A.I., Scientific Discovery, and Realism

Epistemologists have debated at length whether scientific discovery is a rational and logical process. If it is, according to the Artificial Intelligence hypothesis, it should be possible to write computer programs able to discover laws or theories; and if such programs were written, this would definitely prove the existence of a logic of discovery. This far, however, all attempts in this direction have been unsuccessful: the programs written by Herbert Simon’s group, indeed, infer famous laws of physics and chemistry; but having found no new law, they cannot properly be considered discovery machines. The programs written in the “Turing tradition”, instead, produced new and useful empirical generalization, but no theoretical discovery, thus failing to prove the logical character of the most significant kind of discoveries. A new cognitivist and connectionist approach by Holland, Holyoak, Nisbett and Thagard, looks more promising. They picture scientific discovery as the construction of mental models of natural systems through analogical and abductive inferences, activated and constrained by an undetermined number of inputs and feedbacks from the environment. The connectionist architecture of mind accounts for the open-ended and intrinsically complex character which makes scientific discovery non programmable and unpredictable. At the same time, the assumption that by analogy and induction we can achieve faithful representations of nature explains the rationality and success of theorization. Reflection on this meta-research program, therefore, shows that a scientific-realist interpretation of scientific practice is required to account for both the rationality of discovery processes and the failure of past attempts to mechanize them. In fact, it might be argued that the Baconian and Millian belief in a logic of discovery was abandoned by logical positivists precisely because they lacked on the one hand a fully realist and cognitivist approach, and on the other hand a sufficiently wide conception of “logic”: they couldn’t foresee procedures which are rule-governed but complex and holistic because influenced but numberless factors escaping human control