Ethics in Practice

Paternalism and Risk

paternalistically justifying some course of action on the grounds that it is in someone's interests, I shall always be searching for some warrant in that person's own value judgments for saying that it is in that person's interests.

"Some warrant" is a loose constraint, to be sure. Occasionally will we find genuine cases of what philosophers call "weakness of will": people being possessed of a powerful, conscious present desire to do something that they nonetheless just cannot bring themselves to do. Then public policy forcing them to realize their own desire, though arguably paternalistic, is transparently justifiable even in terms of people's own subjective values. More often, though, the subjective value to which we are appealing is one which is present only in an inchoate form, or will only arise later, or can be appreciated only in retrospect.

Paternalism is clearly paternalistic in imposing those more weakly-held subjective values upon people in preference to their more strongly held ones. But, equally clearly, it is less offensively paternalistic thanks to this crucial fact: at least it deals strictly in terms of values that are or will be subjectively present, at some point or another and to some extent or another, in the person concerned.

1 The Scope of Paternalism

When we are talking about public policies (and maybe even when we are talking of private, familial relations), paternalism surely can only be justified for the "big decisions" in people's lives. No one, except possibly parents and perhaps not even they, would propose to stop you from buying candy bars on a whim, under the influence of seductive advertising and at some marginal cost to your dental health.

So far as public policy is concerned, certainly, to be a fitting subject for public paternalism a decision must first of all involve high stakes. Life-and-death issues most conspicuously qualify. But so do those that substantially shape your subsequent life prospects. Decisions to drop out of school or to begin taking drugs involve high stakes of roughly that sort. If the decision is also substantially irreversible -

returning to school is unlikely, the drug is addictive - then that further bolsters the case for paternalistic intervention.

The point in both cases is that people would not have a chance to benefit by learning from their mistakes. If the stakes are so high that losing the gamble once will kill you, then there is no opportunity for subsequent learning. Similarly, if the decision is irreversible, you might know better next time but be unable to benefit from your new wisdom.

2 Evaluating Preferences

The case for paternalism, as I have cast it, is that the public officials might better respect your own preferences than you would have done through your own actions. That is to say that public officials are engaged in evaluating your (surface) preferences, judging them according to some standard of your own (deeper) preferences.

Public officials should refrain from paternalistic interference, and allow you to act without state interference, only if they are convinced that you are acting on:

•relevant preferences;

•settled preferences;

•preferred preferences; and, perhaps,

•your own preferences.

In what follows, I shall consider each of those requirements in turn. My running example will be the problem of smoking and policies to control it. Nothing turns on the peculiarities of that example, though. There are many others like it in relevant respects.

It often helps, in arguments like this, to apply generalities to particular cases. So, in what follows, I shall further focus in on the case of one particular smoker, Rose Cipollone. Her

her case special is that she (or more precisely her heir) was the first to win a court case against the tobacco companies whose products killed her.

In summarizing the evidence presented at that trial, the judge described the facts of the case as follows.

Rose ... Cipollone ... began to smoke at age 16, ... while she was still in high school. She testified that she began to smoke because she saw people smoking in the movies, in advertisements, and looked upon it as something "cool, glamorous and grown-up" to do. She began smoking Chesterfields... primarily because ofadvertising of "pretty girls and movie stars," and because Chesterfields were described ... as "mild." ...

Mrs. Cipollone attempted to quit smoking while pregnant with her first child ... , but even then she would sneak cigarettes. While she was in labor she smoked an entire pack of cigarettes, provided to her at her request by her doctor, and after the birth ... she resumed smoking. She smoked a minimum of a pack a day and as much as two packs a day.

In 1955, she switched ... to L&M cigarettes ... because ... she believed that the filter would trap whatever was "bad" for her in cigarette smoking. She relied upon advertisements which supported that contention. She ... switched to Virgina Slims ... because the cigarettes were glamorous and long, and were associated with beautiful women ~ and the liberated woman....

Because she developed a smoker's cough and heard reports that smoking caused cancer, she tried to cut down her smoking. These attempts were unsuccessful. ...

Mrs. Cipollone switched to lower tar and nicotine cigarettes based upon advertising from which she concluded that those cigarettes were safe or safer ... [and] upon the recommendation of her family physician. In 1981 her cancer was diagnosed, and even though her doctors advised her to stop she was unable to do so. She even told her doctors and her husband that she had quit when she had not, and she continued to smoke until June of 1982 when her lung was removed. Even thereafter she smoked occasionally ~ in hiding. She stopped smoking in 1983 when her cancer had metasized and she was diagnosed as fatally ill.

Permissible Paternalism

This sad history contains many of the features that I shall be arguing makes paternalism most permissible.

Relevant preferences

The case against paternalism consists in the simple proposition that, morally, we ought to respect people's own choices in matters that affect themselves and by-and-large only themselves. But there are many questions we first might legitimately ask about those preferences, without in any way questioning this fundamental principle of respecting people's autonomy.

One is simply whether the preferences in play are genuinely relevant to the decision at hand. Often they are not. Laymen often make purely factual mistakes in their means~ends reasoning. They think ~ or indeed, as in the case of Rose Cipollone, are led by false advertising to suppose ~ that an activity is safe when it is not. They think that an activity like smoking is glamorous, when the true facts of the matter are that smoking may well cause circulatory problems requiring the distinctly unglamorous amputation of an arm or leg.

When people make purely factual mistakes like that, we might legitimately override their surface preferences (the preference to smoke) in the name of their own deeper preferences (to stay alive and bodily intact). Public policies designed to prevent youngsters from taking up smoking when they want to, or to make it harder (more expensive or inconvenient) for existing smokers to continue smoking when they want to, may be paternalistic in the sense of running contrary to people's own manifest choices in the matter. But this overriding of their choices is grounded in their own deeper preferences, so such paternalism would be minimally offensive from a moral point of view.

Settled preferences

We might ask, further, whether the preferences being manifested are "settled" preferences or whether they are merely transitory phases people are going through. It may be morally permissible to let people commit euthanasia voluntarily, if we are sure they really want to

Paternalism and Risk

die. But if we think that they may subsequently change their minds, then we have good grounds for supposing that we should stop them.

The same may well be true with smoking policy. While Rose Cipollone herself thought smoking was both glamorous and safe, youngsters beginning to smoke today typically know better. But many of them still say that they would prefer a shorter but more glamorous life, and that they are therefore more than happy to accept the risks that smoking entails. Say what they may at age sixteen, though, we cannot help supposing that they will think differently when pigeons eventually come home to roost. The risk-courting preferences of youth are a characteristic product of a peculiarly dare-devil phase that virtually all of them will, like their predecessors, certainly grow out of.

Insofar as people's preferences are not settled - insofar as they choose one option now, yet at some later time may wish that they had chosen another - we have another ground for permissible paternalism. Policy-makers dedicated to respecting people's own choices have, in effect, two of the person's own choices to choose between. How such conflicts should be settled is hard to say. We might weigh the strength or duration of the preferences, how well they fit with the person's other preferences, and so on.

Whatever else we do, though, we clearly ought not privilege one preference over another just because it got there first. Morally, it is permissible for policy-makers to ignore one of a person's present preferences (to smoke, for example) in deference to another that is virtually certain later to emerge (as was Rose Cipollone's wish to live, once she had cancer).

Preferred preferences

A third case for permissible paternalism turns on the observation that people have not only multiple and conflicting preferences but also preferences for preferences. Rose Cipollone wanted to smoke. But, judging from her frequent (albeit failed) attempts to quit, she also wanted not to want to smoke.

In this respect, it might be said, Rose Cipollone's history is representative of smokers more generally. The US Surgeon General reports

that some 90 percent of regular smokers have tried and failed to quit. That recidivism rate has led the World Health Organization to rank nicotine as an addictive substance on a par with heroin itself.

That classification is richly confirmed by the stories that smokers themselves tell about their failed attempts to quit. Rose Cipollone tried to quit while pregnant, only to end up smoking an entire pack in the delivery room. She tried to quit once her cancer was diagnosed, and once again after her lung was taken out, even then only to end up sneaking an occasional smoke.

In cases like this - where people want to stop some activity, try to stop it but find that they cannot stop - public policy that helps them do so can hardly be said to be paternalistic in any morally offensive respect. It overrides people's preferences, to be sure. But the preferences which it overrides are ones which people themselves wish they did not have.

The preferences which it respects - the preferences to stop smoking (like preferences of reformed alcoholics to stay off drink, or of the obese to lose weight) - are, in contrast, preferences that the people concerned themselves prefer. They would themselves rank those preferences above their own occasional inclinations to backslide. In helping them to implement their own preferred preferences, we are only respecting people's own priorities.

Your own preferences

Finally, before automatically respecting people's choices, we ought to make sure that they are really their own choices. We respect people's choices because in that way we manifest respect for them as persons. But if the choices in question were literally someone else's - the results of a post-hypnotic suggestion, for example - then clearly there that logic would provide no reason for our respecting those preferences.

Some people say that the effects of advertising are rather like that. No doubt there is a certain informational content to advertising. But that is not all there is in it. When Rose Cipollone read the tar and nicotine content in advertisements, what she was getting was information. What she was getting when looking at

the accompanying pictures of movie stars and glamorous, liberated women was something else altogether.

Using the power of subliminal suggestion, advertising implants preferences in people in a way that largely or wholly by-passes their judgment. Insofar as it does so, the resulting preferences are not authentically that person's own. And those implanted preferences are not entitled to the respect that is rightly reserved for a person's authentic preferences, in consequence.

Such thoughts might lead some to say that we should therefore ignore altogether advertisinginduced preferences in framing our public policy. I demur. There is just too much force in the rejoinder that, "Wherever those preferences came from in the first instance, they are mine now." If we want our policies to respect people by (among other things) respecting their preferences, then we will have to respect all of those preferences with which people now associate themselves.

Even admitting the force of that rejoinder, though, there is much that still might be done to curb the preference-shaping activities of, for example, the tobacco industry. Even those who say "they're my preferences now" would presumably have preferred, ahead of time, to make up their own minds in the matter. So there we have a case, couched in terms of people's own (past) preferences, for severely restricting the advertising and promotion of products - especially ones which people will later regret having grown to like, but which they will later be unable to resist.

3 Conclusions

What, in practical policy terms, follows from all that? Well, in the case of smoking, which has served as my running example, we might ban the sale of tobacco altogether or turn it into a drug available only on prescription to registered users. Or, less dramatically, we might make cigarettes difficult and expensive to obtain - especially for youngsters, whose purchases are particularly price-sensitive. We might ban all promotional advertising of tobacco products, designed as it is to attract new users. We

Permissible Paternalism

might prohibit smoking in all offices, restaurants, and other public places, thus making it harder for smokers to find a place to partake and providing a further inducement for them to quit.

All of those policies would be good for smokers themselves. They would enjoy a longer life expectancy and a higher quality of life if they stopped smoking. But that is to talk the language of interests rather than of rights and choices. In those latter terms, all those policies clearly go against smokers' manifest preferences, in one sense or another. Smokers want to keep smoking. They do not want to pay more or drive further to get their cigarettes. They want to be able to take comfort in advertisements constantly telling them how glamorous their smoking is.

In other more important senses, though, such policies can be justified even in terms of the preferences of smokers themselves. They do not want to die, as a quarter of them eventually will (and ten to fifteen years before their time) of smoking-related diseases; it is only false beliefs or wishful thinking that make smokers think that continued smoking is consistent with that desire not to avoid a premature death. At the moment they may think that the benefits of smoking outweigh the costs, but they will almost certainly revise that view once those costs are eventually sheeted home. The vast majority of smokers would like to stop smoking but, being addicted, find it very hard now to do so.

Like Rose Cipollone, certainly in her dying days and intermittently even from her early adulthood, most smokers themselves would say that they would have been better off never starting. Many even agree that they would welcome anything (like a workplace ban on smoking) that might now make them stop. Given the internally conflicting preferences here in play, smokers also harbor at one and the same time preferences pointing in the opposite direction; that is what might make helping them to stop seem unacceptably paternalistic. But in terms of other of their preferences - and ones that deserve clear precedence, at that - doing so is perfectly well warranted.

Paternalism and Risk

Smoking is unusual, perhaps, in presenting a case for permissible paternalism on all four of the fronts here canvassed. Most activities might qualify under only one or two of the headings. However, that may well be enough. My point here is not that paternalism is always permissible but merely that it may always be.

In the discourse of liberal democracies, the charge of paternalism is typically taken to be a knock-down objection to any policy. If I am right, that knee-jerk response is wrong. When confronted with the charge of paternalism, it should always be open to us to say, "Sure, this proposal is paternalistic - but is the

paternalism in view permissible or impermissible, good or bad?" More often than not, I think we will find, paternalism might prove perfectly defensible along the lines sketched here.

Further Reading

Goodin, Robert E. "The Ethics of Smoking." Ethics 99 (April 1989), 575-624.

Goodin, Robert E. No Smoking: The Ethical Issues

(Chicago and London: University of Chicago Press, 1989).

Some Liberal Constraints

Bans on guns are typically considered a "liberal" policy. We will argue, however, that broad bans on firearms are in fact not liberal policies at all. The policy of a state that disarms its citizenry conflicts with more than one of the fundamental principles of liberalism.

The degree and nature of the conflict between liberalism and gun bans depends, however, on how one conceives ofliberalism. In this regard, gun bans serve as a means of illustrating the disparity between two fundamentally different versions of liberalism, which we shall call wide and narrow liberalism. We shall try to show that a complete ban on the private possession of firearms is impermissible on either view; in fact any meaningful restriction is difficult to justify in the context of wide liberalism. Narrow liberalism, on the other hand, permits more restriction, but, if applied consistently, is unlikely to allow ones that will result in a significant decrease in violent crime.

The assumption motivating most calls for bans on private possession of guns is that a causal relationship exists between the number of guns in the private sector and the number of victims of violent crime: an increase in the number of guns (in some sense) causes an increase in violent crime. I The empirical literature on this issue is baffling, at least to those not trained in mathematics and social science. Let

us assume, for the sake of the argument, that the alleged causal relationship between guns and crime really exists. Is this sufficient to justify a government ban on firearms? In a liberal state, the answer is simple: it is no. In a consistently liberal system, it is considered highly problematic to dispose of the rights and liberties of citizens - where these rights and liberties are believed by their owners to be important - simply and solely because the community can extract a benefit from doing so.

Consider the following example. It is very obvious that we could prevent a great many deaths from AIDS by enacting a policy reputedly followed in Cuba: that of simply rounding up everyone known to have the disease and isolating them in special camps until they no longer carry the disease (presumably, because they are dead). We do not have such a policy, though we know perfectly well that thousands of people will die because we do not have it. The reason we do not have it is that we, or most of us, think that incarceration is a bad way to treat sick people. It violates their rights. Of course, there are many ways to explain why we tend to think this, and why we tend to find this thought so decisive. One perfectly good way to explain it, however, is to say that this country, unlike Cuba, is a liberal democracy: from a liberal point of view, the mere fact that a policy could save lives is not a sufficient reason for adopting it. The policy itself must

Paternalism and Risk

be morally permissible. In a consistently liberal polity, the pursuit of all social goals, including the goal of saving thousands from dying horribly and pointlessly, is constrained.

Liberalism, however, is distinguished not merely by the fact that its pursuit of the good is constrained, but by the particular set of constraints it recognizes: an action, including a government policy, is considered unjust, and consequently unacceptable, if it violates one or more of these liberal constraints. There are a number of principles that could be included in this set, but our focus here will be on three of them - autonomy, neutrality, and equality - and most importantly on autonomy and equality.

The concept of autonomy can be formulated in many different ways, with many different degrees of stringency, but what the formulations have in common is the general notion that individuals should control their own lives and be the instruments of their own acts of will. The individual has certain fundamental rights against interference by others. One way to formulate a usable version of the autonomy principle is to define a part of the individual's life as "private," in virtue of the fact that it includes behavior that has no significant effects on others without their consent, and to declare that, within this private domain, individuals may do whatever they wish. This is the method used by Mill and many others after him. This way of interpreting the idea of autonomy can be called "minimal," in that it constrains government policy less than the other interpretations do. Probably the most "maximal" interpretation is the one associated with the Lockean notion of self-ownership.2 On this view, a person owns his or her self and, by the same token, the product of the labor of that self.

The principle of autonomy is a major source of the familiar liberal animus against paternalism. Government acts paternalistically when it makes or restricts important choices for individuals in order to do those same individuals some good. Such a policy involves interfering with individual conduct even when it affects no one but the individual agent and, consequently, tends to run afoul of even the most minimal version of the autonomy constraint. Liberals have at times carved out exceptions to the au-

tonomy principle, formulating "soft" forms of paternalism which allow interference when the individual conduct involved is seriously nonvoluntary, or when the rights and liberties disposed of are trivial. 3 Because liberalism rests on the principle of autonomy, however, it cannot go very far in justifying policies that are paternalistic.

The second liberal constraint is the principle of neutrality. It holds that the justification for state action must be neutral between particular conceptions of the good life. It is an indirect constraint on government action, in that what it constrains is not the actions but the reasons that are given for them. The policies of liberal states have many sideeffects and, no doubt, some of them create conditions in which certain conceptions of the good can no longer be pursued. For instance, there may be values that can only be achieved by pursuing the way of life of a samurai warrior, an eighteenth century aristocrat, or a Medieval knight - ways of life that were extinguished by liberal institutions. From a liberal point of view, this effect can be just, but only if these policies can be defended on other grounds, apart from the fact that they tend to "stamp out" ways of life that liberals do not appreciate.

As was the case with the principle of autonomy, the equality constraint is open to widely different formulations, and substantially different ones tend to mark the differences between different sorts of liberalism. Common to all the versions of this principle is the idea that the state should treat its subjects as equals. At a minimum, it means that governments must respect equally the rights of its citizens. They must not discriminate against some citizens and in favor of others. No one is to be either above or below the law. It also means that government must not function as an instrument by which the strong take advantage of the weak.

This idea - that governments must take the rights of citizens equally seriously - can be called the minimal interpretation of the equality constraint. It states something that all liberals believe. For instance, if it could be shown that state lotteries tend to take money from the relatively poor and uneducated (perhaps because such people tend to have a comparatively

shaky grasp of probability theory) and tend to put this money into the hands of the relatively rich and educated, all liberals would feel that this is a weighty argument against state lotteries. They would all see such an arrangement as unjust because of the way in which it arbitrarily discriminates against people who are less well endowed and in favor of those who already enjoy advantages.

It is possible, however, to interpret the principle of equality in ways that go beyond this minimal version of it: one can interpret it, as we shall say, extraminimally. There are many ways to do this, but all tend, to one degree or other, to claim that the liberal state is committed, not merely to respecting equally the rights of its citizens, but to equalizing the value of the rights that each person has. The forms that this idea takes range from the idea that inequalities in the distribution of goods, though allowable, are subject to egalitarian constraints, to the idea that resources available in a society should be divided equally among its members through some scheme of redistribution.

As we have already suggested, one can envision quite different varieties of liberalism depending on how one interprets the principles of autonomy and equality. On the one hand, one could adopt an extremely extraminimal interpretation of the principle of equality, which would allow extensive state efforts to equalize the conditions of its citizens. Expecting this state activity to cut into the liberties of the individual, one might then adopt a relatively minimal interpretation of the autonomy constraint. On the other hand, one might adopt an extremely extraminimal interpretation of the principle of autonomy and, expecting this to limit aggressive redistribution policies, one could adopt a relatively minimal interpretation of the principle of equality. Because it allows wider liberty of action, we call the latter sort of position "wide liberalism." The former, for analogous reasons, we will call "narrow liberalism.,,4

As we shall see, it makes a difference, as far as the issue of bans on firearms is concerned, whether one is a wide or a narrow liberal. But we will also argue that it does not make as much difference as one might think. No position

The Liberal Basis of the Right to Bear Arms

could be called liberal that did not accept, in one form or other, all three of the principles we have discussed. We will argue that all three of these areas of concern tend to militate, though in different degrees and in different ways, against bans on firearms.

2 Firearms and Autonomy

Today, the very idea that the possession of a gun - a mere technological device prized by hobbyists and lunatics - is a right, like freedom of speech, freedom of religion, and the right against self-incrimination, strikes many people as silly. Nonetheless, such a conclusion is more or less forced on us by all plausible extraminimal versions of the autonomy constraint, including even the mildest of them.

This may sound like a strange claim, but it becomes much less so when we clearly understand certain of its concrete implications. To this end, consider the case of Ms. Jackson of Atlanta, Georgia:

A College Park woman shot and killed an armed man she says was trying to carjack her van with her and her one-year-old daughter inside, police said Monday....

Jackson told police that the gunman accosted her as she drove into the parking lot of an apartment complex on Camp Creek Parkway. She had planned to watch a broadcast of the Evander Holyfield-Mike Tyson fight with friends at the complex.

She fired after the man pointed a revolver at her and ordered her to "move over," she told the police. She offered to take her daughter and give up the van, but the man refused, police said.

"She was pleading with the guy to let her take the baby and leave the van, but he blocked the door," said College Park Detective Reed Pollard. "She was protecting herself and the baby."

Jackson, who told the police she bought the .44 caliber handgun in September after her home was burglarized, said she fired a shot from the gun, which she kept concealed in a canvas bag beside her car seat. "She

Paternalism and Risk

didn't try to remove it," Pollard said. "She just fired.,,5

Considering the fact that Ms. Jackson's would-be abductor was threatening her with lethal force at the time she killed him, and considering also what his most likely motive was for refusing to simply steal her car and let her go free, it seems obvious that she had a right to do what she did. She has a right to selfdefense.

Though this right is not protected by the minimal interpretation of the autonomy constraint (since the effect her act has on her assailant puts it well outside the domain of privacy) the reasons for adding it to the rights that are protected by this constraint seem to be as great as any could be. The interest that is protected by Ms. Jackson's right to defend herself is life itself. The interests protected by other rights recognized by liberals - including not only freedom of speech, but also freedom of religion, the right against self-incrimination, and many others - are no more important than this one. In many cases, they are a good deal less important.

Of course, Ms. Jackson's action has an effect on her assailant, and effects on others can give rise to powerful claims from those others that their rights have been violated. Such claims are potential reasons for not adding the act to the list of rights that are protected by a fundamental principle. Here, however, the strength of such reasons are as low as they can be. Though the effect suffered by her assailant is a powerfully negative one, his attack on her clearly cancels any claims he might have had against her use of force against him.

If we suppose, then, that Ms. Jackson has any rights to act on her own volition, outside the domain of privacy, then she must have this right, a right of self-defense. Further, if this supposition settles the question of her right of self-defense, it also settles the question of whether she has a right to use her gun to shoot her assailant. 6

So far, then, it would seem that, even a very modestly extraminimal interpretation of the autonomy constraint would have to imply that Ms. Jackson has a right to use her weapon to

defend herself, a right of non-interference that would be violated by having her weapon confiscated. The case we have made for this claim does make a certain assumption. Though it may require a lengthy inquiry to determine precisely what form the assumption should be given, it clearly must include the idea that one violates a right (in this case, Ms. Jackson's right of selfdefense) if one coercively prevents them from using the only, or the best, means to exercising that right (here, Ms. Jackson's handgun). It would be interesting to discuss whether (or why) such a principle is true, but it is not necessary to do so here, since the same principle is recognized by liberals in other contexts. Most of them would agree, for instance, that the government would be violating our rights to freedom of expression if it made possessing a computer modem a crime punishable by a term in prison. The same would be true if the law allowed the police to give out modem permits in the event that they decide an individual citizen has a "valid" reason to possess one. On the face of it, the same principle would seem to apply to an agent, including a representative of the state, who takes Ms. Jackson's gun from her.

Liberals who support gun bans would probably urge at this point that there is an obvious difference between a modem and a gun. A gun, unlike a modem, is a weapon, and advocates of gun bans argue that guns are substantially more dangerous than other weapons. Firearms allow people to commit offenses they could not commit with other weapons, such as knives, clubs, and fists. They enable persons normally in a position of weakness to use the threat of harm to take the money, possessions, and even lives of their victims.

If the presence of guns in one's environment does indeed increase such dangers, and it is at least prima facie plausible to say that it does, the liberal autonomy constraint itself seems to support banning some or all of them. All liberals agree that the principle of autonomy, unlike the principle of equality, applies fully to individuals as well as to states. It is permissible for individuals to treat others in substantially unequal ways, but serious violations of the autonomy of others (at least of sane, innocent adults) is not permissible. Obviously, killing or maiming

others is ordinarily such an impermissible violation. Consequently, the conclusion, which we tentatively suggested a moment ago, that even a modest autonomy constraint would imply that Ms. Jackson has a right against having her weapon confiscated, stands in need of further support. More needs to be said before we can draw such a conclusion.

3 Risk

No one would deny that a liberal state, even the relatively constrained state of wide liberalism, may prohibit activities that kill or maim others. Of course, a state that bans firearms like the one used by Ms. Jackson would not merely be prohibiting actions that actual~y do that sort of harm: they would rather be prohibiting an activity (owning a gun, or a gun of a certain sort) on the grounds that it creates a risk that such harm will be done. However, prohibiting activities that create such risks is itself something that a wide liberal state may do. Such a state may prohibit me from storing dynamite in my basement or driving while intoxicated, even when (luckily) these activities do not kill or maim anyone. One reason for this, and a sufficient one for our purposes, is that such risky activities are themselves violations of the autonomy of others.

Having admitted this, however, we claim that the mere fact of owning a gun, at least a gun like Ms. Jackson's .44, does not belong in this category: it does not create the sort of risk that justifies prohibition in the context of wide liberalism. In particular, it is starkly different from the two examples of risky activities that we just mentioned.

First, just as a gun is obviously different from a modem, it is also, though perhaps less obviously, different from dynamite. Dynamite is an unstable substance, which can be detonated, under some circumstances, by a mere tap. Its unpredictability, together with the sheer scale of the destruction caused when it does explode, justifies us in classifying dynamite as a substance that cannot be handled entirely safely.

In the relevant respect, guns are as different from dynamite as can be imagined. Like clocks,

The Liberal Basis of the Right to Bear Arms

guns are (ignoring an exception which we will discuss shortly) precision instruments: they are designed to function precisely, and for more or less the same reasons that clocks are. People have clocks so that they will know exactly what time it is, and not approximately what time it is. Similarly, they have guns so they will be able to hit a target, and not nearby objects. The function of a gun is not simply to provide lethal force, but to provide precisely controlled lethal force. Partly because of this fact, it is a surprisingly simple matter to handle a gun safely. As millions of Americans know from their firearms safety training, there are a few easy to follow rules which, if they are followed, will guarantee that unplanned detonations will not occur.

The same sorts of considerations suffice to show that possessing a gun is utterly different from drunk driving. Unlike drunk driving, gun ownership is not behavior that creates a significant likelihood of accidental injury. In 1993 there were 0.656 accidental deaths due to firearms per 100,000 firearms in the United States: far less than the more than 21 accidental deaths attributable to motor vehicles for that year, per 100,000 motor vehicles. 7 According to one report, the total number of accidental gun deaths due to guns is smaller than the number attributable to medical error.8

However, there is a sense in which the rate of accidental deaths from guns is actually a minor issue: the alleged risk involved in gun ownership that most often inspires proposals to ban guns probably has little or nothing to do with accidental death. The most important problem raised in the scholarly debate concerning bans on guns is intentional death: most of the antigun literature is sharply focused on the idea that guns make deliberate acts of violence more likely. This is the sort of risk that makes them especially dangerous objects, allegedly justifying placing them under a ban of some sort.

We maintain, however, that, within the limits of wide liberalism, even of a very moderate conception of those limits, this sort of risk is not a legitimate reason for banning guns. To return to the case of Ms. Jackson: the notion that she might use her gun to attack someone impermissibly might conceivably justify