The Long Arc of Legality: Hobbes, Kelsen, Hart David Dyzenhaus

1. Introduction

Twenty-first-century polarization in political thought has left some twentieth-century jurisprudes in an awkward position. Legal philosophers, like so many others today, appear drawn to black-and-white thinking. The positivistic inclination in the United States has reached a fevered pitch, with some returning to an Austinian mindset while others favor a mélange of analytical and normative positivism along with breathtakingly narrow conceptions of legal sources, to arrive at the kind of rigid framework ¹ that positivists deemed straw persons. On the other side are stunningly confident displays of law-morality monism ² leaning left and dazzlingly right-leaning classical views. ³ Many of those weaned on the Hart/Fuller and Hart/Dworkin debates are feeling left out in the cold. We tend to assume at least the desirability of a middle position that acknowledges the prima facie bindingness of positive law while recognizing the centrality of normative ideals in expounding that law and guarding against its especially abhorrent edges. Our questions tend to be about whether such a middle position is achievable, and, if so, at what philosophical cost. As I count myself in this group, I was eager to read David Dyzenhaus’ newest book, an effort by one of today’s most distinguished legal philosophers to bridge that divide. The Long Arc of Legality: Hobbes, Kelsen, Hart was no disappointment.

The Long Arc of Legality is hugely ambitious, and its originality matches its ambition. Its aim is to demonstrate that in key places within the theoretical positions of those sometimes regarded as the leading positivists of prior eras—Hobbes, Hart, and Kelsen—one finds views that are fundamentally natural law-like (and in that sense anti-positivistic) and that the union of their anti-positivistic views reveals the core truth of a form of natural law theory. Happily, on Dyzenhaus’ view, this form of natural law theory permits one to accept driving principles of both Dworkinians and legal positivists: that there is such a thing as positive law as distinct from morality, and that positive law’s demands upon us give every member of the political community reasons to follow it.

The most concise version of the view constructed and defended in The Long Arc of Legality is found in a passage in his introductory chapter. On this view—unlike positivism—there are “principles … discoverable by human reason.” (16)

But [these principles] are not to be found in some natural order of things which tends towards some kind of telos or end, as thinkers from Aristotle through Aquinas to John Finnis have supposed. Rather, these principles are entirely internal to a legal order in that no legal order worthy of the name could fail to instantiate them. Moreover, conformity to these principles will sustain a relationship of reciprocity between ruler and ruled, something like Hobbes’s relation between protection and obedience. (16-17)

It is hard to say how many readers will be convinced by Dyzenhaus to accept what is in some sense a twenty-first-century, Fullerian version of Kelsen or a Hobbesian version of Fuller. Nonetheless, the elaborate argumentation and structure of this account are both illuminating and impressive, and several of the intermediate conclusions of analysis are powerfully defended. Most significantly, Dyzenhaus’ remarkably deep and erudite book re-engages key jurisprudential issues that many philosophers of law have unjustifiably marginalized or treated as special topics. Dyzenhaus rightly depicts the authority of law, the role of courts and other legal actors in interpreting the law, the relationship among national and international legal systems, and the relationship of law and morality, as deeply intertwined issues that a principled philosophical theory of law must address. His book should stand as an impressive reminder of the depth and breadth of a natural law tradition that is utterly distinct from the classical and Thomistic traditions, which continue to serve as the foil for many positivistic philosophers and non-philosophical lawyers.

2. The Challenge Articulated

The central problem of jurisprudence, Dyzenhaus contends, is the problem of authority. He explains it this way: how can law be understood such that there is an answer to the person who asks “‘But, how can that be law for me?’” (2) or ‘Why is this rule’s being law a good reason for me to comply with it?’ More broadly, it is ‘Why should a person accept that putative obligations imposed by law are obligations binding upon him or her?’ (5, 101, 112-13)

The first step in following Dyzenhaus’ train of thought is recognizing that he believes there are constraints on what will count as an adequate answer that make the challenge more difficult, and, indeed, reveal that he is exploring an extremely broad domain of the subject of jurisprudence. An adequate treatment of his version of the problem of authority must, simultaneously, offer an account of the ways in which law is fundamentally institutional, and, relatedly, it must offer an account of the ways in which something’s being so as a matter of law is basically different from its being so as matter of morality (79-87). Additionally, it must be able to solve “the problem of very unjust law” (45): whether judges are obligated to apply a very unjust law and if not, why not (41-78). The book is thus best seen as searching for a jurisprudential theory that will answer a single question with three aspects: How is the law institutionally rooted such that: (i) its demands are not simply the same as the demands of morality; (ii) there are adequate normative grounds—and not just sanction-based incentives—for complying with the law; and (iii) the great injustice of a putative law counts as a substantial legal reason not to apply the putative law?

3. Meta-Norms, the Nub of the Proposal

Dyzenhaus’ answer, drawing upon Hobbes and the social contract tradition, is this: Law is, by its very nature, a product of a political community committed to providing its members with mutually advantageous and crucially important norms of conduct that are coercively enforceable and publicly knowable (90ff). ⁴ Hobbes’s famous Leviathan critique of the “Foole” ⁵ does indeed provide the analytical core of a natural law argument for the rational defensibility (and indeed necessity) of individual compliance with the law, and its fundamental normative claim on the ordinary person (97-99). The conceded inadequacy, in justification, of both actual social contract consent and hypothetical social contract consent, should not lead us to abandon the picture at the core of Hobbes’s account. Rather, it should lead us to realize that there is a political constitution-like idea at the root of what has somewhat misleadingly (if understandably) been taken to be a kind of contractual justification for law’s authority (102-03).

When Hobbes’s political theoretic argument is understood in a jurisprudential register—as an account of the status and bindingness of the law, and an account with a natural law aspect—it becomes clear that he was at some level recognizing the existence of what I shall call ‘meta-norms’, norms that tell us roughly what rules of conduct would have to look like to function successfully in permitting a peaceful political community. On Dyzenhaus’ view, it turns out that Hart’s account of ‘rules of recognition’ and Kelsen’s account of the Grundnorm are (or are closely connected with) highly sophisticated accounts of these constitution-like meta-norms of the legal system. While each of those theorists is in part right to be positivistic—to acknowledge, as a matter of fact, the actual acceptance of such meta-norms in society—neither Hart nor Kelsen regards his account as finished or complete by virtue of having established the fact of acceptance. Hart, for his part, recognizes a wide array of significant natural law points having been set out by Fuller, and more generally, acknowledges a natural law core in any system that would count as a legal system (68-69). ⁶

Dyzenhaus claims to find in Kelsen an even more robust foundation for natural law theory, and in particular, for a rejection of the separation of law and morality. While the natural law twist is partly derived from the fact that Kelsen’s later work expressly defended a citizen’s participation in a democratic political system, that is not its main source. Rather, the principal argument utilized by Dyzenhaus calls for an excursion into Kelsen’s view of international law (225-96). Here, Kelsen maintained the view that international legal norms and domestic norms were all part of one system. This monism stands in contrast to the dualism of Hart and others, who countenanced two separate systems—international law and national law—capable of conflicting with one another. The same arguments, Dyzenhaus suggests, could have and did lead Kelsen to the view that there is a unity between the law and “‘the highest ethical idea.’” (234) ⁷ The resultant monism stands in stark contrast to the separationism commonly associated with Hart. ⁸

4. Static Versus Dynamic Conceptions of Law

Midway through The Long Arc of Legality, Dyzenhaus develops the idea that jurisprudence should understand law in a manner that is dynamic, rather than static:

[T]he fundamental divide in philosophy of law is neither between legal positivism and natural law theory, nor between theories of law and theories of adjudication. Rather, the divide is between static theories of law, as espoused by Bentham, Austin, Hart and Raz, and dynamic theories, as espoused by Hobbes, Kelsen, Radbruch, Fuller and Dworkin, with Kelsen the philosopher of law who set out the fullest account of such a theory. (22-23)

While a static theory regards “‘law as a system of rules … ready for application without regard to the process of their creation,’” (20-21) ⁹ a dynamic theory

tries to capture a crucial feature of the modern legal state: it is the state which engages in the fully law-governed production of legal norms. It thus differs from a static theory in that it includes the dynamic process of legal change within the scope of philosophy of law while a static theory consigns change to some extra-legal space. (20, footnote omitted)

Three appendices provide an unusual denouement. Apparently feeling remiss or vulnerable to criticism for having entirely excluded his Toronto private law theory colleagues from the book, he offers a short appendix to situate himself relative to their work (431-34). A similar gesture is provided with respect to a category of theorists one might have thought especially congenial to Dyzenhaus—inclusive legal positivists, like Canadian Wil Waluchow and like Hart himself—as well as the exclusive legal positivists who for the past decades have dominated Oxford jurisprudence (423-29). In what is arguably a more conspicuous and consequential omission—the failure to address John Finnis, the most prominent English-speaking exponent of traditional natural law theory over the past half-century—Dyzenhaus also provides a separate Appendix; the spirit of his Finnis commentary is (unsurprisingly) more pointed than congenial (435-43). At a minimum, the appendices remind us that there are many other ways to be a natural law theorist or something quite like it.

Each of Dyzenhaus’ chapters warrants a review of its own, both because of the elaborateness and originality of the argument and because of the significance of the (typically controversial) conclusion he intends to draw from it. For present purposes, I shall focus on what I view to be the most novel and promising strand running through the book.

As the outline above indicated, Dyzenhaus’ version of the problem of authority involves defending a position that from various angles appears self-contradictory. One offering a theory of law’s authority must not only defend the moral authority of norms while admitting that the content of the norms often seems to diverge from morality. One must also do so in a manner that leaves room for divergence from those norms when certain kinds of normative shortcomings (‘extreme injustice’) occur. And one must do so in a context in which it is quite obvious that part of the defense will have to do with consistency and clarity of the legal norms and the importance of certain kinds of deference to public authorities.

Dyzenhaus’ solution is not just the banal recognition—from Plato to Hobbes—that improvement on the nastiness of the state of nature is a low bar but an important one. It is more broadly that the enforceability and publicness that are critical to having law are but examples of features that are needed to make a legal system sufficiently well-constructed to merit rational allegiance, and that there are indeed meta-norms (constitutional or rule-of-recognition norms) that must achieve allegiance, too, if a peaceful and just society is to be possible. The justifiability of having such meta-norms, when understood as key to a Hobbesian theory commonly labelled ‘social contract’, is commonly understood as a matter of normative political theory (284, 383, 394). Dyzenhaus maintains that Hobbes himself actually understood them as in some sense constitutional (419). However, rather than concluding—as many do—that such constitutional law is really just in a sense normative political theory, not law, Dyzenhaus concludes that what some people might call ‘normative political theory’ is really law (286, 252, 270). From this vantage point, Dyzenhaus’ searching critical interpretations of Hart, Kelsen, Fuller, and Dworkin depict them as theorists who: (i) maintain that law is not possible without active acceptance and engagement with higher-level norms by judges; (ii) see in those higher-level norms core components of constitutionality and normative political theoretic commitment; and (iii) regard those higher-level norms as being law, while nevertheless having a distinctive and moral quality.

The importance of the ‘dynamic’ aspect of legal theory emerges precisely in connection with the topic of the last sentence. The meta-norms governing legal officials, for Kelsen, are fundamentally practical and action-guiding norms. ¹⁰ Moreover, it is critical that they be enmeshed in a consistent manner (in one sense) with the substantive first-order legal norms themselves. What is important, for Kelsen, is that the domain of norms is in a certain sense complete, but not because of the perfect capacity of lawmakers or constitutionalists to anticipate every possible state of affairs or new law. They are complete in the sense that the actions taken under them are in some sense constitutive of what the norms of the legal system are. That the application of the law is itself part of the law that is being depicted is what Dyzenhaus means by calling it ‘dynamic’ (435-43).

Dyzenhaus’ book—perhaps because of (and not in spite of) the challenges it presents to readers—invites engagement. For that reason, and because (pardon the cliché) imitation is the highest form of flattery, I conclude with an effort to see if I have ‘caught on’. Below is a brief exploration of how Dyzenhaus’ dynamism might be applied to yet another prominent positivist, the late John Gardner.

5. Static versus Dynamic: A Brief Case Study on Gardner

Gardner’s elegant article, “Legal Positivism: 5 ½ Myths” ¹¹ articulates and responds to several common but allegedly misguided objections to legal positivism, which Gardner defines as commitment to his version of the sources thesis:

(LP*) In any legal system, whether a given norm is legally valid, and hence whether it forms part of the law of that system, depends on its sources, not its merits (where its merits, in the relevant sense, include the merits of its sources). ¹²

A very common set of objections entertained by Gardner is that judges who embrace legal positivism will allegedly be precluded (in a variety of ways) from reaching plausible results when it comes to the interpretation and application of the law.

In some quarters legal positivists are thought to be committed to a distinctive view about the proper way of adjudicating cases, according to which judges should not have regard to the merits of cases when deciding them. This conclusion generally comes of combining an endorsement of (LP*) with the widespread assumption that judges are under a professional (i.e. a role-based moral) obligation to decide cases only by applying valid legal norms to them. ¹³

Gardner’s response is that it is patently true that no set of valid legal norms could cover every case that comes before a court, and that, therefore, it is inevitable that judges sometimes should have regard to the merits of the case. ¹⁴ More precisely, Gardner indicates that leading positivists obviously do not inveigh against merits-based adjudication sometimes, but to the contrary recognize its necessity. In this sense (and others), the article replicates Hart’s famous rebuttal of the anti-formalist critique of positivism in his 1957 Holmes Lecture, and Raz’s early critique of Dworkin’s treatment of legal principles. ¹⁵

Remarkably—in light of Dyzenhaus’ project and Gardner’s status as a more recent leading positivist—Gardner’s manner of expressing the alleged absurdity of the objection is evocative of Kelsenian ideas central to The Long Arc of Legality. Gardner writes:

The simplest way to challenge [the proposition that judges should not have regard for the merits of cases when deciding them] is to rely on its systematic and unavoidable collision with another pressing professional obligation of judges, namely their obligation not to refuse to decide any case that is brought before them and that lies within their jurisdiction. If judges are professionally bound to decide cases only by applying valid legal norms to them, the argument goes, then there are necessarily some cases that they should refuse to decide, for there are necessarily some cases not decidable only by applying valid legal norms. ¹⁶

Gardner acknowledges his agreement with Kelsen here, but this is exactly the Kelsenian view that Dyzenhaus sees as characterizing the dynamic nature of law, not its static nature. Dyzenhaus might thus latch onto Gardner’s phrase and ask why what judges are doing as part of their “pressing professional obligation” is not itself part of the law. That is of course exactly Dworkin’s view and part of his longstanding response to the similar objections of Raz; he claimed Raz’s insistence on reserving the word ‘law’ for only the pre-application norms identifiable without reference to values to be indefensibly rigid and ipse dixit, especially in the context of the acknowledgement that in deciding cases, courts must and do reason with moral concepts. ¹⁷

It is notable that Gardner thought he was refuting Dworkin’s “legislative/retroactivity” critique of Hart and Raz when he drew a distinction between judges engaging in legal reasoning to fill gaps in the law and judges legislating from the bench. ¹⁸ Dworkin was of course supportive of a right answer thesis about legal questions that Raz and Hart would have viewed as ‘open’; the positivists’ alleged need to admit there was legislation from the bench, he argued, generated a retroactivity and separation of powers problem of a sort that he (Dworkin) did not need to confront. ¹⁹ Gardner replied that the positivist is not vulnerable to this criticism, because legal reasoning to fill gaps is distinct from legislation, and thus does not generate a separation-of-powers objection. ²⁰ He similarly argued that there was no rule-of-law retroactivity problem so long as it was not retroactive legislation; “the only morally credible rule-of-law ban on retroactive legislation is just that, namely a ban on retroactive legislation, not a ban on the retroactive change of legal norms even when that change is made in accordance with law.” ²¹

Although I share Gardner’s view that retroactivity via legal reasoning may not trigger rule-of-law problems in the way retroactive legislation does, Gardner himself presented no justification for this view, no foundation for the ‘moral credibility’ claim. The obvious answer is that while legal reasoning fills gaps that were in the existing law (understood in a static sense), legal reasoning is in a critical sense application of the law (understood in a dynamic sense). Gardner’s analysis of the superior position of the “legal reasoning” version of gap-filling (compared to legislation from the bench) on the rule-of-law debate evades retroactivity, he says, because the legal reasoning is “in accordance with law.” I agree with that suggestion, but the reasoning’s being “in accordance with law” goes beyond mere consistency, and connotes that there is a more robust way that indicates that the law is being applied.

The upshot of the prior analysis is this: The natural Dyzenhausian response here would be to say that Gardner’s correct identification of legal reasoning within the professional judicial obligation is all the more reason to understand the process of adjudication itself as part of the law. That is what those with a dynamic view of the law do. It is true that Gardner concedes judges are complying with professional obligations, and true that he regarded the justifiability of professional obligations to be a matter of role-based morality. Like Hart in “Positivism and the Separation of Law and Morals,” however, Gardner does not tell us what that morality is, why such obligations need to be complied with, and what the normative significance of ‘role’ is.

The Long Arc of Legality gives rise to a Fullerian- and Dworkinian-inspired argument that Gardner, too, was on the cusp of being a natural law theorist. If—as Gardner suggests—the legal system operates in part only because judges are committed to compliance with these ‘professional obligations’, and if—as Dyzenhaus would argue (drawing from the beginning of his ‘arc’ with Hobbes)—those professional obligations turn out to be rooted in a kind of constitutional ground-norm that calls on judges to play a role that legitimizes the authority of the legal system, then one can imagine Dyzenhaus moving Gardner from the positivist side to the natural law side: It would largely be a matter of rejecting the stipulation that the law must be depicting something static, not dynamic.

6. Conclusion

Although my brief Dyzenhaus-like critique of Gardner on legal reasoning was meant to be more provocative than persuasive, there is a larger and perhaps more personal point to be made about the place of The Long Arc of Legality and Dyzenhaus’ work in contemporary analytic jurisprudence. Just how to classify the process of legal reasoning in jurisprudence—the extent to which one should classify it as law itself or as professionally constrained normative judgment by those occupying a particular role—is a question likely perceived quite differently by those working in different areas of substantive law. While Dyzenhaus himself is a public law expert, we ought not to overlook the fact that his institution (University of Toronto) is arguably the epicenter of private law jurisprudence in the English-speaking world. The Toronto school offers a mould that is more dynamic than static, takes both Fuller and Kelsen seriously, and generates a well-developed version of non-Thomistic natural law theory. As I have argued in prior work and as Weinrib and Ripstein’s work highlights, for private law theorists it is not plausible that one characterize what constitutes the law without understanding legal reasoning to be a part of it. That is plainly a central theme of a line of American jurisprudential thinking from Cardozo through Fuller and Dworkin that draws heavily upon the common law, and indeed part of the reason that John Goldberg and I have taken Cardozo’s common law thinking to be of central jurisprudential importance. ²² This suggests that despite the extraordinary intellectual energy and brilliance that Dyzenhaus poured into The Long Arc of Legality and despite his thoughtful appendix on his Toronto colleagues, the arc of legality may be even longer and wider than he suspected.