In this article I sketch G. N. Lewis's views on chemical bonding and Linus Pauling's attempt to preserve Lewis's insights within a quantum-mechanical theory of the bond. I then set out two broad conceptions of the chemical bond, the structural and the energetic views, which differ on the extent in which they preserve anything like the classical chemical bond in the modern quantum-mechanical understanding of molecular structure.

Recent claims, mainly from computer scientists, concerning a largely automated and model-free data-intensive science have been criticized by several philosophers of science. The debate suffers from lack of detail regarding the actual methods used in data-intensive science and in which ways these presuppose theoretical assumptions. I examine two widely used algorithms, classificatory trees and nonparametric regression, and argue that they are theory laden in an external sense, regarding the framing of research questions, but not in an internal sense, concerning the causal structure of the examined phenomenon. With respect to the novelty of data-intensive science, I draw an analogy to exploratory experimentation.

In this essay I first provide an analysis of various community concepts. Second, I evaluate two of the most serious challenges to the existence of communities—gradient and paleoecological analysis respectively—arguing that, properly understood, neither threatens the existence of communities construed interactively. Finally, I apply the same interactive approach to ecosystem ecology, arguing that ecosystems may exist robustly as well.

Sometimes a theory is interpreted realistically—i.e., as literally true—whereas sometimes a theory is interpreted instrumentalistically—i.e., as merely a convenient device for summarizing, systematizing, deducing, etc., a given body of observable facts. This paper is part of a program aimed at determining the basis on which scientists decide on which of these interpretations to accept a theory. I proceed by examining one case: the nineteenth-century debates about the existence of atoms. I argue that there was a gradual transition from an instrumentalist to a realistic acceptance of the atomic theory, because of gradual increases in its predictive power, the “testedness” of its hypotheses, the “determinateness” of its quantities, and because of resolutions of doubts about the acceptability of its basic explanatory concepts.

The paper aims to put certain basic mathematical elements and operations into an empirical perspective, evaluate the empirical status of various analytic operations widely used within psychology and suggest alternatives to procedures criticized as inadequate.

Experimentation shows the “manyness” of items to be a perceptual quality for both young children and animals and that natural operations are performed by naive children analogous to those performed by persons tutored in arithmetic. Number, counting, arithmetic operations therefore can make distinctions that are not inevitably arbitrary, and conceptual operations can obviously have a status as natural events with psychology. If the elements and conceptual operations involved in mathematical systems were not inherent in physiological process, various primitive discriminations could not be possible. Also, since some calculi have a natural status in a given empirical circumstance, the axioms of others can not be satisfied. Therefore the psychologist when acting as an empirical scientist seeks a calculus having a structure whose elements are isomorphic with natural units of stimulus and response and whose operations are isomorphic with whatever natural processes are involved.

Measurement poses a special problem for the empirical scientist. It concerns but a single class of natural qualities and this only in a limited way. The concept of quantity has a natural counterpart but quantity and measurement are not wholly analogous. Measurement is defined, as H. S. Leonard suggests, as a theoretical activity. Measurement theory has a formal structure but empirical end. Measurement hypothesizes about the position of a particular quality within a definite range of qualities. It therefore is beholden to definite empirical restrictions.

Some hypotheses-making systems use terms and relations per se as the context and starting point for dealing with discriminable events. Such procedures are “transcendent.” In empirical science, questions are part of problem-solving activity and their reference is naturally restricted. In providing description and explanation, psychological researchers frequently use calculi in a transcendent way. This results in theories that are only quasi-empirical and “half” true.

The roles measurement plays in psychology are discussed. Of particular concern are those cases in which the results of measuring or a theory of measurement are used to define the “real” units, or the “real” relations involved in problematic psychological events, and thence to describe and explain behaviors of interest. Metaphysical or ontological usages of measurement sometimes occur. The implication of these arguments with regard to a view of empirical science is discussed.

This article examines the role of experimental generalizations and physical laws in neuroscientific explanations, using Hodgkin and Huxley's electrophysiological model from 1952 as a test case. I show that the fact that the model was partly fitted to experimental data did not affect its explanatory status, nor did the false mechanistic assumptions made by Hodgkin and Huxley. The model satisfies two important criteria of explanatory status: it contains invariant generalizations and it is modular (both in James Woodward's sense). Further, I argue that there is a sense in which the explanatory heteronomy thesis holds true for this case.

Bogen and Woodward characterized data as embedded in the context in which they are produced (‘local’) and claims about phenomena as retaining their significance beyond that context (‘nonlocal’). This view does not fit sciences such as biology, which successfully disseminate data via packaging processes that include appropriate labels, vehicles, and human interventions. These processes enhance the evidential scope of data and ensure that claims about phenomena are understood in the same way across research communities. I conclude that the degree of locality of both data and claims about phenomena varies depending on the packaging used to make them travel and on the research setting in which they are used.

This paper addresses the problem of how scientific knowledge, which is always locally generated, becomes accepted in other sites. The analysis suggests that there are a small number of strategies that enable scientists to resituate knowledge and that these strategies are generic: they are not restricted to specific disciplines or modes of doing science but rather are found in a variety of different forms across the sciences.

This paper examines a promising probabilistic theory of singular causation developed by David Lewis. I argue that Lewis' theory must be made more sophisticated to deal with certain counterexamples involving pre-emption. These counterexamples appear to show that in the usual case singular causation requires an unbroken causal process to link cause with effect. I propose a new probabilistic account of singular causation, within the framework developed by Lewis, which captures this intuition.

A variety of recent philosophical discussions, particularly on topics relating to complexity, have begun to reemploy the concept of ‘emergence’. Although multiple concepts of ‘emergence’ are available, little effort has been made to systematically distinguish them. In this paper, I provide a taxonomy of higher-order properties that (inter alia) distinguishes three classes of emergent properties: (1) ontologically basic properties of complex entities, such as the mythical vital properties, (2) fully configurational properties, such as mental properties as they are conceived of by functionalists and computationalists, and (3) highly configurational/holistic properties, such as the higher-level patterns characteristic of complex dynamical systems. Or more simply: emergence as ontological liberality, emergence as multiple realizability, and emergence as interactive complexity.

In this paper I want to cast doubt on the claim that there is a legitimate process of reasoning to the best explanation which can serve as an alternative to either straightforward inductive reasoning or a combination of inductive and deductive reasoning. I shall argue a) that paradigmatic cases of acceptable arguments to the best explanation must be considered enthymemes and b) that when the suppressed premises are made explicit we have all of the premises we need to present either a straightforward inductive argument or an argument employing both induction and deduction.

This paper starts by looking at the coincidence of surprising behavior on the nanolevel in both matter and simulation. It uses this coincidence to argue that the simulation approach opens up a pragmatic mode of understanding oriented toward design rules and based on a new instrumental access to complex models. Calculations, and their variation by means of explorative numerical experimentation and visualization, can give a feeling for a model's behavior and the ability to control phenomena, even if the model itself remains epistemically opaque. Thus, the investigation of simulation in nanoscience provides a good example of how science is adapting to a new instrument: computer simulation.

In Gemes (1990) I examined certain formal versions of hypothetico-deductivism (H-D) showing that they have the unacceptable consequence that “Abe is a white raven” confirms “All ravens are black”! In Gemes (1992) I developed a new notion of content that could save H-D from this bizarre consequence. In this paper, I argue that more traditional formulations of H-D also need recourse to this new notion of content. I present a new account of the vexing notion of the natural axiomatization of a theory. The notion is used to construct a form of H-D that allows for the type of selective confirmation without which Glymour (1980a,b) claims H-D is hopeless.

The article revisits the large n (sample size) problem as it relates to the Jeffreys-Lindley paradox to compare the frequentist, Bayesian, and likelihoodist approaches to inference and evidence. It is argued that what is fallacious is to interpret a rejection of as providing the same evidence for a particular alternative , irrespective of n; this is an example of the fallacy of rejection. Moreover, the Bayesian and likelihoodist approaches are shown to be susceptible to the fallacy of acceptance. The key difference is that in frequentist testing the severity evaluation circumvents both fallacies but no such principled remedy exists for the other approaches.

Traditionally, forces are causes of a special sort. Forces have been conceived to be the direct or immediate causes of things. Other sorts of causes act indirectly by producing forces which are transmitted in various ways to produce various effects. However, forces are supposed to act directly without the mediation of anything else.

But forces, so conceived, appear to be occult. They are mysterious, because we have no clear conception of what they are, as opposed to what they are postulated to do; and they seem to be hidden from direct observations. There is, therefore, strong initial motivation for trying to eliminate forces from physics. Furthermore, as we shall explain, powerful arguments can be mounted to show that theories with forces can always be recast as theories without them. Hence it seems that forces should be eliminated, in the interests of simplicity.

We argue, however, that forces should not be eliminated—just differently construed. For the effect of elimination is to leave us without any adequate account of the causal relationships forces were postulated to explain. And this would remain the case, even if forces could be identified with some merely dispositional properties of physical systems. In our view, forces are species of the causal relation itself, and as such have a different ontological status from the sorts of entities normally considered to be related as causes to effects.

Many of the “counterintuitive” features of relativistic quantum field theory have their formal root in the Reeh-Schlieder theorem, which in particular entails that local operations applied to the vacuum state can produce any state of the entire field. It is of great interest then that I. E. Segal and, more recently, G. Fleming (in a paper entitled “Reeh-Schlieder meets Newton-Wigner”) have proposed an alternative “Newton-Wigner” localization scheme that avoids the Reeh-Schlieder theorem. In this paper, I reconstruct the Newton-Wigner localization scheme and clarify the limited extent to which it avoids the counterintuitive consequences of the Reeh-Schlieder theorem. I also argue that there is no coherent interpretation of the Newton-Wigner localization scheme that renders it free from act-outcome correlations at spacelike separation.

While orthodox (Neyman-Pearson) statistical tests enjoy widespread use in science, the philosophical controversy over their appropriateness for obtaining scientific knowledge remains unresolved. I shall suggest an explanation and a resolution of this controversy. The source of the controversy, I argue, is that orthodox tests are typically interpreted as rules for making optimal decisions as to how to behave–-where optimality is measured by the frequency of errors the test would commit in a long series of trials. Most philosophers of statistics, however, view the task of statistical methods as providing appropriate measures of the evidential-strength that data affords hypotheses. Since tests appropriate for the behavioral-decision task fail to provide measures of evidential-strength, philosophers of statistics claim the use of orthodox tests in science is misleading and unjustified. What critics of orthodox tests overlook, I argue, is that the primary function of statistical tests in science is neither to decide how to behave nor to assign measures of evidential strength to hypotheses. Rather, tests provide a tool for using incomplete data to learn about the process that generated it. This they do, I show, by providing a standard for distinguishing differences (between observed and hypothesized results) due to accidental or trivial errors from those due to systematic or substantively important discrepancies. I propose a reinterpretation of a commonly used orthodox test to make this learning model of tests explicit.

When natural selection theory was presented, much active philosophical debate, in which Darwin himself participated, centered on its hypothetical nature, its explanatory power, and Darwin's methodology. Upon first examination, Darwin's support of his theory seems to consist of a set of claims pertaining to various aspects of explanatory success. I analyze the support of his method and theory given in the Origin of Species and private correspondence, and conclude that an interpretation focusing on the explanatory strengths of natural selection theory accurately reflects neither Darwin's own self-consciously held views, nor the nature of his support. Darwin's methodological and philosophical arguments were at once consistently empiricist and more sophisticated than such interpretations credit to him.

Network analysis is increasingly used to discover and represent the organization of complex systems. Focusing on examples from neuroscience in particular, I argue that whether network models explain, how they explain, and how much they explain cannot be answered for network models generally but must be answered by specifying an explanandum, by addressing how the model is applied to the system, and by specifying which kinds of relations count as explanatory.

What has science actually achieved? A theory of achievement should (1) define what has been achieved, (2) describe the means or methods used in science, and (3) explain how such methods lead to such achievements. Predictive accuracy is one truth-related achievement of science, and there is an explanation of why common scientific practices (of trading off simplicity and fit) tend to increase predictive accuracy. Akaike's explanation for the success of AIC is limited to interpolative predictive accuracy. But therein lies the strength of the general framework, for it also provides a clear formulation of many open problems of research.