1 For a detailed discussion of these matters, see David Braybrooke's Meeting Needs: Studies in Moral, Political, and Legal Philosophy (Princeton: Princeton University Press 1987). I hope to escape leaving 'needs' undefined because the parameters of the definition will partly be determined by the system itself. The secondary utilitarian element of the system represents the claims of preference, informed or otherwise. What are stipulated as 'needs' in the system should be such as to render plausible that degree of precedence, secured by the system's rules, of needs over preferences. In this context, basic needs become those things which are so important that they must be satisfied before other sorts of utility are attended to. 'One might hope that such needs will be Universal, but one might have to settle for needs which are culturally and historically determined.

2 Needs-based and rights-based theories can be related by appealing to the commonplace view that there are rights to the satisfaction of basic needs. Some people might even think that the existence of a basic need is a sufficient (though perhaps ultimately consequentialist) ground of a moral right to the means of satisfying it. (Braybrooke, 136, is reluctant to countenance this.) Such people might find tools for elaborating their viewpoint in Sumner's, L.W. The Moral Foundation of Rights (Oxford: Clarendon Press 1987), esp. 175-205.Google Scholar

3 See Derek, Parfit Reasons and Persons (Oxford: Clarendon Press 1984) 356Google Scholar. All subsequent references to 'Parfit' are to this work.

4 Or a principle of average utility: in same-number choices it doesn't matter which you pick. Later on I will be using the principle of average utility, and for simplicity's sake the reader should keep in mind that such a principle could be introduced to S without changing anything else.

5 There is no reason why more extensive liberties, or for that matter autonomy, could not be considered needs. This would resolve the more straightforward conflicts between autonomy and utility to the satisfaction of some rights advocates. But even making autonomy the sole basic need would not satisfy strong rights proponents: it will turn out that (i) autonomy is susceptible to 'maximization' in the sense that alternatives which make more people autonomous will be preferred; (ii) conflicts between the autonomy of two persons will be 'resolved' by doing whatever maximizes utility. Roughly speaking, a need for autonomy will satisfy only those for whom autonomy is something valuable in itself, but not for that reason inalienable or incommensurable with other things. For a defense of this position, see James Griffin, 'Towards a Substantive Theory of Rights,' in Frey, R.G. ed., Utility and Rights (Minneapolis: University of Minnesota Press 1984) 137-60Google Scholar.

6 In this sketch I have not attempted a complete or bullet-proof formalization of the rules. Rule (i), as stated, seems to block any argument someone might actually want to make to prefer M to N or vice versa. However, a complete Rule (i) would add that v(XYZ) is also not greater or less than [v(X) + v(Y) + v(Z)]. It also would add conditions to deal with cases involving any number of needs, not just three, and with comparisons between different numbers of individual and combined needs.

7 This system, prior to modification for 'different-number cases,' differs from the one suggested in David Braybrooke's Meeting Needs primarily in its treatment of conflicts among both needs and those who have them. His rules, as far as I can see, are indeterminate when it comes to choosing between satisfying some people's need for B and more people's need for A, or in the distribution of needs packages such as the last one discussed under Ru1e (iii). Moreover, Braybrooke seems to say that, in cases of unresolved conflict between needs, there is a breakdown in the system: it is indeterminate at such points. In my system, there is determinacy because the 'Principle of Precedence' (for needs over matters of preference only) is suspended in such cases, and the principle of utility takes over. Not everyone will prefer this! However, my system adjudicates more conflicts of needs than does Braybrooke's, so the breakdowns occur less often. Cf. Braybrooke, Meeting Needs (Princeton: Princeton University Press 1987), esp. 164CrossRefGoogle Scholar, 171.

8 See Parfit's Third Hell case below.

9 I would like to thank Manal Stamboulie and Wayne Sumner for their comments on earlier versions of this paper. David Braybrooke's remarks in his capacity as a refusing-to-remain-anonymous reviewer were also extremely pertinent and helpful!