Critique of Pure Reason

by Immanuel Kant

  Whatever, therefore, our concept of an object may contain, we must always step outside it, in order to attribute to it existence. With objects of the senses, this takes place through their connection with any one of my perceptions, according to empirical laws; with objects of pure thought, however, there is no means of knowing their existence, because it would have to be known entirely a priori, while our consciousness of every kind of existence, whether immediately by perception, or by conclusions which connect something with perception, belongs entirely to the unity of experience, and any existence outside that field, though it cannot be declared to be absolutely impossible, is a presupposition that cannot be justified by anything.

  The concept of a Supreme Being is, in many respects, a very useful idea, but, being an idea only, it is quite incapable of increasing, by itself alone, our knowledge with regard to what exists. It cannot even do so much as to inform us any further as to its possibility. The analytical characteristic of possibility, which consists in the absence of contradiction in mere positions (realities), cannot be denied to it; but the connection of all real properties in one and the same thing is a synthesis the possibility of which we cannot judge a priori because these realities are not given to us as such, and because, even if this were so, no judgment whatever takes place, it being necessary to look for the characteristic of the possibility of synthetical knowledge in experience only, to which the object of an idea can never belong. Thus we see that the celebrated Leibniz is far from having achieved what he thought he had, namely, to understand a priori the possibility of so sublime an ideal Being.

  Time and labour therefore are lost on the famous ontological (Cartesian) proof of the existence of a Supreme Being from mere concepts; and a man might as well imagine that he could become richer in knowledge by mere ideas, as a merchant in capital, if, in order to improve his position, he were to add a few noughts to his cash account.

  The Ideal of Pure Reason

  Section V

  Of the Impossibility of a Cosmological Proof of the Existence of God

  It was something quite unnatural, and a mere innovation of scholastic wisdom, to attempt to pick out of an entirely arbitrary idea the existence of the object corresponding to it. Such an attempt would never have been made, if there had not existed beforehand a need of our reason of admitting for existence in general something necessary, to which we may ascend and in which we may rest; and if, as that necessity must be unconditioned and a priori certain, reason had not been forced to seek a concept which, if possible, should satisfy such a demand and give us a knowledge of an existence entirely a priori. Such a concept was supposed to exist in the idea of an ens realissimum, and that idea was therefore used for a more definite knowledge of that, the existence of which one had admitted or been persuaded of independently, namely, of the necessary Being. This very natural procedure of reason was carefully concealed, and instead of ending with that concept, an attempt was made to begin with it, and thus to derive from it the necessity of existence, which it was only meant to supplement. Hence arose that unfortunate ontological proof, which satisfies neither the demands of our natural and healthy understanding, nor the requirements of the schools.

  The cosmological proof, which we have now to examine, retains the connection of absolute necessity with the highest reality, but instead of concluding, like the former, from the highest reality necessity in existence, it concludes from the given unconditioned necessity of any being, its unlimited reality. It thus brings everything at least into the groove of a natural, though I know not whether of a really or only apparently rational syllogism, which carries the greatest conviction, not only for the common, but also for the speculative understanding, and has evidently drawn the first outline of all proofs of natural theology, which have been followed at all times, and will be followed in future also, however much they may be hidden and disguised. We shall now proceed to exhibit and to examine this cosmological proof which Leibniz calls also the proof a contingentia mundi.

  It runs as follows: If there exists anything, there must exist an absolutely necessary Being also. Now I, at least, exist; therefore there exists an absolutely necessary Being. The minor contains an experience, the major the conclusion from experience in general to the existence of the necessary.6 This proof therefore begins with experience, and is not entirely a priori, or ontological; and, as the object of all possible experience is called the world, this proof is called the cosmological proof. As it takes no account of any peculiar property of the objects of experience, by which this world of ours may differ from any other possible world, it is distinguished, in its name also, from the physico-theological proof, which employs as arguments, observations of the peculiar property of this our world of sense.

  The proof then proceeds as follows: The necessary Being can be determined in one way only, that is, by one only of all possible opposite predicates; it must therefore be determined completely by its own concept. Now, there is only one concept of a thing possible, which a priori completely determines it, namely, that of the ens realissimum. It follows, therefore, that the concept of the ens realissimum is the only one by which a necessary Being can be thought, and therefore it is concluded that a highest Being exists by necessity.

  There are so many sophistical propositions in this cosmological argument, that it really seems as if speculative reason had spent all her dialectical skill in order to produce the greatest possible transcendental illusion. Before examining it, we shall draw up a list of them, by which reason has put forward an old argument disguised as a new one, in order to appeal to the agreement of two witnesses, one supplied by pure reason, the other by experience, while in reality there is only one, namely, the first, who changes his dress and voice in order to be taken for a second. In order to have a secure foundation, this proof takes its stand on experience, and pretends to be different from the ontological proof, which places its whole confidence in pure concepts a priori only. The cosmological proof, however, uses that experience only in order to make one step, namely, to the existence of a necessary Being in general. What properties that Being may have, can never be learnt from the empirical argument, and for that purpose reason takes leave of it altogether, and tries to find out, from among concepts only, what properties an absolutely necessary Being ought to possess, i.e. which among all possible things contains in itself the requisite conditions (requisita) of absolute necessity. This requisite is believed by reason to exist in the concept of an ens realissimum only, and reason concludes at once that this must be the absolutely necessary Being. In this conclusion it is simply assumed that the concept of a being of the highest reality is perfectly adequate to the concept of absolute necessity in existence; so that the latter might be concluded from the former. This is the same proposition as that maintained in the ontological argument, and is simply taken over into the cosmological proof, nay, made its foundation, although the intention was to avoid it. For it is clear that absolute necessity is an existence from mere concepts. If, then, I say that the concept of the ens realissimum is such a concept, and is the only concept adequate to necessary existence, I am bound to admit that the latter may be deduced from the former. The whole conclusive strength of the so-called cosmological proof rests therefore in reality on the ontological proof from mere concepts, while the appeal to experience is quite superfluous, and, though it may lead us on to the concept of absolute necessity, it cannot demonstrate it with any definite object. For as soon as we intend to do this, we must at once abandon all experience, and try to find out which among the pure concepts may contain the conditions of the possibility of an absolutely necessary Being. But if in this way the possibility of such a Being has been perceived, its existence also has been proved: for what we are really saying is this, that under all possible things there is one which carries with it absolute necessity, or that this Being exists with absolute necessity.

  Sophisms in arguments are most easily discovered, if they are put forward in a correct scholastic form. This we shall
now proceed to do.

  If the proposition is right, that every absolutely necessary Being is, at the same time, the most real Being (and this is the nervus probandi of the cosmological proof), it must, like all affirmative judgments, be capable of conversion, at least per accidens. This would give us the proposition that some entia realissima are at the same time absolutely necessary beings. One ens realissimum, however, does not differ from any other on any point, and what applies to one, applies also to all. In this case, therefore, I may employ absolute conversion, and say, that every ens realissimum is a necessary Being. As this proposition is determined by its concepts a priori only, it follows that the mere concept of the ens realissimum must carry with it its absolute necessity; and this, which was maintained by the ontological proof, and not recognized by the cosmological, forms really the foundation of the conclusions of the latter, though in a disguised form.

  We thus see that the second road taken by speculative reason, in order to prove the existence of the highest Being, is not only as illusory as the first, but commits in in addition an ignoratio elenchi, promising to lead us by a new path, but after a short circuit bringing us back to the old one, which we had abandoned for its sake.

  I said before that a whole nest of dialectical assumptions was hidden in that cosmological proof, and that transcendental criticism might easily detect and destroy it. I shall here enumerate them only, leaving it to the experience of the reader to follow up the fallacies and remove them.

  We find, first, the transcendental principle of inferring a cause from the accidental. This principle, that everything contingent must have a cause, is valid in the world of sense only, and has not even a meaning outside it. For the purely intellectual concept of the contingent cannot produce a synthetical proposition like that of causality, and the principle of causality has no meaning and no criterion of its use, except in the world of sense, while here it is meant to help us beyond the world of sense.

  Secondly. The inference of a first cause, based on the impossibility of an infinite ascending series of given causes in this world of sense,—an inference which the principles of the use of reason do not allow us to draw even in experience, while here we extend that principle beyond experience, whither that series can never be prolonged.

  Thirdly. The false self-satisfaction of reason with regard to the completion of that series, brought about by removing in the end every kind of condition, without which, nevertheless, no concept of necessity is possible, and by then, when any definite concepts have become impossible, accepting this as a completion of our concept.

  Fourthly. The mistaking the logical possibility of a concept of all united reality (without any internal contradiction) for the transcendental, which requires a principle for the practicability of such a synthesis, such principle however being applicable to the field of possible experience only, etc.

  The trick of the cosmological proof consists only in trying to avoid the proof of the existence of a necessary Being a priori by mere concepts. Such a proof would have to be ontological, and of this we feel ourselves quite incapable. For this reason we take a real existence (of any experience whatever), and conclude from it, as well as may be, some absolutely necessary condition of it. In that case there is no necessity for explaining its possibility, because, if it has been proved that it exists, the question as to its possibility is unnecessary. If then we want to determine that necessary Being more accurately, according to its nature, we do not seek what is sufficient to make us understand from its concept the necessity of its existence. If we could do this, no empirical presupposition would be necessary. No, we only seek the negative condition (conditio sine qua non), without which a Being would not be absolutely necessary. Now, in every other kind of syllogisms leading from a given effect to its cause, this might well be feasible. In our case, however, it happens unfortunately that the condition which is required for absolute necessity exists in one single Being only, which, therefore, would have to contain in its concept all that is required for absolute necessity, and that renders a conclusion a priori, with regard to such necessity, possible. I ought therefore to be able to reason conversely, namely, that everything is absolutely necessary, if that concept (of the highest reality) belongs to it. If I cannot do this (and I must confess that I cannot, if I wish to avoid the ontological proof), I have suffered shipwreck on my new course, and have come back again from where I started. The concept of the highest Being may satisfy all questions a priori which can be asked regarding the internal determinations of a thing, and it is therefore an ideal, without an equal, because the general concept distinguishes it at the same time as an individual being among all possible things. But it does not satisfy the really important question regarding its own existence; and if some one who admitted the existence of a necessary Being were to ask us which of all things in the world could be regarded as such, we could not answer: This here is the necessary Being.

  It may be allowable to admit the existence of a Being entirely sufficient to serve as the cause of all possible effects, simply in order to assist reason in her search for unity of causes. But to go so far as to say that such a Being exists necessarily, is no longer the modest language of an admissible hypothesis, but the bold assurance of apodictic certainty; for the knowledge of that which is absolutely necessary must itself possess absolute necessity.

  The whole problem of the transcendental Ideal is this, either to find a concept compatible with absolute necessity, or to find the absolute necessity compatible with the concept of anything. If the one is possible, the other must be so also, for reason recognises that only as absolutely necessary which is necessary according to its concept. Both these tasks baffle our attempts at satisfying our understanding on this point, and likewise our endeavours to comfort it with regard to its impotence.

  That unconditioned necessity, which we require as the last support of all things, is the true abyss of human reason. Eternity itself, however terrible and sublime it may have been depicted by Haller, is far from producing the same giddy impression, for it only measures the duration of things, but does not support them. We cannot put off the thought, nor can we support it, that a Being, which we represent to ourselves as the highest among all possible beings, should say to himself, I am from eternity to eternity, there is nothing beside me, except that which is something through my will,—but whence am I? Here all sinks away from under us, and the highest perfection, like the smallest, passes without support before the eyes of speculative reason, which finds no difficulty in making the one as well as the other to disappear without the slightest impediment.

  Many powers of nature, which manifest their existence by certain effects, remain perfectly inscrutable to us, because we cannot follow them up far enough by observation. The transcendental object, which forms the foundation of all phenomena, and with it the ground of our sensibility having this rather than any other supreme conditions, is and always will be inscrutable. The thing no doubt is given, but it is incomprehensible. An ideal of pure reason, however, cannot be called inscrutable, because it cannot produce any credentials of its reality beyond the requirement of reason to perfect all synthetical unity by means of it. As, therefore, it is not even given as an object that can be thought, it cannot be said to be, as such, inscrutable; but, being a mere idea, it must find in the nature of reason its place and its solution, and in that sense be capable of scrutiny. For it is the very essence of reason that we are able to give an account of all our concepts, opinions, and assertions either on objective or, if they are a mere illusion, on subjective grounds.

  Discovery and Explanation of the Dialectical Illusion in all Transcendental Proofs of the Existence of a Necessary Being

  Both proofs, hitherto attempted, were transcendental, that is, independent of empirical principles. For although the cosmological proof assumes for its foundation an experience in general, it does not rest on any particular quality of it, but on pure principles of reason, with reference to an existence given by the empirical co
nsciousness in general, and abandons even that guidance in order to derive its support from pure concepts only. What then in these transcendental proofs is the cause of the dialectical, but natural, illusion which connects the concepts of necessity and of the highest reality, and realises and hypostasises that which can only be an idea? What is the cause that renders it inevitable to admit something as necessary in itself among existing things, and yet makes us shrink back from the existence of such a Being as from an abyss? What is to be done that reason should understand itself on this point, and, escaping from the wavering state of hesitatingly approving or disapproving, acquire a calm insight into the matter?