¹ ‘The Subjection of John Stuart Mill’, Philosophy, lxviii (1993)Google Scholar.

² The Subjection of Women, CW, xxi. 304 f. And: ‘I deny that anyone knows, or can know, the nature of the two sexes, as long as they have only been seen in their present relation to one another …What is now called the nature of women is an eminently artificial thing – the result of forced repression in some directions, unnatural stimu-lation in others’. The Subjection of Women, CW, xxi. 276.References to Mill are to the Collected Works of John Stuart Mill, ed. Robson, John, Toronto, 1963–91Google Scholar.

³ Brecher, Bob, ‘Why Patronize Feminists? A Reply to Stove on Mill’, Philosophy, lxviii (1993)Google Scholar; Thiel, Inari, ‘On Stove on Mill’, Philosophy, Ixix (1994)Google Scholar; Downing, F. Gerald, ‘A Cynical Response to the Subjection of Women’, Philosophy, Ixix (1994)Google Scholar.

⁴ Utilitas, x (1996)Google Scholar.

⁵ Stove, 8.

⁶ Thiel, 100.

⁷ Brown, 126.

⁸ Roger Crisp adopts a similar position. He expresses doubts concerning some aspects of Mill's feminism, but finds no problem with the basic coherence of his position, merely noting that ‘[as] a strict empiricist, ready only to accept the evidence of experience, Mill says that we can know nothing of the nature of women, other than, presumably because we are aware of the forces of distortion at work, that their present state is not their natural one’. Mill on Utilitarianism, London, 1997, p. 208Google Scholar. However, the ‘presumably’ here might suggest reservations on Crisp's part, either as to what Mill's reasons in fact are, or what they are worth.

⁹ The Subjection of Women, CW, xxi. 276 f.

¹⁰ Ibid., p. 313.

¹¹ Mill's use of the natural/artificial distinction is ambiguous. Sometimes ‘artificial’ is used as above, to denote ‘forced’ as opposed to ‘spontaneous’ or ‘natural’ human development. At other times he opposes the ‘artificial’ to ‘instinct and impulse’ and uses it to denote the humanly natural, and essentially social, process of ‘educating the natural passions’ of which he is, of course, very much in favour. E.g.: ‘Allowing every–thing to be an instinct which anybody ever asserted to be one, it remains true that nearly any respectable attribute of humanity is the result, not of instinct, but of victory over instinct; and that there is hardly anything valuable in the natural man except capacities – a whole world of possibilities, all of them dependent upon eminently artificial discipline for being realised’. Nature, CW, x. 393. Soble's mistaken view, that Mill is committed to a policy of, as far as possible, removing women from the influence of society as such, arises from a failure to note this distinction. Soble, Alan, ‘The Epistemology of the Natural and the Social in Mill's The Subjection of Women’, Mill Newsletter, xvi (1981)Google Scholar.

¹² ‘If men had ever been found is society without women, or women without men …something might have been positively known about the mental and moral differences which may be inherent in the nature of each’. The Subjection of Women, CW, xxi. 276. The supposition is clearly intended as a reductio of the claim that we know what the natural differences are.

¹³ Ibid., pp. 264 f.

¹⁴ Principles of Political Economy, CW, ii, bk. 2, ch. 1.

¹⁵ The Subjection of Women, CW, xxi. 270.

¹⁶ Ibid., p. 269.

¹⁷ Ibid., p. 270.

¹⁸ Ibid., pp. 271 f.

¹⁹ This is, of course, a salient difficulty associated with so–called ‘positive’ conceptions of freedom, whereas Mill is typically regarded as an exponent of ‘negative’ liberty. In fact Mill's position is decidedly equivocal. It is possible to distinguish seven concepts of freedom in Mill, three of which are clearly ‘positive’. See Smith, G. W., ‘The Logic of J. S. Mill on Freedom’, Political Studies, xxviii (1980)Google Scholar.

²⁰ Just how ‘thick’ this conception of female nature needs to be is perhaps debatable. It might be argued that all Mill needs to do to justify a policy of female emancipation is to put the onus of proof on those who claim that subordination is enabling rather than inhibiting for women – to show that women would not prosper more in a situation of sexual equality with men. The moral weight that can be properly attached to ‘equalizing’ claims of this kind will be considered in sect. IV. Here it is sufficient to note that the strategy involves, at the very least, the claim that we have reason to believe that women are in relevant respects like men. And this clearly infringes IC. On the question of whether identifying constraints necessarily involves assumptions about human nature see e.g. Benn, S. I. and Weinstein, W. L., ‘Being Free to Act, and Being a Free Man’, Mind, lxxx (1971)CrossRefGoogle Scholar; Smith, G. W., ‘Slavery, Contentment and Social Freedom’, Philosophical Quarterly, xxvii (1977)Google Scholar; and Gray, J., ‘Freedom, Slavery and Contentment’; in Liberalisms (London, 1989), pp. 69–88Google Scholar.

²¹ Brown, 124–6.

²² The Subjection of Women, CW, xxi. 313. See also n. 27; and The Subjection of Women, CW xxi 280, 304 f.

²³ Brown, 126.

²⁴ Thus: ‘so far as the whole course of human improvement up to this time, the whole stream of modern tendencies, warrants any inference on the subject [of the unnatural-ness of female subordination], it is, that this relic of the past is discordant with the future, and must necessarily disappear’. The Subjection of Women, CW, xxi. 272. And: ‘It may be asserted without scruple, that no other class of dependents have had their character so entirely distorted from its natural proportions by their relation with their masters; for, if conquered and slave races have been, in some respects, more forcibly repressed, whatever in them that has not been crushed down by an iron heel has generally been let alone, and if left with any liberty of development, it has developed itself according to its own laws; but in the case of women, a hot-house and stove cultivation has always been carried on of some of the capabilities of their nature, for the benefit and pleasure of their masters’.Ibid., p. 276.

²⁵ Thus: ‘History … shows the extraordinary susceptibility of human nature to external influences and the extreme variations of those of its manifestations which are supposed to be most universal and uniform.'’ Ibid., p. 277.

²⁶ A System of Logic, CW, viii. 860–74.

²⁷ ‘The profoundest knowledge of the laws of the formation of character is indis-pensable to entitle anyone to affirm even that there is any difference, much more what the difference is, between the two sexes considered as moral and rational beings; and since no one, as yet, has that knowledge … no one is thus far entitled to any positive opinion on the subject. Conjectures are all that can at present be made; conjectures more or less probable, according as more or less authorized by such knowledge as we yet have of the laws of psychology, as applied to the formation of character’. The Subjection of Women, CW, xxi. 277 f.

²⁸ The Subjection of Women, CW, xxi. 277.

²⁹ Collini, S., ‘Introduction’, CW, xxiGoogle Scholar. xix–xxi; and cf. , Mill's defence of racial equality in The Negro Question, CW, xxi. 92–4Google Scholar.

³⁰ Mill's use of examples can sometimes suggest that his argumentative strategy is much cruder. Thus: ‘It cannot be inferred to be impossible that a woman should be a Homer or an Aristotle because no women has yet actually produced works comparable to theirs … This negative fact at most leaves the question uncertain … but it quite certain that a women can be a Queen Elizabeth, or a Deborah, or a Joan of Arc, since this is not an inference, but a fact’. The Subjection of Women, CW, xxi. 302. This might be construed as an attempt to refute the naturalist simply by appeal to the impossibility of proving a universal negative. But it is a straw person. The naturalist need not claim that women never equal the best performances of men, only that if they do not, the causes are au fond natural rather than social. And though the ‘positive’ fact here is admittedly undeniable, as evidence in favour of the proposition that the causes thereof are social it clearly involves an inference and, on Mill's own terms, a conjectural one at that.

³¹ See Smith, G. W., ‘Freedom and Virtue in Politics: Aspects of Character, Circum-stances and Utility from Helvétius to J. S. Mill’, Utilitas, i (1989)Google Scholar, and ‘Enlightenment Psychology and Individuality: the Roots of J. S. Mill's Conception of the Self’, Enlighten-ment and Dissent, xi (1992)Google Scholar.

³² Feuer, L. S., ‘John Stuart Mill as a Sociologist: the Unwritten Ethology’, in Robson, John M. and Laine, Michael (eds.), James & John Stuart Mill: Papers of the Centenary Conference, Toronto, 1976, pp. 86–110Google Scholar.

³³ Mill's ‘two greatest theoretical errors as a scientific thinker’, according to Bain, are his ‘doctrine of natural equality’ and his ‘disregard of the physical conditions of our mental life’. J. S. Mill, London, 1882, pp. 146–8Google Scholar.

³⁴ Stephen cites Mill's ‘especially significant’ comment in his Autobiography: ‘I have long felt that the prevailing tendency to regard all the marked distinctions of human character as innate, and to ignore the irresistible proofs that by far the greater part of these differences, whether between individuals, races or sexes, are such as not only might, but naturally would be, produced by differences in circumstances, is one of the chief hindrances to the rational treatment of great social questions, and one of the greatest stumbling-blocks to human improvement’. ‘An Attempted Philosophy of His-tory’, Fortnightly Review, N.S., xxvii (1880), 677 fGoogle Scholar.

³⁵ See Smith, , ‘Logic’, 246 fGoogle Scholar.

³⁶ This is not, of course, to say that Mill does not have other theoretical irons in the fire here too, only that agnostic liberalism represents a significant theoretical com-mitment for Mill, and that the implications need to be drawn out.

³⁷ The Subjection of Women, CW, xxi. 326.

³⁸ See e.g. Smith, G. W. (ed.), John Stuart Mill's Social and Political Thought: Critical Assessments, 4 vols., London, 1998, vol. 1, pt. 3Google Scholar, ‘Utility and Rules’, passim.

³⁹ The Subjection of Women, CW, xxi. 275.

⁴⁰ Utilitarianism, CW, x. 275.

⁴¹ Thus, the criminal laws of evidence place a particularly heavy burden of proof of guilt upon the prosecution, but even here the test, for obvious practical reasons, is limited to ‘reasonable doubt’.

⁴² Stove, 9.