Are they immutable principles which constitute part of the fabric of the Universe, not made by man, but awaiting discovery by the human intellect? Or are they expressions of changing human attitudes, choices, demands or feelings? These are crude formulations of two extremes in moral philosophy . . . In what follows we shall seek to evade these difficulties.
—H. L. A. Hart Footnote 1
I. HART'S METAETHICAL AMBIVALENCE
Philosophers go “norm” in just the same way as cows go “moo,” the grouse goes;Footnote 2 to this, we might add—“legal philosophers too.” Moral requirements are taken to be “normative” or “binding.”Footnote 3 And this feature of morality, it is thought, is something that any respectable metaethical account had better adequately explain.Footnote 4 There is also another sense of the term “normative” at play in the literature—philosophers use the term “normative” to designate as semantically “normative” any statement purporting to convey what is “correct” or “required” in relation to some standard.Footnote 5 Owing to the potentially wide catchment area of this sense of “normative,” there is a whole world of standards that, though “normative” in the wide sense of the term, might fail to be so in the sense of normative qua moral bindingness.Footnote 6 Normativity in both the above senses becomes relevant in discussions of the so-called “normativity of law.” A comprehensive project of accounting for the normativity of law must explain: (a) how legal statements are semantically normative, and (b) how legal requirements that are conveyed by semantically normative legal statements come to be morally binding.Footnote 7 The two limbs are interconnected. Though our concern here will primarily be with the second limb of the project, some points at which it intersects with the first limb of the project will also assume relevance. Any attempt at determining how legal requirements come to be morally binding would presuppose taking a position on the metaethical debate about the nature of moral bindingness.Footnote 8
For the most part, two camps have dominated the metaethical landscape since the turn of the twentieth century, the cognitivists and the noncognitivists—the two metaethical “extremes” Hart alludes to in the epigraph. As Simon Blackburn figuratively puts it, the cognitivist camp asserts that our moral judgments are the children of Apollo, the god of reason, and the noncognitivists, that our moral judgments are the children of Dionysus, the god of passion, involving, as they do, passion, sentiment, attitudes, or “feelings,” as Hart put it.Footnote 9 The cognitivist Apollonian camp has an impinging picture of morality whereby moral truths impinge on us. The noncognitivist Dionysian camp argues that we project our sentiments onto the world as moral judgments.Footnote 10 The source of “normativity” or moral bindingness on the impinging model is objective moral requirements, while the source of normativity or moral bindingness on the projectivist model is subjective sentiments, feelings, attitudes. To adopt an expression Hart uses elsewhere, the Apollonian camp accuses the Dionysian camp of “blindness” to moral truths, with the latter retorting that the former has been “seeing ghosts” all along.Footnote 11
Notwithstanding the centrality of this metaethical debate to the project of accounting for the “normativity of law,” in The Concept of Law Hart resolutely refused to be drawn into it, declaring, instead, his metaethical nonalignment. That did not, however, prevent him from making some illuminating observations about the semantic limb of the project of accounting for the normativity of law.Footnote 12 But Hart's metaethical ambivalence was to disable him from presenting a clearly articulated and intelligible account of the other limb of the project, which is to say, the normativity of law qua moral bindingness. A symptom of this ambivalence, it will be claimed, was the paradox Hart confesses to in the closing pages of Essays on Bentham, in Chapter X, titled “Commands and Authoritative Reasons.” The paradox took the following shape. Hart holds that legal directives are “content independent” and “peremptory” reasons for action. However, he combines this with a semantic proposition ostensibly dissonant with it—that statements of legal duties, which convey these legal requirements, have “nothing directly to do with the subject's reasons for action.”Footnote 13 Hart realizes that the latter proposition sits oddly in a chapter named “Commands and Authoritative Legal Reasons” and, more worryingly, directly contradicts the chapter's central idea of legal directives constituting “content independent peremptory” reasons for its subjects. Hart admits that this position in which he finds himself may seem to be “paradoxical or even a sign of confusion.” He confesses that the paradox may have been caused because of his insufficient “grasp of many complexities surrounding the matter.”Footnote 14 Interestingly, even at this stage—Essays on Bentham (1982) was Hart's last major work—he can be seen to retain the long held metaethical nonalignment that finds expression in the epigraph above.Footnote 15
The central aim of this paper is to examine what causes the paradox Hart confesses to in Essays on Bentham. Broadly, the paper hypothesizes that this paradox was the result of incompatible philosophical commitments that had crept into Hart's account owing to his metaethical ambivalence. Bringing the hypothesis into sharper relief, it will be claimed that Hart's paradox was the result of eclectically combining, with disregard to their compatibility, elements from the two metaethically disparate models of “normativity of law”: the impinging model based on a cognitivist metaethic and the projectivist model based on a noncognitivist metaethic.
As a result of his metaethical ambivalence, it will be argued that Hart, at different points in time, advances two markedly different versions of the rule of recognition: one congruous with the impinging model, and the other incongruous with it.Footnote 16 Tracing the evolution of the rule of recognition from its earliest avatar in Legal and Moral Obligation (1958) through The Concept of Law (1961) to Essays on Bentham (1982), it will be argued that the version of the rule of recognition in Essays on Bentham, just as the version found in Legal and Moral Obligation, belonged to the impinging phase. This impinging version of the rule of recognition constrained Hart—due its inherent metaethical configuration—to commit to metaethical cognitivism and therefore postulate a “source” of normativity congruous with the impinging Apollonian picture of morality. Despite this, however, Hart was reluctant to postulate a corresponding cognitivist “source” of normativity. The “source” of normativity he himself consistently seems comfortable with—as we will see, Hart seems inclined to make an “attitude” the source of normativity—was congruous with the projectivist account: the noncognitivist, Dionysian picture of morality. Hart, it will be argued, was seeking to yoke together incompatible elements from two incommensurable metaethical models of normativity. This paper hypothesizes that the paradox that Hart finds himself in has its genesis in his metaethical ambivalence. Substantiating this hypothesis will be the principal task this paper undertakes. As a preliminary to that, however, something needs to be said about cognitivism, noncognitivism, and the impinging and projectivist models of normativity of law; the methods of operation of the two models need to be contrasted; and finally, the incompatible elements from two incommensurable metaethical models, the combination of which causes the paradox—which I claim hobbled Hart's account—need to be outlined. Given the challenges in satisfactorily accomplishing all of this a within the scope of the present paper, I will present the models of normativity of law, their method of operation, and the contrast between them somewhat elliptically, through a discussion of their instantiations in the literature. For this I will chose one candidate for each of the impinging and projectivist models: Kelsen for the impinging model, and the Scandinavian Legal Realists for the projectivist model.Footnote 17
I will begin in Section II with a brief metaethical primer on the two models of normativity or bindingness of morality and adumbrate the antinomical consequences of eclectically yoking together elements from the two models. Section III traces the genealogy of the rule of recognition, stretching from Legal and Moral Obligation (1958) through The Concept of Law (1961) to Essays on Bentham (1982), and argues that the version of the rule of recognition in Essays on Bentham is congruous with the impinging model alone. In Section IV, I will elaborate on the central hypothesis of this paper—that Hart's “paradox” stems from deeper underlying metaethical problems he never fully appreciated, which were a consequence of stacking up incongruous elements from the two models of normativity without sufficient regard for their compatibility. In Section V, it will be argued that the diagnosis Hart attempts shortly after confessing to the paradox altogether misses the mark as it does not address the underlying metaethical tangle. Section VI concludes by briefly touching upon the sort of metaethical reconfiguration that could be effected in Hart's account to dissolve the paradox.
Before proceeding further, it is of some importance from the point of view of delineating the scope of the hypothesis advanced here that two caveats are put in. The purpose of the first of the caveats would be to say something to allay a possible disquiet that the central hypothesis of this paper is likely to cause to anyone with a degree of familiarity with the Hartian oeuvre. How could Hart, for whom normativity had nothing to do with moral bindingness, have tied himself in such a metaethical tangle? Was not one of Hart's lasting contributions to jurisprudence finding a way of talking about normativity of law and obligation without the need to invoke moral bindingness while also avoiding the reductionism of classical legal positivism? The suggestion that Hart's account of normativity of law could trip on a metaethical hurdle would seem so implausible as to be “not even wrong,” to employ Wolfgang Pauli's famous derisive snub. In response, it must be admitted straightaway that if our focus were exclusively on The Concept of Law, there would perhaps be no metaethical problem of the kind under discussion here to contend with. It is beyond doubt that in The Concept of Law, Hart seems to concern himself primarily with what has been described here as the “semantic project” of accounting for the normativity of law: one of the two limbs of a comprehensive project of accounting for the normativity of law.Footnote 18 How he accomplishes this project has been fleshed out in some detail elsewhere.Footnote 19 In The Concept of Law, Hart does not explicitly pose the question of normativity qua moral bindingness, much less get into metaethical tangles with it. It must be remembered, however, that this, after all, is the tale of two Harts: one is the Hart of The Concept of Law; the other is the Hart of Essays on Bentham. The problems in Essays on Bentham must be understood on their own terms. It will not do to rely on the Hart of The Concept of Law to define out of existence some problem the Hart of Essays on Bentham faces. Indeed, it is argued here that some of the metaethical problems Hart himself identifies in Essays on Bentham—it is of some significance to note that he identifies them as metaethical problems—are missing in The Concept of Law. But just what is it that happens in Essays on Bentham? It will be argued—and the arguments will be backed up by citing chapter and verse—that in Essays on Bentham, Hart, running against the grain of much of his oeuvre up until that point, poses the problem of normativity qua moral bindingness and wants the rule of recognition to play a role in addressing it. To this end, he wanted the rule of recognition to do things that it never did in The Concept of Law, which is to say, he wanted it to be the source not just of normativity in the semantic sense, but also of normativity qua bindingness. Even though he never uses the term moral bindingness, it will be argued that the questions he tackles and the moves he makes are consistent only with someone who is seeking to make the rule of recognition a source of moral bindingness. The second caveat would bring to notice that the metaethical eclecticism discussed here has no relation to the hybridity in Hart's account of internal legal statements that was initially identified by Joseph Raz and subsequently brought into prominence by Kevin Toh's influential work in the last decade.Footnote 20 Toh argues that Hart's internal legal statements that expressed the speaker's approval of the rule of recognition were a “hybrid” that combined cognitivist as well as noncognitivist components. Toh's project relates to the semantic limb of the project of accounting for the normativity of law, and it does not purport to throw light on the paradox in Essays on Bentham, which, it is argued here, was the result of a metaethical eclecticism concerning models of moral bindingness—which matter pertains to the other limb of the project of accounting for the normativity of law. Footnote 21
II. THE RELEVANCE OF METAETHICS
On the cognitivist picture, moral judgments involve belief in objective moral facts.Footnote 22 On this view, in making moral judgments, we are responding to moral facts that are ontologically independent of the judgment and constitute the “source” of normativity. The “bindingness” or “normativity” of morality, on this account—the impinging model—is to be understood as the truth or correctness of moral judgments in relation to the objective moral facts.Footnote 23 On this model, as Blackburn describes it, moral facts “impinge” on the agent, and it is for the agent to cognize them.Footnote 24 The semantic model intertwined with this position is moral descriptivism, which is the thesis that moral statements purport to describe objective moral facts, which are the source of normativity. This relationship between the moral judgment and the objective moral facts also yields the idea of reason-givingness. To say that there are normative or binding objective moral facts is tantamount to holding that there are reason-giving moral facts.Footnote 25 The semantic model underlying the cognitivist model can also be understood as claiming that moral statements purport to state objective moral reasons for action. Cognitivism branches out into moral realism,Footnote 26 constructivism,Footnote 27 and voluntarism.Footnote 28 While all of the aforesaid versions of cognitivism share the semantic assumption (that moral statements purport to denote objective facts) and the psychological assumption (that moral statements express belief in objective moral facts), they differ on ontology, i.e., on the nature and status of these moral facts that the moral judgments purportedly track. The truth-makers on each of these branches vary—from ontologically given moral facts, as in the case of moral realism, to the say-so of an entity, as in the case of voluntarism.
On the noncognitivist picture, moral judgments do not involve any cognition of objective moral facts or moral truths. All there is to morality, noncognitivists claim, are acts of moral valuation that involve expressions of noncognitive (or conative) mental states such as attitudes, sentiments, or emotions.Footnote 29 The noncognitivists argue there is a noncontingent connection between our moral judgments and motivation, a connection that would be inexplicable if moral judgments involved beliefs in moral facts as the cognitivists claim they do.Footnote 30 The motivational connection can only be explained, noncognitivists argue, if moral judgments are understood as expressing motivationally laden attitudes. Accordingly, “normativity” or “bindingness” on this model is understood as the motivational pull exerted by moral judgments. On this picture, normativity does not impinge on us. Rather, it is we who “project” normativity by expressions of motivationally laden attitudes.Footnote 31 The “source” of normativity or bindingness on this projectivist model is the attitude of the subject. The projectivist model understands normativity as the motivational pull exerted by motivationally laden attitudes, the expression of which is a moral judgment. The idea of truth-makers or objective moral requirements constituting the source of normativity has no place on this model. It follows that the idea of reason-givingness too has no place on this account.Footnote 32
The metaethical debate outlined here has an immediate bearing on discussions about the “normativity of law.” One of the significant enquiries under the rubric of the “normativity of law” (the second limb) revolves around the question: how do legal requirements come to be morally binding?Footnote 33 Any attempt at answering this question necessarily presupposes taking a position on the metaethical debate about the nature of moral bindingness. As there are two plausible ways of understanding “moral bindingness,” it stands to reason that there are two plausible ways of understanding “normativity of law” and accounting for it. One is the account that rests on an impinging model; the other is the account resting on a projectivist model. The semantic thesis congruous with the impinging model of normativity of law would be a replica of the semantic model of the impinging metaethical model—statements of legal obligation would purport to denote objective moral reasons for action. These statements would be truth-apt in relation to the kind of truth-maker underlying the model. The impinging model makes it the case that legal requirements come endowed with normativity (reason-givingness) and the statements of legal obligation conveying those requirements relay that normativity and reason-givingness. Therefore, statements of legal obligation on the impinging model must purport to state these moral reasons for action. But what if they did not? This would lead to antinomy and the account that suffers it would wobble.Footnote 34 The projectivist model would not countenance this semantic model, on which legal statements purport to state moral reasons for action, as it would be incongruous with the projectivist model. The question of precisely what shape the semantic model in a projectivist model need assume need not detain us here.Footnote 35 It would suffice to say for our purposes that the model cannot be one that holds that statements of legal obligation purport to state moral reasons for action. While there are many shapes the impinging model can take, we will here focus on one instantiation of the impinging model, namely, Kelsen's account of normativity of law, and contrast it with the instantiation of the projectivist model, namely, the Scandinavian Legal Realists’ account of normativity of law.
Kelsen equated the ideas of “normativity” and “validity.” Valid law, according to Kelsen, was normative. For Kelsen, “normativity” meant “bindingness” of law, which is to say, the idea that the law ought to be obeyed.Footnote 36 In his theory, the Grundnorm was the source of normativity, transmitting bindingness to all norms of the system. Alf Ross argued that the “obligation to obey” that Kelsen's Grundnorm purported to impose could not but mean “moral obligation,” however hesitant Kelsen—the self-styled legal positivist—would be to acknowledge it as the fount of moral obligation or moral bindingness.Footnote 37
Binding force it is said, means you are duty-bound to obey the law. What kind of duty is meant here? It seems obvious that the duty to obey the law cannot mean here a legal duty or obligation. . . A [legal] duty is always a duty to behave in a certain way. In this case the required behaviour is “to obey the law”. . . if the duty to obey the prescriptions of the legal system is to mean something different from the obligation prescribed directly by this system, then the difference cannot consist in the required behaviour—what we are bound to do—but must consist exclusively in how we are bound.Footnote 38
According to Ross, the rule prescribing an obligation to obey cannot be another legal requirement, as no legal requirement can create causa sui an obligation to obey itself. Accordingly, Ross reasoned, “it is a moral duty toward the system, not a legal duty conforming to the system”;Footnote 39 Ross's reading of Kelsen is of older vintage. Karl Olivecrona argued that Kelsen's account treated law's bindingness as stemming from a “higher” standard, and that such a model necessarily presupposed objective moral values, which was a presupposition typically associated with natural law thought, although Kelsen himself never expressly made such a claim.Footnote 40 The only real difference Olivecrona found between unalloyed natural law theory and Kelsen's version was the identity of the supreme objective standard from where law derives it binding power.Footnote 41 The natural law model Olivecrona specifically seems to have had in mind, when he compared Kelsen to the natural lawyers, was the voluntarist one. Voluntarism is a cognitivist position in that it claims that moral judgments purport to denote moral facts.Footnote 42 However, according to the Voluntarists, it requires an entity—God or Sovereign—to impose them. The truth is fixed by the say-so of such entity.Footnote 43 The say-so of this entity at the helm of the system creates moral obligations to obey. Another way of putting this is that in voluntarism, the “truth-maker” is the say-so of some entity.Footnote 44 Early modern natural law accounts since the enlightenment, including Samuel Pufendorf's and Thomas Hobbes's, followed the voluntarist template.Footnote 45 Pufendorf did not find any values ontologically given in nature.Footnote 46 They had to be “imposed” by the will (voluntas) of an intelligent being.Footnote 47 Hobbes maintained that moral “obligation must come from law, and law from the will of the legislating sovereign.”Footnote 48 Both Olivecrona and Ross saw Kelsen's model of normativity as an iteration of the voluntarists’ impinging model of bindingness typically pressed into service by natural law theorists, the only difference being that the Grundnorm, which created the obligation to obey the law, replaced the say-so of the entity at the helm of the system. Olivecrona's observations about the workings of the voluntarist natural law model, which he saw reflected in Kelsen's model of normativity, are illuminating for much of what is being argued in this paper.
A command cannot be conceived as being obligatory unless issued by an authority possessing the right to command. This authority cannot, however, create such a right through its own commands. The right can only be based on a higher principle whose validity is independent of the authority in question. This leads us straight into the sphere of natural law. The right of the state to command must be derived from principles of natural law, whether this name is used or not.Footnote 49
Olivecrona's argument remains relevant for any account of moral bindingness of law that presupposes a right in the lawgiver to issue commands or a power in him to morally bind subjects, and a corresponding moral duty of obedience in the subject presupposes a “higher principle”—something that we have described here as being presupposed by the impinging model of moral bindingness.Footnote 50 Those who were inclined to read Kelsen along voluntarist lines were also quick to point out that there arose in Kelsen's account a problem typically associated with voluntarism.Footnote 51 Voluntarist accounts treat an entity's say-so as fixing the binding objective moral standard but do not satisfactorily explain what exactly ultimately grounds or justifies obedience to the entity's say-so.Footnote 52 Similarly, the critics believed, the question why one ought to obey the Grundnorm is never addressed in Kelsen's account.Footnote 53 Scandinavian Legal Realists thought that Kelsen's account, although not professedly endorsing the natural law position, nevertheless retains natural law's—what they found questionable—metaethical assumptions. The nub of the methodological error, Hagerstrom located in presupposing some objective standard that “endows law with binding power.”Footnote 54 Translated to terms employed here, Hagerstrom and the Scandivanian Legal Realists found the methaethical assumptions of the impinging model to be faulty. And the solution, they argued, was jettisoning that model of moral bindingness altogether and looking for a model that was metaphysically less profligate: one that did not explain moral bindingness by reference to objective moral standards. The Scandinavian Legal Realists sought to accomplish this by “naturalizing” the normativity of law, for which they relied on a noncognitivist metaethic.Footnote 55 The Scandinavian Legal Realists argued that the moral bindingness of law was to be understood not in accordance with an objective moral standard that impinges on the subject—which their metaphysical austerity did not permit them to countenance—but rather, in accordance with the projectivist model of bindingness, which is to say, in terms of a noncognitivist moral stance taken by subjects toward the law.Footnote 56 Not only did noncognitivism play a very central role in the Scandinavian Legal Realists’ project but, in fact, twentieth-century noncognitivism too was influenced, to some extent, by the Scandinavian Legal Realists, Hagerstrom in particular.Footnote 57
Given that the two models of the “normativity of law” are diametrically opposed, a putative model that hopes to eclectically combine incompatible elements from the two is bound to run into antinomy.Footnote 58 It would not be practicable to set out here the many shapes an antinomy of this kind could assume. Suffice it to say, for our present purposes, that an antinomy is inevitable when a putative model claims that the source of normativity is a conative attitude, but envisages that the “source” of normativity operates as the fount of reason-givingness as it does in the impinging model. It is just such an antinomy that we will find underlying Hart's paradox.
III. TWO VERSIONS OF THE RULE OF RECOGNITION IN THREE PHASES
The rule of recognition always remained at the very core of Hart's legal philosophy and his project of accounting for the normativity of law. What did vary, however, was the metaethical valence of the rule of recognition. Hart's thoughts on the nature of the rule of recognition and the normativity of law can be seen as having proceeded in three distinct phases: the first phase being Legal and Moral Obligation (1958); the second, The Concept of Law (1961); and the third, Essays on Bentham (1982). As a result of his metaethical ambivalence, Hart at different times seems committed to two different models of the rule of recognition: one congruous with the impinging model, and the other incongruous with it. The impinging version of the rule of recognition purports to act just as the Kelsenian Grundnorm does, which is to say, it purports to transmit bindingness to laws deriving its validity from it, and creates an obligation to obey such laws. The impinging version of the rule of recognition is to be found in Legal and Moral Obligation (1958) and Essays on Bentham (1982). His portrayal of legal directives as “peremptory” “content independent” reasons for action in Essays on Bentham belonged to one such “impinging” phase. The idea of a lawgiver's directives giving reasons for action could only ever be congruous with the impinging model of “normativity of law.”
Between Legal and Moral Obligation (1958) and Essays on Bentham (1982), Hart wrote The Concept of Law (1961), where the rule of recognition does not operate as the impinging Kelsenian Grundnorm does, and is thus non-impinging in nature.Footnote 59 However, even in The Concept of Law, Hart very nearly had an impinging rule of recognition like he did in Legal and Moral Obligation and Essays on Bentham. The expression “nearly” is used advisedly. Although Hart looks set in The Concept of Law to make his rule of recognition act as an impinging device, like the Kelsenian Grundnorm, he eventually desists from doing so.
A. The Rex Model in The Concept of Law
Hart introduces the rule of recognition in The Concept of Law with a discussion about the simple legal system with Rex at its helm.Footnote 60 Hart tells us that in a simple legal system composed of Rex and his subjects, the citizens “accept” the rule of recognition—“what Rex says is to be done.”Footnote 61 Hart adds that such acceptance of the rule of recognition is “normative.” In the Rex model—where the rule of recognition is “accepted” by the citizens—what Rex says, Hart argues, will not only be valid but will also generate “obligations” of obedience.Footnote 62 Rex will not only specify what is to be done but will have the “right” to do so.Footnote 63 The “acceptance” of such a rule of recognition, says Hart, will invest Rex with “authority.”Footnote 64 It could be argued that not only does the rule of recognition in the Rex model set out the criteria for valid law but, like the Kelsenian Grundnorm, it has an additional justificatory task—it also purports to generate an obligation to obey Rex's directives.Footnote 65
Hart, however, does not eventually employ this impinging model of the rule of recognition in The Concept of Law. The new model of the rule of recognition that actually holds the field in the remainder of The Concept of Law is of the form “whatever Rex enacts is law.”Footnote 66 It no longer reads—“what Rex says is to be done.” This formulation rule of recognition only sets out the criteria for “identifying” valid law; it does not perform the “justificatory” task that the rule of recognition in the Rex model had assumed. Footnote 67 But why does Hart abandon the impinging model of the rule of recognition in the remainder of The Concept of Law? Briefly tracing the intellectual history of the rule of recognition will be indispensable in order to understand Hart's move.
B. The Impinging Rex Model in Legal and Moral Obligation
Hart first introduced the idea of the rule of recognition in Legal and Moral Obligation (1958).Footnote 68 The rule of recognition we find here is no different in nature from the one in the Rex model in The Concept of Law—its content being what the legislator says “is to be obeyed.” Like in the Rex model, the rule of recognition here imposes an obligation on the subjects to obey the lawgiver, and is reminiscent of Kelsen's impinging Grundnorm. For reasons discussed later in the section, there is little sense in understanding this obligation imposed by the rule of recognition as anything other than a moral obligation. In Legal and Moral Obligation, Hart makes explicit the affinity of his rule of recognition with Kelsen's Grundnorm. He prefaces his account of the rule of recognition with a discussion of Kelsen's Grundnorm. He agrees with Kelsen's criticism of the Bentham-Austin model, and also says nothing to dispute the role that the basic norm plays in the legal system. The only point on which he differs from Kelsen is latter's insistence that the basic norm is a “hypothesis” or is “postulated.” Hart thought such vocabulary to be confusing as it needlessly brought in “a set of quasi mathematical expressions.”Footnote 69 While appearing to favor retaining the structure and function of the basic norm, Hart proposes that the basic norm needs to be grounded in the fact that it is “accepted” in a practice of the citizens:
If a group of persons behaves in the way we have distinguished as accepting what a legislator says as constituting a standard of behaviour, then they accept the rule that he is to be obeyed and then the rule that his word is law exists.Footnote 70
In the above passage, Hart sets out two different roles for the basic norm or what he calls the “rule of recognition.” The first prong of the rule of recognition as formulated by Hart has it that the “legislator is to be obeyed” and the second prong has it that the “legislator's word is law.” This dual-pronged rule of recognition—which also requires obedience to the legislator—bears close resemblance to the Kelsenian basic norm that stipulated that “any law deriving its validity from the rule of recognition ought to be obeyed.” The Kelsenian Grundnorm, it will be recollected, was the fount not only of validity but also of normativity (bindingness) and therefore performed both the roles set out by Hart for the rule of recognition. It must be noted that there is a great qualitative difference between a basic norm that sets out “what the law is” alone and one that not only sets out what the law is but also specifies that the “law ought to be obeyed.”Footnote 71 Twenty-five years later, Hart goes on to use the very same model of the rule of recognition in Essays on Bentham. Also, this is just the model that Hart sets out to advance in The Concept of Law but does not eventually do so. Hart would most certainly have employed it, had it not been for a fortuitous intervention by Graham Hughes in 1960, which we will now briefly turn to.Footnote 72
C. Hughes's Intervention and Hart's Rejection of the Rex Model
In 1960, Graham Hughes criticized the Rex model of the rule of recognition to be found in Legal and Moral Obligation for reasons that had nothing to do with either the rule of recognition's impinging nature or its affinity to the Kelsenian Grundnorm. Hughes's criticism was that Hart's model of the “acceptance” of the rule of recognition was practically implausible in complex modern legal systems as it demanded of the citizens knowledge of constitutional law to an extent well beyond their grasp.Footnote 73 Hart appears to have taken this criticism to heart. Hart admitted that while the citizens in the simple system, such as Rex's, may be cognizant of the rule of recognition, citizens in complex legal systems are ordinarily in no position to be cognizant of it. Thus, while the citizens in the simple system such as Rex's may “accept” the rule of recognition, citizens in complex legal systems were in no position to “accept” it. Hart grants that this model of acceptance would “involve putting into the heads of ordinary citizens an understanding of constitutional matters which they might not have.”Footnote 74
Consequently, Hart breaks up the “acceptance” requirement between the citizen and the legal officials. The officials manifest their “acceptance” by identifying and applying rules in accordance with the rule of recognition, and citizens manifest their acceptance largely by “acquiescence” in the results of these official operations.Footnote 75 When citizens and officials “accept” in this fashion, we have the conditions necessary and sufficient for the existence of a legal system. It should be noted that here Hart does not expect the officials to morally endorse the rule of recognition. His emphasis is on the fact that it is only officials who can be cognizant of the rule of recognition and as long as they identify and enforce laws in accordance with it, the rule of recognition is sustained.Footnote 76
But the one upshot of Hughes's objection was something that Hughes did not speak of and Hart did not seriously emphasize. When Hart tweaks the existence conditions of the rule of recognition, in response to Hughes's objection, he introduces an altogether different model of the rule of recognition. The new model of the rule of recognition is of the form, “whatever Rex enacts is Law.”Footnote 77 It no longer reads, “what Rex says is to be done.” This new rule of recognition that holds the field in the remainder of The Concept of Law only sets out the criteria for “identifying” valid law; it does not perform the justificatory task that the rule of recognition in the Rex model did.Footnote 78 The Kelsen-inspired normativity that Hart had built into the rule of recognition in the previous Rex model is dropped. In other words, this rule of recognition stops being an impinging one. The rule of recognition that Hart had envisioned in the Rex model would have made the rule of recognition do the exact normative load bearing that the Kelsenian Grundnorm did. Hart would probably have continued walking on that road had it not been for the fortuitous intervention by Hughes. As a result, The Concept of Law is free from this Rex model. However, as we shall see in the next section, in Essays on Bentham, Hart returns to this impinging Kelsenian model of the rule of recognition.
This still leaves us with a puzzle: Why does Hart drop the impinging model of the rule of recognition while tweaking the existence conditions of the rule of recognition? Why should the fact that it is the practice of officials rather than of citizens that grounds the rule of recognition in a complex legal system lead Hart to drop the impinging nature of the rule of recognition? Hart never gets around to exploring the reasons for doing so anywhere in the The Concept of Law. He only returns to the theme many years later, in the last few pages of Essays on Bentham, where he tries to diagnose the self-confessed “paradox.” We will return to discussing the significance of this move in Section V, where it will be argued that Hart misdiagnosed the reasons for the paradox.
D. Return to the Rex Model in Essays on Bentham
In Essays on Bentham, Hart atavistically reverts to the impinging Rex model of the rule of recognition, which he had endorsed in Legal and Moral Obligation, and had floated in The Concept of Law, before going on to reject it. In Essays on Bentham, explaining his motivations for re-introducing the Rex model, Hart tells us that he seeks to employ the Hobbesian voluntarist model in explaining the “normativity of law.” Along with it, he also introduces the idea of legal directives being “content independent,” “peremptory” reasons for action: a set of ideas that were clearly absent from his philosophical repertoire in The Concept of Law.Footnote 79
[T]his modern analysis of meaning was first suggested to me by Hobbes who said some simple but illuminating things about commands and the similarity and differences between commands and covenants as sources of obligation or as obligation creating acts. But I do not think I should have seen the full importance of Hobbes’ remarks on these topics had I not had the benefit of the work of Joseph Raz on what he terms ‘exclusionary reasons’ which resembles in many respects the notion which I have taken from Hobbes [i.e., peremptory reasons].Footnote 80
A legal directive is a content independent reason for action, which is to say, the commander's directives are meant “to function as a reason” for the subject “independently of the nature or character of the actions to be done.”Footnote 81 And the legal directive is a peremptory reason in that “the commander's expression of will is. . . not intended to function in the hearer's deliberation as a reason for doing the act, not even as the strongest or dominant reason. . .the commander intends to cut off or exclude it [deliberation about reasons].”Footnote 82 Hart then goes on to advance a model of the rule of recognition, which like the Rex model and the Kelsenian Grundnorm would render it an impinging device:
The general recognition in a society of the commander's words as reasons for action is equivalent to a social rule. . .this rule might be formulated as the rule that the commander is to be obeyed and so would appear as the rule imposing obligations on the subjects. Regarded in another way as conferring authority on the commander. . .Footnote 83
Note the phraseology of the rule of recognition here—“the commander is to be obeyed.” This resembles the phraseology of the rule of recognition in the Rex model in Legal and Moral Obligation and in the abandoned Rex model in The Concept of Law, and indeed that of Kelsen's Grundnorm. Footnote 84 Hart makes it plain that this rule of recognition imposes an obligation of obedience on the subjects. He also stresses the correlative authority it confers on the lawgiver. The laws made by the commander are not only valid but also are authoritative, are normative, and give rise to an obligation of obedience for the subjects.
E. The Impinging Rule of Recognition in Essays on Bentham
Now, how is one to understand such an obligation imposed by the rule of recognition? Could this just be a legal obligation? Or is there warrant for understanding such “obligations of obedience” as moral obligations, though Hart never specifically refers to them as such? Is there any sense in understanding them as strictly legal obligations? In this context, the comments of Ross and Olivecrona on the “obligation of obedience” imposed by Kelsen's Grundnorm are pertinent, for their reasoning also holds salience in the case of the “obligation of obedience” imposed by any foundational rule such as Hart's rule of recognition in Essays on Bentham. In making this argument, Ross's and Olivecrona's treatment of Kelsen is not simply being carried over to Hart's case on the back of the assumption that Hart's rule of recognition is, in some sense, similar to Kelsen's Grundnorm. The point on which they are being invoked is their argument that no legal requirement can create causa sui an obligation to obey itself. This power, they argued, must emanate from some higher moral principle—the source of normativity—whether or not such higher principle is openly invoked. The rule of recognition in Essays on Bentham being impinging in nature purports to endow the lawgiver with the normative power—or “right” in Olivecrona's terminology—to issue directives and purports to create a corresponding moral duty of obedience in the subjects. The “obligation of obedience” that the rule of recognition in Essays on Bentham purports to impose must then be understood as a moral obligation, though Hart nowhere describes it as such.Footnote 85 Hart's invocation of Hobbes as an early harbinger of his account in Essays on Bentham is very instructive, and it shows us that Hart must have been mindful that the obligation of obedience his model in Essays on Bentham purported to set out was—at least ostensibly, if not more—a moral obligation. Consider the following passage in Essays on Bentham, redolent of Olivecrona's argument against Kelsen where Hart seems to recognize that the ideas he borrows from Hobbes presuppose a moral obligation to obey the lawgiver, something that Bentham was fervently opposed to:
Hobbes, like Bentham thought that all laws were commands of a Sovereign but, unlike Bentham thought the commands were laws only if given to those who were under a prior obligation to obey, and his account of this prior obligation was that it arose from the subjects covenant or contract to obey the commander. . . Bentham however would have none of this prior obligation to obey nor of the social contract alleged to generate it nor of the idea that in making laws the Sovereign was exercising the right or normative power. . .Footnote 86
The above passage is one of the two places in Essays on Bentham Hart hints at having been aware that his account of normativity of law operated in manner that purported to generate moral reasons for obedience. The other is where Hart confesses to the paradox in his account. What Hart finds “paradoxical” or “a sign of confusion” is the dissonance between two propositions in his account. The first proposition (let us call this Proposition A) is: laws are “content independent peremptory” reasons for action, the central theme of Chapter X of Essays on Bentham. And that is in dissonance with the semantic proposition (let us call this proposition B): statements of legal obligation “have nothing directly to do with the subject's reasons for action.”Footnote 87 In Chapter VI of Essays on Bentham (pages 159–160), Hart clarifies that the statements of legal obligation “do not refer to actions which they [subjects] have a categorical reason to do.”Footnote 88 Hart holds—which is borne out by his comment on what he takes to be Raz's “cognitive” account of dutyFootnote 89 —that statements of legal obligation could refer to actions that subjects have categorical reasons to do, only on an account, such as Raz's, which equated legal obligation with moral obligation. Hart, however, remains strictly and consistently against any equation of legal with moral obligation.Footnote 90 Even in Chapter X (pages 266–267) Hart makes clear that he means to stick to the very same semantic proposition he advanced in Chapter VI of Essays on Bentham. From the discussion in Chapter VI (pages 159–160) and Chapter X (pages 266–267) it becomes abundantly clear that the “reasons” referred to in proposition B (the semantic proposition) are objective moral reasons because only on such a reading would a statement of legal obligation amount to a statement of moral obligation. In other words, Hart's semantic proposition must be understood to read: “statements of legal obligation do not state any moral reasons for action.” Now, given that the reference to “reasons” in proposition B (the semantic proposition) is to moral reasons, the proposition would not be dissonant with proposition A (the proposition that laws are “content independent peremptory reasons”) unless “reasons” in proposition A referred to objective moral reasons. It is only then that a statement of legal obligation, relaying a reason-giving directive, would amount to a statement of moral obligation, which is something that Hart was opposed to.Footnote 91 In fact, there would be no paradox if the “content independent peremptory” reasons were not to be read as moral reasons.
The immediate cause of the paradox was that Hart endorsed a model of the rule of recognition that operated along the lines of the impinging model and purportedly generated moral reasons for obedience, but at the same time denied the semantic model that would have been congruous with such a model, namely, the model where statements of legal obligation conveyed a moral obligation. At the surface, this paradox is of a piece with the antinomy attributed to Kelsen by Bulygin.Footnote 92 In the next section we will see how Hart's metaethical ambivalence lay behind this paradox. It will be argued that Hart's paradox arose from eclectically trying to fit a projectivist model of normativity into an impinging mold.
IV. HART'S ECLECTICISM AND THE PARADOX
Let us begin with the crucial metaethical question Hart puts to himself but never goes on to engage satisfactorily: How could the directives of a lawgiver ever give rise to reasons for action without the grounding of an “ulterior” moral standard? Echoing Olivecrona, Hart puts to himself the all-important metaethical question:
How can an artifact of the human will such as a command or compliance with a legislative procedure either in itself be or be believed to be a reason for action? Surely, the critic may urge, such products of the human will could only be such a reason if there were some artificial ulterior reason for taking the former as guides to action. . .Footnote 93
Hart seems to have been aware that his impinging model of normativity ostensibly requires an “ulterior” objective moral standard to act as truth-maker and be the “source” of normativity, without which there could be nothing to generate the reason-givingness that his model presupposes. However, oddly, while Hart's model of normativity openly invokes the idea of reason-givingness, he also seems uneasy with the idea of some objective moral standard as the “source” of normativity.Footnote 94 What then on Hart's account is the “source” of normativity if not an ulterior objective moral standard? Though Hart does not specifically provide us with an answer, we are in a position to decipher it from Hart's description of the rule of recognition that he gives earlier in the chapter. The “source” of normativity, for Hart, it turns out, is not some objective truth-maker but is “acceptance” by citizens.
Such a standing recognition (which may be motivated by any of a variety of ultimate reasons) of a commander's words as generally constituting a content independent peremptory reason for acting is a distinctive normative attitude . . . and in my view this is the nucleus of a whole group of related normative phenomena, including. . . the general notion of authority. . . Footnote 95
In the above passage, Hart makes it clear that the source of authority on his model is the “recognition” of the commander's words as constituting content independent reasons. He also makes it abundantly clear that this “recognition” is an “attitude”: a normative attitude. Here, in Essays on Bentham, Hart uses the terms “acceptance,” “recognition,” and “normative attitude” interchangeably.Footnote 96 He points out, in no uncertain terms, that this “attitude” or “acceptance” is the “source” of normativity and authority.
Hart seems to have held the view that his model does not need an ulterior moral standard as the source of normativity because the attitude toward, or acceptance of, the rule of recognition by citizens serves to constitute the “source” of normativity. Hart thus gives the impression of having wanted to use the idea of reason-givingness of the impinging model, while at the same time denying the idea of an objective moral standard as the “source” of normativity. However, the idea of reason-givingness is inextricably intertwined with the idea of objective moral standards that act as truth-makers.Footnote 97 Having rejected the idea of an objective moral standard as the source of normativity, Hart could not consistently have advanced a model of normativity that employed the idea of reason-givingness.Footnote 98
Hart's move of seeking to explain authority in terms of a normative attitude taken by the citizen is consistent with the projectivist model alone. The projectivist model understands authority in terms of the motivational attitudinal states involved in making a moral judgment. On the projectivist model, for the law to have authority for an agent is for her to take a noncognitivist moral attitude of approval or acceptance toward it. It is only on a projectivist account that an attitude could be thought to constitute the source of “normativity” or “authority.” However, there is considerable difficulty in straightforwardly fitting Hart's model of normativity in Essays on Bentham into a projectivist mold. The idea of reason-givingness, which Hart stresses repeatedly in his account in Essays on Bentham, is alien to the projectivist model; it is congruous with the impinging model alone. In other words, Hart seems to have wanted a projectivist model with the projectivist source of normativity but that in every other respect operated like the impinging model and gave reasons for action. It is this metaethical eclecticism that was the real cause of the paradox in his account.Footnote 99 The projectivist model explains “normativity of law” entirely in terms of moral attitudes taken by subjects toward the law.Footnote 100 Despite having wanted to advance a model of normativity that had a projectivist source of normativity, Hart does not seem to have been sensitive to the distinctive method of operation of the projectivist model. From Hart's discussion, it would appear that he wanted to advance a model of normativity with a projectivist “source,” but wanted it to generate reason-givingness.
V. HART'S WRONG DIAGNOSIS
Shortly after confessing that he did not have a “sufficient grasp of many complexities” surrounding the problem, Hart seeks to diagnose why he ended up with the paradox. Hart supposes that the paradox came about entirely because of his having argued on the one hand that judges need not “accept” legal directives as content independent and peremptory reasons for action while having argued on the other hand that the citizens accept the lawgivers’ directives as reasons for action.Footnote 101 In other words, the paradox, Hart thinks, arose because of his having “whittled down” the nature of the acceptance; in simple legal systems there is acceptance by subjects of the commanders’ directives as content independent peremptory reasons, whereas there is no such acceptance (by judges or officials) in a complex legal system.
Hart further supposes that the “whittling down” may have come about as a result of the transition from the simple legal system of Rex to a complex legal system. He says the whittling down is something “for which I had made express provision on p. 257. . . that the introduction into the simple society of specialized law-applying and law-enforcing agencies would mean the institutionalization of the recognition of the commander's authority. . . and this would transform the situation depicted in the model.”Footnote 102 But why should the complexity of the system affect the model?
Where courts with these functions exist the normative attitude consisting in the recognition of the commands as content independent peremptory reasons for action and as standards for the assessment of conduct as right or wrong is itself institutionalized as defining public standards of correct adjudication and a duty to conform to these standards is attached to the office of the judge and assumed by individual judges when they take up that office.Footnote 103
Hart's explanation of what happens in the transition from the simple to the complex legal system is the following. In the simple legal system of Rex, citizens “accept” or “recognize” the words of the commander as constituting content independent peremptory reasons for action. The situation, however, is transformed in a complex legal system. A complex legal system has courts, and where the courts identify and enforce the law, in accordance with the rule of recognition, the legal system is in place. But judges may enforce the law for any reason whatsoever, indeed oftentimes their position as judges, and the settled institutional practice they come into when they assume office, precludes them from deciding on the moral soundness of the law they are enforcing. In other words, the judges who enforce the law need not accept the laws as content independent peremptory reasons. This is why, Hart thinks, the “whittling down” in the nature of the acceptance takes place.Footnote 104 And this, according to Hart, explains the paradox he finds himself faced with.
We are now in a better position to appreciate why Hart thought the transition from the simple to the complex legal system led to a “whittling down” of the nature of acceptance, thereby, in his view, causing the paradox. Hart's explanation for the “whittling down” of the acceptance is based on one significant assumption that he never explicitly articulates: that citizens in a simple legal system (unlike the judges in a complex system) morally “accept” the rule of recognition. But this as we shall now see was an unwarranted assumption as even in a simple legal system the citizens do not necessarily morally accept the rule of recognition, and hence there is no “whittling down” in the nature of acceptance in the transition from the simple to the complex system.
The rule of recognition—whether or not it is charged with the additional justificatory function—constitutes one of the most crucial existence conditions of a legal system. Now, the sort of “acceptance” that constitutes the rule of recognition (regardless of whether its nature be made impinging or nonimpinging) in the simple legal system of Rex cannot be moral acceptance. Insisting on the “acceptance” constituting a rule of recognition necessarily being moral acceptance would have the effect of rendering the existence conditions for the legal system too onerous. Hart would then have had to insist that a simple legal system with Rex as the king exists only if citizens (or a sizeable number of them) morally accept Rex as the lawgiver. It is quite obvious that Hart never wanted to do this. In Legal and Moral Obligation, Hart had specifically ruled out that the “acceptance” constituting the rule of recognition in the simple Rex model was moral acceptance.Footnote 105 Again, in Essays on Bentham, Hart reiterates this point when he argues that the acceptance of the citizens in a simple legal system may be motivated by “any reason whatsoever.” Given this nature of “acceptance” constituting the rule of recognition, it would have been possible for a citizen in a simple legal system to identify and state to others what a directive of Rex requires without morally accepting it, just as Hart thought it possible for a judge in a complex legal system to identify what is legally required and apply the law without any moral endorsement. Hart was mistaken in thinking that this was a phenomenon exclusive to judges in a complex legal system alone brought about by the “institutionalization” of the “acceptance.” Thus, it is not true even for a simple legal system that citizens necessarily morally “accept” the rule of recognition. We would then have to acknowledge that there is, in reality, no “whittling down” of the nature of acceptance of the rule of recognition in the move from the simple legal system of Rex to the complex one, as Hart supposed there was.
VI. UNTYING THE KNOT
The following is an excerpt from the notebook that Hart maintained contemporaneously with his writing of The Concept of Law:
There is something odd, possibly something mistaken in position which I have taken v.2 that the essence/structure of a legal system rests on acceptance of central/certain authority as binding. . .Perhaps all I need to convey is that the obligation is strictly obligation i.e.: narrower than belief in moral goodness. But? There is a muddle (in me) I suspect here.Footnote 106
If the argument advanced in this paper is right, the only place in The Concept of Law where Hart comes anywhere close to taking the position he writes of as “mistaken” and “odd” is the case of the simple model of Rex where he pictures citizens accepting the “impinging” rule of recognition that the directives of Rex are authoritative and binding—the passage also makes it tolerably clear that bindingness here meant moral bindingness.Footnote 107 The impinging rule of recognition on this picture would have created moral obligations of obedience in citizens. Having a legal obligation would have necessarily been equivalent to having a moral obligation. Could the above entry in his notebook have something to do with his having dropped the impinging rule of recognition in the transition from the simple to the complex legal system in The Concept of Law? There is no way of telling with any certainty. What we can tell with some degree of certainty, however, is that in The Concept of Law he goes on to advance a notion of normativity on which obligation is not identical to moral obligation. Of crucial importance in that is the rule of recognition that lays down the criteria for valid law but does not purport to impose an obligation of moral obedience.
What is also clear is that what Hart's account in Essays on Bentham needs in order to dissolve the paradox is a firm metaethical alignment in favor of one of the two models of normativity and a harmonious orientation of all elements in his account to the chosen model. Would this choice have to be arbitrary? Perhaps not. As John Finnis has shown us, Hart's patent metaethical silence notwithstanding, his latent metaethical preference was loud and clear.Footnote 108
