I. Introduction
Libertarianism is a controversial political theory. But it is often presented as resting upon uncontroversial, indeed commonsense, moral foundations. Footnote 1 Don’t Hurt People and Don’t Take Their Stuff is the title of one recent and popular libertarian manifesto, the idea apparently being that once you have accepted this obvious moral principle, all that is standing between you and full-fledged libertarianism is a failure to think through the full set of its logical implications. What distinguishes libertarians from other people, on this view, is not the basic moral principles they accept, but the consistency with which they apply these principles across a wide range of cases, where others fear to follow. Footnote 2
Indeed, many libertarians see the whole edifice of their political theory as resting almost entirely upon a single moral principle — the so-called Nonaggression Principle (NAP). Footnote 3 This principle, in its canonical formulation by Murray Rothbard, holds that “no man or group of men may aggress against the person or property of anyone else.” Footnote 4 Combine that principle with another principle legitimating property in external resources, Rothbard held, and “the right of free exchange and free contract . . . the right of bequest, and . . . the entire property rights structure of the market economy” follow as a matter of simple logic. Footnote 5
How’s that? Well, according to Rothbard, “aggression” can be defined as “the initiation or threat of physical violence against the person or property of anyone else.” Footnote 6 On this definition, taking someone else’s property by use or threat of force is a violation of the NAP, and hence impermissible. So, too, is forcefully preventing people from doing whatever (nonaggressive) things they might want to do with their own bodies or property. So far, none of this is terribly radical. Most of us think it would be wrong to threaten our neighbor with violence in order to get her to hand over her money, or to prevent her from wearing a particular shirt that you do not like. Libertarians, however, apply the principle not just to the personal behavior of private citizens, but to the state and its agents as well. And they hold that it establishes not merely a strong presumption against activities involving the initiation of physical force, but an absolute moral prohibition. Thus, if my taking your money without your consent is theft, so too is taxation. If my threatening to lock you away if you smoke something I think you shouldn’t is impermissible aggression, then so too are drug prohibition laws. Follow this logic far enough, and pretty quickly almost all the activities of the modern regulatory/welfare state, if not the state itself, are seen as aggressive and hence, according to the NAP, absolutely impermissible. Footnote 7
It would, to put it mildly, be philosophically impressive if this kind of argument turned out to work. Commit yourself to the idea that it is wrong to punch your neighbor in the nose and, voilà!, you are thereby committed to legalizing kidney sales and abolishing the welfare state. Philosophers have a special place in their heart for this kind of argument, proceeding as it does from obvious premises to seemingly absurd conclusions. And as someone with strong libertarian sympathies myself, I must admit to finding something deeply compelling in it.
Unfortunately, the argument simply does not work. At least, it does not work quite in the way some of its more strident proponents hope and think that it does. Libertarians are on to something in identifying aggression as a serious moral wrong. And they are most definitely on to something in their belief that a norm against aggression ought to apply to the institutions of government with the same stringency that it applies to individuals acting in their private capacity. But these claims are significantly more modest than those that libertarians try to build into or draw out of the NAP. And so, too, are the conclusions they warrant. Taking seriously a strong norm against aggression might move us in a broadly libertarian direction, but it will not be enough to justify the more radical conclusions that libertarians have drawn, such as that any state provision of public goods or a social safety net are necessarily illegitimate, or, at the extreme, that the state itself is a criminal enterprise that ought to be abolished. Footnote 8 Those conclusions might turn out to be justified on other grounds. But a norm against aggression cannot do the job by itself.
In this essay, I identify two broad categories of problem for the NAP and its role in libertarian theory. The first has to do with what libertarians mean when they talk about prohibiting aggression. In Section II, I note, following in the footsteps of many other critics of libertarianism, that libertarians who talk about “aggression” are using that term in a moralized way. In Section III, I contrast this moralized understanding of aggression with what I take to be the ordinary, nonmoralized understanding, embodied in what I call the “First Punch Theory of Aggression.” So the first problem is that, in spite of its name, the NAP is less about the prohibition of aggression, as that phenomenon is normally understood, and more about the protection of property rights. The second problem has to do with the absolutist nature of the NAP as it is often understood by libertarians. Section IV explains what it is for a moral prohibition to be absolutist, and identifies three significant problems for an absolutist understanding of the NAP. Finally, in Section V, I try to rebuild a more defensible version of the NAP that does not suffer from these two problems, and to show how far such a principle might go in the direction of a robust libertarian political theory. Such a principle, I argue, takes us not quite as far as strict libertarians like Rothbard, Rand, and Nozick would like to go. But it does take us a good way in the direction of classical liberalism, and this itself is enough to constitute a significant challenge to the philosophical and practical status quo.
Before proceeding with this argument, however, I want to include a few words to clarify its target and its motivation. In particular, I want to emphasize, first, that I am not arguing against libertarianism per se, but against one particular kind of argument for libertarianism. Libertarianism is best understood as a family of political theories united by rough agreement on a set of normative beliefs, empirical generalizations, and methodological approaches. Footnote 9 But while there is considerable overlap among libertarians, there is also a great deal of pluralism — more, I suspect, than has generally been recognized. Thus, not only is it possible to defend libertarianism on grounds other than an appeal to the NAP, it is indeed quite common to do so. This is especially true of the most well-known academic libertarians. Neither Eric Mack nor Jan Narveson base their libertarianism on an appeal to the NAP, even though their versions of libertarianism come closest in terms of their conclusions to the sort of radical positions espoused by Rand and Rothbard. Footnote 10 And the same is even more obviously true of other libertarians (more properly classical liberals) like Friedrich Hayek, Loren Lomasky, Richard Epstein, and David Schmidtz. Footnote 11 All of these individuals no doubt think that there is a very strong moral presumption against initiating violence against innocent others. But, as we shall see later in this essay, there is a great deal of difference between this (obvious) principle, and the NAP understood as a foundational and absolute moral rule.
So I do not want to claim too much for this essay. Even if my arguments are successful, they do not show that libertarianism is false. At best, they show that one particular way of arguing for it is unsound. And indeed, as someone who identifies as libertarian myself, my hope in knocking down this one unsound argument is that doing so will clear the field for better arguments for a position I think has considerable merit.
But I do not want to claim too little for this essay, either. Even if the NAP has not played a dominant role in academic philosophical defenses of libertarianism, it has played some role. Robert Nozick’s libertarian theory, for instance, is largely an exploration of the implications of a theory of rights as side constraints, where the content of those side constraints is understood as a libertarian constraint “that prohibits aggression against another.” Footnote 12 And Michael Huemer’s recent book identifies a “nonaggression principle in interpersonal ethics” as one of the three foundations of his libertarian position. Footnote 13
Moreover, outside of the academy, the NAP has played a tremendous role in shaping more popular understandings of libertarianism. Apart from Rothbard, few individuals played as significant a role in defining and popularizing libertarian ideas in the twentieth century (and beyond) as Ayn Rand. Footnote 14 And Rand, like Rothbard, believed that a prohibition on aggression was a core principle of interpersonal morality. Footnote 15
It is therefore worth taking the NAP seriously. If academic libertarian philosophers have generally not gone the route of Rothbard and Rand, it is worth explaining why. And if the more popular understandings of libertarianism (held by both supporters and critics of the view alike) rest on a mistake, then that is certainly worth pointing out too, regardless of whether one is interested in burying libertarianism or rebuilding it.
II. Aggression and Property Rights
Consider the following hypothetical example:
Trespass: George is walking across a field when David jumps out from behind a bush, grabs him, and gently but forcefully removes him from the area.
In this example, it certainly looks as though David is the one committing an act of aggression against George. After all, George is simply walking along, minding his own business, when David comes from out of nowhere, physically lays hands on him, and drags him off the field.
But suppose that the field in this example is David’s property, and that George is knowingly trespassing across it. Footnote 16 In that case it is clear that most libertarians would say that it is actually George who has committed aggression. David’s violence, they would say, is simply an appropriate response to George’s impermissible intrusion upon his land.
What this shows, I think, is that for libertarians the concept of “aggression” is not a morally neutral one. Which of two parties to a conflict can properly be said to have committed aggression against the other does not depend solely on natural, observable facts such as who touched whom first. Instead, what counts as aggression is a function of the underlying property rights. If David owns the field, then George was the one who committed aggression by violating David’s property rights, regardless of the fact that David “threw the first punch,” so to speak. If, in contrast, it is George who owns the field, then David was committing aggression by trespassing on George’s property, and/or by physically assaulting George’s person.
Whether an act counts as aggressive or not thus depends, for the libertarian, on who holds the relevant property rights. But we cannot know who holds the relevant property rights, on this view, without making a moral judgment. The fact that the law holds that somebody has a property right in some object is neither here nor there for the libertarian, for if the law merely asserts the existence of a property right that comports with the moral standard of natural justice, then it is superfluous, while if it asserts the existence of a right that conflicts with that standard, it is illegitimate and void. Footnote 17 Either way, the important question as far as determining what counts as aggression is concerned is not what the law says about who owns what, but who is morally entitled to claim ownership over what.
The libertarian concept of “aggression” is thus a moralized concept, insofar as we cannot determine what counts as aggression without making a moral judgment about who has legitimate property rights in what. Footnote 18
Now, in and of itself, there is nothing necessarily objectionable about this. Moralized concepts are not problematic as such. But our use of them can lead to problems, if we are not careful. Footnote 19
One potential problem has to do with the danger of logical circularity in the libertarian argument for strong rights of private property. It is tempting, and natural, for libertarians to argue that property rights are necessary to protect people from aggression, and that the legitimacy of property rights is thus something that follows, logically, from the NAP. But if what counts as “aggression” is, as we have seen, itself a function of the underlying property rights, then this sort of argument will not work. For it will amount to saying that property rights follow logically from the NAP, which itself logically follows from property rights! On a moralized understanding of aggression, the argument from the NAP to property rights presupposes the truth of the very conclusion it is purported to establish. Footnote 20
That does not mean, of course, that libertarians cannot adhere to a moralized concept of aggression and still support strong rights of private property. It just means that this one particular argument for property rights won’t work. Given the strong intuitive appeal of that argument, that’s a cost. But it might be one that, all things considered, libertarians rightly judge to be worth bearing.
There are, however, two other potential problems with this particular way of talking about aggression. The first is that it turns the NAP into a principle that, in and of itself, is utterly devoid of substantive content. After all, the NAP simply tells us that aggression is impermissible. But on the moralized understanding of aggression, what that actually means is determined entirely by the underlying theory of property to which the NAP is attached. The underlying property rights tell us who is entitled to what, and presumably also what sorts of acts are to count as a violation of those entitlements. Moreover, because the relevant property rights for libertarians are moral rights, they already have their normativity built into them. In other words, part of what it means to say that Jones has a property right in a ham sandwich is that others are not morally permitted to take or use that sandwich without Jones’s consent. But if all that is built in to the underlying theory of property rights already, then what does the NAP add? And what is to stop various non-libertarian theories from endorsing the NAP as well? Even a radical egalitarian could agree that people ought not to aggress against the legitimate property of others — he or she will just have a very different view about what legitimate property rights people actually have!
III. The First Punch Theory of Aggression
The above comments might be taken to suggest that the problem with the NAP is simply that “aggression” has no fixed meaning, and so depends entirely for its content on the underlying theory of property rights to which it is attached. And this is indeed a problem for libertarian approaches that insist upon a moralized definition of aggression where the meaning of the term is fixed almost entirely by the underlying theory of property rights. But another problem emerges when we recognize that “aggression” does have something of a fixed meaning in ordinary language, and that on this meaning, libertarians are not necessarily opposed to aggression at all.
In light of our discussion of the Trespass example, let’s call the theory implicit in the ordinary use of the term “aggression” the “First Punch Theory of Aggression.” On this theory, to aggress against another person is to commit something like the common law offense of battery, or perhaps assault, without the justification of self-defense. Footnote 21 So, my punching you in the nose is aggression, as is my grabbing your wallet out of your hand. And so, perhaps, is my threatening to punch you in the nose unless you hand over your wallet of your own accord.
Libertarians often talk about the NAP in a way that presupposes something like the First Punch Theory of Aggression. Recall, for instance, Rothbard’s definition of aggression as “the initiation or threat of physical violence against the person or property of anyone else.” Footnote 22 Similarly, Ayn Rand writes that “the basic principle of the Objectivist ethics is: no man may initiate the use of physical force against others,” and that “man’s rights can be violated only by the use of physical force.” Footnote 23
By linking the concept of aggression to direct, physical violence, libertarians make the NAP seem like an appealing, commonsense moral principle. After all, most of us think that initiating direct physical violence against others is wrong, at least in the vast majority of circumstances. Footnote 24 So if all the NAP is saying is that such initiations of violence are not permissible, then the NAP looks like a pretty plausible moral principle.
But, of course, that is not all the NAP is saying, at least as it is normally understood by the libertarians who endorse it. Understanding aggression as the initiation of direct, physical violence makes sense of the fact that libertarians want to prohibit various crimes against persons such as battery and rape. But it does not seem to make nearly as much sense of the libertarian desire to prohibit or punish crimes against property such as trespass, breach of contract, and fraud. In Trespass, George might be violating David’s property rights by walking across his field without his consent, but it would be quite a stretch to describe him as initiating physical violence or force against either David or David’s property. George is no doubt doing something wrong, but the libertarian analysis seems to misdiagnose the nature of the offense.
Similarly, consider the following case:
Fraud: Sharon sells a painting to Tom, claiming that it is an original Monet, but knowing that it is not.
Most libertarians would claim that Sharon has violated Tom’s property rights by defrauding him. And most non-libertarians would agree. But while Sharon’s behavior is uncontroversially wrong, it would be grossly implausible to characterize it as involving physical force or violence. Footnote 25
What this shows, together with Trespass, is that not all violations of property rights can be properly understood as involving the initiation of physical force. And so, in supporting property rights, libertarians are in a certain respect doing more than simply supporting a prohibition of aggression, as aggression is normally understood.
In another respect, however, libertarian defenders of property rights are supporting something less than a prohibition of aggression. For if property rights are enforced by direct, physical violence, but their violation does not (necessarily) involve the initiation of direct, physical violence, then libertarians who support the enforcement of property rights are actually supporting the initiation of physical violence. Footnote 26 Kicking George off of David’s land, or seizing Tom’s money out of Sharon’s hands, involves the use of physical violence to redress a violation of property rights that did not itself involve the use of physical violence. In these cases, it is the defenders of property rights who are throwing the first punch. Footnote 27 Perhaps they are not wrong to do so. But even if that is true (as I suspect it is), that simply shows that throwing the first punch is, notwithstanding libertarian claims to the contrary, not always wrong.
It is this kind of concern, I think, that leads many of libertarianism’s most influential critics to question the degree to which libertarians are genuinely committed to individual freedom and noninterference. Footnote 28 Critics like G. A. Cohen argue that libertarians say that they do not like people being bossed around by physical force and the threat of force. But they do not seem to mind force at all when it is used in the service of property. So powerful is this objection, in fact, that even some notable libertarians have occasionally found it compelling. Something very much like it, for instance, seems to have led the great libertarian theorist, Herbert Spencer, to question the compatibility of libertarian freedom with the institution of private property in land. Footnote 29
All of this poses a problem for the libertarian. To be clear, however, the problem is not necessarily that libertarians are wrong to endorse property rights. It is not even that they are wrong to endorse property rights on the basis of considerations about freedom and aggression. The problem is that the nonstandard, moralized notion of “aggression” they use cannot bear the moral weight that libertarians seem to want to put on it.
Much of the NAP’s appeal stems from the way it trades on the ordinary, nonmoralized, First Punch Theory of Aggression of aggression. But, in reality, the NAP is not actually about aggression in this ordinary sense at all. It is about property rights. What libertarians really mean when they say that aggression is impermissible is not that it is always impermissible to throw the first punch. It’s that it is impermissible to throw the first punch unless you are doing so in order to defend property rights. Maybe that is a defensible moral claim. Maybe it’s not. But the way libertarians talk about aggression and property obscures the fact that this is precisely the claim being made, and thus hinders our ability to subject the claim to the philosophical scrutiny it deserves.
Where does that leave the NAP as a foundation for libertarian morality? Well, as we have seen, on the moralized understanding of aggression, the NAP cannot serve as a foundation, since it is necessarily parasitic on a prior theory of property rights that does all the real substantive philosophical work.
What about the nonmoralized conception of aggression? Could a libertarian theory be built off a NAP, where aggression is understood according to something like the First Punch Theory of Aggression? Since the First Punch Theory of Aggression is not parasitic on a prior theory of property, it does seem possible that it could serve the role of a bedrock moral principle. What is less clear, however, is whether it could play this role for libertarianism. Since, as we have seen, the enforcement of property rights at least sometimes involves the initiation of aggression in the First Punch sense, it is not obvious that a NAP based on this understanding of aggression will be consistent with the strong rights of property characteristic of the libertarian view. We will return to this question in Section V. For now, let us turn to a different problem with the NAP.
IV. Absolutism
Libertarians like Rand and Rothbard do not merely hold that aggression is sometimes wrong. Nor do they hold that the wrongness of aggression is merely one fact among others to be taken into account in determining the permissibility of an action. Both of these claims are quite plausible, and indeed, are almost certainly claims to which most non-libertarians would assent. But they are not the claims that libertarian defenders of the NAP make.
Instead, this type of libertarian claims that aggression is always wrong, and that its wrongness is so weighty as to override all, or almost all, competing moral considerations. Consider, by way of an example, the libertarian view of the welfare state. Libertarians who adhere to the NAP sometimes argue against the desirability of the welfare state by arguing that it is ineffective or counterproductive in its efforts to aid the poor. Footnote 30 But regardless of how much libertarians might believe in these arguments, and regardless of whether those arguments might actually be sound, there is nevertheless an important sense in which they are entirely beside the point. This is because for the libertarian who subscribes to the NAP, it simply does not matter whether the welfare state is effective, or whether the poor would be better off without it. Efficiency and the relief of poverty might be genuinely valuable moral goals. But aggression is a trump card. And because the welfare state relies on the threat of violence to obtain its funding, it is by that very fact morally illegitimate irrespective of how attractive or unattractive it might be on other moral dimensions. Footnote 31
In this respect, the libertarian prohibition of aggression is a relatively absolutist one. In the most basic sense, a prohibition is absolutist if it requires compliance in all circumstances, such that it can never be justifiably infringed and must be fulfilled without any exceptions. Footnote 32 Beyond this basic understanding, there are a variety of somewhat more technical ways in which the idea of absolutism might be understood. Footnote 33 Fortunately, for our purposes, the differences do not appear to matter. Most statements of the libertarian NAP seem to be absolutist in all of the relevant senses. Still, it is worth thinking carefully about the different ways in which a theory might be absolutist, if only so that we can better understand the different ways in which libertarianism might be modified to avoid some of absolutism’s more unpalatable implications.
First, a theory can be absolutist if the rules that constitute the theory admit of relatively few, or relatively insignificant, exceptions. This kind of absolutism has to do with the content of the rules. Is lying wrong in every conceivable situation? Or is it okay if it is done in order to save someone’s life? What about to spare someone’s feelings?
Second, a theory can be absolutist if the moral force of its rules is sufficient to override all or most other competing moral considerations. Perhaps we have a duty not to lie but also a duty of beneficence. What happens if these duties come into conflict? To the extent that the rule against lying serves as a kind of “trump card” in such conflicts, it is an absolutist rule.
Third, and closely related to the first sense, a rule will be absolutist if it is relatively unlimited in scope. A rule might not apply in certain situations, not because those cases are exceptions to the rule, but because the rule applies only within a certain specified context, and this situation falls outside of that context. So, for instance, a group might hold that the prohibition on lying is meant to apply only to interactions among members of the group, and not to interactions between members and outsiders.
Fourth and finally, a theory can be absolutist in the sense that its deontic rules do not admit of thresholds at the lower end. Footnote 34 A rule that is not absolutist in this sense might incorporate certain de minimis principles, allowing that my dropping a single unwanted hair onto your property does not meet the requisite threshold for trespass. Absolutist rules admit no such principles.
Moral prohibitions on aggression need not be absolutist. One could endorse a general moral presumption against aggression, for instance, while nevertheless (i) holding it to be a presumption that could be overridden in a wide range of important cases if there were sufficiently weighty moral reasons to do so, or (ii) holding it to be limited in scope or by a consequentially-specified threshold. There are even some libertarians who endorse precisely such a nonabsolutist view. Footnote 35
Most libertarians who endorse the NAP, however, endorse a version of the principle that is relatively absolutist in nature. Rothbard, for instance, explicitly states that the right to be free from invasion is “absolute,” and holds that even if one were to find oneself on a lifeboat with too many people and not enough seats, the NAP would still obtain in its full, rigid form. Footnote 36 Robert Nozick, who at one point describes himself as supporting a “libertarian side constraint that prohibits aggression against another,” is explicitly agnostic on the question of absolutism, writing that “the question of whether side constraints are absolute, or whether they may be violated in order to avoid catastrophic moral horror . . . is one I hope largely to avoid.” Footnote 37 He then proceeds to avoid the question not just largely but entirely; but what he doesn’t say in the remainder of the book is at least as important as what he does say. Nowhere in the book is there any mention of aggression being justified in order to provide funding for interstate highways, or clean air, or other forms of economic public goods. And nowhere is there any suggestion that aggression might be justified in order to provide even the most minimal of social safety nets. Footnote 38
A. There is always an exception
An absolutist prohibition on aggression seems problematic for at least three reasons. The first problem is common to all absolutist prohibitions — namely, for any action X that the prohibition says it would be wrong to do, it is relatively easy to come up with cases in which X seems morally permissible, if not morally obligatory. Lying to others might well be morally impermissible, as a general rule, but lying to the Nazis to save the life of an innocent Jew is certainly not impermissible, and might very well be a moral duty.
Similarly, committing acts of aggression against others might be generally wrong. But do we really believe, on reflection, that it is always wrong, in all circumstances? Cases in which a relatively minor act of aggression can produce a very large individual or social benefit appear to pose a challenge for this thought. Consider these two famous philosophical thought experiments:
Jim and the Indians: Jim finds himself in the central square of a small South American town, where a military officer is about to execute a random group of twenty Indians as a warning to antigovernment protestors. In honor of Jim’s arrival, however, the captain offers Jim the opportunity to free nineteen of the Indians — but only by killing one of them himself. Footnote 39
The Trolley Problem: Five people are tied to a trolley track, and will be run over and killed by the speeding trolley heading their way if nothing is done. You can divert the trolley from its course by flipping a switch, but doing so will send the trolley down another track to which one person is tied, and he will be killed instead. Footnote 40
Most people have a very strong intuition that it is at least permissible to kill one Indian in the first case, and to flip the switch in the second. Footnote 41 But killing an innocent person, whether with a gun or the flip of a switch, is an act of aggression if anything is. The believer in an absolutist version of the NAP must therefore hold that both of these acts of killing are impermissible, and indeed that they would remain impermissible no matter how many lives one might save by performing them. Footnote 42 As Murray Rothbard writes, in the context of another example, “The consequence . . . would not influence such a libertarian, the ‘absolutist’ libertarian, in the slightest. Dedicated to justice and logical consistency, the natural-rights libertarian cheerfully admits to being ‘doctrinaire,’ to being, in short, an unabashed follower of his own doctrines.” Footnote 43 If aggression is absolutely wrong, then it is wrong no matter how bad the consequences of nonaggression might be.
The examples of Jim and the Indians and the Trolley Problem are, of course, highly contrived. Footnote 44 But the libertarian commitment to an absolutist principle of nonaggression has counterintuitive implications in more realistic cases as well. Suppose, for example, that the government was considering a policy that would provide lifesaving vaccines to young children in the developing world, and that it would fund this policy by imposing a very small tax on the very wealthy. Those who are subject to the tax, it seems reasonable to suppose, will suffer only a very small setback to their personal and financial well-being as a result of the tax. Perhaps they won’t even notice it. On the positive side, in contrast, many young children could be expected to avoid premature, painful death.
When presented with examples like this, libertarians’ first instinct is often to rebel against the factual claims stipulated in the example. Well-intentioned policies, they point out, often produce unintended but disastrous consequences. And we can’t even count on governmental policy to be all that well-intentioned, anyway, since public choice economics teaches us that government agents generally produce policies that serve their own self-interest and the interests of their most valuable political supporters. A new vaccination program is more likely to be a bit of political pork for the pharmaceutical industry than it is to do any real good for those in need.
In the context of a debate over public policy, these objections might have considerable merit. In the present context, however, they are irrelevant. For our purpose in discussing the vaccination example is not to determine whether such a policy would be justifiable in an all-things-considered sense, but to gain a better understanding of the implications of the NAP as a moral principle. The relevant question, then, is what implications the NAP would have if the stipulated facts of the case were true. If, as I believe, the NAP would have the implausible implication that even an effective and virtually-costless vaccination program would be absolutely morally impermissible, then this tells against the plausibility of the NAP as a moral principle, whether or not any vaccination programs in the real world are effective or as cheaply provided in the way stipulated by the example. Footnote 45
B. Is nonaggression all there is?
Libertarians generally view the NAP not merely as one source of rights among others, but as the only source of rights. And, from the position of an absolutist libertarian, this makes a certain kind of sense. After all, if there were other sources of rights besides the NAP, then it might be possible for those rights to conflict with the rights generated by the NAP. And within the context of an absolutist morality, there is simply no room for this sort of moral conflict. Footnote 46
So, for example, libertarians believe that we have a right not to be subject to the initiation of physical violence. Suppose, however, they believed that in addition to this right, we also have a right that others rescue us, at least in so-called “easy rescue” situations. The more rights the better, right? Well, no. To have a right is, in the normal case, to have a certain kind of enforceable claim on others. Thus, if you have a right to be rescued, part of what this usually entails is that it is morally permissible for you (or an agent acting on your behalf) to use force to get others to rescue you. But someone who merely fails to rescue you has not, by that failure, initiated force against you. Therefore your use of force against him would constitute an initiation of force. But this, of course, is ruled out by the NAP. The two rights are thus logically incompossible. Footnote 47
Regarding the NAP as the only source of rights (or, at least, the only source of enforceable rights) is, thus, necessary to maintaining the absolutist status of that principle. But it is also a second source of that principle’s deep implausibility. For, consider just how much it rules out.
Libertarians almost always say, for instance, that they are opposed not merely to the initiation of force, but to the initiation of force or fraud. Sometimes, they base this claim on the belief that fraud is really a kind of force. Footnote 48 But this claim cannot stand up to philosophical scrutiny. In our earlier case, Fraud, Sharon is no doubt doing something wrong to Tom. But nothing she does in that example bears even the remotest resemblance to physical violence. If, however, fraud is not a kind of force, then a prohibition on fraud will itself involve the initiation of force, by precisely the same logic as that involved in our discussion of the right to an easy rescue above. Footnote 49 Intuitively, it seems obvious that fraudulent transfers ought to be prohibited. But the libertarian adherent of an absolutist version of the NAP cannot embrace this conclusion, and must, it seems, grant the permissibility of fraudulent transfers.
Of course, almost no libertarians actually do grant the permissibility of fraudulent transfers. But this fact, by itself, doesn’t tell us anything about whether they are logically required to do so, by their own basic moral commitments. And, anyway, there are other cases in which libertarians do explicitly grant the permissibility of intuitively immoral actions, simply because the absolutist form of the NAP leaves no room for any additional and potentially competing rights.
Take, for example, the issue of children. Most of us believe that parents have a moral obligation to provide for the basic needs of their young children. Most of us believe, moreover, that these obligations are enforceable ones. A parent who allowed his or her one-year-old child to starve to death would be violating the rights of that child, and others would be acting justifiably if they were to force entry into the parent’s property to provide the child with lifesaving food. Footnote 50
Such an enforceable obligation, however, seems incompatible with the absolutist formulation of the NAP. And at least one prominent libertarian has recognized this, and decided to bite the rather sizeable bullet such recognition forced into his mouth.
Applying our theory to parents and children, this means that a parent does not have the right to aggress against his children, but also that the parent should not have a legal obligation to feed, clothe, or educate his children, since such obligations would entail positive acts coerced upon the parent and depriving the parent of his rights. The parent therefore may not murder or mutilate his child, and the law properly outlaws a parent from doing so. But the parent should have the legal right not to feed the child, i.e., to allow it to die. The law, therefore, may not properly compel the parent to feed a child or to keep it alive. Footnote 51
In making this point, Rothbard cites the late nineteenth-, early twentieth-century American libertarian, Benjamin Tucker, who wrote similarly that “under equal freedom, as it [the child] develops individuality and independence, it is entitled to immunity from assault or invasion, and that is all. If the parent neglects to support it, he does not thereby oblige anyone else to support it.” Footnote 52 Libertarians like Rothbard and Tucker can grant — and Rothbard does grant — that parents might have a moral obligation to feed their children. But if they are to cling to the absolutism of the NAP, they must, and do, reject the claim that this moral obligation is a properly enforceable one. For if children’s rights are enforceable, then force may properly be used against parents who have done nothing violent themselves. Rothbard’s position on children, as deeply counterintuitive and repugnant as it may be, thus seems to be the correct logical implication of a commitment to an absolutist understanding of the NAP.
C. Aggression everywhere! All the time! Footnote 53
One way of thinking about the libertarian prohibition on aggression, suggested by Robert Nozick, is to imagine individual rights as a kind of “line (or hyper-plane) [that] circumscribes an area in moral space around an individual.” Footnote 54 Actions that “transgress the boundary or encroach upon the circumscribed area” are to count as infringements of those rights. And, in general, such infringements may be justly prohibited. Murder, theft, and assault all involve the crossing of an individual’s moral boundaries without his or her consent, and are all therefore properly criminalized by the libertarian minimal state.
On its face, a view such as this would seem to provide a very strong basis for opposing environmental pollution. And indeed, several prominent libertarians have drawn precisely this conclusion. So, according to former Reagan White House official Martin Anderson,
Just as one does not have the right to drop a bag of garbage on his neighbor’s lawn, so does one not have the right to place any garbage in the air or the water or the earth, if it in any way violates the property rights of others. Footnote 55
Similarly, Murray Rothbard writes that the “vital fact” regarding air pollution is that
the polluter sends unwanted and unbidden pollutants — from smoke to nuclear fallout to sulfur oxides — through the air and into the lungs of innocent victims, as well as onto their material property. All such emanations which injure person or property constitute aggression against the private property of the victims. Air pollution, after all, is just as much aggression as committing arson against another’s property or injuring him physically. Air pollution that injures others is aggression pure and simple. Footnote 56
So far, so good. In fact, it might seem as though libertarianism’s strong stance against aggression serves, in this case, to render the theory more morally attractive by providing much stronger grounds for opposing pollution than alternative political ideologies, such as Rawlsian liberalism, are capable of providing. Footnote 57 Insofar as industrial pollution violates the property rights of landowners downriver, nearby foresters whose trees are damaged by acid rain, Footnote 58 or individual persons whose lungs are intruded upon by airborne pollutants, that pollution constitutes aggression and should be prohibited. Indeed, as long as we make the not unreasonable assumption that some of the harms associated with global climate change constitute a violation of property rights, there is no reason why the libertarian framework could not be extended to address even this most difficult and challenging of environmental problems. Footnote 59
Here again, though, the absolutism of the NAP proves to be its undoing. For if all aggression is to be regarded as a violation of rights, no matter how trivial, then the same must be true of all pollution that infringes upon the property rights of others. David Friedman, himself a libertarian but a critic of the kind of natural-rights position espoused by Locke, Nozick, and Rothbard, was the first to point out the problem:
[C]arbon dioxide is a pollutant. It is also an end product of human metabolism. If I have no right to impose a single molecule of pollution on anyone else’s property, then I must get the permission of all my neighbors to breathe. Unless I promise not to exhale. Footnote 60
Light pollution, too, constitutes a physical, nonconsensual invasion of one’s neighbor’s property. This is obvious enough if we consider, as Friedman goes on to ask us to do, a case in which you fire a thousand megawatt laser beam at your neighbor’s front door. But is shining a flashlight at your house any less a violation of your right against nonconsensual boundary crossings? Is lighting a match within eyesight of your property? Footnote 61
If any unwanted “boundary crossing” against one’s person or property constitutes aggression, and if all aggression, no matter how trivial, constitutes a violation of rights, then it is difficult to avoid the conclusion that almost everything we do in a modern, civilized society involves the violation of rights. Every time you drive your car to work, you are contributing to air pollution that will invade others’ property in their bodies and their land; every time you talk within earshot of another person’s property, you are creating air pollution that causes physical vibrations in people’s eardrums without their consent; every time you turn the lights on in your house, you launch an assault of photons into your neighbors’ property. The only way to avoid aggressing against others, it seems, is to avoid social coexistence altogether.
Libertarians have sought to avoid the troubling implications of their view on questions of pollution in various ways. Nozick did so by apparently giving up on the traditional libertarian idea of individual rights being protected by a strict property rule and adopting instead the position that they are protected by a weaker liability rule — one that allows boundary crossings in a certain range of cases so long as compensation is paid, and which therefore holds that pollution should be permitted whenever the [social] benefits are greater than the costs.” Footnote 62 Rothbard, in turn, argued that pollution should be understood as a nuisance (rather than a trespass) that should be prohibited only if objective harm could be demonstrated, and only if a strict causal connection between the source of pollution and its victim could be demonstrated beyond a reasonable doubt. Footnote 63
Both of these approaches, however, are problematic in several important respects. Regarding Nozick’s proposed solution, we might ask in what sense do you own your property if others can use it as they wish so long as they compensate you for it afterwards? Would slavery cease to be wrong “if only we could be sure that the slave is receiving at least as much compensation as would have induced him to agree to enslavement”? Footnote 64 And regarding Rothbard, why should the demonstration of harm be required for a legal cause of action when it comes to the kinds of “invisible” and “insensible” boundary crossings that constitute nuisance, but not when it comes to the visible, sensible boundary crossings involved in trespass? If only harmful boundary crossings are to be prohibited, then why shouldn’t we hold my harmlessly walking across your property without your consent to be permissible? And if harmless but unwanted trespass is to be prohibited, why not harmless but unwanted nuisances too? Footnote 65
The problem here is a deep one. An absolutist prohibition on aggression understood as the physical invasion of other people’s bodies or property seems most at home in the kind of “micro-level” examples so often employed by libertarians, involving discrete interactions between one person and another. Footnote 66 It is not clear that such a prohibition can function in the sort of world revealed to us by environmental philosophers — one marked by the complex interaction of large numbers of dispersed individuals, in which the long-term and long-range effects of any one person’s actions are marked by the kind of non-linearities and threshold effects that characterize the problem of environmental pollution. In a populous and complexly interrelated world in which almost all of our actions produce physical effects on others without their consent, an absolutist prohibition on aggression seems an impossibly high standard to meet.
V. What ’ s Left for Nonaggression?
I have been arguing in this essay that there are good reasons to reject the Nonaggression Principle as it is usually understood by libertarians. But the problem with the libertarian NAP is not its identification of aggression as a serious moral wrong. The problems are, first, that it really has more to do with property rights than it does with aggression on the ordinary meaning of that term, and second, that it takes the wrongness of aggression to be an absolute moral rule, rather than a serious, but potentially overridable, moral principle. Footnote 67 A version of the NAP could be identified that avoids both of these problems. Such a principle would be quite different from the principle that Rothbard articulated, and that has come to dominate both popular and academic understandings of libertarianism since then. But it would be a more philosophically defensible principle, and one that would retain a significantly libertarian edge. In this section, I will sketch what such a principle might look like, and what some of its implications might be.
A. Distinguishing aggression from the violation of property rights
Such a nonaggression principle would, in the first instance, be about aggression in the ordinary sense of that term, described in Section III as the “First Punch Theory of Aggression” of aggression. It would, in other words, be a norm that has as its immediate object the use of direct physical force against other human beings. It is wrong to physically attack innocent persons; it is wrong to rape them; it is wrong to murder them; it is wrong to kidnap them. And a NAP based on the First Punch Theory of Aggression would prohibit the initiatory use of force in these cases, among others.
A norm against aggression understood in this sense would seem to have a strong basis in the traditional libertarian commitment to self-ownership. Footnote 68 Each individual owns his own body, and therefore has the right to exclude others from using it without his consent. And this is precisely what a norm against aggression does. Footnote 69
There is a sense, then, in which the sort of NAP I am proposing is, like the more common libertarian NAP, based upon a moralized understanding of aggression. After all, whether a given act counts as aggression will depend on whether it violates a person’s ownership rights over his or her body. And that, of course, is a moral question. Consider the following case:
Labor Theft: Jerry tells Mike that if Mike does some physical labor on Jerry’s car, Jerry will pay him $100 at the end of the day. Mike does the work, but at the end of the day Jerry refuses to pay him.
Suppose Mike were to use a reasonable amount of physical force to prevent Jerry from driving away without paying him. Should we classify this as “aggression”? I think not. And the reason is that such physical force would not violate Jerry’s rights over his body. In committing an act of fraud, Jerry has forfeited the right not to have a certain amount of physical force used against him. Of course, Mike is using physical force. And, indeed, if we limit our focus to physical force alone, then Mike is the one who initiates force. But he is not the aggressor in this case, because the force he uses is in response to Jerry’s nonviolent violation of his rights.
The real problem with the libertarian NAP, then, is not merely that it involves a moralized conception of aggression. The problem is that it runs together different parts of morality, conflating physical force against a person’s body with the violation of a person’s property rights in external goods, and thereby conveying the misleading impression that they are both wrong to the same degree, in the same way, and for the same reasons.
The truth is that physical violence against persons is one thing, and the violation of property rights in external resources is something else. “Everything is what it is.” Footnote 70 A (nonabsolute) norm prohibiting the initiation of physical violence against persons is relatively uncontroversial. But the justification of property rights is a more difficult and contentious business. We cannot avoid this difficulty by pretending the task is easier than it actually is, by pretending that a support for property is just the same thing as an opposition to physical violence.
So where does that leave the case for property rights in external resources? If we take self-ownership and an opposition to physical force against people’s bodies as a starting point, how far can those principles take us in the direction of a theory of property in nonbodily resources?
It can probably take us at least part of the way. As mentioned above, Footnote 71 the common-law prohibition of battery covers not merely the use of force against persons, but also the use of force against certain properties very closely connected with persons, such as a hat or a purse. And this seems sensible. There is good reason to regard property that is intimately connected with a person as an extension of the person himself. Someone who slowly and deliberately walks up and knocks your hat off your head is committing an act of aggression against you, as is someone who attempts to snatch your purse from your arm, or the acorn from your mouth. We ought to categorize these acts not merely as wrongs, but as wrongs of roughly the same sort as a punch or a shove.
What about property that is less directly connected with one’s body? Here, the implications of a nonmoralized NAP are less clear. The libertarian who wants to get a theory of nonpersonal property from the NAP faces two problems: the first is that some violations of property rights don’t seem to violate the NAP at all; the second is that (in at least some cases) the establishment of property rights does.
To take the first point first, the Fraud case that we saw earlier can reasonably be construed as involving the violation of property rights, but not the initiation of physical force. And there are many cases of this sort. Here is another.
Embezzlement: Samir writes a program to embezzle money from his office’s computer, moving money from the business’s account into his own personal account.
In this example, Samir is not using physical force against anybody. Nor is he threatening to use physical force. He is, to redescribe the example with a nontechnological analogy, simply picking up money that has not been sufficiently locked-up by its owner. In doing so, he almost certainly acts immorally, and in a way that both is and ought to be illegal. But he does not violate the NAP, at least if we understand “aggression” in the ordinary sense.
Of course, other cases involving the violation of nonpersonal property will involve aggression. Consider:
Extortion: The local mob boss tells a business owner that unless she voluntarily pays him a large amount of “protection” money, he is going to start taking money out of her cash register himself. If she resists his attempts to do so, he tells her, she is likely to meet with an “unfortunate accident.”
In this case, the mob boss’s request for money is backed by the threat that if the store owner says no, he is just going to take it anyway. And if she tries to stop him from doing that, he is going to physically injure her. So the request for money is ultimately backed up by a threat of physical violence against her person. And this, according to even the nonmoralized version of the NAP, is impermissible.
Before we become too optimistic about the prospects for this strategy, however, we should consider the second problem for the attempt to derive property rights from the NAP: namely, that the establishment of such rights is itself a violation of the NAP. Consider the following scenario.
Appropriation: John puts a fence around a piece of land that had previously been open to use by all, builds a hut, and plants some crops. John tells Henry, who had sometimes used the land for foraging in the past, that if he crosses the fence without John’s permission, John will use whatever physical force is necessary and appropriate to remove him.
In this case, it looks like it is the act of establishing a property right itself that is aggressive, in the ordinary sense of that term. John’s claim to a property right in his land is backed by the threat of physical violence. And the same, of course, is true of other property rights as well, whether they were acquired by acts of original appropriation, trade, or gift. So at the same time that a commitment to nonaggression tells against attempts to acquire the property of others by threats of violence, it also seems to tell against the establishment and maintenance of property rights themselves.
How serious are these problems for libertarianism? To my mind, they only constitute a worrisome objection if we insist a priori that the NAP ought to be enough all by itself to resolve all questions concerning the morality of property and theft. And I see no reason to think such insistence is warranted. Nonviolent theft is wrong. But there is no reason to think that the explanation of its wrongness ought to be the same explanation we would give for the wrongness of direct violent assaults on a person’s body. Why assume that vastly different kinds of moral questions need to be addressed by a single moral principle?
Similarly, there seems to be no reason to assume from the outset that the wrongness of aggression ought to settle the question of property rights, all by itself. To claim a property right is, to be sure, to claim the right to threaten or use physical force against others to defend that right. And, all else being equal, initiating force against others in that way is a bad thing to do. But all else isn’t necessarily equal. It would, for instance, be very strange indeed if the justification of a system of property rights had nothing whatsoever to do with the sorts of lives such a system allowed people subject to it to live, or the kinds of social problems it helped them to avoid. Footnote 72
We might assume that the NAP ought to settle these questions all by itself if we wanted to guarantee right at the outset that there would never be any possibility of conflict between the principle of nonaggression and some other moral principle. Like a small, Western town, we might say, we’ve only got room for one sheriff ‘round here. But once we drop the assumption that the NAP ought to be understood as an absolute moral rule, the possibility of conflict becomes much less troubling. And, as we saw in Section IV, there is very good reason to drop that assumption.
B. Nonabsolutism
A principle prohibiting aggression should neither hold in all cases, nor have infinite moral weight. A defensible version of the NAP will therefore be nonabsolutist. And because it is nonabsolutist, it will therefore only be one element of a pluralistic moral theory. After all, once we have accepted that a prohibition on aggression is going to have exceptions, it just doesn’t make any sense to insist that we limit the size of our moral universe to a single moral rule. In fact, we arguably need the other principles to see just where the exceptions to the NAP lie.
A defensible NAP will therefore issue a moral presumption against aggression, and indeed perhaps a very strong presumption. But it is a presumption that can be overridden given sufficiently weighty moral considerations on the other side. The cases discussed in Section IV of this essay constitute an illustrative but non-exhaustive list of situations in which the presumption might be overridden in just this way. Yes, aggression is bad, but if it is necessary to save the life of a toddler from a neglectful (but nonaggressive) parent, then it is justifiable. Yes, aggression is bad, but if the aggression at issue takes the form of a relatively harmless form of pollution that is the byproduct of some personally or socially valuable activity, then again, it is justifiable.
This sort of moral balancing will cause discomfort in those who expect clear, unambiguous guidance from their moral principles. But here, as elsewhere, the problem is with the expectation, not with the principle’s inability to satisfy it. Our principles don’t always provide us with clear, unambiguous guidance in cases of intrapersonal morality, either — about how to balance the sometimes conflicting values of work and family, independence and loyalty, courage and prudence. Why should we expect the business of interpersonal morality to be any easier?
C. How libertarian is it?
The sort of NAP I have defended here makes no pretense of being a complete moral theory. Nor does it pretend to be a complete political theory, or even a completely sufficient ground for deriving a theory of political rights. We should not expect, then, that we will be able to derive libertarianism from it, in the way that Rothbard thinks he can derive libertarianism from the “twin axioms” of nonaggression and homesteading. Whether libertarianism turns out to be defensible will turn not just on the correct understanding of the NAP, but on the way in which the NAP is integrated with the numerous other principles that mark our moral landscape.
One of the major tensions that any libertarian theory based on the NAP will have to resolve is the tension between “right-libertarian” theories that allow for strong rights of property in external resources such as land, and “left-libertarian” theories that do not. Footnote 73 Since, as we have seen, rights of private property themselves can be seen as involving violations of the NAP, there is at least a relatively strong presumption against those rights for anyone who professes allegiance to that principle. Whether or not that presumption can be overridden by considerations internal to the NAP, or external to it, is an issue of the utmost importance. Footnote 74
Supposing, however, that some case can be made for rights of private property, the NAP will push in a strongly libertarian direction regarding the protection of those rights. After all, we have seen above that the NAP condemns the sort of action taken in Extortion. But compare Extortion with this strikingly similar case:
Taxation: An agent of the Internal Revenue Service approaches a local business owner and demands that she pay a large amount of money in taxes. If she refuses, the agent will place an order to have the money seized from her bank account. And if she tries to avoid that, the agent will call the police to seize her, arrest her, and lock her away in jail.
If the mob boss in Extortion acts wrongly, then there is at least a very strong presumption that the IRS agent in Taxation does too. Perhaps this presumption can be overridden. After all, the IRS agent works for a democratically elected government that purports to serve the common good. Whether that is enough to render the coercion she employs justifiable is too large a question to be answered here. But the libertarian view that actions that are wrong when done by private individuals are wrong when they are done by democratic governments too is a powerful one, and one that puts a very heavy justificatory burden on defenders of the state. Michael Huemer’s recent book makes a strong argument that this burden has not yet been met. Footnote 75 I suspect he may be correct.
Still, one might doubt whether a theory can really be libertarian if it takes such an ambivalent stance on the justification of property rights, and the consequent illegitimacy of taxation and redistribution. This concern is understandable. It is, nevertheless, based on two misunderstandings of libertarian thought.
First, it is important to remember that despite the emphasis placed on it by thinkers like Murray Rothbard, the NAP is not, and never has been, the only resource in the libertarian intellectual toolkit. A close study of the libertarian intellectual tradition reveals a theory that is far more pluralistic than is commonly supposed. Footnote 76 In addition to an opposition to aggression, libertarian thought is characterized by (1) an emphasis on the importance of spontaneous order as a source of social organization, Footnote 77 (2) a deep skepticism of authority, especially as manifested in political power, Footnote 78 (3) an appreciation for the mutually-beneficial, wealth-creating, and information-conveying power of voluntary market exchange, Footnote 79 and (4) a commitment to individualism and personal responsibility. Footnote 80 If the libertarian case for private property, and against government redistribution, had to rely on the NAP alone, then it would indeed be on shaky ground. Fortunately for the libertarian, it does not.
Second, it is important to remember that while support for rights to nonpersonal property is an important and distinctive element of the libertarian position, it is not the sole distinctive element. Even if it has no implications for the legitimacy of nonpersonal property, the libertarian commitment to self-ownership still has considerable practical bite, and tells strongly against the legitimacy of paternalism and physical aggression. This, by itself, is enough to render the libertarian position radically distinct from most mainstream political ideologies, and this radicalism is made clear in libertarians’ principled opposition to the war on drugs, to the criminalization of prostitution and other victimless crimes, and to nondefensive military action.
Suitably understood, the NAP is a defensible moral principle. And it is a defensible libertarian moral principle. Its implications for property rights in nonpersonal resources might be less than perfectly clear. But the libertarian case for property rights is about more than just the NAP. And libertarianism as a political theory is about more than just property rights.
