Hypotheses testing in adaptive logics: an application to medical diagnosis

In this article we introduce an adaptive logic: lata^r, which together with a contraction procedure retro and a way to distinguish kinds of premises (knowledge, hypotheses, observations), serves as a formal setting for hypotheses generation and testing in the empirical sciences, in our case medicine, more in particular, medical diagnosis en neurology. As any other adaptive logic, lata^r has a dynamic proof theory and allows for a line of a proof to be marked when it is found that it no longer observes the conditions under which it was obtained in the first place. In addition, our lata^r combines deductive and abductive steps in its dynamic proofs. Adaptive logics serve as a model for the contruction of diagnoses in neurology. As opposed to other abductive models, this one takes into account the fact that diagnostic hypotheses are both produced by an abductive rule and just by assertion, specially when a medical doctor aims at refuting a hypothesis.