SECTION I. System of Cosmological Ideas

That We may be able to enumerate with systematic precision these ideas according to a principle, we must remark, in the first place, that it is from the understanding alone that pure and transcendental conceptions take their origin; that the reason does not properly give birth to any conception, but only frees the conception of the understanding from the unavoidable limitation of a possible experience, and thus endeavours to raise it above the empirical, though it must still be in connection with it. This happens from the fact that, for a given conditioned, reason demands absolute totality on the side of the conditions (to which the understanding submits all phenomena), and thus makes of the category a transcendental idea. This it does that it may be able to give absolute completeness to the empirical synthesis, by continuing it to the unconditioned (which is not to be found in experience, but only in the idea). Reason requires this according to the principle: If the conditioned is given the whole of the conditions, and consequently the absolutely unconditioned, is also given, whereby alone the former was possible. First, then, the transcendental ideas are properly nothing but categories elevated to the unconditioned; and they may be arranged in a table according to the titles of the latter. But, secondly, all the categories are not available for this purpose, but only those in which the synthesis constitutes a series- of conditions subordinated to, not co-ordinated with, each other. Absolute totality is required of reason only in so far as concerns the ascending series of the conditions of a conditioned; not, consequently, when the question relates to the descending series of consequences, or to the aggregate of the co-ordinated conditions of these consequences. For, in relation to a given conditioned, conditions are presupposed and considered to be given along with it. On the other hand, as the consequences do not render possible their conditions, but rather presuppose them- in the consideration of the procession of consequences (or in the descent from the given condition to the conditioned), we may be quite unconcerned whether the series ceases or not; and their totality is not a necessary demand of reason.

Thus we cogitate- and necessarily- a given time completely elapsed up to a given moment, although that time is not determinable by us. But as regards time future, which is not the condition of arriving at the present, in order to conceive it; it is quite indifferent whether we consider future time as ceasing at some point, or as prolonging itself to infinity. Take, for example, the series m, n, o, in which n is given as conditioned in relation to m, but at the same time as the condition of o, and let the series proceed upwards from the conditioned n to m (l, k, i, etc.), and also downwards from the condition n to the conditioned o (p, q, r, etc.)- I must presuppose the former series, to be able to consider n as given, and n is according to reason (the totality of conditions) possible only by means of that series. But its possibility does not rest on the following series o, p, q, r, which for this reason cannot be regarded as given, but only as capable of being given (dabilis).

I shall term the synthesis of the series on the side of the conditions- from that nearest to the given phenomenon up to the more remote- regressive; that which proceeds on the side of the conditioned, from the immediate consequence to the more remote, I shall call the progressive synthesis. The former proceeds in antecedentia, the latter in consequentia. The cosmological ideas are therefore occupied with the totality of the regressive synthesis, and proceed in antecedentia, not in consequentia. When the latter takes place, it is an arbitrary and not a necessary problem of pure reason; for we require, for the complete understanding of what is given in a phenomenon, not the consequences which succeed, but the grounds or principles which precede.

In order to construct the table of ideas in correspondence with the table of categories, we take first the two primitive quanta of all our intuitions, time and space. Time is in itself a series (and the formal condition of all series), and hence, in relation to a given present, we must distinguish a priori in it the antecedentia as conditions (time past) from the consequentia (time future). Consequently, the transcendental idea of the absolute totality of the series of the conditions of a given conditioned, relates merely to all past time. According to the idea of reason, the whole past time, as the condition of the given moment, is necessarily cogitated as given. But, as regards space, there exists in it no distinction between progressus and regressus; for it is an aggregate and not a series- its parts existing together at the same time. I can consider a given point of time in relation to past time only as conditioned, because this given moment comes into existence only through the past time rather through the passing of the preceding time. But as the parts of space are not subordinated, but co-ordinated to each other, one part cannot be the condition of the possibility of the other; and space is not in itself, like time, a series. But the synthesis of the manifold parts of space- (the syntheses whereby we apprehend space)- is nevertheless successive; it takes place, therefore, in time, and contains a series. And as in this series of aggregated spaces (for example, the feet in a rood), beginning with a given portion of space, those which continue to be annexed form the condition of the limits of the former- the measurement of a space must also be regarded as a synthesis of the series of the conditions of a given conditioned. It differs, however, in this respect from that of time, that the side of the conditioned is not in itself distinguishable from the side of the condition; and, consequently, regressus and progressus in space seem to be identical. But, inasmuch as one part of space is not given, but only limited, by and through another, we must also consider every limited space as conditioned, in so far as it presupposes some other space as the condition of its limitation, and so on. As regards limitation, therefore, our procedure in space is also a regressus, and the transcendental idea of the absolute totality of the synthesis in a series of conditions applies to space also; and I am entitled to demand the absolute totality of the phenomenal synthesis in space as well as in time. Whether my demand can be satisfied is a question to be answered in the sequel.

Secondly, the real in space- that is, matter- is conditioned. Its internal conditions are its parts, and the parts of parts its remote conditions; so that in this case we find a regressive synthesis, the absolute totality of which is a demand of reason. But this cannot be obtained otherwise than by a complete division of parts, whereby the real in matter becomes either nothing or that which is not matter, that is to say, the simple. Consequently we find here also a series of conditions and a progress to the unconditioned.

Thirdly, as regards the categories of a real relation between phenomena, the category of substance and its accidents is not suitable for the formation of a transcendental idea; that is to say, reason has no ground, in regard to it, to proceed regressively with conditions. For accidents (in so far as they inhere in a substance) are co-ordinated with each other, and do not constitute a series. And, in relation to substance, they are not properly subordinated to it, but are the mode of existence of the substance itself. The conception of the substantial might nevertheless seem to be an idea of the transcendental reason. But, as this signifies nothing more than the conception of an object in general, which subsists in so far as we cogitate in it merely a transcendental subject without any predicates; and as the question here is of an unconditioned in the series of phenomena- it is clear that the substantial can form no member thereof. The same holds good of substances in community, which are mere aggregates and do not form a series. For they are not subordinated to each other as conditions of the possibility of each other; which, however, may be affirmed of spaces, the limits of which are never determined in themselves, but always by some other space. It is, therefore, only in the category of causality that we can find a series of causes to a given effect, and in which we ascend from the latter, as the conditioned, to the former as the conditions, and thus answer the question of reason.

Fourthly, the conceptions of the possible, the actual, and the necessary do not conduct us to any series- excepting only in so far as the contingent in existence must always be regarded as conditioned, and as indicating, according to a law of the understanding, a condition, under which it is necessary to rise to a higher, till in the totality of the series, reason arrives at unconditioned necessity.

There are, accordingly, only four cosmological ideas, corresponding with the four titles of the categories. For we can select only such as necessarily furnish us with a series in the synthesis of the manifold.


The absolute Completeness

of the


of the given totality of all phenomena.


The absolute Completeness

of the


of given totality in a phenomenon.


The absolute Completeness

of the


of a phenomenon.


The absolute Completeness


of what is changeable in a phenomenon.

We must here remark, in the first place, that the idea of absolute totality relates to nothing but the exposition of phenomena, and therefore not to the pure conception of a totality of things. Phenomena are here, therefore, regarded as given, and reason requires the absolute completeness of the conditions of their possibility, in so far as these conditions constitute a series- consequently an absolutely (that is, in every respect) complete synthesis, whereby a phenomenon can be explained according to the laws of the understanding.

Secondly, it is properly the unconditioned alone that reason seeks in this serially and regressively conducted synthesis of conditions. It wishes, to speak in another way, to attain to completeness in the series of premisses, so as to render it unnecessary to presuppose others. This unconditioned is always contained in the absolute totality of the series, when we endeavour to form a representation of it in thought. But this absolutely complete synthesis is itself but an idea; for it is impossible, at least before hand, to know whether any such synthesis is possible in the case of phenomena. When we represent all existence in thought by means of pure conceptions of the understanding, without any conditions of sensuous intuition, we may say with justice that for a given conditioned the whole series of conditions subordinated to each other is also given; for the former is only given through the latter. But we find in the case of phenomena a particular limitation of the mode in which conditions are given, that is, through the successive synthesis of the manifold of intuition, which must be complete in the regress. Now whether this completeness is sensuously possible, is a problem. But the idea of it lies in the reason- be it possible or impossible to connect with the idea adequate empirical conceptions. Therefore, as in the absolute totality of the regressive synthesis of the manifold in a phenomenon (following the guidance of the categories, which represent it as a series of conditions to a given conditioned) the unconditioned is necessarily contained- it being still left unascertained whether and how this totality exists; reason sets out from the idea of totality, although its proper and final aim is the unconditioned- of the whole series, or of a part thereof.

This unconditioned may be cogitated- either as existing only in the entire series, all the members of which therefore would be without exception conditioned and only the totality absolutely unconditioned- and in this case the regressus is called infinite; or the absolutely unconditioned is only a part of the series, to which the other members are subordinated, but which Is not itself submitted to any other condition.* In the former case the series is a parte priori unlimited (without beginning), that is, infinite, and nevertheless completely given. But the regress in it is never completed, and can only be called potentially infinite. In the second case there exists a first in the series. This first is called, in relation to past time, the beginning of the world; in relation to space, the limit of the world; in relation to the parts of a given limited whole, the simple; in relation to causes, absolute spontaneity (liberty); and in relation to the existence of changeable things, absolute physical necessity.

*The absolute totality of the series of conditions to a given conditioned is always unconditioned; because beyond it there exist no other conditions, on which it might depend. But the absolute totality of such a series is only an idea, or rather a problematical conception, the possibility of which must be investigated- particularly in relation to the mode in which the unconditioned, as the transcendental idea which is the real subject of inquiry, may be contained therein.

We possess two expressions, world and nature, which are generally interchanged. The first denotes the mathematical total of all phenomena and the totality of their synthesis- in its progress by means of composition, as well as by division. And the world is termed nature,* when it is regarded as a dynamical whole- when our attention is not directed to the aggregation in space and time, for the purpose of cogitating it as a quantity, but to the unity in the existence of phenomena. In this case the condition of that which happens is called a cause; the unconditioned causality of the cause in a phenomenon is termed liberty; the conditioned cause is called in a more limited sense a natural cause. The conditioned in existence is termed contingent, and the unconditioned necessary. The unconditioned necessity of phenomena may be called natural necessity.

*Nature, understood adjective (formaliter), signifies the complex of the determinations of a thing, connected according to an internal principle of causality. On the other hand, we understand by nature, substantive (materialiter), the sum total of phenomena, in so far as they, by virtue of an internal principle of causality, are connected with each other throughout. In the former sense we speak of the nature of liquid matter, of fire, etc., and employ the word only adjective; while, if speaking of the objects of nature, we have in our minds the idea of a subsisting whole.

The ideas which we are at present engaged in discussing I have called cosmological ideas; partly because by the term world is understood the entire content of all phenomena, and our ideas are directed solely to the unconditioned among phenomena; partly also, because world, in the transcendental sense, signifies the absolute totality of the content of existing things, and we are directing our attention only to the completeness of the synthesis- although, properly, only in regression. In regard to the fact that these ideas are all transcendent. and, although they do not transcend phenomena as regards their mode, but are concerned solely with the world of sense (and not with noumena), nevertheless carry their synthesis to a degree far above all possible experience- it still seems to me that we can, with perfect propriety, designate them cosmical conceptions. As regards the distinction between the mathematically and the dynamically unconditioned which is the aim of the regression of the synthesis, I should call the two former, in a more limited signification, cosmical conceptions, the remaining two transcendent physical conceptions. This distinction does not at present seem to be of particular importance, but we shall afterwards find it to be of some value.

converted to HTML for RBJones.com by RBJ
first edition 1994/12/23 last modified 1999/8/29