Utilitarianism seems to require us to sacrifice a person if doing so will produce a net increase in the amount of utility. This feature of utilitarianism is extremely unattractive. The puzzle is how to reject this requirement without rejecting the plausible claim that we are often wise to trade lesser amounts of utility for greater amounts. I argue that such a position is not as paradoxical as it may appear, so long as we understand the relationship between the value of utility and the value of persons in a certain way. I argue that the traditional utilitarian position assumes an inadequate view of this relationship. I suggest a more plausible view of this relationship, one which implies that we may not sacrifice a person merely in order to produce a net gain in utility, where that utility does not result from the saving of any other persons' lives.

This paper criticizes the conception of applied ethics as the top-down application of a theory to practical issues. It is argued that a theory such as utilitarianism cannot override our intuitive moral perceptions. We cannot be radically mistaken about the kinds of considerations which count as practical reasons, and it is the task of theoretical ethics to articulate the basic kinds of considerations which we appeal to in practical discussions. Dworkin's model of doing ethics ‘from the inside out’ is used to illustrate the appropriate role for theory in a broader sense. In conclusion, some sceptical questions are raised about how far theoretical ethics can contribute to public policy, especially if this requires a consensus.

Sidgwick's first explicit statement of the utilitarian position, in an essay presented to the Metaphysical Society in 1873, provides a lucid overview of the errors to be avoided and the terms to be clarified in any adequate account of the subject. As a précis of the comprehensive treatment of utilitarianism that would soon appear in The Methods of Ethics, this essay should serve as a useful guide to that work.

In Kantian Consequentialism, David Cummiskey argues that the central ideas of Kant's moral philosophy provide claims about value which, if applied consistently, lead to consequentialist normative principles. While Kant himself was not a consequentialist, Cummiskey thinks he should have been, given his fundamental positions in ethics. I argue that Cummiskey is mistaken. Cummiskey's argument relies on a non-Kantian idea about value, namely that value can be defined, and objects with value identified, conceptually prior to and independent of the choices that a rational agent would make. The contrasting Kantian concept of value is that to possess value is to be the object of (one sort or other) of rational choice. Inasmuch as Cummiskey gives no reason to reject the Kantian account of value in favour of his own (consequentialist) account, his argument does not establish that Kant's ethics inevitably leads to normative consequentialism.

In ‘Utility of Religion’, Mill argues that a wholly naturalistic religion of humanity would promote individual and social welfare better than supernatural religions like Christianity; in ‘Theism’, however, Mill defends the salutary effects of hope in an afterlife. While commentators have acknowledged this discrepancy, they have not examined the utilitarian value of what Mill terms ‘illusions’. In this essay, I explain Mill's case against the utility of supernatural religious belief and then argue that Mill cannot dismiss the utility of hope in an ultimate justice since it need not pervert the intellect or morality. There are thus utilitarian grounds to support some supernatural illusions, which undermines Mill's defence of an exclusively naturalistic religion. I conclude with the suggestion that while the utility of religious belief leads Mill toward William James's view, they disagree about whether supernatural religious sentiment has any unique, intrinsic force.

Non-naturalism has a shady reputation. This reputation is undeserved, at least in the case of one variety of non-naturalism – the variety Sidgwick offers. In section I, I present Sidgwick's view, distinguishing it from views with which it is often lumped. In II and III, I defend Sidgwick against recent objections to non-naturalism from motivation and supervenience. In IV, I briefly consider objections which brought about the downfall of non-naturalism at the middle of the century. In V, I consider the role Sidgwick's arguments for non-naturalism play in Methods I.3. In VI, I contrast Sidgwick's attitude toward analytic metaethics to that of Moore and the non-cognitivists.

Mill's feminism has been attacked as being logically incoherent. The general verdict has been that Mill can easily be defended from the charge. However, both sides in the debate have ignored the fact that his feminism is part of a broader theory of liberal empiricism. Placing The Subjection of Women in this context re–opens the question of its logical credentials and reveals a basic weakness in Millian feminism.

The resilience of utilitarian ethics in the face of unremitting criticism can be explained in part by its use of various strategies of indirect utilitarianism. The success of these strategies throws up a distinctive problem: how can one measure the utility of moral rules, large-scale social institutions or character traits distinctive of virtues? Reading Hume as a utilitarian of sorts in his treatment of justice (and rejecting contractarian readings), I explain his conservative endorsement of entrenched social practices as a consequence of his broadly functionalist approach. I claim that this account has enough merit to ground conservativism in ethics as a satisfactory default position. Projects for reform rather than established institutions are the proper object of utilitarian assessment, thus finessing the problem of measurement I opened up initially.

This article develops an unconventional perspective on the utilitarianism of Bentham and Mill in at least four areas. First, it is shown that both authors conceived of utility as irreducibly multi-dimensional, and that Bentham in particular was very much aware of the ambiguity that multi-dimensionality imposes upon optimal choice under the greatest happiness principle. Secondly, I argue that any attribution of intrinsic worth to any form of human behaviour violates the first principles of Bentham's and Mill's utilitarianism, and that this renders both authors immune to the claim by G. E. Moore that they committed a ‘naturalistic fallacy’. Thirdly, in light of these contentions, I find no flaw in Mill's ‘proof of utility’. Fourthly, I use the notion of intrapersonal utility weights to provide an interpretation of Mill's qualitative hedonism that is entirely consistent with his value monism.

Although people generally agree that innocent targets of culpable aggression are justified in harming the aggressors in self-defence, there is considerable disagreement regarding whether innocents are justified in defending themselves when their doing so would harm other innocent people. I argue in this essay that harming innocent aggressors and active innocent threats in self-defence is indeed justified under certain conditions, but that defensive actions in such cases are justified as permissions rather than as claim rights. This justification therefore differs from that of self-defence against culpable aggressors, since defensive acts of the latter type are justified as claim rights rather than mere permissions. I argue, however, that the two justifications are alike in that both rest on considerations of distributive justice.

This essay, one of the last that Frankena wrote, provides a scrupulously detailed exploration of the various possible meanings of one of Sidgwick's most famous footnotes in the Methods Long intrigued by what Sidgwick had in mind when he said that he would explain how it came about that for moderns it is not tautologous to claim that one's own good is one's only reasonable ultimate end, Frankena uses this note as a point of departure for a penetrating review of Sidgwick's insights and ambiguities on the differences between ancient and modern ethics.

In this article I shall undertake a preliminary exploration of the notion of second best. I shall follow a three-step strategy. First, I shall introduce some applications of the theorem of the second best in different fields of philosophy and social sciences. Secondly, I shall make several conceptual distinctions related to the theorem. I aim to show that there are certain theoretical results that are similar but not identical to the theorem of the second best, and that the notion of second best is often used quite loosely. Finally, I shall try to shed some light on the question of how the theorem is description-sensitive. The overall aim of the paper is to pay attention to the generality of the problem of the second best and stress the importance of the theorem in political theory in particular.

Apart from a short introduction, this contribution consists of a translation of Tadeusz Kotarbinski's ‘Utilitarianism and The Ethics of Pity’ (1914). In that concise and relatively unknown early note, written before he embarked on his long and influential career as a nominalist logician and philosopher of science, Kotarbinski formulates four astonishingly ‘modern’ objections to utilitarianism. Unlike Christian ‘ethics of pity’, utilitarian ethics (i) disregards the normative importance of the distinction between preventing suffering and promoting happiness, (ii) leaves no room for supererogation, (iii) leaves no room for agent-relativity in morality in so far as it disallows ‘inefficient’ self-sacrifice, and (iv) rejects the possibility of genuine ethical dilemmas. To what extent was Kotarbinski a pioneer in his critique? This question is posed but not answered.

Sidgwick maintains, plausibly, that the concept of a person's good is a normative one and takes for granted that it is normative for the agent's own choice and action. I argue that the normativity of a person's good must be understood in relation to concern for someone for that person's own sake. A person's good, I suggest, is what one should (rationally) want for that person in so far as one cares about him, or what one should want for him for his sake. I examine Sidgwick's defence of the axioms of rational prudence and argue that it is powerless to convince anyone who lacks self-concern or thinks he has no reason to care for himself. To the extent that Sidgwick is persuasive, I argue, it is because he insinuates an assumption of self-concern. Similarly, Sidgwick's defence of the axiom of rational benevolence tacitly assumes, not just impersonality, but equal concern.

This paper discusses various possible interpretations of a complex footnote in Mill's Utilitarianism.

Sher's notion of deserved punishment has unacceptable implications. It does not justify punishing some serious wrongdoers, who are unwilling to commit lesser wrongs, more severely than minor offenders. It requires victim-inflicted punishments which repeat the wrongdoings, with the roles reversed. But if Sher moves away from such victim-inflicted punishments, then his theory should treat wrongdoers like tort-feasors who have to pay monetary compensations to their victims.

This paper examines T. H. Green's and Henry Sidgwick's differing views of desireand the will, and connectedly, their differing views of an individual's good and freedom. It is argued that Sidgwick makes effective criticisms of Green, but that important elements in Green's idealist view of an individual's good and freedom survive the criticism and remain significant today. It is also suggested that Sidgwick's own account of an individual's good is unclear in an important way.

This paper examines the undervalued role of Herbert Spencer in Sidgwick's thinking. Sidgwick recognized Spencer's utilitarianism, but criticized him on the ground that he tried to deduce utilitarianism from evolutionary theory. In analysing these criticisms, this paper concludes that Spencer's deductive methodology was in fact closer to Sidgwick's empiricist position than Sidgwick realized. The real source of Sidgwick's unhappiness withSpencer lies with the substance of Spencer's utilitarianism, namely its espousal of indefeasible moral rights.

There are several different forms of comparability involving prudential values. Comparisons of values in the abstract, of realizations of some value, and of options which realize values, are distinct, and related, though not mutually exclusive. Furthermore, if rough equality is thought of as an evaluative relation in terms of which comparisons can be made, it does not imply incomparability. If it involves epistemic vagueness, this does not imply incomparability, since our not knowing which relation holds does not imply that no relation holds. Finally, if it is true to some degree that some relation holds, there may be some fuzzy ordering. Given that we may not be able coherently to talk about degrees of truth, I propose an open-ended comparability condition, which takes account of rough equality, epistemic vagueness and fuzziness.