Autonomy and the Justification of Paternalism
by Oliver Friendship (March 2020)
Brain Trust (Conference), Jack Levine, 1935
In liberal democracies such as America, there is an inherent presumption against paternalistic policies. This is because paternalism is seen as destroying personal autonomy, and “appears to offend a fundamental tenet of liberal societies: namely, that the individual is best placed to know what is in his or her interests”. However, while this knee-jerk aversion to paternalism undoubtedly helps protect individual liberty in many instances, that does not mean that all forms of paternalism are inherently wrongful, or necessarily autonomy-limiting. Indeed, in this essay, I will first establish a conception of what constitutes voluntary choice, to then demonstrate that the preservation of autonomy can be better served with paternalism when a choice is non-voluntary, or would substantially negatively impact the voluntariness of future decisions. I will then argue that autonomy is an intrinsic good and that, because paternalism better serves autonomy in many instances, therefore state paternalism is justified under certain circumstances. From this analysis, and although considerations other than autonomy, which are obviously important to the argument regarding the justifiability of paternalistic policies, are not discussed, a reasonable understanding of when state paternalism is justified, and when it is not, can be gleaned.
To demonstrate how paternalistic policies, which are defined as restrictions on action justified solely with regard to the supposed good or welfare of the individual whose action is being restricted, can, under certain circumstances, better serve autonomy than non-intervention, it is first necessary to outline a conception of “voluntary choice”. This is because, behind the notion of autonomy, lies an ideal of a self-governing individual; and “choice comes naturally into the picture as the concrete embodiment of this ideal”. Distinguishing voluntary from involuntary choice is thus the same as distinguishing autonomous from non-autonomous decision-making.
The foremost account of voluntary choice is that of Joel Feinberg. Feinberg establishes five criteria that an individual must satisfy for their choice to be “perfectly voluntary,” or fully autonomous. They must be:
(1) mentally competent,
(3) not subtly manipulated,
(4) not making a decision because of ignorance or mistaken belief, and
(5) not making a decision in circumstances that are distorting, such as because of drug use.
However, “to require that a voluntary act … satisfy allm [criteria] fully would be to apply an impossibly difficult ideal standard, one that would hardly ever be satisfied”. Thus, voluntariness, and hence autonomy, should be treated as a variable concept, with a “voluntariness-threshold” depending on the action taking place; where more risky, irreversible and important decisions require the elements of perfectly voluntary choice to be more fully satisfied for the decision to be considered autonomous.
It is worth noting at this point that some individuals would object to Feinberg’s variable scale in favour of a rigid model of voluntariness; for example Aristotle, who thought a choice was voluntary if it was both made without coercion, and with sufficient relevant knowledge. Such models are more easily applicable than Feinberg’s variable scale, but their inflexibility means that there are inevitable counterexamples that make every single one of them morally unacceptable in some scenario. In Aristotle’s case, it is the fact that his rigid theory allows for choices made by uncoerced, ostensibly informed, psychiatric patients and drug-addicts to be considered voluntary and totally unproblematic.
By employing Feinberg’s superior variable model, it is apparent that autonomy can be better served with paternalism when an individual’s decision does not meet the threshold of voluntary choice. What’s key in this scenario is that paternalistic policies that interfere with choices that do not meet the voluntariness threshold do not destroy autonomy, because, as outlined above, only sufficiently voluntary choices are autonomous. Moreover, paternalism can actually create autonomy in these instances, by helping individuals more fully satisfy the elements of perfectly voluntary choice so that they can meet the threshold for autonomous decision-making on matters that they previously could not.
This point is best illustrated using the example of an individual with a severe mental illness; something uncontroversially recognized by Feinberg, among others, as impeding their ability to meet the necessary threshold of voluntary choice for many important decisions. In such a situation, paternalistic policies, such as compulsory treatment, often allow individuals to minimise the adverse impact of their mental illness on the competency of their decision-making. Therefore, such paternalistic intervention in the non-autonomous decisions of mentally ill individuals can enable them to better satisfy the elements of perfectly voluntary choice, and hence meet the threshold of voluntary choice for more decisions than they were able to before; thus allowing them to exercise autonomy more. The same concept also applies in many other scenarios involving non-autonomous decision-making, such as compulsory drug rehabilitation programs. In this way, the proliferation and preservation of autonomy can actually be better served with paternalism.
However, the issue of whether paternalism better serves autonomy for a decision that does meet the threshold of voluntary choice is a little more complex, and far more controversial. Enacting paternalistic policies when a choice is sufficiently voluntary, and when an individual is exercising autonomy, is “liberty-limiting … [and] a purported justification of the agent’s presumptively wrongful conduct”. Such paternalistic intervention therefore necessarily inhibits autonomy in the process of restricting sufficiently voluntary choice, meaning that in instances where such choices do not impact the individual’s capacity for future autonomous decision-making, such as gun ownership, the preservation of autonomy is not better served with paternalism. It is thus reasonable for citizens of liberal democratic societies to fight against state paternalism of this form.
However, using the fact that paternalistic intervention in a sufficiently voluntary decision destroys the autonomy of the decision-maker in the moment of making the choice, many individuals, often of a libertarian bent, conclude that any and all applications of paternalism relating to sufficiently voluntary choices are autonomy-limiting. However, this view is misguided, because paternalistic restrictions can better serve autonomy when an individual’s sufficiently voluntary choice would substantially negatively impact their future capacity for autonomy, by depriving them of the ability to satisfy some of the elements of perfectly voluntary choice. Take a person’s, theoretically, sufficiently voluntary decision to take a highly addictive and destructive drug for the first time. While it is true in the moment of drug-taking that paternalistic restrictions on the action are autonomy-limiting, the action itself will severely impede the drug-taker’s capacity for future autonomous decision making. Thus, the restriction of the action actually preserves the ability of the individual to satisfy the elements of perfectly voluntary choice, and hence better serves autonomy.
While this approach is obviously nuanced with regard to the correct “autonomy time-horizon” to apply for when the autonomy of a person’s immediate voluntary decision is offset by its longer-term impacts: there are situations, like that sketched above, where the initial paternalistic restriction of autonomy is clearly outweighed by the future protection of autonomy under any reasonable time-horizon. Therefore, in these instances, the preservation of autonomy is better served through paternalism. Importantly, the conceptual framework outlined above demonstrates that the simultaneous aversion of many Americans to gun control, and support for strong anti-drug laws, is not hypocritical when the focus of a person’s reasoning is individual liberty and autonomy.
However, the fact that, under certain circumstances, autonomy is better served with paternalism is not a justification for state paternalism if it cannot be demonstrated that autonomy is a good that should be promoted within society. And, maybe somewhat surprisingly, some academics argue the opposite, contending that autonomy has no value in itself. This argument commonly takes the form of a thought-experiment based on Aldous Huxley’s novel Brave New World. In the work, the wants of the characters are decided for them, and they are without the autonomy to freely travel, read, or express their opinions. However, “this assault on autonomy does not … take away their happiness”; because they can still get enjoyment out of their externally programmed desires. And so, this argument concludes, as it is entirely theoretically possible for someone to be happy without acting autonomously, therefore autonomy is not necessary for happiness, and is thus not a good in itself.
However, this objection fails because it implicitly assumes that only something that is instrumentally valuable for human happiness is a good. This assumption shows a disregard for the fact that humans are purposive beings, to which the sense of actually choosing a course of action and following it is very important; somewhat irrespective of the results. Indeed, this fact makes autonomy the moral basis of human identity, as something that transforms an episodic existence governed by circumstances into the life of a self-directed individual. Crucially, “it is this shaping of our lives … which has value in itself”. Instead of simply desiring the experience of happiness as decided for by others, humans, who innately value lived experience and truly doing, “want to be a certain way [and] to be a certain sort of person”. To not decide your fate autonomously is, in Robert Nozick’s famous phrase, “a kind of suicide”. Thus, as autonomy is an intrinsic good, and because, as shown above, paternalism better serves autonomy under certain circumstances, therefore state paternalism can be justified in many cases.
Overall, I have demonstrated that when an individual’s choice is non-autonomous, or would significantly adversely impact the autonomy of their future decisions, then paternalism can create and safeguard autonomy. Hence, and somewhat counterintuitively, it can be established that under certain circumstances the preservation of autonomy is better served with paternalism. This conclusion, coupled with the fact that autonomy is an intrinsic good, means that state paternalism can be justified in many instances. Thus, while there will likely remain a healthy aversion against paternalistic policies in liberal democracies such as America, it is important to understand that state paternalism can actually aid in the promotion of the noble ideal of individual autonomy, and is not always opposed to the principles that Americans rightly hold dear.
 New, Bill, 1999, “Paternalism and Public Policy,” Economics and Philosophy 15(1): 63-83. Pg. 63.
 Dan-Cohen, Meir, 1992, “Conceptions of Choice and Conceptions of Autonomy,” Ethics 102(2): 221-243. Pg. 232.
 Feinberg, Joel, 1986, “Harm to Self: The Moral Limits of the Criminal Law,” Oxford University Press. Pg. 115.
 Ibid., pg.115.
 Pope, Thaddeus, 2004, “Counting the Dragon’s Teeth and Claws: The Definition of Hard Paternalism,” Georgia State University Law Review 20(3): 659-722. Pg. 714.
 Young, Robert, 1982, “The Value of Autonomy,” The Philosophical Quarterly 32(126): 35-44. Pg. 37.
 Ibid., pg. 38.
 Nozick, Robert, 1974, “Anarchy, State and Utopia,” New York: Basic Books. Pg. 43.
 Ibid., pg. 43.
Follow NER on Twitter @NERIconoclast