Authority and Independence

I. Introduction

Which is more important: deciding correctly or deciding for yourself? Presumably, there is no general answer. Sometimes what matters most is making the right decision, as when deciding which course of medical treatment to pursue. Sometimes, making one’s own decision matters a great deal—choosing one’s partner or occupation, to take the most obvious examples. As these two familiar objectives do not always coincide, a way to navigate this question is called for. Making the right decision sometimes requires abandoning one’s independent choice and relying on experts and external authorities, and deciding for oneself will presumably entail some non-optimal choices.

On one approach such tradeoffs pose a practical challenge, not a theoretical quandary. According to this view, what we should do is determined by the reasons we have (I shall therefore label it The Rationalist View). Though reasons are often difficult to decipher, the possibility that the best way to adhere to the reasons we have is by relying on others poses no theoretical difficulty. There is a variety of reasons for which one might benefit from relying on external support in one’s decision-making. Ignorance, for example, is a reason for relying on experts like doctors, mechanics, and accountants. The opportunity to rely on the expertise of others is one of the benefits of a social division of labor. Overcoming weakness of the will is another benefit of relying on others, whether persons, like coaches and trainers, or instruments, like alarm clocks and traffic lights. These are, as Jon Elster helpfully put it, “technique[s] for achieving rationality by indirect means.”¹ ‘Indirect’ because they do not operate directly through one’s own judgment but rather replace it with adherence to the dictate of someone (or something) else.

But what about freedom? Does relying on others not conflict with an individual’s freedom or autonomy? From one perspective, it need not. If the ends were set and the means were chosen by oneself, then, though indirectly, one is still acting on one’s own judgment and freedom remains fully intact. Moreover, since “binding oneself,” as Elster says, “is a privileged way of resolving the problem of weakness of will,” it can be seen as a way of being one’s own master. Thus, the sailors restraining Ulysses to the mast are not an obstruction but rather instruments of his will. Experts and authorities are means by which individuals can improve their conduct relative to how they would fare if they were to act on their own. This doesn’t take away from their autonomy and may even enhance it.

But things are not so simple. Certain choices, it seems, require direct connection to the individual’s own judgment. Presumably, it might make sense to consult with people who know you well before choosing an occupation or a spouse. But it does not seem appropriate to defer to their judgment on such matters, even if it is superior to yours. Forming a meaningful intimate relationship with one’s spouse presumably has something to do with the partnership being the object of one’s own choice and not the outcome of submission to the directive of a third party. Or suppose Catholic doctrine is correct and the Pope has infallible access to eternal truths. Still, it seems that we need not subject our every decision to the Pope’s direction.² This apparently has something to do with the value of living life by one’s own lights, being one’s own master, or something along these lines. If too many of a person’s choices are delegated to others—even if based on one’s own judgment that they are superior at making such choices—something seems to have gone wrong.

Adherents of the rationalist view see all this as fully compatible with their position. What matters is making the right decisions, but sometimes the right decision is the one you make yourself. Some choices should not be delegated to others, either because they possess certain features which make deciding them for oneself significant, or to maintain general control over one’s life. Otherwise, rationality ought to be maximized and outcomes ought to be optimized, so one should rely on indirect instruments when they are the most effective means to one’s ends. So long as enough of our decisions and all (or most) of our special decisions (those we should make for ourselves) are made by us, everything else can be governed by indirect means. So says the rationalist.

One of the virtues of the rationalist view is its seamless integration with an instrumental account of authority—the apparent battleground between getting things right and deciding for oneself. A compelling reason to follow an authority is that it makes it more likely to do the right thing. On the other hand, obeying another seems a paradigm of giving up self-direction. Instrumentalism dissolves the tension. As Joseph Raz puts it, “fundamentally there is nothing special” about authority as a source of obligations. It is just another way of achieving our goals by indirect means and thereby a means of self-authorship.

This wholesome view, however, rests on a set of assumptions often left implicit. My aim in this paper is to flesh out these assumptions, upon which certain modifications of the view are required. More importantly, the tension between independence and authority, between getting things right and deciding for yourself, resurfaces. The bar for justifying authority is higher than instrumentalist views make it seem. And even when it is met (and obeying authority is rational), the aim of independence—namely self-rule or self-direction—might be jeopardized.

I begin with a presentation of the rationalist view, stressing one of its key assumptions which has received less notice, and then analyze the way it aims to accommodate independence through its response to the challenge of outsourcing self-government, namely by restricting independence to specific domains and types of decisions. I argue that this response has undesirable paternalistic implications and, more importantly, that the rationale underlying it—namely, the value of creating one’s own life—cannot be confined to specific domains and decisions. In contrast to the rationalist conception of self-creation as authorship, stressing life-shaping rational decisions, I propose the notion of ownership, asking which decisions and acts are attributable to a person. A life is lived, not authored—it is the cumulative effect of myriad choices and acts, big and small, reasoned and not reasoned, spanning all spheres of life. Making them one’s own is not about conforming to reason and can systematically conflict with deference to authority.

II. The Rationalist View

Joseph Raz devised the most elaborate and carefully argued rationalist account of authority. According to his Service Conception,

The normal way to establish that a person has authority over another person involves showing that the alleged subject is likely better to comply with reasons which apply to him (other than the alleged authoritative directives) if he accepts the directives of the alleged authority as authoritatively binding and tries to follow them, rather than by trying to follow the reasons which apply to him directly.³

This is Raz’s famous Normal Justification Thesis (NJT). Authority, the idea is, can only be legitimate when it serves its purported subjects by facilitating their adherence to the reasons which apply to them anyway (independent of its directives). Authority, or subjecting our will to the judgment of others, is “simply one device, one method, through the use of which people can achieve the goal (telos) of their capacity for rational action, albeit not through its direct use.”⁴

But authority is not like other devices. Setting alarm clocks and seeking advice, for instance, do not constrain one’s will in a meaningful way. Once bound, Ulysses’s will is clearly constrained but this allows him to realize another, arguably deeper set of desires, namely not to fall prey to the sirens.⁵ Another kind of subjection is normative. When valid, promises, contracts, and authorities create normative constraints. Upon being voluntarily undertaken, promises and contracts give rise to obligations, while authority, or at least particular directives it may issue, can create obligations even without voluntary submission.

Furthermore, authority involves deference, a surrender of one’s judgment. Promises and requests create obligations, or, in Raz’s terminology, exclusionary reasons. Like authoritative orders, they “exclude considerations of the recipient’s present desires.”⁶ But when the law, the police, or a military superior issue a command, not only present desires are excluded; most other considerations are excluded as well, including the addressee’s tastes, preferences, values, and wishes.⁷ In particular, judgments pertaining to the situation at hand—the reasons that were supposed to be considered by the authority—are excluded from consideration by the subject. “One who commands is … trying to create a situation in which the addressee will do wrong to act on the balance of reasons. He is replacing his authority for the addressee’s judgment on the balance.”⁸ Practical authorities, as opposed to theoretical, do not only inform us of the reasons we have, leaving it up to us to decide what to do. Rather, they tell us what to do, aiming to replace our judgment with theirs. We are “not allowed to second guess the wisdom or advisability of the authority’s directives.”⁹ This deference to the judgment of another is what distinguishes authoritative directives from other forms of self-binding, such as promises, requests, or advice.¹⁰

How can this be compatible with the idea that we should decide things for ourselves? Raz addresses this by adding a second condition to his account of authority, which he calls The Independence Condition:

that the matters regarding which the first condition is met are such that with respect to them it is better to conform to reason than to decide for oneself, unaided by authority.¹¹

When is this condition satisfied? Or, to put it the other way around, when is it better to decide for oneself than to (fully) conform to reason? Raz mentions three considerations which may underlie the value of deciding for oneself. First, we might favor independent choice in order to cultivate decision-making capability and self-reliance (this applies primarily, though not exclusively, to children). Another, non-instrumental, reason to favor independent choice even over better outcomes has to do with the value of certain states and relations. The natural examples, which we have already mentioned, are choosing one’s occupation or one’s spouse: even if a parent or a friend is better suited to identify the ideal spouse, we are inclined to think that one’s own preference should decide such matters. The reason might be that intimate personal relationships are more valuable when voluntarily and independently chosen.¹² A third consideration in favor of independent choice is a general concern: “We are not fully ourselves if too many of our decisions are not taken by us.”¹³

None of this threatens the possibility of legitimate authority on Raz’s account for two reasons. First, these three considerations leave plenty of practical matters susceptible to the purview of authorities. Second, all three cases are perfectly consistent with the Service Conception’s essential approach.

The former case for self-reliance (parents and children) is instrumental where the end is to secure what conformity with reason will, in the long run, secure; the latter case (marriage) depends on the fact that there are reasons that can only be satisfied by independent action. Both of them trace the concerns behind independence back to concerns with satisfying reasons.¹⁴

This rationale is easily applied to the third consideration in favor of independence. Since we have good reason to be ourselves, we have good reason not to delegate decisions on some important questions to others. Deciding them for ourselves is therefore compatible with the Service Conception’s instrumental approach.

Two things should be noted at this point. The force of the first claim—that most of our practical affairs are left untouched by these considerations—depends on whether these considerations are exhaustive. If there are other, more extensive reasons to value independence, the room for authority will narrow accordingly. Secondly, some of Raz’s comments suggest that he thinks the competing considerations of conformity to reason and independence neatly align with discrete realms of human activity. In certain areas—like the choice of relations or occupation—we care about independence.¹⁵ In others, it’s all about optimizing outcomes.

A similar picture emerges from Steven Wall’s account of autonomy. Wall rejects Robert Paul Wolff’s philosophical anarchism according to which “the autonomous person must not surrender decision making authority to anyone on any occasion.”¹⁶ Wall argues that “if this view were correct, then the autonomous person could not be fully reasonable.”¹⁷ But if it is important that we decide for ourselves, then refraining from the surrender of judgment even to a purported authority who can make better decisions need not be unreasonable.¹⁸ Wall might reply that reasonability is a prudential notion—taking the best means to one’s aims, and deciding for ourselves is not a practical end. But this hardly settles the matter, since if it is valuable in and of itself, then independence can conflict with reasonability. As Raz recognizes, the two concerns “may be radically different, and the cases for conformity and independence may be incommensurate.”¹⁹

Yet Raz is unperturbed by the possibility of conflict between conformity with reason and independence. This worry, he claims, “seems to be premised on the thought that the concerns that underlie reasons with which we should conform and those that underlie the reason to act independently of authority have nothing to do with each other. But that is not so.”²⁰ The reason Raz thinks it is not so is because of what I will call The Rationalist Assumption:

We value the ability to exercise one’s judgment and to rely on it in action, but it is a capacity we value because of its purpose, which is, by its very nature, to secure conformity with reason . . . the primary value of our general ability to act by our own judgment derives from the concern to conform to reasons.²¹

What Raz seems to be saying here is that although conforming to our practical reasons and maintaining independence are separate concerns, we should not be too worried about their potential conflict because the latter is inherently subordinate to the former. Action by its very nature is directed at the attainment of certain (practical) ends. Acting independently is valuable when it facilitates the attainment of these ends. Deciding for ourselves can facilitate our conformity with reason either instrumentally (because we are best positioned to identify these reasons), or directly (because we have strong reasons to decide for ourselves on the matter at hand). In either case, the two concerns—conformity with reasons and independence—do not come into conflict. Thus, for example, since we have good reasons to cultivate our self-reliance we have a reason to decide for ourselves sometimes. Even if the particular decisions will lead to sub-optimal results, the overall conformity with reasons will be greater than if we always obey others because it will improve our self-reliance. Putting it more generally, the value of acting on our own judgment consists in the fact that it is sometimes the best way to conform to the reasons that apply to us given our rational ends. This assumption, I shall argue, is less innocent than it seems.

III. Outsourcing Self-Government

If deciding for ourselves is only valuable because, and to the extent that it allows us to conform to reasons we have independent of it, we should be willing to give it up whenever a more effective way to conform to these reasons is available. This is the challenge Mikhail Valdman posed a few years ago.²² Since relying on more competent experts is the rational thing to do, we should submit to the authority of a wise and benevolent decider who can help us realize our goals and aims more effectively. Our intuitive reluctance to cede control to such an authority, or “outsourcing self-government,” stems not from the value we attach to deciding for ourselves, Valdman argues, but rather “from our concern with how this decider will go about improving our welfare.”²³ The worry is that “our decider will attempt to improve our welfare in a way that does not respect our deepest commitments.”²⁴ But if it can be guaranteed that the decider will decide on the basis of our own values and commitments, then there is no reason not to subject oneself to her authority, which will only enhance one’s practical rationality. “Nothing of intrinsic value will be lost by outsourcing self-government to a more competent decider, provided that she will run our life in accordance with our deepest commitments.”²⁵

One of the targets of Valdman’s argument is Steven Wall, who claims that “it is intrinsically good for people to take charge of their own affairs and lead their lives on their own terms.”²⁶ In response, Wall claims that Valdman’s argument rests on a false assumption, namely “that there is a sharp distinction between our evaluative stance—the goals and concerns that matter to us—and the decisions that we make to pursue or realize those concerns.”²⁷ In fact, Wall claims, our evaluative stance is developed over time, in the course of making decisions. The point to note is that although he rejects Valdman’s conclusion, Wall shares his rationalist assumption. He accepts that making decisions for oneself, detached from its contribution to the construction of one’s evaluative stance, is not intrinsically valuable.²⁸ Otherwise, as we have already seen, it would be in tension with practical reasonableness. And autonomy should not “require us to be practically unreasonable.”²⁹

But what about our evaluative stance itself? What if we could cede decision-making authority over it too to a wiser decider (say, the infallible Pope)? If autonomy must not entail a requirement to be less reasonable, then shouldn’t we submit even the formation of our evaluative stance to the authority of wiser experts? Valdman is not reluctant to allow this. He admits that, all else being equal, we would prefer a decider who takes our goals and values and merely makes us act in accordance with them more effectively, than one who changes them or ignores them altogether, “even if [it] would produce for us less net welfare than [the latter].” The reason is because “we want not just to fare well but to fare well on our own terms.”³⁰ However, “our own terms” is for Valdman synonymous with “our deepest commitments.” Thus, while he concedes that “there is value in there being a deep connection between a life’s contents and the person who is living it,” for Valdman, “this connection is cashed out in terms of a tracking relationship between one’s deepest commitments and one’s life’s contents.”³¹ In other words, so long as we live in accordance with our deep commitments, it does not matter how these commitments were formed. “there is no intrinsic value in having [final] authority—even over decisions that concern the direction and shape of one’s life.”³² If a wise and benevolent decider could form better ones for us (perhaps even impose them upon us?) this would not be objectionable. In fact, it would be preferable.

But this doesn’t seem right. When it comes to being one’s own person—being fully ourselves, living a self-directed life, a life of one’s own making—what matters is not only, or even primarily, that one’s conduct align with one’s core values and commitments or one’s evaluative stance, but that it be the outcome of one’s own choices.³³ As Wall recognizes “autonomy is not only a matter of a person living a life that is congruent with his understanding of what is valuable and worth doing, but also a matter of him actually making choices about how to lead his life.”³⁴ If tracking your evaluative stance, or having one that is yours in the first place, is so important, then, presumably, there must be more than just any causal connection between you and your evaluative stance (such as having a wise decider create it for you on the basis of your character, needs, behavior, etc.). Thus, for Wall there is value in making one’s own decisions. But—and this is a crucial point—this value is restricted to decisions contributing to the formation of one’s evaluative stance because only they have a bearing on one’s autonomy. “An autonomous person,” Wall says, is not alienated from “the goals and pursuits that have come to define his life… He develops a normative conception of himself—an evaluative stance—and this stance guides him in his selection and pursuit of goals.”³⁵

Surely different mechanisms can be imagined by which experts can impose their decisions without individuals feeling alienated from them (many cases of subjection to charismatic authorities are experienced as authentic surrender, revealing one’s “true self,” rather than alienation). The notion of alienation Wall is employing here is therefore not psychological; it consists not in what individuals might feel, but in whether their goals and pursuits are a consequence of their own choices. According to Wall, then, being autonomous entails not being alienated from one’s goals and pursuits, which requires that one form one’s own evaluative stance. In other words, it is not only the congruence between the content of choice and some feature of the person (their values, plans, or ideals) that matters, but their causal involvement in bringing it about.³⁶ On this view, deciding for oneself is valuable, but—and this is the key caveat—this value is confined to the limited domain of one’s evaluative stance and conditional on compliance with the demands of reason.

Though he doesn’t employ the term “evaluative stance,” Raz also restricts the value of independent choice to fundamental, long-term plans and commitments: “autonomy and positive liberty bear directly on relatively pervasive goals and relationships and affect more restricted options only inasmuch as they affect one’s ability to pursue the more pervasive ones.”³⁷ By contrast, “denying someone a certain choice of ice-cream is generally admitted to be insignificant to the degree of autonomy enjoyed by that person.”³⁸

To resist Valdman’s independence-negating conclusion, then, rationalists reject his second premise—that coherence with one’s core commitments suffices for autonomy—but they share his first premise, which restricts the discussion to one’s core commitments and values. In other words, they assert the value of independence, conceding the importance of deciding for ourselves rather than just having our conduct cohere with our values, but restrict this to a qualified domain of choice.

This might seem arbitrary. If alienation from one’s core commitments is bad, why is alienation from other elements of one’s psychological and practical life not a problem? The answer, as we just saw, is that autonomy is determined at the level of a person’s evaluative stance, the one that “guides him in his selection and pursuit of goals.”³⁹ Authoring your own life, or being your own person, is realized in forming your evaluative stance, your core values and commitments.

This approach is not confined to perfectionists. Consider Rawls’s conception of the person as “a human life lived according to a plan.”⁴⁰ This view, which Rawls adopts from Josiah Royce, also associates autonomy with some restricted domain of life-shaping decisions. As Samuel Freeman summarizes, “parties to Rawls’s social contract, as free persons, have a fundamental interest in (what Rawls later calls) their rational autonomy, i.e., their freedom to shape their life plans, revise their final ends, and rationally pursue a conception of the good that is of their choosing.”⁴¹

Now, if independence is valuable to the extent that it facilitates conformity to reason, and if deciding for oneself is important only when it comes to “life-shaping” decisions, plans, or evaluative stances, then robbing people of their free choice in other matters is unobjectionable, especially if it helps them better conform to the reasons that apply to them. But this doesn’t seem right. As Raz says himself: “It is intolerable that we should have no influence over the choice of our occupation or of our friends. But it is equally unacceptable that we should not be able to decide on trivia such as when to wash and when to comb our hair.”⁴²

To explain the problem with interferences in such trivial matters and, more generally, violations of negative liberty which do not directly diminish autonomy, Raz appeals to their expressive effects: “Coercing another may express contempt, or at any rate disrespect for his autonomy.”⁴³ But this need not be the case. If the rationalist view is correct, paternalistic interferences can be driven by the aim of enhancing autonomy, liberating their subjects from urges, weaknesses, and beliefs that encumber their exercise of autonomy. Moreover, even when interference expresses disrespect for autonomy, it need not undermine it.⁴⁴ Furthermore, expressing disrespect and even contempt for someone’s autonomy may not be nice, but it does not seem like the kind of consideration which can underwrite a principled prohibition on interference. Benefits to individual or social welfare can presumably justify some measure of disrespect. Finally, paternalistic interventions seem objectionable not (or not merely) because they are disrespectful or offensive, but because they violate their target’s freedom. In fact, what makes them disrespectful and offensive is the fact that they violate freedom, or a person’s sovereignty over themselves.⁴⁵

It appears, then, that autonomy thus understood does not support the robust opposition to paternalism typical of liberalism.⁴⁶ This indeed is Wall’s conclusion: “it does not look like symbolic considerations can be successfully invoked to rule out the permissibility of well-crafted coercive government measures designed to help people lead valuable lives.”⁴⁷

Raz and Wall are not the only ones saddled with these paternalist implications. Note that they follow not from their perfectionism, but from their rationalism.⁴⁸ They are, therefore, shared by non-perfectionist rationalists. On Rawls’s view, a person’s good is defined by that person’s rational life plan, which is “the plan that would be decided upon as the outcome of careful reflection in which the agent reviewed, in the light of all the relevant facts, what it would be like to carry out these plans and thereby ascertained the course of action that would best realize his more fundamental desires.”⁴⁹ This is clearly a demanding standard and one we often fail to meet.⁵⁰ Thus, if he is to resist paternalism, Rawls must assume that conforming to one’s rational life plan is not a condition for independence. In other words, independent choice must be valuable over and above maximizing conformity with one’s good.

It may appear as if this is precisely what the argument for the first principle of justice and its priority entails. But this argument applies only to the liberties Rawls identifies as “basic,” which include liberty of conscience, political liberties, and freedom of the person, not the liberty to decide mundane questions for oneself (such as what flavor of ice-cream to have, or whether to smoke cigarettes).⁵¹ Secondly, Rawls’s justification of the priority of liberty in fact rests on the idea of rational life-plans. Behind the veil of ignorance, Rawls argues, rational individuals will choose to prioritize liberty of conscience, for example, because only such an arrangement will guarantee protection of their liberty to exercise their “non-negotiable” religious convictions,⁵² if they happen to have such. But eating ice-cream and smoking marijuana presumably need not be among the fundamental aims involved in any reasonable conception of the good. Rawls’s argument for the basic liberties rests on the fact that the interests protected by them “are not mere preferences,” whereas paternalist interferences often pertain precisely to mere preferences. Thirdly, paternalist interferences are conducive to the fulfillment of their subject’s life-plan, which, recall, is defined in ideal terms, i.e. as the best life for that individual, not in terms of one’s actual choices, let alone preferences.

Rawls acknowledges that some forms of paternalism are acceptable from behind a veil of ignorance. It is rational for individuals placed behind such a veil

to protect themselves against their own irrational inclinations by consenting to a scheme of penalties that may give them a sufficient motive to avoid foolish actions and by accepting certain impositions designed to undo the unfortunate consequences of their imprudent behavior.⁵³

The reasonability of such hypothetical consent, remember, warrants coercive interventions even absent actual consent. If an external authority (something like Valdman’s experts) can bring one into greater conformity with one’s rational life plan, then there seems to be no reason for Rawls to object. It is of course possible to claim that given the value of independent choice, of deciding things for oneself, it would be rational to choose principles that prohibit paternalism (perhaps excepting circumstances in which one is incapacitated). This avenue, however, is not open to Rawls as it assumes a comprehensive moral doctrine. In conclusion, restricting the value of independent choice to pervasive values and relationships, to a person’s evaluative stance, or, to use Rawls’s locution, to rational life plans,⁵⁴ opens the door to rampant paternalism.

Let me reformulate the issue and take stock of the argument so far. The general picture of the relation between right choice and independent choice that emerges is the following. The basic assumption is that if an option is preferable for a person over some other option, then, presumably, choosing this option is better than both unchosen options. Let us call this the value of choice assumption:

[value of choice] if x>y, then x*>x and x*>y (where x and y are options available to s, x>y signifies that doing x is preferable to doing y for s, and * indicates the voluntary choice of that option by s).

Notice that the value of choice is not generally true as it has just been formulated. When an option is immoral, for example, choosing it might make things worse (x>y and y>x*). That is, if x is morally repugnant, it might be better to x unthinkingly or under duress than to choose to x voluntarily.

Setting aside such cases, we can distinguish between two types of cases to which the value of choice might apply:

Type-A y<x, but, at the same time, y*>x.

These are cases in which the voluntary choice of an otherwise inferior option increases its post-choice value, so to speak, above that of an otherwise superior option. This gives the following ordering of options: y<x<y*<x*.

The second type of cases are ones in which making the right decision (x) matters more than making the less optimal decision for oneself (e.g., disposal of hazardous materials).

Type-B y<x and y*<x

The preference ordering in such cases is: y<y*<x<x*.⁵⁵

To resist wholesale outsourcing of self-government, adherents of the rationalist assumption grant special status to particular domains with respect to which independent choice is valuable (one’s evaluative stance, in Wall’s terminology, one’s pervasive goals and relationships in Raz’s, or one’s rational life plans in Rawls’s). Such life-shaping decisions are regarded as Type-A decisions. All other matters, however, are classified as Type-B. In both kinds of cases paternalist interferences are addressed: In Type-A cases we should choose for ourselves as this is what reason dictates (the better choice is always the one we choose for ourselves, whether it is x or y). In Type-B cases the symbolic harms of coercion and manipulation will usually decide against interference.

This response encounters three challenges. First, its resistance to paternalism for Type-B cases is not satisfactory because paternalist interferences do not necessarily have the offensive symbolic significance it identifies. And second, even when they do, the symbolic aspect cannot carry the liberal weight the theory puts on it, namely the symbolic or expressive offense does not necessarily entail that paternalism ought to be ruled out sufficiently broadly. Finally, this view commits the wrong-kind-of-reasons fallacy. The problem with paternalism is not (or not just) that it gets in the way of choosing our life plans or formulating our evaluative stance, but that it violates our independence. This last point suggests that independence, or deciding for ourselves, may be valuable more broadly than rationalists allow. In other words, our intuitive objection to paternalism even in mundane and imprudent choices suggests that independence matters in such contexts as well. This takes us to the central task, namely, tackling the rationalist view of independence and its value directly, not by illuminating its implications, but by analyzing its assumptions.

IV. The Value of Independence

We have encountered three types of considerations which might underlie the value of deciding for ourselves. One is to improve our ability to make decisions and develop our self-reliance. Mill famously touted the instrumental benefits of exercising individual choice: “The human faculties of perception, judgment, discriminative feeling, mental activity, and even moral preference, are exercised only in making a choice.”⁵⁶

This instrumental defense of independent choice does not go very far. For one thing, it’s not clear that the cultivation of self-reliance and decision-making capabilities requires the effective exercise of these faculties, rather than employing them hypothetically, as an intellectual exercise. Furthermore, developing these abilities presumably requires some experience in decision-making, not that every decision be independent and not that important decisions be independent. Thus, it will not rule out outsourcing self-government and exercising decision-making within the bounds determined by a competent authority (this is precisely what we often do with children).

The other two virtues of independence are non-instrumental. One is retail, applying to particular choices. Benefiting from the value of some activities and relationships partly relies on their being the object of people’s own choice, not just fitting their preferences or optimizing outcomes for them. Thus, as Raz, Wall and most others (pace Valdman) concede, even if one’s parent or friend is a better judge of one’s character and better suited to choose one’s spouse or occupation, we are inclined to think that one’s own choice should decide the matter. What explains this? One idea, mentioned earlier, is that there are certain desirable features of relationships based on one’s own choice. Part of what constitutes the value of (some) personal relationships is the fact that they are voluntarily and independently chosen and sustained by the parties.

This account, however, which seems to be endorsed by rationalists, militates against their broader view. The rationale behind it extends not just to the relationship as a whole but to interactions within it as well. After all, what is a relationship other than a series of activities, emotions and interactions?

Debra Satz noted that,

many of our actions have a special meaning for us precisely because we chose them. Think about buying a birthday gift for a devoted friend. Even if I could hire someone to make the choice and purchase for me, I may want to do it myself as a way of expressing and communicating my own feelings.⁵⁷

What separates gifts from commercial exchange is the priority of the expressive aspect over the instrumental.⁵⁸ In market exchange we are typically interested in the uses to which we can put the goods exchanged (or money paid for them). Gifts can be useful, but their primary significance—for both giver and receiver—lies in what they express. The sincere expression of emotions and attitudes, and perhaps the giver’s values and commitments, precludes performing these actions under duress or by deference to the directive of another. This is the basis of Hobbes’s claim about the relevant kind of freedom: “when we say a gift is free, there is not meant any liberty of the gift, but of the giver, that was not bound by any law, or covenant to give it.”⁵⁹ An object awarded under legal obligation, is not a genuine gift. The same goes for helping your friend or listening to her troubles. You can pay someone else to do these things, possibly better than you, but it is not the same as doing them yourself.⁶⁰

This concern seems quite pervasive with respect to relationships which are not purely instrumental, where we have an interest in the relationship itself, not merely in the benefits it provides. In such contexts, independence is often critical not only in choosing the relationship but also in everyday, mundane interactions within it. The distinction between choice of a relationship and choices within it is artificial and arbitrary. If your decision to propose to your spouse should be your own, then so should buying her a gift or preparing her favorite dish. If it is important that we choose our partners and our friends ourselves, it is also important that we conduct our relationships with them by our own lights—which is to say, by our own choices—and not by adherence to the decisions other impose on us, even if they reflect our “deep values and commitments.” (Imagine that your spouse’s relations to you—buying you gifts, spending time with you, sleeping with you—is conducted under the direction of a Valdmanesque “decider”.) This provides a broader basis for independence than Mill’s instrumental justification or the rationalist notion of choosing pervasive relationships.

Rationalists might concede the point yet argue that its scope is rather limited as it presumably applies primarily to personal relations. Arguably, it is not so limited since it also applies to many other meaningful contacts and interactions. But I shall not dwell on this as there is another, more expansive consideration.

The second kind of non-instrumental reason for valuing independence we mentioned is wholesale, concerning individuals’ overall decision making. The idea, as Raz puts it, is that “We are not fully ourselves if too many of our decisions are not taken by us.”⁶¹ Here too lurks a more expansive idea. Being fully ourselves is presumably a gradable notion.⁶² If it is important that we make our own decisions to “be fully ourselves,” then, arguably, the more decisions we make the more fully we are ourselves. And, if it is important to be fully ourselves, then the more fully we are ourselves the better, so the more decisions we make for ourselves the better.

What we observed with respect to meaningful relationships applies to other meaningful endeavors. Here too a distinction between long-term, once-and-for-all decisions, plans and commitments, and the choices that go into their execution is artificial and arbitrary. Consider one’s occupation. It is widely agreed that the choice of occupation should be among the choices people make for themselves. But arguably, the formation of one’s professional life consists not just in choosing in which department to enroll or which position to take, but also in the countless acts performed over the course of one’s studies and in the execution of one’s job. ⁶³ Just as the quality of a relationship is not fixed by the decision to enter into it, but in its actual conduct, so the meaning and worth of one’s professional life consists in the slew of activities, both chosen and unchosen, that comprise its actual unfolding. More broadly, forming an evaluative stance, to use Wall’s terminology, is not like signing a contract or pledging allegiance. Rather, it is an ongoing form of conduct manifested in, and constituted by, many decisions, some momentous and some trivial and mundane, and, presumably, by behavior not anchored in decisions at all. It is not just choosing some goals, adopting certain values, or, to use Rawls’s terminology, drafting a life plan that matter, but also the countless unremarkable choices and acts that go into their execution.

The shape of a life is a cumulative outcome, not the unfolding of a plan or the acting out of a script. It is carved by incremental modifications produced by myriad choices, some big, some small and many of medium magnitude. Some directly pertain to major plans, core value and meaningful relationships, and many relate to them only indirectly and inadvertently, or not at all. Thus, if the aim is “to make one’s own life,” to use Raz’s terminology, then not only momentous decisions, but also many mundane, everyday decisions should be independent. If it is the general course of one’s life that matters, “controlling” or “shaping” one’s own destiny, then ordinary, short-term, trivial choices seem no less, and sometimes more, influential over the trajectory of one’s life than reflectively endorsed, pervasive goals and core commitments. If it is one’s character or personality that matters, then, again, there is no reason to give the latter a privileged status. Bernard Williams once mocked what he described as “an external view of one’s own life, as something like a given rectangle that has to be optimally filled in.”⁶⁴ Living a life is not just sketching the rectangle, but also filling it in. In fact, the rectangle isn’t sketched, it emerges in the end, when the totality of one’s actions add up to the shape of one’s life.

To be sure, this does not entail that independence is always important or that it is equally important in all contexts and domains. But it does show that the value of independence cannot be segregated to specific practical domains or types of choices. Independence has domain-neutral, pro tanto (though not decisive) value.⁶⁵

V. Actors not Authors

So far, I have argued that the distinction between ‘big,’ ‘life-shaping’ momentous decisions—choosing life-plans, subscribing to core values and forging meaningful relations—and mundane acts of living them out is superficial and artificial, at least insofar as the ideal of shaping one’s own life is concerned. Rationalists might at this point concede that independence is more pervasive than they suggest but still invoke their principal line of defense, namely that there is no conflict between authority and independence since “the concerns that underlie reasons with which we should conform and those that underlie the reason to act independently of authority” are the same—conformity with reason, as the rationalist assumption posits.⁶⁶ But the observations underlying our argument challenge this assumption.

“The ruling idea behind the ideal of personal autonomy,” Raz says, “is that people should make their own lives.”⁶⁷ But if the significance of endorsing values and life-plans is supposed to derive from the value of shaping one’s life, why single out rational, reflective choices of values and life-plans? What shapes our lives are not only choices related to life-projects, core values, and basic commitments and not only rational decisions.⁶⁸ Many of our most fateful choices can hardly be described as rational decisions at all. Some are small, seemingly trivial choices, of the kind that Sydney Morgenbesser and Edna Ullmann-Margalit called “picking.”⁶⁹ Even when we don’t pick arbitrarily but choose for reasons, the outcomes are determined by elements outside our control and beyond our foresight. Your decision to study philosophy will alter your life in ways you can neither anticipate nor assess in advance. Whether you chose it rationally and reflectively or not, it is hard to regard the outcome—the ultimate shape of your professional life—as an object of your choice. You can choose your spouse, but you can hardly predict how your life together will unfold. The latter is an outcome of your choice, not its object.

In other words, being the intentional planner of your life is not the only thing that matters, but also being the one who brought it about, the one who actually made the choices that resulted in it having the shape that it has. As we saw above, even rationalists agree that it is not just the congruence with one’s values or plans that matters, but the causal efficacy of one’s agency—being the one who gave shape to one’s life. But then this is also the case when picking without reasons.

As Ullmann-Margalit and Morgenbesser showed, there’s a similar phenomenon at the opposite end of the spectrum. Some decisions are so big that reasons ultimately run out. Deciding to immigrate or to have a child may involve much reflection, but such things are rarely decided by reasons. Sometimes they are inescapable reactions to a powerful experience. You travel to Damascus to arrest followers of Jesus and suddenly something happens, a transformative experience to which you cannot but surrender, and you are baptized. A prisoner is executed and his mate sees God hanging from the gallows, irredeemably losing his faith.⁷⁰ Conversions of this kind can totally upend a life, but can hardly be described as responsive to reasons, at least not in the standard reflective sense. If they can be described as choice at all, they do not involve decision.

Even when they do not occur by way of conversion, big, life-shaping choices are not always a consequence of reasoned decision. Following Ullmann-Margalit, it can be argued that big decisions alter one’s life-plans, evaluative stance, core values and commitments, and therefore cannot be rationalized by them.⁷¹ But even those who dispute this normative claim, will presumably not dispute the descriptive claim that often such choices are not decided by weighing reasons.

Ferdinand, the protagonist of Céline’s Voyage au bout de la nuit, enlists and goes to war, irreversibly altering the course of his life. But as Céline describes it, his choice is barely a decision, certainly not a reasoned decision. While he is busy fervidly proselytizing to his friend, Arthur, that war is a sham, a cynical ploy by the rich and powerful at the expense of the impoverished and gullible, a regiment marches by on its way to the front.

But just then, who should come marching past the cafe where we’re sitting but a regiment with the colonel up front on his horse, looking nice and friendly, a fine figure of a man! Enthusiasm lifted me to my feet.

“I’ll just go see if that’s the way it is!” I sing out to Arthur, and off I go to enlist, on the double.⁷²

This is not a case of conversion, but neither is it reasoned judgment. It is a spontaneous surge of enthusiasm, which will define the course of Ferdinand’s life. It is hardly a reasoned choice, yet it is paradigmatically his choice.

It is worth recalling how familiar and ordinary such experiences are, in life and, consequently, in literature. Consider Emma Bovary, to take an example from a different context. When Emma agreed to go riding with Rudolphe, who seduces and then abandons her, she did not choose to upend her life. And yet, she was charting her own course, from its ecstatic heights to its miserable end.⁷³ Such examples abound.

Making one’s own life concerns which actions are attributable to a person; being a rational chooser turns on reacting properly to reasons. These are distinct issues which should not be conflated. If you like, it is the difference between an author and an owner, laying out the script of one’s life and living it. Ferdinand did not plan his misadventure and Emma Bovary did not author the script of her affair and ultimate demise. These were the consequences not the objects of their choices. Even when decisions are reflective and reasoned they do not make one an author as one does not choose their implications and outcomes. One chooses which path to take, not where it will lead and what one will encounter along the way. A life is lived, it is not authored.

Our rational ambitions should not obscure the simple reality that, as even a rationalist like Aristotle stressed, “non-rational passions, and therefore actions expressing them, are no less human than reason.”⁷⁴ Unreasoned reactions, spur-of-the-moment decisions, and other behaviors that are not rationally chosen are no less determinative of how one’s life will go than reasoned choices and they are no less one’s own. If the goal is to be the maker of one’s own life, then, presumably, independence matters not only with respect to rational, reasoned decisions, but also when choice is spontaneous, unreflective, whimsical, passionate, capricious, or impulsive. In other words, the Rationalist Assumption is false—we value independence (not only “the exercise of judgment”) not just because it secures conformity with reason. Sometimes it is valuable when there are no reasons or even when it defies reason. Independence, or acting by one’s own lights, better captures the idea of making one’s own life, endorsed by rational liberals,⁷⁵ than autonomy, understood in terms of responsiveness to reasons.

Taken one way, the Service Conception of authority is well suited to accommodate the value of independent choice. When it matters, choosing for oneself is yet another reason and hence another reason that should be reflected by a legitimate authority. If it is important that you choose your occupation for yourself over and above choosing the occupation you are best suited for or the one that will deliver the most desirable outcomes for you, then a purported authority must leave this choice up to you and if it tries to order you, it thereby violates the NJT and you are under no obligation to obey its orders, as it constitutes no valid authority for you.⁷⁶

Yet, this view of independence as one of the reasons the directives of legitimate authorities ought to reflect sits uncomfortably within the Service Conception as it is articulated by Raz. There are two sources of tension. One is a technical matter: If independence is among the reasons a purported authority has to reflect in order to meet the NJT, then the Independence Condition is redundant. Independence on this reading is not a separate condition supplementing the NJT. Rather it is an element of the NJT. In other words, independence is not an additional condition which a purported authority meeting the NJT also has to satisfy, but one of the conditions such an authority must satisfy if it is to meet the NJT.

Rationalists can concede this technical point and admit that Independence is in fact not a separate condition on a par with the NJT but rather a particularly interesting element of the NJT, singled out because it merits special attention. Thus understood, the Service Conception provides a useful framework within which the question of legitimate authority can be worked out. It tells us “what has to be the case if some people have authority over others,” namely “only if there are sufficient reasons for the latter to be subject to duties at the say-so of the former.”⁷⁷ It doesn’t tell us if and when sufficient reasons obtain. But the ambitions of the theory apparently exceed this. It aims to provide not just a solution to this theoretical question, but also the means to address what Raz calls the moral question: “how can it ever be that one has a duty to subject one’s will and judgment to those of another?”⁷⁸ Now if Independence is not a separate condition, then the Service Conception’s answer to this question is rather modest: to be legitimate, authorities ought to serve their would-be subjects by improving their conformity with the reasons they have independent of the authorities’ directives. But why should the authority’s directives not be counted among those reasons? If they are reasons which apply to the subject and the NJT entails that a legitimate authority’s directives ought to improve one’s conformity with all the reasons that apply to him or her, then why should these reasons not count? The answer, apparently, is because there are no such reasons—the mere say so of another can never give rise to duties, because such duties are not “consistent with one’s standing as a person.”⁷⁹ This is no doubt an important contribution, capturing the idea that there are no natural authorities, that no individual or institution has authority over others merely by virtue of who or what they are. But as an answer to the moral question it is incomplete as it does not indicate which kinds of considerations might count against authority and which do not. It leaves room for an unlimited number of reasons which—like independence—might trump authority quite extensively.

This suggests another reading of the NJT, whereby the NJT does not cover all of the kinds of reasons that might apply to the would-be subject. Rather, the reasons with which an authority ought to facilitate conformity are prudential reasons, or reasons having to do with attaining the optimal outcomes given one’s practical ends. Thus understood, the question why the reasons issued by the authority’s directives should not be taken into account receives a substantive answer: a legitimate authority is one which helps its subject better adhere to the reasons that apply to her in light of the purposes she has set for herself. But now, to exclude those cases in which obedience improves the achievement of one’s aims yet deciding for oneself is more important, a second condition is called for, namely the Independence Condition.⁸⁰

According to the first interpretation, call it the “unrestricted interpretation” of the NJT, the class of reasons to which obedience to authority must improve one’s conformity (call it set R) includes independence (r_i) (that is: r_i∈R). According to the restricted interpretation, independence is a separate condition, not included in the set of reasons conformity with which authority ought to improve (r_i∉R). On the unrestricted reading, then, the NJT says that an authority is legitimate iff:

[UNRESTRICTED] the subject would better conform to the full set of reasons, R, if he intends to be guided by the authority’s directives than if he does not.⁸¹

According to the restricted interpretation it says that an authority is legitimate iff:

[RESTRICTED] the subject would better conform to the restricted set of reasons, R- r_i, if he intends to be guided by the authority’s directives than if he does not.

The restricted interpretation seems better equipped to address the moral question. The restrictions determining which reasons are members of R will determine which considerations influence the legitimacy of an alleged authority and which do not. The justification for these restrictions will provide the account for the basis of legitimate authority. On this interpretation independence can be a separate concern, not one of the reasons an authority ought to reflect. This fits Raz’s reference to “two concerns, one satisfied by conformity with reasons, the other by acting on one’s own judgment.”⁸² On this interpretation we have practical aims—the optimal rational outcomes of our actions given our goals, or some such—and also other aims, among them independence. But then, the possibility that these ends might pull in different directions is perfectly coherent. In fact, it is to be expected. Authority is about conforming to reason; independence is about being the maker or the owner of one’s life. Independence excludes deference and submission; authority requires it.⁸³

Amending the theory to fit the unrestricted NJT seems like the better option but renders the instrumental account merely a framework for posing the question of authority, not sufficient for solving it, since it depends on the breadth and strength of the value of independence. Sometimes it is more important to decide for yourself—independence is then a higher-order reason to act on one’s judgement irrespective of the balance of (all other) reasons. Raz is correct when he rejects the claim that “the concerns that underlie reasons with which we should conform and those that underlie the reason to act independently of authority” have nothing to do with each other. Both are reasons pertaining to our actions. But they remain in conflict even if they are not “incommensurate.” It is inaccurate to say that the “value of our general ability to act by our own judgment derives from the concern to conform to reasons” (the Rationalist Assumption). Rather, acting on our own judgement is valuable as such (because it is valuable to be the maker of one’s own life etc.), and by virtue of this it gives us reasons to decide for ourselves. I’ve argued that this value cannot be segregated to specific domains and types of choice and that it is not reducible to right reasons. Reasons of independence are pervasive and can conflict with other, prudential reasons very broadly. Emma Bovary could have resigned herself to the parish priest’s or to her chaste mother-in-law’s authority. Suppose that obeying their direction would have led to outcomes more compatible with her core values and commitments, not only objectively, but also by her own all-things-considered retrospective judgment. Should she have subjected herself to such authorities? Perhaps you think that in the interest of marital fidelity she should have. But, setting aside moral obligations, can it be denied that her independence would have been compromised by such submission? Such a life would be, in one important sense, less her own.

Deference to authority, then, jeopardizes something of value even when justified. The relation between authority and independence, in other words, is not one of happy collaboration, but a tense one of mutual exclusion. The easy reconciliation of authority with autonomy—“One no more abandons reason or forfeits one’s autonomy if one follows the commands of authority than if one follows trends on the stock exchange”—does not apply to independence.⁸⁴ The reason is best expressed in what Raz says shortly afterwards: “accepting authority inevitably involves giving up one’s right to act on one’s judgment on the balance of reasons.”⁸⁵

VI. Conclusions

I have argued that, properly understood, grounding the value of independent choice in the idea of making one’s own life entails that this value is not limited to one type of choices (rational or reasoned) and cannot be segregated to discrete domains of choice (core values, commitments, or life plans). Making one’s own life concerns not which reasons guide us but what is attributable to us. One can reject the idea that actually making the choices is what matters and join Valdman in claiming that it doesn’t matter who makes the choice so long as it coheres with one’s values. Some of us, including Wall and apparently Raz and Rawls, find this interpretation of the idea of living life by one’s own lights misguided. It has the implausible implication that there’s no loss of (valuable) freedom in wholesale subjection to a wise decider. For those of us who reject this idea and grant that making independent choices matters and not just conformity with one’s values, commitments, or plans, the value of independence, properly understood, is pervasive. Even trivial choices, like the choice of ice-cream, or when to wash and when to comb our hair should be up to us and even when they are not reasoned. This is why coercion in such matters is pro tanto wrong—even if it doesn’t express contempt or deplete one’s range of valuable options and even if it guides us to better conform to reason. This may seem unintuitive in its own way (does it really matter that you choose your ice-cream for yourself?). To the extent that it does, I think this should be attributed to the fact that the choice itself in this case is not of great significance. In other words, it is not that independence is not important in such instances, but that the choice itself is not very significant. One indication of this is the familiar fact that even though the choice isn’t very important to us, we resent anyone imposing it on us. We can allow that the value of independent choice can vary with the gravity of the choice without denying that it is valuable independently of its contribution to the realization of one’s other values, commitments, or plans, or facilitates one’s conformity with reason. Not only paternalist coercion, but also prudential submission to authority comes at a price, paid in the currency of independence.

To be sure, this does not mean that authority is never legitimate, that deference to the judgment of others is never the right thing to do. But it does mean that the value of independence cannot be confined to special spheres of activity and is not subordinate to other ends. If independence is intrinsically valuable, then the conflict between authority and independence is systemic, even if not ubiquitous.

Acknowledgements

For helpful discussion and comments, I am grateful to participants in SciencePo’s Théory Politique Seminar, the Big Ideas Workshop at the University of Bayreuth, Tel Aviv University’s Law Faculty seminar and The Centre for Moral and Political Philosophy at The Hebrew University of Jerusalem, where earlier versions of this paper had been presented.

Competing Interests

The author has no competing interests to declare.