some wider principle, the series of premises necessarily resting upon some ultimate proposition or propositions, for whose truth no other ground could exist beyond their own immediate clearness. This conception of the nature of validity and of rational procedure, as we have already pointed out, was made inevitable for rationalism by the representative theory of ideas. Just because the truth of ideas consisted in their correspondence to the reality which they represented, there could, in the last resort, be no test of truth except intuition.

Now, so far as his ideal of scientific procedure went, Kant was a thoroughgoing rationalist. He was not as he remarks in the preface to the second edition of the Critique of Pure Reason— opposed to the dogmatical procedure of reason, since science must always derive its proofs from pure principles a priori. It is only necessary to inquire in what way, and by what right, reason has become possessed of such principles. Mathematics typified in his mind, as in Descartes's, the ideal of scientific method; and this ideal was further confirmed by the recent development of mathematical physics. The fact of the existence of a body of a priori judgments he assumed without question. Profoundly as Kant was stirred by the analysis of Hume, Hume's scepticism left him untouched. Human knowledge is so he believed—unassailably founded on universal and necessary

truths.

As to the character of these truths, however, Kant had become convinced that the a priori premises on which the sciences are founded must be synthetic propositions. It will be remembered that this was a question on which rationalists had not been agreed. Descartes, to whom the issue had not clearly presented itself, admitted both analytic and synthetic propositions among intuitive truths. Spinoza, too, had included synthetic propositions as axioms among the first principles of his system. Hobbes and Leibniz, however, had, for differing reasons, made the attempt to base deductive science solely on definitions. Now it was evident to Kant, that this latter procedure was impossible. From

analytic propositions alone no new truths could be deduced. They can, as Kant remarks, serve only "to form the chain of the method, and not as principles." Furthermore, not only did he recognize that metaphysics and natural science contained synthetic principles, but he was equally convinced that geometry and even arithmetic were based upon such principles. Now the mathematical sciences were the only ones to which Hume had allowed demonstrative certainty, as being based upon the direct comparison of ideas—all judgments involving the notion of cause being only of various degrees of probability or 'moral' certainty. But Kant found that Hume's criticism of the causal relation turned upon its synthetic character; so that, although Hume himself had never formulated the distinction between analytic and synthetic judgments—and, indeed, the distinction is wholly foreign to his thought-his criticism needed only to be generalized in order to apply with equal cogency to the principles of mathematics.1 It was in this way that Kant's reading of Hume reacted so sharply upon his inbred rationalism. It brought into relief the fundamental difficulty of rationalism and empiricism alike: What warrant can exist for universal relations between terms essentially disparate?

It was, then, with a clear recognition of this difficulty, that Kant was led to formulate the problem, How are synthetic judgments a priori possible? And yet, despite this insight, he failed to realize that a solution of the problem must involve a transformation of the whole scheme of rationalistic logic. His purpose was not to destroy but to fulfill rationalism.

The solution which Kant believed himself to have discovered lies in the fact, that a priori principles are the indispensable conditions of the unity of experience. They are a priori, i. e., immediately certain and logically independent of all other knowl

1As an indication of Kant's rationalistic bias, we may cite the remark (in the Introduction to the Critique of Pure Reason, 2d ed.), that if Hume had thus generalized his criticism, his good sense must have forced him to reject both the logical consequences of the criticism and the fundamental premises from which they were drawn.

edge, precisely because they are involved in every possible bit of empirical knowledge. Their necessity lies in the indispensability of the function which they perform in experience. If it is their universality which serves as the basis of all valid knowledge, they themselves are reciprocally justified by the whole system of experience. All this implies unmistakably an important limitation upon the dogmatic conception of irreversible logical priority. This appears in the fact, that, as a consequence of the Kantian treatment of the a priori as the form of thought, its legitimate application must be restricted within the limits of possible experience. That is to say, a priori principles are not true in that they severally and independently correspond to reality, else a limitation upon them would be unthinkable. A type of truth thus emerges in the critical philosophy, which is not conceived as a relation between the world of thoughts, on the one hand, and a world of reality, on the other. This new truth is a concept which, like any of the categories, is itself applicable only within experience. Moreover, the truth of the a priori principles is no longer a matter of conformity to objects, either phenomenal or noumenal. Kant himself expresses this in his suggestion, that, instead of assuming as had previously been done, that our cognition must conform to objects, we make the assumption, that objects must conform to our mode of cognition. On the other hand, if the objects of empirical experience are determined only by conformity to the laws of our intelligence, the a priori principles of experience become knowledge only by application to those objects. The correspondence between concept and object which thus results is, therefore, a secondary matter, rather the consequence than the ground of the truth of the a priori principles. What does at once determine and constitute their truth is precisely the function they perform. Considered apart from this function, indeed, they are not true, for they are not knowledge at all, but mere "cobwebs of the brain," as Kant calls them.

That Kant did not realize the full significance of the changes

he had wrought in the logic of rationalism, or make consistent use of his own new conceptions, we shall attempt to make clear as we proceed. Here we need only call attention to the fact, that he insists upon preserving the traditional rationalistic idea of truth alongside of the revolutionary one, though only as a rubric beneath which all is blank—an ideal unattainable by human intelligence. Knowledge of the thing-in-itself, if such knowledge there were, could alone exemplify this truth. Here alone could be found an object absolutely independent of the ideas which refer to it, and to which, as their eternal standard, they mustif they are to be true-submissively conform. And, on the other hand, it is evident, that only in truth of the representative type could the thing-in-itself be revealed to us. For this is an object which lies outside of human experience, and hence can be present to it only by representatives. No such relation as obtains between the phenomenal object and its idea can obtain here. But the impossibility of any representation is equally evident. Any ascertainable resemblance is out of the question. And one cannot postulate an identity of any of the relations between ideas and those which make up the structure of the thing-in-itself; for all the former are limited in their legitimate application to the phenomenal world, while the thing-in-itself may be structureless. For the same reason a theory of secondary qualities is ruled out, even the conceptions of unity and multiplicity having no warranted application beyond experience. Thus the thing-in-itself remains unknowable, and the traditional conception of truth is without exemplification.

When we pass on to consider the conception of reality in Kant, we are at first struck by the apparent fact, that the new conception of truth which he has introduced has not had any effect here at all. To be sure, just as he distinguishes between two possible orders of knowledge (one of which we lack), so he distinguishes between two kinds of reality, the reality of the thingin-itself and that of the phenomenon. The former is the self

subsistence of an orthodox rationalistic substance, with, to be sure, the important defect, that it is unknowable; and analogy would lead us to expect that the latter would represent the critical standpoint. But such does not at once appear to be the case. The reality of the phenomenon, as Kant treats it, is rather suggestive of empiricism. In his own phrase, it is "that in the phenomenon which corresponds to the sensation." When one speaks of the reality of anything which is not at the moment perceived, that can only mean that it is connected, by means of the analogies of experience, with what is so perceived, so that it coheres with it in a single larger whole. The absent phenomenon thus owes its reality to the present phenomenon-a singular and most instructive parallel to Hume's doctrine of belief.

But when we pause to reflect upon the nature of the coherence with present reality, which gives reality to the absent, we see that here too the critical theory has worked its transformation. This coherence is not a mere association reducible to the contiguity of mutually independent elements. It is the organization of experience under categories. To put the matter differently, the older notion of reality has developed for Kant into two intimately united, but nevertheless formally distinct, factors,— reality, in the sense above defined, and objectivity. When, therefore, we would rightly estimate the significance of Kant's realitas phænomenon, we must recall that only an object can be thus real; and that an object is an object only by reason of its internal (and external) organization. We must, then, even add that it is not simply the absent phenomenon which owes its reality to the work of thought, but the present phenomenon as well, since it is only as an object that it could be real. That is to say, apart from the thought-activity, nothing would be present save an utterly meaningless image, to which the attribute of reality would have no application whatsoever.

We have already seen that Kant's critical problem, How are synthetic judgments a priori possible? arose through his recognition