The so-called problem of irrelevant conjunction has been seen as a serious challenge for theories of confirmation. It involves the consequences of conjoining irrelevant statements to a hypothesis that is confirmed by some piece of evidence. Following Hawthorne and Fitelson, we reconstruct the problem with reference to Bayesian confirmation theory. Then we extend it to the case of conjoining irrelevant statements to a hypothesis that is disconfirmed by some piece of evidence. As a consequence, we obtain and formally present a novel and more troublesome problem of irrelevant conjunction. We conclude by indicating a possible solution based on a measure-sensitive approach and by critically discussing a major alternative way to address the problem.

Today, mechanisms and mechanistic explanation are very popular in philosophy of science and are deemed a welcome alternative to laws of nature and deductive-nomological explanation. Starting from Mitchell's pragmatic notion of laws, I cast doubt on their status as a genuine alternative. I argue that (1) all complex-systems mechanisms ontologically must rely on stable regularities, while (2) the reverse need not hold. Analogously, (3) models of mechanisms must incorporate pragmatic laws, while (4) such laws themselves need not always refer to underlying mechanisms. Finally, I show that Mitchell's account is more encompassing than the mechanistic account.

According to Marr, a computational-level theory consists of two elements, the what and the why. This article highlights the distinct role of the Why element in the computational analysis of vision. Three theses are advanced: (a) that the Why element plays an explanatory role in computational-level theories, (b) that its goal is to explain why the computed function (specified by the What element) is appropriate for a given visual task, and (c) that the explanation consists in showing that the functional relations between the representing cells are similar to the “external” mathematical relations between the entities that these cells represent.

In evolutionary biology changes in population structure are explained by citing trait fitness distribution. I distinguish three interpretations of fitness explanations—the Two-Factor Model, the Single-Factor Model, and the Statistical Interpretation—and argue for the last of these. These interpretations differ in their degrees of causal commitment. The first two hold that trait fitness distribution causes population change. Trait fitness explanations, according to these interpretations, are causal explanations. The last maintains that trait fitness distribution correlates with population change but does not cause it. My defense of the Statistical Interpretation relies on a distinctive feature of causation. Causes conform to the Sure Thing Principle. Trait fitness distributions, I argue, do not.

Hierarchical Bayesian models (HBMs) provide an account of Bayesian inference in a hierarchically structured hypothesis space. Scientific theories are plausibly regarded as organized into hierarchies in many cases, with higher levels sometimes called ‘paradigms’ and lower levels encoding more specific or concrete hypotheses. Therefore, HBMs provide a useful model for scientific theory change, showing how higher-level theory change may be driven by the impact of evidence on lower levels. HBMs capture features described in the Kuhnian tradition, particularly the idea that higher-level theories guide learning at lower levels. In addition, they help resolve certain issues for Bayesians, such as scientific preference for simplicity and the problem of new theories.

Bayesian probabilistic explication of inductive inference conflates neutrality of supporting evidence for some hypothesis H (“not supporting H”) with disfavoring evidence (“supporting not-H”). This expressive inadequacy leads to spurious results that are artifacts of a poor choice of inductive logic. I illustrate how such artifacts have arisen in simple inductive inferences in cosmology. In the inductive disjunctive fallacy, neutral support for many possibilities is spuriously converted into strong support for their disjunction. The Bayesian “doomsday argument” is shown to rely entirely on a similar artifact.

The paper defends the doctrine that Linnaean taxa, including species, have essences that are, at least partly, underlying intrinsic, mostly genetic, properties. The consensus among philosophers of biology is that such essentialism is deeply wrong, indeed incompatible with Darwinism. I argue that biological generalizations about the morphology, physiology, and behavior of species require structural explanations that must advert to these essential properties. The paper concludes by summarizing my responses to the objection that, according to current “species concepts,” species are relational, and to the objection that essentialism cannot accommodate features of Darwinism associated with variation and change.

Kurt Gödel criticizes Rudolf Carnap's conventionalism on the grounds that it relies on an empiricist admissibility condition, which, if applied, runs afoul of his second incompleteness theorem. Thomas Ricketts and Michael Friedman respond to Gödel's critique by denying that Carnap is committed to Gödel's admissibility criterion; in effect, they are denying that Carnap is committed to any empirical constraint in the application of his principle of tolerance. I argue in response that Carnap is indeed committed to an empirical requirement vis-à-vis tolerance, a fact that becomes clear upon closer scrutiny of Carnap's relevant writings.

Critics of the ideal of value-free science often assume that they must reject the distinction between epistemic and nonepistemic values. I argue that this assumption is mistaken and that the distinction can be used to clarify and defend the argument from inductive risk, which challenges the value-free ideal. I develop the idea that the characteristic feature of epistemic values is that they promote, either intrinsically or extrinsically, the attainment of truths. This proposal is shown to answer common objections to the distinction and provide a principled basis for separating legitimate from illegitimate influences of nonepistemic values in scientific inference.

In this article and its sequel, we derive Bayesianism from the following norm: Accuracy—an agent ought to minimize the inaccuracy of her partial beliefs. In this article, we make this norm mathematically precise. We describe epistemic dilemmas an agent might face if she attempts to follow Accuracy and show that the only measures of inaccuracy that do not create these dilemmas are the quadratic inaccuracy measures. In the sequel, we derive Bayesianism from Accuracy and show that Jeffrey Conditionalization violates Accuracy unless Rigidity is assumed. We describe the alternative updating rule that Accuracy mandates in the absence of Rigidity.

The architects of the modern synthesis banned essentialism from evolutionary theory. This rejection of essentialism was motivated by Darwin's theory of natural selection, and the continuity of evolutionary transformation. Contemporary evolutionary biology witnesses a renaissance of essentialism in three contexts: “origin essentialism” with respect to species and supraspecific taxa, the bar coding of species on the basis of discontinuities of DNA variation between populations, and the search for laws of evolutionary developmental biology. Such “new essentialism” in contemporary biology must be of a new kind that accommodates relational (extrinsic) properties as historical essences and cluster concepts of natural kinds.

Whether meaning is compositional has been a major issue in linguistics and formal philosophy of language for the last 2 decades. Semantic holism is widely and plausibly considered as an objection to the principle of semantic compositionality therein. It comes as a surprise that the holistic peculiarities of scientific language have been rarely addressed in formal accounts so far, given that semantic holism has its roots in the philosophy of science. For this reason, a model-theoretic approach to semantic holism in the language of science is presented here. This approach preserves compositionality to a large extent.

Rayleigh and Ramsay discovered the inert gas argon in the atmospheric air in 1895 using a carefully designed sequence of experiments guided by an informal statistical analysis of the resulting data. The primary objective of this article is to revisit this remarkable historical episode in order to make a case that the error-statistical perspective can be used to bring out and systematize (not to reconstruct) these scientists't resourceful ways and strategies for detecting and eliminating error, as well as dealing with Duhemian ambiguities and underdetermination problems as they arose in the context of their local research settings.

I argue that a certain type of naturalist should not accept a prominent version of the no-miracles argument (NMA). First, scientists (usually) do not accept explanations whose explanans-statements neither generate novel predictions nor unify apparently disparate established claims. Second, scientific realism (as it appears in the NMA) is an explanans that makes no new predictions and fails to unify disparate established claims. Third, many proponents of the NMA explicitly adopt a naturalism that forbids philosophy of science from using any methods not employed by science itself. Therefore, such naturalistic philosophers of science should not accept the version of scientific realism that appears in the NMA.

When a technique purports to provide information that is not available to the unaided senses, it is natural to think that the only way to validate that technique is by appealing to a theory of the processes that lead from the object of study to the raw data. In fact, scientists have a variety of strategies for validating their techniques. Those strategies can yield multiple independent arguments that support the validity of the technique. Thus, it is possible to produce a robust body of data with a single technique. I illustrate and support these claims with a historical case study.

A common argument against explanatory reductionism is that higher-level explanations are sometimes or always preferable because they are more general than reductive explanations. Here I challenge two basic assumptions that are needed for that argument to succeed. It cannot be assumed that higher-level explanations are more general than their lower-level alternatives or that higher-level explanations are general in the right way to be explanatory. I suggest a novel form of pluralism regarding levels of explanation, according to which explanations at different levels are preferable in different circumstances because they offer different types of generality, which are appropriate in different circumstances of explanation.

The causal theory of reference is often taken to provide a solution to the problems, such as incomparability and referential discontinuity, that the meaning-change thesis raised. I show that Kuhn successfully questioned the causal theory and Putnam's idea that reference is determined via the sameness relation of essences that holds between a sample and other members of a kind in all possible worlds. Putnam's single ‘essential’ properties may be necessary but not sufficient to determine membership in a kind category. Kuhn argued that extension is fixed by similarity-dissimilarity relations that are liable to change in taxonomic reorganizations of science.

The received view in the philosophy of biology is that biological taxa (species and higher taxa) do not have essences. Recently, some philosophers (Boyd, Devitt, Griffiths, LaPorte, Okasha, and Wilson) have suggested new forms of biological essentialism. They argue that according to these new forms of essentialism, biological taxa do have essences. This article critically evaluates the new biological essentialism. This article's thesis is that the costs of adopting the new biological essentialism are many, yet the benefits are none, so there is no compelling reason to resurrect essentialism concerning biological taxa.

In this article and its prequel, we derive Bayesianism from the following norm: Accuracy—an agent ought to minimize the inaccuracy of her partial beliefs. In the prequel, we make the norm mathematically precise; in this article, we derive its consequences. We show that the two core tenets of Bayesianism follow from Accuracy, while the characteristic claim of Objective Bayesianism follows from Accuracy together with an extra assumption. Finally, we show that Jeffrey Conditionalization violates Accuracy unless Rigidity is assumed, and we describe the alternative updating rule that Accuracy mandates in the absence of Rigidity.