Most patterns of an organism develop reproducibly and predictably. Thus, most biological patterns are largely predetermined by the nature of the zygote and by the nature of the surrounding world. Some ontogenetic patterns can also be considered to be preformed. Eighteenth and nineteenth century definitions of ‘preformation’ suggested that all aspects of a precursor pattern—its elements and its configuration—are preserved during development. Today, the idea of preformed configurations has been lost. To revive this lost idea, we offer the following biologically contemporary definition for preformed ontogenetic patterns: preformed patterns are those patterns with topologies that have been conserved during their ontogenies. This definition is presented in precise mathematical language, and its application is demonstrated in a system of developing, abstract patterns.

In this paper I show that Goodman's theory of projectibility, although partly successful, is inadequate since it fails to take into consideration the “approximate” nature of certain scientific hypotheses.

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.

Barker and Achinstein misread Goodman's definitions of ‘grue’ and ‘bleen’. If we stick to Goodman's definition of ‘grue’ as applying “to all things examined before t just in case they are green but to other things just in case they are blue” (my italics), and his parallel definition of ‘bleen’, then Barker and Achinstein's arguments are seen to be irrelevant. The result is to by-pass the question whether Mr. Grue sees things as grue rather than as green while showing that it is possible for human conceptual schemes to employ different sensory terms. These two issues are separate.

The question “what is (are) the unit(s) of selection” can be interpreted in three different ways. These interpretations are discussed and it is shown that they prompt different answers; such units are shown to be individuals in the context of the given interpretation. One of these interpretations is argued, by examples, not always to have an unambiguously correct answer. An alternative approach to this question is sketched.

This paper responds to a recent claim by Shrader-Frechette that current particle physics, with its essentially atomist paradigm, is in a state of Kuhnian crisis. We respond to Shrader-Frechette's claim in two ways: first, we argue directly against much of the evidence used by Shrader-Frechette as indicators of Kuhnian crisis; second, we question Shrader-Frechette's application of Kuhnian categories to current research in general, pointing out the dangers inherent in such an analysis.

A philosophically important but largely overlooked cognitive theory is examined, one that provides information on which inferences an agent will make from his beliefs. Such a theory of feasible inferences is indispensable in a complete cognitive psychology, in particular, for predicting the agent's actions on the basis of rationality conditions and attributed beliefs and desires. However, very little of the feasibility theory which applies to a typical human being can be shown a priori to apply to all agents. The logical competence required of a rational agent seems to have a cluster structure: it cannot be the case that an agent is able to make no inferences, but an agent can be unable to make any particular one.

This paper explores the status of the von Neumann-Lüders state transition rule (the “projection postulate”) within “real-logic” quantum logic. The entire discussion proceeds from a reading of the Lüders rule according to which, although idealized in applying only to “minimally disturbing” measurements, it nevertheless makes empirical claims and is not a purely mathematical theorem. An argument (due to Friedman and Putnam) is examined to the effect that QL has an explanatory advantage over Copenhagen and other interpretations which relativize truth-value assignments to experimental arrangements. Two versions of QL, the lattice-theoretic (LT) and partial-Boolean-algebra (PBA), are considered. It turns out that the projection postulate is intimately connected with the choice of conditional connective for QL. The effect of the projection postulate is obtained with the Sasaki conditional. However, this choice is found to require extra assumptions, on both the LT and PBA versions, which are either just as ad hoc as the projection postulate itself or indefensible from within the real-logic QL framework.

It is argued that the debate over the positivist theory of historical explanation has made only a limited contribution to our understanding of how historians should defend the explanations they propose importantly because both positivists and their critics tacitly accepted two assumptions. The first assumption is that if the positivist analysis of historical explanation is correct, then historians ought to attempt to defend covering laws for each of the explanations they propose. The second is that unless a historian can justify an explanation that he proposes, then his preference for that explanation is not rationally defensible. It is argued that the first assumption is false and that if in order to justify an explanation, one must justify a covering law for it, then the second assumption is also false. A program for investigating how historians should defend their explanations is suggested.

In Philosophical Problems of Statistical Inference, Seidenfeld argues that the Neyman-Pearson (NP) theory of confidence intervals is inadequate for a theory of inductive inference because, for a given situation, the ‘best’ NP confidence interval, [CIλ], sometimes yields intervals which are trivial (i.e., tautologous). I argue that (1) Seidenfeld's criticism of trivial intervals is based upon illegitimately interpreting confidence levels as measures of final precision; (2) for the situation which Seidenfeld considers, the ‘best’ NP confidence interval is not [CIλ] as Seidenfeld suggests, but rather a one-sided interval [CI0]; and since [CI0] never yields trivial intervals, NP theory escapes Seidenfeld's criticism entirely; (3) Seidenfeld's criterion of non-triviality is inadequate, for it leads him to judge an alternative confidence interval, [CIalt.], superior to [CIλ] although [CIalt.] results in counterintuitive inferences. I conclude that Seidenfeld has not shown that the NP theory of confidence intervals is inadequate for a theory of inductive inference.

On pp. 55–58 of Philosophical Problems of Statistical Inference (Seidenfeld 1979), I argue that in light of unsatisfactory after-trial properties of “best” Neyman-Pearson confidence intervals, we can strengthen a traditional criticism of the orthodox N-P theory. The criticism is that, once particular data become available, we see that the pre-trial concern for tests of maximum power (and for their derivative confidence intervals of shortest expected length) may then misrepresent the conclusion of such a test (or interval estimate). Specifically, I offer a statistical example where there exists a Uniformly Most Powerful test (a UMP-test), a test of highest N-P credentials, which generates a system of “best” confidence intervals (the [CIλ] interval system) with exact confidence coefficients. But the [CIλ] intervals have the unsatisfactory feature that, for a recognizable set of outcomes, the interval estimates cover all parameter values consistent with the data, at strictly less than 100% confidence.

It is shown that the uncertainty principle has nothing directly to do with the non-localisability of position and momentum for an individual system on the quantum logical view. The product Δx·Δp for localisation of the ranges of position and momentum of an individual system → ∞, while the quantities ΔX and ΔP in the uncertainty principle ΔX·ΔP ≥ ħ/2, must be given a statistical interpretation on the quantum logical view.

Glymour has developed an account of the confirmation of scientific hypotheses which he advocates as an alternative to the hypothetico-deductive and Bayesian accounts. This account is subject to a counter-example which may be accomodated by a slight modification. So modified it describes an important dimension of confirmation. If the modification of Glymour's account is slightly extended, both the resulting account and the hypothetico-deductive account may be seen as special cases of a Bayesian theory which is immune to Glymour's criticisms.