Is Plausible Reasoning a Sensible Alternative for Inductive-Statistical Reasoning? 1 Ricardo S. Silvestre 1 , Tarcísio H. C. Pequeno 2 1 Department of Philosophy, University of Montreal, 2910 Édouard-Montpetit, Montréal, QC, H3T 1J7, Canada ricardo.silvestre@umontreal.ca 2 Department of Computer Science, Federal University of Ceará, Bloco 910, Campus to Pici, Fortaleza-Ceará, 60455-760, Brazil thcp@ufc.br Abstract. The general purpose of this paper is to show a practical instance of how philosophy can benefit from some ideas, methods and techniques devel- oped in the field of Artificial Intelligence (AI). It has to do with some recent claims [4] that some of the most traditional philosophical problems have been raised and, in some sense, solved by AI researchers. The philosophical prob- lem we will deal with here is the representation of non-deductive intra- theoretic scientific inferences. We start by showing the flaws with the most traditional solution for this problem found in philosophy: Hempel’s Inductive- Statistical (I-S) model [5]. After we present a new formal model based on previous works motivated by reasoning needs in Artificial Intelligence [11] and show that since it does not suffer from the problems identified in the I-S model, it has great chances to be successful in the task of satisfactorily repre- senting the non-deductive intra-theoretic scientific inferences. 1. Introduction In the introduction of a somewhat philosophical book of essays on Artificial Intelli- gence [4], the editors spouse the thesis that in the field of AI “traditional philosophi- cal questions have received sharper formulations and surprising answers”, adding that “… important problems that the philosophical tradition overlooked have been raised and solved [in AI]”. They go as far as claiming that “Were they reborn into a modern university, Plato and Aristotle and Leibniz would most suitably take up ap- pointments in the department of computer science.” Even recognizing a certain amount of over enthusiasm and exaggeration in those affirmations, the fact is that there are evident similarities and parallels between some problems concretely faced in AI practice with some classic ones dealt with within philosophical investigation. 1 This work is partially supported by CNPq trough the LOCIA (Logic, Science and Artificial Intelligence) Project.