Few issues in jurisprudence have received so much attention in recent years as whether citizens have a distinctive moral duty to obey the law. Yet the differences among the disputants might well seem slender to the unprofessional eye. No one holds that the duty is absolute: even its most passionate advocates allow that it is sometimes morally permissible to disobey the law,! as when abolitionists aided runaway slaves before the American Civil War. But neither does anyone advocate open or frequent disobedience. Those who doubt the supposed duty yet hold that we very often have a strong moral reason to do what the law requires independently of its commands, e.g., not to assault, cheat or rob others. The doubters allow that we are obligated to obey whenever the law has established patterns of conduct that are dangerous to depart from, such as driving to the left in Great Britain. They believe that disobedience is permissible only when there is no independent moral reason to obey or when the weight of independent reasons favors disobedience; and they do not suppose that in reasonably just societies these conditions obtain often. Finally, those who are sceptical about the duty of obedience nonetheless prize the great social benefits that quite obviously can only be achieved through government; and they believe that one which is reasonably just deserve! s its citizens' cooperation and support. It therefore seems probable that the putative duty's advocates and disbelievers alike would virtually always agree in their judgments of particular illegal conduct--or at least that any differences between them would not flow from their disagreement about the philosophical issue.

Despite the debate's apparent lack of practical significance, many philosophers and academic lawyers yet disagree hotly about whether there is a "prima facie" duty to obey the law. I shall later conclude that their controversy is primarily metaethical rather than political, being fueled by disagreement about the very point of positing prima facie duties. Let us look first at some particular arguments.

Philosophical worries about the precise contours of the duty citizens owe to the state date back at least to Plato. But the particular claim that there is a pr! ima facie duty to obey the law was first voiced in 1930, by the great British classicist and moral intuitionist, W. D. Ross, in The Right and the Good.

Twentieth century intuitionism is part of an older family of metaethical theories that ascribe to humanity a common moral faculty. (Older siblings include the moral sentiment theories of Hutcheson or Hume and the moral rationalisms of Aquinas or Richard Price.) Ross devised the distinction between prima facie duty and absolute duty in the hope of solving the problem bedeviling all such theories: viz., that of setting out the principles that explain the moral faculty's deliverances about particular kinds of morally relevant factual circumstances. Ross's intuitionistic contemporaries generally agreed that Sidgwick had shown the futility of any attempt to frame exceptionless general principles of rightness or wrongness--that every promising candidate will turn out either to be inconsistent with our firm intutions ab! out examples or else be a disguised tautology such as "Murder is wrong" or "Justice is giving every man his due." But Ross thought it possible to frame absolute principles of what he called "prima facie" rightness or wrongness. His distinction is often explicated by the practice of promising: it is obvious to everyone that it is sometimes permissible to break promises; but it is also plausible to suppose that this is wrong "other things being equal"--by which is meant some such notion as "wrong unless justified by one's thereby fulfilling some moral consideration of equal or greater weight." (See Thomson 1990, ch. 12, for the best intuitionistic account of the moral constraint of promises.) Ross offered a list of our separate prima facie duties, comprising inter alia those of fidelity, gratitude, beneficence and non-maleficence. Somewhat as an afterthought, he suggested that there is also a duty to obey the law:

Ross did not repeat these arguments in his later, longer book, The Foundations of Ethics (1938). His suggestion found little attention until 1955, when it was taken up by H. L. A. Hart in his seminal article, "Are There Any Natural Rights." Hart offered a fresh argument in support of the duty of obedience, based upon his formulation of what has come to be known as the principle of fair play:

John Rawls refined Hart's argument in a series of influential articles. By 1964 Ross's suggestion had become a philosophical commonplace, so much so that Rawls could confidently assert:

This consensus was broken in 1973, when an alternative position began to be developed by a number of philosophers, including the present author. (See inter alia: Smith 1973; Sartorious 1975; Raz 1979; Simmons 1979; Woozley 1979; Greenawalt 1987).! I shall now sketch what I take to be a broad area of agreement among those who reject the duty of obedience (but the reader is cautioned that I cannot include many necessary qualifications and subtleties). Firstly, most of us acknowledge the existence of legitimate authority; but, unlike many political theorists, we analyze the concept of authority without reference to a duty of obedience. For instance, in my early paper on the topic, I offered a perhaps overly simple definition, which counts a government as possessing legitimate authority when it has a moral right (in the sense of "that which is morally permissible") to coerce its citizens' obedience. Kent Greenawalt has refined this definition by adding further conditions, e.g., that if a govenment has legitimate authority its citizens are virtually always obligated not to interfere with enforcement of its commands. (Greenawalt 1987, p. 55). But he, too, finds it unnecessary to posit a prima facie duty of obedience to every! law in order to account for governmental authority.

Secondly, we doubters have in various ways criticised the sundry arguments that have been offered in support of the duty to obey the law. We do not launch frontal assaults upon them; rather we attempt to show that, when their underlying principles are properly understood, it is evident that the factual situation of most citizens fails to trigger their conditions of application. Thus, we do not deny that a genuine, voluntary consent to obey every law would found a general duty to obey; rather we follow Hume in denying that most citizens of any nationality have performed acts which constitute such consent. (1948 [1777]). Similarly, after analyzing the scope of the obligations of fair play, gratitude, rule and act utilitarianism, we have concluded that these principles do not reach every situation in which the law requires us to act. (The best comprehensive account of these arguments is in Greenawalt 1987.)

Lastly, we doubters have pointed out that the supposed general obligation to obey the law must be redundant in every normative theo! ry that has any plausibility whatsoever. That is because every such theory must contain proscriptions against assault, reckless endangerment, fraud, breach of serious promises, etc., whose conjunction arguably specifies each important moral interest that we are bound to respect. Hence, regardless of whether a theory provides specifically for obedience to the law, it will imply trivially that there is a prima facie duty to obey whenever disobedience puts an important moral interest at risk. Moreover, we contend, when we do have an obligation of obedience its weight is exhausted by the collective weight of the independent moral reasons that point in the same direction. Our argument is broadly speaking intuitionistic, being primarily based upon our considered moral judgments about particularly described examples. We point out that no one supposes that mere illegality gives rise to moral concern: no one condemns the prudent driver who slowly and safely runs through a lengthy stop ! light at an empty rural intersection at two in the morning. Nor does anyone suppose that illegality worsens what is independently wrong: no one would say that the practice of husbands raping their wives has recently been made more reprehensible by having belatedly been made illegal. (Here we echo Blackstone: "Neither do divine or natural duties . . . receive any stronger sanction from being also declared to be duties by the law of the land" (1793, p. 54).) Hence, we conclude, the supposed general obligation of obedience plays no useful explanatory role in normative theory, and so there is no good reason to accept it.

Many philosophers have not been not convinced by these arguments. Virtually no one seems to believe that any of the classic arguments from rule or act utilitarianism, gratitude, consent or fair play, can be refurbished so as to yield a general prima facie duty of obedience. (But see Klosko 1987, and Walker 1988.) Instead, fresh arguments have been o! ffered. (Finnis 1980; Honore 1981; Mackie 1981; Dworkin 1986.) What has not been noticed is that there is a metaethics supposed in these arguments that is vastly different from Ross's intuitionism--one so different as to raise doubt that they affix the same meaning as did he to "prima facie duty." Summed briefly, the difference is this: Most who favor the prima facie duty of obedience conceive of normative theory as catechistic and perhaps even as political. They believe that philosophers ought to aim at formulating a set of principles that the rest of humanity might accept and articulately employ in arriving at their considered moral judgments. However, Ross had no such ambition at all. He spoke primarily to other philosophers, and he did not expect to find catechumens there or in the public at large. Rather than preaching at humanity, he assumed that we all have a faculty that permits us to discern moral truth. He thought that the proper task of normative theory is to explai! n the principles that the moral faculty employs. These different metaethical visions yield different conclusions about whether there is a prima facie obligation to obey the law. Hence, what appears to be a dispute about politics is in reality a dispute about the proper end of normative theory. It is no wonder then that the disputants differ so slenderly over particular instances of illegal conduct.

Apart from their common project of framing moral principles for general adoption, catechistic philosophers are a diverse metaethical lot. They comprise the moral realist, John Finnis, an avowed defender of Thomistic natural law theory, who attempts to delimit the obligations of humankind by speculating about which principles would best promote the common good were ordinary people to accept and to act upon them. (Finnis 1980, esp. pp. 303- 308). The rule-utilitarian prescriptiv! ist, Richard Hare, is also in their company. (Hare 1981). And they include the moral sceptic, John Mackie, whose argument for the obligation of obedience shall be our exemplar from the class. (Mackie 1981).

The first sentence of Mackie's book, Ethics: Inventing Right and Wrong, boldly proclaimed "There are no objective [moral] values" (1977, p. 15). But he nonetheless offered arguments in normative theory, whose point he thought is to invent a morality that will allow humanity to florish peacefully were it generally accepted. (Id., p. 193) It need not be created wholly new: all moralities contain restrictions upon the free use of violence, theft, promise-breaking, etc.; and it is evidently necessary that the one to be recommended to society must have some such content if it is to promote human welfare. Still, we cannot suppose that existing moralities adequately serve the goal of human florishing: e.g., ". . . some more traditional obligations traditionally atta! ched to status, not created by contract, are dispensible; patriotism . . . may have outlived its usefulness" (Id., p. 123). Mackie's conception of normative theory requires philosophers always to determine the consequences of general belief in a large variety of alternative moral principles, and to settle upon that set whose acceptance would best make us florish. (Once a list is complete one supposes that its adherents would then try to persuade us all to adopt their revised morality. However, Mackie does not even address how they ought set about this further task.)

What then about the duty of obedience? In an article devoted to the topic, Mackie began by announcing without a shred of argument or evidence that "the dominant conventional morality of our present society" recognizes its existence (1981, p. 144). However, he placed no weight upon this, but focused instead upon whether:

He returned an affirmative answer, on this ground:

If we accept Mackie's methodological presuppositions, it is hard to demur from his conclusion, given the obvious practical empirical truth that general obedience to law is essential to human well being. The objections that I set out earlier against the obligation of obedience have no! force against him. He is unfazed by the fact that the obligation can't be derived from other moral principles, such as gratitude or fair play, because he contends that it follows directly from the goal of human florishing. Neither is it a telling argument against him that the duty plays no useful explanatory role in normative theory, because he rejects the supposition that there is any moral reality for philosophers to explain. Nor will he be impressed by the suggestion that the obligation does not fit our intuitions about particular kinds of cases, because he rejects the intuitionistic metaethics that lies behind this style of argument. Lastly, he will not care that the duty is redundant in any colorable normative theory, because he does not aim at explanatory economy. He may respond that redundancy is even a virtue in normative theory, because it is desirable that conduct tending to promote human flourishing be morally overdetermined. Since obedience to the law promotes thi! s end, it is well that ordinary people accept more than one principle that will inspire them so to behave. Greenawalt criticises Mackie's argument primarily on the ground that the available evidence does not show that recognition of an obligation of obedience is necessary to sustain adequate compliance with the law. (Greenawalt 1987, pp. 179-85.) But this too seems weak: if recognition of the duty will do no harm--Greenawalt does not allege that it will--, and if it may do some good by reinforcing citizens’ tendencies to abide by the law, it seems that we ought to place it in our catechism of principles, even if we do not know that its recognition is essential to achieving the ameliorating aim of normative theory. Given Mackie's catechistic metaethics, he is clearly right to endorse a prima facie duty of obedience to law. Let us now consider how the duty fares from a Rossian point of view.

It is often said that metaethics and normative ethics are wholly independent of one another. That commonplace captures one important truth: philosophers' moral sympathies do vary independently of their metaethics; the entire political spectrum can be found in philosophers of every stripe. But the claim of independence hides the more important truth that a philosopher's metaethics sets constraints upon the form of the principles her normative theory can recognize and upon the arguments that may be deployed in their support.

Since it is impossible to argue for all one's premises, every philosophical theory must rest in the end upon intuition--upon the theorist's unsupported bare beliefs, which she hopes her readers will share. But the role that intuition plays differs greatly among metaethical theories. Catechistic philosophers typically suppose that an acceptable normative theorymust be one satisfactory to philosophy, and so presuppose t! hat only philosophers' intuitions (perhaps restricted to those of "logic" or "language") have authority for normative theory. (C.f., Hare 1981, §§ 1.3, 1.6). Ross thought that philosophers' substantive moral intuitions are authoritative, but only because he supposed that everyone's are. (Ross 1939, p. 1f.). Like Judith Jarvis Thomson, our most celebrated contemporary intuitionist, Ross was a metaethical rationalist: he believed that fundamen- tal moral principles are necessary truths and that virtually all humanity has a faculty of reason by which it can recognize such truths in favorable circumstances. As did many great philosophers before him (e.g., Aristotle, Cicero, Aquinas, Hume, Adam Smith, Kant), Ross supposed that the moral faculty is fungible. Let us call this assumption "commonalism." Let us also adopt this model of the common moral faculty: whether Reason, Sentiment or something else again, it is a mentalistic "black box" into which nonmoral beliefs are fe! d as stimulii and from which moral conclusions issue. (A psychologically accurate model must posit reciprocal causal relations between nonmoral and moral belief. But since normative theory concerns the conditions of passage from premises of empirical fact to moral conclusions, the one-way model will suffice for our purposes.)

Since commonalists presume that the deliverances of the moral faculty are true and that everyone has one, they also suppose that the proper task of normative theory is to explore and explain "common sense" morality, but is emphatically not to change it. (E.g., Kant held that ordinary folk can perfectly well discover their duties without aid of philosophers, and that philosophy's only practical office is to explain the basis of ordinary practical reason in necessary truth, which may help those who understand this to be less tempted to follow specious arguments and inclinations contrary to duty. (Kant 1964 [1785], p. 71-73).) Hence, setting v! ery different goals for normative theory, commonalism implies constraints upon its principles and arguments that are very different from those of catechistic metaethics.

Firstly, commonalists suppose that the primary constraint upon normative theory is that it adequately explain those of our own and others' moral intuitions that are made in circumstances conducive to reliability. Roughly, these are: that we are not prey to false relevant nonmoral belief (e.g., about whether Sally hit John); and that we are not relevantly subject to influences likely to corrupt judgment (e.g, family feeling, racial prejudice). Commonalist methodology, which relies heavily upon our intuitions about hypothetical (often fantastical) examples, produces intuitions that are maximally reliable. That is because our personal interests are then disengaged and we are immunized from nonmoral error, since the facts upon which the moral faculty works are stipulated. (But we must note as a cave! at that philosophers' theories sometimes badly skew their intuitions. E.g., Gilbert Harman believes that morality is merely a matter of group convention, which prompted him to the spectacularly counterintuitive claim that "it would be a misuse of language to say of hardened professional criminals that it is morally wrong of them to steal from others or that they ought morally not to kill people" (1977, p. 113.)).

Secondly, because commonalists assume that we have somehow all acquired an inchoate knowledge of a common morality, they do not suppose that normative principles are subject to learnability constraints or that they must be easily understood by the vulgar. (Those troubled about how anyone could recognize moral truth but not be able to understand a summation of its principles should reflect upon the familiar fact that few--if any--have a complete articulate understanding of the principles of English grammar but many have inchoately mastered them. (C.f., S! mith 1979)). And indeed, the normative theories offered by leading commonalists, such as Hume and Kant, or more lately Ross and Thomson, are exceedingly abstract and fearsomely complex. In contrast, the principles of catechistic theories are constrained by what not-too-terribly-well-educated people can be expected to apply. Catechistic principles must therefore be aphoristic, so that they can effectively be inscribed upon and kept before the minds even of the dull. (C.f., Hare 1981, § 2.4). (Arguments in support of catechistic principles may of course be as complicated as one pleases, since they are addressed primarily to philosophers.)

Thirdly, unlike catechists, commonalists should never argue for moral principles by appealing to supposed favorable consequences of ordinary persons accepting and acting upon them. For again, they hold that ordinary citizens' considered moral judgments are as reliable as any philosopher's and have equal authority for normati! ve theory. Hence, commonalists (in professional moments) ought not even attempt to prescribe political principles for nonphilosophers to follow. (Indeed, given the complexity of their theories, they would be naive to expect wide readership among nonphilosophers.) They should instead attempt to discern what principles we would settle upon in ideal conditions of judgment: i.e., if we knew all relevant nonmoral facts (but philosophers should be diffident about whether their training gives them competence to speak authoritatively upon matters of empirical fact), and if we were reasoning consistently, were uninfluenced by invidious bias, and had adequate time for reflection and consultation with others engaged in the same difficult task. In consequence, commonalists must forswear arguing for a general obligation of obedience from any such premise as Mackie's, viz., that humanity would better florish were we all schooled in a catechism containing this principle of duty. The principl! es of the commonalist enterprise imply that speculation about the consequences of humanity's general articulate acceptance of this (or any other moral principle) has no probative force in normative theory.

Commonalists will therefore rightly reject the catechists' principal argument for a distinctive prima facie duty of obedience. Nor is there any other argument towards this conclusion that their metaethical principles can accord any weight. To the contrary, as was noted above, the appeal to our moral intuitions shows that the existence and weight of a moral reason to obey the law is always a function of some independent moral reason which argues in this direction. And that same appeal shows also that we sometimes have no moral reason whatever to do what the law requires, as when one outstays the parking meter by five minutes at a time when there is ample available parking. Constraints of explanatory adequacy and economy, to which catechists are indifferent, oug! ht decisively incline commonalists to exclude the duty of obedience from their normative theories.

It appears then that whether a normative theorist ought recognize a distinctive duty of obedience depends upon her metaethics: catechists should; commonalists should not. Hence, the question can receive no decisive answer except in the context of a full-blown metaethical theory. It is a mistake to suppose, as I once did, that it may be addressed discretely. (Smith 1973).

Nonetheless, since I believe that some version of commonalism is true and that catechistic theories are wrongheaded, I hold to my earlier doubts about the duty of obedience. I cannot here defend commonalism. (But see, Smith, 1979 and Smith, 1992). Still, I can offer these summary criticisms of catechistic theories as a promissary note to be paid in full on some other occasion: Firstly, although many philosophers have thought that ! their professional office is to correct society's erroneous moral beliefs, virtually everyone else has ignored their schooling. It is evident upon brief reflection that anyone who sets herself this goal dooms herself to frustration. Philosophy is difficult reading, and few have either time or inclination to wade through it. Most nonphilosophers who do--virtually all of them academics in other disciplines--doubt that a philosophical education confers any special store of wisdom. But if nonphilosophers won't defer or even listen to us, what is the point of our trying to correct their opinions?

Secondly, despite the interest and subtlety of various catechists' arguments, there is scant intrinsic plausibility to the supposition that ordinary people require philosophers' aid to discover what they ought individually and collectively to do. When voiced by philosophers it seems suspiciously self-aggrandizing. But its primary implausibility lies in its making normative e! thics into something wholly anamolous. The historically dominant conception of analytical philosophy is that its task is explanatory, not hortatory or prescriptive. Its "philosophy of" branches generally (e.g., language, law, science, mathematics, etc.) do not undertake to tell speakers, judges, scientists, mathematicians, etc., how they ought behave in these roles. Rather philosophers attempt to formulate explanatory theories about what speakers, etc., actually do and how they do it, and about the nature of the various entities with which they work (e.g., meanings, laws, scientific theories, numbers, etc.). No one supposes that native speakers of English need study philosophy of language to speak it correctly, that mathematical or scientific proofs would be improved were mathematicians and scientists all proficient in the "philosophies of" their respective disciplines, or that constitutional decisions would be more just were Supreme Court Justices required to be schooled in a! nalytical jurisprudence. Why would normative ethics break this pattern? In default of a convincing answer to this question we should reject all catechistic metaethics out of hand--and with them the supposed distinctive duty of obedience.

M. B. E. Smith

Smith College

Nothampton, MA 01063

Blackstone, W., Commentaries on the Laws of England, 12th ed. (London: Strahan & Woodfall, 1793).

Devlin, P., The Enforcement of Morals (New York: Oxford U. Pr., 1963).

Dworkin, R., Law’s Empire (Cambridge, MA: Harvard U. Pr., 1986).

Finnis, J., Natural Law and Natural Right (New York: Oxford U. Pr., 1980).

Greenawalt, K., Conflicts of Law and Morality (New York: Oxford U. Pr., 1987).

Hare, R., Moral Thinking (New York: Oxford U. Pr. 1981).

Harman, G., The Nature of Morality, (New York: Oxford U. Pr. 1977).

Hart, H., ! `Are there any natural rights,' Philosophical Review 64 (1955) 175- 191.

Hume, D., `Of the original contract,' [1777], H. Aiken, ed., Hume's Moral and Political Philosophy (Darien: Hafner, 1948), 356-372.

Honore, T., `Must we obey? Necessity as a ground of obedience,' Virginia Law Review, 67 (1981), 39-61.

Kant, I., (1785) Groundwork of the Metaphysic of Morals; trans. H. J. Paton (New York: Harper & Row, 1964).

Klosko, G., `The principle of fairness,' Ethics 97 (1987), 353-362.

Mackie, J. L., Ethics: Inventing Right and Wrong (Harmondsworth: Penguin Books, 1977).

Mackie, J. L., `Obligations to obey the law,' Virginia Law Review, 76 (1981), 143-158.

Rawls, J., `Legal obligations and the duty of fair play,' S. Hook ed., Law and Philosophy (New York: N.Y.U. Pr., 1963), 3-18.

Raz, J., `The obligation to obey the law,' The Authority of Law (New York: Oxford U. Pr., 1979) 233-249.

Ross, W. The Right and the Good (Ox! ford: Oxford U. Pr. 1930).

Ross, W., The Foundations of Ethics (Oxford: Oxford U. Pr. 1938).

Sartorius, R., Individual Conduct and Social Norms (Encino: Dickinsen, 1975)

Simmons, A., Moral Principles and Political Obligations (Princeton: Princeton U. Pr. 1979).

Smith, M., `Is there a prima facie obligation to obey the law?,' Yale L. J. 82 (1973), 950-76.

Smith M., `Ethical intuitionism and naturalism: a reconciliation,' Canadian Journal of Philosophy 9 (1979), 609-29.

Smith, M., `The best intuitionistic theory yet! A review of J. Thomson, The Realm of Rights,' Criminal Justice Ethics (Summer/fall 1992), 85-97.

Thomson, J., The Realm of Rights (Cambridge, MA.: Harvard U. Pr. 1990)

Walker, A., `Political obligation and the argument from gratitude,' Philosophy &. Public Affairs 17 (1988), 191-211.

Woozley, A., Law and Obedience (Chapel Hill: U. North Carolina Pr. 1979).