In the "transcendental" exposition of space and time Kant follows this "metaphysical" exposition, which had to prove their non-empirical, and non-discursive, hence their a priori and intuitive, character, with the proof that only such an explanation of space and time could make it conceivable how synthetic cognitions a priori can arise from them. The principles of mathematics are of this kind. The synthetic character of geometrical truths is explained by the intuitive nature of space, their apodictic character by its apriority, and their objective reality or applicability to empirical objects by the fact that space is the condition of (external) perception. The like is true of arithmetic and time.

If space were a mere concept, no proposition could be derived from it which should go beyond the concept and extend our knowledge of its properties. The possibility of such extension or synthesis in mathematics depends on the fact that spatial concepts can always be presented or "constructed" in intuition. The geometrical axiom that in the triangle the sum of two sides is greater than the third is derived from intuition, by describing the triangle in imagination or, actually, on the board. Here the object is given through the cognition and not before it.—If space and time were empirical representations the knowledge obtained from them would lack necessity, which, as a matter of fact, it possesses in a marked degree. While experience teaches us only that something is thus or so, and not that it could not be otherwise, the axioms, (space has only three dimensions, time only one; only one straight line is possible between two points), nay, all the propositions of mathematics are strictly universal and apodictically certain: we are entirely relieved from the necessity of measuring all triangles in the world in order to find out whether the sum of their angles is equal to two right angles, and we do not need, as in the case of judgments of experience, to add the limitation, so far as it is yet known there are no exceptions to this rule. The apriority is the ratio essendi of the strict necessity involved in the "it must be so" (des Soseinmüssens), while the latter is the ratio cognoscendi of the former. Now since the necessity of mathematical judgments can only be explained through the ideality of space, this doctrine is perfectly certain, not merely a probable hypothesis.—The validity of mathematical principles for all objects of perception, finally, is based on the fact that they are rules under which alone experience is possible for us. It should be mentioned, further, that the conceptions of change and motion (change of place) are possible only through and in the representation of time. No concept could make intelligible the possibility of change, that is, of the connection of contradictory predicates in one and the same thing, but the intuition of succession easily succeeds in accomplishing it.

The argument is followed by conclusions and explanations based upon it; (1) Space is the form of the outer, time of the inner, sense. Through the outer sense external objects are given to us, and through the inner sense our own inner states. But since all representations, whether they have external things for their objects or not, belong in themselves, as mental determinations, to our inner state, time is the formal condition of all phenomena in general, directly of internal (psychical) phenomena, and, thereby, indirectly of external phenomena also. (2) The validity of the relations of space and time cognizable a priori is established for all objects of possible experience, but is limited to these. They are valid for all phenomena (for all things which at any time may be given to our senses), but only for these, not for things as they are in themselves. They have "empirical reality, but, at the same time, transcendental ideality." As external phenomena all things are beside one another in space, and all phenomena whatever are in time and of necessity under temporal relations; in regard to all things which can occur in our experience, and in so far as they can occur, space and time are objectively, therefore empirically, real. But they do not possess absolute reality (neither subsistent reality nor the reality of inherence); for if we abstract from our sensuous intuition both vanish, and, apart from the subject (N.B., the transcendental subject, concerning which more below), they are naught. It is only from man's point of view that we can speak of space, and of extended, moveable, changeable things; for we can know nothing concerning the intuitions of other thinking beings, we have no means of discovering whether they are bound by the same conditions which limit our intuitions, and which for us are universally valid. (3) Nothing which is intuited in space is a thing in itself. What we call external objects are nothing but mere representations of our sensibility, whose true correlative, the thing in itself, cannot be known by ever so deep penetration into the phenomenon; such properties as belong to things in themselves can never be given to us through the senses. Similarly nothing that is intuited in time is a thing in itself, so that we intuit ourselves only as we appear to ourselves, and not as we are.

The merely empirical reality of space and time, the limitation of their validity to phenomena, leaves the certainty of knowledge within the limits of experience intact; for we are equally certain of it, whether these forms necessarily belong to things in themselves, or only to our intuitions of things. The assertion of their absolute reality, on the other hand, involves us in sheer absurdities (that is, it necessitates the assumption of two infinite nonentities which exist, but without being anything real, merely in order to comprehend all reality, and on one of which even our own existence would be dependent), in view of which the origin of so peculiar a theory as the idealism of Berkeley appears intelligible. The critical theory of space and time is so far from being identical with, or akin to, the theory of Berkeley, that it furnishes the best and only defense against the latter. If anyone assumes the absolute or transcendental reality of these forms, it is impossible for him to prevent everything, including even our own existence, from being changed thereby into mere illusion. But the critical philosopher is far from degrading bodies to mere illusion; external phenomena are just as real for him as internal phenomena, though only as phenomena, it is true, as (possible) representations.

Phenomenon and illusion are not the same. The transcendental distinction between phenomena and things in themselves must not be confused with the distinction common to ordinary life and to physics, in accordance with which we call the rainbow a mere appearance (better, illusion), but the combination of sun and rain which gives rise to this illusion the thing in itself, as that which in universal experience and in all different positions with respect to the senses, is thus and not otherwise determined in intuition, or that which essentially belongs to the intuition of the object, and is valid for every human sensibility (in antithesis to that which only contingently belongs to it, and is valid only for a special position or organization of this or that sense). Similarly an object always appears to grow smaller as its distance increases, while in itself it is and remains of some fixed size. And this use of words is perfectly correct, in the physical or empirical sense of "in itself"; but in the transcendental sense the raindrops, also, together with their form and size, are themselves mere phenomena, the "in itself" of which remains entirely unknown to us. Kant, moreover, does not wish to see the subjectivity of the forms of intuition placed on a level with the subjectivity of sensations or explained by this, though he accepts it as a fact long established. The sensations of color, of tone, of temperature are, no doubt, like the representation of space in that they belong only to the subjective constitution of the sensibility, and can be attributed to objects only in relation to our senses. But the great difference between the two is that these sense qualities may be different in different persons (the color of the rose may seem different to each eye), or may fail to harmonize with any human sense; that they are not a priori in the same strict sense as space and time, and consequently afford no knowledge of the objects of possible experience independently of perception; and that they are connected with the phenomenon only as the contingently added effects of a particular organization, while space, as the condition of external objects, necessarily belongs to the phenomenon or intuition of them. It is through space alone that it is possible for things to be external objects for us. The subjectivity of sensation is individual, while that of space and time is general or universal to mankind; the former is empirical, individually different, and contingent, the latter a priori and necessary. Space alone, not sensation, is a conditio sine qua non of external perception. Space and time are the sole a priori elements of the sensibility; all other sensuous concepts, even motion and change, presuppose perception; the movable in space and the succession of properties in an existing thing are empirical data.

In confirmation of the theory that all objects of the senses are mere phenomena, the fact is adduced that (with the exception of the will and the feelings, which are not cognitions) nothing is given us through the senses but representations of relations, while a thing in itself cannot be known by mere relations. The phenomenon is a sum total of mere relations. In regard to matter we know only extension, motion, and the laws of this motion or forces (attraction, repulsion, impenetrability), but all these are merely relations of the thing to something, else, that is, external relations. Where is the inner side which underlies this exterior, and which belongs to the object in itself? This is never to be found in the phenomenon, and no matter how far the observation and analysis of nature may advance (a work with unlimited horizons!) they reach nothing but portions of space occupied by matter and effects which matter exercises, that is, nothing beyond that which is comparatively internal, and which, in its turn, consists of external relations. The absolutely inner side of matter is a mere fancy; and if the complaint that the "inner side" of things is concealed from us is to mean that we do not comprehend what the things which appear to us may be in themselves, it is unjust and irrational, for it demands that we should be able to intuit without senses, in other words, that we should be other than men. The transcendent questions concerning the noumenon of things are unanswerable; we know ourselves, even, only as phenomena! A phenomenon consists in nothing but the relation of something in general to the senses.

It is indubitable that something corresponds to phenomena, which, by affecting our sensibility, occasions sensations in us, and thereby phenomena. The very word, the very concept, "phenomenon", indicates a relation to something which is not phenomenon, to an object not dependent on the sensibility. What this may be continues hidden from us, for knowledge is impossible without intuition. Things in themselves are unknowable. Nevertheless the idea (it must be confessed, the entirely empty idea) of this "transcendental object", as an indeterminate somewhat = x which underlies phenomena, is not only allowable, but, as a limiting concept, unavoidable in order to confine the pretensions of sense to the only field which is accessible to it, that is, to the field of phenomena.

The inference "space and time are nothing but representations and representations are in us, therefore space and time as well as all phenomena in them, bodies with their forces and motions, are in us," does not accurately express Kant's position, for he might justly reply that, according to him, bodies as phenomena are in different parts in space from that which we assign to ourselves, and thus without us; that space is the form of external intuition, and through it external objects arise for us from sensations; but that, in regard to the things in themselves which affect us, we are entirely ignorant whether they are within or without us.

It can easily be shown by literal quotations that there were distinct tendencies in Kant, especially in the first edition of his principal work, toward a radical idealism which doubts or denies not merely the cognizability, but also the existence of objects external to the subject and its representations, and which degrades the thing in itself to a mere thought in us, or completely does away with it (e.g., "The representation of an object as a thing in general is not only insufficient, but, … independently of empirical conditions, in itself contradictory "). But these expressions indicate only a momentary inclination toward such a view, not a binding avowal of it, and they are outweighed by those in which idealism is more or less energetically rejected. That which according to Kant exists outside the representation of the individual is twofold: (1) the unknown things in themselves with their problematical characteristics, as the ground of phenomena; (2) the phenomena "themselves" with their knowable immanent laws, and their relations in space and time, as possible representations. When I turn my glance away from the rose its redness vanishes, since this predicate belongs to it only in so far and so long as it acts in the light on my visual apparatus. What, then, is left? That thing in itself, of course, which, when it appears to me, calls forth in me the intuition of the rose. But there is still something else remaining—the phenomenon of the rose, with its size, its form, and its motion in the wind. For these are predicates which must be attributed to the phenomenon itself as the object of my representation. If the rose, as determined in space and time, vanished when I turned my head away, it could not, unless intuited by a subject, experience or exert effects in space and time, could not lose its leaves in the wind and strew the ground with its petals. Perception and thought inform me not merely concerning events of which I am a witness, but also of others which have occurred, or which will occur, in my absence. The process of stripping the leaves from the rose has actually taken place as a phenomenon and does not first become real by my subsequent representation of it or inference to it. The things and events of the phenomenal world exist both before and after my perception, and are something distinct from my subjective and momentary representations of them. The space and time, however, in which they exist and happen are not furnished by the intuiting individual, but by the supra-individual, transcendental consciousness or generic reason of the race. The phenomenon thus stands midway between its objective ground (the absolute thing in itself) and the subject, whose common product it is, as a relative thing in itself, as a reality which is independent of the contingent and changing representation of the individual, empirical subject, which is dependent for its form on the transcendental subject, and which is the only reality accessible to us, yet entirely valid for us. The phenomenal world is not a contingent and individual phenomenon, but one necessary for all beings organized as we are, a phenomenon for humanity. My representations are not the phenomena themselves, but images and signs through which I cognize phenomena, i.e., real things as they are for me and for every man (not as they are in themselves). The reality of phenomena consists in the fact that they can be perceived by men, and the objective validity of my knowledge of them in the fact that every man must agree in it. The laws which the understanding (not the individual understanding!) imposes upon nature hold for phenomena, because they hold for every man. Objectivity is universal validity. If the world of phenomena which is intuited and known by us wears a different appearance from the world of things in themselves, this does not justify us in declaring it to be mere seeming and dreaming; a dream which all dream together, and which all must dream, is not a dream, but reality. As we must represent the world> so it is, though for us, of course, and not in itself.

Many places in Kant's works seem to argue against the intermediate position here ascribed to the world of phenomena—according to which it is less than things in themselves and more than subjective representation—which, since they explain the phenomenon as a mere representation, leave room for only two factors (on the one hand, the thing in itself = that in the thing which cannot be represented; on the other, the thing for me = my representation of the thing). In fact, the distinction between the phenomenon "itself" and the representation which the individual now has of it and now does not have, is far from being everywhere adhered to with desirable clearness; and wherever it is impossible to substitute that which has been represented and that which may be represented or possible intuitions for "mere representations in me," we must acknowledge that there is a departure from the standpoint which is assumed in some places with the greatest distinctness. The latter finds unequivocal expression, among other places, in the "Analogies of Experience" and the "Deduction of the Pure Concepts of the Understanding," § 2, No. 4 (first edition). The second of these passages speaks of one and the same universal experience, in which all perceptions are represented in thoroughgoing and regular connection, and of the thoroughgoing affinity of phenomena as the basis of the possibility of the association of representations. This affinity is ascribed to the objects of the senses, not to the representations, whose association is rather the result of the affinity, and not to the things in themselves, in regard to which the understanding has no legislative power.