I. The “Reverse” Theory

There is a passage in Anarchy, State, and Utopia (1974) where Robert Nozick voices the following general misgiving about the claim that people have affirmative rights to various things, such as social assistance or welfare:

It may seem a nice idea to base some human rights on material needs, but Nozick is saying that, in the real world, the resources that would have to be used to satisfy these needs may already be owned by private individuals. And if all the relevant resources are privately owned, there is simply nothing to be done: the property rights that particular individuals have over particular things “fill the space of rights, leaving no room for general rights to be in a certain material condition.”

Nozick's argument here assumes that claims based on need occupy a relatively superficial role in a general theory of economic entitlement. It is as though we first figure out who owns what, applying the principles that determine property rights, and then figure out whose needs are left unsatisfied and what is to be done about them. On this approach, welfare rights (if there are any) live only in the interstices of property. Some of what Nozick says suggests that this subordination of welfare to property is more or less unavoidable: “Things come into the world already attached to people having entitlements over them” (ASU, 160). He thinks this is clearest in the case of body parts: you may need these kidneys or this retina, but my entitlement to them is necessarily prior to yours, for we cannot even grasp my status as a person without comprising in that status a rightful claim to the limbs, cells, and organs that make me who I am (ASU, 206). Nozick thinks that the subordination of welfare to property holds for some external objects as well: “Isn't it implausible that how holdings are produced and come to exist has no effect at all on who should hold what?” (ASU, 155). The trouble with welfare rights, on this account, is that they “treat objects as if they appeared from nowhere, out of nothing” (ASU, 160). “[I]s this the appropriate model,” he asks, “for thinking about how the things people produce are to be distributed?” (ASU, 198).

The virtue of the passage that I quoted at the beginning of this essay (ASU, 238) is that it forces us to confront issues of moral priority in the economic sphere. All sides to the debate can benefit from this. Often in philosophical discussions about welfare, we assume that it is already known who owns what and that the question is simply whether people can be forced to contribute out of their own wealth to relieve the needs of the poor. Nozick observes correctly that this way of arguing underestimates the challenge of welfare rights. The idea that a person's right to use or receive resources might be based generally on need, rather than on some proprietorial relation between this person and this thing, poses a threat to the conventional understanding of property. Property entitlements will never be secure if they are subject to challenge on this basis. So, he thinks, we have to choose: Either particular people have acquired property entitlements in particular things, in which case others cannot have welfare rights that can be enforced against them, or people in general have welfare rights to material and economic security, and others cannot have property rights that can be used to resist the demands of these welfare rights. One or the other of these statements can be true, but not both. We do not have to accept Nozick's resolution of this conflict in order to agree with him about the importance of confronting it honestly and explicitly.

Nozick's position is that “particular rights over things fill the space of rights, leaving no room for general rights to be in a certain material condition.” He maintains that the alternative is to stipulate the general rights first, and then try to fit property entitlements around them. “[T]o my knowledge,” he says, “no serious attempt has been made to state this ‘reverse’ theory” (ASU, 238). Is this right? Has no one made a serious attempt to use welfare rights to define the environment in which property rights operate? Has no one made a serious attempt to construct the mirror image of Nozick's view—that is, a conception in which welfare rights fill up most of the space of rights in respect of material things, leaving little room (or less room than one would intuitively think) for rights to private property?

In the “reverse” approach, rights based on needs would be fundamental in our theory of human access to and use of resources. And instead of making these rights the basis of a duty of charity incumbent upon existing property-holders, we would take them as a basis for calling property arrangements into question, so that the existence of unsatisfied need would become an objection not just to the way property rights are being exercised (e.g., selfishly, thoughtlessly, etc.) but to their very shape and distribution. In proclaiming welfare rights, we would not be begging property-holders to be a little more generous. Rather, we would be asking this question: By what right do they claim to hold something as exclusively their own in the face of others' needs? We would insist that property answer at the tribunal of need, not the other way round.

Philosophers have toyed with versions of such a “reverse” approach as part of the reaction to Anarchy, State, and Utopia. In “Property Rights and Interests” (1979), Virginia Held compares acquisitive economic activity to a game, “where winners and losers compete on friendly terms as if engaged in a sporting event.” The idea is that winners are allowed to keep their winnings and “losers may prefer having had a chance to play the game and especially having a chance to play again and win, than to … be assured all along of an even apportionment of all proceeds.”²

But Held did not think that this game-playing could possibly trump the serious business of securing welfare rights:

Elsewhere she suggests, provocatively, that welfare rights are really the only property rights there are. People are entitled to what is needed for a decent life and adequate self-development. “Beyond this there are no moral rights to property.” Held concedes that people do have property interests that go beyond this, but she denies that there is anything wrong (as there would be in the case of rights) with dealing with these in a utilitarian way.⁴

This seems to me a fine example of what Nozick would call the “reverse” theory.

More notoriously, John Rawls's argument in A Theory of Justice (1971) has more or less exactly the structure that is required for the “reverse” theory. This should not be a surprise. Despite the “no serious attempt” rhetoric, Nozick criticizes Rawls's book on exactly this score, for relegating property entitlements to relatively superficial levels (ASU, 206):

Any attempt to generate anything like property entitlements based on original position premises would lead, says Nozick, to approximations at best: “It is difficult to see how such attempts could derive … the particular convolution of historical entitlement principles” (ASU, 202). This is what Rawls says too. Rawls's remarks on property are very sparse, but they indicate that property rights are relegated to an entirely derivative status in what is otherwise a well-developed theory of human claims in respect of material resources. It is true that Rawls's first principle of equal liberty comprises the right to hold personal property.⁵

But like many such rights, this right to hold property has nothing to do with the existence-conditions for property entitlements. It is simply a right not to be excluded from the list of those who may hold property if there are any property rights to be held (as women and slaves sometimes have been excluded in the past).⁶ Whether there are to be private property rights in significant resources is for Rawls an open question depending on how his principles of justice bear on the circumstances of a particular society. Though “[c]ertain institutional forms are embedded within the conception of justice,”⁷ private property is not one of them. Rawls does work with an example of “a property-owning democracy.”⁸ But this is just an illustration; it “is not intended to prejudge the choice of regime in particular cases.”⁹ A just society might be capitalist or socialist:

The parameters of this assessment are the principles of justice, which include rights of exactly the sort that Nozick classified as general rights to various things (ASU, 238), such as equality of opportunity and the rights defined by the difference principle. Now, the difference principle—the Rawlsian principle that permits social and economic inequalities provided that they are to the benefit of those who are worst off¹¹

—may not specifically define rights to welfare. But it is a principle governing the manner in which the basic structure generates social and economic inequalities, and it insists that the structure must be arranged in a way that maximizes the expectations of members of the worst-off group. As legal scholar Frank Michelman insists—in a fine essay that examines exactly the extent to which Rawls's work makes welfare rights fundamental—the difference principle mandates “an outcome-oriented appraisal of the pattern and makeup of distributive ‘shares’ precipitated by economic, political, and other societal processes.”¹² I do not mean to deny the procedural aspect of Rawls's conception.¹³ But in the illustration that Rawls provides of the operation of such a system, institutions to secure the provision of a social minimum feature prominently.¹⁴ Welfare institutions can operate procedurally too;¹⁵ procedural justice is not associated exclusively with the operations of the market. As Michelman puts it, in Rawls's illustration market procedures (if there are any) must be complemented by procedures designed to secure a social minimum, “which must be provided in order that the residual market determination of distributive shares may be considered just.”¹⁶

So I think that, viewed in this way, Rawls's argument answers exactly to Nozick's specifications for the “reverse” theory. The basic structure of a Rawlsian society is not adjusted to secure the integrity of anything like property entitlements. Instead, it is adjusted to respect the elementary claims that people have in regard to the material wherewithal for basic survival, flourishing (under conditions of equality), and self-respect. Then, given principles of this kind as parameters, we examine what room there is in the circumstances of any actual society for private property rights in goods of various sorts. The basic rights associated with welfare and equality fill up the space of rights first, and then property rights, if there are any, have to contort themselves to fit this Rawlsian substructure.

Did Nozick forget this when he said “no serious attempt has been made to state this ‘reverse’ theory”? I do not know. He certainly regarded Rawls's conception as “serious”:

Nozick thought indeed it was possible to explain why one should not work within Rawls's conception. He said that before we go ahead with a Rawlsian approach, we must be in possession of reasons for turning our backs on the concept of historical entitlement:

On the basis of the passage I cited at the beginning of this essay (ASU, 238), this seems right. Presumably, however, the epistemic burden is symmetrical. We would be ill advised to accept a Nozickian theory of historical entitlement (or to begin fleshing out principles that resonate with our proprietorial and libertarian intuitions) unless we were sure that no adequate theory of the Rawlsian kind—no adequate “reverse” theory—was to be gotten.

This last little exchange—“Don't go down the welfarist road until you are sure no entitlement theory is to be gotten,” and “Don't go down the entitlement road until you are sure no welfarist theory is to be gotten”—suggests that it might be wise to develop a theory of entitlement and a theory of welfare hand in hand. We would try to ground property entitlements to vindicate our proprietorial intuitions, but we would do so in a way that also respected the impulse to assert welfare rights or rights to a social minimum. There would be no assumption that any one body of rights came into the arena first, to throw its weight around and fill up the space of rights, before the other was allowed to enter, to pick up the leavings. We would develop our property/welfare theory holistically.

Whether Nozick would regard this as a way of constructing the “reverse” theory, I do not know. It certainly has the characteristic of not permitting the construction or recognition of property rights to the extent that they conflict with certain imperatives of human welfare and survival. Whatever we call it, we do not need to look very far for instances of this kind of “reverse”/symmetrical approach. The first example I shall provide is the approach to property taken by John Locke. My second example is Robert Nozick himself, in some of the features that his theory shares with Locke's.

Now citing these as examples might seem cheap or preposterous—preposterous because such imputations plainly contradict the tenor and the received understanding of these philosophers' works, and cheap if imputing this approach to Locke and Nozick were nothing but a provocation based on a throwaway line here and there. I am sensitive to these considerations. I shall try to show that the imputations are based on elements central to the arguments of Locke's Two Treatises of Government and Nozick's Anarchy, State, and Utopia, respectively. And I shall try to mitigate the outrageousness of these imputations by showing how they illustrate not only what I am calling the “reverse” position, but also the general importance of integrating claims about welfare rights with claims about property in the overall structure of a theory of justice.

II. John Locke on Charity

No one will deny that John Locke was a believer in property,¹⁹

and his theory of the acquisition of property rights has something like the structure of what Nozick calls a historical entitlement approach. I will not go into the details, for they are very well known. It is sufficient to cite the locus classicus of Locke on justice in acquisition:

This argument is applied to land as well as to objects.²¹

It is supposed to establish entitlements that are resilient against expropriation by the state, that form the basis for a market economy, and that probably generate considerable inequality under modern conditions.²²

However, Locke's account of acquisition is not unconditional. He says that appropriation by labor is permissible “at least where there is enough, and as good left in common for others.”²³

This is commonly referred to as “the sufficiency proviso,” or even just “the Lockean proviso” (and that is what I shall call it).²⁴ The Lockean proviso is plainly relevant to our discussion. However, since Nozick adopts a version of it, I shall postpone my consideration of it until Section III, below, where I consider Nozick's conception of historical entitlement as an instance of the “reverse”/symmetrical theory.

Too much emphasis on the Lockean proviso belies a much more fundamental condition on his theory of property. This is what I shall call “the principle of charity.”²⁵

The principle of charity requires property-owners in every economy to cede control of some of their surplus possessions, so they can be used to satisfy the pressing needs of the very poor when the latter have no way of surviving otherwise. I think the existence and importance of this condition shows that Locke held a version of the “reverse”/symmetrical theory: when considerations of desperate need are present, property rights must give way, because the fundamental definition of property rights is, in the last analysis, organized around the principle of satisfying need.

The principle of charity is not introduced in the Second Treatise chapter on property.²⁶

It is introduced in the course of Locke's argument against Sir Robert Filmer in the First Treatise. There Locke insists that anyone who is in desperate need has “a Title to so much out of another's Plenty, as will keep him from extream want where he has no means to subsist otherwise.”²⁷ It is not an original doctrine: there are versions of it in Aquinas and in most natural law theories.²⁸ If it is taken at face value, it changes the complexion of Locke's theory of property quite significantly, mitigating the “disproportionate and unequal Possession of the Earth,” which a market economy generates,²⁹ by providing in effect a right to a “safety net” for the poorest members of society.

The immediate context of Locke's introduction of the principle of charity is his attack on Filmer's claim that God gave the world and everything in it to Adam and his male heirs, and thus secured for them an entitlement to absolute political authority. Now Locke denied that there was in fact any such specific donation of resources by God to some humans at the expense of others.³⁰

But his strategy in the First Treatise is that of a lawyer—resisting each step in his opponent's argument by showing successively that even if the previous step is conceded, the next step does not follow, and so on.³¹ So he insists that even if there was a donation, it was a donation of property not sovereignty. Property could become absolute power only if the proprietor of the world's resources were able to exploit his wealth to coerce political obedience. Locke dismisses this out of hand: “The most specious thing to be said, is, that he that is Proprietor of the whole World, may deny all the rest of Mankind Food, and so at his pleasure starve them, if they will not acknowledge his Soveraignty and Obey his Will.”³²

What are Locke's reasons for insisting that property rights cannot be exploited in this way? He offers a utilitarian argument,³³

but his main argument is based on elementary natural right:

The language of this passage is interesting. In modern liberal theory, charity is not usually seen as a matter of right: to give of one's wealth to the poor is something one ought to do, but not something one owes to any assignable individual. It is often presented as a paradigm of a moral duty that ought not to be enforced.³⁵

Poor people should be grateful for charity, but they may not demand it or claim it or justify it as rightfully theirs. Some of what Locke says is suggestive of this tradition. He distinguishes the demands of charity from those of justice, drawing at least at a verbal level the distinction that modern liberals have made so much of. But, in order for his argument against Filmer to succeed, Locke needs to be able to show not only that it is wrong to withhold charitable assistance, but that the rich man has no right to do so, no right even to offer assistance subject to conditions such as political submission. For the most part, then, the Lockean language is emphatically stronger: though charity is contrasted with justice, still charity “cannot justly be denyed.” If it cannot justly be denied, then anyone who denies it cannot claim to be exercising his property rights. The sin of uncharitableness simply vitiates the exercise of the rights in question. A man in need has “a Right” to another's surplus goods: indeed, Locke talks of his having “a Title,” in a way that suggests that this too is to be regarded as a property entitlement.

What about the fact that charity is regarded by Locke as a specifically Christian virtue? Shouldn't this be a reason for not enforcing it in a liberal theory? Apparently not, according to Locke. He does make a point of saying that many aspects of religion cannot be enforced by the state.³⁶

The clearest example is Christ's commandment to believe in him as the Messiah,³⁷ which, as Locke argues in A Letter Concerning Toleration (1689), cannot and should not be enforced. For that case, however, we have specific reasons for nonenforcement: (1) belief is not subject to the will, and coercion works only through the will; and (2) what anyone believes (e.g., about Jesus being the Messiah) does not affect the well-being of anyone else.³⁸ Neither condition applies to the principle of charity. Indeed, it is noteworthy that when Locke writes in The Reasonableness of Christianity (1695) about the difference between what Christ requires us to believe (where uncoerced sincerity is of the essence) and what he requires us simply to do, it is charity Locke mentions as an illustration of the latter.³⁹ The matter is complicated by one passage in The Reasonableness of Christianity which suggests that charity may be a virtue that state power could be used to uphold but should not: Jesus' commandment to the rich young man to “sell all you have and give to the poor.”⁴⁰ But Locke reads this as Christ's test to see whether the young man in question really would be willing to follow his commandments, not as an imperative addressed to all of us:⁴¹

Selling all you have and giving it to the poor would be a form of what I want to call radical charity—giving away what you actually have a moral right to keep, giving away enough to impoverish yourself—and that is not what Locke is arguing for in the First Treatise. Radical charity may be a particular requirement imposed on particular people, but it is not intended as a general command. Radical charity is not required by the principle we are discussing. The principle of charity from the First Treatise affects the rich man only in “the Surplusage of his Goods.”⁴³

The doctrine of charity is set out in Locke's First Treatise, but it is not mentioned explicitly in the Second Treatise.⁴⁴

We know that the two treatises were written at different times.⁴⁵ Is there any reason to believe that the author of the famous chapter on property in the Second Treatise (chapter 5) intended the theory he set out there to be qualified by the doctrine I have been discussing?

It is tempting for modern theorists who claim Lockean inspiration in their thinking about property but oppose the principle of charity to insist on a separation between the Second Treatise doctrine and the First Treatise doctrine in this regard. The First Treatise doctrine—which, contrary to what its title may suggest, was produced later by Locke—they might treat as an aberration, unwarranted by the premises of his original theory.⁴⁶

In general, I think this approach to textual exegesis is quite unsound, motivated as it is more or less entirely by a desire to find a theory of a certain shape in Locke and by an irritation at having to consider aspects of his thought on these matters that indicate a more complex view. For there is nothing actually in the text of the Second Treatise that precludes the principle of charity. The textual evidence for the independence of the two treatises in regard to charity is just Locke's silence on the issue in the Second Treatise chapter on property: there are several places where one would expect him to mention such a proviso (if he believed in it), and he does not, at least not in so many words. But this is quite inconclusive. We do not take Locke's silence on inheritance and bequest in the Second Treatise chapter on property to indicate a belief that property rights were not inheritable or subject to passage by bequest; instead, we see that the very extensive discussion of inheritance and bequest in the First Treatise naturally adds to and complements the discussion in the Second Treatise.⁴⁷ The fact is that there was no particular need to discuss inheritance and bequest in chapter five of the Second Treatise, but argumentatively it is required for Locke's account of heritable political authority in chapter 9 of the First Treatise. And similarly, there was no particular need to discuss the principle of charity in chapter five of the Second Treatise, though it could be deployed for political advantage in the attack on absolutism in chapter 4 of the First Treatise. In any other context, the fact that different doctrines are mentioned in different chapters would not be adduced as grounds for separating them from one another. It is only the fact that the principle of charity is seen as an annoyance for a certain sort of “Lockean” theory that requires me to spend time defending the claim that two consistent sets of propositions written by the same author and published in different parts of the same book should be regarded as parts of the same theory (unless very good reason is shown why they should not be).

I used to think there was a substantial difficulty in accommodating the charity doctrine with the Second Treatise theory of property. The basis on which Locke argues for individual acquisition is the idea of mixing one's labor with resources or land: “joyn[ing] to it something that is his own, and thereby [making] it his Property.”⁴⁸

On a literal interpretation of this argument, the laborer transfers to the object the inviolable right he has in his own person. Now, it is hard to see how that right can be overridden by the mere fact of another's need. Either the object contains the owner's labor or it does not: the other person's need cannot, as it were, drive the labor out. I used to think that what was going on here was that Locke was running two incompatible lines of argument to justify private property in the Second Treatise: one was entitlement-based, deriving from the exclusive right one has in one's person and one's labor,⁴⁹ and the other was based broadly on the notion of human need and what must happen in order for mankind to prosper.⁵⁰ The provisos and limitations seemed to make sense primarily in relation to the second of these lines of argument, and the same would be true of the principle of charity.

But I no longer think it makes sense to see the mixing of labor as a separate line of argument, less hospitable to modification by charity or the other provisos. I think the labor theory works only against a background of labor's significance in God's overall plan for the survival of the human beings he has created. Laboring is not just something we happen to do to resources, nor is being permitted to help oneself by labor just a divine indulgence of the self-interested inclinations of an acquisitive being. It is the naturally requisite next step following our creation, once we accept that we were created subservient to God's design “that Man should live and abide for sometime upon the Face of the Earth, and not that so curious and wonderful a piece of Workmanship by its own Negligence, or want of Necessaries, should perish again, presently after a few moments continuance.”⁵¹

If labor were not presented against this background, if it were just something I happen to own, there would be no answer to an embarrassing question that Nozick poses: “[W]hy isn't mixing what I own with what I don't own a way of losing what I own rather than a way of gaining what I don't?” (ASU, 174–75). Any account of the significance of labor rich enough to answer this question will have to hook up with the account based on need, and therefore will naturally be subject to the principle of charity.

Bearing this in mind, I think we can say that the substantive case for importing the First Treatise doctrine of charity into the Second Treatise theory of property revolves around the premises of Lockean natural law. The basic principle of the law of nature, said Locke, is the preservation of as many people as possible: “Everyone as he is bound to preserve himself, … so by the like reason when his own Preservation comes not in competition, ought he, as much as he can, to preserve the rest of Mankind.”⁵²

This sounds like a positive duty to do whatever will promote the preservation of as many people as possible. Locke makes it clear that it goes far beyond mere nonaggression. It includes the power to pardon offenders; it even includes a duty to infringe property rights in an emergency, like “pull[ing] down an innocent Man's House to stop the Fire, when the next to it is burning….“⁵³ Anyway, it has never been clear to me that the duty of charity—at least as Locke presents it—is a positive as opposed to a negative duty. Paragraph 42 of the First Treatise denies that property owners have the right to withhold their surplus goods from the poor. Thus, Locke seems to be committed not to a view about giving but to the view that neither the rich person nor civil society on his behalf is entitled to resist the poor when the poor attempt to seize his surplus goods. It is not a question of forcing the rich to do anything: it is enough that they be compelled simply to stand back and let the poor take what (on account of their “pressing Needs”) is rightfully theirs.⁵⁴

More specifically, at the beginning of the chapter on property, Locke tells us in very general terms that “[m]en, being once born, have a right to their Preservation, and consequently to Meat and Drink, and such other things, as Nature affords for their subsistence….”⁵⁵

Now, this might be construed as mere rhetoric, except that Locke quickly puts it in play to do some rather important work. For when he defends the legitimacy of unilateral appropriation, he raises the specter of starvation as a basis for his rebuttal of any consent requirement:

The prospect of starvation simply short-circuits any complaint based on the illegitimacy of appropriation without consent. Since the rights that are rebutted in this way are real rights—each person's right to partake in God's original donation—Locke is saying, in effect, that the right to withhold consent can be trumped by desperate need. He cannot consistently maintain that without also maintaining what I have called the principle of charity.

One last hurdle remains. Those who want to portray an uncharitable Locke are fond of telling us about the John Locke who was the author of an essay on the poor law—a draft of a representation concerning methods of employing the poor. That draft certainly seems devoid of the sort of compassion that one would associate with a generous spirit of charity. Locke talks about “begging drones” and “superfluous brandy shops,” and he suggests that the idle poor should be whipped and mutilated if they go begging instead of doing the work assigned to them. Even little children should be given two or three hours of labor useful to the parish per day.⁵⁷

I want to say two things about this. First, we should bear in mind that the paper on the poor law does assume that the poor have a right to subsistence. “[E]veryone must have meat, drink, clothing, and firing.”⁵⁸

Secondly, Locke took very seriously the injunction to labor, and a poor man with an offer of gainful employment was not a person who had no means to subsist (within the meaning of the principle of charity). We have already seen that Locke is in favor of charity being enforced, but not radical charity.⁵⁹ Well, now we have a second form of radical charity that Locke is not prepared to enforce: this would be charity to someone who refuses to work when he could work, charity to someone who does not make any efforts to provide for himself or work to subsidize the cost of his provision.

I hope this textual and historical account of Locke's principle of charity has not been tedious. What I want to establish is that Locke believed very strongly that human survival and flourishing (and indeed the increase of a surviving and flourishing population) were the basic moral imperatives of economic life. In Section I of this essay, I discussed which moral considerations come on stage first and establish themselves as the substructure, which other rights and principles must then work around. Well, in Locke's theory, the substructure comprises the principle of preservation and increase. No rights exist in conflict with this substructure, and the whole argument in favor of property entitlement and acquisition is predicated on these first principles:

If this is the foundation of property, then property cannot be used to oppose other consequences that flow from these first principles. Elementary rights to material assistance in conditions of urgent need flow directly, and they must condition the property rights, not the other way round.

III. Nozick on Welfare

Now I want to return to the Lockean proviso and the work that it does in Robert Nozick's theory. Nozick is committed to the following proposition:

What is this, if it is not a statement of what Nozick called the “reverse” theory (ASU, 238)? It gives priority to the right not to have one's material situation worsened, whether that situation consists in holding property rights or just in having access of some kind to the resources needed for a decent life. It gives these rights priority in exactly the sense that the “reverse” theory is supposed to give priority to welfare rights: property entitlements must work around them; no property entitlements are recognized if they are incompatible with these rights.

One can quibble with this—does the Lockean proviso really generate rights?—and I will explore the quibbles in a moment. For now, just observe how strongly this condition operates in Nozick's theory. It applies to acquisition: if I want to be the first to take a given resource into possession as a private holding, I may do so (by engaging in whatever the principle of justice in acquisition says: labor, occupancy, whatever), but only so long as the situation of others is not thereby worsened. The condition also applies to the exercise of one's property rights in whatever one has appropriated: “a person may not appropriate the only water hole in the desert and charge what he will” (ASU, 180), for previously the situation was that other people could come to the water hole and drink for free. And the condition renders one's property rights and their exercise vulnerable to revision in the light of changing circumstances: “[n]or may he charge what he will if he possesses one [water hole], and unfortunately it happens that all the water holes in the desert dry up, except for his” (ASU, 180).⁶²

It also affects justice in transfer:

Thus, the proviso affects property entitlements generally, even those at a great distance from the original acquisitions that got historical entitlement underway: “Each owner's title to his holding includes the historical shadow of the Lockean proviso on appropriation” (ASU, 180).

Despite this “shadow,” Nozick believes that the Lockean proviso can be satisfied fairly easily. He argues against any interpretation of it that would limit acquisition to cases where there is no impact on other people's acquisitive opportunities (so that there is enough, and as good, left in common for others to appropriate).⁶³

If people are no worse off overall as a result of others' acquisitions, they have no ground for legitimate complaint (ASU, 176). This means we can appeal to “the various familiar social considerations favoring private property” on grounds of efficiency (ASU, 177)⁶⁴—considerations that leave him confident “that the free operation of a market system will not actually run afoul of the Lockean proviso” (ASU, 182). I shall not quibble with this, though it is worth noting the honesty of Nozick's observation that the claim is an empirical matter and he may be wrong (ASU, 182). As we shall see, Nozick's objections to actually existing market economies have more to do with the brutal injustice associated with their establishment and with the particular property rights held by the rich than with any violation of the Lockean proviso (ASU, 230–31).

Does the ease with which the Lockean proviso can be satisfied in a market society based on property entitlements mean that we cannot use Nozick's embrace of the proviso as a ground for attributing to him what we have called the “reverse” theory? Not at all. A principle may be fundamental and it may define the substructure of rights, in Nozick's sense (ASU, 238), without being very demanding. All that is necessary to sustain the proposition that Nozick's theory is a “reverse” theory is to show that it is not the case that property rights fill up the space of rights leaving no room for rights such as a right not to have one's material condition worsened. The Lockean proviso shows this. Does it make any difference that the Lockean proviso is part and parcel of Nozick's conception of historical entitlement so that property rights (subject, admittedly, to that proviso) fill up the space of rights leaving no room for any other right to be in a given material condition? No, that is just being silly. At most it shows that Nozick is adopting what I referred to at the end of Section I as the “symmetrical” approach—developing his theory of historical entitlement hand in hand with a theory of people's rights to be in a certain material condition.

But does Nozick's version of the Lockean proviso actually enshrine rights to be in a certain material condition? There is certainly a difference between the proviso and Locke's principle of charity. The latter establishes that each person has a right to survive, and hence “a Title to so much out of another's Plenty, as will keep him from extream want, where he has no means to subsist otherwise….”⁶⁵

Nozick's version of the proviso respects only a claim not to be shifted from a certain baseline of well-being. It is a baseline defined (rather problematically—ASU, 177) as the condition one would be in but for the appropriation(s) or holding(s) whose legitimacy is now called into question. Moreover, it is not a claim to be and to remain at that level of well-being (however it is defined): it is only a claim not to be moved down from it by an appropriation (or the equivalent of an appropriation). If one's well-being declines for other reasons, the Lockean proviso does not generate a claim to have it restored. These are important points. But they do not affect the underlying argument about the shape of Nozick's approach. His theory begins with the assumption that the formation and justification of property entitlements must be responsive to concerns about well-being that are not embodied in property rights. In this sense, it is not true that on Nozick's account only property rights define the “substructure” of rights.

The Lockean proviso is not an incidental accretion to Nozick's conception that he could easily abandon. Without such a proviso, any principle of justice in acquisition (and anything founded on it, such as a system of ongoing property rights in a market economy) becomes morally very problematic, since it involves the unilateral imposition of duties (albeit negative duties) on people without their consent.⁶⁶

Few are willing to say that the imposition of such duties can be justified utterly without reference to the interests or needs of those whom they exclude from access to resources.⁶⁷

That is very general. More specifically in Nozick's case, we see the importance of the Lockean proviso in assimilating the justification of initial acquisition to the kind of justification used for principles of justice in transfer. In the latter case, we justify transfers on grounds related to their representing Pareto improvements: because the transactions are voluntary, no one thinks that he loses in a market transfer. Therefore, so long as the parties are rational, a market transfer (barring externalities) makes at least some better off without making anyone worse off. In initial acquisition, by contrast, it looks as though there will be losers: I acquire this field as my property; and since you are excluded, you no longer have the access that you had before. The Lockean proviso ameliorates this problem. Now we say that an acquisition is legitimate only if it leaves no one worse off, that is, only if it is a Pareto improvement. Acquisition is thus assimilated to transfer so far as the principle of its justification is concerned.

By the way, it is interesting to note how Nozick's use of the idea of Pareto improvement differs from that of scholars involved in the Economic Analysis of Law (EAL).⁶⁸

When a change is justified as an improvement according to EAL, it is because it is thought that although some people benefit and some suffer from the change, the gains of those who benefit are sufficient to compensate the losses of those who suffer. This is the famous Kaldor-Hicks criterion of social improvement: a change that improves the situation of some by worsening the situation of others, is permitted if the winners win more than enough to compensate the losers.⁶⁹ For EAL, this gives the change in question the respectability of a Pareto improvement. But although a Kaldor-Hicks improvement is like a Pareto improvement in certain respects (e.g., in the amount of wealth-gain required), it is not relevantly like a Pareto improvement in the respects that ought to matter when we think about justice. True, in a Kaldor-Hicks improvement, we ensure that the winner gains enough to compensate the loser, so that no one need be worse off. But this assessment is not associated, in any way, with a practical determination actually to assign the loser any portion of what the winner gains. The preferences of the losing party are certainly taken into account in considering whether the change is an improvement: it must be the case that the loser would prefer a portion of the winner's gain to what he could secure in the absence of the change. But justice is not interested in facts about preferences per se. Justice is interested in facts about the satisfaction of preferences, and with a Kaldor-Hicks change there need be no actual satisfaction of this preference of the loser. Similarly, the legal imposition of a Kaldor-Hicks improvement has none of the consensual respectability of a Pareto improvement. When we impose a Kaldor-Hicks improvement, we are not in any way honoring the voluntary consent of the losing party. That the loser would consent (under condition C) adds nothing to the respectability of a change that involves no actual consent (because condition C is not satisfied).⁷⁰ To oppose a Kaldor-Hicks improvement in the name of justice is therefore not to oppose justice to human welfare, nor is it to oppose justice to people's revealed preferences, nor is it to oppose justice to a scheme that makes everyone better off. It is, rather, to insist on the importance of respecting actual individuals with their actual preferences in the actual world; and it is to oppose the imposition on individuals of actual losses for which nothing but hypothetical compensation is envisaged. Now Nozick will have none of this. He famously rejected the utilitarian idea that some people's gains can make up for other people's losses: what happens is that some people gain and others lose; nothing more (ASU, 33). And he is insistent on this point specifically in connection with the Lockean proviso:

This is a striking difference between Nozick and the EAL: Nozick's account has integrity, theirs does not.

This is but one example of the integrity of Nozick's argumentation in Anarchy, State, and Utopia. When the book appeared, left-liberals were so shocked that many of them complained the book was written simply to promote a political agenda. Brian Barry's review was typical and a now-famous passage of it ran as follows:

I think this sort of abuse is quite unwarranted. If Nozick were standing for office and these were his policies, I guess there would be reason to vote for some other candidate. In fact, he was not in a position to starve or humiliate anyone or to eliminate any transfer payments through the state, nor was he was aspiring to any office where his “proposals” to this effect would be a matter of political concern. He was, as Barry reminds us, occupying a professorial chair in a philosophy department. He was not in fact proposing anything, or calling for anything, or indeed urging anything, except this: that we might try thinking about justice in a slightly different way. As Nozick pointed out in his preface to Anarchy, State, and Utopia:

Nozick's integrity probably explains why his work has had more influence on the Left than on the Right. For those of us on the Left, it was Nozick (rather than, say, F. A. Hayek) who undermined our confidence in simple formulae of equality and who convinced us that “no end-state principle or distributional patterned principle of justice can be continuously realized without continuous interference with people's lives” (ASU, 163). Nozick laid down the challenge that we thought we needed to answer. (I am not saying we succeeded in answering it, but it was Nozick's challenge that sharpened our wits.) However, Nozick argued too honorably to be of much use to the triumphant Right in the 1980s and 1990s. He was never prepared to say that the historical-entitlement critique of equality and welfarism in his book amounted to a defense of actually existing market institutions, nor would he pretend that a Lockean defense of property could go any distance toward legitimizing contemporary disparities of wealth in (for example) the United States. On the contrary, he thought it undeniable that contemporary holdings would be condemned as unjust by any remotely plausible conception of historical entitlement. (The point of Nozick's argument in chapters 7 and 8 of Anarchy, State, and Utopia was that egalitarians were condemning the existing distribution for the wrong reason—that is, simply as unequal—rather than on account of the violence, fraud, expropriation, ethnic cleansing, state corruption, and so on, involved in the history of most holdings of property in America.) Once actual historical injustice was established, then the burden fell on the part of his conception dealing with the rectification of injustice. And for Nozick, it was an open question whether the actual operation of (say) Rawls's difference principle might approximate the operation of a plausible process for rectifying historical injustice. That, Nozick thought (quite rightly), would involve addressing some very difficult questions about time, counterfactuals, second-best principles, and so on:

This was not what defenders of free markets and opponents of welfare wanted to hear in the 1980s and 1990s.

IV. Welfare and Justice

It is often said of welfare rights that they are impracticable or too expensive or too demanding. Some critics even argue that socioeconomic rights violate the logical principle “Ought implies can.” Many nations, they say, do not have the resources to provide even minimal economic security for masses of their citizens, and since states differ considerably in this regard, it hardly makes sense to regard economic provision as a matter of universal human entitlement.⁷²

In some cases, sheer scarcity will trigger the principle “Ought implies can.” But the fact that this happens in some cases does not undermine the universalist thesis. The “ought” remains in the background, in all cases, commanding each country to do as much as it can in this regard, and its full force will be triggered for a given country whenever it acquires the ability to fulfill it. Moreover, in many cases, the alleged impossibility stems not from brute scarcity as such but from an assumption that the existing distribution of property is to remain largely undisturbed. When a conservative government in the West says in response to some plea for welfare provision, “The money simply isn't there,” what is usually meant is that it would be either impolitic or improper to try and raise it from existing property-holders and income earners by taxation. On the one hand, if the objection is simply that raising the money would be impolitic, then really the “ought” of welfare rights is being frustrated, not by the “can't” of impracticability, but by the “shan't” of selfishness and political ambition. If, on the other hand, the objection is that the money is not the government's to distribute, then we are back where we began, with our question about the relative priority to be accorded to property rights as opposed to welfare rights. Either way, resting the matter simply on “Ought implies can” underestimates the challenge posed by these rights to the underlying distribution of wealth and income.

Still, someone may press the question: Aren't these socioeconomic rights awfully demanding? They may say: The first generation of rights claims—civil and political rights—require only that we and our governments refrain from various acts of tyranny and oppression. They are “negative” rights correlative to duties of omission, whereas so-called second-generation rights—socioeconomic rights—are correlative to positive duties of assistance. An advantage of negative rights is that they never conflict with one another, for one can perform an infinite number of omissions at any given moment. With positive rights, by contrast, we inevitably have to consider the inherent scarcity of the resources and the services that are called for.

There are many things wrong with this argument. For one thing, the correlation of first- and second-generation rights with the distinction between negative and positive rights simply will not stand up. Many first-generation rights require the positive establishment and maintenance of certain frameworks, and all of them make costly claims upon scarce police and forensic resources. Even those first-generation rights that can be conceived in the first instance in terms of negative freedom—such as free speech and freedom of worship—still require more from a government than that it simply stay its hand. We set up a government to protect our rights, not simply to respect them. All serious political theorists recognize that protection is costly,⁷³

and that therefore first-generation rights as much as socioeconomic entitlements raise hard questions about priorities for us all.

It follows that where resources are scarce relative to human wants, any system of rights or entitlements will seem demanding to those who are constrained by it. If an economic system includes provision for welfare assistance, it may seem overly demanding to taxpayers. But if it does not include such provision, then the system of property rights in such an economy will seem overly demanding to the poor, requiring as it does that they refrain from making use of various resources that they may need. As usual, the question is not whether we are to have a system of demanding rights, but how the costs of the demands are to be distributed. That, of course, is why the question that Nozick raised in the passage cited at the outset of this essay is so important.⁷⁴

Thinking about scarcity and practicability does have one important advantage: it forces proponents of second-generation rights to take seriously issues of distributive justice. It is an unhappy feature of the language of rights that it expresses demanding moral claims in a sort of “line-item” way, presenting each individual's case peremptorily, as though it brooked no denial, no balancing, no compromise. This feature of rights has always troubled those who are sensitive to the fact that individuals live in a social environment where the things that they may reasonably expect must be adjusted constantly to reflect similar claims and expectations on the part of others. If we want a framework that more adequately reflects the necessarily compromised character of rights claims (civil rights, property rights, welfare rights), we must turn instead to the modern discussion of justice. The theory of justice—not Rawls's theory necessarily, but theorizing about justice—is the proper matrix for reconciling conflicting claims of right. Modern theorists have not written nearly enough about the relation between liberal theories of rights and liberal theories of justice. Let me end with a gesture toward such an account.

Familiar claims of right will figure both as inputs and as outputs in our theorizing about justice. They function as inputs inasmuch as they help constitute our initial sense of the individual claims that need to be considered. That people are normally taken to have the right of freedom of worship, the right to participate in politics, and the right to work, helps us frame the problems of justice that a society of such persons gives rise to. We expect that many familiar claims of right will also figure among the outputs of an adequate theory of justice. We hope the institutional structure that is set up in this way is one that will be able to offer certain guarantees, though they may not be as strong or as unqualified as those we had in mind before we remembered the need to balance our own claims against those made by others. For example, we may go in with a sense that respecting a human being means respecting his freedom of worship, but come out (of the balancing process) with a religious guarantee that is perhaps somewhat modified by a recognition that children, for example, must be given an adequate education and not left at the mercy of their parents' convictions, and that certain aspects of religious practice must on occasion give way to independently grounded considerations of public administration.

The same is true in regard to socioeconomic claims. We go into the discussion of justice with the sense that people are not just the disembodied wraiths of liberal ideology, but needy individuals subject to vicissitudes of embodiment and materiality: we must have food and shelter, we must work, we get sick, we grow old, we are often dependent, and so on. A preliminary sense of the importance of all this—I mean its importance to each of us as individuals, not just its importance to society—is conveyed by provisions such as the second-generation rights we have been discussing.

However, things are less clear at the output end. Consider Rawls's theory, again, as an example. Rawls is anxious that the task of a theory of justice not be understood simply as an allocation of distributive shares: who gets what, when, and how? “We must not assume,” he says, “that there is much similarity from the standpoint of justice between an administrative allotment of goods to specific persons and the appropriate design of society.”⁷⁵

Thus, for example, the “difference principle” is not to be interpreted as dictating that the worst-off group be given a certain share of resources. The effect of the principle is that when we are designing or (more likely) evaluating and reforming the network of rules and procedures that constitute the institutional structure of society, we should do so in a way that is oriented toward the advantage of the worst-off group. The institutions should be designed to operate on the assumption that when the system is working, outcomes are evaluated purely procedurally. We are not to meddle with the outcomes of a just institutional structure even if we think that by doing so we could make the array of outcomes even more just from a distributive point of view.

Now, here is a curiosity. F. A. Hayek has taken the comments that I have just cited as grounds for claiming that the difference between Rawls's approach and his own approach is “more verbal than substantial.”⁷⁶

Rawls's work, he says, has been “wrongly … interpreted as lending support to socialist demands.”⁷⁷ Hayek takes Rawls to agree with his own central claim that “[j]ustice is not concerned with those unintended consequences of a spontaneous order [such as a market] which have not been deliberately brought about by anybody.”⁷⁸

I think that Hayek is mistaken about this and that he exaggerates the implications of Rawls's refusal to consider the justice of particular allocations of goods. Rawls's position may be illuminated as follows. Suppose that, on one occasion, the institutions of our economy happen to yield a distribution of wealth, D₁, that is judged inferior in terms of the difference principle to another distribution, D₂. Should we immediately interfere and reallocate wealth so that we change D₁ into D₂? Rawls's answer, like Hayek's, is no. For Hayek the matter ends there. But for Rawls there is a further question to be addressed: Can we change the institutional structure (including the structure of property entitlements, if there are any) so as to render it more likely in the future that the normal operation of our economy will yield distributions like D₂ rather than D₁? The answer to this question may also be no, because the proposed change might be incompatible with institutional virtues such as publicity, stability, and the rule of law.⁷⁹

The new institutions that are being suggested may not, as it were, hold up, as institutions. Still—and this is what Hayek overlooks—the answer is not necessarily no. If change is possible and if the resulting institutional structure would be viable, stable, and so on, then we are required as a matter of justice to implement it, for the difference principle just is the requirement that we arrange (and, if necessary, rearrange) our institutions so that social and economic inequalities are to the greatest benefit of the least advantaged. So, for example, if we find ourselves in a market society that lacks basic welfare provision, we have to consider whether the institutional structure of a market economy would be wrecked qua institutional structure by the addition of what Rawls calls a “transfer branch,” charged with administering a social minimum. Would that make it impossible for the economic structure as a whole to operate predictably, publicly, impersonally, and in accordance with other institutional virtues? Hayek has devoted a large part of his life to arguing that it would—that the modern regulated welfare state is incompatible with the rule of law.⁸⁰ It is pretty clear that Rawls disagrees with him about that. But the deeper disagreement is that Rawls thinks it is the job of a theory of justice to select principles for evaluating economic institutions along exactly these lines, whereas Hayek simply denies that this is a legitimate concern about justice.

I have taken this digression into the Hayek-Rawls misunderstanding because I want to stress that in a theory of justice such as Rawls's we cannot guarantee that socioeconomic rights will emerge in a familiar or predictable form. As an abstract matter we can say, with the drafters of Article 25 of the Universal Declaration of Human Rights, that everyone has “the right to a standard of living adequate for the health and well-being of himself and his family.” But this may not necessarily emerge as a specific legal or constitutional guarantee: a just society may not have a rule to this effect, nor even any particular agency charged with administering this standard. There may be a variety of provisions and arrangements—tax breaks, educational opportunities, rent control (or its abolition), unemployment insurance schemes, etc.—all of which taken together may represent the best that can be done in an institutional framework to honor the underlying claim for the individuals in whose behalf it can be made.

This fact may upset welfarists who think it more important that a right be proclaimed than that it be secured. It may upset those who are uninterested in the viability of social institutions and in rule-of-law values. But people who strike these attitudes should reflect a little on the role that rights are supposed to play in our social and political evaluations. When we think about rights, we must do our thinking in political philosophy not just moral philosophy, and this means being willing to consider the embodiment of various moral concerns in the institutional arrangements of human society. Once we see this as the distinctive perspective of political philosophy, we begin to see the inadequacy of the approach that leaves rights as simply programmatic claims, without considering how they fit together as a system and how the rights of one person might be made compatible (concretely, not just in theory) with the rights of another.

The approach I have suggested may upset those who want to maintain at all costs that each individual right is absolute, and who in consequence refuse to face issues of scarcity, conflict, and moral priority. (Some say this about property and some may say it about welfare rights.) This insistence is perhaps understandable: the modern preoccupation with rights is partly a response to the trade-offs that characterize utilitarian calculations; rights were supposed to express limits on what could be done to individuals for the sake of the greater benefit of others; but if rights themselves involve conflicts and balancing, it looks as though there is no getting away from the casuistry and complex moral calculations that were thought to be the hallmark of more blatantly consequentialist theories. I have some sympathy with this view. However, the insistence on absolutism does not make the conflicts go away; it does not make the situations that appear to call for trade-offs disappear. Those situations are not something that consequentialists and their fellow travelers have perversely invented in order to embarrass moral absolutists. There is a danger that, if we insist on the absoluteness of rights, we may end up with no rights at all—or no rights that embody the idea of real concern for the individuals whose rights they are.

The proper response to situations that pose conflicts and dilemmas is not to deny that they exist or to protect the term “rights” from being contaminated by them. It is to face up to the challenges they pose with the clearest sense we can get of the intrinsic worth of each individual's well-being and independence in a world that is also inhabited by others. Unless that sense becomes hopelessly attenuated, it is exactly what we should identify as our theory of rights. Rights are not to be thought of as moral absolutes, standing tragically unemployed, on the sidelines of a world riven by distasteful conflict and hard choices. On the contrary, they are supposed to be our best, most honest, and most respectful response to such a world.

Let me return, finally, to Robert Nozick. With his emphasis on rights as side constraints (ASU, 26 ff.), it is tempting to associate Nozick with the sort of absolutism I have just been attacking. And it is tempting to read his critical comments about welfare rights (ASU, 238) as an indirect consequence of this absolutism. I do not just mean that his absolutism about property leads him to reject welfare rights; I mean that he might be thought to reject welfare rights partly because of what they would have to amount to qua rights. In fact, I think Nozick has contributed immeasurably to our ability to think these things through carefully in a manner that is appropriate to justice and not merely to line-item claims of right. I hope I have been able to show in this essay that this is true of his discussion about “filling the space of rights” in Anarchy, State, and Utopia (238), and true also of his discussion of the Lockean proviso.