<p>The idea here is to derive properties and categories of knowledge and technology transfer from a theoretical model of the science system, for which we draw on the work of the German sociologists Niklas Luhmann and Rudolf Stichweh. The aim is to determine the characteristics of transfer that are already defined by the autopoiesis of the science system. We will explain why transfer can be understood as the collection of input–output processes across the boundaries of the academic system and can be associated with the science system’s exchange of services with other subsystems of society, but is independent of the fulfilment of the science system’s function. With this distinction between “function” and “service,” we can address two apparent contradictions: How is the paradigm of openness that determines transfer compatible with the closedness that underlies the autopoiesis of science, and how can transfer across system boundaries and multipolar innovation networks be brought together conceptually? This paper distinguishes between three process classes of transfer. While the first two process classes precede or follow a real research process (inside-out and outside-in), the third process class is of the form out → in → out and follows the scheme that a solution is sought for a specific problem that serves external purposes. Then, ten different categories of knowledge and technology transfer are compiled and discussed for the three process classes derived. They show how complex transfer structures can be constructed from elementary processes without having to sacrifice the idea of a system boundary, which is indispensable for the functioning of science, which casts doubt on the assertion often made in the literature that this boundary is dissolving (“blurring of boundaries”). In the end, we also address other important practical and theoretical implications of these considerations.</p>