- Chapter 1 - Natural Law Theory Mark C. Murphy Natural law theorists claim that, necessarily, law is a rational standard for conduct: it is a standard that agents have strong, even decisive, reasons to comply with. This is the central thesis from which their developed theory of law takes its starting point. My aim here is to make clear how we might understand natural law theory's central thesis, how it can be deployed in a fruitful theory of law (see can there be a theory of law?) and why one might take it to be true. I will proceed by first examining briefly the way that this thesis surfaces in the work of Thomas Aquinas, the paradigmatic natural law theorist: aside from providing a salutary glimpse of the history of natural law theorizing, this will help us to see in Aquinas's work the ambiguities and tensions that form the problematic of recent natural law thought. I will then proceed analytically, examining some of the various formulations that the natural law thesis might take, considering the extent to which each of these formulations is incompatible with the legal positivism (see legal positivism) with which natural law theory is typically contrasted, and asking what sorts of arguments can be offered for the natural law thesis. It will no doubt be wondered why a thesis that concerns a connection between law and reasons for action bears the seemingly uninforma-tive label "natural law theory." It bears this label because the most historically important defender of this central thesis is Thomas Aquinas, and Aquinas identified the principles of rational conduct for human beings as the principles of the natural law. Thus, given Aquinas's theory of reasons for action, the thesis in question can be stated as asserting a connection between human law and natural law. A danger with this label, of course, is that one might confuse theses of Aquinas's theory of practical rationality with theses of his theory of law, and take objections to one of these theories to constitute objections to the other. A different label might have been better at describing the view at the level of abstraction that we will treat it. But the label "natural law theory" has stuck, and I will not attempt to detach it here. Aquinas's Theory of Natural Law Brian Bix has remarked that it is, in general, a bad idea to read texts on law from the distant past with the assumption that the concerns of the authors of those texts are the concerns of contemporary analytical jurisprudence (Bix 1996:227). Bix, and others, have suggested that this is particularly true of Aquinas: Aquinas, they write, was not interested in providing a descriptive theory of the nature of human law; he was, rather, concerned to provide a theory of political obligation (see legal and moral obligation), that is, an account of the source and limits of the moral requirement to comply with the demands of law (see, for example, Bix 2002:63; Soper 1983:1181). Aquinas was concerned with the problem of political obligation. But that does not mean that he was not also concerned to provide a correct description of what law essentially is. Here is a helpful comparison. In the work in 15 Natural Law Theory which Aquinas's mature thoughts on law are to be found, the Summa Theologiae,1 the set of questions that is labeled by commentators the "Treatise on Law" (ST Iallae 90-107) is preceded by sets of questions labeled by commentators the "Treatise on Virtue" (ST Iallae 55-70) and the "Treatise on Vice" (ST Iallae 71-89). These considerations of virtue and vice are primarily descriptive - they are meant to provide an account of the concept, nature, and causes of virtue and vice. This is a speculative, not a practical, enterprise, however much one may draw upon Aquinas's answers here to get a better grip on (for example) how tasks of moral education ought to be carried out. So just because Aquinas later draws practical conclusions about the law and the requirement of obedience to it, that does not mean that he was not trying to come up with a theory of (human) law - an account of law that is both necessarily true and which provides an explanation of it (see can there be a theory of law?). That Aquinas is indeed concerned with the task of providing an adequate descriptive theory of law is clear when we examine the structure of his argument to the conclusion that human law is a rational standard for conduct. This conclusion is a straightforward implication of his view that all law is a rational standard for conduct. The thesis that all law is a rational standard for conduct is defended in the first article of the first question of the Treatise on Law (ST Iallae 90, 1) and it is a thesis that applies not only to human law but to the (for the most part unknowable) eternal law, that law by which God exercises providence over all creation (ST Iallae 91, 1). No practical issues are being addressed and no such issue has even been raised. Only later does Aquinas make the further argument that human law is capable of binding in conscience (ST Iallae 96,4), and only much later does Aquinas provide a full account of obedience to authority, including political authority (ST Ilallae 104-105). There is little reason to follow Bix and others in holding that Aquinas's theory of law is primarily a theory of the obligation to obey it. To return, then, to the natural law theorist's central thesis, and Aquinas's defense of it. Why does Aquinas think that all law is a rational standard for conduct? Law is a sort of rule and measure of acts, according to which one is induced to act or restrained from acting, for lex (law) is said to be from ligare (to bind) because obligat (it binds) one to act. But the rule and measure of human acts is reason, which is the first principle of human acts,... for it belongs to reason to order things to the end, which is the first principle in practical matters, according to the Philosopher [that is, Aristode]. However, that which is the principle of any given genus is the measure and rule of that genus, like unity in the genus of number and the first motion in the genus of motion. Hence it follows that law is something pertaining to reason. (ST Iallae 90,1) Though this argument is couched in unfamiliar terms, its gist is, I think, plain enough. Aquinas's idea is that, no matter what else we think about law, we agree that it consists in rules, mandatory standards by which our conduct is to be assessed. Furthermore, the sort of assessment involved is essentially practical: the standard that law sets is a standard by which one is "induced to act or restrained from acting." But the only standards that can induce rational beings to act, qua rational beings, are rational standards. So law necessarily is a rational standard for conduct. Aquinas's full, famous definition of law is that law "is nothing other than [1] an ordinance of reason [2] for the common good, [3] issued by one who has care of the community, and [4] promulgated" (ST Iallae 90,4). The latter three elements of this definition are subordinate to the first element, in that Aquinas employs the claim that law is an ordinance of reason to show that law is for the common good, issued by one who has care of the community, and promulgated. Why does Aquinas write that law must be for the common good? Because law is a rule not concerning an individual qua individual, but for the governance of group conduct; and just as what determines reasonable conduct of an individual is that individual's good, what determines reasonable conduct for members of a group is the common good of that group (ST Iallae 90, 2). Why does Aquinas write that law can be made only by one who has care of the community? Because while anyone can make suggestions about how it is reasonable to order group conduct, only one who is charged with making such 16 Mark C. Murphy determinations can render an authoritative ruling on what is to be done, thereby setting the standard that members of that group must follow (ST Iallae 90, 3). Why does Aquinas write that law must be promulgated? Because rational beings cannot act on a rational standard as such unless they have the means to become aware of the existence of the standard, its status as authoritative, and its content, and the promulgation of the rule provides for this awareness (ST Iallae 90,4). The essential character of both the nonpositive and the positive elements of law are explained through the master thesis that law is a rational standard for conduct (cf. Finnis 1996: 205). Aquinas is aiming at descriptive adequacy in providing his theory of law. It is not a statement that, or of the conditions under which, people are obligated to obey the law, but an account of what law is: a rational standard. But it turns out that the content of this descriptively adequate statement of what law is entails that one must draw on one's normative views, whether theorized or not, in order to provide a fuller, more descriptively adequate account of law (cf. Finnis 1980:16). From the fact that law necessarily is a rational standard for conduct it follows that (in some sense, to be explored further in the following section) a rule that cannot be a rational standard for conduct for beings like us cannot be law for beings like us. But which rules can be rational standards for conduct for beings like us is something that cannot be grasped without drawing on one's normative views. Here is an analogy. Suppose that I attempt to build a "reason-backed rule" machine. When a person pulls the handle of the machine, the machine is supposed to display on its screen a rule, in the handle-puller's language, that the handle-puller has strong reason to comply with. Now, it is an accurate description of the machine to say that its function is to exhibit rules that those who operate it have strong reason to comply with. But if one is going to provide a fuller account of when the machine is functioning as it is designed to function and when it is not, one is going to have to draw on one's views, theorized or not, about what one has strong reasons to do. If you pull the handle on the machine and it displays "you should give one-third of your income to Oxfam," you will not be able to say whether the machine is functioning as designed unless you can say whether one in circumstances such as yours has reasons to make sacrifices of this sort. The same holds true with law, on Aquinas's natural law view. In offering further claims on the nature of law, Aquinas draws upon a wide range of his normative beliefs, some already defended in the Summa, some later to be defended in the Summa, some undefended in the Summa but assumed in virtue of the context of the work. (The Summa is a teaching tool for those training for the religious life: Aquinas's intended primary audience consists of those who share his Christian commitments, some of which are normative commitments.) These normative beliefs concern, in part, what agents have reason to do. Aquinas relies on this stock of claims about reasons for action in defending more specific theses about what the essential features of law are. Here is one example of how Aquinas draws on claims about reasons for action in drawing more specific conclusions about the nature of law. Aquinas holds that there is a "natural law," consisting of the fundamental principles of practical rationality, which govern all human conduct, individual and collective (ST Iallae 91,2; 94,1-6). (It is important to keep in mind that there are natural law moral theories and natural law Legal theories. The two are logically separable: one can affirm either one while rejecting the other. For a quick overview of natural law moral theory, see Murphy 2001: 1-3, and Murphy 2002.) All reasons for action are rooted in the natural law. Thus one of the conclusions that Aquinas can reach, given the abstract connection between human law and reasons for action previously established, is that all human law is rooted in the natural law (ST Iallae 95,2). This does not mean, Aquinas emphasizes, that all human law simply reproduces the contours of natural practical rationality (ST Iallae 96, 2-3); while some of it does (for example, laws against murder, rape, etc.), some of it goes beyond the natural law by fixing, by making determinate, the vague requirements of the natural law (for example, "drive no more than 65 miles per hour" determines the vague "when driving, proceed at a reasonable speed") (ST Iallae 95,2). So Aquinas is clear that the human law is not just a mirror that reflects in whole or in part the 17 Natural Law Theory demands of the natural law. The view that the natural law theorist holds that cp-ing's being independently morally required is necessary for cp-ing's being legally required (or, even worse, necessary and sufficient for cp-ing's being legally required) is a common caricature of natural law theory, but Aquinas's emphasis on the way that human law can make determinate what the principles of practical rationality leaves indeterminate shows that he does not hold that view. Finnis has also responded to this caricature explicitly (see his 1980: 28), but it continues to be attributed to the natural law view. Consider, for example, the following argument from Jules Coleman and Brian Leiter: [According to natural law theory,] in order to be law, a norm must be required by morality. Morality has authority, in the sense that the fact that a norm is a requirement of morality gives agents a (perhaps overriding) reason to comply with it. If morality has authority, and legal norms are necessarily moral, then law has authority too. This argument for the authority of law, however, is actually fatal to it, because it makes law's authority redundant on morality's____Natural law theory, then, fails to account for the authority of law. (Coleman and Leiter 1996: 244) This argument assumes the premise that the natural law theorist claims that cp-ing's being independently morally required is necessary for cp-ing's being legally required. But Aquinas rejects this thesis, as does the natural law jurisprudential tradition generally. Here is another example of how Aquinas draws on independent theses about reasons for action in drawing specific conclusions about law. Aquinas accepts as a matter of Christian moral orthodoxy, and later argues in philosophical/theological terms, that there are some moral absolutes, norms that it is unreasonable for one ever to violate (ST Ilallae 33, 2). There can never be adequate reason to kill innocent persons (ST Ilallae 64, 6), or to blaspheme (ST Ilallae 13, 2). But it follows, given the connection between law and reasons for action, that a rule, promulgated by one who has care of the community, that requires one to kill the innocent or to blaspheme would fall outside the definition of law that Aquinas offers (ST Iallae 96,4). Aquinas's natural law account of human law, influential as it has been in defining the natural law program, is marked by ambiguity and unclar-ity at its core. First off, how are we to understand the claim that law is a rational standard for conduct? Does it follow from this claim that - as many critics of natural law theory have supposed - wicked or unjust or otherwise unreasonable rules cannot be valid law? And if it does not follow from the natural law thesis that unreasonable rules cannot be valid law, what on earth does it mean to claim that, necessarily, law is a rational standard for conduct? Second: against Aquinas, it seems as if there are plenty of systems of rules that in some way apply to rational beings for which there does not exist an internal connection between those standards and reasons for action. Consider, for example, rules of games, or certain outdated codes of honor; it does not seem that it is essential to these systems of rules that there be sufficient reason for rational beings to comply with them. Why, then, should we think that this particular kind of system of rules, a legal system, exhibits this internal connection between law and reasons for action? Can we get an adequate account of the warrant for claiming that there is indeed this internal connection? The Meaning of the Natural Law Thesis How should we understand Aquinas's natural law thesis? In asking this question, I am not primarily asking how we ought to interpret Aquinas's texts, but rather what is the best way to formulate the connection between law and reasons for action that Aquinas and natural law theorists that followed him were impressed by. The formula that we are to understand is: necessarily, law is a rational standard for conduct. The most straightforward understanding of this thesis - and the understanding that was fixed upon by critics of the natural law view in order to expose it as absurd - is an understanding on which necessarily, law is a rational standard for conduct is a proposition of the same form as necessarily, a square has four and only four sides. Just as 18 Mark C. Murphy a figure with five sides simply is not a square, this strong reading of the natural law thesis - I will henceforth call it the Strong Reading, or the Strong Natural Law Thesis - holds that a rule that is not a rational standard for conduct is no law at all. Legality is strictly limited by rationality: lex sine rationem non est lex. Why do I formulate the Strong Reading as lex sine rationem non est lex (that is, law without [adequate] reason is no law at all) rather than as the better known natural law slogan, lex iniusta non est lex (an unjust law is no law at all)? The latter is sometimes attributed to Augustine, sometimes to Aquinas, but as Kretzmann notes, that precise formulation occurs in neither Augustine's nor Aquinas's work (Kretzmann 1988: 100-1). It continues to be common to formulate the natural law thesis in terms of a connection between law and justice, or between law and morality more generally. I have chosen to formulate the view in terms of the connection between law and reasons for action because it is clear that the tradition of the natural law theorizing connects law with practical rationality generally, and that same tradition has treated a failure with respect to justice as simply one way that a purported law can fail to be backed by decisive reasons for compliance. It is of course controversial to characterize injustice as simply a species of rational failure, but it is uncontroversial that this is how Aquinas saw it (ST Ilallae 58, 4), and it is because Aquinas saw unjust action as rationally defective action that he was willing to affirm claims very like "lex iniusta non est lex." The Strong Reading of the natural law thesis is the usual target of positivist criticism. As John Austin wrote, To say that human laws which conflict with the Divine law are not binding, that is to say, are not laws, is to talk stark nonsense. The most pernicious laws, and therefore those which are most opposed to the will of God, have been and are continually enforced as laws by judicial tribunals. (Austin [1832] 1995, Lecture V: 158) Presumably Austin would say the same about the formulation of the natural law thesis under consideration here: to say that laws inadequately backed by reasons for compliance are not laws is to talk stark nonsense. Again, and much more recently, here is Brian Bix on the Strong Natural Law Thesis: The basic point is that the concept of "legal validity" is closely tied to what is recognized as binding in a given society and what the state enforces, and it seems fairly clear that there are plenty of societies where immoral laws are recognized as binding and enforced. Someone might answer that these immoral laws are not really legally valid, and the officials are making a mistake when they treat the rules as if they were legally valid. However, this is just to play games with words, and confusing games at that. "Legal validity" is the term we use to refer to whatever'is conventionally recognized as binding; to say that all the officials could be wrong about what is legally valid is close to nonsense. (Bix 2002: 72-3) Even self-labeled natural law theorists have endorsed objections of these sorts. John Finnis, whose work has clearly been at the forefront of the revival of natural law theory in the late twentieth century, has written that the Strong Reading is "pure nonsense, flatly self-contradictory" (Finnis 1980: 364); and Robert George has remarked that the fact that Aquinas was perfectly willing to talk about unjust laws shows that the paradigmatic natural law position does not affirm the Strong Reading, for to affirm the Strong Reading while being willing to refer to "unjust law" would be inconsistent (George 2000: 1641). There are two distinguishable criticisms here. One of these is the "self-contradiction" criticism: the Strong Natural Law Thesis either is internally inconsistent or is inconsistent with other claims that natural law theorists are willing to affirm. The other is the "officials' say-so" criticism: the Strong Natural Law Thesis is inconsistent with the practice of legal officials. How serious are these criticisms for the Strong Reading? The "self-contradiction" criticism is far from decisive. It need not be stark nonsense to affirm claims of the form "a_X is no X at all." David Lyons has noted that "counterfeit dollars are no dollars at all" is simply true (Lyons 1984: 62). One might also add that "glass diamonds are no diamonds at all" is simply true. The cases in which "a_X is no X at all" makes perfect sense are 19 Natural Law Theory those in which the blank is filled with an alienans, a certain class of adjective (Geach 1956: 33—4). "Fake" is always an alienans: fake Rolexes are not Rolexes, fake dog doo is not dog doo, fake flowers are not flowers. "Counterfeit" is always an alienans as well. But there are some adjectives that count as instances of the alienans only with respect to particular nouns: while "glass" is obviously not always an alienans (glass sculptures are sculptures), it can be (glass diamonds are not diamonds). The strong natural law theorist can hold that "unable to serve as a rational standard" is, when applied to "law," an alienans, and thus avoid the charge that the Strong Reading is incoherent. (See also Kretzmann 1988.) The "officials' say-so" objection is also far from decisive. Bix's claim is that since the consensus of legal officials is that there are laws that are inadequately backed by reasons for compliance, it would be flying in the face of the word of experts and indeed courting incoherence to assert the contrary. But it can hardly be a criterion for the truth of a legal theory that it make impossible divergence between official say-so and the implications of that theory. On Austin's general jurisprudence, every law is a command, issued by a sovereign and backed by a sanction (Austin [1832] 1995, Lecture 1:21). A sanction is a credible threat of harm to a subject attendant on a violation of the order (Austin [1832] 1995, Lecture 1:22). It follows from Austin's view that there is no law that is not backed by a sanction. But, possibly, all of the legal officials in some society might hold that some particular norm, a norm unbacked by a sanction, is law. If Austin's view is true, law without sanction is no law at all. Thus Austinian positivism violates Bix's constraint. Even on a more sophisticated view like Hart's, Bix's constraint is violated. On Hart's general jurisprudence, whether something is law in a given society depends on whether it is recognized as such by the rule of recognition, the usually tremendously complex rule that guides legal officials in making, identifying, and applying law (Hart [1961] 1994: 94-5). It follows from Hart's view that there is no law that is not acknowledged as such by the rule of recognition. But, possibly, all of the legal officials in some society might hold that some particular norm, a norm not acknowledged by the rule of recogni- tion, is law. (The rule of recognition might hold that if norm N was part of the originally adopted constitution, then it is law; but they might all hold a false view about whether norm N was part of the originally adopted constitution.) If Hart's view is true, law unacknowledged by the rule of recognition is no law at all. Thus Hartian positivism violates Bix's constraint. Now, one might say that the actual(zs opposed to the merely possible) practice of legal officials is not at odds with the Austinian or Hartian view. While I have extremely strong doubts about the former and strong doubts about the latter, we can note, first of all, that the actual convergence is not enough to rescue the incoherence claim: whether the view is incoherent cannot depend on contingent facts. The incoherence charge could be revised, even in the face of this sort of contingency, by holding that law is conventional and that therefore to deny that officials' say-so is dispositive is to assert an incoherent view. But that law is conventional is a substantial claim, indeed the substantial claim that the natural law theorist is concerned to deny (or, better, to qualify). While it is often hard to tell when a claim that a rebuttal is question-begging is warranted, this would seem to be one of those warranted cases: any appeal to the status oflaw as conventional to rescue the claim that the officials' say-so argument is decisive would beg the question against the natural law theorist. One can, of course, still make the point that the say-so of legal officials is not to be gainsaid in a theory of law. But that is a much weaker point, as much weaker as an appeal to authority is weaker than a reductio ad absurdum. Suppose though, that one continues to be suspicious of the Strong Reading of the natural law thesis, noting that officials' say-so seems to run contrary to the view. What alternative formulations of the view are available? One alternative is that suggested by those who would hold that the primary concern of Aquinas in the Treatise on Law is to provide an account of political obligation: on this view, the claim that law is necessarily a rational standard is a disguised normative claim. On this formulation, what we may call the Moral Reading of the natural law thesis, the natural law theorist's central thesis is just a dramatic way of saying that one ought to obey the law only when it is adequately reasonable. As George proposes, 20 Mark C. Murphy "What is being asserted by natural law theorists [is]... that the moral obligatoriness which may attach to positive law is conditional in nature" (George 1996: viii). All that the natural law theorist wants to do in affirming a connection between law and reasons is to issue a reminder that adherence to some laws would constitute such a departure from reasonableness that there could not be adequate reason to obey them; the only law that merits our obedience is law that meets a certain minimum standard of reasonableness. Whatever the intrinsic merits of this claim, I will immediately put it to the side as a candidate formulation of the natural law thesis. If the Moral Reading were all there is to the natural law thesis, the natural law theorist would have almost no one to disagree with in the entire history of philosophy. And whatever other desiderata a formulation of the natural law thesis must satisfy, a candidate formulation must be one that preserves the status of natural law theory as a contentious position. There is, however, a contentious natural law position that is nevertheless not prone to some of the initial deep misgivings to which the strong formulation is prone. Recall that the strong formulation is to be understood in such a way that necessarily, law is a rational standard for conduct is a proposition of the same form as necessarily, squares have four and only four sides. A weaker but still interesting version of the natural law thesis -call it the Weak Reading, or the Weak Natural Law Thesis - affirms that necessarily, law is a rational standard while holding that it is not of the same form as necessarily, squares have four and only four sides; rather, it is of the same form as necessarily, cheetahs are fast runners. A figure with only three sides is no square at all; but it is not true that an animal that is not a fast runner cannot be a cheetah. Rather, an animal that is not a fast runner either is not a cheetah or is a defective cheetah. The necessity attaches to the kind cheetah rather than to individual cheetahs: while one might fail to be a fast runner while remaining a cheetah, belonging to the kind cheetah sets a standard such that those that are not fast runners fall short as cheetahs (cf. Thompson 1995 and Foot 2001: 30; Robert Alexy makes such a distinction, which he labels a distinction between "classifica-tory" and "qualificatory" connections between properties and kind-membership, and employs it with respect to theories of law in Alexy 1998:214 and 1999: 24-5). This seems to be the approach taken by John Finnis, the most influential contemporary defender of natural law theory. Finnis roundly rejects the natural law thesis in its stronger formulation, labeling the Strong Reading paradoxical and inconsistent and incoherent and self-contradictory (Finnis 1980: 364—5). But he affirms the Weak Reading. According to Finnis, regardless of whether one is inclined to take a natural law view in jurisprudence, it is a mistake to look for necessary and sufficient conditions for legality (Finnis 1980: 6,9-11). Rather, we ought to proceed by looking for the conditions that define the central, paradigmatic case of legality (Finnis 1980: 9-11). On Finnis's view, the paradigmatic case of legality is the rule or norm that is not only socially grounded but also grounded in a correct understanding of what reasons for action agents have. While there may be laws that are unreasonable for agents to follow, Finnis allows, these laws are laws only in a secondary, derivative, incomplete sense. Their status as laws is parasitic on the primary, fundamental, complete sense of law, that notion of law on which laws bind rational agents to compliance (Finnis 1980: 14). (I say more on Finnis's argument for this view in the final section below.) It might be supposed that the Weak Reading of the natural law thesis just is the Moral Reading that I set to the side as trivial. If this were the case, it would surely be damaging to defenders of the Weak Reading. But these readings are not identical. The defender of the Weak Reading wants to make a claim about what counts as a defect'm law -and the conditions under which some objectionable (or even otherwise unobjectionable) aspect of a thing counts as a defect in it are very specific, tied to the kind of thing at stake. It is, after all, a commonplace that a feature of some object can be objectionable without that feature's being a defect in that thing. The flourishing of the rodent in my attic is objectionable, all right, but I wouldn't presume to claim that its flourishing makes that rodent defective. Similarly, all the defender of the Moral Reading can say is that there is some way in which unreasonable laws are objectionable; the Moral Reading of the natural law thesis does not itself make the further claim that 21 Natural Law Theory law that is not rationally binding is defective as law. Thus the defender of the Weak Reading has an extra argumentative burden, that of showing that law is the kind of thing that is backed by decisive reasons, so that an individual law unbacked by decisive reasons is substandard. We should also note that while the distinction between the Strong and Weak Natural Law Theses - between a view on which reasons for action are connected to the legal validity of a norm, and a view on which those reasons are connected to legal nondefectiveness - is very important, and thus the distinction that I will focus on for the remainder of this chapter, it is not the only relevant distinction one could make. One could distinguish among natural law theories on the basis of the strength or sort of reasons for action to which legal validity or legal nondefectiveness is allegedly connected. For example: while I have focused on how we ought to understand claims like lex sine rationem non est lex - does it mean that unreasonable laws really lack validity, or does it mean that while such can be legally valid, they are in some way defective as law or perversions of law? - one might also focus on the nature and extent of the departure from reasonableness involved. Assuming for a moment the Strong Reading, one might ask, that is, whether any unreasonableness in law is sufficient to undermine legal validity, or whether perhaps the unreasonableness must reach some extreme pitch before legal validity is precluded. Thus, for example, Gustav Radbruch's famous formula is about legal validity but kicks in only in cases of severe injustice: on Radbruch's view, enactments the injustice of which are at "an intolerable level" have "no claim at all to legal status" (Radbruch 1946, cited in Alexy 1999: 16; see also continental perspectives on natural law theory and legal positivism). He would surely have said, though, that any level of injustice in law makes it legally defective, even if not necessarily legally invalid. Natural Law Theory and Legal Positivism Legal positivism has defined itself by setting itself in contrast with natural law theory. This is as true of Austin's and Bentham's positivist views as it is of Hart's and Raz's. But we have seen that the natural law view - like the positivist view - admits of a variety of formulations. To what extent is the opposition between natural law theory and legal positivism a real opposition? Suppose that we take as the generic legal positivist thesis the view that the status of some social rule as law is logically and metaphysically independent of the status of that social rule as a rational standard of conduct. This is close to what Austin had in mind when he delivered his path-breaking lectures on jurisprudence: "The existence of law is one thing; its merit or demerit is another" (Austin [1832] 1995, Lecture V: 157). It is close to the thesis that Hart defends, in contrast to the natural law position, in "Positivism and the Separation of Law and Morals": "In the absence of an expressed constitutional or legal provision, it could not follow from the mere fact that a rule violated standards of morality that it was not a rule of law" (Hart [1958] 1983: 55). It is close to Coleman's Separation Thesis, which on his view defines the positivist outlook: "There exists at least one conceivable rule of recognition (and therefore one possible legal system) that does not specify truth as a moral principle among the truth conditions for any proposition of law" (Coleman 1982: 141). It is, with proper qualifications, entailed by Raz's Sources Thesis: "A jurisprudential theory is acceptable only if its tests for identifying the content of the law and determining its existence depend exclusively on facts of human behaviour capable of being described in value-neutral terms, and applied without recourse to moral argument" (Raz 1979d: 39^0). If we take this to be the generic positivist position, it is obvious that there is no incompatibility between the Moral Reading of the natural law thesis and the positivist standpoint. The positiv-ists, after all, were concerned to defend their position on the nature of law not merely for the sake of conceptual clarity but also for reasons of moral psychology: by demystifying law, one will be less likely to obey simply because it is the law and more likely to obey only when there is adequate reason to do so. (For a critique of this line of argument for positivism, see Soper 1987.) This is entirely consistent with, and even complemen- 22 Mark C. Murphy tary to, the Moral Reading's insistence that law is to be obeyed only when it falls within the domain of the reasonable. (As I mentioned above, it is the overwhelming plausibility of the Moral Reading that is its undoing: it is so plausible it is uninteresting and nondistinctive.) On the other hand, the Strong Reading of the natural law thesis is just as clearly incompatible with generic positivism. For the positivist wants at least to take a stand on legal validity: social rules can be legally valid though there be far from adequate reason to comply with them. Austin and Bentham took as their primary targets Blackstone's seeming affirmation of the Strong Natural Law Thesis, and it is the seeming affirmation of versions of the Strong Natural Law Thesis by Radbruch and by Fuller that Hart took as his primary target. So the strongest version of natural law theory is necessarily at odds with positivist views. With respect to the Weak Reading, matters are less clear. One is tempted to say that the Weak Natural Law Thesis, according to which any law either is an adequate rational standard for conduct or is defective, is entirely compatible with the positivist thesis. For the Weak Reading does not deny that there can be valid law that only an unreasonable person would comply with. It says only that, be it valid, it nevertheless falls short of some standard internality to legality. This is a view endorsed both by Finnis and by MacCormick (who describes himself as a positivist; MacCormick 1992: 108). But while positivists have been willing to allow that their views require them to employ evaluative judgments in providing their theory of law (for example, judgments about what phenomena are more important than others in the categorization of human institutions), some may be less sanguine about the notion that the provision of an adequate theory of law requires one to take a stand on highly disputable and disputed questions of practical reasonableness. Thus the Weak Natural Law Thesis might well be taken to be a departure from the positivist program, even if it is compatible with the most influential formulations - for example, Austin's, Bentham's, Hart's, Raz's, and Coleman's - of first-order positivist theses. Defending the Natural Law Thesis On the basis of what sorts of arguments can the natural law view be defended? I will put to the side the Moral Reading of the natural law thesis: it is too uninteresting and uncontroversial to bother with. It is the Strong and Weak Readings, both of which aim to provide an account of the nature of law, that are of interest here. I will proceed by pursuing two argumentative strategies: the 'legal point of view' argument, initially defended by Finnis in his 1980 Natural Law and Natural Rights and continually reaffirmed by him since then, and the "function" argument, defended (with reservations) by Michael Moore in a couple of recent papers. Finnis's argument for the natural law thesis is inspired by Hart's methodology in The Concept of Law (see legal positivism). We should not, Finnis writes, hope to provide an account of the necessary and sufficient conditions for law, such that some legal systems and individual norms and decisions in cases will count as law through exemplifying these conditions, whereas the remainder will not. Rather, we should hope for an account that provides us with the central, paradigm instances of law and legality. With this account, we will be able to classify some social systems and social norms as clearly law, some as entirely extralegal, and some as simply falling short of or distinct from the central case in one or another specific way. So the task of the legal theorist is to provide the central case of law. But centrality is an evaluative notion, and this particular evaluative notion is always from a point of view. The question, then, is whether there is a point of view that is privileged within legal theory. Again Finnis, following Hart, holds that there is such a privileged point of view: it is the point of view of those who take the internal point of view with respect to a legal system. People who take the internal point of view with respect to a legal system are those who take its rules as such to be a guide to their conduct. Hart emphasizes that he does not mean to privilege any particular motivation or rationale for taking the internal point of view: those that 23 Natural Law Theory treat the law as a standard for conduct based on moral considerations and those that treat it as such based on "calculations of long-term interest" or "disinterested interest in others" or "an unreflecting inherited or traditional attitude" or "the mere wish to do as others do" all equally take the internal point of view (Hart [1961] 1994: 203). But Finnis argues that the internal point of view, as characterized by Hart, is not sufficiently differentiated for analytical purposes. The argument is by elimination: none of these species of the internal point of view, save the point of view of the person who obeys the law because it is a matter of moral requirement, can be the privileged legal point of view, for "All these considerations and attitudes are manifestly deviant, diluted or watered-down instances of the practical viewpoint that brings law into being as a significantly differentiated type of social order and maintains it as such" (Finnis 1980: 14). The central legal viewpoint is that in which legal systems are seen as morally worthy, worth bringing about and preserving, and in which the demands of law are justified and binding; and indeed the clearest case of this central viewpoint is that of the person whose moral judgment is correct (Finnis 1980: 15-16). Given this most privileged point of view, it is clear that law in its central or "focal" meaning will be law that is a rational standard for conduct. This strategy is meant only to establish the Weak Natural Law Thesis, and it is obvious that it can establish no more than that: its appeal to the central, paradigmatic notion of law is not meant to preclude the presence of a limited, technical sense of legal validity, a sense explicable without reference to moral or practical considerations. But it is hard to see why we would follow Finnis even this far in his extension of Hart's methodology on the basis of this argument. Hart has good reason for taking the burden of proof to be on those who wish to make some particular version of the internal point of view more privileged. For while his arguments against the legal realists show that legal theory must account for the datum that people can take the internal point of view with respect to a system of legal norms (Hart [1961] 1994: 88-91), this datum just is that people treat the existence of legal rules as reasons or constituent parts of reasons for action. The datum does not, however, extend further to the basis on which they so treat those norms. Far from the internal point of view just being an "amalgam" of different viewpoints, Hart's undifferentiated take has a clear rationale, and so is not unstable; it is up to Finnis to destabilize it. But nothing he says in the crucial stretch of argument discussed above succeeds in destabilizing it. The law tends not to care a whit for the motivations that one has for complying with it; and while Finnis appeals to the great efficacy of some points of view in generating a legal system, one might rightly retort both that the tasks of explaining how a legal system comes into being and explaining what it is for a legal system to be in place are, while interestingly related, different questions and that there are some points of view that may have greater efficacy in generating and sustaining a legal system than that of the person of full practical reasonableness - for example, that of the person who holds a false tribal or nationalistic morality. By so closely identifying the task of characterizing law with the task of saying what a fully practical reasonable person should be interested in when dealing with the law, Finnis's view seems to become simply applied ethics - he is asking what features of the law the fully reasonable citizen, or the fully reasonable judge, should be interested in responding to, and in particular what features of the law are such, when present, for the fully reasonable citizen or judge to treat the law as authoritative. But this seems to make Finnis's view too much like the uninteresting Moral Reading, leaving his critics to wonder what all the fuss was about natural law theory (Bix 1996: 226). A more promising line of argument, to my mind, takes as its starting point the common notion of function. According to this line of argument, once we see that some legal systems or individual legal norms have functions, and see what those functions are, we should recognize that those systems and norms have nondefective-ness conditions that include the presence of reasons for action. One might worry that this sort of argument for the natural law thesis is doomed to triviality: what could be easier, one might ask, than to assign a morally charged function to law, and then, on the basis of such an 24 Mark C. Murphy ascription, hold that law that does not perform this function, or perform it satisfactorily, is either no law at all or is law only defectively? It is obvious that no interesting argument for the natural law thesis that proceeds from the idea that the law has a function can follow this pattern. But the ascription of a function to an object is a much more constrained matter than such an argument would suggest. I cannot simply assign the function "keeping New Haven populated" to law professors, and then declare that law school faculty that do not reside in New Haven are no law professors at all, or are law professors only defectively. What are the conditions that must be met to ascribe a function to some object or institution, and how can these be brought to bear to show that one or another formulation of the natural law thesis is correct? Roughly, and not at all originally, and not entirely uncontroversially, we can say that for an object or institution x to have the function of cp-ing, the following conditions must be satisfied: (characteristic activity) x is the kind of thing that cp-s {goalproductivity) x's cp-ing tends to bring about some end-state S (teleology) x cp-s because x's cp-ing tends to bring about some end-state S (value) S exhibits some relevant variety of goodness. There is reason to think that each of these conditions is individually necessary; and there is reason to think that they are jointly sufficient. A heart has a characteristic activity: it pumps. Its pumping tends to bring about the circulation of the blood; and, indeed, the heart pumps because its pumping contributes to the circulation of the blood. (This is so in two ways: in animals with hearts there is a feedback loop such that the circulation of the blood is in part what causes the heart to be able to continue pumping; and the very structure and activity of the heart was selected because of efficiency in causing the circulation of the blood.) Some would take these first three conditions to be jointly sufficient, but it seems to me that it is also important that the circulation of the blood is beneficial for the animal. As Mark Bedau has noted, a stick pinned against a rock in a stream by the backwash that very stick has created may exhibit the first three features: it is pinned against a rock, its being pinned against a rock causes the backwash, and it is pinned against the rock because its being pinned against the rock causes the backwash. But no one would be tempted by the view that it is the stick's function to be pinned against the rock (Bedau 1992: 786). One way to accommodate such cases is to emphasize that functions are ascribed when there is, in some sense, a good realized through the activity: either an end sought out by the designer of the object, or simply the self-maintenance of the thing in question, or the like. To show, then, that the natural law thesis is true in virtue of the law's function (or one of the law's functions) one needs to show that these various conditions are satisfied, and that a particular legal system or law fails to perform its function when it fails to serve as a rational standard for conduct. An instance of this strategy is the argument offered by Moore. Moore suggests that the essence of law might reside in its function rather than in any distinctive set of structures. To find out what law's function is, we look at the sorts of cases that we pretheoretically label instances of legal systems and laws, and try to identify some distinctive good that they serve; we can then use that tentative identification of a distinctive end served by law to identify other instances of laws and legal systems. If it turns out that there is some good distinctively served by law, and that law can serve this good only if those under that law are practically required to comply with it, then we have reason for thinking that the natural law thesis is true. Indeed, Moore suggests that this argument, if successful, would be sufficient to establish the Strong Natural Law Thesis (see Moore 1992, 2001). Moore worries about whether there is any distinctive end that law serves: he doubts that there is any such distinctive end - though he notes some candidates, such as John Finnis's notion of the common good - and thinks that if there is no such distinctive end, then we must give up on the idea that law is to be understood in terms of its function. But this is too hasty. For recall that the ascription of a function to some thing brings into play not just the goal brought about by the thing's activity (S) but the characteristic activity 25 Natural Law Theory of that thing (. Radbruch, Gustav. 1946. Gesetzliches Unrecht und ubergesetzliches Recht. In Gustav Radbruch, Rechtsphilosophie, 8th edn., ed. Erik Wolf and Hans-Peter Schneider. Stuttgart: Koehler, 339-50. Raz, Joseph. 1979a. The Authority of Law. Oxford: Oxford University Press. Raz, Joseph. 1979b. The claims of law. In Joseph Raz, The Authority of Law. Oxford: Oxford University Press, 28-33. Raz, Joseph. 1979c. The obligation to obey the law. In Joseph Raz, The Authority of Law. Oxford: Oxford University Press, 233—49. Raz, Joseph. 1979d. The social thesis. In Joseph Raz, The Authority of Law. Oxford: Oxford University Press, 37-52. Raz, Joseph. 1985. Authority, law, and morality. Monist 68: 295-324. Raz, Joseph. 1986. The Morality of Freedom. Oxford: Oxford University Press. Simmons, A. John. 1979. Moral Principles and Political Obligations. Princeton, NJ: Princeton University Press. Smith, M. B. E. 1973. Is there a prima facie obligation to obey the law? Tale Law Journal 82: 950-76. Soper, Philip. 1983. Legal theory and the problem of definition. University of Chicago Law Review 50: 1170-200. Soper, Philip. 1987. Choosing a legal theory on moral grounds. Social Philosophy and Policy 4: 31—48. Thompson, Michael. 1995. The representation of life. In Rosalind Hursthouse, Gavin Lawrence, and Warren Quinn (eds.), Virtues and Reasons. Oxford: Oxford University Press, 247-96. 28