1 Kant's works have been referenced in the following way, using Akademie pagination:

2 It need not be a psychological fact about us that we always have this as our end, but to the extent we do not, we are rejecting an end we rationally ought to have. So much seems the point of Kant's argument for a duty not to commit suicide in his discussions of the Categorical Imperative at GW 422 and 429.

3 Herman, B. The Practice of Moral Judgment (Cambridge, MA: Harvard University Press 1993), 65Google Scholar. My interpretation of Kant's views on beneficence draws to a great degree on Herman's persuasive treatment of them.

4 Allen Buchanan argues that Kant's ‘impossibility in willing’ rationale supports the much stronger conclusion that we cannot will that others ever fail to provide aid — that is, that on any occasion in which aid is needed, providing it is morally obligatory; thus there is in fact no discretion or latitude allowed in beneficence (Buchanan, A. ‘The Right to a Decent Minimum of Health Care,’ Philosophy and Public Affairs 13 (1984) 55–78, at 74-5)Google ScholarPubMed.

5 Compare Herman: ‘One might view the idea of taking another's ends as my own not in the sense that I should be prepared to act in his place (I act for him; I get for him what he wants when he cannot) but, rather, in the sense that I support his status as a pursuer of ends, so that I am prepared to do what is necessary to help him maintain that status. We might say “I help him pursue-his-ends” and not “I help him in the pursuit of his ends.” This interpretation … leads me to view the well-being of another as something more than the (passive) satisfaction of his desires. What I support is the other's active and successful pursuit of his self-defined goals’ (70).

6 Kant apparently had worries along these lines himself. He expresses reservations about state welfare provision along with his personal concern for the poor in Reflexion 8000: ‘The helpless poor must be supported and, when they are children, be well cared-for. Why? because we are men and not beasts. This (follows) flows not from the rights of the poor as citizens but from their needs as men. Not just the innocent; for then there would be few. Who shall support them? The question is not whether it is the State or the citizen, for when the State provides for them, the citizen provides for them as well, but instead, whether from the free will of the citizens or through coercion they will be supported — as a gift or as a tax. The latter brings out a competition for provisions among the candidates; it is a means of acquisition and also a claim to entitlement (titulus der Anspruche). And who shall determine helplessness?’ (GS XIX, 578)

7 Rosen, A. Kant's Theory of Justice (Ithaca, NY: Cornell University Press 1993), 191Google Scholar

8 The instrumental view is defended by Gregor, Mary Laws of Freedom (Oxford: Blackwell 1963), 36Google Scholar; Williams, Howard Kant's Political Philosophy (Oxford: Blackwell 1983), 193ff.Google Scholar; Kersting, Wolfgang ‘Kant's Concept of the State,’ Essays on Kant's Political Philosophy Williams, H. ed. (Cardiff: University of Wales Press 1992), 164n.Google Scholar; and Aune, Bruce Kant's Theory of Morals (Princeton, NJ: Princeton University Press 1979), 157, 160Google Scholar. Pogge, Thomas takes the instrumental rationale to be primary, but argues that the ‘enlightenment’ of the populace might support a more expansive notion (‘Kant's Theory of Justice,’ Kant-Studien 7 [1988] 407–33, at 422n)Google Scholar. Murphy, Jeffrie like Rosen, takes Kant to justify the public provision of welfare from beneficence (Kant: The Philosophy of Right [London: Macmillan 1970], 146)CrossRefGoogle Scholar. A similar rationale for welfare is offered by Gewirth, Alan although it is not ascribed to Kant (‘Private Philanthropy and Positive Rights,’ Social Philosophy and Policy 4 [1985] 60–7)Google Scholar. Grcic, Joseph (‘Kant on Revolution and Economic Inequality,’ Kant-Studien 77 [1987] 447–57, at 455)Google Scholar. and Linden, Harry van der (Kantian Ethics and Socialism [Indianapolis: Hackett 1988], 203)Google Scholar offer justifications for welfare which ostensibly depend on Kant's principles of autonomy; however, I think they (like Pogge’ s second rationale) are best understood as resting on freedom, which is the approach Mulholland, Leslie takes (Kant's System of Rights [New York: Columbia University Press 1990]).Google Scholar I consider this line of argument in Section V below.

9 If it were the duty of the ruler to tax the wealthy to provide for the poor, we would expect Kant to say so, rather than merely adverting to an ‘authorization’ to do so. I thank an anonymous referee for pointing this out.

10 I thank an anonymous referee for suggesting this line of thought. An argument to this effect is made (in a non-Kantian context) by Sterba, James (‘From Liberty to Welfare,’ Ethics 105 [1994] 64–98, at 78f)CrossRefGoogle Scholar; Allen Buchanan argues for a ‘decent minimum of health care’ on just these grounds (‘The Right to a Decent Minimum of Health Care,’ 70ff.; see also Buchanan, A. ‘Justice and Charity,’ Ethics 97 [1987] 558–75)CrossRefGoogle Scholar.

11 At points Rosen suggests something like this argument (cf. Rosen, 188, 204), but it gets tangled with (i) the idea that the state as a moral person has its own duty of beneficence, or (ii) the idea that individual ends of self-preservation justify the general will in providing public welfare (200-2, 206-7), or (iii) the idea that a social minimum is necessary as a condition of external freedom. The arguments from self-preservation and freedom I consider below.

12 See Beito, David “‘This Enormous Army”: The Mutual Aid Tradition of American Frtemal Societies before the Twentieth Century,’ Social Philosophy and Policy 14 (1997) 20–38CrossRefGoogle Scholar; and Davies, Stephen ‘Two Conceptions of Welfare: Voluntarism and Incorporationism,’ Social Philosophy and Policy 14 (1997) 39–68CrossRefGoogle Scholar, for evidence to this effect in British and American history of the eighteenth and nineteenth centuries.

13 Rousseau distinguishes the general will, which ‘considers only the general interest,’ from the ‘will of all,’ which ‘considers private interest and is merely the sum of private wills’ (Rousseau, J.-J. Basic Political Writings trans. Cress, D. [Indianapolis: Hackett 1987], 155)Google Scholar. As Mulholland observes (333), this is not Kant's conception of the general will, as it lacks the a priori nature which is crucial for Kant. My understanding of Kant's account of the general will is greatly indebted to Mulholland.

14 See Mulholland (266ff.) for a fuller discussion of this ‘deduction.’

15 Kant says a kind of ‘despotism’ results from ‘an executive power in which all citizens make decisions about and, if need be, against one (who therefore does not agree); consequently, all, who are not quite all, decide, so that the general will contradicts both itself and freedom’ (PP 352).

16 Rosen notes that Kant assumes that every rational being seeks the maximum amount of external freedom and the minimum restrictions on liberty, so that ‘maximum equal liberty is one of the rational ends of the general will, or of a people considered as a whole’ (137). Cf. also Kersting (‘Kant's Concept of the State,’ 151).

17 Rosen attempts to avoid this problem by claiming that since the welfare institutions are an expression of the general will, they are not coercively imposed. He claims that ‘In a Kantian republic, governments are popularly elected and the policies they adopt, including those concerned with social welfare and taxation, are not, if honestly presented at election time, coercively imposed on an unwilling citizenry …. Their governing by popular consent means that their policies bear the imprint of the popular will …. they can be represented, within the context of the liberal conception of electoral consent, as voluntary restrictions imposed by a people on its own freedom’ (Rosen, 204). But there is no support for this in anything Kant says; worse, it manifestly rests on the misconception of the general will rejected above.

18 David Schmidtz observes that even if our minimal needs for self-preservation are rationally required, having them assured by the state may not be. As he puts the point, ‘That we are better off having food, clothing, and shelter is beyond question; that we are better off having guaranteed provision of such things is not’ (Schmidtz, D. ‘Guarantees,’ Social Philosophy and Policy 14 [1997] 1–19, at 1)CrossRefGoogle Scholar.

19 Hillel Steiner claims, for instance, that to be rightful initial distributions of property must be the object of social contract. He proposes that a refinement on Henry George's strategy to collectivize and distribute land rents may be required by such a contract (Steiner, H. ‘Liberty and Equality,’ Political Studies 29 [1981] 555–69)CrossRefGoogle Scholar. While Steiner does not purport to interpret Kant's political theory, the conception of freedom he articulates is a good fit for this line of argument. Other arguments from liberty to welfare are offered by J. Sterba, ‘From Liberty to Welfareߣ; Cohen, G.A. ‘Capitalism, Freedom, and the Proletariat,’ The Idea of Freedom: Essays in Honour of Isaiah Berlin Ryan, A. ed. (Oxford; Oxford University Press 1979)Google Scholar; and Loevinsohn, Ernest ‘Liberty and the Redistribution of Property,’ Philosophy and Public Affairs 6 (1976) 226–39.Google Scholar

20 Kersting, W. ‘Politics, Freedom, and Order: Kant's Political Philosophy,’ Cambridge Companion to Kant Guyer, P. ed. (New York: Cambridge University Press 1992), 351Google Scholar

21 The move from freedom to equality and vice versa is permissible in Kant's system because, as Kersting observes (‘Kant's Concept of the State,’ 152), for Kant liberty and equality are ‘mutually definable.‘

22 As does Kant's third principle of law, independence (TP 290), which, as Kersting himself argues, must be understood as an ‘economic characteristic’ (‘Kant's Concept of the State,’ 153), though Kersting argues that the standing of this ‘contingent’ principle with freedom and equality cannot be justified within Kantian theory.

23 Kant says that someone trying to ‘wrest the apple from my hand’ is wronging me ‘with regard to what is internally mine (freedom)’ (MM 248): ‘if I am holding a thing (and so physically connected with it), someone who affects it without my consent (e.g. snatches an apple from my hand) affects and diminishes what is internally mine (my freedom), so that his maxim is in direct contradiction with the axiom of Right’ (MM 250). Kant's idea seems to be that our innate right to freedom extends to our bodies and even to the land — the space on earth — we must occupy as embodied beings: he says ‘dragging me from my resting place’ is morally on par with wresting the apple from my hand (MM 248).

24 ‘Now all men rightfully possess the area of the Earth wherein Nature or chance has placed them without their will, and so according to an innate right (before all acts establishing a right) in the possession of the earth on which they arrive, as the supreme condition of the possibility of use of the same so long as this is straightforwardly necessary for the bare preservation of their existence’ (GS XXIII, 318).

25 It might be argued by the defender of welfare rights that this betrays a crucial defect in Kant's conception of innate right, that a better conception would entail that others are obligated to provide for us if that is necessary for self-preservation. In response note, first, that even if this is right, it is not in any event the view Kant actually held, and second, that given his overall moral theory he had reason to reject this stronger thesis: it would unilaterally impose on others an even stronger obligation than what is imposed by the view of innate right he actually held.

26 There are those, of course, who through no fault of their own are not able to participate in such cooperative arrangements (e.g. through physical disabilities), hence whose means of self-preservation are problematic. But it is important to see that an appeal to the innate right to freedom will not solve that problem, since such persons would likewise be incapable of securing their own survival absent the property rights structure that civil society implements. The argument from freedom is of no help to such cases, which often motivate concern for public welfare in the first place. By way of contrast, the right to welfare defended by James Sterba is ‘conditional on people doing all that they legitimately can to provide for themselves’ (Sterba, 83n.; cf. also 77); in the system he defends crippling disabilities would ipso facto trigger that right. But Kant's innate rights are not sensitive to these sorts of incapacities, nor do they impose the burden of support on others that Sterba's view does. Whether such cases are satisfactorily addressed by Kant's instrumental rationale for welfare is a further question.

27 I am grateful for helpful comments on earlier versions of this paper to Thomas Christiano, Avery Kolers, Andrews Reath, Steven Scalet, David Schmidtz, Houston Smit, and anonymous referees, and for financial support to the Institute for Humane Studies and the Lambe Charitable Foundation.