In this paper I give a thorough account of the history of the open question argument. I have provide Moore’s original impetus for it and its traditional formulation. I then examine the Cornell Realists’ objection to that original formulation and showed that their objection does indeed show the open question argument to be incorrect in its conclusions. Having presented the history of the open question argument and having assessed the most challenging objections to it, I turn to Andrew Altman’s powerful reconstruction of the open question argument in order to see how well, if at all, it sidesteps the objections leveled against the classical formulation. I then argue that while Altman does present the most coherent defense of the open question argument available, I conclude that insofar as he has rested upon a commitment to Carnap's philosophy of language over a Fregean semantic and an untenable rendering of post-Kripkean philosophy of language as it concerns rigid designation, we must reject his reformulation. Given that rigid designation itself undermines Altman’s position, I conclude that the open question is still in need of a defense before it can regain its position as a major player in the discipline of ethics.

The thesis aims at analyzing the concepts of risk and safety as well as the class of concepts to which they belong, thick concepts, focusing in particular on the normative aspects involved. Essay I analyzes thick concepts, i.e. concepts such as cruelty and kindness that seem to combine descriptive and evaluative features. The traditional account, in which thick concepts are analyzed as the conjunction of a factual description and an evaluation, is criticized. Instead, it is argued that the descriptive and evaluative aspects must be understood as a whole. Furthermore, it is argued that the two main worries evoked against non-naturalism – that non-naturalism cannot account for disagreement and that it is not genuinely explanatory – can be met. Essay II investigates the utilization of the Kripke/Putnam causal theory of reference in relation to the Open Question Argument. It is argued that the Open Question Argument suitably interpreted provides prima facie evidence against the claim that moral kinds are natural kinds, and that the causal theory, as interpreted by leading naturalist defenders, actually underscores this conclusion. Essay III utilizes the interpretation of the Open Question Argument argued for in the previous essay in order to argue against naturalistic reduction of risk, i.e. reduction of risk into natural concepts such as probability and harm. Three different normative aspects of risk and safety are put forward – epistemic uncertainty, distributive normativity and border normativity – and it is argued that these normative aspects cannot be reduced to a natural measure. Essay IV provides a conceptual analysis of safety in the context of societal decision-making, and argues for a notion that explicitly includes epistemic uncertainty, the degree to which we are uncertain of our knowledge of the situation at hand. Some formal versions of a comparative safety concept are also proposed. Essay V puts forward a normative critique against a common argument, viz. the claim that the public should follow the experts’ advice in recommending an activity whenever the experts have the best knowledge of the risk involved. The importance of safety in risk acceptance together with considerations from epistemic uncertainty makes the claim incorrect even after including plausible limitations to exclude ‘external’ considerations. Furthermore, it is shown that the scope of the objection covers risk assessment as well as risk management. Essay VI provides a systematized account of safety engineering practices that clarifies their relation to the goal of safety engineering, namely to increase safety. A list of 24 principles referred to in the literature of safety engineering is provided, divided into four major categories. It is argued that important aspects of these methods can be better understood with the help of the distinction between risk and uncertainty, in addition to the common distinction between risk and probability. / QC 20100803

Une des questions centrales de la métaéthique est celle de savoir si les propriétés morales sont des propriétés naturelles ou non-naturelles. Ce mémoire fait valoir que nous ferions bien de remettre en question une constellation d’arguments en faveur du non-naturalisme moral : l’argument de la question ouverte, l’intuition normative et l’argument du gouffre. L’influent argument de la question ouverte de Moore, d’abord, ne soutient le non-naturalisme que s’il commet une pétition de principe. L’intuition normative commet ou bien le sophisme d’inférer sur la base de sa différence spécifique qu’une chose n’appartient pas à un genre donné, ou bien sous-estime la panoplie de propriétés naturelles qui possèdent les caractéristiques censées être distinctives des propriétés morales et normatives. L’argument du gouffre, quant à lui, sous-estime l’ubiquité du fossé logique et conceptuel censé marquer une discontinuité métaphysique profonde entre les domaines normatif et naturel. / One of the burning questions among metaethical realists is whether moral facts and properties are natural or non-natural. In this thesis, I argue that we should treat a family of arguments for non-naturalism with considerable scepticism: the Open Question Argument, the Normative Intuition, and the argument from the Is-Ought Gap. Moore’s famous Open Question Argument only supports moral non-naturalism if it begs the question against the modest (non-reductionist) naturalist. As for the Normative Intuition, it either commits the fallacy of inferring on the basis of a thing’s specific difference that it does not belong to the genus it putatively belongs to, or it underestimates the breadth of natural properties that possess the features which non-naturalists allege are distinctive of moral and normative properties. The argument from the Is-Ought Gap, for its part, underestimates the ubiquity of the logical and conceptual gap that allegedly marks a deep metaphysical discontinuity between the normative and natural domains.