Summary Research consists of choosing a problem, proposing and testing problem solutions, and presenting the results. In its central moment — conjectures and testing — science must be autonomous in order to be successful. Securing this autonomy by organizational means is one of the tasks of research policy. Research needs to justify itself only when the researcher makes a claim to the resources of others. To discuss problems of justification of governmental support, it is imperative to distinguish between basic and (...) applied research. In basic research the problem choice is guided only by considerations of scientific interest , while in applied research the problems emanate from extra-scientific concerns since applied research is by definition a means for tackling concrete practical problems. Deciding on external criteria (e.g. deciding whether to support more energy research or more cancer research) is a genuine political problem rather than a problem of research policy. Thus applied research can be justified by referring to the benefit the expected results will yield, but basic research requires a completely different justification. One such justification is the argument that applied research requires a certain overhead in basic research. Securing an adequate balance between funding basic and applied research within the problem area concerned is another task of research policy. (E.g. in the area of cancer whether to support more molecular biology research or more clinical research.) When the overhead argument is not applicable, there still remain several possible justifications for basic research. These are examined in the paper. When setting priorities for basic research within a certain discipline, one apparently must turn to the scientific community itself. It may well be that the researches, who constitute the only expertise available for this task, have to rely on tacit knowledge . If so, this remains afaute de mieux procedure since articulated criteria would be preferable to intuitive procedures. The purpose of methodological reflection on research policy making is to supply intellectual instruments for making the discussion about substantive problems more rational. The substantive problems can be tackled only through the close cooperation of research policy makers and researchers. Thus methodology should not limit the degrees of freedom of either but increase them. (shrink)