Abstract
The concept of the hypothesis is omnipresent across all discipline of science, be it natural sciences or humanities. Despite the prominent role in everyday scientific work, its true meaning and logical position are remarkably elusive. In this essay, I argue that within the concept of the hypothesis, a twofold collision of categories occurs. First, the hypothesis needs to be connected to current, already achieved knowledge. At the same time, it may not be directly deduced from this knowledge: Therefore, its formulation has a necessary non-rational aspect. Second, in this dichotomic state, the hypothesis is in the moment of its formulation immediately revoked because the hypothesis is not meant to be permanent. Either, its connection to present knowledge survives if the hypothesis is later shown to be true. Or, its non-rational aspect survives if the hypothesis is disproven. Both these cases are thought as potential outcomes when the hypothesis is formulated. Hence, in this formulation, rationality and non-rationality collide twofold. The collision of these two categories cannot be resolved within the position of the categories rationality and non-rationality. If either of the two colliding categories are emphasised, paradoxes arise, for example in the question of responsibility for scientific research (ethics of conviction vs. ethics of responsibility). The hypothesis must, therefore, be traced back to its radically individual origin: The researcher who formulated it – nature itself does not know any hypothesis.