² It should be pointed out, all the same, that enormous difficulties stand in the way of any purely syntactical definition of the moral ‘ought’. For we may use the term ‘ought’ whenever we wish to refer to a norm, and there are norms whose existence does not derive from human interest, and yet which are not moral. If I say ‘The eclipse ought to occur at 3.20’, then, grammatically speaking, I have uttered a categorical ‘ought’. But I have not made a moral judgement. Any view that attempts to overcome this objection by ruling that ‘ought’ is simply ambiguous should be regarded with suspicion: are we so sure that we can tell when a term has more than one sense? (On these questions see Roger Wertheimer's interesting, if somewhat negative, discussion in The Significance of Sense, Cornell, 1972.)Google Scholar In the course of this paper I suggest tests for the moral use of ‘ought’ that are not merely syntactic; but these tests emerge as a consequence of our moral philosophy, and cannot be included in its premises.

³ Philosophers of a Quinean cast of mind will of course insist that the same is true of any judgement. But it is not necessary to decide this issue, since, even for the Quinean, there will be ‘constraints’ imposed by one's scheme of concepts, and the question then becomes whether these constraints apply in the case of moral judgements, or whether they are unique to the empirical (or ‘scientific’) realm.

⁴ For developments of this idea see: Stig Kanger, ‘New Foundations for Ethical Theory’, and Hintikka, Jaakko, ‘Some main Problems of Deontic Logic’, both in Deontic Logic: Introductory and Systematic Readings, ed. Hilpinen, Risto (Dordrecht, 1971).Google Scholar Perhaps the most elegant statement of the formal analogy between the concept of obligation and the alethic modalities occurs in Montague, Richard, ‘Logical Necessity, Physical Necessity, Ethics, and Quantifiers’, Inquiry, 1960.Google Scholar

⁵ Hintikka, , op. cit., pp. 73–5.Google Scholar

⁶ On this point, see Williams, B. A. O., ‘Ethical Consistency’, P.A.S.S., 1965.Google Scholar

⁷ Examples are suggested by the discussion in Casey, John, ‘Actions and Consequences’, in Morality and Moral Reasoning, ed. Casey, John (London, 1971).Google Scholar

⁸ Cf. Foundations of the Metaphysic of Morals, pp. 454 ff. (Original pagination.)Google Scholar

⁹ It is important to see, however, that the logical possibility of irresoluble dilemmas is admitted, as soon as we accept that there are or could be imperatives that are categorical. Therefore, if deontic logic is to be what it claims to be, it must accommodate dilemmas, and not dismiss them.

¹⁰ Cf. Hare, R. M., The Language of Morals (Oxford, 1952) Pt. II.Google Scholar

¹¹ Williams, , op. cit., pp. 123–4.Google Scholar

¹² We may conclude at once, therefore, that Utilitarianism provides no direct answer to our problem.

¹³ Bennett, J. F.: ‘The Simplicity of the Soul’, J. Phil. 1967, see esp. p. 654.Google Scholar

¹⁴ Some philosophers (including Aristotle) have spoken of the conclusion of practical reasoning as an action. However, an action may fall under more than one description, but will not be justified by the practical syllogism under each of these descriptions. This suggests that we should speak of the intention, rather than the act, as the true conclusion.

¹⁵ Kant, , op. cit., pp. 417–18.Google Scholar Hill, Thomas E. Jr., in ‘The Hypothetical Imperative’, Phil. Rev. 1973, gives reasons for doubting that this was Kant's view.Google Scholar

¹⁶ Hare, , op. cit., p. 37.Google Scholar

¹⁷ Thus modern philosophers have often argued that the form of reasoning ‘X wants Y, X believes that Z is the means to Y, therefore X decides on Z’ is logically conclusive, by surreptitiously interpreting ‘X wants Y’ as ‘X decides on Y’. Examples are: Davidson, Donald, ‘How is Weakness of the Will Possible’ in Moral Concepts, ed. Feinberg, Joel (London, 1969)Google Scholar; and Von Wright, G. H., The Varieties of Goodness (London, 1963) pp. 168–71.Google Scholar

¹⁸ Cf. Kenny's discussion of long-term and short-term desires: Kenny, Anthony: Action, Emotion and Will (London, 1963) p. 124.Google Scholar

¹⁹ The strategy of this argument is, of course, based on Aristotle's Nichomachean Ethics.

²⁰ See Aristotle, op. cit., 1126a.

²¹ Aristotle, op. cit., 1097b; Kant, , op. cit., pp. 415–16.Google Scholar

²² A point worth emphasising against the crasser kinds of Utilitarianism, particularly those that lean on the Theory of Preference.

²³ Aquinas: Summa Theologica, Ia 2ae 30, 4.