The idea that science aspires to and routinely achieves truths about the world has been challenged in recent writings. Rather than beginning with a theory of scientific development, or of scientific explanation, we begin with a consideration of truth claims in ordinary discourse, particularly with Davidson's truth-functional semantics. Next we consider the way in which some framework features of ordinary language discourse are extended to and modified in scientific discourse. Two areas are treated in more detail: quantum theory, and the peculiar problem of semantic entailment it involves; and quantum field theory. These supply a basis for criticizing some historicist and logicist treatments of the truth of scientific claims.

Models that fail to satisfy the Markov condition are unstable because changes in state variable values may cause changes in the values of background variables, and these changes in background lead to predictive error. Such error arises because non-Markovian models fail to track the causal relations generating the values of response variables. This has implications for discussions of the level of selection: under certain plausible conditoins most standard models of group selection will not satisfy the Markov condition when fit to data from real populations. These models neither correctly represent the causal structure generating nor correctly explain the phenomena of interest.

Rosenberg (1983), in his comments on our article (Sober and Lewontin 1982) concerning the units of selection controversy, has matters precisely backwards. He chides us for conflating a causal theory with a mathematical model, whereas this is precisely the error we argued is at the core of the position defended by Williams (1966) and Dawkins (1976). Their idea—genic selectionism—confuses the computational adequacy of a mathematical model with a substantive causal theory. The main point of our paper was simply this: the fact that evolution by natural selection can be “represented” in terms of selection coefficients that attach to single genes does not imply that selection is always selection for or against the possession of those single genes.

While practitioners think highly of randomized studies, some philosophers argue that there is no epistemic reason to randomize. Here I show that their arguments do not entail their conclusion. Moreover, I provide novel reasons for randomizing in the context of interventional studies. The overall discussion provides a unified framework for assessing baseline balance, one that holds for interventional and observational studies alike. The upshot: Practitioners’ strong preference for randomized studies can be defended in some cases, while still offering a nuanced approach to evidence appraisal, one where not all nonrandomized studies are treated equally.

How can we make informed decisions about whom to trust given expert disagreement? Can experts on both sides be reasonable in holding conflicting views? Epistemologists have engaged the issue of reasonable expert disagreement generally; here I consider a particular expert dispute in physics, given conflicting accounts from Harry Collins and Allan Franklin, over Joseph Weber's alleged detection of gravitational waves. Finding common ground between Collins and Franklin, I offer a characterization of the gravity wave dispute as both social and evidential. While experimental evidence alone may not have forced resolution of the dispute, there were also credibility-based reasons warranting epistemic trust and distrust. Thus we see how social factors can have evidential significance and how expert disagreement can be reasonable.

Although a great deal has been written concerning the verifiability of empirical, declarative sentences, yet some further clarification in this area seems possible and desirable. I think it is important to determine what sort of property ‘verified by’ is. Obviously it is a semiotical property. Only sentences, never matters of fact, are verified. But where, in the area of semiotics, is it to be placed? To use Charles Morris's useful and now somewhat traditional classification, is it a syntactic, semantic, or pragmatic (i.e. psychological) predicate? My answer is that it falls within semantics. Before developing this, however, I should like to point out why I do not place it in either of the other two subdivisions.

Frankel (1976) has argued that the theory of causality developed by Rom Harré (Harré 1970) and his colleague Edward Madden (Harré and Madden 1975) is incoherent, since the proposal that causal claims are naturally necessary leads to a vicious infinite regression

“which ends by requiring that for any causal claim to be accorded the status of natural necessity an infinite number of causal claims must be accorded the status of natural necessities.” (Frankel 1976, p. 560)

The renormalization group (RG) has been characterized as merely a coarse-graining procedure that does not illuminate the microscopic content of quantum field theory (QFT) but merely gets us from that content, as given by axiomatic QFT, to macroscopic predictions. I argue that in the constructive field theory tradition, RG techniques do illuminate the microscopic dynamics of a QFT, which are not automatically given by axiomatic QFT. RG techniques in constructive field theory are also rigorous, so one cannot object to their foundational import on grounds of lack of rigor.

Naturalized metaphysicians defend the thesis that science licenses metaphysics, such that only metaphysical results based on the best science are legitimate. This view is problematic, due to the fact that the reasons they identify for such license are apparently self-defeating. Anjan Chakravartty defends a revised approach to understanding the licensing relation. I argue that the proposed response is a step forward on behalf of naturalizing metaphysics but still does not take seriously the contention that science involves, inextricably, a contribution from the a priori. I conclude by considering what options the naturalized metaphysician is left with.

The Variety of Evidence Thesis (VET) says that (ceteris paribus) the more diverse (or varied) of two bodies of evidence is the more confirmatory of a hypothesis H. Two recent types of Bayesian explication of VET account for the intuitive force of VET by defining variety as some function of the probabilities of the propositions which collectively constitute a body of evidence. I show that these two accounts of VET are not tracking a meaningful property of bodies of evidence, but rather are tracking artifacts of how those bodies of evidence are described. According to each account, whether a body of evidence is more varied than another depends on how the bodies are split into parts. Furthermore, for each type of account there exists a way to redescribe the total evidence such that any two totalities are equally varied.

Children acquire complex concepts like dog earlier than simple concepts like brown, even though our best neuroscientific theories suggest that learning the former is harder than learning the latter and, thus, should take more time (Markus Werning). This is the complex-first paradox. We present a novel solution to the complex-first paradox. Our solution builds on a generalization of Fei Xu and Joshua B. Tenenbaum’s Bayesian model of word learning. By focusing on a rational theory of concept learning, we show that it is easier to infer the meaning of complex concepts than that of simple concepts.

According to a standard account of evidence, one piece of information is stronger evidence for an hypothesis than is another iff the probability of the hypothesis on the one is greater than it is on the other. This condition, I argue, is neither necessary nor sufficient because various factors can strengthen the evidence for an hypothesis without increasing (and even decreasing) its probability. Contrary to what probabilists claim, I show that this obtains even if a probability function can take these evidential factors into account in ways they suggest and yield a unique probability value. Nor will the problem be solved by appealing to second-order probabilities.

In this paper I attempt to trace certain aspects of the evolution of theories of space-time and mechanics as revealed by a brief comparative study of Newtonian theory, Robb's theory, general relativity, and Milne's kinematical relativity. The first object is to emphasise how each theory leaves us in a position in which the succeeding one appears as a perfectly natural next step in the development of ideas. The second object is to show how, in spite of superficial differences in character, the theories all necessarily possess the same general structure constituted by the presence of hypotheses, from which certain general mathematical relations are deduced, which in their turn are used to predict relations between observable quantities.

After maintaining in (Sobel 1997) that certain cyclical preferences can be reasonable, the following questions were tabled: “in what circumstances and under what assumptions … these … preferences of George's would turn him into a ”money pump,” and in what circumstances and under what assumptions they would not do that.” (47) Two pumps afford answers. George, who is sufficiently reasonable and well-informed to use backward induction, has, for this reason, nothing to fear from the first pump, but the second, of a nonstandard design by Wlodek Rabinowicz, ruins him just because he is subject to the constraints of backward induction. Could he be helped by extending his options from actions in a single round, to courses of actions starting in that round? Possibly not. He would be safe if he were capable of resolute choices of couses of action. Even so, he is in trouble through no fault of his, if, as I think, not having that capacity, though a misfortune, is not a fault.

Suppose that the axioms of conjoint measurement hold for quantities having two independent components and that the axioms of extensive measurement hold for each of these components separately. In a recent paper, Luce shows that if a certain axiom relates the two measurement systems, then the conjoint measure on each component is a power function of the extensive measure on that component. Luce supposes that each component set contains all “rational fractions” of each element in that set; in this note we present an alternative form of the axiom relating the measurement systems that enables us to prove Luce's result without requiring that such “rational fractions” exist.

In order to assess the tenacity of psychoanalysts in continuing to use a concept of psychic energy, it is advisable to consider whether, as they sometimes claim, the concepts of energy, force, and work in psychoanalysis are akin to those in the natural sciences. Strong disanalogies suggest that the psychoanalytic concepts are quite different and used equivocally even within psychoanalysis. However, they may not be subject to the objections which certain critical psychoanalysts have raised.

It has long been known that the practice of testing all hypotheses at the same level (such as 0.05), regardless of the distribution of the data, is not consistent with Bayesian expected utility maximization. According to de Finetti's “Dutch Book” argument, procedures that are not consistent with expected utility maximization are incoherent and they lead to gambles that are sure to lose no matter what happens. In this paper, we use a method to measure the rate at which incoherent procedures are sure to lose, so that we can distinguish slightly incoherent procedures from grossly incoherent ones. We present an analysis of testing a simple hypothesis against a simple alternative as a case-study of how the method can work.

In The Nature of Selection, Elliott Sober argued that natural selection is in principle powerless to explain why any individual organism has the traits it does rather than the very same individual having different traits. In this note, I argue that in a recent and prominent critique of Sober’s position, Bence Nanay talks past that position rather than addressing it.