I assume that most of the controversy about the topic of paternalism involves the question of whether and under what circumstances it is justified. An answer must address two preliminary matters: what paternalism is, and what would qualify as a justification for it. Since each of these background issues could consume an entire essay, I will attempt to say as little as possible about them in order to move directly to what I hope is more central. But these matters cannot be avoided altogether. I further assume that legal philosophers are less likely to be interested in paternalism per se than in paternalism in the domain of law. Thus I will concentrate on the special problems that arise in justifying legal paternalism. As we will see, these problems are formidable. Little of the progress that moral philosophers have achieved in justifying paternalism in personal relations is readily adaptable to legal contexts.
1. What Paternalism Is and What Counts as a Justification for It
On most occasions in which A is justified in not allowing B to act according to his preferences, A’s objective is to protect persons who might be harmed by B’s behavior. When A treats B paternalistically, however, B is prevented from acting because of the adverse effects on B himself. As a very rough approximation, A treats B paternalistically when A interferes with B’s freedom for B’s own good—to protect or promote B’s health, safety, economic interests or moral well-being.
When philosophers attempt to be more precise about the nature of paternalism, however, the details of their definitions vary considerably (Archer (1990)). They debate whether particular examples that deviate from the foregoing approximation do or do not qualify as genuine instances of paternalism. I will mention two (of many) such debates briefly. Still, for reasons that will become clear, I avoid commitment about how paternalism should ultimately be defined.
First, some philosophers point out that paternalism need not involve two parties. In some examples, only one party is needed. Suppose that Jane understands her proclivity to gain weight, so she locks her refrigerator at night. Is her act of making food inaccessible an instance of paternalism in which Jane plays the role of both A (her present self) and B (her later self whose preferences she disregards)? In other examples, three or more parties are involved. Suppose that parent A forbids his oldest son C from sharing cigarettes with his youngest son B on the grounds that tobacco is bad for B’s health. This case is unlike a typical situation in which A coerces C to prevent him from harming B, since B may be eager to engage in the proscribed transaction. In this case, we again are invited to conclude, A treats B paternalistically. Paternalism in this case is said to be impure or indirect (Dworkin (1972)), since A interferes directly with the liberty of C, and only indirectly with that of B, the intended beneficiary.
A second debate about the nature of paternalism stems from the claim that the paternalist need not interfere in anyone’s freedom (Gert and Culver (1976)). Suppose, for example, that doctors in an emergency room are presented with an unconscious victim whose life can be saved only by surgery. Since the patient remains unconscious throughout the procedure, it is hard to conclude that the surgery interferes with her liberty. Still, it seems apparent that the doctor treats the patient paternalistically.
So-called libertarian paternalism poses a possible counter example to the claim that paternalists interfere with liberty (Sunstein (2006)). Libertarian paternalism works primarily by designing default rules to correct for well–known cognitive biases and volitional lapses, thereby minimizing the likelihood that persons will make decisions that are contrary to their own interest (Trout (2005)). Consider the following two examples. Rather than explicitly choosing to participate in an efficient company health plan, employees might be enrolled automatically unless they opt out. Seat belts might be constructed to buckle immediately upon closing a car door, although occupants could unbuckle them if they choose. Are these provisions really paternalistic? As long as persons can change these default rules, no interference with choice has occurred. Notice that individuals “can” alter the default rule in two senses. First, persons face no penalty if they elect not to comply. Second, opting out is not onerous, requiring a mere stroke of the pen or push of a button. When these two conditions are satisfied, it seems more accurate to construe these rules as designed to influence persons to pursue their self-interest.
Are these two kinds of nonstandard cases genuine counter examples to my rough approximation, requiring that my definition be modified? Probably. But I am skeptical that a “right answer” to such questions can be defended. That is, I am unsure that we can decide whether each nonstandard case is or is not a “real” instance of paternalism. I suspect that some cases are simply better or worse examples than others. We should not expect decisive reasons for or against categorizing an example as an instance of paternalism. Ultimately, a philosopher who insists on a yes-or-no answer to the question of whether a given case is an instance of paternalism must resort to stipulation. The more important issue, I submit, is whether the behavior of A (or the default rule in question) is justified in the foregoing examples. Generalizations about whether and under what circumstances paradigmatic cases of paternalism are justified may or may not be defensible when applied to nonstandard situations.
Whatever else paternalism is taken to be, I suggest it should be understood as a reason or motivation for acting (Grill (2007)). An interference with liberty should not be construed as paternalistic in virtue of its effects. Consider two kinds of case in which effect and motivation diverge. Suppose that A somehow succeeds in advancing B’s interests even though A’s reason for interfering in B’s freedom is to enrich himself. Or suppose that A fails to advance B’s interests despite A’s best efforts to do so. I believe that the latter but not the former case should be categorized as an instance of paternalism. This categorization follows from defining paternalism in terms of the motives that lead A to interfere with B, rather than by how A’s interference actually affects B’s welfare. Although consequences may be important in deciding whether paternalism is justified, I believe they are unimportant in understanding what paternalism is. The decision to construe paternalism as a reason or motivation has several important implications that I will later discuss, with special emphasis on the law.
It may be helpful to present a concrete, ordinary case from which to generalize about the nature and justifiability of paternalism as well as to highlight the difficulties of defending it in the legal domain. The best examples are drawn from personal relations. When asked about his preference, four-year-old Bobby objects to taking his oral flu vaccine. His father has tried to persuade him, but Bobby is unmoved. I stipulate that the father’s only reason for wanting Bobby to take his vaccine is to safeguard his child’s health. It seems reasonable for his father not to permit Bobby to eat ice cream, his favorite dessert, unless he takes the vaccine. This example not only describes a situation in which Bobby is treated paternalistically, but also represents a relatively clear case in which the treatment is justified. In any event, I make these two assumptions about this case.
Five criteria conspire to make this example a relatively clear case of justified paternalism. First, the intrusion involves a minor interference in Bobby’s liberty. Bobby is not beaten or deprived of something of great significance to induce him to change his behavior. Second, the objective sought by his father is obviously valuable. No one contests the importance of health and the ill effects of contracting flu. Third, the means chosen are likely to promote this end. The vaccine is effective, and Bobby is likely to agree to take the vaccine if his desire for ice cream is strong. Fourth, Bobby himself is not in a favorable position to make the right decision. Children have notorious cognitive and volitional deficiencies relative to competent adults that prevent them from recognizing their best interests or from acting appropriately when they do. Fifth, his father stands in an ideal relationship to Bobby to treat him paternalistically. Parents have special duties to protect and enhance the welfare of their children. My example satisfies each of these five criteria, and the easiest cases of justified paternalism do so as well. I do not insist that any of these conditions is necessary to justify paternalism. I simply allege that particular examples will prove more difficult to justify as they deviate from this paradigm.
2. Can Laws Be Paternalistic?
If paternalism is construed as a particular reason for failing to regard the preference of a person as decisive, disagreements about whether a given law is or is not paternalistic may be misguided and futile. Laws do not seem to be the kinds of things that can be paternalistic; only reasons can be paternalistic. Still, claims that a law is paternalistic are frequently voiced by philosophers and should not be dismissed as confusions (Goldman and Goldman (1990)). How should these claims be understood? Perhaps a law should be categorized as paternalistic when it exists for paternalistic reasons.
Several difficulties plague this attempt to understand how a law can be paternalistic. What does it mean to “exist for paternalistic reasons”? Many of these difficulties are familiar to legal philosophers who defend theories of statutory or constitutional interpretation that attach significance to legislative intent. The first kind of problem arises even when a legislator is a single individual, like a monarch. A legislator is not generally required to disclose his motivation for enacting legislation. In most cases, his reason must be inferred from various sources, many of which can be ambiguous or even contradictory. In addition, a legislator may have no single reason to favor a given law; he may have several distinct purposes in mind. The second kind of problem arises in a modern democracy in which laws are enacted by a number of representatives. How can we combine the separate motives of several individuals into the purpose of a group? Should we count only those legislators who voted in favor of the law, or should we also consider the reasons of those who opposed it? Even if we agree about which legislators should count, these persons are bound to have diverse reasons for enacting the statute. One legislator may vote in favor of a bill for paternalistic reasons, another may vote for the bill even though he rejects this paternalistic reason, while yet a third may support the bill for reasons having nothing to do with its merits. How can we possibly decide whether to classify such a law as paternalistic? Finally, the persistence of law complicates endeavors to categorize it. All of the legislators who passed a statute may be dead. Suppose that most of the legislators who enacted a law did so for paternalistic reasons, but this rationale is now widely discredited and legislators have non-paternalistic reasons for not repealing the law. Since the law persists for altogether different reasons than those that led to its enactment, should we continue to classify the law according to the original motivation of those who created it? In light of these (and other) difficulties, we might despair of any prospects of identifying “the reason” for a law.
Problems in categorizing laws as paternalistic arise whenever particular examples are discussed. Consider a statute forbidding child labor. Presumably, many legislators favored such a law in order to protect the welfare of children. An additional reason for the statute, of course, is to preserve the jobs and wages of adults who otherwise would be forced to compete with underage workers. As far as I can tell, there simply is no definitive answer to the question of whether such a law is or is not paternalistic. I suspect that every actual law that anyone has ever been inclined to categorize as paternalistic resembles the example of child labor in this respect. Here, then, is the first of many difficulties that arise in attempts to apply philosophical insights about paternalism in personal relationships to legal contexts. Admittedly, on some occasions, the motivations for interference can be complicated and unclear even in personal relationships. Flu vaccines, for example, help to prevent other persons from contracting the disease. I assume, however, that Bobby’s father may be confident about his reason for requiring his son to take the vaccine. The difficulties in classifying a given interference as paternalistic are seldom as formidable in personal relationships as in legal contexts.
This fact will become important when we turn to whether and under what circumstances paternalism is justified. Most philosophers, it is fair to say, have relatively strong intuitions against the justifiability of paternalism, at least when it is imposed on sane adults. Suppose that a philosopher becomes persuaded of a theory according to which paternalism is never justified, or is justified only under narrowly specified conditions. Of what value is his theory in assessing legislation in the real world? It is hard to see how his theory could be used to condemn any existing law. That is, he cannot demand that any particular law must be repealed because he is convinced that his theory is correct. All that he is entitled to conclude is that such a law is unjustified insofar as it exists for paternalistic reasons.
Nonetheless, a theory about the conditions under which paternalism is justified would have some limited relevance in legal contexts. Such a theory might constrain the considerations to which legislators are allowed to appeal in their deliberations about whether to support a given piece of legislation. That is, a legislator may be persuaded that some kinds of reasons he might otherwise cite in support of a proposed law should not be permitted to count in its favor. Alternatively, such a theory might constrain the considerations to which citizens are allowed to appeal in making judgments about whether a law is justified (Waldron (1993)). After all, citizens in a democracy should be encouraged to make their own judgments about whether laws are good or bad. A theory about the conditions under which paternalism is acceptable as a rationale could be useful for these purposes.
Despite these enormous obstacles, I remain hopeful that the label “paternalistic” can meaningfully be applied to given laws, and that a theory of when paternalism is justified in personal relations is relevant to the issue of whether such laws are justified. Perhaps the most promising proposal for interpreting the claim that a given law is paternalistic is that “the most plausible rationale” or “the best rationale” in favor of the law is to limit the freedom of persons for their own good. According to this proposal, a law can be paternalistic even though no one, past or present, ever thought to defend it on paternalistic grounds. “The best rationale” of a law can be something other than the reasons that actually led anyone to support it. Of course, commentators are bound to disagree about whether one rationale is more plausible than another. But those who conclude that the best reason in favor of a given law is paternalistic are committed to categorizing it as paternalistic. Even though I tend to favor this approach, we should remain cautious and tentative when we attach the label “paternalistic” to a given law.
Thus far, I have pointed out some of the difficulties in applying a definition of paternalism to particular laws. In addition, our understanding of what paternalism is becomes less clear as we move from nonlegal to legal contexts. My central basis for this claim is that laws are necessarily general and applicable to groups of persons whose circumstances differ widely. The fact that a law is applicable to a group of persons— whereas a paradigm case of paternalism in a personal relationship is applicable only to a single individual—gives rise to many of the perplexities in understanding the nature and justification of legal paternalism. Suppose we (somehow) decide that legislators have enacted a given law to limit the freedom of persons for their own good. Even so, we are likely to find that the law does not disregard the preferences of each and every person whose liberty is restricted. Some limitations of liberty are best construed as devices that enable persons to attain their antecedent objective. I am doubtful that we should deem a rationale as paternalistic when it enables a person to achieve an objective that he recognizes as desirable but is unlikely to attain in the absence of the legislation that limits his liberty (Dworkin (1972)).