MORAL OBLIGATION 217has an obligation to do x can be raised andanswered. The only "objective" point that can beestablished is whether a person's case comes undera rule, and whether he will go on to say that he hasan obligation, but not whether he is right or wrongin going on the way he does.Hart's account of obligation is thus a sophis?ticated Attitudinal (Emotive) Theory. He explainsthe peculiarities of obligation claims as a combina?tion of two quite different ingredients. The first isa fairly straightforward empirical, sociological one,namely, the ascertainable fact that a person's casecomes under a rule of obligation which is actually"accepted" by a significant portion of his group,regardless of the nature and cogency of theirreasons for accepting it.16 The second is tied to thespeaker's "acceptance" of the rule, from the"internal point of view."My objection to Hart's theory can be put in theform of a dilemma: Can "outsiders" have obliga?tions or can they not ? If we say they cannot, wehave plainly failed to give a correct account ofobligation claims. For such claims involve direc?tives purporting to be "unconditionally" applicableand sound, and possessing a peculiarly strongbinding force. But if "outsiders" cannot haveobligations, then the applicability of obligationclaims is conditional on a person's actually "accep?ting" a given rule.17If we wished to convince an "outsider" that he isobligated to do a certain thing, we should have toargue, not that he actually had this obligation but,absurdly, that he ought to accept the obligationrules of the group so that he can have obligations. Infact, of course, it is sufficient to provide adequatereasons in support of the obligation rule. For whensuch reasons are available, nonacceptance of therule is irrelevant since in that case it has beenshown that the rule ought to be accepted.Embracing the other horn of the dilemma, say?ing that "outsiders" also have obligations, reducesHart's position to that of J. Austin.18 For then,since the directives embodied in these obligationrules are not regarded as requiring justification butmerely as being widely accepted, they are brutecompulsions for those who do not happen toaccept them.One last point. Hart's main concern in giving a"content-independent account" of legal obligationis to avoid landing in the camp of the Natural Lawtheorists and so having to maintain that iniquitoussocial rules cannot be law.19 No such consequenceneed follow from the rejection of Hart's view thatobligation-claims (apart from drawing attention tothe fact that a person comes under a rule ofobligation) merely express a certain attitude of"acceptance" toward these rules. To show this, letus consider argument (C) :(i) One has a (moral) obligation to obey thelaw as such.(2) R ("Do x") is a valid law in society S.(3) So there is a presumption that people living in Shave a (legal) obligation to obey R.(4) But Jones lives in S and his case comes underR.(5) So there is a presumption that Jones has a (legal)obligation to obey R.(6) So, other things being equal, Jones (morally)ought to obey R.We avoid the consequence to which Natural Lawtheory is committed by insisting on the presumptivenature of the steps from (2) to (3) and from (4) to(5). Laws with a certain content may not give riseto obligations. But provided the content of R is notobjectionable on moral grounds, the presumptionthat Jones has a (legal) obligation to obey R holdsgood. Hence there is no incompatibility betweenthe claim (A), which Natural Law theorists rightlyinsist on, that unless (1) is sound, no valid law cangive rise to obligations, (B), which I insist on, thatsome valid laws do not give rise to (genetic) legal,hence not to (binding) moral obligations, and (C),which positivistically inclined philosophers, likeHart, insist on, that we are justified in assertingthat, unless the contrary is proved, the fact thatsince a certain valid law enjoins the doing of x,those whose case comes under it may be thoughtand said to have a (genetic) legal obligation to do x.Perhaps the most plausible version of the Will16 Hart's explanation of "acceptance" is in terms of how people regard deviations from the rule. In The Concept of Law hespeaks indifferently of their regarding it as a signal (p. 87), or a reason (p. 88), or a justification (pp. 54-55) for a hostile reaction tothe rule-deviants, yet surely whereas a signal does not purport to have any cogency, a justification purports to have a very highdegree. The second and characteristic ingredient (pp. 86, 88) is an attitudinal one, namely, that only "insiders," only those who"accept" the rule in question, will actually use obligation language (thereby expressing their acceptance of the rule), though theymay apply it equally to "outsiders."17 For further details on this point, see Pt. Ill, sect. 3 below.18 Cf. Pt. II above.19 Cf. Hart, op. cit., pp. 206-207.2l8 AMERICAN PHILOSOPHICAL QUARTERLYTheory is that which employs as its obligationcreating device The Promise. Here, the giver andthe addressee of the directive are one and the sameperson. The person to whom the promise (not thedirective) is given, thereby receives the right to setin motion, if necessary, whatever machinery thereis for exacting fulfillment of the promise. The in?ference from the nonmoral premiss, that someonein certain circumstances uttered certain words con?stituting a direction to himself, to the conclusionthat he morally ought to follow the directive,seems wholly licensed by what is necessarily in?volved in a promise, and so to be analytic. All thesame, the validity of the inference plainly dependson two sorts of assumptions, assumptions concerningthe inferences licensed by the social institutionitself, and assumptions concerning the power of asocial institution to license such steps. If the movefrom "Jones promised to do x" to "Jones has anobligation to do x" is defeasible, and if the defeatingcondition is itself moral in nature, say, "Immoralpromises do not give rise to obligations," then thesocial institution of promising is an obligationcreating device which rests on moral convictions.Hence in the case of some social institutions, say,child-marriage vows or slave labor contracts, whichrest on no moral convictions, these "promises" donot give rise to genuine obligations. We shouldtherefore admit that what gives rise to moral obli?gations in the case of promises is not saying, accord?ing to the rules of the institution, "I promise," butthe fact, where it is fact, that it is wrong to breakour promises.In view of the extensive discussions this topic hasrecently received, it would be tedious to go over theground once more. I therefore simply assume thatthe Will Theory fails even in the case of the mostpromising candidate, The Promise.20I conclude that even the most persuasive versionsof the Will Theory are untenable. Their chiefweakness lies in their attempt to derive obligationsanalytically from the expression of some will. Theconclusions so derived have inevitably been shornof all normative content. Hence, there is a tendencyto smuggle in surreptitiously a supplementary prin?ciple such as that a will so expressed (promise,accepted social rule, law) should be regarded asbinding, or that a will so expressing directives hasauthority to bind, or that the will of such a personis certain to give directives which have a claim tobe regarded as binding. But in the absence of suchsupplementary principles, the will models em?ployed produce at best directives which it may beexcusable, perhaps wise (but not obligatory), tofollow.IllAnother group of theories, which I call ideolo?gical theories, attempt to derive obligation-claimsfrom the conclusions of some form of practicalreasoning. Consider a simple case. Jones is anadipose man with a certain heart condition whichrequires that he lose weight. Someone offers thefollowing piece of advice : stop eating bread andsalami. He might back up his advice as follows inargument (D) :( i ) If anyone with a certain sort of heart con?dition and suffering from adiposity is tominimize the risk of a heart attack, he mustlose weight.(2) If anyone who suffers from this condition andalso regularly eats bread and salami is tolose weight, he must stop eating bread andsalami.(3) But you have a heart condition of this sort,you suffer from adiposity, and you regularlyeat things, and it is your end to minimize therisk of a heart attack.(4) So, other things being equal, you must stopeating bread and salami.Such arguments are closely parallel to those wenoted above, in connection with obligationcreating social devices, such as commands andrules. Here, too, we find general directives statingor implying applicability conditions [i.e., (1) and(2)]; assertions to the effect that these conditionsof applicability are satisfied in the case of a parti?cular person [i.e., (3)]; and a conclusion whichapplies the directive, that is, the consequent of ( 1 )and (2), to that particular person. We must, how?ever, note five important points of difference.(a) The major premisses contain not only adirective, but also a statement of a "connection in?0 For a fuller discussion of both the assumptions I called in question, cf., John Searle, "How to Derive 'Ought' From 'Is',"The Philosophical Review, vol. 73 (1964), pp. 43-58; Roger Montague, " 'Ought' From 'Is','' Australasian Journal of Philosophy, vol.43 (*965)> PP- I44_I67; Evan K. Jobe, "On Deriving 'Ought' From 'Is'," Analysis, vol. 25 (1965), pp. 179-181 ; W. D. Hudson,"The 'Is-Ought' Controversy," Analysis, vol. 25 (1965), pp. 191-195; Antony Flew, "On Not Deriving 'Ought' From *Is',"Analysis, vol. 24 (1964), pp. 25-32; James E. McGlellan and B. Paul Komisar, "On Deriving 'Ought' From 'Is'," Analysis,vol. 24 (1964), pp. 32-37 ; James and Judith Thomson, "How Not to Derive 'Ought' From 'Is'," The Philosophical Review, vol. 7361964), pp. 512-516.MORAL OBLIGATION 219nature." This connection may be of various kinds,and assertions to the effect that it holds can betrue or false. Thus, if a heart attack is bound tooccur unless the agent loses weight or remains com?pletely motionless and if the latter is impracticable,then his losing weight is a necessary condition of thenon-occurrence of a heart attack. In that case, wecan say that to avoid a heart attack the agent mustlose weight. If it is possible for him to remainmotionless, then losing weight is not a necessarybut at best a sufficient condition of the nonoccurrence of a heart attack. If it is at any ratea sufficient condition of reducing the chances of aheart attack, then we can say that to avoid (orreduce the risk of) a heart attack, he can lose weight.If losing weight is thought greatly preferable tolying still, this can be indicated by saying that hewould be well-advised or that he should lose weight.This sort of argument is designed to support anend-promoting directive which is asked for andgiven when a person has a certain end and does notknow how to attain it. It is assumed that when thedirective is given, so is a solution to the ques?tioner's problem. Hence, the conclusion which canbe formulated either in terms of an imperative or interms of words such as "can," "must," "should,"or "ought," is felt to be a directive even when it isnot in imperative form. Conversely, even when for?mulated in the imperative, the conclusion is feltto be capable of being true or false, for it impliesthat there is a connection in nature between thebehavior delineated in the conclusion, and the endattributed to the agent in the minor premiss. Theonly difference between the imperative and theother formulations is that the former says nothingwhatever about the nature of this connection, andso indicates nothing about the merits of the solu?tion or the chances of success in following it.(c) The conclusion must therefore be read intwo ways, as an end-promoting directive to some?one seeking the solution of a problem, and asdescriptive of a certain sort of connection in nature,namely, a possible means-end connection. Ourconclusion, "Stop eating . . ." or "You must stopeating . . ." is interpreted as a directive if, sup?posing there to be a discrepancy between the direc?tive and Jones's behavior ; it is he or his behavior thatis criticized on that ground. It is read as descriptiveif supposing there to be no such discrepancy andJones still does not attain his end, namely, to loseweight, it is the conclusion which is criticized on thatground. Of course, just how soon he must loseweight and how much and under what conditionsDdepends on the degree of explicitnethe means-end connection is stated ision. The fact that the conclusiondescriptively at all shows that the cnature, asserted in the major premtaken as the ground of the conclusiona directive, and that the merit of ttherefore based on whether or not tholds.The main conclusions I wish to derive from thisare two : (i) that the reason why it is right to inter?pret the remarks "Stop eating . . ." or "You must(should) stop . . ." as directives is that they areoffered in a context in which the person addressedis assumed to have a certain end to which theaction delineated is implied to be a means. Neitherthe grammatical form nor the occurrence of thewords "must," "should," or "ought" are the rea?sons for it, since the substitution for these words ofthe word "can," which clearly has no directiveforce by itself, does not eliminate the directiveforce of the remark as a whole, (ii) That the func?tion of the words "can," "must," and "should" insuch remarks is rather to indicate the precisenature of the connection between the actionrecommended and the end of the person addressed.(d) We can now state one important differencebetween two ways of interpreting a directive suchas, "If you cannot stay in bed, stop eating breadand salami." Taken as an order or command, it isan end-setting directive : the if-clause specifying thecondition under which the directive is to befollowed. In that case, the question of its soundnesscannot arise, but that of its authorization can.But taken as a piece of advice, it is an endpromoting directive, the if-clause stating theapplicability-condition. In that case the questionof its authorization cannot arise, but that of itssoundness can. Of course, a remark may have to betaken in both ways, e.g., a doctor's order by anarmy doctor to a sick recruit.(e) As in the case of commands and rules, we donot find the word "obligation" in the conclusionsof practical arguments of this form. To deriveobligation-claims, we should therefore have tointroduce an additional premiss, such as (5) or (6)in argument (D) :(5) One has an obligation to do what one must,can, and should do in order to attain one'send.(6) One has an obligation to do what someone(soundly) advises one to do in order to attainone's end.220 AMERICAN PHILOSOPHICAL QUARTERLYPlainly it would be even harder to establish such ageneral proposition than it would be to establishthe corresponding (A-4), (B-6), or (C-6).21However, at least the words "must" and "ought"naturally occur in the conclusions of such argu?ments. It is therefore tempting to identify the con?clusions of such practical arguments with obliga?tion-claims. Thus von Wright says, "To show whysomething is an obligation founded on interest, isnot to show that it is something we ("really,""innermost") want to do, but that it is something wehave to do for the sake of that which we want (to be,to do, to have, to happen)."22 Von Wright thusimplies that when we have to do something for thesake of some end, we have some kind of obligationto do that thing. It is the fact that we have to dosomething for the sake of that which we want (tobe, to do, to have, to happen) which makes this anobligation; not, as Hume wrongly thought, thefact that there is something we really, innermostwant to do. However, this fact simply does notmake a line of action obligatory. It does not evennecessarily make it something we ought to do. AsHart points out,23 in this sort of situation we maybe obliged to do this thing, but we are not obligatedto do it.24To return to our argument (D), even the word"ought" in the conclusion introduces an importantnew element. For to say, in a context such as wehave examined, that Jones ought to stop eatingbread and salami, is to imply not merely that theaction recommended is the best way of attaining acertain possible end, but also that it is an endwhich the agent would be well advised to adopt.If the agent, realizing that unless he runs he willmiss the train, says to himself, "I must run," heleaves open the question of whether or not he iswell advised to pursue the end he has, namely, tocatch that train. If he says, "I ought to run," heimplies that he would be well-advised to pursuethat end.Thus, whereas the occurrence of words such as"can," "must" and even "should," does not implyanything about the advisability of the pursuit ofalternative courses of action to alternative ends,the occurrence of the word "ought" does have suchimplications. But even claims made by means of"ought" do not necessarily ascribe obligations. Themain task for Teleological Theories is therefore toprovide support for propositions such as (D) (5)and (6). I briefly examine three popular theoriesto expose the essential weaknesses of this model.(III-i) Rational Egoism, interpreted as atheory of obligation, maintains that one has anobligation to do whatever and only what one oughtto do, and that there is only one thing one oughtto do, namely, to pursue those ends the attainmentof which would be in one's best interest. Althoughthis theory has not been held by any great philos?opher, a brief discussion of it will help to makeclearer the strengths and weaknesses of the moresophisticated theories in this group. All theories ofthis group make two important moves. The first isto distinguish between those ends or goals which aperson finds himself having and those which heacquired as a result of deliberation. Standing backfrom our own past actions, noting the circum?stances under which we have come to have, andthen to pursue and sometimes to reach certain ends,we find that quite frequently we afterwardsregretted having pursued and reached some ofthese ends, because of their bad consequences orthe loss of the better things we might have hadinstead. Egoism recommends that we work out forourselves in the light of our knowledge, our predi?lections, preferences, likes and dislikes, our capa?cities, talents, energies, and skills, our opportunitiesand resources, a life plan whose realization wouldmake our life as rich and worthwhile as possible,and that we then plan the steps necessary for itsrealization. Of course, as we grow older and wiser,we may have to modify this plan in the light of ourchanged insights or the changed circumstances.But at any given time we should, with the aid ofsuch a plan, be able to judge not only what will bethe best means to an end we then find ourselves21 Cf. n. 3 above.22 Varieties of Goodness, op. cit., p. 170.23 Hart, The Concept of Law, op. cit., pp. 80-81.24 In his interesting book, Practical Reasoning (Oxford, Clarendon Press, 1963), chap. XII, David P. Gauthier, after showing insome detail, and in my opinion quite correctly, what it is to be obliged to do something then, wrongly, explains having anobligation as a special case of being obliged, namely, being obliged to someone by some "obliging factor" (p. 184). I suspectthat Gauthier was misled by a failure to distinguish between "being obliged by someone to do something," which is not a case ofhaving an obligation to do it, and "being obliged to somone for something" which is such a case. Because of this failure, he didnot see that not only is having an obligation not a special case of being obliged (by someone to do something), but beingobliged (to someone for something) is a very special case of having an obligation. Hence being obliged cannot be used to explainhaving an obligation.MORAL OBLIGATION 221having, but also whether the pursuit and attain?ment of such an end fits best into our life plan, orwhether instead we should refrain from pursuingthat end and pursue another more suitable endinstead. At any given time what we ought to do iswhat is the best way of attaining those ends, theattainment of which best realizes a life plan whoserealization will make our life as rewarding andmeaningful as possible.The second move is to offer an account of whatmakes a fact a reason for doing a certain thing. Inany system of practical reasoning it is importantthat we be told not only how to reason with whatfacts, but also why we should reason in just that way,which will often require us to resist inclination.Such a method of reasoning must have some appealto, or attraction for, the reasoner or else it will notbe used. Rational Egoism offers a very attractivemethod, so attractive that it is never challenged,but serves rather as a paradigm, perhaps the onlyparadigm of practical reasoning. On this model, thecriterion of a reason for doing something is thatfollowing it is in the agent's own best interest or forhis own greatest good. Thus, a fact, F, is properly areason for JV to do x, if N, on being apprized of Fand therefore doing x, is thereby promoting his bestinterest. Spelled out, this means, as we have seen,that in acting because of F, he is promoting an end,the attainment of which helps to realize a life planwhich will make his life as rewarding and meaning?ful as possible. The model thus provides an ex?tremely strong justification for this type of practicalreasoning, resting on the acknowledged superiorityof this mode of action over action on impulse. Itmay be virtually impossible to establish whether agiven fact is a good reason for doing something, inthis system, but it is hardly possible to challengethe framework. Insofar as we are willing to curbour impulses at all, are ready at all to refrain frompursuing ends we find ourselves having, we shallbe open to the consideration that the pursuit ofother goals would make for a richer, more reward?ing, more worthwhile life. We should regard aperson as worse than quaint if he seriously askedwhy he should do what is in his best interest.It must be granted that Rational Egoism cangive a plausible answer to questions of the form,"What ought I to do"? For it can answer notmerely questions about what is a possible, the onlypossible, or the best way of attaining an end someone happens to have, but also to the questionwhich of the many ends he might pursue wouldthe best one to pursue. When we say "You oughthave been there, it was unbelievable" or "yoought to invest in Fixed Trusts" or "You ougnot to take on any more speaking engagementswe may support these claims in just the wayRational Egoism indicates. Yet clearly these anot cases of someone's having an obligation. Forsomeone has an obligation to do something, thenfollows that it would be wrong for him not to do ibut that does not seem to follow from thesremarks.(III-2) Utilitarianism25 is similar to RationalEgoism, for it uses the same ingredients, but itserves them up in a different mixture. Act-Utili?tarians say that one ought to pursue those endswhose attainment would be for the greatest good ofthe greatest number. An individual, therefore, hasto consider not only his own life plan and what isnecessary to realize it, but also the life plan ofeveryone else. Reasoning under this sytem, thus,also requires him not to pursue those of his endswhose attainment, though for his own greatestgood, is not for the greatest good of the greatestnumber. Act Utilitarianism can claim to be muchcloser to our everyday moral convictions thanRational Egoism. We believe that moralityrequires us often to do things which are not in ourbest interest and to refrain from doing thingswhich would be in our best interest. RationalEgoism flies in the face of this deep-seated convic?tion. Egoism is moreover necessarily useless as away of deciding which of two people ought torefrain from pursuing an end which is his interest,in circumstances in which it is impossible for bothof them to attain their ends. In these cases, ActUtilitarianism can sometimes yield a decision.However, Act Utilitarianism has a much weakeranswer to the question why we should use itsmethod of reasoning. For to the question why theyshould prefer the greatest good of the greatestnumber to their own greatest good, Act Utili?tarians must reply either that people ought to aimat the greatest good of the greatest number or thatthey in fact do. The former falls wholly outside theirown conceptual framework and reduces Utilitari?anism to Intuitionism, i.e., to the absence of anytheory of obligation. The latter is often false, andso reduces the applicability of obligation-claims to25 Cf., e.g., J. J. C. Smart, An Outline of a System of Utilitarian Ethics (Melbourne, Melbourne University Press, 1961) ; R. B.Brandt, "Toward a Credible Form of Utilitarianism" in Morality and the Language of Conduct, ed. Hector Neri Casta?eda andGeorge Nakhnikian (Detroit, Wayne State University Press, 1963).D*222 AMERICAN PHILOSOPHICAL QUARTERLYthose who are psychologically so made that theydo in fact pursue these ends. Like Hart's "insiders,"Utilitarians must admit that those with a differentpsychological make-up need not, perhaps ought notto use its system of practical reasoning. On thisview, too, morality is the system of reasoning usedby and applicable to only those "insiders" who infact have the utilitarian end. On this theory, amorality is much like the code of some coterie orreligion.From our point of view, the only differencebetween Act and Rule Utilitarianism is that insteadof the pursuit of ends, the latter speaks of acting inaccordance with rules. Its formula for determiningone's obligations is: One ought to enter on thatcourse which, if made a general rule, would pro?mote at least as great a good of at least as great anumber as any other course open to him, if madeuniversal. This modification of Utilitarianism maywell yield results still closer to our every-day moralconvictions, but it is otherwise open to the sameobjections as Act-Utilitarianism.(III-3) The Basic Weakness of All Teleological Theories Well Formulated by Kant.As he points out, all such practical reasoning isbased on "hypothetical imperatives" and so lacksthe peculiar force of obligation-claims which arecategorical. Following Kant, one might try toexplain the force of obligation-claims as a certainproperty of some imperatives, namely, being ina certain sense unconditional, such that "we can?not be free from the precept if we give up thepurpose." The precepts which apply to a persondo so irrespective of whether or not he has theend to whose attainment the precept is a means.But though important, possession of this featureis only a necessary, not a sufficient condition.There are directives which satisfy it withoutbeing obligations, such as the doctor's order to ouradipose man. That order may well be shown to beapplicable to Jones under the conditions outlined ;its applicability is not conditional upon his in factwanting to reduce: he ought to have this endwhether or not he in fact has it; and it may besound. Jones cannot be "free from" the precept ifhe gives up the end. Nevertheless, this is not anobligation. For surely despite all this, Jones isentitled, has a right, to go on eating bread andsalami if he wants to. But he cannot have both anobligation to stop eating bread and salami and alsoa right to go on doing so if he wants to.We need to distinguish two types of universalapplicability and soundness : (i) where it is the caseeither in fact or of logical necessity, that all personshave the end, e.g., "You ought to do this if youvalue your happiness"; (ii) where the end inquestion is one which some persons sometimes donot have, but which any person always ought tohave, e.g., "You ought to do this if you value yourhealth." In case (i) a person would be free from theprecept by ceasing to have the purpose and soKant's unconditionally criterion rightly disposes inthis case. But in case (ii) he cannot be free byceasing to have the purpose and so that criterionby itself would, wrongly, uphold this case as anobligation.We can represent the argument (E) of such acrude "Kantian" theory in this way:(i) One has an obligation to act in accordancewith certain precepts, namely, those deline?ating actions which are means to ends oneought to have, whether or not one actuallyhas them;(2) that is, ends from which one cannot be freeeven if one could give up the purpose whichthey serve ;(3) that is, ends which are unconditional, and ina certain sense universally applicable andsound.(4) But health is such an end, "an ought throughand through."26(5) S? precepts delineating actions which aremeans to the preservation of one's healthgive rise to obligations to follow them.Now, a strong case can really be made for (4) :because there is no single end or combination ofends (except perhaps their totality) whose aban?donment would make the maintenance of one'shealth unnecessary or useless. There is thus a per?fectly good sense (though not the sense which Kanthas in mind), in which the maxim "preserve yourhealth" is always applicable and sound, hence un?conditional, hence categorical. All the same, the factthat an argument for doing what preserves one'shealth is conclusive, does not make doing itobligatory.IVOur brief review of some theories of obligationexposes their failure to explain and justify, in manycases even to attend to, the peculiar and seeminglyobnoxious binding force attaching to obligation26 To use Falk's useful term. Cf. W. D. Falk, "Morality, Self, and Others" in Morality and the Language of Conduct, ibMORAL OBLIGATION 223claims. This conceals the justifiability of challeng?ing those who make obligation claims until theycan demonstrate their bindingness. But thereceived theories cannot meet such challenges. Inanswer to the claim that a directive issuing fromsome Authorized Will could not amount to anobligation unless it were wrong not to follow direc?tives from such a source, the Will Theory can atmost say that such a claim is meaningless withinthe Will Theory. But this derives directives withseemingly obnoxious binding force from a theorywhich has deprived itself of all possibilities ofjustification. And much the same is true, mutatismutandis, for Teleological Theories. For a directivebased on a certain type of practical argumentcould not have moral binding force and so couldnot be obligatory unless it were wrong not tofollow directives arrived at by such a type of prac?tical argument. The recived theories distract atten?tion from these questions by stressing the emotiveappeal of such a method for generating obligationclaims ("its rationality," "its autonomy," "its dig?nity") or the appeal to individuals of the particularobligations generated by such methods (thatsuicide is wrong, idleness is wrong, etc.).How, then, can we avoid this epistemologicalcircularity? Clearly, the procedure must be this.We must first show wherein the peculiar bindingforce implied in obligation claims consists. We mustthen demonstrate the need for claims with such abinding force. And lastly we must indicate a methodfor generating directives with such a binding force.(IV?i) The Binding Force of ObligationClaims. As argued earlier, the binding force ofobligation claims is moral. We must now ask whatthis means. My answer is: "A directive has moralbinding force" means, "It is not solely the addres?see's business to decide whether or not to follow thedirective." What, on my theory, gives directivesmoral binding force is not that they are sufficientlyspecific to permit us to establish them conclusivelyor unconditionally; not even that when they con?flict with others, they are, or are rightly, regardedas overriding; it is, rather the fact that they con?cern themselves with issues and problems whosesolution is not solely the agent*s business but also thatof others who have a legitimate concern aboutwhether or not the person to whom such a directiveapplies follows it or not. When directives are of thissort it is justifiable for a society to take suitablemeasures to ensure that its members follow them.This seems to me the true kernel of the conviction,which runs through most versions of the WillTheory, that the binding force of obligation claimslies in the fact that the addressees of such direc?tives know that they are liable to incur the sanctionif they do not follow them.What is meant by the phrase, "not solely theagent's business to decide whether to follow thedirective" ? As children grow up, more and more ofthe things they have been directed to do at certaintimes and in certain ways are left to their own dis?cretion. Eventually, whether they follow such direc?tives becomes solely their business. At some stage,there comes an end to a mother's authority to seeto it that her son, using the appropriate system ofprudential and moral reasoning, keeps the rightcompany, drinks the right drinks in the right quan?tities, and regularly attends the right church. Allthe same, we do not think there is an end to every?one's authority in regard to all directives. In regardto some, e.g., the law, it is not solely even anadult's business to decide whether or not to followsuch directives.That it is not solely his business to decide onsome issue means that not all kinds of interferencewith his conduct in these matters are necessarily un?justifiable. Some may even be desirable. By "inter?ference" I mean pressures which are in them?selves obnoxious. I have in mind methods such ashandcuffing, jailing, fining, and possibly simply"condemning," which require justification. I donot mean perfectly legitimate pressures such asreasoning or pleading with a person, which amother may, perhaps should, exert even on anadult son.However, the importance of ensuring compliancewith certain directives may justify even interference.To show that some directive has moral bindingforce, it is not, therefore, enough to show that in agiven society a person's failure to comply withsuch a certain directive is followed by group inter?ference or that the group regards departure fromsuch directives as a signal for hostility, to use Hart'sphrase. If that is the case in a community, and if thecommunity regards such hostility as justified, per?haps its absence as undesirable, then such adirective is indeed regarded as having moral bindingforce. And since such a practice tends to increaseconformity, tends to act as a social pressure backingup the directive, we might say that, in a sense, sucha directive has moral binding force. All the same,this answer which might be quite adequate in asociological context, is not adequate in an epistemological one. The question we must answer iswhether it is rightly so regarded. But are we not now224 AMERICAN PHILOSOPHICAL QUARTERLYmoving in a circle ? Directives, we said, have moralbinding force if the question of whether to followthem is not solely the agent's business. It is notsolely the agent's business if it is permissible, per?haps desirable, for certain people to interfere forthe purpose of ensuring conformity with the direc?tive. But does not this mean simply that it ought tobe someone's job or duty to see to it that addresseesof such directives follow them ? And does not thismean that those, whose job it is to see to this, havean obligation to do these things, which means in turnthat they come under directives with moral bindingforce, which implies that someone else has anobligation to see that they follow these directives,and so on ad infinitum ?This looks like a vicious regress but it is not. Itsappearance of viciousness may well be one of thereasons why some philosophers feel that "moralbinding force" should be analyzed in terms of theactual social pressures operative in a community,or at most the community's actual beliefs of theappropriateness of such pressures, but never interms of the truth of such beliefs. The realizationthat there is nothing vicious about this regress mayhelp to make my own answer more acceptable. AsI construe the claim " 'Do x* has moral bindingforce," it implies that "Do #" is a directive inregard to which there ought to be a person whose jobit is to ensure that all those to whom addressees ofthe directive applies follow it. But this means onlythat it is desirable that there should be such a personwith such a job, if that were necessary to ensure thatthe directives are followed. And this may be truewhether or not there is such a person. As the line ofsuch supervisors lengthens, the likelihood increasesthat the remoter ones will carry out their duties ofensuring that those under them do so, withouthaving themselves to be supervised by yet furtherranks in the system. Thus, " 'Do *' has a moralbinding force" does not imply either that there isin fact a person who has such a job giving him theobligation to see to it that the appropriate personsdo x, or that a certain person has an obligation tosee to it that someone is appointed to such a job, letalone that everyone morally ought to take this jobupon themselves. And so there is no vicious regress.(IV?2) The Justification of Moral Direc?tives. If my account of directives with moral bind?ing force is correct, then they are regarded as otherpeople's business, and so as licensing interference.Their existence thus narrows what a person is freeto do, and so it is in need of justification.Here is a sketch of such a justification. My argument takes a number of things for granted. Itassumes, for instance, that a morality is an opera?tive system of practical reasoning, that is, a systemof general end-setting directives which the membersof the group are taught in their childhood as partof the conventional wisdom of the group. Systemsof practical reasoning differ from each other inrespect to the way in which such directives can besupported. The best-understood system is that ofself-interest. In that system, a general directive,such as "Protect your health," would be supportedby showing that they are ways of attaining states ofaffairs (such as being able to do things one wantsto do and to enjoy life) which are necessary condi?tions of the good life for the person in question.Such self-interested directives are capable of cominginto conflict with a person's inclination. For a per?son to be able to follow such directives he must notonly be taught what they are but must be trainedin childhood to follow them even when they gocounter to his inclinations. Such training willrequire certain forms of pedagogic interferenceduring childhood. This we regard as justified onaccount of the great benefits which such trainingbestows on the individual, and because we know ofno other way of enabling him to derive suchbenefits. Of course, since the benefits so derivedcome to him, the question of whether he followssuch directives is solely his business, hence seeingto it that he later follows them is not justifiable.The main distinction to bear in mind is thatbetween the proof of, or support for, the directivesof self-interest, and the need (inherent in any formof practical reasoning) and the justification for in?terfering with a person's inclinations and desires.The first tells us what such directives are, and whythey should be followed. The second tells us theextent to which they should be inculcated and en?forced.A system of general moral directives can bejustified along similar lines, though we must bearin mind the differences as well as the similarities.The main similarities are first that the support forgeneral moral directives presumably must lie in thefact that following them leads to certain desirablestates of affairs, and secondly that, since they mustbe capable of coming into conflict with inclinationsand desires, there is a need to train the young sothat they know what such directives are, and areable to follow them when they conflict with inclina?tion and desire. The main differences are thefollowing: Moral directives must be regarded ascapable of overriding not only inclination but alsoMORAL OBLIGATION 225the directives of self-interest when a person cannotfollow both, hence it will be difficult to train theyoung to follow moral directives. This explains,though it does not by itself justify, the practice ofembodying some of our most important moraldirectives (e.g., "Thou shalt not kill") in our legalsystem and supporting them by severe and sup?posedly effective sanctions. We do, however,believe in the justifiability of this practice of seeingto it that people follow moral directives even whenthey are grown up and have learned what is rightand what is wrong. And we support this belief bysaying that whether or not a person follows moraldirectives is not solely his business but otherpeople's as well. And if this is true, then theseothers are entitled to see to it that people followmoral directives and, within limits, to take measureswhich are necessary to achieve this purpose.We have brought to light an impoitant differ?ence between self-interest and moral directives.Both imply that the practice of teaching thesedirectives, and using pedagogic molding techniquesto make it possible for the young to follow them ifthey want to, can be justified by the improvedlives of those involved. But there the similarityends. For moral directives must be such as to berightly regarded as overriding not only inclinationbut also self-interest; the question of whether or notsomeone follows them must not be solely hisbusiness ; and so the social practice of seeing to itthat everyone follows them must be justified(beyond the training period). Lastly, whereas weknow quite well how we support or prove, and sohow we formulate, the directives of self-interest,it is more difficult to do this for moral directives.Hence in the case of self-interested directives wecan spell out why it is that they are rightlyregarded as overriding inclination, and why societyis justified in inculcating in the young an appro?priate attitude toward these directives during theirtraining but not later, and why it is that the mat?ter of whether or not a person follows them issolely his business and no one else's, hence why noone has what are often misleadingly called "obli?gations to himself."27 But in the case of moraldirectives it is not always easy to spell this out. Whatthen is the formula for constructing moral direc?tives? We already have the clues we need. We canderive it from the premiss that moral directivesmust have a content such as to yield two things:(a) support for saying that they are rightlyregarded as overriding inclination and self-interest,and (b) support for saying that whether or notsomeone follows them is not solely his business.Let us try to derive such a formula. We caneliminate the solitary desert islander, for in his casewhether he follows these directives must be solelyhis business. In fact, the situation we must en?visage is precisely that envisaged by Hobbes: agroup of people following inclination except whenself-interest conflicts with it, and having a will ofthe requisite quality to follow reason rather thaninclination. Such people are then confronted bythe question of whether these principles of actionare adequate for the good life or not. Hobbesmakes an excellent case for saying that they are not.It is based on the following sort of argument.(i) Human needs, wants, and aspirations can bebetter satisfied under conditions of proximity,specialization, and cooperation, than in isolation.(ii) However, the scarcity of goods and the result?ing conflicts of interest, as well as the fear generatedby the justified belief that others will follow aggres?sive inclinations, and the mutually conflictingdirectives of self-interest, tend to lead to harmfuland wasteful expenditures of resources and an un?bearable climate of life, (iii) If there were availablefor guidance a set of directives regarded as over?riding the directives of self-interest and applicableon those occasions when following the latter wouldlead to such harmful and wasteful behavior, theclimate of life and so life for everyone would beimproved, (iv) However, if it is true, as is widelybelieved, that people tend not to do what theythink is contrary to their interest, they will betempted, even after training, not to follow thedirectives of morality but those of self-interest whenthe two conflict, (v) But since the behavior of aperson who yields to the temptation to follow selfinterest and to ignore moral directives will, ipsofacto, detrimentally affect another person's interests,the question of whether or not he follows moraldirectives is ipso facto not solely his business butsomeone else's as well, namely, the business of theperson whose interest would be adversely affected.And since such behavior, unless prevented, wouldadversely affect the climate of life, whether or notpeople follow moral directives, is everyone's business.27 Elsewhere I have dismissed this view much too cavalierly. There I was able to show merely that it is impossible to enterwith oneself into the sort of temporary moral relationship into which one can enter with another. But this impossibility is perfectlycompatible with one's having an obligation to do a certain thing simply because it is the best thing for one. On my present view,one cannot have such an obligation because it is no one else's business whether or not one does what is best for oneself.226 AMERICAN PHILOSOPHICAL QUARTERLY(IV?3) The Moral "Calculus." In section(IV?2) we gave the outline of an argument show?ing the need, for a group of people following eitherinclination and/or a system of self-interested direc?tives, to have a set of directives overriding selfinterested ones, in regard to which the decision ofwhether or not to follow them is not solely thebusiness of those to whom they apply. This outlinejustification shows that what creates the need forsuch an overriding system is the fact that followingself-interested directives will often lead to harmfuland wasteful conflict and, if this occurs generally,to a climate of life which fully deserves the title"cold war." These considerations yield up themissing item, the calculus we must employ toarrive at directives satisfying the criteria for beingmoral. Such a formula must spell out how wedetermine the content of the directive and thecircumstances under which it applies.It will be remembered that such directives mustspell out what is to be done in circumstances whentwo people, following directives of self-interest,could not both attain their end and would be driveninto mutually harmful or wasteful efforts to attaintheir own end while preventing the other fromattaining his, and where such efforts would beundesirable, at any rate if prosecuted "with noholds barred." The content must indicate whichone of the interests is to give way or what com?promise is to be made.To understand the peculiarity of moral rules, wemust however bear in mind a second requirement.Where interests conflict, there are many possibleregulations dealing with the conflict. The directiveembodying the regulation would not be properlymoral (as opposed to being legal or conventional)unless it purported to be the best possible way ofregulating such a conflict. Hence, other thingsbeing equal, general moral directives are open tocritical scrutiny and to modification as social con?ditions change and our knowledge of consequencesincreases. The rationale for this second requirementis, of course, that since moral directives from thenature of the case will override someone's concernevery time they apply, everyone must have (as faras possible) the same good reason for accepting sucha general regulation.I offer a simple example : A and B are interestedin renting apartments. Both have looked at a greatmany of them, but have liked a certain apartmentbest. A was the first to inspect it and to pay adeposit. B saw it shortly afterwards, and to secureit, offers to pay a higher rent. The moral rule hereis that the giving of a deposit binds the landlord andso excludes other prospective tenants. All concernedhave the same good reasons for accepting this solu?tion as always overriding. For although on thisoccasion, the rule excludes B and thus goes against2?'s interest, it is still in Z?'s interest just as much asin A9s that there should be such a rule, which willprotect him on future occasions when he is the firstto find a suitable apartment. Hence allocatingobligations is in the end justified by the benefiteveryone receives from this practice.In this essay I have advanced the following majortheses :( i ) All obligation claims are subclasses of generaldirectives with moral binding force, and so are anintegral part of a morality even though some, e.g.,promissory or legal obligations, assign tasks which,but for being thus assigned, would not be moraltasks.(2) The binding force which is characteristic ofsuch moral directives can be characterized by say?ing that the question of whether to follow suchdirectives or not is not solely the business of those towhom they apply.(3) This characterization of the binding force ofsuch moral directives provides the answers to threeimportant ethical questions :(a) Why is it desirable that certain general direc?tives should be universally taught and stringentlyenforced? The answer is that on their beinggenerally followed depends the general climate oflife which is the springboard from which an indi?vidual can, in accordance with his abilities andtastes, build a worthwhile life for himself; andbecause in the absence of such enforcement thelikelihood of their not being followed would bevery great.(b) What is the proper subject-matter for generalmoral directives? The answer is: the general andauthoritative adjudication between conflictingtypes of individual interests and concerns.(c) How can the content of such directives be cor?rectly formulated? The over-all principles onwhich such formulae are constructed is that itshould be the best solution, i.e., the one whichprovides for each of those whose concerns areaffected, as far as possible, an equally good reasonfor accepting this adjudication between conflictingconcerns.