A deontologist might refuse to kill one to save five, and yet take a one-in-a-million risk of killing one to avert a one-in-a-million risk of five dying. No prominent decision theory can accommodate this pair of preferences, as weighting outcomes by a uniform probability (or risk-weighted probability) preserves their ranking.
Risk-sensitive decision theories (which permit agents to have non-neutral attitudes towards risk) require them to have the same attitude towards all kinds of risk. But there are good reasons to think that if it can be rational to have a general risk attitude across all domains, then it can be rational to have domain-specific risk attitudes: to be, say, risk averse for epistemic goods, risk neutral for pleasure, and risk seeking for aesthetic value. I develop the formal resources to model these attitudes.
I argue that the characteristic feature of deontology is a particular kind of domain-specific risk attitude. Specifically, the deontologist thinks duty violations are worse than consequentialist harms, but is comparatively risk-seeking with respect to duty violations relative to consequentialist harms.
This paper examines whether doxastic dilemma pose a challenge to pragmatism. A doxastic dilemma arises when epistemic reasons support believing p, while practical reasons support believing not-p, raising the question of what one ought to believe all things considered. Evidentialists argue that pragmatists must answer this question and determine which belief is normatively required. This challenge, however, depends on a belief principle: if one ought to believe p, then one ought not to believe not-p. I argue that pragmatists need not accept this principle. By considering cases in which practical reasons conflict with one another, I show that even incompatible belief contents may each be supported by distinct reasons and thus possess their own normative standing. Therefore, the fact that one cannot simultaneously hold both beliefs does not show that only one has normative support. Doxastic dilemmas, thus, do not successfully undermine pragmatism.
The problem of justification has long been the central question of philosophy. At stake is what accounts as sufficient justification for claiming that a statement genuinely refers to the object in question. Unless we address this ontological gap between language and the external world, we cannot confidently claim that our statements are directly related to the object we aim to describe. With this, this paper proposes a redefinition of the problem of justification by examining the persistent gap between language and the external world. Traditional accounts of epistemic justification often assume a relatively direct relation between language and reality. However, linguistic mediation complicates this relation: our access to the world is structured through concepts, interpretive frameworks, and socially conditioned practices of meaning. By clarifying the distinction between linguistic representation and worldly causation, this study redefines justification not primarily as a property of the world or isolated beliefs, but what counts as a meaningful and responsible representation of the world. By situating justification within our linguistic practices, this approach offers a more precise account of how beliefs can be answerable to a world that is never accessed except through language, concepts, and interpretations.
It's widely reported that the classical Bayesian norms of rationality break down in the face 'higher-order evidence'. On first pass, this needn’t be particularly troubling. Classical Bayesianism has many well-documented limitations. However, these limitations don't undermine its application across a wide range of cases. So, they don't directly challenge Bayesian orthodoxy. On close inspection, the problem runs deeper though. Hedden & Dorst (2022) have recently argued that almost all evidence is (in part) ‘higher-order evidence’. If they are right, and it is in fact the case that the classical Bayesian norms break down in the face of ‘higher-order evidence’, it follows that the classical Bayesian norms break down in almost all cases. In this paper, I defuse this problem. I show that it rests on the conflation of two distinct kinds of 'higher-order evidence’. Once we properly attend to the distinction between these two kinds of higher-order evidence we can accept Hedden & Dorst’s (2022) initial observation without endorsing wide-spread violations of classical Bayesianism. Hence, we can make a more accurate assessment of the scope and limits of Bayesian epistemology.
Following Dorst (2026), this paper develops a higher-order account of epistemic vagueness. I begin by reconstructing Williamson’s epistemicist account of vagueness, focusing on the margin principle. I argue that epistemic vagueness cannot be explained by first-order indiscriminability alone, since such an explanation is self-undermining. A satisfactory account therefore appeals to higher-order uncertainty. I then clarify that the margin principle rests on two assumptions: the indiscriminability of just-noticeable differences (JNDs) and gradualness. I suggest that the former must be restricted, while the latter is incompatible with empirical data and unnatural. Finally, I argue that higher-order uncertainty provides a better explanation of non-gradual patterns in epistemic vagueness, and that this explanation is supported by psychological studies on discrimination and metacognition.
