British Journal for the Philosophy of Science

@thebjps.bsky.social

5957 Followers 1022 Following

Latest issue: https://www.journals.uchicago.edu/toc/bjps/current BJPS Review of Books: https://www.thebsps.org/reviewofbooks/ BJPS Short Reads: https://www.thebsps.org/short-reads/ Podcast: https://podcasters.spotify.com/pod/show/bjps

A man and a woman stand in front of a PowerPoint presentation
A sign for University of Chicago journals next to leaflets for the BJPS
A woman stands in front of a PowerPoint, which displays the title of her talk
Robin Hendry wearing a blue shirt and gesticulating in the approved academic manner. He looks happy enough.
A man stands with his hand on the face in front on a PowerPoint slide calling for people to defend Exeter University
One back and white dog and one white and brown dog sit on a couch, with a sign reading ‘ask me anything’ between them
BSPS Keynote 11.30am, Thursday 23 July (UK time) Eleanor Knox, 'The Scale and Context-Relativity of Spacetime Structure' ABSTRACT. Many theories of quantum gravity appear to imply that the spacetime of general relativity will be effective or emergent. This talk explores the consequences of such a view, and argues that, on any plausible account on which general relativistic spacetime emerges, other relativistic and non-relativistic spacetime structures will emerge in relevant contexts and at relevant scales. This has surprising and under-appreciated consequences—for example, that, at certain scales, binary black holes live in Newtonian spacetime.
BSPS Keynote 11.30am, Wednesday 22 July (UK time) Tarja Knuuttila, 'Revisiting Artefactualism: The Model–Target Coupling Displaced' ABSTRACT. Models are epistemic artefacts, constructed and put to many uses, including understanding, prediction, data mining, and design. In many cases, the key to their epistemic value is erotetic: they are built to address pending theoretical or empirical questions, and that is where their epistemic value lies, not in a model’s correspondence to a determinable real-world target system. I defend this thesis and dispel a double bind in the account’s recent reception. It has been claimed, on the one hand, that the artefactual account is covertly representational and, on the other hand, that as a non-representational account, it cannot supply the content that property attribution and model-world comparison require. These charges pull in opposite directions and cannot both land. The artefactual account does attribute content to models, located in the interpreted representational modes and media used in their construction, but does not assume that the epistemic value of models is due to the representational relationship between model and target. Artefactualism stays neutral between different theories of representation, though closer to deflationary, inferentialist accounts. No longer organised around the model-target relation as its unit of analysis, the account clarifies the use of cross-domain model templates, the modal reach of modelling, and models without determinable targets, such as large language models and synthetic models.
Presidential address Robin Hendry Metaphysics and Science: A Call for History 4pm, Tuesday 21 July (UK time) ABSTRACT. Scientists and philosophers down the ages have argued that there should be no overlap between science and metaphysics. Others have claimed, to the contrary, that science needs a foundation that only metaphysics can provide. I would challenge both these lines of thought, but the first is the main target of this talk. Consider the following claims, which are widely accepted: (1) The Universe was created in a great explosion whose effects remain visible; (2) Biological species are the product of a chancy process involving mutation and natural selection; (3) What happens in our minds is the outcome of neural processes. Not only are these all metaphysical (in a broad sense): scientific research is also relevant to them, even if they aren’t simple to decide. If we understand metaphysics broadly to be concerned with general and abstract questions of existence and reality, science is clearly concerned with metaphysics. But how good is science at answering metaphysical questions like these, and under what conditions? I address that question historically, by looking at two well-known failures (causal determinism and Newtonian mechanics; a claimed refutation of vitalism); and some important but neglected successes (the achievements of 19th-century chemistry). I conclude by arguing that metaphysical claims can be supported by science if they have had an active, intimate and indispensable heuristic and conceptual role in the development of science. Otherwise, it’s too soon to tell.
ABSTRACT. In general, visual experiences represent determinately. And visual experiences, generally, represent properties of distal objects like their colour, shape, and size, but they do not, generally, represent properties of proximal states like that of incoming light or the retina. By making perceptual constancies central to perceptual representation, Schulte extends Neander’s causal–informational teleosemantic theory in order to accommodate these facts. However, by appealing to the psychophysics and chemistry of how light-related properties interact to produce stimulation to the visual system and how the visual system processes such input to produce experiences, I argue that Schulte’s theory fails to accommodate the facts of distality and determinacy.
ABSTRACT. Mental capacities for perceiving, remembering, thinking, and planning involve the processing of structured mental representations. A compositional semantics of such representations would explain how the content of any given representation is determined by the contents of its constituents and their mode of combination. While many have argued that semantic theories of mental representations would have broad value for understanding the mind, there have been few attempts to develop such theories in a systematic and empirically constrained way. This article contributes to that end by developing a semantics for a ‘fragment’ of our mental representational system: the visual system’s representations of the bounding contours of objects. At least three distinct kinds of composition are involved in such representations: ‘concatenation’, ‘feature composition’, and ‘contour composition’. I sketch the constraints on and semantics of each of these. This account has three principal payoffs. First, it models a working framework for compositionally ascribing structure and content to perceptual representations, while highlighting core kinds of evidence that bear on such ascriptions. Second, it shows how a compositional semantics of perception can be compatible with holistic, or Gestalt, phenomena, which are often taken to show that the whole percept is other than the sum of its parts. Finally, the account illuminates the format of a key type of perceptual representation, bringing out the ways in which contour representations exhibit domain-specific form of the sort that is typical of structured icons such as diagrams and maps, in contrast to typical discursive representations of logic and language.
ABSTRACT. Much attention in philosophy of science has been devoted to explicating highly prized concepts such as explanation, theory, or model, among many others, resulting in a plurality of nuanced philosophical accounts (for example, the deductive-nomological account, the causal account, the unification account, and the mechanistic account of explanation). The rationale for this enterprise is to be found in the central epistemic roles that such concepts are taken to play in science. But do these concepts actually play such significant roles? In this article, we investigate the actual usage of epistemic concepts in the practice of science by analysing terminological occurrence patterns in scientific publications. Narrowing down the study to six major epistemic concepts (theory, model, mechanism, explanation, understanding, and prediction), we use text-mining methods to quantify actual terminological usage and relationships in a corpus of 73,771 full-text scientific articles of the biological and medical sciences (from the BioMed Central database). The resulting terminological cartographies partly validate select philosophical intuitions, but also suggest notable differences between philosophical reconstructions and the actual roles that epistemic concepts appear to be playing in the scientific discourse. We also investigate the incidence of disciplinary context.
ABSTRACT. This article defines and defends a notion of teleological function that is fit to figure in explanations concerning how organic systems, and the items that compose them, are able to perform certain activities, such as surviving and reproducing or pumping blood. According to this notion, a teleological function of an item (such as the heart) is a typical way that items of that type contribute to some containing system’s ability to do some activity. An account of what it is for an item to contribute to a containing system’s ability to perform an activity is provided. I argue that the view acquires its normative status in virtue of the fact that it obeys a function–accident distinction and obeys a function–dysfunction distinction; that the view is ahistorical; and, that its ahistoricity provides it with an advantage over one of its main competitors.
ABSTRACT. We develop and apply a multi-dimensional account of explanatory depth towards a comparative analysis of inflationary and bouncing paradigms in primordial cosmology. Our analysis builds on earlier work due to Azhar and Loeb that establishes initial conditions fine-tuning as a dimension of explanatory depth relevant to debates in contemporary cosmology. We propose dynamical fine-tuning and autonomy as two further dimensions of depth in the context of problems with instability and trans-Planckian modes that afflict bouncing and inflationary approaches, respectively. In the context of the latter issue, we argue that the recently formulated trans-Planckian censorship conjecture leads to a trade-off for inflationary models between dynamical fine-tuning and autonomy. We conclude with the suggestion that explanatory preference with regard to the different dimensions of depth is best understood in terms of differing attitudes towards heuristics for future model building.
ABSTRACT. Grice, Lewis, and Skyrms propose similar distinctions between kinds of meaning. The meaning of terms in human language, as Lewis and Skyrms had it, is ‘conventional’. Skyrms presents models showing how it is possible for conventional meaning to evolve in a population without reliance on pre-existing meaning. But one might think of conventionality as coming in degrees, based on whether the evolutionary process begins with ‘natural saliences’. We propose a theory of natural salience and several extensions of Skyrms’s models to capture this notion. These models reveal that natural saliences can hinder, as well as help, the evolution of language.
ABSTRACT. Drawing on a passage from Ramsey’s ‘Truth and Probability’, we formulate a simple, plausible constraint on evaluating the accuracy of credences: the calibration test. We show that any additive, continuous accuracy measure that passes the calibration test will be strictly proper. Strictly proper accuracy measures are known to support the touchstone results of accuracy-first epistemology, for example, vindications of probabilism and conditionalization. We show that our use of calibration is an improvement on previous such appeals by showing how it answers or sidesteps problems that have been raised for previous work in this area.
ABSTRACT. Most existing proposals to explain the temporal asymmetries we see around us are sited within an approach to physics based on time evolution, and thus they typically put the asymmetry in at the beginning of time in the form of a special initial state. But there may be other possibilities for explaining temporal asymmetries if we don’t presuppose the time-evolution paradigm. In this article, I explore one such possibility, based on Kent’s ‘final-measurement’ interpretation of quantum mechanics. I argue that this approach potentially has the resources to explain the electromagnetic asymmetry, the thermodynamic asymmetry, the coarse-graining asymmetry, the fork asymmetry, the record asymmetry, and the cosmological asymmetry, and that the explanations it offers may be better than explanations appealing to a special initial state. My hope is that this example will encourage further exploration of novel approaches to temporal asymmetry outside of the time-evolution paradigm.
ABSTRACT. It is an underappreciated fact within the philosophical literature on laws of nature that many scientific laws require the aid of a supporting cast of additional modelling ingredients (such as boundary conditions, material parameters, interfacial stipulations, rigidity constraints, and so on) in order to perform their traditional role in scientific inquiry. In this article, I suggest that this underappreciated fact spells trouble for some recent reformulations of Lewis’s best systems account of laws of nature. Under the auspices of ‘pragmatic Humeanism’, several philosophers have recently argued that the criteria of strength and simplicity that lay at the heart of Lewis’s original formulation should be replaced with alternatives that are more sensitive to the role that laws play in scientific practice. Although the criteria that these philosophers put forward differ in a variety of ways, they are primarily concerned with the ability of laws to furnish us with predictions and encode information. This, I suggest, is a problem. If it is true that many scientific laws do not on their own perform some of the roles with which they are traditionally associated, then they are unlikely in isolation to make meaningful contributions to the predictive strength of a system or encode information about particular systems. Such laws are thus unlikely to end up in the best system, and so these accounts will have trouble conferring lawhood upon them.
ABSTRACT. Dimensional approaches in psychiatry involve a boundary between health and pathology understood as a difference in degree rather than kind, so that no qualitative distinction can be made between health and pathology. However, this approach raises epistemological and ontological issues. First, the major sources of evidence used are connected to different conceptualizations of how dimensional traits extend across health and pathology. Second, it relies on two core assumptions: that there is continuity from health to pathology at the symptomatic level, and that such continuity reflects an underlying continuity in the genetic liability for pathological conditions. Third, this view implies a linear relationship between genotype and phenotype. Drawing on epigenetics and developmental biology, we argue instead that mental disorders are stable, dynamic clusters that differ qualitatively from ‘healthy states’. This has both theoretical advantages and implications for clinical practice
ABSTRACT. This article develops and advocates a rule for assigning self-locating credences in quantum branching scenarios, called indexed branch-counting. It is argued that indexed branch-counting can be justified on both accuracy-theoretic grounds and on the grounds that it satisfies a requirement of exchangeability for probability assignments. Since indexed branch-counting diverges from the Born rule, this poses trouble for Everettian approaches to probability. The article also addresses a common argument against branch-counting, namely, that the rule is incoherent in light of putative vagueness in the number of branches. Finally, the article addresses a recent proposal from Saunders that aims to reconcile branch-counting with the Born rule, arguing that the proposal faces challenges.
Table of Contents Crossing the Threshold: An Epigenetic Alternative to Dimensional Accounts of Mental Disorders By Davide Serpico and Valentina Petrolini Conquering Mount Everett: Branch Counting Versus the Born Rule By Jake Khawaja Laws of Nature and Their Supporting Casts By Travis McKenna Disappearing Without a Trace: The Arrows of Time in Kent’s Solution to the Lorentzian Quantum Reality Problem By Emily Adlam Consequences of Calibration By J. Robert G. Williams and Richard Pettigrew Naturalizing Natural Salience By Jacob VanDrunen and Daniel A. Herrmann Explanatory Depth in Primordial Cosmology: A Comparative Study of Inflationary and Bouncing Paradigms By William J. Wolf and Karim P. Y. Thébault Mechanistic Explanations and Teleological Functions By Andrew Rubner Charting the Territories of Epistemic Concepts in the Practice of Science: A Text-Mining Approach By Christophe Malaterre and Martin Léonard Contours of Vision: Towards a Compositional Semantics of Perception By Kevin J. Lande Distal Content in Informational Teleosemantics: Challenges from Colour Constancy and Colour Chemistry By Lance Balthazar
ABSTRACT. The application of the notion of `observable' from gauge theory to diffeomorphism-invariant theories—and, most relevantly, to general relativity—has led to numerous conceptual and technical issues for interpreting classical theories and building quantum versions of them. I distinguish between two senses of gauge transformation, local and global, and argue that the notion of `observable' appears more naturally in the local sense of gauge transformation. Then I argue that diffeomorphism invariance can be understood as a gauge symmetry only from a global point of view and hence that the concept of `observable' applies only in a restricted manner. This has the consequence that some popular claims in the literature, such as that the physical content of diffeomorphism-invariant models is encoded in correlations, are unfounded.
Khalifa, Millson & Risjord have proposed a ‘thoroughgoing’ inferentialist account of scientific representation, claiming this account overcomes the ‘smuggling’ objection: that inferentialist accounts require implicitly relying on some more fundamental relation between the model and target, such as similarity, isomorphism or denotation. Their proposal attempts to avoid the objection by exporting Brandom’s expressivist theory of reference from the philosophy of language to the case of scientific representation. I focus only on denotation here. Using the DEKI account of scientific representation, which assumes denotation to be a crucial element of representation, I critically analyse this proposal. I argue that their proposal does not successfully address the objection and that it is also open to further objections. I conclude that denotation is a required element for a working account of scientific representation and that thoroughgoing inferentialism remains unsatisfactory.
ABSTRACT. As well as deterministic and probabilistic causal species, there’s a third hitherto unanalysed species: underdeterministic causal phenomena (non-deterministic yet non-probabilistic). I develop the prerequisite framework for any such theory. This framework is like the deterministic structural-equations framework but with one consequential difference—an equation can return multiple values. This allows me to define causal possibility and necessity and corresponding notions of interventionist might- and would-counterfactuals. I also define conditional independence, which obeys the graphoid axioms, and prove that underdeterministic models satisfy the causal Markov condition. This underdeterministic framework can causally model decision-making under bounded uncertainty, games with multiple equilibria, infinite fair lotteries, and any other non-deterministic situation where indeterminacies are essentially non-probabilistic or where there is reason not to use probabilities.
ABSTRACT. Organisms and other biological entities get things wrong, while the entities of pure physics, such as atoms and inorganic molecules, do not make mistakes. We set out a novel framework for understanding biology and its demarcation from physics: mistake-making. We distinguish biological mistakes from mere failures. We then propose a definition of mistakes that is compatible with various views about functions. The definition of mistake-making is agential, since mistakes do not just happen—in the sense analysed here—but are made. We offer a definition of the minimal biological agent. Two key features of our theory are, first, that it is a supplement to, not a replacement for, existing general frameworks within which biology is understood and practised. Second, it is designed to be experimentally productive, and we end with case studies showing mistake theory can be useful in the generation of research questions and novel hypotheses for biologists.
Average number of referee invitations per paper: 6 Average time for referees to return a report: 33 days Average days from first referee invitation to receipt of both referee reports: 60 days
ABSTRACT. It is often thought that conventional theories of particle physics do not deserve realist commitment, despite their outstanding empirical success. A number of effective realists argue that we should distinguish between the low- and high-energy claims of particle theory; that we can and should be realist about the former but not the latter. I present a reductio ad absurdum against the naive extension of this proposal to the most empirically successful theories, such as quantum electrodynamics (QED). By considering two replies to this argument, I distinguish two forms of effective realism. A conservative form hews closely to traditional forms of realism and the resources of this tradition allow conservative effective realism to avoid the reductio; however, this form of effective realism is left without a positive account of QED. A more radical form can account for QED, but it requires substantial development, along with a revision of the terms of the realist debate.
ABSTRACT. According to mainstream philosophical views causal explanation in biology and neuroscience is mechanistic. As the term ‘mechanism’ gets regular use in these fields it is unsurprising that philosophers consider it important to scientific explanation. What is surprising is that they consider it the only causal term of importance. This article provides an analysis of a new causal concept—it examines the cascade concept in science and the causal structure it refers to. I argue that this concept is importantly different from the notion of mechanism and that this difference matters for our understanding of causation and explanation in science.
ABSTRACT. Sometimes when a physical theory is abandoned, its theoretical entities continue to be used. What should scientific realists say about these entities? The usual answer is that they are useful fictions. We offer a different answer. We start by distinguishing horizontal reduction from vertical reduction. We describe a verticalization procedure that transforms horizontal reductions into vertical reductions. We identify a sense in which the higher-level theory describes distinct subject matters from the lower-level theory, enabling in certain cases the higher-level theory to retain distinctive explanatory power even in the presence of reduction. We suggest that theoretical entities from superseded theories should be retained in a scientific realist account when, reinterpreted as higher-level abstractions, those theories and their entities perform distinctive work in providing the best explanation for less fundamental phenomena of interest. Or: a good relict is an emergent relict.
ABSTRACT. The mechanistic approach has been proposed as a way of integrating different levels of explanation in psychiatry. It has also been put forward as a non-essentialist and yet realist model for psychiatric classification, which does not privilege a specific level at which the relevant mechanisms are to be found. I argue that the mechanistic view of inter-field integration cannot address the challenge of integrating both variables and mechanisms of different kinds, which in turn casts doubt on its potential to provide a foundation for psychiatric classification. The main obstacle, I contend, is the variations in styles of mechanistic modelling one finds in the fields concerned with different levels of explanation. I outline an alternative view of explanation and classification of mental disorders, which is still mechanistic in spirit, but does away with a substantive ideal of integration.
ABSTRACT. Some cultural evolution theorists claim that humans tend to imitate the traits of people who are similar to themselves. They also suggest selection shaped human psychology to attend to similarities and weigh them when learning from others. Little formal modelling explores the conditions under which this argument is valid. I develop models that suggest that that it depends on assumptions about the function of social roles. On the default view, social roles facilitate coordination but the model is not very robust in terms of the initial conditions, parameter settings or population structure. But if social roles facilitate the division of labour, then similarity-biased learning evolves under a wide range of conditions. These results improve our understanding inequality. Evolutionary bargaining models have been proposed to explain inequality. These models use similarity-biased learning assumptions. I suggest improvements to that research programme in light of these results.
ABSTRACT. It has become popular to start with Lewis’s BSA of natural law and then develop its Humeanism in a more pragmatic spirit by putting greater weight on the laws’ role in helping us to meet our goals despite our epistemic limitations. I examine three key versions of pragmatic Humeanism and argue they all encounter significant obstacles in accounting for the laws’ relation to counterfactuals. Some obstacles appear when we examine whether not only the laws’ truth but also their lawhood is preserved under counterfactual antecedents, along with whether the laws’ truth is preserved under nested counterfactuals. Others arise from the roles that counterfactuals play in science, the nature of metaphysical necessity, and the relation between metaphysical and scientific explanation. Although I cannot offer a complete evaluation of pragmatic Humeanism’s success here, I conclude that pragmatic Humeanism does not yet deserve credit for nicely explaining the laws’ relation to counterfactuals.
ABSTRACT. Recent experimental evidence suggests that human subjects use a win–stay/lose–shift with inertia learning dynamics (WSLSwI) to establish signalling conventions in the context of Lewis signalling games. Here we consider the virtues and vices of this low-rationality dynamics. Most saliently, WSLSwI is much faster than simple reinforcement learning in establishing conventions. It is also more reliable in producing optimal signalling systems. And it exhibits a high degree of the stability characteristic of a reinforcement dynamics. We consider how increasing inertia may increase speed in finding optimal conventions under this dynamics, the virtues of cognitive diversity, and how the dynamics meshes with high-order rationality. In brief, WSLSwI is extremely well suited to establishing conventions. That human subjects use it for this purpose is a remarkable adaptation.
British Journal for the Philosophy of Science Volume 77, Issue 1 Table of Contents In Defence of Science By Jochen Briesen The Efficacy of Human Learning in Lewis Signalling Games By Calvin T Cochran and Jeffrey A Barrett Can Pragmatic Humeanism Account for the Counterfactual Invariance of Natural Laws? By Marc Lange When Is Similarity-Biased Social Learning Adaptively Advantageous? By Daniel Saunders Challenging the Mechanistic View of Integration in Psychiatry By Caterina Marchionni Theoretical Relicts By Katie Robertson and Alastair Wilson Cascade Versus Mechanism By Lauren N Ross Effective and Selective Realisms By John Dougherty Biological Mistakes By David Oderberg, Jonathan Hill, Christopher Austin, Ingo Bojak, François Cinotti, and Jonathan M Gibbins Editors’ Choice Article: On the Ecological and Internal Rationality of Bayesian Conditionalization By Olav B Vassend The Underdeterministic Framework By Tomasz Wysocki
ABSTRACT. According to Bayesians, agents should respond to evidence by conditionalizing their prior degrees of belief on what they learn. A major aim of this article is to demonstrate that there are common scenarios where Bayesian conditionalization is less rational—from both an ecological and an internal perspective—than other theoretically well-motivated belief updating strategies, even in simple situations and even for an ‘ideal’ agent who is computationally unbounded. The examples also serve to demarcate the conditions under which Bayesian conditionalization may be expected to be ecologically optimal. A second aim of the article is to argue for a broader notion of rationality than what is typically assumed in formal epistemology. On this broader understanding of rationality, classical decision theoretic principles such as expected utility maximization play a less important role.
ABSTRACT. Much of the literature on values in science is framed around the idea of a shift in the status of the so-called value-free ideal (VFI) for science: having been widely accepted in the past, it has now come to be thoroughly rejected. Instead, the commonly asserted view now is that there is a new problem for philosophy of science to address, namely, distinguishing appropriate from inappropriate value-influence on science. While understanding the many intricate ways that values can influence science is an important task for philosophy, we argue that this framing—that there is a single phenomenon to be understood or a single problem to be solved—is misleading and liable to distort inquiry. This framing creates a chimera.
ABSTRACT. Recent studies allege that large language models (LLMs) exhibit a range of cognitive biases familiar from human cognition. I argue that the case for many biases is weaker than it may appear. Using case studies of knowledge effects in the Wason selection task, availability bias in relation extraction, and anchoring bias in code generation, I show how a range of vindicatory strategies traditionally used to vindicate apparent biases in humans can be used to push back against allegations of bias in LLMs. I discuss implications for the role of cognitive bias in evaluating LLM performance, the rationality of human cognition, and future work on cognitive bias in LLMs.
ABSTRACT. Several authors have recently argued against realism about dark matter due to it being empirically unconfirmed or too conceptually thin. In response, Vaynberg has convincingly argued in favour of dark matter realism based on the Bullet Cluster observations. However, anti-realist concerns about conceptual thinness or future empirical detection may linger. I argue that these can be diffused by distinguishing between two different dark matter concepts, which I call astrophysical dark matter and fundamental dark matter. I submit that anti-realist concerns about future dark matter detection conflate empirical confirmation for fundamental dark matter with empirical confirmation for astrophysical dark matter. I further argue that the resulting realist commitment to astrophysical dark matter is more substantive than dark matter anti-realists recognize. I end with dispelling some concerns about unconceived alternatives.
ABSTRACT. A counterpossible is a counterfactual with an impossible antecedent. Jenny has argued that relative computability theory provides examples of false counterpossibles. If Jenny were right, this would be highly significant, since it would follow that the standard analysis of counterfactuals, according to which counterpossibles are all vacuously true, is incorrect. In this paper, we argue against the claim that computability theory provides examples of false counterpossibles. We distinguish two ways of reading the alleged false counterpossibles. Under the first reading, they are indeed false but, as we argue, they are not genuine counterpossibles. Under the second reading, they are genuine counterpossibles, but they are true. There is a way to interpret the alleged false counterpossibles as false, and there is a way to interpret them as counterpossibles, but under neither of these two interpretations are they false counterpossibles.
ABSTRACT. There is a question of whether de-idealization is needed for justified use of---for 'checking'---idealizations. We argue that the standard philosophical account of de-idealization has become too idealized, but that this does not preclude the possibility of justificatory practices that show how models can be used to make inferences about the world. In turn, motivated by examples in physics, we provide a more expansive and practice-driven account of de-idealization by relaxing the standards for closeness to more realistic theoretical items, identifying at least three kinds of procedure for de-idealization: intra-model, inter-model, and measurement de-idealizations. These examples highlight how idealizations can be---and indeed have been---scrutinized within physics without appealing to the philosopher's idealized notion of de-idealization.
ABSTRACT. Work on the evolutionary origins of morality emphasizes one individual helping another without external pressure, threat or enticement, since external pressure can be dismissed as self-interested. This focus, we maintain, limits the range of moral theories under consideration and stacks the evidential deck in favour of theories that characterize morality as a kind of niceness, usually based on emotional capacities such as empathy. We argue that demand sharing—in which animals share after solicitation, request, or demand—should be seen as central in the evolution of morality. The resulting research programme would be more theoretically open-ended. It may shift the literature towards second-personal moral theories that take conflict to be central. It may also shift the literature towards moral psychologies that recognize the complexity of mixed, conflicting moral emotions. The question then is how conflict gave rise to demands that create moral interaction.
ABSTRACT. If a proposition is typically true, given your evidence, then you should believe that proposition; or so I argue here. In particular, I propose and defend a principle of rationality—call it the typical principle—that links rational belief to facts about what is typical. As I show, this principle avoids several problems that other, seemingly similar principles face. And as I also show, in many cases this principle implies the verdicts of the principal principle. Ultimately, the typical principle may be the more fundamental of the two.
ABSTRACT. Recent work in judgement and decision-making has stressed that institutions, like individuals, often rely on decision-making heuristics. But most of the institutional decision-making heuristics studied to date are highly firm- and industry-specific. This contrasts to the individual case, where many heuristics are general-purpose rules suitable for a wide range of decision problems. Are there also general-purpose heuristics for institutional decision-making? In this article, I argue that a number of methods recently developed for decision-making under deep uncertainty have a good claim to be understood as general-purpose decision-making heuristics suitable for a broad range of institutional decision problems.
ABSTRACT. Machine learning sits at the intersection of statistics and computer science, raising the question of its underlying methodology. While much has been made of the link between the process of learning from data and induction, the falsificationist component of machine learning has received less attention. We argue that falsification is central to machine learning methodology. It is often thought that machine learning algorithms infer general prediction rules from past observations. But these algorithms can also be described as choosing one prediction rule from an entire class of functions. In particular, the algorithm that determines the weights of an artificial neural network operates by risk minimization, rejecting prediction rules that lack empirical adequacy. It also exhibits a behaviour of implicit regularization, pushing hypothesis choice towards simpler prediction rules. Taking both aspects together, we give a falsificationist account of artificial neural networks.
ABSTRACT. Although the urgency of our present environmental crises is not in doubt, claims that we are in the midst of a sixth mass extinction are highly controversial. Our aims here are, first, to shed philosophical light on this scientific debate—the many conceptual and methodological challenges involved in answering this scientific question—and, second, to offer new perspectives on what the value of asking this question has been, and if that value persists. We argue that a straightforward answer to the question of whether we are experiencing a mass extinction is undermined by the challenges in defining ‘mass extinction’, uncertainties in past and present diversity assessments, and data incommensurabilities. Further, an excessive focus on the mass extinction framing can be unhelpful to conservation efforts, leading us to miss out on the many other valuable insights that Earth’s deep time can offer in guiding our future.
I diagnose the problem in quantum thermodynamics of generalizing classical work into the quantum domain. I begin with the no-go theorem: no universal measurement scheme for quantum work satisfies two intuitive, classically consilient desiderata. I assess this incompatibility as stemming from the measurement problem. Decoherence restores compatibility for all practical purposes but raises questions about what ‘universality' should mean and whether any measurement scheme can be universal. I consider a different standard of universality—in terms of ontology—by defining a trajectory-based notion of quantum work using the quantum potential. While this preserves the classical role of work as the integral of forces over distances and evades the tension of the no-go theorem, consilience fails elsewhere; no single quantum work concept seems capable of preserving all classical features, raising questions for what it takes for successful generalization of the concept to quantum thermodynamics.
ABSTRACT. Newton's gravitational theory is a hard case for scientific realists, as despite its overwhelming empirical success, what the theory tells us about the nature of gravitational forces seems to be inconsistent with our current best theory of gravity, namely, the general theory of relativity. After demonstrating the limitations of existing responses to that challenge, I argue that a better response can be given by questioning a premise that has so far been taken for granted in the debate: the assumption that all inertial forces are fictitious. Rejecting this premise is justified by distinguishing different types of inertial forces, some of which are indispensable in a sufficiently strong sense to justify realism about them.
ABSTRACT. We consider the class of physical theories whose dynamics are given by natural equations, which are partial differential equations determined by a functor from the category of n-manifolds, for some n, to the category of fibre bundles, satisfying certain further conditions. We show how the theory of natural equations clarifies several important foundational issues, including the status and meaning of minimal coupling, symmetries of theories, and background structure. We also state and prove a fundamental result about the initial value problem for natural equations.
ABSTRACT. There is a widely held view on measurement inferences that defends the prohibition that we should not make inferences from averages taken with ordinal scales (versus quantitative scales; interval or ratio). This prohibition is general—applying to all ordinal scales—and is sometimes endorsed without qualification. Adhering to it dramatically limits research in the social and biomedical sciences. I provide a Bayesian analysis of this problem, determining when measurements from ordinal scales can be used to confirm hypotheses about relative group averages. I illustrate with the alleged paradigm ordinal scale—the Mohs scale of mineral hardness—arguing that it has been mischaracterized in the literature. The prohibition, I conclude, cannot be upheld, even in a qualified sense. The beliefs needed to make average comparisons are less demanding than those appropriate for quantitative scales.
ABSTRACT. When they agree on which data streams are algorithmically random, two Bayesian agents beginning the learning process with different priors may be seen as having compatible beliefs about the global uniformity of nature. This is because the algorithmically random data streams are of necessity globally regular: they are precisely the sequences that satisfy certain important statistical laws. By virtue of agreeing on which data streams are algorithmically random, two Bayesian agents can thus be taken to concur on which global regularities they expect to see in the data. I show that this type of compatibility between priors suffices to ensure that two computable Bayesian agents will reach inter-subjective agreement with increasing information. Thus, when shared by computable Bayesian learners with different subjective priors, the beliefs about uniformity captured by algorithmic randomness provably lead to merging of opinions.
ABSTRACT. There are two important traditions in the philosophy of induction. According to one, inductive arguments are warranted by rules. Bayesianism is the most popular view within this tradition. According to the other, inductive arguments are not warranted by rules but by matters of fact. Norton’s material theory of induction is one such influential view within this tradition. Despite their limitations, both approaches illuminate important aspects inductive support. Here I present a hybrid theory of induction, in which I acknowledge and articulate the role of both rules and matters of fact in our understanding. Rules of induction accurately describe relations of inductive support when they are warranted, and a rule of induction is warranted if the right facts about the matter of the induction obtain. This provides a functional and accurate account of inductive support that can make sense of how different rules of induction coexist and tackle problems in epistemology.
ABSTRACT. There are two kinds of quantum fluctuations relevant to cosmology that we focus on here: those that form the seeds for structure formation in the early universe and those giving rise to Boltzmann brains in the late universe. Boltzmann brains are random agglomerates of particles that, by extreme coincidence, form functioning brains. Unlikely as these coincidences are, they seem to be predicted to occur in a quantum universe as vacuum fluctuations if the universe continues to exist for an infinite (or just very long) time. Indeed, they are predicted to occur over and over, forming the majority of all brains in the history of the universe. We provide an introduction to the Bohmian version of quantum theory and explain why in this version, the undesirable kind of fluctuation leading to Boltzmann brains does not occur (or at least not often), while the desirable kind of fluctuation responsible for structure formation in the early universe does.
ABSTRACT. Biochemists commonly ascribe functions to biomolecules and classify them accordingly, as has been noted in the recent literature on biochemical kinds. But while a lot has been written on biological and psychological functions, less has been said about biochemical functions. I consider whether a biological or a chemical understanding of function can be applied to biochemical functions, illustrating the issue with the example of vitamin B12. I argue that an adequate characterization of biochemical function cannot be provided by either a biological or a chemical understanding alone. Instead, I argue that if we accept functional attribution to biochemical molecules, then biochemical function is constituted by chemical dispositional properties that causally contribute to selected biological processes.
ABSTRACT. Weyl famously argued that if space were discrete, then Euclidean geometry could not hold even approximately. Since then, many philosophers have responded to this argument by advancing alternative accounts of discrete geometry that recover approximately Euclidean space. However, they have missed an importantly flawed assumption in Weyl’s argument: physical geometry is determined by fundamental spacetime structures independently from dynamical laws. In this article, I aim to show its falsity through two rigorous examples: random walks in statistical physics and quantum mechanics.
ABSTRACT. Episodic and semantic memory are often taken to be fundamentally different mental systems, and contemporary philosophers often pursue research questions about episodic memory, in particular, in isolation from semantic memory. This article challenges that assumption, and puts pressure on philosophical approaches to memory that break off episodic memory as its own stand-alone topic. I present and systematize psychological and neuroscientific theories of semanticization, the thesis that memory content tends to drift from episodic to semantic in structure over time and exposure to an environment. Semanticization, I argue, is a long-term interconnection between episodic and semantic systems that requires approaching both the content and function of these two memory systems as a whole. Thus we have a reason to reject projects by Martin, which aims to carve out a uniquely episodic memory content, and Michaelian, which pairs episodic memory to its own unique function. Instead, seeing declarative memory as a single system with two facets or even a continuum of features allows for deeper insight into both content and function.