In Identifying Future-Proof Science (OUP 2022) I argue that we can confidently identify many scientific claims that are future-proof: they will last forever (so long as science continues). Examples include the evolution of human beings from fish, the fact that the Milky Way is a spiral galaxy, and “oxygen atoms are heavier than hydrogen atoms”. Whilst claims about truth in science are usually associated with scientific realism, it is crucial to note that most anti-realists will also agree with such examples, whether on the grounds that they concern in-principle observables, on the grounds that we are rightly confident that there are no plausible unconceived alternatives, or on other grounds. But how should we go about identifying future-proof science? This appears to be a new question for philosophers of science, and not an unimportant one. It unites traditional ‘realists’ and ‘anti-realists’, usefully demonstrating a point of consensus amongst philosophers of science: we all agree that there are many established scientific facts, including facts about things that have never been observed. Even philosophers who stress that “history shows that scientific truths are perishable” (Oreskes 2019, Why Trust Science?) think that there are many scientific truths that are here to stay, such as ‘smoking causes cancer’ and human-caused climate change. Kyle Stanford, for example, believes in many ‘establish scientific facts’, including our knowledge of fossil origins (Stanford 2011). Thus I argue that philosophers should never have presented themselves as polarised on two sides of a ‘science and truth’ debate. The labels ‘realism’ and ‘antirealism’ are mostly unhelpful, and should be left behind. The interesting question concerns how we identify the scientific facts. It is argued that the best way to identify future-proof science is to avoid any attempt to analyse the relevant first-order scientific evidence (novel predictive success, unifying explanations, etc.), instead focusing purely on second-order evidence. Specifically, a scientific claim is future-proof when the relevant scientific community is large, international, and diverse, and at least 95% of that community would describe the claim as a ‘scientific fact’. In the entire history of science, no claim meeting these criteria has ever been overturned, despite enormous opportunity for that to happen (were it ever going to happen). There are important consequences for school education: If this is indeed the way to identify future-proof science, then the vast majority of school-leavers will have hardly any of the requisite skills, since schools systems around the world completely neglect to teach children how to judge the second-order evidence for scientific claims.

In this talk, I attend to three main tasks. First, I locate the main rationale for my perspectival realism in what I call historical naturalism (drawing on Massimi 2022, Ch 8). I argue that our realist commitments originate from a thoroughgoing naturalistic stance. However, by contrast with classical ways in which naturalism has been portrayed in the literature (starting from Quine 1968), I point out the need to enlarge naturalism to encompass our scientific history as a way of better understanding how we came to carve the world with the kinds we know and love. Our natural kinds, I argue, are the product of our scientific history that is redefining the very idea of what ‘naturalness’ means. My second task is to give one major highlight of perspectival realism: how to rethink the ontology of natural kinds in light of historical naturalism. Here I shall bring my Neurathian approach to natural kinds in dialogue with the approaches of my co-symposiasts by highlighting relevant affinities with Chang’s view of natural kinds born out of epistemic iterations, Chirimuuta’s haptic realism with the notion of ‘ideal patterns’ and Vickers’ future-proof facts and associated commitment to predicting novel phenomena. I defend an ontology of phenomena and explain how I see natural kinds as groupings of phenomena. This way of rethinking natural kinds has the advantage of avoiding the classic problems about reference discontinuity / conceptual change at one hand, and ‘eternal natural kinds’ at the other hand. My third and last task is to articulate the reasons why I see such a shift in realist commitments as crucial for delivering a pluralist and inclusive view of scientific knowledge production, where past theories and past achievements are not just either celebrated in the hagiography of the winners or throw in the dustbin of history. Instead, they are an intrinsic part of how we reliably came to know the world as being this way. A focus on historically and culturally situated scientific perspectives, combined with an inclusive notion of ‘epistemic communities’, allows one to reassess scientific knowledge production not as the repository of an elite community of scientists. Perspectival realism celebrates the social and collaborative nature of scientific knowledge and embeds a plurality of situated epistemic communities in the very fabric of scientific knowledge production.

The way that I seek to redirect the realism debate is away from the question of the reality of unobservable posits of scientific theories and models, and towards the question of whether those theories and models should be interpreted realistically. This makes it easier to include within the realism debate sciences of relatively large and observable items, as are many branches of biology. But it is not a simple trade of the ontological question of realism for a semantic one. My contribution will focus on computational neuroscience. In this discipline, models are normally interpreted as representing computations actually performed by parts of the brain. Semantically, this interpretation is literal and realistic. Ontologically, it supposes that the structure represented mathematically as a computation (i.e. a series of state transitions) there in the brain processes. I call this supposition of a structural similarity (homomorphism) between model and target, formal realism. This stands in contrast to an alternative way to interpret the model which I call haptic realism (Chirimuuta 2016). The view here is that whatever processes exist in the brain are vastly more complicated than the structures represented in the computational models, and that the aim of modelling is to achieve an acceptable simplification of those processes. Thus, the success of the research is more a matter of structuring than of discovering pre-existing structures. Ultimately, the realism debate is motivated by curiosity about what it is that the best scientific representations have to tell us about the world: is this thing really as presented in the model? Thus, I argue that the contrast between formal realism vs. haptic realism is a good template for framing the realism debate when discussing the implications of sciences of extremely complex macro and mesoscopic systems, such as the nervous system, and generalising to elsewhere in biology, including ecology, as well as the physical sciences of large complex systems such as climate and geological formations. Haptic realism does not suppose that the structures given in scientific models are fully constructed or mind-dependent, but that there is an ineliminable human component in all scientific representations, due to the fact that they can never depict the full complexity of their target systems and as such are the result of human decisions about how to simplify. The acceptability of certain simplifications (abstractions and idealisations) over others is due to a number of factors, including predictive accuracy, mathematical/computational tractability, and the envisaged technological applications of the model. Formal realism supposes that scientific representations are, at their best, a clear-view window onto mind-independent nature, whereas haptic realism maintains that this is an unrealistic way to describe the practices and achievements of science.

My re-conception of realism is based on new pragmatist notions of knowledge, truth and reality, which are elaborated in the forthcoming book Realism for Realistic People. These notions are designed for better understanding and facilitation of scientific and quotidian practices. I focus on “active knowledge,” which consists in knowing how to do things. Active knowledge both enables and utilizes propositional knowledge. The quality of active knowledge consists in the “operational coherence” of epistemic activities. Operational coherence is about designing our activities so that they make sense as plans for achieving our aims, and it is a notion deeply connected with the interpretive dimensions of Chirimuuta’s haptic realism. I re-conceive the very notions of reality and truth in terms of operational coherence, thereby rendering them as concepts operative in actual practices: roughly speaking, true propositions facilitate operationally coherent activities, which deal in real entities. Empirical truth is not a matter of correspondence to an inaccessible sort of mind-independent reality; the correspondence achieved in real practices is among accessible realities that are “mind-framed” yet not “mind-controlled.” My main interest in reconceiving realism is to turn it into an operational doctrine that we can actually put into practice, and in keeping with the best scientific practices. I take realism in and about science as “activist realism”: a commitment to do whatever we can in order to improve knowledge. And I take this in a realistic spirit, focusing on the search for what we can actually do in a process of continual learning. There are some implications of activist realism that would be contrary to the instincts of standard scientific realists, and I will highlight three of them in this presentation. (1) Following the imperative of progress inherent in activist realism naturally results in a plurality of systems of practice, each with its real entities and its true propositions. The link with Massimi’s perspectivism is evident, including the notion of natural kinds that she develops in this symposium. (2) Activist realism eliminates the unproductive opposition between realism and empiricism. The kind of naturalism advanced by Massimi also connects naturally with both realism (in my sense) and empiricism (in the usual sense). (3) The activist stance allows us to condemn those who work against empirical learning, while not claiming for ourselves supernatural access to “external reality.” The drive toward continual empirical learning may place activist realism into an interesting tension with Vickers’ preference for “future-proof” facts: should realists be eager to engage in new learning that may overturn the most secure-seeming facts of today?