Minkowski geometry is axiomatized in terms of the asymmetric binary relation of optical connectibility, using ten first-order axioms and the second-order continuity axiom. An axiom system in terms of the symmetric binary optical connection relation is also presented. The present development is much simpler than the corresponding work of Robb, upon which it is modeled.

This essay is, primarily, a discussion of four results about the principle of maximizing entropy (MAXENT) and its connections with Bayesian theory. Result1 provides a restricted equivalence between the two: where the Bayesian model for MAXENT inference uses an “a priori“ probability that is uniform, and where all MAXENT constraints are limited to 0–1 expectations for simple indicator-variables. The other three results report on an inability to extend the equivalence beyond these specialized constraints. Result2 established a sensitivity of MAXENT inference to the choice of the algebra of possibilities even though all empirical constraints imposed on the MAXENT solution are satisfied in each measure space considered. The resulting MAXENT distribution is not invariant over the choice of measure space. Thus, old and familiar problems with the Laplacian principle of Insufficient Reason also plague MAXENT theory. Result3 builds upon the findings of Friedman and Shimony (1971; 1973) and demonstrates the absence of an exchangeable, Bayesian model for predictive MAXENT distributions when the MAXENT constraints are interpreted according to Jaynes's (1978) prescription for his (1963) Brandeis Dice problem. Lastly, Result4 generalizes the Friedman and Shimony objection to cross-entropy (Kullback-information) shifts subject to a constraint of a new odds-ratio for two disjoint events.

There have been two principal paradigms for the formulation of the causal relation–logical implication and functional relationship. In this paper, I present a case for preferring the functional relationship formulation and then discuss how the functional relationship formulation may be implemented in the probabilistic case in a manner analogous to the way others have implemented the logical implication formulation in the probabilistic case. I show how the “local independence” assumption found in many models used in the behavioral and social sciences is a consequence of the functional relation concept of probabilistic causality. I also discuss why causal relations are underdetermined in experience and are to be dealt with pragmatically in a hypothesis-testing framework.

The importance for realism of the concept of truthlikeness was first stressed by Popper. Popper himself not only mapped out a program for defining truth-likeness (in terms of falsity content and truth content) but produced the first definitions within this program. These were shown to be inadequate. But the program lingered on, and the most recent attempt to revive it is that of Newton-Smith. His attempt is a failure, not because of some minor defect or technical flaw in his particular account but rather because the program incorporates a fundamental flaw. However, realists need not despair. There already exists an entirely different program not subject to these criticisms.

In response to recent recognition of the complexities of scientific change, discussion of the objectivity and the rationality of science has focused on criteria of theory choice. This paper addresses instead the rationality of scientific decisions at the level of ongoing research. It argues that whether or not a realist view of theories is compatible with the historical discontinuities of scientific change, certain realist assumptions are crucial to the rationality of research. The researcher must presume that questions about the existence and the properties of at least some of the “unobservable“ entities he theorizes about or experiments on are answerable on the basis of his work. The rationale of research cannot be understood solely in terms of the desiderata of instrumental utility or the empirical adequacy of theories.

A major theme in biological research is the quest for mechanism, embodied in explanatory reductionism: the interpretation of phenomena through links to the entities and laws of more fundamental sciences. For example, the form of Starling's Law of the Heart, relating contractile force to heart volume, follows from the sliding-filament hypothesis of muscle contraction, a molecular concept. Although alternative mechanisms for muscle contraction and cardiac regulation could be deduced from biochemical principles, the formulation provides clear correspondence with the phenomena and a deeper understanding of cardiac function. Moreover, even if the reduction provides no better formulation of the phenomenon, of the macro-regularity, it is still prized as a nearer approach to physical reality, as a step toward realizing fundamental unities and codifying general principles, and as a door to new experimental opportunities.

In the course of defending both a unified model of intertheoretic relations in science and scientific realism, Paul Churchland has attempted to reinvigorate eliminative materialism. Churchland's eliminativism operates on three claims:

(1) that some intertheoretic contexts involve incommensurable theories,

(2) that such contexts invariably require the elimination of one theory or the other, and

(3) that the relation of psychology and neuroscience is just such a context.

I argue that a more detailed account of intertheoretic relations, which distinguishes between the relations that hold between successive theories at a particular level of analysis over time and those that hold between theories at different levels of analysis at the same time, offers grounds for denying Churchland's second and third claims and, therefore, undermines his eliminativism. The paper concludes by suggesting why it is, nonetheless, not unreasonable, given this more detailed model of intertheoretic relations, to expect the eventual elimination of common sense psychology.

This paper contributes to the recently renewed debate over methodological individualism (MI) by carefully sorting out various individualist claims and by making use of recent work on reduction and explanation outside the social sciences. My major focus is on individualist claims about reduction and explanation. I argue that reductionist versions of MI fail for much the same reasons that mental predicates cannot be reduced to physical predicates and that attempts to establish reducibility by weakening the requirements for reduction also fail. I consider and reject a number of explanatory theses, among them the claims that any adequate theory must refer only to individuals and that individualist theory suffices to explain fully. The latter claim, I argue, is not entailed by the supervenience of social facts on individual facts. Lastly, I argue that there is one individualist restriction on explanation which is far more plausible and significant than one would initially suspect.

Realism about progress–that progress consists in the obtaining of some metaphysical relation between a sequence of theories and the world–is often thought to be required by realism about the representational character of theories. The purpose of this paper is to show how one can be a realist about theories without being a realist about progress. The result is a view that cannot be at odds with what the history of science shows.

A system as complex as man can be viewed from many sides, and the one or the other axis can be selected to form a theoretical image. Partial truths will emerge, and their mutual intermeshing will gradually raise truth to higher levels. … It has always proven prejudicial to attribute general validity to a partial truth. Yet, a partial truth could not have been reached without overstating its value. Thus the history of truth is intimately interconnected with the history of emotion, but the latter has remained in the dark.

Robert Musil, The Man Without Qualities

The paper deals with the semantics of mathematical notation. In arithmetic, for example, the syntactic shape of a formula represents a particular way of specifying, arriving at, or constructing an arithmetical object (that is, a number, a function, or a truth value). A general definition of this sense of “construction“ is proposed and compared with related notions, in particular with Frege's concept of “function” and Carnap's concept of “intensional isomorphism.“ It is argued that constructions constitute the proper subject matter of both logic and mathematics, and that a coherent semantic account of mathematical formulas cannot be given without assuming that they serve as names of constructions.

It is possible for a causal factor to raise the probability of a second factor in some situations while lowering the probability of the second factor in other situations. Must a genuine cause always raise the probability of a genuine effect of it? When it does not always do so, an “interaction“ with some third factor may be the reason. I discuss causal interaction from the perspectives of Giere's counterfactual characterization of probabilistic causal connection (1979, 1980) and the “contextual unanimity” model developed by, among others, Cartwright (1979) and Skyrms (1980). I argue that the contextual unanimity theory must exercise care, in a new way that seems to have gone unnoticed, in order to adequately accommodate the phenomenon, and that the counterfactual theory must be substantially revised; although it will still, pending clarification of a second kind of revision, be unable to accommodate a kind of interaction exemplified in cases like those described by Sober (1982).

This book consists thematically of three broad sections: (1) a lengthy introduction (approximately one third of the book) in which Grünbaum critically assesses the hermeneutic construal of psychoanalysis, as represented in the work of Habermas, G. S. Klein, and Ricoeur; (2) a critical examination of Popper's assessment of both psychoanalysis and inductivism; and (3) a logical analysis of core psychoanalytic ideas that constitute the foundation for much of psychoanalytic theory. This last section is, in my view, the heart of the book and therefore, it is that section on which I will focus in this review.

General discussions of the appropriate relations between history and philosophy of science must be complemented by examinations of particular studies involving both fields. Martin Curd's attempt to illuminate the rationality of theory change through analysis of the Copernican Revolution is such a study; his work is undercut by serious flaws and actually displays an ahistorical approach. The result misleads both about the Copernican Revolution and the general problem of theory change in science. The study does illustrate several types of failing that can vitiate efforts to bring historical considerations into philosophical discussion, namely, pitfalls in the characterization of theories, arguments, range of choices, and criteria.

One of us (Eells 1982) has defended traditional evidential decision theory against prima facie Newcomb counterexamples by assuming that a common cause forms a conjunctive fork with its joint effects. In this paper, the evidential theory is defended without this assumption. The suggested rationale shows that the theory's assumptions are not about the nature of causality, but about the nature of rational deliberation. These presuppositions are weak enough for the argument to count as a strong justification of the evidential theory.

The psychofunctionalist claim that psychological terms can be defined through the use of an experimental theory has been criticized on the grounds that it is “chauvinistic“: that it denies mentality to any creature of which the selected theory is false. I analyze the “argument from science fiction” that is thought to establish this conclusion, and show that its plausibility rests on a scope ambiguity in formulations of functional definitions. One formulation is indeed chauvinistic, but an alternative rendering is not, and is perfectly consistent with ascribing mentality to creatures of which the selected psychological theory is false. This alternative interpretation of psychofunctionalism is set out in detail, defended from several objections, and finally tied to the semantics of ordinary language psychological terms.

Inertial frames and Lorentz transformations have a preferred status in the special theory of relativity (STR). Lorentz transformations, in turn, embody Einstein's convention that the velocity of light is isotropic, a convention that is necessary for the establishment of a standard signal synchrony. If the preferred status of Lorentz transformations in STR is not due to some particular bias introduced by a convention on signal synchronism, but to the fact that the Lorentz transformation group is the symmetry group of the theory, then the signal synchronism is not a matter of convention but rather a matter of fact. In order to explore the conventionalist thesis, that within the frame of STR isotropy in the velocity of light and, hence, signal synchronism is a matter of convention, we need a generalized Lorentz transformation group that does not embody Einstein's isotropy convention, and upon which STR can be based. We present here a new approach to the resulting search for a generalized STR, which is well suited for establishing some well-known results of Winnie as well as some new results.

In a number of recent papers I have been developing the theory of “nomic probability,“ which is supposed to be the kind of probability involved in statistical laws of nature. One of the main principles of this theory is an acceptance rule explicitly designed to handle the lottery paradox. This paper shows that the rule can also handle the paradox of the preface. The solution proceeds in part by pointing out a surprising connection between the paradox of the preface and the gambler's fallacy.

Hausman has argued that Mill in the Logic demands verification of qualified, inexact statements if they are to be considered lawlike. This puts Mill in line with a reasonable interpretation of what modern microeconomists are about, but requires the additional hypothesis that Mill abandoned his earlier stress on modal truth in his 1836 essay on the method of economics. The paper maintains that neither textual nor contextual evidence supports this hypothesis. Moreover, it is superfluous if one attends carefully to how Mill conceived economic science, which occupied a peculiar, somewhat isolated place in his own views on the deductive method and on verification.

One central debate in cognitive science is over imagery. Do images constitute, or constitute evidence for, a distinctive, depictive form of mental representation? The most sophisticated advocacy of this view has been developed by Kosslyn and his coworkers. This paper focuses on his position and argues (i) that though Kosslyn has not developed a satisfactory account of depiction, there is nothing in principle unintelligible about the idea of depictive neural representation, but (ii) Kosslyn's model of imagery rescues the intelligibility of pictorialism at the cost of its explanatory power.