1 This notion of unitariness is not easily characterized. For instance, the universal precept ‘Always adopt the practices of your host country’ is syntactically very simple but is not unitary in my sense. Evidently the unitary/non-unitary distinction admitsof a large grey area, and for that reason it might be better to treat such terms as scaling adjectives. Thus we should speak of a moraltheory, not as being unitary/relativist, but as being more unitary/less relativist than another.

² Pace Williams, Bernard, who makes much of this point in ‘The Truth in Relativism’, Proceedings of the Aristotelian Society 75 (1974–1975), 215-228, and reprinted in Moral Luck (Cambridge University Press, 1981).Google Scholar

³ At least this would seem to be so if the condition is understood in the way in which Hare, for example, does. It may prove to be more difficult to get past Kant, in view of the latter's exceedingly severe strictures on the possible role of inclinations as morally relevant features. (I am grateful to Suzanne Stern-Gillet for this point.)

⁴ There is nothing logically suspect about the notion of an uncommanded command, and we need not fear that moral concepts will lose their force altogether if their imperative associations evaporate. After all, many of our abstract concepts evolved via metaphor from ideas relating to the will (witness abstract itself), and are none the worse for having lost all their volitional significance.