A. Participants

Who are to count as participants in the practice of law? When he addresses this issue, Hart usually refers casually to “officials of the system.” Other times he distinguishes judges from other officials but treats them collectively as participants in the rule of recognition,Footnote ⁷ while elsewhere he focuses on courts, leaving it unclear whether he also thinks that judges are somehow more important than other officials in determining the content of the rule of recognition.Footnote ⁸ Importantly, though, he invites us to distinguish the requisite attitudes expected from legal subjects and legal officials. He insists that in order for law to exist, it suffices that the bulk of the population by and large obeys the law. In an oft-quoted passage from The Concept of Law, he readily acknowledges that:

Contrary to legal subjects, we cannot speak of a legal system in Hart's sense, unless the officials of the system are regarded as participants in the rule of recognition in the sense that “there should be a unified or shared official acceptance of the rule of recognition containing the system's criteria of validity”;Footnote ¹⁰ put differently, unless officials accept the rule of recognition “as a public, common standard of correct judicial decision, and not as something which each judge does for his part only.”Footnote ¹¹

Hart's descendants depart from this view in two important respects. First, they make much of different types of participation in the practice of law. Second, in many cases they make more room for meaningful participation of legal subjects. I give a couple of examples now, but their importance becomes apparent only once we see, in the next section, how the different roles are associated with different forms of interaction. An illustration of the first point of departure from Hart's account is Scott Shapiro's version of legal conventionalism. Shapiro contends that (1) the practice of law involves a variety of officials performing different roles or occupying different positions in a cooperative enterprise; and (2) this practice of officials has a hierarchical structure. At the top are those participants who assume a position of authority over others, that is, those who have the power to dictate how others must act in matters relating to the joint endeavor. Shapiro is here alluding to the distinction between legislatures and courts. It is courts, he says, that are “subject to the authority of the legislature, for they are required to apply the rules enacted by them [sic]. Indeed, the legislature has the authority it has precisely because judges are committed to treating their directives as reasons for action and modifying their intentions and actions accordingly.”Footnote ¹²

As to the second point of departure, the most forceful advocate of the inclusion of legal subjects in the practice of law is Gerald Postema. He maintains that the point of the practice of law has to do with the guidance of legal subjects (whether his more specific construal of law's point is correct need not detain us here). From this he argues that officials ought to calibrate their acts and decisions not only with a view to the acts and decisions of other officials but also with a view to the patterns of expectation and understanding latent in the society, to which the rules they enunciate are meant to apply. There exists, as he puts it, interdependence between officials and the populace.Footnote ¹³ In fact, he vehemently rejects the suggestion that officials need heed the “beliefs, attitudes and expectations” only of other officials without any concern for ordinary citizens. He writes that:

As suggested above, in order to appreciate the significance of this differentiation of types of participants, we need to explore the kind of interaction contemporary legal conventionalists claim is involved in the practice of law. It is to this topic we now turn.

B. Interaction

What is the nature of the practice of law? Once again, let me contrast the contemporary position to Hart's. In The Concept of Law, Hart is not particularly interested in describing the form of interaction that gives rise to and sustains the practice of law. It is sufficient for his purposes to assert that there is a rule of recognition—and hence also a legal system—just in case all participants apply the same criteria of legal validity and develop the proper attitude, which he calls acceptance, toward this practice of applying the same criteria. Once a settled pattern of behavior of this sort is in place, he thinks, newcomers have merely to attend to it in order to figure out what their duties are under the practice.Footnote ¹⁵ But he does not elaborate on how the settled practice comes about, nor does he take care to analyze the structure of the reasons that support the attitude of acceptance. Of course, to contrast his view to the natural-law position, he maintains that these reasons are not necessarily moral. He allows instead that “allegiance to the system may be based on many different considerations: calculations of long-term interest; disinterested interest in others; an unreflecting inherited or traditional attitude; or the mere wish to do as others do.”Footnote ¹⁶ These disparate reasons then generate the attitude of acceptance that is mapped on the settled practice.

But being primarily concerned with how practices may give rise to obligations of the proper kind, contemporary legal conventionalism cannot afford Hart's explanatory minimalism. It has to flesh out the relationships between participants in legal practice that it takes to be the foundation of legal obligation, in a way that makes it possible for such relationships to generate obligations. The two ideas most commonly invoked are coordination and joint activity.

As is well documented, the coordination model takes its lead from David Lewis's account of convention.Footnote ¹⁷ It supposes that legal practice is a pattern of normative expectations that evolves as a response to a set of recurring coordination problems. We have law because we seek to “guide actions and structure social interaction, thereby providing rational agents with reasons by which they can direct their own behavior,” but on the way to doing this, we come up against a host of coordination problems.Footnote ¹⁸ And law is a way to settle on a solution and thus resolve these coordination problems.

More specifically, law structures social interaction in one or—according to some theorists—two ways. First, it is understood to structure the interaction between legal officials. On this account, legal officials seek to ensure that they will apply the same criteria in determining whether a standard is a legal rule. This, it is argued, is a coordination problem. Legal officials will settle on one scheme of such criteria if they know that this is the scheme that the rest of the officials will also adhere to. The fact that others will abide by it is at least part of the reason why they ought to go along, provided, of course, that they have a reason to find some common scheme of criteria for membership in the legal system in the first place.

Second, recall that for Postema the practice of law is primarily oriented to the guidance of legal subjects' behavior. This generates the obligation on officials to try to situate the rules by which they intend to guide legal subjects within the context of the patterns of understanding and expectation that are already embedded in the wider society. Rules must make sense to their addressees, and once it is accepted that legal subjects are participants in the practice of law, it is a short step to say that officials must, in order to achieve coordination of the requisite type, also take into account the patterns of the subjects' reasoning so as to ensure that what they direct legal subjects to do is at least intelligible to them.Footnote ¹⁹

In light of the serious doubts that have been raised as to the applicability of the coordination model to the legal context, the second strand of legal conventionalism drops the idea that the practice of officials aims to provide a solution to a coordination problem altogether. Instead, it takes its lead from contemporary accounts of joint activity. One of the main sources of inspiration is Michael Bratman's influential theory of shared cooperative activity. Interaction for Bratman takes the general form of “responsiveness to the intentions and actions of the other” with an eye to realizing a certain common endeavor and in the knowledge that others have the same responsiveness (this element is captured by Bratman's conditions of mutual responsiveness and commitment to the joint activity).Footnote ²⁰ Responsiveness requires that participants form and act on individual plans in pursuit of the common goal that—at the very least—do not conflict in a way that makes joint pursuit of the goal impossible. In other words, responsiveness requires participants to strive to make their plans “mesh.” In its most demanding version, interaction also involves a commitment to mutual support, whereby each of the participants in the joint activity undertakes to help others in the completion of the task assigned them.

Legal conventionalists who sign up to this account of joint activity see law as a collective enterprise of officials whose aim is—glossing over differences between the various accounts—to create and maintain a system of law, where this aim crucially includes the application of a common set of criteria for determining a standard's membership in the legal system. Participation in this enterprise gives rise to an obligation on the part of officials to “mesh” their actions with those of others so as jointly to achieve the common aim. On this view, unless judges take into account how their colleagues determine what is to count as law, they will fail to fix on a common set of criteria, and thus the legal system will break apart.

Now, presumably the meshing of individual action may be achieved in many ways. One particularly effective strategy that Scott Shapiro explores is to bestow upon one of the participants the power authoritatively to determine how tasks are to be distributed among the group. When this happens, the common endeavor becomes a “joint activity with authority” (JIAA). In this case, the participant holding the position of authority “intends his or her orders to be reasons for the subject to adopt the content of the order as a subplan as well as reasons to revise the subject's subplans so that they mesh with the orders” and “[t]he subject intends the authority's orders to be reasons to adopt the content of the orders as subplans as well as reasons to revise his or her subplans so that they mesh with the orders.”Footnote ²¹ Shapiro calls this type of interaction, which is shaped by authoritative intervention, “vertical interlock.” The concept of vertical interlock characteristic of JIAAs is central in Shapiro's account of the interaction between different types of participants. We suggest above how it is supposed to work in the legal case. Shapiro takes the joint activity of law to be constituted by the cooperation of different types of officials to the exclusion—in principle—of ordinary citizens. Of these officials, judges have their contributions to the joint activity determined in large part through orders by another type of officials, legislators. Judges regard these orders as reason-giving.

It is important to note that neither the coordination nor the joint-activity version of legal conventionalism insists upon total concord among participants. In effect, both versions make ample room for disagreement. First, they maintain that the practice of law may be sustained even against the backdrop of disagreements over ultimate values and goals;Footnote ²² second, they maintain that participants need not agree on all matters pertaining to their joint activity itself. Most conventionalist accounts in fact insist that so long as participants share a sufficiently strong sense of the joint activity they engage in and of what kinds of mutual accommodation are necessary for it, they may disagree over the requisite course of action in specific circumstances.Footnote ²³ Jules Coleman goes as far as to suggest that the pattern of the joint enterprise itself is subject to continuous negotiation, with disagreements over it being even “settled by an appeal to substantive moral argument.”Footnote ²⁴ Negotiation will be needed precisely because and insofar as participants will not agree with respect to what it takes to carry on with the joint enterprise. However, this bargaining process still takes place “within a particular framework,” the one supplied by one's engagement in the joint activity of officials.

For his part, Shapiro explains the existence of disagreement in joint activity by pointing out that the content of the joint activity and the assignment of tasks it entails are not always transparent to participants. Due to this lack of transparency, disagreement is likely to arise over how to read off one's rights and duties under the practice from its structure.Footnote ²⁵ Importantly, Shapiro, following Bratman, has even allowed for the existence of some level of competition within the framework of a joint activity. Individual participants or groups may take diverse views about the goals to be pursued or about the means to be employed in pursuit of common goals.Footnote ²⁶ Shapiro reasonably supposes that his conception of JIAA may account for this type of competition-fraught joint activity. One of the functions of authorities within the framework of joint activity, he argues, is to resolve competition or at least channel it in ways that do not make it detrimental to the pursuit of the joint activity. Of course, he acknowledges that apart from the types of competition that are accommodated through the introduction of an authority, there are other types that are allowed to play out and indeed are constitutive of the joint activity. Bratman's example for this situation is chess. When we play chess, we do not intend that our intentions mesh “all the way down.” No doubt we would be unable to play a complex game such as chess—actually any game, complex or not, insofar as it is a rule-governed activity—if we did not agree on some ground rules and we were not assured that others would follow them. But we would also not be playing a competitive game such as chess unless we competed over winning it.Footnote ²⁷

Unlike Bratman, other theorists in the second camp opt for what they see as a less demanding conception of joint activity that drops the requirement of mutual support (as a necessary condition of joint activity in general, though perhaps still characteristic of certain types of joint activity). According to proponents of such minimalist conceptions,Footnote ²⁸ the possibility of joint activity is based solely on the existence among participants of the appropriate participatory intention. This is “an intention to do my part of a collective act, where my part is defined as the task I ought to perform if we are to be successful in realizing a shared goal.”Footnote ²⁹ Arguably, this participatory intention is compatible with very low levels of responsiveness. Thus we can say of an employee in a large corporation that he is engaged in a joint activity with his fellow employees, although he merely carries out his assigned task without having to heed how the others perform theirs, just by virtue of the fact that he does what he is supposed to do to contribute to the collective goal. Still, even in such a case I am inclined to think that there is room to speak of some form of interaction, at least in the sense of a prior division of labor and the existence of an expectation that I will not obstruct my fellow employees in the performance of the task assigned them.

We can see how such a minimalist view maps on certain familiar aspects of legal practice. Once an official has performed his assigned task, as defined by the rules that shape his office, his act acquires institutional significance and hence must be reckoned with by other officials. To use Kutz's terminology, it constitutes “grounds for deliberative response.”Footnote ³⁰

To sum up: the conventionalist package contains the idea that law is a practice of interaction between participants occupying different roles within it (who may—and in some variants of legal conventionalism must—include citizens). The idea of such a practice is compatible with the existence of a certain measure of disagreement over its content (and maybe also with a degree of competition between participants).

A. The Straight-to-Morality Assumption

I think both arguments rest on a fundamental assumption about the implications of natural-law theory that I call the straight-to-morality assumption. The following passage is indicative:

I believe that the straight-to-morality assumption poses a formidable challenge to natural lawyers. It does not just threaten to break down the compatibility of natural-law theory and the “conventionalist package.” Actually, on at least one interpretation, it would probably provide a knockdown argument against natural-law theory as a distinctive account of law. But for present purposes, I just want to examine what “looking directly to moral theory” might be taken to mean and whether it has the consequences that Shapiro thinks for a natural-law approach to the joint activity of law.

Let us start by getting a better idea of what Shapiro himself understands by the straight-to-morality assumption; it would help to connect the assumption with the organizational and the Stachanov argument. So, supposing that participants under the natural-law approach must “look directly to moral theory,” what is the problem with it? One idea might be this: the organizational argument seems to suggest that when participants look directly to moral theory, they will not engage in the joint activity unless there is some way for the personal moral convictions of each to agree all the way up, so to speak. So, for instance, if some of the participants are, say, consequentialists and some deontologists, a way must be found to reconcile the consequentialists' concern for the good, impartially conceived, with the deontologists' concern for individual integrity. Now, although much moral philosophy is devoted to bringing together consequentialism and deontology, it is implausible to suppose that the very possibility of joint activities can be made to stand and fall on the success of this otherwise admirable philosophical project.

Such an absurd position, however, cannot be attributed to natural-law theory. The success of most ordinary joint activities depends not on individual plans agreeing all the way up but on their meshing at some appropriate level. As Shapiro puts it: “Members of a group need not agree on every matter related to the sought-after outcome in order to work together in order to achieve it. Each participant can intend to engage in a joint activity but do so for very different reasons.”Footnote ⁵² But since “deep meshing” of subplans is not required in the easy cases, why insist upon it in the hard ones? Stated differently, if, at some appropriate level, both the deontologist and the consequentialist have reasons to engage in the joint activity and work together, why must they stick to a counterproductive and stubborn strategy? Shapiro himself acknowledges this when he writes that on the natural-law approach, participants “might concur on shallower goals.”Footnote ⁵³

Here is another suggestion: even if both the consequentialists and the deontologists eventually find a shared basis for engaging in the joint activity of law, disagreements are likely to arise. Maybe the consequentialists do not give claims of individual rights as much weight as the deontologists do, the result being that in certain cases one group would decide for the defendant and the other for the plaintiff. Now, a good deal of disagreement will be endemic among people who “look directly to moral theory.” But as we saw in the exposition of the conventionalist package, the existence of disagreement is not in itself at odds with the very notion of a joint activity. Besides, people disagree about the nature of the project they are engaging in and about the acts they ought to perform in its pursuit even when they do not have to invoke moral arguments to make a case. They disagree whether their sailing trip should include more stops in popular resorts or in secluded bays. And when they do, they do not thereby manifest a lack of the requisite intention to carry on the joint activity.Footnote ⁵⁴ To be sure, the stakes are typically higher when the disagreement is morally motivated. But even those who disagree on moral grounds may still share enough to stick it out. This is not to say that a joint activity will always be sustained even in the face of the most deep-rooted disagreement. Sometimes participants in a joint activity will disagree so radically that they will feel they are better off opting out of the joint endeavor altogether. But again, this point applies to moral and nonmoral disagreement alike. If my girlfriend does not want to spend the summer holidays in the mountains with me, I have many friends who do.Footnote ⁵⁵

Consider a third possibility. The Stakhanov argument says that by accepting that one's participation in the joint activity is to be governed by one's own personal moral convictions, participants under the natural-law approach are required to place considerable trust in each other's motives. Without such trust, they would have no reason to give each other such a “tremendous amount of discretion” over the content of the plans that constitute the joint activity. This statement, we see below, gives us a more fruitful lead about how “looking directly to moral theory” may be understood by theorists like Shapiro. Of course, Shapiro does not mean to say that moral theory itself gives participants tremendous discretion. Moral theory is more likely to issue requirements that those who look directly to it are bound to follow. It may leave many moral issues indeterminate, thus licensing a choice from among several options. But it is unlikely that Shapiro has discretion of this circumscribed sort in mind, if only because it does not seem to be tremendous.Footnote ⁵⁶

Perhaps, then, what is in play here is the sense of discretion that Dworkin has labeled “weak.” He writes of it that “[s]ometimes we use ‘discretion’ in a weak sense, simply to say that for some reason the standards an official must apply cannot be applied mechanically but demand the use of judgment.”Footnote ⁵⁷ True, under the natural-law approach, many, indeed all, questions about the content of the joint activity of law and the rights and duties flowing from it are apparently left to each participant's personal judgment of the merits. For every course of action, participants must on this view ask: Is this the morally right thing to do? Or more precisely: Does this course of action pass the appropriate moral test of legality? If they answer affirmatively, they will conclude that they must pursue it. Otherwise, they will not (all else being equal).Footnote ⁵⁸ But again, the question arises: If it is true that natural-law theory grants participants in the joint activity of law “weak discretion,” what is wrong with it?

I think the suggestion is that under the natural-law approach, when participants are figuring out what to do, they are presumably totally indifferent to what their fellow participants do. This strategy, it seems, undermines the point of having a joint activity. Joint activities of the kind legal conventionalists are interested in make collective action possible by somehow constraining the behavior of participants. Either for the purpose of coordination or for the purpose of carrying out a joint enterprise, participants are understood to undertake to heed each other's actions and fine-tune their own behavior accordingly. This central assumption, as we seen above, is captured in Bratman's condition of mutual responsiveness. In Section II, we review some of the modes of interaction reflecting this assumption that legal conventionalists have envisaged. Importantly, some legal conventionalists invite us to understand authority structures as themselves modes of interaction whereby participants, in order more effectively to “organize” their behavior, agree to treat the decisions of one among them as binding. The various forms of interaction provide participants guidance when they are uninformed and assurance when they are uncertain, and make each other's actions more predictable. Having participants look directly to moral theory, on the contrary, seems to introduce a good deal of instability in the joint activity, since it putatively encourages participants to ignore others and orient themselves to the course of action that by their own lights appears morally optimal. Where mutual responsiveness ought to have called the shots, now reigns the individual quest for justice.

Here is a different, broader, and perhaps more suggestive way of putting the same point: it is a central desideratum of theories of law that they demonstrate the way in which the law makes—or claims or strives to make—a practical difference to the practical reasoning of legal subjects.Footnote ⁵⁹ Of course, in the context of Shapiro's objection, we are talking about officials and only secondarily about legal subjects. But even though the law probably does not make the same kind of practical difference in the two cases, it is implausible to suppose that it does not make a practical difference to officials as well. Take the case of a judge. We tend to think that his role is crucially determined by the decisions of the legislature. We casually assume that it makes a huge difference to the way he decides a case if the legislature passes a law that is pertinent to it. Notice how the idea of vertical interlock (and more generally the idea of plan-meshing) tracks the same general theoretical concern. It is this aspect of legal practice that, on the view we are considering, sits uneasily with natural-law theory. If the official is guided by his sense of justice, it seems attenuated to say that he need in any way be interested in what other officials have said and done—let alone that what they have said and done may make a practical difference to his institutional duties.

B. Morality and the Joint Activity of Law

Now, this might well be what Shapiro means when he writes that under the natural-law approach, participants look directly to moral theory. But we still need to know whether this is what the natural-law approach is really committed to and whether this is what “looking directly to moral theory” really means. To score an easy point first: A modest natural lawyer might maintain that although the purpose of law is moral and hence we cannot make sense of what is going on in legal practice, unless we see it in the light of certain morally significant goals that we achieve by having it, still this does not mean that individual participants need have any recourse to moral considerations when they are deliberating what it takes to go on with the practice.

This is a position that I think a natural lawyer can conceivably hold, but I am setting it aside because I want to examine whether the robust version of natural-law theory is also immune from Shapiro's charge. Besides, I would not be offering much by way of reconciliation if it were tailored to theories that are shading into legal positivism anyway. Suppose, then, a theory advocates that participants in the joint activity of law cannot but take moral considerations into account in order to figure out what duties and rights they have under that practice. Would such a theory be committed to saying that each participant decides what to do totally unconstrained by what his fellow participants do?

Actually, it is Shapiro himself who suggests a negative answer to this question. In his discussion of substantive goals in stable systems, he writes that:

Of course, the term “looking directly to moral theory” does not convey this indirect strategy of pursuing some goal, but, terms aside, there seems to be no reason why natural lawyers should be indifferent to considerations of efficiency or dismissive of the two-level indirect pursuit of substantive goals that Shapiro refers to in the passage just quoted.

In what follows, I elaborate on this insight and outline a conception of natural-law theory that places it front and center. To this effect, I wish first to introduce a distinction that is important for the argument I am developing. This is the distinction between two different parameters of good governance.Footnote ⁶¹

Take a decision by an institution that is assigned the power to dictate what we ought to do—call it an agent of governance. One dimension along which we assess this decision has to do with its content. This part of our assessment is governed by first-order considerations or considerations of content. First-order considerations determine whether the decision dictates a right or wise or desirable course of action. If the decision distributes benefits and burdens, does it do so justly? If it threatens sanctions for some type of behavior, is the behavior seriously antisocial? Does it respect our rights?

The second dimension, by contrast, concerns the person or institution that has made the decision, the way the person has been appointed or the institution set up, the reasons we have for giving that person or that institution the power to make this sort of decision, and so forth. This dimension is governed by second-order considerations or considerations of institutional design, such as procedural fairness and democratic accountability.Footnote ⁶² So, for instance, part of our reason for giving democratic legislatures such enormous power over our lives has to do with our belief that we are governed well when we are governed by institutions that are accountable to the people in the way that democratic legislatures are.

In the light of the examples I give here, it is easy to mistake the distinction between first- and second-order considerations with the distinction between substantive and procedural values. But although there is bound to be substantial overlap between the two distinctions, they need to be kept apart. To see why, consider the following example. The constitutional assembly enacts a provision that says that the legislature ought to be elected by popular election. Relative to this provision, the relevant first-order consideration is the procedural value of popular election. Our second-order assessment of this provision will revolve around the moral and political standing of the constitutional assembly, whose members may or may not be popularly elected.

What is the appropriate response to second-order considerations, especially insofar as they apply to political institutions? I take it that the character of the response depends on the role one performs—or the position one occupies—in an arrangement that includes agents of governance, but I do elaborate on this point here.Footnote ⁶³ Rather, I focus only on the response that attaches to the role of a judge in a legal system with courts and a legislature. (I assume that the legislature is democratic and that it is a good thing to be governed by democratic institutions.)Footnote ⁶⁴ I try to show that familiar aspects of the courts-legislature relationship, and especially those that Shapiro refers to as “vertical interlock” between different institutional actors in the joint activity of law, can be explained by appeal to democracy as a second-order consideration. I call this value-laden approach of the joint activity of law the sideways account, to indicate the fact that, although it starts from conventionalist ideas, it diverges from the standard legal-conventionalist account of those ideas and gives them a robust natural-law twist.Footnote ⁶⁵

The judge in our example of a legal system with courts and a democratic legislature may, of course, take pride in the fact that he is part of a legal system with a democratic legislature, and such a reaction, too, would be perfectly in order.Footnote ⁶⁶ But the value of democracy is not like the value of a beautiful painting. It is not something just to be looked at. When we live in a democratic state, we do democracy a good service if, among other things, we have the life of our political community effectively governed in accordance with democratic principles. Now, because the judge is a state official and occupies a critical place in the political order, he is expected to contribute to the promotion of democracy in a special way. He is expected, when delivering a judgment, to implement and possibly also further the decisions that have been made by the democratically elected legislature. He may think that on some particular occasion the balance of first-order considerations actually supports a solution that is different from the one dictated by the legislature, but the second-order consideration of democracy gives him reason—though perhaps not always an overriding one—to prefer the latter to the former. By assuming the role of a judge in the legal order he has, to use Rawls's terminology, “abdicated his title” freely to act on the balance of first-order moral reasons.Footnote ⁶⁷ For convenience, I henceforth say that the judge has a duty of respect toward the legislature, by which I mean at a minimum that as far as his official capacity is concerned, he must do his share to give effect to the legislature's decisions.

Now, let me try to connect this account of the judicial role to the problem at hand. How does the duty of respect help us rebut the straight-to-morality assumption? The answer is that the duty of respect crucially shapes the moral test that determines the judge's duties and powers under a natural-law account. The proper formula is not “What does justice require in this case?” but rather “What is in my power to do given that I participate in the joint activity of law and owe a duty of respect to the legislature?” If on the natural-law approach the judge's role were governed by the former question, Shapiro would be right to argue that natural law makes the judge—or any other participant in the practice of law—indifferent to the behavior of his fellow participants. The latter question, by contrast, blocks such a conclusion. For it makes evident that in order to perform his role adequately, the judge must always look over his shoulder to see whether the legislature has decided something that is relevant to the case before him. If he finds in the record of legislative decisions something pertinent, he must give it effect. His reason will not just be that he happens to take part in a joint activity. It is, rather, the more complex reason that insofar as the joint activity he is taking part in is relevantly governed by the principle of democracy, it is pro tanto a morally worthwhile activity to which he has a moral duty to contribute.Footnote ⁶⁸ And the duty this more complex reason gives rise to will be very similar in content to the duty to have one's individual plans mesh in the relevant way with the acts and decisions of one's fellow participants in a joint activity, that is, the type of duty legal conventionalists routinely make reference to.

In short, while it is true that on this view judges look to moral theory, it is still not accurate to say that they look directly to it. By contrast, their reliance on moral theory is crucially mediated by second-order considerations. Thus the idea of second-order considerations goes a long way toward explaining how the law makes a practical difference; it does so because the duties of participants in legal practice are determined by first- and second-order considerations alike.

Note how the preceding remarks help us acquire a better sense of my earlier claim that the natural-law approach does not presuppose any kind of parsimonious “weak description.” It may be recalled that according to this claim, the natural lawyer offers a value-laden explanation of one's duty to contribute to the joint activity of law that is alternative to and competitive with the positivist “weakest description.” Now we can see why. For the natural-lawyer values are constitutive of the joint activity. In fact they are constitutive of it in a way that weaves together the two major themes in the conventionalist package: interaction and plurality of roles. Take the value of democracy as an illustration. In the account of judicial duty put forward in the foregoing paragraphs, democracy has been employed to distinguish both one type of participant from another (legislatures embody this value, while courts do not—not, that is, in the same way) and the type of interaction between the two (by furnishing a moral reason for one to give effect to the decisions of another). It is (at least in part)Footnote ⁶⁹ because the judge has strong democracy-based reasons to heed democratically reached decisions that he has a duty to “mesh” his official behavior with that of the other participants in the joint activity, notably the democratic legislature.

By contrast, I have said nothing so far about another prominent theme in the conventionalist package, namely, the place of citizens in legal practice. To do this at all satisfactorily would take us far from our present concerns. Suffice it to say for the time being that second-order considerations, as is the value of democracy, must have a grip not only on officials such as judges but on private citizens as well. Compliance of the latter with democratically reached decisions may plausibly be thought of as one appropriate response to the value of democracy. It is a different matter—and one heavily depending on the contingencies of specific legal systems—in what further ways this normative grip of second-order considerations is cashed out in terms of citizens' participation in legal practice. But there is, it seems to me, nothing in the structure of the account proposed here that precludes such participation.Footnote ⁷⁰

C. Some Objections

It may be objected to my project that the value of democracy, though clearly relevant to the legitimacy of legislatures in a legal system, is too abstract to account for the content of the duty of courts to apply statutes. Unless we already have the concept of law application, we cannot derive it from some lofty political ideal like democracy. The political ideal might boost our confidence to pursue the task of law application when the law we are to apply has democratic pedigree, but it cannot constitute that task. In response, it must first be pointed out that reference to courts as law-applying institutions, though admittedly embedded in common usage, seems to me to load the dice in the positivists' favor. It presupposes that there is a body of norms that come into existence by virtue of their provenance from the right source and must then be brought to bear on individual cases by the courts. Apart from that, it is not obvious why the value of democracy might not be thought up to the task. Quite the contrary, the link between valuing democracy and giving effect to democratically reached political decisions is both firm and straightforward. That is what I argue judges are called to do. And clearly judges are very well placed to give effect to such decisions. The duty of respect readily follows from these two premises.

But perhaps the objection can be put in a weaker and more plausible form. Now, the objection does not say that democracy cannot make sense of the courts' duty to give effect to democratic decisions at all, but that as a matter of fact courts often (perhaps typically) do not need to invoke democracy at all to justify their giving effect to democratic decisions. Imagine a constitutional provision that dictates: “Courts ought to give effect to the decisions of the legislature.” In such a case, it seems, the value of democracy plays no role in the determination of judicial duty. Instead of resorting to democracy, judges can simply follow the constitutional provision. In this paper I do not make any claims about the function of rules in legal reasoning, nor do I intend to do so. So rather than address the objection head-on, I will try to deflect it. Here is how. Even if sound, the objection does not question the compatibility of natural-law theory with the conventionalist package. For it draws on only a certain view about the nature of rules; it remains silent about the nature of joint activities. Even so, it is a helpful objection because it highlights that the joint activity of law is a project that current participants share not only with their contemporaries but also with those having taken part in it in the past. And the latter, of course, include the constitutional assembly. The sideways account must be able to make sense of this cross-temporal aspect of the joint activity of law. Here I can only indicate how the argument might go. In order to explain the duty of respect the judge owes the constitutional provision, the sideways account would probably have to appeal to reasons having to do with the appropriate body to make this sort of momentous decision and the proper procedure to follow when making it.

Generalizing from the example just discussed, it is worth adding one remark for the sake of clarification. Throughout I am using democracy to illustrate how the idea of second-order considerations can help us account for the constraints in individual action that arise in the context of the joint activity of law in a manner consistent with the tenets of natural law. But of course democracy is not the only second-order consideration that constrains the exercise of institutional power, nor should we suppose that unless we find democracy embedded in a system of governance, that system does not merit being called law on the natural-law account. From this, a couple of conclusions may be inferred. First, in sufficiently complex legal systems, democracy is bound to coexist and often compete with other second-order ideals. In fleshing out the judge's duty of respect in such legal systems, we must bring all these ideals to bear. Second, where democracy is absent, there may be other second-order moral considerations sustaining the joint activity of law and the duty of respect that makes such activity possible. This is not to deny the possibility that a given system of governance is for a variety of reasons such that no scheme of second-order moral considerations may plausibly be said to support a duty of respect on the part of participants. A clear-headed proponent of robust natural-law theory should be prepared to claim that such a system of governance does not generate any type of legal obligation.Footnote ⁷¹

A different objection might be that the value of democracy, even if sufficiently concrete, is not strong enough. At best, democracy gives us only a very weak reason to give effect to democratically reached political decisions. But this is not at all how judges understand their duty to apply statutory law. They believe that application of statutory law actually lies at the core of their institutional role. The suggestion is that if the foundation of the duty of respect is shown to be weak, this would imply that judges may ignore democratic political decisions without much regret and that they may regularly do this in favor of stronger moral reasons such as the reason to do justice in particular cases. This line of reasoning is backed by the familiar thought that in extensive enterprises, occasional violations of one's duties arising from participation in the enterprise do not have serious detrimental effect on its success.

There are a number of counterarguments to this objection. First, my interest is primarily with the foundation of the duty of respect, not its strength or its scope. Accordingly, it suffices to establish that in many cases, the value of democracy can plausibly be said to ground a duty of respect on the part of the judge to assist in carrying out the democratic legislature's decisions. Moreover, it is not part of my argument that the duty of respect exclusively turns on the pernicious consequences of noncompliance—for instance, a state of affairs where the authority of democratic institutions is diminished in the eyes of legal subjects—and thus disappears in the absence of such consequences. There are many reasons, it seems to me, to uphold democratically reached decisions apart from such consequence-driven considerations. If this is true, then maybe the duty of respect is not so weak after all.

Second, it must be insisted that however weak the value of democracy is understood to be in general, it carries more weight as a consideration applying to officials of democratic regimes. As noted above, the position of officials in the legal system makes them particularly apt to give effect to democracy. Still more important, on the account I have presented, their office is defined precisely by appeal to the kind of values they are supposed to serve in the proportions that they must serve them. To be a judge means to act in ways sanctioned by the conception of the judicial role in the legal system. Of course, we can imagine a judge abusing his position, and there is nothing incoherent in this. But even abuses of judicial power make sense only against the backdrop of an understanding of what that power involves in the normal case. And this understanding, I am trying to show, may be thought of as determined by normative parameters such as the value of democracy discussed in this section.

Third, even if we face the objection head-on, we ought to acknowledge that it stings with equal force any conception of our obligations under extensive enterprises, whether these enterprises are defined in terms of a set of moral considerations, as in the natural-law approach, or in value-neutral terms, as in the common legal-conventionalist position. There is nothing in the argument to suggest that the natural lawyer fares worse in providing a sufficiently strong ground for participation in the joint activity of law as compared to his legal-conventionalist counterpart. In fact, it seems that the former can probably do better than the latter. For whereas the promotion of democracy is a cause that could imaginably enlist quite a few willing officials, the cause of “establishing and maintaining a unified system of rules” is rather unlikely to spark an equal measure of enthusiasm.

D. Logistics and Trust

Equipped with what has been said so far, we can now revisit Shapiro's two arguments mentioned at the end of the previous section. So is a value-laden approach to the joint activity of law vulnerable to the organizational and Stakhanov arguments? Let us take each of them in turn. I claimed above that if the organizational argument is taken to suggest that under the sideways account morally motivated participants have to agree all the way up to first principles, it is being unfair to natural law. It is also being unfair if it assumes that morally motivated participants in a joint activity must be committed to advancing the whole of their moral agendas. The sideways account reinforces these claims. Instead of their own moral views, I said above, judges have good second-order reasons to defer to and further the moral views expressed in democratically reached legislative decisions. Of course, when they determine the import of those second-order reasons, they cannot but rely on their own moral view on the matter. The difference is that this time the exercise of moral judgment is not part of a lonely quest for justice that will only accidentally converge with the like quests of other quixotic loners. Rather, it serves the purpose of identifying the way in which the acts of one's fellow participants in the joint activity matter for one's own contribution in a common project. The aim of organizing the joint activity is at its heart.

What largely got the sideways account rolling is Shapiro's claim that morally motivated participants presumably enjoy wide discretion in determining the content of their contribution to the common project, which, recall, is the main premise of the Stakhanov argument. This claim has been shown to be erroneous. Under the sideways account, second-order considerations direct participants to mold their contributions in the light of the acts and decisions of their fellow participants. Undoubtedly, the invocation of second-order considerations does not change the fact that in the end it is moral judgment that calls the shots. In this sense, then, participants still enjoy “weak discretion.” But does this make the requirement of mutual trust extremely demanding, as the Stakhanov argument supposes? I think not; at least not more demanding than in other value-neutral types of joint activity. The reason is that the demand for plan-meshing, which is so central to the success of the joint activity, is not left to the goodwill of each individual but is grounded in moral principles that morally motivated participants ought to uphold.