Wolff declares that the notion of legitimate political authority is self-contradictory or, as he puts it, "inherently incoherent." (602)¹ He recognizes, of course, that large numbers of people accept the claims to authority that states make. But he points out that this confers only a de facto legitimacy upon those states, and he maintains that in fact there can no more be a state that truly enjoys de jure legitimacy than there can be a round square or a married bachelor. Wolff characterizes himself, accordingly, as a philosophical anarchist:

On the basis of a lengthy reflection upon the concept of de jure legitimate authority, I have come to the conclusion that philosophical anarchism is true. That is to say, I believe that there is not, and there could not be a state that has a right to command and whose subjects have a binding obligation to obey. (607)

A contractual democracy is legitimate, to be sure, for it is founded upon the citizen's promise to obey its commands. Indeed, any state is legitimate which is founded upon such a promise. (69)

He does it by confusing two distinct questions. On the one hand, there is the question whether there are conceivable conditions under which an authority enjoys de jure legitimacy. On the other hand, there is the question of whether it is morally justifiable to bring about those. conditions, and thus to establish an authority that has a genuine right to command and whose subjects have a binding obligation to obey. Wolff is led, by considering the relation between political authority and individual autonomy, to give a negative answer to the second question. The inconsistency to which I have called attention arises when he mistakes this for a negative answer to the first.

Wolff offers no argument that tends in any way to support the thesis that the concept of de jure legitimate authority is logically incoherent. His discussion of political authority is actually directed against its desirability, rather than against the possibility of its being legitimate. This suggests that of the two contradictory views he advances, he would in the end choose to adhere to the quite plausible though somewhat banal one expressed in the second passage I have quoted. As Wolff there recognizes, a person clearly does acquire a moral obligation to obey someone's commands by promising explicitly or implicitly to do so. There can be, accordingly, states whose subjects have an obligation to obey.²

So Wolff believes, however inconsistently with his other beliefs, that all it takes to provide de jure legitimacy to a political authority is what it has commonly been thought to take: roughly, the consent of the governed. Given this belief on his part, there is an inaccuracy in his representation of himself as a philosophical anarchist. In the first quoted passage, Wolff identifies philosophical anarchism as an answer to the question of whether there is or could be a legitimate political authority. On this question, however, the view to which he seems most fundamentally committed is an utterly conventional one. If he is an anarchist, it is not in virtue of his position on the issue with which he says anarchism is concerned.

Rather, it is in virtue of his separate claim that all political authority, whether legitimated by consent or not, is morally indefensible. His discussion of politics is not, despite what he himself says about it, devoted to making the point that there is a conceptual error in the supposition that political authority can be genuinely legitimate. Its thrust is that there is a moral error in conferring de jure legitimacy on an authority by binding oneself with a promise to submit to it. Wolffs anarchism, then, must consist in the view that the promise that establishes an authority as truly legitimate ought not to be made. As we shall see, however, he does not consistently hold to this view either.

Wolffs defense of anarchism rests upon his account of autonomy and its relation to authority. This account is, in certain fundamental respects, cripplingly muddled. To begin with, Wolff misconstrues the extent to which the notion of autonomy is pertinent to the analysis of political relationships.

He characterizes the autonomous man as one who decides for himself both what he ought to do and by what principles his decisions are to be guided. And he contrasts autonomous men with men who permit their actions to be determined by decisions that are made by others. Being autonomous, as Wolff understands it, is essentially a matter of acting in accordance with "commands" that one issues to oneself, rather than of acting in submission to the commands of another.

In order to make responsible decisions concerning what he will do, clearly, the autonomous man needs to be well informed. In illustration of this point, Wolff says that "the contemporary American citizen . . . has an obligation to master enough modern science to enable him to follow debates about nuclear policy and come to an independent conclusion." (17) Now it may well be that American citizens have this obligation. But their having it is not derivable, as Wolff supposes, from their obligation to be autonomous. For the alternative to making up one's own mind about nuclear policy is not to accept the commands of another concerning what one will do.

When the government of the United States decides whether to develop a first-strike nuclear capacity, the sequel is not that it proceeds to issue commands to all its citizens. An American citizen, unless his work is specifically relevant to the implementation of nuclear policy, is not confronted with a requirement to comply in some way with a governmental decision of this kind. If our autonomy depends upon how we reach decisions concerning what we will do, the question of autonomy appears not to arise here at all.

To be sure, an autonomous man may have to decide whether to pay the taxes that the government of the United States claims from him, and in order to make this decision he will have to evaluate the uses to which his tax money will be put, including its use in implementing the government's nuclear policy. But this is probably not pertinent to what Wolff has in mind. If it were, he would not have made his point with reference to the obligations of "the contemporary American citizen." Numerous people besides American citizens pay taxes to the United States, some American citizens pay no taxes, and many of those who do pay taxes have no choice in the matter since their taxes are withheld from the incomes they receive.

Wolff also cannot rely on the fact that American citizens have the right to vote for those who determine national policy. A responsible decision concerning how to vote does indeed require the exercise of independent judgment about the issues at stake in the election, of which nuclear policy may be one. But a person can, if he chooses, refrain from voting altogether. Wolff says nothing that would support the notion that failing to vote involves a sacrifice of autonomy or an otherwise avoidable submission to the will of another. On the contrary, he implies that it does not, since he quite emphatically denies that the autonomy of an American citizen is protected by the provisions that entitle him to vote for his representatives in government.³

A responsible person will certainly, if he is to have an opinion concerning nuclear policy, decide for himself what his opinion is to be. But he may prefer not to have an opinion on this subject at all. Moreover, he may have good reason for this preference. There are other important things in life, and it may be entirely reasonable for a person to devote himself to them at the expense of his knowledgeabihty concerning nuclear policy. Let us suppose for the sake of argument, however, that someone's failure to inform himself about nuclear policy would be, in a given case, objectionable. It is not at all obvious that what would make it objectionable would be its inconsistency with the person's obligation to be autonomous. A government's formulation of its nuclear policy is fundamentally a matter of deciding how the common resources of the governed are to be employed. Whatever obligation an individual has to inform himself about such things appears to derive primarily from his obligation to exercise responsible surveillance and control, insofar as he is able, over the uses to which his resources are put. Perhaps Wolff can show that this obligation comes down in some way to the obligation to avoid submission to the commands of another, but it does not readily seem to do so.

The most serious difficulties in Wolffs work are his inconsistencies in the use of the concept of autonomy. He tells us that being autonomous "means making the final decisions about what one should do." (15) We must begin by asking what precisely it is, in his view, to make the "final" decisions about what one should do, and just what sort of thing he understands being autonomous to preclude.

"The autonomous man . . . may do what another tells him," Wolff explains, "but not because he has been told to do it"; a man forfeits his autonomy when he decides "to obey the commands of another without making any attempt to determine for himself whether what is commanded is good or wise." (14) These statements might easily lead us to suppose that Wolff understands autonomy as implying that an autonomous man never decides to comply with a command unless he himself approves the course of action he has been told to follow -- i.e., unless he has reasons for following that course of action that would have led him to decide to follow it even if he had not been commanded to do so. And then it would be pretty clear why Wolff believes that a society of autonomous men would have to be an anarchistic one.

But this is evidently not what Wolff really means by autonomy, as what he says next makes clear:

If someone in my environment is issuing what are intended as commands, and if he or others expect those commands to be obeyed, that fact will be taken account of in my deliberations. I may decide that I ought to do what that person is commanding me to do, and it may even be that his issuing the command is the factor in the situation which makes it desirable forme to do so. For exmaple, if I am on a sinking ship and the captain is giving orders for manning the lifeboats, and if everyone else is obeying the captain because he is the captain, I may decide that under the circumstances I had better do what he says, since the confusion caused by disobeying him would be generally harmful. But insofar as I make such a decision, I am not obeying his command: that is, I am not acknowledging him as having authority over me. I would make the same decision, for exactly the sgme reasons, if one of the passengers had started to issue "orders" and had in the confusion come to be obeyed. (15-16, emphasis in original)

How is it, then, that the man in this example maintains his autonomy? The explanation lies, presumably, in the fact that he does not join in manning the lifeboats only because he has been commanded to do so. He does not, as the other passengers are described as doing, take the mere issuance of the captain's command as sufficient to determine his compliance. When he decides to comply with the command, it is only because he believes that certain additional conditions are satisfied -- conditions that his own judgment leads him to regard as being of critical importance. He reserves to himself the final decision concerning what he ought to do, in that he judges for himself that these additional conditions -- especially that failure to comply would be generally harmful -- are, in the situation at hand, both decisive and fulfilled. Wolff's insistence that the man's compliance differs from obedience is based, in short, upon the fact that the man's compliance is not blind or uncritical. What preserves the man's autonomy, despite his decision to comply with the captain's command, is that he does not consider himself bound unconditionally to do whatever he is told by the captain to do.

This way of understanding Wolffs conception of autonomy is confirmed by another of his examples, which he offers as illustrating a limited forfeiture of autonomy. "When I place myself in the hands of my doctor," Wolff says, "I commit myslef to whatever course of treatment he prescribes." (1 5) The forfeiture of autonomy here is a limited one, since it involves giving up independence of judgment only with regard to matters of health. The commitment in question does amount to a forfeiture of autonomy, however, just because it is tantamount to an unreserved decision to comply with the doctor's recommendations concerning health. If a patient does have this (rather questionable) attitude toward his doctor, then he will accept as a sufficient reason for performing some action the fact that his doctor prescribes the action as medically desirable. The patient forfeits his autonomy because his commitment to do what the doctor says is unconditional.

But given that this is how Wolff understands autonomy and its forfeiture, his strictures against political authority are wholly unconvincing and, indeed, largely beside the point. The acceptance of political authority -- the readiness to do what one is told to do even though one might not otherwise regard the prescribed course of action as the best available -- entails no blindly unreserved commitment to unreflective obedience. The whole point of having a constitutional government -- whether democratic or of some other variety -- is, after all, precisely to render the authority of those who issue political commands conditional. Wolffs argument for anarchism is that the autonomy of the citizen is incompatible with the acceptance of political authority. As Wolff himself understands autonomy, however, it is incompatible only with a kind of absolute acceptance of authority that is by no means required of the citizens of nonanarchistic constitutional states.

It is elementary that the citizen of a constitutional state is supposed to reserve to himself, just as the man in the lifeboat example does, the final decision concerning whether or not to comply with some command. For the decisions and commands of government are construed to be legitimate only if they are arrived at in certain specified ways and if they honor certain specific requirements and prohibitions. The citizen is understood to enjoy the right and to have the responsibility to satisfy himself that the commands he receives fulfill -- with respect both to procedure and to substance -- the conditions upon which the legitimacy of the political authority is acknowledged by the constitution to depend.

When a person consents to the authority of a constitutionally limited government, he does not commit himself to doing whatever he is told to do by the people who make up the government. He is certainly not expected to emulate the obedient passengers in the lifeboat example and to do what some official commands just because the man is an official. He is expected to comply only if the official's commands satisfy various procedural requirements and conform with various principles of limit and exclusion. The citizen does agree, of course, to comply with decisions that he himself may think wrong or unwise. But his agreement to do this is conditional, and it is he himself who is to judge whether the conditions of legitimacy are fulfilled. The sort of critical consent to authority in question here is at least as compatible with the individual's autonomy as is the compliance of the man on Wolff's sinking ship. It is possible, to be sure, to conceive a greater degree of autonomy than is enjoyed by either the man in Wolff's example or the responsible citizen of a constitutional state. But the limited and conditional acceptance of authority that is agreed to in both cases hardly amounts to the wholesale abandonment of human dignity that Wolff alleges to be implicit in any social arrangement other than anarchism.

In order to support his thesis that anarchism alone is morally acceptable, Wolff introduces the exotic notion, which he himself often disclaims, that it is morally unthinkable to tolerate any limitation of autonomy whatever. This makes it easy to reach an anarchist position. At great length, Wolff argues that no representative or majoritarian system of government can resolve the natural conflict between autonomy and authority. For the views of a person's political representatives, or of the majority of voters in the electorate to which he belongs, may not coincide with his own. And his promise to obey the laws enacted under a representative or majoritarian system of government is therefore tantamount to a commitment to do what he is told to do even when he himself regards another course of action as preferable. From this, Wolff concludes that a person who promises to obey the laws of such a government thereby sacrifices his autonomy.

Now it is certainly true that laws made by representatives or majorities are unlikely to coincide with the preferences of all voters. This goes without saying. But how is it supposed to be pertinent to the problem of autonomy? It would be pertinent only if being autonomous required that a man comply with a command only when it commanded what he himself thought it best to do. It is not pertinent at all if autonomy is understood to require merely that a man accept authority conditionally and that he reserve to himself the final judgment concerning whether or not the conditions of his compliance have been met. As we have seen in the preceding section, Wolff sometimes speaks of being autonomous in terms of the second requirement. When he wishes to make anarchism seem plausible, however, he argues on the basis of the first notion of what being autonomous requires. Throughout his discussion, he switches from one of these ways of understanding what autonomy requires to the other, as it suits his purpose to do so.

Wolffs claim that being autonomous and accepting political authority are incompatible is hardly deniable, if being autonomous means being absolutely autonomous. It has no plausibility at all, on the other hand, if it means what it means in his discussion of the lifeboats episode --that is, if a degree of autonomy less than absolute is conceded to be morally tolerable. Similarly, Wolffs idea that it is deeply immoral to sacrifice autonomy is drearily unimpressive if it means that one should never settle for anything less than absolute autonomy, while it is entirely reasonable if it means that one should never abandon one's autonomy altogether. Wolffs inconsistency in his use of the notion of autonomy is designed to permit his argument against authority to appear both plausible and bold. The argument is only plausible, however, when it is understood in a way that makes it commonplace; and it is bold only when it is understood in a way that renders it altogether implausible.

Although he professes otherwise, it seems clear that Wolff knows as well as the next man that anarchism is not the only appropriate political viewpoint for someone who takes autonomy seriously. This appears most plainly in his discussion of unanimous direct democracy which, he concedes, provides an exception to his general claim that autonomy and authority are incompatible. The point about unanimous direct democracy is this: if no one is commanded to do anything except on the basis of a unanimous agreement in which he himself has joined, then no one is commanded to do anything but what he himself has said he thinks ought to be done; and then each person is in fact an author of the commands to which he is subject, and no sacrifice of autonomy whatever is involved in his compliance. But now, consider how Wolff elaborates this point:

A community may agree unanimously on some principle of compulsory arbitration by which economic conflicts are to be settled. An individual who has voted for these principles may then find himself personally disadvantaged by their application in a particular case. Thinking the principles fair, and knowing that he voted for them, he will (hopefully) acknowledge his moral obligation to accept their operation even though he would dearly like not to be subject to them. He will recognise the principles as his own.... More precisely, this individual will have a moral obligation to obey the commands of the mediation board or arbitration council, whatever it decides, because the principles which guide it issue from his own will. Thus the board will have authority over him (i.e., a right to be obeyed) while he retains his moral autonomy. (24-25, emphasis in original)

This passage, apart from its easily dispensable restriction to economic conflicts, gives a routine account of the source of political obligation in any constitutional state. Wolff presents it as a description of an arrangement that is generally impractical and whose interest is essentially theoretical -- unanimous direct democracy. But this is quite gratuitously misleading. To speak of "unanimous direct democracy" suggests that everyone votes on everything and that no decision is valid if anyone dissents from it. Wolff plays on this meaning of the term when he asserts that, while unanimous direct democracy does satisfy his criteria for a morally acceptable form of government, it is too unwieldy to be considered a viable political alternative. But the "unanimous direct democracy" he describes in the quoted passage is not unwieldy or impractical at all. Indeed, it is very much like what we have, or are in principle supposed to have, now.

There is no question, in Wolff's description, of everyone voting on everything or of decisions being binding only if they reflect the unanimous agreement of the electorate. The only thing direct and unanimous in his account is the consent to procedures for reaching decisions. Once that consent has been given, those who have given it are -- hopefully! -- already to acknowledge that the decision-makers specified by the procedures in question have a right to be obeyed, whatever they decide -- i.e., even if their decisions strike the citizen as neither good nor wise. The unanimity at issue here is not a matter of the remarkable outcome of a voting procedure that might have come out differently. Those who give their consent to a set of procedures are by definition unanimous in doing so. Talking about a condition of unanimity in this context is, accordingly something of a red herring.

There is no trace of anarchist thinking in what Wolff says in the passage in question. What, he there commends is nothing other than a constitutional government whose authority is conceived to derive from the consent of the governed -- a social arrangement in which procedures are specified for reaching decisions by which all those who consent to the constitution thereby agree to be bound. In the situation Wolff envisages, the individual does not retain absolute autonomy; on the contrary, he binds himself to accept decisions with which he disagrees. Wolff characterizes him nonetheless as being morally autonomous and thus endorses a limitation of the scope of individual autonomy such that the person's autonomy extends not to the content of commands but to the procedures by which commands are legitimated. In this endorsement, he does no more than to subscribe to what is surely the most conventional and widely accepted justification of constitutional political authority.

1. Three-digit numerals refer to the pages of Wolff's essay "On Violence," Journal of Philosophy 66 (1969). Two-digit numerals refer to his book In Defense of Anarchism (New York: Harper Torchbooks, 1970).

2. While the state's right to command is perhaps not strictly entailed by the obligation of its subjects to obey, their promise to obey may reasonable be construed as conveying this right.

3. Cf. pp. 27-34.