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

by Norman Kemp Smith

As has already been stated, the wider problem concerning the principle of causality is developed only in the Treatise; the problem regarding the concept of causality is discussed both in the Treatise and in the Enquiry. An appreciation of the wider problem is required, however, in order to set this second problem in its true light, for it is only through its connection with the wider issue that Hume’s reduction of the concept of causality to a merely instinctive, non-rational expectation acquires its full significance. Hume’s analysis then amounts, as Kant was the first to realise, to an attack upon the objective validity of all constructive thinking. Not only rationalism, but even such metaphysics as may claim to base its conclusions upon the teaching of experience, is thereby rendered altogether impossible. The issue is crucial, and must be honestly faced, before metaphysical conclusions, no matter what their specific character may be, whether a priori or empirical, can legitimately be drawn. If we may not assert that an event must have some cause, even the right to enquire for a cause must first be justified. And if so fundamental a principle as that of causality is not self-evident, are there any principles which can make this claim?

The account which we have so far given of Hume’s argument covers only that part of it which is directed against the rationalist position, and which was therefore so influential in turning Kant on to the line of his Critical speculations. But Hume attacked with equal vigour the empiricist standpoint; and as this aspect of his teaching, constituting as it did an integral part of Kant’s own philosophy, must undoubtedly have helped to confirm Kant in his early rationalist convictions, we may profitably dwell upon it at some length. In opposition to the empiricists, Hume argues that experience is incapable of justifying any inference in regard to matters of fact. It cannot serve as a basis from which we can inductively extend our knowledge of facts beyond what the senses and memory reveal. Inductive inference, when so employed, necessarily involves a petitio principii; we assume the very point we profess to have proved.

The argument by which Hume establishes this important contention is as follows. All inductive reasoning from experience presupposes the validity of belief in causal connection. For when we have no knowledge of causes, we have no justification for asserting the continuance of uniformities. Now it has been shown that we have no experience of any necessary relation between so-called causes and their effects. The most that experience can supply are sequences which repeat themselves. In regarding the sequences as causal, and so as universally constant, we make an assertion for which experience gives no support, and to which no amount of repeated experience, recalled in memory, can add one jot of real evidence. To argue that because the sequences have remained constant in a great number of repeated experiences, they are therefore more likely to remain constant, is to assume that constancy in the past is a ground for inferring it in the future; and that is the very point which demands proof. In drawing the conclusion we virtually assume that there is a necessary connection, i.e. an absolutely constant relation, between events. But since no single experience of causal sequence affords ground for inferring that the sequence will continue in the future, no number of repeated experiences, recalled in memory, can contribute to the strengthening of the inference. It is meaningless to talk even of likelihood or probability. The fact that the sun has without a single known exception arisen each day in the past does not (if we accept the argument disproving all knowledge of necessary connection) constitute proof that it will rise to-morrow.

“None but a fool or a madman will be unaffected in his expectations or natural beliefs by this constancy, but he is no philosopher who accepts this as in the nature of evidence.”[1781]

Since, for all that we know to the contrary, bodies may change their nature and mode of action at any moment, it is vain to pretend that we are scientifically assured of the future because of the past.

“My practice, you say, refutes my doubts.[1782] But you mistake the purport of my question. As an agent, I am quite satisfied in the point; but as a philosopher, who has some share of curiosity, I will not say scepticism, I want to learn the foundation of this inference. No reading, no enquiry has yet been able to remove my difficulty or give me satisfaction in a matter of such importance. Can I do better than propose the difficulty to the public, even though, perhaps, I have small hopes of obtaining a solution? We shall at least, by this means, be sensible of our ignorance, if we do not augment our knowledge.”[1783]

Kant was the first, after thirty years, to take up this challenge. Experience is no source of evidence until the causal postulate has been independently proved. Only if the principle of causality can be established prior to all specific experience, only if we can predetermine experience as necessarily conforming to it, are empirical arguments valid at all. Hume’s enquiry thus directly leads to the later, no less than to the earlier form of Kant’s epoch-making question.[1784] In its earlier formulation it referred only to a priori judgments; in its wider application it was found to arise with equal cogency in connection with empirical judgments. And as thus extended, it generated the problem: How is sense-experience, regarded as a form of knowledge, possible at all?[1785] By showing that the principle of causality has neither intuitive nor demonstrative validity, Hume cuts the ground from under the rationalists; by showing that sense-experience cannot by itself yield conclusions which are objectively valid, he at the same time destroys the empiricist position. In this latter contention Kant stands in complete agreement with Hume. That the sensuously given is incapable of grounding even probable inferences, is a fundamental presupposition (never discussed, but always explicitly assumed) of the Critical philosophy. It was by challenging the sufficiency of Hume’s other line of argument, that which is directed against the rationalists, that Kant discovered a way of escape from the sceptical dilemma. The conditions of experience can be proved by a transcendental method, which, though a priori in character, does not lie open to Hume’s sceptical objections. Each single experience involves rational principles, and consequently even a single empirical observation may suffice to justify an inductive inference. Experience conforms to the demands of pure a priori thought; and can legitimately be construed in accordance with them.

We may now pass to the philosophy in which Kant was educated. It gave to his thinking that rationalist trend, to which, in spite of all counter-influences, he never ceased to remain true.[1786] It also contributed to his philosophy several of its constructive principles. Only two rationalist systems need be considered, those of Leibniz and of Wolff. Kant, by his own admission,[1787] had been baffled in his attempts (probably not very persevering) to master Spinoza’s philosophy. It was with Wolff’s system that he was most familiar; but both directly and indirectly, both in his early years and in the ’seventies, the incomparably deeper teaching of Leibniz must have exercised upon him a profoundly formative influence. In defining the points of agreement and of difference between Hume and Leibniz,[1788] we have already outlined Leibniz’s general view of the nature and powers of pure thought, and may therefore at once proceed to the relevant detail of his main tenets.

Upon two fundamental points Leibniz stands in opposition to Spinoza. He seeks to maintain the reality of the contingent or accidental. These terms are indeed, as he conceives them, synonymous with the actual. Necessity rules only in the sphere of the possible. Contingency or freedom is the differentiating characteristic of the real. This point of view is bound up with his second contention, namely, that the real is a kingdom of ends. It is through divine choice of the best among the possible worlds that the actual present order has arisen. There are thus two principles which determine the real: the principle of contradiction which legislates with absolute universality, and the principle of the best, or, otherwise formulated, of sufficient reason, which differentiates reality from truth, limiting thought, in order that, without violating logic, it may freely satisfy the moral needs. Leibniz thus vindicates against Spinoza the reality of freedom and the existence of ends.

Though Leibniz agrees with Spinoza that the philosophically perfect method would be to start from an adequate concept of the Divine Being, and to deduce from His attributes the whole nature of finite reality, he regards our concept of God as being too imperfect to allow of such procedure. We are compelled to resort to experience, and by analysis to search out the various concepts which it involves. By the study of these concepts and their interrelations, we determine, in obedience to the law of contradiction, the nature of the possible. The real, in contradistinction from the possible, involves, however, the notion of ends. The existence of these ends can never be determined by logical, but only by moral considerations. The chief problem of philosophical method is, therefore, to discover the exact relation in which the logical and the teleological, the necessary and the contingent, stand to one another.