This article develops a new version of instrumentalism, in light of progress in the realism debate in recent decades, and thereby defends the view that instrumentalism remains a viable philosophical position on science. The key idea is that talk of unobservable objects should be taken literally only when those objects are assigned properties (or described in terms of analogies involving things) with which we are experientially (or otherwise) acquainted. This is derivative from the instrumentalist tradition insofar as the distinction between unobservable and observable is taken to have significance with respect to meaning.

I propose a general alethic theory of epistemic risk according to which the riskiness of an agent’s credence function encodes her relative sensitivity to different types of graded error. After motivating and mathematically developing this approach, I show that the epistemic risk function is a scaled reflection of expected inaccuracy (a quantity also known as generalized information entropy). This duality between risk and information enables us to explore the relationship between attitudes to epistemic risk, the choice of scoring rules in epistemic utility theory, and the selection of priors in Bayesian epistemology more generally (including the Laplacean principle of indifference).

Recent discussions in the philosophy of science have devoted considerable attention to the analysis of conceptual issues relating to the methodology of explanation and prediction in the sciences. Part of this literature has been devoted to clarifying the very ideas of explanation and prediction. But the discussion has also ranged over various related topics, including the status of laws to be used for explanatory and predictive purposes, the logical interrelationships between explanatory and predictive reasonings, the differences in the strategy of explanatory argumentation in different branches of science, the nature and possibility of teleological explanation, etc. The aim of the present article is to examine the issues involved in such questions from the specialized perspective afforded by one particular kind of physical systems—namely, systems, here to be characterized as discrete state systems, whose behavior has been studied extensively in the scientific literature under the general heading of Markov chains. These systems have been chosen as our focus because their behavior over time can be analyzed at once with great ease and with extraordinary precision.

Phylogenetic trees are meant to represent the genealogical history of life and apparently derive their justification from the existence of the tree of life and the fact that evolutionary processes are treelike. However, there are a number of problems for these assumptions. Here it is argued that once we understand the important role that phylogenetic trees play as models that contain idealizations, we can accept these criticisms and deny the reality of the tree while justifying the continued use of trees in phylogenetic theory and preserving nearly all of what defenders of trees have called the “importance of tree thinking.”

I argue that a phenotypic trait can be an adaptation to a particular environmental condition, as against others, only if the environmental condition and the phenotype interactively cause fitness. Models of interactive environmental causes of fitness generally require that environments be individuated by explicit representation rather than by measures of environmental quality, although the latter understanding of ‘environment’ is more prominent in the philosophy of biology. Hence, talk of adaptations to some but not other environmental conditions relies on conceptions of ‘environment’ importantly different from that commonly presupposed in philosophy of biology.

Historical sciences like evolutionary biology reconstruct past events by using the traces that the past has bequeathed to the present. Markov chain theory entails that the passage of time reduces the amount of information that the present provides about the past. Here we use a Moran process framework to show that some evolutionary processes destroy information faster than others. Our results connect with Darwin’s principle that adaptive similarities provide scant evidence of common ancestry whereas neutral and deleterious similarities do better. We also describe how the branching in phylogenetic trees affects the information that the present supplies about the past.

This study purports to demonstrate the possibility of applying logical analysis in the field of jurisprudence, and the usefulness of this method for exhibiting some essential features of the law.

logical analysis applies to language systems. To carry out the logical analysis of a language, is to construct a simplified model language “in close connection with” the given language, and to study the conditions of validity of the sentences of this model language. The given language is subjected to a kind of purifying chemical process. To deal with this simplified language seems to be a detour; it is, however, the most direct way to an ultimate insight into the logical structure of the given language.

We reject a widespread objectivism about kinds of evolutionary groups in favor of a new conventionalism. Surprisingly, being any one kind of evolutionary group typically depends on which of many incompatible values are taken by suppressed variables. This novel pluralism underlies almost any single evolutionary group concept, unlike familiar pluralisms claiming that multiple concepts of certain sorts are legitimate. Consequently, we must help objective facts determine which candidate evolutionary groups satisfy the definition of a given evolutionary group concept, regardless of whether we also help determine the legitimacy of that concept’s applications.

The central concept of Carnap's probabilistic theory of induction is a triadic relation, c(h, e), the probability or degree of confirmation of the hypothesis, h, on evidence, e. The relation is a purely logical one. The value of c can be computed from a knowledge of h, of e, of the structure of the language, and of the inductive rule to be employed.

Much of Grünbaum's work may be regarded as a careful development and systematic elaboration of the Riemann–Poincaré thesis of the conventionality of congruence, the thesis that the continuous manifolds of space, time, and space-time are intrinsically metrically amorphous, i.e. are devoid of intrinsic metrics. Therefore, to appreciate Grünbaum's philosophical contributions, one must have a clear understanding of what he means by an intrinsic metric. The second and fourth sections of this paper are exegetical; in them we try to piece together, from his sundry remarks about intrinsic metrics, what Grünbaum means by the term ‘intrinsic metric.’ We shall argue that, the customary carefulness and precision of Grünbaum's writings notwithstanding, there are residual unclarities and difficulties which beset his conception of an intrinsic metric. In the fifth section we shall propose an explication of Grünbaum's notion of an intrinsic metric which seems on the whole faithful to Grünbaum's intuitions and insights and which also seems capable of performing the philosophical services which his work demands of that notion. The third section is a digression on Zeno's metrical paradox of extension.

Individualists claim that wide explanations in psychology are problematic. I argue that wide psychological explanations sometimes have greater explanatory power than individualistic explanations. The aspects of explanatory power I focus on are causal depth and theoretical appropriateness. Reflection on the depth and appropriateness of other wide explanations of behavior, such as evolutionary explanations, clarifies why wide psychological explanations sometimes have more causal depth and theoretical appropriateness than narrow psychological explanations. I also argue for the rejection of eliminative materialism.

Mathematicians, physicists, and philosophers of physics often look to the symmetries of an object for insight into the structure and constitution of the object. My aim in this article is to explain why this practice is successful. In order to do so, I present a collection of results that are closely related to (and, in a sense, generalizations of) Beth’s and Svenonius’s theorems.

Over the last 20 years, several philosophers have developed a new approach to biological functions, the organizational (or systems-theoretic) approach. This is not a single theory but a family of theories based on the idea that a trait token can acquire a function by virtue of the way it contributes to a complex, organized system and thereby to its own continued persistence as a token. I argue that the organizational approach faces a serious liberality objection. I examine three different ways organizational theorists have tried to avoid that objection and show how they fail.

Criteria of adequacy for scientific representation of the phenomena pertain to accuracy and truth. But that representation is selective and may require distortion even in the selected parameters; this point is intimately connected with the fact that representation is intentional, and its adequacy relative to its particular purpose. Since observation and measurement are perspectival and the appearances to be saved are perspectival measurement outcomes, the question whether this “saving” is an explanatory relation provides a new focus for the realist/antirealist debate. The Born rule and von Neumann's “collapse” postulate in quantum mechanics provide a telling case for this question in recent physics.

Regularity is the thesis that all contingent propositions should be assigned probabilities strictly between zero and one. I will prove on cardinality grounds that if the domain is large enough, a regular probability assignment is impossible, even if we expand the range of values that probabilities can take, including, for instance, hyperreal values, and significantly weaken the axioms of probability.

There are several interpretations of the argument structure of Darwin's Origin of Species, representing Covering-Law, Inference-to-the-Best-Explanation, and (more recently) Semantic models. I argue that while all three types of interpretation enjoy some textual support, none succeeds in capturing the overall strategy of the Origin, consistent with Darwin's claim that it is ‘one long argument’. I provide detailed criticisms of all three current models, and then offer an alternative interpretation based on the view that there are three main argument strategies in the Origin, all supporting the ‘causal efficacy’ of Darwin's theory. This interpretation provides both a more unified treatment of the text, and some important implications concerning the relation between general philosophical models of scientific theory support and specific historical cases.

Proponents of genic selectionism have claimed that evolutionary processes normally viewed as selection on individuals can be “represented” as selection on alleles. This paper discusses the relationship between mathematical questions about the formal requirements upon state spaces necessary for the representation of different types of evolutionary processes and causal questions about the units of selection in such processes.

In this paper I describe the history of the surreal trajectories problem and argue that in fact it is not a problem for Bohm's theory. More specifically, I argue that one can take the particle trajectories predicted by Bohm's theory to be the actual trajectories that particles follow and that there is no reason to suppose that good particle detectors are somehow fooled in the context of the surreal trajectories experiments. Rather than showing that Bohm's theory predicts the wrong particle trajectories or that it somehow prevents one from making reliable measurements, such experiments ultimately reveal the special role played by position and the fundamental incompatibility between Bohm's theory and relativity. They also provide a striking example of the theory-ladenness of observation.

In his The Uses of Argument (Cambridge University Press, 1958), S. Toulmin presents serious charges against ordinary logical theory, e.g., that it does not distinguish between analytic or formally valid or conclusive or warrant-using arguments, that the distinction between premises and conclusion is a bad oversimplification, that “major premise” conceals the distinction between inference-warrant and inference-backings, that logicians have been mistakenly working under an ideal of geometrical form.

The paper argues that none of the charges is proven, that most of them cannot be proven, that Toulmin's new logic is at best only vaguely hinted at and that his suggestions are positively obscure or mistaken.

Isaac Levi and I have different views of probability and decision making. Here, without addressing the merits, I will try to answer some questions recently asked by Levi (1985) about what my view is, and how it relates to his.