Einstein told Heisenberg that only theory decides what can be measured. In theory there are the concepts which decide the meaning of measurements. They always come with preconceptions, in terms of which they must be initially formulated. In the sensible efforts which are being made to develop the contemporaneous methodology of the natural sciences the following principle can be perceived as the direction for activities in this domain: through logic to physics and to biology. It can be expressed in other words i.e. in the statement that the conception of performing physical measurements must be preceded by taking into consideration and theoretically analyze the concepts of the respective physical phenomena and the definitions of the respective physical quantities. Shortly, theory precedes measuring and suggests what can be measured. Therefore, Kant’s undertaking of investigating the metaphysical foundations of the natural sciences is not obsolete, as the empiricist philosophy of science would have stated. Moreover, the empiricism of the twentieth century did not succeed in its efforts to establish the demarcation between empirical science and metaphysics.

The philosophy of the natural sciences of contemporaneous century can be considered to be intentionally retrospective philosophy. A purpose of my paper is exploring and considering the genesis of theories of the natural sciences. The question on the foundations of the knowledge of the natural sciences remains of great meaning. The exact sciences are fascinated by reality and they rest on that knowing all about reality from experience. In the phases of normal science philosophy is not necessary for the factual progress, even impedes that. The seminal question how scientific revolutions are possible is better understood in the context of comparison with biological evolution. According to Konrad Lorenz, there occur fulgurations, flash-like mergers of other structures into a more complex new structure which accomplishes a new simplicity of performance at a higher level of integration, the new level from the previously mentioned structures being unpredictable. Therefore, from the philosophical point of view most important in science are not the always and everywhere applicable methods of science but its unique factual problems and results. At first sight we suppose that a closed theory will appear to be both the more amenable to a fundamental explanation and the more generally valid the later it occurs in the succession of theories. In particular, this may lead us to quantum theory. In this case we suspect that nowadays the best criterion for acceptance of a philosophy of science is whether it can make quantum theory understandable. The above question on the possibility of constructing closed theories has the same epistemological structure as Kant’s question: how is experience possible at all? As regards content, in accordance with our understanding of theory; both questions are synonymous. Experience in the Kant's sense can be expressed conceptually. We have learned above that scientific concepts obtain a precise meaning only in the context of a closed theory.