The results achieved by science are never more than approximations in the sense that the units, the bricks with which the house is built, are liable to be rejected, and the constructions achieved are subject to revision.

The point however which I wish to emphasize is that the scientist is satisfied of the truth, the reality of its partial results. Newton, for instance, in formulating the law of gravitation has, so to speak, marked off a strip of reality. The ground covered cannot be lost; when some natural law is enunciated, the proper conditions for its discovery and verification having been observed, a sure footing in reality has been gained, science standing to this extent on terra firma, though beyond the domain within which the law applies the phenomena may be heaving and billowing like the sea.

Now the question I am intent upon is, Why is it possible for science to be content with partial acquisition? Why does it profess to know positively a part without knowing the whole? And why can ethics not take a step without an ideal of the whole?

Kant’s chief purpose in the Critique of Pure Reason was to vindicate the certainty of the physical knowledge of a part as being compatible with total ignorance of the whole. The older metaphysics was engaged in the attempt to supply the whole, to sketch it out in order to give certainty to the part that is within the reach of science. The older metaphysics said to science: You have in hand the conditioned, but remember the conditioned depends on the unconditioned. Unless, therefore, you round out what you possess, with the help of the unconditioned, the certainty you seem to have within the field of the conditioned disappears. Again, science traces causes, and the older metaphysicians insisted that the whole chain of causes hangs in air unless it be attached to a first cause. Now Kant’s Critique of Pure Reason really amounts in nuce to this: you do not require the whole in order to explain the part. Link the partial phenomena together in a certain way, a way dependent on the joint action of the space and time intuitions and the categories, and you will gain the desired certainty. The certainty is in the linkage. We may add link to link of the chain of reality without troubling to consider by what piers it is supported or on what shore the piers rest—if indeed there be piers and shores at all. The bridge hangs over the River of Time and we can safely travel on it. How we get on to this bridge we do not know, and where we shall leave it we cannot know either.

It is a mistake to speak of Kant as a rationalist pure and simple. When he expelled the older metaphysics he antagonized pure rationalism. The older metaphysics held that the mere existence of the conditioned proves the existence of the unconditioned, requires the unconditioned. In Kant’s answer to this lies the gist of his enterprise in philosophy: You are quite right, he says, that the idea of the conditioned requires the idea of the unconditioned, logically, rationally. But observe well, nature is not just logical or rational. There is an irrational element in it, namely, extended manifold and temporal sequence. Juxtaposition and sequence are irrational, because, if I interpret him rightly, in the case of each the relation presented to the mind is that of parts outside each other—in the one case alongside, in the other before and after; while in the logical or rational relation the parts are implicit in the whole as in the case of the premises of a syllogism and the conclusion, the relation of a genus to the species, the universal to the particular.

We have in nature, according to Kant, a partnership between the irrational and the rational factors. And thereupon he proceeds to argue that we impose laws on nature, understanding thereby that we get hold of reality or objectivity in so far as we are able to imprint the rational element upon the irrational. The positing of the thing per se, which has proved a stumbling-block to many, is no more than a confession that we shall never succeed entirely in this business of subjecting the irrational to the rational factor. The thing per se is the X that remains over when the rational function has done its utmost. A thing, a real object, is that which is imprinted with, penetrated with, rationality. The manifolds of space and time, of juxtaposition and sequence are incapable of completely receiving this imprint, that is, of completely responding to our quest for reality, and this their incompetency is expressed in the notion of the thing per se.

To return to the main question as to the difference between the method by which science proceeds and the reverse method prescribed to ethics, I ask, Why is absolute knowledge of nature impossible? The answer is, Because absolute knowledge would mean the completely rational construction of nature, and this is prevented by the irrational element existing in it. But why has the relative knowledge we possess the character of certainty? Why are we sure of the law of gravitation? Why are we justified in saying that science within certain limits plants her foot on terra solida? Because at certain points the sense data do coincide with the rational requirements. There are recurrent phenomena of such a kind, coupled together in such a way, that each is capable of mathematical measurement, and that the sequence of the one after the other can therefore be predicted.

Nature might have been arranged quite otherwise. The time spans might have been so long, as to prevent our observing the recurrences. A day-fly cannot observe the periodicity of the earth’s revolution around its axis. The fact however that there is this partial correspondence between human rationality and the unknown nature of things is a bare fact, incapable of explanation.[24] The answer, then, I take it, is: our knowledge of nature is relative, which means incompletely rational, because of the foreign element in nature unamenable to the operation of the rational, the synthetic, function. This relative knowledge is none the less certain, that is, in some sense absolute, because of the partial coincidence of the phenomena of nature and the synthetic processes of the mind.

With this degree of certainty we must perforce content ourselves, in dealing with outside nature. In trying to understand and interpret that which is not ourselves, we hit upon barriers which cannot be transcended, upon a foreign factor which opposes itself to our endeavors. But it is otherwise in the sphere of conduct. Here, if there is to be certainty at all, in regard to right as distinguished from wrong, if there is to be such a thing as right in the strict sense, we cannot content ourselves with the paradoxical, relative-absolute just described. For here we not merely interpret but act, and we must possess an ideal plan of the whole if we are to be certain of our rightness in any particular part of conduct. For in conduct there is no such partial coincidence between the rational and the irrational as in the case of physical law. There is not a single partial rule of conduct, neither “Thou shalt not kill” nor “Thou shalt not lie,” nor any other that, taken by itself, is of itself ethically right. It may be right, it may be wrong. It takes its ethical quality from the plan of conduct as a whole, and without reference to the whole it is devoid of rightness.[25]

I have thus indicated the ground of the distinction between the method of science and the method of ethics, a distinction, it is true, to which Kant himself did not adhere. Partial coincidence of the rational with the irrational is expressed in physical law; absence of such concurrence destroys any attempt to build up an ethical theory on the empirical method. We cannot plant our feet on the part, gaining there the sense of certainty: we must creatively conceive the ideal of the whole and educe every partial mode of ethical conduct from that.