The Background, Failure, and Aftermath of Nozick’s Attempt to Justify the State

A state is any organization that claims and exercises a monopoly on central uses of coercive force over a broad geographic area.^[1] The state is a modern invention. For most of human history, there was no state, no government of any kind. Today, states are nearly ubiquitous. Very few pieces of dry land on Earth are unclaimed by any state.^[2] The state now seems almost as inescapable to us as water to a fish.

Perhaps for this reason, the question of the state’s justification as such is very far from mainstream concerns. The question of whether any state is ever justified is raised only by those interested in abstract political philosophy.

Different states claim different powers and issue and enforce different laws. Some of these states are unjustified—we all agree on that. Suppose that a given state is justified. It obviously does not follow from this that its powers are unlimited. It does not even follow that citizens in its territory have any special political obligation to obey the laws that it issues and enforces—obligations arising from the political situation beyond any moral and practical obligation related to the content of the laws themselves.^[3]

The fundamental question of political philosophy is whether any state is justified. It is the question of whether a state can exist without thereby being immoral in some fashion. By far the most impressive and influential modern attempt to justify the state is in the first part of Robert Nozick’s (1974) Anarchy, State, and Utopia. There, Nozick presents a complicated argument attempting to justify the minimal (night watchman) state. He does not think that any state with powers beyond the minimal state can be justified.

Nozick’s book is one of the most widely read and discussed works of contemporary political philosophy, probably second only to John Rawls’s (1971) A Theory of Justice.^[4] Despite this, the philosophical literature on Nozick’s book has focused mainly on the second part, in which Nozick critically discusses the theories of Rawls and others who argue for a more-than-minimal state.^[5] There’s an obvious reason for this. Most political philosophers—and most philosophers—think it is obvious that even states far more extensive than Nozick’s minimal state are justified.

In sharp contrast, from the start, Nozick’s attempt to justify the state received significant attention in libertarian circles. In fact, in many respects, Nozick’s discussion grew out of an in-progress debate between minarchism and anarchism in libertarian political circles. I will explain this background history before analyzing Nozick’s argument and arguing that it has two possible readings and fails on both. I will also argue that Nozick’s failure casts doubt on the compatibility of natural rights and the state, and that a clear understanding of his attempt and its failure also sheds light on Nozick’s later (apparent?) apostasy from libertarianism.

The Historical Background to Anarchy, State, and Utopia

The crucial intellectual background to Nozick’s attempt to justify the state is not found in academic political philosophy but rather in libertarian political debates. To fully understand Nozick’s arguments, we need to understand this background.

One of the largest intellectual influences on broadly libertarian circles in the 1960s was the novelist Ayn Rand.^[6] Through her novels, assorted later writings, and public lectures by her and her protégés—many organized through the then-popular Natheniel Branden Institute from 1958 until 1968—Rand advocated for nonaggression, individualism, and a minimal state.

Concurrently, there was growing anarchist opposition to Rand’s “minarchist” position. This was a kind of right-wing anarchism that focused on individualism and property rights, not the more traditional left-wing, collectivist form of anarchism. The major figure among these “free-market anarchists” or “anarcho-capitalists” was economist Murray Rothbard.^[7] During the 1950s, an informal libertarian group—the Circle Bastiat—formed around Rothbard. This group interacted with and sometimes feuded with the informal group around Rand—the Collective. Eventually, Rothbard and other Circle members were excluded from interactions with the increasingly cultlike Collective for being insufficiently deferential to Rand.^[8]

Others were coming to anarchist conclusions at around the same time, including the young Roy A. Childs Jr., who would become a central figure in libertarian media until his early death in 1992. In 1969, the then twenty-year-old Childs wrote an open letter to Ayn Rand, boldly beginning, “Dear Miss Rand: The purpose of this letter is to convert you to free market anarchism” (Childs 1994, 145). Childs’s central charge was direct: The existence of the state is incompatible with Rand’s moral prohibition on the initiation of force.

The argument is simple but compelling. Suppose that a citizen judges it best to use a service that competes with the state. Perhaps this citizen even sets the service up himself. What is Rand’s minimal state to do? If the state uses its might to shut down the service and enforce its monopoly, then it has initiated force and is thus immoral and unjustified. If the state does not use aggression to shut down the service but instead allows the alternative service to exist and allows its citizens to patronize it, then it is not a state at all because it has foregone its monopoly and is simply competing with alternatives in a free market. Therefore, any supposed state is either unjust or not really a state at all.

Even stated briskly, the tension is very difficult to deny. Rand’s response was characteristic: She unsubscribed Childs from her newsletter without further comment.^[9] While the tone of Childs’s letter was a bit much, his arguments warranted reflection. Several people who would later become influential free-market anarchists—apparently including George H. Smith—were originally won over to the anarchist cause by Childs’s letter.^[10] Rand’s closest followers were not convinced, but anarchism was becoming more and more widely accepted in the nascent libertarian movement.^[11]

According to the anarchist position of Rothbard and Childs, in the anarchist state of nature the various services legitimately provided by the state are instead provided by the free market. In this way, the coercive monopoly of the state is avoided. Childs lays out the situation in his open letter: “We advocate competing agencies of protection, defense, and retaliation; in short, we claim that the free market can supply all of man’s needs—including the protection and defense of his values. . . . We merely recognize protection, defense, and retaliation for what they are: namely, scarce services, which, because they are scarce, can be offered on a market at a price. We see it as immoral to initiate force against another to prevent him from patronizing his own court system, etc.” (Childs 1994, 152). Anyone familiar with Anarchy, State, and Utopia will immediately recognize the connection between this and Nozick’s discussion in part 1.

Influential anarchist libertarian books began to appear shortly before Nozick’s book was published. In fiction, 1966 saw the appearance of The Moon Is a Harsh Mistress, Robert A. Heinlein’s science fiction tale of libertarian revolution on a lunar colony.^[12] In nonfiction, the most notable works were The Market for Liberty, by Linda and Morris Tannehill (1970), and The Machinery of Freedom, by David Friedman (1973). In the early 1970s, anarchism was in the (libertarian) air. But Nozick rarely flags the connection between his discussion and then-raging minarchist-versus-anarchist debates. The central indication of it comes at the end of the acknowledgments: “It was a long conversation about six years ago with Murray Rothbard that stimulated my interest in individualist anarchist theory. Even longer ago, arguments with Bruce Goldberg led me to take libertarian views seriously enough to want to refute them, and so to pursue the subject further. The result is before the reader” (Nozick 1974, xxv–xxvi). In addition to having been a graduate student in Princeton’s philosophy program with Nozick, Goldberg was also a member of Rothbard’s Circle Bastiat. From all available information, his intellectual influence on Nozick was significant (see Raico 2002).

With this background, it is unsurprising that Nozick starts from assumptions shared by Rothbard, Childs, and other free-market anarchists of the early 1970s. The precise logic of his strategy will be discussed in the next section, but it is already clear that Nozick was attempting to provide the detailed minarchist response to the anarchist challenge that Rand and her followers were unwilling or unable to provide.

In stark contrast, Nozick’s starting assumptions were not shared by most academic political philosophers. Thomas Nagel (1975) famously described Nozick’s approach as “libertarianism without foundations” to highlight the fact that Nozick did not start from first principles to develop a theory of natural rights. At first glance, this charge is unfair for two reasons. First, Nozick does say several things that illuminate the natural rights he begins with (see Bader 2010). Second, this criticism holds Nozick to an unreasonable standard that orthodox political philosophers are themselves (rightly) never held to.^[13]

Nozick is attempting to justify the state while admitting natural rights. Moreover, he is obviously attempting to provide a detailed response to the claims by Rothbard, Childs, and other anarchists that natural rights are inherently incompatible with the existence of the state. It is no wonder, then, that orthodox political philosophers were often puzzled (and troubled) by the first part of Nozick’s book, while libertarian anarchists immediately recognized what Nozick was attempting. In 1977, the first ever issue of the Journal of Libertarian Studies featured replies to Nozick from many prominent libertarian anarchists, including Rothbard, Childs, and Randy Barnett.

The dependence of Nozick’s work on the libertarian thinking of the time has been recognized before. Witness Jeffrey Paul’s (1981, 4–5) comment in his introduction to the collection Reading Nozick: “His immediate predecessors in libertarian social philosophy established the theoretical agenda for Nozick’s own work. The American economist Murray Rothbard, for example, has in the twentieth century revived Benjamin Tucker’s and Lysander Spooner’s arguments for individualist anarchism. This American triumvirate is Nozick’s invisible collective antagonist as he strives in the first part of Anarchy, State, and Utopia to establish the propriety of the minimal state and its desirability over its anarchist competitors.”^[14] As Paul noted, these and other antagonists are largely “invisible” in Nozick’s main text—hence the need for the historical context offered by this section. With this background now in place, let us, like the anarchists of a half century ago, start evaluating Nozick’s bold attempt to refute anarchism.

The Logic of Nozick’s Strategy

Nozick’s strategy is to start from a state-of-nature (anarchic) situation and describe a process by which a minimal state could—or would?—come to be without any rights violations. But it is unclear exactly what a just-so story like this is supposed to prove.

The overall logic of Nozick’s strategy is one of the most puzzling aspects of his attempt to justify the state. Nozick’s (1974, 5) explicit statements about his methodology are extremely brief and unsatisfying: “If one could show that the state would be superior even to this most favored situation of anarchy, the best that could realistically be hoped for, or would arise by a process involving no morally impermissible steps, or would be an improvement if it arose, this would provide a rationale for the state’s existence; it would justify the state.” There are three disjuncts here, and it seems that Nozick thinks that any of the three would suffice to justify the state in some sense. The disjunct that Nozick seems to opt for in Anarchy, State, and Utopia is the middle one, showing that, from a state of anarchy, the state would arise without any moral violations—in particular, without any individual’s rights being violated.^[15]

I say “seems to opt for” because Nozick’s story is extremely complex, and when developing it, he isn’t always explicit about what he is claiming. At the very least, he describes a process by which, he argues, a state could arise from anarchy by way of an invisible-hand process. But could is weaker than would, and the idea that Nozick was merely arguing that it is possible for a state to be justified must be rejected.^[16]

To see why, suppose that Nozick had been aiming to refute only the following position:

Strong anarchism: Necessarily, the state is unjustified—that is, in any possible situation the state in said situation is unjustified.

This is quite easy to refute, given the following principle:

Consent justification: Necessarily, an individual’s clearheaded and informed consent to the operations of an institution justify the operations of said institution with respect to the consenting individual.

The idea of clearheaded and informed consent could stand some elaboration, and there may be cases where it is indeterminate whether such consent has been given.^[17] What is more, there is room for debate about whether the operations justified by consent are restricted—can one consent to being tortured or to being a slave? With these points noted, some version of this principle is both extremely plausible and is accepted by both Nozick and his anarchist opponents.^[18]

The essential point here is that we can sanction what would otherwise be violations of our rights—moral boundary crossings—with clearheaded and informed consent. This will be discussed further below, but, modulo the murky issues already noted, it is something that virtually all natural-rights theorists accept.

Now, consider a situation where all adults in a geographic area of claimed state control offer clearheaded and informed consent to the state’s operations. By consent justification, the state is justified in this situation. And since this situation is obviously possible in every relevant sense, strong anarchism has been refuted. Voilà. Notice too that this same argument will also justify a state that is far more than minimal unless severe constraints are added to the consent justification principle.

If Nozick’s aim was only to refute strong anarchism, he didn’t need the complex just-so story he lays out in his book. So, we might argue, appealing to a form of charity, that Nozick was not merely trying to show that it is possible for a state to be justified but perhaps had a different modal conclusion in his sights. For instance, he might instead have been targeting

Moderate anarchism: Necessarily, a state is unjustified in any possible situation in which relevant consent to that state’s operation is not universal in its geographic area.

Consent is relevant when, absent the consent, the state’s operations would involve rights violations. There is obviously no argument against moderate anarchism from consent justification and universal consent hypotheticals. To argue against moderate anarchism instead requires a hypothetical where the state is justified without there being universal consent in the hypothetical situation. And unlike strong anarchism, moderate anarchism is a position that real-world anarchists—including the libertarian anarchists in Nozick’s orbit—actually accepted.

Refuting moderate anarchism would be a significant accomplishment, showing that at least some states are not inherently unjustified absent universal consent. But although significant, it would still be a merely modal result, establishing only that there are possible situations in which a minimal state is justified, not that any actual state—even if minimal—is justified. That is, it does not establish any version of

Archism: A state of kind K is justified in our actual situation.

Moreover, read most naturally, Nozick’s strategy could at best establish only that a minimal state would be justified if it arose without any morally impermissible steps. But this does not even suffice to establish that a minimal state would be justified after a minarchist takeover. This problem led Murray Rothbard (1977) to dismiss the relevance of Nozick’s argument and to quip that, by his own lights, Nozick should argue for anarchy since anarchy is apparently a required preliminary situation for any justified minimal state.

Given the frustrating lack of details Nozick provides about the overall logic of his strategy, this reaction is understandable. However, as the preceding Nozick quote suggests, Rothbard has the counterfactual portion of Nozick’s strategy wrong. Nozick implicitly accepts a would-to-actuality principle that I make explicit as follows:

Counterfactual justification: If a state would have arisen out of the state of nature without any morally impermissible steps, then that state is justified in our actual situation.

This principle connects Nozick’s hypothetical reasoning to actuality. Nozick himself does not formulate this principle or anything like it, outside of the disjunct in the passage quoted above. The counterfactual logic of Nozick’s strategy was first clearly highlighted by Ralf Bader (2017).

Bader argues that Nozick was trying to determine what would have happened in the nearest possible world without any morally impermissible steps—that is, without any rights violations. This is impossible to determine in precise detail, but Bader argues that, with respect to the state, all that is relevant is which institutions would have arisen in the nearest counterfactual world without rights violations. The institutions that would have come about if no rights had been violated are actually justified—that is, justified in reality, not just in the counterfactual situation, by something like counterfactual justification.^[19]

Bader believes that this explains the invisible-hand approach Nozick takes as well as why Nozick privileges the minimal state—both moves are to ensure that the described situation is as close to reality as possible. Worlds with more overt mechanisms, or with more extensive states, are not as similar to reality, in the relevant sense, as worlds where a minimal state arises through an invisible-hand process.

This connects Nozick’s justification of the state to the principles of rectification in his well-known (perhaps infamous) entitlement-based theory of justice. Principles of rectification concern what is to be done when rights are violated. All of us, not just libertarians, recognize that what would have happened absent rights violations is relevant to this issue. If you steal my car, rectifying things would presumably involve returning the car to me, if possible. This is because absent your violation of my property rights via theft, the car would have remained in my possession.

Institutional injustices are rectified when our actual institutions match the institutions that would have arisen in the situation most similar to reality in which no rights were violated. Thus, in arguing about what would have happened absent rights violations, Nozick is also saying something relevant to what should happen, in actuality, in terms of state institutions. There are some technical questions about the precise implementation of this strategy, but here they can be left to one side.^[20]

Given all this, there are two reasonable readings of Nozick’s argument for the state:

The weak reading: Nozick’s goal is to refute moderate anarchism by showing that a minimal state could be justified without universal consent.

The strong reading: Nozick’s goal is to establish minimal state archism (archism for a minimal state) using counterfactual justification by showing that a minimal state would have arisen out of the state of nature without any morally impermissible steps.

Given Nozick’s lack of explicitness, I am not sure that he was clearly opting for one over the other. Ultimately, it does not matter which of these strategies, if either, Nozick himself intended. What matters is only whether his attempt to justify the state succeeds on either reading.

Nozick’s Attempt to Justify the State

With its argumentative logic now clarified, let’s go over Nozick’s hypothetical story. He starts by assuming a state of nature. Despite this starting point, Nozick’s approach is far from common Hobbesian practical arguments against the desirability of the state of nature. This is likely partially because libertarian anarchists had already convinced Nozick that the case against anarchy was not that easy.

From this starting position, Nozick attempts to show that the minimal state would—and therefore could—come to be without any rights violations, even in the absence of universal consent. This hypothetical progression from anarchy to the minimal state can be usefully broken into three distinct stages.

Stage 1: From the state of nature, various general protection agencies form that sell rights protection and enforcement to clients.

Stage 2: In a given geographic area, either one of the protection agencies formed at stage 1 eventually becomes dominant by successfully out-battling the other protection agencies or an allied federation of protection agencies emerges, turning the allied agencies into a de facto dominant agency. Either case effectively results in a single dominant agency in each given geographic area. These two cases can be treated in the same manner, and Nozick’s terminology freely switches between dominant agency and dominant association throughout his discussion.

Stage 3: The dominant agency in an area shields its clients from risky rights enforcement procedures, thus denying nonclients the ability to freely choose their own enforcement procedures in disputes with clients of the dominant agency. To avoid rights violations, the dominant agency also compensates these nonclients for this restriction on their enforcement options by offering the agency’s services to them, possibly at a reduced rate.

Through stage 2 and the start of stage 3, we have what Nozick calls an ultraminimal state—a powerful organization that does not approach universal coverage for all citizen candidates in its geographic area. After stage 3, the dominant agency will have moved from this to a minimal state by extending its protections to some nonclients—that is, to some potential clients in its geographic area who do not initially consent to becoming its clients.^[21]

This would extend the dominant agency’s coverage beyond its client base, but the coverage does not extend to all disputing parties in its geographic area. Nozick (1974, 109–110) argues that this does not disqualify the dominant agency from being a state: “The dominant protective agency’s domain does not extend to quarrels of nonclients among themselves. . . . This does not show that the dominant protective association is not a state. A state, too, could abstain from disputes where all concerned parties chose to opt out of the state’s apparatus. . . . And shouldn’t (and mustn’t) each state allow that option to its citizens?” Despite this claim, we might think this opens the door to the counterargument that Nozick’s minimal state is not actually a state, since the state’s claimed monopoly on force might, pace Nozick, require interference in at least some violent disputes between nonclients. The problem becomes even more acute when it is noticed that Nozick, at least in passing, suggests that even nonclients in disputes with clients might refuse and thus be denied the dominant agency’s services. These points are worth keeping in mind, but Nozick’s position is broadly in keeping with standard definitions of the state, so let us grant it for the sake of argument.

Nozick (1974, 108) ultimately argues that after stage 3, the dominant agency exercises something close to a de facto monopoly on force in its geographic area: “Since the dominant protective association judges its own procedures to be both reliable and fair, and believes this to be generally known, it will not allow anyone to defend against them; that is, it will punish anyone who does so. The dominant protective association will act freely on its own understanding of the situation, whereas no one else will be able to do so with impunity. Although no monopoly is claimed, the dominant agency does occupy a unique position by virtue of its power. It, and it alone, enforces prohibitions on others’ procedures of justice, as it sees fit.” If all this is correct, then even the prima facie persuasiveness of Nozick’s attempt hangs on whether all his steps are morally permissible.

Given the ethical assumptions shared by almost all libertarians, including both Nozick and his anarchist opponents, this reduces to the question of whether anyone’s natural rights are violated at any of Nozick’s three stages. The main potential weak point comes at stage 3. In prohibiting nonclients from choosing their own enforcement procedures in disputes with clients, the dominant agency seems to be flagrantly violating their rights.

Nozick is aware that this is the crucial point in his argument, so he spends virtually the entirety of chapter 4 discussing the issue and the entirety of chapter 5 applying his conclusions. He argues for principles of risk and compensation that together ensure that no rights are violated through stage 3. He believes that issues of risk (even including fear of risky activity) need special treatment in the natural-rights tradition: “Actions that risk crossing another’s boundary pose serious problems for a natural-rights position” (Nozick 1974, 74). He ultimately argues on several related grounds—one based in the claim that we have procedural rights not to subject ourselves to unreliable arbitration procedures. The clients of the dominant agency have transferred these procedural rights to the dominant agency so that the dominant agency can protect its clients’ rights by shielding them from risky procedures.

His discussion does not dwell on it, but Nozick requires something stronger than this. The dominant agency has the right to protect its clients from procedures that are believed to be risky or even not known not to be risky. In other words, Nozick must apply protection of these transferred rights in an internalist rather than an externalist fashion, or else his hypothetical story would stall on which procedures were actually risky.

If things were left there, protecting the procedural rights of clients would require violating the rights of nonclients, and we would simply have an unavoidable conflict of rights. But Nozick argues that in certain circumstances, compensation for a loss serves not to simply redress what would otherwise be a rights violation but to make it the case that no rights violation has taken place. As you might suspect, this is the crucial point in the argument.

Nozick (1974, 57) clarifies the crucial notion of compensation: “Something fully compensates a person for a loss if and only if it makes him no worse off than he would otherwise have been.” This requires that there be some facts, and some way of accessing said facts, about how “worse off” a person is made by some interaction. As we’ll see, this is another point that many anarchists—particularly those influenced by Austrian economic theory—are skeptical about.

Nozick (1974, 87) explicitly admits that he has no exact statement to offer of the principles he is appealing to at this crucial step in his justification of the state: “Nor need we state the principle exactly. We need only claim the correctness of some principles, such as the principle of compensation, requiring those imposing a prohibition on risky activities to compensate those disadvantaged through having those risky activities prohibited to them. I am not completely comfortable presenting and later using a principle whose details have not been worked out fully, even though the undeveloped aspects of the principle do not appear to be relevant to the issues upon which we shall wield it.” Sometimes, those who wish to impose risky procedures, but are prohibited from doing so, are disadvantaged by the prohibition. If they are compensated for these disadvantages, then there has been no rights violation by the prohibiting party. Nozick (1974, 102–3) thinks this leads to the following situation: “This dominant protective association will prohibit anyone from applying to its members any procedure about which insufficient information is available as to its reliability and fairness. It also will prohibit anyone from applying to its members an unreliable or unfair procedure; which means, since they are applying the principle and have the muscle to do so, that others are prohibited from applying to the protective association’s members any procedure the protective association deems unfair or unreliable.” The crucial question at this point is what compensation is offered to these disadvantaged parties.

Nozick’s (1974, 112–13) answer—the answer that moves us from the ultraminimal state to the minimal state—is that these parties are compensated with an offer of the protection agency’s own services, in some cases at a below-market rate: “Thus the dominant protective agency must supply the independents—that is, everyone it prohibits from self-help enforcement against its clients on the grounds that their procedures of enforcement are unreliable or unfair—with protective services against its clients; it may have to provide some persons services for a fee that is less than the price of these services.” In this way, the ultraminimal state extends to a minimal state by providing services to nonclients in its geographic area. As I mentioned in passing, Nozick briefly allows that some may refuse this compensation. He seems to believe that, in this situation, even a refused offer of compensation for a loss means there was no rights violation.

If all this works—a big if!—Nozick has taken us from a state of nature to the state with no morally impermissible steps. We have an organization that exercises something close to a de facto monopoly on force within its geographic area, extending its services even to many who did not initially consent to receiving them.

Given the conclusions of the previous section, there are two relevant ways to understand each of Nozick’s stages. According to the weak reading, we need only that each stage could happen. According to the strong reading, we need that each stage would happen. If each of Nozick’s stages can be defended on either of these readings, we have a significant antianarchist conclusion. And if it can be defended under the strong reading, we would have a legitimate justification of the (minimal) state.

Problems with Nozick’s Attempt

I have broken Nozick’s strategy into three stages with two distinct readings of each stage. Let us go stage by stage and reading by reading. When we do, we will see that the challenges are—to understate things—severe.

Stage 1 is fairly solid on the weak reading—things could have developed in that direction. But stage 1 is highly questionable on the strong reading since it requires that things would have developed in this direction in the most similar nonactual situation without rights violations. Some have countered that nothing like Nozick’s one-stop-shopping protection agencies would form from the state of nature.^[22] It seems much more plausible that various more specialized protection services would be offered and that protection packages would be bundled together, which is what we actually see in private security and other markets for protection and insurance.

Stage 2 also seems solid on the weak reading. A single protection agency could come to dominate in each geographic area, and this does not seem to inherently involve rights violations.^[23] But again, on the strong reading, this step is doubtful. Murray Rothbard (1977, 48) comments: “The Nozick contention that a dominant agency would develop in each geographical are, then, is an example of an illegitimate a priori attempt to decide what the free market would do, and it is an attempt that flies in the face of concrete historical and institutional knowledge. Certainly, a dominant agency could conceivably emerge in a particular geographic area, but it is not very likely.” Rather than a single dominant agency arising, various stalemates and stable markets would be more likely.^[24] What is worse, Nozick’s claim that a federation or agreement of various agencies to cooperate counts as a monopoly is simply an abuse of language. As Childs (1977) quipped, we might as well say that all farmers in the world together have a monopoly on food production—this is simply not what the word monopoly means.

Stages 1 and 2 of Nozick’s attempt succeed on the weak reading but not on the strong reading, which already means that Nozick does not succeed in establishing minimal state anarchism. Whether he succeeds in refuting even moderate anarchism will depend on the success of stage 3 under the weak reading.

Stage 3 is by far the most philosophically involved of Nozick’s stages. I will argue that it fails even on the weak reading. The central problem is that Nozick’s risk and compensation principles are highly dubious. These two principles are at the very heart of Nozick’s strategy, and chapter 4—the most analytically intricate chapter in the book—is entirely devoted to untangling some of the background to these principles. Unsurprisingly, these discussions have drawn the most fire from anarchist critics of Anarchy, State, and Utopia.

Especially worrying is that Nozick’s principles allow for prohibiting risky procedures on the basis of fear. Murray Rothbard (1977, 49) comments: “Once one can use force against someone because of his ‘risky’ activities, then the sky is the limit, in short there is then virtually no limit to aggression against the rights of others. Once we permit someone’s ‘fear’ of the ‘risky’ activities of others to lead to coercive action, then any tyranny becomes justified, and Nozick’s ‘minimal’ state quickly becomes the ‘maximal’ State.” Nozick allows the dominant protection agency to shield its clients from being subject to procedures that it does not deem nonrisky, and in so doing he allows far too much.^[25]

There is also a sense in which he might allow far too little. For if a competing agency advances a procedure that the dominant agency deems nonrisky, even within Nozick’s framework, the dominant agency cannot shield its clients from those alternative procedures. This puts pressure on the existence of even a de facto monopoly on coercive force at stage 3. But let’s focus on Rothbard’s criticism since, if correct, it would show that, even if Nozick’s justification works, it allows far too much. Nozick allows coercive action when we fear risky activities. Rothbard and other more traditional natural-rights theorists believe that this would sanction almost anything.

How might Nozick respond to this charge? He might appeal to unavoidable uncertainty and claim that we are always merely weighing the probabilities. According to this way of thinking, when someone points a gun at you, they haven’t (yet) violated your rights. When you act to stop them, you are—he might claim—acting on a version of the risk principle. The idea behind this reply is for Nozick to argue that his differences with Rothbard are simply a matter of degree and that we all face the question of which risks we will allow ourselves to be subjected to.

Something like this is probably the best that can be said in favor of Nozick, but it is ultimately unconvincing. The first problem is that there is a real difference between certainty of risk and uncertainty of risk. Nozick explicitly allows the dominant agency to shield its client from procedures “about which insufficient information is available as to its reliability and fairness.” So, it is not just known risks that allow coercion. Nozick also allows coercion in cases where the risks are unknown. When we consider the wide latitude this would allow, it is difficult not to sympathize with Rothbard’s complaint.

The second problem is even more fundamental. Obviously, life involves unavoidable epistemic uncertainties. I am not completely certain that the gun the home-invading robber is pointing at me is real, that it is loaded even if real, or that the safety is off even if it is both real and loaded. The problem is that this kind of merely epistemic uncertainty does not speak in favor of Nozick’s exceptions for risk.

There remains a difference in kind between actions that can be reasonably judged as initiating boundary crossing (such as when the armed robber attempts to enter my home) and actions that raise the probability of a boundary being crossed (such as when a group of known robbers get together for pizza).^[26] This is a bright line that is of crucial importance, and Nozick has blurred it in an unacceptable way. The tightened-up version of Rothbard’s charge is that there is no principled way of keeping Nozick’s allowances for unconsented-to boundary crossings limited to the very few situations in which he himself allows them in Anarchy, State, and Utopia.

What I am implicitly suggesting here is that much of what is covered by a prohibition on risky behavior would be better covered by appealing to the folk notion of an aggressive action and counting the initiation of an aggressive action as thereby activating standard self-defense rights.^[27] Crucially, I don’t think that we should attempt to define aggressive actions using a threshold for the probability of boundary crossing.^[28]

We have to be careful with terminology here. According to the Nozick-inspired way of talking that I prefer, violence and coercion involve boundary crossing. Rights protect boundaries and rights violations all involve boundary crossings. This is not to say, however, that all boundary crossings are thereby rights violations. We could say that, but it would introduce awkwardness into this particular dialectic. It is better to say that clearheaded and informed consent can sanction boundary crossings and that consented-to boundary crossings are not rights violations. The remaining question left open by this terminology is whether all unconsented-to boundary crossings are rights violations. With a couple caveats that will be discussed below, Rothbard and I say yes, while Nozick says no.

Eric Mack has suggested that the debate between anarchism and minarchism can insightfully be described as a debate about when the state’s unavoidable coercion is rights-violating (pers. comm., July 5, 2024). This description is similar to my framing in terms of boundary crossings. The role of Nozick’s exception principles for compensated risk is in sanctioning certain instances of coercion—that is, unconsented-to boundary crossings.

I have argued that the risk principle is problematic, but the risk principle is intertwined with the compensation principle. Nozick needs his compensation principle to argue that the prohibitions the dominant protection agency imposes on nonclients are not ultimately rights violations. Yet not only is the compensation principle extremely problematic, but even if it could be maintained, it would not vindicate the compensation that the dominant protection agency offers. Both of these points have been made in the literature. Let us take each in turn.

The compensation principle faces problems of general motivation as well as problems of detailed formulation. The problems of general motivation turn on the fact that once he has endorsed the compensation principle, Nozick has effectively abandoned his starting idea of rights as side constraints. Here is one statement of this point from a libertarian anarchist: “Compensation is not a sufficient condition for justifying or permitting a violation of rights. Contrary to Nozick’s principle of compensation, all violations of rights should be prohibited. That’s what right means. The only way rights are abdicated is by consent of the rights holder” (Barnett 1977, 20).^[29] And here is one from a nonlibertarian (left-wing) anarchist:

In short, Nozick’s real problem is that given his extremely strong theory of individual rights, side constraints, and so forth, he ought in all consistency to come to the conclusion that no unconsented-to-boundary-crossings (i.e., rights violations) are permissible, regardless of compensation. But that is a crazy conclusion, as he realizes. If accepted, it would immobilize us all, making us much like a bizarre gathering of morally musclebound rights freaks, lovely to look at, but unable to lift a finger for fear of encroaching on one another’s moral space. So Nozick compromises. Of course, once he starts, only his intuition, or the degree of his moral finickiness, tells him when and where to stop, and how much pay in order to achieve what a mathematically sophisticated Anglo-Saxon of the eighth century might have called a discounted wergelt raising the tribe to its previous indifference curve. (Wolff 1977, 17)

There is a sense in which this charge misreads Nozick. As I noted above, Nozick does not agree that unconsented-to boundary crossings are inherently rights violations. And his compensation principle emphatically does not entail that any apparent rights violation is washed away provided that compensation is offered after the fact.^[30]

The idea is instead that in certain situations where there may be a risk of boundary crossing, unconsented-to boundary crossings that mitigate those potential risks do not themselves count as rights violations if compensation is offered to those disadvantaged by the boundary crossing.

This clears Nozick on the letter of the charge but perhaps only on a technicality. This issue is crucial, so let me belabor it. Nozick’s starting picture is that rights are side constraints that place moral boundaries. Rights violations involve boundary crossings. The uncompromising position is that only a person’s consent morally allows the crossing of these boundaries:

Absolute rights: A person’s natural rights are moral boundaries such that any unconsented-to boundary crossing is morally wrong—only clearheaded and informed consent morally sanctions boundary crossing.

As stated, this is a bit too uncompromising, even for those who stress the crucial role of consent. I will note three qualifications accepted even by many natural-rights theorists, mainly to set them aside.

The first concerns situations where consent is simply not possible. Our future children cannot consent to being brought into existence. Despite this, it would be strange to argue that having children is morally wrong because it violates the rights of the child.^[31] Likewise, a man who passes out on an airplane did not give consent for the doctor in the seat next to his to touch him in the attempt to save his life. Yet he surely would have given consent, if he could have, at least in normal circumstances. These are the kinds of cases in which an appeal to merely hypothetical consent is reasonable.

The second qualification concerns cases of imminent catastrophe. Professor Evil designs a machine that will destroy reality at the press of a button. Can we allow him this power? Natural-rights theorists disagree on whether exceptions should be made for such cases. Nozick (1974, 30) explicitly punts on the issue in Anarchy, State, and Utopia: “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.”

The third, somewhat related, qualification concerns cases where a minor rights violation staves off a major rights violation. Perhaps when the axe murderer is chasing me, my only path to escape involves cutting slightly across your property line and thus violating your property rights.

I think that virtually every natural-rights theorist recognizes the need for qualifications of the first type, but there is disagreement on qualifications of the second and third types. Here, let us focus on a bounded principle of absolute rights, saying that within certain bounds rights are absolute, by making all three of these qualifications explicit:

Bounded absolute rights: A person’s natural rights are moral boundaries such that any unconsented-to boundary crossing is morally wrong if consent was possible, imminent moral catastrophe does not loom, and the boundary crossing is not done for the purpose of avoiding a different boundary crossing.

That is, what would otherwise be a morally objectionable unconsented-to boundary crossing is possibly sanctioned only by (1) hypothetical consent when explicit consent is impossible, (2) imminent moral catastrophe, and (3) avoidance of other unconsented-to boundary crossings. This is not a principle that I myself endorse, since I am skeptical of clause (3) and at least somewhat skeptical of clause (1).^[32] However, I include these clauses to exclude some irrelevant edge cases from the discussion to follow.

Nozick explicitly departs even from bounded absolute rights. As Wolff notes in the preceding quote, Nozick realizes that absolute rights, and even bounded absolute rights, make the state, by its nature, unjustified. So, he compromises, and this compromise leads him to something like:

Restricted rights: There are additional situations that morally sanction unconsented-to boundary crossings, beyond the exceptions set out in bounded absolute rights.

In short, Nozick allows additional exceptions to our natural rights. This is what is packaged together in his risk and compensation principles. I call them together exception principles since they offer an exception to our natural rights, even within the boundaries within which virtually all other natural-rights theorists see our rights as absolute.

For anarchists, this is the point at which the serpent enters Eden. When natural-rights libertarians claim that the state is inherently immoral, they are invoking at least bounded absolute rights, not restricted rights. And as the Barnett quote above illustrates, for many libertarians, the very meaning of rights requires invoking bounded absolute rights, and maybe even absolute rights, but certainly not restricted rights. Nozick opposes this trend, and once again the issue is not that he is committed to moral boundaries being overridden by broadly external factors, such as catastrophic moral consequences. It is instead internal factors that allow boundary crossings—that is, factors based on Nozick’s analysis of risk and compensation.^[33]

Nozick’s exception principles depart from the bulk of the natural-rights tradition by applying to activities that, though not aggressive actions in any sense, pose a risk of boundary crossing.^[34] In special circumstances such as these, Nozick claims that the boundary crossings do not violate rights if just compensation is offered. There is definitely tension here. By requiring compensation for those who are disadvantaged, Nozick seems to implicitly recognize that something morally questionable has happened. For all these reasons, his move to restricted rights marks a significant departure from the standard natural-rights tradition.^[35]

The problems with this are analogous to the problems with consequentialist accounts of rights. Once we morally sanction unconsented-to boundary crossings within the normal bounds where rights are typically seen as absolute, rights then play a different role in our ethical thinking. Such “rights” no longer put up moral boundaries that can only be crossed without wrongdoing when there is informed consent. Now, we and the dominant protection agency acting in our stead are allowed to coerce others in certain circumstances. Unlike Rothbard, Childs, Barnett, and most other individualist anarchists, Nozick rejects even bounded absolute rights in favor of restricted rights and thereby morally sanctions certain unconsented-to boundary crossings. This is necessary for any natural-rights justification of the state, since many of us do not consent to the activities of the state.

The end of Wolff’s quote from above brings us from these general problems to problems of detail. Nozick has to assume that there are facts about what counts as just compensation for a given boundary crossing. These are facts that are, by their nature, not determined by actual or possible free-market exchanges. The whole point is that there is no consent but there must still be some kind of just price for adequate compensation. As Wolff indicates above, Nozick appeals to the tool of indifference curves. Rothbard calls this “pretentious,” and indeed it is.^[36] Moreover, it is also potentially dubious. Barnett comments:

This “just price” fallacy permeates the whole of Nozick’s discussion of “compensation.” It confuses the morally permissible exchange for the penalty for violating a right which is compensation. If someone violates another’s rights, the victim is entitled to compensation to make up for the transgression. This simply means he is entitled to what was taken from him. We don’t pretend that money is the equivalent or even “fair price” for the loss of life or limb. We say only that some attempt must be made to restore to that victim what was taken from him as far as humanly possible. The crucial distinction here is while voluntarily paying a purchase price makes an exchange permissible, compensation does not make an aggression permissible or justified. (Barnett 1977, 20)

Let’s ignore Barnett’s apparent suggestion that Nozick always and everywhere allows compensation to smooth over what would otherwise be a rights violation. As I have stressed, he does not. He instead allows this only in certain situations—those involving certain kinds of unacceptable risk.

If the cleaned-up version of Barnett’s criticism works, it works even if we follow Nozick in accepting restricted rights. Thus refined, the criticism is simply that Nozick’s additional exception situations are ill-defined in requiring an incoherent and market-independent notion of a just price. There are various theories of value and price that might allow for a coherent definition of a just price absent free consent, but few if any of them are appealing to libertarians. Things are worse than that, for, as we saw above, Nozick counts offered compensation as just even when it is clearheadedly and explicitly refused. It is fair to say that, at the very least, Nozick’s principle of compensation is saddled with problematic theoretical baggage.

Finally, even if we assume that the idea of a just price not determined by the market is in good standing, the compensation that Nozick’s dominant protection agency offers is absurd. Roy Childs (1977, 27) comments: “What is he willing to offer us as compensation for being so prohibited? He is generous to a fault. He will give us nothing less than the State. Should one wish to reject this admittedly generous offer, it would be responded that he cannot reject it. It is foisted upon one whether one likes it or not, whether one is willing to accept the State as compensation or not.”

This is not strictly accurate—as noted above, Nozick seems, in passing, to allow that some nonclients may refuse the offer of reduced-rate services.^[37] Nozick is forced to say that when the state crosses your boundaries and makes you an offer that is in fact just compensation, then it has not violated your rights even if you disagree that the compensation is just and refuse the offer.

There are also practical considerations that threaten to undo the entire counterfactual supposition. What happens once the dominant agency offers its services at below-market rates to initial holdouts? Nozick briefly claims that this would not decrease the appeal of the dominant agency’s services, but this is doubtful. Nozick has, in effect, provided an incentive for holding out to get a lower rate. Presumably, this would make being an initial client less desirable. The costs incurred by offering holdouts services at a below-market rate might also require the agency to raise the price for clients. Clients might quit in protest or in an attempt to get the services at the nonclient rate. Or the raised price on clients might make room for successful market competition. All these practical points undermine at least the strong reading of stage 3 and possibly—depending on how powerful they ultimately are—the weak reading as well.

Stage 3 is the crucial stage in Nozick’s construction, but it is riddled with problems. It involves rejecting absolute rights and even bounded absolute rights in favor of restricted rights. It involves a version of the restricted rights principle that assumes the coherence of a nonmarket notion of a just price. It also involves an offer of compensation that is not just, even assuming restricted rights and the coherence of Nozick’s just price exemption principles—unless the just price is so secure that it can even trump an explicit lack of consent. And finally, it is practically unstable even assuming solutions to all the other problems we have canvassed.

To sum up, on the weak reading, Nozick’s approach succeeds at stages 1 and 2, but fails at stage 3. On the strong reading, his approach fails at every stage. Thus, Nozick has neither refuted moderate anarchism nor established minimal state anarchism. Moreover, because of the nature of the failure, especially at stage 3, there is no way to patch Nozick’s strategy to refute moderate anarchism, to say nothing of establishing minimal state anarchism.

This last point holds assuming, as most libertarian natural-rights theorists do, that rights are absolute, not restricted. If we accept absolute rights or even bounded absolute rights, then the state is unjustified. If a version of restricted rights is endorsed instead, we face different problems. There is a worry that Nozick’s exception situations are ad hoc and tailored to allow only a minimal state. Even when so tailored, they arguably allow much more, since it is hard to see how to keep the exception cases limited to those Nozick envisions. Along this path, we will find a Nozick-style justification for a far more than minimal state.^[38]

Even if all the other problems are set aside and ignored, the final stretch of Nozick’s path is a razor’s edge for libertarians. Fall off one side, and Nozick’s argument for the state completely breaks down. Fall off the other, and there is no way to keep the exceptions bottled up, and more and more of the state’s would-be rights violations get excused. I believe the logical situation here helps to answer a historical puzzle by shedding light on Nozick’s final remarks on political philosophy—remarks that apparently contain a somewhat puzzling recantation of his earlier libertarian position.

Nozick’s (Apparent?) Change of Mind

For many years after publishing Anarchy, State, and Utopia in 1974, Nozick left political philosophy alone and even tried to distance himself from it, writing in the introduction to 1997’s Socratic Puzzles that “it is disconcerting to be known primarily for an early work. Others have identified me as a ‘political philosopher,’ but I have never thought of myself in those terms. . . . I have not responded to the sizable literature on Anarchy, State, and Utopia, or followed it closely. I did not want to spend my life writing ‘The Son of Anarchy, State, and Utopia,’ ‘The Return of the Son of . . . ,’ etc.” (Nozick 1997, 1–2). It is hard not to sympathize with this.

In 1989, Nozick finally revisited political philosophy in print in an eleven-page chapter in The Examined Life titled “The Zigzag of Politics,” in which he explicitly abandons his earlier libertarian views. There is some general awareness of this in the philosophical community, but rarely is it acknowledged how completely he seemed to reject his earlier views.

Nozick’s (1989, 286–88) central statement of his new position is as follows:

Democratic institutions and the liberties coordinate with them are not simply effective means toward controlling the powers of government and directing these toward matters of joint concern; they themselves express and symbolize, in a pointed and official way, our equal human dignity, our autonomy and powers of self-direction. . . . That symbolism is important to us. Within the operation of democratic institutions, too, we want expressions of the values that concern us and bind us together. The libertarian position I once propounded now seems to me seriously inadequate, in part because it did not fully knit the humane considerations and joint cooperative activities it left room for more closely into its fabric. It neglected symbolic importance of an official political concern with issues or problems, as a way of marking their importance or urgency, and hence of expressing, intensifying, channeling, encouraging, and validating our private actions and concerns toward them. . . . The libertarian view looked solely at the purpose of government, not at its meaning; hence, it took an unduly narrow view of purpose, too.

The role of symbolic elements in rationality is something that Nozick later explored more fully in his 1993 book The Nature of Rationality. In a footnote appended to this passage, Nozick (1989, 287) specifies the connection to his earlier position: “In these remarks I do not mean to be working out an alternative theory to the one in Anarchy, State, and Utopia, or to be maintaining as much of that theory as possible consistent with the current material either; I am just indicating one major area—there may be others—where that theory went wrong.” The position outlined there is very difficult to make sense of, and there are few specifics in the chapter. At one point, Nozick (288–89) discusses the symbolic value of social programs in light of those who do not consent to funding them: “To be sure, this joint public affirmation is not simply verbal; those spoken for may have to pay taxes to help support the programs it involves. (That a fig leaf was created to cover the shame of their unconcern does not mean that they do not have to help pay for it.)” The symbolic value of certain joint public actions supposedly justifies using taxation—that is, for libertarians, theft and partial slavery—to fund them. Remember, this is Robert Nozick, supposed libertarian superhero, writing this. This is extremely jarring, since this kind of talk, though standard in progressive circles, is nearly unheard of among libertarians.

Nozick indicates some flexibility by making allowances for opting out of policies morally objected to. His example involves public money going to abortion. He notes that this leads to the problem—or potential problem—of too many people opting out, so he suggests a 5 percent tax bonus going toward other things could be paid by these holdouts—effectively, a monetary penalty for sincere moral objections! Nozick does not discuss this or the obvious voluntaryist rejoinder that if there are too many opt-outs to voluntarily pay for a policy, then the policy should go unimplemented. Later in this brief chapter, Nozick even comes out in favor of civil rights laws and restrictions on freedom of speech and association. What in the world happened here?

A crucial clue comes when Nozick (1989, 291) briefly discusses whether these policies involve rights violations: “It is not necessary to decide whether there is a right to discriminate that gets overridden when such discrimination is prevalent enough to constitute a significant burden to a group, or whether there is no such right yet some rare discriminations are too trivial in their effects to warrant a systematic legal interventions which too have their costs and effects.” Here, he is essentially admitting that if many different factors can override rights, then rights don’t matter at all—it isn’t even important to determine whether there is a right that is being overridden.

In effect, Nozick is endorsing one of the central criticisms of his attempt to justify the state—namely, that if rights can be overridden within standard bounds, then rights don’t matter at all. It doesn’t even matter if they exist. If we accept a Nozick-style principle of restricted rights instead of absolute rights or bounded absolute rights, then it is very difficult to support a non–ad hoc picture in which rights retain their central place in our moral theorizing.^[39]

If this is what you think about rights, then you are better described as rejecting rights altogether. But even those who think this claim is too strong must still admit that Nozick never endorsed the kind of absolute and exceptionless rights that most libertarians endorse, even within reasonably circumscribed limits.

In this section, I have quoted heavily from Nozick’s only late discussion of political philosophy. This was necessary in order to illustrate how far he diverged from his earlier libertarian position. But even in libertarian circles, Nozick’s recantation is often dismissed by citing late interviews in which he claimed to have moderated but not abandoned his libertarian views. The most well-known example comes from a 2001 interview with Julian Sanchez. When asked whether, in Nozick’s (2001a) book Invariances, he had gone back to applying the label of “libertarian” to himself, Nozick (2001b) replied, “Yes. But I never stopped self-applying. What I was really saying in The Examined Life was that I was no longer as hardcore a libertarian as I had been before. But the rumors of my deviation (or apostasy!) from libertarianism were much exaggerated.”^[40] Initially, this seems very difficult to credit, but I think we have the pieces in place to dissolve the tension between this claim and his apparent recantation.

After the influence of Goldberg, Nozick always considered himself to be a libertarian. He did so because respect for personal rights and liberty remained extremely important in his ethical and political thinking. Yet, even when he wrote Anarchy, State, and Utopia, he did not view rights in the way that most natural-rights theorists and most libertarian anarchists did. He instead accepted a version of the restricted rights principle, with—he believed—fairly constrained exception cases. But as his nonmainstream critics on both the Left and the Right were quick to point out, once you restrict rights in this fashion, it is hard to keep to a principled position.

Small differences in how we set the dials on acceptable risk and just compensation will lead to vast differences in what is morally allowed. And this sets up internal tensions between rights as boundaries and cost-benefit analysis. These internal tensions ultimately destabilize the entire enterprise. In the limit, rights become just one more thing thrown into a broader cost-benefit analysis—and then, as Rothbard and his compatriots warned, the sky’s the limit.^[41]

Seen from this perspective, the only real shift in Nozick’s thinking between Anarchy, State, and Utopia in 1974 and Invariances in 2001 concerned the extent of these exceptions and their ramifications for political philosophy. The early Nozick tried to argue that the exceptions allowed a minimal state but nothing else, while the later Nozick began to realize that his reasoning to these exceptions could not be so tightly contained. And, as his values shifted over the years, the porous boundaries imposed by so-called rights became a feature, not a bug. This is not a substantial change of mind but instead the chickens coming home to roost.

Whether the move from the Nozick of 1974 to the Nozick of 1989 and 2001 is seen as progress or regress depends on one’s point of view. From my perspective, we have a story of libertarian paradise lost. Nozick might have remained a libertarian the entire time, but if so, he moved from libertarian-minarchist adjacent to anarcho-capitalism to something more like a classical liberal who gives significant but limited weight to personal liberty.

Past, Present, and Future

Nozick’s reputation in political circles is divided. Among mainstream and predominantly left-wing academics, he is regarded as the libertarian thinker.^[42] But among actual libertarians, Nozick is not nearly as intellectually influential.^[43] Moreover, the intellectual movements that were and remain more influential among libertarians—Ayn Rand and the Objectivists, on one hand, and Murray Rothbard and the Austrians, on the other—were movements that Nozick had little sympathy for and criticized in print (see Nozick 1971; 1977).

Though Nozick frames his discussion in terms of rights, he does not endorse either of the popular proposed foundations for libertarian rights that crucially include property rights—namely, the self-ownership theory and the labor-mixing theory.^[44] These theories are usually developed to support some version of absolute rights or at least bounded absolute rights. Of course, there are serious questions about whether any such theory of rights can be successfully defended.^[45] There are also questions about whether any of these foundational views support property rights of the kind that libertarians see as foundational.^[46] I am not endorsing any version of these theories here but simply noting a crucial philosophical difference between Nozick and many other libertarian thinkers.

Obviously, Nozick’s failure to justify the state does not conclusively show that all related attempts will similarly fail, but the nature of his failure does suggest this. To block the natural-rights case against the state, rights must be weakened in some way. Statists are virtually forced to endorse some form of restricted rights. Nozick made a mighty attempt to make restricted rights palatable to traditional natural-rights theorists, but as we have seen, his weakenings were almost universally rejected by natural-rights theorists on plausible and principled grounds.^[47] Moreover, the early libertarian critics saw Nozick’s exception principles as already containing the seeds of a state far more than minimal. Thus planted, those seeds eventually grew into the zigzag of politics. Nozick’s beautiful gift to libertarians was, in the end, a Greek one after all.