This paper is concerned with a recent argument of Jerry Fodor's to the effect that the frame problem in artificial intelligence is in principle insoluble. Fodor's argument is based on his contention that the mind is divided between encapsulated modular systems for information processing and ‘central systems’ for non-demonstrative inference. I argue that positing central systems is methodologically unsound, and in fact involves a muddle that bears a strong family resemblance to the basic error in dualism. I therefore conclude that Fodor's position on the frame problem should be rejected.

Recent impasses reached by nuclear theory, a recurrent desire to interpret quantum phenomena in mechanistic terms and a certain loss of glory suffered by the general theory of relativity make it appear timely, perhaps, to review discarded attempts at common-sense explanations of atomic and relativistic physics, making sure that they have not been too hastily rejected. Mr. Hartley has been a persistent and circumspect advocate of a mechanical view which, even if it cannot claim to solve all problems in the universal flash of brilliance for which physicists and philosophers have waited in vain for decades, raises worthwhile points for consideration. The nature of this task and the preoccupation of technical journals with imminent matters singles out our journal as the proper carrier for Mr. Hartley's thoughts. They are here presented without mathematical detail but in a manner suggestive of the unity of his treatment. Technical details are discussed in some of the items of the bibliography.

Some argue that time’s causal arrow is grounded in an underlying thermodynamic asymmetry. Often, this is tied to Humean skepticism that causes produce their effects, in any robust sense of ‘produce’. Conversely, those who advocate stronger notions of natural necessity often reject thermodynamic reductions of time’s causal arrow. Against these traditional pairings, I argue that ‘reduction-plus-production’ is coherent. Reductionists looking to invoke robust production can insist that there are metaphysical constraints on the signs of objects’ velocities in any state, given other—including far later—states’ properties. The Past Hypothesis may thus be a metaphysical condition, not a physical law.

Replacing the value-free ideal (VFI) for science requires attention to the broader understanding of how science in society should function. In public spaces, science needed to project the VFI in norms for science advising, science education, and science communication. This resulted in the independent science advisor model and a focus on science literacy for science education and communication. Attending to these broader implications of the VFI that structure science and society relationships is crucial if we are to properly replace the VFI with a better ideal.

Scientists may sometimes generalize from their samples to broader populations when they have not yet sufficiently supported this generalization. Do such hasty generalizations also occur in experimental philosophy? To check, we analyzed 171 experimental philosophy studies published between 2017 and 2023. We found that most studies tested only Western populations but generalized beyond them without justification. There was also no evidence that studies with broader conclusions had larger, more diverse samples, but they nonetheless had higher citation impact. Our analyses reveal important methodological limitations of many experimental philosophy studies and suggest that philosophical training may not protect against hasty generalizations.

There is a well-known equivalence between avoiding accuracy dominance and having probabilistically coherent credences (see, e.g., de Finetti 1974; Joyce 2009; Predd et al. 2009; Pettigrew 2016). However, this equivalence has been established only when the set of propositions on which credence functions are defined is finite. In this article, I establish connections between accuracy dominance and coherence when credence functions are defined on an infinite set of propositions. In particular, I establish the necessary results to extend the classic accuracy argument for probabilism to certain classes of infinite sets of propositions, including countably infinite partitions.

Influential classical and recent approaches to explicate confirmation, explanation, or explanatory power define these relations or degrees between hypotheses and evidence. This holds for both deductive and Bayesian approaches. However, this neglects the role of data, which for many everyday and scientific examples cannot simply be classified as evidence. I present arguments to sharply distinguish data from evidence in Bayesian approaches. Taking into account this distinction, we can rewrite Schupbach and Sprenger’s measure of explanatory power and show the strengths of this adaptation by applying it to the example of the detection of gravitational waves in physics.

Models typically have actual, existing targets. However, some models are viewed as having non-actual targets. I argue that this interpretation comes at various costs and propose an alternative that fares better along two dimensions: (1) agreement with practice and (2) ontological and epistemological parsimony. My proposal is that many of these models actually have actual targets.

This article characterizes various species concepts in terms of set-theoretic models that license biological inferences and illustrates the logical connections among different species concepts. Species in this construal are abstract models, rather than biological or even tangible entities, and relate to individual organisms via representation, rather than the membership or mereological whole/part relationship. The proposal sheds new light on vexed issues of species and situates them within broader philosophical contexts of model selection, scientific representation, and scientific realism.

Many modern commentators on inertial phenomena hold (or just assume) that there is no “problem of inertia”, on the grounds that either (a) no explanation is needed for such phenomena or (b) the explanation is already at hand. My purpose here is to comment on both views, defending the thesis that the problem is real and still unsolved.

Biologists have proposed a variety of explanations for extravagant sexual displays, and controversies over explanations define the history of sexual selection research. Recently, Richard Prum has defended Darwin’s idea that the evolution of sexual displays is driven by arbitrary and nonadaptive preferences of potential mates. Prum argues that this explanation should be the null model for sexual selection research. I show that if we adopted Prum’s proposal, the inferences we could be justified in making are more modest than he claims. I also discuss problems with estimating model parameters that at present prevent his proposal from being useful in practice.

Some issues raised by parapsychological phenomena (psi) are examined in the light of their implications for a philosophy of science. It is shown that the kinds of problems psi poses for science vary with the way one conceives of science as well as one's conception of psi. It is suggested that psi may be a product of the fact that all of our scientific concepts are abstractions and therefore oversimplifications. This raises the possibility that our best conceptual technique for dealing with psi is a non-discursive symbolism, since this would not demand “classes” (oversimplification). Implications of this approach are considered.

Using the ‘Aharonov approach’, Linden and colleagues purportedly prove that reaching equilibrium is a universal property of quantum systems. Such a proof would constitute a very significant result in the foundations of statistical mechanics. I argue that, as it stands, this proof is not sound. However, based on the their theorems, I construct an argument for the conclusion that an arbitrary small subsystem of a large quantum system typically tends toward and remains in, or close to, equilibrium. This is the central result of the article. In the final part of the article, I defend the Aharonov approach against anti-interventionist criticisms.

In a recent article (1980) G. H. Merrill has defended realism against an argument devised by Hilary Putnam. My first aim is to show that Merrill's defence is inadequate. I shall also argue that the proper conclusion of Putnam's argument is somewhat different from the conclusion Putnam himself offers.

I uncover here a conflict in Kant’s natural philosophy. His matter theory and laws of mechanics are in tension. Kant’s laws are fit for particles but are too narrow to handle continuous bodies, which his doctrine of matter demands. To fix this defect, Kant ultimately must ground the Torque Law; that is, the impressed torque equals the change in angular momentum. But that grounding requires a premise—the symmetry of the stress tensor—that Kant denies himself. I argue that his problem would not arise if he had kept his early theory of matter as made of mass points, or “physical monads.”

This is one of a pair of discussion notes comparing some features of the account of causation in Wolfgang Spohn’s Laws of Belief with the “interventionist” account in James Woodward’s Making Things Happen. Despite striking similarities there are also important differences. These include (i) the “epistemic” orientation of Spohn’s account as opposed to the worldly or “ontic” orientation of the interventionist account, (ii) Spohn’s focus on token-level causal claims in contrast to the primary interventionist focus on type-level claims, (iii) the role of temporal information in accounting for causal asymmetry, and (iv) the extent to which causal claims are “frame relative.”

In this article I bring the recent philosophical literature on theoretical equivalence to bear on dualities in physics. Focusing on electromagnetic duality, which is a simple example of S-duality in string theory, I show that the duality fits naturally into at least one framework for assessing equivalence—that of categorical equivalence—but that it fails to meet a necessary condition for equivalence on that account. The reason is that the duality does not preserve “empirical content” in the required sense; instead, it takes models to models with “dual” empirical content. I conclude by discussing how one might react to this.