Towards an Account of the State

Robert Bass

We have been familiar with the territorial state, in one form or another, for at least several centuries. Depending on just how it is characterized, it is arguable that it has persisted for the last several thousand years.

Still, the territorial state is not an institution that, like the family (perhaps), appears whenever humans live together. Biologically modern human beings have been around for somewhere between 30,000 and 100,000 years. The earliest examples that might plausibly be claimed to be territorial states date back no more than about 10,000 years. And even in those early cases, much of the world's population did not live within those states. Given the facts that most of the human tenure on this planet preceded the origin of the territorial state and that many have lived outside territorial states since they originated, it is likely that most of the people who have ever lived have not lived as members or citizens of any state.

We can begin with brief discussion of conquest theories of the origin of the state. These are of interest, first, because conquest theories seem well-supported both by such historical records as we have available and by theoretical arguments that no other mechanism adequately accounts for the appearance of the state. Robert Carneiro, for example, says

[A]ll ... voluntaristic theories of the rise of the state founder on the same rock: The demonstrated inability of autonomous political units to relinquish their sovereignty in the absence of overriding external constraints. We see this inability manifested again and again by political units ranging from tiny villages to great empires. Indeed, one can scan the pages of history without finding a single genuine exception to this rule. Thus, in order to account for the origin of the state we must set aside voluntaristic theories and look elsewhere. (Carneiro, p. 734)

More recently, in similar terms, Mancur Olson writes:

There is by now a huge theoretical and empirical literature on this point, and the great preponderance of this literature agrees that ... very large groups are not able to achieve collective goals through voluntary collective action.

Thus we should not be surprised that while there have been lots of writings about the desirability of "social contracts" to obtain the benefits of law and order, no one has ever found a large society that obtained a peaceful order or other public goods through an agreement among the individuals in the society. (Olson, p. 568)

Second, if one of the conquest theories (or some variant upon one of them) is correct, it would provide a straightforward explanation of the connections between states and territories. If it is true that the origin of states from conditions of statelessness occurs (generally or typically) by means of conquest, then there is nothing surprising in the connection of states with territory. "Territory" is just the land area that the conquerors conquered - that is, over which and over the inhabitants of which they have established effective military control.

Some conquest theory may provide a correct account of the origin of states, but it is less satisfying as an account of what it is to be a state. Origins do not determine outcomes; to think otherwise is to commit a genetic fallacy. What is needed is something more like a systematic account. In developing that account, we should remember that states are political institutions that we are able, in general, to recognize even without an account. We are not simply in the dark about them. Hence, an adequate account will be constrained by the need to accommodate as many as possible of our pre-analytic judgments as to what is or is not a state.

I. Sovereignty Accounts

One attempt in the direction of defining what it is for something to be a state has focused on the notion of sovereignty. A state was said to be sovereign within its territory in the sense that it could enforce its will without answering to any other powers. Implicit in this was the claim that some reasonably sharp distinction could be drawn between the state and the population over which it had jurisdiction. That is, in order to enforce its will, there had to be, in addition to the state, some population upon whom its will was enforced. Additionally, it was argued, by Jean Bodin and, later, by Hobbes, that there could be no limits to the authority of the state to enforce its will because any such limit would represent a rival authority and hence, the absence of sovereignty.

The reference to rival authorities highlights a point that was included in but not explained by the sovereignty account. For, of course, it has widely been recognized that states have had, in some sense, to deal with rival authorities. They have had to deal with each other. To begin to make the sovereignty definition work, one has to refer to the territory within which the state exercises its unlimited sovereignty.

However, in the modern world, many entities that are commonly regarded as states do not meet these requirements. Many governments exhibit division of powers among different branches, constitutional limitations upon power, and some sort of ultimate appeal to an electorate. There seems to be in these cases no body of persons nor any identifiable institution which is simply, even within a territory, in a position to enforce its will without being answerable to any other group or institution.

II. Recognition Accounts

Another attempt that has been made is to characterize states in terms of recognition. A state would be something that other states (generally) recognize as a state. Thus, for example, a transnational corporation such as Exxon, though it controls greater resources than many small countries, does not receive recognition or representation at the United Nations. On the other hand, the few city blocks in Rome known as the Vatican does receive representation.

One may be tempted to think, however, that there are regress or circularity problems with a recognition account. It may seem, that is, that "state" is being defined in terms of "state" or that an infinite historical regress of states is required, so that every state could be recognized as such by pre-existing states.

Neither of these problems is really troublesome. What is actually required is that there should be some political arrangement or institution - a conqueror of a territory and its population, say - which exists in proximity to at least one other conqueror of a neighboring territory and population. These conquerors need not, ab initio, have called themselves or the arrangements which they headed "states." However, they could still recognize the similarities in the arrangements, recognize (if that was the case) that neither was in a position to conquer the other, etc. Then, they could apply the term "state" (or some equivalent) to each other and to any other that happened to meet the same criteria. Over time, however, the character of states may change so that what gets recognized by a consensus or near-consensus of existing states no longer has all (or maybe even very many) of the properties of the initial arrangements.

Since the circularity and regress problems can be avoided, a recognition account may do a better job of picking out the institutions that are commonly called states than any definition that focuses attention upon some independently identifiable characteristic or characteristics of those institutions.

Still, one would like to have some more substantive account. Among other things, a recognition account doesn't seem to answer questions about how much recognition is required. (What if half the already recognized states recognized some new entity as a state while half did not?) Second, there are questions about how recognition is signalled. Does willingness to negotiate a binding agreement with some entity some of the actions of which are not legal under the laws of any of the states claiming jurisdiction amount to recognition? (Consider, for example, Israel's negotiations with the PLO or the case a few years ago in which the Columbian government negotiated with a cocaine cartel.)

Perhaps a more substantive account wouldn't track the common usage of the term quite as well as a recognition account, but it might have more explanatory power with respect to the features of the cases it covered.¹

III. The Weberian Definition

Certainly, the best-known attempt to supply a substantive account is Max Weber's:

An imperatively coordinated corporate group will be called "political" if and in so far as the enforcement of its order is carried out continually within a given territorial area by the application and threat of physical force on the part of the administrative staff. A compulsory political association with continuous organization will be called a "state" if and in so far as its administrative staff successfully upholds a claim to a monopoly of the legitimate use of physical force in the enforcement of its order.

It is not possible to define a political corporate group, including the state, in terms of the end to which its corporate action is devoted . . . . [I]t is possible to define the "political" character of a corporate group only in terms of the means peculiar to it, the use of force.²

To understand this, it must be recognized that Weber is using "legitimate" as a sociological rather than as a normative term. That is, he is not claiming that only groups that claim and exercise a normatively correct monopoly on the use of force qualify as states. Rather, by "legitimacy", he is referring to widespread acceptance among the relevant population of the organization's monopoly on the use of force.

Still, Weber's definition seems to require too much in at least two respects. First, there is some ambiguity in what is required for "successfully uphold[ing] a claim to a monopoly of the legitimate use of physical force in the enforcement of its order." In many states, there may be occasions on which the administrative staff successfully enforces its order but in which, with respect to that particular occasion, it is not regarded as legitimate by the population within its borders (or - to take account of actions that a government may take under conditions of secrecy - would not be regarded as legitimate if it were known). Thus, the state may be regarded as legitimate while at least some of its actions are not. A satisfactory account should make the distinction between a (sociologically) legitimate state and (sociologically) legitimate actions of that state.

Second, there are probably few states that actually either claim or exercise a monopoly on legitimate use of force. Almost all allow other parties to use force under some circumstances, e.g., in immediate self-defense.

Third, it may be noted that Weber assumes that there is a connection between states and territory without offering an explanation. But it is not simply self-evident that people who live on contiguous plots of land must recognize and accept the legitimacy of the same holder of a monopoly on the use of force (and therefore, of course, not obvious that it must be a monopoly).³

IV. Nozick's Weberian Account

These difficulties can probably be repaired. Indeed, I shall offer a proposal to repair them below. First, however, I want to examine Robert Nozick's attempt in Anarchy, State, and Utopia to describe a possible origin for the state that satisfies the Weberian definition (or a plausibly improved version thereof) without involving initial conquest and addresses, if it is successful, each of the three problems noted. That is, it first attempts to explain why a state of a certain sort is actually (normatively) legitimate. Presumably, sociological legitimacy can be expected to follow upon that.⁴ Second, it explains how it acquires a monopoly on legitimate use of force (or, better, on deciding and enforcing its judgment concerning what is a legitimate use of force). Third, it explains the territorial dimension of states. If states did come into being in the way that he describes, they would have to have associated territories.

The outlines of Nozick's account are well-known. He begins with people living in a Lockean state of nature without any government. They are assumed to have rights to life, liberty and property. But, following Locke, he says that there are "inconveniences of the state of nature." In order

[t]o understand precisely what civil government remedies, we must do more than repeat Locke's list of the inconveniences of the state of nature. We also must consider what arrangements might be made within a state of nature to deal with these inconveniences .... Only after the full resources of the state of nature are brought into play, namely all those voluntary arrangements and agreements persons might reach acting within their rights ... will we be in a position to see how serious are the inconveniences that yet remain to be remedied by the state, and to estimate whether the remedy is worse than the disease. (Nozick, pp. 10f.)

Broadly speaking, his argument from this point is economic. His claim is that the "voluntary arrangements and agreements persons might reach acting within their rights" in a state of nature will themselves set in motion a series of processes that results in the emergence of the state. There is no stable equilibrium in the state of nature prior to the arrival of what he terms the ultraminimal state.

There is a further process, this time driven not by economic considerations but by the moral rectitude of the personnel of the ultraminimal state, which is said to lead from it to the minimal state - which, within its territory, "is limited to protecting all its citizens against violence, theft, and fraud, and to the enforcement of contracts, and so on". (Nozick, p. 26)

The argument hinges on three transitions: first, from the state of nature to the emergence of a dominant protective agency, second, the transformation of a dominant agency into an ultraminimal state which prohibits competition in provision of protective services, and third, from the ultraminimal state to the minimal state which provides protective services to all within its geographical area. I shall, for the most part, deal with the first and only briefly address the second.

Nozick begins his discussion at the stage at which there are no institutions for dealing with the problems in the state of nature arising out of interpersonal conflict. Even assuming, as he does, that people in a state of nature will generally or typically respect Lockean rights, there are bound to be some violators, and, just as importantly, there are bound to be conflicts between parties who honestly disagree about what their respective rights require. Moreover,

men who judge in their own case will always give themselves the benefit of the doubt and assume that they are in the right. They will overestimate the amount of harm or damage that they have suffered, and passions will lead them to attempt to punish others more than proportionately and to exact excessive compensation .... Thus private and personal enforcement of one's rights ... leads to feuds, to an endless series of acts of retaliation and exactions of compensation. And there is no firm way to settle such a dispute, to end it and to have both parties know it is ended. (Nozick, p. 11)

As a (first) remedy available in the state of nature,⁵ Nozick proposes the formation of protective associations, the members of which would combine against aggressors: "all will answer the call of any member for defense or for the enforcement of his rights." (Nozick, p. 12)

A set of considerations, which I will not here examine further, is offered for the conclusion that these protective associations would evolve into protective agencies offering protection, judicial and enforcement services to clients.

A crucially important phase of his argument concerns the transition from a situation in which there are a number of distinct protective agencies competing in the same geographical area to the situation in which there is, in an area, a single dominant agency. The problem that drives this transition process is conflict between protective agencies when they have some disagreement with respect to a case in which both are involved.

Oddly, Nozick only considers situations in which two agencies are in conflict, though his initial assumption was that there were "several different protective agencies." (Nozick, p. 15) Perhaps, he is assuming (mistakenly) that all disputes have only two parties and that it is safe to generalize from the outcomes of isolated two-party disputes to the general outcomes when there are multiple agencies.

"Only three possibilities," he says, "are worth considering." In the first case, one agency is consistently victorious over the other. Thus, clients transfer their patronage to that agency, which is then dominant within the area.⁶ In the second, "[o]ne agency has its power centered in one ... area, the other in another. Each wins the battles fought close to its center of power, with some gradient being established. People who deal with one ... but live under the power of the other either move ... or shift their patronage..." (Nozick, p. 16) Thus, the claim is made that "[i]n neither of these two cases does there remain very much geographic interspersal. Only one protective agency operates over a given area." (Nozick, p.16) I have doubts about both of these, derived largely from the fact that Nozick excludes from consideration any of the effects that might be consequent upon the presence of more than two protective agencies in an area. But, in many ways, the most interesting of the possibilities he considers is the third:

The two agencies fight evenly and often. They win and lose about equally, and their interspersed members have frequent dealings and disputes with each other. Or perhaps without fighting or after only a few skirmishes the agencies realize that such battling will occur continually in the absence of preventive measures. In any case ... [they] agree to resolve peacefully those disputes about which they reach differing judgments. They agree to set up, and abide by the decisions of, some third judge or court to which they can turn when their respective judgments differ .... Thus emerges a system of appeals courts and agreed upon rules about jurisdiction and the conflict of laws. Though different agencies operate, there is one unified federal judicial system of which they all are components. (Nozick, p. 16)

If the omission of other agencies rendered his arguments questionable before, here it is glaring. Suppose that there were even one more agency involved in a region where the agencies come into conflict on approximately equal terms. Then there could be at least three distinct "court systems" - one to resolve conflicts between each of the three possible agency pairs. If there were three agencies, A, B, and C, there could be a court to which A and B take disputes, one to which A and C take disputes and one to which B and C take disputes. All three agencies would only need to agree on a single court system for those cases in which they were all involved. Even there, there might be subclassifications of disputes that ended up in different courts - one, say, for contractual disputes and another for dealing with violent crimes. But, if that were the case, there would be neither a "unified federal judicial system" nor any semblance of a dominant agency with something like a geographical monopoly.

To this point, I have been assuming that Nozick overlooked the relevance of additional protective agencies to his argument. But there is an interesting and common mistake that might explain the omission and may have led him to think that additional agencies could not make a difference to the outcome. (I do not know that he makes this mistake; I have found no direct textual evidence for it.)

Consider adjudicable disputes, defined as disputes which can be settled by some court decision. Assume that it is desirable that adjudicable disputes be settled by some court decision (rather than by fighting or remaining unsettled). It is often said (correctly) that there must then be a final court of appeal. This admits of two different interpretations. To begin with the incorrect interpretation:

1) There must be a final court of appeal for all adjudicable disputes.

But that gets the quantifiers in the wrong order. (If it were correct, it would be an argument for having a single world supreme court.) The correct order is:

2) For each adjudicable dispute, there must be a final court of appeal.

The second interpretation, of course, carries no implication that all adjudicable disputes, in order to be judicially settled, will have to be settled by the same court (or by a court that is part of the same judicial system). Also, interestingly, it leaves open the possibility of a solution to the problem that the final court of appeal may end up as a partial or biased "judge in its own case" if it should happen to become involved in a dispute. If there need be no final court of appeal for all disputes, then every dispute can, in principle, be brought before some independent court.

I think it is plain that the arguments for the emergence of a dominant agency in a geographical area are unsuccessful. But suppose we follow Nozick a bit further by assuming that somehow a dominant agency has arisen.⁷ He argues that one of the things that the agency may properly protect its clients against is the risk that some other agency may apply unreliable judicial procedures to them and perhaps punish or penalize them (or increase the risk of punishing or penalizing them) for offenses of which they are not guilty. There is some quite dense discussion of this matter in his fourth chapter, "Prohibition, Compensation, and Risk." For the present, I will not endeavor to discuss it in any exegetical detail; I shall simply assume that he gets things right.⁸

His general conclusion, based on that discussion, is that the dominant protective agency may prohibit the risky activities of other protective agencies, provided that the clients of the other agencies are compensated for being deprived of protective services.⁹ It is this prohibition of competition in protective services that marks the transition from a (mere) dominant agency to the ultraminimal state.

Even granting that the dominant protective agency is entitled to prohibit (what it regards as) unreliable procedures applied to its clients, however, Nozick has not shown how the dominant protective agency can (legitimately) transform itself into an ultraminimal state. For, of course, the non-dominant agencies do not deal solely with conflicts between their clients and the dominant agency. They also apply their (supposedly unreliable) procedures in disputes between or among their own clients (who apparently are so obtuse as to prefer them to the more reliable procedures of the dominant agency). And Nozick has offered no reason that the dominant agency would be entitled to prohibit the application of unreliable procedures to those who consent to it. For that matter, he has offered no reason that they can't be applied in inter-agency conflicts, provided they are not applied to clients of the dominant agency. Thus, even if a dominant agency does emerge from the state of nature, Nozick has not shown how a (normatively) legitimate state can arise.

V. A Neo-Weberian Hypothesis

I would like briefly to outline a modification of Weber's account of the state, one which I have found fruitful and which - though it cannot be fully supported here - seems to directly or indirectly deal with each of the three problems mentioned with respect to Weber's definition.

We can begin by considering law. I shall take it that the primary meaning of law is, at least, that it imposes enforceable requirements.¹⁰

What does "enforceability" mean here? There are several components that I wish to distinguish. We can begin with what it does not mean: It does not mean "possible to be enforced." There are many things that can be enforced but which are not thereby matters of law. The mugger, for example, may force me to turn over my wallet. Some things, perhaps, cannot be enforced - such as admiration or love.

Second, to say that a requirement is enforceable does not mean that it is always or usually or commonly or (even) ever enforced. Rather it means that it may, under appropriate circumstances, be enforced. However, with respect to a given requirement, there may be many possibilities of arbitration, etc., short of enforcement. If these are sufficient to secure compliance with the requirement (or rectification after a breach) in a few or most or even in all cases, then the enforcement itself need not occur.

Third, with a forthcoming qualification, I take "enforceable" to be analogous to "desirable." An enforceable requirement is one that it is proper to enforce. The qualification is that I take "proper to enforce" in the way that Weber took "legitimacy." It is not (directly) a normative but rather a sociological classification. A requirement is (sociologically) proper to enforce if, in a relevant population (where the relevant population includes those to whom the requirement is applied), it is widely agreed to be (in some sense)¹¹ proper to enforce.¹²

Now we can say something about law more generally. On one hand, it has identifiable content; on another, that identifiable content has an institutional embodiment. The content of law is, with respect to a given population, the system of enforceable requirements (though, as mentioned in note 10, that system may include elements [laws] that are not themselves enforceable requirements). The embodiment will be the institution or set of institutions defined (at least in part - they may have other functions) by the fact that they provide assurance of compliance with the content of law (to the extent that such assurance can be provided at an acceptable cost), address and deal with adjudicating disputes with respect to legal requirements and, if necessary, enforce compliance or rectification in the event of breach of those requirements.

This account of law can be called "horizontal" in the following sense: It does not view law as necessarily proceeding from a sovereign to subjects or from state to citizenry. Rather, it may arise from among people to order their mutual interaction. When a significantly large number of people - large enough that not many can base expectations with regard to the behavior of many others upon personal trust built out of long familiarity - interact over a significant period of time with respect to matters in which there are large stakes, there will be law (whatever the details of its implementation). There will be enforceable requirements recognized among those people. Otherwise, people who could not rely just on personal trust would not risk the large stakes and the large-scale interaction would not occur.

It will be noted that, in providing an account of law, I have not, so far, relied upon or assumed the operations of government or the state. That was deliberate: I intend to characterize the state in terms of law. What is the relation of government to the content of the law and to the institutions which embody the law for a given population? There is an obvious but mistaken answer to each half of this question: that the state is the (sole) source of law and that the state is or provides the (sole) law-embodying institution.

The first is mistaken because sources of law other than the state are frequently recognized by states. The common law is perhaps the best known example, but many portions of statutory law as well are simply codifications of pre-existing practices. (For example, the Uniform Commercial Code is, in substantial part, such a codification.)

The second is mistaken because there have been (and are) institutions other than the state that have fulfilled the role of assuring, adjudicating and enforcing compliance with the law. The best known of these are probably arbitration associations, but there are many other examples.¹³

Some direction toward a better account of the relation between law and the state can be found by considering (a much simplified account of) a set of distinctions introduced by H.L.A. Hart.

First, it can be noted that not all rules of law are of the same kind. Some of them apply directly to behavior: they require certain actions, performances, transfers of possession, permit certain activities, forbid certain other activities, performances, etc. These can be called primary rules of law. But there is a different class of rules of law, those which "specify the ways in which the primary rules of law may be conclusively ascertained, introduced, eliminated, varied, and the fact of their violation conclusively determined." (Hart, p. 92) These are the secondary rules of law.

Among the secondary rules are rules of recognition that specify how some other rule of law is to be recognized as such or that specify which rule of law takes precedence in the event of apparent conflict. These will typically identify some factual feature that rules of law (of the appropriate class) possess - that it was (or they were) passed by a legislature, established through judicial precedent, that it accords with or violates a constitution, etc.

Among the rules of recognition, there is some supreme rule of recognition which, in the event of conflict, overrides but is not overridden by any other. This, obviously, cannot be identified in terms of some further rule of recognition and so, depends only upon acceptance, directly, by those who are institutionally required to employ it in the identification of other rules of law, and indirectly (at least) by the population which accepts what that institution identifies (or those institutions identify) as rules of law. With this background, I'm prepared to propose a definition of the state or government. It is that institution, organization or agency, in a given society that is widely accepted¹⁴ as possessing a monopoly upon secondary rules of law, especially upon the recognition rules (with the exception of the supreme recognition rule).

This allows the possibility that the state will recognize sources of law and institutional embodiments of law other than itself but reserves to the state, so long as it retains the widespread acceptance upon which it depends, the final say on what is or is not law. So, more briefly (if misleadingly when the qualifications are not understood), the state is a monopolist of law.¹⁵

Comments? I'd love to hear.

References

Benson, Bruce. The Enterprise of Law. San Francisco: Pacific Research Institute for Public Policy, 1990.

Carneiro, Robert L. "A Theory of the Origin of the State" in Science 169 (August 1970), 733-738.

Childs, Roy A., Jr. "The Invisible Hand Strikes Back" in Liberty Against Power ed. Joan Kennedy Taylor. San Francisco: Fox and Wilkes, 1994.

Hart, H.L.A. The Concept of Law. Oxford: Clarendon Press, 1961.

Mack, Eric. "Nozick on Unproductivity: The Unintended Consequences" in Reading Nozick ed. Jeffrey Paul. Totowa: Rowman and Allanheld, 1981.

Nozick, Robert. Anarchy, State, and Utopia. New York: Basic Books, 1974.

Reiman, Jeffrey H. In Defense of Political Philosophy. New York: Harper and Row, 1972.

Notes

1 In the cases in which the substantive account differed from a recognition account - which, it is to be hoped, would be few - it would be possible to amend the substantive account so as to allow for recognition (or absence of recognition) by the entities that qualify as states under the substantive account.

2 Reiman, p. 20. The two paragraphs are excerpts from a longer passage in Weber's The Theory of Social and Economic Organization. All emphases are from the original.

3 If it were true, all states would have to be of continental dimensions - at least if the continent did not have unpopulated regions separating the states.

Consider, for example, those who live on the borders of a state's jurisdiction. There seem to be no particular difficulties with people living side by side on the northern U.S. border and the southern Canadian border while recognizing different states as entitled to their allegiance. If this is only a special case, it is still worth asking why it is a special case and what is different in ordinary cases where such an arrangement could not or would not be expected to work.

4 He assumes that the sequence of events that will result in the emergence of the state will result from efforts on the part of individuals to protect their Lockean rights and that, in the course of this process, the organization that comes to be the state does not in fact have to violate Lockean rights.

5 As an aside, there is a possible remedy for some of these problems that Nozick doesn't consider: performance bonding. The fact that feuding parties cannot trust each other to terminate a feud does not mean that they cannot trust anybody. One party might assure another that he honestly intends to end a feud by turning over assets to a mutually trusted third party who will, in turn, transfer them to the other party if the first party does not adhere to his agreement.

6 Again, what has happened to the other agencies in the area? Even if all inter-agency conflict followed this pattern, the result need not be a single dominant agency. There might be non-transitive relations of reliable victory. (A regularly defeats B which regularly defeats C which regularly defeats A.)

7 I have not, after all, argued that a dominant agency can't arise, just that Nozick hasn't shown that one would.

8 I do not in fact think he gets matters quite right, but I shall not be relying upon any claims that he does not. For some useful critical discussion, see Eric Mack's "Nozick on Unproductivity: The Unintended Consequences." Also worth attention is Roy Childs' elegant argument in "The Invisible Hand Strikes Back" that if a Nozickian minimal state did somehow come into being, a rights-respecting, invisible-hand process would convert it first into an ultraminimal state, then to a dominant agency and finally to just one protective agency among others.

9 The compensation that he proposes is that they be provided with protective services by the dominant agency either free or at a discount

10 There may, of course, be laws that are not associated with enforcement. The U.S. Congress, for example, may proclaim a "national day of mourning (or whatever)" without attaching any sanctions to non-mourning. Second, there may be "dead-letter laws" which have been duly enacted with sanctions attached, but are no longer enforced. Third - a very large class - there are laws which, by creating some legal status, enable a person to do something which he otherwise could not have done, e.g., the law of corporations or the law of agency. I assume that such laws make sense in a system in which there are laws associated with enforcement but not otherwise.

In some cases, such as international law, literal enforcement may never (or almost never) occur because there are significant (and very large) costs associated with literal enforcement. Then, sanctions short of enforcement may be employed indefinitely. This does not, I think, show that the provisions of international law are not or are not regarded as enforceable, just that the costs of enforcement are too high relative to any likely gains from enforcement.

11 I do not mean to pre-judge the sense or senses in which a population may regard a law as proper to enforce. They may range from, at one extreme, judgments that its enforcement is morally required to, at the other extreme, pragmatic judgments that its enforcement is no worse than any of the available alternatives.

12 Can a requirement be sociologically proper to enforce if there is not widespread agreement in the relevant population upon the propriety (in some sense) of its enforcement? I think that depends on whether or not it is part of a system of enforceable requirements which itself enjoys sociological propriety. If it is, it may be. Otherwise, in isolation, it will not be.

13 See Benson, pp. 179-268.

14 This should be understood to be widespread acceptance of the legitimacy of the institution's possession of a monopoly over secondary rules of law.

15 There is, to be sure, some fuzziness in this definition, appealing as it does in two areas to widespread acceptance. (How widespread must acceptance be?) In defense, I offer three considerations: (1) I know of no account, other than simple stipulative definitions, that is able to avoid such fuzziness. (2) The fact that an account leaves some matters fuzzy, does not mean that there are none that it settles clearly. (3) More important than whether an account admits fuzziness is whether it locates fuzziness in the right places. It is an objection to an account that it renders indeterminate matters that are, apart from the account, clear. It is not an objection that it leaves matters indeterminate where, apart from the account, we are also inclined to judge matters indeterminate. Though it would take far too long to argue the case here, I would claim that what I've proposed leaves the fuzziness in the right places.