I

One of the tasks Kramer sets for himself in his excellent new book, Where Law and Morality Meet, is to resolve the debate between incorporationists and exclusive legal positivists in favor of the former.¹

Kramer rightly points out the constant danger in this debate of reducing the disagreement to a terminological one, particularly since both exclusivists and incorporationists agree that moral principles provide binding reasons for judges to settle hard cases in certain ways rather than others. Disagreement arises, however, as to the proper description of the activity a judge engages in when she settles a hard case on the basis of some moral principle. Whereas the exclusivist claims that the judge is always making new law in such cases, the incorporationist argues that it is at least possible that the judge is applying already existing law. Whether this is a correct account of the judge's action depends on the rule of recognition partly constitutive of the legal system in which she is an official. If that rule is incorporationist, then the judge's action consists in the application of existing law, while if the rule is not incorporationist, the judge makes law when she authoritatively settles the case.

The contest between exclusivism and incorporationism cannot be settled, therefore, by pointing out that legal officials often treat moral principles as binding reasons for settling disputes. Kramer claims, however, that by focusing on a property of moral principles other than this one, a property that moral principles have in common with customary norms, the dispute can be settled decisively in the incorporationist's favor. I begin, then, with a description of Kramer's general argument for the claim that any theorist willing to acknowledge that custom can serve as a source of law ought to acknowledge that correctness as a moral principle can do so as well. I then turn to Kramer's specific application of this argument to Raz.

Kramer claims to find a weak spot in the exclusivists' position in their adherence to the following two claims. First, exclusivists explicitly recognize custom as a possible source of law; that is, a given norm can qualify as legal in virtue of its being a customary norm if the rule of recognition to which officials in the legal system in question adhere instructs them so to treat customary norms. Second, exclusivists argue that the appeal by judges to moral principles when settling cases ought to be conceived of along the same lines as judges' appeals to the laws of foreign countries or the rules of sporting associations when settling cases. Though a judge ought to treat the latter norms as settling the case before her—that is, she ought to give legal effect to those norms—her assignment of such a function to those norms does not thereby entail their incorporation into the law of the judge's legal system. Rather, they remain external to it. Raz contends that legal officials and philosophers ought to adopt an analogous position with respect to moral principles; though judges may be under a duty to apply those principles to settle certain cases (e.g., cases that cannot be settled via the application of existing social source–based law), those principles are not thereby incorporated into the judges' legal system.

Kramer then poses the question: “What distinguishes customary norms from the laws of foreign states, such that the former can be incorporated into the norms of a legal system, while the latter cannot?”²

He argues that the key difference between norms of these two types is that customs are free-floating, by which he means that they “do not issue from clear cut channels of authority”—that is, they do not result from deliberate acts of creation by officials—and “they are not subject to deliberate and abrupt modification through set procedures.”³ Foreign laws, in contrast, are not free-floating. It seems reasonable to conclude, then, that what makes customary norms eligible for incorporation into a legal system is that they are free-floating. Foreign laws, in contrast, cannot be incorporated as law because they lack this property. When we turn to moral principles, however, we find that they meet Kramer's definition of being free-floating; they are, in that sense, akin to customary norms and distinct from foreign laws. Kramer concludes, therefore, that if it is possible to incorporate customary norms into law in virtue of their being free-floating, so too it is possible to incorporate moral norms into law in virtue of their being free-floating. Thus as Kramer understands him, Raz errs when he groups moral principles with foreign laws as necessarily external to the norms of a legal system; instead, moral principles ought to be grouped together with customary norms as norms it is possible to incorporate into the laws of a particular legal system.

Kramer sums up his argument as follows:

As I argue below, exclusivists are not committed by their endorsement of custom as a source of law to viewing free-floatingness as a sufficient or even a necessary condition for legal validity. First, however, I set out Kramer's specific argument against Raz, which allegedly demonstrates the incompatibility of Raz's remarks regarding customary law and incorporationism.

I should note at the outset that in the criticism of incorporationism to which Kramer refers, Raz takes its main thesis to be the claim that “all that is derivable from the law (with the help of other true premises) is law.”⁵

Call this the derivability thesis. Kramer denies that the derivability thesis is part of incorporationism as he understands it, a point with which I believe other defenders of that theory would agree.⁶ This issue is irrelevant to my argument here, since both Kramer and I focus on a remark that Raz makes in the course of refuting the derivability thesis, not the refutation itself. I argue that this remark misrepresents Raz's position and suggest an alternative statement he ought to have made in its place. Though Raz's refutation of the derivability thesis is not at stake here, in order to lend credence to my suggested replacement I will indicate its compatibility with Raz's critique of that thesis.

Kramer quotes Raz as objecting to incorporationism or, more precisely, the derivability thesis, because “it makes the law include standards which are inconsistent with its mediating role, for they were never endorsed by the law-making institutions on whose authority they are supposed to rest.”⁷

He then argues that Raz ought to reject custom as a source of law on the exact same grounds, since “like moral principles, customary norms in their origins are not endorsed or created by the law-making institutions of the society within which they emerge.”⁸ However, Raz explicitly endorses custom as a possible source of law in the very same article in which he advances the above criticism of the derivability thesis. Obviously Raz might revise his claim regarding custom. But, not surprisingly, Kramer suggests that he abandon his criticism of incorporationism instead, stating that “if the absence of any original endorsement is no bar to the status of customary norms as laws, it should be no bar to such a status for moral principles.”⁹

In sum, Kramer's argument from custom against exclusivism and in favor of incorporationism has two components. The first involves an argument demonstrating that like customary norms, moral principles are free-floating, a feature that distinguishes them from foreign laws and the rules of sporting associations and that qualifies them for incorporation into law (in conjunction with an incorporationist rule of recognition). The second component involves highlighting an apparent inconsistency in Raz's legal philosophy, namely, his willingness to endorse custom as a possible source of law, on the one hand, and his denial of the same status to moral principles, on the other, for reasons that would seem to apply equally well to both. In the next section, I refute each of these arguments, beginning with Kramer's challenge to Raz.

II

Despite what he says in the quotation on which Kramer fastens, Raz does not require endorsement by law-making institutions in order for a norm to qualify as law. Rather, what he requires is that a norm be capable of being authoritative. To do this, a norm must meet two conditions. First, it must be, or at least be presented as, someone's view of how those (allegedly) subject to the norm ought to conduct themselves. This condition reflects the requirement that in order to be authoritative, a norm must be capable of mediating between agents and the reasons that apply to them (the norm's dependent reasons, as Raz calls them). In order for it to play this role, a norm must meet a second condition, namely that agents be able to identify the norm without recourse to the reasons on which it depends. Raz clearly believes that despite not being endorsed by any law-making institutions, customary norms can meet these two conditions.¹⁰

There is textual support in Raz's writings for two accounts of how customary norms can be authoritative. They differ with respect to whose view the customary norm can be said to express regarding how those subject to the norm ought to behave, or in other words, who is the author of the norm. In what follows I set out both accounts; which one accurately reflects Raz's view is not essential, since both accounts entail that customary norms can be authoritative, a conclusion that is the key to defending Raz against Kramer's criticism.

Consider the following familiar account of how a customary norm might come to exist in a particular society. At time t1, some agents in this society engage in actions of a certain type C, the type of action that will eventually become a matter of customary practice. These agents engage in actions of the type in question for certain reasons (exactly which reasons and whether they are good reasons are irrelevant). Over several generations, new members of the society begin to conform to the evolving practice of ϕing when conditions A, B, and C are met. Crucially, conditions A, B, and C are not the reasons that led the agents at time t1 to ϕ, even though it is those reasons, and not A, B, and C, that justify ϕing (or, in a case where ϕing is not in fact justified, were mistakenly believed by the agents at time t1 to justify ϕing). Suppose that at time t7, seven generations after time t1, there exists a practice in this community of ϕing whenever conditions A, B, and C are met and that the justification members of this community offer for their behavior in these conditions is something like “ϕing is what you ought to do whenever conditions A, B, and C are met.” Furthermore, members of the community (and perhaps even outsiders) are criticized for failing to ϕ in the appropriate conditions. In these circumstances, a customary rule exists in this community. Clearly this rule mediates between agents and the reasons that apply to them, and insofar as recognizing that conditions A, B, and C obtain does not require that agents consider the reasons upon which the norm depends, the customary norm also meets the second requirement for being the kind of norm that could be authoritative. Whether members of the community ought to follow the rule depends on whether they are in fact more likely to act on the balance of reasons by doing so than by trying to determine for themselves what the balance of reasons favors (or requires) in these circumstances.¹¹

But our concern here is not with the conditions under which agents ought to treat norms as authoritative but rather with the conditions a norm must meet if it is even to be capable of being authoritative.

An agent may deliberately set out to create a customary norm, though whether she succeeds depends on the actions and attitudes of others, something over which the agent can normally exercise only limited control. In this respect, customary norms differ from law, where it is possible for a legal official to create a legally valid norm even if no one (or at least very few people) to whom it is addressed treat it as authoritative. But it is also possible for customary norms to arise without any deliberate effort on the part of any agent to create an authoritative rule; a shared belief (or at least a widespread belief) amongst a group of agents that a certain sort of behavior is required and that deviations from that standard call for justification suffices for the existence of a customary rule.¹²

Such a state of affairs may obtain without any intentional effort to bring it about. What is crucial, however, is that once a customary norm exists, it can mediate between agents who engage in the practice and the reasons there are (or at least were) for engaging in acts of that type in the first place.

The possibility that a customary norm might exist and yet have no author would seem to conflict with Raz's claim that “a directive can be authoritatively binding only if it is, or is at least presented as, someone's view of how its subjects ought to behave.”¹³

Raz adds the caveat that a directive need only be presented as someone's view of how those the directive addresses ought to behave in order to cover cases of insincerity; that is, cases where someone aims to create the impression that those over whom he claims authority ought to ϕ, though in fact he believes that the balance of reasons that apply to them favor (or require) their γing. Perhaps this caveat can be extended further, to encompass cases where a norm is presented as if it expresses the view of some agent or agents regarding how those addressed by the norm ought to conduct themselves, though in fact no agent ever intended it as such. After all, customary norms often present themselves as the wisdom of earlier generations, or that of some great lawgiver from the mists of prehistory, even when they are not in fact the product of a deliberate act of creation or endorsement by any particular agent or agents.¹⁴ Indeed, in some of his early remarks on law and authority, Raz himself considers the question of whether customary norms can be authoritative though they have no author, and answers in the affirmative:

It certainly seems possible for a customary norm to provide a protected (or preemptive) reason for action, despite its not being the product of any deliberate act of creation. As the above discussion illustrates, customary norms can mediate between agents and the reasons that apply to them, a role that Raz claims requires their providing preemptive reasons for action. But perhaps the crucial question is whether an authorless customary norm could ever enjoy legitimate authority. That is, could the customary norm's “claim” to direct the behavior of those it addresses ever be justified?

I suggest that at least on Raz's understanding of what usually justifies a claim to authority—the normal justification thesis (NJT)—a customary norm could enjoy legitimate authority. The NJT involves showing that “the alleged subject is likely better to comply with reasons which apply to him (other than the alleged authoritative directives) if he accepts the directives … as authoritatively binding, and tries to follow them, than if he tries to follow the reasons which apply to him directly.”¹⁶

It seems plausible to think that customary norms are the product of a process that can reasonably be thought to provide norms that sometimes meet the standard set out by the NJT. Indeed, many conservative political philosophers would seem to be committed to such a conclusion. Of course, customary norms will not always enjoy legitimate authority, but then the same is true of legislation and judicial decisions. And we must recall that customary norms can and do change over time, often in response to their failure to meet the NJT, a point I return to below. I tentatively conclude, therefore, that despite their lack of an author, customary norms possess those features necessary for being authoritative and that it is reasonable to think that in some cases they actually are authoritative.

Though I think the above description of customary norms is consistent with Raz's analysis of what is involved in a norm's being capable of being authoritative, it is not clear to me whether he would endorse it. Raz writes that though a law may not be the product of a single intention, every aspect of that law is the product of some agent's intention.¹⁷

Yet Raz clearly wishes to acknowledge custom as a possible source of law. Thus it seems we face a conflict between the above analysis of custom as capable of being authoritative even if authorless and the claim that all law is the product of some agent's intention.

This conflict can be avoided by the second analysis of how custom can be authoritative. Before we move on to that account, however, consider the context in which Raz makes his remark about every law being the product of some agent's intention. This statement is part of his response to the following critique of Raz's analysis of law in terms of authority:

The main point of this critique is that our practice of attributing authorship to legal norms that have evolved over time does not support the claim that law mediates (and must mediate) between agents and the reasons that apply to them. But a secondary point is that relying on the mediating role of authority becomes an empty phrase when rules have evolved over time. If true, this point would seem to undermine the claim that customary norms can be authoritative, since paradigmatic customary norms are just rules that evolve over time. But I see no reason to think that this claim is true. Modifications to a customary norm that occur over time are usually made in response to perceived shortcomings of the norm; specifically, a perceived discrepancy between what is involved in acting in conformity with the norm and what is involved in acting on the balance of reasons on which the norm depends. Prominent among the causes of such modifications will be changes in the environment inhabited by (some) members of the community in which the norm is (or has been) operational or a change in the technology available to (some) members of this community. Modifications to the norm, then, will be the result of intentional actions, though once again the actors whose conduct is of the type that later becomes customary may not intend by their conduct to modify the norm (though it should be noted that they may have just such an intention). As long as it is plausible to think that customary norms evolve in a way that tracks, or more accurately, is believed to track, the reasons that apply to those subject to the norms, we should not conclude that relying on the mediating role of authority is merely an empty phrase when applied to norms that evolve through the activities of many hands over a long time. And if not, then perhaps we need not follow Raz in claiming that every aspect of every law is the product of some agent's intention.

The second analysis of how customary norms can be authoritative begins with Raz's claim that while customary norms are not “normally generated by people intending to make law … [such norms] can hardly avoid reflecting the judgment of the bulk of the population on how people in the relevant circumstances should act.”¹⁹

Customary norms play the requisite mediating role because they instruct a person regarding what most people in his community believe he ought to do, and he can identify their instructions to him without having to discover the reasons upon which their instructions depend. Unlike the first analysis of custom offered above, this analysis can identify an author for specific customary norms; they are instructions to an agent from his contemporaries in their community. So understood, customary norms clearly meet the first condition necessary for being capable of being authoritative: they are someone's view of how those subject to the norm ought to behave. Moreover, it is consistent with Raz's claim that every aspect of every law is the product of some agent's intention; whatever determinate content a given customary norm has, it has as a result of bulk of the community believing that it has such content.

Of course, members of the community may not view customary norms as directives they (collectively) issue to themselves (individually). Rather, while they may view themselves as justified in criticizing someone for diverging from the customary practice (i.e., in enforcing the norm), they are likely to view the custom as having authority in virtue of the practice having been done that way in the past. They may even mistakenly view it as having been deliberately created by former leaders or officials in their community and think its authority rests on the authority of those agents. Raz can respond, however, that what makes the custom a standard of behavior deviation from which merits criticism in this community is the belief by contemporary members of the community that it is, not that those in the past intended it as such (if they did). Though contemporary members of the community may not be the originators of the customary norm, its continued existence depends on their attitude toward it. In a weak sense, then, customary norms are endorsed by many agents in the community.²⁰

In short, what makes a customary norm an authoritative standard is that it is so viewed by the bulk of the community, regardless of their own self-understanding of why they view it as such.

Regardless of which of these two accounts of how customary norms can be authoritative we accept, the crucial point is that such norms are capable of being authoritative: they can mediate between agents and the reasons that apply to them, and they can be identified without recourse to the reasons on which they depend. This is as it must be if we are to reconcile Raz's analysis of the nature of law with his explicit claim that customary norms can be legally valid. But what, then, to are we to make of Raz's claim that in order to be law, standards (norms) must be endorsed by the lawmaking institutions on whose authority they are supposed to rest?

I suggest that we simply ignore this remark as not accurately reflecting Raz's own position. Instead, Raz should have offered the following disjunctive test of legal validity: A norm qualifies as legal if and only if it is (1) endorsed by the lawmaking institutions on whose authority it rests; or (2) it is part of the political community's customs—standards of right conduct that exist only because most members of the political community believe they do, and behave accordingly, and that are capable of being authoritative insofar as they reflect the judgment of the bulk of the population as to how people in the relevant circumstances should act. Either condition is sufficient for legality when officials of a given legal system treat its fulfillment as a binding basis for their judgments. Clearly such a disjunctive test for legality will permit the incorporation of customary norms into law while at the same time denying the status of law to moral principles qua moral principles.²¹

Of course, a test for legality must be argued for; it does not suffice as a response to Kramer that we stipulate a test that excludes moral principles from law while allowing for the incorporation of custom. But Raz has an argument for defending the above disjunctive test of legal validity as superior to Kramer's own test, which treats being free-floating as sufficient for legality: Raz's test is consistent with the requirement that norms be capable of being authoritative, while Kramer's test is not.²²

The disjunctive test for legal validity that I attribute to Raz in the preceding paragraph is fully compatible with the point he wishes to make against the claim that “all that is derivable from the law (with the help of other true premises) is law.”²³

Raz contends that this claim is incompatible with the thesis that in order for a norm to be legal, it must be capable of being authoritative. It does not follow from a norm having been endorsed by lawmaking institutions that every claim derivable from it would be so endorsed as well. When legal officials endorse or create a norm, they do so (or at least claim to do so) for the purpose of directing their subjects' behavior. Legal norms express officials' beliefs regarding which action in a given circumstances the balance of reasons favors.²⁴ However, Raz claims that “people do not believe in all that is entailed by their beliefs.”²⁵ Assuming Raz is correct about this, it is not the case that all of the norms that are derivable from the law (with the help of other true premises) can be authoritative—that is, capable of mediating between agents and the reasons that apply to them—since they will not all express officials' beliefs about what their subjects ought to do. An identical argument applies as well to customary norms. It simply does not follow from a norm reflecting the judgment of the bulk of the population regarding how people in the relevant circumstances should act that every norm derivable from it (with the help of other true premises) similarly reflects the judgment of the bulk of the population regarding how people should act. Thus Raz's argument against the derivability thesis remains untouched by the proposed revision to his statement regarding the sufficient conditions for legal validity.

In sum, Kramer contends that Raz's commitment to custom as a possible source of law is incompatible with remarks the latter makes while criticizing incorporationism, or more accurately, the derivability thesis. The problem lies with Raz's claim that in order to qualify as law, a norm must be endorsed by the lawmaking institutions on whose authority the norm is supposed to rest. Kramer suggests that Raz retain his commitment to custom as a source of law and that he resolve the aforementioned incompatibility by dropping the requirement that a norm be endorsed by some lawmaking institution in order to qualify as legal. Kramer appears to believe that Raz, having done so, will have no reason to object to the possibility of norms counting as legally valid in virtue of their correctness as moral principles without the endorsement of any legal official (or anyone else, for that matter).

In response, I have argued that the remark Kramer fastens onto misrepresents Raz's position. Raz is not committed to the view that the only way a norm can qualify as law is for it to be endorsed by the lawmaking institutions on whose authority it depends. Rather, he is committed to the claim that in order to be law, a norm must be capable of being authoritative, and that in order to be authoritative, it must be capable of mediating between an agent and the reasons that apply to her. Norms endorsed by lawmaking institutions meet this condition, but so too do customary norms. Thus the following disjunction more accurately states Raz's test for legal validity. Norms treated as binding by legal officials qualify as laws of that system if and only if they satisfy one of the two following disjunctive tests: either they are created and/or endorsed by officials in lawmaking institutions, or they are customary norms of the political community.²⁶

Both disjuncts meet Raz's requirement that as a necessary condition of legality, norms be capable of being authoritative. This requirement provides in turn a justification for Raz's recognition of custom as a possible source of law despite customary norms not being endorsed by any legal official, as well as his denial of this status to moral principles.

III

As I indicate in Section I, Kramer believes that Raz treats customary norms but not foreign laws or the rules of sporting associations as eligible for incorporation into the law of a given legal system. Kramer then argues that it is because customary norms are free-floating while foreign laws are not that the former but not the latter can be incorporated. Since moral norms are also free-floating, and since Raz clearly accepts custom as a source of law, consistency requires that he recognize morality as a source of law as well.

The arguments of the previous section demonstrate that it is because customary norms are capable of being authoritative while moral norms are not that Raz can (and does) consistently reject morality as a possible source of law while granting this status to custom. But even if being free-floating is not a sufficient condition for a customary norm's being eligible for incorporation as law, might it be a necessary condition for such status? Were Kramer correct in attributing to Raz the view that foreign law cannot be incorporated into law, then possessing the property of being free-floating might account for why customary norms are eligible for incorporation while foreign laws are not. But in fact, it does not appear that Raz wishes to make such a strong claim.

The point Raz wants to make when discussing foreign laws is that we cannot infer from a legal system having a rule that gives legal effect to some standard that the standard has thereby been made part of the legal system.²⁷

Raz uses conflict-of-law doctrines that direct a court to apply foreign law when settling certain types of disputes to illustrate this point. For example, that a U.S. court is directed to apply Italian law when settling certain cases brought before it does not necessarily entail that the Italian law in question has thereby become part of U.S. law. But this illustration and the line of argument it supports are perfectly consistent with the view that foreign law can sometimes be incorporated into the law of a different legal system. Indeed, Raz himself identifies in the United Kingdom's Human Rights Act 1998 (which came fully into force in 2000) an example of what he believes to be the incorporation of foreign law—specifically, various rights enumerated in the European Convention on Human Rights—into the law of the United Kingdom. More important than his judgment in this particular case, however, is the fact that there is nothing in Raz's theory of law that rules out the incorporation of foreign law. It is both necessary and sufficient for a norm's being eligible to be incorporated as law that it is capable of being authoritative, and this is a property that both foreign laws and the rules of sporting associations typically have.

Interestingly, however, while Raz's conception of law does not entail that being free-floating is a necessary condition for the incorporation of a norm as law, he does make some brief remarks that suggest the lack of this property can be a weighty (though perhaps not conclusive) consideration against it. Raz writes that:

A plausible, though perhaps only partial, interpretation of this claim is that reasons of political morality and/or political psychology will often count against the incorporation of foreign laws. The capacity of the law to serve as a focal point for individuals' identification with the political community will likely be reduced the more the law incorporates foreign law. This is so because such law may be subject to deliberate and abrupt modification by agents outside the political community; it will, in that sense, not be free-floating. It would not be surprising if citizens of the United Kingdom were to experience a sense of alienation from their political community were they to find significant aspects of their lives regulated by Japanese law, subject to modification by Japanese legal officials over whom they exercise no authority, and who are not even members of the same political community.²⁹

Similarly, while there is nothing in Raz's conception of law that excludes the incorporation of various customary norms extant in some other political community (i.e., a political community other than the one partially organized by the legal system in question), considerations of political morality and/or political psychology will usually count strongly against taking such action.

IV

So far I have focused my discussion on defending Raz against Kramer's criticism. Yet Kramer believes that his argument applies to all those he identifies as exclusive legal positivists. Therefore I conclude this paper by briefly demonstrating that another of these legal philosophers—Scott J. Shapiro—also has the resources within his theory of law to rebut Kramer's argument.

Shapiro argues that in order to be law, a norm or rule must be capable of guiding an agent's conduct, a function it can fulfill only if it can make a practical difference to an agent's deliberation.³⁰

To do so, the rule must be capable of “motivating agents to act differently from how they might have without [the rule's] guidance.”³¹ That is, if a person would reach the same conclusion regarding what he ought to do in every case to which a certain rule applies, regardless of whether he followed the rule or deliberated directly on the merits of the potential actions open to him, then the rule in question would not be one that could make a practical difference to this agent's motivation. As Shapiro puts the point, the rule makes a practical difference to an agent's deliberation only if, in counterfactual circumstances in which the agent does not seek to guide his conduct by following the rule, his conduct might diverge from that required by the rule.³²

Shapiro argues that rules pronounced valid under an incorporationist rule of recognition—one that directs legal officials to treat moral correctness as sufficient for legal validity—could not make a practical difference to the reasoning of legal officials. The crucial point is that there can be no case in which the decision the judge renders depends upon the primary rules validated under this rule of recognition. Since the judge is already committed to a rule of recognition that directs her to reach decisions through the application of correct moral principles, a rule validated by this rule of recognition (i.e., one deemed valid because it is morally correct) cannot result in the judge making a decision that she would otherwise not have reached—that is, a decision that she was not already committed to making given her prior commitment to an incorporationist rule of recognition.³³

The same is not true for a rule of recognition that treats only social sources, including customary practice, as sources of valid law. The decisions rendered by a judge committed to this rule of recognition do depend upon the rules validated under it, in that the correct settlement of a particular case depends upon what the existing customary practice actually is.³⁴

Given that customary practice can evolve over time or cease to exist altogether, the rules validated by a rule of recognition that treats custom as a source of law can also change over time. It follows that we can imagine a legal official committed to the same rule of recognition reaching different decisions in two cases that are similar in all the relevant respects except that the customary practices of the community in question have changed between the time when the first case and the second case were settled. It is due to the possibility of such a change in customary practices that norms validated by a rule of recognition that treats custom as a source of law are capable of making a practical difference to legal officials' deliberation.³⁵

Kramer claims that since all exclusive legal positivists recognize that customary norms can be laws of the legal systems in which they are operative independent of their being enacted or endorsed by officials in lawmaking institutions, they ought to grant the same recognition to moral principles. It should now be clear why Shapiro does not act inconsistently when he refuses to do so. Customary norms but not moral ones can make a practical difference to legal officials' deliberation (when validated by an appropriate rule of recognition), and therefore Shapiro treats custom but not morality as a possible source of law. (It is worth noting as well that foreign laws can make a practical difference to legal officials' deliberation, and so there appears to be nothing in Shapiro's account of law that would prohibit him from treating foreign law as a source of law.) Of course, Shapiro may be mistaken in thinking that an essential feature of law is that it be able to make a practical difference to an agent's deliberation.³⁶

However, he is not committed by his treatment of custom as a possible source of law to granting the same status to morality, as Kramer implies that he is.