Matteo Antonelli

This special issue of Isonomia showcases a selection of papers presented at the Triennial Conference of the SILFS (Italian Association for Logic and Philosophy of Sciences) held at the University of Urbino on 4-7 September 2023. A capital event in the association’s life, the Triennial Conference also represents a unique venue for researchers in the field to present their work, share their ideas, and interact with the larger community of scholars, as well as with unusually wide audiences of academics and non-academics. Then, as happens very frequently, the quality and originality of the papers presented encourages the organisers to output a volume of proceedings, and this has also been the case this time. The topics addressed by the authors mainly revolve around five research clusters: 1) cognitive sciences and AI (ACCIAI, ALFIERI-FLERES-RAFFA, BIANCHINI, GALLI), 2) general philosophy of science (ALAI, CRUPI, MARCACCI), 3) philosophy of physics (FANO, GIANNETTO, ROMANO), 4) philosophy of mathematics (PICCOLOMINI D’ARAGONA), 5) the philosophy of other sciences (CARLINI). But this classification, done for internal purposes, is far from being exhaustive and definitive, as several papers will also meet the descriptors for more than one cluster. This is hardly surprising, given the deeply interdisciplinary character of the issues tackled by the authors in their contributions. As editors of this collection, we pride ourselves on having selected works that reflect researchers’ unwavering interest in the discipline’s core topics (scientific reasoning, prediction and confirmation, paradigms, the philosophy of relativity and of quantum mechanics) as well as works on topics arising in nascent, but already thriving, areas such as the philosophy of AI and environmental philosophy. Overall, we believe that the contributions in this volume testify to the vitality of our disciplines, and to their constant evolution, in a way which is not always, if ever, made perceptible by other kinds of scientific publications. In what follows, we describe in further detail the contents of each contribution. Vincenzo Crupi’s paper, Logical predictivism: How to fix use-novelty and vindicate the Copernican Revolution, challenges the claim that the gradual preference for Copernicanism over the Ptolemaic system was the consequence of ‘epistemic luck’. Through introducing a view called logical predictivism, hinged, in turn, on a re-evaluation of the notion of ‘use-novelty’, Crupi maintains that there are solid grounds to assert that, in fact, Copernicus’ views instantiated a more sound and successful scientific methodology than Ptolemy’s. In his article, Mercury’s perihelion anomaly as a use-novel confirmation of general relativity, Vincenzo Fano reassesses the notion of ‘use-novel onfirmation (prediction)’ in philosophy of science by reviewing Alai’s definition of the concept and using Mercury’s perihelion anomaly in general relativity as a case study. Fano argues that, although Mercury’s perihelion anomaly fits quite well with Alai’s rendition of the notion, not all aspects of Einstein’s reasoning about, and use of, Mercury’s perihelion example straightforwardly and automatically fall under the criteria laid out by Alai. With Flavia Marcacci’s paper, Novel “Old Facts”, Old “Novel Facts” and the Periodisation as an Epistemological Practice, we go back to the issue of the nature and essence of the Copernican revolution. Marcacci crucially argues that the debate on how much the latter thrived on the use of facts, be they “old” or “new”, is considerably restructured by carefully looking at the periodisation of the discoveries of the relevant pieces of evidence, a fact hardly taken into account, and one should add, almost invariably neglected, by the debate in the last few decades. Enrico Giannetto’s Whitehead’s Relational Special Relativity. A Natural Philosophy of Time discusses a reformulation of Einstein’s special relativity due to Alfred North Whitehead. A vigorous opponent of the belief in the independent reality of space-time, Whitehead construed physical reality originally as being based on a succession of temporal events, something which ultimately led him to produce the purely relational version of special relativity discussed in the paper. Mario Alai reviews various objections to the No-Miracle Argument (NMA) and the refinements it has undergone in order to fend them. A recent objection is that, when put in a probabilistic form, the argument commits the “base-rate fallacy”: that the probability of a startling novel prediction is antecedently very low, but very high in the light of a hypothesis H, does not significantly raise the conditional probability of H. This is because, given the empirical underdetermination of hypotheses, the prior probability that H is true is negligible. Alai answers that the prior probability of hypotheses is not negligible, because in science they are not chosen randomly, but gradually generated bottom-up with strong empirical constraints and rigorous top-down controls. Antonio Piccolomini d’Aragona’s paper, A note on a Kuhnian-Lakatosian reading of the debate between realism and constructivism in logic, aims to offer a new account of the opposition between constructivism and realism in mathematics. The former is taken by Piccolomini to instantiate Lakatos’ notion of “research programme”, whilst the latter seems to better fit in with Kuhn’s notion of “paradigm”. This helps the author to bring to the fore the main conceptual opposition between these two philosophical orientations, namely, between the rigidity of realism and the flexibility of constructivism. The paper also contains an examination of the issue, central to both the Lakatosian and the Kuhnian approach, whether “revolutions” really take place in mathematics. In Getting Even with Cognitive Science, Alessandro Acciai and Alessio Plebe probe the epistemological stakes of importing the methods of empirical psychology to study Neural Language Models (NLMs). They argue that borrowing methods from experimental psychology can be useful to carry out the investigation of NLMs’ “minds”, and, as a consequence, also to advance the study of mind, in general. In Robots and Global Challenges: What We Need to Question for a More Sustainable Robotics, Ilaria Alfieri, Antonio Fleres and Maria Raffa reframe the notion of sustainability in robotics through taking into consideration three fundamental questions concerning the environmental and social dimensions of robots. More specifically, the authors challenge prevailing assumptions about robotic embodiment, assess active inference as a computational framework for more sustainable implementations, and consider ethical concerns through the lens of social robotics for sustainability. Francesco Bianchini’s paper, Evaluating and measuring intelligence in Neural Language Models: a methodological approach, proposes a new methodological approach to assessing AI systems – especially LLMs – in the context of user interaction. The paper also raises fundamental questions about AI evaluation and the development of new analytical frameworks for AI systems which may focus on their capabilities and on the theoretical and practical grounds for classifying them as intelligent. Stefano Carlini’s, Umwelt and cities: Explanatory and Pragmatic Usefulness, uses Jakob von Uexküll’s notion of Umwelt to assess the impact of urbanization on cities. The author first presents the “selectionist” and the “constructionist” interpretations of the concept, then proceeds to show that both integrate into the notion of urban ecology and finally clarifies how this integration is useful to understand urban fauna’s behaviour. Carlini’s proposal also has practical consequences, insofar as it aims to formulate strategies of intervention for the management of urban species. Giovanni Galli’s article, Scientific Realism and Understanding with Deep Learning Models, examines the value of scientific realism in the context of the use of deep learning models (DLMs) for scientific understanding. The author defends a deployment realism framework: when AI models are reliable and accurate in practice, their success justifies a belief in the reality of the entities and processes they predict. Galli also advocates the role of AlphaFold DLMs as powerful tools for scientific inquiry, and claims that their ability to “understand” may merely be a consequence of their predictive power. Davide Romano, in Multi-Field as a determinable, defends the view that the multi-field – a realist interpretation of the wave function in quantum mechanics – is a determinable, namely, a physical object characterized by indeterminate values with respect to some properties. The paper then proceeds to suggest that the multi-field can also be characterized in terms of a determinable-based, object-level, account of metaphysical indeterminacy.