Libertarians sometimes take their analogies too far. From the plausible insight that taxation is like forced labor certain respects, they jump to the entirely implausible conclusion that taxation is forced labor. To libertarians, this move appears to be a penetrating insight into the “essence” of taxation. To non-libertarians, it appears to be a kind of stubborn blindness to the multitude of respects in which the phenomena are different. Other examples, unfortunately, abound.

Take, for example, the oft-repeated libertarian claim that passing a law that requires him to do something is like pointing a gun at someone’s head and telling him to it. (See here and here for two recent examples from my recent exchange with Bryan Caplan) As with other analogies taken too far, there is a kernel of insight here. The insight is that laws are, ultimately, backed by the threat of coercion. The government does not merely request that you pay your taxes. It demands it. And if you refuse and are persistent and effective in your resistance, then at some point men with guns will show up at your door to compel your compliance or to lock you away.

As important as this insight is (and I do not deny its importance), libertarians nevertheless take it too far when they identify the coercive authority of law with a gunman issuing orders. To see this, I want to take a simple pair of examples dealing with a different sort of authority and then draw out two lessons from them.

Example 1: A parent orders his eight year old child to go to his room and stay there for a while as punishment for misbehavior. The parent is willing to use a moderate level of coercion to enforce this command, should it be necessary.

Example 2: A parent points a gun at his eight year old child’s head and tells him to stay in his room for a while as punishment for misbehavior.

The first and most obvious lesson of these examples is that even if laws are coercive, not all coercion is the moral equivalent of pointing a gun at someone’s head. The parent in example 1 is acting in a way that most of us would probably not judge as wrong, while the parent in example 2 clearly is. Alternatively, even if you think the parent is acting wrongly in both cases, the wrong in example 2 is clearly of a different and more serious kind. So not all coercion is morally equivalent.

Can we say anything more about this difference? I think we can, but let’s first move to a different pair of examples that illustrate the thesis in question more precisely.

Example 3: The city government posts a stop sign at an intersection in your neighborhood.

Example 4: I walk up to your car at that same intersection, point a gun at your head, and tell you to “stop.”

Let us grant that both examples 3 and 4 involve the use of coercion. Even still, an important difference between them has to do with the range of options that the two actions leave open to their target, and hence, I think, with the degree to which they involve interference with the target’s negative liberty. In example 4, my action leaves you with precisely two options: stop, or be shot. In example 3, however, your range of options is considerably broader (though still narrower than it would have been had no law been passed). You can stop, or (assuming you are being watched) you can pay a fine. Or, if you don’t want to pay the fine, you can go court and try to appeal. If the appeal doesn’t work, you can, I suppose, either pay the fine or perhaps have your wages garnished. If you somehow escape discharging your financial debt in either of these ways, you can go to jail peacefully when the police come to get you. If you resist and the police use force to try to bring you to jail, you can resist that force to the utmost, or you can give in. If you resist the force to the utmost, then and only then might you get shot. More likely you’ll be pepper sprayed or beaten.

That’s no picnic, to be sure, but it’s certainly a less substantial an infringement upon your liberty than getting shot. And the fact is, the threat of physical violence manifests itself only after you have made a number of decisions that could have gone otherwise, and that would have involved a much less substantial infringement upon your liberty. As a result, your ability to control your own life is substantially greater in example 3 than in example 4.

Another morally relevant distinction between examples 3 and 4 has to do with the form of the restriction. Example 4 involves a single, isolated command whereas example 4 involves a general rule. Liberal theorists have long thought that there is something significant about the generality of law, though they have disagreed about what precisely it is. Friedrich Hayek, for example, sometimes wrote as though truly general laws would not really be coercive at all. And this is quite implausible. What is plausible, on the other hand, is his contention that the generality of laws lessens the impact of their coerciveness on our freedom.

The coercion which a government must still use for this end is reduced to a minimum and made as innocuous as possible by restraining it through known general rules, so that in most instances the individual need never be coerced unless he has placed himself in a position where he knows he will be coerced. Even when coercion is not avoidable, it is deprived of its most harmful effects by being confined to limited and foreseeable duties, or at least made independent of the arbitrary will of another person. Being made impersonal and dependent upon general, abstract rules, whose effect on particular individuals cannot be foreseen at the time they are laid down, even the coercive acts of government become data on which the individual can base his own plans. Coercion according to known rules, which is generally the result of circumstances in which the person to be coerced has placed himself, then becomes an instrument assisting the individuals in the pursuit of their own ends and not a means to be used for the ends of others.(The Constitution of Liberty, p. 21)

So laws, even if coercive, and even if morally wrong, are not the moral equivalent of pointing a gun at someone’s head. Which leads to the second lesson, a lesson at which I will only gesture here. That lesson is that it is a mistake to think of coercion as the essence of law. What is essential to law is its claim to authority. Coercion is important, but both in terms of the way the law perceives itself and in terms of the way it is perceived by the majority of its subjects, coercion is at best of secondary significance compared to the law’s claim to authority.

The same is true of the parent in example 1. Perhaps coercion does lurk in the background behind the parent’s command. But far more important than coercion in understanding the parent-child relationship is the parent’s claim to, and the child’s acceptance of, the parent’s authority to order.

In some ways, this distinction receives its most famous exposition in H.L.A. Hart’s masterpiece of jurisprudence, The Concept of Law, especially chapter 5. In the first half of his book, Hart engages in a thorough critique of the theory of law espoused in John Austin’s The Province of Jurisprudence Determined. The core of Austin’s theory is that law is nothing more than an order backed by a threat, given by someone to whom others are in a habit of obedience, and who himself is in a habit of obedience to no one. Hart’s critique of this position is one of the most devastating takedowns in the history of 20th century analytic philosophy, and libertarians (who are often intuitively attracted to something like the Austinian theory of law) would do well to read it. This recent paper (gated) by Christopher Morris, develops the distinction between coercion and authority in greater depth, and is also well taking a look at.

To be clear, I am not claiming that state’s actually have authority in any normative sense. Hart was a legal positivist, and so whatever notion of authority flows out of his writings is a purely descriptive one, though one that describes, in part, the normative attitudes that citizens and state officials typically have. Nor am I claiming that state coercion is unimportant, or nothing to worry about, or something we shouldn’t have much, much less of. State coercion is a problem. But to solve the problem, we have to accurately understand it. And the gunmen metaphors to which libertarians so often resort are not very helpful.