Defending Liberty: Essays in Honor of David Gordon

CHAPTER 5

Epidemics as a Problem for Liberty

Lester H. Hunt and R. Kevin Hill

First, do no harm

—Attributed to Hippocrates

How should we respond to epidemics of serious disease?1 Some possible responses pose serious ethical problems for liberty, while others do not. Probably the ideal response to any such disease is to be found in pharmaceutical interventions (PIs): vaccines and other medicines that either reduce the likelihood that one will contract the disease or, once it is contracted, either cure it, alleviate its symptoms, or otherwise makes it less severe than it would have been. Though PIs can raise ethical issues (for instance, if a “live virus” vaccine actually causes people to suffer from the disease in a certain percentage of cases), these are not the sort of issues that will concern us here. Again, there are a range of non-pharmaceutical interventions (NPIs) that likewise do not raise problems of this sort. These include, rather obviously, individual, voluntary wearing of masks in order to avoid either becoming infected or infecting others and individual, voluntary social distancing practiced for the same reasons. These are obviously exercises of freedom, not infringements of it. The ethical problems we have in mind are raised by any action or policy that does appear to infringe liberty: since every such infringement requires a justification, those that cannot be adequately justified are unjust and morally wrong. Epidemic policies that raise this sort of issue include the entire range of coercive non-pharmaceutical interventions (CNPIs), such as: 1) required wearing of masks, 2) required social distancing, 3) isolation (that is, confining people who are infectious to a limited area, such as their homes) and quarantine (similarly restricting the movements of those who might have been exposed to the disease while waiting to see if they become sick), 4) erection of a cordon sanitaire (limiting entrance into or exit from a given geographical area, such as a city or an entire country), 5) various coercive measures designed to limit the size of crowds (such as requiring staggered hours of operation for certain businesses), 6) closing down some or all public transportation, 7) banning large meetings, with or without exceptions (for instance, for religious services), 8) closing schools, 9) closing all businesses and government offices except for ones declared to be “essential,” 10) requiring everyone to stay in their homes except for certain purposes (such as to do “essential” work, or to get medical help or food). Our question here will be: which CNPIs are justified and which are not? Obviously, we will not be able to discuss adequately all ten of the CNPIs we just described, but we do hope to sketch a set of principles that might plausibly distinguish between the justified and the unjustified.

We should emphasize that we will be addressing ethical issues only. We will not be commenting on which CNPIs are epidemiologically effective and which are not. Discussions of the opposite sort, where only epidemiological issues and considerations are raised, with no explicit acknowledgment of ethical ones, are common. This raises the possibility that some would object to our approach. “Isn’t it irresponsible,” they might ask, “to worry about ethical niceties when there is a deadly epidemic on the loose? Surely, when thousands, even millions of lives are at risk, only epidemiological considerations are relevant!” We think the purely epidemiological perspective that this objection seems to assume is one that normal human beings never actually employ. Consider, as evidence for this, the simple fact that, in recent debates about policy responses to the SARS-CoV-2 virus, no one has mentioned the following possible policy: simply round up those who are infected, execute them, and dispose of their bodies using some safe and sanitary method. Surely, it is very unlikely that anything like this even occurred to anyone while seriously thinking about policies that might be a good idea. Why not? Such a policy might kill more people than the disease would have killed, but that is not obvious, at least if the policy were carried out sufficiently early and quickly. More important, where does our revulsion at the thought of these surplus deaths come from? Since they are not caused by disease, epidemiology has absolutely nothing to say about them. Surely the source of this recoil of ours lies in ethical concerns about human life. The point of view that does not even consider such cruel policies is epidemiologically impure. It is, to some extent, a moral point of view.

So far, of course, this is consistent with the idea that, where many lives are at stake, the only ethically relevant consideration is the minimization of fatalities—all other ethical considerations are irrelevant and concern with them is frivolous and irresponsible. One version of this idea is the suggestion that the correct ethical view of epidemics is to apply what Ayn Rand called “the ethics of emergencies”2:

It is important to differentiate between the rules of conduct in an emergency situation and the rules of conduct in the normal conditions of human existence. … An emergency is an unchosen, unexpected event, limited in time, that creates conditions under which human survival is impossible—such as a flood, an earthquake, a fire, a shipwreck. In an emergency situation, men’s primary goal is to combat the disaster, escape the danger and restore normal conditions (to reach dry land, to put out the fire, etc.). … Men can live on land, but not in water or in a raging fire. Since men are not omnipotent, it is metaphysically possible for unforeseeable disasters to strike them, in which case their only task is to return to those conditions under which their lives can continue. By its nature, an emergency situation is temporary; if it were to last, men would perish.3

The idea here seems to be that if an individual, community, or the whole world falls into a situation in which humans cannot survive—her examples being finding oneself immersed in a body of water without a viable watercraft or in the midst of a raging fire—the sole imperative is to do what is needed to get out of that situation. Other principles and values may be set aside, if necessary.

This idea is certainly a defensible one. If the city is about to be destroyed by a catastrophic fire, we will dynamite blocks of buildings without asking the owners’ permission. If the Titanic is sinking and there are not enough lifeboats for everybody, we will not use the property rights of the owners as a guide to deciding who gets in and who does not: these rights would not even mean that the owners would be allowed to get in. But this idea, plausible as it is, does not apply to epidemics per se. To take the case that at the moment is closest to home: As we write, COVID-19 has spread around the world. This is a serious emergency, but it is not the sort of catastrophic emergency that Rand describes. It does not threaten the survival of the human race. The present emergency is not midnight on the Titanic. If there is a good argument for the notion that deadly epidemics, per se, justify setting aside all ethical considerations other than the preservation of life, this is not it.

Again, consider the following argument, which is related to the Rand-derived one but is more plausible as applied to the present crisis: “Whenever we have a chance to save a large number of lives, we should set aside ethical considerations other than the preservation of human life insofar as doing so will save lives. The most important right is the right to life. If we have to violate lesser rights in order to defend against violations of a greater one, that is justified.” This argument is also dubious, especially if “preservation of human life” means net preservation. Suppose we have captured a terrorist who has planted a ticking time bomb somewhere in the city. Our technical advisors tell us that the strategy most likely to reveal the location of the bomb is to kill members of the terrorist’s family, one at a time, in some very vivid and convincing way, until we either run out of victims or the terrorist reveals the location of the bomb. Surely it is not frivolous to say that such a strategy would be impermissible. Such a strategy would also be, in our current legal system, against the law, as would be the strategy of torturing the terrorist himself/herself. Recognizing constraints that rule out such a course of action are well within the realm of common-sense morality.

At this point, it seems a natural move to amend this argument: “Of course we mustn’t kill innocent members of the terrorist’s family: that would violate the right to life. The entire purpose of the policy recommended here is to defend against violations of that right.” But notice that the right to life, in that event, is being treated as a constraint that can require us to forego a greater quantity of the very thing—people living rather than becoming the victims of homicide—that the right is apparently supposed to protect. Why? As Robert Nozick pointed out some years ago, rights as constraints express the Kantian idea that human beings are not to be used as mere means: in this case, it bars us from killing one person simply to benefit other individuals, even if the benefit is preventing more than one death.4 But there is precisely the same sort of reason for recognizing the various other rights, the ones that are being set aside here in defense of the right to life. This argument would justify us in exploiting individuals for the benefit of others in every way short of killing them. It seems irrational to accept such a thing and yet stop short of killing for the greater good. Since that is an unacceptable alternative, it surely seems best to treat all basic human rights, and not the right to life alone, as constraints.

In what follows, we will seek to understand which CNPIs are consistent with the basic principles of a free society. Our goal is to limit CNPIs to ones that are morally defensible and as friendly to human freedom as is reasonably possible.

We will begin by looking at Anglo-American common law, in particular the law of torts and of contracts as a possible source of CNPIs that might deal satisfactorily with epidemics. This seems a good place to begin the search for freedom-friendly policies. Consider the nature of tort law: dealing with wrongful use of force against persons and property is one of its central concerns. This can be seen easily by considering the intentional torts: trespass to land, trespass to chattels (non-consensual but temporary use of another’s personal property), and conversion (theft) involve violation of property rights. Assault (which in tort law means threatening battery), battery (nonconsensual harmful bodily contact), false imprisonment, and wrongful death: all these implicate one’s rights to life, liberty, and bodily integrity. There are other intentional torts (e.g., defamation, invasion of privacy, intentional or negligent infliction of emotional distress, etc.), but they raise ancillary issues which are not relevant to the question of the permissibility of CNPIs. In addition to intentional torts, an act may be tortious if it involves negligence, that is, acting without sufficient care in a way that causes harm. There is also a special category of torts involving strict liability for harms caused, regardless of whether sufficient care is taken, but the types of activities which qualify for this treatment are very limited and unusual (e.g., the use of explosives). The rights recognized by tort law are person-and-property rights, immunities against coercive interference which the individual possesses against other individuals. There are no victimless torts. Taking dangerous drugs or practicing prostitution are not torts, nor is reading the Koran, The Communist Manifesto, or The Tropic of Cancer. If force is the province of tort law, contract law deals with its ugly little brother: fraud. What is most important for our purposes, contract law also empowers citizens to, by mutual consent, create new rights and obligations for themselves and their contracting partners, rights and obligations that have the force of law. This immeasurably expands the freedom of individuals to create solutions to mutual problems that would be unavailable under a predetermined set of rights and obligations imposed from above. Together, the rights recognized by tort and contract law are the sorts of rights that would be recognized by a polity that sees its function as protecting citizens from force and fraud while enabling them to freely cooperate in pursuing their diverse visions of the good life. These are the sorts of rights that can be the basis of a free society.

Could such a system produce CNPIs that would provide a sufficient basis for a satisfactory policy response to epidemics? We begin with some obvious facts and some obvious questions. Clearly, infectious diseases can cause bodily damage and even death. Further, these are inflicted on the sufferer by other people. Bodily injury and death caused by other people is just the sort of thing that tort law is about. The obvious questions, then, are: Can spreading an infectious disease be a tort and, if it could, would that be satisfactory as the coercive element of society’s response to epidemics?

There were at one time several United States jurisdictions which had laws prohibiting the intentional or negligent transmission of HIV. Since this is just an instance of common law battery, such laws appear to be consistent with the sort of framework we are using here. But it is worth noting that during most of the time such criminal statutes were in operation, infection was readily ascertainable through testing, casual transmission was almost impossible, the kinds of activities which could lead to transmission were voluntary, and in the event of infection, contact tracing is at least feasible. Lastly, effective treatments for HIV infection are now available, making it easier to avoid having the disease. Given that such a law could be made to work, we can also infer that a tort suit for battery in such cases would also be possible. And in that case, it is very hard to see what could possibly justify further CNPIs to prevent the spread of HIV where these background facts obtain, as they have done now for many years. Under these kinds of facts—that is, difficulty of transmission, the consensual nature of the acts that enable transmission, ease of testing, feasibility of contact tracing, and availability of effective treatments—the state can confine itself to enforcing the rights that exist under tort law and has no need for more drastic coercive measures.

Other fact patterns are possible, and some of them are relevantly different. Contrast the case of HIV with that of the common cold. With the latter, there is no testing possible, and thus until people are symptomatic, they have no way of knowing whether they are infected and contagious or not. Transmission is easy and casual, rendering contact tracing almost impossible. Does this kind of case justify isolating—coercively isolating—the infected? Clearly not, not because the harms of the common cold are redressable through tort law, but because the harms of the common cold are trivial or, as the law sometimes says, de minimis. There is no point in redressing it: it’s not worth the candle in administrative costs to do so.

The interesting kind of case for our purposes, however, are infectious diseases which combine elements from both sorts of cases. Consider COVID-19. It is potentially lethal, and quite a bit more quickly than HIV is, and thus cannot be ignored like the common cold. But like the common cold, for many weeks into the pandemic of 2019–20, testing was not widely available in the United States, making it impossible to know whether one should isolate oneself. Transmission was extremely easy under normal social conditions, and it was apparently possible either while one was asymptomatic or before symptoms became sufficiently noticeable to arouse concern. There is still at the time of writing no vaccine, and available treatments are of limited effectiveness. Obviously, the market is not yet in a position to solve this problem with PIs, though it may eventually do so.

Is tort law able to address the problem? Not in any obvious way, for two reasons. First is the problem of the unknown tortfeasor: given the ease and character of transmission, who will you sue when you become ill? A reverse class action against the whole of society? How will you prove your case? Second is the problem of individual insignificance. Some harms are the cumulative result of repeated applications of some harmful event or agent. There could be a number of agents who contributed to this effect, no one of which is sufficient to bring the harm about, while all of them combined are sufficient. This can be true of toxins, which is one thing that stands in the way of effectively dealing with pollution as a tort. It can also be true of infectious diseases. In virus-borne diseases the size of the inoculum, the load of viral particles introduced by contact with carriers, is a predictor of whether one catches the disease, how severe it is, and whether it is fatal. An initial dose below a certain level might even bring about immunity. There is evidence that this is true of COVID-19.5 This compounds the problem of the unknown tortfeasor, as individual contributions might be very small, making those who contributed harder to find, and it adds another problem as well: if a given individual’s contribution to the harmful effect is actually too small to cause it, then how do we justify holding that individual responsible for it, even if we can identify them?

We will argue, nonetheless, that the case for satisfactory CNPIs is not lost, not even if we limit the state, as we propose to do, to protecting the rights that individuals have under the common law of torts and contracts. In addition to staying as far as possible within the bounds of these person-and-property rights, we will forego any advantage that may be derived from another source: namely, utilitarianism. Utilitarianism proposes to choose between actions, policies, or rules by aggregating the good or bad effects of everyone affected by it and choosing the one with the most favorable aggregate result. It has a built-in propensity to sacrifice the interests and liberty of the individual to this aggregate good. Since there is no super-individual entity that enjoys this aggregate good, this means sacrificing the good of the individual, not to some superior being that might be entitled to sacrifices, but simply to other individuals. Common law rights are a much more liberty-friendly framework.

How are they related to coercive measures against infectious diseases? They, after all, are rights, and they include rights against unconsented invasions of one’s body that result in injury or death. Mary Mallon, better known as “Typhoid Mary,” was an asymptomatic carrier of typhoid fever. She is known to have infected at least forty-seven people, all of whom had a far more severe form of the disease than she did, and to have killed at least three.6 Those people had rights against being injured or killed, by anyone, including Mary Mallon. Nonetheless, she did inflict these things. That she did not do so maliciously, and probably never understood that she had the disease,7 is relevant to whether she should have been punished or held liable for damages. But unless there is some argument that her ignorance and lack of malice caused her victims’ rights against her to wink out of existence, they are not relevant to whether she violated their rights. Further, these rights seem to be just the sort of right that justifies the use of coercion. Suppose that I know that Typhoid Mary is about to come into some possibly infectious contact with me (perhaps she is about to touch me in some hazardous way with her contaminated hands). Suppose in addition that the only way I can avoid this contact—perhaps she falsely believes she is not infected, as the historical Mary probably did, and that she is not a carrier—is to use some sort of physical coercion against her. Surely I would not be wrong in using coercion. This is simply an instance of the right of self-defense. In a case like this, it would be self-defense against what Nozick called an “innocent threat”8: someone who is on a course to kill or injure one through no fault of theirs. Surely coercion against innocent threats is justified. The only alternative to such a view would be, in a case like Mary Mallon’s, to accept a duty to submit to a substantial risk of infection, and this seems too onerous a duty by any reasonable standard. The right of self-defense, understood in this way, is not based on the guilt of the threatener, but on the existential imperative of the defender: he or she must be free to defend life and liberty against imminent threats.9

Though the case for isolating the infectious seems so far to be straightforward, even within the narrow ethical constraints we have accepted, it does raise at least one thorny side-issue. This is the nature of the “hazard” or “risk” that can justify forcible self-defense and coercive state actions like forced isolation. It is natural to think of this risk in purely probabilistic terms, as the enhanced likelihood of suffering injury or death. The problem is that among thinkers that take a strongly pro-liberty position there is some disagreement about whether risk, conceived in this probabilistic way, can justify coercion by itself. Nozick held that one may legitimately use force to defend against risk. He took it as obvious that one may legitimately use force against someone who tries to play Russian roulette on one’s head, even if the weapon had 100,000 chambers, only one of which is loaded. He also thought that the same sort of consideration can justify the state in prohibiting a manufacturer from using an efficient but risky manufacturing process that imposes danger on people in the vicinity of the factory. In addition, he believed, the state may rightly prohibit someone, for reasons of risk-prevention, from doing some valuable activity that others are allowed to do—his example is an epileptic who wishes to drive a car—though in that case it has an obligation to compensate them for the disadvantage such a prohibition would impose on them.10 Murray Rothbard, on the other hand, argued that merely exposing another person to risk of harm (apparently regardless of the degree of the risk) is not something that can be resisted by force. The legitimate occasions for forcible defense are limited to imminent violent attacks which include at least one overt act. He argued that the risk-alone view lay behind one of the principal arguments for alcohol prohibition in the U.S., which was that imbibing alcohol made it more likely that one would do something violent and that prohibiting it was simply a defensive action against violence. We can imagine, he said, that someone could make an argument today for making vitamin supplements compulsory on the grounds that certain vitamin deficiencies make people more irritable, and that irritability makes it more likely that one will be violent. Even if the facts alleged were the case, he held that they would not be just grounds for compulsion.11

These two ideas are probably not as different as they at first appear. Limiting Nozick’s idea that risk can justify coercion so that it does not apply to being a risky sort of person (a consumer of alcohol, say, or a non-consumer of supplements) seems like a friendly amendment and indeed seems to be a move he does make elsewhere in his book.12 It also seems like a good idea, since the practice of coercing people who may never do anything wrong simply for being, so to speak, a wrong sort of person, can easily lead to abuse and may well be intrinsically unjust. On the other hand, the overt acts that Rothbard has in mind include ones that merely threaten—have not yet actually done—harm.13 This would mean that the reason such an act can be coercively prohibited is the risk of harm it poses to others. This would mean that the risk produced by an act can justify coercively preventing it. In what follows, we will assume this to be correct: actions that pose an imminent risk of harm or death justify a coercive response.

This idea is sufficient to justify isolating the infectious in certain circumstances. We have a right to use coercion to fend off potentially infectious contacts, and it is likewise just for the state to exercise this right in our behalf. Such defensive state action would take the form of prohibiting such contact, with some sort of penalty attached, even if the penalty consists merely in official disapproval of acts that violate the prohibition. As with any use of coercion, it is important that it be limited to the mildest form capable of achieving the appropriate level of deterrence. Depending on the degree of harmfulness of the disease, its level of contagiousness, and the trustworthiness of the carrier, the best practice might be to simply ask detected carriers to self-isolate. This practice has often been followed in the current COVID-19 epidemic. On the other hand, officials might rightly decide that we cannot afford to trust a carrier of a disease to follow the proper practices. This is how officials saw the case of Mary Mallon. Originally, after she was found to have infected at least twenty-two people and killed one, she was confined for a period of three years. Then, on humanitarian grounds, a judge freed her after extracting from her a promise to permanently quit her erstwhile profession as a cook. She had by that time achieved a reputation as a capable cook for prosperous families in the New York area. Unfortunately, food preparation is one of the main ways in which typhoid fever is spread. After her release, she worked for a while in a laundry, but apparently decided she could not make a living away from her original profession, so she returned to it. By the time she was apprehended again, she had infected twenty-five more victims and killed two others. She was sent back to her original place of isolation.14 Mallon’s case was admittedly an extreme one—a highly unreliable carrier of a highly contagious and very dangerous disease—but this extreme case is perhaps enough to bring out the obvious truth of a crucial principle: that coerced confinement can be necessary to sufficiently deter a carrier from committing acts of battery and would in such cases be justified. It is sometimes the only way we can with sufficient assurance prevent such right-violating acts.

Of course, this only means that compulsory isolation is sometimes justified. How do we draw the line between cases in which it is justified and those in which it is not? As we have already suggested, both degree of contagiousness and degree of harmfulness (especially, lethality) are relevant, as is the reliability of people identified as carriers. Isolating someone for having a cold (highly contagious but not harmful enough), as for having AIDS before there were effective treatments (very deadly but not very contagious), would be morally wrong: such measures would amount to depriving someone of their liberty without just cause. This surely applies to Cuba’s notorious policy, during the AIDS epidemic of the ‘eighties, of subjecting everyone found to be HIV-positive to a life sentence in a system of special prison camps.15 Beyond commenting about extreme cases such as these, there seems to be little that can be established in a philosophical essay like this one about how these three considerations should be weighed in deciding cases and formulating policies as to where to draw the line.

The same sorts of considerations that justify coercion in the case of isolating the infectious also justify two further CNPIs without which such isolation might be of little avail as a defense against dangerous diseases. These are contact-tracing—that is, requiring people believed on good evidence to be carriers to divulge the names of people with whom they have had potentially infectious contact—and compulsory testing of this secondary category of individuals. Depending on the circumstances and the nature of the infection, “testing” might comprise simply looking for patently observable symptoms or limited-term quarantining to find out whether symptoms become observable as well as, if available, laboratory tests such as the polymerase chain reaction test for the SARS-CoV-2 virus. In all these cases, the degree of certainty needed—will this person cause sufficiently serious physical harm?—is lower than that required to justify the use of coercion involved in punishing someone for a crime. Even in criminal court, the standard falls short of absolute certainty—“beyond a reasonable doubt” does not mean beyond any doubt one could have—but it is considerably higher than the standard for justifying defensive coercion, such as isolating the infectious and these two ancillary practices. This is partly because criminal matters are about events that have happened, whereas defensive measures are remedies for things that have not happened—yet—and may indeed never happen. Not everyone for whom Mary Mallon prepared food developed typhoid fever. In human affairs, the degree of certainty that is attainable for anticipating future events is typically much lower than the degree available for judging of the past. More importantly for our purposes, the ethical considerations are very different, and in relevant ways. Those who produce decisions in a criminal proceeding have a duty to avoid the very grave injustice of punishing the innocent, so far as that is possible. This places a heavy burden on those who decide the case to find that the evidence conclusively shows that the accused does indeed deserve punishment. On the other hand, defensive coercion is not a punishment and is not based on the judgment that the individual being coerced deserves it: if it were, coercion against innocent threats would be unjust. If a reliable test shows that someone is a carrier of a dangerous disease, that by itself justifies some sort of coercive response. The reason is that the grounds for coercion do not lie in the moral responsibility of the person coerced, but in the rights of the person who is being defended against harm. As we have said, those who are not yet infected do not have a duty to wait for carriers to injure or kill them before coercive remedies become available. All that need be shown is that a given person poses a significant risk to others, and that is established by the reliability of the test and by our knowledge of the disease. With the practices of contact-tracing and compulsory testing of contacts, we are also responding to mere risk of harm, rather than actual harm. In these cases, the level of risk is somewhat lower than in the case of those already known to be carriers, but the degree of coercion is also lower, often much lower. The grounds for these CNPIs are obviously strong, just as it is for isolating the infectious—provided, to repeat, that the disease is sufficiently contagious and harmful and the infectious person cannot be relied on to avoid transmitting the disease to others.

That these practices are justified, even from a rigorously proliberty point of view, might not surprise many people. Less obvious, and frankly more interesting, are the issues raised by CNPIs directed, not at the infectious and their contacts, but at the general population. What about requirements of mask-wearing and social distancing? Banning large gatherings? Locking down a major portion of the economy? As we suggested at the outset, there are a wide array of such policies that might be considered and, indeed, throughout history many such CNPIs have been used. What are the relevant principles and guidelines for evaluating such measures?

First, it is important to realize that, in an epidemic, anyone who is not a member of one’s own household represents some finite risk of infection. You could even say that this is what makes it an epidemic. This simple fact already justifies some CNPIs directed at the general population. Depending on the particular disease and the way in which it is transmitted, there are often simple precautions that individuals can take to substantially reduce the risk they are imposing on others. For instance, typhoid fever is communicated mainly via the feces-and-urine-to-mouth route, especially via food preparation with unwashed hands, and COVID-19 is spread mainly through droplets expelled and sent through the air with one’s breath. Consequently, a significant portion of the risk can be avoided by simple measures such as, in the one case, hand-washing before preparing food and, in the other, staying six feet or more from people who are not members of one’s household and, when this is not possible, wearing effective mouth-and-nose masks. Measures that—depending on the circumstances—either require or urge individuals to take such safety-steps are just, even in many cases morally obligatory. Admittedly, even in an epidemic the risk posed by other individuals is often small, and the risk avoided by these measures is even smaller, but this is a risk of killing someone or inflicting serious bodily injury. In this way, failing to take such measures resembles Nozick’s example of playing Russian roulette with a one hundred thousand chambered revolver. While Nozick’s revolver is a fanciful imaginary object, the risk of infecting someone, unfortunately, is not. It is also, intuitively, the sort of risk that can justify coercion. Indeed, failure to enact such measures can constitute a very serious moral wrong. That said, it is well to keep in mind, that the penalties by which this sort of CNPI is backed up can simply amount to official disapproval of violators. Given the ease of compliance, and the fact that compliance protects not only friends and loved ones but, in some cases, one’s own health and safety, such mild penalties may be sufficient.16

How does this sort of justification apply to another class of CNPIs directed at the general population: namely, ones aimed at preventing people from congregating in large groups by, for instance, banning meetings of more than a certain number of people, or by closing schools? Large gatherings can be spreader events: One survey of the literature found 72 large gatherings that were found to spread infectious diseases, such as measles, mumps, and Influenza A H1N1. However, these all seem to be events—such as summer camps where people live in close quarters—in which people were not taking the sorts of precautions that we have just been discussing.17 Perhaps the strongest reason for CNPIs that prohibit people from congregating in large numbers is that these are situations in which it is difficult to make sure that people do take basic precautions and consequently create opportunities for each person to have numerous potentially contaminating contacts in a short period of time. This sort of consideration might be strongest as a justification for school closings, since children are notoriously bad at social distancing. In addition, officials might know that in the adult population, even apart from large gatherings, many people simply are not consistently taking reasonable precautions, and meeting bans might be an attractive way to minimize the effects of their careless behavior.

If this is the rationale for such bans, they do not seem to be entirely consistent with the sort of justification that we have just presented for requiring basic hygienic precautions. That justification was individualistic, in that the principles and reasons given were applied to individuals and not to groups as such. The question – does this act violate rights?—was treated as providing the answer to the further question: May we prohibit this act? An affirmative answer to the first question was taken as giving an affirmative answer to the second one. Here the justification for coercion is that it prevents individuals from doing a rights-violating act, but the coercive measure does so by prohibiting everyone from attending gatherings of more than a certain size. This presents a problem for the rigorously individualistic point of view: We seem to be contemplating a policy of deterring a rights-violating act by coercing people who would otherwise commit that act and also people who are completely innocent, would behave responsibly and—let us assume, in order to avoid raising more than one issue at a time—would as individuals pose no threat to anyone. We seem to be deterring the disease-spreading threat posed by the group attending the event, as a group. But this would include prohibiting individuals who (we are assuming) would not be committing a right-violating act. Thus, the CNPI of placing a ban on congregating in large numbers cannot be justified in the way that we have justified requiring basic hygienic measures.

This is a problem because, to cite just one reason, there can be circumstances in which those same responsible individuals might well want a rule that coerces them in precisely the way such a ban would. After all, a spreader event inserts more of the disease organisms into the general population, which raises the probability that any given person, including some responsible individuals, will be interacting with them. Given that it is impossible for an individual to know who is a carrier and who is not and, further that it is likely to be impossible to protect oneself perfectly against infection from unknown carriers, it seems that there must be a plausible ethical justification for banning large gatherings in some circumstances. But how exactly would this justification go, and exactly what sorts of measures would it justify?

At this point, an approach that some people would find appealing would be to opt for utilitarian reasoning: gathering bans are justified when the aggregate total benefit for everyone affected by it, in the form of infection abatement, exceeds the total aggregate cost of abridging everyone’s freedom of assembly. We have already said that we will be avoiding such appeals to the collective good here. In addition to the reason that we have already given for doing so, there is an additional reason that applies to this particular ethical issue: we have already seen reason for thinking that meeting bans might not involve, at least not in every case, the sacrifice of the interests of some individuals for the benefit of others. Does this mean that there can be an individualist defense of gathering bans? We think it does, but to adequately explain why, we need to first make some general comments on theories of rights.

It has long been a commonplace in discussions of rights that a right can be protected either by a property rule or by a liability rule.18 For rights that place limits on the behavior of others, a property rule says those others may not cross that limit except with the permission of the holder of the right, while a liability rule says they may not cross it unless due compensation is paid to the holder. In the stereotypical case, compensation is paid with a literal payment—that is, with money. To use Joel Feinberg’s example: caught in a blizzard, you break into a mountain cabin and take life-saving shelter there, unavoidably damaging the cabin in the process; later you identify the owners and send them a check.19 Your taking steps to fully compensate the owners contributes to making your damaging their property an innocent act. But this is not the only form that compensation can take. As Richard Epstein pointed out some years ago, if we recognize that compensation can include “implicit in-kind compensation,” we can see that compensation can be brought about by rules of conduct as well as by actions. More precisely, it is possible for a rule, which imposes costs on those to whom it applies, to also compensate them for those costs. His example is the “live and let-live rule,” which states, roughly, that one should tolerate certain low-level nuisances produced by others’ use of their property—such as smoke from a neighbor’s barbecue pit or smells produced by emptying a cess-pool—because the same rule gives one the same freedom to use one’s own property, incidentally, producing these same nuisances oneself. The nuisances that the rule imposes on one are trifling, and the freedom of action and enjoyment of one’s own property one gains in return are substantial, more than compensating for those trifling costs.20

It is easy to see how a ban on large gatherings during an epidemic can provide sufficient implicit in-kind compensation. Depending on how contagious the disease is and how harmful, even a slight decrease in the probability of infection can be a substantial benefit, even for people who take reasonable precautions but who—as may well be the case—cannot achieve complete safety as they make their way through the general population. On the other hand, the cost might well be trivial by comparison. Perhaps the gardening club has to meet as two smaller groups rather than one larger one, or the steering committee meeting has to take place virtually, on the internet. Under a liability rule approach to rights, a large gathering ban, under fact patterns such as these, can be fully justified.

At this point, there are some problems we need to solve, if possible. One involves the problematic nature of the liability rule approach itself. The other has to do with a serious ethical complication that arises when the notion of implicit in-kind compensation is applied to actual cases.

Eric Mack has subjected the liability rule conception of rights to some searching criticisms. First, compensation addresses the harm done by violating a right, but not the wrong done. Harm and wrong are very different things. Compensation might erase the harm caused by violating a right, but if it was a right at all, violating it is wrong. Second, the liability rule conception is inconsistent with Nozick’s attractive explanation of rights, which is that they express the idea of the dignity of the individual human being by blocking the individual from being used as a tool. If we violate someone’s rights because we can gain enough by it to afford compensating them for the damage done, we are using them and paying them for it, but we are still using them. Third, it is also inconsistent with Nozick’s answer to H. L. A. Hart’s principle of fairness, the principle that individuals who have taken on certain burdens in the course of some benefit-conferring activity have a right against all beneficiaries of that activity that they share in those burdens, on the grounds that “[o]ne cannot, whatever one’s purposes, just act so as to give people benefits and then demand (or seize) payments.”21 Why isn’t this simply advance compensation, Mack asks, which makes the forced imposition of these burdens blameless? Fourth and last, under a liability rule theory of rights, all forced exchanges that leave all parties at least as well off as before would be permissible. For instance, suppose that a man is captured in Africa and shipped to America to be a slave for life, and that his tribe is subsequently wiped out by a horrific tropical disease. He lives out a terribly hard life, but that is (let us suppose) better than dying of the disease. That would on this theory mean that what the slave traders did was morally permissible.

These are certainly weighty objections, especially the second one, which charges, plausibly, that a liability rule conception of rights is inconsistent with the very basis of rights in the idea of human dignity. But three of them—the second through the fourth—are directed at the liability rule idea as a theory of rights in general, or at least rights that constrain one’s treatment by others. That is, they are directed at the notion that all such rights may be infringed provided only that there is due compensation for the harm done. That is why the liability rule approach is inconsistent with human dignity: if the liability idea is conceived as a quite general understanding of rights, then rights can no longer serve as blocks to being used. We propose to work around the last three objections by severely restricting the application of the liability rule approach to the context in which Nozick himself applies it: to the prohibition of risky activities and, further, only to some risky activities.22 As we will use it, the notion of liability rules will have a narrowly circumscribed application, with many blocks to using persons as tools left standing.

As Nozick has pointed out, prohibiting-but-compensating an action is a sort of middle ground between permitting it and simply prohibiting it.23 Where in this range of appropriate responses do risky activities belong? Clearly, an infected person’s failing to take appropriate precautions is (again, given appropriate levels of contagiousness and harmfulness) simply prohibitable. Such a failure is analogous to Nozick’s example of playing unconsented Russian roulette on another’s head with a pistol with 100,000 chambers. In some historic epidemics the analogous weapon would have far fewer chambers than that. A person who, in an epidemic, does not know whether they have been infected or not, and yet does not take precautionary measures is at one remove from that: it is like playing Russian roulette without knowing whether the gun is loaded or not. We have been assuming that a CNPI requiring them to take precautions is legitimate, though perhaps the somewhat attenuated nature of the wrong that the CNPI would be prohibiting should be reflected in relatively mild penalties for non-compliance. Now consider the case of responsible individuals who wish to attend a large gathering and, again, suppose that their precautions would indeed mean that they would not themselves contribute to the spread of the disease. May we legitimately ban the gathering, knowing that this would mean prohibiting them from attending? This is behavior that is one step further removed from the category of the simply prohibitable. On the one hand, we can (depending on the disease) have good enough reason for thinking that the gathering would increase the risk of infection for the general population, and that we can only avert this increment of risk by banning it. On the other hand, in the situation we are now assuming for the sake of the argument – this would have to include prohibiting attendance by people who are not actually exposing others to the risk, along with those who are. Coercing people who pose no risk, and giving risk abatement as the reason for it, would ordinarily be wrong, except that here there is (we are supposing) substantial risk that can be avoided in no other way.

This seems an appropriate occasion for the middle-level response of prohibit but compensate. What Nozick says of the rationale for his own use of the idea seems applicable here, despite the fact that he applies it to cases quite different from ours, such as the epileptic driver:

[P]rohibition with compensation … is not a “split the difference” compromise between two equally attractive alternative positions, one of which is correct but we don’t know which. Rather, it seems to me to be the correct position that fits the (moral) vector resultant of the opposing weighty considerations, each of which must be taken into account somehow.24

Since we have no practicable way of identifying these unoffending but banned individuals, the implicit in-kind compensation, brought about by the same rule that we are imposing on them, is apparently the only available way compensation can be brought about. Clearly this is entirely possible. Because the increment of risk abatement is distributed through the general population, this benefit falls upon the banned but unoffending individuals along with everyone else. Depending on the nature of the disease being fought, these benefits can be significant and the cost of the ban comparatively trivial. In that case, these individuals receive adequate implicit in-kind compensation. We have not interfered with their liberty of action without good reason. And because these reasons are adequate even from the point of view of their own interests, these interests have not been sacrificed for the benefit of others. This seems to answer the first of Mack’s objections, that compensation erases harm but not the wrong that causes it. The attendance of the banned but unoffending individuals represents a special case, a moral middle ground between the permissible and the simply prohibitable.

It is important to realize that this way of justifying CNPIs places a certain sort of constraint on them. CNPIs that prohibit the innocent and unoffending but cannot be justified in this way are, so far, unjustified. An example taken from actual events might help clarify what this means in practice. The Madison Marriott West in Middleton Wisconsin is a ten story, two hundred and ninety-two room hotel. It is also the second largest convention center in the Madison area, with a number of large meeting rooms and auditoriums. It specializes in large-scale events, often involving hundreds of people. This means—given that a large gathering ban is still in effect as we write, seven months after its onset—that its core business is now illegal in the state of Wisconsin. Before the ban, it had around one hundred twenty-five employees. After it, the number has fluctuated between ten and fifteen. There is no janitorial or cleaning staff at all. The elaborate array of meeting rooms lies empty, though still requiring tax and mortgage payments. Though it is impossible to gather accurate numbers about this, it is a safe bet, given the state of the economy, that nearly all of the over one hundred people who lost their jobs (which included all of the lowest paid employees) are still unemployed as we write this. Did the ban afford these people adequate implicit in-kind compensation for the heavy burdens it imposed on them? It is impossible to give a definitive answer to that question here, but it is surely a very realistic possibility that it did not. The benefit of the rule that falls on everyone is merely the avoidance of a certain interval of risk, and like every other safety measure, its value is finite and may well be very small compared to the economic and psychological burdens of seven months (so far) of unemployment. The fact that the risk thereby abated is a risk of injury, suffering, and death does not mean that its value will be enormous. The same thing is true of smoke detectors in one’s house, yet consumers still do consider the price of detectors when deciding how many (if any) to buy. The evils avoided have to be discounted by the probabilities involved.

However, supposing in a particular epidemic the people who become unemployed due to a ban on large gatherings are not afforded adequate implicit in-kind compensation, that does not mean (yet) that such bans are unjustified. Such in-kind compensation is not the only kind of compensation available to us. We can also compensate them by paying them directly, and finance these payments with taxes collected from those who benefit from the ban, which of course is the general population of the area affected by the epidemic. The interesting question of how much compensation is sufficient is one we will not be able to discuss here.25 It should suffice to say that compensating them adequately is the only way to avoid coercively exploiting them for the good of others. It might happen that the taxpayers signal (in the voting booth or by some other means) that they find the amount of the compensation too large and object to paying it, but that would be good evidence that the ban is a bad idea because the benefits it affords for the many are not worth the costs it imposes on the few.

It is probably obvious by now that the approach we have just taken to bans on large gatherings can be used to decide whether any CNPI is morally justified. First, consider whether the CNPI prohibits individuals from doing some act that violates the rights of others by exposing them to risk of infection by a (sufficiently serious) disease. If the answer is yes, the measure is so far justified. Second, in order to prohibit these risky actions, must it also prohibit acts that do not pose such a risk? If the answer is no, the measure is justified. If it is yes, it is unjustified unless one of two things is also true: either the measure itself affords these additional prohibited individuals adequate implicit in-kind compensation, or they are compensated directly at the expense of the people who do benefit from the measure. For the sake of illustration, we will look at two such measures, one easy to justify, and the other very difficult.

During the Influenza Epidemic of 1918, the New York Board of Health established a timetable mandating staggered opening and closing hours for businesses and government offices in order to decrease rush hour congestion, especially on public transportation. The object was to slow the rate at which the deadly disease spread.26 Such a mandate, which applies to everyone doing business in the city, will necessarily coerce a number of people who, as individuals, are not a threat to others. But the costs it imposes on them are light, amounting to no more than inconveniences. Given how contagious and how deadly the “great influenza” was—informed opinion estimates that it killed between 2.5 and 5 five per cent of the population of the Earth27—if it was at all effective at slowing the spread of infection, everyone was adequately compensated for their inconveniences by the same mandate that imposed them. If that is so, the measure is justified.

A much more complex set of issues is presented by a policy of “lockdown,” like the one imposed by various state governments during the COVID-19 epidemic: namely, forced closures of businesses and offices that the government has declared “non-essential.” There are at least two sorts of issues raised by such policies. One is obvious, given things that we have already said. The number of those unemployed surged by more than 14 million, during recent widespread lockdowns, from 6.2 million in February 2020 to 20.5 million in May.28 One study found that in May through September of that year, 8 million Americans slipped into poverty.29 Some of this damage was an inevitable consequence of the disease, as people sought to protect themselves from the virus by avoiding places of business, such as restaurants. But some of it was deliberately caused when the government declared many people’s places of employment non-essential. Did these measures afford these people adequate in-kind compensation for losing their jobs? It is certainly arguable that, if the disease involved is as contagious and as harmful as the Black Death (bubonic plague) of 1347–50, such a thing would indeed be the case. Any significant decrease in a substantial likelihood of catching a disease that is nearly always fatal would seem to compensate one for being unemployed. But COVID-19, though very contagious and nothing to make light of, is far less deadly than the Black Death was. If the increment of safety from this disease is not enough to compensate the many victims of the shutdown policy, that would mean that they were forced to take a net loss of well-being in order to produce a net gain for others. This would be an injustice. In that case, by analogy with the case we have already made in regard to bans on large gatherings, the injustice can be erased by compensating these victims.

However, if there is good enough reason that the policy would provide sufficient benefit in terms of risk avoidance, then commonsense decency would seem to require that responsible officials make this claim in public and make a case for it. Otherwise, even if the victims are not sure they are being forcibly exploited for the good of others, they will get the clear impression that those responsible for the policy do not think it matters whether the victims are being exploited or not. The idea that it is unjust to force some people to accept welfare losses simply in order to provide benefits to others is hardly a radical one: it is simply a matter of common-sense morality. If officials do not seem to think that the case for the fairness of the lockdown needs to be addressed, that will no doubt be noticed by those who are forced by the policy to accept real hardship. To claim that the policy is justified by the fact that the livelihoods of these people are not “essential” may amount to backing up an injustice with an insult.

The fairness problem raised by the COVID lockdowns is exacerbated by the fact that the harm done by the disease, and consequently the benefits of avoiding it, vary dramatically from one age-group to another. Here is one account of (as we write) a very recent set of numbers from the Centers for Disease Control:

The CDC’s latest death counts indicate that the crude case fatality rate is around 28 percent for patients 85 or older and 18 percent for 75-to-84-year-olds. That rate falls to about 8 percent for 65-to-74-year-olds, 2 percent for 50-to-64-year-olds, 0.6 percent for patients in their 40s, 0.2 percent for patients in their 30s, 0.06 percent for patients in their late teens and early 20s, 0.02 percent for 5-to-17-year-olds, and 0.04 percent for children 4 and younger.30

If we assume that death rate is a good proxy for the harmfulness of the disease (which in this case includes other effects besides being killed), then the harmfulness of the disease, and consequently the value of reducing the risk of catching it, will vary widely with age. The likelihood that workers in their early 20s will die of the disease, if they test positive for it, is 0.06 per cent. The value of lockdown for them is a function of this number, discounted by a downward shift, probably of only a few per cent, of the likelihood that they will indeed test positive. This is a benefit, but is it more valuable than the months of employment that have been taken from these people in order to produce it? Perhaps a case can be made that it is, but it is far from obvious that it will be successful. Also very worrying, from an ethical point of view, is the distribution of costs and benefits among age groups. The lion’s share of the risk of death from infection—and consequently the lion’s share of the benefit from lockdown—falls on people aged 65 and over. But these are the people who are least likely to directly bear the costs of the lockdown, as they are unlikely to be part of the work force. They are unlikely to be among the 14 million who lost their jobs. This is a very serious problem with the fairness of the lockdown policy. It may well be that those who benefit from such a policy owe compensation to those who bore, and continue to bear, the costs.

Finally, we should say a few words about a likely source of resistance to the position we have taken here. The principles we have sketched out here are rather restrictive on state action. On any very likely account of the facts, the widespread lockdowns enacted in the United States early in the COVID-19 epidemic might very well be ruled out, and the brutally unfair way they were administered certainly would be. Does that mean these principles are too restrictive? Given how difficult it would be, according to our position, to justify something like the lockdowns that were widely in effect in the US during the outbreak, someone who is not quite convinced by our arguments might be more impressed by the lives that might be saved by such measures. Though we admittedly have no way to lay such concerns to rest here, we do think we can make one fairly simple point that might blunt their effect.

First, a true story: Seven years ago, a friend of ours was in a west African country, serving as an advisor to its government regarding responses to a deadly Ebola outbreak that was spreading there. When he was done, he flew back to the US and was surprised when he was called off the plane by customs officials and taken directly to a medical examination (something that had not happened to him in Africa). They somehow knew, without his volunteering the information, that he had been in west Africa. They gave him a radio-connected thermometer, with instructions to take his temperature daily, which the device would transmit to a public health official. Two days later, in Florida, he was visited in his hotel by a state official, who told him that Florida did not believe the federal guidelines in these matters were strong enough, and subjected him to another examination. A few years later, in March 2020, he was in Europe, where many responsible individuals feared that the rapidly spreading COVID-19 epidemic would swamp their hospitals with cases. When he landed in New York again, he was again surprised—this time, because nothing happened. No officials appeared to monitor his health, no one seemed to care that he was arriving from a public health hot spot. There seemed to be no trace of the earlier responses to the preceding pandemic.

Apparently, the institutional memory of the last alarming epidemic seems to have been very weak. Unfortunately, this was hardly unusual. As COVID-19 surged to pandemic levels, nation states around the world were surprised and unprepared, almost as if they had not expected another serious epidemic to reach them. There were at least two exceptions, however. Taiwan began screening passengers arriving from Wuhan, China, where the COVID-19 was first identified, on December 31, 2019, while it was still the subject of limited reporting in the West. Taiwanese medical workers confirmed their first case on January 21, two days before Wuhan itself went into full lockdown. Taiwan continued to pursue its policy of testing—testing people they had some reason to think might be infected, as opposed to mass indiscriminate testing such as was practiced in China—and rigorous contact-tracing and quarantining of contacts. The reason Taiwan was able to act so swiftly was instructive: The island nation was hit hard by an earlier epidemic of respiratory disease from China, Severe Acute Respiratory Syndrome (SARS), in 2003. After that disturbing episode, the government set up the Central Epidemic Command Center, to coordinate various government offices, and invested in tests, personal protective equipment such as masks, as well as in effective contact-tracing.31

Another early responder to COVID-19 was South Korea. In their case, their readiness for a quick response was a result of the nation’s experience with Middle Eastern Respiratory Syndrome (MERS) in 2015. The death toll from that disease (39 deaths) was widely regarded as a national disgrace and led the National Assembly to amend their Infectious Disease Control and Prevention Act to allow the government to collect personal information from people who test positive for epidemic diseases for purposes of efficient contact-tracing and testing of contacts. The information was to be garnered from such sources as surveillance-camera footage, smartphone location data, and credit card purchase records. In addition, during the current COVID epidemic, they used these data to transmit, via smartphone messaging, information about the location and movements of known or likely carriers of the disease, with the identity of the individuals concealed, to people known (by means of their phones) to have been nearby. This enabled people to avoid known hotspots and monitor their own health so they could take the necessary steps if they found they had been infected.32 The policy on large gatherings has generally shifted with changing circumstances, banning various sorts of gatherings—such as religious meetings—when and if they become spreader events. In both countries, moreover, wearing masks is common.

The simple point we wish to make is that by means of early action Taiwan and South Korea have avoided the general lockdowns that have been so common in other countries, and have avoided closing their borders—and yet they have done relatively well in combating the disease. This is especially true of Taiwan, which at this writing has recently marked its 200th day in a row without a single locally transmitted case.33 There is evidence, then, that testing, contact-tracing, isolation and quarantine—the most clearly ethical CNPIs if judged by the person-and-property rights we have been assuming—are alternatives to the draconian lockdowns that have been widely practiced and defended, and that they are comparatively effective ones.34 Admittedly this may well only be true if such measures are pursued efficiently and, most important, early in the progress of the epidemic. But surely such promptness and efficiency are themselves ethically imperative. For the nation-states of the world—at least ones that have highly active ties of trade and travel with countries like China—to continue to act like the previous deadly epidemic will be the last one is simply irresponsible.35

1By “epidemic” we mean an infectious disease that has spread throughout the population of a given geographical area or is in the process of doing so. A pandemic is an epidemic that has affected an entire country, several countries, or the whole world. As this distinction is irrelevant to the sorts of issues we will discuss here, we will use only the more inclusive of the two terms, epidemic, unless the extensive spread of the disease is relevant.

2James Peron, “Rattigan, Rand and Pandemic Emergencies,” The Radical Center, accessed 9/14/2020, https://medium.com/the-radical-center/rattigan-rand-and-pandemic-emergencies-a2591cdf732f.

3Ayn Rand, “The Ethics of Emergencies,” in The Virtue of Selfishness (New York: New American Library, 1964), p. 47.

4Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), pp. 30–31.

5Amanda D’Ambrosio, “Does Virus Dose or Load Predict How Sick You Get With COVID-19?—Initial Exposure, Strength of Virus Infection Both Seen as Contributors to Illness Severity” Medpage Today, September 18, 2020, accessed 10/19/2020, https://www.medpagetoday.com/infectiousdisease/covid19/88692.

6Judith Walzer Leavitt, Typhoid Mary: Captive to the Public’s Health (Boston: Beacon Press, 1996), pp. xvii–xviii.

7Leavitt, Typhoid Mary, pp. 11, 162–63.

8Nozick, Anarchy, pp. 34–35.

9Eric Mack, “Elbow Room for Self-defense,” Social Philosophy and Policy (2016): pp. 18–39.

10Nozick, Anarchy, pp. 79–84.

11Murray Rothbard, The Ethics of Liberty (New York: New York University Press, 1998 [orig. pub 1982]), p. 78.

12See Nozick, Anarchy, pp. 142–46. There he discusses preventive restraint—the practice of preventing people from doing things (such as buying a gun) deemed to make them more dangerous people—and preventive detention—which is the practice of imprisoning people because they are deemed significantly more likely than the average person to commit a crime. He says, among other things, that such practices might seem legitimate if we see some people as incapable of making a decision or incapable of deciding against acting wrongly. Otherwise, coercion is ruled out if no rights violation will occur unless the coerced person makes some (further) decision to do wrong. In addition, “even if “ coercion in such cases is permissible, it would only be so if the coerced person is compensated for any disadvantages the coercion would impose upon them.

13“It is important to insist, however, that the threat of aggression be palpable, immediate, and direct; in short, that it be embodied in the initiation of an overt act.” Rothbard, Ethics, p. 78.

14Leavitt, Typhoid Mary, p. xviii.

15The regime called them sanitariums, which internally is what they were, except that, like a prison, inmates were barred from leaving. It may well be, as the New York Times tells us, that this policy “ had a huge damping effect on the early epidemic,” but this merely illustrates a point we have already made, that epidemiological considerations are not the only ones that are relevant here. These facilities eventually closed due to budgeting problems when the Soviet Union imploded, and Soviet foreign aid collapsed. See Donald G. McNeil Jr., “Cuba’s Fortresses against a Viral Foe,” New York Times, May 7, 2012, https://www.nytimes.com/2012/05/08/health/cubas-aids-sanitariums-fortresses-against-a-viral-foe.html, and Rebecca Sananes, “Love, Loss, and Beauty Pageants: Inside a Cuban HIV Sanitarium,” March 26, 2016, NPR, https://www.npr.org/sections/goatsandsoda/2016/03/26/471765424/love-loss-and-beauty-pageants-inside-a-cuban-hiv-sanitarium.

16During the COVID-19 epidemic, the state of California passed a typical six feet or face mask rule, but failed to attach any penalties for non-compliance. In Plumas County, where one of the present co-authors lives, local businesses put signs saying that they are complying with the guidelines and that people without face coverings would not be admitted. Anecdotally, compliance with the guidelines in public places appears to be over 90%. See California Department of Public Health, “Guidance for the Use of Face Coverings,” accessed 9/25/2020, https://www.cdph.ca.gov/Programs/CID/DCDC/CDPH%20Document%20Library/COVID-19/Guidance-for-Face-Coverings_06-18-2020.pdf.

17Jeanette J. Rainey, Tiffani Phelps, Jianrong Shi, “Mass Gatherings and Respiratory Disease Outbreaks in the United States—Should We Be Worried? Results from a Systematic Literature Review and Analysis of the National Outbreak Reporting System,” Plos One, August 18, 2016, accessed 9/26/2020, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0160378.

18The distinction was introduced in Guido Calabresi and A. Douglas Melamed, “Property Rules, Liability, and Inalienability,” Harvard Law Review 85, no. 6 (1972): pp. 1089–1128.

19Joel Feinberg, “Voluntary Euthanasia and the Inalienable Right to Life,” Philosophy and Public Affairs 7, no. 2 (1978): p. 102.

20Richard A. Epstein, “Nuisance Law: Corrective Justice and Its Utilitarian Constraints,” The Journal of Legal Studies 8, no. 1 (1979): pp. 82–87.

21Nozick, Anarchy, p. 95.

22Following earlier discussions by Peter Railton and David Sobel, Mack claims that Nozick opts for a quite general liability rule conception of rights in “Prohibition, Compensation, and Risk,” Ch. 4 of Anarchy, pp. 54–87. Mack, “Elbow Room,” p. 204, n. 11. See also Eric Mack, “Nozickian Arguments for the More-Than-Minimal State,” in The Cambridge Companion to Nozick’s Anarchy, State and Utopia, eds. Ralf Bader and John Meadowcroft (Cambridge: Cambridge University Press, 2011), pp. 100–03. The interpretation presented in the latter paper seems to be that the relevant part of the argument of Ch. 4 goes like this: first, Nozick asserts, apparently arbitrarily, that rights are only protected by liability rules and then presents two arguments—one concerning difficulties in finding the correct amount for the compensation, and one showing that compensation does not prevent the fear one experiences about the possibility of some rights-violating acts—to the conclusion that some rights should be treated as if they are property rule rights. This is pretty clearly a misinterpretation. When Nozick asks (Anarchy, p. 57) whether those who transgress the boundaries of others are “permitted to perform such actions provided that they compensate the person whose boundary has been crossed?” (italics in the original) it is a real question, not a rhetorical one, and the arguments about problems in calculating compensation amounts and about fear are partial answers to it, and are explicitly presented as such. At the end of this part of Ch. 4 (p. 71), he says: “Other considerations converge to this result [that is, that not all rights can be liability rule rights]: a system that permits boundary crossing, provided compensation is paid, embodies the use of persons as means….” He gives a reason for rejecting the general liability rule theory of rights that is identical to the one presented by Mack in his second objection to that theory.

23Nozick, Anarchy, p. 83.

24Nozick, Anarchy, p. 146.

25See Nozick, Anarchy, Ch. 4, passim. Nozick discusses three different types or levels of compensation. Full compensation for a particular cost is an amount sufficient to make one as well off as one would be if the cost had not been imposed. Market compensation is the amount that the individual accepts, or would have accepted, as an inducement to voluntarily accept the cost. Disadvantage compensation is an amount sufficient to make one as well off as those in the general population who have not had this particular cost imposed on them. Given that the sort of rule we are considering here probably imposes costs that everyone must share, where these costs are distinct from the special ones for which this individual is being compensated, disadvantage compensation is apt to be less than full compensation, which is very likely to be less (possibly much less) than market compensation.

26See Francesco Aimone, “The 1918 Influenza Epidemic in New York City: A Review of the Public Health Response,” Public Health Reports, 125, Supplement 3 (2010): pp. 71–79. Available online at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2862336/.

27Laura Spinney, Pale Rider: The Spanish Flu of 1918 and How It Changed the World (New York: Public Affairs, 2017), p. 4.

28Rakesh Kochhar, “Unemployment rose higher in three months of COVID-19 than it did in two years of the Great Recession,” Pew Research Center, June 11, 2020, accessed 10/13/2020, https://www.pewresearch.org/fact-tank/2020/06/11/unemployment-rose-higher-in-three-months-of-covid-19-than-it-did-in-two-years-of-the-great-recession/.

29Zachary Parolin, Megan Curran, Jordan Matsudaira, Jane Waldfogel, Christopher Wimer, “Monthly Poverty Rates in the United States during the COVID-19 Pandemic.” Poverty and Social Policy Working Papers, October 15, 2020, accessed 10/17/2020, https://static1.squarespace.com/static/5743308460b5e922a25a6dc7/t/5f87c59e4cd0011fabd38973/1602733471158/COVID-Projecting-Poverty-Monthly-CPSP-2020.pdf.

30Jacob Sullum, “The Latest CDC Estimates of COVID-19’s Infection Fatality Rate Vary Dramatically With Age,” Reason.com, September 29, 2020, accessed 10/14/2020, https://reason.com/2020/09/29/the-latest-cdc-estimates-of-covid-19s-infection-fatality-rate-vary-dramatically-with-age/.

31Joshua Berlinger, “Taiwan just went 200 days without a locally transmitted Covid-19 case. Here’s how they did it,” CNN World, October 30, 2020, accessed 11/4/2020, https://www.cnn.com/2020/10/29/asia/taiwan-covid-19-intl-hnk/index.html.

32Wudan Yan, “What Should the U.S. Learn from South Korea’s Covid-19 Success?” Undark, October 5, 2020, accessed 11/5/2020, https://undark.org/2020/10/05/south-korea-covid-19-success/.

33See footnote 31, above.

34It is true that South Korea as we write has had increased trouble with a second wave of the COVID epidemic, due in significant part to public resistance and disobedience to some large gathering bans—especially bans of in-person religious meetings—but this in a way reinforces the point we are suggesting here. It seems likely such disobedience would be greater if the government had also attempted to lockdown a major portion of the economy, which of course they have not done. See Choe Sang-Hun, “New Covid-19 Outbreaks Test South Korea’s Strategy,” The New York Times, October 2, 2020, accessed 11/6/2020.

35We would like to thank Imtiaz Moosa and Nathaniel Hunt for discussing these issues with us, and Ross Levatter and Scott Horton for useful comments on an earlier draft.