I

The problem of the normativity of law, as it has come down to us from the work of H. L. A. Hart, is that of explaining what it is for a community to treat laws as furnishing reasons to behave in certain ways and to criticize those who deviate.Footnote ¹ Arguably the most important part of explaining the legal practice is to provide an accurate characterization of people's commitments to laws. Such a characterization can be deemed the basic building block of any plausible explanation of the normative nature of the legal practice.Footnote ²

At one point in his admirably clear and fair-minded survey of recent legal philosophy, Jeremy Waldron says that people's commitments to laws are ultimately predicated on others’ like commitments. Waldron asserts this view in the context of discussing an imaginary community of Freedonia:

Let me call this characterization of the normative commitments involved in the legal practice the predication thesis.

There is a version of the predication thesis that is correct, and this correct version captures an important aspect of our commitments to laws. But there are also some subtly different versions that are inaccurate and misleading, and a version of it that is currently quite commonly held by legal philosophers, especially some prominent legal positivists, has led the debate about the normativity of law into a theoretical cul-de-sac. I want to propose an alternative way to understand what Waldron is pointing to, an understanding that will make the problem of the normativity of law more tractable than it has seemed recently. In particular, I shall argue, this new version of the predication thesis is necessary to deal successfully with a hitherto-unnoticed problem that arises from the possibility of persistent controversies about legal systems’ most fundamental criteria of legal validity.

I will begin with a brief description of the currently prevalent version of the predication thesis, which I will call ‘conventionalism’. I will then argue that conventionalism cannot adequately handle an objection by Ronald Dworkin based on the possibility of persistent controversies about the most fundamental criteria of legal validity. I realize that Dworkin's objection has been a topic of a very longstanding debate, already discussed to utter exhaustion on all sides. But I will argue for what may initially appear an implausible proposition – that the true significance of the possibility of persistent fundamental legal controversies has not so far been appreciated by Dworkin or those who have reacted to his arguments. Briefly stated, what such controversies threaten is the conception of our legal practice as consisting of genuinely normative, or rule-mediated, interactions. I will argue that once we appreciate this possible implication of persistent fundamental legal controversies, we will come to see that we need a fundamental revision in our understanding of our commitments to laws. And the revised understanding of our commitments will amount to a new version of the predication thesis that is correct.

I must beg the reader's patience and indulgence in the early sections of this article. In order to set up the new problem I have in mind in a non-question-begging way, I have found it necessary to cover some well-traversed legal philosophical territory in those sections. But even as I do, I will be pointing to some little-noticed landmarks, and taking issue, either explicitly or implicitly, with some standard cartographical claims.

II

Hart famously argued that the basic psychological phenomenon that we should concentrate on to explain the normativity of law is neither the community members’ fear of a superior's sanctions nor their habit of obedience, as Jeremy Bentham and John Austin believed, but instead a distinctive conative psychological attitude on the part of the members that he called acceptance of rules. An acceptance of a rule by a person, according to Hart, consists of that person's dispositions to regulate his conduct in accordance with the rule, to justify his own and others’ conduct by appeals to that rule, and to criticize his own and others’ non-compliance by appeals to that rule.Footnote ⁴

An observer is warranted in concluding that a rule exists or prevails in a community insofar as the bulk of that community accept the relevant rule in the sense just described, and follow the rule as a result. What distinguishes laws (or legal rules) from other rules, according to Hart, is that laws of any single community come in a package of systematized rules with some specific kinds of higher-order rules – what Hart calls secondary rules – that govern the operation of the rules within the package. These higher-order rules include the rule of recognition which regulates the identification of the rules of the system; the rules of change which regulate any modification in the rules of the system; and the rules of adjudication which regulate settling of disputes about the content and application of the rules of the system.Footnote ⁵ An observer is warranted in concluding that a legal system exists or prevails in a community if some powerful subset of the members of the community (call them the ‘officials’ of that community) accept the secondary rules in the sense indicated in the preceding paragraph and follow them as a result, and the rest of the community at least comply with the rules that are valid according to the rule of recognition prevailing in that community.

Hart's claim that a rule exists or prevails among the members of a community if a sizeable portion of them accept the relevant rule and follow it as a result does not amount to a claim that that rule is a conventional rule. According to Hart's account of rule-existence, for example, a rule against spousal rape exists in the current Dutch society. But that is certainly not to say that that rule is a conventional rule in that society. To consider a rule a conventional one clearly requires more than to think that the existence of that rule consists of some convergence of behavior and a widespread conative psychological attitudes in favor of such behavior. Not only must a rule be followed and accepted by the members of a community in that sense, it must be accepted and followed by them as a matter of convention. So what is it for a rule to be accepted and followed as a matter of convention? In the posthumously published postscript to The Concept of Law, Hart says that rules are conventional ‘if the general conformity of a group to them is part of the reasons which its individual members have for acceptance’.Footnote ⁶

What Hart says here is ambiguous. By ‘reasons’ he may have in mind the motives that would be mentioned in explanations of a person's behavior. Or he may have in mind the considerations to which the person himself would appeal to justify his behavior. A person may be motivated to act in a particular way – e.g. to maintain a certain distance in face-to-face conversations – by the fact that a general conformity to act in that way prevails among his peers, whether or not he would appeal to that fact to justify his behavior, and whether or not he even recognizes that fact. But Hart's motivation for speaking of conventional rules is to address Dworkin's arguments to the effect that Hart's legal theory can explain only conventional practices, and the kind of conventional practices that Dworkin has in mind are those in which the relevant persons treat the existence of a convention as furnishing justifying reasons in support of certain kinds of behavior.Footnote ⁷ We can infer then that, according to Hart, rules are conventional if the general conformity of a group to them is part of the justifying reasons that its members have for acceptance, and to which they would appeal to justify that acceptance or behavior complying with the relevant rules.

What is surprising is that many legal philosophers believe that what Hart calls rules of recognition are conventional rules in this last sense. Dworkin was the first to articulate this conception of rules of recognition, and others have followed him.Footnote ⁸ What is even more surprising is that in the postscript to The Concept of Law, Hart himself says that ‘certain important legal rules including the rule of recognition’ are conventional rules.Footnote ⁹ Both of these facts are surprising. First, there is hardly any intimation of such a conception of rules of recognition in the first edition of The Concept of Law and Hart's subsequent writings before the postscript.Footnote ¹⁰ Second, the conventionalist view of rules of recognition is in noticeable tension with a very important element of Hart's legal theory – namely, that acceptances of laws generally and of a rule of recognition in particular on the part of the members of a community may be based on a myriad of different reasons.Footnote ¹¹ Hart makes this last point repeatedly in reaction to a natural law view that such acceptances necessarily amount to moral endorsements. But the same point can be made in reaction to the view that the acceptances are necessarily convention-driven in the sense that Hart specifies in the postscript.Footnote ¹²

Waldron seems to be reflecting the common conventionalist understanding of rules of recognition when he says that ‘the behaviour of other Freedonians will figure among the reasons each Freedonian has for acting and for monitoring self and others’ in accordance with the rule for assembling regularly in the marketplace.Footnote ¹³ And the version of the predication thesis that this common understanding suggests is the following. A community's members’ normative commitments to laws are ultimately predicated on other members’ like commitments in the sense that: (i) the members are committed to individual legal rules partly because those rules are valid according to a rule of recognition; and (ii) the members (or their officials) are committed to that particular rule of recognition partly because others are committed to the same rule of recognition. I will hereinafter use the label ‘conventionalism’ to refer to this version of the predication thesis.Footnote ¹⁴

III

Over the years, Dworkin has consistently made a particular criticism of what he has taken to be Hart's legal theory in particular, and of legal positivist theories more generally. From early on, he has thought of rules of recognition as conventional rules, in very much the sense that Hart specifies in his postscript, and has argued that legal systems can exist even in the absence of such conventional rules of recognition. One point that Dworkin has made is that flourishing legal systems can exist when the most fundamental laws are jointly accepted not as a matter of convention, but out of independent normative convictions on the part of at least some of the relevant officials. Another point is that flourishing legal systems can and actually do exist even when there is no real consensus but instead deep and persisting controversies about the most fundamental laws.Footnote ¹⁵ The distinction between these two points has often been obscured by an assumption, widely shared, to the effect that normative claims, and especially moral ones, are invariably more controversial and consensus-eluding than any factual claims.Footnote ¹⁶ But once we dispense with that questionable assumption, we can see that the two points are distinct.

The first point raises the question of whether a community's members and its officials can be motivated by their moral and other normative convictions in their acceptance of laws, and this question has been the subject of a lively recent and ongoing debate. It is notable that even those legal philosophers who have argued that legal positivism can accommodate the fact that people sometimes accept laws out of their moral and other normative convictions – the so-called inclusive or soft legal positivists – have taken the view that that is the case only if there is a convention among the officials of the relevant community of recognizing moral and other normative considerations as legitimate grounds of legal decision-making.Footnote ¹⁷ This means that, as these conventionalist legal philosophers see things, the ultimate criteria of legal validity in any legal system are accepted as a matter of convention – that is, people accept them because their fellows accept them. But this position obviously fails to respond adequately to Dworkin's first point. Although it accommodates the fact that some laws may be accepted by people out of their independent normative convictions, it does not accommodate the fact that the most fundamental laws may be accepted out of their independent normative convictions.

What is more, these attempts by legal positivists are quite unnecessary. For Dworkin's first point is a challenge to conventionalism, but does not pose any problem for Hart's original understanding of the existence of rules and of laws. As I pointed out before, Hart repeatedly took the view that the members of a community and their officials may accept laws, and a rule of recognition more particularly, for a myriad of reasons including in many cases out of their independent normative convictions. As a matter of fact, Dworkin's insistence that the relevant independent normative convictions are necessarily moral convictions is doubtful, and Hart has rightly objected to it.Footnote ¹⁸ In taking the position that necessarily people or their officials accept the most fundamental laws as a matter of convention, conventionalists are making the same kind of mistake that Dworkin is making.

Dworkin's second point – namely, that flourishing legal systems can and actually do exist even when there is no real consensus but instead deep and persisting controversies about the ultimate criteria of legal validity – also poses an acute challenge for conventionalism. It is difficult to maintain that people or their officials accept the most fundamental laws of their legal system as a matter of convention when there is no convention, or any other kind of convergent practice, of following those laws. The standard conventionalist response to this second point by Dworkin seems to be: (i) to concede that there can be some deep and persistent legal controversies; but (ii) to insist that at the most fundamental level there is a convention or some other convergent practice the existence of which motivates people or their officials to accept laws.Footnote ¹⁹

Such responses are clearly inadequate. Dworkin's second point is based on the possibility of persistent controversies about the ultimate criteria of legality in legal systems. Conventionalists have systematically and rather conveniently underestimated this objection by allowing controversies only up to the penultimate level. The following discussion by Marmor is quite indicative:

But at least in common law legal systems, legal claims as radical and controversial as the one that Marmor is considering would not be out of bounds. The competing legal claims that constitute the controversies of the sort that Dworkin has in mind – e.g. about the proper role of the original meaning in interpretations of constitutional provisions – can be radically novel ones about a legal system's most fundamental criteria of legality. Certainly, Chief Justice Marshall's recognition of the doctrine of judicial review in the famous 1803 American Supreme Court case of Marbury v. Madison was at least as radical and initially controversial as the imagined recognition of the doctrine of precedent in a German court.Footnote ²¹

To insist, in the face of these considerations, that people invariably accept the most fundamental laws of their legal system as a matter of convention is in the end no more useful, and no less Ptolemaic, than to insist in the face of all contrary considerations that all laws are commands.

IV

Unlike Dworkin's first point, his second point poses a real challenge to Hart's original understanding of the existence of rules and of laws. One of the existence-conditions of a legal system in a community, according to that original understanding, is the joint acceptance by the community's officials of a rule of recognition. The possibility of a legal system with no such jointly accepted rule of recognition obviously poses a problem for such an account. In my opinion, however, Dworkin's second point reveals a different and far more serious problem having to do with the nature of the psychological attitude of acceptance. I will get to that more serious problem beginning in the next section after outlining in this section how Hart could have revised his existence conditions of laws to deal with controversies about rules of recognition.

In his discussion of international law in chapter 10 of The Concept of Law, Hart already has provided us with much guidance as to how we could go about addressing the problem about the existence of laws in the absence of a jointly accepted rule of recognition.Footnote ²² Despite the fact that the rules of international law are not held together by a commonly accepted rule of recognition, Hart insists on the genuinely legal nature of such rules because they display other salient features of law. In particular, a set of commonly accepted secondary rules of international law including a rule of recognition, though they are currently non-existent, would not be an absurdity if they were to exist.Footnote ²³ In comparison, for example, such secondary rules of morality would be absurd.

Following Hart's lead, we may take the position that a legal system, or at least a legal regime, exists or prevails in a community if: (i) its members regulate their conduct by a set of rules; and (ii) modifications of such rules, resolutions of disputes about contents of such rules, and authoritative determinations about the applicability of such rules are all treated by the members as amenable to being carried out by appeals to a set of jointly accepted secondary rules. Importantly, (ii) can obtain even when there is no current joint acceptance in the relevant community of any secondary rules. The case for an attribution of a legal system to the relevant community would be strengthened considerably if we were to observe the following further fact: (iii) despite persistent disagreements, the members of the community or their officials carry on deliberations aimed at discovery and maintenance of a jointly accepted ultimate rule of legal validity.

I offer this proposal here only tentatively. Something like it would be needed to amend Hart's original understanding of the existence of laws in light of Dworkin's second point about the possibility of legal systems in which there are persistent controversies about the most fundamental laws. I suspect that with some industry and ingenuity a robustly plausible account along its line could be devised.

There is a far deeper and more serious challenge that Dworkin's second point poses. It is my view that the whole rationale behind the adoption of the picture of legal interactions as being conducted by means of acceptances of and appeals to rules could be compromised by the possibility and actuality of persistent fundamental legal disagreements. What would be required to address this problem is a revision of our conception of the attitude of acceptance of rules. And the revision I have in mind would in effect entail a new version of the predication thesis that is quite different from conventionalism. But I am getting ahead of myself. Let me explain the problem in the course of the next three sections, before going on to sketch what I consider an adequate solution in section VIII.

V

I will henceforth call interactions among a group of people conducted on the basis of acceptances of and appeals to rules ‘rule-mediated interactions’, and any practice consisting of such interactions a ‘rule-mediated practice’. Let me begin by arguing that in characterizing the legal practice as a rule-mediated practice, Hart was characterizing those who take part in the legal practice as committed to a particular norm, in addition to any other norms to which they are committed. That is the case because to take part in any rule-mediated practice is to commit oneself to the norm I have in mind – namely, a norm calling for interpersonal interactions only on the basis of reasons or considerations that are commonly acceptable.

This commitment is displayed in the dispositions of a person who takes part in a rule-mediated practice. Let us say that I take a position on a particular normative issue out of an acceptance of a certain rule. If I were challenged about my position on this issue, or about my conduct motivated by my acceptance, I would appeal to the rule that I accept. In doing so, I would be assuming that my challenger would deem the rule correct and as justifying my position. But if he were not persuaded, then I would be disposed to appeal to some higher-order rules that the challenger would find acceptable, and which would in turn justify the rule that I initially appealed to and also my position. And similar observations could be made of an instance in which I challenge another person's normative position or action by appealing to a rule. To take part in a rule-mediated practice, as Hart describes such a practice, is to possess dispositions to interact with others on the basis of reasons or considerations that are commonly acceptable.

There would likely be limits to how much effort I will put in to find common ground with my audience. But the existence of such a limit would not imply that an attribution to me of a commitment to finding common ground with my audience is unwarranted. It is not my claim that this normative commitment is a commitment to a norm that is overriding or exclusionary. Those who have this commitment may have other commitments that compete with, and that at times (or even usually) override or outweigh, this particular commitment. And they may also believe or presuppose that a full or thorough realization of the norm would be possible only in very favorable circumstances. I believe that people's commitment to the norm of interacting on the basis of commonly accepted reasons or considerations actually has all of these characteristics, and for this reason may be appropriately described as a commitment to an ideal. People who are committed to an ideal need not be motivated by that commitment whenever opportunities present themselves;Footnote ²⁴ only idealists among them are expected to be so. There is an oft-mentioned ideal of rule of law, which is usually contrasted with rule of men.Footnote ²⁵ I suspect that the ideal of rule of law is a particular version of the ideal of interacting with others on the basis of reasons or considerations that are commonly acceptable. This latter ideal can be appropriately called the ideal of rule of rule. In sum, what I propose is that those who take part in rule-mediated practices in general, and in the legal practice in particular, should be characterized as committed to the ideal of rule of rule.Footnote ²⁶

The fact that the legal practice is a rule-mediated practice, which is quite different in kind from a coercion-based practice that Bentham and Austin conceived the legal practice to be, does not imply that the legal practice excludes all coercive measures. Those who take part in rule-mediated practices, and are thereby committed to the ideal of rule of rule, may in some instances resort to coercive and quasi-coercive measures such as cajoling, browbeating, manipulation, etc. That would happen when various commitments that those people have outweigh their commitment to the rule-of-rule ideal. But it is one thing to think, as Bentham and Austin did, that coercive interactions are paradigmatic, or at the heart, of the legal practice; and it is quite another to think, as I do following Hart, that at times the interactions of people who are committed to the rule-of-rule ideal fall short of that ideal, and in such cases they resort to coercive and quasi-coercive measures.

VI

It would be difficult to be committed to an ideal if one were to believe that the ideal could not be realized. The difficulty here is not merely psychological. In fact, it would be less than rational – that is, normatively criticizable – to maintain one's commitment to an ideal if one were to come to believe that the ideal cannot be realized.Footnote ²⁷ I remarked in the preceding section that people who are committed to the rule-of-rule ideal may believe that the full realization of it may be possible only in very favorable circumstances. Be that as it may, such people have to believe that the ideal can be realized, if only in very favorable circumstances; or at least they cannot believe that the ideal is unrealizable.

We may say that the possibility of the realization of the rule-of-rule ideal is a necessary presupposition of a commitment to that ideal. The relevant possibility here is obviously not logical, metaphysical, or even physical possibility. People would not think that it makes sense to be committed to the rule-of-rule ideal, or any other ideal for that matter, if only a logical, metaphysical, or physical possibility of realizing the ideal were thought to obtain. Instead, what would make such a commitment warranted is some natural possibility, where nature is conceived broadly to include the laws of biology, psychology, sociology, and much else. Given, roughly, what human beings, their society, and their physical environment are like, could the ideal of rule of rule be realized? Those committed to the ideal presuppose that the answer is positive.

Another tricky issue has to do with what kind of realization is realization enough. A small group taking part in a rule-mediated practice with very simple rules – e.g. a practice of playing a very simple game – may be able to interact with each other only on the basis of reasons or considerations that are acceptable to all. But such a thorough realization of the ideal would elude any large group taking part in a rule-mediated practice of any real complexity. In a context of the latter sort, the members of a group may decide to delegate the task of determining the acceptability of reasons or considerations to a subgroup consisting of specialists, and thereafter defer to the judgments of those specialists in making determinations of acceptability. A certain counterfactual condition may be built into their deference so that frequent serious misalignments in the judgments of the non-specialists and of the designated specialists would result in a withholding of deference in general, or at least designation of some other members as specialists. In the event that such a division of labor prevails among a group, a realization of the rule-of-rule ideal among the specialists may legitimately be deemed a sufficient realization of the ideal among the whole group.Footnote ²⁸

Certainly, the legal practice is a rule-mediated practice of great complexity. And it is unsurprising to find in any actual community governed by laws that some members are designated as specialists who determine the correct legal rules that are to govern that community, and that the rest of the community defer to the specialists’ judgments as to what the legal rules of that community are. The officials of a legal system are such specialists. It follows that a realization of the rule-of-rule ideal among the officials of a legal system, along with the appropriate paying of deference by the non-officials to the legal judgments of the officials, would amount to sufficient realization of the ideal in that legal system. The possibility of such a realization of the rule-of-rule ideal is a necessary presupposition of a commitment to that ideal in a legal system. Dworkin's second point – that flourishing legal systems can and actually do exist even when there is no real consensus but instead deep and persisting controversies about the ultimate criteria of legal validity – can be thought to pose a problem for this presupposition of realizability.

A community in which the rule-of-rule ideal is realized among its officials would be one in which the officials agree on the most fundamental criteria of legal validity in their legal system. It follows that in order for the members of a community to maintain a presupposition of the realizability of the rule-of-rule ideal in a legal system, they must be able to believe that their officials can come to a consensus on the fundamental criteria of legal validity in their legal system. If the prospect of such official consensus were real, each member of the community, in his interactions with other members, would be able to assume in good faith that any legal judgment that he makes, if correct, would ultimately be acceptable to other members. He would be able to assume that any legal disagreement that may arise between him and others could be settled eventually by appealing to the ultimate criteria of legal validity that are commonly accepted by the officials of his community, or the criteria that would be accepted by the officials in reasonably favorable circumstances.

On the other hand, if there were no real possibility of the officials coming to a consensus on the most fundamental laws of their legal system, then the presupposition of realizability would obviously be put in jeopardy, and with it the commitment to the ideal of rule of rule.

VII

There are some reasons to doubt the prospect or possibility of such official consensus about fundamental laws in modern legal systems. It was John Rawls's view that even when reasonably favorable conditions obtain, there are a number of factors that prevent citizens of a liberal democratic society from agreeing on a single comprehensive ethical doctrine. These factors, which Rawls somewhat curiously calls ‘the burdens of judgment’, include: (i) the complexities of evidence and the difficulties in making correct inferences from them; (ii) diversity of human experiences and the way that experiences shape and shade our ethical reasoning; (iii) underdetermination of ethical conclusions by available evidence; and (iv) the incompatibilities of normative conceptual schemes and the resulting inability of any society to accommodate more than some of them at once.Footnote ²⁹ These burdens of judgment, we notice, are not unique to ethical reasoning. In fact, (i)–(iii) do not even mark off normative reasoning in general from non-normative reasoning. But if we were to conceive normative reasoning as something different from ordinary empirical reasoning (plus mathematical and logical reasoning) as many influential meta-normative theories do, if we were to think of normative judgments as synthetic a priori judgments grounded in but not entailed by empirical conclusions,Footnote ³⁰ then the burdens of judgment that Rawls enumerates appear especially troublesome for normative reasoning. In addition to the difficulties involved in inferring theoretical conclusions from the relevant evidence, there would be additional difficulties in inferring normative conclusions from the relevant considerations including the aforementioned theoretical conclusions. In sum, it would not be surprising to see persistent normative disputes among a large group of people.

In his writings leading up to and including Political Liberalism, Rawls opined that although citizens of a modern liberal democratic society would be unable to agree on a single comprehensive ethical doctrine, nevertheless, they, motivated by their respective comprehensive positions, would be able to come to have an overlapping consensus on a single political conception of justice. In his last published writings,Footnote ³¹ however, Rawls arrives at a more pessimistic conclusion that agreement even on a political conception of justice would likely be beyond reach. Even reasonable people, in very favorable circumstances, freely exercising human reason, would most likely come to disagree on both comprehensive doctrines and political conceptions of justice. This is the fact of reasonable pluralism – or the particularly acute version of it that Rawls eventually recognized – that Rawls believed that any ‘realistically utopian’ political philosophy must face and deal with.

Legal philosophy is not political philosophy. But any plausible philosophical explanation of the normative practice of law must also cope with the fact of reasonable pluralism. If the burdens of judgment make it the case that no comprehensive ethical doctrine or even a political conception of justice could win a common acceptance by the citizens of a liberal democratic society, then it is natural to expect that no single rule of recognition could win a common acceptance by such citizens or even by their officials.Footnote ³² And whatever the scope of our explanatory ambitions, it includes the goal of explaining legal systems that prevail in modern liberal democratic societies of the sort that Rawls discusses. We can speak of the fact of reasonable legal pluralism (or more specifically the fact of reasonable recognitional pluralism), and Dworkin's second point can be seen as bringing that fact to our attention.

The members of a community, shorn of any real prospect of reaching a consensus on the most fundamental criteria of legal validity in their legal system, would have difficulty maintaining the presupposition that the rule-of-rule ideal can be realized. And that in turn would mean that they could not in good faith remain committed to that ideal. Such people would no longer put a premium on interacting with each other on the basis of commonly acceptable considerations. No longer could coercive and quasi-coercive interactions be considered peripheral aspects of their practice. And this means that despite their individual acceptances of and appeals to rules, such people's legal interactions would ultimately be no different in kind from the interactions that are depicted in Bentham's and Austin's earlier legal positivist theories. Here, an important and arguably the most fundamental of the purposes for moving from the earlier legal positivist picture composed out of commands and obedience to that of acceptances of and appeals to rules seems to have been defeated. This is the problem that Dworkin's second point brings to our attention.

VIII

In a recent article, Michael Bratman says that it is ‘useful in the philosophy of action to see certain features of human agency as (at least, implicit) responses to pervasive and systematic problems that human agents face’.Footnote ³³ This is the case in the philosophy of law as well. We can think of the fact of reasonable legal pluralism as a pervasive and systematic problem that human agents living in modern democratic societies face. And as Bratman argues for various conclusions in the philosophy of action by portraying certain features of human agents as their (implicit) responses to a number of problems they face, so we can argue for a new view about the nature of people's commitments to laws as their (implicit) response to the problem posed by the fact of reasonable legal pluralism.

I believe that we can both be cognizant of the fact of reasonable legal pluralism, and maintain the conception of the legal practice as a genuinely normative, rule-mediated, practice, quite different from the coercion-based practice that Bentham and Austin pictured. What is required is a fundamental revision in our conception of the attitude of acceptance of rules. The new view I will advocate can be conceived as a part of the ‘deep structure of our . . . practical thinking’Footnote ³⁴ that we have implicitly developed to deal with the fact of reasonable legal pluralism. What I will be presenting in this last substantive section of the article is a mere sketch of the solution to the problem that I have outlined in the preceding three sections. Much more will need to be said to fill out the proposal.

The gist of the problem that I outlined in the three preceding sections can be formulated as follows. A person who takes a particular normative position N out of his acceptance of some rule R, as Hart conceives the psychological attitude of acceptance, is influenced and guided by the normative force of R in adopting N. And he is influenced and guided by that normative force when he demands, by appealing to R, that his interlocutors adopt N as well. Now, his interlocutors may be unsusceptible to the normative force of R here and now and refuse to adopt N. There may be some higher-order rule that they accept in common, and to which the first person could appeal to persuade his interlocutors to accept R and thereby also come to adopt N. But in the absence of such commonly acceptable higher-order rules, the person's attempts to influence his interlocutors would amount to quasi-coercive prodding. And the fact of reasonable legal pluralism means that, ultimately, people living in modern democratic societies cannot in the end appeal to commonly acceptable rules in their legal interactions.

What this formulation of the problem suggests is that, in order to maintain a genuinely rule-mediated practice, the members of a community who seek to influence each other cannot rely only on the normative force of the rules that they accept. For given the fact of reasonable legal pluralism, sole reliance on the normative force of the relevant rules would oftentimes result in their interactions degenerating into instances of mutual prodding. So what can such people do? How can they maintain a genuinely normative practice that the legal practice aspires to be in the face of the fact of reasonable legal pluralism? What is worth exploring, in my opinion, is the possibility of people relying not only on the normative force of the rules they accept, but also on the fact that their interlocutors are disposed to accept and appeal to certain rules. This last possibility calls for a new conception of the psychological attitude of acceptance of rules.

To facilitate comparisons between the acceptances of rules as Hart conceives them, and the acceptances of rules that I have in mind, let me label them as simple acceptances and plural acceptances, respectively. (The reason for the latter label should become obvious in due course.) According to Hart's picture, a person may come to accept a certain rule R (in the simple sense) for a wide variety of reasons. Once he does, he is influenced and regulated by R's normative force in reaching certain legal conclusions, and when he appeals to it he again relies on its normative force to influence others. On the other hand, a person who has a plural acceptance of some rule S, as I am conceiving that psychological attitude, initially comes to accept S because he deems it likely to elicit acceptances by many others in his community, in addition to any other reasons he had for accepting it. He is influenced and guided by the normative force of S in reaching certain legal conclusions, and he also relies on S's normative force in his attempts to influence others. In these respects, he is like his counterpart with a simple acceptance of a rule. But what distinguishes him is that he is disposed to withdraw his commitment to S should others not take up (sufficiently) his proposal to abide by S, and to make his commitment to an alternative rule should he sense that others are more likely to accept that alternative. This willingness on the part of the person who has a plural acceptance of a rule is not without limits; there would be a range of rules that such a person is willing to commit to, and some rules would be outside this range. Furthermore, his willingness to accommodate and to be responsive to others’ normative dispositions would depend on others having and manifesting the same dispositions.

In sum, a person who has a plural acceptance of a rule does not rely solely on the normative force of the rule that he accepts in his attempts to influence others, but relies also on others’ dispositions to accept certain rules. In his attempts to influence others, he manifests a certain amount of deference towards others insofar as they too are similarly deferential towards him. As a matter of fact, given this deference on his part, it would not be quite accurate to characterize his endeavors merely as attempts to influence others. His endeavors are better characterized as attempts to reach common ground with those whom he is addressing insofar as they too are trying to reach common ground with him. It can be said that such a person's acceptance of a rule is different from the simple acceptance of a rule, of the sort that Hart discusses, in that his acceptance aims at a shared acceptance of a rule. What a person aims at in having a plural acceptance of a rule is to come to accept the relevant rule jointly with others who are similarly aiming at a joint acceptance of a rule. For this reason, each person's acceptance of a rule is initially formed and thereafter maintained on the assumption that the relevant rule is acceptable and would be accepted by others. Such an acceptance of a rule aimed at a shared acceptance can be appropriately called a ‘plural acceptance of a rule’ given that its content can be approximated as: ‘Let us abide by this rule!’Footnote ³⁵

The new conception of the psychological attitude of acceptance of a rule, if adopted as the new basic building block of an explanation of the legal practice, would significantly alter our understanding of that practice. Notice that even in the face of the fact of reasonable legal pluralism, the members of a community who accept rules in the sense that I have specified, invite each other to share in the acceptance of those rules, and adjust their acceptances in light of what is acceptable to each other, would have an easier time maintaining belief in the full realizability of the rule-of-rule ideal in reasonably favorable circumstances. Even people who at one time accept different and conflicting conceptions of the rule of recognition of their legal system can in good faith believe in the prospect of coming to share a single conception of that rule if the acceptances of rules that are involved are plural acceptances. They can in effect believe in, and be guided by, the possibility of a full realization of the rule-of-rule ideal. Of course, such full realization still will not be guaranteed. But such guarantee is not necessary. What is necessary is that there is not any in-principle reason to think that a full realization of the ideal is impossible.

With the ability to maintain the crucial presupposition comes the ability to maintain the commitment to the ideal of rule of rule. And it is this commitment on the part of the members of a community that distinguishes their genuinely normative practice from convergent practices maintained by coercive and quasi-coercive measures.

IX

It would obviously take much more than what I have said in the preceding section to offer a compelling case for the view that people's commitments to laws, especially their commitments to (what they conceive to be) the most fundamental laws of their legal system, are plural acceptances of rules. But in this article, I have been more concerned with setting out a problem (and with creating a dialectical space for that problem) than with offering a solution. I hope to have opportunities to offer greater details and arguments for my solution elsewhere.Footnote ³⁶

I began this article by promising a new version of the predication thesis, and I would like in closing to follow through on that promise. According to the predication thesis, people's commitments to laws in general, and more particularly to what they deem the most fundamental laws of their legal system, are predicated on others in their community having like commitments. In rejecting conventionalism, we have rejected a very tempting and widespread way of interpreting what the predication thesis articulates. But we are now in a position to appreciate that there may be an alternative and more plausible version of the thesis. People's commitments to laws are predicated on others’ like commitments in the sense that those who form plural acceptances of legal rules maintain those acceptances only on the condition that the rules that they accept would be acceptable to others in their community. And the relevant others are those who, in their acceptances of rules, are similarly deferential towards others. Notice that this new version of the predication thesis allows for commitments to rules that are not currently backed up by a convention or some similar kind of practice. A person may accept a rule with the aim or in the hope of instigating a shared acceptance of that rule, as well as in response to an existing shared acceptance of that rule. This particular feature enables the new version of the predication thesis to escape the criticisms of the conventionalist version I outlined before. That is a considerable advantage.Footnote ³⁷