If I were compelled to summarize the libertarian philosophy’s distinguishing feature while standing on one foot, I’d say the following: Every person owes it to all other persons not to aggress them. This is known as the nonaggression principle, or NAP.

What is the nature of this obligation?

The first thing to notice is that it is unchosen. I never agreed not to aggress against others. Others never agreed not to aggress against me. So if I struck you and you objected, you would not accept as my defense, “I never agreed not to strike you.”

Even an explicit agreement rests on an unchosen obligation. Let’s say you lent me five dollars, I refused to repay the loan, and when you demanded repayment, I said, “Why am I obligated to repay the money?” You would probably reply, “Because you agreed to repay me.” If I replied, “True, but when did I agree to abide by my agreements?,” what would you say? If you said that failure to repay constituted aggression, and I replied that I never agreed not to aggress against you, we’d be back where we started.

Of course this would point the way to absurdity — an infinite regress of agreements to keep my agreements. We would get nowhere. There has to be a starting point.

If I were to ask, “Why do we owe it to others not to aggress against them,” what would you say? I presume some answer rooted in facts would be offered because the alternative would be to say this principle has no basis whatsoever, that it’s just a free-floating principle, like an iceberg. That would amount to saying the principle has no binding force. It’s just a whim, which might not be shared by others. In other words, if a nonlibertarian demands to know why he is bound by the unchosen NAP, libertarians will have answers. Their answers will differ — some will be more robust than others — but they will have answers. At least I hope so.

Now if we have an unchosen obligation not to aggress against others and that obligation is rooted in certain facts, this raises a new question: Might the facts that impose the unchosen obligation not to aggress also impose other obligations? If one unchosen obligation can be shown to exist, why couldn’t the same foundation in which that one is rooted produce others?

To the question “Why do we owe it to others not to aggress against them,” I would respond along these lines: because we individually should treat other persons respectfully, that is, as ends in themselves and not merely as means to our own ends. But some libertarians would reject that as too broad because it seems to obligate us to more than just nonaggression. They might answer the question this way: “Because one may use force against another only in defense or retaliation against someone who initiated the use of force.” But this can’t be sufficient because it amounts to a circular argument: To say that one may use force only in response to aggression is in effect merely to restate the nonaggression principle. One shouldn’t aggress because one shouldn’t aggress. But the NAP can hardly justify itself.

So we need a real justification for the NAP, and the one I’ve offered seems like a good start. The NAP is an implication of the obligation to treat persons respectfully, as ends and not merely as means. Of course this also requires justification. Why should we treat other persons respectfully?

Many libertarians, though certainly not all, approach the question of just conduct — specifically, as it relates to the use of force — from egoistic considerations, such as those provided by Ayn Rand. They say we should never aggress against others because doing so would be contrary to our self-interest: the dishonesty required by a life of injustice would be psychologically damaging, and we’d eventually run out of victims.

Socrates and Plato saw a problem with the first part of this answer. If one could act unjustly toward others while appearing to be just, could unjust conduct serve one’s self-interest? Egoistic libertarians can be asked the same question. What if you could lead an unjust life with a guarantee of the appearance of justice? Must dishonesty be damaging? The same people who would say yes to that question, however, would also say that a person who spins a complicated web of lies to keep the Nazis from learning he is harboring Jews in his attic won’t suffer such damage. If that person can escape harm, why not the unjust liar? Saying that one set of lies is for a good cause doesn’t strike me as an adequate answer. How would a good cause save someone from the harm of “faking reality”?

So it seems that a simple self-interest model doesn’t take us where we want to go: to the unchosen obligation to respect people’s freedom, or more broadly, to treat persons as ends and not merely as means. I would be a little uneasy if a libertarian told me that it is only his self-interest that prevents him from clubbing me on the noggin and making off with my wallet.

And yet, self-interest still might provide an answer. Roderick Long tackles this problem in his extended essay “Reason and Value: Aristotle versus Rand” (PDF). What Long shows, to my satisfaction at least, is that Rand’s notion of self-interest as expressed in her nonfiction essays is too flimsy to support the libertarian prohibition on aggression and the general injunction to treat people respectfully. To be more precise, Long shows that Rand’s explicit writings on ethics are a tangle of at least three different and inconsistent defenses for the nonaggression principle (one of them Kantian — how’s that for irony?).

Before we get to this, however, we must invoke an important distinction that Long emphasizes: instrumental versus constitutive means to an end. An instrumental means is external to the end. A constitutive means is intrinsic to the end; we can’t imagine the end without it. Long uses the example of a man dressing up for evening out (where “dressing up” includes a necktie). Shopping for a tie is an instrumental means. Wearing the tie is a constitutive means — it is part of what we mean by “dressing up.” One can dress up without shopping for a tie, but one cannot dress up without wearing a tie.

We can look at justice, which includes respect for other persons’ rights, in both ways. Does respect for their rights serve our self-interest merely because we would earn good reputations and others will cooperate with us? (This is Thomas Hobbes’s position.) Or is respecting their rights also a constituent of living a good human life? The answer is crucial. In the first case, one’s self-interest could be served by acting unjustly so long as one could appear to be just. In the second case, one could not flourish by acting unjustly even if one could go undetected. As Socrates suggested, it is preferable to live justly with a reputation for injustice than to live unjustly with a reputation for justice.

Long shows that Rand has both instrumental and constitutive elements in her nonfiction writing on ethics; in some places she says a person’s goal should be survival, while in other places she speaks of survival “qua man.” It isn’t entirely clear whether individuals should aim at the longest possible life regardless of the type of life or at a particular type of life regardless of its length. (Her novels appear to take the latter position — suicide is even contemplated by heroic characters.) If it’s the first, then violating someone’s rights might occasionally be to one’s self-interest. Imagine that at 4 a.m. you pass an alley in a deserted part of town where a man is passed out and a hundred-dollar bill is sticking out of his pocket. The chances of getting caught are zero. Do you take the money? If not, why not? An instrumental model of justice should say to take the money. A constitutive model would not.

It might be said that a rational person acts on rational principles even if in particular cases his or her self-interest is not served. But Long points out that such “rule egoism” ends up being no egoism at all, since the rule is followed regardless of its consequences. This approach is deontological, not teleological, as Rand would want it. So the reply is inadequate.

What are the grounds for accepting the constitutive model of virtue, including justice? Turning to Aristotle, Long writes,

For Aristotle, a human being is essentially a logikon animal and a politikon animal.…

To be a rational animal is to be a language-using animal, a conversing animal, a discursive animal. And to live a human life is thus to live a life centered around discourse.

Our nature as logikon is thus closely allied with our nature as politikon. To be apolitikon animal is not simply to be an animal that lives in groups or sets up governments; it is to cooperate with others on the basis of discourse about shared ends.…

Being politikon is for Aristotle an expression of being logikon; just as logikon animals naturally conduct their private affairs through reason rather than through unreflective passion, so they naturally conduct their common affairs through public discourse and rational persuasion, rather than through violence.…

Thus, Long adds, “To violate the rights of others, then, is to lessen one’s humanity.… To trample on the rights of others is never in our self-interest, because well-being cannot [quoting Aristotle] ‘come about for those who rob and use force.’”

One’s goal is to flourish by achieving excellence in those things that make one human — Aristotle says that “the task of man is a certain life, and this an activity and actions of soul with logos.” One cannot flourish if one lives in a nonhuman way. If this sounds like Rand, it’s because her fictional characters understand it, even if her nonfiction essays do not express it unambiguously.

Long concludes,

A truly human life, then, will be a life characterized by reason and intelligent cooperation. (Bees may cooperate after a fashion, but not on the basis of discourse about shared ends.) To a logikon animal, reason has value not only as an instrumental means to other goals but as an intrinsic and constitutive part of a fully human life; and the same holds true for cooperation. The logikon animal, insofar as it genuinely expresses logos, will not deal on cooperative terms with others merely because doing so makes others more likely to contribute instrumentally to the agent’s good; rather, the agent will see a life of cooperation with others as an essential part of his own good.

Aristotle’s book on friendship in the Nicomachean Ethics beautifully elaborates on this point.

If this is right, we owe respect to others’ humanity, via respect for their rights, because the activity manifesting that respect is a constituent of our own flourishing as logikon and politikon animals. We owe it to ourselves to owe it to others. This Aristotelian insight points to an interpersonal moral realm in which the basic interests of others meld in important ways with our own. “To the extent that we are logikon animals,” Long writes, “participation in a human community, together with a shared pursuit of the human good, is a constitutive part of a truly human life.”

But does this show that we owe anything more than nonaggression? It seems so. We abstain from aggressing against others because, as logikon and politikon animals, we flourish by engaging the humanity of other individuals. Clearly, abstaining from aggression is not the only way to engage their humanity, just as aggression is not the only way to deny their humanity. Thus these Aristotelian considerations entail the obligation to treat others respectfully broadly.

One last question remains: Is this obligation broadly to treat other persons as ends and not merely as means a libertarian matter? It is, at least in this way: The obligation broadly to treat other persons as ends and not merely as means is validated by the same set of facts that validate the nonaggression principle. Nonaggression is simply one application of respect. Thus a libertarian society in which people generally thought that nonaggression was all they owed others would be a society that should fear for its future viability qua libertarian society.

Finally, I’m sure libertarians do not have to be reminded that nonaggressive affronts against persons may be responded to only in nonaggressive ways. Neither governmental nor private force may be deployed to counter peaceful offenses. Why not? Because the rule of proportionality dictates that force may be used only to meet force. In other words, some obligations are enforceable and others are not.

(While thinking about this article, I profited mightily by conversations with Gary Chartier.)