I. WHAT IS FORCE?
One of the basic premises of the entire libertarian political philosophy is that the use of force is justified only in retaliation against those who have initiated its use.
In order to make this proposition compatible with other views held by libertarians, it is necessary to extend the use of the word "force" beyond the traditional one, in the strictly physical sense, in which one's muscular movements cause bodily injury or death to another—let us call this the "muscular" sense of the word "force." For example, 1) every libertarian believes that you are entitled to eject someone forcibly from your property if you do not wish him to be there or to call the police so that they can do so; and similarly with the theft or attempted theft of any of your possessions, such as your watch. If the word "force" were limited to the strictly muscular sense described above, then it would be you and not the thief or trespasser who was initiating the use of force by apprehending him—after all he was not using force against you, he was only stealing your watch or walking on your yard. Yet all libertarians would defend the use of force in defending property rights, although they might disagree about the degree of its use which is justified. Similarly, 2) every libertarian believes in the enforcement of contracts after they have been voluntarily entered into: if I have not repaid you the money I have contracted to return to you by a certain time, you have the right to invoke the law, and I may be fined or even imprisoned, although I have not used force (in the muscular sense) against you in any way. (It is sometimes said that the term "aggression" would be a better term to use than "force," but its meaning is not clear as it stands: how, in not repaying my debt to you, have I aggressed against you? It could turn out that I have or that I haven't, depending on what meaning is assigned to the word "aggressed.")
It is important, then, if we are to commit ourselves to the principle that the only justifiable use of force is in retaliation against its initiation, that we be clear as to what we mean by "the use of force." Let us attempt to come at this concept by means of a few examples.
1. I see a man beating his wife and conclude that this is a paradigm case of the use of force, until I discover that she is a masochist who wanted to be beaten and actually asked for it. I now conclude that this action is not properly described as the use of force, since the concept of force involves the element of unwillingness. (Or I may, if I wish, still call it the use of force but qualify this statement by saying that it is a force willingly received and then add that it is only the use of force unwillingly—to the recipient—that the libertarian is talking about when he condemns the use of force. There is no use of force when there is no victim of it, and the person who wants to be beaten is not properly described as a victim of the beating but only as a recipient of it.)
2. I intend to strike the punching bag but you step in the way and I hit your nose instead. Have I used force? As long as you did not wish to be hit in the nose, you are a victim of it, in spite of the fact that my hitting you was accidental. One could then distinguish the accidental from the intentional use of force, and attach greater condemnation and greater penalties to the use of the intentional than to the accidental kind, although I see no reason why both should not count as instances of the use of force.
3. An employer fires an employee, for whatever reason. Liberals and Marxists are fond of citing this kind of thing as an example of force or coercion, but this example will not detain any libertarian for a moment. Unless there was a contract specifying that the employee was guaranteed his job for a certain period, the employee is free to leave it for whatever reason at any time, and the employer is free to terminate the employee's services at any time. (Sometimes there is a contractual agreement that the employer must give 30 days' notice before dismissing the employee and the employee must give similar notice before he quits.) In each case, it is simply an example of the nonrenewal of a voluntary arrangement between them; no use of force is involved.
4. Professor Hayek uses the example of two men who settle in the desert some miles apart, each attempting to make a living there. Both succeed to an extent, but after the expenditure of a year of labor and considerable money, the first man, A, runs out of water. A goes to the second man, B, and asks to buy some water; B refuses. A asks to borrow some, and again B refuses. Obviously A can't continue to live there without water; the refusal of water amounts to either a death sentence for A or the necessity of leaving the desert. Hayek appears to conclude that since water is an essential commodity B is coercing A by refusing him water. Libertarians would not concur in this judgment; the fact that A ran out of water is not B's responsibility. Perhaps (morally) B should give or sell him some water, but A doesn't have a right to B's water, and in any case B did not use force against A. Instead of saying that B forced A to leave the desert, it seems to me much more accurate to say that "the desert has beaten him."
Of course the example could easily be changed so as to include force. Suppose that A attempts at gunpoint to make B give him the water; then he has surely used force (or threat of force). There is no doubt that this is force, and moreover that it is force not used in retaliation against the initiation of force. Thus, if one thinks that A's action is justified, he cannot consistently also think that the initiation of force is never justified.
What then is the use of force? In LIBERTARIANISM (p. 23) I defined "force" (following Tannehill) as "behavior which requires the unwilling involvement of another person or persons." If I hit you over the head, I have used force against you unless you have consented to be hit over the head. If you steal my watch from me when I am out of the house, you have used force against me because I am an unwilling victim of your action, though this would not be so if I consented to your taking it. A merchant defrauds me, and since I am an unwilling victim of his action, it comes under the heading of force.
I still think that this definition is on the right track, but it is somewhat tricky, and there are implications of it that I had not sufficiently considered.
1. Suppose I invite a group of people to my house and they become engaged in a bitter quarrel. I intensely dislike being in this situation; I don't want to be in it. Am I an unwilling victim of it? Not quite. The fact that I don't want to be in the situation is not the same as saying that I am an unwilling victim of it. I perform every day actions which I don't really want to perform but do because they are a part of my (voluntarily undertaken) job. Nobody wants to wash the dishes, but some people still do it and not through the use of force. The fact that I find the quarrel at my party unpleasant, I surely didn't want it to happen, doesn't mean that I am unwillingly involved in it: for example, I can still choose to beat a hasty retreat from the house and leave the quarrelers to themselves. If I were physically unable to leave the house, e.g., if someone pushed me against the wall so that I couldn't get out, then it would be correct to say that force had been used. (Still, the word "unwilling" is vague; there is a sense in which I am unwillingly involved in the guests' actions, though it could hardly be said that force was involved.)
2. If you steal my watch, you have used force against me in that I am an unwilling victim of your action. But now suppose that you put your wristwatch on and drive away in your car; and suppose also that I don't want you to have your watch, because I want it for myself. I am unwilling for you to retain your watch; I am only willing for myself to have it. Can I say then that I am an unwilling victim of your action? Since I covet your watch, I am unwilling for you to retain it. But it would surely be ridiculous to say that you have used force by retaining in your possession the watch that you yourself have bought!
If we say (as presumably we want to say) that you are using force when you steal my watch from me, but not that you are using force when you simply retain in your possession your own watch, and if we also say that I am unwillingly involved on both occasions, then it seems that something further must be involved in use of force than the element of unwillingness. What we want to say is that in spite of the unwillingness (by me) which is present both times, the first is a case of the use of force and the second is not, because in the first case you are violating my rights and in the second case you are not. In other words, what is and is not to count as force depends on a prior theory of property rights (and doubtless of rights in general). In other words, when libertarians talk about the use of force, they can define it only in the context of their theory of rights.
This situation, of course, is not unprecedented in the history of thought: many terms, in the sciences for example, are theory-laden or theory-dependent in that they are defined only in the context of a certain theory. But it is nevertheless somewhat disturbing to find that we cannot define "force" "on its own," so to speak—that libertarians cannot say what they mean by "force" until they have first developed a view of human rights. Many libertarians will doubtless welcome this consequence, saying that it is what they had suspected would happen all along if one wants to use a common-sensically vague term like "force" in a more precise sense than is employed in everyday speech, just as the word "energy" can be defined precisely in physics because it has lost some (though not all) of the meaning it has in everyday expressions like "I don't have much energy today."
On the other hand, some libertarians, wishing to stick more closely to ordinary usage, will find attractive the prospect of returning to the "good old garden variety" sense, what we have called the "muscular" sense, of "force." Returning to this common usage, they would then say that a person is not justified in using force unless someone has 1) initiated the use of (muscular) force against him, or 2) violated a contract or other agreement with him, or 3) defrauded him, or 4) taken something that belongs to him, or 5) trespassed on his property, etc. These libertarian contentions would still be as true as ever, but we would not have to surrender the simple precept that there is no justification for the use of force against a person other than the prior initiation of force by him. We would have to say that there are circumstances other than the initiation of force in which retaliation is justified.
II. IS THE INITIATION OF FORCE EVER JUSTIFIABLE?
Whether or not one defines "force" in such a way that the trespasser, the thief, and the contract-violator can be said to have initiated the use of force, there are other problems. There are cases, admittedly more unusual ones, in which a person quite clearly has initiated the use of force, even in its narrower "muscular" sense, and in which some libertarians, though not all, would hold that he was justified in doing so. Some would hold, for example, that a starving man would be justified in snatching a loaf of bread from a wealthy man without the latter's consent, if doing this was necessary to prevent starvation; they would say that the starving man's life was a greater value than the wealthy man's possession, although admittedly the starving man was violating the other's property rights.
Or suppose that a physician has discovered a cure for cancer and refuses to give (or even to sell) it to you to save your wife's life; some libertarians, though not all, would hold that you would be justified in taking it from him by force. Here there is no way at all in which the physician could be said to have "aggressed" against you; he perfected the formula himself, he did it only for his own use and that of his family, and surely he has a right (though not a duty) to dispose of it in whatever way he chooses, including not imparting the secret to anyone else.
If you say the man was doing wrong in stealing the secret in order to save his wife, would you still say so if the stakes were made higher? Suppose that only an initiatory act of force would suffice to stop a plague (e.g., by universal inoculation, even upon those who because of primitive superstititions refused to be inoculated, and if not inoculated would be spreaders of the disease)? This situation is actually encountered with members of Jehovah's Witnesses, who refuse to have themselves or their children receive inoculations and blood transfusions, even if this results in their death. Or suppose that only an initiatory act of force would suffice to save a civilization? What if the man who refuses to divulge the formula (or to be inoculated) is a depressive, or a candidate for suicide, and doesn't care a whit about the fate of civilization? Just make the stakes high enough, one is tempted to say, and the answer seems clear: the higher the stakes are, the less plausible the rule of never initiating the use of force.
In fact, one doesn't have to make the stakes all that high. An actress is on stage about to receive the Academy Award. Millions of people are watching the event on television. At that moment her drunken husband (himself an invited guest at the proceedings) arrives in the auditorium, staggers onto the stage, and makes a blubbering fool of himself. Surely it would not be immoral to suggest that some member of the audience would have done well to give him a sock in the jaw before he got on the stage, in order to prevent the occurrence of a scene that would be the ultimate in embarrassment. Yet if anyone had done so, it would be he and not the drunken husband who was the initiator of force. Hundreds of similar examples could be adduced without much trouble.
Of course, though he didn't initiate force in the muscular sense (the man who stopped him initiated it against him), he could plausibly be said to have violated an implicit contractual agreement, not by being there (he was after all invited), but by being on the stage at that time and in that condition. But there are many other cases in which there is no such violation, say—
A ship has sunk, and 25 men are in a lifeboat designed for less than ten; a storm is approaching and all on board will sink unless the lifeboat is lightened by at least ten men. Since no one volunteers, the captain commands certain men to jump overboard. This is surely the initiation of force against these men; yet the only alternative is that they all perish. Is such initiation of force unjustified in these circumstances? If Mr. A, one of the ten, refuses the command, is he doing wrong? Lifeboat situations are admittedly peculiar, and the ethics of lifeboat situations is (as some libertarians have pointed out) admittedly different from that of civilized life; perhaps libertarians would hold that the captain is justified in making his request, and also that Mr. A is justified in refusing it. In any case, the example is food for thought for those libertarians who hold that the initiation of force is never justified in any circumstances whatever.
The following example is not from a lifeboat-type situation, and many cases like it have actually occurred. Before there was a vaccine for typhoid fever, a problem arose concerning the disposition of persons who, though not themselves ill with the disease, were the carriers of it to others; these were the "Typhoid Mary" cases of the earlier years of this century. It was dangerous to have a Typhoid Mary in the community, since if one of them was permitted to mingle freely with other people, large numbers of these people would become infected with the disease and often die from it. The question is, was it justifiable for the other members of a community to commit a Typhoid Mary to permanent isolation from them (as with lepers to a leper colony)? Here there was no question of "punishment," since punishment implies that a crime has been committed, and Typhoid Mary had committed no crime; she was just the unfortunate carrier of a dangerous disease. Was it justifiable for others to initiate the use of force against her though she had committed no wrong? If not, then many members of the community would fall ill and die through being infected by her. But if it is, then it cannot also be consistently claimed that the initiation of force is never justified.
III. FORCE VS. THREAT OF FORCE
Important though these questions are, and however libertarians may resolve them, I shall now concentrate on some problems that arise when one considers the retaliatory use of force in what might be called the "good old garden variety" sense, the "muscular" sense, of "force," the sense in which the man who starts swinging his fists at someone in a bar (who may have provoked him verbally but has taken no action against him) is initiating the use of force.
There are several questions which seem to me to require more detailed exploration than they usually receive: 1) must one wait until force has actually been used, or is the threat of force sufficient to justify the use of it in retaliation? 2) exactly when can one be said to be initiating it? 3) against whom is it justifiable to use it? 4) how much of it should be used? and 5) who should administer it?
As to the first question: in many cases, if you wait until force has actually been used against you, it will be too late. If you do nothing until he actually shoots, you will be dead, at which time it will be impossible to retaliate. It seems most implausible to insist that you must wait until force has actually been used; if this were adopted as a general principle, i.e., if it could be relied on that everyone would wait until he was actually attacked, the ranks of the peaceful people of the world would be decimated and their assailants would inherit the earth. Clearly some provision will have to be made to justify response to the threat of force.
But how serious must the threat be? If someone says to you at the height of a quarrel that he'll kill you the next time he gets a chance, is this justification for your dispatching him on the stop? To know how serious the threat is, you must know something about the background, character, and habits of the person who does the threatening: whether he is apt to threaten but not mean it, whether he means it at the time but is likely to cool off quickly, or even if he doesn't cool off whether he is likely to carry it through (some people are more talk than action), and so on. You have to make the best estimate you can of the probability that the threat will actually materialize if you do nothing to prevent it. A man who killed someone who was target-practicing, and who takes the mere physical proximity of firearms as a sufficient threat, is surely "jumping the gun" on retaliation. On the other hand, if a psychotic with a knife comes down the fire escape from the apartment above you every night and threatens to kill you and your children while they are asleep, and you go to the police and they say they can do nothing until a crime has actually been committed, you will feel quite justified in believing that some kind of preventive action should be taken to forestall the catastrophe. Yet where is one to draw the line with this? If a man starts to light a match inside a gunpowder plant, should you take a pessimistic view of his intentions?
Various criteria have been suggested. It has been said that you should wait until you see the whites of their eyes. This of course was intended to apply to an enemy army confronted by day (or by moonlight with the light facing them); and however useful it may once have been, it is ludicrous in an atomic age. In the case of individual confrontation with guns, the tradition of the Wild West was that the man must wait until the opponent reaches for his gun—if you shoot before that it is you who initiated the fight, and if you wait until after that it's self-defense (hence the importance of a quick draw, to shoot him first and yet have it be considered self-defense!). Of course, if he was reaching for his handkerchief at just that moment and you thought he was reaching for his gun, that's just your tough luck. But all these are only approximate and rather arbitrary criteria for very limited classes of situations. The point that is illustrated by all of them, a point which is not generally recognized, is that as a rule the retaliatory use of force must be to some extent anticipatory.
Now the very thought of anticipatory action, taken before the other person has actually done anything is quite repugnant to libertarians; it smacks of mock trials and shooting on a flimsy pretext. To take just one example: some psychologists claim to be able to predict with over 95% accuracy by means of psychological tests which children below ten will become juvenile delinquents by the time they are 15; and they have suggested putting these pre-delinquent children in special institutions, against their own and their parents' wishes if necessary, so as to prevent the crimes that these predelinquents otherwise would have committed. There is simply too much of 1984 in this. "We hereby choose to incarcerate you, not because you've done anything yet, but because we think you will." "Preventive detention" of all sorts is the perfect tool for any dictator. And yet, if the peaceful people are not to be killed off and replaced by their assailants, something in the way of "anticipatory use of force" must be countenanced, because of the simple fact of reality that waiting any longer may be fatal. If a mob is attacking City Hall, the police need not wait until they have destroyed lives and property before forcibly dispersing them.
Clearly the circumstances in which an act classifiable as the-warding-off-of-probable-aggression will vary so much from case to case that there is no one formula that will encompass all of them, other than the rather vague one that "the probability of being attacked (or otherwise aggressed against) must be high." To make the difficulties yet greater, often one must make this estimate of probability in a split second, before one has had any time to weigh the evidence.
IV. WHAT IS INITIATION?
How do we tell who has initiated the aggressive act? There are times when the answer is obvious: someone gets drunk and starts a fight with an innocent bystander. But already the case is more complex when the situation changes a bit: Mr. A says some insulting things about Mr. B's wife, and in order to stop him B gives A a punch in the jaw. Here presumably the libertarian position would be that B's action is the initiation of force and therefore wrong, albeit understandable and perhaps to a degree excusable (although many people, presumably nonlibertarians, would disagree even here and say that B was totally justified in his action even though he was initiating the use of force). But now suppose that Mr. C raises his fist at Mr. D, and D incorrectly interprets this as a sign that C is going to hit him, so D hits C instead. Who has now initiated the aggression? C in fact was not going to hit D, and D didn't know it and it seemed to him, perhaps because he was a bit paranoid anyway, that C was.
Who started the family quarrel? The husband says to visitors in the presence of his wife, "My wife doesn't want me to tell you this, but before I married her.…" and tells a story he knows his wife doesn't want him to tell. She becomes angry and, being extremely jealous, she cuts the legs off of a new pair of pants he wants to wear that evening. A bitter quarrel follows. Did he start it by telling a story from the past that he knew she didn't like, or did she start it by using the scissors on his pants? "After all," he says, "she's insanely jealous, and she ought to get over it." "But you knew I was that way and still you told that story," she says. "Must I respect every hangup you've got?" he retorts. As any psychiatrist will testify, both were aggressors and aggressed against. The event just escalated from a small and insignificant beginning.
When the parties involved are not individuals but governments (or individuals acting in their capacity as government leaders), it is the exception rather than the rule to be able to brand one part or the other as the aggressor. Doubtless the Nazis initiated aggression in 1936 when they occupied the Rhineland; but traditionally the Rhineland had been German territory, and most Germans felt that this territory had been taken from them by the Versailles Treaty. Indeed, Germany's principal aim throughout the 1930s, when the Allied Powers had promised to rectify the savage Versailles Treaty but did not, was to take upon themselves the job of the rectification; Germany considered herself only to be retaliating against the aggression committed through the Versailles Treaty by the Allies. And before that—well, it becomes even more hopelessly entangled. The assassination of the Archduke at Sarajevo in 1914 set off a chain reaction of events in the different countries, each of which was considered by the leaders of the country in question to be a defensive measure against incipient or threatened aggression by others. History of course abounds with such examples.
Mary Queen of Scots had made no overt move against Queen Elizabeth I; but Elizabeth finally (on the insistence of her advisers) had her killed. Was it the initiation of aggression or was it self-defense? Mary had followers throughout England and Scotland who would surely have killed Elizabeth and set themselves in power. It was clearly the one or the other. Exiled, Mary would have been a constant and increasing threat. She would surely have killed Elizabeth and her entire staff without a tremor of regret if she had had the chance. Who was initiating the use of force?
You inherit a kingdom. You have great plans for the founding of a true libertarian society, in which the government shall never initiate aggression against anyone. But there is a band of rebels near the northern border who threaten war if you are installed as chief leader. They are misguided, of course; they don't understand your plans and intentions. You want peace and prosperity for everyone, the rebels included. But meanwhile they are a threat to you and your entire regime; they will kill you and topple the first really libertarian society in history if they are given the chance. So, reluctantly, you decide that in the interests of everyone in the libertarian society you are beginning to install, you have to oppose them by force of arms—there is no other way, for they won't listen to reason. It is self-defense on your part, of course. If only they weren't so blind, you wouldn't have to start your regime with such a dirty job. So you do it—and as a result their numbers increase, the aggressive acts they suspected you would perform all along have now become a reality, and the rebellion grows. Were you initiating aggression, or was it self-defense?
In relations between nations, it is usually assumed that the first nation to cross the borders of another is the aggressor. But the situation is not so simple as this; when nations are involved, the situation is immensely more complicated:
…Usually the "aggressor" is simple-mindedly branded as the first state that crosses another state's borders with troops. The "collective security" principle holds that all the nations of the world are then duty-bound to get together and use force majeure against the "aggressor," and to defeat his evil designs. In practice, in our century, the U.S. has taken upon itself the "collective security" role, the White Knight in shining armor that sets out to defend the entire world against the Bad Dragon of aggression.
The fallacies and dangers in this doctrine abound at every hand. The first problem is the simplistic definition of "aggression." The analogy, usually implicit…, is always taken from aggression by one individual upon another. If Smith is seen to be jumping on and stealing a watch from Jones, then Smith can easily be labelled the "aggressor," and police may be called upon to defend Jones and apprehend the criminal for return of the loot. But while we might be able to say easily that Jones deserved to have the watch and that therefore Smith was an aggressor, the same can scarcely be said for state X which has been invaded by state Y. For to call state Y an "aggressor" per se must mean that the present territorial boundaries of state X are somehow morally and rightfully its own, in the same way that Jones' watch is rightfully his own. But since national territories have invariably been acquired by previous aggression rather than by voluntary social contract, to leap automatically to the defense of the invaded State is an absurdity. On what moral grounds are we to cry "halt" and thereby ratify every aggression previous to, say, December 1971, as legitimate and moral?
The Israelis claim that the territory now occupied by the state of Israel is their ancestral home and was taken from them by conquest and occupied by interlopers (Arabs, Turks, etc.) during the intervening centuries—so that they are now reclaiming what is their own. (There is a problem here, since the original occupants are long since dead, how the harm can be undone by the retaking of the territories by the great-great-great-etc. grandchildren of the original victims.) The Arabs, for their part, claim that the territory has been their home for many centuries and that it has now been taken away from them by the Israelis. Who is right? Both are right, in that both peoples have at various times occupied this area for long periods. Both are wrong, in that neither people occupied the territory originally: both groups took it by force from others who occupied it before them, in still earlier acts of aggression (Israelites took it from the Canaanites, for example, who probably took it from some people before that). What about those who occupied it for the first time, before anyone else had it? No one knows who they are; and in any case they are long since dead, probably killed to the last man by the first invaders who took it over. There is simply no way in which such a situation can be rectified after the damage has been done, any more than a murdered man can be brought to life again in the interests of justice. To point to one group or the other and say that they are the initiators of aggression is grotesque: they are all tarred with the same brush.
(To be concluded in REASON's January 1973 issue)
John Hospers is Professor of Philosophy and Director of the School of Philosophy at the University of Southern California. He received his Ph.D. in philosophy from Columbia University and has written many books and articles, including the authoritative volume, LIBERTARIANISM, which was recently published in paperback by Reason Press. A widely respected philosopher, Professor Hospers was nominated by the Libertarian Party as its candidate for President of the United States, and he is presently conducting an active campaign for the presidency.
NOTES AND REFERENCES
 F. von Hayek, THE CONSTITUTION OF LIBERTY (University of Chicago Press, 1960), p. 136.
 Murray Rothbard, "The U.N. and the Law," LIBERTARIAN FORUM, December 1971, p. 1.
 See, on this point, John Hospers, "Property," THE PERSONALIST, Vol. 53, Summer 1972, pp. 163-73.
This article originally appeared in print under the headline "Some Problems About Punishment and the Retaliatory Use of Force".