I. Feinberg on the Limits of Criminalization

The main issue for Feinberg is whether there are principles which govern the question of what conduct falls within the legitimate powers of a state to regulate by means of the criminal law. Of course, the state may seek to change, alter, and limit the behavior of its citizens by many means other than criminalizing the conduct. It can do so by providing incentives for people to act one way rather than another (e.g., if you have children you get a break on your income taxes). It can do so by making it costly to engage in certain conduct (e.g., if you smoke you will have to pay a very large tax on each pack of cigarettes). It can do so by using the law of torts to make it the case that if you act in certain ways you are open to civil suits which may cost you a lot of money (e.g., you can be held liable for the torts of those to whom you serve alcoholic beverages). It can do so by making it more difficult to enter into contracts, thus limiting the ability to secure goods by mutual agreement (e.g., minors may not be sued for breach of contract, gambling debts are not enforceable). It can do so by providing warnings and information about dangerous products (e.g., explosives or toxic substances). It can do so by tying the receipt of federal funds to states' adopting certain laws (e.g., only states that require cyclists to wear helmets may get federal highway funds). Those who think that when the state cannot legitimately criminalize conduct it is helpless to try to change behavior are as mistaken as those who think that if John Stuart Mill's argument is correct that the state should not punish people for self-destructive behavior, we are left without resources of criticism, social pressure, and ostracism to use on those who act foolishly, irrationally, or imprudently.

Nevertheless, there are types of conduct which we believe the state may legitimately seek to forbid or require by means of criminal sanctions, and Feinberg believes there are what he calls “liberty-limiting” principles which distinguish between those acts which the state may legitimately criminalize and those it may not. For Feinberg, these principles are a “harm to others principle” and an “offense principle.” The harm principle states that it is always a good reason in support of penal legislation that it would probably be effective in preventing harm to persons other than the one prohibited from acting. The offense principle states that it is always a good reason in support of penal legislation that it would probably prevent serious offense to persons other than the one prohibited from acting. These exhaust the set of liberty-limiting principles. Conduct may not be sanctioned if it merely causes harm to the actor, or (crucially, as we shall see) if the harm it causes to others is consented to by the others.

Conduct may not be sanctioned merely because it is wrongful. It must, in addition, cause harm to others. Even if it is wrongful, and does cause harm to others, it may not be sanctioned unless it wrongs others, i.e., there are specific individuals who have what Feinberg calls a “grievance.” We shall see later what that amounts to.

There has been a remarkable consensus that whatever other principles might be required for an adequate theory of criminalization, some form of a harm to others principle is required. Indeed, the only exception I know of is a recent essay by Arthur Ripstein which seeks to defend what he calls the “sovereignty principle.”Footnote ²

Most of the debates about the harm to others principle concern how to specify the term “harm”: What are the links between our agency and harm, e.g., are we required to prevent harm, or may we also seek to promote benefit? Do only acts cause harm or may we harm by omission? What are the limits of harm, i.e., can we harm the dead, can we harm those whom we bring into existence? How do we delimit the cases of harm that are appropriate for state interference, i.e., does the harm have to be a violation of a right, does it have to be wrongful, does it have to harm particular persons? And so on.

In this essay, I am going to look at Feinberg's views about harms and point out some problems and puzzles in his theory. Some of these are ones he himself has recognized and attempted to deal with. Some are ones that he did not recognize or did not deal with adequately.

As is well-known, Feinberg distinguishes between two types of harms. The first involves any setback to interest. The second is a wrong, a violation of persons' rights. It is the latter conception which is regulated by the harm to others principle. Given this conception, there is conceptual space for the idea of “non-harmful wrongs” and “non-wrongful harms.” The former are violations of rights that do not set back the wronged party's interests, i.e., they involve harmless wrongdoing. The latter are setbacks to interests but are not violations of the wronged party's rights. The most common examples of the latter are harms to which the harmed party consented.

I shall now discuss some puzzles and problems that arise in Feinberg's treatment of the issues of harm and wrongdoing.

II. Problems with Feinberg's Account

It might be thought that if I poke you in the eye with a sharp stick causing pain but no lasting injury to the eye I have harmed you. But, for Feinberg, this is not the case. He distinguishes between harmful conditions and all “the various unhappy and unwanted physical and mental states that are not states of harm in themselves.”Footnote ³

Among the disliked but not harmful states are “hurts.” The other two categories are offenses and a residual category which includes things like embarrassment, boredom, etc. For Feinberg, the exact details of classification are not important. What is important is the contrast between harmful conditions and all the undesirable states which are harmful only when their presence is sufficient to impede an interest. A state of extreme and unrelenting pain can interfere with the pursuit of various goals and objectives and is thus a harm as well as hurtful.

It is because of this distinction that Feinberg has to go beyond what might be called “minimal liberalism,” the view that only the harm to others principle justifies restrictions on liberty. A “moderate liberalism” upholds an offense principle as well.

The interesting theoretical issue is why Feinberg denies that we have an interest in not being hurt as such. Even on his own view of what are interests he lists a certain amount of “freedom from interference.”Footnote ⁴ Since most inflictions of pain do involve interference with someone's body by others, it is unclear why we do not have a derived interest in being free from being hurt by others. One might still think, however, that even if we include the infliction of pain under the harm to others principle we still need the offense principle to include as harms things like being exposed to sounds and sights which are emotionally disturbing. So this issue is not of great theoretical moment.

A second issue is of more theoretical interest. According to Feinberg, the fact that a person's character has been worsened, coarsened, corrupted, damaged, etc. does not count as a harm to the person. If a wicked person has no ulterior interest in having a good character, and if such a character is not in his other interests, then the depraved character is not a harm to him, and even if he becomes worse, he does not necessarily become worse off. This view has important implications for what I call “moral paternalism.” By this I mean interference with a person against his will to prevent him from becoming (morally) worse off.

If Feinberg's view is correct, then moral paternalism is impossible in some cases. If a person does not value having a good moral character, or if having such a character does not have contingent connections with goods he does value, i.e., if being a better person does not make him happier, then we cannot improve his (moral) welfare by restricting his freedom. We can, however, as I have argued elsewhere invoke a different kind of principle which is intermediate between legal moralism—the view that it is permissible for the state to interfere simply on the grounds that the conduct in question is immoral—and moral paternalism. Unlike the interference prescribed by the former, we do not interfere simply because the acts interfered with are immoral. Unlike the interference advocated by the latter, we do not interfere simply to make the person better off (morally). Rather we interfere to make the person a better person even if he is not better off as a result.

The next oddity to notice in Feinberg's conception of harm is that it does not include various cases of unfairness, in particular the case of exploitative injustices. We now begin to approach the area of harmless wrongdoing since some of these injustices will not cause harm to those who suffer them. This will be for either of two reasons. First, these injustices might not cause harm because they do not set back the interests of those who are exploited. Second, they might not cause harm because those who are exploited may have consented to them freely, i.e., the Volenti principle is satisfied.Footnote ⁵ Actually it should be noted that this discussion is one of two ways in which Feinberg invokes the Volenti principle. Here he says that the exploitative injustices “may not harm their victims … because the exploitees have freely consented to them.” This seems to interpret the Volenti principle as “to he who has consented no harm is done.” But this contradicts the introduction of the Volenti principle in Volume I, Harm to Others, where he says that “There is no doubt that a person can be harmed … by … the conduct of others to which he fully consents.” He then goes on to consider “the more interesting question … whether a person can ever be wronged by conduct to which he has fully consented.” And he introduces the Volenti principle as the principle that “to one who has consented no wrong is done.”Footnote ⁶

Assuming the latter is the view that Feinberg wants to defend, he is faced with the issue of whether it is permissible to criminalize coercive, mutually beneficial exploitation. An example would be a case in which you fall into a river and cannot swim. I come by and offer to pull you out for a very large sum of money. You agree, I rescue you, and you pay me.

Under the harm to others principle, this conduct cannot be criminalized. Far from harming you, I have saved your life. Yet, at the least, we might think that such a contract should not be enforceable—even though one can foresee that fewer people will be rescued under a regime in which such contracts are not enforceable. Feinberg believes that since no one has been harmed, the only principle which might justify criminalizing this conduct is legal moralism. But since, in his view, the Volenti principle operates here, the exploitee has no complaint, and therefore it is illegitimate to interfere. The issue of whether unfairness is a distinct principle or can be encompassed under the harm to others principle is one that Feinberg discusses sensitively in the case of noncoercive exploitation. Sometimes he finds a harm; sometimes he says that although evil, such exploitation cannot be criminalized. The use of the Volenti principle, therefore, requires further examination.

III. Problem Cases for the Volenti Principle

“To one who has consented no wrong has been done.” This principle, used in conjunction with the harm to others principle, plays an important role in Feinberg's arguments about harmless wrongdoing. In particular, it plays an important role in attacking legal moralism. The most important use is against the counterexample to his view posed by Irving Kristol's proposal to have gladiatorial contests in Yankee Stadium open to consenting adult audiences.Footnote ⁷ Kristol assumes that our intuitions are that such contests should not be allowed. If the only legitimate liberty-limiting principles, however, are the harm to others and offense principles, then how can such contests be restricted? The only people present who could complain about being offended are there voluntarily and freely, so the offense principle does not apply. But what about the harm to others principle? Are not the gladiators who lose the contest, and their lives, seriously harmed? Are they not the victims of assault and murder?

Feinberg thinks not. “The primary evil relied upon by the legal moralist is not that anyone was harmed (i.e., injured and wronged) for no one was, and not that anyone was injured even without being wronged, since that “otherwise evil” is nullified by consent, and there would be an even greater evil, indeed a wrong, if consent were overruled.”Footnote ⁸

Note that although Feinberg is attributing this view to the legal moralist, he is actually doing so because he (Feinberg) believes it to be the correct view.

Now there are a number of distinct notions being employed here. I will break them down so they can be examined separately.

1. 1. A moralized notion of “harm” which is an amalgam of injury plus wrong. One does not count as being “harmed” in this sense, either because one is not injured or because one is not wronged.

2. 2. A notion of injury distinct from being wronged which can be nullified by consent.

So, one can be injured but not harmed (no wrong). One can be wronged but not harmed (no injury). Consent can nullify injury. Consent can nullify being wronged.

From this follows the idea of harmless wrongdoing. To be harmed one must both have one's interests set back and be wronged. If one's interests are not set back one is not harmed even if one has been wronged. This is one example of harmless wrongdoing. Even if one's interests are set back and there has been wrongdoing one may not have been wronged. This is another example of harmless wrongdoing. It is the latter kind of harmless wrongdoing that I am mainly concerned with in this paper.

Now let us return to the gladiatorial contest example. Here is an outline of the argumentative strategy used by Feinberg.

Now, in fact, there is an alternative strategy to hold on to bold liberalism. One could give up the Volenti principle. In the gladiator case, this would involve bringing the relevant conduct under the purview of the harm to others principle by denying that the mere fact that one has consented to some harm, i.e., the risk of death, makes it a case of harmless wrongdoing. If, as some have contended, there are certain rights which cannot be alienated, then one retains the right not to be harmed (even if one has consented to the harm). So the issue becomes one of relative costs to the theory. Does Feinberg lose more than he gains by abandoning the Volenti principle?

The first thing to note is that it would surely be a mistake to abandon the Volenti principle in general. It is clear that consent is a moral power to alter the normative status of various acts. What would otherwise be theft is a gift if I consent to your possession of a good of mine. What would otherwise be rape is merely sex if mutual consent is present. What would otherwise be battery is simply surgery if I consent to its performance on me. Feinberg is surely right to argue that often the issue of whether someone has been wronged depends on whether, and how freely, she has consented to what is done to her.

The Volenti principle is an essential part of his arguments against paternalism as well. Otherwise, the manufacturer and distributor of cigarettes to willing purchasers is on a par with the person who puts poison in your food.

The essential question, therefore, is not whether the Volenti principle is ever valid but whether it has limits. Are there kinds of acts which are still wrongs even if they have been (freely) consented to? Are there limits to what a person may (permissibly) consent to?

We should recall that even Mill, that stalwart opponent of paternalism, made an exception in the case of voluntary slavery. His argument is in effect that the decision to become a slave represents not simply a decision to abandon autonomy temporarily (retaining the possibility of regaining it) but to enter a situation in which one cannot rethink the decision and decide to regain one's autonomy. Given Mill's theory of value according to which the development of autonomous individuals is required for utility, a utilitarian theory cannot accept such a permanent abandonment of future autonomy. Whatever the nature of the argument, it is clearly an instance in which Mill abandons the Volenti principle. Consent in this type of case cannot perform its normal task of immunizing what is done against legal sanction.Footnote ⁹

It is true that Mill's argument in this case does not concern the use of criminal law in order to threaten, but rather the use of civil law to deny enforcement of a contract. But I see no reason to suppose that Mill could not support the former as an additional tool to prevent voluntary slavery.

Arthur Ripstein in a recent article on the harm to others principle argues for a limitation on the Volenti principle in terms of sovereignty.Footnote ¹⁰ Ripstein means by sovereignty, roughly, that every person is entitled to use her powers as she sees fit provided that others may do the same. Here is his argument:

The problem with this argument is that it seems too strong. For it rules out cases of physician-assisted suicide and euthanasia which I believe can be legitimate in certain circumstances. Ripstein seeks to avoid this difficulty by making a distinction between physician-assisted suicide and euthanasia.

My objection to this is that it gives moral weight to the issue of who performs the last causal act in somebody's death, and this does not seem correct. For instance, Doctor Jack Kevorkian had a machine hooked up to the patients whose suicide he assisted, and asked them to press the button to release the lethal medication. But it has always seemed to me that there cannot be a moral difference between those cases and the case in which, say because the patient is paralyzed, Kevorkian has to push the button. If we hold everything else constant, i.e., continued life is not a benefit to the patient, she freely consents to what is being done, she has the right to change her mind at the last moment, and so on, then the difference in who takes the last causal step to securing her death cannot make a moral difference.

I want to present two more cases before turning to argument.

1. (1) Dwarf-tossing. This is a “bar sport” in which dwarfs are fitted out with padding and a helmet, and have a handle attached to their backs. The contest is to see who can throw them the farthest.

2. (2) German cannibal case. A German man advertised for someone willing to be killed and consumed. A person answered the advertisement. He was killed—after some preliminaries which are too grisly to relate—hung on a meat hook, and consumed over a period of ten months.

The issue for me becomes whether one can find limits to the Volenti principle which enable us to reconcile the following views:

1. A. It is legitimate to forbid slavery contracts and gladiatorial contests, and perhaps cases like the German cannibal case and dwarf-tossing.

2. B. It is legitimate, in some circumstances, legally to permit physician-assisted suicide and euthanasia.

3. C. The Volenti principle is valid for many cases of harm to which people consent.

Why is the Volenti principle valid at all? Why does it seem plausible in many cases? The obvious explanation is that consent is a moral power to change the moral status of acts. But it is not clear exactly what the explanation of that power is? There are alternative models.

On one model, by consenting I give up something and transfer it to you. I had the right to use this hat. I had the sole power to determine how it would be used. By giving you the hat, by consenting to your possession of it, I give up my right to determine use and transfer it to you. It is a corollary of this transfer that I no longer have a right to complain when you destroy the hat or wear it in a way that I do not approve.

On another model, that of Kant, consent involves two people “uniting their wills.” It is not a matter of my giving up something. So when I consent to your doing something to me, when you perform the action it is as if we are acting jointly. I have the same responsibility for what you do, as if I were doing it to myself.

On a third model, consenting just is giving up one's right to complain about the crossing of a moral boundary.

These models can have various moral implications. For example, on the third model it does not follow from the fact that you have waived the right to complain that others may not have the right to complain on your behalf. The fact that you are willing to undergo debasing and humiliating conduct (e.g., dwarf-tossing) does not settle the issue of whether we can criticize you and the others for such an agreement. Whether we can, in addition, make it legally difficult for the consensual action to take place will depend on whether some principle such as moral paternalism or legal moralism can be justified. Recall that the former is the view that it is permissible for the state to interfere with consensual acts in order to make the parties morally better off. The latter is the view that it is permissible for the state to interfere simply on the grounds that the conduct in question is immoral.

Suppose we accept the view that the source of the change of moral status is the waiving of one's right to complain about what is done to one. This has a weaker and a stronger version. The weaker is just that one no longer has the status to complain, but it may still be that one's rights have been violated. The stronger is that one no longer has the status to complain because one's rights cannot be violated if one waives them. The Volenti principle seems better explained by the latter reading. It is clearly the one that Feinberg uses in his arguments. The law is limited to the protection of people's rights. Only if there is someone who has been wronged may the state intervene; mere wrongdoing is not enough. What the law may enforce is grievance morality. Satisfaction of the Volenti principle removes the case from the purview of grievance morality.

If we accept Feinberg's premises, then to make them compatible with interference with gladiators and dwarfs we must interpret the Volenti principle in such a way that it is not satisfied in these cases. This means that there are some acts, fully consensual, which leave rights intact (although it may be the case that the person has forfeited her right to complain about the violation). This, of course, is just the doctrine of inalienable rights.

Now “inalienability” comes in weaker and stronger versions as well. The weaker version is that one cannot abandon one's rights permanently, i.e., effect a general abandonment of one's rights. The stronger version is that one cannot abandon them even on a single occasion. Note that the gladiator case differs from the dwarf case just in the fact that the former involves permanent abandonment (at least it risks that) whereas the latter might only involve abandoning one's rights to dignity on a single occasion.

Another ambiguity is between rights which can neither be forfeited nor waived, and those which can be forfeited but not waived. Thus, one might have a right to life which could be forfeited by one's attempting to take the life of another, but not waived, e.g., by consenting to be killed by another.

Finally we must distinguish between waiving the exercise of a right one continues to possess, and waiving the right itself. If I ask your assistance in committing suicide I do not wave my right to life; I relinquish the exercise of that right.

What kind of theory could explain the judgments in the four cases—gladiatorial contest, voluntary slavery, dwarf-tossing, and justified euthanasia?

First, no inalienable rights theory will do. If the right to life is inalienable, then both the gladiator and the euthanasia cases must be classified as violations of the individual's right to life. For similar reasons, a theory such as Ripstein's according to which assigning a right to kill one to another person is incoherent because it is inconsistent with the idea that the individual must retain the legal personality that is a requirement of his “uniting his will” with the other party. If the judgment about euthanasia is correct then rational persons must, in certain circumstances, be entitled to lift the moral barrier against being killed by others.

Second, it does not look plausible that a single theory explains both the slavery and dwarf-tossing cases. The intuition behind not allowing people to enter into slavery contracts is that it is wrong for people to renounce their autonomy on a permanent and irrevocable basis. The intuition behind the dwarf-tossing case is that it is wrong for people to allow themselves to be degraded in certain ways. Of course, if one is prepared to allow as a single theory a view which says that we must weigh the reasons for and against in each case and judge which reasons are weightier, then there could be a single theory. But it would not be informative.

If we accept Feinberg's view that what is done by others to you, with your consent, negates any rights violations (owed to you) then if one is seeking to justify the target cases one must argue against what I shall call the “Feinberg premise”:

(FP) The only wrongs it is legitimate to criminalize are those which violate the rights of some particular victim.

As we have seen, when he talks about “harmless wrongdoing” he really means to cover “harmful wrongdoings which do not wrong any person.”

What is Feinberg's argument for the Feinberg premise? Here are the relevant texts that I could find.

The argument of the first quote assumes that the only reason for using coercion is to produce a good for someone. But this begs the question.

The argument of the second assumes that the only reason for using coercion is to protect someone against harm. But this begs the question.

The argument of the third quote simply repeats in different ways what a grievance morality consists of. It provides no additional argument why only grievance morality can justify coercion. In addition, it equates things which should be kept distinct. Whether or not one is entitled to complain about something depends on many different considerations. The person who launches a deadly attack on you is not entitled to complain if you oppose him with deadly force. Nevertheless, if he is a so-called innocent threat, it is at least arguable that you have no right to use such force. Someone might not be entitled to complain about some treatment because he himself advocates and defends such treatment. But that does not show we are wrong to treat him in that way.

In truth, unlike Feinberg's argument against paternalism, which involves a fair bit of theory about the nature of personal sovereignty, and the way drawing the boundaries as the paternalist suggests is in conflict with a conception (an ideal!) of the autonomous person, his argument here consists largely of making various illuminating distinctions about kinds of evils, and then simply asserting that some of them are not serious enough to warrant preventing by coercion.

The general form of his argument is that, given the importance of personal liberty, if we cannot justify a restriction of liberty by pointing to someone who can complain, then we cannot restrict liberty. But there are various ideals that are at least as important to us as some of our minor grievances. If it is wrong to show contempt for dwarfs by using them like inanimate objects because it is always wrong to be contemptuous of persons on the grounds of their physical characteristics, and if the fact that the dwarfs freely consent does not make the exhibition any less contemptuous, then why is there not as good a reason for criminalizing the conduct as some violation of someone's rights? This, of course, is not an argument but a rhetorical question. In the absence of some argument against it, however, it is entitled to prevail until defeated.