A Commentary to Kant's 'Critique of Pure Reason'

by Norman Kemp Smith

The principle of causality thus conditions consciousness of objective succession, and Hume, in asserting that we are conscious of the succession of events, therefore admits all that need be assumed in order to prove the principle. The reason why Hume failed to recognise this, is that he ignored the distinction between consciousness of the subjective order of our apprehensions and consciousness of the objective sequence of events. Yet that is a distinction upon which his own position rested. For he teaches that determination of causal laws, sufficiently certain to serve the purposes alike of practical life and of natural science, can be obtained through observation of those sequences which remain constant. Such is also the position of all empiricists. They hold that causal relation is discovered by comparison of given sequences. Kant’s contention is that the apprehension of change as change, and therefore ultimately the apprehension even of an arbitrarily determined order of subjective succession,[1190] presupposes, and is only possible through, an application of the category of causality. The primary function of the understanding does not consist in the clarification of our representation of an event, but in making such representation possible at all.[1191] The primary field of exercise for the understanding lies not in the realm of reflective comparison, but in the more fundamental sphere of creative synthesis.[1192] In determining the nature of the given it predetermines the principles to which all reflection upon the given must conform. The discursive activities of scientific reflection are secondary to, and conditioned by, the transcendental processes which generate the experience of ordinary consciousness. Only an experience which conforms to the causal principle can serve as foundation either for the empirical judgments of sense experience, or for that ever-increasing body of scientific knowledge into which their content is progressively translated. The principle of causality is applicable to everything experienced, for the sufficient reason that experience is itself possible only in terms of it. This conclusion finds its most emphatic and adequate statement in the Methodology.

“...through concepts of understanding pure reason establishes secure principles, not however directly from concepts, but always only indirectly through relation of these concepts to something altogether contingent, namely, possible experience. For when such experience (i.e. something as object of possible experience) is presupposed, the principles are apodictically certain, though by themselves (directly) a priori they cannot even be recognised at all. Thus no one can acquire insight into the proposition that everything which happens has its cause, merely from the concepts involved. It is not, therefore, a dogma, although from another point of view, namely, from that of the sole field of its possible employment, i.e. experience, it can be proved with complete apodictic certainty. But though it needs proof, it should be entitled a principle, not a theorem, because it has the peculiar character that it makes possible the very experience which is its own ground of proof, and in this experience must always itself be presupposed.”[1193]

Before making further comment upon Kant’s central argument, it is advisable to consider the varying statements which Kant has given of it. We may take his successive proofs in the order in which they occur in the first edition.

First Proof.[1194]—The argument is developed in terms of Kant’s early doctrine of the transcendental object. The only points specially characteristic of the statement here given of that doctrine consist (a) in the emphasis with which it is asserted that representations can be experienced only in succession to one another, and that they can never stand in the relation of coexistence,[1195] and (b) in the almost complete ignoring of the transcendental object as source or ground of the rule in terms of which the successive representations are organised. (a) This is a point common to the arguments of all three Analogies. In the first and third the problem is how, from representations merely successive, permanence and coexistence can be determined. In the second Analogy the problem is how from representations invariably successive a distinction can be drawn between the subjectively determined order of our apprehensions and the objective sequence of events. Or in other words: how under such conditions we can recognise an order as given, and so as prescribing the order in which it must be apprehended. Or to state the same point in still another manner: how we can distinguish between an arbitrary or reversible order and an imposed or fixed order, and so come to apprehend the subjective order of our apprehensions as in certain cases controlled by, and explicable only through, the objective sequence of events.[1196]

(b) The reason why the transcendental object, as source of the determinate and prescribed order of the given events, falls into the background in this passage is that Kant is concerned only with the general principle or category by means of which the order is apprehended as necessary. That principle has a subjective origin even though the particular sequences of concrete events have by means of that concept to be conceived as inexorably determined by their noumenal conditions.[1197] The principle accounts for the comprehension of the order as objective, and that is the only point with which Kant is here immediately concerned. That the assertion of the subjective origin of the category is not inconsistent with recognition of the imposed order of the given has already been shown above.[1198] Kant’s own illustration, in this section, of the ship sailing down stream shows that he was prepared to assume without question that they are compatible. His argument is, however, obscure, owing to his failure to distinguish between the two senses in which the term ‘rule’ may be employed. The term may signify either the universal and merely formal principle that every event must have a cause, or it may be used to denote the fixed order in which concrete events are presented to sense-perception. The latter order need not represent a series the members of which are causally connected with one another, but only one that is due to causal necessities. Thus the successive positions of a ship sailing down stream are not interrelated as cause and effect, and yet in order to be apprehended as objectively successive must be conceived as causally conditioned. The term ‘rule’ has very different meanings in the two cases. ‘Rule’ in the first sense is of subjective origin. It is formal, and can never be given. It is read into the given. ‘Rule’ in the second sense is given merely, and being due to noumenal conditions constitutes the material element in natural science, the empirical content of some particular causal law. Owing to Kant’s failure explicitly to distinguish between these two very different connotations of the term, such a sentence as the following is ambiguous: “That in appearance which contains the condition of this necessary rule of apprehension is the object.” Kant may mean that the prescribed order of the concrete events is due to the transcendental object; but in that case it is not given as necessary. Necessity, as he constantly insists, is the one thing that can never be given. The sentence is also misleading through its use of the term ‘appearance.’ That term has no legitimate place in a passage inspired by the doctrine of the transcendental object; there can be no such middle term between subjective representations and the thing in itself. As Kant himself states,[1199] appearance defined in terms of that doctrine is “nothing save a complex of representations.”

There is a very essential difference in the view which Kant takes of the causal relation according as he is proceeding upon subjectivist or upon phenomenalist lines. From the one point of view appearances are representations merely, and accordingly are entirely devoid of causal efficacy. They are not causes and effects of one another. They have not the independence or self-persistence necessary for the exercise of dynamical energy or even for the reception of modifications. Being “states of the identical self,” all causal relation, dynamically conceived, must lie solely in their noumenal conditions. Causality reduces to the thought of necessitated (not necessitating) sequence. It is, as Kant has suggested in A 181 = B 224, a mere ‘analogy’ in terms of which we apply the logical relation of ground and consequence[1200] to the interpretation of our subjective representations, and so view them as grounded not in one another but exclusively in the thing in itself. Causality in the strict sense, i.e. dynamical agency, can be looked for only in the noumenal sphere.

Caird, while adopting this explanation of the term ‘analogy,’[1201] is, as might be expected from his Hegelian standpoint, extremely indefinite and non-committal as to whether or not empirical objects can be genuine causes. Riehl, notwithstanding his professedly realistic interpretation of Kant, adopts the above subjectivist view of natural causation. So also do Benno Erdmann and Paulsen. The latter[1202] speaks with no uncertain voice.

“Causality in the phenomenal world signifies for Kant, as for Hume, nothing but regularity in the sequence of phenomena. Real causal efficiency cannot of course occur here, for phenomena are ideational products. As such they can no more produce an effect than concepts can.”

The corresponding phenomenalist view of the causal relation receives no quite definite formulation either in this section or elsewhere in the Critique, but may be gathered from the general trend of Kant’s phenomenalist teaching.[1203] It is somewhat as follows. The term ‘analogy’ is viewed as having a meaning very different from that above suggested. The causal relation is not a mere analogy from the logical relation of ground and consequence; it is the representation of genuinely dynamical activities in the objects apprehended. Those objects are not mere states of the self, subjective representations. They are part of an independent order which in the form known to us is a phenomenalist transcript of a deeper reality. If the causal relation is the analogy of anything distinguishable from itself, it is an analogon or interpretation of dynamical powers exercised by things in themselves,[1204] not of the merely logical relation between premisses and conclusion. The objects of representation may exercise powers which representations as such can never be conceived as possessing. Between the individual’s subjective states and things in themselves stands the phenomenal world of the natural sciences. Its function, whether as directly experienced through sense-perception or as conceptually reconstructed through scientific hypothesis, is to stand as the representative in human consciousness of that noumenal realm in which all existence is ultimately rooted. The causal interactions of material bodies in space are as essentially constitutive of those bodies as are any of their quantitative properties. Causal relation, even in the phenomenal sphere, must not be identified with mere conformity to law. The true and complete purpose of the natural sciences is not to be found in the Berkeleian or sceptical ideal of simplification, but in the older and sounder conception of causal explanation. That, at least, is the view which Kant invariably defends whenever he has occasion to discuss the principles of physical science.

Second Proof.[1205]—The argument of the first proof is here developed in indirect fashion. In the absence of any rule prescribing necessary sequence, no distinction can be made between subjective and objective succession. The justification for such a rule lies therefore, not in an inductive inference from repeated experience, but in its necessity for the possibility of experience. It is an expression of the synthetic unity in which experience consists.