Truth and Knowledge
II. Kant's Basic Epistemological Question
KANT IS GENERALLY CONSIDERED to be the founder of epistemology in the modern sense. However, the history of philosophy before Kant contains a number of investigations which must be considered as more than mere beginnings of such a science. Volkelt points to this in his standard work on epistemology, saying that critical treatments of this science began as early as Locke.  However, discussions which to-day come under the heading of epistemology  can be found as far back as in the philosophy of ancient Greece. Kant then went into every aspect of all the relevant problems, and innumerable thinkers following in his footsteps went over the ground so thoroughly that in their works or in Kant's are to be found repetitions of all earlier attempts to solve these problems. Thus where a factual rather than a historical study of epistemology is concerned, there is no danger of omitting anything important if one considers only the period since the appearance of Kant's Critique of Pure Reason.  All earlier achievements in this field have been repeated since Kant.
Kant's fundamental question concerning epistemology is: How are synthetical judgments a priori possible? Let us consider whether or not this question is free of presuppositions. Kant formulates it because he believes that we can arrive at certain, unconditional knowledge only if we can prove the validity of synthetical judgments a priori. He says:
“In the solution of the above problem is comprehended at the same time the possibility of the use of pure reason in the foundation and construction of all sciences which contain theoretical knowledge a priori of objects.”  “Upon the solution of this problem depends the existence or downfall of the science of metaphysics.” 
Is this problem as Kant formulates it, free of all presuppositions? Not at all, for it says that a system of absolute, certain knowledge can be erected only on a foundation of judgments that are synthetical and acquired independently of all experience. Kant calls a judgment “synthetical” where the concept of the predicate brings to the concept of the subject something which lies completely outside the subject — “although it stands in connection with the subject,”  by contrast, in analytical judgment, the predicate merely expresses something which is already contained (though hidden) in the subject. It would be out of place here to go into the extremely acute objections made by Johannes Rehmke  to this classification of judgments. For our present purpose it will suffice to recognize that we can arrive at true knowledge only through judgments which add one concept to another in such a way that the content of the second was not already contained — at least for us — in the first. If, with Kant, we wish to call this category of judgment synthetical, then it must be agreed that knowledge in the form of judgment can only be attained when the connection between predicate and subject is synthetical in this sense. But the position is different in regard to the second part of Kant's question, which demands that these judgments must be acquired a priori, i.e., independent of all experience. After all, it is conceivable that such judgments might not exist at all. A theory of knowledge must leave open, to begin with, the question of whether we can arrive at a judgment solely by means of experience, or by some other means as well. Indeed, to an unprejudiced mind it must seem that for something to be independent of experience in this way is impossible. For whatever object we are concerned to know, we must become aware of it directly and individually, that is, it must become experience. We acquire mathematical judgment too, only through direct experience of particular single examples. This is the case even if we regard them, with Otto Liebmann  as rooted in a certain faculty of our consciousness. In this case, we must say: This or that proposition must be valid, for, if its truth were denied, consciousness would be denied as well; but we could only grasp its content, as knowledge, through experience in exactly the same way as we experience a process in outer nature. Irrespective of whether the content of such a proposition contains elements which guarantee its absolute validity or whether it is certain for other reasons, the fact remains that we cannot make it our own unless at some stage it becomes experience for us. This is the first objection to Kant's question.
The second consists in the fact that at the beginning of a theoretical investigation of knowledge, one ought not to maintain that no valid and absolute knowledge can be obtained by means of experience. For it is quite conceivable that experience itself could contain some characteristic feature which would guarantee the validity of insight gained by means of it.
Two presuppositions are thus contained in Kant's formulation of the question. One presupposition is that we need other means of gaining knowledge besides experience, and the second is that all knowledge gained through experience is only approximately valid. It does not occur to Kant that these principles need proof, that they are open to doubt. They are prejudices which he simply takes over from dogmatic philosophy and then uses as the basis of his critical investigations. Dogmatic philosophy assumes them to be valid, and simply uses them to arrive at knowledge accordingly; Kant makes the same assumptions and merely inquires under what conditions they are valid. But suppose they are not valid at all? In that case, the edifice of Kant's doctrine has no foundation whatever.
All that Kant brings forward in the five paragraphs preceding his actual formulation of the problem, is an attempt to prove that mathematical judgments are synthetical (an attempt which Robert Zimmermann,  if he does not refute it, at least shows it to be highly questionable). But the two assumptions discussed above are retained as scientific prejudices. In the Critique of Pure Reason  it is said:
“Experience no doubt teaches us that this or that object is constituted in such and such a manner, but not that it could not possibly exist otherwise.” “Experience never exhibits strict and absolute, but only assumed and comparative universality (by induction).”
In Prolegomena  we find it said:
“Firstly, as regards the sources of metaphysical knowledge, the very conception of the latter shows that these cannot be empirical. Its principles (under which not merely its axioms, but also its fundamental conceptions are included) must consequently never be derived from experience, since it is not physical but metaphysical knowledge, i.e., knowledge beyond experience, that is wanted.”
And finally Kant says:
“Before all, be it observed, that proper mathematical propositions are always judgments a priori, and not empirical, because they carry along with them the conception of necessity, which cannot be given by experience. If this be demurred to, it matters not; I will then limit my assertion to pure mathematics, the very conception of which implies that it consists of knowledge altogether non-empirical and a priori.” 
No matter where we open the Critique of Pure Reason we find that all the investigations pursued in it are based on these dogmatic principles. Cohen  and Stadler  attempt to prove that Kant has established the a priori nature of mathematical and purely scientific principles. However, all that the Critique of Pure Reason attempts to show can be summed up as follows: Mathematics and pure natural science are a priori sciences; from this it follows that the form of all experiences must be inherent in the subject itself. Therefore, the only thing left that is empirically given is the material of sensations. This is built up into a system of experiences, the form of which is inherent in the subject. The formal truths of a priori theories have meaning and significance only as principles which regulate the material of sensation; they make experience possible, but do not go further than experience. However, these formal truths are the synthetical judgment a priori, and they must — as condition necessary for experience — extend as far as experience itself. The Critique of Pure Reason does not at all prove that mathematics and pure science are a priori sciences but only establishes their sphere of validity, pre-supposing that their truths are acquired independently of experience. Kant, in fact, avoids discussing the question of proof of the a priori sciences in that he simply excludes that section of mathematics (see conclusion of Kant's last statement quoted above) where even in his own opinion the a priori nature is open to doubt; and he limits himself to that section where he believes proof can be inferred from the concepts alone. Even Johannes Volkelt finds that:
“Kant starts from the positive assumption that a necessary and universal knowledge exists as an actual fact. These presuppositions which Kant never specifically attempted to prove, are so contrary to a proper critical theory of knowledge that one must seriously ask oneself whether the Critique of Pure Reason is valid as critical epistemology.”
Volkelt does find that there are good reasons for answering this question affirmatively, but he adds: “The critical conviction of Kant's theory of knowledge is nevertheless seriously disturbed by this dogmatic assumption.”  It is evident from this that Volkelt, too, finds that the Critique of Pure Reason as a theory of knowledge, is not free of presuppositions.
O. Liebmann, Hölder, Windelband, Ueberweg, Ed. v. Hartmann  and Kuno Fischer,  hold essentially similar views on this point, namely, that Kant bases his whole argument on the assumption that knowledge of pure mathematics and natural science is acquired a priori.
That we acquire knowledge independently of all experience, and that the insight gained from experience is of general value only to a limited extent, can only be conclusions derived from some other investigation. These assertions must definitely be preceded by an examination both of the nature of experience and of knowledge. Examination of experience could lead to the first principle; examination of knowledge, to the second.
In reply to these criticisms of Kant's critique of reason, it could be said that every theory of knowledge must first lead the reader to where the starting point, free of all presuppositions, is to be found. For what we possess as knowledge at any moment in our life is far removed from this point, and we must first be led back to it artificially. In actual fact, it is a necessity for every epistemologist to come to such a purely didactic arrangement concerning the starting point of this science. But this must always be limited merely to showing to what extent the starting point for cognition really is the absolute start; it must be presented in purely self evident, analytical sentences and, unlike Kant's argument, contain no assertions which will influence the content of the subsequent discussion. It is also incumbent on the epistemologist to show that his starting point is really free of all presuppositions. All this, however, has nothing to do with the nature of the starting point itself, but is quite independent of it and makes no assertions about it. Even when he begins to teach mathematics, the teacher must try to convince the pupil that certain truths are to be understood as axioms. But no one would assert that the content of the axioms is made dependent on these preliminary considerations. [In the chapter titled “The Starting Point of Epistemology,” I shall show to what extent my discussion fulfils these conditions.] In exactly the same way the epistemologist must show in his introductory remarks how one can arrive at a starting point free of all presuppositions; yet the actual content of this starting point must be quite independent of these considerations. However, anyone who, like Kant, makes definite, dogmatic assertions at the very outset, is certainly very far from fulfilling these conditions when he introduces his theory of knowledge.