APA Newsletters

Spring 2001
Volume 00, Number 2

Newsletter on Philosophy and Law

Articles

Previous Article | Index | Next Article

The Paradoxes of Authority

Scott J. Shapiro

In his In Defense of Anarchism, Robert Paul Wolff argued that legitimate authority and moral autonomy are logically incompatible.2 In this essay, I would like to examine his argument. I will claim that his attack on authority is much more powerful and subtle than is typically acknowledged. I will then contrast his argument with another well-known anarchist critique of authority, one that aims to show that authority and rationality are incompatible. I will conclude by demonstrating that solutions to one paradox are not necessarily solutions to the other.

A. Authority and Autonomy

Wolff begins his discussion by distinguishing between power and authority.3 To have power is to have the ability to compel others to do as one wants. To have authority is to have the right to rule. A gunman has power, but he does not have authority. He can coerce his victim to cooperate by threat of force, but he is unable to impose the obligation to comply. Unlike authority, power cannot be honored in the breach: one can cheat at one's taxes, but one cannot cheat a thief.

As Wolff points out, someone can have authority in one of two senses.4 One can have authority by possessing the moral right to rule. The exercise of such a right, if it exists, genuinely gives rise to moral obligations to obey. A ruler can, therefore, claim authority and fail to have it in this sense. The Supreme Soviet Legislature claimed the authority to rule the Soviet Union, but it lacked the moral right to do so. It lacked legitimate, or de jure, authority. The Supreme Soviet did enjoy a measure of acceptance by many of the Soviet people. They believed that the Supreme Soviet possessed de jure authority, although they were mistaken-it merely had de facto authority.

Wolff is primarily interested in the phenomenon of de jure authority. This is so for two reasons. First, the concept of de facto authority logically presupposes the concept of de jure authority. One who possesses de facto authority is someone whose claims to de jure authority are believed by a significant portion of its subjects. Second, it is uncontested that de facto authority exists. The philosophical anarchist is interested in whether de jure authority exists, because she wants to show that the moral obligation to obey the law can never obtain.

To have the right to rule, according to Wolff, is to have the right to be obeyed. To obey an authoritative command is to perform the act commanded for the reason that it was commanded. Commands, therefore, differ from arguments.5 An argument is meant to persuade. It attempts to convince the person that they ought to act in a certain way and it does this by presenting to the interlocutor the reasons that make the action worthy. One who issues a command, on the other hand, does not intend to convince the subject of the wisdom of his order. The commander does not give reasons why the action commanded is worthy of obedience, but rather demands that his command be taken as a conclusive reason for obedience.

It is possible, therefore, for someone to conform to a command without obeying it.6 This happens when the command makes the subject aware that he has reasons for performing the act commanded and he acts for these reasons, rather than because of the command. To acknowledge someone's claim to authority, according to Wolff, is to recognize that their right resides in their person.7 They possess their power in virtue of who they are, rather than in virtue of what they command.

Having set out his conception of authority, Wolff proceeds to present his account of moral autonomy. For Wolff, an autonomous person is not someone who is merely responsible for her actions. Rather, such a person also takes responsibility for her actions.8 A person takes responsibility whenever she attempts to determine what she ought to do. An autonomous agent, according to Wolff, is a deliberating agent.

Unlike many who have seen autonomy as a necessary condition for moral responsibility, or as the capacity to choose, Wolff treats it as an independent moral duty. Every person is charged with examining every aspect of his moral life: he must constantly gather new information, scrutinize his motives, critique his desires and evaluate his options in light of this reflection. One who acts without assessing the merits of so acting fails to take responsibility for his actions, and to this extent, is violating their duty to act autonomously.9

It follows from Wolff's definition of moral autonomy that no one can obey authority and remain autonomous. A person obeys a command when he conforms for the reason that another has so commanded. An autonomous person, however, never acts for the reason that another has so commanded. He acts only when he is convinced that, on the merits, action is appropriate. Hence, an autonomous agent can never submit to another's authority. As Wolff puts it: "The defining mark of the state is authority, the right to be ruled. The primary obligation of man is autonomy, the refusal to be ruled."10

Although the autonomous agent cannot obey authority, Wolff is quick to add that he does not necessarily disobey authority.11 If the autonomous agent thinks there are good moral reasons to pay taxes, then he will believe that he should pay his taxes. But that person does not accept the obligation because the law requires him to pay his taxes. He believes that he should pay his taxes because he believes this to be the right thing to do independent of the law's demands.

B. Preliminary Assessment

Wolff's argument appears to be valid: given his premises, his conclusion seems to follow. However, his premises are dubious. For example, it is not obvious that one should think of autonomy as a duty. To whom do we owe this duty? It is strange to think that I am morally bound to act autonomously for another's sake. Why would anyone care why I act correctly, as long as I act correctly?

Even if one does accept that there is a duty to act autonomously, it is doubtful that Wolff's formulation should be accepted. Why must a person deliberate about every moral action? Shouldn't he defer to another's judgment when that judgment is better than his? The idea that a person must weigh the balance of reasons every time a moral decision arises is not only dangerous in cases of informational asymmetries or cognitive disabilities but is also terribly wasteful. Surely one could focus his energies in a more productive use than constant deliberation.

To be sure, there is such a thing as over-reliance on authority. To cede too much decision-making to others is both foolhardy and morally irresponsible. Moreover, the more one depends on another's judgment, the greater the chance that one will lose the ability to make judgments for oneself and the more vulnerable one becomes to manipulation. Finally, the faculties of judgment and self-reflection are distinctively human capabilities, the exercises of which contribute in an essential way to human flourishing. To sacrifice them is, in some real sense, to forfeit one's humanity. This "dehumanizing" effect of authority especially concerned William Godwin, the first "modern" philosophical anarchist.

Man is the ornament of the universe, only in proportion as he consults his judgment. … But, where I make the voluntary surrender of my understanding, and commit my conscience to another man's keeping, the consequence is clear. I then become the most mischievous and pernicious of animals. I annihilate my individuality as a man, and dispose my force as an animal to him among my neighbors, who shall happen to excel in imposture and artifice, and to be least under restraint from the scruples of integrity and justice.12

While the dangers of reliance on authority are real, it is important not to exaggerate them. The world is simply too complex for anyone to live one's life completely unaided by experts of one kind or another. Even Wolff admits that "[t]here are great, perhaps insurmountable, obstacles to the achievement of a complete and rational autonomy in the modern world."13 Complete autonomy, in Wolff's sense, is simply not an option. If authority is inconsistent with autonomy, then so much the worse for autonomy.

It is interesting to note that Kant himself did not see a clash between authority and autonomy. He famously argued that those subject to authority ought to question its demands, but this public use of reason should not prevent them from acting on them. Enlightenment is precluded both when authority demands blind obedience and when subjects do not respond with unconditional compliance. "The citizen cannot refuse to pay the taxes imposed upon him; presumptuous criticism of such taxes, where someone is called upon to pay them, may be punished as an outrage which could lead to general insubordination. Nonetheless, the same citizen does not contravene his civil obligations if, as a learned individual, he publicly voices his thoughts on the impropriety or even injustice of such fiscal measures."14

Wolff's formulation of the anarchist's challenge is unconvincing because his understanding of autonomy is implausible. We should, however, be careful not to dismiss Wolff's argument too quickly, for on any credible conception of autonomy, the tension between it and authority is hard to ignore. After all, "autonomy" literally means "self law-giving." The autonomous person does not act simply because another has told him to do so-he acts only when convinced that action is appropriate. To be autonomous, in other words, involves taking oneself as the ultimate authority on moral questions. This commitment seems to leave no logical space for external authorities to occupy. As the proverb goes, one cannot serve two masters.

With this in mind, I think it is possible to give a more charitable reading to Wolff's objections. We should first distinguish, in a way that Wolff fails to do, between two different features of authoritative directives. We can say, following H.L.A. Hart, that authoritative directives are both "peremptory" and "content-independent" reasons for action.15 A "peremptory" reason is a reason that preempts deliberation. A command is a peremptory reason, in that once the command has been issued, the time for deliberation comes to an end and the time for action begins. The obeying agent no longer considers the merits of following the command and simply acts as he is directed.

A directive is a "content-independent" reason when it gives an agent a reason to comply irrespective of whether the agent has a reason to act on its content. The fact that such a directive requires that an act be done is, by itself, a reason to perform that act. One who obeys a command, therefore, treats the command as a content-independent reason, because he complies for the reason that he was commanded, not because he has reasons to act on the content of the command. For example, if Jim takes out the garbage because his father commanded him to do so, then he is treating the command as a content-independent reason.

Content-independent reasons for complying with a directive should be contrasted with "content-dependent" reasons. A content-dependent reason is a reason for conforming to a directive because the directive has a certain content. If the garbage smells, Jim will have a reason for taking out the garbage that is independent of the fact that his father commanded him to do so. By taking out the garbage, he will have removed an unpleasant odor from the house. Jim, therefore, has two reasons to listen to his father's command: the command is a content-independent reason, while the unpleasant odor is a content-dependent reason.

Although Wolff appears to object solely to the peremptory nature of authority, I think that it is the combination of peremptoriness and content-independence that offends him. Authority and autonomy clash not simply because one who obeys does not deliberate. The problem is also that such a person believes that the fact that he was ordered to act in a certain way gives him a reason to so act. He takes the will of another as his reason, indeed the only reason, rather than the merits of the case at hand.16 Such a person, therefore, will think that he has a ready defense to any charge of improper behavior. While the person will agree that he performed a wrongful act, he will plead that the reasonableness of his actions must be viewed in a content-independent manner: whether he had reason to follow orders cannot be judged based on the content of those orders. It is the fact that he was commanded to act, rather than what he was commanded to do, which gave him a conclusive reason to do as he did.

An autonomous person, by contrast, never treats a command as a content-independent and peremptory [hereinafter "CIP"] reason for action. The demands of authority mean nothing to the autonomous agent, for such a person never allows his will to be determined by the will of another. She cares solely about the act commanded, not the command itself, and will acquiesce only when convinced that there are good reasons to act on the content of the command. According to this interpretation, autonomy and authority are incompatible because obedience to authority requires acting on CIP reasons, whereas the autonomous person does not acknowledge the existence of such reasons.

One benefit of seeing Wolff's argument in this way is that our previous objections are no longer sufficient to meet his challenge. Autonomy is not conceived as a separate duty that morality imposes upon us and that we owe to others. To say that everyone should act in a morally autonomous manner is to make a claim about the space of reasons. Autonomous agent are those who recognize that the only reasons that exist are either content-dependent or non-peremptory ones. Moral autonomy is important because it is important that people act on reasons and not act on non-reasons.

Moreover, on this account, reliance on experts does not necessarily lead to heteronomy. While expert advice is a CIP reason for believing that the expert is correct - one believes what the expert says simply because the expert has said it - the purpose of giving advice is to alert the advisee that the recommended course is supported by the balance of content-dependent reasons. When the advisee draws this inference and acts on it, the agent will be acting for content-dependent reasons, even if he does not know what they are.

Most importantly, this interpretation shows that the philosophical anarchist's anxiety about authority is not frivolous: their worry is that people will treat authoritative directives as CIP reasons for action and, in so doing, fail to take the appropriate responsibility for their actions. They will attempt to justify their conduct by pleading that they were "just following orders." This type of defense not only seems cowardly, but strictly speaking irrelevant. How can an act be made acceptable simply because someone else says that it is acceptable? Authorities may have the power to change positive law, but no one (maybe not even God) has the ability to change the moral law. As Godwin put the point: "The most crowded forum, or the most venerable senate, cannot make one proposition a rule of justice that was not substantially so, previously to their decision."17

Yet, the philosophical anarchist reminds us, without admitting that Nuremberg defenses are sometimes good justifications, it is hard to see how de jure authority is possible. The legitimacy of authority stands or falls on whether a subject can justify his actions by pleading that he was "just following orders." Legitimate authority is possible, in other words, only when CIP reasons are possible.

C. Authority and Rationality

It is sometimes thought that Wolff's challenge to authority is merely a special case of a more general paradox, one that purports to show the incompatibility of authority and rationality. The general argument is familiar: Consider any directive issued by an authority and any action required by that directive. Either the balance of reasons supports that action or it does not. If the balance of reasons supports the action, an agent should conform to the directive, but not because conformity is required by the directive, rather because agents should always act according to the balance of reasons. On the other hand, if the balance of reasons does not support the action, then an agent should not conform to the directive because agents should never act against the balance of reasons. It would seem, therefore, that authoritative directives can never be reasons for action-if a directive gave the right result, the directive would be irrelevant; if the directive gave the wrong result, then the obedience to the directive would be unreasonable.

Since authoritative directives can never be reasons for action, it follows that rational agents can never obey authority. The proof: Rational agents always aim to act on undefeated reasons and act in accordance with that aim. If an agent were to obey an authority, they would either have to believe that they had an undefeated reason to obey or believe that they didn't have an undefeated reason but would have obeyed anyway. If the former were true, then the agent would have irrational beliefs, given that according to the first argument, authoritative directives can never be reasons for action. If the latter were true, then the agent would not be acting in accordance with the aim of acting on undefeated reasons. Hence, it seems that rational agents can never obey authority.

If the above arguments are sound, it would follow that moral agents can never rationally guide their conduct by authoritative directives. Since morality requires that agents act on the balance of moral reasons, obedience to authority can never be rationally justified for moral agents: whenever a directive required an action supported by the balance of moral reasons, that directive would be morally irrelevant; otherwise, it would be morally pernicious. Authoritative directives can never be moral reasons for action and, hence, it would be irrational for any moral agent to obey authority.

Such "derivative" arguments are possible because rationality is essentially a formal ideal. Rationality does not mandate conformity to any particular standard-it simply requires that an agent live up to the standards that the agent judges he should live up to. The paradox attempts to show the incompatibility of rationality and authority by demonstrating that authoritative directives will clash with any normative standard: either the directive in question conforms to the given standard, in which case it is redundant, or it conflicts with the standard, in which case the standard requires nonconformity. To generate a contradiction between authority and any specific normative standard, one need only plug the standard into the equation and out will pop the desired reductio. While Wolff's challenge appears to be such a derivative argument, it is important to see that it is not. This becomes evident when it is noted that the concept of rationality is at right angles to that of autonomy. To be rational is to aim to act on undefeated reasons and to act in accordance with that aim. To be autonomous, by contrast, is to aim to act on non-CIP reasons and to act in accordance with that aim. It does not follow, therefore, that a rational agent is an autonomous agent. If an agent believes that he has an undefeated CIP reason for action, then he will be acting rationally but not autonomously if he acts for this reason. Conversely, autonomous agents are not necessarily rational. If an agent acts on a content-dependent reason that, by his own lights, is defeated, then he will be acting autonomously but irrationally.

It should come as no surprise, therefore, that one can challenge the rationality of authority and not its effects on autonomy. The paradox of authority and rationality attempts to show the impossibility of having an undefeated reason to obey authority and, hence, the irrationality of believing that one can have such a reason. It does not attack the content-independent and/or peremptory nature of authoritative directives. Likewise, one can object to authority because it engenders heteronomy, not irrationality. The problem with obedience, according to our interpretation of Wolff's challenge, is that authoritative directives are not CIP reasons, not that it would be incoherent to believe that they are undefeated reasons.

Given that these critiques differ from each other, one should not expect that the solution to one challenge will constitute an effective reply to the other. To see this, consider the following response to the paradox of authority and rationality: The social contract theory, the response begins, is a coherent theory of political obligation. A rational agent might regard it as true, even if it is not true. Assume, then, that an agent accepts the social contract theory as true. According to this agent, someone possesses legitimate authority over another when the latter has consented to be ruled by the former. Because the consent generates a promissory obligation to abide by the demands of the authority, any directive issued gives subjects who have consented a reason to act in accordance with it. Assume that this person consents to be governed by an authority. He will now regard any directive issued as a reason for action. Consequently, it might be rational for him to comply with a command whose content, by his own lights, is not supported by the balance of content-dependent reasons. From that agent's perspective, even though the balance of content-dependent reasons would not support conformity, the balance of all reasons-content-dependent reasons as well as the content-independent reason-might tilt in the direction of obedience. In this way, it may be rational to obey authority even when they are wrong about the content of their directives.

This response, however, will not work against Wolff's challenge. Wolff's argument, as we have seen, is predicated on the idea that there are no such things as CIP reasons for action. The above response would, therefore, beg the question. After all, consent itself purports to be a CIP reason for action. Subjects who consent to be governed by an authority are obligated, under the social contract theory, simply in virtue of their consent. One cannot show how a CIP reason is possible by producing another (alleged) CIP reason. One must first establish that my will can give me a reason to act against the balance of reasons. But if authorities lack the power to change the moral law, how can I have the power to do so?

I don't mean to imply that Wolff's autonomy paradox is harder to answer than the rationality paradox. They are simply different critiques and, as such, each may require different solutions. Unfortunately, those who respond to philosophical anarchism do not always make it clear to which paradox they are responding.

Endnotes

1. The following essay is excerpted from a much longer piece entitled "Authority" to be published in The Oxford Handbook of Jurisprudence and the Philosophy of Law (Jules Coleman and Scott Shapiro eds., forthcoming)

2. See Robert Paul Wolff, In Defense of Anarchism (1970). The relevant sections of Wolff's monograph are excerpted as "The Conflict between Authority and Autonomy" in Authority (Joseph Raz ed., 1990). All citations in this essay will be made with reference to this excerpt.

3. Wolff, supra note 2, 20.

4. Id. at 21.

5. Id. at 22.

6. Id.

7. Id. But see discussion in Section B., infra.

8. Id. at 25.

9. Id. at 28-29.

10. Id. at 29.

11. Id.

12. William Godwin, Enquiry concerning Political Justice 122 (K. Carter ed., 1971).

13. Id.

14. Kant, "An answer to the question: What is enlightenment?" in Kant: Political Writings (H. Reiss ed., 1989).

15. H.L.A. Hart, "Commands and Authoritative Reasons" in Essays on Bentham, 253 (1982).

16. See, e.g., Wolff, supra note 2, 26 ("The autonomous man, insofar as he is autonomous, is not subject to the will of another. He may do what another tells him, but not because he had been told to do it." (emphasis in original)).

17. Godwin, supra note 12, 88.

Previous Article | Index | Next Article

APA Newsletters

Spring 2001 Volume 00, Number 2

Spring 2001
Volume 00, Number 2