1 Urmson, J.O., “The Interpretation of the Moral Philosophy of J.S. Mill”, Philosophical Quarterly, 3 (1953), 33–39.CrossRefGoogle Scholar

2 Urmson cites E. F. Carritt as holding the “received view”. See The Theory of Morals (London, 1952), chapter IV. See also Moore, G.E., Principia Ethica, (Cambridge, reprinted 1968), p. 106.Google Scholar Since Urmson, many writers have defended the received view. The terms “act-” and “rule-utilitarianism” were introduced, I believe, by Brandt, R.B., Ethical Theory (Englewood Cliffs, N.J., 1956), pp. 380–406.Google Scholar

3 See, e.g., Mabbott, J.D., “Interpretations of Mill's Utilitarianism”, Philosophical Quarterly, 6 (1956), 115·120,Google Scholar esp. pp. 115-116; Mandelbaum, Maurice, “Two Moot Issues in Mill's Utilitarianism”, in Schneewind, J.B., Mill: A Collection of Critical Essays (Notre Dame, 1969), 206-233,Google Scholar esp. pp. 206-207; Sosa, Ernest, “Mill's Utilitarianism” in Smith, James M. and Ernest Sosa, Mill's Utilitarianism: Text and Criticism (Belmont, Cal., 1969), 154-172,Google Scholar esp. p. 154. However, Sosa seems to recognize that Mill maybe a utilitarian whose theory is of neither of these types. See pp. 160·161.

4 A systematic discussion of the relations. between act-utilitarianisms (or “simple-utilitarianisms“), rule-utilitarianisms and forms of utilitarian generalization is found in Lyons, David, Forms and Limits of Utilitarianism (Oxford, 1965).CrossRefGoogle Scholar

5 Or, better, “in violation of any set of rules to which no set of rules is preferred on the utilitarian ground specified”. This formulation allows for the possibility of ties. Here and throughout the paper I ignore the desirability of qualifying utilitarian principles so that the rightness and wrongness of actions turn, not on the actual value of their consequences, or on the actual value of certain states of affairs obtaining, but on their expected value. Compare the definitions of R.B. Brandt, op. cit., pp. 380 ff., and 396 ff.; and of David Lyons, op. cit., pp. 3, 9, 11, 25. Theories I call “act-utilitarian”, Lyons calls theories of “simple-utilitarianism”.

6 I should qualify this by admitting that it may be possible to devise a rule-utilitarianism which is extensionally equivalent to Mill's theory, although I see no obvious way of doing this. Nothing in my argument turns on whether this is possible.

7 See D.P. Dryer, “Mill's Utilitarianism”, an introductory essay in John Stuart Mill, Essays on Ethics, Religion and Society, ed. J.M. Robinson (Toronto, 1969), which is volume X of Collected Works of John Stuart Mill, lxii-cxiv, esp. pp. xcviii and cii. Dryer also attributes to Mill a form of “iterated-utilitarianism”, but he and I differ on the details. I arrived at my view independently of Dryer. See below, part 3.

8 Mill, John Stuart, Utilitarianism, chapter V, paragraph 14.Google Scholar I will cite chapter and paragraph numbers in giving references to this work. Subsequent references will appear in the text.

9 Lyons, David, “Mill's Theory of Morality”, Nous, 10 (1976), p. 106,CrossRefGoogle Scholar also pp. 102-5 (hereafter cited as Lyons (1976)). Lyons’ argument was anticipated by Sosa, op. cit., pp. 160-161.

10 Lyons (1976), p. 103.

11 Mill frequently uses “expedient” interchangeably with “productive of utility” or even with “required by considerations of utility”, and “expediency” interchangeably with “general utility”. See Utilitarianism, chapter V, paragraphs, 1, 2, 6, 10, 13, 16, 24, 32, 36, 37, 38. In chapter II, third paragraph from the end, Mill discusses uses of “expedient” on which the expedient is opposed to the useful, or to that which is productive of utility. Here he suggests that anything which is expedient in the highest degree, or which is expedient simpliciter, and not merely expedient for some parochial or temporary interest, is productive of more utility than any alternative would be. It is true that he characterizes these restricted uses as uses on which the expedient is “opposed to the right”. But this does not require that what is maximally expedient simpliciter be identified with the right, and so it does not require an act-utilitarian reading. In the official discussion of chapter V, the simply expedient is, I will suggest, treated as neither opposed to the right, nor as identical with the right, but as including the right. In what follows, I will write that an action A is “maximally expedient” in a situation Just in case there is no maximal alternative to A and no alternative to A which would produce consequences of equal value to those which performing A would produce. ·

12 J. O. Urmson, op. cit., David Lyons, (1976), pp. 109, 112, 114, 115. However, in “Mill's Theory of Justice”, Lyons is careful to point out that on his reading, there is a contrast between Mill's view and “ordinary rule-utilitarianism”. Mill would accept a rule-utilitarian principle, but not as his basic principle. For Mill, the basic utilitarian principle is “a claim about ultimate ends”. Mill's “analysis of obligation”, together with this basic utilitarian principle, leads to a rule-utilitarian position (pp. 10-11). The difference between Mill's view and the ordinary rule-utilitarian view, Lyons says, is in “the way Mill's moral theory develops” (p. 11). Now, on Mill's analysis of obligation, the “turning point” of the distinction between moral wrongness and mere inexpediency lies in the fact that “moral Judgements … concern the Justification of sanctions“- (p. 9). On Lyons’ reading, Mill's remarks about the “turning point” of this distinction are a basic part of his analysis of moral obligation and are more central to his position than the derivative rule-utilitarian principle he accepts. Hence, Mill's remarks can be taken at face value, and Lyons’ reading escapes the present difficulty. Nevertheless, Lyons’ interpretation is subject to the difficulty I indicate in the next paragraph. See Goldman, A.L. and Kim, J., eds., Values and Morals, (Dordrecht, Holland, 1978), pp. 1–20.CrossRefGoogle Scholar

13 For sake of simplicity, I delete the quantifiers and reference to circumstances in subsequent formulations.

14 See also V, 17. I am taking the “ought” of clause (b) to be a “non-moral ought”. D.G. Brown seems to object to such a reading in his “Mill on Liberty and Morality”, The Philosophical Review , 81 (1972), p. 156 (hereafter cited as Brown, (1972)). In support of his objection, Brown cites a remark Mill makes in a letter of 28 November 1859 to Dr. W.G. Ward. See The Letters of John Stuart Mill, ed. by Elliot, H.S.R., Vol. I, p. 229.Google Scholar However, Mill's remark is inconclusive. Mill says “It appears to me that to [those who have a true moral feeling] the word ought means that if they act otherwise they shall be punished by this internal and perfectly disinterested feeling” (my italics). He goes on to suggest that he takes himself to have given an account of the moral feelings in Utilitarianism. He does not say that the word “ought” has the indicated meaning wherever it appears in this essay. Moreover, an attempt to use this “definition” to elucidate clause (b) would make it quite implausible. Whether S is morally required to do A would turn in part on whether anyone would be punished by his conscience if S were not punished in any way for failing to do A.

15 Lyons, David, “Human Rights and the General Welfare”, Philosophy and Public Affairs, 6 (1977), p. 123Google Scholar (hereafter cited as Lyons (1977)).

16 One might hypothesize that Mill was led to this view by his principle of liberty. In On Liberty Mill contended that a person's liberty to do something can Justifiably be interfered with only if his doing that thing would be harmful to others. Suppose he thought that one ought to suffer the reproaches of conscience for doing something harmful to others. Then he could have concluded that a person ought to be punished by law or public opinion for doing something only if he ought also to suffer the reproaches of conscience (see Brown (1972), and the Library of Liberal Arts edition of On Liberty, C.V. Shields, ed., p. 13).

Mill's having this view could partly explain why, even if he held the conscience theory, he might not feel the need to emphasize that it is punishment by conscience which is appropriate in the case of wrongful behaviour, as opposed to punishment by law or public opinion. Thus, there is a passage in which Mill appears to summarize the theory of obligation he developed in Utilitarianism but does not mention the “reproaches of conscience”. This passage is in a note to his edition of James Mill's Analysis of the Phenomena of the Human Mind. See the second edition (1869, reprinted New York, 1967), vol. II, pp. 324-6, esp. p. 325. This passage was pointed out to me by D.C. Brown, and it is potentially embarassing. However, what is said in Utilitarianism must take priority over what is said in a note to someone else's work.

17 Letter of 28 November 1859, Eliot, H.S.R., ed., The Letters of John Stuart Mill, vol. 1 (London, 1910), pp. 227–231,Google Scholar esp. p. 229. Cited in Brown (1972), p. 156.

18 D.P. Dryer, op. cit., p. cv, also pp. xcviii, and cii.

19 An iterated theory is also attributed to Mill by D.G. Brown in “Mill's Criterion of Wrong Conduct”, submitted to Dialogue. Brown would have his theory meet the second of these criteria, but it fails to meet the first.

20 For example, see Lewis, David, Counterfactuals (Cambridge, Mass., 1973)Google Scholar, According to this theory (pp. 48-9), “If it were the case that α then it would be the case that β is true at a possible world i if and only if either (1) no world accessible from i is one at which α is true, or (2) there is a world k accessible from i such that, for any world J which is at least as similar to the world i as is the world k, “if α then β” is true at J. The analysis is relativized to a given “comparative similarity system” which assigns to each world i an ordering of worlds with respect to their “comparative similarity” to i, and a set of worlds “accessible from” i.

21 W.V.O. Quine explains the device of “quasi-quotation” used here in Mathematical Logic (Cambridge, Mass., 1965), pp. 33-37.

22 Lewis discusses this assumption in Counterfactuals, pp. 19-21. He discusses the notion of a possible world and the notion that possible worlds can be ranked with respect to their overall similarity to any possible world in in Counterfactuals, pp. 48-9, 84-95. It may be, as Lewis suggests, that different ran kings of possible worlds with respect to their similarity to a world i can be given, depending on which respects of similarity or difference are taken to be most important for theoretical purposes. If so, I would have to abandon the simplifying assumption made in the text that there is a unique pair of sets of worlds, ,for any world i and sentence α. The analysis would have to be relativized to some “comparative similarity system” and hence to some pair. See Lewis, pp. 48-9.

23 See Mill, J .S., “Remarks on Bentham's Moral Philosophy”, reprinted in Smith, James M. and Sosa, Ernest, Mill's Utilitarianism (Belmont, Cal., 1969), 20–30,Google Scholar esp. pp. 24-5.

24 I owe this case to Steven Davis and similar cases to Philip Hanson, David Zimmerman, and others. These people and Raymond Jennings suggested approaches to solving the difficulty.

25 This would be a case of generic spill-over, even though the term might seem inappropriate here.

26 Rodger Beehler pointed out this problem.

27 Rawls, John, A Theory of Justice (Cambridge, Mass., 1971), pp. 157Google Scholar ff. I have briefly discussed Rawls’ position on this issue in relation to his own theory in “Justice and the Difference Principle”, Canadian Journal of Philosophy, 4 (1974), 229-240.

28 See Lyons (1977), pp. 128.129.

29 Although my interpretation of Mill differs from David Lyons', I was led to my view by reflecting on Lyons’ lectures on Mill given at Cornell in 1974. I benefited greatly from discussing the ideas developed here with D.G. Brown. For their helpful comments, I would also like to thank Philip Hanson and those who contributed to discussions of a shorter version of this paper when it was read to the British Columbia Tri-University Conference, January 1978, and to the 1978 Congress of the Canadian Philosophical Association in London, Ontario.