot determine how far reason can be trusted, or why reason can be trusted only that far and no further. This confusion can be cleared up, and all future relapses prevented, only by formally establishing the boundary of reason’s use on the basis of principle.

We cannot indeed, beyond all possible experience, form a definite notion of what things in themselves may be. Yet we are not at liberty to abstain entirely from inquiring into them; for experience never satisfies reason fully, but in answering questions, refers us further and further back, and leaves us dissatisfied with regard to their complete solution. This any one may gather from the Dialectics of pure reason, which therefore has its good subjective grounds. Having acquired, as regards the nature of our soul, a clear conception of the subject, and having come to the conviction, that its manifestations cannot be explained materialistically, who can refrain from asking what the soul really is, and, if no concept of experience suffices for the purpose, from accounting for it by a concept of reason (that of a simple immaterial being), though we cannot by any means prove its objective reality? Who can satisfy himself with mere empirical knowledge in all the cosmological questions of the duration and of the quantity of the world, of freedom or of natural necessity, since every answer given on principles of experience begets a fresh question, which likewise requires its answer and thereby clearly shows the insufficiency of all physical modes of explanation to satisfy reason? Finally, who does not see in the thorough-going contingency and dependence of all his thoughts and assumptions on mere principles of experience, the impossibility of stopping there? And who does not feel himself compelled, notwithstanding all interdictions against losing himself in transcendent ideas, to seek rest and contentment beyond all the concepts which he can vindicate by experience, in the concept of a Being, the possibility of which we cannot conceive, but at the same time cannot be refuted, because it relates to a mere being of the understanding, and without it reason must needs remain forever dissatisfied?

We cannot form a definite idea of what things in themselves may be beyond all possible experience. Yet we cannot simply refrain from asking about them. Experience never fully satisfies reason. As it answers questions, it leads us further and further back and leaves us dissatisfied with the lack of a complete answer. Anyone can see this in the Dialectics of pure reason, which therefore has sound subjective grounds. Once we have a clear idea of the soul’s nature, we may become convinced that its manifestations cannot be explained materialistically. Who, then, can refrain from asking what the soul really is? And if no concept drawn from experience is adequate, who can refrain from explaining it through a concept of reason? This is the concept of a simple immaterial being, though we cannot prove in any way that it has objective reality. Who can be satisfied with knowledge from experience alone when facing the cosmological questions? These concern the duration and size of the world, and freedom or natural necessity. Every answer based on experience gives rise to another question, which also demands an answer. This clearly shows that all physical explanations fail to satisfy reason. Finally, who does not see that we cannot stop with principles of experience alone? All our thoughts and assumptions based on such principles are thoroughly contingent and dependent. And who does not feel compelled to seek rest and contentment beyond all the concepts that experience can support? We feel compelled to do this despite every warning against losing ourselves in transcendent ideas. We seek rest in the concept of a Being whose possibility we cannot conceive, but which also cannot be refuted. This is because the concept refers to a mere being of the understanding. Without this concept, reason must remain forever dissatisfied.

Bounds (in extended beings) always presuppose a space existing outside a certain definite place, and in closing it; limits do not require this, but are mere negations, which affect a quantity, so far as it is not absolutely complete. But our reason, as it were, sees in its surroundings a space for the cognition of things in themselves, though we can never have definite notions of them, and are limited to appearances only.

Bounds (in extended beings) always assume a space outside and surrounding a particular region. Limits do not require such a space. They are merely negations that apply to a quantity insofar as it is not absolutely complete. But our reason, as it were, sees room around it for knowledge of things in themselves. We can never have definite concepts of things in themselves, however, and are limited to appearances alone.

As long as the cognition of reason is homogeneous, definite bounds to it are inconceivable. In mathematics and in natural philosophy human reason admits of limits, but not of bounds, viz., that something indeed lies without it, at which it can never arrive, but not that it will at any point find completion in its internal progress. The enlarging of our views in mathematics, and the possibility of new discoveries, are infinite; and the same is the case with the discovery of new properties of nature, of new powers and laws, by continued experience and its rational combination. But limits cannot be mistaken here, for mathematics refers to appearances only, and what cannot be an object of sensuous contemplation, such as the concepts of metaphysics and of morals, lies entirely without its sphere, and it can never lead to them; neither does it require them. It is therefore not a continual progress and an approximation towards these sciences, and there is not, as it were, any point or line of contact. Natural science will never reveal to us the internal constitution of things, which though not appearance, yet can serve as the ultimate ground of explaining appearance. Nor does that science require this for its physical explanations. Nay even if such grounds should be offered from other sources (for instance, the influence of immaterial beings), they must be rejected and not used in the progress of its explanations. For these explanations must only be grounded upon that which as an object of sense can belong to experience, and be brought into connexion with our actual perceptions and empirical laws.

As long as the knowledge gained by reason is of the same kind, we cannot conceive of definite bounds to it. In mathematics and natural science, human reason has limits, but not bounds. Something lies outside these fields that reason can never reach. But reason’s progress within them will never reach a point of completion. We can expand our understanding of mathematics and make new discoveries without end. The same is true of discovering new properties, powers, and laws of nature through further experience and rational connections between experiences. But we cannot mistake the limits here. Mathematics deals only with appearances. Anything that cannot be an object of sensory perception lies entirely outside its sphere. The concepts of metaphysics and morals are examples. Mathematics can never lead to these concepts, nor does it need them. Mathematics therefore does not continually progress toward these sciences or come closer to them. There is, so to speak, no point or line where they meet. Natural science will never reveal the inner constitution of things. This inner constitution is not an appearance, though it can serve as the ultimate basis for explaining appearances. Nor does natural science need it for physical explanations. Even if such explanations were offered from other sources, such as the influence of immaterial beings, natural science must reject them. It must not use them in developing its explanations. For these explanations must be based only on what can belong to experience as an object of the senses. They must be based on what can be connected with our actual perceptions and empirical laws.

But metaphysics leads us towards bounds in the dialectical attempts of pure reason (not undertaken arbitrarily or wantonly, but stimulated thereto by the nature of reason itself). And the transcendental Ideas, as they do not admit of evasion, and are never capable of realisation, serve to point out to us actually not only the bounds of the pure use of reason, but also the way to determine them. Such is the end and the use of this natural predisposition of our reason, which has brought forth metaphysics as its favorite child, whose generation, like every other in the world, is not to be ascribed to blind chance, but to an original germ, wisely organised for great ends. For metaphysics, in its fundamental features, perhaps more than any other science, is placed in us by nature itself, and cannot be considered the production of an arbitrary choice or a casual enlargement in the progress of experience from which it is quite disparate.

But metaphysics leads us toward limits through the dialectical attempts of pure reason. These attempts are not arbitrary or reckless, but arise from the nature of reason itself. We cannot avoid the transcendental Ideas, and we can never realize them. Yet they show us not only the limits of the pure use of reason, but also how to determine them. This is the purpose and use of this natural tendency of our reason. It has given birth to metaphysics as its favorite child. Like every other birth in the world, this one is not the result of blind chance. It comes from an original seed, wisely organized for great purposes. For nature itself has placed the fundamental features of metaphysics in us, perhaps more than those of any other science. Metaphysics is therefore not the product of an arbitrary choice or an accidental expansion of experience. It is quite different from experience.

Reason with all its concepts and laws of the understanding, which suffice for empirical use, i.e., within the sensible world, finds in itself no satisfaction because ever-recurring questions deprive us of all hope of their complete solution. The transcendental ideas, which have that completion in view, are such problems of reason. But it sees clearly, that the sensuous world cannot contain this completion, neither consequently can all the concepts, which serve merely for understanding the world of sense, such as space and time, and whatever we have adduced under the name of pure concepts of the understanding. The sensuous world is nothing but a chain of appearances connected according to universal laws; it has therefore no subsistence by itself; it is not the thing in itself, and consequently must point to that which contains the basis of this experience, to beings which cannot be cognised merely as phenomena, but as things in themselves. In the cognition of them alone reason can hope to satisfy its desire of completeness in proceeding from the conditioned to its conditions.

Reason finds no satisfaction in its concepts and laws of the understanding, though they suffice for empirical use, within the sensible world. For questions keep returning, depriving us of all hope of solving them completely. The transcendental ideas are problems of reason that aim at such completion. But reason sees clearly that the sensible world cannot contain this completion. Nor can the concepts that serve only to understand that world. These include space and time, and everything we have called pure concepts of the understanding. The sensible world is nothing but a chain of appearances connected by universal laws. It therefore has no existence by itself. It is not the thing in itself. It must consequently point to what contains the basis of this experience: beings that cannot be known merely as phenomena, but as things in themselves. Only by knowing these beings can reason hope to satisfy its desire for completeness as it moves from the conditioned to its conditions.

We have above (§§ 33, 34) indicated the limits of reason with regard to all cognition of mere creations of thought. Now, since the transcendental ideas have urged us to approach them, and thus have led us, as it were, to the spot where the occupied space (viz., experience) touches the void (that of which we can know nothing, viz., noumena), we can determine the bounds of pure reason. For in all bounds there is something positive (e.g., a surface is the boundary of corporeal space, and is therefore itself a space, a line is a space, which is the boundary of the surface, a point the boundary of the line, but yet always a place in space), whereas limits contain mere negations. The limits pointed out in those paragraphs are not enough after we have discovered that beyond them there still lies something (though we can never cognise what it is in itself). For the question now is, What is the attitude of our reason in this connexion of what we know with what we do not, and never shall, know? This is an actual connexion of a known thing with one quite unknown (and which will always remain so), and though what is unknown should not become the least more known—which we cannot even hope—yet the notion of this connexion must be definite, and capable of being rendered distinct.

Above (§§ 33, 34), we identified the limits of reason in knowing things that are mere creations of thought. The transcendental ideas have now urged us to approach these limits. They have led us, so to speak, to the place where occupied space, namely experience, touches the void. This void consists of noumena, things about which we can know nothing. We can therefore determine the bounds of pure reason. Every bound has something positive in it. For example, a surface is the boundary of a physical space and is therefore itself a space. A line is a space that forms the boundary of a surface. A point forms the boundary of a line but is still always a place in space. Limits, by contrast, contain only negations. The limits identified in those paragraphs are not enough. For we have discovered that something still lies beyond them, though we can never know what it is in itself. The question now is: How does our reason relate what we know to what we do not and never will know? This is an actual connection between something known and something entirely unknown, which will always remain unknown. The unknown cannot become even slightly better known, and we cannot even hope that it will. Yet, though the unknown remains unknown, our conception of this connection must be definite and capable of being made distinct.

We must therefore accept an immaterial being, a world of understanding, and a Supreme Being (all mere noumena), because in them only, as things in themselves, reason finds that completion and satisfaction, which it can never hope for in the derivation of appearances from their homogeneous grounds, and because these actually have reference to something distinct from them (and totally heterogeneous), as appearances always presuppose an object in itself, and therefore suggest its existence whether we can know more of it or not.

We must therefore accept an immaterial being, a world of understanding, and a Supreme Being. All of these are mere noumena. Only in them, as things in themselves, does reason find completion and satisfaction. Reason can never hope to find these by tracing appearances back to their similar grounds. We must also accept these noumena because appearances actually refer to something distinct from them and completely different in kind. Appearances always presuppose an object in itself. They therefore suggest that such an object exists, whether or not we can know anything more about it.

But as we can never cognise these beings of understanding as they are in themselves, that is, definitely, yet must assume them as regards the sensible world, and connect them with it by reason, we are at least able to think this connexion by means of such concepts as express their relation to the world of sense. Yet if we represent to ourselves a being of the understanding by nothing but pure concepts of the understanding, we then indeed represent nothing definite to ourselves, consequently our concept has no significance; but if we think it by properties borrowed from the sensuous world, it is no longer a being of understanding, but is conceived as an appearance, and belongs to the sensible world. Let us take an instance from the notion of the Supreme Being.

But we can never know these beings of the understanding as they are in themselves, that is, know them definitely. Yet we must assume that they exist in relation to the sensible world and connect them with it through reason. We can at least think of this connection through concepts that express their relation to the world of sense. But if we represent a being of the understanding using only pure concepts of the understanding, we represent nothing definite. Our concept therefore has no meaning. On the other hand, if we think of such a being through properties borrowed from the world of sense, it is no longer a being of the understanding. Instead, we conceive of it as an appearance, and it belongs to the sensible world. Let us take an example from the idea of the Supreme Being.

Our deistic conception is quite a pure concept of reason, but represents only a thing containing all realities, without being able to determine any one of them; because for that purpose an example must be taken from the world of sense, in which case we should have an object of sense only, not something quite heterogeneous, which can never be an object of sense. Suppose I attribute to the Supreme Being understanding, for instance; I have no concept of an understanding other than my own, one that must receive its perceptions (Anschauung) by the senses, and which is occupied in bringing them under rules of the unity of consciousness. Then the elements of my concept would always lie in the appearance; I should however by the insufficiency of the appearance be necessitated to go beyond them to the concept of a being which neither depends upon appearance, nor is bound up with them as conditions of its determination. But if I separate understanding from sensibility to obtain a pure understanding, then nothing remains but the mere form of thinking without perception (Anschauung), by which form alone I can cognise nothing definite, and consequently no object. For that purpose I should conceive another understanding, such as would directly perceive its objects,³⁹ but of which I have not the least notion; because the human understanding is discursive, and can [not directly perceive, it can] only cognise by means of general concepts. And the very same difficulties arise if we attribute a will to the Supreme Being; for we have this concept only by drawing it from our internal experience, and therefore from our dependence for satisfaction upon objects whose existence we require; and so the notion rests upon sensibility, which is absolutely incompatible with the pure concept of the Supreme Being.

Our deistic conception is a completely pure concept of reason. But it represents only a thing containing all realities, without being able to determine any one of them. To determine one, we would have to take an example from the world of sense. Then we would have only an object of sense, not something entirely different that can never be an object of sense. Suppose, for example, that I attribute understanding to the Supreme Being. I have no concept of an understanding other than my own. My understanding must receive its perceptions (Anschauung) through the senses. It works to bring those perceptions under rules of the unity of consciousness. The elements of my concept would therefore always lie in appearances. Yet appearances are inadequate, so I would have to go beyond them. I would have to form the concept of a being that does not depend on appearances. Nor would appearances be conditions that determine this being. But suppose I separate understanding from sensibility to obtain a pure understanding. Then nothing remains but the mere form of thinking without perception (Anschauung). Through this form alone, I can know nothing definite and therefore no object. To know an object, I would have to conceive of another understanding, one that directly perceives its objects.³⁹ But I have not the slightest notion of such an understanding. Human understanding is discursive. It cannot perceive directly, but can know only through general concepts. The same difficulties arise if we attribute a will to the Supreme Being. We have this concept only because we draw it from our inner experience. We therefore draw it from our dependence for satisfaction on objects whose existence we need. So the notion of will rests on sensibility, which is absolutely incompatible with the pure concept of the Supreme Being.

Hume's objections to deism are weak, and affect only the proofs, and not the deistic assertion itself. But as regards theism, which depends on a stricter determination of the concept of the Supreme Being which in deism is merely transcendent, they are very strong, and as this concept is formed, in certain (in fact in all common) cases irrefutable. Hume always insists, that by the mere concept of an original being, to which we apply only ontological predicates (eternity, omnipresence, omnipotence), we think nothing definite, and that properties which can yield a concept in concreto must be superadded; that it is not enough to say, it is Cause, but we must explain the nature of its causality, for example, that of an understanding and of a will. He then begins his attacks on the essential point itself, i.e., theism, as he had previously directed his battery only against the proofs of deism, an attack which is not very dangerous to it in its consequences. All his dangerous arguments refer to anthropomorphism, which he holds to be inseparable from theism, and to make it absurd in itself; but if the former be abandoned, the latter must vanish with it, and nothing remain but deism, of which nothing can come, which is of no value, and which cannot serve as any foundation to religion or morals. If this anthropomorphism were really unavoidable, no proofs whatever of the existence of a Supreme Being, even were they all granted, could determine for us the concept of this Being without involving us in contradictions.

Hume’s objections to deism are weak. They challenge only the proofs of deism, not its claim itself. But his objections to theism are very strong. Theism depends on a more specific conception of the Supreme Being, who in deism is merely transcendent. As that concept is formed in certain cases (in fact, in all common cases), Hume’s objections are irrefutable. Hume always insists that the mere concept of an original being gives us no definite thought. This is so when we attribute only ontological properties to that being, such as eternity, omnipresence, and omnipotence. We must add properties that give us a concrete concept. It is not enough to say that the original being is a cause. We must explain the nature of its causality, for example, as that of an understanding and a will. Hume then turns his attacks on the essential point itself, theism. Previously, he had directed his attacks only against the proofs of deism. That attack is not very dangerous to deism in its consequences. All his dangerous arguments concern anthropomorphism. He holds that anthropomorphism is inseparable from theism and makes theism absurd in itself. But if anthropomorphism is abandoned, theism must vanish with it. Nothing would remain but deism. Nothing can come of deism. It has no value and cannot serve as a foundation for religion or morals. If anthropomorphism were really unavoidable, no proofs of a Supreme Being’s existence could give us a definite concept of that being without involving us in contradictions. This would be so even if all those proofs were granted.

If we connect with the command to avoid all transcendent judgments of pure reason, the command (which apparently conflicts with it) to proceed to concepts that lie beyond the field of its immanent (empirical) use, we discover that both can subsist together, but only at the boundary of all lawful use of reason. For this boundary belongs as well to the field of experience, as to that of the creations of thought, and we are thereby taught, as well, how these so remarkable ideas serve merely for marking the bounds of human reason. On the one hand they give warning not boundlessly to extend cognition of experience, as if nothing but world⁴⁰ remained for us to cognise, and yet, on the other hand, not to transgress the bounds of experience, and to think of judging about things beyond them, as things in themselves.

Suppose we combine the command to avoid all transcendent judgments of pure reason with another command. This other command tells us to proceed to concepts beyond reason’s immanent (empirical) use, and it seems to conflict with the first. We discover that both commands can stand together, but only at the boundary of all lawful use of reason. This boundary belongs both to the field of experience and to the field of creations of thought. In this way, we also learn how these remarkable ideas serve only to mark the limits of human reason. On the one hand, they warn us not to extend knowledge gained from experience without limit, as if nothing but the world⁴⁰ remained for us to know. On the other hand, they warn us not to cross the bounds of experience and attempt to judge things beyond them as things in themselves.

But we stop at this boundary if we limit our judgment merely to the relation which the world may have to a Being whose very concept lies beyond all the knowledge which we can attain within the world. For we then do not attribute to the Supreme Being any of the properties in themselves, by which we represent objects of experience, and thereby avoid dogmatic anthropomorphism; but we attribute them to his relation to the world, and allow ourselves a symbolical anthropomorphism, which in fact concerns language only, and not the object itself.

But we stop at this boundary if we limit our judgment to the relation the world may have to a Being. The very concept of this Being lies beyond all the knowledge we can gain within the world. For then we do not attribute to the Supreme Being any of the properties we use to represent objects of experience. In this way, we avoid dogmatic anthropomorphism. Instead, we attribute those properties to his relation to the world. We allow ourselves a symbolic anthropomorphism, which concerns only language, not the object itself.

If I say that we are compelled to consider the world, as if it were the work of a Supreme Understanding and Will, I really say nothing more, than that a watch, a ship, a regiment, bears the same relation to the watchmaker, the shipbuilder, the commanding officer, as the world of sense (or whatever constitutes the substratum of this complex of appearances) does to the Unknown, which I do not hereby cognise as it is in itself, but as it is for me or in relation to the world, of which I am a part.

If I say that we must regard the world as if it were the work of a Supreme Understanding and Will, I mean only this. The world of sense stands in the same relation to the Unknown as a watch to its maker, a ship to its builder, or a regiment to its commanding officer. The same is true of whatever underlies the whole collection of appearances. In saying this, I do not know the Unknown as it is in itself, but as it is for me or in relation to the world, of which I am a part.

§ 58. Such a cognition is one of analogy, and does not signify (as is commonly understood) an imperfect similarity of two things, but a perfect similarity of relations between two quite dissimilar things.⁴¹ By means of this analogy, however, there remains a concept of the Supreme Being sufficiently determined for us, though we have left out everything that could deter mine it absolutely or in itself; for we determine it as regards the world and as regards ourselves, and more do we not require. The attacks which Hume makes upon those who would determine this concept absolutely, by taking the materials for so doing from themselves and the world, do not affect us; and he cannot object to us, that we have nothing left if we give up the objective anthropomorphism of the concept of the Supreme Being.

§ 58. Such knowledge is based on analogy. Analogy does not mean an imperfect similarity between two things, as people commonly understand it. It means a perfect similarity between the relations of two quite dissimilar things.⁴¹ This analogy still gives us a sufficiently definite concept of the Supreme Being. It does so even though we have left out everything that could define the Supreme Being absolutely or in itself. For we define the Supreme Being in relation to the world and to ourselves, and we need no more. Hume attacks those who would define this concept absolutely, drawing on themselves and the world to do so. His attacks do not affect us. Nor can he object that we have nothing left if we give up the objective anthropomorphism of the concept of the Supreme Being.

For let us assume at the outset (as Hume in his dialogues makes Philo grant Cleanthes), as a necessary hypothesis, the deistical concept of the First Being, in which this Being is thought by the mere ontological predicates of substance, of cause, etc. This must be done, because reason, actuated in the sensible world by mere conditions, which are themselves always conditional, cannot otherwise have any satisfaction, and it therefore can be done without falling into anthropomorphism (which transfers predicates from the world of sense to a Being quite distinct from the world), because those predicates are mere categories, which, though they do not give a determinate concept of God, yet give a concept not limited to any conditions of sensibility. Thus nothing can prevent our predicating of this Being a causality through reason with regard to the world, and thus passing to theism, without being obliged to attribute to God in himself this kind of reason, as a property inhering in him. For as to the former, the only possible way of prosecuting the use of reason (as regards all possible experience, in complete harmony with itself) in the world of sense to the highest point, is to assume a supreme reason as a cause of all the connexions in the world. Such a principle must be quite advantageous to reason and can hurt it nowhere in its application to nature. As to the latter, reason is thereby not transferred as a property to the First Being in himself, but only to his relation to the world of sense, and so anthropomorphism is entirely avoided. For nothing is considered here but the cause of the form of reason which is perceived everywhere in the world, and reason is attributed to the Supreme Being, so far as it contains the ground of this form of reason in the world, but according to analogy only, that is, so far as this expression shows merely the relation, which the Supreme Cause unknown to us has to the world, in order to determine everything in it conformably to reason in the highest degree. We are thereby kept from using reason as an attribute for the purpose of conceiving God, but instead of conceiving the world in such a manner as is necessary to have the greatest possible use of reason according to principle. We thereby acknowledge that the Supreme Being is quite inscrutable and even unthinkable in any definite way as to what he is in himself. We are thereby kept, on the one hand, from making a transcendent use of the concepts which we have of reason as an efficient cause (by means of the will), in order to determine the Divine Nature by properties, which are only borrowed from human nature, and from losing ourselves in gross and extravagant notions, and on the other hand from deluging the contemplation of the world with hyperphysical modes of explanation according to our notions of human reason, which we transfer to God, and so losing for this contemplation its proper application, according to which it should be a rational study of mere nature, and not a presumptuous derivation of its appearances from a Supreme Reason. The expression suited to our feeble notions is, that we conceive the world as if it came, as to its existence and internal plan, from a Supreme Reason, by which notion we both cognise the constitution, which belongs to the world itself, yet without pretending to determine the nature of its cause in itself, and on the other hand, we transfer the ground of this constitution (of the form of reason in the world) upon the relation of the Supreme Cause to the world, without finding the world sufficient by itself for that purpose.⁴²

Let us begin by assuming, as a necessary hypothesis, the deistic concept of the First Being. In Hume’s dialogues, Philo grants this assumption to Cleanthes. This concept understands the First Being only through ontological predicates such as substance and cause. We must make this assumption because reason cannot otherwise be satisfied. In the world of sense, reason deals only with conditions that are themselves always conditional. We can make the assumption without falling into anthropomorphism, which transfers predicates from the world of sense to a Being entirely distinct from that world. The predicates we use are mere categories. Although they do not give us a determinate concept of God, they give us one that is not limited by any conditions of sensibility. Nothing, then, prevents us from attributing causality through reason to this Being in relation to the world. We can thus move on to theism without having to attribute this kind of reason to God in himself, as a property belonging to him. In the world of sense, we can bring the use of reason to its highest point only by assuming a supreme reason as the cause of all connections in the world. This allows reason to remain completely consistent with itself throughout all possible experience. Such a principle must be entirely beneficial to reason and cannot harm its application to nature in any way. At the same time, we do not transfer reason as a property to the First Being in himself. We attribute reason only to his relation to the world of sense, and so avoid anthropomorphism entirely. Here we consider only the cause of the form of reason perceived everywhere in the world. We attribute reason to the Supreme Being insofar as he contains the ground of this form of reason in the world. But we do so only by analogy. The expression shows merely the relation that the Supreme Cause, unknown to us, has to the world, so as to determine everything in it in the highest possible conformity with reason. Thus we are prevented from using reason as an attribute to conceive God. Instead, we use it to conceive the world as we must if we are to make the greatest possible use of reason according to principle. We thereby acknowledge that the Supreme Being is entirely inscrutable and cannot even be thought of in any definite way as he is in himself. This approach keeps us from using our concepts of reason as an efficient cause, acting through the will, in a transcendent way. We do not use them to define the Divine Nature by properties borrowed solely from human nature, and so lose ourselves in crude and extravagant notions. It also keeps us from flooding our study of the world with explanations that go beyond nature, based on our notions of human reason transferred to God. Such explanations would cause our study to lose its proper use. It should be a rational study of nature alone, not a presumptuous derivation of its appearances from a Supreme Reason. The expression suited to our limited understanding is that we conceive the world as if its existence and internal plan came from a Supreme Reason. In this way, we come to know the constitution that belongs to the world itself, without claiming to define the nature of its cause in itself. At the same time, we place the ground of this constitution, the form of reason in the world, in the relation of the Supreme Cause to the world. We do not find the world sufficient by itself to provide that ground.⁴²

Thus the difficulties which seem to oppose theism disappear by combining with Hume's principle—"not to carry the use of reason dogmatically beyond the field of all possible experience"—this other principle, which he quite overlooked: "not to consider the field of experience as one which bounds itself in the eye of our reason." The Critique of Pure Reason here points out the true mean between dogmatism, which Hume combats, and skepticism, which he would substitute for it—a mean which is not like other means that we find advisable to determine for ourselves as it were mechanically (by adopting something from one side and something from the other), and by which nobody is taught a better way, but such a one as can be accurately determined on principles.

Thus, the difficulties that seem to stand in the way of theism disappear when we combine two principles. The first is Hume's principle: "Do not use reason dogmatically beyond the field of all possible experience." The second is a principle Hume entirely overlooked: "Do not regard the field of experience as setting its own limits for our reason." The Critique of Pure Reason shows the true middle way between dogmatism, which Hume opposes, and skepticism, which he would put in its place. This middle way is unlike others we choose for ourselves as if by a mechanical method. Such a method takes something from each side but teaches no one a better way. The middle way shown by the Critique, however, can be precisely determined by principles.

§ 59. At the beginning of this annotation I made use of the metaphor of a boundary, in order to establish the limits of reason in regard to its suitable use. The world of sense contains merely appearances, which are not things in themselves, but the understanding must assume these latter ones, viz., noumena. In our reason both are comprised, and the question is, How does reason proceed to set boundaries to the understanding as regards both these fields? Experience, which contains all that belongs to the sensuous world, does not bound itself; it only proceeds in every case from the conditioned to some other equally conditioned object. Its boundary must lie quite without it, and this field is that of the pure beings of the understanding. But this field, so far as the determination of the nature of these beings is concerned, is an empty space for us, and if dogmatically-determined concepts alone are in question, we cannot pass out of the field of possible experience. But as a boundary itself is something positive, which belongs as well to that which lies within, as to the space that lies without the given complex, it is still an actual positive cognition, which reason only acquires by enlarging itself to this boundary, yet without attempting to pass it; because it there finds itself in the presence of an empty space, in which it can conceive forms of things, but not things themselves. But the setting of a boundary to the field of the understanding by something, which is otherwise unknown to it, is still a cognition which belongs to reason even at this standpoint, and by which it is neither confined within the sensible, nor straying without it, but only refers, as befits the knowledge of a boundary, to the relation between that which lies without it, and that which is contained within it.

§ 59. At the beginning of this annotation, I used the metaphor of a boundary to establish the limits of reason’s proper use. The world of sense contains only appearances, which are not things in themselves. But the understanding must assume that things in themselves, namely noumena, exist. Our reason includes both appearances and noumena. The question is how reason sets boundaries to the understanding in relation to both fields. Experience contains everything that belongs to the world of sense, but it does not limit itself. In every case, it only moves from one conditioned object to another equally conditioned object. Its boundary must lie entirely outside it, in the field of pure beings of the understanding. But this field is empty for us when it comes to determining the nature of those beings. If the only concepts in question are dogmatically determined ones, we cannot leave the field of possible experience. A boundary itself, however, is something positive. It belongs both to what lies within the given whole and to the space outside it. So reason gains genuine, positive knowledge by extending itself to this boundary, without trying to cross it. At the boundary, reason finds itself facing an empty space, where it can conceive forms of things but not things themselves. The field of the understanding is therefore bounded by something otherwise unknown to it. Even from this standpoint, reason knows that the field has this boundary. This knowledge neither confines reason within the sensible world nor lets it wander outside it. Rather, as knowledge of a boundary should, it concerns only the relation between what lies outside the boundary and what lies within it.

Natural theology is such a concept at the boundary of human reason, being constrained to look beyond this boundary to the Idea of a Supreme Being (and, for practical purposes to that of an intelligible world also), not in order to determine anything relatively to this pure creation of the understanding, which lies beyond the world of sense, but in order to guide the use of reason within it according to principles of the greatest possible (theoretical as well as practical) unity. For this purpose we make use of the reference of the world of sense to an independent reason, as the cause of all its connexions. Thereby we do not purely invent a being, but, as beyond the sensible world there must be something that can only be thought by the pure understanding, we determine that something in this particular way, though only of course according to analogy.

Natural theology is a concept at the boundary of human reason. Reason is compelled to look beyond this boundary to the Idea of a Supreme Being. For practical purposes, it must also look to the Idea of an intelligible world. It does not do this to establish anything about these pure creations of the understanding, which lie beyond the world of sense. Rather, it does so to guide its use within the world of sense, according to principles of the greatest possible theoretical and practical unity. For this purpose, we relate the world of sense to an independent reason as the cause of all its connections. In doing so, we do not simply invent a being. Beyond the world of sense, there must be something that can be thought of only by the pure understanding. We describe that something in this particular way, though of course only by analogy.

And thus there remains our original proposition, which is the résumé of the whole Critique: "that reason by all its a priori principles never teaches us anything more than objects of possible experience, and even of these nothing more than can be cognised in experience." But this limitation does not prevent reason leading us to the objective boundary of experience, viz., to the reference to something which is not itself an object of experience, but is the ground of all experience. Reason does not however teach us anything concerning the thing in itself: it only instructs us as regards its own complete and highest use in the field of possible experience. But this is all that can be reasonably desired in the present case, and with which we have cause to be satisfied.

Our original proposition therefore remains. It sums up the whole Critique: "Through all its a priori principles, reason teaches us only about objects of possible experience. Even about these objects, it teaches us only what can be known through experience." But this limit does not prevent reason from leading us to the objective boundary of experience. There, reason points to something that is not itself an object of experience but is the ground of all experience. However, reason teaches us nothing about the thing in itself. It only teaches us about its own fullest and highest use in the field of possible experience. But this is all we can reasonably ask for in this case, and we have reason to be satisfied with it.

§ 60. Thus we have fully exhibited metaphysics as it is actually given in the natural predisposition of human reason, and in that which constitutes the essential end of its pursuit, according to its subjective possibility. Though we have found, that this merely natural use of such a predisposition of our reason, if no discipline arising only from a scientific critique bridles and sets limits to it, involves us in transcendent, either apparently or really conflicting, dialectical syllogisms; and this fallacious metaphysics is not only unnecessary as regards the promotion of our knowledge of nature, but even disadvantageous to it: there yet remains a problem worthy of solution, which is to find out the natural ends intended by this disposition to transcendent concepts in our reason, because everything that lies in nature must be originally intended for some useful purpose.

§ 60. We have now fully shown how metaphysics is possible as a disposition of human reason and in its essential aim. We have done so in terms of its subjective possibility. But we have found that the merely natural use of this disposition leads us into transcendent dialectical arguments. These arguments conflict with one another, either apparently or in reality. This happens unless a discipline arising only from a scientific critique restrains the disposition and sets limits to it. Such fallacious metaphysics is not only unnecessary for advancing our knowledge of nature, but even harmful to it. Yet a problem remains that is worth solving. We must discover the natural ends intended by our reason's disposition toward transcendent concepts. For everything in nature must originally be intended for some useful purpose.

Such an inquiry is of a doubtful nature; and I acknowledge, that what I can say about it is conjecture only, like every speculation about the first ends of nature. The question does not concern the objective validity of metaphysical judgments, but our natural predisposition to them, and therefore does not belong to the system of metaphysics but to anthropology.

Such an inquiry is uncertain. I admit that what I can say about it is only conjecture, like all speculation about nature's original purposes. The question does not concern the objective validity of metaphysical judgments, but our natural tendency to make them. It therefore belongs to anthropology, not to the system of metaphysics.

When I compare all the transcendental Ideas, the totality of which constitutes the particular problem of natural pure reason, compelling it to quit the mere contemplation of nature, to transcend all possible experience, and in this endeavor to produce the thing (be it knowledge or fiction) called metaphysics, I think I perceive that the aim of this natural tendency is, to free our notions from the fetters of experience and from the limits of the mere contemplation of nature so far as at least to open to us a field containing mere objects for the pure understanding, which no sensibility can reach, not indeed for the purpose of speculatively occupying ourselves with them (for there we can find no ground to stand on), but because practical principles, which, without finding some such scope for their necessary expectation and hope, could not expand to the universality which reason unavoidably requires from a moral point of view.

When I compare all the transcendental Ideas, I think I can see the aim of a natural tendency of reason. Together, these Ideas make up the distinctive problem of natural pure reason. They compel reason to leave the mere study of nature and go beyond all possible experience. In this effort, reason produces what we call metaphysics, whether it is knowledge or fiction. The aim of this natural tendency seems to be to free our concepts from the bonds of experience and the limits of merely studying nature. It frees them at least enough to open up a field of objects for the pure understanding alone, beyond the reach of our senses. The purpose is not to speculate about these objects, for there we can find no ground to stand on. Rather, practical principles need such a field for their necessary expectations and hopes. Without it, they could not reach the universality that reason inevitably requires from a moral point of view.

So I find that the Psychological Idea (however little it may reveal to me the nature of the human soul, which is higher than all concepts of experience), shows the insufficiency of these concepts plainly enough, and thereby deters me from materialism, the psychological notion of which is unfit for any explanation of nature, and besides confines reason in practical respects. The Cosmological Ideas, by the obvious insufficiency of all possible cognition of nature to satisfy reason in its lawful inquiry, serve in the same manner to keep us from naturalism, which asserts nature to be sufficient for itself. Finally, all natural necessity in the sensible world is conditional, as it always presupposes the dependence of things upon others, and unconditional necessity must be sought only in the unity of a cause different from the world of sense. But as the causality of this cause, in its turn, were it merely nature, could never render the existence of the contingent (as its consequent) comprehensible, reason frees itself by means of the Theological Idea from fatalism, (both as a blind natural necessity in the coherence of nature itself, without a first principle, and as a blind causality of this principle itself), and leads to the concept of a cause possessing freedom, or of a Supreme Intelligence. Thus the transcendental Ideas serve, if not to instruct us positively, at least to destroy the rash assertions of Materialism, of Naturalism, and of Fatalism, and thus to afford scope for the moral Ideas beyond the field of speculation. These considerations, I should think, explain in some measure the natural predisposition of which I spoke.

I find that the Psychological Idea reveals little about the nature of the human soul, which lies beyond all concepts drawn from experience. But the Psychological Idea shows clearly enough that these concepts are inadequate. It therefore discourages me from accepting materialism. Materialism’s psychological conception cannot explain nature and also restricts reason in practical matters. Similarly, the Cosmological Ideas keep us from naturalism, which claims that nature is sufficient in itself. They do so by showing that all possible knowledge of nature is obviously insufficient to satisfy reason in its legitimate inquiry. Finally, all natural necessity in the world of the senses is conditional, since it always presupposes that things depend on other things. Unconditional necessity must be sought only in the unity of a cause distinct from the world of the senses. But if the causality of this cause were itself merely natural, it could never make the resulting existence of contingent things comprehensible. Through the Theological Idea, reason therefore frees itself from fatalism and leads to the concept of a cause possessing freedom, or a Supreme Intelligence. Fatalism takes two forms: blind natural necessity within the interconnected world of nature, without a first principle, and the blind causality of that first principle itself. Thus, even if the transcendental Ideas do not give us positive knowledge, they at least destroy the rash claims of Materialism, Naturalism, and Fatalism. They thereby make room for moral Ideas beyond the field of speculation. I think these considerations explain, to some extent, the natural predisposition I mentioned.

The practical value, which a merely speculative science may have, lies without the bounds of this science, and can therefore be considered as a scholion merely, and like all scholia does not form part of the science itself. This application however surely lies within the bounds of philosophy, especially of philosophy drawn from the pure sources of reason, where its speculative use in metaphysics must necessarily be at unity with its practical use in morals. Hence the unavoidable dialectics of pure reason, considered in metaphysics, as a natural tendency, deserves to be explained not as an illusion merely, which is to be removed, but also, if possible, as a natural provision as regards its end, though this duty, a work of supererogation, cannot justly be assigned to metaphysics proper.

The practical value that a purely speculative science may have lies outside that science. It can therefore be considered only as a scholion, or supplementary note. Like all scholia, it is not part of the science itself. Applying such a science, however, certainly falls within philosophy. This is especially true of philosophy drawn from the pure sources of reason. In such philosophy, the speculative use of reason in metaphysics must necessarily be consistent with its practical use in morals. Pure reason has an unavoidable tendency toward dialectic, which metaphysics examines. This dialectic deserves to be explained not merely as an illusion to be removed. If possible, it should also be explained as something nature has provided for a purpose. Even so, this task goes beyond what is required and cannot justly be assigned to metaphysics itself.

The solutions of these questions which are treated in the chapter on the Regulative Use of the Ideas of Pure Reason⁴³ should be considered a second scholion which however has a greater affinity with the subject of metaphysics. For there certain rational principles are expounded which determine a priori the order of nature or rather of the understanding, which seeks nature's laws through experience. They seem to be constitutive and legislative with regard to experience, though they spring from pure reason, which cannot be considered, like the understanding, as a principle of possible experience. Now whether or not this harmony rests upon the fact, that just as nature does not inhere in appearances or in their source (the sensibility) itself, but only in so far as the latter is in relation to the understanding, as also a systematic unity in applying the understanding to bring about an entirety of all possible experience can only belong to the understanding when in relation to reason; and whether or not experience is in this way mediately subordinate to the legislation of reason: may be discussed by those who desire to trace the nature of reason even beyond its use in metaphysics, into the general principles of a history of nature; I have represented this task as important, but not attempted its solution, in the book itself.⁴⁴

The solutions to these questions are discussed in the chapter on the Regulative Use of the Ideas of Pure Reason⁴³. They should be considered a second scholion, though one more closely related to metaphysics. That chapter explains certain principles of reason that determine a priori the order of nature. More precisely, they determine the order of the understanding, which seeks nature’s laws through experience. These principles seem to constitute and legislate for experience, though they arise from pure reason. Unlike the understanding, pure reason cannot be considered a principle of possible experience. Nature does not inhere in appearances or in their source, sensibility, itself. It exists only insofar as sensibility is related to the understanding. Likewise, systematic unity in the use of the understanding to achieve a whole of all possible experience may belong to the understanding only in relation to reason. This may or may not explain the harmony between these principles of reason and experience. Experience may or may not, in this way, be indirectly subject to the legislation of reason. Those who wish to trace the nature of reason beyond its use in metaphysics, into the general principles of a history of nature, may discuss these questions. In the book itself, I presented this task as important, but did not attempt to solve it.⁴⁴

And thus I conclude the analytical solution of the main question which I had proposed: How is metaphysics in general possible? by ascending from the data of its actual use in its consequences, to the grounds of its possibility.

This concludes my analytical answer to the main question I proposed: How is metaphysics possible in general? I have worked from the consequences of its actual use back to the grounds of its possibility.