Mechanisms are usually viewed as hierarchical, with lower levels of a mechanism influencing, and decomposing, its higher-level behavior. To draw quantitative predictions from a model of a mechanism, the model must capture this hierarchical aspect. Recursive Bayesian networks (RBNs) were put forward by Lorenzo Casini et al. as a means to model mechanistic hierarchies by decomposing variables into their constituting causal networks. The proposal was criticized by Alexander Gebharter. He proposes an alternative formalism, which instead decomposes arrows. Here, I defend RBNs from the criticism and argue that they offer a better representation of mechanistic hierarchies than the rival account.

We show that as a chain of confirmation becomes longer, confirmation dwindles under screening-off. For example, if E confirms H1, H1 confirms H2, and H1 screens off E from H2, then the degree to which E confirms H2 is less than the degree to which E confirms H1. Although there are many measures of confirmation, our result holds on any measure that satisfies the Weak Law of Likelihood. We apply our result to testimony cases, relate it to the Data-Processing Inequality in information theory, and extend it in two respects so that it covers a broader range of cases.

In this paper, I investigate the nature of a priori biological laws in connection with the idea that laws must be empirical. I argue that the epistemic functions of a priori biological laws in biology are the same as those of empirical laws in physics. Thus, the requirement that laws be empirical is idle in connection with how laws operate in science. This result presents a choice between sticking with an unmotivated philosophical requirement and taking the functional equivalence of laws seriously and modifying our philosophical account. I favor the latter.

The aim of this paper is to introduce the instrumental variables technique to the discussion about causal inference in econometrics. I show that it may lead to causally incorrect conclusions unless some fairly strong causal background assumptions are made, assumptions which are usually left implicit by econometricians. These assumptions are very similar to, albeit not identical with, James Woodward's definition of an ‘intervention’. I discuss similarities and differences of the two points of view and argue that—understood as a practical method of causal inference—the set presented here is superior.

Von Neumann argued by means of a thought experiment involving measurements of spin observables that the quantum mechanical quantity$S_{VN}=-k\mathrm{Tr}\,(\rho \mathrm{log}\,\rho) $ is conceptually equivalent to thermodynamic entropy. We analyze Von Neumann's thought experiment and show that his argument fails.

One way in which philosophy of science can perform a valuable normative function for science is by showing characteristic errors made in scientific research programs and proposing ways in which such errors can be avoided or corrected. This paper examines two errors that have commonly plagued research in biology and psychology: 1) functional localization errors that arise when parts of a complex system are assigned functions which these parts are not themselves able to perform, and 2) vacuous functional explanations in which one provides an analysis that does account for the inputs and outputs of a system but does not employ the same set of functions to produce this output as does the natural system. These two kinds of error usually arise when researchers limit their investigation to one type of evidence. Historically, correction of these errors has awaited researchers who have employed the opposite type of evidence. This paper explores the tendency to commit these errors by examining examples from historical and contemporary science and proposes a dialectical process through which researchers can avoid or correct such errors in the future.

In a recent paper, Michael Friedman and Hilary Putnam argued that the Lüders rule is ad hoc from the point of view of the Copenhagen interpretation but that it receives a natural explanation within realist quantum logic as a probability conditionalization rule. Geoffrey Hellman maintains that quantum logic cannot give a non-circular explanation of the rule, while Jeffrey Bub argues that the rule is not ad hoc within the Copenhagen interpretation. As I see it, all four are wrong. Given that there is to be a projection postulate, there are at least two natural arguments which the Copenhagen advocate can offer on behalf of the Lüders rule, contrary to Friedman and Putnam. However, the argument which Bub offers is not a good one. At the same time, contrary to Hellman, quantum logic really does provide an explanation of the Lüders rule, and one which is superior to that of the Copenhagen account, since it provides an understanding of why there should be a projection postulate at all.

Two Bayesian approaches to choosing between statistical models are contrasted. One of these is an approach which Bayesian statisticians regularly use for motivating the use of AIC, BIC, and other similar model selection criteria, and the other one is a new approach which has recently been proposed by Bandyopadhayay, Boik, and Basu. The latter approach is criticized, and the basic ideas of the former approach are presented in a way that makes them accessible to a philosophical audience. It is also pointed out that the former approach establishes a new, philosophically interesting connection between the notions of simplicity and informativeness.

This is a critical discussion of Paul Humphreys's fusion view of emergence, focusing on the basal loss feature of his ontology. The discussion yields some general morals for special science ontology.

In this paper I argue that certain methodological conventions within psychophysics reflect a continued uncertainty over the reliability of subjects’ introspective judgements. Firstly, different psychophysical methods do not rely equally on the introspective capabilities of experimental subjects. I contrast “minimally introspective” tasks with “introspection-reliant” ones. It is only in the latter that introspection can be said to have a nontrivial role in the subjects’ performance. Secondly, my distinction maps onto a number of important methodological divisions in vision science. The introspection-reliant categorization captures many of the tasks typically considered less able to yield useful information regarding the processes underlying visual sensation.

This article considers claims that biology should seek general theories similar to those found in physics but argues for an alternative framework for biological theories as collections of prototypical interlevel models that can be extrapolated by analogy to different organisms. This position is exemplified in the development of the Hodgkin-Huxley giant squid model for action potentials, which uses equations in specialized ways. This model is viewed as an “emergent unifier.” Such unifiers, which require various simplifications, involve the types of heuristics discussed in Wimsatt's writings on reduction, but with a twist. Here, the heuristics are used to generate emergent rather than reductive explanations.

In this article I review attempts to define the term “ad hoc hypothesis,” focusing on the efforts of, among others, Karl Popper, Jarrett Leplin, and Gerald Holton. I conclude that the term is unhelpful; what is “ad hoc” seems to be a judgment made by particular scientists not on the basis of any well-established definition but rather on their individual aesthetic senses. Further, a hypothesis considered ad hoc can apparently be retroactively declared non–ad hoc on the basis of subsequent data, rendering the term meaningless.

This article considers the temporal dimension of data processing and use and the ways in which it affects the production and interpretation of knowledge claims. I start by distinguishing the time at which data collection, dissemination, and analysis occur (Data time, or Dt) from the time in which the phenomena for which data serve as evidence operate (Phenomena time, or Pt). Building on the analysis of two examples of data reuse from modeling and experimental practices in biology, I then argue that Dt affects how researchers (1) select and interpret data as evidence and (2) identify and understand phenomena.

In a recent PSA paper (2001a), as well as some other papers (1995, 2000, 2001b) and a book chapter (1999, Chapter 7), Stathis Psillos raised a number of objections against structural realism. The aim of this paper is threefold: (1) to evaluate part of Psillos’ offence on the Russellian version of epistemic structural realism (ESR); (2) to elaborate more fully what Russellian ESR involves; and (3) to suggest improvements where it is indeed failing.

This paper discusses features of an epistemically valuable form of evidence that raise troubles for received and new epistemological treatments of experimental evidence.

Reproduction is central to biology and evolution. Standard concepts of reproduction are drawn from animals. Nonstandard examples of reproduction can be found in unicellular eukaryotes that distribute their reproductive strategies across multiple generations, and in mutualistic systems that combine different modes of reproduction across multiple lineages. Examining multigenerational and multilineal reproducers and how they align fitness has implications for conceptualizing units of evolution.

The fact that the goals and methods of science, as well as its empirical conclusions, are subject to change, is shown to allow at once for: (a) the objectivity of warrant for knowledge claims; (b) the absence of a priori standards from epistemology; (c) the normative character of epistemology; and (d) the rationality of axiological innovation. In particular, Laudan's attempt to make axiological constraints undercut epistemic realism is confuted.

Scientific realists have argued that the truth(likeness) of our theories provides the only explanation for the success of science. I consider alternative explanations proposed by antirealists. I endorse Leplin’s contention that neither van Fraassen’s Darwinist explanation nor Laudan’s methodological explanation provides the sort of explanatory alternative which is called for in this debate. Fine’s suggestion— that the empirical adequacy of our theories already explains their success—is more promising for antirealists. Leplin claims that this putative explanation collapses into realism on one reading and into vacuity on another reading. But his analysis conflates three doctrines into two, and one of the three avoids both realism and vacuity.

Although on opposite sides of the logic of discovery debate, Laudan and Simon share a thesis of divorce between discovery (invention) and justification (appraisal); but unlike some other authors, they do not base their respective versions of the divorce-thesis on the empirical/logical distinction. Laudan argues that, in contemporary science, invention is irrelevant to appraisal, and that this irrelevance renders epistemically pointless the inventionist program. Simon uses his divorce-thesis to defend his account of invention, which he claims to be non-inductive—so evading the problem of induction. Underlying both authors' positions are inadequate conceptions of inductive inference. Laudan here ignores the role in contemporary science of plausibility arguments, which provide a crucial link between invention and appraisal, and thence an epistemic rationale for inventionism. Simon's account of invention does covertly call upon inductive principles from the context of appraisal, and this is what gives his program epistemic import; otherwise he would be vulnerable to Laudan's “no rationale” critique. The tensions in both authors reveal the falsity of the divorce-thesis, and the essential function of induction in both appraisal and invention of hypotheses.