According to the probabilistic relevance account of confirmation, E confirms H relative to background knowledge K just in case $\mathrm{P}(\mathrm{H}/\mathrm{K}\&\mathrm{E})>\mathrm{P}(\mathrm{H}/\mathrm{K})$. This requires an inequality between the rational degree of belief in H determined relative to two bodies of total knowledge that are such that one (K&E) includes the other (K) as a proper part. In this article, I argue that it is quite plausible that there are no two possible bodies of total knowledge for ideally rational agents meeting this requirement. Hence, the positive relevance account may have to be rejected.

Confirmation is a graded notion: evidence can confirm a hypothesis to a greater or lesser degree. There has been debate about how to measure degree of confirmation. Starting from the observation that we would like evidence to be a discriminating indicator of truth, we conduct computer simulations to determine how well the various known measures of confirmation predict the extent to which a given piece of evidence fulfills that role, given a hypothesis of interest. The outcomes show that some measures are markedly better indicators of truth than others.

Sometimes retracted or refuted scientific information is used and propagated long after it is understood to be misleading. Likewise, retracted news items may spread and persist, despite being publicly established as false. In this article, we explore the dynamics of retraction using agent-based models of epistemic networks, to see why false beliefs might persist despite retraction. We find that, often, those who received false information simply fail to receive retractions because of social dynamics. Surprisingly, delaying retraction can increase its impact. Further, retractions are most successful when issued by the original source of misinformation, rather than a separate source.

Many policy decisions take input from collections of scientific models. Such decisions face significant and often poorly understood uncertainty. We rework the so-called confidence approach to tackle decision-making under severe uncertainty with multiple models, and we illustrate the approach with a case study: insurance pricing using hurricane models. The confidence approach has important consequences for this case and offers a powerful framework for a wide class of problems. We end by discussing different ways in which model ensembles can feed information into the approach, appropriate to different collections of models.

This article examines the role of historical science in clarifying the causal structure of complex natural processes. I reject the pervasive view that historical science does not uncover natural regularities. To show why, I consider an important methodological distinction in geology between uniformitarianism and actualism; methodological actualism, the preferred method of geologists, often relies on historical reconstructions to test the stability of currently observed processes. I provide several case studies that illustrate this, including one that highlights how historical narratives can improve predictive models.

I will argue that Kochen-Specker arguments do not provide an algebraic proof for quantum contextuality since, for the argument to be effective, (1) operators must be uniquely associated with measurements and (2) commuting operators must represent simultaneous measurements. However, in all Kochen-Specker arguments discussed in the literature either 1 or 2 is not met. Arguments meeting 1 contain commuting operators that do not represent simultaneous measurements and hence fail to physically justify the functional composition principle. Arguments meeting 2 associate some operators with more than one measurement and hence need to invoke an extra assumption different from noncontextuality.

I argue that explanations for time asymmetry in terms of a ‘Past Hypothesis’ face serious new difficulties. First I strengthen grounds for existing criticism by outlining three categories of criticism that put into question essential requirements of the proposal. Then I provide a new argument showing that any time-independent measure on the space of models of the universe must break a gauge symmetry. The Past Hypothesis then faces a new dilemma: reject a gauge symmetry and introduce a distinction without difference or reject the time independence of the measure and lose explanatory power.

Callender and Cohen have proposed to apply a “Gricean strategy” to the constitution problem of scientific representation. They suggest that scientific representation can be reduced to stipulation by epistemic agents. This account has been criticized for not making a distinction between symbolic and epistemic representation and not taking into account the communal aspects of representation. These criticisms would not apply if Grice’s actual strategy were properly employed. I transpose Grice’s strategy to epistemic representation. The main novelty of the resulting account is a distinction between contextual representational use and general representational status, which I address using the notion of indexicality.

Olav Benjamin Vassend proposes two solutions to the “interpretive problem” of assigning nonzero probabilities to hypotheses that are known to be false. He argues that the verisimilitude interpretation (probability expresses the degree of belief that the hypothesis is closest to the truth) and the counterfactual interpretation (probability is conditional on a false supposition) are equivalent. While Vassend’s intuition about these two solutions is basically correct, the technical details of his treatment need elaboration and correction. Appropriate tools for combining verisimilitude and Bayesian probabilities can be found in my Truthlikeness.