ONE-SYSTEM INTEGRITY AND THE LEGAL DOMAIN OF MORALITY

Abstract

According to contemporary nonpositivist theories, legal obligations are a subset of our genuine moral obligations. Debates within nonpositivism then turn on how we delimit the legal “domain” of morality. Recently, nonpositivist theories have come under criticism on two grounds. First, that they are underinclusive, because they cannot explain why paradigmatically “legal” obligations are such. Second, that they are overinclusive, because they count as “legal” certain moral obligations that are plainly nonlegal. This paper undertakes both a ground-clearing exercise for and a defense of nonpositivism. It argues, in particular, that Dworkin's claims about the legal domain of morality in his later work are often mischaracterized by critics, because these critics fail to read these claims in light of his earlier theory of “Law as Integrity.” A nonpositivist theory that unifies Dworkin's earlier and later work, I argue, deals with the criticisms leveled at nonpositivist theories better than other nonpositivist competitors.

I. INTRODUCTION

The last number of years has seen significant developments in nonpositivist, or antipositivist, general jurisprudence.¹ Contemporary nonpositivist theories take what is sometimes described as a “one-system” view of law and morality. That is, they reject the view that legal and moral norms are part of different normative systems, and posit instead that legal obligations are best understood as a subset of our genuine moral obligations, just as family obligations or promissory obligations are types of moral obligation. Debates within nonpositivism then turn on how we make sense of the legal “domain” of morality—whether, for example, there is anything morally distinct about legal obligations.²

Recently, prominent nonpositivist theories have come under criticism on two grounds.³ First, that they are underinclusive, because they lack the explanatory resources to tell us why paradigmatically “legal” obligations are such. Second, that they are overinclusive, because they count as “legal” certain moral obligations that are plainly nonlegal. In this paper, I analyze whether nonpositivist theories can meet this emerging line of criticism.

The central claim of the paper is that Ronald Dworkin's theory of “Law as Integrity,” when understood as a one-system nonpositivist theory grounded in an account of the moral value of legal practice, answers the criticisms effectively.⁴ The theory does so more successfully than other nonpositivist theories, I suggest, because it is grounded in an account of the moral value of legal practice that is more specific and developed than other theories.

The paper is structured as follows. In Section II, I analyze Mark Greenberg's “Moral Impact Theory” of law, one of the most influential contemporary nonpositivist theories.⁵ I argue that claims that the Moral Impact Theory is underinclusive rest on an important misunderstanding about nonpositivism generally. The Moral Impact Theory, when properly understood, successfully deals with these arguments. In its present formulation it struggles, however, to deal with the charge of overinclusivity. The difficulty that the theory has in this regard points to a deeper issue, which is that the theory struggles to explain what, if anything, makes legal obligations distinctively legal to begin with. This, I argue, is because it relies on a stipulation about the nature of legal systems that is unconvincing, at least without further argument.

In Section III, I present an alternative theory of the legal domain of morality, based on Dworkin's work. I argue, following Hillary Nye, that it is a mistake to understand Dworkin's earlier work in “two-systems” terms.⁶ This mistake is prevalent in the literature. More specifically, I argue that the theory of Law as Integrity developed by Dworkin in Law's Empire is best understood in one-system terms, as a theory about the legal domain of morality, resting on an account of the moral value of legal practice. On this view, the moral value of legal practice lies in the constitution of a political community in which coercion is regulated in a morally justifiable way.

My reason for making this argument, however, is not just to set the record straight about Dworkinian nonpositivism.⁷ Rather, I wish to argue that this particular version of nonpositivism more successfully deals with claims of under- and overinclusivity than other prominent nonpositivist theories. In Sections IV and V, I consider whether the one-system version of Law as Integrity can withstand claims of under- and overinclusivity. I argue that it does so successfully in both instances.

The distinct contribution of this paper, then, is twofold. First, it builds on the one-system understanding of Dworkin's work articulated by Nye, by offering a more fleshed out conception of “One-System Integrity.” Second, it shows that One-System Integrity deals more convincingly with positivist criticisms than do other one-system versions of nonpositivism.

II. THE MORAL IMPACT THEORY: CLAIMS OF OVER- AND UNDERINCLUSIVITY

A. The Moral Impact Theory

Mark Greenberg's answer to the question of how to demarcate the legal domain of morality begins like this: legal obligations are just those moral obligations that obtain in virtue of the actions of legal institutions.⁸ The sorts of fact patterns that ground legal obligations are familiar enough. Legislatures pass enactments, members of the executive pass secondary legislation, and judges make decisions in common law cases. According to Greenberg, the difference to our “moral profile” made by these actions is what we call our “legal obligations.”⁹

In the United Kingdom, the Road Traffic Regulation Act 1984 provides that the speed limit on roads designated as “restricted roads” is 30 mph.¹⁰ The Act also provides that a “restricted road” is one with street lighting provided by lamps not more than 200 yards apart.¹¹ By taking certain actions—gathering in Westminster, voting in favor of the 1984 Act, etc.—Parliament changed the moral obligations of those within its jurisdiction. If I find myself driving on a road lit by street lamps 100 yards apart, then I am morally forbidden from driving faster than 30 mph.¹²

This Act also provides that a particular member of the executive may pass an order to raise or lower the speed limit.¹³ Assuming that morality has run its usual course here, we can say that the actions of Parliament in enacting the original Parent Act also made it the case that certain further legal actions (in the form of ministerial orders) could alter our moral profile further. More precisely, the Parent Act determines that subsequent secondary legislation can count toward altering our moral profile in some way that it would not have done absent the Parent Act. A minister declaring that the speed limit is now 15 mph would have no effect on our moral profile had the Parent Act not said that it can have an effect. And the fact that the Parent Act said that it would have an effect is itself a fact that is only relevant because morality makes it so.

Importantly, these changes to our moral situation do not come about because the legislature has some special normative power to change that moral situation. There is nothing morally privileged about the communicative or linguistic content of a statute, and the creation of legal obligation does not come about because of any institution's will.¹⁴ Indeed, members of a legislature might fail entirely to bring about the moral effect that they foresaw. The fact that the legislature acted in a certain way, and the fact that the enactments that it passes bear certain words, are just facts like any others. Morality determines what weight they have in determining our obligations. Typically, we will say that moral facts about democratic governance make it the case that parliamentary enactments generate obligations that more or less correspond to their communicative content. But this is just shorthand for a complex moral argument.

This explanation gives us criteria for demarcating the legal domain of morality, but plainly it is not yet enough to guard against claims of overinclusivity. Intuitively, we would not describe every obligation that results from the actions of legal institutions as a “legal” obligation. The enactment of a statute stating that a certain minority group is to be persecuted would not only fail to generate genuine obligations to persecute this minority group; it would also generate obligations to resist that statute and help the group. Greenberg calls these “paradoxical” obligations.¹⁵ This obligation to resist results from the actions of legal institutions, but it would seem counterintuitive to call it a “legal” obligation.

It is worth noting that while it might seem counterintuitive to call such paradoxical obligations “legal,” one could choose to bite this particular bullet and claim that the category of “legal” obligations is much wider than we had previously thought. To my knowledge no nonpositivist theorist has publicly committed to this position, and I do not do so here, but it is on the table. These sorts of taxonomical questions are simply not as important for nonpositivist theories as they are for positivist ones.¹⁶

It seems to me, however, that nonpositivism should be able to say what, if anything, distinguishes legal obligations generally from other moral obligations. Understanding whether a particular domain of morality has its own operative moral principles is what allows us to work out the precise content of our rights and obligations within those domains. If law is just a branch of morality, then we should approach it as philosophers do other branches of morality and try to understand it as best we can. While it may be less important for nonpositivist theories to engage with taxonomical questions of whether any particular obligation either is or is not legal, failure to deal with claims of under- and overinclusivity could be symptoms of a deeper issue, indicating that we don't understand the practice as well as we might hope.

For Greenberg, there is something about the nature of legal institutions, and the process through which legal obligations come about, that makes legal obligations special. According to him, the reason why a “paradoxical” change to our moral profile does not count as a legal obligation is that it has not come about in the “legally proper way.”¹⁷ Greenberg does not give a full account of the legally proper way, but he does give one important feature. This is that the legally proper way is tied to what he views as a fact about the nature of legal systems. He states: “[w]e have an intuitive understanding of the legally proper way for a legal system to generate obligations, and we can articulate it theoretically by appealing to what legal systems are for or are supposed to do.”¹⁸ What are legal systems supposed to do? According to Greenberg:

[I]t is part of the nature of law that a legal system is supposed to change our moral obligations in order to improve our moral situation – not of course, that legal systems always improve our moral situation, but that they are defective as legal systems to the extent that they do not.¹⁹

And again:

If a legal system is, by its nature, supposed to change moral obligations, it is not surprising that the central feature of law – its content – is made up of the moral obligations that the legal system brings about. Moreover, the view that a legal system is supposed, not merely to change moral obligations, but to do so in a way that improves the moral situation will, as we will see, play an important role in determining which of the moral obligations that result from actions of legal institutions are legal obligations.²⁰

It seems that what counts as a moral change made in the legally proper way will depend on whether that moral change was made by an institution that is supposed to create obligations that improve our moral situation overall.

Here then is one way of explaining what is distinct about law as a domain of morality. Legal obligations are distinct from other moral obligations in obtaining in virtue of the actions of institutions that are supposed to change our normative circumstances in ways that improve the moral situation overall. This claim about what legal institutions are supposed to do provides a nondefectiveness condition: legal systems are defective to the extent that the moral obligations they generate do not improve our moral situation overall. This is why certain changes to our moral profile, such as the generation of a moral obligation to resist a statutory directive, have not come about in the legally proper way, and so are not “legal” obligations.

I consider these aspects of the Moral Impact Theory in greater detail below, when assessing whether the theory convincingly excludes paradigmatically nonlegal moral obligations. I argue that it does not, because the claim that legal systems are supposed to generate obligations that improve our moral situation is not one that can be stipulated without substantive moral argument.

First, however, I wish to examine the claim that the Moral Impact Theory is underinclusive. These claims apply to nonpositivism more generally, but have been targeted at the Moral Impact Theory, so I will focus on those specific claims. The counterargument that I offer, however, is one that applies to all one-system theories. I argue that this particular criticism rests on an important mistake. Highlighting this error will be important in helping us clear some ground in the debates around nonpositivist theories and the ways that they attempt to demarcate the legal domain of morality.

B. Is the Moral Impact Theory Underinclusive?

1. Conflicting Legal and Moral Obligations

Hasan Dindjer, in a careful and sophisticated critique of the Moral Impact Theory, argues that this theory fails to explain the existence of legal obligations that are defeated by conflicting moral obligations.²¹ In this sense it is underinclusive. Greenberg specifically claims that legal obligations are the “all-things-considered” moral obligations that obtain in virtue of the actions of legal institutions.²² The problem, Dindjer says, is that we can think of various examples where it seems like our all-things-considered obligations diverge from our legal obligations, even where our legal obligations came about in the legally proper way. For example, we might find ourselves morally required to break the speed limit in order to transport a sick relative to hospital. Surely it cannot follow that we have no legal obligation to follow the speed limit?²³ Yet that is what seems to follow on Greenberg's theory.

The legally proper way gives us little help here, since the obligation that we generally have to obey the speed limit came about in a perfectly ordinary and unproblematic manner. The problem here is that a legal institution has acted in a way that has given rise to moral obligations, but it appears to be pro tanto rather than all-things-considered obligation. This seems to create a problem for Greenberg, because he stipulates that legal obligations are all-things-considered.

One strategy here would be to concede the point and say that legal obligations are those pro tanto moral obligations that obtain in virtue of the actions of legal institutions. According to Dindjer, however, the pro tanto variation, which he associates with Hershovitz, is equally vulnerable to this criticism. This is because we can think of examples where it seems unnatural to speak of any moral obligation existing, yet where it would seem equally unnatural to say that there exists no legal obligation. He says, for example, that it would seem odd to say that in the case of “an obviously deserted country road with no danger present, an experienced driver invariably violates an obligation in driving before the traffic lights have turned to green, or not indicating before a turn.”²⁴ Here, Dindjer believes, we intuitively think of ourselves as having a legal obligation to obey the speed limit even where we have no moral obligation to do so, whether all-things-considered or pro tanto.

In my view, this objection rests on an important misunderstanding about nonpositivist theories, one that leads critics to beg the question in favor of positivism. This is important since this is relevant for our consideration of all nonpositivist theories, not just the Moral Impact Theory or Hershovitz's theory.²⁵ Highlighting this misunderstanding will help us to see that objections to nonpositivism frequently beg the question in favor of positivism by assuming that important premises, which nonpositivism would reject, are axiomatic.

2. The Context-Sensitivity of Legal Obligations

The misunderstanding in question relates to the question of whether or not legal obligations are context-sensitive. One reason that Greenberg's insistence that legal obligations are all-things-considered obligations makes trouble for his theory, according to Dindjer, is that:

[A]ll-things-considered moral incidents are highly context-dependent in a way that legal incidents are often not . . . The point is that when the law is not sensitive to context in this way — as where it sets a numerical speed limit — one’s all-things-considered moral obligations may remain highly context-sensitive. This is a deep contrast in the structure of legal and moral obligation.²⁶

That legal obligations are always and necessarily context-sensitive, however, is a claim that is very much open to nonpositivists. The imperviousness of legal obligations to context is not a premise that can be assumed without begging the question. Moral obligations are always context-sensitive. To ask whether legal obligations are context-sensitive, then, is just to ask whether legal obligations are genuine moral obligations or not. This is precisely what competing jurisprudential theories disagree over; an answer to it cannot be assumed as a way of testing the plausibility of nonpositivism. The answer to the question of whether or not legal obligations are context-sensitive must follow from a jurisprudential theory.

In the example above, the legislature enacting a statute setting a particular speed limit is just an event that happens in the world. The moral obligations that obtain as a result of this statute are very much context-sensitive. They couldn't but be; morality is always context-sensitive. It is perfectly plausible for a nonpositivist theory to say, for example, that we generally have a legal obligation to obey the speed limit in most instances, but not all the time. Trevor Allan puts the point neatly: “A legal proposition obtains whatever authority it possesses from the cogency of the reasons that justify it. It cannot be more than a summary and provisional statement of how the balance of reasons lies in a specific context or in particular circumstances.”²⁷

This context-sensitivity is at the heart of Dworkin's early discussion of legal principles. A key feature of principles is that their significance fluctuates depending on the particular factual context and on what other principles are in play.²⁸ These principles determine what impact a statute has on our legal obligations. Since the principles are context-sensitive, so are the obligations that obtain in virtue of the statute on which the principles act. In his discussion of Riggs v Palmer, for instance, Dworkin notes that “the court cited the principle that no man may profit from his own wrong as a background standard against which to read the statute of wills and in this way justified a new interpretation of that statute.”²⁹ Each new application of this statute demands a fresh interpretation, because the “background standard” (the principle that no one may profit from their own wrong) is context-sensitive.

The deep difference in the structure of legal and moral obligation that Dindjer identifies, then, is a contestable conclusion about the nature of law, not a paradigm aspect of the concept of law that any conception must explain. On the nonpositivist view, legal obligations are genuine moral ones, and therefore sensitive to context. On one positivist view, legal obligations are “perspectival” obligations to obey the norms created by legal institutions.³⁰ Whichever of these views we prefer, it begs the question to begin with the assumption that legal obligations are insensitive to context.

Dindjer's point is perhaps that we do not experience our legal obligations as context-sensitive. It is true that the person who broke the speed limit on the way to the hospital might think, “Oh dear, I broke the law just now.”³¹ Psychological facts like this, it might be said, speak to the fact that we often believe that we can be the holders of legal obligations that we do not have an all-things-considered obligation to follow. I do not think that we should set too much store by such psychological factors, which rarely track intuitions about deep-structure moral debates.

To see this, consider the limited role we would consider such experiential facts to have in other debates in moral philosophy. Suppose, for example, that you kill an attacker in self-defense. You may feel intense guilt. You may even feel that you have violated your attacker's right to life, although you were justified in doing so. Within moral philosophy, there are different ways of thinking about this scenario. “Specificationism” holds that the content of a right is not fixed, but rather depends on the context in which that right operates.³² While the right to life affords normative protection against being killed in most circumstances, specificationism holds that that right affords no such protection to an attacker who is threatening the life of another. A “generalist” theory, on the other hand, might hold that killing the attacker in self-defense is all-things-considered justified despite that action violating the attacker's right to life.³³ The point for present purposes is just this: your feelings on the matter do not help us to choose between these moral theories, because the sorts of deep-structure moral points being adjudicated are not ones on which we generally have intuitions. Whether one violates a “legal” obligation when breaking the speed limit on the way to the hospital, similarly, is a complex philosophical question, and our intuitions are of limited help in finding an answer.

Dindjer makes this very point in a different context, when disputing Greenberg's claim that we have an intuitive sense of law being made in the legally proper way: “[t]he notion [of the legally proper way] is a technical one, that people do not think or talk about as such; nor is it clear that we have tacit views about it.”³⁴ We do think and talk about having legal rights and obligations, but behind our talk and tacit views, these too are highly technical philosophical concepts. It is not unreasonable to think that our common usage of “legal” obligation sometimes does not track the best understanding of that concept, and that the reactive attitudes that result from acting in certain ways may not track whether we have acted “illegally.”

It is true that it is a reality of legal practice that ostensible legal obligations are often enforced regardless of the all-things-considered moral position. One can still receive a speeding ticket even when rushing to the hospital. Again, though, whether we describe this as the enforcement of a legal obligation or simply as a use of force is an open question. The objection that the label “legal” should attach to these obligations is premised on a contestable theory of legal obligation, one that demands justification. It cannot be assumed from the outset. Dworkin puts the point pithily: “vocabulary should follow political argument, not the other way around.”³⁵

This is important, because appealing to the context-sensitive nature of legal obligation is precisely the means by which Greenberg, and nonpositivism more generally, can deflate claims of underinclusivity. It is open to him to argue that the person with an all-things-considered moral obligation to break the speed limit on the way to the hospital has no legal obligation to obey the speed limit in this instance. It might sound a little odd to say that legal obligations are context-sensitive, but this is little more than a semantic quirk. If one views legal obligations as context-sensitive, which is a perfectly plausible moral position, then it is entirely sensible to say that the speeder violated no legal obligation here.³⁶

3. Context-Sensitivity, Eliminativism, and Clear Guidance

An objection to my characterization of legal obligation as context-sensitive might be that this renders law incapable of providing coherent, predictable guidance. If legal obligations are sensitive to context, then how can individuals confidently orient themselves in accordance with the law?³⁷

Liam Murphy deploys this argument against “eliminativist” jurisprudential theories.³⁸ It will be useful to consider the argument in this context, because the debate between Murphy and eliminativists maps onto the present discussion of whether viewing legal obligations as context-sensitive makes law incapable of guiding action.

Eliminativist theories are ones that do without a “doctrinal” concept of law.³⁹ Rather than rely on talk of what “the law” of a particular place is, eliminativism encourages us to focus on other questions, such as evaluative questions about the value of legality, “folk” questions about how people typically think and speak about law, predictive questions about what legal officials will do, etc.⁴⁰ The Moral Impact Theory, inasmuch as it uses legal obligations to refer to genuine moral obligations, seems eliminativist in spirit.

Murphy poses several challenges to eliminativism.⁴¹ For present purposes, I will focus only on the claim that it does away with any possibility of law providing clear, predictable guidance to individuals. Murphy argues that we need to know what the law is in order to know how to act, and so it is implausible to do away with the doctrinal concept of law. It would be no good, he says, to say simply that we are using the word “law” in different ways; one of us as a prediction about what legal officials will do and another as a normative statement about what they should do, for example.⁴²

It should hopefully be clear that this sort of concern maps onto discussions about the context-sensitivity of legal obligations. The claim that legal obligations are not sensitive to context derives from the view that there is a doctrinal concept of law, something insensitive to context, and it is this doctrinal concept that offers clear guidance on what we are supposed to do. The claim that legal obligations are context-sensitive is true if we view legal obligations as those genuine moral obligations that obtain in virtue of institutional action. The concern that viewing legal obligations as context-sensitive does away with the capacity of law to offer clear guidance, then, is a concern with thinking of legal obligation in a way that is untethered to law in a doctrinal sense.⁴³

Kornhauser argues that Murphy's concern is misplaced, because removing the doctrinal concept of law from the picture (and we can add, viewing legal obligation as context-sensitive) does not make any practical difference to how legal systems function or how we interact with them. We can still look to legal materials (statutes, cases, etc.), we can predict how legal institutions and officials will behave, and we can make judgements about what our all-things-considered moral obligations are. Individuals can still decide how to act based on the legal materials, which will form the basis for the decisions of officials.⁴⁴

Kornhauser's argument seems to me to be correct, and it explains why we need not be detained by the concern that thinking of legal obligations as context-sensitive will restrict the ability of legal practice to offer clear guidance to individuals. Law's action-guiding function is not lost when we think of legal obligations as context-sensitive moral obligations, because this changes nothing about the practical circumstances in which individuals find themselves. Individuals can still look to the legal materials, make predictions about what legal officials will do, and hold moral convictions about what those officials should do.

We might say that this departs from our intuitions or usual language usage.⁴⁵ As I have argued above, however, it is not clear how much store we should put by intuitions about something as philosophically complex as legal obligation. If I speed a sick child to the hospital and anticipate criminal punishment, I might well think of myself as “breaking the law,” just as I might think of myself as “breaking a promise” not to commit murder. But it is difficult to see why this sort of internalized linguistic formulation must necessarily track any deeper philosophical truth about the obligations in question. Both the eliminativist position against which Murphy argues, and the contextual-legal-obligation position that I set out here, are in effect invitations to rethink the way that we talk about law in line with deeper philosophical insights.⁴⁶

C. The Moral Impact Theory, Overinclusivity, and the Nature of Legal Systems

While charges of underinclusivity can be deflated, the strategy that Greenberg uses to deflect criticisms of overinclusivity is less convincing. Greenberg's claim, by way of reminder, is that it is part of the nature of legal systems that they are supposed to create obligations that improve the moral situation overall. Greenberg doesn't specify what sort of claim this is. One could read it as a conceptual claim, for example, but this is unlikely given the skepticism Greenberg expresses elsewhere about the utility of conceptual analysis in explaining the determination of legal facts.⁴⁷ It seems most likely that we should understand his claim as an attempt to offer part of a “real definition” about legal systems; a definition, that is, about the essence of legal systems, rather than a linguistic or conceptual definition of a legal system.⁴⁸ To say that it is in the nature of legal systems that they are supposed to change our circumstances in ways that improve the moral situation overall, on this view, is to make a claim not about the concept of a legal system, but about the thing that the concept picks out.

From this, a nondefectiveness condition follows: a legal system that creates obligations that do not improve our moral situation (or that make it worse) is defective qua legal system.⁴⁹ Certain of the obligations that obtain in virtue of the actions of legal institutions do not count as legal on the grounds that they do not come about in the “legally proper way,” which seems to mean that they come about in a way that goes against what legal systems are supposed to do.

These aspects of the Moral Impact Theory are not fully developed, as Greenberg acknowledges. I wish to argue, however, that there is reason to doubt that they can be developed in an explanatorily useful way. To draw this point out, I will consider whether it would make sense to talk about other obligation-generating practices as “improving our moral situation.” First, however, I will articulate the nature of the argument that I make about these aspects of the Moral Impact Theory, and how my argument relates to claims, made by others, that the Moral Impact Theory is overinclusive.

1. The Real Problem with Overinclusivity

Before discussing the notion of “improving the moral situation” in more detail, it is worth pausing to try to pin down exactly how we should understand claims of “overinclusivity.” One version of the criticism is simply that there are certain moral obligations that obtain in virtue of the actions of legal institutions, but that it would be silly to call these obligations “legal.” The critique is that the Moral Impact Theory is deficient to the extent that it fails to explain why these paradigmatically nonlegal obligations are not legal.⁵⁰

To my mind, this critique is not particularly compelling. As I have said already, these sorts of taxonomical questions are just less important for nonpositivist theories than positivist ones. And as I argued in the previous section, there are limits to what our intuitions and ordinary language can tell us about these deep philosophical questions. There is no reason, then, why Greenberg could not simply bite the bullet and accept, for example, that “paradoxical” obligations are legal. Again, taxonomy should follow theory, not the other way around.

There is, however, a stronger version of the overinclusivity critique. The question of what makes something a legal obligation precedes questions about whether specific obligations are or are not legal. It may be that the failure to offer a coherent account of why certain obligations are not legal is symptomatic of a deeper issue, which is that the theory fails to tell us what makes something a legal obligation to begin with.

It is this version of the critique that I will explore here. The problem, I argue, is that Greenberg tries to exclude certain obligations from the legal domain of morality without offering an account of what is morally distinct about legal obligations. Instead, he relies on the device of the “legally proper way” and the notion of improving the moral situation to explain why certain obligations do not count as legal. That legal systems are supposed to improve our moral situation, however, is not something that we can stipulate as a fact about the nature of legal practice. Rather, that is a claim that must follow from an account of the moral principles operative in the domain. If that is the case, however, then the notion of “improving the moral situation” is redundant, for reasons set out below. It follows that not only does the Moral Impact Theory fail to explain why certain obligations are not legal; it fails to tell us anything significant about what makes a legal obligation distinctively legal, at least without further development.⁵¹

To draw this argument out, I will consider analogies with other obligation-generating practices.

2. Friendship, Family, and Promises

The idea that legal obligations are supposed to improve our moral situation is not, it seems to me, intuitively obvious. There are plenty of practices that are supposed to create obligations that make us better off, but not all such practices are necessarily supposed to create obligations that make us morally better off.

Consider first friendship. Friendships give rise to obligations, and we could plausibly say that these obligations are supposed to make us better off, in the sense that they make us happier, or more fulfilled, or more well-rounded people. If you and I become friends, and so come to owe certain obligations to each other, that would, it seems to me, be an all-around better state of affairs than for us not to be friends. It would be odd, however, to say that we are morally better off in virtue of such friendship obligations. Our moral situation isn't necessarily better than what it was before; it's just different.

Parenthood obligations provide a similar example. The birth of a child is an event that changes its parents’ normative situation, creating new rights and obligations for them. The birth of a child might constitute obligations to spend less time with one's friends, or spend less money on oneself, for example.⁵² These are genuine moral obligations, but again it would seem odd in the extreme to say that we are morally better off because we have them.

Moreover, not only does it seem unusual to speak of obligation-generating practices as making us morally better off; it also seems explanatorily redundant to do so. Sure, we might say that the normative world in which parents owe obligations to their children is better overall, morally speaking, than a world in which parents owe no obligations to their children. This is true, but trivially so. It doesn't tell us anything particularly interesting about the practice, or about what is distinct about parenthood obligations, or about why some obligations that obtain in virtue of having children are not parenthood obligations. What we should be looking for, surely, when trying to figure out which obligations are parenthood obligations, is a theory of the moral rules governing the practice.

Promising is another obligation-generating practice that presents an instructive parallel. Again it would seem redundant to claim that it is in the nature of the practice of promising that it creates obligations that improve our moral situation. The operative moral principles in play are what tell us what counts as a promise, when we should keep our promises, etc.

Take for example Scanlon's view of promising. For Scanlon, promissory obligations are grounded in the more general moral obligation not to manipulate others.⁵³ More broadly, the operative moral rules, for Scanlon, are those that no one could reasonably reject.⁵⁴ Armed with this account of the moral underpinnings of the practice, we can explain why promises create obligations, and why certain types of ostensible promise, such as immoral promises, do not create promissory obligations. For example, we can say, roughly, that an immoral promise is not really a promise because a purported duty to keep an immoral promise could not be justified by a moral rule that no one could reasonably reject. The operative moral rules directly provide constraints on what counts as a promissory obligation. It is difficult to see what would be gained by adding that these obligations make us morally better off.

The picture doesn't change if we take a different account of the morality of promising. Suppose promising is a normative power to transfer to others the power to decide how we should act, grounded in the interest that we have in forming intimate relationships with others.⁵⁵ Again, this account of the moral underpinnings of the practice directly explains what counts as a promise and what does not, and when and why promises are binding. By promising to rob a bank, for instance, I cannot transfer to the promisee the right to decide that I rob the bank, because I have no moral right to rob banks to begin with.⁵⁶ We can add that promises are supposed to create obligations that improve the moral situation, because a promise that failed to do so would undermine our ability to form intimate relationships. But again this seems largely redundant.⁵⁷ The operative moral principles are what tell us what we need to know about the practice.

In all of the examples given above, the moral accounts of the practices are what explain what obligations count as friendship obligations, family obligations, promissory obligations, etc. If we removed the account of the operative moral rules, and left only the claim that friendship or family or promising improves the moral situation overall, we would lose the ability to say what counts as a family obligation, a friendship obligation, or a promissory obligation. But if we include the account of the moral underpinnings of the practice, then the addition of the claim that these practices improve our moral situation is redundant.

3. Law

As with practices considered above, there is something odd about describing legal obligations as improving our moral situation. Having obligations not to assault each other, to drive on a certain side of the road, not to renege on contracts, or not to trade shares in a company based on insider information, etc., are all valuable aspects of legal practice. We might be better off having such obligations.⁵⁸ The question is whether the changes in our situation that these obligations bring about are best understood as moral improvements. Perhaps they are, but this seems to me to lack the intuitive obviousness that would justify stipulating it as a fact about the nature of legal practice. Rather, it seems like the sort of claim that should follow from a more substantive account of the moral principles governing the practice.

However, if we provide an account of the moral principles governing legal practice, then it is difficult to see how the idea of law improving the moral situation could be anything but redundant. As with the other practices considered, the moral underpinnings of the practice can directly explain why we hold the obligations that we do in virtue of legal practice, and therefore which obligations count as legal. That is, if a nondefectiveness condition obtains in virtue of a moral explanation of the practice, then the nondefectiveness condition does not really do any work in explaining why certain obligations are not legal. The moral account of the practice does that work directly.

For example, in Dworkin's account, legal practice is grounded in the constitution of a political community in which all are treated with equal concern. From this, it follows that legal obligations are those whose enforcement in court is justified by principles of justice drawn from relevantly similar past decisions.⁵⁹ It follows that a moral obligation is only properly thought of as a legal obligation if its coercive enforcement would be justifiable. We could add that we are morally better off having such obligations than not having them, but it is difficult to see the explanatory upshot.

Just as with friendship, family, and promising, it is difficult to see how the claim that law “improves the moral situation overall” could tell us anything of interest about legal practice, unless that claim is nested in a broader account of the moral principles underpinning the practice. If it is nested in such an account, however, then it becomes redundant. It is the operative moral principles—not the idea of moral improvement and the nondefectiveness condition of the legally proper way—that explain what makes certain obligations distinctively legal.

It is difficult to assess the Moral Impact Theory's success in guarding against overinclusivity (if we think that is important), because the device introduced to do the theoretical work is underspecific. Greenberg is aware of this and notes that future work would need to flesh out the notion of the “legally proper way.”⁶⁰ I have attempted to show here, however, that there is a deeper issue. Whatever way the notion of “improving the moral situation overall” is fleshed out, the attempt to plug this notion into the theory as a fact about the nature of legal systems, without relying on a substantive normative account of legal practice, is unconvincing. The most obvious way to argue for the claim that law is supposed to create obligations that improve the moral situation would be to argue that this follows from an account of law's moral underpinnings. But this would render the nondefectiveness condition explanatorily redundant.

In the next section, I will consider an alternative method of demarcating the legal domain of morality, based on Dworkin's work. This theory adopts precisely the sort of strategy that I suggest is missing from the Moral Impact Theory. That is, it argues that legal obligations are morally distinct from other sorts of obligations.

III. LAW'S VALUE AND CITIZENS’ VULNERABILITY: COERCION AND INTEGRITY

In this section, I set out an alternative account of the legal domain of morality, based on Dworkin's theory of Law as Integrity. In the next two sections, I consider whether this theory deals with charges of under- and overinclusivity, respectively, in a more satisfactory way than the Moral Impact Theory. First, however, some exegesis is necessary, because there is some ambiguity around whether Dworkin's earlier work can be understood in one-system terms. I say “exegesis,” but my primary aim is not to say what Dworkin meant, but rather to give the best understanding of his theory. It is, I think, in the spirit of his interpretivist project to seek the best constructive interpretation of his theory, rather than to rely solely on the interpretation of its author.

In Justice for Hedgehogs, Dworkin states that he views legal rights as a subset of moral rights and that he now believes that legal rights are “those that people are entitled to enforce on demand” in courts.⁶¹ Law, then, is the branch of morality concerned with the rights and obligations that can justifiably be enforced by legal institutions. On one reading, Dworkin is adopting an entirely new view to the one he held in previous work. He views legal obligations as moral obligations when previously he did not, and in an effort to avoid overinclusivity, he stipulates that legal rights are those rights that can be enforced in court. Some have claimed, and Dworkin himself conceded, that in earlier work he adopted a “two systems view” of law and morality.⁶² The assertion in Justice for Hedgehogs that legal rights are genuine moral rights, then, is something of an about-face.⁶³

Critics have subsequently taken issue with Dworkin's stipulation that legal rights are those that can be enforced in court. Some find this stipulation unsatisfactory. Greenberg, for instance, says that he views this “very different” position as a “version of the Moral Impact Theory that restricts legal rights and obligations to those that should be enforced by courts.”⁶⁴ He seems to view the means by which Dworkin draws these boundaries, however, as arbitrary. Greenberg claims that this “Judicial Enforcement Theory” is circular, since an account of law should explain why legal norms are judicially enforceable, not simply define them according to that enforceability.⁶⁵ Others have been similarly critical of Dworkin's account of the legal domain of morality.⁶⁶

Against this trend, Hillary Nye argues that it follows from Dworkin's longstanding “anti-metaphysical” commitments that his theory of law has always been a one-system one.⁶⁷ Throughout his career, Dworkin rejected the idea that there is any way to make “external” claims about the domain of value, claims, that is, about what morality “really is,” as distinct from first-order claims about what morality requires.⁶⁸ Dworkin, Nye argues, understands law in the same way.⁶⁹ There are no external claims, on the Dworkinian view, about what law “really is,” as a conceptual matter. There are only interpretive claims internal to the practice. When we recognize this, it makes little sense to think that Dworkin was ever committed to a two-systems view. Rather, the view of law articulated in Justice for Hedgehogs is not “a dramatic shift, but a different way of expressing his longstanding emphasis on legality and the thought that legal theory is an inherently normative endeavor.”⁷⁰

In my view, this reading of Dworkin's work is convincing. I wish to contribute here to this understanding of Dworkinian nonpositivism by fleshing out a one-system version of Dworkin's theory of Law as Integrity. When read as a one-system theory, Law as Integrity offers an account of the moral value of legal practice that motivates the view that the subset of our moral obligations whose enforcement in court is justifiable is morally distinct from the rest of our moral obligations. This moral distinction is what motivates the thesis that it is this subset that we should think of as legal obligations.⁷¹ The value of legal practice, on this theory, lies in the constitution, through legal practice, of a political community in which coercion is regulated in a morally justifiable way.

My approach here is akin to one of reflective equilibrium. The explicitly one-system theory of Justice for Hedgehogs, focused on justifiable enforcement in court, provides the best way of understanding the earlier theory of Law as Integrity. This understanding of Law as Integrity, in turn, shows that the “Judicial Enforcement Theory” set out in Justice for Hedgehogs is not arbitrary. The criticisms leveled by Greenberg, Hershovitz, and others at the Judicial Enforcement Theory are uncharitable, precisely because they fail to consider this in light of the earlier theory of Law as Integrity.⁷²

Understanding the relationship between law and coercion is key to this account of the legal domain of morality. Nicos Stavropoulos emphasizes the centrality of coercion to Dworkinian nonpositivism.⁷³ When we win in court, we are entitled to call on the state to use its monopoly on coercive force on our behalf against other members of our political community. When we lose, we are made vulnerable to that coercive force being brought to bear on us on behalf of the political community. It is true that in almost all of our daily interactions with law, coercion does not obviously seem to feature. We regularly stop at red lights, carry out employment obligations that we have contracted into, and pay for goods that we want to own, all without any threats against us if we fail to do these things in the right way.⁷⁴ But the threat of coercion being brought to bear on us on behalf of the community is always present. If I don't stop at a red light or if I shoplift, the police are permitted by law to arrest me. A court can lawfully order my employer to hand over money to me if she has failed to do so on her own. Coercive enforcement is a central feature of our legal practice.⁷⁵

This is not to say that law creates coercion. Even in law's absence, the possibility of coercion by others would be unavoidable. The possibility of finding oneself subject to coercion is an inescapable aspect of living with others.⁷⁶ Law does, however, allow us to call on state institutions to coerce others on our behalf. Legal practice, then, is the practice of centralizing coercion and regulating its use in morally justifiable ways.⁷⁷ Any conception of law, it follows, is partly an attempt to explain the conditions of moral justification for the use of this coercion.

The key Dworkinian claim is this: when it regulates coercion in a particular way, legal practice constitutes a special form of political community in which “associative obligations” obtain. These are the “special responsibilities social practice attaches to membership in some biological or social group.”⁷⁸ This is how we begin to demarcate the legal domain of morality on the Dworkinian account: legal obligations are a special kind of associative obligation that we owe to other members of our political community, obligations that arise as a result of the value of constituting a particular kind of morally valuable political community. While these tenets of Dworkin's theory are well known, they merit unpacking so they can be understood in one-system terms.

Dworkin's strategy for demonstrating that legal obligations are associative in nature is to examine the character of “associative obligations” in other domains, and then try to show that political obligations share these features. The first key feature of associative obligations is that they are “special,” i.e., holding only between members of the association.⁷⁹ Second, they are “personal,” in the sense that they run from each individual to each of the other individuals within the group.⁸⁰ An academic may owe certain obligations to her university as a corporate entity. When we speak of obligations of friendship, however, we mean obligations that hold among individual friends. Even in a group of friends, the obligations run from each and to all (though they may differ in strength depending on the particular history of each friendship within the group). Finally, associative obligations only obtain where members of a group suppose that their obligations derive from a responsibility of equal concern for the well-being of the other members of the group.⁸¹ That is, the group must suppose that each person's role and life are equally valuable and important.

Characterizing legal obligations as associative, on this view, is key to explaining how genuine moral obligations can obtain in virtue of legal practice. A political society in which associative obligations obtain becomes a special type of community, which Dworkin calls a “true political community,” one committed to the equal standing of all members.⁸² This sort of community is morally valuable, because it treats members as enjoying equal standing within it, and so expresses a special ideal of political equality. Under specific conditions, legal practice gives rise to these sorts of obligations.

What standards must a political community meet in order to generate associative obligations through its legal practice? This is where the political value of “integrity” enters the picture. A political community becomes a “true community,” Dworkin says, when it decides what its members owe one another by drawing on a coherent scheme of principle that governs all members in the same way. Collective force can be justified, in such a community, only when licensed by principles drawn from past decisions about when such force is justified.⁸³ In this way, members affirm an ideal of equal concern for one another.

A community that enforces rights and obligations by asking what principles of justice underpinned similar cases has the four characteristics of associative obligations.⁸⁴ In particular, it embodies the requirement of equal concern: “integrity assumes that each person is as worthy as any other, that each must be treated with equal concern according to some coherent conception of what that means.”⁸⁵ A legal practice that enforces obligations in accordance with integrity is thus constitutive of a special, morally valuable type of political community, one that ensures that citizens are treated equally in the coercive enforcement of rights and responsibilities. In this way “collective decisions are matters of obligation and not bare power” in a political community that practices integrity.⁸⁶

This takes us back to the criticism of Dworkin's later claim that legal obligations are those obligations whose enforcement in court is justifiable. We are now in a position to see why that attempt to delimit the legal domain of morality is not arbitrary. When we consider the enforcement condition in light of the story of integrity laid out here, we see that this criticism is uncharitable. It fails to consider the moral explanation underpinning the decision. Rights and obligations that can be enforced in court are morally distinct from other moral rights and obligations. The fact that coercive enforcement attaches to these rights and obligations means that their content is determined by the value of integrity.⁸⁷ And these obligations have different grounds to other moral obligations: they ultimately obtain in virtue of the moral value in constituting a political community in which all are treated with equal concern.

One might disagree on substantive grounds with this thesis. That is, one might think that the moral difference between judicially enforceable obligations and other types of political obligation is not great enough to merit the distinction. But that argument requires substantive engagement with the moral explanation underpinning the distinction. Otherwise the complaint that we misuse the label “legal” by applying it only to enforceable rights is just a taxonomical, or semantic claim.⁸⁸

We have a candidate answer, then, to the question of how to delimit the legal domain of morality. This candidate is a reading of Dworkin's theory of Law as Integrity that views that theory in one-system terms. I will refer to this view as “One-System Integrity” in the remainder of this article.⁸⁹ According to One-System Integrity, legal practice can, under specific circumstances, constitute a morally valuable kind of political community. This community is constituted when coercion is regulated in a way in which all are treated with equal concern and respect. The conditions for the constitution of such a community are fulfilled when participants are able to have rights and obligations coercively enforced against each other in line with the value of integrity, or principled consistency. Our legal obligations, then, are the ones that we are entitled to have enforced in court, and the precise content of these obligations is determined by the value of integrity. This is what makes those obligations different from other, nonlegal moral obligations. In the next two sections, I consider whether this strategy is successful in dealing with charges of under- and overinclusivity.

IV. IS ONE-SYSTEM INTEGRITY UNDERINCLUSIVE?

In Section II.B I considered the claim that nonpositivist theories are underinclusive because they fail to account for ostensibly legal obligations that clash with all-things-considered moral obligations. I argued that this concern was overblown, and rested on a question-begging assumption about the structure of legal obligation. Nonpositivist theories generally are entitled to claim that legal obligations are context-sensitive, such that one can have, for example, a legal obligation to obey the speed limit in most circumstances, but not when rushing one's sick mother to the hospital.

Under One-System Integrity, the question is whether it would be justifiable to coercively enforce the pro tanto obligation to obey the speed limit, even if the speeder was under no all-things-considered obligation in the particular circumstances. Could such enforcement be justified by principles of justice underpinning previous decisions? If it could, then this is a legal obligation. If it could not, then the theory has the resources to deny that it is a legal obligation.

The failure to see that legal obligations can be context-sensitive (or at least that that argument is on the table) influences other claims of underinclusivity that Dindjer makes about Dworkin's focus on judicial enforcement. He argues that relying on the enforceability criterion leaves us unable to account for legal obligations that are not enforced because they are not justiciable.⁹⁰ He says:

Although courts sometimes appeal to these doctrines as an imprecise label for the nonexistence of a legal duty, they are also used more accurately to indicate that a court may not adjudicate genuine legal duties, such as certain duties governments owe in armed conflicts, or when acting in coordination with foreign states.⁹¹

Consider for example the well-known GCHQ case in the UK, which established that the exercise of the royal prerogative is subject to judicial review.⁹² Here the royal prerogative had been used to prevent employees of the security services from joining trade unions. The Court, while establishing that the prerogative was subject to judicial review, refused to exercise judicial review in this particular instance because of the national security concerns involved. Dindjer's point is that we would not say, as a result of cases like this, that there is no legal right to have the prerogative judicially reviewed. Rather, we would say that the legal right is not enforced.

Again, however, this criticism rests on the failure to see that it is open to nonpositivism to view legal rights and obligations as context-sensitive. I have argued in Section II.B that it begs the question in favor of positivism to assume otherwise. It is perfectly sensible to argue, from a nonpositivist perspective, that there exists no legal right to have the prerogative reviewed in this particular set of circumstances. On this view, saying that there is a “legal right to” anything is just a shorthand; an approximation of a right whose precise content changes depending on the specific factual circumstances in which it is engaged, and depending on what other principles are engaged. Even saying that there is no legal right to have the prerogative reviewed in “national security” cases would only be an approximation of a far more complicated moral statement. What counts as a “national security” case, and when the use of the prerogative in such a context can be reviewed, are moral questions with which judges engage. The “national security” context, then, is shorthand for a set of particular moral considerations that pertain to whether coercive enforcement can justifiably be called upon in these particular circumstances.

A different charge of underinclusivity is found in the literature on “underenforced” constitutional obligations. Serious attention was first brought to these in a seminal essay by Lawrence Sager.⁹³ He points out that there seem to be occasions in which the US Supreme Court, for reasons of institutional restraint, fails to “enforce a provision of the Constitution to its full conceptual boundaries.”⁹⁴ Certain constitutional provisions, Sager claims, create rights and obligations that are both (i) legal, and (ii) not properly enforceable by courts. For example, Section 5 of the Fourteenth Amendment to the US Constitution provides: “Congress shall have the power to enforce, by appropriate legislation, the provisions of this article.”

Sager points out that in a number of Supreme Court cases, the Court has held that Congress has the authority to enact legislation designed to further and uphold the Fourteenth Amendment.⁹⁵ This authority is derived from Section 5 of that Amendment. The “full scope” of the Fourteenth Amendment, then, includes a judicially enforceable right, and congressional authority to enact legislation furthering Equal Protection beyond the courts. Both of these aspects, Sager thinks it important to say, are legal aspects of the Fourteenth Amendment. The second aspect, however, is not one that is enforceable in court. Whether or not this is the correct constitutional interpretation of the Fourteenth Amendment is not something that I can answer here. However, let us grant for now that it is. If such an obligation exists, should we call it a legal obligation?

Sager thinks that we would lose something important if we confined our conception of legal obligation only to those obligations to whose enforcement we are entitled. This is why he objects to the distinction that Dworkin draws in Justice for Hedgehogs between “legal” and “legislative” rights, both of which he views as subcategories of “political rights.”⁹⁶

It isn't obvious, however, that we lose anything of importance by denying the label “legal” to unenforced rights. When we understand that the theoretical compartmentalization of the rights that are enforced in court is motivated by a deeper moral theory of legal practice, the importance of the label “legal” dissipates. According to One-System Integrity, the rights that we can justifiably have enforced (“legal rights”) and the underenforced obligations Sager is interested in (what Dworkin calls “legislative rights”) are distinct types of moral right.

We do not need to think that “legal” rights are any more important than “legislative” rights, just as we need not think our family obligations are necessarily more important than our promises to our friends. There are important and comparatively unimportant family obligations, and important and comparatively unimportant promissory obligations. Similarly, there are more and less important legal rights and more and less important legislative rights. The label “law” has enormous importance if we view legal obligations as some form of nonmoral norm that claims peremptory force over our own moral calculations. But when we view legal rights as themselves a species of moral right, the importance of the label dissipates. The instinct to attach great importance to a label with the historic solemnity of “law” is understandable, but there is little substance behind the instinct, by the lights of One-System Integrity.

The claim that One-System Integrity fails to account for underenforced rights, then, ultimately collapses into a semantic debate. One-System Integrity makes a moral distinction between rights whose enforcement is justifiable and those whose enforcement would not be. Whether or not we should attach the label “legal” to one or another set of obligations is just not an important question when we accept the basic premises of nonpositivism. There are legal institutions, and there are various obligations that flow from the actions of these institutions. The fact that some of these obligations invite legitimate enforcement demarcates an interesting and morally distinct subclass. But nothing really turns on attaching the label “legal” to that subclass. The obligations will still be interesting and morally distinct, whatever label we use. When we understand both enforceable and unenforceable rights as a branch of morality, whether or not something is a “legal-political-right” or a “nonlegal-political-right” really isn't that important.⁹⁷

V. IS ONE-SYSTEM INTEGRITY OVERINCLUSIVE?

In Section III, I argued that the Moral Impact Theory struggled to explain why not every obligation that resulted from the actions of legal institutions is a legal obligation. This was symptomatic, I argued, of the Moral Impact Theory's broader failure to explain what, if anything, is distinctive about legal obligations. One-System Integrity offers a clearer explanation of why certain moral obligations are not legal, because it is grounded in an account of the moral underpinnings of legal practice.

Again, legal obligations, on this view, are those obligations whose enforcement in court is justifiable, and this justifiability is determined by the principle of integrity. It seems to me that this offers a compelling way of excluding nonlegal moral norms from the legal domain of morality. Take, for example, the “paradoxical” obligations that concerned Greenberg. These occur when a wicked legal institution generates obligations of resistance. Greenberg introduces the device of the legally proper way partly to explain why such obligations are not legal obligations. On the view of One-System Integrity, we can say that these obligations are not legal, despite their obtaining in virtue of the actions of legal institutions, because their enforcement could not be justified by the operative principles of political morality (i.e., those picked out by integrity).⁹⁸

Or take obligations that arise too far “downstream” of the actions of legal institutions to plausibly call the obligations “legal.” When the government relaxed the restrictions around Covid-19 to permit visits to care homes, that may have given rise to an obligation to visit one's grandmother. We would be unlikely to say that one had a legal obligation to visit her, even though the obligation obtained in virtue of the actions of legal institutions. On the Moral Impact Theory, the legally proper way is a catchall device for these sorts of scenarios. We saw, however, that this device rested on an underdeveloped claim about the nature of legal systems. One-System Integrity offers a neater way of excluding such obligations. The obligation to visit one's grandmother is not a legal obligation because the enforcement of that obligation in court could not be justified by principles of justice drawn from relevantly similar past cases.

There may be other cases that I have not considered here. In general, however, it seems to me that the story offered by One-System Integrity does a good job of demarcating the legal domain of morality. It specifies the legal subset of our moral obligations by offering an account of the moral value of legal practice, and the morally relevant coercion to which participants in that practice are made vulnerable. When we divorce the enforcement criterion to which Dworkin signals in Justice for Hedgehogs from his earlier theory, that criterion can seem arbitrary. When we consider it in the light of that earlier theory, however, we see that it rests on a fleshed out theory of political legitimacy. We might still disagree with this view on the grounds that it does not offer the best moral explanation of legal practice. But our disagreement will then be a first-order moral one, which is precisely the sort of disagreement that nonpositivist theories should have with one another.

VI. CONCLUSION

The argument in this paper is intended both as a ground-clearing exercise and as a defense of nonpositivism as a theoretical agenda. Recent criticisms of contemporary nonpositivism have raised difficult challenges for those theories. While I have argued that the argument from underinclusivity can be deflated, the argument from overinclusivity expose deeper issues in certain nonpositivist theories, such as the Moral Impact Theory.

In considering how well nonpositivist theories deal with these challenges, however, critics have largely excluded from consideration the theory that to my mind does the best job. By drawing a line between “new” nonpositivist theories and older ones, critics have considered a diluted and uncharitable version of Dworkin's account of legal obligations as those obligations whose enforcement in court is justified. This Judicial Enforcement Theory, divorced from the rest of Dworkin's political and legal theory, becomes a straw man. I have argued that Dworkin's jurisprudential theory is best understood as a one-system view that rests on an account of the moral value of legal practice in constituting a political community in which coercion is regulated in a morally justifiable manner. This theory has the theoretical resources to guard against both under- and overinclusivity.

I hope that the argument in this paper can be taken forward in two ways. First, I hope that both critics of nonpositivism and nonpositivist theorists engaging in internal debates will engage with Law as Integrity as a one-system view. Second, I hope that my analysis shows that the most effective way for nonpositivism to demarcate the legal domain of morality is with a fleshed out theory of the moral principles that are operative in that domain. The further development of competing accounts of the moral underpinnings of legal practice would, I believe, take general jurisprudence in a fruitful direction.