Abstracts: symposia

The default view of the Second Law is that entropy increase is an objective feature of our world. However, entropy is a property of macrostates or probability distributions, and on a purely abstract level there are infinite possible ways of partitioning our universe’s phase space into regions or cells over which different sets of macrostates or probability distributions can be defined. Each choice of partitioning will tell its own entropic story, and this suggests a kind of perspectivalism: our own entropy gradient characterises not the universe itself, but rather our particular physical coupling to it. This perspectival thesis is supported to the extent to which radically different couplings can elicit radically different types of entropy gradients in the eye of an observer. This symposium aims to explore this issue by examining various aspects of coarse-graining, from causal constraints to biological, psychological, and even social factors. By bringing these considerations into the foreground, the authors hope to ease the tension between objectivist and subjectivist views of coarse-graining, and ultimately of the thermodynamic arrow.

David Papineau (King’s College London): Unifying Causal and Thermodynamic Asymmetry  

Recent work has highlighted some striking similarities between the asymmetry of causation and the asymmetry of entropy increase. One increasingly popular way to analyse causation is in terms of acyclic causal models composed of deterministic structural equations. In many contexts, the asymmetry of dependent and independent variables in these equations is taken as given. But at the same time, it is widely recognised that the crucial assumptions needed for causal inference — the Causal Markov Condition and the Faithfulness Condition — can be licensed by the assumption that the exogenous variables in causal models are probabilistically independent. This has led a number of theorists to propose that such exogenous independence is the key to causal asymmetry: a causal model will correctly capture causal order just in case its exogenous terms are probabilistically independent. The currently dominant philosophical account of thermodynamic asymmetry is David Albert and Barry Loewer’s “Mentaculus”. However, in a series of recent papers, David Wallace has in effect argued that their "principle” theory ignores the “constructive” techniques that practising scientists use to analyse the behaviour of non-equilibrium branch systems. These techniques involve applying fundamental dynamics to repeated coarse-grainings that eliminate correlations between elements of the system, as for instance in Boltzmann’s H-theorem, the more general “BBGKY hierarchy”, and standard treatments of Brownian motion. So both species of asymmetry seem to derive from a combination of symmetric laws with probabilistic independence requirements on boundary conditions. This talk will compare the two accounts, and in particular ask whether their similarity carries some deeper significance.

Saakshi Dulani (Johns Hopkins University): Value in Coarse-Graining

There has been much debate over the statistical definition and interpretation of thermodynamic entropy, particularly between Boltzmannians and Gibbsians. The former favour a macrostate-based definition of entropy while the latter favour a probability-based one. The disagreement is usually framed as a continuation of the age-old divide between objectivists and subjectivists, in which Boltzmann entropy is characterized as ontic (about the world), and therefore objective, whereas Gibbs entropy is characterized as epistemic (about an observer), and therefore subjective. To help remedy this gridlock, I subvert the assumption that Boltzmann and Gibbs entropy cleanly align with objectivist and subjectivist interpretations respectively. I argue instead that both types of entropy share objective and subjective features. I go on to propose that entropy broadly serves as an interface between the observer and the world, reflecting how we carve up reality and coarse-grain our representations. Indeed, an interface view of entropy has gained traction in resource-theoretic and interventionist accounts of thermodynamics/statistical mechanics. They emphasize how a choice of coarse-graining is perspectival, relative to one’s empirical ‘capacities’, ‘control’, and ‘interests’. Yet I contend that such accounts are incomplete because they stop short of acknowledging extra-empirical values, which are implicitly baked into these perspectival notions. I then argue that since perspectival notions are value-laden, the notions of coarse-graining, and by extension, entropy, are value-laden too. Finally, I discuss some provocative consequences of my account. I draw inspiration from the literature on agency, emergence/reduction, and social epistemology to highlight challenges for totalitarian reductionist projects. For example, the Mentaculus proposal, which is Boltzmannian in spirit, seeks to derive an objective, non-fundamental reality from a fundamental ontic base. But this aspiration neglects to appreciate how coarse-graining itself is value-laden, embedding subjective elements like epistemic standpoints. I suggest that when we coarse-grain different levels of reality, any attempt at reduction between levels fails since each level encodes its own set of values. 

Athamos Stradis (University of Cologne): The Causal and Thermodynamic Arrows: Two to Tango

Our universe exhibits a causal arrow whose typical relata are localised macroscopic objects: Billy and Suzy throwing rocks, rocks smashing windows, and so on. It is widely supposed that these relata need an entropy gradient in order to exist, for not much happens in equilibrium. Hence, causal structure requires the thermodynamic arrow. However, standard qualitative explanations of the latter make no overt reference to causation. The thermodynamic arrow is portrayed as simply ‘there’ in the background, aloof from the causal entanglements of rocks, windows, and angry parents. This notion of a one-way relationship poses a puzzle: if the thermodynamic arrow is truly autonomous, then why should causal structure – and by extension, a causal arrow – exist at all? This entire aspect of our universe starts to look like a contingent mystery. But I will argue that this notion is false, for the relationship is very much two-way. Just as the thermodynamic arrow provides a suitable entropic climate for causal structure, causation ratifies the macrostates underlying that gradient as objectively salient. Once we recognise that causal structure is baked into the thermodynamic arrow – and not merely icing on top – the puzzle dissolves. This result supplements more quantitative derivations of the thermodynamic arrow that enlist causally asymmetric assumptions like the Stoßzahlansatz by explaining why this particular entropy gradient is physically significant: the relevant variables participate in causal structure, hence why they’re observable in principle. My analysis therefore bridges the two previous talks by suggesting that causation might be the objectivist tool that prunes back the plurality of possible coarse-grainings, and by extension the plurality of valid entropy gradients, endorsed by fully fledged perspectivalism.

Concepts like ‘sex’, ‘ancestry’, and ‘diversity’ often play fraught but important roles in popular understandings of the health and life sciences, scientific practice, and policy. Although philosophers have long recognized that such concepts may be ambiguous and come with significant baggage, many have proven remarkably resilient and difficult to change. Philosophers of the life sciences have treated conceptual instability as a feature of pluralism, often finding benefits to imprecision. This symposium provides a renewed look at imprecise concepts in the life sciences, with a focus on loaded concepts whose stabilization carries ethical and social risks. The goal is to clarify how imprecise concepts are stabilized and entrenched in scientific practice, and the potential consequences of this process. Central questions to be addressed include: How – and by whom – are biology concepts articulated and negotiated between science and society? What makes biological concepts stable and legitimate objects of inquiry, despite calls for changes or substitutions? Speakers will address how diverse social and scientific practices such as measurement, policymaking, and media contribute to resistance to conceptual change and the risks of methodological inertia.

Marina DiMarco (Washington University in St. Louis): Imprecise Sexes

Sexes stand accused of many forms of indeterminacy: as used in science, ‘male’ and ‘female’ might have multiple meanings, perhaps in different contexts; they might under-specify other relevant details; their boundaries may be blurry, and so on. In light of this, one stream of philosophical work on sexes has been concerned to achieve, or restore, precision and clarity about their meaning(s) as a scientific concept(s) (Richardson 2022), or to abandon them if this can’t be adequately salvaged (Watkins & DiMarco 2025). Against this current, however, other philosophers of science hold that conceptual ambiguity can sometimes be legitimate, ineliminable, or even a feature of scientific concepts (e.g., Neto 2020). In this talk I’ll ask (1) what this means for calls to clarify scientific uses of ‘sex’ and/or preserve their ambiguity, and (2) what we might learn about conceptual indeterminacy in general from careful attention to sexes and the scientific and social practices in which they are embedded. One consideration that cuts across these questions is that sexes are what I call “sticky” kinds: kinds for which social and political baggage, and/or a particular logic of research questions (in Elisabeth Lloyd’s 2015 sense), have proven difficult to shed. Because sexes are sticky, I'll argue that forms of imprecision whose benefits depend on transparency about the nature of ambiguity, and/or users' ability to navigate multiple, distinct practices and conceptualizations, are relatively risky. Optimistically, education and advocacy about the uses of sex concepts both in and out of science might help reduce these risks. This could make conceptual imprecision and its benefits for science an additional motivation for social change.

Mia Miyagi (Brown University): What Constitutes a Genomic Sex Difference?

Human geneticists imagine large genomic datasets as informative about sex differences on multiple temporal and biological scales. Patterns of genetic diversity or ‘ancestry’ are used to infer historical differences in the population history of sexes, and relationships between phenotype measures and genotype data in biobanks are used to describe sex specific ‘genetic architectures’ in traits ranging from cardiovascular health to facial attractiveness (Wilson Sayres 2018, Ober et al. 2008, Hu et al. 2019). What work must sex as a category of genetic analysis do in order for all of these applications to coexist? In this talk, we’ll examine what shared assumptions about the nature of sex undergird the diverse quantitative methods used in these research areas, as well as how ‘quality control’ practices around sex in genomics might restrict the reproducibility or portability of genomic sex difference claims. In addition, I will argue that population genetic simulation frameworks offer an opportunity to quantify the possible impacts of imprecision around ‘sex’ and ‘sex differences’, and use one such simulation program to show that the way sex (and in particular intrasexual variation) is operationalized has a measurable impact on sex-specific demographic inference (Miyagi et al. 2025). I contend that cases such as this can function as areas of translation to animate concerns about conceptual imprecision for a scientific audience.

Yasmin Haddad (Université du Québec à Montréal): The Ancestry Paradox

‘Ancestry’ is a central organizing concept in human population genetics. It plays a key explanatory and classificatory role, structures major modeling practices, and underwrites replacement of race in contemporary genomic research. Yet, despite this centrality, ancestry remains methodologically under-specified and conceptually vague: it is operationalized in multiple, often incompatible ways across scientific contexts, with no single agreed-upon definition. This talk addresses a paradox arising from this tension: how can a concept that is imprecise and multiply operationalized nonetheless become stabilized, entrenched, and epistemically authoritative in scientific practice? I argue this paradox can be resolved by shifting attention away from ancestry as a representational category and towards the practices, assumptions, and methods that constitute it as a legitimate object of inquiry.

I begin by examining why ancestry has become indispensable in human population genetics. I show that its appeal rests largely on the claim that it is more objective than race, insofar as it is taken to track real patterns of genetic similarity and difference. I then assess this claim through the lens of prominent theories of scientific objectivity, arguing that existing accounts – whether factual or social – struggle to explain how ancestry secures its epistemic authority despite its conceptual instability.

As an alternative, I develop an account of constitutive objectivity, according to which objectivity emerges through the critical stabilization of the very practices and modelling assumptions that bring a concept into scientific use. Ultimately, constitutive objectivity offers a general framework for understanding how imprecise concepts can nonetheless function objectively in model-based sciences.

Lauren Wilson (University of Exeter): Diversity within Genetic Studies of Complex Traits

This talk investigates the question “What do researchers mean when they talk about and advocate for increasing diversity in genetic studies of complex traits? Over the last decade there has been an alarm raised, most genomic studies are performed on white populations of European ancestry. To prevent disparity in access to the potential benefits of genomic studies of health, there have been continued calls to diversify the populations studied, with emphasis placed on increasing racial diversity, a proxy for genetic ancestry (Wojcik et al. 2019). But how is the concept of diversity operating in these discussions? Previous philosophical work has explored epistemic merits of conceptual complexity and precision within the life sciences (Brigandt and Love 2012; DiMarco 2025). My work builds on this foundation to show how the imprecise concept of diversity operates in scientific practice within genetic studies of complex traits related to health. Understanding the ways diversity is being conceptualized in this area enables scientists and policy makers to more precisely and reflexively advocate for the types of diversity that really matter for addressing research questions.

In my talk, I taxonomize the relationship between diversity, aims of research, and values. I identify three common ways researchers conceptualize diversity within genomic studies of complex traits, racial diversity, diversity of genetic ancestry, or diversity of single nucleotide polymorphisms (SNP’s). I then identify three main research aims increasing diversity is meant to facilitate. Epistemic aims include increasing diversity to address ongoing methodological issues and to facilitate research on specific research questions. For example, increasing diversity may help researchers better control for the effects of population structure on their results or allow researchers to make progress on specific research questions, such as exploring issues of structural racism. A non-epistemic aim of increasing diversity is justice. Increasing diversity corrects for under-representation of certain marginalized groups in studies and ensures that these groups have access to any future benefits.

My taxonomy can help researchers assess if they are focusing on the kind of diversity that will allow them to achieve their research aims. For example, to facilitate the use of genetic methodologies to study the effects of structural racism, sample data sets must increase racial diversity. This mapping can also help identify the ways focusing on increasing racial diversity, diversity of genetic ancestry, or diversity of SNP’s may not be sufficient to address important research aims.

It is an open question whether diversity will improve genetic and genomic studies. However, my analysis shows that when researchers call for increasing diversity in genetic studies there is slippage between racial diversity, diversity of genetic ancestry, and genetic diversity. Answering the question of whether increasing diversity in genomic studies of complex traits will help researchers achieve their aims requires understanding what form of diversity researchers are interested in and what issues they hope to address.

This symposium investigates and explores applications of AI tools in molecular studies and their philosophical implications. By focusing on AI models used in current chemistry and biochemistry, we will consider how using them influences core debates in philosophy of science, such as reductionism, realism and theory-ladedness. Our symposium will discuss these topics in three talks. The first talk will consider the manifold hypothesis in ML and will focus on chemical case studies to explore its consequences. The second talk will focus on how AI is used in chemistry for predictions and how this impacts realism. The last talk will consider AlphaFold and the protein folding problem in reductionist terms. Going beyond general discussions on AI and ML, our symposium puts forward talks that consider the actual use of AI in molecular sciences and provides novel contributions to the related debates in philosophy of science.

Fintan Mallory (Durham University) and Mel Andrews (Princeton University): Constructing Chemical Manifolds

A considerable amount of research in computational data analysis is aimed at generating low dimensional representations of the dataset under study. This method, familiar from traditional, linear statistical techniques such as PCA and ICA has formed the basis of many advances in machine learning (Meilǎ & Zhang, 2025). The advances of recent years involve projecting the dataset onto a lower-dimensional non-linear manifold whose dimensions account for most of the variance in the data. This method has supported theoretical developments in chemistry (Tribello et al. 2012), astronomy (Chen et al. 2015) and neuroscience (Perich et al. 2025). This has led to the reconceptualization of vast swathes of science as ‘manifold learning’ within the machine learning community. This new ideology is supported with appeal to the ‘manifold hypothesis’, a metatheoretical claim that the data within any dataset lies on a lower-dimensional implicit manifold. These manifolds are often taken to be real structures within the world whose shape is open to discovery by researchers. For example, the possible configurations of an ethanol molecule are claimed to lie on the surface of a two-dimensional torus while malonaldehyde lies on a manifold of 3 dimensions (Koelle et al. 2022). While the manifold hypothesis has been presented as a substantive metaphysical thesis, some philosophers of science suggest that it largely serves a rhetorical function to justify the epistemic capture of theoretically informed, expertise-driven science by machine learning researchers. The goal of this talk is to make sense of the manifold hypothesis within a pragmatist approach to the philosophy of science with a particular focus on case studies from chemistry. Rather than trying to delineate which features of this structure are due to objective features of reality and which are the product of convention, we will show how manifold structure is constructed by theory-laden methods of data curation and analysis. The talk will present several formulations of the manifold hypothesis of varying strength and discuss what would be evidence in support of them. It will then present a specific case study, showing how the toroidal structure of the ethanol manifold is constructed within scientific practices.

Vanessa Seifert (University of Athens): Could Ai Undermine Standard Scientific Realism?

Atoms and molecules have constituted a paradigmatic case in defending the reality of unobservable entities. When Jean Perrin (1916) proclaimed to have calculated Avogadro’s number by thirteen different methods, philosophers of science took this to show how to empirically establish the truth of scientific theories and the reality of the unobservable entities. However, the way AI models are developed to predict novel chemical substances and compounds undermines the idea that the predictive success of science empirically supports the truth of its theories. This is because there are at least some AI models whose predictive success is not based on incorporating or following an underlying theory. Instead, their success is based on the analysis of empirical data.

To spell this out, I present two AI models that are developed in order to predict chemical substances: the generative intelligence framework developed by Krishnan et al. (2025) at MIT and the ML algorithm IMPRESSION developed by Gerrard et al. (2020) at the University of Bristol. Based on the analysis of these examples, I identify two ways by which AI models are trained to make predictions. The first — and so-called standard way— is by ‘feeding’ them empirical data, with additionally (though not necessarily) imposing theoretical constraints (this is the case with Krishnan et al 2025). The second is by ‘feeding’ them theoretical data that have been produced by quantum chemical computations (this is the case with IMPRESSION).

These two ways of training AI models shows that the predictive success of our models cannot always be accounted for by the truth of an underlying theory. This is particularly evident by the standard way by which AI models are trained which does not directly involve incorporating theoretical hypotheses or theoretical data. This is particularly problematic for the defence of scientific realism in terms of the No Miracles Argument (NMA) which takes novel predictions in science to be best explained by the truth of our underlying theories.

In light of this, I examine whether classic ways of defending standard scientific realism (namely via the NMA) can be maintained in light of AI’s use for the prediction and discovery of chemical substances and compounds. I argue that the way by which AI models are developed and — specifically— trained to make their predictions undermines arguments that are based on the predictive success of our scientific models.

Francesca Bellazzi (Durham University): Does AlphaFold Solve The Protein Folding Problem?

The problem of protein folding has been an ongoing concern for biochemists, molecular biologists, and structural biologists. This problem concerns the relation between sequence and structure in proteins and whether we can predict the native structure starting from the amino acid sequences. Experimental techniques to register such data include X-ray crystallography, nuclear magnetic resonance, spectroscopy, and electron microscopy. In recent years, computational tools to predict such structures from their amino acids have been developed and AI tools like AlphaFold2 allowed for significant improvements. Specifically, AlphaFold2 has been deemed as one of the greatest contributions of AI to science and an important scientific breakthrough in 21stscience (Yang et al. 2023). One of the main advantages of using these tools is the possibility to inquire into proteins’ functionality, protein-protein interactions, and ultimately it can aid drug discovery and further research. But is AlphaFold really solving the protein folding problem?

This talk will answer this question and investigate whether AlphaFold solves the protein folding problem. I will suggest an interpretation of the protein folding problem in terms of reductionism: it can be framed in terms of reducing the structure of the protein to its sequence. I will then argue that, while AlphaFold is successful in predicting the structure of the protein, it does not support the reduction of structure to sequence, and so it does not solve the protein folding problem in these terms. This does not exclude other forms of reductionism though, such as one in which structure is reducible to a combination of physic-chemical information and evolutionary constraints. This analysis has consequences also for the kind of inferences we can make based on AlphaFold regarding proteins’ functionality and bioavailability and the characterization of proteins as biochemical kinds.

First, it will introduce the protein folding problem and its history and how it can be seen as a problem of reduction. Then it will go through the details of AlphaFold and how the tool is said to do de novo prediction, claiming to solve one of the aspects of protein folding problem. This will offer the grounds to argue that AlphaFold does not solve the protein folding problem as a problem of reduction of structure to sequence, but rather it supports the prediction of structure based on a combination of information on physical constraints and evolutionary constraints. I will conclude by exploring whether we can frame a form of reductionism in biochemical practice that is not bottom-up, but rather considers multiple parameters.

In cases of Science-Based Science Scepticism, scientists themselves (rather than members of the public) claim that some scientific claims or theories, supported by the work of other scientists, ought not to be used in some policy context. Consider, for example, how some expert epidemiologists argued against the use of various models for developing lockdown policies during the Covid-19 pandemic (Melnick and Ioannidis 2020); or how some climate scientists object to basing policies on projections concerning deeply uncertain tipping points in the climate system (Kopp et al. 2025).

Building on recent work on values in science, we distinguish two questions: first, whether some claim counts as an instance of good science by disciplinary standards; and second, whether some scientific claim is “good enough” to play a proper role in policy advice. Precisely because these questions are easily confused, many real-life cases may conflate them. What may seem to be straightforward cases of epistemic disagreement involve a complicated cluster of epistemic and normative claims and presuppositions.

Through case-studies of Science-Based Science Scepticism and “good enough” science, this symposium will provide tools both for better understanding actual debates and for normative reform.

Mathias Frisch (Leibniz University Hannover) and Hannah Hilligardt (University of Bern): Climate Tipping Points: scientific disagreement in the face of uncertain evidence 

In October 2024 forty-four climate scientists submitted an “Open Letter by Climate Scientists to the Nordic Council of Ministers” warning that the Atlantic Meridional Overturning Circulation (AMOC) is increasingly at risk of passing a tipping point with potentially devastating consequences especially for Nordic countries. The threat posed by an AMOC collapse is large enough, according to the letter’s signatories, that it justifies communicating what at this point in time still is deeply uncertain evidence concerning the possibility of an AMOC collapse and that this evidence ought to inform climate policy decisions. Yet as climate tipping points have received a growing amount of public attention, research on tipping points remains one of the most controversial areas of climate science. While many climate scientists would agree with the signatories of the of the open letter argue that tipping point risks, such as an AMOC collapse, ought to be communicated to policy-makers (and perhaps even with a sense of urgency), others warn that dramatic proclamations about an AMOC collapse are too speculative and risk diverting attention from more “immediate dangers posed by increased greenhouse gas emissions” (Kopp 2024).

In this talk I will argue that the debate concerning tipping points revolves only partly around epistemic disagreements about the evidential situation. To a significant extent climate scientists agree on the evidential status of research on climate tipping points. Those invoking potential climate tipping points in climate policy discussions largely agree with their opponents on the deep uncertainties characterizing our understanding of these tipping points, but disagree on the implications for policy advice. Rather the disagreement concerns appropriate evidential standards for policy purposes and the legitimacy of appealing to deeply uncertain knowledge (and, in particular, uncertain knowledge concerning what appear to be low-likelihood events) in policy advice. Moreover, whether certain models projecting the behavior of a climate tipping point are “good enough” for policy advice depends in subtle ways on how the models and the evidence derived from them is interpreted and on what the decision framework is within which the evidence concerning tipping points is evaluated. Thus, I will argue, we can understand the disagreement concerning tipping points as at least partly a normative disagreement about which decision framework is appropriate in the face of catastrophic risks and a concomitant disagreement about how properly to interpret the evidence provided by tipping point research. This, I will further argue, implies that the disagreement among climate scientists cannot to be settled on purely scientific grounds and ought to be negotiated in ways that centrally involve the input from stakeholders and policymakers.

Katherine Furman (University of Liverpool): Evidential Scarcity as a Route to Science Based Science Scepticism

Evidence hierarchies play a dominant role in medicine and in policy decision-making, despite the sustained critiques against them. These hierarchies tell us how to produce evidence, how to prioritise it, and they give us clear guidance for decision-making – do whatever the “best” evidence says. Importantly, hierarchies also set parameters on what can count as a reasonable disagreement between experts. Lower-level evidence cannot be reasonably used in an argument against higher-level evidence, and this limits the scope of science-based science scepticism.

Built into the structure of the evidence hierarchy approach is the assumption that evidence is plentiful, and we just need to make choices about what to do with our evidential bounty. But sometimes evidence is scarce. For instance, at the start of the Covid pandemic, when there was very little scientific understanding of the disease. We also see this in instances of economic policy making that impacts rural communities, and about which our best social science evidence is patchy.

Some sciences are comfortable with the experience of evidential scarcity. Archaeologists, palaeontologists and astrobiologists all typically have very little access to the evidence they want and need to develop and test their theories. As a result, they have to find creative ways of managing what little evidence they have. For instance, one way is to engage in analogical reasoning. Here, a well-understood phenomenon is extended to hypothesise about a new target. This is an effective way of extending what little evidence you have.

We saw the same kind of analogical reasoning used in policy-orientated epidemiological modelling at the start of the Covid pandemic. In this case, there was very little evidence on modes of disease transmission, infection rates, fatality rates, and the consequences for those who experienced the disease and survived. Here, evidence was scarce. By engaging in analogical reasoning about diseases that were well-understood – such as various strains of influenza – epidemiological models were developed (and predictably criticised). To engage in successful analogical thinking, there need to be good reasons to think that the well-understood phenomenon appropriately resembles the new target, and these reasons are up for critique. This opens up space for expert-disagreement, and intra-scientific scepticism.

The point of this talk is that evidential scarcity is a reality in many policy-orientated sciences. This will require creative scientific thinking about what to do with limited evidence, and this predictably opens up pathways to science-based science scepticism.

Stephen John (University of Cambridge): Cancer screening: when is the evidence “good enough”?

Many experts claim that routine screening of asymptomatic individuals for early cancer is an effective intervention. Other experts disagree, suggesting that even highly popular screening modalities – such as routine age-based mammographies for early detection of breast cancer – do “more harm than good”. One way of understanding these debates is as involving more fundamental disagreements in medical epidemiology; for example, over the relative weighting of “expert judgment” versus RCT-generated evidence.

Clearly, to a large extent, debates over cancer screening do revolve around questions of effectiveness. However, as this paper shows, they also often involve a slightly different set of concerns: often, both proponents and opponents of screening agree on the evidential situation. In particular, they often agree on the uncertainties around cancer screening. Where they disagree is on the practical relevance of these uncertainties: opponents of screening tend to argue that it is wrong to introduce screening programmes when we are uncertain that they will do more good than harm; proponents respond that we are certain enough that the policies will be, all things considered, a positive.

The first half of this paper does two things. First, developing the analysis above, it argues that we can model debates over screening as involving an implicit normative disagreement over when medical claims are “certain enough” for use in policy. Second, it then provides a way of understanding the basis for this disagreement over “certain enough” medical evidence, as stemming from a broader disagreement between traditional public health epidemiology – characterised by a broadly Utilitarian commitment to maximise net social welfare – and the perspective of clinical ethics, characterised by a strong commitment to the principle of non-maleficenece (“first, do no harm”). In effect, I argue that what appears to be a disagreement over the evidence for and against screening effectiveness is, in fact, better understood as a (very peculiar) version of a far more general tension in medical practice.

In the second half of the paper, I turn to more normative questions: if we think of the “screening wars” as involving a disagreement over when the science is “good enough” (rather than over whether the science is “good”), what does this imply for policymaking? I suggest we ought to interpret this phenomenon in terms of what I call “ethical trespassing”: just as the “epistemic trespasser” makes epistemic claims in areas which are beyond her expertise, the “ethical trespasser” makes ethical judgments in areas where she lacks authority to impose her ethical views on others. Tentatively, I suggest that at least sometimes, proponents of Evidence Based Medicine are guilty of “ethical trespassing”: they inappropriately import ethical principles from the clinic to the domain of policy-making. While this does not show that we must or must not implement cancer screening, it does imply the need for robust mechanisms to ensure that decisions reflect appropriate values.

Torbjørn Gundersen (University College of the Norwegian Correctional Service): Science advice, objectivity and good enough evidence

Public epistemic trust in scientific advice is partly based on its compliance with a set of scientific principles, such as objectivity, rigour, and neutrality. Accordingly, the guidelines of science advisory bodies often explicitly state that the process of providing advice must be based on such scientific norms. At the same time, adherence to objectivity and related principles poses several concrete challenges for scientists in providing relevant, informative, and responsible advice to their audiences. First, an emphasis on accuracy and rigour may lead scientists to refrain from including potentially policy-relevant results and sources in their expert assessments, thereby ignoring more recent findings from the grey literature and open-access archives that are not yet well established in the peer-reviewed literature. Second, a commitment to scientific norms such as criticism and scepticism may lead scientists to be overly epistemically cautious. A case in point is the claim that the Intergovernmental Panel on Climate Change’s (IPCC) adherence to scientific norms has led the panel to be too conservative in its assessment of the impacts of climate change, due to concerns about false positives (Oppenheimer 2021; Brysse 2013), thereby setting the evidential bar too high for public policy (Lloyd et al. 2021).

These examples indicate a need for scientists to interpret and apply scientific principles properly when they are part of a science advisory body. Here, I argue that objectivity and other scientific principles should not be understood in the same way in science advisory contexts as they are in scientific research. Rather, they must be interpreted in a context-sensitive manner that also takes into account factors such as the institutional role of science advisory bodies, their relevant audiences, and political disagreements.

A fundamental challenge in guiding science advice by scientific principles is that there may be disagreement among scientists about what these principles mean. Indeed, both the philosophical literature and scientists’ own views display a plurality of interpretations of objectivity. I aim to shed light on how scientific principles ought to be understood as guidelines in science advice by reconsidering the two challenges mentioned above, namely decisions about good enough sources and appropriate evidential standards. I conclude by arguing that adherence to scientific principles such as objectivity, rigour, and neutrality is not merely an indicator of trust or a device for signalling credibility, but that these principles, under the proper interpretation, should be considered central conditions for epistemically trustworthy scientific advice (for epistemic, ethical, and democratic reasons).

The philosophical debate over the nature of mental illness is often framed as a dispute between 'naturalists' and 'normativists'. This framing obscures the fact that naturalists also appeal to norms in understanding mental illness. The difference lies in the type of norm invoked: while normativists appeal to evaluative or social norms, naturalists appeal to biological norms, typically understood as outcomes of natural selection. This symposium approaches norms and normativity in psychiatry and mental health from a new angle. Moving beyond the constructed dichotomy of the traditional debate, we seek to explore how distinct sources of normativity (ethical, biological, social and epistemic) intersect in psychiatry and mental health. To this end, the symposium brings together scholars working on diverse forms of normativity in the context of psychiatry.

Harriet Fagerberg (University of Cambridge): Three Kinds of Brain Function

There has been much debate in philosophy of biology over whether functions in general, or brain function in particular, should be understood as selected effects. However, even assuming a selected effects theory, the implications for the brain are by no means straightforward. This paper argues that we should recognise three distinct sources of normativity in the brain. This raises the question of which norms, or which norm violations, feature in mental illness.

On the selected effects theory, a function is something that a trait did which caused that trait’s selection (Millikan, 1989; Neander, 1991). For the brain, the most straightforward implication is that the brain has proper functions as a consequence of inter-generational genetic selection on an evolutionary timeline (Neander, 2017). However, recent work suggests that the brain also has ontogenetic selected effects functions acquired through neural selection mechanisms such as synapse selection, whereby the brain overproduces synaptic connections and selectively retains those which prove useful (Garson, 2024; 2019; 2011). Moreover, some argue that many neural functions which appear to result from genetic selection are in fact the products of cultural selection (or ‘cognitive gadgets’) (Heyes, 2018).

This picture yields three distinct sources of functional normativity in the brain. First, there are the brain's 'evolutionary functions', brought about by genetic selection. Second, there are 'ontogenetic functions' acquired through neural selection within the individual organism's lifetime. Third, there are what one might call 'cultural functions' acquired through cumulative cultural evolution.

These functional norms interact in interesting ways. The mechanisms of neural selection set the stage for what kinds of ontogenetic functions we can acquire, but are themselves genetically selected traits with distinct functional profiles of their own. Similarly, some capacities enabling cultural learning are likely outcomes of genetic selection, even if the specific 'cognitive gadgets' acquired by them are not (Heyes, 2012). More puzzlingly, these distinct sources of normativity can dissociate and even conflict. In previous work, I have shown that Garson's Generalised Theory yields cases where a single neural trait is simultaneously functional in ontogenetic terms and dysfunctional in evolutionary terms, and vice versa (Fagerberg, Forthcoming; 2022). Similar conflicts can be imagined between cultural and evolutionary functions, and indeed between cultural and ontogenetic ones.

A great deal of debate has historically centred on whether mental disorders should be understood as types of neural dysfunction (Borsboom, 2019; Insel and Cuthbert, 2015). Against this backdrop, we can re-frame the problem: functions or dysfunctions relative to what normative standard? I argue that viewing mental illness through the lens of three distinct sources of functional normativity serves to disambiguate the debate, and speaks in favour of a less general selected effects theory on which different types of selected effects are assumed to serve distinct theoretical roles.

Sera Schwarz (Yale University): Which kinds of problems are psychiatric problems?

Most philosophers of psychiatry take mental health problems to represent some kind of deviation from relevant norms. But there is still widespread disagreement over which particular kinds of norms these are, or properly should be (see, e.g., Szasz 1961; Boorse 1976; Wakefield 1992; Evans 2025; Schwarz 2025). In other words, people disagree about which kinds of problems mental illnesses represent. In my talk, I will not directly address this first-order disagreement. Instead, I aim to develop and defend a metatheoretical claim about how best to approach it. In particular, I will argue that, in many cases, theoretical and clinical reasoning about the nature of particular psychiatric problems - that is, about the kind of problem a particular psychiatric problem is - can and should be guided by robustly ethical criteria. In other words, I will argue that there is significant latitude for ethical considerations to inform our thinking about the kind of normative deviation a particular psychiatric condition represents in some particular case.

This claim might seem anodyne. But, on reflection, it should also seem at least a little surprising. It’s tempting to think - and many philosophers have indeed assumed - that any respectable answer to the second-order question (“how should we arbitrate between different ways of thinking about psychiatric problems?”) will simply follow from our best account of the relevant empirical or metaphysical facts (Boorse, 1976; Murphy 2005). If this is right, our understanding of what psychiatric conditions “really are”, or at least our understanding of the kinds of processes psychiatric taxonomies seek to track, would determine how we think about their normative character. For example, if psychiatric disorders turn out to be natural kinds, this would be reason to think of them as deviations from biological norms; but if they turn out to be human kinds, they would be better understood as deviations from social or moral norms.

I will suggest, however, that we should resist this intuitive view: it is often the case that ethics can and should guide our psychiatric metaphysics, rather than vice versa. First, I argue that the empirical facts systematically underdetermine the character of psychiatric problems. In other words, claims about the way the world is need not strictly determine which kind of norm-violation a psychiatric condition represents. Which kind of norm-violation a particular condition represents, however, has important implications for the epistemically and metaphysically appropriate way to describe, explain, and reason about it. I then argue that, in the face of normative pluralism, there are overriding ethical reasons to either privilege or resist some of the possible characterizations of a problem. I conclude by suggesting that the ethical stakes are particularly high when considering whether to characterize psychiatric problems as distinctively psychological problems; and I offer some reasons for preferring to engage psychological norms in characterizing and explaining these problems, even if they might just as well be characterized, so far as the empirical or causal facts go, as biological or social problems.

Sebastian Rodriguez Duque (University of Cambridge): Values in mental health service delivery: An embedded philosophy case study in the use of mental health measures with youth

In this article I explore the role of values in the use of questionnaires for the use of measurement in the delivery of youth mental health. I substantiate my claims through my ongoing collaboration with philosophers of science at McGill University, and health-outcomes researchers at the University of British Columbia and Foundry, a youth wellness and mental health organization in BC, Canada.

In mental health, measurement-based-care (MBC) has received substantial funding and resources in countries like Canada. MBC establishes a baseline and is used to track changes over time in the mental health status of a patient (Sarvet, 2020; Van Cleave et al., 2012). This helps providers and clients identify areas of concern and understand the effectiveness of their interventions. In mental health, MBC specifically refers to the use of patient-reported outcome measures (Kroenke & Unutzer, 2017). These are subjective measures of health-related quality of life that purport to measure what really matters to patients from their own point of view (McClimans, 2024). In mental health, they most often use DSM categories to model the target measurement concept (i.e. the construct or measurand). Psychometrics is the science for measuring experiences, attitudes and mental traits. The quality of psychometric instruments is usually evaluated through technical methods testing for reliability and validity. In the physical sciences, whether a measure is valid – whether it measures what it purports to measure – depends in part on the coordination of robust modeling of the measurand, and the empirical constraints of measurement procedures. However, in mental health there is no robust modeling of the measurand (e.g. depression). For example, the PHQ-9 simply adopts the 9 items for Major Depressive Episode of the DSM as its model for depression. Moreover, there is also less clarity on how the measurement procedures (e.g. the response behaviour of someone filling out a questionnaire and the attainment of a score) are linked to some magnitude of the construct.

In this article I argue that current psychometric validation is unfit for the purpose of the clinical interpretation of questionnaires in mental health. I argue that emphasis on questionnaire standardization of traditional validation makes the use of measurement questionnaires overly rigid in a clinical context. Using the example of the PHQ-9, I argue that the rigorous interpretation of measurement scores can only be validated pragmatically, within a given context and for a given purpose. I argue that the main reason for the above is the necessary role of non-epistemic values in the development, use and interpretation of psychometric questionnaires. The value-ladeness of the construct (i.e. Major Depression), the values implicit in the questionnaire items, the values of service providers and clients and the consequences of the interpretation of measurement scores interact and require that a measurement practice remains dynamic and collaborative. However, this runs counter the need for standardization to build broader systems of care, and it has implications for the validity of measurement data in mental health research more broadly when relying on psychometrics. I consider these points briefly to conclude.

Justin Garson (City University of New York): Toward an Error Theory of Psychiatry

Over the last decade, a new generation of critics of psychiatry has emerged, most notably in the UK, where movements such as the Critical Psychiatry Network have gained traction (Longden and Read 2017; Davies 2021; Moncrieff 2023). These critics argue that psychiatry, in its current form, is doing more harm than good. Their concerns include overmedication, overdiagnosis, the obscuring of social causes of distress, the failure to engage meaningfully with lived experience, and pharmaceutical influence over treatment.

These new critiques tend to focus on contingent features of psychiatry, features that, in principle, could be improved or reformed. In what follows, I want to identify and defend a different style of critique, what I call an error theory of psychiatry. This approach does not merely target psychiatry’s contingent flaws (e.g., institutional coercion, diagnostic inflation, or neglect of lived experience). Instead, it seeks to locate the critique in something essential to psychiatry. In other words, it seeks to critique psychiatry on the grounds of what it is. Put differently, it aims to identify a core ontological assumption that psychiatry, as such, is committed to, and to show that this assumption is false and potentially harmful.

In psychiatry, the most notable error theorist – and perhaps the only such theorist – is Thomas Szasz (1960). He argued that psychiatry, as such, is committed to the existence of mental disorders; but, he held, mental disorders do not exist. For him, the very category involved a contradiction in terms. I don’t accept Szasz’s critique, because I don’t believe the notion of a mental disorder is inherently contradictory. My own error theory begins elsewhere: not with psychiatry’s commitment to mental disorders, but with its identity as a branch of medicine.

Psychiatry, as such, is a medical discipline. That is what distinguishes psychiatry from other mental health professions. As such, it is committed to the proposition that some mental health problems are medical problems – problems that fall under the jurisdiction of medicine. I hold, furthermore, that what it is to be a medical problem is to involve some kind of inner dysfunction. Vaguely put, what makes something a medical problem is, in part, that it stems from something in the individual that is not working “as it ought” – something is failing to function appropriately, or functioning contrary to nature. If my reasoning holds, then psychiatry, as such, is committed to the view that mental health problems stem from internal dysfunctions, and are therefore the kinds of things doctors should treat (Garson forthcoming).

I believe, moreover, that this view is generally false. Evidence suggests that many mental health problems reflect normal functioning of the cognitive system. They are ordinary responses to life’s challenges, or instances of evolved cognitive diversity, not pathological breakdowns (Garson 2022; 2024). Because this assumption – that mental health problems are medical disorders – is essential to psychiatry, rejecting it entails rejecting psychiatry as such or seriously restricting its purview. I’ll close by outlining what I see as non-psychiatric mental health interventions that can better assist people in distress.

Alastair Wilson (University of Leeds): Grounding Explanations in Physics

Naturalistic metaphysics looks to physics for guidance about fundamental reality. Uncontroversially, physics tells us about the material contents of the physical world. Can physics also tell us about relations of determination inherent in the physical world? In this talk I defend the view that grounding (the non-causal yet objective explanatory determination of one fact by another) is physical in character: that is, physical theories (properly interpreted) attribute patterns of ground to the natural world. I disentangle some different questions about grounding which we might use physics to answer: the extensional question, the intensional question, and the hyperintensional question. In response to the extensional question, I sketch a scientific epistemology of ground. I then offer a naturalistic account of determination in the physical world that makes use of a single primitive relation of objective explanatory determination - physical grounding - which is mediated by physical principles. I motivate the physical grounding account by describing some preliminary applications: to classical physics, to quantum physics, and to spacetime physics.

Kal Kalewold (University of Leeds): Race Vs. Racialization-Based Explanations in Social Science

This talk addresses the nature and role of racialization in scientific explanations. Debates between racial anti-realists and social realists about race center on whether race is indispensable to the social sciences. Realists argue that race is part of the causal structure of the social world. Consequently, race is necessary to explain a host of phenomena including discrimination, wealth inequality, and health disparities. Recently, anti-realists have charged that the realist case is explanatorily superfluous and circular (Khalifa and Lauer 2021, Singh and Wodak 2023). Social race realists, the objection goes, explain race in terms of racialization and justify race on grounds that it explains the phenomena that constitute racialization.  I defend the realist view from the circularity and superfluity charge. Whether race has explanatory value will depend on considerations that apply to all special science kinds.

Juha Saatsi (University of Leeds): Explaining self-knowledge and deflating the personal/sub-personal distinction

This talk uses open awareness mindfulness meditation as a case-study of Robert Rupert's (forthcoming) naturalistic 'coherent coordinated contribution' account of self-knowledge. Rupert's account deflates the personal/sub-personal distinction that is central to most philosophical accounts of the self and self-knowledge, and reduces self-knowledge to states of the cognitive system that bear information about itself that facilitates improvements in cognition-driven performance by enhancing the system's coherence. This theory, I argue, can provide a good explanation of how meditation can work to improve self-knowledge and catalyse self-improvement without introspection or other personal-level investigation.

The concept of a ‘geometric object’ is central to modern physics and philosophy of physics. And yet, the notion of a geometric object—let alone the history and philosophy of the concept—remains remarkably poorly understood. This symposium, taking place on the centenary of the notion of a geometric object as proposed by the mathematician and physicist J. A. Schouten, aims to expose to philosophers the rich and philosophically fecund history of the geometric object concept, to situate it in is proper context with respect to antecedent work in mathematics (notably, Klein’s ‘Erlangen programme’ for geometry) and modern work on related topics (in particular, modern work by philosophers of physics on ‘natural theories’ and on ‘Kleinian methods’), and to push the programme in exciting new directions with respect to inter alia understanding Noether’s theorems and explicating the metaphysics of gauge fields and spinor fields. As such, the symposium represents a fully integrated project in philosophy, history, physics, and mathematics.

Ruward Mulder (University of California, Irvine): How to be a good geometer: the geometric objects concepts through the eyes of Jan Arnoldus Schouten

This talk will narrate the origins of the concept of geometric object through the eyes of Dutch geometer and expert on Lie groups Jan Arnoldus Schouten, through his published work, correspondence (in particular with E ́lie Cartan), and a range of Dutch inaugural lectures, public talks, and newspaper articles (e.g. Schouten 1920, 1939, 1949, 1951) from the 1920s–1940s (hitherto untranslated and often not publicly available). These archival sources, held at the KNAW, the Centrum Wiskunde & Informatica in Amsterdam, and the per- sonal archive of Gerard Alberts, reveal Schouten’s early grasp of the collapse of Klein’s Erlangen programme right at the advent of general relativity, an appreciation of the constructive interaction between the ‘geometric’ displacement theory and the more analytic ‘object theory’, but also a certain ambivalence towards the abstractness of the geometric object.

The 1920s history of geometric objects is not so obvious as one might think: the mathematical consolidation and generalization of parallel transport, embedding of manifolds, and the associated group structures were all still in flux (cf. Alberts 2000; Cogliati and Mastrolia 2018; Giovanelli 2021). Three partially independent strands coexisted (cf. Struik 1989): (i) Schouten’s programme of generalizing Levi-Civita’s parallel transport, (ii) the Princeton school’s ‘path geometry’ (constant direction curves) associated with Eisenhart, and (iii) Cartan’s exterior differential calculus. March (2025) has recently emphasised that the early concept of geometric objects was initially broader than in later formalizations, including also embedded submanifolds, provided these could be coordinatized and transformed consistently.

This talk highlights two specific aspects of Schouten’s development and assessment of geometric objects. First, the abstractization of geometry. Schouten’s dissertation was systematizing tensor calculus in full generality, with every object—within the Kleinian paradigm—written in explicit component form with respect to specific coordinates: that involves a heavy use of sub- and super-scripts. This gained him some infamy (cf. Alberts 1998, p. 32): Weyl spoke of ‘Orgien des Formalismus’, and Brouwer (the leading Dutch mathematician at the time) is alleged to have said—using a Dutch slur—“Isn’t that the guy who translated geometry into jabbering [hottentots]”. According to Struik (1989), Schouten himself reflects on his dissertation that he could strangle (erdrosseln) its author. Through collaborations with Cartan (Schouten and Cartan 1926a,b) and correspondence with Ricci, Schouten developed and embraced index-free formulations, noting its pragmatic notational reform and intellectual emancipation, while also wary of perpetrating an abstractization of geometry.

Second, in the wake of the above, the debate about the priority of displacement theory versus object theory came to the forefront at the 1936 Oslo conference, following Schouten and Haantjes’ (1936) consolidation of geometric objects to local fields with pointwise numerical components (cf. March 2025). Looking back in later lectures, Schouten explicitly described the ensuing disagreement as driven by an “emotional element” (het emotioneele element): what geometers feel to be genuinely geometric. Schouten here identifies—following the quote above—this emotional element with a changing valuation of the important problems within a field. This characterization reminds one of the changing of paradigms and the associated tension within an epistemic community: dis- agreement about the subject matter of geometry—and what it means to be geometer.

Eleanor March (University of California, Irvine): Geometric objects as the successor to the Erlangen programme

In contemporary philosophical discussions of geometric objects, it is common to see the geometric objects programme mentioned alongside (or even assimilated under) Klein’s Erlangen programme and its modern successors (see e.g. Duerr (2019) and Read (2022)), according to which a geometric structure is characterised in terms of a privileged class of symmetry transformations acting on the space of interest (Wallace 2019). In the modern context of the reformulation of the geometric objects programme in terms of ‘natural bundles’ (Kolar et al. 1993), this connection reappears under the guise of the ‘finite-order theorem’ (Palais and Terng 1977), which states (very roughly) that every natural bundle can be recovered as an associated bundle to some (possibly higher-order) principal frame bundle, and thereby can be specified in terms of a particular group action (the group transformations between frames at some order) on its typical fibre.

Despite this, the historico-conceptual relationship between the Erlangen pro- gramme and geometric objects programme is almost entirely unexplored—and indeed, the extent of the connection between them remains somewhat murky. In some detail: whereas the Erlangen programme sought to characterise geometric structures in terms of a class of symmetry transformations acting on the space of interest, the coordinate transformations which feature in the definition of a geometric object in general do not act on the entire base manifold, but rather only in the neighbourhood of a point. Moreover, these coordinate transformations properly understood do not (in the first instance) act on the points of the manifold itself, but rather on the coordinate charts associated to that manifold.

In this talk, I redress this gap in the literature by providing a thoroughgoing assessment of the Kleinian roots of the geometric objects programme, focussing on the reasoning which led Schouten from the group-theoretic principles of the Erlangen programme to his (1926) formulation of the ‘problem of geometry’, which laid the foundations for the subsequent proliferation of work on geometric objects (Schouten and Haantjes 1936; Schouten and van Dantzig 1935; Veblen 1929; Veblen and Thomas 1926; Veblen and Whitehead 1932). In particular, I discuss the extent to which the geometric objects programme is best understood as a route to ‘localising’ the Erlangen programme, and the extent to which the programme was understood in this way by Schouten and his contemporaries. Along the way, I trace several interesting philosophical connections, including the influence of Cartan’s work as a bridge between the Erlangen programme and geometric objects programme.

Silvester Borsboom (The Mathematics Department and Radboud Center for Natural Philosophy in Nijmegen): Noether’s theorems, invariants, and geometric objects

In 1914–19, Felix Klein gave a series of lectures in Göttingen on the history and development of mathematics. At the same time as he gave these lectures, Klein corresponded with the great Emmy Noether on conservation laws in general relativity: exchanges which, together with those with Hilbert, led to Noether’s famous two theorems (and their converses) regarding the relationships between conserved quantities (‘invariants’), symmetries and constraints. (See Noether (1918) for the original article, Kosmann-Schwarzbach (2011) for the history of Noether’s theorems, Rowe (2022) and Sauer (2025) for background on the Noether–Klein exchange, and Read and Teh (2022) for recent philosophical work on Noether’s theorems.) As such, the Erlangen programme’s influence could also be felt on this cornerstone of modern physics.

These connections are tantalising and worth investigating unto themselves. In addition, however, when considering Noether’s original work (Noether 1918) it is interesting to note that she herself clearly separated her results into two theorems, referred to now as Noether’s first and second theorems. The first is most familiar and provides the link between conservation laws and symmetries. The second is usually applied in the context of gauge theories and states that a localizable symmetry transformation yields a constraint rather than a conserved quantity (for details see Binz et al. (1988)). Interestingly, the concept of a gauge (German: Eich) symmetry was first introduced by Weyl in the context in GR in the same year as Noether’s seminal paper, which hardly seems coincidental, though Weyl re-applied the notion to the context of quantum theory only in 1929 (Brading 2002).

A constraint is in some ways like a localized conservation law, in the sense that both are statements fixing the momentum conjugate to some generalized position coordinate. This naturally raises the question: is Noether’s second theorem just a ‘localized’ version of her first, and does this mirror the ‘localization’ of the Erlangen programme occurring in the work of Schouten and others? In Noether’s first theorem, a global symmetry yields a conserved charge. In the second theorem, the symmetry is promoted to a function of spacetime, and the resulting ‘conservation law’ is upgraded to a redundant degree of freedom, i.e. a quantity whose value must be fixed at every point to ensure gauge invariance. This shift mirrors the transition from the original globalized Erlangen programme to the localized theory of geometric objects, and in turn to the theory of principal fibre bundles. My purpose in this talk is to unpack the relationships between Noether’s two theorems in terms of these analogies with the Erlangen and geometric objects programmes, and to highlight the importance of this interaction for the development of a distinct notion of gauge symmetry as opposed to ‘normal’ symmetries.

James Read (University of Oxford): Geometric objects and Kleinian presentations

In recent years, there has been a significant upturn of interest in ‘Kleinian methods’ for articulating the geometrical commitments of spacetime theories. This resurgence of interest was initiated by Wallace (2019), who argued that a methodology akin to that of Klein’s Erlangen programme—that geometrical structures be characterised as the invariants of certain group transformations—is legitimate and indeed often advantageous as a means of articulating the geo- metrical commitments of a physical theory.

In reaction to Wallace (2019), Barrett and Manchak (2024a,b) have argued that ‘Kleinian presentations’ of generic spacetimes are not possible. (This claim is made in the context of general relativity, but it extends to a significantly broader class of spacetime theories.) And in turn in response to this, Gomes et al. (2024) have argued that “[u]nsurprisingly, a structured space that lacks symmetries cannot be characterised in terms of its symmetry group and therefore cannot be given a [Kleinian] presentation” in the sense of Barrett and Manchak (2024a,b), but that a pointwise version of ‘Kleinian methods’ (i.e., a ‘localisation’ of Kleinian methods to apply point-by-point in the differentiable manifold representing spacetime) can recover generic spacetime structures—even those which lack symmetries.

Quite apart from the many technical considerations and questions which arise regarding these exchanges, at this point a number of historical–conceptual questions arise regarding this recent work:

1.     Which of these two modern approaches (i.e., that of Barrett and Manchak (2024a,b) on the one hand versus that of Gomes et al. (2024) on the other)—if either—is in fact most akin to the original Erlangen programme?

2.     Given that the geometric objects programme itself was motivated by the desire to ‘localise’ the Erlangen programme and therefore provide a powerful and general means of characterising geometrical structures, to what extent do the interactions between these modern authors parallel discussions which arose one century earlier?

3.     Later developments in the geometric objects programme, such as those found in Kucharzewski and Kuczma (1964), draw a number of subtle dis- tinctions which again overlap with recent discussions (and in particular distinctions drawn by Gomes et al. (2024)). So again: to what extent can one draw parallels between recent developments and this earlier work?

In this talk, I will answer these questions, arguing that (i) there are strong parallels between modern discussions of ‘Kleinian methods’ and this earlier work on geometric objects, (ii) one can secure a better appreciation for this recent wave of work (and in particular what is novel in it) by situating it in its historical context, (iii) there has been an evident case of ‘Kuhn loss’ here, and many of the insights into Kleinian methods secured by the geometric objects programme have sadly been lost to time, but can—in the spirit of integrated history and philosophy of science—be brought fruitfully to bear on the modern discussions in order to identify contours which seem not yet to have been recognised, and as such to drive these debates forward in promising new directions.

John Dougherty (The Munich Center for Mathematical Philosophy): Geometric objects and the puzzle of spinors

The geometric objects programme is organised around transformation rules: geometric objects on a space are characterised by the way they transform under changes of coordinates on that space (for modern presentations of this way of understanding geometric objects, see Duerr (2019) and Read (2022)). Beginning with the work of Nijenhuis (1972), this characteristic property of geometric objects was formalized in the theory of ‘natural bundles’, canonical resources on which include Kolar et al. (1993). Natural bundles make precise the sense in which coordinate transformations on some space act on geometric objects on that space.

Although many modern theories of physics can be formulated in the language of natural bundles, many seemingly cannot. For example, Yang–Mills theory is not natural, but rather ‘gauge natural’ (see Fatibene and Francaviglia (2003) and Kolar et al. (1993) and for philosophical discussion March and Weatherall (forthcoming); for more on ‘natural theories’, see Weatherall and March (2025)). Pitts (2012), moreover, has argued that ‘non-linear’ geometric objects (that is, geometric objects with non-linear transformation rules between their coordinate representations) cannot straightforwardly be represented in a coordinate-free formalism. Perhaps the biggest elephant in the room, however, is the case of spinor fields: spinors are not geometric objects (see Pitts (2012)), and yet they represent a significant range of matter (i.e. fermionic fields) in the universe. In the language of natural bundles and generalisations thereof, spinors are neither natural nor gauge natural.

The purpose of this talk is to take up the challenges with which we are confronted by the above issues. In particular, the following tasks will be addressed:

1.     To demonstrate, pace Pitts (2012), that all geometric object fields—whether linear or not—can be incorporated into the framework of natural bundles.

2.     To take up the challenge to Kleinian methods and the geometric objects program presented by spinors by explaining how the transformation rules characterising such objects can be understood in a coordinate-invariant way using natural bundle methods.

3.     To explicate the metaphysical content of natural bundles and theories formulated therewith—which include those theories describing much of the matter in the universe, as mentioned above.

Ultimately the goal of this work is to push forward the limits of the geometric objects programme with respect to invariant presentations of both non-linear geometric object fields and spinor fields, and—even more importantly—to characterise what the world is like according to such theories (to repeat: a matter of considerable metaphysical import, given that much matter in the actual world is represented by spinor fields!).

Philosophical debates over the use of race in medicine are rich and ongoing, intersecting with longstanding concerns in the philosophy of medicine and the metaphysics of race. The existing literature is primarily concerned with population-level disparities, resulting in debates structured around the question of the causal or ontological status of race as producer of health disparities. However, this focus neglects epistemic and ethical features specific to the individual clinical encounter. Differences in the epistemic situation of the epidemiologist versus the clinician, as well as distinct ethical considerations, mean that what makes race informative as a variable in population-level research can come apart from what makes race informative in clinical settings. Therefore, the focus on the population level in the current philosophical literature limits its ability to guide the use of race in the clinic.

Azita Chellappoo (The Open University) and Zinhle Mncube (University of Massachusetts, Amherst): Race, Informativeness, and the Limits of Population-Level Evidence in Clinical Reasoning

The assumption that population-level racial health disparities mean that race is informative in clinical decision-making is widespread in current philosophical debates, particularly in arguments that rely on epidemiological evidence to suggest that race cannot be eliminated from the clinic. However, this assumption neglects distinctive features of clinical reasoning, as well as differences between the epistemic aims of various aspects of clinical practice. Population-level differences in incidence or prevalence of disease between racial groups do not guarantee that race will be informative in guiding clinical decision-making. Furthermore, conditions for when race is likely to be informative differ depending on the clinical aim. Existing literature often focuses on the calculation of individual risk from population-level estimates. However, this is far from the only task that clinicians perform. To illustrate the distinct epistemic roles that race plays in clinical practice, we analyse the use of race in (i) diagnosis and (ii) measurement of function.

With respect to diagnosis, we argue that the iterative character of diagnostic reasoning, and the use of testing and treatment thresholds, limit the informativeness of race in generating and pursuing diagnostic explanations. Turning to race correction in clinical measurements such as spirometry, we argue that these practices are best understood as model-based estimates of current physiological function rather than as forms of diagnosis or risk prediction, and therefore as sites where population-level correlations are most easily mistaken for evidence about individual bodies.

Drawing these cases together, we identify and critique what we call the Population-Level Informativeness Assumption: the presumption that population-level predictiveness licenses individual-level clinical inference. We conclude by proposing a task-sensitive framework for assessing when, if ever, race is epistemically informative in the clinic.

Joanna Karolina Malinowska (Adam Mickiewicz University): Race on Drug Labels: Pharmacogenomics, Epistemic Shortcuts, and Clinical Translation

Pharmacogenomics aims to individualise prescribing by linking drug response to specific genotypes and metaboliser phenotypes. Yet drug labels and prescribing information routinely summarise pharmacogenomic evidence using broad ethnoracial categories (e.g. “White/Caucasian”, “Black/African American”, “Asian”), typically as prevalence claims about allele frequencies or metaboliser status. This paper analyses that labelling practice as a clinical interface where population-level findings are converted into ostensibly actionable cues for decisions about particular patients.

Racialised prevalence statements on labels provide negligible evidential value for patient-level decision-making and, in some settings, can be actively dangerous. The relevant target of inference is genotype/phenotype (e.g. poor, intermediate, or ultrarapid metaboliser status), but coarse ethnoracial categories are weak proxies. Within-category variance in pharmacogenetic traits is large and clinically salient; the standard tripartite frame compresses and selectively omits global variation; and, crucially, in clinical encounters the categories are operationalised through self-identification or clinician perception—neither of which tracks the causal variables that matter for drug response. The predictable result is patient-level misclassification: clinicians may over- or under-estimate risk, pursue inappropriate screening heuristics, or treat category membership as if it were evidence of the relevant genotype, thereby undermining the very promise of precision medicine.

The paper then explains why this epistemically thin, but clinically potent, content persists. It traces a pipeline in which heterogeneous population descriptors in studies are repeatedly forced into legacy classificatory schemes by regulatory guidance, global databases, and clinical data standards, with further simplification driven by comparability pressures and dossier logics. At the point of uptake, these categories can also become economically actionable: label-based proxy reasoning can incentivise the expansion and marketing of pharmacogenomic testing as a “solution” to category-defined risk, even when the label’s category statistics themselves do not justify patient-level inferences. In this way, the label functions not as a neutral summary of evidence, but as a regulatory–clinical artefact that (re)biologises ethnoracial categories while simultaneously shaping clinical practice, patient burden, and resource allocation.

The broader implication for debates on race in the clinic is that evaluating race-based clinical reasoning requires analysing not only population evidence, but also the infrastructures and incentive environments that translate that evidence into patient-facing decisions.

Zeshan Qureshi (University of Cambridge) and Mehrunisha Suleman (University of Oxford): Racial Concordance in Clinical Encounters

The concept of racial concordance, where clinicians and patients share the same racial background, has garnered attention in recent years for its potential to improve healthcare outcomes. Studies suggest that racial concordance can lead to better communication, increased trust, improved patient satisfaction, and better clinical outcomes. However, the underlying assumptions and implications of racial concordance merit deeper philosophical examination, particularly in the context of debates over the nature of race and the role of racial categories in medicine.

This paper explores the philosophical dimensions of racial concordance in clinical encounters, addressing both its epistemic and ethical implications. Central to this discussion is the tension between understanding race as a socially constructed category and its continued relevance in medical practice. If race is indeed a social construct with no biological basis, as many philosophers and social scientists argue, then what justifies the use of racial concordance as a criterion for improving medical care? Does reliance on racial concordance inadvertently reify race, reinforcing the very categories that are often seen as socially and politically problematic?

Drawing on contemporary debates in the philosophy of race, we argue that racial concordance can be understood as a pragmatic strategy in the face of systemic racism and health disparities. Racial concordance may enhance epistemic justice by fostering an environment in which patients' lived experiences are more likely to be understood and validated. This, in turn, may mitigate the epistemic injustices that often arise in racially discordant clinical encounters, where patients' symptoms and concerns are more likely to be misunderstood or dismissed. This seems attractive relative to the current alternative of ongoing disengagement and discrimination experienced by minoritised racial groups.

However, the ethical implications of promoting racial concordance are complex. While it may offer short-term benefits in improving patient care, it also risks entrenching racial categories in ways that could perpetuate long-term inequalities. For instance, if racial concordance is seen as necessary for effective care, this could lead to a segregated healthcare system where racial matching becomes the norm rather than the exception. This possibility raises important questions about the kind of society we are promoting through such practices, and whether the goal should instead be to train all clinicians to provide culturally competent care, regardless of their own racial background.

In conclusion, this paper argues that while racial concordance may serve as a useful tool in addressing immediate disparities in healthcare outcomes, it must be approached with caution. Philosophically, it requires us to navigate the delicate balance between acknowledging the reality of race as a social construct and the pragmatic need to address the lived consequences of racialisation in medical practice. Racial concordance may be necessary in the short term in settings of high levels of distrust and racism that have led to poor outcomes for racialised minority groups. Ultimately, the goal should be to move toward a healthcare system that does not rely on racial concordance but instead fosters an environment where all patients, regardless of race, can receive equitable and empathetic care.

This symposium introduces topics and philosophical questions arising from contemporary quantum biology. Although there has been speculation about the relevance of quantum theory for biological phenomena since the early 20th century, much of the empirical evidence for it has only come in the 21st. The first talk summarises the state of the art of the physics, chemistry, and biology relevant for the most thoroughly investigated quantum biological phenomenon: avian magnetoreception, the ability of certain birds to detect magnetic fields, which they use for navigation during migration. The second talk rebuts three common conceptual objections to the relevance or coherence of quantum biology. Properly understood, skepticism towards reductionism, the putatively different scales for quantum physics and biology, and the physical phenomenon of decoherence do not nullify the possibility and relevance of quantum biology. The third talk proposes a positive account of what makes quantum biology distinctive, involving explanatory connections between non-adjacent levels of reality. After revealing flaws with extant proposals, it indicates how this account might be analogised to other sciences to complicate the general framework of a leveled conception of reality in the metaphysics of science.

Peter Hore (University of Oxford): Avian navigation: A quantum-biological compass sense

Small migratory songbirds are formidable navigators: weighing less than 50 g, they fly extraordinary distances between their breeding and wintering grounds, alone and at night, ultimately with centimetre precision. To find their way they use the sun and the stars, olfaction and landmarks, but they can also perceive the direction of the Earth’s magnetic field. Despite more than 50 years of research, the biophysical mechanism of this remarkable compass sense remains obscure.

The Earth’s magnetic field is feeble – 10 to 100 times smaller than that of a fridge magnet – and interacts very weakly with biological tissue. A classical thermodynamic argument, for example, suggests that the interaction of a molecule with the geomagnetic field is about a million times too small to have a significant effect on a chemical transformation. Nevertheless, there is a class of chemical reaction intermediates, known as radical pairs, whose reactivity can be tuned by exploiting the quantum properties of their electron spins. Pairs of organic radicals are very often formed in a coherent superposition state that can persist for as long as a microsecond, giving ample time for a weak magnetic field to alter the ensuing coherent spin dynamics and hence change the yields of the reaction products.

Theoretical considerations supported by experimental evidence suggest that radical pair chemistry could be at the heart of the avian magnetic compass sense. The primary sensors are thought to be cryptochrome flavoproteins located in photoreceptor cells in the birds’ eyes. Although many of the details are far from clear, it appears that light-induced intra-protein electron-transfer reactions produce a magnetically sensitive flavin-tryptophan radical pair and hence a signalling state of the protein whose quantum yield encodes the direction of the magnetic field. Over the last 20 years, the cryptochrome hypothesis has become a central part of the field of Quantum Biology.

The presentation will start with a brief outline of the physics, chemistry, and biology behind the cryptochrome hypothesis, emphasizing the quantum aspects of the mechanism. This will be followed by a description of on-going behavioural tests on Eurasian blackcaps (Sylvia atricapilla) that arguably provide the strongest evidence to date for a cryptochrome-based mechanism. It appears that calculations of the quantized energy levels of flavin-containing radical pairs predict the ability of a migratory bird to orient using its magnetic compass.

Margarida Hermida (University of Salzburg): Quantum biology, scale, and physical explanations in biology

In philosophy of science, quantum biology sits somewhat uneasily within a traditional conception of scientific fields. There are three reasons for this. Firstly, there is, among most philosophers of biology, a general distrust concerning physical explanations in biology. Secondly, there is the idea that quantum physics cannot explain biological phenomena because it works at the wrong scale. And thirdly, it is often thought that even if quantum phenomena do take place in biological systems, they are most probably quickly lost due to decoherence, hence cannot have biological significance. In what follows, I address each of these worries in turn, and show that quantum biology has important philosophical implications.

Fears of unchecked reductionism loom large when considering the importance of physical explanations in biology, but the worry is misplaced. Despite occasional claims to the contrary, physical explanations are pervasive in many areas of biology, but they do not threaten its autonomy as a discipline. On the contrary, the progressive unification of biology with the physical sciences, particularly in the fields of biochemistry, molecular biology, and biophysics, is responsible for much of its extraordinary success from the mid-twentieth century onwards.

The worry that events at the quantum scale ordinarily have no effect on living things is rebutted by discoveries in quantum biology, such as the relevance of quantum tunnelling for respiration, quantum coherence in photosynthesis, and radical pair formation in bird navigation systems. Quantum phenomena are deeply implicated in how living things obtain free energy to power their activities; hence they are fundamental for biology. However, something like a ‘principle of the primacy of molecular explanations’ seems to be operational in biology, with most explanations sought at the scale of macromolecules. Pushback against this implicit principle has mainly come from the anti-reductionist and organicist camp, whose proponents emphasise the importance of the whole organism in biological explanations. But as it turns out, structures and processes both above and below the molecular scale can be explanatory. It really is an empirical question which entities and scales are explanatorily relevant concerning each particular biological phenomenon.

Finally, the environment of the cell is thought to be too wet and crowded for quantum phenomena to be of relevance, since they are thought to be immediately lost to decoherence. But the interior of a cell is not a structureless soup of chemicals. Where quantum events are important, they are spatially constrained to particular locations within the cell, where interactions are restricted; and conditions for their occurrence are very precisely regulated. For example, the distances between adjacent redox centres within respiratory complexes are maintained within tight margins by natural selection, given the crucial – literally life-and-death – importance of respiration. This in turn has led to co-evolution between mitochondrial and nuclear genomes, with far-reaching evolutionary consequences for complex life on Earth.

Samuel C. Fletcher (University of Oxford): What is quantum biology?

Quantum biology is often taken as an interdisciplinary field of research that applies models and concepts from quantum mechanics to explain biological phenomena. After recalling some characteristic putative examples of quantum biological phenomena, including avian navigation and photosynthesis, I show how extant accounts of what quantum biology consists in have not adequately characterised it. I propose my own account, arguing not only that it remedies the deficiencies of other proposals, but also that it suggests a descriptive and normative research program in the philosophy of interdisciplinary science. The key to these is the failure of a kind of explanatory screening-off between levels of reality that may be of independent interest.

So far, quantum biologists have attempted to characterise their field in one of three ways. First, they delimit it by biological phenomena for which classical physics is inadequate. But the phenomena to be explained are biological, not purely physical, and is not generally in the domain of classical physics. Second, they describe it as limning how Nature’s knowledge of quantum theory gives its organisms a biological advantage. But that too presupposes erroneously a pre-quantum or classical comparator which is absent. Third, they list biological phenomena whose explanation or description deploys characteristically quantum phenomena, such as tunnelling or entanglement. But there is little agreement on what should appear on this list.

My positive proposal focuses instead on the way explanations are improved by including quantum mechanical facts: I say that quantum biology is an interdisciplinary research field concerned with biological phenomena, explanation of which is improved by adding facts from quantum physics, because explanation only with the resources of biology does not render the phenomena unsurprising (or is otherwise unsatisfactory). So it goes, for instance, with the birds: Explanation of their ability to migrate is improved by including facts from quantum physics. Merely appealing to avian magnetoreception is not satisfactory because this is such a surprising ability, itself in need of explanation.

The phenomena that quantum biology concerns instantiate a pattern that could be of greater prevalence than previously recognised. There is an implicit expectation in much of science that the disciplines study levels of reality organised in a hierarchy with the following feature: facts at an intermediate level screen off facts at a lower level from those at a higher level. For instance, chemical facts screen off physical facts from explanations of biological facts, in the sense that once one conditions on the chemical facts, conditioning further on physical facts make no difference to the explanations of biological facts. Quantum biology shows that this screening off does not always obtain, as it concerns precisely situations where conditioning on (quantum) physical facts improves explanations of biological facts. Further investigation is needed to determine how widespread this phenomenon is and the extent to which it challenges central or peripheral features of the leveled conception of reality.

Before asking whether aesthetic sensibilities promote or undermine science, we must first understand what such sensibilities do (or could do) for scientific knowledge-generation. The aim of this symposium is to identify and characterise the plurality of functions that aesthetic sensibilities, engagement, emotions, and judgements might play across various scientific contexts and stages of inquiry. In moving beyond familiar discussions of theory-choice and beauty, the papers collectively develop varied perspectives on aesthetics in science, clarifying when aesthetics facilitates creativity, guides evaluation, shapes research agendas, and reinforces the influence of social values in science.

Luana Poliseli (University of Manchester): Overcoming the “fear “of aesthetic values in science by understanding its distinct dimensions

Aesthetic engagements in science are diverse, conceptually plural, and often difficult to identify, extending well beyond traditional associations of beauty (i.e. McAllister 1996) with theoretical or experimental success (see Ivanova & French 2020; Ivanova & Murphy 2023). Drawing on contemporary debates in the aesthetics of science (Ivanova 2017), I argue that aesthetic features operate at multiple stages and levels of scientific activity, from momentary felt responses to methodological norms and communicative practices. By (non-exhaustively) grouping kinds of aesthetic engagement according to their purposes and characteristics, I seek to advance understanding of how they contribute to scientific reasoning and decision-making. By intentionally situating these engagements within the history and philosophy of biology and contemporary transdisciplinary research practices, I show that aesthetic and artistic dimensions have been perennially present in scientific contexts, particularly in relation to nature and the life sciences, and have evolved alongside modern scientific practices.

Expanding on Poliseli (2024), I’ll identify distinct dimensions of aesthetic engagement in and of science: emotional, methodological, sensorial, intuitive, and artistic. Rather than treating science as a purely logical or objective enterprise, I will underscore the ambiguities, heterogeneity, and cognitive limitations inherent in knowledge production, underscoring that scientific understanding is shaped not only by empirical and theoretical considerations but also by sensuous, affective, and interpretive experiences. In doing so, it repositions aesthetics as central to reflections on scientific practice and argues for its potential incorporation into scientists’ epistemic toolboxes, supporting more inclusive and responsive ways of understanding research practices in contemporary contexts.

Alice Murphy (University of St Andrews): Aesthetic and Social Values in Science: The Case of “Invasive” Species

In this talk, I show how aesthetic values play a significant but largely overlooked role in Invasion Science. “Invasive” species are widely regarded as a major threat to ecosystems, often linked to biodiversity loss and the extinction of native species. Yet Invasion Science, the field dedicated to studying their causes, impacts, and management, has come under increasing scrutiny. Critics have questioned the emphasis on “nativeness” in assessing a species’ potential harmfulness, the use of militaristic language employed in scientific literature, policy guidance, and public awareness projects, and whether the purported threat of introduced species has been overstated.

A small body of literature in philosophy of science has uncovered social values embedded in the theory, practice, and language choices of Invasion Science (Larson 2005; Eliott-Graves 2016; Guiaşu and Tindale 2018; Frank 2021). But the presence of aesthetic values, and their function in the discipline, has received little attention. I seek to remedy this. Aesthetic concerns can be traced back at least as far as Elton’s (1958) The Ecology of Invasions by Animals and Plants and continue to shape the field. Aesthetic considerations arise in various aspects of Invasion Science, motivating research agendas, legitimising management initiatives, and shaping assessments of ecological change in scientific and public-facing contexts.

This, I argue, demonstrates how aesthetic values are intertwined with, and can reinforce, social values. More broadly, this case study illustrates how aesthetic values and judgements function across multiple stages of scientific activity beyond familiar discussions of theory choice. I argue that attending to these roles is crucial for accounts that seek to address the “new demarcation problem” (Holman and Wilholt 2022), i.e., the issue of isolating legitimate from illegitimate value-influence.

Mariona Miyata-Sturm (Oxford University): The Psychological Role of Aesthetic Evaluation in Science

What can explain the fact that aesthetic considerations often support knowledge acquisition? I argue that a ‘metacognitive’ account makes the best sense of the epistemic function(s) of aesthetics in science, according to which aesthetic feelings are feedback from working with theory and evidence which reflect certain epistemically relevant features of the theory and how it combines with available evidence.

Metacognitive processes monitor first-order cognitive states and processes like remembering, perceiving, or decision-making, and are often based on feelings, like feelings of error or understanding (Arango-Munoz and Michaelian 2014; Ackerman and Thompson 2017). Intriguingly, many of the features which trigger positive metacognitive feelings appear in descriptions of what scientists find beautiful, such as simplicity and symmetry. This has inspired a handful of philosophers and psychologists to explain aesthetic feelings as upshots of metacognitive processes (Schwarz 2018; Todd 2017; Miyata-Sturm 2024). Here I present a new development in this research programme, explaining the connection between theoretical virtues, specific feelings, and the features which trigger them, and arguing that scientists’ aesthetic evaluation is typically based on how theory and evidence combine and not solely on intrinsic features of theories. My hypothesis is that scientists’ aesthetic experiences are often epistemically useful because they are metacognitive feedback responding to features of epistemic value and signals of the quality of our cognitive engagement with theory and evidence.

This account can explain the persistent tendency across the sciences to rely on aesthetic considerations in theory evaluation (McAllister 1996; Vaidyanathan et al. 2023). Despite the lack of a direct connection between beauty and truth, aesthetic feelings can play the typically useful epistemic role that scientists often expect them to play because they are triggered by epistemically relevant features of a theory and the evidence that bears on it, such as richness of detail and that it enables an easy overview. The epistemic value of aesthetic feelings, and the judgements that are based on them, is due to the moderate reliability of the metacognitive processes generating these feelings (Koriat 2008). By making metacognitive monitoring central to the explanation of the epistemic value of aesthetics in science, we can explain why aesthetic evaluation appears in epistemic contexts in the first place, how aesthetic feelings can be generally useful signals of epistemic value, and why they nonetheless relatively often lead scientists astray. I illustrate the explanatory potential of the account with examples from the earth sciences, showing how it shines a new light on the role aesthetic feelings play in scientific reasoning and decision-making.

Adrian Currie (University of Exeter): (Revolutionary) Science Should Be Funny

“Solving a puzzle or riddle, for example, provides us with pleasure and even delight as we find a new way of seeing or thinking. When the feeling of mirth is experienced in subversive humor, the audience enters play mode, if it is not already in it, and is more open to challenges to their fundamental beliefs… “ (Kramer 2020, 160).

I’m going to argue that if we take the philosophy of humour seriously, there’s a clear role that humour could, should, and maybe sometimes does, play in science. There’s a lot of work on humour in science. Most of it concerns communication and pedagogy (e.g. Berge 2017). There’s also a bit on how humor is used to police group dynamics between scientists (Wylie 2021). But humor (I think) should play a more direct role in knowledge-production as well. Incongruity theories have it that humour operates by subverting expectations in a way that generates the emotion of mirth (Morreall 1987). These subversions, it has been argued, can lead to revelation: jolting us from everyday, mundane thinking (Marra 2017). They can further do so in a safe way: humour invites us into a space to play, potentially enabling creativity. I’ll argue that when scientists need to think outside of ‘normal’ science: developing new theories, interventions, models and so on, it seems plausible in principle that humour could facilitate ‘new ways of seeing or thinking’ via subversive revelation. This also suggests a connection between the emotion of mirth and Helen de Cruz’s discussion of wonder in science (de Cruz 2024). Perhaps, like wonder, mirth is sometimes an epistemic emotion which enables the rethinking and resituating necessary for revolutionary science.