On the Fourfold Root of the Principle of Sufficient...

One of the chief objects of the "Critique of Pure Reason" is to show the universal validity, for all experience, of the causal law, its à priority, and, as a necessary consequence of this, its restriction to possible experience. Nevertheless, I cannot assent to the proof there given of the à priority of the principle, which is substantially this:—"The synthesis of the manifold by the imagination, which is necessary for all empirical knowledge, gives succession, but not yet determinate succession: that is, it leaves undetermined which of two states perceived was the first, not only in my imagination, but in the object itself. But definite order in this succession—through which alone what we perceive becomes experience, or, in other words, authorizes us to form objectively valid judgments—is first brought into it by the purely intellectual conception of cause and effect. Thus the principle of causal relation is the condition which renders experience possible, and, as such, it is given us à priori."[98]

According to this, the order in which changes succeed each other in real objects becomes known to us as objective only by their causality. This assertion Kant repeats and explains in the "Critique of Pure Reason," especially in his "Second Analogy of Experience,"[99] and again at the conclusion of his "Third Analogy," and I request every one who desires to understand what I am now about to say, to read these passages. In them he affirms everywhere that the objectivity of the succession of representations—which he defines as their correspondence with the succession of real objects—is only known through the rule by which they follow upon one another: that is, through the law of causality; that my mere apprehension consequently leaves the objective relation between phenomena following one another quite undetermined: since I merely apprehend the succession of my own representations, but the succession in my apprehension does not authorize me to form any judgment whatever as to the succession in the object, unless that judgment be based upon causality; and since, besides, I might invert the order in which these perceptions follow each other in my apprehension, there being nothing which determines them as objective. To illustrate this assertion, Kant brings forward the instance of a house, whose parts we may consider in any order we like, from top to bottom, or from bottom to top; the determination of succession being in this case purely subjective and not founded upon an object, because it depends upon our pleasure.

In opposition to this instance, he brings forward the perception of a ship sailing down a river, which we see successively lower and lower down the stream, which perception of the successively varying positions of the ship cannot be changed by the looker-on. In this latter case, therefore, he derives the subjective following in his own apprehension from the objective following in the phenomenon, and on this account he calls it an event. Now I maintain, on the contrary, that there is no difference at all between these two cases, that both are events, and that our knowledge of both is objective: that is to say, it is knowledge of changes in real objects recognized as such by the Subject. Both are changes of relative position in two bodies. In the first case, one of these bodies is a part of the observer's own organism, the eye, and the other is the house, with respect to the different parts of which the eye successively alters its position. In the second, it is the ship which alters its position towards the stream; therefore the change occurs between two bodies. Both are events, the only difference being that, in the first, the change has its starting-point in the observer's own body, from whose sensations undoubtedly all his perceptions originally proceed, but which is nevertheless an object among objects, and in consequence obeys the laws of the objective, material world. For the observer, as a purely cognising individual, any movement of his body is simply an empirically perceived fact.

It would be just as possible in the second as in the first instance, to invert the order of succession in the change, were it as easy for the observer to move the ship up the stream as to alter the direction of his own eyes. For Kant infers the successive perception of different parts of the house to be neither objective nor an event, because it depends upon his own will. But the movement of his eyes in the direction from roof to basement is one event, and in the direction from basement to roof another event, just as much as the sailing of the ship. There is no difference whatever here, nor is there any difference either, as to their being or not being events, between my passing a troop of soldiers and their passing me. If we fix our eyes on a ship sailing close by the shore on which we are standing, it soon seems as if it were the ship that stood still and the shore that moved. Now, in this instance we are mistaken, it is true, as to the cause of the relative change of position, since we attribute it to a wrong cause; the real succession in the relative positions of our body towards the ship is nevertheless quite rightly and objectively recognised by us.

Even Kant himself would not have believed that there was any difference, had he borne in mind that his own body was an object among objects, and that the succession in his empirical perceptions depended upon the succession of the impressions received from other objects by his body, and was therefore an objective succession: that is to say, one which takes place among objects directly (if not indirectly) and independently of the will of the Subject, and which may therefore be quite well recognised without any causal connection between the objects acting successively on his body.

Kant says, Time cannot be perceived; therefore no succession of representations can be empirically perceived as objective: i.e. can be distinguished as changes in phenomena from the changes of mere subjective representations. The causal law, being a rule according to which states follow one another, is the only means by which the objectivity of a change can be known. Now, the result of his assertion would be, that no succession in Time could be perceived by us as objective, excepting that of cause and effect, and that every other succession of phenomena we perceive, would only be determined so, and not otherwise, by our own will. In contradiction to all this I must adduce the fact, that it is quite possible for phenomena to follow upon one another without following from one another. Nor is the law of causality by any means prejudiced by this; for it remains certain that each change is the effect of another change, this being firmly established à priori; only each change not only follows upon the single one which is its cause, but upon all the other changes which occur simultaneously with that cause, and with which that cause stands in no causal connection whatever. It is not perceived by me exactly in the regular order of causal succession, but in quite a different order, which is, however, no less objective on that account, and which differs widely from any subjective succession depending on my caprice, such as, for instance, the pictures of my imagination.

The succession, in Time, of events which stand in no causal connection with each other is precisely what we call contingency.[100] Just as I am leaving my house, a tile happens to fall from the roof which strikes me; now, there is no causal connection whatever between my going out and the falling of the tile; yet the order of their succession—that is, that my going out preceded the falling of the tile—is objectively determined in my apprehension, not subjectively by my will, by which that order would otherwise have most likely been inverted. The order in which tones follow each other in a musical composition is likewise objectively determined, not subjectively by me, the listener; yet who would think of asserting that musical tones follow one another according to the law of cause and effect? Even the succession of day and night is undoubtedly known to us as an objective one, but we as certainly do not look upon them as causes and effects of one another; and as to their common cause, the whole world was in error till Copernicus came; yet the correct knowledge of their succession was not in the least disturbed by that error. Hume's hypothesis, by the way, also finds its refutation through this; since the following of day and night upon each other—the most ancient of all successions and the one least liable to exception—has never yet misled anyone into taking them for cause and effect of each other.

Elsewhere Kant asserts, that a representation only shows reality (which, I conclude, means that it is distinguished from a mere mental image) by our recognising its necessary connection with other representations subject to rule (the causal law) and its place in a determined order of the time-relations of our representations. But of how few representations are we able to know the place assigned to them by the law of causality in the chain of causes and effects! Yet we are never embarrassed to distinguish objective from subjective representations: real, from imaginary objects. When asleep, we are unable to make this distinction, for our brain is then isolated from the peripherical nervous system, and thereby from external influences. In our dreams therefore, we take imaginary for real things, and it is only when we awaken: that is, when our nervous sensibility, and through this the outer world, once more comes within our consciousness, that we become aware of our mistake; still, even in our dreams, so long as they last, the causal law holds good, only an impossible material is often substituted for the usual one.

We might almost think that Kant was influenced by Leibnitz in writing the passage we have quoted, however much he differs from him in all the rest of his philosophy; especially if we consider that Leibnitz expresses precisely similar views, when, for instance, he says: "La vérité des choses sensibles ne consiste que dans la liaison des phénomènes, qui doit avoir sa raison, et c'est ce qui les distingue des songes. —— Le vrai Critérion, en matière des objets des sens, est la liaison des phénomènes, qui garantit les vérités de fait, à l'egard des choses sensibles hors de nous."[101]

It is clear that in proving the à priority and the necessity of the causal law by the fact that the objective succession of changes is known to us only by means of that law, and that, in so far, causality is a condition for all experience, Kant fell into a very singular error, and one which is indeed so palpable, that the only way we can account for it is, by supposing him to have become so absorbed in the à priori part of our knowledge, that he lost sight of what would have been evident to anyone else. The only correct demonstration of the à priority of the causal law is given by me in § 21 of the present work. That à priority finds its confirmation every moment in the infallible security with which we expect experience to tally with the causal law: that is to say, in the apodeictic certainty we ascribe to it, a certainty which differs from every other founded on induction—the certainty, for instance, of empirically known laws of Nature—in that we can conceive no exception to the causal law anywhere within the world of experience. We can, for instance, conceive that in an exceptional case the law of gravitation might cease to act, but not that this could happen without a cause.

Kant and Hume have fallen into opposite errors in their proofs. Hume asserts that all consequence is mere sequence; whereas Kant affirms that all sequence must necessarily be consequence. Pure Understanding, it is true, can only conceive consequence (causal result), and is no more able to conceive mere sequence than to conceive the difference between right and left, which, like sequence, is only to be grasped by means of pure Sensibility. Empirical knowledge of the following of events in Time is, indeed, just as possible as empirical knowledge of juxtaposition of things in Space (this Kant denies elsewhere), but the way in which things follow upon one another in general in Time can no more be explained, than the way in which one thing follows from another (as the effect of a cause): the former knowledge is given and conditioned by pure Sensibility; the latter, by pure Understanding. But in asserting that knowledge of the objective succession of phenomena can only be attained by means of the causal law, Kant commits the same error with which he reproaches Leibnitz:[102] that of "intellectualising the forms of Sensibility."—My view of succession is the following one. We derive our knowledge of the bare possibility of succession from the form of Time, which belongs to pure Sensibility. The succession of real objects, whose form is precisely Time, we know empirically, consequently as actual.

But it is through the Understanding alone, by means of Causality, that we gain knowledge of the necessity of a succession of two states: that is, of a change; and even the fact that we are able to conceive the necessity of a succession at all, proves already that the causal law is not known to us empirically, but given us à priori. The Principle of Sufficient Reason is the general expression for the fundamental form of the necessary connection between all our objects, i.e. representations, which lies in the innermost depths of our cognitive faculty: it is the form common to all representations, and the only source of the conception of necessity, which contains absolutely nothing else in it and no other import, than that of the following of the consequence, when its reason has been established. Now, the reason why this principle determines the order of succession in Time in the class of representations we are now investigating, in which it figures as the law of causality, is, that Time is the form of these representations, therefore the necessary connection appears here as the rule of succession. In other forms of the principle of sufficient reason, the necessary connection it always demands will appear under quite different forms from that of Time, therefore not as succession; still it always retains the character of a necessary connection, by which the identity of the principle under all its forms, or rather the unity of the root of all the laws of which that principle is the common expression, reveals itself.

If Kant's assertion were correct, which I dispute, our only way of knowing the reality of succession would be through its necessity; but this would presuppose an Understanding that embraced all the series of causes and effects at once, consequently an omniscient Understanding. Kant has burdened the Understanding with an impossibility, merely in order to have less need of Sensibility.

How can we reconcile Kant's assertion that our only means of knowing the objective reality of succession is by the necessity with which effect follows cause, with his other assertion[103] that succession in Time is our only empirical criterion for determining which of two states is cause, and which effect. Who does not see the most obvious circle here?

If we knew objectiveness of succession through Causality, we should never be able to think it otherwise than as Causality, and then it would be nothing else than Causality. For, if it were anything else, it would have other distinctive signs by which to be recognised; now this is just what Kant denies. Accordingly, if Kant were right, we could not say: "This state is the effect of that one, wherefore it follows it;" for following and being an effect, would be one and the same thing, and this proposition a tautology. Besides, if we do away with all distinction between following upon and following from, we once more yield the point to Hume, who declared all consequence to be mere sequence and therefore denied that distinction likewise.

Kant's proof would, consequently, be reduced to this: that, empirically, we only know actuality of succession; but as besides we recognise necessity of succession in certain series of occurrences, and even know before all experience that every possible occurrence must have a fixed place in some one of these series, the reality and the à priority of the causal law follow as a matter of course, the only correct proof of the latter being the one I have given in § 21 of this work.

Parallel with the Kantian theory: that the causal nexus alone renders objective succession and our knowledge of it possible, there runs another: that coexistence and our knowledge of it are only possible through reciprocity. In the "Critique of Pure Reason" they are presented under the title: "Third Analogy of Experience." Here Kant goes so far as to say that "the co-existence of phenomena, which exercise no reciprocal action on one another, but are separated by a perfectly empty space, could never become an object of possible perception"[104] (which, by the way, would be a proof à priori that there is no empty space between the fixed stars), and that "the light which plays between our eyes and celestial bodies"—an expression conveying surreptitiously the thought, that this starlight not only acts upon our eyes, but is acted upon by them also—"produces an intercommunity between us and them, and proves the co-existence of the latter." Now, even empirically, this last assertion is false; since the sight of a fixed star by no means proves its coexistence simultaneously with its spectator, but, at most, its existence some years, nay even some centuries before. Besides, this second Kantian theory stands and falls with the first, only it is far more easily detected; and the nullity of the whole conception of reciprocity has been shown in § 20.

Not without considerable hesitation did I thus venture (in 1813) to attack a theory which had been universally received as a demonstrated truth, is repeated even now in the latest publications,[107] and forms a chief point in the doctrine of one for whose profound wisdom I have the greatest reverence and admiration; one to whom, indeed, I owe so much, that his spirit might truly say to me, in the words of Homer: