I

H.L.A. Hart's discussion of international law in chapter 10 of The Concept of Law has earned him few friends. In one respect this is surprising; after all, Hart devotes a considerable portion of the chapter to rebutting his fellow legal-positivist John Austin's claim that international law is not law properly so-called but (only) positive morality. Why, then, do so many remain unhappy with Hart's treatment of international law? In a word, because he characterizes it as primitive. More precisely, Hart maintains that international law resembles the simple legal order that regulates a primitive society, one that “we are accustomed to contrast with a developed legal system,” and especially, in The Concept of Law, with the municipal legal system of a well-functioning state.Footnote ¹

For some, the mere use of the word “primitive” to describe international law casts a good deal of suspicion not only on Hart's discussion of it but also on his entire analysis of law.Footnote ² For others, the problem lies with the fact that this very analysis provides no warrant for the conclusion that international law is akin to a primitive social order; that is, for Hart's assertion “that international law not only lacks the secondary rules of change and adjudication which provide for legislature and courts, but also a unifying rule of recognition specifying ‘sources’ of law and providing general criteria for the identification of its rules.”Footnote ³ Hart's arguments in chapter 10, writes Jeremy Waldron, “become careless and their application thoughtless in regard to law in this area,” and the result is an embarrassingly inadequate account of the nature and status of international law.Footnote ⁴ Mehrdad Payandeh also claims to identify multiple shortcomings in Hart's application of his own ideas to international law and suggests that his “insistence that international law does not constitute a legal system seems almost as problematic as Austin's insistence that international law is not law at all.”Footnote ⁵ Even Jean d'Aspremont, who draws a great deal from Hart's analysis of law in developing his own formalist theory of international law's sources, describes Hart's discussion in chapter 10 as a “disappointing and unconvincing portrayal of international law as a very primitive set of rules.”Footnote ⁶

At best, then, Hart's allegedly careless ruminations on international law constitute a missed opportunity to improve the quality of his own jurisprudential theorizing as well as that of his students and readers. At worst, despite his intentions Hart bears some blame for the persistence of skepticism vis-à-vis international law's genuine status as law and the relatively small number of analytical legal theorists engaged with this “issue of the hour.”Footnote ⁷

In this essay I argue that antipathy to Hart's treatment of international law rests on two misunderstandings. The first concerns the nature of the distinction, or rather distinctions, that Hart draws between a primitive and an advanced society. Not only in chapter 10 but throughout The Concept of Law, Hart fails to separate clearly and consistently two bases for differentiating a simple or primitive social order from a complex or advanced one. One is the absence or presence of secondary rules tout court. The other is the absence or presence of a division of labor in identifying, altering, applying, and enforcing the general social rules that govern and partly constitute a given society. This specialization in the performance of governance tasks is realized in the practice of a specific class of secondary rules, namely those that create or constitute legislative, judicial, and executive offices and institutions.Footnote ⁸ Waldron, Payandeh, and others assume that Hart has the first of the two aforementioned distinctions in mind when he characterizes international law as primitive, and they critique his account on that basis. I contend, however, that when Hart describes international law as akin to a simple social order, he employs the second basis for distinguishing a primitive from an advanced society. Read in this light, many of Hart's claims in chapter 10 were compelling when he wrote The Concept of Law, and, to a not-insignificant degree, remain so today.Footnote ⁹

The second confusion that afflicts critics of Hart's remarks on international law follows from a failure to distinguish carefully between two different functions that reference to the rule of recognition serves in Hart's analysis of law. One is epistemic; the (or a) rule of recognition is a norm that actors can use to identify the law of a particular society. The other is ontological; here, “rule of recognition” refers not to a norm but to the practice of holding accountable that makes it the case that rules R1, R2, and so on are rules of a particular society. Whereas Hart's critics rely on the epistemic role played by a rule of recognition and so reject his claims that international law lacks a rule of recognition and is not a legal system, I argue instead that these claims should be interpreted ontologically. A legal system in the sense that concerns Hart in chapter 10 requires a hierarchy of agents, a feature that international law largely does without. That is why he maintains we can form no ontological “rule” of recognition for international law, or at least not an informative “rule,” since there is nothing much to say about what makes L1, L2, and so on international law other than that states use them to hold themselves and each other accountable.Footnote ¹⁰

In the next section I offer a careful reading of passages from chapters 5 and 6 in The Concept of Law that support the following two conclusions. First, the absence of a division of labor in the performance of governance tasks—that is, identifying, altering, applying, and enforcing law—is equally, if not more, central to Hart's understanding of what makes a society primitive as is the absence of any secondary rules at all. Second, and relatedly, it is primarily in terms of the presence of such a division of labor and the implications it has for the ontology of law that Hart understands the idea of a legal system and the ideas of a rule of recognition and legal validity that accompany it. These conclusions provide the basis for the interpretation of Hart's remarks on international law that I offer in Section III, one that I explicate by way of contrasting it with Waldron's and Payandeh's mistaken reading of chapter 10.

In the final section I offer rebuttals to several other criticisms of Hart premised on a misunderstanding of his characterization of international law. Furthermore, I suggest some of the ways in which embracing that characterization might enhance both our theoretical understanding of how international law contributes to the production of social order as well as the quality of our reflection on international law as a means for advancing or realizing justice.

II

From the very outset, Hart fails to distinguish clearly and consistently between the absence of secondary rules and the absence of specialization in the performance of governance tasks as a basis for characterizing a social order as primitive. He begins his exposition of the idea of a primitive society with the following claim:

Here Hart does not maintain that primitive communities are ones characterized only by the existence of duty-conferring rules. Rather, he claims only that in such a society social rules exist and contribute to the production of social order in virtue of the group members in general taking the internal point of view to those rules; that is, their “use [of] the rules as standards for the appraisal of their own and others’ behaviour.”Footnote ¹² In order to avoid certain confusions that he fears will follow from labeling as “customary” a social order that lacks any specialization in the performance of governance tasks, Hart proposes in the next sentence to refer “to such a social structure as one of primary rules of obligation.”Footnote ¹³ Alas, in doing so, he appears to shift from distinguishing between a primitive and an advanced society on the basis of the absence or presence of specialization in governance, or what I call the specialization distinction, to distinguishing between the two types of societies on the basis of the absence or presence of secondary rules tout court, or what I call the functional distinction.Footnote ¹⁴

The ensuing discussion of the defects of uncertainty, rigidity, and inefficiency that characterize a primitive society and the contribution that rules of recognition, change, and adjudication make to ameliorating these defects further muddies the fact that there are two ways to distinguish primitive from advanced societies at issue. The functional distinction figures most clearly in Hart's description of a primitive society suffering from the rigid or static character of its rules:

The implication seems clear: if a society possesses rules of change, then it is not a primitive society, a point Hart's critics press against him when they point to the existence of such rules in international law as a reason to reject his description of it as primitive. Read in context, however, even this passage offers some reason to think that Hart had both the functional and the specialization distinction in mind. He continues:

The acknowledgment that this extreme case has likely never been realized in any actual community, together with Hart's earlier reference to studies of actual primitive communities, suggests a recognition that communities without specialization in modifying the law, that is, a legislature or courts, nevertheless can and do possess rules of change; for example, rules that empower members “to release those bound from performance or to transfer to others the benefits which would accrue from performance.”Footnote ¹⁷ The contrast Hart draws between “changing the general rules” and changing “the obligations which arise under the rules in particular cases” also implies that the absence of a legislature or courts marks one way to distinguish a primitive society from an advanced one, with the absence of any rules of change at all providing a separate way of drawing such a distinction. Granted, in one sense the creation of a legislative body need not involve marking out a special set of agents who make or modify the rules; the legislators may consist of all a community's members. Nevertheless, it introduces such specialization “in embryonic form” by introducing the categories of official and subject that Hart maintains are absent from a primitive community.

Moreover, the deliberate change of general rules that constitutes legislation differs from the process of customary rule change in two related respects. First, the actors whose practice of holding accountable produces a change to a society's customary norm do not understand their actions as the deliberate creation of a new norm but rather as making explicit a normative standard to which the community is (or has already become) committed. Second, the customary rule-making process is successful only when the new rule is generally practiced within the community, whereas the legislative law-making process makes possible successful rule change in advance of members of the society in question using the rules to hold one another accountable.Footnote ¹⁸ I return to all of these points below.

Hart's discussion of inefficiency focuses entirely on the shortcomings endemic to any society that does not employ a division of labor in carrying out the tasks of adjudication and enforcement. With respect to adjudication, Hart writes, “disputes as to whether an admitted rule has or has not been violated will always occur and will, in any but the smallest societies, continue interminably, if there is no agency specially empowered to ascertain finally and authoritatively, the fact of violations.”Footnote ¹⁹ As for enforcement, Hart describes as a weakness of a “simple form of life” the fact that “punishments for violations of the rules, and other forms of social pressure involving physical effort or the use of force, are not administered by a special agency but are left to the individuals affected or the group at large.”Footnote ²⁰ These defects in the capacity of rules to contribute to the production of social order can be mitigated by the adoption of rules that create specialized offices (e.g., judges) who enjoy sole authority to ascertain whether a violation of the rules has occurred and “the exclusive power to direct the application of penalties by other officials.”Footnote ²¹ While Hart focuses on the characterization of these rules as secondary rules, the above description of the defects they address and the kind of solutions to them Hart identifies clearly implies that in these passages he conceives of specialization in the tasks of adjudication and enforcement as the property that distinguishes an advanced society from a simple or primitive one.

Hart's discussion of the role(s) played by the rule of recognition is perhaps the most vexing feature of The Concept of Law. Nevertheless, a compelling case can be made that the specialization distinction, far more than the functional distinction, informs Hart's claim that “possession” of a rule of recognition distinguishes a primitive from an advanced society. Hart identifies as the first defect of a primitive social order the fact that when “doubts arise as to what the rules are or as to the precise scope of some given rule, there will be no procedure for settling this doubt, either by reference to an authoritative text or to an official whose declarations on this point are authoritative.”Footnote ²²

Reference to the existence of an official obviously implies a division of labor between rulers and ruled. But so, too, does reference to an authoritative text, since such a text must have an author, someone or some group whose judgment regarding the rules and their scope is taken as authoritative by other members of the community. Hart says as much when he writes that the crucial first step from the prelegal to the legal—or perhaps he ought to have said, from a primitive rule-guided social order to an advanced one—is not “the mere reduction to writing of hitherto unwritten rules” but “the acknowledgement of reference to the writing or inscription as authoritative, i.e. as the proper way of disposing of doubts as to the existence of the rule.”Footnote ²³

Why a first but, by implication, not a final step from prelegal to legal? What does Hart mean when he describes even the simplest rule of recognition, such as reference to a list or text viewed as authoritative, as providing “in embryonic form the idea of a legal system?”Footnote ²⁴ Why is it that, “in the simple operation of identifying a given rule as possessing the required feature of being an item on an authoritative list of rules we have the germ of the idea of legal validity?”Footnote ²⁵ The answer to these questions, I submit, is that we have introduced the possibility of not only specialization in the identification of law but a division of labor in the task of sustaining the rules that constitute it. That task is performed by the practice of holding accountable that Hart characterizes in terms of adopting the internal point of view to the social rules that regulate the affairs of members of a given society. It is the presence of this division of labor, at least to some significant degree, that makes a society advanced in the sense that concerns Hart, and so its absence that makes a society primitive.

The foregoing discussion reveals an ambiguity in Hart's discussion of the idea of a “legal system.” In some passages Hart contrasts a legal system with a mere set of laws or primary rules of obligation. The former differs from the latter in virtue of its including a rule of recognition that addresses doubts over the scope of various rules by arranging them in an order of superiority; for example, on the basis of their source.Footnote ²⁶ In this case the (or a) rule of recognition refers to a genuine rule, one that actors can use to guide their conduct. Specifically, the rule presents itself as, and may well be, epistemically authoritative, a description of the practice of holding accountable that constitutes the normative social order in question that both members and nonmembers can use to describe and predict the conduct of practitioners.Footnote ²⁷ Even a society that lacks any specialization in the performance of governance tasks can have a rule (or more than one rule) of recognition that serves this aim; for example, one that depicts norms restricting violence against people as having action-guiding priority over norms that restrict violence against property.

Other passages in which Hart invokes the concept of a legal system, however, are better read as claims regarding the absence or presence of a division of labor in the task of sustaining the law by adopting the internal point of view to the rules that constitute it. Hart employs the phrase “rule of recognition” in these passages as well, but the phrase is a misnomer in that what Hart refers to is not a norm that officials use to determine what to do or to believe but the social fact constituted by officials’ practice of holding themselves and one another accountable. Talk of a rule of recognition here is ontological, a description of what makes it the case that rules R1, R2, Rn are rules of the society in question. Put another way, it is an account of the truth conditions for candidate rules of recognition that play an epistemic role, that is, statements that purport to provide actors with authoritative descriptions of officials’ practice of holding themselves and one another accountable.Footnote ²⁸

When Hart describes as necessary and sufficient for the existence of a legal system in a given society general obedience to the law on the part of private citizens, that is, subjects, and its officials’ “effective acceptance” of common rules for identifying, changing, and adjudicating the law, he offers a response to the ontological question “what makes law?”Footnote ²⁹ But why think satisfaction of both conditions is necessary and not merely sufficient for the existence of law, or more accurately, of a legal system? The answer is that only when both conditions are satisfied (to some considerable degree) will a society have achieved a division of labor in sustaining the practice of holding accountable that constitutes law that marks the transition from a primitive to an advanced social order. As Hart writes:

If we ignore the unfortunate phrase “which consists only of primary rules,” there can be no doubt that for Hart, one critical, indeed “most fruitful,” way to distinguish an advanced legal system from a simple legal order is on the basis of specialization in the performance of governance tasks.Footnote ³¹

In sum, when Hart invokes the concept of a rule of recognition that serves the epistemic function of identifying a society's rules and their scope, reference to a “legal system” refers to a hierarchy of rules. In contrast, when Hart invokes the concept of a rule of recognition as part of his ontological account of what makes law, reference to a “legal system” refers to a hierarchy of actors, of officials and subjects or rulers and ruled.

III

With a clearer understanding of the two (or perhaps three) ways Hart distinguishes between a primitive and an advanced social order, as well as the two different roles that reference to a rule of recognition plays in Hart's analysis of law, we can now explicate and assess Hart's characterization of international law in chapter 10. An explanation of how and why criticisms of that characterization fail to hit their mark provides an especially perspicuous method for carrying out this task.

Waldron develops his objections to Hart's views on international law by evaluating “Hart's claim that the international legal order is a primitive legal system, consisting of nothing but primary rules.”Footnote ³² He begins by pointing out that international law includes secondary rules of adjudication, such as those set out in the statute creating the International Court of Justice (ICJ), as well as secondary rules of change, such as those constitutive of the treaty-making process. Clearly Hart was aware of these features of international law; indeed, he explicitly refers to both of them in chapter 10. Therefore Waldron attempts to extract what he takes to be Hart's reasons for nevertheless describing international law as a relatively simple social order. In the case of the ICJ, Waldron maintains Hart's reason is the lack of compulsory jurisdiction, while in the case of treaty-making, the reason is the fact that the resulting rules do not apply to nonparties.Footnote ³³ Though Waldron concedes that the absence of compulsory jurisdiction marks a significant disanalogy between a municipal legal system and international law, he points out that this does not warrant the inference that international law contains no rules of adjudication.Footnote ³⁴ As for the fact that treaties do not apply to nonparties, Waldron writes: “it is not clear why the point about ‘binding states that are not parties’ should be jurisprudentially so important,” and in particular, why Hart should apparently treat it as a necessary condition for the existence of rules of change.Footnote ³⁵

Waldron follows Payandeh in attributing Hart's “poor reasoning” to an unwarranted assumption that the functions that secondary rules perform, namely mitigating uncertainty, rigidity, and inefficiency, require the existence of the specific institutions found in a municipal legal system, such as that of a legislature or courts with wide-ranging compulsory jurisdiction:Footnote ³⁶

The problem with both Waldron's and Payandeh's critiques is that “the main distinction” Hart draws between a primitive and an advanced society, or at least the one he employs when he describes international law as akin to a simple social order, is not the functional one that Waldron and Payandeh assume. Rather, it is the absence or presence, respectively, of a division of labor in the carrying out of those tasks in virtue of which a legal system can rightly be said to exist.

Why accept this claim? Start with the fact that Hart tells us so in the introductory section of chapter 10, where he writes: “it is indeed arguable, as we shall show, that international law not only lacks the secondary rules of change and adjudication which provide for legislature and courts.”Footnote ³⁸ The claim here concerns the absence of specific kinds of secondary rules, not the absence of secondary rules tout court. Recall as well that when Hart first introduces the idea of a simple society that lacks rules of change, he distinguishes between the absence of even an embryonic form of legislation, that is, rules constitutive of a process that makes possible deliberate change to general rules and that introduce the germ of the idea of a distinction between ruler and ruled, and rules that make possible changes to “obligations which arise under the rules in particular cases.”Footnote ³⁹ The secondary rules that constitute treaty-making are examples of this second type of rules of change; they alter obligations that obtain under the general rules of international law only for the particular states that sign (and, if necessary, ratify) them.Footnote ⁴⁰

As Hart notes in the final paragraph of chapter 10, were international law to evolve so that multilateral treaties were generally recognized as binding upon states that were not parties to them, then the norms that comprise international law would include secondary rules that provide for legislation, that is, the ability to make deliberate changes to general rules in advance of any actor using them to hold himself or others accountable. It would also include the distinct categories of rulers (or officials) and ruled (or subjects), and so constitute a vertical organization that contrasts with the common description of international society as a horizontal order. Thus we have an answer to Waldron's question quoted above regarding the jurisprudential significance of “binding states that are not parties” to a treaty; such a practice is sufficient for the existence of a division of labor in making changes to the general rules that regulate members of international society, and it is the development of specialization in this and other governance tasks that transforms a given society from primitive to advanced.

Consider, now, Hart's remarks on international adjudication and enforcement.Footnote ⁴¹ He concedes that with respect to each of these components of governance, there is some analogy between municipal and international law. In the case of enforcement, international law includes “secondary rules specifying or at least limiting the penalties for violation,” which might make some contribution to reducing “the smouldering vendettas which may result from self-help.”Footnote ⁴² But to the extent these vendettas occur primarily because of the “absence of an official monopoly on sanctions,” with an emphasis on both monopoly and official(s), the fact that international law relies almost exclusively on self-help does indeed make it quite like a simple social order.Footnote ⁴³ Payandeh's observation that at present all states enjoy a right to enforce a limited number of international legal norms poses little challenge to this conclusion, both empirically and in terms of its fit with Hart's description of a primitive society as one in which enforcement is “left to the individuals affected or to the group at large.”Footnote ⁴⁴

Hart does briefly argue earlier in chapter 10 that reliance on self-help to enforce the law may pose less of a threat to the existence of a stable social order in the international context than it does in the domestic one.Footnote ⁴⁵ But he also points out that the same facts that explain why there is less need for an advanced legal system to produce a stable social order explain why it is less likely that such a system will develop in the international context, that is, why some and perhaps even many international actors will not judge it to be in their interest, all things considered. Note that this is an empirical claim and conditional on certain facts that may turn out to be contingent. It is consistent, for example, with the possibility that economic or environmental changes may lead states to engage in greater transfers of jurisdiction to international organizations (and the officials that populate them) than was true when Hart wrote The Concept of Law.Footnote ⁴⁶

With respect to adjudication, international law clearly differs somewhat from a primitive society in that it contains secondary rules, “empowering individuals to make authoritative determinations of the question whether, on a particular occasion, a primary rule has been broken.”Footnote ⁴⁷ Recall, however, that the defect that Hart believes courts and other adjudicatory bodies serve to mitigate is the fact that, “disputes as to whether an admitted rule has or has not been violated will . . . continue interminably, if there is no agency specially empowered to ascertain finally, and authoritatively, the fact of violation.”Footnote ⁴⁸ Insofar as courts without compulsory jurisdiction lack this authority except in cases where parties to a dispute voluntarily place themselves under it, Hart likely concluded that their existence marks only a small advance in international law's contribution to the production of social order.

He may have been mistaken on this score; arguably, a number of ICJ decisions have made significant contributions to how states understand their legal obligations, rights, and so on, and in so doing have shaped their conduct in ways that would not have occurred in the absence of a World Court. Moreover, the proliferation within international law over the past few decades of adjudicatory bodies with compulsory jurisdiction, not to mention increasing engagement with international law by domestic courts, is surely the area of governance in which international law has moved furthest away from a simple social order (or, if you prefer, closest to an advanced one).Footnote ⁴⁹ Nevertheless, if Hart's interest in the development of a division of labor in governance was motivated by the belief that it makes law a more effective tool for social control, then it was not implausible for him to conclude that a society that possesses only a court without compulsory jurisdiction was in this respect more like a social order that lacked any division of labor in adjudication than one equipped with compulsory courts of wide-ranging jurisdiction.

It is worth emphasizing here a point that both Hart and his critics make, though they then often ignore it: the property of being a primitive or advanced society is scalar, not bimodal.Footnote ⁵⁰ While Hart does sometimes group primitive and international law together in a way that suggests they are equivalent, at other times he offers a more nuanced characterization of international law. For example, in the introduction to chapter 10 he writes:

The argument in the preceding paragraph takes this talk of resemblance seriously, not least because it directs our attention away from the metaphysical-conceptual parlor game of determining, for its own sake, what is law, and toward the “more fruitful” and empirically informed study of how rules and institutions function to produce social order.

Hart takes up the question of whether international law contains a rule of recognition as a response to Kelsen's a priori assertion that it must. I discuss some of the details of that argument below. Here I consider the position Hart stakes out at its conclusion, namely that, “there is no basic rule providing general criteria of validity for the rules of international law, and that the rules which are in fact operative constitute not a system but a set of rules, among which are the rules providing for the binding force of treaties.”Footnote ⁵²

In the previous section I distinguish two functions that reference to a rule of recognition plays in Hart's analysis of law or a legal system. The first is an epistemic function: a rule of recognition provides (or purports to provide) an epistemically authoritative description of what counts as law in a given legal order; that is, an account of the sources of law in that society, possibly including a description of which laws take priority over others in the event of (apparent) conflict between them.Footnote ⁵³ Understood in these terms, the claim that international law possesses no rule of recognition is false, as both Payandeh and Waldron point out. For example, both note that Article 38(1) of the ICJ Statute and the Vienna Convention on the Law of Treaties provide authoritative means for identifying international law.Footnote ⁵⁴ As for systematizing international legal norms in the sense of identifying rules that regulate conflicts between them (or, perhaps better, specify their scope so that they do not conflict), Payandeh identifies the existence of a number of international rules that serve this function (albeit not on the basis of a law's source).Footnote ⁵⁵

Furthermore, both Payandeh and Waldron take Hart to task for making it a condition for international law to qualify as a legal system that it manifest a level of systematicity in this respect that does not obtain in many municipal legal orders, especially in the case of constitutional law.Footnote ⁵⁶ While all of these claims are correct, they fail to address Hart's assertion that international law lacks a rule of recognition and that therefore it does not count as a legal system. That is because in making this claim, Hart invokes the ontological, not the epistemic, function that reference to the rule of recognition plays in his analysis of law.Footnote ⁵⁷

Recall that in chapter 6 Hart maintains that a legal system exists when citizens or subjects generally obey the law while officials use it as a critical common standard of behavior. This “composite character” is what distinguishes a legal system from “a simpler decentralized pre-legal form of social structure.”Footnote ⁵⁸ A legal system, then, is a social order characterized by a significant division of labor in sustaining the practice of holding accountable that constitutes a given society's law. Put another way, what makes law (i.e., accounts for its existence) in a society that possesses a legal system is the practice of secondary rules of change, adjudication, and enforcement by one set of actors (officials) who rule over another set of actors (citizens).Footnote ⁵⁹ Given this characterization of what it is for a society to possess a legal system, Hart's assertion that the rules that comprise international law constitute not a system but only a set seems quite plausible. Indeed, neither Waldron nor Payandeh deny it, nor do most international legal theorists except those who mistakenly think international law's status as law depends on its possessing such a division of labor.

Of course, it is precisely this mistake that Hart aims to rebut in the final section of chapter 10.Footnote ⁶⁰ Thus Payandeh errs when he claims that “Hart offers no compelling reason why a legal system would have to closely resemble the archetype of the municipal legal order of a modern constitutional state.”Footnote ⁶¹ To count as a legal system in the sense that concerns Hart here just is to possess the division of labor in governance (e.g., a legislature, courts, etc.) realized in modern states and social orders that resemble them in this respect (e.g., the Roman Empire, or the Catholic Church).Footnote ⁶²

Hart not only denies international law the status of a legal system, he also asserts that it lacks a rule of recognition. Even if we restrict our interest to the ontological question of what accounts for the existence of law, this claim appears false. For we can formulate the following “rule” of recognition for a primitive society, a description of the conditions under which rules regulate (some of) the conduct of its members: rule R1, R2, and so on are rules of primitive society P if and only if its members “use the rules as standards for the appraisal of their own and others’ behaviour.”Footnote ⁶³ But in fact Hart does not deny that we can formulate such a rule, only the utility of doing so. Speaking of the “strange basic norm which has been suggested for international law: ‘States should behave as they have customarily behaved,’” Hart writes: “we may be persuaded to treat as a basic rule, something which is an empty repetition of the mere fact that the society concerned (whether of individuals or states) observes certain standards of conduct as obligatory rules”Footnote ⁶⁴

In contrast, where a society is characterized by a division of labor in the performance of governance tasks, the formulation of a “rule” of recognition shines new light on how law contributes to the production of social order by calling our attention to the fact that it is the conduct of officials, actors occupying offices constituted by rules that empower them to make, apply, and enforce the law, that accounts for law's existence. Likewise, in a primitive society the claim, “This is a valid law” amounts to nothing more than the empty claim, “this is a law of our society because it is a law of our society (i.e., because we treat it as such),” whereas once a division of labor in governance occurs, the claim “This is a valid law” can be theoretically informative; for example, “This is a valid law because it was enacted by Parliament.” Thus with some minor rephrasing we can better convey the idea Hart meant to express in the passage quoted above: there is no theoretically useful basic rule in the international legal order of the sort that could provide informative general criteria of validity. This reflects the fact that there is relatively little division of labor in the performance of the governance tasks that constitute international law. It is, in this respect, not a legal system and therefore bears a closer resemblance to a simple or primitive social order than to an advanced social order like the one realized in a well-functioning modern state.Footnote ⁶⁵

The foregoing interpretation of Hart's claim that international law lacks a rule of recognition and does not constitute a legal system casts an illuminating light on some of the remarks he makes in the course of his criticism of Kelsen. Admittedly this interpretation requires a departure from the text and, in particular, a focus on agents rather than rules. Nevertheless, I suggest that the arguments presented so far in this article warrant such a departure. To begin, rather than assume that Hart understands “rule of recognition” and “validity” epistemically and so conclude that he is talking past Kelsen, we should instead take Hart to be engaging directly with Kelsen by making an ontological claim regarding the existence conditions for norms.Footnote ⁶⁶ Contra Kelsen, Hart asserts that the existence of international legal norms or rules does not require a Grundnorm:

It is true, Hart concedes, that in this:

Take this last sentence first. I suggest that it should be read as follows: in a legal system, that is, a social order characterized by specialization in the performance of governance tasks, observers and participants can state correctly that R1 is a rule of that society in advance of any actor using it to hold him- or herself or another accountable. We can make this claim because there exists, and we can point to, a practice of legislation (including, in some cases, by courts) in virtue of which it is possible to make deliberate changes to the society's general rules. This is not possible in a simpler society, one that relies solely on a customary process of rule formation, in which successful rule change occurs only when most members of the society in question use or acknowledge the propriety of others using R1 to hold themselves and others accountable.Footnote ⁶⁹ Of course, we can formulate a rule that tells us how to identify new customary rules in advance of their being practiced; for example, we ought to believe in the existence of a customary norm C if (1) actors generally conduct themselves in ways that comport with it (or try to hide the fact when they do not), and (2) we have reason to believe they do so out of a sense of obligation or because they believe C is a norm that applies to them.Footnote ⁷⁰ However, Hart's claim here is ontological, not epistemic, or so I contend.

Turning now to the other sentences in the above quotation, the existence of a rule-governed social order does not require a hierarchy of agents, because the practice of holding accountable that constitutes such an order can be (and often has been) dispersed among all the members of the society in question. It is a mistake, therefore, to assume that we can, let alone must, formulate a theoretically useful “rule” of recognition, one that provides us with an informative description of what makes it the case that R1, R2, and so on are valid or “binding” rules for every rule-guided social order. Rather, such a “rule” “is not a necessity, but a luxury, found in advanced social systems”;Footnote ⁷¹ that is, in social systems characterized by a division of labor in the performance of governance tasks.

Liam Murphy speaks for many when he writes:

My goal in this section is to demonstrate that this disconnect is owed to Hart's presentation of his ideas but not to those ideas themselves.

IV

The defense of Hart's take on international law presented above has two important implications. First, in setting the record straight with respect to what Hart meant when he characterized international law as akin to a primitive society, it provides a basis for correcting certain other erroneous attacks on Hart and his analysis of law. One might hope (against hope?) that this contributes at least in some small degree to legal theorists’ collective effort to improve our theoretical understanding of law as well as our moral engagement with it. Second, and relatedly, Hart's discussion of international law highlights an important dimension of law's ability to contribute to the production of social order. In doing so, it adds further impetus to an already flourishing interdisciplinary investigation of international law's effectiveness. But perhaps more important, an appreciation of Hart's depiction of international law, properly understood, may encourage those legal and political theorists who argue for international legal action or reform to be more attentive to the differences between the international legal order and the legal system of a moderately well-functioning state.

Waldron considers ever so briefly the possibility that Hart means exactly what he says when he states that international law resembles a primitive social order in virtue of its lacking “secondary rules of change and adjudication which provide for legislature and courts.”Footnote ⁷³ His quick dismissal of it on the grounds of triviality is doubly mistaken, however.Footnote ⁷⁴ First, it is no small matter to concede that international law is largely devoid of specialization in the performance of governance tasks, given that one of Hart's goals in chapter 10 is to rebut those who argue or assume that the presence of such specialization is at least a necessary condition for classifying international law with municipal law as law properly so-called. Second, if a central aim of The Concept of Law is to provide an account of how law contributes to the production of social order, then it seems quite plausible to hypothesize that any particular legal order's ability to perform that task will depend to a significant degree on the extent of specialization in the conduct of governance tasks that characterizes it. Perhaps that observation is so obvious as to be trivial, but the many calls for international action and reform to the international legal order that display a failure to take seriously or even recognize the limits of international law's ability to shape the social world suggest otherwise. Moreover, Hart's observation that municipal law includes power-conferring rules is similarly trivial in that it does not expose us to a phenomenon with which we are unfamiliar but only calls our attention to a fact that we have momentarily lost sight of under the influence of bad legal theory.

The bad legal theory that Hart criticizes in his discussion of international law, the conceptual claim regarding law that threatens to inhibit productive theoretical inquiry and sound moral deliberation, is the position that only if it possesses those features characteristic of an advanced municipal legal system, such as a legislature and compulsory courts, can international law provide a degree of social control. That he makes such an argument renders somewhat ironic the fact that Hart is so often tarred and feathered for the alleged crime of drawing unwarranted conclusions regarding law tout court on the basis of reflection on municipal law alone.

Hart is also innocent of two further charges Waldron levels against him. The first, suggested by the title of the paper in which Waldron criticizes Hart's take on international law, is that the latter believes international law to be “a relatively small and unimportant part of jurisprudence.” Waldron plays fast and loose with the text of The Concept of Law here. What Hart actually says is that, “only a relatively small and unimportant part of the most famous and controversial theories of law are [sic] concerned with the propriety of using the expressions ‘primitive law’ or ‘international law’ to describe the cases to which they are conventionally applied.”Footnote ⁷⁵ Read in context, Hart clearly intends to dismiss as unimportant the question of whether rival theories of law sanction talk of international law as law, for, as he repeats several times in The Concept of Law, Hart is not concerned with definitions. More important, Hart's discussion of international law makes a vital contribution to the jurisprudential task he undertakes in The Concept of Law, namely making some headway in explaining how law contributes to the production of social order.

Specifically, in chapter 10 Hart offers arguments intended to reinforce three conclusions he has defended earlier in the book. First, the command theory's depiction of law distorts our understanding of it and so ought to be rejected.Footnote ⁷⁶ Second, law differs in important ways from morality, and so we ought to distinguish carefully between the question “What is law?” and any number of moral questions we might ask about it. While the answers to some of those questions might reveal certain necessary connections between law and morality, this is not the case with respect to the very existence of law.Footnote ⁷⁷ Third and finally, a division of labor in making, applying, enforcing, and identifying law and so in sustaining the existence of a given legal order likely enhances law's ability to contribute to the production of social order but it is not a necessary condition for the existence of a law-governed social order.Footnote ⁷⁸ No doubt it would have been nice had Hart told us more—say, about exactly how differences in the type of agents and circumstances subject to regulation by social norms affect the content of the rules and the kind of institutions that regulate (and partly constitute) particular societies. But insofar as Hart wrote The Concept of Law as an introduction to legal philosophy, carried out largely with the intention of clearing away mistaken legal theories that he thought impeded either theoretical inquiry or moral deliberation, he can hardly be faulted for his failure to do so.

The second baseless criticism Waldron levels against Hart is that the latter's discussion of international law is a cause of the fact that contemporary legal philosophers have contributed little to the debate over the nature and importance of international law sparked by actions undertaken by various officials in the George W. Bush administration as well as by academic works such as Jack Goldsmith and Eric Posner's Limits of International Law.Footnote ⁷⁹ Even if we accept arguendo that philosophers fail to engage with this debate-cum-political campaign, it is hard to see why we should attribute that fact to Hart's take on international law. After all, to the extent Goldsmith, Posner, John Yoo, and others understand (international) law along the lines of the command theory, Hart explicitly addresses and criticizes their position. And to the extent that these theorists challenge not international law's status as genuine law but the conditions under which it actually provides those it addresses with reasons to comply with it, Hart does not much engage with that question.Footnote ⁸⁰ Hart's decision not to take up in any detail in The Concept of Law the topics of political obligation and law's legitimate authority has not impeded contemporary legal and political philosophers’ engagement with them. Nor is it clear that reflection on international law's de jure legitimacy has or will generate new accounts of what makes law legitimate.Footnote ⁸¹

Brian Tamanaha seizes on Waldron's critique of Hart's depiction of international law to reiterate two long-standing objections to Hart's analysis of law advanced by sociolegal jurisprudents.Footnote ⁸² The first is Hart's characterization in chapter 6 of a simple (i.e., primitive) social order as a “prelegal” form of social structure. By implication, and ignoring any talk of resemblance as opposed to identity, sociolegal jurisprudents have concluded that Hart must also characterize international law as prelegal. The second and related objection is that in developing his analysis of the concept of law Hart accords an unwarranted priority to the municipal law of a well-functioning modern state, or worse yet restricts the concept of law to such a legal system, a conclusion suggested by the fact that as a matter of consistency Hart must depict international law as a “prelegal” social order.Footnote ⁸³ As should now be clear, the first of these two objections is misguided, and to the extent the second objection rests on the first, it, too, should be rejected. Properly understood, Hart's characterization in chapter 6 of a simple social order as “prelegal” means only that such a society lacks a division of labor in the performance of governance tasks, whereas his argument in the last section of chapter 10 aims to establish the propriety of categorizing international law as law properly so-called despite its being “prelegal” in this sense.Footnote ⁸⁴ Given his poor choice of words Hart undoubtedly bears considerable responsibility for these mistaken objections. Nevertheless, they ought to be discarded in light of the argument presented here.

What about Payandeh's assertion, noted in the previous section, that for purposes of assessing the (relative) simplicity or complexity of the international legal order, it is “more convincing” to focus on how it functions to address challenges of uncertainty, rigidity, and inefficiency than to ask “whether international law encompasses legislative, judicative, and executive structures comparable to the municipal system in form.”Footnote ⁸⁵ Unfortunately, Payandeh never explicitly states the standard by which we can judge as more or less convincing the depiction of international law as primitive. However, his exploration of the rules and mechanisms international law contains for addressing uncertainty, rigidity, and inefficiency suggests that it is shedding light on international law's ability to contribute to the production of social order that provides the standard Payandeh implicitly invokes. If so, then Payandeh's description of these rules and mechanisms largely supports Hart's characterization of international law as akin to a primitive social order, once that claim is properly understood. This is so because Payandeh documents the relative paucity of specialization in the performance of governance tasks that comprise international law and points to various important respects in which this distinguishes the effects international law can have on social order from those produced by the municipal law of a well-functioning state.Footnote ⁸⁶ It is rather ironic, then, that Payandeh repeatedly chastises Hart for implying that international law is inferior to municipal law, when the only respect in which Hart asserts this is the case is precisely the one that Payandeh also identifies.

Payandeh worries that Hart's characterization of international law as “a less developed and thereby inferior set of social rules” contributes, or at least could contribute, to a diminution of its legitimacy.Footnote ⁸⁷ For example, it might “lead political decision-makers—even if only subconsciously—to be more inclined to disregard the rules of international law when non-compliance is in their interest,” whereas domestic judges may ask “why should inferior international law trump conflicting norms of domestic law?”Footnote ⁸⁸ Yet here again, Hart is international law's friend, not foe. The importance of a particular system of social rules lies not just in how well it serves to produce social order but also in the matters it regulates. No necessary inference is warranted, therefore, from the relative paucity of specialization and hierarchical structure that characterizes international law to any conclusion regarding the role it ought to play in agents’ deliberations. This is the vitally important point with which Hart concludes his discussion of international law in chapter 10. Having argued that with respect to its “formal structure” international law is indeed inferior to municipal law, Hart points out that with respect to its content, “no other social rules are so close to municipal law as those of international law.”Footnote ⁸⁹ Like Bentham, Hart concludes that this fact warrants the treatment of both as examples of law properly so-called.Footnote ⁹⁰

By his own account, Hart sought in chapter 10 of The Concept of Law to defend only the modest claim that classifying international law with municipal law as law properly so-called was unlikely to “obstruct any practical or theoretical aim.”Footnote ⁹¹ I suggest a bolder claim is warranted: foregrounding the resemblance international law bears to a simple social order provides both theoretical and practical/moral benefits. The former consists in an improved understanding of how international law produces social order, including the limits of its ability to do so. The sustained cross-disciplinary engagement over the past few decades between international relations scholars and international legal theorists on the question of international law's effectiveness illustrates this point, as both rationalism and constructivism have been put to good use to illustrate some of the myriad ways in which international law regulates (or fails to regulate) the conduct of members of a largely horizontal social order.Footnote ⁹²

Though there is a place even in this explanatory endeavor for the deployment of conceptual apparatus and skills in which philosophers enjoy a comparative advantage, they are likely to play a more prominent role when the conversation shifts from a concern with explanation and prediction to a concern with justification. For philosophers, then, the biggest payoff from an engagement with Hart's reflections on international law may be the reminder from “one of their own” that law's utility as a tool for realizing some desired state of affairs varies quite considerably depending on whether it is the law of a well-functioning state or international law. Two important points follow. First, when they turn to questions of implementation, theorists of global justice should look to a broad range of institutional mechanisms to advance the goals or standards they defend and recognize that law, international or domestic, may sometimes provide a very weak tool for affecting the change they desire. Second, attempts to deploy international legal norms for argumentative purposes, as well as calls for reform of international law, ought to rest on a proper understanding of how it contributes to the production of social order, namely one that takes as its starting point the absence (by and large) of a legislature, compulsory courts, and centralized enforcement. Arguably, this is so not only practically speaking but also morally as well.Footnote ⁹³

The concept of a social rule is undoubtedly the centerpiece of Hart's analysis of law, and his characterization of law in terms of different kinds of rules marks a critical advance over Austin's reductive account. Perhaps it should not be surprising, then, that Hart relies so heavily on the language of rules when distinguishing primitive from advanced societies in The Concept of Law. As I demonstrate here, however, talk of rules often obscures another phenomenon that appears to have been of equal if not greater concern to Hart when he invoked that distinction, namely a division of labor in making, applying, enforcing, and identifying law and so in sustaining the social practice that constitutes a particular legal system. Understood in those terms, Hart's characterization of international law as resembling a simple or primitive legal order was true when he wrote it, and to a lesser but still considerable degree, it remains true today.