It is a commonplace of nineteenth-century German intellectual history that with the collapse of post-Kantian Idealism, philosophy ceded its claim of scientificity to the positive sciences.[31] The special sciences made rapid and startling progress, continually adding to the store of human knowledge. Philosophy, meanwhile, seemed to have lost its way, and many thinkers pinned their hopes for a rehabilitation of philosophy on a return to Kant. The Marburg School in particular interpreted this to mean that philosophy should orient itself with respect to the sciences, rather than strive—as German Idealism had—to forge a scientific system of its own, independent from the results of the positive sciences: only in this attenuated sense would philosophy be “scientific.” For all that, Natorp by no means conceived philosophy as a humble handmaiden. On the contrary, its task is to discover and establish the highest principle(s) of rational understanding, and thereby the principles not only of the sciences, but also of ethics and aesthetics, in short, of all the domains of human culture.[32] It must however take science as its primary object of inquiry because science represents the paradigm of knowing (Erkenntnis). Only a critique of science can therefore elucidate rationality or, as the Marburgers call it, the “logic” of thinking, for it is only in science that we can most reliably witness thinking at work, successfully achieving knowledge. Thus the question of the “concept of science” becomes “the chief question of logic and the foundational question of philosophy” (Cohen 1902: 445). The Marburgers identify the unifying principle of science and ethics in particular[33] as the concept of law (Gesetz), and, as Cohen puts it, “it is the business of logic to determine the meaning of law, or rather, the meanings of law” (Cohen 1902: 445).

Natorp and Cohen find general affirmation of this train of thought in Kant's project of seeking the conditions of possibility of the “fact” of mathematical natural science.[34] However, the similarities end there, for by rejecting or modifying several basic aspects of Kant's philosophy, they also end up with a radically transformed conception of the nature of scientific experience and the meaning of knowledge. The first of their modifications stems from an anti-psychologistic critique of Kant himself, namely of what they see as a confusion in the first Critique between the task of a transcendental grounding of the sciences and that of a transcendental logic of human cognition.[35] The former is in their view the genuine critical enterprise, for it promises to reveal the autonomous sources of objective knowledge, whereas the latter threatens to trace science back to psychological, and therefore contingent, subjective (albeit a priori) wellsprings.[36] Second, they deny any scientific role to intuition as conceived by Kant, either pure or empirical. Partly this is a result of their anti-psychologism, which forbids them from grounding the objectivity of science in the subjective faculties of cognition; but it is also because they see, with Kant, the essence of thinking in its activity and spontaneity, whereas intuition (at least as defined by Kant) is passive and affective.[37] Hence, intuition thus conceived threatens to introduce a heteronomous, and therefore rationally unacceptable, factor into science. Finally, the Marburgers follow their German Idealist predecessors in dismissing all talk of things in themselves, conceived as things existing independently of knowledge. We can see how these three important modifications of Kant's philosophy stem from the same basic concern with rational autonomy. For reason to be autonomous, its activity must be spontaneous; but this spontaneity cannot be conceived of psychologically, because human cognition as a matter of fact has a passive, and therefore heteronomous, intuitive element, namely sensibility. Furthermore, things in themselves can play no explanatory role here because they are ex hypothesi alien to reason.

These modifications have two radical consequences for Kantian doctrine, consequences that characterize the Marburgers' own theory of science and cognition. The first is a new conception of science; the second is a new conception of the categories (see Section 4). It might seem that science, as the achievement of an autonomous rationality, must fail to be objectively true of the world, if reason's autonomy implies that it can have no intuitive, receptive link to the world via sensibility. How in general could the rationally constructed system be related to the constraints of experience? How in particular could physics, the science of motion in space and time, be possible if the pure forms of intuition, space and time, were banished from science?[38]

If Natorp often seems to embrace the troubling thesis that science is not of the phenomenal world, this is because he holds, first, that the meaning of “phenomenon” is problematic; and second, that the aspect of science relevant to philosophy has nothing to do with its relation to a phenomenal realm. In this he follows Cohen's dictum:

Not the stars in the heavens are the objects which [the transcendental] method teaches us to contemplate in order to know them; rather, it is the astronomical calculations, those facts of scientific reality which are the “actuality” that needs to be explained…. What is the foundation of the reality which is given in such facts? What are the conditions of that certainty from which visible actuality takes its reality? The laws are the facts, and [hence] the objects [of our investigation]; not the star-things. (Cohen 1877: 27, f.)[39]

The point is that the scientific or epistemic value of, say, astronomy, is not to be found in what is given and observable by the senses, but rather in the mathematical exactness of its equations. These alone constitute and underwrite the truth-value of astronomy's propositions, and they are solely the achievements of reason's activity. As noted above, the essential characteristic of science lies in its objectivity, and that objectivity is rooted in its lawfulness. It is this formal feature of objectivity that constitutes the philosophical interest in science, not the material content of a particular science's theorems; in other words, the philosophical question is: “How is this lawfulness possible?” This question is distinct from the psychological question, “What are the psychological laws that make it possible for me (as a psychophysical being) to observe a star?” or the astronomical question, “What are the laws governing the ’being’ of this star in its states and properties?”

Hence, it is not so much the case that science on the Marburgers' conception loses all traction on the phenomenal or “actual” world, as that they are asking an entirely different question. While for Kant himself such traction is the only warrant that we are cognizing a genuine object, for Natorp the nexus of science and apparent reality is irrelevant to the spontaneous, legislating factor of science that is the activity of reason alone and therefore of paramount interest to philosophy. How such essentially subjective application of categories to sensible phenomena in fact happens is a problem of psychology, not philosophy, to investigate. Thus we must be very careful in interpreting the Marburg talk of “scientific experience.” Yes, it is not experience in general or psychological (subjective) experience (Erlebnis), but scientific (objective) experience (Erfahrung) which is the “fact” whose transcendental sources philosophy is to seek.[40] Yet we must not in turn take this scientific “experience” to mean “experience cleansed by experiment” (to paraphrase Helmholtz[41]): experiment by definition obviously remains empirical. Rather, by “scientific experience” Natorp just means the “legislative” act of categorial “Grundlegung,” of “hypothesis.”[42]

Now by the end of the nineteenth century, it was obvious to any informed observer of science that its categorial structures were in fact hypothetical and dynamic: the fact of scientific experience could no longer be taken as the essentially complete edifice of Newtonian physics, as Kant had done. In Natorp's rewording of Kant, science is not a factum at all, but a fieri, i.e., not an accomplished deed, but an ongoing doing.[43] Hence, what makes science scientific—i.e., productive of genuine knowledge—cannot possibly be founded on a set of fixed (physical) principles, analogous to mathematical axioms the certainty of which somehow flows into its theorems. Instead, the Marburgers argue, its scientificity can only reside in its method, i.e., in the regular and regulated manner of its progress. And since its scientificity is equivalent to objectivity or lawfulness, transcendental critique must determine the relation of lawfulness to method.

http://plato.stanford.edu/entries/natorp/#2