1 ‘Most of us believe that there are certain people to whom we have special obligations …. Common-sense morality largely consists in such obligations,’ says Parfit, Derek ‘Prudence, Morality and the Prisoner's Dilemma,’ Proceedings of the British Academy 65 (1979), 539–64,Google Scholar 556. C. D. Broad similarly observes that ‘the common opinion certainly is that a person has a more urgent duty to benefit those who stand in certain relations to him, e.g. his children or his parents, than to benefit others who do not; and that this special urgency depends directly on these special relationships,’ see ‘Self and Others,’ Broad's Critical Essays in Moral Philosophy, ed. D.R. Cheney (London: Allen & Unwin 1971), 262-82 at 265. Here, of course, Broad and Parfit are merely reaffirming the insight of Sidgwick, Henry The Methods of Ethics, 7th ed. (London: Macmillan 1907),Google Scholar Bk. 3, Ch. 4, Sec. 2 and Bk. 4, Ch. 3.

Common sense morality tends to slide over any distinction between ‘special obligations’ and ‘special duties,’ using those terms interchangeably; and it also tends to use those terms where something like ‘special permissions’ might strictly speaking be more appropriate. For our purposes, however, this sloppy ordinary usage will suffice: nothing in our argument depends upon these confusions or upon correcting them. A.R. White alleges some interesting differences between duties and obligations in Rights (Oxford: Oxford University Press 1984), chapter 4.

2 ‘When rights arise out of special transactions between individuals or out of some special relationship in which they stand to each other, both the persons who have the right and those who have the corresponding obligation are limited to the parties to the special transaction or relationship. I call such rights special rights to distinguish them from those moral rights which are thought of as rights against (i.e., as imposing obligations upon) everyone …’ Hart, H. L.A. ‘Are There Any Natural Rights?’ Philosophical Review 64 (1955), 175–91,CrossRefGoogle Scholar 183.

3 ‘In weighing the pros and cons of a possible line of conduct, I hold myself entitled to give extra weight to a cost or benefit which is to accrue to someone with a special relation to me. The most special relation of all is identity — I am entitled to put my thumb on the moral scales on behalf of my own interests,’ says Bennett, Jonathan ‘Morality and Consequences,’ Tanner Lectures on Human Values, vol. 2, ed. McMurrin, S.M. (Cambridge: Cambridge University Press 1981), 45–116, 78;Google Scholar see similarly Fried, Charles ‘The Lawyer as Friend,’ Yale Law Journal 85 (1976), 1060–89,CrossRefGoogle Scholar 1070. All this, of course, merely echoes Bernard Williams’ long-standing concern with the preservation of one's integrity; see ‘A Critique of Utilitarianism,’ Utilitarianism: For & Against (Cambridge: Cambridge University Press 1973) and Moral Luck (Cambridge: Cambridge University Press 1981), esp. Ch. 1. Davis, Nancy ‘Utilitarianism and Responsibility,’ Ratio 22 (1980), 15–35,Google Scholar 25 ff. convincingly demonstrates that Williams’ objections to utilitarianism are traceable to its agent-impartiality, rather than to any distinctions between positive and negative responsibility. See too Scheffler, Samuel The Rejection of Consequentialism (Oxford: Clarendon Press 1982).Google Scholar

4 ‘As a parent, one is entitled, if not obligated, to prefer the interests of one's own children over those of children generally. That is to say, it is regarded as appropriate for a parent to allocate excessive goods to his or her own children, even though other children may have substantially more pressing and genuine needs for the same items.’ Wasserstrom, Richard ‘Lawyers as Professionals,’ Human Rights 5 (1975), 1–24,Google Scholar 4. ‘[W]e have special obligations [to] … the people to whom we stand in certain relations — such as our children, parents, pupils, patients, members of our own trade union, or those whom we represent. We believe we ought to help these people in certain ways. We should try to protect them from certain kinds of harm, and should try to give them certain kinds of benefit …. Carrying out these obligations has priority over helping strangers.’ Parfit, ‘Prudence, Morality and the Prisoner's Dilemma,’ 556.

5 Non-conclusive duties are sometimes described as prima facie. We avoid this way of speaking. A prima facie obligation suggests, not a call which would be conclusive if other things were equal, but a call which as things are appears (at least on first reckoning) to be conclusive. The difference is that between a potentially conclusive duty — our notion — and an apparently conclusive one.

6 Self-relativised duties are discussed under the title of agent-relativity in: Derek Parfit, ‘Prudence, Morality and the Prisoner's Dilemma,’ 555 ff.; Nagel, Thomas ‘The Limits of Objectivity,’ Tanner Lectures on Human Values, vol. 1, ed. McMurrin, S.M. (Cambridge: Cambridge University Press 1980), 75–139;Google Scholar and Sen, Amartya ‘Rights and Agency,’ Philosophy & Public Affairs 11 (1982), 3–39.Google Scholar

7 Someone may argue that every other obligation is dependent on the duty to do what is right. It is true, indeed it is logically true, that for any moral call on A to see that-p, that demand holds only if he is called upon to do what is right. The very necessity of this connection however rules it out as a case of dependence. Where the duty to see that-p is dependent on another duty to see that-q, we require that it should be no more than contingently true that A is called upon in the first regard only if he is called upon in the second. It must be logically possible that he should be bidden to see that-p without being obliged to see that-q. Dependence supposes distinctness and otherwise doubt is cast on whether the duties really are distinct.

8 The imposition of special obligations may even be susceptible of a consequentialist justification which would protect common sense against the consequentialist difficulty mentioned at the end of the introduction. Suppose that the imposition is justified by the attainment, inter alia, of the result C. This does not mean that the special obligations are dependent, after all, on a distinct obligation: the duty to try to bring about C. It may be that C, like spontaneity, is a goal which cannot be prescribed for intentional pursuit, and cannot therefore be laid down as a duty, without its achievement being compromised. See Elster, Jon Sour Grapes (Cambridge: Cambridge University Press 1983) ch. 2,CrossRefGoogle Scholar for a discussion of such goals. Bernard Williams appears to regard integrity as just such a justifying consequence; see his Moral Luck, page 49. For discussion of a related case, see Pettit, Philip ‘Satisficing Consequentialism,’ Proceedings of the Aristotelian Society, Supp. Vol. 58 (1984), 268–72.Google Scholar The case is generalised in Philip Pettit and Geoffrey Brennan, ‘Preemptive Consequentialism,’ unpublished ms.

9 As against this consideration, it may be urged that the dependence is not the direct variety discussed but an indirect sort such as rule-consequentialists invoke. A has the special duty of giving his children a good start, not because he is called upon himself to see that as many as possible do so, but rather because he is subject to a general rule the point of which is to ensure that result. The rule is: everyone ought to provide especially for his own children. This only postpones the problem. Consider the case where if A complies with the rule he will, for whatever perversity of circumstance, defeat the result sought: if he provides especially for his own children, two or more other fathers will be unable to provide for theirs. So long as the side-effects are not overwhelming, he will be bound in such a case to break the rule, neglecting his own offspring for the sake of maximising parental favouritism overall. This runs counter to our common sense intuition and shows that the ruleconsequentialist representation would denature the special obligations it is designed to save. The approach for which we shall argue does not denature the obligations in this way.

10 Broad, ‘Self and Others,’ 262, identifies two principles endorsed by Sidgwick and defines ethical neutralism in terms of them. The two principles correspond roughly to the two premises in the argument given by us below; hence the title of the argument. The case we have given the neutralist fits most naturally with consequentialism. Notice though that on one interpretation of Kant (not however the one suggested in the text below) the Kantian may also qualify as a neutralist. This Kantian sees the pursuit of obligation as always serving the adumbration of an ideal order.

11 There is also a consequentialist form. The consequentialist defender of special duties might say that we do not ascribe a duty to someone, A, to see that p, just because it is desirable that-p. We also require that it is desirable that someone in A's position should be blamed and perhaps punished if he fails to see that-p. This requirement will be satisfied only in some types of case, where the sanction of blame and punishment is likely to ensure that the desirable p-result is maximised and where the sanction does not have too many undesirable side-effects. In other sorts of cases we will not regard A's seeing that-p as the exercise of duty, only as an act of superogatory virtue. This defense of special duties is disingenuous, however. The explanation which it offers of common sense morality explains the morality away. Consider the effect that it ought to have on moral reasoning in one's case. I see something desirable which I can achieve. I remember that it will not be said to be my duty to achieve it, since there would be no utility in blaming me or my likes for failing in such a thing. Will I be any the less inclined to acknowledge that there is a call on me — a non-conclusive obligation — to see to the matter? Surely not, at least if I am sincere. The consequentialist argues, not that the desirable fails to engender duty, only that it is not always desirable to say that it does.

12 Sen, ‘Rights and Agency,’ mentions the first and sympathetically explores the second.

13 It will become clear, when the criterion of responsibility is spelled out, why there is no need (as there was in point 1 of the neutralist argument) to add the rider: given that A satisfies the appropriate qualifying conditions. A will be responsible for p only if he satisfies those conditions.

14 This is due to the danger of misunderstandings and oversights. Others, expecting that you will abide by the conventions, will focus their attention and resources on their own peculiar responsibilities. Hence, they may not notice that you have failed to discharge yours; or, if they do, they may already be otherwise committed. We assume that one agent will always be salient. If there are two or more close contenders, then it may be necessary to break the tie and restore uniqueness by reference to a lottery-like procedure.

15 It may be mentioned in passing that the most important sub-class of RD duties is probably that of duties of non-malevolence. It is required of everyone that he acts so that no harm is brought upon innocent people. Because they are RD obligations, it is misleading to categorise them, as is often done, with duties of benevolence, which are UI or perhaps UD in nature. If duties of nonmalevolence are thought to require that a person does not bring harm directly – other than by omission – on other people and nothing else, then they are cast as RI duties and would constitute a further sort of special obligation.

16 The classic source of the tradition in English literature is Bradley's Ethical Studies (2nd ed, Oxford University Press 1927). See also Hollis, Martin ‘Dirty Hands,’ British Journal of Political Science 12 (1982), 385–98.CrossRefGoogle Scholar For a comment on Bradley see Pettit, Philip ‘Social Holism and Moral Theory’ Proceedings of the Aristotelian Society, Vol. 86, 1985/6, pp. 173–97.CrossRefGoogle Scholar

17 ‘A consequential principle will say merely that such and such is (or is not) to be brought about: it leaves no room for any essential mention of who is (or is not) to bring it about.’ Mackie, J.L. Ethics (Harmondsworth: Penguin 1977), 157.Google Scholar Nagel, Thomas ‘Subjective and Objective,’ Mortal Questions (Cambridge: Cambridge University Press 1979), 196–213,Google Scholar 205 similarly contrasts ‘two ways of looking at the world. On the one hand there is the position that one's decisions should be tested ultimately from an external point of view, to which one appears as just one person among others … On the other hand … life is always the life of a particular person, and cannot be lived sub specie aeternitatis.’ Nagel confesses that in The Possibility of Altruism (Oxford: Oxford University Press 1970) he had defended a version of the former position; but he says that he has since come to see the attractions of the latter view as well.

18 Some, by virtue of their special roles as teachers or preachers or moral tutors, might also have a special duty to do so. But this is over and above the general responsibility that everyone shares.

19 In either case, it has the interesting feature of being reflexive. That is to say, it applies to the responsibility that it itself involves: everyone has the duty to act so that everyone is aware of his responsibility for acting so that everyone is aware of his responsibilities. There is no regress here, only selfsubsumption.

20 The state offers formal, legal instruments by which a person can make commitments and incur the consequent duties; and should he fail to dispatch these commitments, the state is the agent that is required to take over. That means, not that the state then has the obligation to act on his behalf, only that it has the obligation to force the person to discharge his commitments, insofar as possible.

21 In preparing this paper we were aided by comments from John Braithwaite, Richard Lindley, Onora O'Neill and Jack Smart. The paper was written during Goodin's tenure as Research Fellow in the Social Justice Project, Research School of Social Sciences, Australian National University. Some similar themes are pursued in Goodin's Protecting the Vulnerable (Chicago: Chicago University Press 1985).