RB 65

The transcendent argument constitutes a procedure that tends to miss a few vital points of scientific argumentation. First, that science is about testing whether or not a proposition is valid, not about supporting a system of beliefs by blurring the distinction between hypothesis and theory. Second, that inductive and deductive argumentation can only possibly unite on a hypothetical basis, under the assumption that if a certain set of facts are true, then a certain conclusion is possible to deduct from the supposed truth.However, this form of argumentation is similar, but not identical, to deductive argumentation, and must thus be distinguished from deductive reasoning proper.What is interesting with hypothetical-deductive argumentation is that it allows the scientist to test the viability of a hypothesis a priori, and the scientist is thus allowed the possibility of rejecting or (for the time being) accepting a hypothesis without further empirical testing. It is, however, imperative to keep in mind that the enumerative inductions, which the transcendentalist view uses to support the inference that reality has a certain necessary structure or content, are based upon empirical facts which only allow inferences and statements regarding regularities in nature or general propositions about nature, but never the establishment of undeniable truths or axioms. Since absolute necessity is a binary condition, not a condition that can be established ex analogia, it follows naturally that deductions made from inductively established premisses, strictly speaking, lose their apodictic validity. In fact, conclusions transcending from particulars to universals are valid on condition that the inductively based premiss for the deductive argument itself holds true. It is thus only possible to apply the deductive method as an instrument of prediction and description if its premisses, at the same time as they are treated as if they were axioms, are held open to constant revision, in other words treated as hypotheses. If not, then this method of prediction rapidly loses its value as a tool of prediction and description, because the predictions and p a r t i 1 , c h a p t e r 4 126