Hostname: page-component-74656f9884-npj9w Total loading time: 0 Render date: 2026-11-09T23:39:59.032Z Has data issue: false hasContentIssue false
  • English
  • Français

THE SHAPE OF LOCKEAN RIGHTS: FAIRNESS, PARETO, MODERATION, AND CONSENT

Published online by Cambridge University Press:  01 January 2005

Richard J. Arneson
Richard J. Arneson
Affiliation:
Philosophy, University of California at San Diego
Rights & Permissions [Opens in a new window]

Abstract

In chapter four of Anarchy, State, and Utopia, Robert Nozick raised interesting questions about whether or not it is ever morally acceptable to act against what are agreed to be an individual's natural moral rights. The pursuit of these questions opens up issues concerning the specific content of these individual rights. This essay explores Nozick's questions by posing examples and using our considered responses to them to specify the shape of individual rights. The exploration provisionally concludes that a conception of individual moral rights quite different from Nozick's looks attractive and merits further development.

Information

Type
Research Article
Information
Social Philosophy and Policy , Volume 22 , Issue 1 , January 2005 , pp. 255 - 285
Copyright
© 2005 Social Philosophy and Policy Foundation

I. Introduction

The Lockean natural rights tradition—including its libertarian branch—is a work in progress. Thirty years after the publication of Anarchy, State, and Utopia (1974), Robert Nozick's classic work of political theory is still regarded by academic philosophers as the authoritative statement of right-wing libertarian Lockeanism in the Ayn Rand mold.1

Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974). Further references to Nozick's book will take the form of parenthetical page citations in the text, using the abbreviation ASU. Nozick criticized arguments purporting to provide a moral foundation for capitalism that he located in Ayn Rand's writings. See Nozick, “On the Randian Argument,” reprinted in Robert Nozick, Socratic Puzzles (Cambridge, MA: Harvard University Press, 1997), 249–64. This essay was first published in 1971.

Despite the classic status of this great book, its tone is not at all magisterial, but improvisational, quirky, tentative, and exploratory. Its author has more questions than answers. On some central foundational questions, he refrains from taking a stand. There is spadework yet to be done on the project of developing the most plausible versions of Lockean and Lockean libertarian views. Prior to doing this work and articulating the sensible alternatives and what can be said for and against them, we are not yet in a position reasonably to opt for any particular version of Lockean theory, or for that matter to decide between the natural rights tradition and rival consequentialisms. This essay aims to explore hard and soft versions of Lockean theory. The exploration aims to persuade the reader to favor the soft versions.

Section II formulates four claims (all asserted by Nozick) and provisionally identifies the Lockean libertarian view with these claims. Section III notes that although Nozick in his 1974 book made scant progress toward providing a justification of his particular doctrine of rights, compared to advocates of rival doctrines, no rights theorist since then has made significant advances on that front, so Nozick's achievement has not been superseded. Section III also rehearses Nozick's view of rights as side constraints. Sections IV and V raise a question that Nozick first posed: Should rights be regarded as specifying ways individuals may not be treated, infringement of which is never, or sometimes, or always morally acceptable, provided full compensation is paid to any victims? Hard libertarianism is defined as a version of Lockean libertarianism that replies “Never!” to this question along with offering strict interpretation and uncompromising affirmation of the four provisional claims detailed in Section II. Sections VI through VIII explore and criticize Nozick's own discussion of the question under review. An alternate phrasing of it is: Is every infringement of an individual natural right a violation of it? The discussion proceeds by raising several examples and interpreting them as counterexamples to hard libertarianism. Sections IX and X discuss examples that provide reasons to embrace a weak or soft interpretation of the Lockean norm of self-ownership. Sections XI and XII introduce further softening. Section XI proposes a Pareto constraint on the content of individual rights. Section XII proposes moderation, the idea that any natural moral right of any person should give way if the consequences for other people if one does not violate the right are sufficiently bad. Moderation says rights are side constraints that give way under pressure of consequences. In other words, people are inviolable, up to a point. Section XIII argues that the soft Lockean position at which we have arrived is not a repudiation but, rather, an intelligible development of the Lockean tradition, and in particular does not reject but only qualifies the claim of self-ownership. Section XIV summarizes the modifications to Nozick's version of libertarianism that this essay defends. Section XV notes that the considerations adduced in favor of soft as opposed to hard versions of Lockeanism consist entirely of descriptions of examples. But Nozick himself has rightly stressed that the evidence for a moral theory that consists in responses to puzzle cases is not to be sneered at.

II. Basics

The fundamental Lockean libertarian view comprises four claims:

  1. Each person has a moral right to do whatever she chooses with whatever she legitimately owns unless her actions would harm nonconsenting others in certain ways that violate their rights.
  2. Each person has the right not to be harmed by others, whether by physical assault, interference with her liberty by coercion or force, physical damage to her person or property, extortion, theft or fraud, breach of contract, libel, or threat of any of these things.
  3. Each adult person legitimately owns herself.
  4. All of these moral rights are forfeitable by misconduct, transferable from their holder to another by mutual consent, and waivable by voluntary consent of their holder.2

    Not all Lockean libertarians accept the claim that one's right to self-ownership is fully transferable and waivable. This claim implies that voluntary slavery contracts may be morally valid. Nozick takes this line, however, and it has the virtue of simplicity, so I include it in the statement of core Lockean libertarianism. See also John Simmons, “Inalienable Rights and Locke's Treatises,” Philosophy and Public Affairs 12, no. 3 (1983): 175–204.

An important derivative element in Lockean theory is that from the premises above, given a world in which material resources are initially unowned, it follows that individuals can acquire extensive private ownership rights over material resources.3

Some readers may balk at my terminology here, on the ground that John Locke's initial premise is that the earth is communally owned. I believe my characterization in the text is correct as applied to Locke. Locke holds that initially the earth is unowned, and that all persons have provisional use rights, which can be supplanted by permanent bequeathable full private ownership rights, given certain conditions. See John Locke, Second Treatise of Government (1690), ed. C. B. Macpherson (Indianapolis, IN: Hackett Publishing Company, 1980), chap. 5. In this essay, I use the terms “Lockeanism,” “Lockean theory,” and “Lockean libertarianism” to refer to a family of views that develop doctrines of individual moral natural rights that are broadly similar to the basic position adumbrated by Locke. Although there are important libertarian strands in Locke's doctrines and arguments, his view is sufficiently different from the libertarianism of Rand and Nozick that it would be misleading to call Locke himself a libertarian. In this essay, “hard libertarianism” refers to an uncompromising version of libertarianism. Libertarianism also comes in softer versions. If one qualifies and weakens soft libertarianism sufficiently, eventually one arrives at the “soft Lockeanism” defended in this essay. To my mind, nothing essential hangs on the terminology, but the position I defend departs significantly from the doctrines standardly associated with libertarianism, so it is probably better to reserve the term “libertarian” for positions closer to the paradigm case. For the purposes of this essay, the paradigm libertarian is Nozick.

The exact specification of this derivation, the characterization of its outcome, and the assessment of its success are crucial and tricky issues for Lockean theory, and are much debated.4

For a recent discussion, see John T. Sanders, “Projects and Property,” in David Schmidtz, ed., Robert Nozick (Cambridge: Cambridge University Press, 2002), 34–58. For skeptical discussions of the issue, see G. A. Cohen, Self-Ownership, Freedom, and Equality (Cambridge: Cambridge University Press, 1995). For discussions of John Locke's dealings with this issue, see A. John Simmons, The Lockean Theory of Rights (Princeton, NJ: Princeton University Press, 1992), chap. 5; and Jeremy Waldron, The Right to Private Property (Oxford: Oxford University Press, 1988). The literature on this topic is voluminous.

For the purposes of this essay, I simply assume that some version of this derivation does succeed and that within Lockean theory strong rights of individual private ownership of material resources are justifiable.5

I am assuming the correctness of the Lockean derivation of private property rights arguendo, in order to concentrate attention on other issues, and not because I believe the derivation is correct. For familiar reasons, I doubt it is correct.

III. Why Side Constraints?

This bare-bones statement of the Lockean idea immediately prompts two closely related questions. The list of the ways in which people have a right not to be harmed is a motley set. One wonders, can ordering principles be found that explain and unify the items on the list and that justify each item's inclusion (or suggest revisions)? A second question is, what is the moral basis of individual rights so conceived? What are the reasons that should persuade reasonable persons to accept Lockean morality as correct?

In a review of Anarchy, State, and Utopia, Thomas Nagel reports with some impatience that Nozick does not make much progress in answering these questions:

To present a serious challenge to other views, a discussion of libertarianism would have to explore the foundations of individual rights and the reasons for and against different conceptions of the relations between those rights and other values that the state may be in a position to promote. But Nozick's book is theoretically insubstantial: it does not take up the main problems….6

Thomas Nagel, “Libertarianism without Foundations,” reprinted in Jeffrey Paul, ed., Reading Nozick (Totowa, NJ: Rowman and Littlefield, 1981), 191–93.

I agree with Nagel that Nozick has not explored the foundations of individual rights, but I deny that this lack marks his book as theoretically insubstantial. Decades later, moral philosophers have made only slight advances in the understanding of the moral foundations of individual rights. Nozick's modest achievement in exposing the beams above the foundations looks more and more impressive with the passage of time.

At the most general level, a morality of individual rights denies that moral principles postulate goals, which all persons equally have reason to pursue, by whatever means would be most effective. Such goals might be agent-relative, in the sense that what specific goal is set for a given individual is relative to that individual. “Each person ought to seek to maximize her own happiness” is an example of an agent-relative goal-oriented principle. “Each person ought to seek to maximize aggregate human happiness” is an example of an agent-neutral goal-oriented principle. An agent-neutral goal-oriented principle postulates the same goal or goals for all persons.

In contrast, a morality of individual rights, as Nozick puts it, imposes side constraints on choice of a course of conduct by an agent. Among the available alternative actions a person might choose at a particular time, rights rule out some options; that is, rights render these options ineligible. The person who would respect rights confines her choice to the reduced set of options, the ones that do not violate anyone's rights. From the standpoint of an individual deciding what to do, rights are commands addressed to her, and moreover addressed to her at this particular time: “Do not now do anything that—now or later—violates anyone's rights!”7

The formulation in the text cannot be quite right as it stands. Suppose I wrongfully launched a slow missile attack at you in the recent past, and now I can choose to launch a countermissile that will prevent this attack from injuring you without violating anyone's rights. A Lockean morality will hold that I must now launch the countermissile that will block my previous attempted rights violation. Hence, against the statement in the text, it is not enough that I choose actions that do not, now or in the future, violate anyone's rights. I must also choose actions that, to the greatest extent possible without introducing any new rights violations, undo the effects of past actions of mine that, left unchecked, will violate someone's rights in the future.

A right specifies a way that a person (or group of persons) should be treated or not treated—for example, not to be physically assaulted. Rights function not as goals to be promoted but as constraints to be respected. If A has a right not to be assaulted, that right does not tell me to act so as to minimize assaults on A. Rather, it tells me that I must not now choose any act that involves my assaulting A (or my inducing or assisting another to assault A). Rights belong to people, the right-holders. If A has a right not to be assaulted, this generates a duty on my part not to assault A, a duty that is owed to A, and one that is waivable by A and transferable by A to others.

As so far characterized, a morality of individual rights could consist entirely of agent-relative positive duties to help people in need and more generally to undertake specified sorts of actions toward specified other people.8

This claim in the text might sound obviously false as a characterization of a morality of rights conceived as a morality of constraints, so some explanation is in order. A morality of goals postulates goals and directs the agent to undertake whatever means are necessary to reach the goals. A morality of constraints says the agent may pursue whatever goals she wishes provided that certain constraints on her choice of goals or courses of action to achieve the goals are respected. The constraints might either enjoin a positive act or direct the agent to refrain from a type of act. Thus, a morality of constraints might assert, “Don't harm your mother!” or “Whatever else you do, help your mother!” Just as the former, negative constraint does not say, “Do whatever is required to bring it about that your mother is not harmed (by another or by you),” the latter, positive constraint does not say, “Do whatever is required to bring it about that your mother is helped (by another or by you).” A positive constraint prescribes an action that must be done, not a goal that must be achieved. In a conciliatory spirit, I add that if the reader wants to insist that a morality of constraints can only consist of negative constraints, the reader is welcome to amend my text accordingly. Nothing I want to assert hangs on this point.

The doctrine decisively does not go in that direction. The core content that fills the individual rights structure is the idea that the negative duty not to harm others in certain ways takes priority over any positive duties to give aid.9

This terminology of negative and positive duties is taken from Philippa Foot, “The Problem of Abortion and the Doctrine of Double Effect,” reprinted in Philippa Foot, Virtues and Vices and Other Essays in Moral Philosophy (Berkeley: University of California Press, 1978), 27. This essay was first published in 1967. According to Foot, negative duties are duties not to harm, more generally to refrain from specified courses of action. Positive duties are duties to aid, to do something for somebody or other.

The libertarian versions of this moral theory are uncompromising in this regard: the moral rights that each adult person initially has are entirely negative—rights not to suffer interference or harm of certain sorts. No adult individual initially has any right to any sort of positive treatment or aid from others. No individual initially has any duty to provide such aid, though by voluntary acts such as binding oneself by contract or doing what brings about the birth of a child, or through inadvertence amounting to negligence, one may according to libertarianism come to have strict duties to provide others with aid, to which (under these circumstances) they have a right. This is the thesis of self-ownership: Each adult person is the full rightful owner of herself, and possesses over herself the full rights to use and abuse that an owner of a piece of property has over that property. Since A owns herself, no other person B has any property rights in A, which would give B a right to dictate to some extent how A should use her own body (beyond the negative constraint not to harm others in certain ways).

Why should we accept that morality rightly construed has a side constraint structure, that it is constituted by side constraints? Also, why should we accept the further claim that the fundamental constraints are negative, not positive? Nozick tries to place this conception in an attractive light by clarifying some implications of one version of it and suggesting that these implications are plausible. In effect, Nozick tries to show that the Lockean structure is acceptable in reflective equilibrium.10

An individual is in a state of reflective equilibrium when she has critically examined pertinent arguments and affirms a set of general moral principles that explain and justify the moral judgments that she endorses about what to do in particular circumstances. The agent's particular and general moral judgments are in equilibrium. For this notion, see John Rawls, A Theory of Justice, rev. ed. (Cambridge, MA: Harvard University Press, 1999), 40–46.

Beyond that, he floats some suggestions about what the moral basis of Lockean rights theory might be, but I agree with Nagel that these suggestions do not advance the discussion very far.11

One suggestion that Nozick makes is that beings who have a capacity for rational long-term agency and meaningful life in virtue of those capacities acquire Lockean rights not to be harmed in specified ways, rightly deemed wrongful. See Nozick, Anarchy, State, and Utopia, 48–50. On its face, the suggestion is incomplete. Since persons whose Lockean rights are violated often succeed in leading meaningful lives, not suffering such violations cannot be a necessary condition for achieving a meaningful life. Anyway, why should possession of a capacity for meaningfulness bring it about that one is endowed with Lockean rights? Nozick recognizes this gap but does not hint at how one might fill it. For criticism of Nozick on this point, see Samuel Scheffler, “Natural Rights, Equality, and the Minimal State,” reprinted in Paul, ed., Reading Nozick, 148–68.

However, this does not mean that Nozick's discussion has been superseded, since no one else has succeeded in advancing the discussion much further in the interim.12

For the suggestion that rights confer the valuable status of inviolability (or at least limited violability) on every person and that this status is valuable, hence we have rights, see Frances Kamm, “Non-consequentialism, the Person as an End-in-Itself, and the Significance of Status,” Philosophy and Public Affairs 21, no. 4 (1992): 354–89. Nagel seems to endorse Kamm's suggestion in Thomas Nagel, “Personal Rights and Public Space,” Philosophy and Public Affairs 24, no. 2 (1995): 89–93. Kasper Lippert-Rasmussen refutes the suggestion in his “Moral Status and the Impermissibility of Minimizing Violations,” Philosophy and Public Affairs 25, no. 4 (1996): 333–51.

This essay retreats from the large theoretical questions raised in the preceding paragraph. I shall instead follow Nozick's lead and explore variants of the particular side constraints and individual rights that Nozick espouses. The aim is to seek a position in the region of his version of libertarianism that, as it were, rounds off its sharp edges and brings us closer to a reflective equilibrium that considers and accommodates examples that tend to elicit nonlibertarian responses in many of us, even those of us who are initially sympathetic toward the spirit of Nozick's project. One recurring suggestion is that within the side constraints framework, there are plausible candidates for the status of constraints other than those on which Nozick concentrated attention. Another is that perhaps morality has a hybrid structure that combines side constraints and goal promotion in some way. The most straightforward hybrid identifies moral goals that each person should pursue to some extent—the goals to be balanced off against constraints in deciding what to do in any case, with constraints having less than infinite weight (less than lexical priority). Another hybrid strategy suggests that for any right of any strength, if the consequences of abiding by it in any particular case would be sufficiently bad, one is morally permitted to do what the right, at the first level, prohibits.

IV. Nozick's Chapter Four Questions

One significant foundational question that Nozick leaves unresolved is the topic of chapter four of Anarchy, State, and Utopia. To introduce the topic, Nozick has us suppose we have some understanding of the content of the moral rights that we naturally possess. Nozick states, “Individuals have rights, and there are things no person or group may do to them (without violating their rights)” (ASU, ix). People have rights, moral claims not to be treated by other people in certain ways. An individual's right is partly constituted by the obligations of others to constrain their behavior by refraining from doing what the right forbids. Nozick suggests that we think of the right as drawing a boundary in moral space. The question then arises, “Are others forbidden to perform actions that transgress the boundary or encroach upon the circumscribed area, or are they permitted to perform such actions provided that they compensate the person whose boundary has been crossed?” (ASU, 57). In another terminology, we might say that actions that do what a person's right specifies one should not do infringe the right. Infringements that are, all things considered, morally wrong (at least partly on the ground that they are infringements) are violations of the right. The issue then arises whether all infringements of rights are violations. Nozick's question then can be rephrased: Are some or all infringements of rights coupled with full compensation to the injured right-bearer morally permissible?13

Nozick identifies full compensation with compensation that leaves the recipient neither glad nor sad that the combination of the rights violation and the compensation for it occurred—a subjective view. An objective version would hold that the individual is fully compensated when she is rendered neither better off nor worse off by the combination of suffering the rights violation and receiving the compensation for it.

Or are any and all infringements of rights morally forbidden (unless the right-bearer consents to what is done, in which case there is no real infringement)?

To make sense of what Nozick is up to in his chapter four discussion, we must suppose that we are thinking through questions about what rights people have and what exactly rights require of people, without already having committed ourselves firmly to principles that entail particular answers to these questions. In Rawlsian terminology, we have not attained reflective equilibrium and we know it. Reflection on some examples and cases persuades some of us that Lockean accounts are on the right track, that some position in this neighborhood is correct, but much remains unsettled. If we take it for granted at the start of the discussion that all people are endowed initially with a particular set of moral rights, of known content and character, and that it is always morally wrong to act against anyone's rights without obtaining the individual's prior consent, Nozick's chapter four questions, which presuppose that the content and character of rights are to some degree unsettled, will appear either trivial or incoherent. Nozick will be read as asking, “Under what circumstances is it morally permissible to do what it is never morally permissible to do, namely, violate rights?”

This essay explores these questions. I shall endeavor to assess the adequacy of the responses and suggestions adduced by Nozick in his pathbreaking discussion and to follow some of his insights to see where they lead. I shall proceed by describing examples, considering possible responses, and formulating principles that would explain and justify the responses and that seem independently plausible.

V. Hard Libertarianism

The position I shall call “hard libertarianism” holds that Lockean natural rights (1) may never permissibly be infringed without the prior consent of the right-holder and (2) may always permissibly be infringed provided that the prior consent of the right-holder has been given. In addition, hard libertarianism affirms unequivocally and without any qualifications the four claims that make up the basics of Lockeanism as presented in Section II of this essay. Soft libertarian/Lockean positions substitute “sometimes” for “never” in (1) and “also” for “always” in (2), and in addition relax some or all of the four basic claims discussed in Section II. Beyond a certain point, substitution and relaxation of this sort renders the term “libertarianism” unapt.14

See note 3 for my usage of the terms “Lockean” and “libertarian” and “soft.”

Hard libertarianism is controversial both in what it allows and in what it prohibits. Consider some examples.

Whim

For no good reason A voluntarily requests that B saw off A's arm. B saws off A's arm. Provided that A is sufficiently morally competent to qualify as a right-bearer, he has the right to waive any right that he has and to set aside the protection to his interests (here, his intact arm) that his Lockean rights afford. If A is neither mentally retarded below an appropriate threshold nor so severely mentally ill that he is not reasonably deemed responsible for his choices, his moral rights include the right to waive any of his other rights.

Hiker

A is a hiker lost in a blizzard in the mountains. He stumbles upon a cabin that is privately owned by B and posted with “No Trespassing” signs. Although the door to the cabin is locked, A could break the lock, enter the cabin, build a fire using the cabin's furniture as fuel, eat the food from the larder, and save his life. Hard libertarianism holds that A is morally prohibited from taking B's property without B's permission even to save his life and even if he fully intends to provide full compensation for costs imposed on B by his taking.

In the face of these examples, the hard libertarian might stand fast by her position. If one were moved to alter hard libertarianism so that it yields a softer verdict on such cases, the next question would be, what sort of norms might be reasonable to embrace to give shape to the accommodations? In this connection it is helpful to pay attention to Nozick's insightful discussion that bears on these matters.

VI. Nozick on the Factors that Might Determine Permissible Boundary Crossing

Nozick mentions several considerations that militate against the proposal to allow any boundary crossing without prior consent provided full compensation is paid to any person whose right is infringed by the boundary crossing and who is injured thereby. I list all that Nozick mentions:

a. A system that allows boundary crossings with full compensation “embodies the use of persons as means” (ASU, 71).

b. “[K]nowing they are being so used, and that their plans and expectations are liable to being thwarted arbitrarily, is a cost to people” (ASU, 71).

c. “[S]ome injuries may not be compensable” (ASU, 71).

d. An agent may not know that she will have the means to pay compensation if injury occurs and compensation is called for (ASU, 71).

e. Some boundary crossings tend to produce widespread fear and anxiety not only in actual victims, but also in people who identify themselves as potential victims (ASU, 65–71).

f. A system that allows boundary crossings with full compensation licenses an unfair distribution of the benefits of what would be voluntary exchange scenarios except that the prospect of involuntary takings renders negotiation toward voluntary exchange a comparatively unattractive prospect. Suppose A owns a car, which B covets. Rather than negotiate with A and pay the price they agree on, B under a system that allows takings in the absence of voluntary consent can simply take the car and pay A the lowest amount of money that would induce A voluntarily to relinquish the car. Normally the price they would agree on would be somewhat higher than this, with the benefits that arise from the fact that B values the car more than A being split between A and B. Nozick points out that in this sort of case, permitting a taking rather than requiring B to negotiate to a voluntary agreement with A seems unfair (ASU, 63–65).

This list can be pared down. Regarding (c), one can just note that the principle that infringement of any right is allowed provided full compensation is paid to all injured victims does not on its face allow infringement of rights in any case when the infringement causes uncompensable injury. Also, the principle on its face forbids people from infringing a right and declining to pay full compensation because they lack the means to pay, so (d) on Nozick's list is otiose. The issue of compensation raises practical questions concerning what sorts of conditions on right infringement should be built into legal rules or social norms. But the theoretical question— what moral principles are correct?—is not touched by this sort of consideration. The proposal to allow all boundary crossings provided full compensation is forthcoming emerges unscathed from the worries about compensation expressed in (c) and (d).

The injunction in (a) against using people as means strikes me as not advancing the discussion and hence as eliminable from the list. Using people as means to one's ends cannot in itself be problematic. This occurs constantly in interactions we all would regard as innocuous. Kant's humanity formulation of the categorical imperative forbids using people (the humanity or rational agency in people) as mere means.15

Immanuel Kant, Groundwork of the Metaphysics of Morals, ed. and trans. Mary Gregor, with an introduction by Christine Korsgaard (Cambridge: Cambridge University Press, 1998), 38. The principle states: “So act that you use humanity, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means.”

What is it to use a person merely as a means to one's ends? If I use you as a means to my ends, but only within limits prescribed by morality, I would say I am not using you merely as a means. Following this suggestion, let us say that the injunction against using people as means may be interpreted as the injunction never to use people as means to one's ends in ways that are unacceptable according to correct moral principles. Which principles are these? The slogan by itself does not say, and moreover, so far as I can see, does not point us toward any particular answer. So scratch entry (a) from the foregoing list of pertinent considerations.

Entry (b) should also be dropped. The worry that one's plans are liable to be interrupted by a boundary crossing is a cost to be compensated like any other. If the concern is that people who never actually suffer having their boundaries crossed might be troubled by anxiety that such things could happen to them, this anxiety is an aspect of the generalized fear consideration (to which I now turn).

Nozick makes the interesting suggestion (e) that certain harms such as assault tend to provoke generalized, widespread fear and anxiety, and that for this reason one should not treat the infliction of the harm along with full compensation as permissible. Of course, people's innocent activities that violate no one else's rights will sometimes cause harm to others, and people may fear suffering such harms. If it is permissible for A to inflict a harm on B, given that the infliction violates no right of B, then it can hardly be impermissible that A might also be inducing anxiety in C that a similar harm might befall him. For example, A might permissibly harm B by successfully wooing the person B ardently wants to marry, or by successfully applying for a job for which B, who is less well credentialed, also applied, and which he would have been granted but for A's application. But if one is inflicting harm by doing something that at the first level is forbidden by a moral right, the situation seems different.

One response that one might put in the mouth of the advocate of the proposal to allow boundary crossings with compensation is that the point about fear is not an objection to the proposal but, rather, indicates a complexity in the idea of paying full compensation. Suppose I assault B and as a consequence C, D, and E become fearful. We might hold that this sort of indirect harm is properly traceable to the infringer of the right, hence full compensation to victims should be understood broadly so that C, D, and E qualify as victims to be compensated. At this point, one might draw a line between reasonable and unreasonable fear responses. Suppose that C and D are in circumstances that are in relevant ways similar to B's, so that it is understandable and reasonable that they come to believe that there is some considerable probability they might be harmed in a way that is similar to what has happened to B. They become alarmed at this prospect. In contrast, E lives in Alaska, far away, or he has a moat with alligators around his house, or for some other reason the fact that I have assaulted B does not provide any grounds for E's altering upward his belief that he might be subject to assault (or provides grounds that reasonably raise the probability only below some threshold that should be tolerable). We might distinguish sharply between the reasonable fear experienced by C and D and the unreasonable fear experienced by E and require the boundary crosser to compensate fully the C's and D's of the world but not the E's. Alternatively, we might discount the cost of fear by the degree to which the person who experiences the fear of the boundary crossing should be held responsible for unreasonably becoming fearful and reduce the required compensation correspondingly. In any event, what Nozick seems to be calling attention to is a fact about appropriate compensation, not an objection to the proposal under review.

Nozick objects to this treatment of fear. He notes that it may in practice be very difficult to determine whether the vague anxiety E experiences is fear of assault or something else, and whether the fear is properly traceable back to any particular boundary crossing or aggregate of boundary crossings. He also notes that even if no boundary crossings of a certain type in fact occur, the bare fact that the rules in play allow boundary crossings of that type provided compensation is forthcoming might elicit fear. Here I am inclined to draw the line. If no boundary crossings of type X have ever occurred, how can I reasonably fear such attacks? This anxiety must be either de minimis or caused by my oversensitivity rather than by any actions of other persons. Or one might hold that B's fear of boundary crossing X in the absence of any instances of boundary crossings of type X, to the degree it is reasonable, must be caused by the fact that boundary crossings of other types occur, so that one should then hold the perpetrators of these other boundary crossings liable to pay for this extended but reasonable anxiety as well. The discussion in this and the two preceding paragraphs then leads to the conclusion that Nozick's consideration (e) also deserves to be eliminated from his list.

To recapitulate: Nozick cited six considerations, (a) through (f), as weighing against the proposal to allow any boundary crossing without prior consent provided that full compensation is paid to any person whose right is infringed and who suffers injury. The foregoing discussion has suggested grounds for dropping all but (f) from his list.

VII. Fair Distribution of the Benefits of Exchange

I now turn to the sole remaining entry on Nozick's list, point (f), concerning the fair distribution of the benefits of voluntary exchange (versus unfair distribution in a system that allows boundary crossings with full compensation). Nozick's point is not that there is an independently determinable fair price for goods and services, which voluntary exchange tends to approximate. He disparages the medieval notion of “just price” as a possible basis for economic regulation. Nozick's point is that if one has full ownership rights in something, it is unfair that one is dispossessed of the something absent one's own voluntary choice to relinquish it. This is the norm that is directly challenged by the responses many of us have to the Hiker example introduced at the end of Section V above. There is more to be said about this example, but further examples may be useful. Consider the following.

Holdout

A proposes to use property he owns to start a widget-making factory, which would be profitable, and beneficial to many. The only feasible process for manufacturing widgets unfortunately spews a type of pollution that unavoidably inflicts small harm on a great many people, the millions of residents of a valley. Each resident suffers a one-dollar loss per year from this pollution. A proposes a scheme of compensation that involves establishing a park that will provide two dollars per year of benefit to each valley resident. Millions agree to this scheme, but one resident, B, refuses to relinquish his right not to be impinged on by harmful pollution unless virtually all of the profits of the factory's operation accrue to him.

Suppose the theorist responds that B has no right to gouge his cooperating and productive neighbors in this way. One might propose a “just price” intuition: If one's involvement in a productive activity amounts only to being slightly physically harmed by it, one has a moral right to be fully compensated for this inconvenience, no more and no less. But this principle sounds suspiciously ad hoc and tailored too specifically to the particular example to be very convincing.

One source of unease with the example might stem from concern that one's right not to be impinged on by others in ways that physically cause damage to one's person or property is not a full, tradable property right. After all, C should not be permitted to sell his right not to be harmed in specified ways in the future to D, who might wish to purchase such rights in the hope that their price might rise in the future. (Or should he? The hard libertarian might simply not blink in the face of any putative counterexample.) But the concern about gouging can arise when the rights at stake are ordinary property rights of an unproblematic sort, as in the following examples.

Spite

A installs a huge false chimney on the second story of his house. Its only point is to annoy his neighbor B by blocking B's view of the surrounding landscape. Or A might threaten to install such a chimney and agree to desist only if B pays A very close to B's reservation price, the maximum he would be willing to give to retain the unimpeded view.16

This example is drawn from the 1855 French legal case Keller v. Doerr. I owe this reference to Gijs Van Donselaar, The Benefit of Another's Pains: Parasitism, Scarcity, Basic Income (Amsterdam: Department of Philosophy, University of Amsterdam, 1997), 2.

Easy Rescue

A is drowning. He is in a leaky boat that is slowly sinking, so there is ample time for negotiation. B happens to arrive on the scene in his boat, and offers to save A's life, a feat that can be accomplished at very small cost to B, if A agrees to transfer all of his financial assets to B in return for this easy rescue. (The example can be amplified with details such as that A is far wealthier than B or the reverse, or that A is far better off than B in aggregate lifetime well-being to this point or the reverse.)

Although Nozick does not discuss these two specific examples, it is clear that he would give no quarter to the position that B is under an enforceable duty to perform easy rescue and hence has no right to charge A whatever he is willing to pay for lifesaving service. However, Nozick discusses more sympathetically some pros and cons that might be applied to cases such as Spite. His thought is that the imperative of prohibiting boundary crossings without prior consent in order to ensure fair division of the benefits of exchange applies only to the category of productive exchange. In the typical exchange, both parties benefit, and this sort of exchange is unproblematically productive. He defines an unproductive exchange as one that satisfies two conditions: (1) one party to the exchange is no better off as a result of it than if the other party did not exist at all, and (2) what this party who is no better off gets from the exchange is the other party's abstention from an activity she would not be motivated to engage in except for the possibility of selling abstention. Another test mentioned by Nozick is that one party to an unproductive exchange would be no worse off if the exchange were prohibited.17

Nozick, Anarchy, State, and Utopia, 84–87.

Where an exchange is in the offing that would be unproductive in this sense, morality does not insist on allowing the exchange to go forward. Depending on further circumstances, it might be right to forbid the activity by one party to the contemplated exchange that would become the object to be exchanged, and it might be right in some cases to require (and in others not to require) compensation accompanying the prohibition. If the first condition for unproductive exchange is satisfied but not the second, call the exchange semiproductive. Nozick suggests it may be morally acceptable to prohibit semiproductive exchange by prohibiting one of the parties from engaging in the activity, abstention from which is being exchanged, provided some compensation is given to the party whose favored activity is prohibited in this way.

Nozick applies this analysis to blackmail. In some cases, a blackmail exchange would be unproductive. In these cases, prohibition of blackmail with no compensation to the would-be blackmailer is appropriate. Other cases are mixed. Suppose A wants to publish secrets about B's life. Their inclusion in a book A is writing would improve the book. It is worth a thousand dollars to B to prevent the publication of these secrets. The value to A of his planned use of this information about B is $500. According to Nozick, A can legitimately charge at most $500 for declining to publish B's secrets. This latter case is one in which the first but not the second condition for unproductive exchange would be satisfied, but once B agrees to pay $500 to A, the latter's attempt to make a deal with B for further payment would be an attempt to bring about an unproductive exchange.

The problem with this line of analysis is that the hypotheticals it relies upon are unruly and cannot be tamed.18

In this connection, see the criticism of Nozick in Eric Mack, “Nozick on Unproductivity: The Unintended Consequences,” in Paul, ed., Reading Nozick, 169–90.

I would be better off if the party with whom I am negotiating an exchange did not exist at all in many circumstances in which this fact does not render the looming exchange in any way problematic or liable to special restriction or prohibition. I am being hired as a consultant to help write A's acceptance speech for the Nobel Prize in chemistry, but if A did not exist at all or had nothing at all to do with me, I would be better off, because I would then be next in line to receive the Nobel Prize for my chemistry achievements, which are good, but less good than A's. This fact does not render my exchange of my speechwriting services for A's cash on mutually agreeable terms a semiproductive exchange. The first condition as Nozick formulates it is satisfied, but this does not have the significance Nozick attributes to it. The same is true if I purchase A's abstention from entering a contest she is sure to win if she enters. My chances of winning are greater if A does not enter the contest, and better for that matter if A did not exist or had nothing at all to do with me.

For much the same reason, one cannot say that A wrongfully causes harm to B by physical interaction with B just in case A renders B worse off than would be the case if A did not exist. Nor can the condition be: just in case A renders B worse off than B would be if she had no interaction at all with A. Consider a case in which A has been B's trading partner for years. These interactions have brought a large profit to B. Today A viciously kicks B in the shin, seriously injuring his leg. B is still a net beneficiary of his interactions with A; nonetheless, A violates B's right not to be physically attacked when A kicks B in the shin.

Perhaps it is possible to revise Nozick's conditions on unproductive exchange while preserving the basic idea—or perhaps not. The judgment that in the Spite case A has no right to erect a false chimney on his own property and that it would not be wrong forcibly to prohibit A from harming B in this way need not wait upon the vicissitudes of further developments of Nozick's theory of unproductive exchange. In the example, A is being malicious. He is acting in a way he knows will harm B purely for the sake of harming B. (We can contrast Spite as specified in the text with an alternate version in which A is a lover of false chimneys and wants to erect this addition to the top of his house in order to improve its appearance by his lights. In the alternate version, A in some sense does the same thing as in the original version, but his intention and motivation are not spiteful, and this fact makes all the difference. In the alternate version, A has a moral right to act as he does and should not be prohibited from doing so.) A plausible interpretation of Lockean rights maintains that intention and motivation can affect the permissibility of what an agent does. An act might be permissible if done with a certain intention or with a certain motive, but the same act—the same physical movement leading to the same physical changes in the world—done with a different intention or motive might be impermissible.

One might object that Spite is a very odd case, from which no general normative implications flow.19

In this essay, I am concerned with what it is morally permissible and impermissible for people to do in various possible examples. I am not trying to devise a practical doctrine of laws and social norms, which might well be designed for ordinary cases, not extraordinary ones. That said, if a hard-libertarian reader is disposed to think that this and other cases I discuss are practically resolvable by writing careful contracts, I recommend consulting the theory of incomplete contracting on the economic importance of the impossibility of writing a complete contract that deals with all contingencies that might arise. See Oliver Hart, Firms, Contracts, and Financial Structure (Oxford: Clarendon Press, 1995), and the additional literature Hart cites.

In Spite there is no plausible motive except malice that would likely explain the homeowner's desire to construct a false chimney on his house. But in the general case, it will not be possible to read off the agent's motive from his behavior. For this reason a regime of rights that rendered the limits of an agent's freedom sensitive to harmed-and-putatively-injured parties' hunches and guesses about the motives of the agent whose doings they find distressing would be an administrative chaos. So we should not understand the limits of people's moral rights in this way.

The objection makes a significant mistake. It fails to distinguish the theoretical question—“What rights do we have?”—from the practical question—“What sort of laws should be established in order to protect individual rights?” Epistemic and administrative difficulties will play a large role in answering the second question but not in answering the first. It will then sometimes make perfect sense for an advocate of individual rights to say, for example, that Smith has a perfect moral right to do X, but nonetheless it is morally correct that a just state should prohibit Smith from doing it, or that Smith has no moral right to do Y but nonetheless a just state should allow her to do it.20

Here I presume that an adequate Lockean morality will integrate considerations of mandatory moral goals to be promoted and required side constraints to be respected, and will not go whole hog for either alternative. The fact that if we tried to establish rules that would require the state to enforce people's subtle right to X, the result would be that people's more important rights would be frustrated is then a relevant consideration for the question, what conception of rights should the state enforce?

VIII. Why Ever Allow Infringements without Prior Consent?

The flip side of the question, “Why not always allow any border crossing coupled with full compensation?” is “Why ever allow such border crossings not licensed by prior consent?” Discussing this issue, Nozick suggests that transaction costs—the costs of negotiating deals—play a key role. The costs of reaching a deal might be large because the number of parties whose agreement is needed may be large. It may be difficult to locate the agents from whom agreement must be obtained, and difficult or expensive to set in place the communications technology needed for negotiation. The process of making a deal may consume resources in many ways. At the limit, it may be physically impossible to negotiate with someone.

When transaction costs are high, insistence on prior consent to any boundary crossing might leave all affected parties worse off than they would be if a more permissive rule were accepted. The transaction-cost consideration also introduces a reason of a quite different sort for sometimes allowing boundary crossing provided compensation is forthcoming. Consider an example in which at the time a choice has to be made, prior consent is impossible to obtain.

Coma

A has suffered a serious injury and needs immediate medical care, including surgery. Deliberately touching a person without her consent fits the technical definition of an assault, and imposing medical care on a person—which can significantly affect the patient's well-being for better or worse—without her consent is a serious breach of her rights. In this case, A is in a temporary coma and can neither give nor withhold consent to the surgery she needs to save her life.

It would violate the letter but not the spirit of hard libertarianism to allow the surgeon to provide medical care to A without first obtaining A's prior consent in situations of this sort. Three features of the situation favor this judgment: the infringement being proposed would be of great benefit to A; she is not capable of giving or withholding consent without delay that would dissipate this benefit; and she would consent if she were capable of giving or withholding consent. (The hard libertarian who is strict on this point would insist that A is out of luck if she has not previously contracted for care that encompasses this contingency or delegated to a specific agent the authority to choose in this contingency.)

The Coma case, from one perspective, is the tip of an iceberg. The underlying difficulty is that insistence on no impingement without consent only imperfectly safeguards an agent's liberty and vital interests. The giving and withholding of consent even under favorable conditions sometimes reflects the agent's distorted assessment or weakness of will. At such times, the agent voluntarily consents to what is bad for self and others.

Notice that some of the cases described previously either do not on their face appear to involve transaction-cost concerns or can be redescribed so it is clear that transaction costs are not driving the judgment that favors permitting infringement. Consider just one of these cases: Easy Rescue requires only a negotiation between two people; the negotiation concerns only the price to be paid by one for a simple service offered by the other, and there is plenty of time to negotiate. True, in any thin market setting like this, if the gap between the buyer's and seller's reservation prices is large, agreeing on a division of the gain from trade can be a challenge.

IX. Chipping Away at Self-Ownership

Easy rescue cases, interpreted as prompting the claim that people are under an enforceable duty to be minimally decent Samaritans, directly challenge the self-ownership thesis. Hard libertarians will resist this challenge, but so will others. To explore the moral basis of self-ownership, one should contrast easy rescue cases with another class of cases that they somewhat resemble. Nozick provides an example:

Accommodation

A is minding his business, harming no one. A allows B onto his property. A has no reason to be distrustful of B. However, B takes out a gun and shoots innocent people from A's window perch. Police arrive and attempt to apprehend B. A shouts to the police that he is simply going about his business and has a perfect right to be where he is, which happens to be in the way of the police as they seek to render B harmless.

Nozick raises the question, Does A in these circumstances have a duty to get out of the way of the police as they seek to incapacitate B, who is wrongfully threatening others?21

Robert Nozick, “War, Terrorism, Reprisals—Drawing Some Moral Lines,” reprinted in Nozick, Socratic Puzzles, 303. This chapter was initially a book review published in 1978.

One might hold that the police have the moral right forcibly to remove A from the premises so that he does not impede the urgent crime-fighting effort. One might also hold that if A has a duty to move out of the way of the police and does not, then it might be permissible, depending on further circumstances, for the police to shoot at B even if doing so risks harming A. Given that A is not accommodating in the way morality requires, it might be the case that the police, killing A foreseeably but unintentionally as they seek to stop B, do not violate any right that A possesses against them. We can perhaps simplify the complex issues the example raises by imagining a state of nature version of it.

State-of-Nature Accommodation

A is fleeing B, an evil aggressor who intends to kill A. Let us assume it would be morally acceptable for A to kill B in self-defense, but unfortunately A in the circumstances is unable to do anything that is likely to save his life by attacking B in self-defense. There is a small alcove in the rocks above, from which C is watching B's chase of A unfold. A can elude B and save his own life by jumping onto the alcove where C is standing. B, a heavy-set aggressor, will not be able to follow. C could move to the rear of the alcove to give A room to land safely. If C does not accommodate A in this way, A has no way to save his own life except by jumping to the alcove, landing where C now stands, and thereby killing him. In the circumstances, all the facts just described are mutual knowledge between A and C: A knows, C knows; A knows that C knows; C knows that A knows; and so on. C does not move to the rear of the alcove.

Consider the position that in these circumstances, C has a moral duty to accommodate by moving from the unowned land that he is currently using but which A now needs as a platform to land on, in order to save his life. If C does not make this accommodating move and A jumps to this spot, killing C, A is not violating any of C's rights. One might adhere to this position without rejecting the thesis of self-ownership. That is to say, each person fully owns herself, but this is of course compatible with there being limits on the rights anyone has to use unowned land. The moral basis for the requirement of accommodation is the proper understanding of the idea that in a state of nature any person is at liberty freely to use any unowned land. If one person wishes to use a particular piece of unowned land and this desired use is not in conflict with any use anyone else wishes to make of the land, the use is morally permissible. A limit on this right of free use is that if two or more individuals wish to use the same piece of land at the same time in conflicting ways, the one who begins using the land first has the right of way unless one would-be user has significantly greater need of it than the other. In the latter case, the person with the significantly greater need has the right to use. If A and B both wish to sunbathe on the same strip of unowned land, and A gets there first, A has the right to use the land in this way. But if C needs to lie down on this particular sunlit land in order to relieve her intense headache, A must give way to C. I suppose there is also a time-limit on the extent to which any individual can use and keep using any piece of unowned land. The longer one has used the land for one's own purposes, the stronger becomes the presumption that one must yield to give others a turn, if a queue of would-be users of the same land forms. I do not claim that free use rights so understood are clear and well defined: their inadequacies, which perhaps warrant their supersession by private property rights, include their unclarity and vagueness. My claim is that our understanding of the nature of free-use rights to unowned land, partial as it is, underwrites our understanding of the idea that in the example of State-of-Nature Accommodation, C's right to the unimpeded use of the alcove where she stands disappears when A arrives on the scene with a pressing need to use that same spot temporarily.

Seen in this light, the Accommodation case and the State-of-Nature Accommodation case, which might look similar, in fact raise different issues. The latter raises questions about how to interpret rights to free use of unowned property and the limits of those rights. (These limits might continue in force after appropriation and cast a shadow on private ownership rights.) The former introduces a rights violator and other persons (the police) who are acting at risk to themselves to apprehend the rights violator, stop him from threatening others, and perhaps punish him in order to deter others from seeking to trample on people's rights. Those persons who act in this way act to preserve the system of rights that all of us are morally bound to uphold. In this sense they act for us, and we owe them cooperation corresponding to the amount of cost and risk they incur in the efficient pursuit of this goal. Refraining from impeding their efforts when we can refrain at moderate cost and risk to ourselves is part of the fair return we owe to those who cooperate to uphold the system of human rights by preventing those who would violate rights to advance their own ends from succeeding in their efforts. Consider in this connection the following example:

Free Rider. A, B, C, D, and E cooperate to provide police protection in their neighborhood. These patrols significantly benefit everyone who lives in this area, including F, by deterring crime. Given the nature of this benefit, there is no question of accepting it or rejecting it; it falls on everyone in the area, willy-nilly. The cooperators demand that F pay one-sixth of the cost of the ongoing police patrols, but F refuses.

In these circumstances, F is dragooned into receiving the benefits of the cooperative scheme that supplies the public good of justice. Is it acceptable to dragoon F into paying his fair share of the costs? (The complication is that given that A, B, C, D, and E are going to provide the good, they have no choice but to provide it to F as well.) The hard libertarian answers “No.” The hard libertarian holds that the individual should be left free to live as he chooses unless his actions harm nonconsenting other people in certain ways that violate their rights. Failing to reciprocate a benefit that another person has conferred on you, in the absence of any prior mutual agreement to reciprocate, does not constitute your harming the other, according to the hard libertarian, in any way that amounts to a violation of the other's rights.

One may appeal to two arguments, alone or in tandem, to support a “Yes” answer, that is, to hold that coercing F to pay his fair share of the scheme is acceptable. One is the Hart-Rawls principle of fairness.22

For a formulation of the Hart-Rawls principle of fairness, references to the relevant prior literature, and sharp skeptical discussion of the principle itself, see Nozick, Anarchy, State, and Utopia, 90–95; A. John Simmons, “The Principle of Fair Play,” reprinted in Simmons, Justification and Legitimacy: Essays on Rights and Obligations (Cambridge: Cambridge University Press, 2001), 1–26; and A. John Simmons, “Fair Play and Political Obligation: Twenty Years Later,” ibid., 27–42.

When others cooperate to supply public goods that benefit a number of people, those who have cooperated have a right that the other beneficiaries pay their share of the costs—to refuse is to be a free rider, which one has no right to be.23

The duty to refrain from being a free rider might be construed broadly, applying to any scheme that supplies significant public goods, or narrowly, applying only to public goods schemes that provide goods of justice (security of enjoyment of all people's natural rights). The rationale for holding to the narrow construal would be that one views the system of Lockean natural rights, properly understood, as of paramount moral importance, swamping other considerations in the determination of what we owe each other.

I construe this principle to continue to apply when there is no voluntary acceptance of benefits received (because of the nature of the good, not because the cooperators have deliberately foisted the good on an unwilling recipient). The hard libertarian reply is that in the absence of some voluntary act that can plausibly be regarded as triggering an obligation, no duties or obligations beyond the negative duties not to harm ever arise for anyone.

The second argument asserts straightaway that each person has a natural duty to promote justice—to do his part, when feasible and not excessively costly, to bring into existence and sustain schemes that protect people's moral rights. In the Free Rider case, the natural duty to promote justice applies to F and, in these circumstances, requires him to pay his fair share of the costs of police protection. The hard libertarian replies either by flatly denying the natural duty to promote justice or in a more subtle way by denying that the general duty to support justice anywhere and everywhere requires F (for example) to contribute to the particular local justice promotion enterprise that his neighbors have devised.24

For this argument, see A. John Simmons, “Associative Political Obligations,” in Simmons, Justification and Legitimacy, 65–92.

In response one might argue that the natural duty to promote justice requires the individual to accept only modest risk and cost in the service of justice, but the contribution that is made must be an efficient use of his resources. Normally this efficiency requirement ties the general duty to promote justice to some local scheme that is operating or is on the horizon. One can more surely and easily promote justice nearby than from afar. This need not always be the case, so the natural duty of justice may allow some Oliver North's and Che Guevara's to be excused from local justice duties because they are contributing sufficiently to justice provision schemes elsewhere.25

I am imagining that Oliver and Che, each in his own way, are deemed to be doing enough to advance the cause of justice in a distant country that the natural duty to promote justice cannot be interpreted as ruling out their disobeying some laws of their home country. This is a concession to the Simmons point cited in the previous footnote.

The important break from hard libertarianism occurs when one accepts the natural duty to promote justice along with the norm against free riding.

One issue is whether the individual has a moral duty to refrain from free riding in circumstances like the Free Rider case. A second issue is whether the duty is enforceable, whether it warrants coercive force. Hard libertarians tend to see coercion as a large evil. I disagree. Being coerced to refrain from acts one has a right to perform is morally odious, but coercion that prevents one from doing what one has no right to do is not so bad. If traffic laws are just, it is unfortunate if I am coerced to obey them, because it would be better if I willingly complied, so that I am not coerced (the state's threats directed at me being idle). But if I am disposed to disobey, and coerced not to do so by effective traffic law enforcement, the forcing of my will here is not a significant bad, and is definitely preferable to the state of affairs in which I am free of coercion and act wrongfully, violating significant rights of other people.

X. Self-Ownership Revisited

Construed in a libertarian spirit, self-ownership asserts that each person fully owns herself and may do with herself whatever she likes so long as she does not thereby harm others. This entitlement includes the right of each person to destroy herself or waste her own life. As such, this same entitlement strictly forbids restriction of a person's freedom for her own good.26

Familiar qualifications to this formulation are needed. I might sign a contract that authorizes you to coerce me to stop eating when I try to indulge in rich desserts, and you might agree to this contract from concern for my good. When I lunge for the fancy chocolate cake and you restrain me, you are restricting my freedom for my own good, but in virtue of the contract, you are not violating any right of mine. The libertarian will also qualify the formulation to allow at least temporary paternalistic restriction of liberty in order to block egregiously nonvoluntary choices, as when I am about to drink poison in the false belief that the glass contains wine.

One might instead hold that each individual has over herself the full rights that a private property owner has over whatever things he owns. These rights might not be unlimited, even setting aside the harm-to-others issue. Locke flirts with a stewardship conception of private ownership rights. He claims that one's ownership over a piece of land lapses if one lets the land go to waste.27

Locke, Second Treatise of Government, chap. 5.

Pressed to its logical limit, the no-waste condition would have it that if anyone else would use your land more productively and efficiently than you would, your right to own it gives way to the right of the more efficient would-be user. Suppose the idea is, rather, that one's uses must meet some acceptable threshold standard of productivity in order to satisfy the no-waste proviso.

By this line of thought, each person's self-ownership rights are also limited by a no-waste requirement. Owning oneself, one has a moral duty not to waste one's life but to make good use of it. A human life is (barring catastrophe) a precious opportunity for good, which ought not to be squandered. This norm leaves each individual vast realms of freedom to live as she chooses, because there are boundless varieties of ways to make something good of one's life. Still, there are limits. The no-waste condition is flagrantly violated by the agent who consents to be mangled for no good reason in the Whim case. Another example would be a self-indulgent, petulant suicide. An agent may violate the no-waste condition deliberately or by mistake. If self-ownership is understood by analogy with land ownership constrained by the no-waste requirement, paternalistic restrictions of an agent's liberty for her own good in extreme cases of mistake or folly are consistent with self-ownership, not wrongful violations of it.28

Some readers will wonder to whom the duty not to waste one's life might conceivably be owed, and who would be entitled to decide that an individual is violating the duty and to intervene forcibly on the individual's behalf. On the first question, the duty not to waste one's life is owed to those—oneself and others—who stand to benefit if no waste occurs. This is a diffuse group. On the second question, in a Lockean theory, anyone is morally entitled to act to enforce people's natural rights. By the same token, anyone should be morally entitled to act to enforce the natural duty of each person not to waste his own life. Attempts at enforcement might spark umbrage on the part of the person who is being judged a life-waster, and quarrels may ensue, so here as in other cases of conflict around the enforcement of rights and duties, there is need for an impartial, reliable, judicious umpire, if one can be found.

XI. In Search of Principles: The Pareto Constraint

We need to search for principles that might unify some of the seemingly disparate responses to the various puzzle cases examined to this point. One possibility is that rights are subject to a Pareto constraint:

A state of affairs that can be improved by making someone better off without making anyone else worse off is morally unacceptable. The specification of individual rights should be adjusted (if possible) so that respecting everyone's individual rights does not produce such a state of affairs.

This Pareto constraint varies in substance depending on the interpretation of “better off” and “worse off.” On a subjective interpretation, one is better off just in case one's preferences are satisfied to a greater extent (with one's own ranking of preferences determining the priority of satisfying each of them). On an objective interpretation, one is better off if one's well-being, assessed by the correct standard, is greater. The objective interpretation obviously relies on controversial value-theory assumptions, but if it can be sustained, the Pareto constraint generates moral considerations of greater strength, and I adopt it here. The appeal of the Pareto constraint in the context of developing Lockean theory is evident in the following example:

Benign Trespasser

A is the absentee owner of a large estate with an orchard. B and C trespass on the land in the autumn, picking ripe fruit in the orchard that would otherwise rot. They do no damage. There is no feasible way to spread the benefits B and C receive to other people without dissipating them. A's absentee status renders it the case that no negotiation between A and B and C can occur that would fix a mutually agreeable price for this incursion on A's land. Here, the Pareto constraint on the interpretation of Lockean rights requires us to take the view that the right of private ownership allows benign trespassing.

Some of the examples discussed so far in this essay that suggest the need to qualify libertarian principles are not explained and justified by the Pareto constraint. Consider the Easy Rescue case. One might try to claim that some would be better off and none worse off in a society in which our conception of individual rights is adjusted so that all who find themselves in serious predicaments and need easy rescue have a right to easy rescue if anyone is in a position to provide it, and each person has a duty to provide easy rescue if he is in circumstances such that by easy rescue he can prevent a person's unwanted serious injury or premature death.

One might claim that from a suitable ex ante standpoint, before it is known who will be in the rescuer role and who will be potential beneficiaries of easy rescues, everyone gains by agreeing to a rule requiring easy rescue. But ex post, once it becomes evident who will in fact be playing what role on what occasions, it is no longer plausible to maintain that a rule mandating easy rescue renders everyone better off than he would be if the rules were not in place. Consider A, who is lucky enough never to need an easy rescue and unlucky enough to find herself sometimes in the rescuer role. She is ex post a net loser. Or consider B, who as it turns out finds himself in a position to extract a large monetary windfall by negotiating the terms of an easy rescue with Bill Gates, who happens to need this service and is extraordinarily flush with cash to pay for it. B is better off over the course of his life in a world in which Lockean natural rights are enforced and the rights are not adjusted by adding a duty of easy rescue. That way, he gains the windfall from Bill Gates, which he does not receive under the regime in which easy rescue is required.

So consider the proposal that initially Lockean rights should be adjusted in order to incorporate an ex ante Pareto requirement. The idea would be that the rules that specify the content of Lockean rights should be qualified as necessary to yield the result that, from an ex ante perspective (before it is known who will gain and who will lose from any rules alteration), one cannot alter the rules further in a way that would improve everyone's expected level of advantage from the operation of the system of rules.29

This example illustrates the ex ante Pareto norm: Suppose that 100,000 people live in an isolated valley and that each year each resident faces a 1 percent chance of suffering an automobile accident that will cost the individual, on the average, $20 if easy rescue is forthcoming and $100 if it is not. (Easy rescue will be forthcoming if and only if it is legally required.) If everyone in the valley is legally required to provide easy rescue if present at an automobile accident, each member faces a 1 percent chance per year of being required to perform this service at a cost to the rescuer, on the average, of $30. Money is linear with utility for valley residents and all have a neutral attitude toward risk. In these circumstances, the introduction of a legal requirement of providing easy rescue provides an ex ante Pareto improvement.

To clarify the proposal, one would need to specify in a more determinate way the relevant ex ante standpoint. Even without clarification, it should be plain that the ex ante Pareto norm radically alters Lockean rights. This point becomes more vivid if attention is switched from Easy Rescue examples to examples that involve the forced utter sacrifice of some persons' interests to benefit others. Surely a Lockean rights perspective that retains its integrity must disallow killing one person when that would be a necessary means to saving two others. But from an ex ante perspective, not knowing whether one is the person to be sacrificed or one of the two to be saved, everyone's prospects are improved by acceptance of a policy that licenses an agency to kill one innocent whenever doing so saves more than one innocent. The ex ante Pareto norm, once accepted as qualifying a set of Lockean rights, transforms the “rights” perspective into something close to aggregative utilitarianism.

If the aim of the discussion in this essay is to explore the alternatives within the Lockean tradition rather than others that radically and brutally depart from it, we should interpret the Pareto norm as ex post rather than ex ante.

The ex post Pareto norm does not unequivocally ratify some of the judgments about examples that I am supposing soft Lockeans will want to ratify. Easy Rescue is one such case. Hiker is another. A rule allowing infringement in circumstances like those of Hiker could be devised that would improve everyone's welfare (compared to the hard-libertarian position that requires the hiker to perish), but if we vary the case by providing the lost hiker access to a pay telephone so negotiation with the cabin owner can occur prior to use of the mountain cabin, any rule that partially expropriates the property right to the cabin for the benefit of those who need emergency aid is worse for the cabin owner, who can benefit by selling temporary use of the cabin to those in need of it. If, given this amended version of the case, the judgment persists that the cabin owner's property right should give way in the face of the lost hiker's need, the basis of the judgment is not the Pareto constraint but a norm of fairness that prohibits gouging the needy in this sort of situation.

XII. Moderation

One possible nonlibertarian response to the Easy Rescue case denies that what warrants the limitation of self-ownership rights in this case has anything specifically to do with self-ownership. The idea would be that any moral right, however sacred, gives way if the consequences of upholding it in a particular case are sufficiently bad. Rights on this view are nonrigid, spongy side constraints.30

The term “spongy side constraint” is borrowed from Judith Jarvis Thomson, who uses it to describe a moderate position on the stringency of moral rights in her The Realm of Rights (Cambridge, MA: Harvard University Press, 1990), 154.

For any individual deciding what to do and faced with a set of available courses of action that might be chosen, rights are side constraints that rule out certain courses of action and render them morally ineligible for choice, but if the consequences of abiding by these constraints would be sufficiently bad, they relax, and options that would otherwise be ineligible become eligible.

In an interesting footnote, Nozick sets aside the issue of whether rights are rigid or nonrigid side constraints. He comments, “The question of whether these side constraints are absolute, or whether they may be violated in order to avoid catastrophic moral horror, and if the latter, what the resulting structure might look like, is one I hope largely to avoid” (ASU, 30). Nozick does not declare a commitment one way or the other on this issue.

An alternative picture that contrasts with the rights-as-spongy-side-constraints metaphor represents rights as rigid but sometimes reconfigured under pressure of consequences. Sufficiently bad consequences can reshape the requirements that rights impose, the resultant rights being sometimes rigid again and sometimes nonrigid. For example, each person has a right of self-ownership, and no duty to use one's body to provide aid to the needy, but if the consequences of insistence on self-ownership would be sufficiently bad, the right dissolves, and may become a strict duty to aid, and correspondingly a right on the part of the needy to be assisted.

Judith Jarvis Thomson has suggested a useful way to characterize the stringency of a right: the degree to which it resists being overridden by counterconsiderations.31

Thomson, The Realm of Rights, chap. 6. An earlier discussion is in Judith Jarvis Thomson, “Some Ruminations on Rights,” reprinted in Paul, ed., Reading Nozick, 130–47. This section of my essay is heavily indebted to Thomson's writings on the issue.

She proposes that a right is more stringent the greater the harm that would ensue for the right-holder if the right is not upheld. This formulation by itself explains the plausible thought that even though I have a right not to be violently assaulted and a right that my extra shirt button not be taken from me without my consent, the two rights are not on a par. They are not on a par in the sense that a greater weight of reasons is required to justify overriding the right against violent assault. A great harm is less easily offset than a small harm. Note also that if by some chance you would suffer grievous injury if your extra shirt button were stolen, the right to retain your extra shirt button would become correspondingly more stringent.

Thomson's formulation needs some tweaking.32

In my essay “Moderate Deontology, Aggregation, and Rights” (typescript available from the author), I consider how one may aggregate goods and bads of various amounts that would accrue to the right-holder and others if one respects a right or not. Some doubt that failure to respect a person's serious right, such as a right not to be killed, could be offset by any aggregation of tiny benefits to others, however numerous these others. I argue against this plausible claim.

Suppose that I would not be harmed at all if you stole from me the hard drugs I own and prize, because without your intervention I would use the drugs to my detriment. My right to the hard drugs I own then has zero stringency according to the formulation of the previous paragraph. However, the theft in these circumstances might be deemed an instance of wrongful paternalism, restriction of a person's liberty against his will for his own good. The individual has an interest in personal sovereignty, in not being subject to such paternalism, even if frustration of the interest does not harm him or reduce his welfare. Also, a person might have an interest in pursuing a valuable goal, the fulfillment of which would not enhance her own welfare. (Call such interests agency interests.) If A steals B's rowboat, which she was about to use to rescue someone in distress, the net effect may be no harm at all to B, who now is able to enjoy her picnic lunch in peace. But the significant loss to B's agency interest, we plausibly suppose, renders B's right to undisturbed possession of her rowboat a stringent right in the circumstances.

To accommodate these possibilities, I shall simply stipulate that a right is more stringent the greater the loss to the right-holder, in terms of the fulfillment of her interests (including welfare, personal sovereignty, and agency interests) if the right is not upheld.

According to the moderate view of moral rights, the stringency of a right depends on what is at stake for the right-holder. If the net effect on other people of infringing the right (acting against it), compared to what would happen to them if the right were not infringed, is positive and sufficiently large, the right is overridden. When a right is overridden, it is at least permissible, and may be mandatory, to infringe it. The ratio of the loss-to-the-right-holder-if-the-right-is-infringed to the net-loss-to-others-if-the-right-is-not-infringed determines whether or not infringement in a particular case is permissible. If the numerical value of the ratio is sufficiently large, the right may permissibly be breached. The simplest view would hold that there is one numerical value of this ratio that always marks the point of permissible infringement of any right. Another view would hold that this numerical value might vary depending on the type of right under consideration. For example, a one-to-three ratio might suffice for relatively inconsequential rights, so that it is morally permissible to steal my extra shirt button in order to prevent three similar thefts. The required ratio might shift for more consequential rights, so that it is not morally permissible to attack and kill one healthy, innocent person even in order to save three other healthy, innocent persons from a similar fate.

A further complication is that considerations of responsibility or deservingness might discount or amplify the value of the losses and gains of affected parties that determine whether a right in given circumstances may be overridden. A limiting case that illustrates this point is a lone individual defending herself against attack by evil aggressors. It's her life or theirs. Given the culpability of the agents who would suffer if the individual's right not to be attacked is upheld, their numbers do not matter. Even if the evil aggressors are legion, their cumulative loss if the individual defending her life manages to kill all of them would not morally outweigh her loss if she were wrongfully killed.

XIII. “Core” Self-Ownership: No Duty to Aid the Needy Per Se

None of the modifications to the Lockean individual rights doctrine that I have suggested contravenes the core of the libertarian self-ownership constraint, which I would formulate in this way: The mere facts that one person is in a position to help another person who is in need of help and that the first person could provide assistance so that what she loses by providing aid is less than what the beneficiary gains are not sufficient to generate a duty to aid. Nor are matters changed if it is stipulated that nothing else the first person could do instead would help the needy to a greater extent (at lesser cost to herself than the gain to the beneficiaries). The mere fact that another is in need does not trigger an enforceable duty to aid.33

Just for the record, I state that I do not myself endorse this statement of core self-ownership. In this essay I am exploring the natural rights tradition from within, not lobbing external criticisms at it.

This core self-ownership idea is consistent with the claim that one's ownership of things other than oneself is less than full. When the ratio of the good that one's possessions would do if used by others to the good those possessions would do if they remained entirely in one's own control exceeds a threshold, the right of ownership relaxes, as in the Hiker case.

Core self-ownership is also consistent with the claim that one has a natural duty to promote justice (the condition in which all people's natural rights are fully respected) to some extent. This duty can be understood so that it generates a duty to cooperate with citizens acting to uphold justice as in the Accommodation case. Core self-ownership is also compatible with the anti-free-rider norm suggested by the Free Rider case.

If core self-ownership is construed weakly as I have been suggesting, various other aspects of concern for fairness might be embraced without triggering conflict with this weak self-ownership constraint. One such aspect is that in thin market settings, where there are few buyers and sellers for a good or service, one does not have the right to charge whatever the traffic will bear. In particular, when the gap between the reservation prices of the would-be buyer and seller is very large, one does not have the right to force one party to an exchange at her reservation price even if one can do this by hard bargaining. This aspect of fairness colors the response to the Holdout, Hiker, and Easy Rescue cases. Notice that the rejection of the idea that in these examples the person in a strong bargaining position may legitimately charge whatever price she can negotiate does not rule out profit-taking in these settings by the person with the bargaining edge. Morality might accept profit-taking (beyond bare compensation for loss) but prohibit gouging or excessive profit-taking in such cases.

Objection: There is another element of self-ownership that equally merits designation as part of the core of the principle, and that straightforwardly conflicts with the judgments supposedly evoked (in my view) by Holdout, Hiker, Free Rider, and Easy Rescue. People are free to lead their lives as they choose (so long as they do not harm others) and may not be used against their will to advance the purposes of others.

Reply: It is wrong from the Lockean standpoint to force people against their will to advance other people's arbitrary ends, ends the coerced persons are not morally required to pursue anyway. The question then is, what ends are morally mandatory? The Lockean nonlibertarian holds that the promotion of justice for all—justice being identified with fulfillment of a Lockean set of rights—is an end that every person morally must adopt and pursue at least to a threshold reasonable extent. In like manner, the freedom to live as one chooses, within broad limits, is compatible with requiring individuals to comply with fairness norms in their dealings with one another.34

For a contrary view, see Loren Lomasky, “Nozick's Libertarian Utopia,” in Schmidtz, ed., Robert Nozick, 59–82. On p. 73, Lomasky writes, “Basic respect for autonomy enables pursuit of individual projects within a liberal order, but autonomy in any recognizable guise is the first casualty of a social arrangement in which everyone effectively owns a piece of everyone.” To my ear this claim sounds wildly exaggerated. Consider a regime in which minimally decent Samaritanism is enforced: everyone has a minimal property right in everyone else that entitles each to command the aid of another in just those circumstances when (a) the aid will save a life well worth living, or bring about some comparably great benefit for its recipient, and (b) the aid can be provided at minimal cost to the giver. (If one likes, one can add a third condition: (c) the potential recipient of aid is not culpably responsible, beyond a certain threshold magnitude, for her dire predicament.) Whatever may be said for or against such a regime, it does not destroy autonomy in any recognizable guise (or in its most sensible guises). Much the same should be said of a regime in which each person has the following limited property right in the bodies of other persons: no one is permitted to be a free rider on cooperative schemes that provide important public goods of justice.

However, it should be noted that a problem is lurking in this neighborhood, one I shall not here attempt to resolve. Suppose one denies a duty to carry out easy rescues. Perhaps this is the point at which core self-ownership is thought to bite. With this denial in place, insistence on fairness conceived as ruling out charging whatever the traffic will bear in thin market settings will then be inconsistent with the Pareto constraint. To see this, note that A might be willing to rescue, but only if she profits by charging B his reservation price for this service, and B would prefer to pay this price rather than perish.

Even moderation (as introduced in Section XII), the norm that of those considered in this essay most radically alters the shape of natural rights theory, is strictly consistent with the core idea of self-ownership as I conceive it. The mere fact that a person is in need does not suffice to give the needy person a right to be assisted by anyone else, nor does it suffice to generate an enforceable duty to aid that falls on anyone else. But if someone is needy in the sense of being threatened with a loss that is excessive in relation to the cost that would have to be incurred by another individual who could avert the loss at least cost to herself, and if the threatened loss itself exceeds some threshold magnitude, then the moderation norm holds that a right to be aided and a corresponding duty to aid do arise.35

There are two requirements here: the ratio between the benefit to the potential recipient of aid and its cost to the potential supplier must be sufficiently favorable, and the net magnitude of gain the recipient of aid would receive must be sufficiently large. One also wants the notion of “the person positioned to be able to supply aid at least cost to herself” to be sensitive to gains and losses over time. If Mother Teresa has already made huge sacrifices of her self-interest for the truly needy, whereas I have to this point in my life made nil sacrifices of this kind, and right now Mother Teresa could save a life at a cost of $10, whereas I could save the same life at the cost of $10.50 (suppose money here is linear with interpersonally comparable welfare), we want the moderate principle of Good Samaritanism to require me, not Mother Teresa, to provide the aid if anyone should in this instance.

XIV. The Principles Considered

Moderation expresses the idea that people are inviolable, up to a point. Individuals have rights that constrain what others may do to them, but the constraints give way when too much evil will ensue if those rights are upheld.

The other revisionary principles that I have invoked to support the judgments about puzzle cases that seem compelling modify the idea of inviolability, or individuals' fundamental entitlements. These principles can all be regarded as partial explications of the idea of fair dealing. It is not fair to begrudge anyone a benefit when allowing that whatever brings the benefit is costless to all others (including oneself). It is not fair to hold out for the maximum possible benefit for oneself when circumstances of thin markets give one a bargaining edge over those with whom one might contract on mutually beneficial terms. It is never fair to act maliciously toward others, aiming at nothing except harm or evil for them. There is also fairness toward oneself. One may not act with malice toward oneself. One has a duty to make something good of one's life—utterly squandering it is unfair to oneself. It is not fair to free ride on the cooperative behavior of others that provides one the public goods associated with secure enjoyment of the system of rights. A closely related point is that it is not fair to shirk one's duty to assume one's reasonable share of the costs of promoting justice (the system of natural rights, modified by these fairness requirements and by the principle of moderation itself). What is unfair in these ways one does not have a moral right to do.

These revisions to Lockean libertarianism all raise the Goldilocks problem.36

In the fable “Goldilocks and the Three Bears,” Goldilocks wanders into the bears' house, encounters three chairs, three beds, and three bowls of soup with different features, and for each of the triples ponders which one is “just right.” Given multiple moral values, each one varying by degree, a moral theory sets their relative value and so determines, for any states of affairs that combine the values in different combinations, which state of affairs is best, which second-best, and so on.

In each case, fairness is identified with a proper or appropriate or reasonable weighing of conflicting principles, but nothing has been said in this essay to identify these optimal balance points. More needs to be said here. A sketch is not a moral theory.

“Fairness” here is just a term that encompasses various moral considerations. Ideally, one would like to have a theory of fairness rather than a collection of intuitions. But lack of a theory does not by itself impugn the intuitions about puzzle cases.

A serious concern that would need to be addressed in a satisfactory treatment is whether or not the revisions proposed in the libertarian position can coalesce into a stable doctrine. Is soft Lockeanism an unstable compromise of what further investigation would reveal to be incompatible elements? Can side constraints and entitlements cohere with outcome-oriented moral concerns? A full answer to these questions would require the development of a complete moral theory. But pending the construction of such a theory, I do not see a deep problem here. “People are inviolable, up to a point” sounds paradoxical, but the underlying idea is coherent: “You may do whatever you choose so long as you do not harm others in ways that violate their negative rights, unless (a) doing what you choose while violating no such right would allow excessively large losses to others' interests, in which case you must cater to their interests, or (b) failing to violate a right of another would lead to excessively large losses to others' interests, in which case you must violate the right.” This is a consequence-constrained side constraint view.

XV. Conclusion

Throughout this essay, I have merely described examples and noted my own responses that take the form of judging the limits of people's rights in the situations as characterized. I have not advanced any argument supporting these responses except to note their inherent plausibility. The way is thus open for the hard libertarian to stand fast by her position and budge not. Against the assertion that the responses that contradict hard libertarianism are plausible, the hard libertarian can respond, “So what?” However, what Nozick says about the Randian position is relevant here:

A large part of the attraction of the Randian view for people is the way it handles particular cases, the kind of considerations it brings to bear, its “sense of life.” For many, the first time they encounter a libertarian view saying that a rational life (with individual rights) is possible and justified is in the writings of Miss Rand, and their finding such a view attractive, right, etc., can easily lead them to think that the particular arguments Miss Rand offers for the view are conclusive or adequate. Here it is not the argument which has led them to accept the view, but rather the way the view codifies, integrates, unifies, extends many of the judgments they want to make, feel are right, and supports their aspirations. If this is so, then one should hold the view so that it is open to challenge from just the sort of data that has provided its main support.37

Robert Nozick, “On the Randian Argument,” reprinted in Nozick, Socratic Puzzles, 249–64; see footnote 16 on p. 386.

I agree, and of course Nozick's own positive doctrine must answer to examples in just this same way. Libertarianism is open to challenge from the responses to cases that seem most compelling to us after scrutiny and reflection. On this basis my provisional conclusion is, “Down with hard libertarianism, up with soft Lockeanism.”38

“Up with soft Lockeanism” means that this doctrine deserves to be regarded as a serious rival of the most plausible consequentialist morality. My own hunch is that consequentialism ultimately wins this competition.

Footnotes

I thank Ellen Frankel Paul for helpful, constructive, and substantive comments on a prior draft of this essay. It goes without saying that her comments outstripped my ability to respond.