Critique of Pure Reason

by Immanuel Kant

  If we review the entire extent of our knowledge supplied to us by the understanding, we shall find that it is the systematising of that knowledge, that is, its cohe-

  rence according to one principle, which forms the proper province of reason. This unity of reason always presupposes an idea, namely, that of the form of a whole of our knowledge, preceding the definite knowledge of its parts, and containing the conditions according to which we are to determine a priori the place of every part and its relation to the rest. Such an idea accordingly demands the complete unity of the knowledge of our understanding, by which that knowledge becomes not only a mere aggregate but a system, connected according to necessary laws. We ought not to say that such an idea is a concept of an object, but only of the complete unity of concepts, so far as that unity can serve as a rule of the understanding. Such concepts of reason are not derived from nature, but we only interrogate nature, according to these ideas, and consider our knowledge as defective so long as it is not adequate to them. We must confess that pure earth, pure water, pure air, etc., are hardly to be met with. Nevertheless we require the concepts of them (which, so far as their perfect purity is concerned, have their origin in reason only) in order to be able to determine properly the share which belongs to every one of these natural causes in phenomena. Thus every kind of matter is referred to earths (as mere weight), to salts and inflammable bodies (as force), and lastly, to water and air as vehicles (or, as it were, machines, by which the former exercise their operations), in order thus, according to the idea of a mechanism, to explain the mutual chemical workings of matter. For, although not openly acknowledged in these terms, such an influence of reason on the classifications of natural philosophers can easily be discovered.

  If reason is the faculty of deducing the particular from the general, the general is either certain in itself and given, or not. In the former case nothing is required but judgment in subsuming, the particular being thus necessarily determined by the general. This I shall call the apodictic use of reason. In the latter case, when the general is admitted as problematical only, and as a mere idea, while the particular is certain, but the universality of the rule applying to it is still a problem, several particular cases, which are all certain, are tested by the rule, whether they submit to it; and in this case, when it appears that all particular cases which can be produced are subjected to it, the rule is concluded to be universal, and from that universality of the rule conclusions are drawn afterwards with regard to all cases, even those that are not given by themselves. This I shall call the hypothetical use of reason.

  The hypothetical use of reason, resting on ideas as problematical concepts, ought not to be used constitutively, as if we could prove by it, judging strictly, the truth of the universal rule, which has been admitted as an hypothesis. For how are we to know all possible cases, which, as subject to the same principle, should prove its universality? The proper hypothetical use of reason is regulative only, and intended to introduce, as much as possible, unity into the particulars of knowledge, and thus to approximate the rule to universality.

  The hypothetical use of reason aims therefore at the systematical unity of the knowledge of the understanding, and that unity is the touchstone of the truth of the rules. On the other hand, that systematical unity (as a mere idea) is only a projected unity, to be considered, not as given in itself, but as a problem only, though helping us to discover a principle for the manifold and particular exercise of the understanding, and thus to lead the understanding to cases also which are not given, and to render it more systematical.

  We have learnt, therefore, that the systematical unity, introduced by reason into the manifold knowledge of the understanding, is a logical principle, intended to help the understanding by means of ideas, where by itself it is insufficient to establish rules, and at the same time to impart to the variety of its rules a certain harmony (or system according to principles), and by it a certain coherence, so far as that is possible. To say, however, whether the nature of the objects or the nature of the understanding which recognises them as objects, were in themselves intended for systematical unity, and whether to a certain extent we may postulate real unity a priori, without any reference to the peculiar interest of reason, maintaining that all possible kinds of knowledge of the understanding (therefore the empirical also) possess such unity and are subject to such general principles from which, in spite of their differences, they can all be derived, would be to apply a transcendental principle of reason, and to render systematical unity necessary, not only subjectively and logically as a method, but objectively also.

  We shall try to illustrate this use of reason by an example. One of the different kinds of unity, according to the concepts of the understanding, is that of the causality of a substance, which we call power. The different manifestations of one and the same substance display at first so much diversity that one feels constrained to admit at first almost as many powers as there are effects. Thus we see, for instance, in the human mind sensation, consciousness, imagination, memory, wit, discrimination, pleasure, desire, etc. At first a simple logical maxim tells us to reduce this apparent diversity as much as possible by discovering, through comparison, hidden identity, and finding out, for instance, whether imagination connected with consciousness, be not memory, wit, discrimination, or, it may be, understanding and reason. The idea of a fundamental power, of which logic knows nothing as to its existence, is thus at least the problem of a systematical representation of the existing diversity of powers. The logical principle of reason requires us to produce this unity as far as possible, and the more we find that manifestations of one or the other power are identical, the more probable does it become that they are only different expressions of one and the same power which, relatively speaking, may be called their fundamental power. The same is done with the others.

  These relatively fundamental powers must again be compared with each other, in order, if possible, by discovering their harmony, to bring them nearer to one only radical, that is, absolute fundamental power. Such a unity, however, is only an hypothesis of reason. It is not maintained that such a unity must really exist, but only that we must look for it in the interest of reason, that is, for the establishment of certain principles for the various rules supplied to us by experience, and thus introduce, if it is possible, systematical unity into our knowledge.

  If, however, we watch the transcendental use of the understanding, we find that the idea of a fundamental power is not only meant as a problem, and for hypotheti- cal use, but claims for itself objective reality, postulating the systematical unity of the diverse powers of a substance, and thus establishing an apodictic principle of reason. For without even having tested the harmony of those diverse powers, nay, even if failing to discover it, after repeated experiments, we still suppose that such a unity exists, and this not only, as in our example, on account of the unity of the substance, but even in cases where very many, though to a certain degree homogeneous, powers are seen, as in matter in general. Here, too, reason presupposes a systematical unity of diverse powers, because particular laws of nature are subject to more general laws, and parsimony in principles is not only considered as an economical rule of reason, but as an essential law of nature.

  And, indeed, it is difficult to understand how a logical principle by which reason demands the unity of rules can exist without a transcendental principle, by which such a systematical unity is admitted as inherent in the objects themselves, and as a priori necessary. For how could reason in its logical application presume to treat the diversity of powers which we see in nature as simply a disguised unity, and to deduce it, as far as possible, from some fundamental power, if it were open to reason to admit equally the diversity of all powers, and to look upon the systematical unity in their derivation as contrary to nature? In doing this reason would run counter to its own destination, and propose as its aim an idea contrary to the constitution of nature. Nor could we say that reason had previously, according to it
s principles, deduced that unity from the contingent character of nature, because this law of reason, compelling her to look for unity, is necessary, and without it we should have no reason at all, and, in the absence of reason, no coherent use of the understanding, and, in the absence of that, no sufficient test of empirical truth;—on which account we must admit the systematical unity of nature as objectively valid and necessary.

  We find this transcendental presupposition concealed in the cleverest way in the principles of philosophers, though they are not aware of it, nor have confessed it to themselves. That all the diversities of particular things do not exclude identity of species, that the various species must be treated as different determinations (varieties) of a few genera, and these again of still higher genera; that therefore we ought to look for a certain systematical unity of all possible empirical concepts, as derivable from higher and more general concepts, this is a rule of the schools or a logical principle without which no use of the understanding would be possible; for we can only conclude the particular from the general, if the general qualities of things form the foundation on which the particular qualities rest.

  That, however, there exists in nature such a unity, is only a supposition of the philosophers, embodied in their well-known scholastic rule, 'entia praeter necessitatem non esse multiplicanda,' 'beginnings or principles should not be multiplied beyond necessity.' It is implied in this, that the nature of things itself offers material for the postulated unity of reason, and that the apparent infinite variety ought not to prevent us from supposing behind it the existence of unity in fundamental properties, from which all diversity is derived by mere determination only. That unity, though it is an idea only, has been at all times so zealously pursued, that there was more ground for moderating than for encouraging the desire for it. It was something when chemists succeeded in reducing all salts to two genera, namely, acids and alkalies; but they tried to consider even this distinction as a variety only, or as a different manifestation of one and the same fundamental element. Different kinds of earths (the material of stones and even of metals) have been reduced gradually to three, at last to two; but not content with this, chemists cannot get rid of the idea that there is behind those varieties but one genus, nay, that there may be even a common principle for the earths and the salts. It might be supposed that this is only an economical trick of reason, for the purpose of saving itself trouble, and a purely hypothetical attempt which, if successful, would impart by that very unity a certain amount of probability to the presupposed principle of explanation. Such a selfish purpose, however, can easily be distinguished from the idea according to which we all presuppose that this unity of reason agrees with nature, and that in this case reason does not beg but bids, although we may be quite unable, as yet, to determine the limits of that unity.

  If there existed among phenomena so great a diversity, not of form, for in this they may be similar, but of contents, that even the sharpest human understanding could not, by a comparison of the one with the other, discover the slightest similarity among them (a case which is quite conceivable), the logical law of genera would have no existence at all, there would be no concept of genus, nor any general concept, nay, no understanding at all, considering that the understanding has to do with concepts only. The logical principle of genera presupposes, therefore, a transcendental one, if it is to be applied to nature, that is, to all objects presented to our senses. According to it, in the manifoldness of a possible experience, some homogeneousness is necessarily supposed (although it many be impossible to determine its degree a priori), because without it, no empirical concepts, and consequently no experience, would be possible.

  The logical principle of genera, which postulates identity, is balanced by another principle, namely, that of species, which requires manifoldness and diversity in things, in spite of their agreement as belonging to the same genus, and which prescribes to the understanding that it should pay no less attention to the one than to the other. This principle, depending on acute observation or on the faculty of distinction, checks the generalising flights of fancy, and reason thus exhibits a twofold and conflicting interest, namely, on the one hand, the interest in the extent (generality) of genera, on the other hand, the interest in the contents (distinction) of the manifoldness of species. In the former case the understanding thinks more under its concepts, in the latter, more in its concepts. This distinction shows itself in the different manner of thought among students of nature, some of them (who are pre-eminently speculative) being almost averse to heterogeneousness, and always intent on the unity of genera; while others, preeminently empirical, are constantly striving to divide nature into so much variety that one might lose almost all hope of being able to judge its phenomena according to general principles.

  This latter tendency of thought is likewise based on a logical principle which aims at the systematical completeness of all knowledge, so that, beginning with the genus and descending to the manifold that may be contained in it, we try to impart extension to our system, as we tried to impart unity to it, when ascending to a genus. For if we only know the sphere of a concept which determines a genus, we can no more judge how far its subdivision may be carried than we can judge how far the divisibility of matter may be carried, by knowing the space it occupies. Hence every genus requires species, and these again sub-species, and as none even of these sub-species is without a sphere (extent as conceptus communis), reason in its utmost extension requires that no species or sub-species should in itself be considered as the lowest. Every species is always a concept containing that only which is common to different things, and as it cannot be completely determined, it cannot be directly referred to an individual, but must always comprehend other concepts, that is, subspecies. This principle of specification might be expressed by entium varietates non temere esse minuendas.

  It is easily seen that this logical law also would be without meaning and incapable of application, unless it were founded on a transcendental law of specification which, though it cannot demand a real infinity of variety in things that are to become our objects (for this would not be justified by the logical principle, which only asserts the indeterminability of the logical sphere with regard to a possible division), yet imposes on the understanding the duty of looking for sub-species under every species, and for smaller varieties for every variety. If there were no lower concepts, there could not be higher concepts. Now the understanding knows all that it knows by concepts only, and hence, however far it may carry the division, never by means of intuition alone, but again and again by lower concepts. In order to know phenomena in their complete determination (which is possible by the understanding only) it is necessary to carry on without stopping the specification of its concepts, and always to proceed to still remaining differences or varieties of which abstraction had been made in forming the concept of the species, and still more in forming that of the genus.

  Nor can this law of specification have been derived from experience, which can never give so far-reaching a prospect. Empirical specification very soon comes to a standstill in the distinction of the manifold, unless it is led by the antecedent transcendental law of specification, as a principle of reason, and impelled to look for and to conjecture still differences, even where they do not appear to the senses. That absorbent earths are of different kinds (chalk and muriatic earths) could only be discovered by an antecedent rule of reason, which required the understanding to look for diversity, because it presupposed such wealth in nature as to feel justified in anticipating such diversity. For it is only under a presupposition of a diversity in nature, and under the condition that its objects should be homogeneous, that we have understanding, because it is this very diversity of all that can be comprehended under a concept which constitutes the use of that concept, and the occupation of the understanding.

  Reason thus prepares the field for the understanding—

  1st. Through the principle of the homogeneousness of the manifold, as arranged under
higher genera.

  2ndly. Through the principle of the variety of the homogeneous in lower species; to which,

  3rdly, it adds a law of the affinity of all concepts, which requires a continual transition from every species to every other species, by a gradual increase of diversity. We may call these the principles of homogeneousness, of specification, and of continuity of forms. The last arises from the union of the two former, after both in ascending to higher genera, and in descending to lower species, the systematical connection in the idea has been completed; so that all diversities are related to each other, because springing from one highest genus, through all degrees of a more and more extended determination.

  We may represent to ourselves the systematical unity under these three logical principles, in the following manner. Every concept may be regarded as a point which, as the standpoint of the spectator, has its own horizon, enclosing a number of things that may be represented, and, as it were, surveyed from that point. Within that horizon, an infinite number of points must exist, each of which has again its own narrower horizon; that is, every species contains sub-species, according to the principle of specification, and the logical horizon consists of smaller horizons (sub-species only), but not of points, which possess no extent (individuals). But for all these different horizons, that is genera, determined by as many concepts, a common horizon may be imagined, in which they may all be surveyed, as from a common centre. This would be the higher genus, while the highest genus would be the universal and true horizon, determined from the standpoint of the highest concept, and comprehending all variety as genera, species, and sub-species.