Kurt Baier, "The Justification of Governmental Authority," The Journal of Philosophy, Vol. 69, No. 20, Sixty-Ninth Annual Meeting of the American Philosophical Association Eastern Division. (Nov. 9, 1972), pp. 700-716. To be presented in an APA symposium on Authority and Autonomy, December 29, 1972. Commentators will be Carl Cohen and Gerald Dworkin; see this journal, this issue, pp. 716 and 716-718, respectively.


Kurt Baier

University of Pittsburgh

Authority is now widely held in contempt, probably none more than that of law and government.1 Some would say that, provided the proper limits of governmental authority are clearly set out and provided governments adhere to these limits, governments can justifiably claim authority over their subjects and may use coercive sanctions against those who refuse to obey. Others find this the most objectionable approach. They might be willing to put up with the tyranny of the majority or even the brutality of the police as long as governments do not attempt to cover up their bloody business with the wig and gown of justice and authority.

Recently, two interesting arguments have been advanced (I shall call them, somewhat arbitrarily, the arguments from freedom and from autonomy) designed to show that what is objectionable are not the abuses of governmental authority but authority itself. The argument from freedom could be formulated as follows:

  1. A government's claim to authority can be justified if and only if there is justification for saying that, simply by issuing decrees on certain matters, a government provides its subjects with adequate reasons to do what it requires of them, reasons which are moreover sufficient to justify the government in imposing coercive sanctions on those who refuse to obey.
  2. But it can provide such reasons if and only if it can, by promulgating such regulations, make it wrong for its subjects to do certain things that were not wrong prior to their promulgation.
  3. But no one, not even God, can do that.
  4. Hence, the use of coercive sanctions to back up the law amounts to forcing people to do what it would not be wrong for them not to do.
  5. But this is a violation of their freedom, hence unjustifiable.
  6. Hence, there can be no legitimate governmental authority.

Against step (ii) of this argument one might, however, raise the objection that legislators need not try to make things wrong by legislation, but could confine themselves to promulgating the rules of morality and to enforcing them. Hence it is not impossible that there should be legitimate governmental authority. This objection is met by the argument from autonomy, which goes somewhat like this:

By declaring ("finding'') authoritatively what it is morally wrong for subjects to do, a government forces its subjects to do certain things, not because they have themselves judged these things to be morally wrong, but because the government has said so. If one recognized a government's authority to do that, one would thereby hand over to the government the final decisions about what one ought to do, that is, relinquish one's autonomy. But one's primary obligation is to achieve the highest degree of autonomy possible; hence this is unjustifiable; and so there can be no justified governmental authority.2

One could perhaps sum up one central point in these two arguments by saying that the belief that there is such a thing as justified governmental authority involves two legalistic mistakes. The first is to think of right/wrong on the model of lawful/unlawful. It is to think that one can by an act of will change the moral character of a type of act, in the same way as a legislator can by his fiat change its legal character. The second is to think of right/wrong on the model of constitutional/unconstitutional. It is to think that one can determine the moral character of a particular act, in the same way as a supreme court judge by an expert verdict can determine its legal character.

In this paper I am concerned to say what difference it makes whether we think of governments as having or as not having justified authority. I want in particular to make clear what would justify governmental authority and what kinds of reason for accepting their directives are provided by governments with legitimate authority. To this end I first give an account of authority in general, indicating briefly some of the important peculiarities of governmental authority. I shall next examine whether what I have just called the first legalistic mistake really is a mistake. And finally I say a few words about the bearing of autonomy on the whole issue.


1. There are three main types of account of authority, none entirely satisfactory. The first sees authority as a type of power or ability. Thus, T. D. Weldon says, " 'authority* means power exercised with the general approval of those concerned,"8 and De Jouvenel conceives of it as "the ability of a man to get his proposals accepted." 4 But to say of a policeman that he has authority to arrest suspects does not mean that he can always get his "proposal" to arrest a suspect accepted by him. The second type of account therefore represents it as a sort of entitlement or legitimation. Thus, Hobbes claims "that by Authority, is always understood a Right of doing any act: and done by Authority, done by Commission, or License from him whose right it is."5 According to J. R. Lucas, "a man, or body of men, has authority if it follows from his saying 'Let X happen' that X ought to happen."6 And R. S. Downie says that "In all cases of authority . . . there is necessarily some sort of legitimacy although there may or may not always in fact be accompanying power."7 But 'In certain matters Rasputin had absolute authority with the Czarina' does not imply that Rasputin had any sort of right to guide the Czarina in these matters, let alone that if he said to her, "Let X happen," it really ought to happen. Perhaps for these reasons, the third theory says that 'authority' has two senses8 or that it is ambiguous (Defense, 5), having both a "de facto" and a "de iure" sense. Although the most plausible, this view, too, is not without its difficulties, if only because there are quite a few different interpretations one might give to the expressions the 'de facto sense' and the 'de iure sense', and it seems equally unsatisfactory to retain all or to eliminate some, and the labels 'de facto' and 'de iure' give no clue or the wrong clue to the relation between the different interpretations.

Suppose, for instance, that Eichmann has de iure authority to organize the Final Solution. Does that mean simply that he has legal license (Hobbes) or a right (Hobbes, Peters) or that he is justified (Peters, Downie, Lucas) or that there is something legitimate about his doing it (Downie) or that his subordinates ought to or are obligated or are at least within their rights to carry out his orders (Lucas)? If legal license really implies any or all of these other things, why is this sense not also a factual sense? Is it not a fact, though of course an "institutional fact," that he did (or did not) have this license? And if having the legal license does not imply having de iure, legitimate, or justified, etc. authority, what is the connection between these different senses lumped together under the heading 'the de iure sense'?

And now suppose that Eichmann has de facto authority. This can mean at least three different things, depending on whether it implies that the subjects actually comply (Benn & Peters, 20), that they believe they ought to comply (Defense, 5), or that they comply because they believe they ought to ("Authority," 96). Clearly, these three "facts" are rather different from one another, and only the first is "purely" factual -- for in the cases of believing that one ought to do something and even doing it because one believes it, there would seem to be implied "some sort of legitimacy" (Downie), i.e., some "de iure" element, even if only indirectly.

I conclude that this third account, too, is unsatisfactory. The conceptions of being de facto and de iure are both too crude and at the same time too narrow to do justice to the complex normative, explanatory, and descriptive elements in the concept of authority.

2. My own thesis is that all authority claims make explicit or tacit reference to, and at the same time imply that a complex relationship holds between, the following five elements which are the main dimensions of the concept of authority:

  1. a range of matters, to be called "the (authority) scope";
  2. a class of utterances, to be called "the authority utterances," bearing on some matters within the authority scope;
  3. a class of persons to be called "the (authority) principals," whose utterances within the authority scope are the authority utterances;
  4. a range of persons, to be called "the (authority) subjects," for whom the authority utterances are meant;
  5. a form of behavior, to be called "compliance behavior," which is behavior appropriate for a person who accepts some authority utterance(s).

Thus, 'Jones is an authority on infectious diseases' specifies one authority principal, Jones, and the authority scope, infectious diseases. But it makes only tacit reference to authority utterances, to the range and identity of the authority subjects, and to the range of compliance behavior. 'Roberts has authority to direct the traffic in this area' explicitly specifies the authority scope and one authority principal, Roberts. It also specifies, though only by implication, the range of the authority subjects (those involved in the traffic in this area), and it implies a good deal about authority utterances and compliance behavior.

The complex relationship, implied by all uses of 'authority' to hold between five elements of these kinds, can be specified in two conditions, to be called the normative and explanatory conditions, respectively. If both these conditions are satisfied, then there holds between five specifiable elements of these types the complex relationship to which all uses of 'authority' refer.

The normative condition says that any use of 'authority' implies a specifiable set of requirements that determines a particular scope, qualifications of principals, subjects, utterances, and compliance behavior, such that, if a given person has the qualifications of a principal relative to this scope and this range of subjects, there is in general sufficient reason of a certain kind9 for such subjects to accept authority utterances, that is, utterances by such principles within this scope.

The explanatory condition says that, if there are persons who believe that a given person has the qualifications of a principal relative to a range of subjects including them and relative to this scope, and if they engage (or tend to engage) in compliance behavior with respect to that person's authority utterances, then their belief that that person has these qualifications must operate with them as a sufficient reason for engaging in such compliance behavior.

Suppose that a policeman is directing the traffic at an intersection; then talk about his authority implies the satisfaction of both the normative and explanatory conditions. If there is no set of qualifications such that, if he satisfies them, then motorists have sufficient reason (or the kind yet to be specified) to follow his directions, then there is nothing to which 'his authority' can refer. Or, supposing there are motorists who believe that he has these qualifications and who follow his directions, if their belief is not for any of them, by itself, a sufficient reason for engaging in compliance behavior, then again there is nothing to which 'his authority' can refer.

We should note that both the normative and the explanatory conditions are formulated as conditionals. The antecedent of the normative condition is that certain persons satisfy the qualifications of a principal relative to a given scope and range of subjects. I shall call that the grounding condition, and I shall speak of grounded authority whenever the grounding condition is satisfied. The antecedent of the explanatory condition is that certain persons who believe that certain others have the qualifications of a principal relative to a certain scope and a range of subjects including them, engage (or tend to engage) in compliance behavior relative to these principals' utterances within this scope. I shall call that the effectiveness condition, and I shall speak of effective authority whenever the effectiveness condition is satisfied. Now, these antecedents of the normative and explanatory conditions may be contrary to fact, and there can therefore be ungrounded and ineffective authority.

Thus 'Jones is an authority on infectious diseases' is an authority idiom which implies that both the effectiveness and the grounding conditions are satisfied, i.e., that the authority is both effective and grounded. For Jones could hardly be an authority on these diseases if his authority were not effective, that is, if there were no people who complied with what he told them and did so because their belief that his authority was grounded was for them a sufficient reason for doing so. But this idiom also implies that his authority really is grounded, not merely believed so. For if it were not, Jones would be a charlatan, and so not an authority on the subject. That he would not be an authority but merely (wrongly) believed one, is shown by what we would have to say when he is unmasked: not that he used to be, though he now no longer is, such an authority, but rather that, though he was believed to be, in reality he never was.

Other idioms imply only that one or another of these conditions is satisfied or leave it open whether either of them is. Thus, 'Some teachers have great authority with their pupils' implies the satisfaction of the effectiveness condition (the effectiveness of the authority ascribed to them), but is noncommittal about the other; whereas 'The Court has constitutional authority to order busing' implies the satisfaction of the grounding condition, but is noncommittal about the other.

It seems there can also be authority that is both ineffective and ungrounded; for we can speak of the authority to, say, arrest drug pushers caught in the act. This implies only that one could specify qualifications for principals and subjects relative to that scope. This seems plausible for this particular scope, though by no means every conceivable scope, e.g., the authority to have those over eighty strangled. But then, relative to a given policeman and a given range of subjects (i.e., persons believing that they have the appropriate principal qualifications), this authority (individuated over the dimension of scope) may be both ineffective (the subjects do not comply) and ungrounded (this policeman does not have the proper qualifications). Of course, when authority is individuated over the dimension of principal (the authority someone has), then it must be either grounded or effective. This fact gives plausibility to the view that there are two senses, a de facto and a de iure sense of 'authority'. But surely it is not plausible to say that in 'the authority to arrest drug pushers' 'authority' is used either in the de facto, the de iure, or both senses, given that such authority may be both ineffective and ungrounded, or that it is used in yet a third sense. Rather, 'authority' is used in the same sense in all cases, though we can always ask whether the authority in question has one or both (and in some cases even either) of two important properties, effectiveness and groundedness.10

Another thing meant by the 'de iure and de facto senses' of 'authority' are two important types of authority,11 natural and conventional, based on two different ways of determining the scope of someone's authority. The scope of natural authority is determined by the natural qualities of the principal, that of conventional authority by the conventional powers attached to the position or role occupied by the principal. Natural authority is acquired by acquiring those qualities which promise excellence of performance. Conventional authority is conferred by occupying a position with conventional powers. Of course, the effectiveness of a person's authority may to some extent depend on his natural qualifications -- perhaps a completely incompetent doctor would have no authority with his patients even if he were the ship's doctor or the regimental surgeon -- but the prestige conferred by the office may well suffice.

Yet another thing philosophers mean by these "two senses of 'authority*" is simply the difference between the two types of conventional authority we ordinarily distinguish by the terms 'de iure* and 'de facto*. That distinction relates to the question of the conventionally proper occupancy, by an authority principal, of a given conventional position. There are two different ordinary ways of drawing this distinction, a narrower and a wider one. The narrower one construes the properties as incompatibles and limits their application to effective conventional authority. In this narrow use, to say that Eichmann's authority to organize the Final Solution was de iure is to say that he had effective authority and that he lawfully occupied the position in question; that it was merely de facto, is to imply that, though he had effective authority, it was not de iure. In the wider use, the properties are compatible and do not mark a distinction within effective conventional authority. In this wider use, saying that Eichmann's authority was de facto means that it was effective, but it is noncommittal about its being de iure; and that it was de iure means that Eichmann came lawfully to hold a position with legal powers corresponding in scope with that of his authority, but it is noncommittal about its being effective.

3. So far I have not said anything about the special kinds of reasons implied in the authority-grounding condition. I believe this condition is such that grounded authority is justified authority: authority such that there is justification both for the principal's having it and for the subject's complying with it. Of course, this does not mean that the grounding conditions are such that they guarantee the principal's not abusing his (properly grounded and so justified) authority. This relationship of the grounding condition to justification significantly distinguishes authority from influence and power.

There would seem to be three requirements on reason if they are to be sufficient reasons for subjects to comply with what they are told by authority principals satisfying the grounding conditions:

  1. they must be principal-related;
  2. they must not be imposed; and
  3. they must be subject-oriented; i.e., complying with them must benefit the subjects, at least as a class.

(i) Reasons are principal-related if only the speaker or someone similarly qualified or positioned can give them to the person in question. In this respect, authority is like influence with and power over, but unlike influence on. If the Godfather asks a mafioso to do something for him, say, to arrange the funeral for his murdered son, this is a case in which the reason for the mafioso to comply is principal-related. It would not be a reason to arrange that sort of funeral for whoever asked him. It is a reason only because the Godfather "has a hold on him," ranging from obligation to the Godfather for past services, via hopes for future protection, to fears of reprisals. Only a person thus or similarly related can provide reasons of this sort.

By contrast, a teacher's influence showing in the works of his disciples will not normally be the result of his influence with or power over them, but will actually be that influence on them. The disciple writes and speaks like the teacher not (or not normally) for principal-related reasons, but because the teacher offered good reasons for the views themselves, reasons anyone could have offered him.

The grounding conditions of authority must be such that its satisfaction provides principal-related reasons. If Jones, the authority on infectious diseases, has authority with his students, then they must indeed believe that he is an expert in the field, but they must if necessary also be willing to accept some of the things he tells them without asking for evidence or proof. If they ask for proof of a particular claim, then they don't accept it on his authority, and if they don't accept any of his claims on his authority, then he does not have authority with them. Their accepting, on his authority, what he tells them, implies that they believe he satisfies the grounding condition and that this in turn provides them with sufficient reason to accept the things he tells them without their insisting on judging those things themselves on the basis of principal-neutral reasons.12 In Jones's case, the grounding conditions will be Jones's expertise on infectious diseases and his readiness to tell them what he believes to be the truth on these matters.

(ii) Reasons are imposed if they lie in the consequences the principal attaches to the subject's compliance or noncompliance. The reasons for the mafioso to comply are imposed reasons. The reason for me to take a raincoat when the forecast is for rain is not imposed. It is characteristic of having a hold on someone, or having power over him, but not of having credit or authority with someone, that the satisfaction of the relevant grounding condition should provide imposed reasons. Having power over someone implies that the principal can and is willing to inflict harm for noncompliance; having influence with him that he has other kinds of hold over him, such as having the ability and the willingness, conditional on compliance, to do favors in the future.

(iii) Reasons are subject-oriented if the rationale for ascribing to someone authority with a certain scope is that those subject to it will benefit from the arrangement, that is, from complying with what he tells them. It is on account of this that the satisfaction of the grounding condition also constitutes the justification of the authority. In this respect, authority importantly differs from power and influence.

It is difficult to say both briefly and correctly what I mean by "subject-oriented reasons," and particularly so in the case of what might be called "constative" authority utterances ('You have a salmonella infection') as opposed to "directive" ones ('Take tetracyclic). For in the constative case it is not so clear as in the case of directives, what constitutes compliance behavior, except to believe what the authority says, but such belief need not be in any way beneficial to subjects. It should, however, be obvious that those for whose benefit this sort of information has been collected (the sick) and those who gather it to help the sick (doctors), will tend to benefit from having -- so as to be able to act on (this is what 'compliance' means here) -- the pronouncements of an authority rather than those of a charlatan or none at all. And it should also be clear that in the case of neither influence nor power do we find reasons that are necessarily subject-oriented even in this weak sense. One would, on the contrary, expect both power and influence utterances to yield principal-oriented, not subject-oriented, reasons for subjects to comply with power or influence utterances.

To understand what "sort of legitimacy" (Downie) is involved in authority, we must understand the difference between being grounded and being justified. The fact that the influence, credit, or weight someone has with someone else is grounded (because, let us say, he has a hold on him, can and is prepared to repay, or is influential) does not imply that it is justified. And the same is true, with a minor qualification, also of power. The qualification, which may come as a surprise, is that power is necessarily grounded but not, of course, therefore justified. A person who does not satisfy the grounding conditions, who does not have the right sort of hold, on people, simply does not have power over them. Thus, in the case of power, the grounding condition is necessarily satisfied. By contrast, in the case of influence, credit, and weight, the compliance condition is necessarily satisfied. There is simply no such thing as ineffective influence, weight, or credit. This distinguishes both power and authority from influence. For, again surprisingly, power may be ineffective, e.g., the legal power to impose fines on corporations violating pollution laws.

What distinguishes authority from both power and influence (apart from the nature of the grounding condition) is the fact that neither the grounding nor the effectiveness condition is essential and that therefore there can be (and indeed are) authority idioms that are noncommittal about whether the authority is effective, and others that are noncommittal about whether it is grounded. It is not, of course, possible for there to be power and influence idioms some of which are noncommittal on the first and also some noncommittal on the second point if, as I said, there is no such thing as ungrounded power and no such thing as ineffective influence.


4. I can take up only one point in the argument from freedom: whether any one can, simply by his word, make something wrong. The first point to note is that there are two different ways of doing so: a direct and an indirect way. When marijuana, abortion, or soliciting in drag, let us say, are switched by our legislators from unlawful to lawful and back to unlawful, this change from one legal property to another is direct, that is, brought about without reliance on any deliberate prior change of another property to which the different legal property already attaches. If I am a pot pusher, or an abortionist, and so on, then I can indirectly make my conduct lawful, by changing my domicile and, as a result, plying my trade in places where it is already lawful to do so. The same distinction between directly and indirectly changing the property of a type of act also applies to the moral properties right and wrong. Under certain circumstances, giving someone $100 is making him a gift. If I have stolen the money from him, then giving him $100 is not a gift but restitution of stolen property. By stealing from him, I can thus indirectly change my act of not giving him $100 from not giving him a gift, which is not wrong, into not returning stolen property, which is wrong. Thus, even if I cannot directly make anything wrong, it seems clear that I can do so indirectly. And so the question arises whether it is not possible for someone, simply by his word, indirectly to make something wrong. And if that is possible, then perhaps the government, simply by its word, can.

Certain expressions, such as 'Sure . . .' or 'You can count on it* and 'Perhaps . . .' or 'But don't count on it', when added to one's declarations of intention serve the purpose of telling the other of the firmness or weakness of one's resolve to do these things. When one's resolve is weak, then in certain circumstances it would be wrong to declare one's intention without indicating that weakness; doubly wrong actually to say that one's resolve is strong, because in so doing (quite apart from telling a lie) one unnecessarily risks causing the other expense, harm, loss, suffering, disappointment, and the like. If one already has made such a declaration of intention, the only way to avoid causing such expense, etc., may well be actually to carry out the declared intentions. Thus, simply by the use of certain words, such as 'You can count on it', one can at will change the character of one's future acts. One can transform them from being merely the not carrying out of one's intentions, which is morally permissible, to being the unnecessary risking or avoidable causing of expense, etc., which is morally wrong. One can thus deliberately, though indirectly, change the moral nature of a certain type of act from being right to being wrong for one to perform.

Now, the legislator, in virtue of his sanctioning powers, has this ability, by his mere word, to change the character of certain types of act: the side of the road you drive on, the kind of gas you use, the smog devices you employ, the amount of water, power, gas, etc, you use, for what purposes and at what times. Because of the fear of the attached sanctions, the legislator's word can turn random conduct into orderly uniform conduct, and, therefore, engaging in it into generally beneficial, engaging in other conduct into generally harmful conduct. Although this point would need to be developed at greater length, I hope I have said enough to show that a government can, merely by its fiat, which directly makes certain types of conduct unlawful, indirectly make them morally wrong. Of course, this does not show that whatever the government makes unlawful, is thereby indirectly made wrong. On this view, only a limited class of conduct could be turned into wrongful conduct in this way. But it is enough to show that there can be justified governmental authority provided its scope has been suitably defined. Unless such a government abuses its authority, then its subjects ought to obey its commands.

But it also seems possible directly to make something wrong merely by one's word. Promising is the obvious example. For by promising I seem to bring into being a new obligation. If I say, "Of course, I'll be there, you know me, but mind you, it's not a promise," I avoid creating any promissory obligation, i.e., any obligation that goes beyond seeing to it that the other will not suffer any loss etc. owing to my assurances. If I promise, then I have an obligation going beyond that, namely, to do as I said I would, irrespective of whether the promisee would incur any losses if I changed my mind. It seems to me that not only is this the way we understand promising, but that such an institution of self-obligation is, within certain limits (excluding the immoral promise), highly desirable and morally unobjectionable. Likewise, a good deal can be said for viewing government as an institution of other-obligation, though of course there it is even more important to be clear about its permissible scope. The most obvious advantage of an institution so viewed is that, as in the case of the promise, the obligation is created independently of whether the "word" has in fact made the conduct in question wrong. Just as the promisor has a (prima facie) obligation to do what he promised whether or not changing his mind would hurt the promisee, so on this interpretation the subject has an obligation to obey the law whether or not it has created a situation in which it would be wrong to do the sort of thing it forbids.

But whether the case for this institution is good or not, what I have said should be enough to show that it might be good, so that there could very well be justified governmental authority with the power, by its word, directly to make certain conduct (prima facie) wrong. The proof that there is no such authority cannot rest solely on the concept of government, authority, and making wrong by fiat. Such a proof would have to be much more complicated, depending on many empirical facts about human needs and the best way to organize their satisfaction. Schemes for the improvement of political institutions that are based on simple a priori proofs of the impossibility of justified governmental authority are as jejune as the perennial philosophers' attempts to improve the institutions designed to deal with crime by a priori proof that, because everything is determined, no one is responsible for anything he does.


5. If, as I tried to show, the argument from freedom fails, then even the soundness of the argument from autonomy could not prove the illegitimacy of all governmental authority -- and, as we shall see, that argument rests on a dubious premise.

It may be thought that one could base one's rejection of political authority solely on the idea of (moral) autonomy (as Wolff appears to do in Part I of Defense), by arguing that recognition of the authority of the state, i.e., the admission that one has a moral obligation to obey the laws of the state "simply because they are the laws of the state" (Defense, 18), implies forfeiture of (moral) autonomy (18/19), that "the primary obligation of man is (moral) autonomy" (18), and that the obligation to be (morally) autonomous must always outweigh the obligation one may demonstrably have to obey certain sorts of laws simply because they are part of the law of the land.

6. But there are two difficulties. The first is that this construes the conflict between authority and autonomy as if it were a conflict between two substantive prima facie duties, of which the second is always more stringent than the first. However, the conflict here is not between substantive but between methodological duties. The argument claims that one has a duty to employ a certain method for determining what one's substantive duty is, but recognizing the authority of the state amounts ipso facto to failing in that methodological duty. However, if what I said in the previous two sections is sound, then recognizing the authority of the state does not necessarily involve the acceptance of a morally objectionable method for determining what one's duty is, since such recognition does not require any particular method. It implies something about substantive, not about methodological duties. I return to this point below (section 8).

7. The second difficulty is to find a suitable sense of 'autonomy' in which it is true both that recognizing the authority of the state is incompatible with being autonomous and that one has a duty to be autonomous. For what sense could that be?

It cannot be literally "self-legislation," for no individual can be literally a self-legislator, in the sense that he submits only to laws that he alone has made for himself, and so "is not subject to the will of another" (Defense, 14). He can at best try to follow only self-made maxims. For while a society, such as a colony, can become autonomous by getting rid of the legislature of the mother country, and perhaps even "more autonomous" by becoming democratic, that is, by assigning every member of the society a share in the making of the law to which each is subject, the individual members of that society cannot thereby become autonomous. For the law to which one is subject is necessarily made largely by others. This remains true, even in the unlikely event that all want, for a significant period, exactly the same laws.

It cannot be judgmental autonomy in general, for, as Wolff rightly points out, it is not even reasonable to be judgmentally autonomous in all fields (15), let alone is it one's supreme duty.

Is it, then, moral autonomy, making the final decisions about what one ought, morally speaking, to do? There would seem to be three points, failure at which would open one to the charge of moral heteronomy: if because of pressure from others or one's own desires one does not carry out what one has decided or judged the morally right thing to do; if from fear or diffidence or other reasons, one does not choose the proper method for arriving at such a judgment or decision; if from fear or similar reasons one arrives at a decision or judgment which a proper employment of that method would not have yielded.

However, if we have such a duty, we can with perfect consistency both perform it and recognize the authority of government. For such a recognition does not amount to an admission that the government has "authoritatively found" on the moral character of a type of conduct. It does not commit one to saying that one was mistaken in thinking that, say, driving on the right was morally neutral, that the government must be right in declaring it wrong. One is no more committed to that than the promisor, who acknowledges that he has an obligation to, say, meet Jones as promised, is committed to the view that by promising he has "authoritatively found" on the wrongness of his not meeting Jones, which he previously had considered morally neutral. Of course, if one were thus committed, one might well be guilty of moral heteronomy on the choice of the wrong method or on faulty use of the right method.

8. But now I want very tentatively to raise doubts about whether all forms of moral heteronomy really are morally objectionable. Is there not such a thing as natural moral authority and legitimate recognition of it, at least on questions of casuistry? Is the heterono-mous peasant who accepts the priest's judgment on what it would be right for him to do in matters of abortion, divorce, or suicide, necessarily failing in a (methodological) duty? What if he tries but fails to understand the learned arguments? Given his information and the state of the art, what better method is available to him? Is reliance on his "conscience" really necessarily preferable? What if "his conscience tells him" to rely on the priest?

And what about conventional moral authority and its recognition? Is not the government, in relation to some moral matters, in much the position the supreme court judge is in relation to constitutionality? Governments should be and often are better informed than individuals about what is needed and what it would be morally desirable for everyone to do, say, in economic, military, or population matters. One could say that, by selecting from many possible arrangements it judges morally desirable, the one it judges the most suitable, the government has both authoritatively found that arrangement most suitable and by its finding made it prima facie wrong for its subjects to act otherwise. Of course, no such findings can be final: reasoned judgments that a law requires what is morally undesirable should lead to its being changed. And no such findings can be absolute: every individual must balance the prima facie duties created by the law against the other prima facie duties to which he is subject.

I have argued that recognition of governmental authority does not commit one to the two legalistic premises involved in the arguments from freedom and autonomy and that, even if it did, this might not show the illegitimacy of all governmental authority, since it is not at all obvious that either of them is mistaken.


1 Cf., e.g., Robert Nisbet, "The Nemesis of Authority," The Intercollegiate Review, viii, 1/2 (Winter/Spring 1972): 5-13.

2 I have taken these arguments from various writings by Robert Paul Wolff and others. I hope I have not misrepresented or weakened them, though I must mention that my use of 'freedom' is very different from Wolff's. In any case, I am not concerned to find loopholes in them, but to isolate what seem to me their most interesting and strongest premises and to examine these. Here are some of the works I have used and profited from: Wolff, In Defense of Anarchism (New York: Harper Torchbooks, 1970), to be referred to as Defense; Wolff, ed., The Rule of Law (New York: Simon 8c Schuster, 1971); Wolff, The Poverty of Liberalism (Boston: Beacon Press, 1968); Gerald Dworkin's review of Defense, Journal of Philosophy lxviii, 18 (Sept. 16, 1971): 561-567; Stanley Bates, "Authority and Autonomy," ibid., lxrx, 7 (April 6, 1972): 175-179; Lisa H. Perkins, "On Reconciling Autonomy and Authority," Ethics, lxxxii, 2 (January 1972): 114-123; David Sobers, "Wolff's Logical Anarchism," ibid.: 173-176.

3 The Vocabulary of Politics (London: Pelican 1953), p. 56.

4 Bertrand de Jouvenel, Sovereignty (Chicago: University Press, 1957), pp. 29-31, Quoted by R. S. Peters, "Authority," Aristotelian Society, Supplement Volume xxxii (1958): 207-224; to be referred to as "Authority."

5 Leviathan (London: Everyman Library, J. M. Dent, 1937), p. 84.

6 J. R. Lucas, The Principles of Politics (Oxford: Clarendon Press, 1966), pp. 16 ff.

7 R. S. Downie, Roles and Values (London: Methuen, 1971), pp. 76 f.

8 Peters, op. cit., pp. 84 ff., also S. I. Benn and Peters, The Principles of Political Thought (New York: Free Press, 1965), pp. 20 ff.

9 This kind will be specified in section 3.

10 I want to record my indebtedness to John Cooper and Charles Teggatz for criticizing and helping me improve earlier versions of this last point, though I fear they may still think I am wrong.

11 Cf. on this distinction, Gareth B. Matthews, "Senses and Kinds," Journal of Philosophy, lxix, 6 (March 23, 1972): 149-157.

12 Authoritarians are those who insist on their subjects' acceptance of what they tell them without giving them principal-neutral reasons, even when there are good reasons for and none against doing so.