CFP: /Philosophies/ Special Issue: The Problem of Induction Throughout the Philosophy of Science

Dear Colleagues,

Any inference from current observations to observations not yet made or to generalizations that entail facts not yet in evidence is widely conceded to invite a Problem of Induction. Some view this as a problem of justification: How can we justify such ampliative inferences? This invites a focus on the relation of justification between theory and evidence, typically deploying the tools of conceptual analysis. Insofar as methodology is relevant, these analyses of justification tend to elevate existing scientific practices to the status of ideals. However, the focus on justification can also invite radical rejections of methodological conventions, such as the defense of non-empirical methods of theory evaluation in theoretical physics. In either case, there is a presumption that a logic of discovery is either impossible or irrelevant.

Others have framed the Problem of Induction in terms of the failure to obtain truth: Going beyond our observations entails the risk of error, and the problem is to ascertain how or if this risk can be eliminated, mitigated, or circumscribed. This invites a focus on method. Proposed solutions to this version of the problem include formal learning theory (logical reliability), meta-inductivism, and probably approximately correct (PAC) learnability. These proposals and their focus on method reopen the door to a non-trivial logic of discovery and are especially salient to the rise of machine learning in science and debates over its foundations and scope. Can machines carry out novel scientifically significant induction of theory from observation? What are the normative constraints on such algorithms? These more recent methodological approaches to the problem of induction have also given rise to the development of new domains of scientific practice, chief among them the burgeoning field of causal discovery.

How have these divergent interpretations of the Problem of Induction and the consequent strategies introduced to resolve or dissolve the problem shaped the recent philosophy of science? What should a philosophy of the new sciences of inductive learning look like? This Special Issue aims to bring to the fore the ways in which the Problem of Induction continues to drive the philosophy of science, and to evaluate the impact of proposed solutions to the problem of induction on both science and its philosophy.

Prof. Benjamin Jantzen
Guest Editor

Keywords

• problem of induction
• confirmation
• inductive skepticism
• meta-induction
• formal learning theory
• logical reliability
• theory choice
• ampliative inference
• machine learning
• logic of discovery

Please contact Special Issue Editor Viktor Ćirković at [email protected] for further information.

Philosophies (ISSN 2409-9287, https://www.mdpi.com/journal/philosophies) is an international peer-reviewed open access journal. The journal has recently been indexed into Emerging Sources Citation Index (ESCI) in Web of Science.

Special Issue Homepage: