APA Newsletters€¦ · Theodore Benditt, Editor Spring 2002 Volume 01, Number 2 APA NEWSLETTER ON Philosophy and Law FROM THE EDITOR Theodore Benditt University of Alabama at Birmingham

Volume 01, Number 2 Spring 2001

APA Newsletters

© 2002 by The American Philosophical Association ISSN: 1067-9464

NEWSLETTER ON PHILOSOPHY AND LAW

FROM THE EDITOR, THEODORE BENDITT

ARTICLES

LAWRENCE B. SOLUM

“A Law of Rules: A Critique and Reconstruction of Justice Scalia’s View of the Rule of Law”

JOAN MCGREGOR

“The Rule of Law and Discretion in ‘Billy Budd’”

NOEL B. REYNOLDS

“Legal Theory and the Rule of Law”

GERALD F. GAUS

“The Legal Coordination Game”

RECENT ARTICLES OF INTEREST—ABSTRACTS

RICHARD A. EPSTEIN

“Privacy, Publication, and the First Amendment: The Dangersof First Amendment Exceptionalism”

PAUL G. MAHONEY

“The Common Law and Economic Growth: Hayek Might Be Right”

THOMAS W. MERRILL AND HENRY E. SMITH

“What Happened to Property in Law and Economics?”

PENNEY LEWIS

“Rights Discourse and Assisted Suicide”

RONALD DWORKIN

“Do Values Conflict? A Hedgehog’s Approach”

EUGENE VOLOKH

“Freedom of Speech and Information Privacy: The Troubling Implicationsof a Right to Stop People from Speaking about You”

LIAM MURPHY

“Beneficience, Law, and Liberty: The Case of Required Rescue”

RECENT BOOKS OF INTEREST

Theodore Benditt, Editor Spring 2002 Volume 01, Number 2

APA NEWSLETTER ON

Philosophy and Law

FROM THE EDITOR

Theodore BendittUniversity of Alabama at Birmingham

I want to acknowledge and thank Professors Julie van Camp(Philosophy, Cal State-Long Beach), Lester Hunt (Philosophy,Wisconsin), and William Edmundson (Law, Georgia State)for contributing some of the abstracts that appear in this issue.We should also take note of the efforts that they and othermembers of the Committee on Philosophy and Law (ClaireFinkelstein (Law, U. Penn.), Joan McGregor (Philosophy,Arizona State), and committee chair David Luban (Law,Georgetown)) have made in developing symposia presentedin conjunction with regional APA meetings.

The Current Issue of the NewsletterThis issue of the Newsletter is on “The Rule of Law.” The ruleof law is a central ideal in our thinking about government. Itproposes, at a minimum, that people should be governed bymeans of general rules promulgated in advance, so that theymay know what is expected of them and conform theirbehavior appropriately. The point of the ideal is to eliminateofficial despotism and even to minimize the need fordiscretion on the part of government officials. Ambiguity andarbitrariness are the evils to be avoided. The rule of law issaid to require that the state and its officials be subject to lawand that the law and its application be public, general, andregular. The rule of law, it is sometimes said, is the law ofrules.

The papers in this issue take up two different rule-of-lawissues, and within each, though not by design, they roughlypresent point-counterpoint positions. The first two deal withthe sorts and levels of constraints that ought to operate ondecision-makers, typically judges: how narrow should legalrules be, and, contrariwise, how much discretion ought thereto be? In his “A Law of Rules: A Critique and Reconstructionof Justice Scalia’s View of the Rule of Law” Lawrence Solumconfronts Justice Scalia’a well-known and influentialarguments for the desirability of establishing narrow rules asopposed to deciding cases by weighing circumstances on acase-by-case basis. He argues that this is too simplistic, for inmany contexts general standards can be more limiting thannarrowly drawn rules. His critique of Justice Scalia, though,is sympathetic, maintaining that his views need to besupplemented by considerations of context and character.

In her article “The Rule of Law and Discretion in BillyBudd ” Joan McGregor maintains that the rule of law isconsistent with equity, the exercise of discretion by judges toachieve justice in particular circumstances. This is particularlyso, she maintains, in sentencing those found guilty of crimes.Discretion has often been thought to be problematic becauseit risks putting into judges’ hands the arbitrary power to whichthe rule of law is supposed to be the solution. The issue,then, is how to distinguish legitimate exercises of discretionfrom arbitrary exercises of power. McGregor argues that thereare suitable moral standards that both in principle and in factlimit judicial power. She makes her case within theframework of Herman Melville’s novella Billy Budd.

The next two articles deal with the rule of law in thecontext of political theory; in particular, they focus on the waysin which law can, or cannot, be understood as a device forcoordinating people’s behavior. In “Legal Theory and the Ruleof Law” Noel Reynolds maintains that the rule of law can beunderstood as a set of conditions that rational actors wouldimpose on any authority they would create to act in their steadin creating and administering legally binding rules. Theauthority and obligation associated with law derive from thisfundamental convention, and the principles of the rule of laware the conditions of that agreement, which become therebygoverning principles to which legislatures, judges, andenforcement agencies can be held in their official actions.These generally recognized standards are inherent in thisconventionalist concept of law in the sense that naturallawyers have wanted, but they arise from a social fact, not abackground moral or political theory, thus bridging thepersistent chasm that divides positivist and natural lawtheorists.

In “The Legal Coordination Game,” on the other hand,Gerald Gaus is critical of the idea that the rule of law can beunderstood as the solution to a coordination problem. Thereis a need, despite our disagreements, for common coursesof action on various matters, because some agreement, evenif it is not the one a given person would prefer, is better thanno agreement at all. The question Gaus examines is whetherthe tools of game theory provide a good model of the way inwhich law coordinates people’s behavior; he answers that itdoes not. In particular, he argues, game theory does notmodel our own legal system with its constitutional limitationson law-making powers, nor does it make a good case forremoving these legislative disabilities.

— APA Newsletter, Spring 2002, Volume 01, Number 2 —

— 104 —

LAW

Future Issues of the NewsletterTopics and editors for the next three issues of the Newsletterare:

Fall, 2002FEMINIST JURISPRUDENCE

Submission Deadline: June 15, 2002Editor: Patricia Smith

Department of PhilosophyCUNY-Bernard Baruch CollegeP.O. Box G-143717 Lexington AvenueNew York, NY [email protected](212) 387-1675

Feminist jurisprudence is the analysis of law from theperspective of obtaining justice and equal treatment forwomen. It contains many different facets, focus points, andtheories. Recently some feminists have been focusing onglobal issues and the idea of women’s rights as human rights.It is interesting that the subordination and even the overtpersecution of women is always a civil internal matter,unsuited to intervention by other nations. The horrors ofAfghanistan pose a warning to concerned men and womeneverywhere. In the face of emerging militant fundamentalistsects, economic exploitation, and domestic violence womenare widely at risk, but national sovereignty apparentlyprecludes intervention. Globalism itself poses interestingdebates. If you are interested in submitting feminist work inany of these areas, please send abstract to:[email protected].

Spring, 2003LEGAL ETHICS

Submission Deadline: January 15, 2003Editor: David Luban

Georgetown University Law Center600 New Jersey Avenue, NWWashington, DC [email protected](202) 662-9000

There are a great many interesting issues having to dowith ethics in the law, and so there are few restrictions as totopic for the newsletter issue on legal ethics. We are, though,interested principally in papers of substantial philosophicalinterest, rather than strictly doctrinal or social-scientificstudies. Anyone interested in submitting a paper of up to5,000 words should contact the guest editor directly.

Fall, 2003SCIENCE IN THE LAW

Submission Deadline: June 15, 2003Editor: Susan Haack

Department of PhilosophyUniversity of MiamiP.O. Box 248054Coral Gables, FL 33124-4670(305) 284-6109

Courts rely increasingly on scientific testimony; recentcontroversies over “junk science” in tort litigation and DNAevidence in criminal cases remind us that scientific evidence

can be not only a powerful tool, but also powerfully confusing.This interaction of law and science prompts a whole rangeof philosophical questions: about the honorific use of“science” as an all-purpose term of epistemic praise, aboutscientific method, textbook versus frontier science,disagreement and the pooling of evidence in scientificcommunities, truth and consensus, statistical andepistemological probabilities, interests and bias in inquiry, andthe strengths and weaknesses of the adversary system as away of determining the truth and arriving at justice. Even theU.S. Supreme Court, in its efforts to ensure that when courtsrely on scientific testimony it is not flimsy speculation butdecent work, has found itself tackling the problem of thedemarcation of science (Daubert, 1993, Kumho, 1999), andquestioning the legitimacy of the distinction betweenmethodology and conclusions (Joiner, 1997).

Announcement: Berger Memorial Prize in Philosophyof LawThe Berger Memorial Prize in the Philosophy of Law, a prizeestablished by the APA in memory of Professor Fred Bergerof the University of California at Davis, will be awarded forthe eighth time in the spring of 2003. The prize was madepossible by gifts to the APA from Professor Berger’s friends,relatives, and colleagues following his untimely death in 1986.

The prize is awarded to an outstanding published articlein philosophy of law by a member of the Association. Articlespublished in 2000 or 2001 are eligible for consideration forthe 2003 prize. Members of the APA Committee on Philosophyand Law who will select the winning article are not eligiblefor consideration. (Eligibility of published articles is governedby the date shown on the publication, not by the date of actualprinting or mailing. Eligibility of authors, however, is extendedto all members of the APA in good standing in 2002-2003,regardless of whether they were members as of the date ofpublication.) Submitted articles may have been published inphilosophy serials, law reviews, political science serials,serials in other related fields, or regularly publishedanthologies such as Nomos or AMINTAPHIL volumes.Nominations may be made by the author, the editor, anotherAPA member, or any other individual. The prize, including acash award of $500, will be presented at the 2003 meeting ofthe Pacific Division of the APA, of which Professor Berger wasan active member. If suitable arrangements can be workedout between the winning author and the Program Committeefor that meeting, it is anticipated that he or she will be invitedto participate in a special symposium on the topic of thewinning article at that meeting, to be held in the San FranciscoBay Area in March 2003.

Nominations for the prize should include three copies ofthe article, with the author’s name and other identifyingreferences removed, and a separate cover letter indicatingthe identity of the author. (Nominators need not verify theauthor’s membership status in the APA, but they may wish tosuggest that non-members whose work they are nominatingcontact the National Office of the APA [302-831-1112] forinformation concerning membership.) If a paper is submittedin a language other than English, that paper should beaccompanied by a good translation into English. Allnominations should be sent to: Berger Memorial PrizeCompetition, American Philosophical Association, Universityof Delaware, Newark, DE 19716. Nominations must bepostmarked no later than June 15, 2002.

— Philosophy and Law —

— 105 —

LAW

ARTICLES

A Law of Rules: A Critique andReconstruction of Justice Scalia’s View ofthe Rule of Law

Lawrence B. Solum*

I. The Rule of Law as a Law of RulesA. Scalia’s Thesis: Appellate Courts Should MinimizeDiscretion

In one of the most influential articles in contemporarypractical jurisprudence, Justice Antonin Scalia of the UnitedStates Supreme Court advances the thesis that the rule of lawrequires a law of rules.1 More particularly, Justice Scaliaadvances the thesis that rule-of-law values favor general rulesover all-of-the-circumstances balancing tests as a tool for theelaboration of legal norms in a common-law system. Generalrules limit judicial discretion and promote the predictabilityand uniformity of the law, whereas case-by-case multifactorbalancing increases judicial discretion and hence results inunpredictable decisions and disparate treatment of litigants.

Scalia advanced this thesis in “the narrow context of lawthat is made by the courts,”2 a context that is especiallyprominent in common law systems where the judicialdecisions can have binding precedential effect. The courtsarticulate this precedential effect as a set of technical legalnorms called the law of stare decisis and the doctrine of lawof the case.3 In the United States, the lawmaking function ofthe courts extends beyond the relatively narrow realm of truecommon-law where the main body of rules is judge-law (withrelatively few statutory modifications) to the broader realmof common-law style interpretation of statutes andconstitutional provisions that are framed in abstract andgeneral terms. Following standard usage, I shall refer tojudicial lawmaking in both the narrow and the broad sensesas common-law adjudication.

B. The Complex Structure of ConstraintScalia observed that the common-law decisions of appellatecourts differ with respect to the degree with which theyconstrain the discretion of future courts that are bound bythem. The example with which he illustrates this point isworth repeating:

In deciding, for example, whether a particularcommercial agreement containing a vertical restraintconstitutes a contract in restraint of trade under theSherman Act, a court may say that under all thecircumstances the particular restraint does notunduly inhibit competition and is therefore lawful;or it may say that no vertical restraints unduly inhibitcompetition, and since this is a vertical restraint it islawful. The former is essentially a discretion-conferring approach; the latter establishes a generalrule of law.4

For ease of reference, I will refer to Scalia’s example asthe Sherman Act Case. This case is highly stylized; we don’thave any particular description of the facts and procedural

posture of the dispute that might give rise to the choice posedby Scalia. It is clear from context, however, that the choice isto be made by an appellate court faced with a record in whichsome producer was accused at trial of violating the ShermanAct by transactionally limiting the conduct of wholesalers orretailers with whom the producer deals. For example, thecase might involve a widget manufacturer requiring that allthe widget retailers with which it deals to adhere to a set priceschedule. Scalia’s case assumes that the appellate court willfind for the producer; the outcome in this sense is not at stake.Scalia imagines that the appellate court is faced with a choicebetween two rationales on which to base its decision andthat the chosen rationale will function as a binding precedenton courts faced with similar cases in the future. The firstalternative is to find for the producer on the ground that “underall the circumstances the particular restraint does not undulyinhibit competition and is therefore lawful:” call this the all-the-circumstances balancing approach. The secondalternative is to find that “no vertical restraints unduly inhibitcompetition:” call this the general-rule approach. Scaliaargues that the former approach is discretion conferring, whilethe latter approach is discretion limiting.

What is the real difference between the two alternativesthat Justice Scalia has sought to capture with the ShermanAct Case? One possibility is that Scalia is pointing toward thewell-known distinction between “rules” and “standards,”5 butthis distinction captures only part of the phenomena withwhich Scalia is concerned. Scalia’s “general rules of law”might be identified with the category of “rules” and Scalia’sall-the-circumstances test might be an example of a“standard.” But this binary opposition between rules andstandards is unsatisfactory, because rules vary in a multiplicityof dimensions that affect judicial constraint.

To understand Scalia’s point we need at least apreliminary understanding of how and why legal norms differwith respect to constraining judicial decisions. A rough andready sense of the complex ways that legal norms differ intheir constraining force might be articulated as follows. Legalnorms can be more or less general or specific, more or lessabstract or concrete, and more or less universal or particular.The operative language of a rule (the words and phrases thatdo the work) can be the subject of widespread interpersonalagreement in judgments about application or of deep andintense controversy in such judgments. These characteristicsof rules affect the extent to which rules are discretionconferring or discretion limiting.

General, abstract, and universal rules will have a widedomain of application, and hence will have a broader scopeof constraint. Take the Sherman Act itself: this federal statuteattaches legal consequences to transactions or relationshipsthat constitute an “unreasonable restraint of trade.” On theone hand, this norm covers a lot of ground; it is applicable toa wide variety of human conduct and hence to a wide varietyof potential legal disputes. On the other hand, the“unreasonable restraint of trade” norm illustrates the fact thatabstract, general, and relatively universal language can leavea great deal to individual judicial decision. “Unreasonable”is an abstract term about which there is considerabledisagreement in judgments. Hence, the statutory languagemay leave a wide range of options open to judicial decision.The broad language constrains judges to apply the statute toa wide variety of conduct, but unless and until the phrase


— 106 —

LAW

“unreasonable restraint of trade” is given more concrete andparticular meanings by judicial interpretation, judges haveconsiderable options for choice in how it is applied.

Specific, concrete, and particular norms have a narrowerscope of application. The second alternative posed in theSherman Act Case seemingly takes this to the limit. JusticeScalia’s hypothetical case posited an appellate decision in aparticular case that holds, “under all the circumstances theparticular restraint does not unduly inhibit competition andis therefore lawful.” This norm is obviously particular,concrete, and specific. At first blush, one might think thatthis norm does not constrain at all. Any future case, it mightbe argued, will be different with respect to somecircumstances, and hence will be unconstrained by thedecision. This is an oversimplification. In actuality, theparticularistic variant of the Sherman Act case would involvevery powerful constraints on discretion.

How might a decision of the Sherman Act Case by an all-of-the-circumstances approach actually constrain discretion?Consider two distinct answers to this question. First, anappellate court that ruled in this particularistic way wouldimpose the most stringent sort of constraint on the discretionof the lower court to which the case was remanded. Therewould be absolutely no room for an intermediate appellatecourt or trial court to reach the conclusion that there was anunreasonable restraint. In comparison, an appellate courtthat held that vertical restraints were per se valid and thenremanded for “further proceedings not inconsistent with thisjudgment,” might well allow for discretionary determinations,for example on the question whether the actions at issue weretruly a vertical restraint or not. Lawyers call this form ofconstraint, “law of the case.” The law-of-the-case doctrinebinds lower courts to follow, after remand, the rule laid downby the appellate court upon appeal.

Second, even the particularistic variant of the ShermanAct Case would likely involve significant constraints on thedecisions of future courts. This is because the surface formof a decision holding “under all the circumstances theparticular restraint does not unduly inhibit competition andis therefore lawful” is misleading. Let us zero in on the phrase“under all the circumstances.” If read literally this wouldmean all the circumstances, including rigid designators suchas the names of the particular firms involved in the dispute,the dates on which the events occurred, and even the colors,shapes, or odors of the products involved. Thus, the under-all-the-circumstances norm might not apply to very similarbusiness practices if the defendant were Microsoft rather thanIBM; or the events occurred in 2002 rather than 1998; or theproduct was Windows rather OS2. In other words thedecision would have no constraining force beyond theparticular set of events, i.e., these transactions between thesefirms in these products on these dates in these physicallocations. One might imagine a lawyer arguing to a trial courtjudges, “All the circumstances means all the circumstances,and so you are free to ignore the Supreme Court’s decision inthe Sherman Act Case. Our case occurred on different datesbetween different firms.”

In the real world, however, such an argument would not“pass the laugh test.” The phrase “under all thecircumstances” is understood to mean “all the legally relevantcircumstances,” and the reasoning in the opinion will makeit clear that some circumstances were relevant and others

were not. Abstractly, we might imagine that the legallyrelevant circumstances included six or seven salient factors,such as relative size and market power of the firms, whetherthe defendant had a large share of the relevant market, andso forth. In future cases in which all of these factors werepresent to the same or a greater degree than they werepresent in the Sherman Act case, the particularistic variant ofthe holding would strongly constrain the decision of lowercourts. In the law of stare decisis, this is called “being on allfours” with the binding precedent. This circumstance, ofbeing relevantly similar in all respects relevant to a multi-factorall-the-circumstances norm, is not so rare as one mightimagine, but whether a subsequent case being on all fourswith a prior case is a rare or a common event, when thishappens, the discretion of the lower court is stronglyconstrained.

As compared to a relatively general and universal rule, arelatively specific and particular legal norm is both less andmore constraining. Specific and particular norms have anarrower scope, but within the narrow scope in which theyoperate the constraint may be deeper and more pervasive.There is no simple function that makes general, abstract, anduniversal rules more constraining than specific, concrete, andparticular ones.

The story is further complicated by the fact that the degreeof interpersonal agreement in judgments about theapplications of a norm is not a simple decreasing (orincreasing) function of abstraction, generality, or universality.Some very general and abstract norms are the subject ofwidespread interpersonal agreement; some very concrete,specific, and particular norms are the focus of intense andpersistent disagreement in judgments. Before we illustratethis point with legal norms, we should observe that thisphenomenon is quite general. Consider the follow pair ofdescriptions: “seven” and “messy office.” “Seven” is a veryabstract, general, and universal term. It applies in countlesscontexts. Although there may be hard cases, there will be aremarkable degree of interpersonal agreement in judgmentsover questions like, “Are there seven clothespins on the line,”or “is seven a prime number?” On the other hand, the phrase“messy office” is relatively particular, concrete, and specific.To borrow some jargon, “messy” is a thick term, one that ishighly embedded in particular social practices. Yet it is nothard to imagine considerable intersubjective disagreementabout the question whether my office or your office is a“messy office.”

This general point about the lack of a simple relationshipbetween constraint and degrees of abstraction, generality, anduniversality can easily be illustrated with legal norms.Consider the following pair of norms. The first norm isprovided by Article III of the United States, which states thatjudges who exercise “the judicial power of the United States”shall hold their positions during “good behavior.” “Goodbehavior” is a relatively abstract and general phrase, but thereis near unanimity about the meaning and scope of applicationof the good behavior clause: it grants the judges of the federalSupreme Court, Courts of Appeals, and District Courts lifetenure in their positions, subject only to removal byimpeachment. The second norm is provided by the eleventhamendment to the United States Constitution: “The Judicialpower of the United States shall not be construed to extendto any suit in law or equity, commenced or prosecuted against


— 107 —

LAW

one of the United States by Citizens of another State, or byCitizens or Subjects of any Foreign State.”6 The eleventhamendment is framed in relatively specific, concrete, andparticularized language, but there has been more than acentury of deep and persistent disagreement in judgmentsabout the applicability of that language to particular cases.7

Here is another example of how ad hoc balancing testsmay actually provide more constraint than general rules. Thelaw of personal jurisdiction determines whether a court willbe able to force a defendant to appear, at the risk of a bindingjudgment in case of default. In United States law, this questionis governed in part by the due process clauses of the fifth andfourteenth amendments to the Constitution. A complex setof rules has emerged out of the Supreme Court’sdetermination that due process is satisfied if the defendanthas such minimum contacts with the forum jurisdiction thatthe assertion of jurisdiction does not offend traditional notionsof fair play and substantial justice.8 Case law has developedthis general “minimum contacts” standard into a complexlyarticulated rule, including two parts. First, courts ask whetherthe defendant has engaged in actions that constitutepurposeful availment of the benefits and protections of thelaws of the forum states. If the answer to this question is yes,the second inquiry involves a multifactor balancing test thatconsiders the plaintiff ’s interest in filing the suit in the chosenforum, the forum jurisdiction’s interest in hearing the case,the burden on the defendant of appearing in the forum, therelative efficiency of the forum considering the location ofwitnesses and evidence, and any other public policy concernsthat might bear on the decision.9 The first inquiry looks like itinvolves a general rule. The second inquiry is clearly amultifactor-balancing test. One might expect that the firstinquiry would be relatively more constrained than the second,but in fact, the opposite is true. There has been three decadesof controversy over the meaning of the purposeful availmentrequirement, but courts almost never disagree about theproper application of the balancing test. This pattern wasdramatically illustrated by the Supreme Court’s decision inAsahi, where the Justices fractured over the meaning of thegeneral rule but agreed almost unanimously about theapplication of the balancing test.10

There is yet another way in which the relationshipbetween the form of legal norms and their ability to constraindiscretion is complex. Disagreements in judgment about theapplication of legal norms can be structured in a variety ofways. Sometimes the whole field of play in is in contention.That is, sometimes there is a radical disagreement about themeaning of a norm. More frequently, there is a zone ofagreement, a limited field of contention. Thus, it might bethe case that classical cartels and other price-fixingagreements between competitors are generally agreed to beunreasonable restraints of trade while there is argument overwhether a manufacturer requiring retailers to sell at aparticular price is unreasonable or not. H.L.A. Hart uses themetaphor of core and penumbra to express this sort ofdisagreement in judgments.11 Assuming we agree on the fieldto which a given legal norm applies, we might imagine thatsome legal norms create relatively bright lines without anysubstantial penumbral zone whereas other rules create largepenumbral zones in which there are substantialdisagreements in judgments about application.

Justice Scalia’s discussion of these matters is quick, buthe recognizes at least some of the complexity that I haveexplicated. Thus, he acknowledges that constraint “is all amatter of degree.”12 Scalia’s thesis is that legal norms shouldconstrain broadly and deeply. Implicit in this thesis is theidea that legal norms should be the subject of widespreadintersubjective agreement in judgments about application.Scalia’s ideal legal norm would be phrased in generallanguage that sweeps broadly but constrains deeply, drawingbright lines and minimizing the size of penumbral zones.Thus, in the Sherman Act case, his assumption is that a per serule that no vertical restraint of trade is unreasonable will applyto many factual situations and that there will be relatively fewdisagreements about the *rules*-rule’s- application. Althoughthere may be some penumbral cases, in which it is unclearwhether a given transaction constitutes a vertical constraint,the per se rule creates a relatively bright line, sorting mostbusiness practices into the clearly-a-vertical-restraint categoryor the clearly-not-a-vertical-restraint category.

C. Scalia’s Argument for Discretion-Constraining Norms

1. Legitimacy Requires the Appearance of EqualTreatment

Scalia’s first argument for preferring discretion-constrainingnorms is that the appearance of equal treatment is requiredfor legitimacy. As he states the argument,

[O]ne of the most substantial…values [favoringjudicial general rules of law that constraindiscretion]…is the appearance of equal treatment.As a motivating force of the human spirit, that valuecannot be overestimated. Parents know that childrenwill accept quite readily all sorts of arbitrarysubstantive dispositions—no television in theafternoon, or no television in the evening, or evenno television at all. But try to let one brother or sisterwatch television when the others do not, and youwill feel the fury of the fundamental sense of justiceunleashed. The Equal Protection Clause epitomizesjustice more than any other provision of theConstitution. And the trouble with the discretion-conferring approach to judicial law making is that itdoes not satisfy this sense of justice very well. Whena case is accorded a different disposition from anearlier one, it is important, if the system of justice isto be respected, not only that the later case bedifferent, but that it be seen to be so. When one isdealing, as my Court often is, with issues so heartfeltthat they are believed by one side or the other to beresolved by the Constitution itself, it does not greatlyappeal to one’s sense of justice to say: “Well, thatearlier case had nine factors, this one has nine plusone.” Much better, even at the expense of the mildsubstantive distortion that any generalizationintroduces, to have a clear, previously enunciatedrule that one can point to in explanation of thedecision.13

Scalia’s argument is not that general rule-like norms doprovide equal treatment required by some norm of proceduralfairness, but they that they are required to create theperception of fairness that is essential to the legitimacy ofjudicial decisions.


— 108 —

LAW

Thus, Scalia’s legitimacy argument recognizes thepossibility that perception and reality might cleave. A simple,general rule might create the appearance of uniform andequal treatment, while subtle disagreements in judgmentsmight really result in cases that were truly alike in relevantrespects being treated differently in fact. Scalia’s firstargument is aimed at the appearance and not the reality ofequal treatment. From this it follows that Scalia is likely to beconcerned with legitimacy as a social fact and not as anormative standard. In other words, general rules tend toincrease citizens’ beliefs that they are being treated fairly andhence that their government is a legitimate source of norms.This is, of course, a claim about sociological fact, and Scaliaprovides almost nothing in the way of support for this claim.14

Scalia does not take his argument to the next step andexplain why perceptions of fairness and legitimacy ought tobe pursued, but we can easily fill this gap in his argument.Perceptions of legitimacy are likely to increase voluntarycompliance with legal norms. Voluntary compliance is likelyto reduce the costs of norm enforcement. In extreme cases,perceptions of illegitimacy may lead to violent resistance tothe law. In ordinary cases, it might lead to passivenoncompliance with judicial orders. It is easy to imagine arobust case for beneficial consequences of perceivedlegitimacy.

2. Predictability and UniformityScalia’s second argument is that general rules of law will doa better job of creating the reality of uniformity of treatmentand hence predictability of results. This argument has afactual premise, that general rules do a better job thanmultifactor totality-of-the-circumstances tests in creatingintersubjective agreements in judgments. His particularwarrant for this claim is based on a set of institutional factsabout the judicial system in the United States. It might beargued that appellate review by the United States SupremeCourt could bring the judgments of lower court judges intoagreement about the application of a totality-of-the-circumstances test. By way of rejoinder, Scalia argues:

The fact is that when we decide a case on the basisof what we have come to call the “totality of thecircumstances” test, it is not we who will be “closingin on the law” in the foreseeable future, but ratherthirteen different courts of appeals—or, if it is afederal issue that can arise in state court litigation aswell, thirteen different courts of appeals and fiftystate supreme courts. To adopt such an approach,in other words, is effectively to conclude thatuniformity is not a particularly important objectivewith respect to the legal question at issue.15

To this factual premise, Scalia adds a normative one.Without uniformity in application, the law is unpredictable.Without predictability, those affected by the law cannotdetermine whether or not their conduct will give rise to legalconsequences: “Rudimentary justice requires that thosesubject to the law must have the means of knowing what itprescribes.”16 This argument is based on the idea that it isunjust or unfair to impose negative legal consequences thatcannot be avoided by persons who take reasonable steps tocomply with the law. More could be said about why this so,but this shallow statement of the argument is likely to beaccepted by those who adhere to a variety of moral theories.

There is, of course, a complementary consequentialistcase for the predictability and certainty of legal normapplication. Predictability and certainty are likely to facilitateplanning and complex social cooperation. The efficiencygains made possible by contract law require that the partiesto an agreement be able to reliably predict whether theiragreement will be enforced. The deterrent effect of criminaland tort law depends in part on the perceived likelihood thatpenal sanctions or tort damages will be imposed.

3. General Rules Constrain Arbitrary Judicial PowerScalia offers a third argument favoring judicial rules, one thatfocuses on the notion that general rules constrain arbitraryjudicial power. On a surface level this argument appears tobe tautological and empty. If Scalia’s institutional and factualarguments are correct, then general rules constrain discretionand totality of the circumstances tests do not. But withoutsaying more, it is not clear why this counts as a rationale forlegal norms that constrain. Thus, the following passageappears to be question begging:

I had always thought that the common-law approachhad at least one thing to be said for it: it was thecourse of judicial restraint, “making” as little law aspossible in order to decide the case at hand. I havecome to doubt whether that is true. For when, inwriting for the majority of the Court, I adopt a generalrule, and say, “This is the basis of our decision,” Inot only constrain lower courts, I constrain myselfas well. If the next case should have such differentfacts that my political or policy preferences regardingthe outcome are quite the opposite, I will be unableto indulge those preferences; I have committedmyself to the governing principle. In the real worldof appellate judging, it displays more judicial restraintto adopt such a course than to announce that, “onbalance,” we think the law was violated here—leaving ourselves free to say in the next case that,“on balance,” it was not.17

The missing normative premise comes and goes very quicklyin the remainder of the paragraph:

It is a commonplace that the one effective checkupon arbitrary judges is criticism by the bar and theacademy. But it is no more possible to demonstratethe inconsistency of two opinions based upon a“totality of the circumstances” test than it is todemonstrate the inconsistency of two jury verdicts.Only by announcing rules do we hedge ourselvesin.

The key word here is arbitrary. Judges need to be hedged in,because if they are not, they may make arbitrary decisions.

Scalia’s argument is still incomplete. We may agree that“arbitrary” decisions are bad ones, but it may be important toknow why this is the case. Scalia could appeal to a variety ofarguments. Some will fold back into his other arguments.Arbitrary decisions may undermine legitimacy and produceunfair unpredictability. A different line of argument might bebased on the idea that unconstrained arbitrary power is likelyto be abused, imposing costs on society at large to serve thepersonal whims or even narrow self-interest of judges.


— 109 —

LAW

4. General Rules Facilitate the Protection of IndividualRights

This argument is closely related to Scalia’s legitimacy claim.The core idea is that general rules are more likely to beperceived as legitimate and hence more likely to “stick” whenthey are politically unpopular:

While announcing a firm rule of decision can thusinhibit courts, strangely enough it can emboldenthem as well. Judges are sometimes called upon tobe courageous, because they must sometimes standup to what is generally supreme in a democracy: thepopular will. Their most significant roles, in oursystem, are to protect the individual criminaldefendant against the occasional excesses of thatpopular will, and to preserve the checks andbalances within our constitutional system that areprecisely designed to inhibit swift and completeaccomplishment of that popular will. Those are taskswhich, properly performed, may earn widespreadrespect and admiration in the long run, but—almostby definition—never in the particular case. Thechances that frail men and women will stand up totheir unpleasant duty are greatly increased if theycan stand behind the solid shield of a firm, clearprinciple enunciated in earlier cases.18

The core idea here is that judges are more likely to takecourageous stands in favor of unpopular but just decisions ifthey are able to justify their actions by general rules.

D. The Law of Rules and Rule of Law ValuesAt this point, it should be clear that Scalia’s arguments for“the law of rules” are similar or identical to the valuestraditionally associated with “the rule of law.” Thus, A.V.Dicey’s historically influential formulation of the rule of lawincorporated three ideas: (1) the supremacy of regular lawas opposed to arbitrary power; (2) equality before the law ofall persons and classes, including government officials; and,(3) the incorporation of constitutional law as a binding partof the ordinary law of the land.19

John Rawls articulated a more elaborate list of therequirements of the rule of law, which he defined as “theregular, impartial, and in this sense fair” administration of“public rules.”20 In schematic form and with some alterations,Rawls offered the following conception of the rule of law:21

1. The Requirement that Compliance Be Possible. Thelegal system should reflect the precept that oughtimplies can.a. The actions which the rules of law require and

forbid should be of a kind which men canreasonably be expected to do and to avoid.

b. Those who enact the laws and issue legal ordersshould do so in good faith, in the sense that theybelieve “a” with respect to the laws and ordersthey promulgate.

c. A legal system should recognize impossibility ofperformance as a defense, or at least amitigating circumstance.

2. The Requirement of Regularity. The legal systemshould reflect the precept that similar cases shouldbe treated similarly.a. Judges should justify the distinctions they make

between persons by reference to the relevantlegal rules and principles.

b. The requirement of consistency should hold forthe interpretation of all rules.

3. The Requirement of Publicity. The legal systemshould reflect the precept that the laws should be public.

a. The laws should be known and expresslypromulgated.

b. The meaning of the laws should be clearlydefined.

4. The Requirement of Generality. Statutes and otherlegal rules should be general in statement and shouldnot be aimed at particular individuals.5. The Requirement of Due Process. The legal systemshould provide fair and orderly procedures for thedetermination of cases.

a. A legal system ought to make provision fororderly and public trials and hearings.

b. A legal system ought to contain rules of evidencethat guarantee rational procedures of inquiry.

c. A legal system ought to provide a processreasonably designed to ascertain the truth.

d. Judges should be independent and impartial,and no person should judge her own case.

Absent so far is the notion that the rule of law requiresthat the government and government officials be subject tothe law.22 Thus, a sixth aspect of the rule of law might beadded to Rawls’s formulation as follows:

6. The Requirement of Government under Law. Actionsby government and government officials should besubject to general and public rules.

a. Government officials should not be above thelaw.

b. The legality of government action should besubject to test by independent courts of law.

Most of Scalia’s justifications for his discretion-constraining general rules approach are parallel to elementsof the rule of law. Scalia’s concern for publicity rests on thesame underlying concern as does the requirement ofpublicity; Scalia’s concern for equal treatment is closelyrelated to the requirement of regularity. Scalia’s argumentthat general rules safeguard against arbitrary decision reflectsDicey’s first principle and Rawls’s requirement of generality.The notion that general rules safeguard judicial independenceis closely related to the requirement of government underlaw.

This completes my exposition Justice Scalia’s claim thatthe rule of law requires a law of rules. In the section thatfollows I will provide a critique and reconstruction of Scalia’sposition. I will expose certain weaknesses in Scalia’sexposition of his thesis, and then attempt to reformulate hisargument in a way that avoids those weaknesses.


— 110 —

LAW

II. A Critique of the Law of RulesWould a law of rules really guarantee the rule of law? Theanswer to this question is surely no. My critique of the law ofrules will proceed in two stages. Initially, I shall offer a thoughtexperiment that illustrates the possibility that a law of rulesmight not constrain. Then, I shall diagnose the thoughtexperiment and identify two conceptual problems withScalia’s thesis.

A. A Thought Experiment: A Tale of Two Legal CulturesLet us imagine two societies that might exist in the distantfuture or far past. Call them Sea-Eless and Lahahndekon.Each has a very distinctive culture.

The legislative body of Sea-Eless believes in general rulesand has put that belief into action in the form of a highlystructured code. Likewise, the judges of Sea-Eless are carefulto base their reasoning on general rules. These judges justifytheir decisions by reference to statutory language. When astatute is general, abstract, and ambiguous, they alwaysannounce a general rule to resolve the ambiguity. Ironically,however, it is very difficult to predict the outcome of a case inthe courts of Sea-Eless unless one knows the particular judgeassigned to the case and the political values of that judge.The judges of Sea-Eless, it turns out, believe that legal rulesare inherently indeterminate. “For every rule, there is acounter-rule,” they say. “Law is politics,” is their favoriteslogan. Moreover, the society-at-large in Sea-Eless is wrackedby deep and persistent political divisions of class, race, andculture. Because judges are appointed for life terms anddifferent parties have controlled the process in differenthistorical periods, it is not surprising that different judges takeradically opposing views as to the meaning and applicationof the general rules. Everyone agrees what the rules are;everyone disagrees about what they mean.

The law of Lahahndekon works very differently. There isvery little statute law, and common law norms are based ona system of precedent. Each judge is committed to decidingonly the case that is before her, and holdings are stated inprecise but narrow terms. The judges of Lahahndekon followprecedent rigorously, although their jurisprudence admits thepossibility that a prior decision might come to be recognizedas a mistake and be overruled. When deciding a hard case,involving an unresolved question of law, these judges agreethat one ought always to ask what holding will produce thegreatest social welfare from an ex-ante or forward-lookingperspective. This requires an evaluation of all the socialinterests involved. Because these judges come from similarbackgrounds, have similar values, and receive similar training,they almost always agree about how particular cases oughtto be decided. In Lahahndekon, there a very few generalrules, but the outcome of particular cases is uniform, certain,and predictable.

In Sea-Eless, there is a law of general rules but no rule oflaw. In Lahahndekon, there is a law of interest balancing, butthe rule of law prevails. If these scenarios are plausible, thenit is clear that that there is no necessary connection betweenthe form of general rules and the substance of rule of lawvalues. The question whether the scenarios are plausible is acomplex one. In order for me to demonstrate their plausibility,I would need to flesh out the bare bones of the thoughtexperiment and provide particular examples of legal rules andjudicial decisions. But even without such a demonstration,

the scenarios are lent prima facie plausibility by the real-worldexamples that I have already adduced, e.g. the law of personaljurisdiction as described above. Rather than pursuing thisstrategy further, I shall now turn to the conceptual problemswith the thesis that the rule of law requires a law of generalrules.

B. Two Conceptual Problems with the Law of RulesThere are two conceptual problems with the thesis that therule of law requires a law of rules. The first problem, I shallcall “the problem of social practice.” This problem is basedon the idea that the same legal forms can take on quitedifferent meanings when embedded in different socialcontexts. The second problem, I shall call “the problem ofcharacter.” This problem is based on the idea that legal actors,e.g. judges or lawyers, have forms of discretion that cannotbe directly controlled by legal forms.

1. The Problem of Social PracticeScalia argues that judges are more constrained by generalrules than they are by balancing tests. That may be true insome contexts but false in others. One reason this is so isthat the constraining power of rules is not solely a matter ofform, the syntax and semantics of the rules, but is also afunction of context, the social practices that govern the waythat legal rules are used. If a given legal culture produceswidespread intersubjective agreement on judgments aboutapplication of a legal rule, then the rule is likely to be seen asconstraining. But if there is a lack of intersubjectiveagreement, then the rule may not constrain. Similarly,balancing tests will be seen as constraining if there isintersubjective agreement about how proper balancing is tobe done. If judges agree on what factors are relevant andhow they are to be weighted, the outcome of a balancingtest can be highly predictable and certain.

The problem that intersubjective agreement poses forScalia’s view of the rule of law is not that his observations arewrong, but rather that his view is partial and incomplete.There are clearly areas of law where balancing tests provideless constraint than general rules. This is especially the casewhere the general rules rely on relatively sharp edgedcomponents upon which there is a high degree ofintersubjective agreement whereas the legal rules allowdifferences in judgment about the relative importance ofvarious factors to come into play.

2. The Problem of CharacterThe second problem is the problem of character.Intersubjective agreement about the meaning of legal normsis a product of many factors. Social context is one of these,and judicial character is another. Judges can relate to theirrole in a variety of ways. It is possible for a judge to regardthe law as an instrument for the achievement of a personalagenda. In each case, the judge could ask herself, “Can I getaway with achieving the result that I prefer, given the relevantlegal norms and the attitudes and institutional capabilities ofthe appellate court to which I am answerable?” A judge likethis may not see general rules as more constraining thanbalancing tests. Faced with a general rule, such a judge mayhunt for plausible ambiguities or turn her attention to othermeans for influencing the result. If the general rule is clearand unambiguous, then the outcome might be manipulatedthrough bad-faith procedural rulings or fact-finding that isinaccurate but will survive appellate review because the


— 111 —

LAW

system defers to trial court judges on factual matters. AsJustice Scalia observed, it is not difficult to devise similarstrategies for manipulating a balancing test.

By way of contrast, a judge might possess the virtue ofjudicial integrity and take the view that her job is to act withfidelity to the relevant legal norms, deciding each case as shebelieves the law requires and setting aside, as far she can,her own personal views about who ought to prevail.23 Such ajudge might find general rules and balancing tests to beequally constraining. Rather than seeking to manufactureambiguities in general rules, she could seek the interpretationof the rule that best coheres with its language, purpose, andhistory of application. Rather than manipulating a balancingtest, she could attempt to determine the method of weighingand balancing various factors that best coheres with purposeof the test and precedents applying it.

As with the problem of intersubjective agreement, theproblem of character is a supplement to Justice Scalia’saccount. Justice Scalia is concerned about constraint, andthe problem of character explores one of the reasons whyconstraint is important. Without constraint, judges who lackthe virtue of judicial integrity can manipulate the law to theirown ends. But the form of legal norms is hardly a completeanswer to the problem of judges who lack integrity. Suchjudges will have all too many opportunities to manipulateoutcomes, whatever the form of the legal rules.

III. Reconstructing the Law of RulesAlthough I have presented my arguments as a critique ofJustice Scalia’s views about the rule of law, I intend thiscritique to be a sympathetic one. Although I will not presentthe arguments here, I believe that a very strong case can bemade for the proposition that the rule-of-law values ofpublicity, certainty, uniformity, and predictability of the laware very important values. Moreover, I agree with JusticeScalia that in many contexts, general rules will better serverule-of-law values than will all-of-the-circumstances balancingtests.

But to Scalia’s points, I would add the following. For legalnorms to be constraining and hence predictable and uniformin application, great care must be taken to shape the legalnorm so as to avoid the zones of intersubjective disagreementabout meaning. This is not a simple matter of looking forgenerality; frequently more general rules will also be moreambiguous and hence less constraining and predictable inapplication. Rather, it is a matter of selecting the tool that isappropriate to the circumstances. As Scalia astutely observes,it will sometimes be the case that the more constraining andpredictable rule will seem less desirable than a rule that wouldbe normatively more attractive—if it were applied by judgeswho agreed about its meaning and who were faithful to itsspirit. Sometimes, a very limited holding will do moreconstraining work than a broad holding. These sorts ofjudgments can only be made in particular contexts. Suchjudgments will only be made well by judges who areexperienced in the practical discipline of law and hence areequipped to predict the effect of their rulings on a complexsystem that is not well captured by abstract generalizations.

Moreover, legal systems must take judicial characterseriously. Ideally, judicial integrity in the sense of fidelity tolaw should rank very high in the hierarchy of criteria for judicialselection. Indeed, I would argue that judicial integrity should

be considered as a more important virtue than the candidate’sagreement with the appointing and confirming officials’ viewsabout controversial questions that the candidate maysomeday decide. Realistically, rules of procedure andpractices of appellate review should take into account thereality that in the legal culture of the United States, manyjudges do take a strongly instrumental attitude toward thelaw. One of the great advantages of crafting the right kind ofgeneral rules is that such rules may help to expose the viceof instrumentalism, but this can only happen if trial judges arenot able to hide their instrumentalism behind biased fact-finding or manipulation of discretionary procedural powers.

The rule of law does not require a law of rules; nor doesa law of rules guarantee the rule of law. The problem ofjudicial constraint is not that simple, and the strategies thatare adequate to advance the predictability and uniformity ofthe law defy easy summary. The rule of law requires soundpractical judgment by judges of integrity. Nothing less willdo, although much more may be necessary.

Endnotes* Professor of Law and William M. Rains Fellow, Loyola Law School,Loyola Marymount University. © 2002 by the author.1. Antonin Scalia, “The Rule of Law as a Law of Rules,” University ofChicago Law Review 56 (1989), p. 1185.2. Scalia, supra note 1, p. 1176.3. Lawrence B. Solum, “Stare Decisis, Law of the Case, and JudicialEstoppel,” Moore’s Federal Practice Vol. 18 (3d edition, 1997), §§134.01 to 134.06, 134-20 to 124-24.4. Scalia, supra note 1, p. 1177.5. For a good discussion, see Ronald Cass, The Rule of Law inAmerica (Harvard University Press: 2001), pp. 4-6; see also LouisKaplow, “Rules versus Standards: An Economic Analysis,” Duke LawJournal 42 (1992), p. 557; Duncan Kennedy, “Form and Substance inPrivate Law Adjudication,” 89 Harvard Law Review (1976), pp. 1687-1713.6. United States Constitution, Amend. 11.7. The depth of the controversy is illustrated by the Supreme Court’sdecision in Seminole Tribe of Florida v. Florida, 517 U.S. 44 (1996).Seminole Tribe was decided by a 5-4 vote, and features a long dissentby Justice Souter that recounts much of the history of disagreementover the meaning of the text. For a sample of the vast secondaryliterature, see Akhil Reed Amar, “Of Sovereignty and Federalism,”Yale Law Journal 96 (1987), p. 1425. Of course, the full story of theeleventh amendment is far more complicated than I have indicatedin text.8. See International Shoe Co. v. Washington, 326 U.S. 310, 316, (1945).9. See generally Burger King Corp. v. Rudzewicz, 471 U.S. 462 (1985);World-Wide Volkswagen Corp. v. Woodson, 444 U.S. 286 (1980).10. Asahi Metal Industry Co., Ltd. v. Superior Court, 480 U.S. 102(1987). Eight of the nine justices signed on to the portion of JusticeO’Connor’s opinion that applied the balancing test. Justice Scaliadid not join this portion of the opinion, but did not write separatelyto indicate the reason for his failure to join. Four Justice signed onto the part of her opinion framed in terms of the categorical andgeneral “purposeful availment” rule; five Justices disagreed with thisportion of the opinion. There were two concurring opinions on thatissue, one by Justice Brennan and one by Justice Stevens, eachconcurring opinion was joined by an overlapping but different set ofJustices.11. See H.L.A. Hart, “Positivism and the Separation of Law andMorals,” Harvard Law Review 71, (1958) p. 607.12. Scalia, supra note 1, p. X13. Scalia, supra note 1, p. 1178.14. Scalia, supra note 1, p. 1178. Scalia’s primary supportingargument is an analogy with rules that parents devise for theirchildren’s watching of television: “Parents know that children will


— 112 —

LAW

accept quite readily all sorts of arbitrary substantive dispositions—no television in the afternoon, or no television in the evening, oreven no television at all. But try to let one brother or sister watchtelevision when the others do not, and you will feel the fury of thefundamental sense of justice unleashed.” Id.15. Scalia, supra note 1, p. 1178-79.16. Scalia, supra note 1, p. 1179.17. Scalia, supra note 1, p. 1179-80.18. Scalia, supra note 1, p. 1180.19. Dicey, Introduction to the Study of the Law of the Constitution10th ed. (London: Macmillan, 1960), 202-03.20. John Rawls, A Theory of Justice (Cambridge: Harvard UniversityPress, 1971), p. 235. Rawls’s account was influenced by Lon Fuller’swork. See Lon Fuller, The Morality of Law, rev. ed. (New Haven: YaleUniversity Press, 1969), pp. 33-94. Joseph Raz has offered anotherinfluential formulation. Raz maintains that the ideal of the rule oflaw is captured in two injunctions: (1) people should be ruled bylaw and obey it, and (2) the law should be such that people will beable to be guided by it. Joseph Raz, “The Rule of Law and its Virtue,”Law Quarterly Review 93 (1977), p. 195.21. One significant change has been made in Rawls’s presentationof the rule of law precepts and their implications. I have substituted“ought” or “should” for “must” throughout. In Rawls’s originalpresentation, the sense that the requirements of the rule of law arenecessary for a system to be characterized as legal comes throughvery strongly. For the purposes of this paper, and indeed, I think forRawls’s own purposes, the question whether there can be lawwithout the rule of law does not need to be answered.22. For a discussion of this idea, see Ronald Dworkin, “Political Judgesand the Rule of Law,” A Matter of Principle (Cambridge: HarvardUniversity Press, 1985), p 9.23. See Lawrence B. Solum, “Equity and the Rule of Law,” in IanShapiro (ed.), Nomos XXXVI: The Rule of Law (New York: New YorkUniversity Press, 1994), p. 120; Lawrence B. Solum, “The Virtues andVices of a Judge: An Aristotelian Guide to Judicial Selection,”Southern California Law Review 61 (1988), p. 1730.

The Rule of Law and Discretion in “BillyBudd”

Joan McGregor*

But the exceptional in the matter moves the heartswithin. Even so too is mine moved. But let not warmhearts betray heads that should be cool. Ashore in acriminal case will an upright judge allow himself offthe bench to be waylaid by some tender kinswomanof the accused seeking to touch him with her tearfulplea? Well the heart here denotes the feminine inman is as that piteous woman, and hard though itbe, she must be ruled out.

Captain Vere speaking to the drumheadcourt, Billy Budd, Herman Melville

Captain Vere in the novella Billy Budd cautions the membersof the drumhead court, who are convened to decide the fateof Billy Budd, not to be swayed by the “feminine in man” andbe tempted to consider the particularities of Billy Budd’scircumstances and his character in meting out the sentence.To do so, according to Vere, is not to follow the “Rule of Law”but to be influenced by emotions which have no place inadjudicating cases. Was Vere correct in instructing his officersthat there is no role for emotions, empathy, and attention toparticulars in adjudication? For Vere, the Rule of Law involvesonly the use of reason and the strict application of the rules

to the facts of the case. The rule in this case is the Mutiny Actwhich states: “In wartime at sea a man-of-war’s man strikeshis superior in grade… [a]part from its effect the blow itselfis…[a] capital crime.” And when asked “Can we not convictand mitigate the penalty?” Vere responds: “Lieutenant, werethat clearly lawful for us under the circumstances considerthe consequences of such clemency.” Vere doesn’t think thatit is lawful and believes that if it were and they did mitigatethe sentence the consequences would be disastrous; thesailors would see it as weakness on the part of the officers,and a swift mutiny would follow. Vere repudiates the use ofwhat he calls “natural justice,” which is what Billy morallydeserves, what justice demands given the totality ofcircumstance. Rather Vere says that their allegiance is not tojustice but to the law. We are not responsible, he admonishes,for the form of the law, but we are merely the agents whoadminister it. “Our vowed responsibility is in this: that howeverpitilessly that law may operate, we nevertheless adhere to itand administer it.” Are Vere’s interpretations correct? Are theRule of Law and the use of the kind of discretion that involvesfocusing on the “narrative” story of the accused in decidingthe appropriate sentence antithetical to one another? What Iwant to consider is whether the Rule of Law is consistentwith judges exercising their discretion to apportionparticularized justice, what is called equity.

To begin to answer that question, we need to knowwhether Vere is correct in his analysis of the “Rule of Law.”The slogan the “The Rule of Law” is rallied around by politicalgroups and its absence is used as the basis of criticism foropposition to dictatorial regimes. Why is the Rule of Lawessential to legitimate political authority? What is the Rule ofLaw? First, the Rule of Law is associated with the phrase‘governed by rules, not by men.’ Why exactly is living underrules better than being governed by men, particularly, whenit looks like, as in the Billy Budd case, following the rule resultsin applying a draconian sentence? What does the Rule of Lawsay about legal interpretation and judicial discretion? Whenjudges exercise discretion are we still living under the Rule ofLaw, or are we then living under the rule of men?

The Rule of Law, according to A.V. Dicey, a famous Englishlegal scholar, entails two fundamental ideas. First, it entails“the absolute supremacy or predominance of regular law asopposed to the influence of arbitrary power, and excludesthe existence of arbitrariness of prerogative or even widediscretionary authority on the part of the government.”Second, it entails “the equal subjection of all classes to theordinary law of the land administered by the ordinary lawcourts” and “excludes the idea of any exemption of officialsor others from the duty of obedience to the law which governsother citizens or from the jurisdiction of the ordinarytribunals.”1

Dicey emphasizes that the Rule of Law is opposed to theuse of arbitrary power. Opposing the use of arbitrary power,law should be impartial and thereby thwart the advancementof individuals’ self-interests. The surest method to bluntarbitrary self-interest is the requirement of general rules.Governing with general rules promotes neutrality,predictability, and stability. What follows from the characterof general rules is that an activity can only be unlawful if it isspecified by a general rule (principle of legality). Anothercorollary of this is that individuals are only legitimatelypunished by government if they have violated a legal rule.


— 113 —

LAW

Though rules play a significant role in the Rule of Law,however, we must be careful not to be too sanguine aboutwhat we get with rules. For instance, unfortunately, rulesthemselves do not guarantee substantive justice. There isnothing in the notion of a rule that guarantees that the contentof the rules will be just. The rules themselves may be unjust,e.g., “All blacks must go to segregated schools.” The fairnessthat rules may guarantee is merely the comparative fairnessof each person being treated the same, possibly equally badly.

Upon further inspection, the nature of rules themselvesdoes not guarantee much at all, for example, nothing aboutthe nature or concept of rules determines that the rules willbe clear. Rules may be difficult to understand or impossibleto obey. Notice too that predictability and stability are afunction of rules not constantly changing. The nature of rules,however, does not preclude them from changing nor does itput stipulations on the amount of change. Rules give uscomparative fairness, when like cases are treated alike.Consequently, rules take some of the arbitrariness ofgovernmental power away although not necessarily much.

It looks as if just having rules is not going to get us muchwithout some formal procedural constraints on the rulesthemselves. Some of these constraints have been exploredby Lon Fuller in his allegorical tale about Rex. 2 The formalconstraints are that the rules are public, no secret laws areallowed. Secondly, the rules must be reasonably clear in theirmeaning so that people can understand what is proscribed.Third, the rules themselves cannot have contradictoryproscriptions. Fourth, the rules should be appliedprospectively, not the opposite, retrospectively (no ex postfacto, this is the principle of legality). Fifth, the rules shouldbe valid for some period of time, in other words, the rulescannot constantly change. Sixth, the rules are possible tocomply with, they don’t demand that people do things thatare not possible. Seventh, the applications of the rules areconsistent with the rules that are promulgated. There mustbe a close relationship between the rules that are stated aheadof time and the enforcement and adjudication of the rules.

The Rule of Law requires governing with general rulesand that the rules to a greater or lesser degree satisfy theseformal requirements. A system of rules that was secret,changed all the time, was inconsistent, was impossible tocomply with, was applied retroactively, and so on, would notbe a system governed under the Rule of Law. Governmentswill, however, vary in the degree to which they satisfy theideal, consequently, the Rule of Law itself may be a matter ofdegree. Needless to say most systems will not fully satisfy theideal, they may satisfy many aspects of it and fail in otherareas. One question will be where is the line below which asystem would no longer be said to be under the Rule of Law.

The Rule of Law requires that legitimate government begovernment by general rules and that the rules conform tothe formal constraints mentioned above. Is this all that isrequired? No. The second idea of Dicey’s was that those rulesapply to everyone equally. This means that the governors arenot above the rules of law. This is important since, asphilosophers such as Immanuel Kant and Montesquieucautioned, humans have a strong natural inclination to exemptthemselves from rules that they will cheerfully apply toeveryone else. If the law makers knew that they were alsothe law enforcers, they would be likely to draft unduly harshor extreme laws and exempt themselves from the laws. If, on

the other hand, they knew that they were also subject to thelaws they made, they would be more tempered in their writingof legislation. The way to subvert human nature and avoidpolitical tyranny is separation of powers. With separation ofpowers, e.g., legislative, executive, and adjudicative, no onesingle branch has all the power.

The Rule of Law excludes arbitrary power by governingwith general rules that satisfy, to some degree, the statedformal requirements and apply to ever yone equally.Nevertheless, given the nature of rules and the separation ofpowers, whenever there is a question about the law, it isjudges who give the official answer about law. In reachingtheir decisions, judges rely upon sources of law such aslegislation and decisions in previous cases. But sometimes itseems that interpreting rules calls for something more than aformal process to determine the right answer. Is interpretinglaw a purely formal endeavor? Does judicial interpretationrequire that judges do more than “mechanically” apply thelaw to the case in front of them? Montesquieu thought thecorrect theory of judicial decision-making was that judges“mechanically” apply the law and if judges went beyond thatthey would be exercising the arbitrary power that the Rule ofLaw was supposed to guard against. Montesquieu’s idea givessupport to the view that Captain Vere takes in Billy Budd’scase. Do judges, when interpreting rules, need somethingmore than a formal process to determine the right answer?Does judicial interpretation require that judges do more than“mechanically” apply the law to the case in front of them?

Judicial decision-making frequently requires somethingbeyond mechanically applying the rules to the case at hand.For a variety of reasons legal cases often require discretion.Some of the kinds of cases that require something beyondmechanically applying the rules are, for example, “cases offirst impression.” Those are cases where there are no bindingprecedent and no cases that are obviously similar. One typeof case of first impression involves cases with novel facts. Forexample, the custody case regarding frozen embryos was anovel set of facts. 3 Other types of cases requiring discretionare ones where the terms of a statute are vague. 4 An oftenused example of a vague term in a statute is the use of ‘vehicle’in the law “No vehicles in the park.” Interpreting ‘vehicle’requires more than consulting a dictionary to understand itsmeaning in this statute. A third type of case that requiresdiscretion on the part of judges is when it is not merely thatthe terms are vague, that their meaning may vary from contextto context, but that the terms refer to concepts—called“contested concepts” 5—that are in dispute. Contestedconcepts are ones like “cruel and unusual,” “due process,”even “reasonable.” There are other types of cases that requirediscretion as well, specifically when sentencing defendants.

Admitting that sometimes judges must use discretion,how is discretion distinguished from the arbitrary power thatthe Rule of Law was meant to protect against? Why is thereso much cynicism about judges exercising discretion? Beforeanswering this question, it is worth mentioning that judicialdiscretion has not always been treated with the suspicion anddisdain with which Montesquieu and others have held it. Theancient view of Plato and Aristotle saw discretion as essentialto the activity of judging. Plato said:

In a country where the regulation of the courts is assatisfactory as can be achieved and the judges-to-be have received a good education and been


— 114 —

LAW

examined by all kinds of tests, it is right and properto grant them complete discretion on all points todo with the punishments or fines that convictedcriminals should suffer.

Plato recognized that an offender may turn out to beparticularly good or bad and the judge adjudicating the casemust give the sanction accordingly.

More recently, wide discretion has been viewed withsuspicion. Why is there a lack of trust in judges’ discretion?Richard Posner, a judge himself, has the following to say aboutjudges:

Politics, personal friendships, ideology, and pureserendipity play too large a role in the appointmentof federal judges to warrant treating the judiciary asa collection of sainted genius-heroes miraculouslyimmune to the tug of self-interest… A federal judgecan be lazy, lack judicial temperament, mistreat hisstaff, berate without reason the lawyers who appearbefore him, be reprimanded for ethical lapses, vergeon or even slide into senility, be continually reversedfor elementary legal mistakes, hold underadvisement for years cases that could be decidedperfectly well in days or weeks, leak confidentialinformation to the press, pursue a nakedly politicalagenda, and misbehave in other ways that might geteven a tenured civil servant or university professorfired; he will retire his office.

This is hardly Plato’s view of judges. Luckily, accordingto Posner, the framers didn’t have Plato’s view either: “theframers of the Constitution tried to design a government thatcould be operated by moral and intellectual mediocrities.”According to Posner, judges crave discretion because “If theyadhered rigidly [to stare decisis], realistically they would notbe voting as often, because voting implies discretion; if thereis no felt choice, there is not pleasure in choosing.” If theydid not exercise discretion in sentencing “the process wouldbe turned into a process of mechanical, but laborious,computation.” Posner’s pessimistic attitude has won out.Consequently, in recent years we have moved in the directionof severely limiting judicial discretion, particularly insentencing.

What more can be said about the present day skepticismabout discretion? There are a number of reasons for themistrust, starting with similar motivation that drives theconditions for the Rule of Law. Remember that Kant andMontesquieu advocated the separation of powers becausethey believed that human nature was corruptible. Sincehuman nature was fallible, we ought to structure ourgovernment accordingly, taking the temptation to abusepower out of individuals’ hands by separating and limitingtheir domains of power. The idea of the fallibility of humanscan be traced back to its theological roots in Christianity—our “Fallen Nature.” The corruptibility of human nature leavesus skeptical that any human judges can acquire suitable skills,training, and knowledge enough to be allowed to operate withunconstrained discretion, to keep from exempting themselvesand their friends, and injecting their own values into theirdecision-making.

Critics of discretion often conceive it as “anything goes”in judicial decision making. Legal Realists claim, for example,that judges decide whatever they want to decide when the

law is unclear (and according to the Realist it is often or alwaysunclear). A decision maker may be said to have discretion if“he is simply not bound by standards set by the authority inquestion.” Legal indeterminacy leads to judges deciding onthe basis of personal or political preferences or interests. This“anything goes” theory presents a great challenge to thelegitimacy of adjudication under law and undermines thenormativity of law. Fortunately, this theory is not intellectuallydefensible.

Another reason for skepticism about discretion came outin the push for federal and state sentencing guidelines. Whatled the way for determinate sentencing and the guidelineswas a growing public outcry about judges being “soft oncriminals.” It was widely thought that judges were too lenienton offenders and the guidelines were to tie the hands of judgesand make them hand out the harsher sentences. Anotherreason for the guidelines was the issue of fairness.Unconstrained discretion permits judges to hand out disparatesentences for the same offense. These discrepancies appearto violate the formal principle of justice that demands thatwe treat like cases alike. Comparative injustice can be aproblem, but there is another problem as well with judgeshaving wide discretionary authority over sentencing.Individuals will not know in advance what the sanction is forviolation of a particular offense since it is based upon judges’ad hoc determinations after the fact. This violates theconstraints that we said were important for the Rule of Law,viz., that there are general rules announced ahead of timeand that their application matches the stated rule, appliedprospectively, and so on. Consequently, granting judges widediscretion in sentencing would be to undermine the essentialingredients of the Rule of Law.

Are these good reasons for the sweeping renunciationof judicial discretion? Not entirely. Conceiving of discretionas “anything goes” judgment, as judges deciding on the basisof personal or political preferences or interests or on the feltidentification with the litigants, may provide a description orexplanation of what a given judge does, but it is not ajustification. For judicial discretion to be justified it must beexercised in a manner consistent with the judicial duty touphold the law. The judicial duty to uphold the law excludessome kinds of reasons from judicial deliberations. Judgeshave a duty, for example, to act impartially, “without respectto persons.” Therefore, judicial duty to uphold the lawexcludes reasons for a decision that would constitute notrespecting persons. Discretion is not a matter of unbridledchoice; it is a matter of judgment that may earn respect frommembers of the legal community and others.

Focusing on the narrow area of judicial discretion insentencing, there are certainly reasons for thinking that judgescan abuse that power. Judges who routinely hand out stiffersentences to minority defendants would be abusing thatdiscretionary power by permitting racial hatred to dictate theirsentencing decisions. Those sentences would violatecomparative justice since race alone has no relevance indistinguishing defendants. Judicial discretion properlyunderstand excludes some reasons from judicialdeliberation–in this case the race of the offender.

Not all discretion in sentencing is abusive. Justicedemands that like cases be treated alike, but what goes alongwith that is that relevant differences between individuals mustbe acknowledged. Those differences in the individuals may


— 115 —

LAW

well lead to different sentences being justified. Justicerequires that we individuate. That means that we treatindividuals according to their relevant individualcharacteristics that affect their desert.

Usually when we talk about justice in punishment, wemean retributive justice. Retribution in punishment meansthat individuals receive the punishment that they deserve andthat punishment is based not only on the offense committedbut the mental state of the offender. There are, however, otherrelevant factors that dictate the just punishment. Individuatingcharacteristics that are properly considered by the judge insentencing are, for example, the age of the offender, whetherit was a first offense, whether the offender is a drug addict,whether the offender came from an abusive family, amongother factors. Using these individuating characteristics insentencing is referred to as equity.

The problem, however, is that in considering the“narrative” story of the person, we seem to diverge from thestrict retributive sentence. Is it just to weigh these factors intothe decision of the sentence? The law pronounces in generalterms what justice in retribution demands. But regularly thelaw is too general and in its generality too harsh in specificcases. Consequently, equity counsels that we consider theparticularities of the situation and temper or fine-tune ourjudgments accordingly. The retributive theory focuses ongiving the person his “just deserts.” Blindly applying the“correct” retributive punishment to the offense (e.g., murderdeserves death) misses essential features that are relevantto the desert of the offender in the particular case. Strictretributive justice presupposes that agents start with the sameabilities and resources to conform their behavior to thedictates of law. In practice, individuals have very different,complex backgrounds that affect the capacity to controlaggression and other impulses, to evaluate choices properly,and thereby to follow the law. Each individual story tells of atangled web of factors and circumstances that shapeindividuals’ capacities to comply with the rules of law. Judgesneed to employ an “empathetic” understanding of theoffender’s narrative, to try to get inside that person’s life, tograsp the obstacles that individual had to overcome.Judgments that do not consider individuals’ narrative storiesare not going to capture the offender’s actual desert. Failureto consider these unique factors means that the judge is nottreating the person fairly before the law. Indeed, interpretingretribution as devoid of a more complicated picture of desertmakes it susceptible to the same as deterrence theory,namely, that by neglecting the unique particularities of theindividual, it does not treat the defendant with respect as anindividual. Exercising discretion in sentencing by consideringthe particularities of this person, her “narrative story,” then isproperly part of a judge’s duty in sentencing. It permitsevaluation of all the relevant circumstances, the uniquenessof each individual person, and then the crafting of anappropriate legal response. Nevertheless, this is not to suggestthat retribution and equity are the only factors that the judgemust consider when sentencing. Judges have to be sensitiveto the deterrent rationale of punishment as well and balancethose considerations with the interests of retribution andequity. Saying that judges have discretion is not to say that“anything goes” in their decision-making. That kind ofunhampered decision-making is not consistent with judges’duty to uphold the law. Judicial duty to uphold the law, can,

however, coexist with discretion when the law is understoodto admit and exclude reasons for sentencing decisions.

Considering the role that equity plays in sentencingmakes apparent how discretion is not only consistent withthe Rule of Law but, since it is required by the separation ofpowers, is indispensable to the Rule of Law. It is as impossiblefor legislators to specify in advance every imaginableexception to a rule of conduct as it is for them to specify inadvance every possible life story and how those should affectsentences. When crafting legislation, legislators must relyupon the judgment or discretion of judges to interpret therules reasonably and give out the appropriate sentence.

Rethinking now our initial concern, what about the caseof Billy Budd? Was it consistent with the Rule of Law for themembers of the drumhead court to interpret the statute andconsider the particularities of Billy Budd’s circumstanceswhen sentencing? From what has been argued, the answeris yes.

Vere interprets the statute as requiring nothing more thanthe presence of the causal connection between Billy’s actionand Claggart’s death. Such an interpretation of a law isinconsistent with retributive ideas about desert. Somedetermination beyond the causal connection is required forthe punishment to be just, particularly when the punishmentis so severe. Billy could have had an epileptic seizure and hitand killed Claggart and under Vere’s interpretation that toowould have been a capital offense. The Rule of Law does notrequire that we interpret statutes “mechanically” or literallyas Vere counsels. Retributivists are interested in the mentalstate of the offender so that the actus reus, Billy strikingClaggart, done knowingly versus recklessly versus negligently,would deserve very different punishments. A reasonableinterpretation of the statute would be: knowingly (withmalicious intent) striking an officer (in the execution of hisoffice) is a capital offense. Do the facts in this case fit thatdescription? They do not. Since Claggart was lying when heaccused Billy of mutinous conspiracy, he could not be “in theexecution of his office.” Abandoned as a baby, Billy had anaive innocence that left him incapable of seeing evil in othersor forming malicious intent. After understanding the complexweb of factors that made up Billy’s life, the circumstances ofthe offense, and Billy ’s mental state, the appropriatepunishment for Billy, his “just deserts,” is not the gravestpenalty that society has to bestow upon a person. Themitigating factors are many, making the so-called strictretributive response for the crime unwarranted, making itunjust.

Notice that if the legislature could determine in advanceall the multitude of factors that contribute to desert, then itwould be possible in principle to do away with equity, withjudicial discretion in sentencing, and merely mechanicallyhand out the “correct” retributive response. The problem isthat it is not possible to come up with all the possible narrativestories that would affect the appropriate punishment, andconsequently, legislators must rely upon judges to arrive atthe equitable response.

A fundamental concern for legal systems is to have respectfor the dignity of each unique individual. One significant waythat respect is accorded to individuals is that legal systemsmust not “mechanically” hand out sentences without regardto the uniqueness of the individual narrative of a person’slife. Judicial discretion, correctly used, respects the


— 116 —

LAW

uniqueness of each individual. Sentences that come out ofthat discretion don’t look at individuals as merely behavior,but rather look at the complex inner life of the actor, thecircumstances in which he acted, to ascertain the appropriatepunishment. Does that mean that anything goes, and judgeshave unlimited discretion, no boundaries to their decision?No, that kind of unprincipled discretion would undermine theRule of Law, amounting to judges having arbitrary power.

In a perfect world, then, judges would determinepunishment on the basis of equity, which refines and tempersthe retributive response. Since we realize that we don’t livein a perfect world, the moral ideal of equity probably wouldn’tbe achieved and permitting too much discretionary spacefor judges to find the equitable response might lead to abuses.In practice, we will want to codify appropriate ranges ofpunishments, thereby limiting the space for equity, but notwholly eliminating it. Additionally, abuses are most egregiouswhen they err on the side of too much punishment;consequently, equity should play a larger role in mitigatingsentences. When the judge empathizes with the offender,understanding the complicated web of factors that lead tothis occurrence, then leniency in punishment may well beappropriate as it was in Billy Budd’s case.

Endnotes* Department of Philosophy, Arizona State University1. V. Dicey, Introduction to the Study of the Law of the Constitution(10th ed., 1961), 203-04.2. Lon Fuller, The Morality of Law (New Haven: Yale University Press,1964).3. See Joel Feinberg, “The Dilemma of Judges who must Interpret‘Immoral Laws’,” in Philosophy of Law (Belmont: WadsworthPublishing Co., 5th ed., 1995).4. Other types of cases of first impression are ones where the factsmay be familiar but bringing a suit under them is novel, as in theinitial unconscionable contract cases.5. We want to know, for example, if a motorized skateboard is a“vehicle” for the purposes of this law. It requires judgment aboutwhat is reasonable given the topic at hand, the standards that are inpractice, and other factors. The problem with mechanicaljurisprudence is that it applies terms always in a literal fashion, whichcan often lead to ridiculous outcomes (“No vehicles in the park”would mean that a WW II tank that a veterans group wanted to putup as a monument in the park would be outlawed).6. Ronald Dworkin, “Hard Cases,” in Taking Rights Seriously(Cambridge, MA: Harvard University Press, 1971)7. Plato Laws, Book Nine8. Richard Posner, Overcoming Law (Cambridge, MA: HarvardUniversity Press, 1995) pp. 110,111.9. Posner, 110.10. Posner, 125.11. Ibid.12. Judges with widely different views agree in rejecting this theoryof discretion: William J. Brennan said: “Justices are not platonicguardians appointed to wield authority according to their personalmoral predilections.” Georgetown University (Oct. 12, 1985). AndRobert Bork said “Not only is moral philosophy typically inadequateto the task but, more fundamentally, there is no legitimating reasonthat I have seen why the rest of us should be governed by the judge’smoral visions.” “The Constitution, Original Intent, and EconomicRights,” San Diego Law Review 23 (1986): 823-72, at 825.13. Aristotle thought so; for him, considering the individuating factors,particularly those that mitigate the sentence or judgment, is to bejust. “[I]t is clear what the equitable is, that it is just, and better thanjust in one sense of the term. We see from this, too, what an equitableman is. A man is equitable who chooses and performs acts of this

sort, who is no stickler for justice in a bad sense, but is satisfiedwith less than his share even thought he has the law on his side.Such a person is equitable, and this trait of character is equity,being a kind of justice and not a distinct trait of character.Nicomachean Ethics 1137b34-1138a3.14. See Martha Nussbaum’s discussion of Seneca in “Equity andMercy” in her Sex and Social Justice (Oxford: Oxford University Press,1999) p. 167. Seneca developed the notion that judges should tempertheir judgments by “both identification and sympatheticunderstanding” of the offender. Seneca thought that by daily self-examination, “[a] person will cease from retributive anger and bemore moderate if he knows that every day he has to come beforehimself as judge.” This kind of self-examination will temper theretributive fervor and the reflection will make one aware of thecomplexity of circumstances and factors that affect one’s actions.15. Notice that Oliver Wendell Holmes and, following his footsteps,Posner thought that the offender should be treated as a thing withno insides to be scrutinized from the internal viewpoint–but simplya machine, merely as behavior that judges must predict. This viewprecludes considering the particular history of the defendant, thereasons that might have made it difficult for him to conform hisbehavior to the law. Holme’s and Posner’s view is nevertheless purelydeterrence based. Richard Posner, The Problems of Jurisprudence(Cambridge, MA: Harvard University Press, 1990), p. 168.16. Consider the following case and how equity might temper theretributive response. Recently in Arizona the county prosecutorcharged and tried a young Hispanic woman for the accidentaldrowning death of her baby in the bathtub. If we merely consider inthe manner of the prosecutor that she committed negligent homicide,then the strict retributive punishment for that offense would be fiveyears in jail. If, however, the judge considered the defendant’s wholenarrative story, that this was a young unmarried woman who lefther child for a few moments to go outside only to return to discoverthis great tragedy—the death of her child—then the judge mightempathetically put herself in the defendant’s position, that is, thatthe defendant will live her whole life with this tragedy and theattendant suffering, and the judge might empathize with themomentary lapse of judgment that caused this horrific outcome (themoral luck involved), and thereby discern that a much lesserpunishment is deserved by this defendant.17. From a deterrence vantage, Vere argued that the sailors knewthe punishment for the offense, and consequently, failure toimplement the penalty would undermine the deterrent effect of thelaw. The problem with the deterrence argument is that all of thiscould have been true whether or not Billy Budd was guilty of theoffense. Nothing about the particularities of the case was requiredother than the fact that Claggart was dead and that the sailors believedthat Billy killed him. But, of course, that is the problem withdeterrence arguments—actual guilt is not required. Deterrencearguments alone do not take seriously the dignity of each individualperson by their failure to take into account relevant facts about thecase, such as whether the defendant was in fact guilty.18. The unbridled choice model does not exclude some reasonsfrom decision-making. For instance, it is not a good reason fordeciding one way that the defendant is a friend or in a member of aparticular racial group or paid you a bribe. Judicial discretion excludessome kinds of reasons from decision-making. The kinds of reasonsthat are included or excluded can and do change. Over the years,for example, our understanding of human psychology will changeour ideas about acceptable reasons for mitigating a sentence. Somereasons clearly are included and others will be open to disputewhether they are or are not relevant to determining, for instance, aperson’s just deserts.


— 117 —

LAW

Legal Theory and the Rule of Law

Noel B. Reynolds*

When future Nobel Prize winner F. A. Hayek wandered out ofeconomics into the fields of political and legal theory in themid-1940s, he was widely ignored and not much appreciated.Hayek was convinced that latent in the western traditions oflaw lay a doctrine of rule of law that he and others couldwield as a powerful new weapon in their war against anexpanding socialism in Europe and America. The first legaltheorist to pick up on the idea was Harvard’s professor ofjurisprudence, Lon L. Fuller, who incorporated Hayek’sprinciples of the rule of law into his 1963 Storrs lectures andlabeled them the “inner morality of law.” Over the next fourdecades, the rule of law gradually attracted increasingattention from positivist and natural law theorists alike, andmost recently it has provided the subject for several significantpublications each year. Perhaps because of the way Hayekused the concept, legal theorists were slow to agree that itshould play a central or important role in legal theory. This,too, has changed as the rule of law has increasingly beenrecognized as a central and essential element of the idea oflaw itself. Where it was once thought to be a morally neutralelement of that concept (Fuller and Oakeshott), it has morerecently been saddled with explicit moral content (Finnis,Dworkin, and most recently Barnett and Allan).

As with most branches of philosophy, legal theory beginswith the real world and the problems of understanding thatarise as humans try to explain their experiences in that world.Law—like language, science, morals, and politics—has beena feature of human societies from ancient times to the present.Law seems like a rather simple feature of those earliersocieties compared to the complex legal systems of our ownday that can hardly be constrained within national boundaries.Legal theorists have yet to reach the levels of agreement thatcharacterize these other fields of philosophical and scientificendeavor as to just what it is that requires explanation. Withthe revival of natural law thinking in the last half century, theconcept of law seems at least as controversial today as it washalf a century ago.

The principal source of continuing disagreement derivesfrom the mid-century positivist insight that law and moralsare separate things. On the one hand, the notion of legalobligation must be recognized and addressed by anysuccessful explanation of the full range of legal phenomena.On the other hand, positivists saw no reason to explain thatwidespread form of obligation as being derived from moraltruths that bind all human beings. Somehow, it was a featureof human interaction that could take different forms acrosstime and space and was not tied to any particular moraltheory. In fact, its very function appeared to be the facilitationof harmonious actions by people who might personally beguided by quite different moral theories—with no requirementthat their moral differences be resolved as a pre-condition ofsuch actions. And unlike moral truth, legal obligations canbe changed by human actions of all kinds. They are socialfacts. Legal obligation did not seem to fit into standard moraltheory.

In this essay I will recommend a conventionalist versionof the positivist concept of law that recognizes the conceptof rule of law as an important and integral feature of the idea

of law itself. I will further explain how this approach can givefull recognition to both the factual and the normative characterof law, but without invoking moral or political theories thatwould entail truth claims beyond those generated by humanaction. To this end, I will begin with an account of law, as weexperience it in human societies—borrowing insights fromboth sociologists and economists.

Law is a human invention that is widespread, but notuniversal. It is always attached to a particular people, forwhom it provides guidance in the daily pursuit of theirinterests. Legal jurisdictions may overlap, and humans makechoices that bring them under different systems of law atdifferent points in time. Theorists have generally come toaccept Hayek’s insight that law is strongest in societies whereindividuals are free to pursue their own life plans with minimalinterference from the state, and that where the officers of thestate have wider discretion to impose their own wills on thecitizenry, without constraint of law, there is correspondinglyless meaningful freedom for individuals. The claims of Hayekand Fuller—that a principal purpose of law is to providestandard and protected ways for people to pursue their owninterests—are now widely accepted. As human societiesgrow in size and complexity, new forms of law are developedapace to accommodate those complexities.

This kind of thinking undermines the older assumptionthat law is a creation and instrument of governments andrulers. While that may be true of particular rules of law, theagreement of a people to bind themselves to the same legalsystem seems to be both logically and chronologically priorto the governments they create to service their systems oflaw. Politics itself turns out to be a product of that agreement.For where there is no law governing human interaction, thereis only war. Politics arise when war is exchanged for law. Ina regime of law, the people can seek, through political activity,the creation and implementation of particular rules of lawthat will protect or advance their individual interests anddreams.

The approach I will take in this essay will be to sketch anaccount of law as convention that will provide empiricalgrounding for the principles of the rule of law. Rather thanargue philosophically for the particular concepts of this theory,I will argue for the preferability of this general theoreticalapproach over the alternatives in terms of its ability to accountfor the major elements of our experience of law in ways thatcorrespond to our actual experience—and particularly for itssuperior account of the notion of rule of law.

I will start with a revised version of the simple notion,endorsed by positivists from Hobbes to Hart to Coleman, thatlaw is grounded in a set of conventions. The first of theseconventions, on this revised account, is an agreementbetween all parties to yield their individual right to veto anyrule that would impose obligations upon them—and to createby this agreement a common authority to promulgate andadminister rules that all are obligated to obey and to enforcethose rules as necessary. Every member of a legal system ispresumed to have accepted this convention. The secondconvention, which is more like Hart’s rule of recognition andwhich can be negotiated independently of the first—but notantecedently—is the form and process of government thatwill bear this authority. This one does not require unanimityabout the specific outcome, but only acceptance of theprocess that produces it. It is the function of a constitution to


— 118 —

LAW

spell out these institutional structures and their relationships.The basic logic of such a fundamental convention assumesthat all parties find the creation of a common authority to bein their interest, and that the agreement has certain conditionsattached to it which guarantee to the extent possible that theauthority will not be turned against them inappropriately. Itis these conditions which are embodied in the principles ofrule of law as spelled out originally by Hayek and more fullybelow. It is by adherence to the rule of law that the citizensand their public officials preserve the original convention andthe obligations to obey the law that flow from it.

The ProblemThe difficulty of accounting for both the factual and normativeaspects of law has long defined the central issue in legaltheorizing. In the 1980s, writers from both sides of this disputeproposed that a mutually acceptable solution might lie in thecharacterization of law as a particular kind of conventionwhich was both a social fact and a framework of reasons foraction. The thesis at issue has been variously labeled thecoordination thesis or the thesis that law is convention. Thebasic ideas are drawn from that branch of formal game theorywhich deals with coordination problems, those problems inwhich all parties stand to gain through collaboration. Certainsocial and linguistic theorists have seen in these non-zero-sum games a helpful explanation for the emergence ofconventions or rules. Given that in this account such rulesarise out of simple factual conditions, if it can also be shownthat these rules provide reasons for action, we may in facthave a promising solution to the central quarrel of legalphilosophy.

Early critics of the coordination thesis, as well asadvocates, focused on coordination analysis as a means ofjustifying particular rules of law, rather than the fundamentalconvention that establishes a community of law. In my view,this was a serious mistake on both sides. Advocates werehard pressed to demonstrate that every rule was a truecoordination solution, and critics were handed too many easytargets. The coordination problem, as defined centuries agoby Thomas Hobbes, is the human condition in the state ofnature without law. The coordination solution to that problemis the creation of an authority. The subsequent rules madeby that authority may or may not meet the formalrequirements of coordination problems and solutions. Butthat doesn’t matter, as their justification need only derive fromtheir origins in valid exercises of the authority. The right placeto focus our interest in coordination theory as it might applyto legal theory is where the classical and contemporarytheorists of political contract have focused. If a coordinationthesis has a chance to bridge the gap between positivist andnatural law theory, it will be by conceiving of the fundamentalarrangement that constitutes the legal society as a conventionarising out of a coordination problem. If Hart’s rule ofrecognition can be seen as being derived from thisconvention, perhaps it can be shown also to have the kind ofnormative force that its critics have denied it.

A. Conventions and CertaintyBefore exploring the particular implications of aconventionalist theory of law, it will help to review whatsociological theorists from Parsons to Luhmann have learnedabout the ideas of convention and rule in the most generalcontext of human action. Mutual adherence to conventional

rules is the technique we use for reducing most of theuncertainty of our social world. By convention we stipulatesets of legitimate expectations that we each have of oneanother’s conduct. This simplifies life greatly. When someonedisappoints these expectations we can blame them withoutneeding to revise our own cognitive models of the world.Equipped with such sets of conventional rules for conduct,we can interact meaningfully with others through language,markets, the law, family mores, morality, systems of religion,or games. Furthermore, such interactions do not necessarilypresuppose any prior experience with or knowledge of theothers with whom we are interacting. Eliminating most ofthe uncertainty in such relationships requires only that theparties accept the same conventions. By adopting stipulatedrules of action as conventions, we cut short the infinite spiralof expectations that must be calculated for human interactionand thereby render human actions and expectationspredictable. As certainty rises, so does the value of theindividual freedom of society members to work toward animprovement of their lives or toward other goals importantto them.

our man-made theories of the natural world provide uswith a structure of laws that can guide our interactions withnature, so also, man-made conventions provide us with astructure of rules for guiding social interaction, or forcoordinating our behavior with the legitimate expectationsof others. This means that in our social lives we have twopossibilities of reaction to disappointment. We can makecognitive adjustments, or we can hold firm in our structures,requiring the world to adjust to our (normative) expectations.This account of the distinction between cognitive andnormative structures is offered by the sociologist Luhmannto clarify elementary law-making mechanisms. Normativeexpectations give us grounds for holding disappointmentsagainst actors. Thus “norms are counterfactually stabilisedbehavioural expectations.”

B. The Source of Legal ObligationThis counterfactual element of norms implies theirunconditional validity and is the major source of the “ought”in the law. Legal obligation expresses this function ofcounterfactual validity. On this view, the “ought” of the lawhas little to do with either background moral or politicaltheories or threats of enforcement. All it means is that eachmember of a legal community has a legitimate expectationthat others act as they ought, i.e., in accordance with theconventional rules, just because they have agreed to theauthority which has produced those rules and not becausethey reflect any ultimate moral truth or other reality. The sameobligation obtains in instances where the rule is arbitrary fromevery moral perspective (e.g., driving on the right or on theleft).

It is worth noting in passing that to attach the law to atheory of moral truth is to lose this advantage—for that wouldtie the “ought” of the law back to a cognitive uncertainty, i.e.,the truth about morality. And to the extent that moral truth isthought to be objective and knowable, determining the“ought” of the law is transformed back into a cognitiveproblem with all the residual uncertainty and ground forcontroversy which that entails.


— 119 —

LAW

C. The Sociological BasisThe idea that human actions are best described withreference to such systems of reciprocal expectations isfundamental to the sociological theory of Talcott Parsons andother social interaction theorists. Parsons’ central thesis wasthat

…as soon as several actors who are each able tochoose the meaning of their action subjectively wantto act in relation to each other in a given situation,the mutual expectations of behavior must beintegrated, and this could happen by aid of thestability of lasting, learnable and internalizablenorms. Otherwise the ‘double contingency’ of thedetermination of meaning between two subjectscould not be overcome, nor could the‘complementarity’ of expectations be established…[E]very lasting interaction presumes norms andwithout them a system could not exist.”

Conventions and Their Assumptions: Insights fromGame TheoryIt is helpful, in trying to grasp the huge magnitudes of theignorance we overcome through the use of rules, to look atthe rule-formation process in the microcosm of two-personrelationships. Schelling’s work in game theory provides someof the most helpful insights into the nature of conventionsand the ubiquitous human activity of forming conventions.Schelling recognized that not all games are zero-sum and thatthere are games in which each player can win only if the otherdoes as well. Schelling became interested in these gamesbecause of the coordination strategies they can produce. Thesuccess of players in such games depends on the degree towhich they can accurately perceive what the other playerexpects them to do so that they can do it and achieve perfectcoordination. The universal strategy of players in such games,regardless of the level of communication allowed betweenthem, is to establish conventional understandings which haveno intrinsic value, but which serve to signal intentions andexpectations to each other. Even in games where no overtcommunication is allowed, players universally find ways ofsuggesting patterns of conduct or rules of behavior to oneanother which greatly improve their joint chances of winningthe game.

From this we learn that even the simplest exercises intwo-person coordination require the establishment of at leastimplicit conventions or rules. A complex society is possibleonly because personal, face-to-face relationships are notrequired for establishing such rules for mutual interaction.Rather, the community provides the means for producing,publishing, perpetuating, and modifying the rules of language,markets, and the law. The rules which can protect us fromrisks of error are largely public goods equally available to allcommunity members.

It is important to point out that the error we seek to avoidin social interactions is not so much the error of predictingwrongly what others will do but misjudging what they willexpect us to expect them to do. Our own actions are basedin expectations of expectations. Rules have their main benefit

at the reflexive level of expectations of expectations,thus creating certainty of expectation from whichfollow the certainty of one’s behaviour and the

predictability of others’ behaviour only secondarily…[C]ertainty in the expectation of expectations,whether it be by aid of purely psychological strategiesor by aid of social norms, constitutes the essentialbasis of all interaction and is much more meaningfulthan the certainty of fulfilling expectations.(Luhmann 1985, 30)

Rules are able to accomplish this in a complex societyfor several reasons. Rules anonymize and take theexpectations into the realm of the impersonal. They are validregardless of who expects or does not expect. Rules aretemporally stable; they do not require renewedascertainment. And rules are so abstract in their factualitythat they regulate the expectations of expectations (Luhmann1985, 30).

Conventional Rules Imply Mutual AdvantageBecause the agreement implicit in conventions is presentmost visibly at the time a convention is born, an understandingof conventions for the purposes of legal theory can benefitfrom accounts of the process of convention formation. Again,Schelling is perhaps the most helpful on this score. It wouldseem intuitively correct to assume that agreements emergein the coordination of action (from the point of view of theactors) when there are opportunities to benefit mutually fromsuch coordination. Because the operation of rules tends toregularize behavior, rules most naturally enter our lives whenthere is advantage to be had from such regularization. Suchareas of life include those where (a) conflict might bedangerous, (b) cooperation would be beneficial, (c)understanding is a condition for amelioration, or (d) one’sbehavior directly impinges upon the welfare of others.

When we speak of conventions as solutions tocoordination problems we specifically think of the problemfrom the perspective of the various actors whose actionsrequire coordination. In such problem situations there aretwo fundamental factors which prevail. First, the gains foreach participant from coordinating action are usually sosignificant that there is incentive on both sides to find asolution. Second, because each actor is a separate individual,his or her specific interests may indicate a somewhat differentsolution than what will be most preferred by any other actor.Thus, any solution reached by agreement among the actorsmust be advantageous in some way to each one, but will notlikely be maximally advantageous to any.

Levels of ConventionGiven the complexity of modern societies, we could not speakof law as convention unless we had an extended concept ofconvention formation that preserved the notions ofparticipation and agreement beyond the point where face-to-face negotiation is possible. The following sketch of atheory of levels of convention formation might help show howthat is possible.

To this point I have only considered examples of simpleconventions where people could come together inarrangements which straightforwardly meet their variousrequirements on some specific issue. Nothing is required forsuch agreements beyond information, initiative, knowledgeof self-interest, and some freedom of action.

The enormous improvements available through suchprimary level conventions soon attract our interest in


— 120 —

LAW

extending this technique into situations which are morecomplex. Simple exchanges or arrangements are not feasiblefor all situations where there are mutual benefits to be hadfrom agreements. Examples of such complicated situations,all of which are exemplified in some context by law, mightinclude:

1. Conventions involving large numbers of participantswhere the time and effort required to includeeveryone in the negotiation are not easily justifiedby the gains to be had.

2. Conventions which benefit a community as a wholebut do not benefit all participants equally.

3. Conventions which are neutral with respect toindividual interests, and which could easily bedifferent, but which are important to have, inwhatever form (e.g., traffic rules).

4. Conventions which produce or constitute publicgoods and thereby provide motivation foranonymous and unilateral cheating.

In a broad sense, all such complicated situations posecoordination problems in the game theory sense of the term,at least if it can be construed broadly. In these situations, allparticipants stand to gain from the establishment of mutualrules, but the means for negotiating such rules are costly orimpossible to manage. For such situations, it does makesense, when the situations obtain for the same definablegroup of participants, to solve the coordination problem inHobbesian fashion by creating an authority which can act notas a member of the community, but in the name of allcommunity members, and can formulate rules to governthese situations. The fundamental and essential conventionon which all the rest stands is the agreement to enter into acommunity of authoritative rules or law. Because of Hobbes’sinsight that “no man can be obligated except by an act of hisown,” this convention entails the daunting requirement thatit be based on unanimous agreement.

The Unanimity RequirementAs daunting as the requirement of unanimity might firstappear, it does have a solution—which in turn resolves thepuzzle about rule of law. Unanimous agreement to the basicconvention of a legal society presupposes some expectedbenefits to all and avoidable injury to none. Many-partyexchanges or agreements reached by unanimous consentretain the same unqualified voluntariness as a freely engagedtwo-party agreement.

But actual unanimity in a law community is impossibleto ascertain. As a result, a substitute notion is needed. Thecommon error of much democratic theory is to compromisefor something like majority rule, which is an essential part ofpublic decision making. But majority rule alone is inadequatefor the creation of universal obligation to obey laws. The trueanalogue to unanimity is rule of law. The conditions ofequality, generality, and prospectivity assumed in the rule oflaw can be construed to preserve the abstract conditions ofunanimous agreement. It only makes sense that in voluntarilyceding their veto right over the process of making rules oflaw that would obligate them in the future, people would insiston certain conditions—that (1) no such laws could single outidentifiable individuals for penalties or benefits, (2) allindividuals would be treated equally under any laws orprocedures, and (3) the rules could not be changed after the

fact or in any other way that might hand discretionary orarbitrary power to those functionaries authorized to apply thepenalties of law to specifiable individuals. On analysis, theseand other similar conditions turn out to be equivalent to thelong-recognized principles of the rule of law, and to includesome others not previously recognized in this connection.

It is at this point in the analysis that the works of F. A.Hayek and Michael Oakeshott are so valuable in spelling outwhat such a form of association would be like in the abstract.But the actual constitutional structure of such a communityand the actual laws it would produce over time would becontingent on the experience, wisdom, interests, and politicalbalance represented in the community. And they could beauthoritative only on the basis of an understood, fundamentalagreement to form and maintain a community of law, madein spite of the inevitable fact that there would be winnersand losers on specific issues. The reasons for engaging insuch a convention include not only the evils it can remove(Hobbes) but also the vastly augmented opportunities forhuman action that it creates (Hume). Nor is such afundamental convention self-enforcing or maintaining, as thecollapse of law under the Third Reich and countless otherexamples demonstrate.

Perhaps the most improbable claim of this theory is thata legal system can reasonably be characterized as resting onthe actual agreement or consent of every member of thesociety that is obligated to obey the law. The defense of thisassumption is only possible for those societies in which adultscan choose to leave—taking all their legally obtained assets—at any time. In such a society, all persons are deemed toconsent to the authority of law, either because they chose tomove into the geographical area served by that legal system,or because they were born there and never chose to leavefor another situation they deemed preferable. This conclusiondoes not require that their choices include objectively optimalalternatives. We live in the real world, and our choices ofregimes are limited to the ones that actually exist. When wechoose one that actually affords us the protections of law,and we take advantage of that law to further our own interestsby participating in the market and the legal society, we givethe members of that society sufficient reasons to expect thatwe will obey the law and that we expect them to do the same.

The agreement of a community to a fundamentalconvention of this kind constitutes a special kind ofcommunity, which subsequently gains definition as itestablishes a body of coincident rules that apply equally toeach member of the community and a set of establishedprocedures for maintaining the magistracy and revising therules. Such communities are legal societies, bound by theirmutual obligations to a authoritative laws and a government.

Constructive UnanimityThe move from actual unanimity in social-decision makingto the establishment of legal authority merely recognizes thepractical impossibility of achieving complete agreement ona day-by-day basis, as well as the temptation to hold out forinordinate concessions that the requirement of unanimitycreates for individuals. The conventionalist theory advancedhere claims that it is rational, given the impossibility ofachieving actual agreement on every rule that will necessarilyaffect all members of the society, to concede one’s veto oversuch rules in exchange for institutions of authority on certain


— 121 —

LAW

conditions that will govern the future exercise of that authority.The creation of legal authority is by definition unanimous assuch authority only exists for those who agree to it and chooseto take advantage of its existence in their lives. Scholars whohave studied the process of social agreement have noted anumber of general characteristics of such agreements. I haveelsewhere shown how these empirically identifiedcharacteristics of actual agreements can be generalized as aset of abstract conditions of agreement.

One of those conditions is that the authority beconstrained by a standard of constructive unanimity, that is,that it be limited in its actions to creating and enforcing rulesand procedures that all members of the society mightreasonably have agreed to in advance. This is not a Rawlsianveil of ignorance. The assumption has to be that peopleknowing their actual preferences, values, needs, and abilitiescould reasonably have agreed to let standards be laid downby public officials. This notion of constructive or ad hocunanimity suggests the following as implicit limits on all legalauthority:

1. Rules cannot violate the deeply held moral andreligious beliefs of citizens. To the extent that theseare matters many people hold to be more importantthan anything else, including life itself, it is notreasonable to expect them to put these at risk inagreements made with others to improve theirsituations in other respects. This may limit the rangeof moral and religious views that can share a singlelegal system. But note that it is a negative restraintonly and does not require complete moral andreligious agreement. It does require religious liberty.

2. Authoritative decisions cannot arbitrarily single outindividuals or groups for particular penalties orbenefits.

3. The public officials themselves, in their private rolesas citizens and in their public roles as magistrates,are subject to all the rules they create.

4. There can only be one set of rules for everyone.There can be no special (privileged or repressed)categories of citizens.

5. Rules cannot be changed after the fact or maderetroactive in their application without the actualconsent of all concerned or compensation to thosenegatively affected.

6. All making and enforcement of rules must beknowable and observable by all citizens.

7. Every citizen must have reasonable access to theprocess by which the rules are formulated andadministered.

8. Every citizen, when accused of rule violations, musthave full opportunity to defend his or her case beforedisinterested judges .

Other conditions of such social conventions could belisted. This short list includes some of the most importantconditions of conventionality. The claim is that anyfundamental convention to be ruled by law implicitly holdsthese as limitations on the authorities it creates. These arenot moral principles, but are factual conditions that protectconstructive unanimity. Given the anticipated benefits, it isreasonable to give up one’s veto to an authority that can beexpected to act within these limitations.

The Principles of the Rule of Law.The link between this kind of rational actor analysis andtraditional natural-law and rule-of-law theories is that mostof these conditions of conventionality can be expressed interms of the widely recognized principles of rule of law,particularly as these have been articulated by Hayek andFuller. Elements 1 and 7 on this list have not been previouslyrecognized in rule of law discussions. The others are directlytranslatable into recognized rule of law principles as follows:

1. The principle of generality—all rules must be generalin scope.

2. The principle of generality (again!)—the rules mustapply to everyone.

3. The principle of equality—there cannot be more thanone class of legal persons.

4. The principle of prospectivity—new rules can onlyapply to the future.

5. The principle of publicity—no secret rules orprosecutions are allowed.

6. The principle of due process—all prosecutions mustfollow established rules which give defendants andplaintiffs full opportunity to defend their actions andtheir interests.

On this analysis, the conditions of conventionality orconstructive unanimity are equivalent to the principles of therule of law. Rule of law is revealed as an implicit norm orstandard for legal communities that understand law asagreement or convention rather than coercion or habit. Ruleof law just means constructive unanimity or conventionality.This includes, but goes far beyond, the view that authorityrequires obedience to law. It also entails a broad set of implicitlimits on all authority. Not just anything can be a law. Thereare implicit standards and limits in much the sense that naturallawyers have always wanted. But these standards are notderived from abstract moral or political principles. Rather,they are derived from social facts, from actual individualchoices made for whatever reasons people (who may bepresumed to act according to their own moral convictionsand interests) may have as they pursue what they considerto be important in their lives. Thus the theory claims totranscend the recognized impasse in legal theory by basinglaw in social fact, while identifying broad standards inherentin law.

The Rule of LawOn the theory sketched in this paper, the principles of rule oflaw become tests of conventionality. They provide a legalsystem with non-conventional, necessary norms for regulatinga system of conventions. They are the substitute for a theoryof justice or rights or principles (Dworkinian). Yet they arenot substantive.

The principles of rule of law in this context need not besuspected of being disguised guardians of Kantian autonomy.People do not have to be Kantians to prefer a society ofconventions to a rule of tyrants, even if it turns out that inprinciple Kantians would favor such an arrangement. So alsomight liberals of other stripes, including utilitarians.

The side benefits are enormous. The first of these is thatby admitting the notion of convention, the logic of consensusinfuses the system of law with a higher order of normativity—the principles of rule of law which become non-ethical and


— 122 —

LAW

non-natural law criteria for a principled criticism of law. Andthese meta-legal principles are sufficiently powerful that theycover most of the ground that the ethical theorists and naturallawyers want included in the critique of law. Furthermore, itmakes the critique internal to the legal system, and availableto judges. It is a means whereby such second order normscan be invoked without any presumption that a particularsystem of ethics may be correct or even preferable.

The weakness from the point of view of more extremenatural lawyers and political theorists will be that there is noabsolute protection of rights. This is correct. The systemcannot maintain itself against a determined, contraryconsensus in the population. Rule of law is a spectacular,but fragile achievement in human societies. Its continuingexistence and effectiveness will always depend on theunderstanding and commitment of citizens to preserve it indifficult times. Hence, the recurring recognition of politicaltheorists—that free societies depend as much on public virtuefor their success as they do on well designed constitutions—is well founded.

Endnotes* Professor, Department of Political Science, Brigham YoungUniversity.1. F.A. Hayek, The Road to Serfdom (London: Routledge, 1944) andThe Constitution of Liberty (Chicago: University of Chicago Press,1960).2. Lon L. Fuller, The Morality of Law, Rev. ed. (New Haven: YaleUniversity Press, 1969).3. Niklas Luhmann, A Sociological Theory of Law (London:Routledge and Kegan Paul, 1985).4. Talcott ParsSons and Edward A. Shils, eds., Toward a GeneralTheory of Action (Cambridge, MA: Harvard University Press, 1951),pp. 14, 105.5. Thomas C. Schelling, Strategy of Conflict (Cambridge: HarvardUniversity Press, 1960).6. Michael Oakeshott has produced the most abstract philosophicaldiscussion. See his On Human Conduct (Oxford: Oxford UniversityPress, 1975) and “The Rule of Law,” On History and Other Essays(Oxford: Basil Blackwell, 1983).7. Noel B. Reynolds, “Law as Convention,” Ratio Juris 2 (March 1989):105–120 and “Pareto Optimality and the Rule of Law,” in Method andMorals in Constitutional Economics (Berlin: Springer, 2002): 237–252.

The Legal Coordination Game

Gerald F. Gaus*

I. The Coordination Model of Political AuthorityJeremy Waldron tells us that “the felt need among membersof a certain group for a common framework or decision orcourse of action on some matter, even in the face ofdisagreement about what the framework, decision or actionshould be, are the circumstances of politics.” Political authorityand the law, Waldron insists, presuppose the circumstancesof politics. We reasonably disagree not only about conceptionsof the good life and value, but about justice and the commongood. However, because we need to act together, we cannotrest content with each going his own way. We thus have todeal with the fact that we reasonably disagree while achievingsome sort of unity of action. Thus our need for politicalauthority and the rule of law.

According Waldron’s interpretation of the circumstancesof politics, then, we (1) feel a need for a common frameworkor action yet (2) have intractable, yet reasonable, differencesas to what that should be. Waldron understands the problemin terms of an impure coordination game as in Figure 1 below.

Figure 1: An Impure Coordination Problem

This is Luce and Raiffa’s “Battle of the Sexes” problem. Rowand Column wish to go out together: he wants to go to thefights with her (<R1, C1>); she wants to go the ballet withhim (<R2, C2>). Either coordination point (<R1, C1>, <R2,C2>) is preferred by both of them to options in which theyfail to coordinate. Thus <R1, C1> and <R2, C2> arecoordinative equilibria.

Waldron clearly thinks all this is important for an analysisof the rule of law, though it is not quite clear just how it isimportant. He tells us that whether such a coordination gamewill be solved

depends on the circumstances of each case,including how much more each prefers his or herown favourite outcome to the less favored[equilibrium point], how likely each thinks it is thatthey will get their favourite outcome by holding out,etc…I do not want to claim that law solves PCs[coordination games] and that is why we shouldrespect it. (LD: 104)

The law, Waldron argues, can make one coordination pointmore salient by attaching sanctions, and so make it less likelythat people will hold out for their favorite outcome. “Butbefore it can do that, the society must have decided which ofthe coordinative strategies to select as the one to be bolsteredin this way. That itself is no mean achievement—and I wantto say that it is by embodying that achievement that lawcommands our respect” (LD: 104).

The idea seems to be that although a specific lawcontributes to coordination by “selecting” a specificcoordination point, it does not necessarily “solve” acoordination problem: the “achievement” of the law is toselect which coordination point should be sought. It is hardto see just what this means—unless we actually coordinateon a point, the selection of it as the one to be pursued doesnot seem much of an achievement—but what is clear is thatWaldron believes that law and legal authority can be modeledon an impure coordination game:

We want to act together in regard to some matter M,but one of us thinks it is important to follow policy Xwhile others think it is important to follow policy Y,and none of us has reason to think any of the othersa better judge of the merits of M than himself…

In these circumstances, the following will not be away of settling on a common policy: each doeswhatever he thinks is important to do about M. Wemust find a way of choosing a single policy in which[we]…can participate despite our disagreements onthe merits (LD: 107)


— 123 —

LAW

As Waldron understands politics, we will debate anddiscuss the merits and demerits of each of the possiblecoordination points (policies); since it is an impurecoordination game, I prefer a different coordination point(<R1, C1>) than do you (<R2, C2>), and so we havesomething to argue about. However, we each prefer anycoordination point to lack of coordination (<R1, C2>, <R2,C1>). In essence, then, Waldron argues that we need tocoordinate on some single reasonable policy, even if it is notthe one that each of us sees as most reasonable.

II. Is Coordination Always Better than Going it Alone?In contrast to John Finnis, who insists that the sense in whichlaw solves a coordination problem is distinct from “the gametheoretical concept of a coordination problem,” Waldronexplicitly draws on game theoretical modeling of coordinationgames, so let’s look closer at these games. Now it is temptingto suppose that the idea of an impure coordination gamenecessarily implies that, while we have disputes about thepreferred coordination point, we must find a way of actingtogether, since any way of coordinating is better than anyuncoordinated outcome. But things are more complicatedthan this. Following David Lewis, let us define a coordinationequilibrium as “a combination in which no one would havebeen better off had any one agent alone acted otherwise,either himself or someone else.” Given this, it is possible tohave a coordination game in which some uncoordinatedpoints (which are not in equilibrium) are better for everyonethan some coordinated points. Consider for example Figure2:

Figure 2: A Coordination Game With a Non-Coordination Point Pareto-superior to a CoordinationEquilibrium

<R3, C3> satisfies Lewis’ definition of coordinationequilibrium; no one can be made better off by a move byeither. Given Column’s play of C3, there is no move that Rowcan make that is better off for anyone; given Row’s play of R3,there is no way Column can move that makes things betteroff for either. Consider <R2, C1>, which is not a coordinationequilibrium; either player can make a unilateral move thatmakes both better off. Yet <R2, C1> is Pareto-superior to <R3,C3>; both players are better off in an uncoordinated non-equilibrium than in the coordinated equilibrium of <R3, C3>.So it does not follow that in every coordination game “goingit alone” is always worse than every way of coordinating.

To be sure, in Figure 2 it is rational to coordinate giventhe play of the other player. <R2, C1> is not a stable option;unilateral defection by either player would move them bothto an equilibrium that Pareto-dominates it. Nevertheless, wesee in Figure 2 that it is not the case that in every impurecoordination game it is always better for everyone tocoordinate than if they were somehow stuck in anuncoordinated outcome. Suppose we found ourselves in <R3,

C3>; this could not be justified on the grounds that, whilesome coordination points are better, it is at least better foreveryone than if somehow there was no law that producedcoordination. Figure 2 does not support the view that legalauthority and the law are worthy of respect just because theyhelp select ways of coordinating, for some ways ofcoordinating are worse for everyone than some ways of failingto coordinate. If we are going to make the coordinationanalysis of law attractive we must add, at the very minimum,the further requirement that to qualify as a legal coordinationgame no coordination equilibrium is Pareto-dominated by anynon-coordination point. That is, if legal authority is to bejustified in terms of a coordination game—if we are to explainwhy it “commands our respect”—we need to specify anadditional requirement of the game: it must never be the casethat rational agents would unanimously agree to move froma coordinated outcome to an uncoordinated point. (As theywould in Figure 2, where a move from <R3, C3> to <R2, C1>would be endorsed by both Row and Column.) To say thatlaw allows us to coordinate but that we would unanimouslyagree to return to an uncoordinated situation hardly seems acompelling case for law. More strongly, but still very plausibly,we should add that in our legal coordination game everycoordination point actually Pareto-dominates every non-coordination point. It must always be the case that it is betterfor everyone (or, at least, not worse for anyone) to act togetherthan to go it alone in any way. Each person does at least aswell in every coordination equilibrium as he would so in anyway of going it alone. Let us call this the Pareto-dominanceof coordination. Unless this holds, we need to comparedifferent ways of going it alone to different coordination points;some people will rationally prefer some cases of an absenceof a legal system to some legal systems. For them, it wouldnot be true that each doing “whatever he thinks is importantto do about M” is suboptimal vis a vis every way ofcoordinating. At least from the perspective of some,coordination would look like a fetish for acting together, whichmakes them worse off than they might have been.

III. Coordination on Civil SocietyThe most plausible version of a coordinative analysis of lawis to see a specific legal system, or political society, as acoordination point, and so the absence of coordination a stateof nature without law or political society. Hobbes’ theory issometimes interpreted in this way. We can depict Hobbes’state of nature as a no-agreement point, and all civil societiesas equilibrium points. The power of Hobbes’ characterizationof the state of nature is that it is so horrible that everyconceivable political society is a coordination equilibrium thatPareto-dominates every non-coordinated point. So, bydepicting horrible no-agreement points, Hobbes can showthat everyone benefits by any coordinated outcome (type ofgovernment). Given that all civil societies are coordinationequilibria, no one has any incentive to defect, understood asleaving a coordinated outcome (a civil society) to return tothe state of nature.

It is often underestimated just how much an effectivecoordination account of civil society depends on a Hobbesian-like state of nature story. As soon as we make the state ofnature a kinder and gentler place, with perhaps Lockean“inconveniences” but not constant war, we immediatelyundermine the claim that all no-agreement points are Pareto-dominated by every social contract. Locke was no game


— 124 —

LAW

theorist, but he clearly recognized this. The Hobbesian canonly get her result if anything is better than no agreement;but if no agreement is inconvenient but not a living hell, thenwe will be more selective about possible coordination points.Some ways of acting together, or some common frameworks,will be seen by some as worse than no agreement. Thesepoints are not, in principle, possible solutions to the legalcoordination game. People are apt to start insisting on clausesto the social contract, excluding “ways of acting together”(types of civil societies) that they rank as worse than the stateof nature. Some may rank regimes without bills of rights asworse than the state of nature. And the less harsh we makethe state of nature, the more civil societies will not beperceived as improvements by some contractors, and so theywould be disqualified as possible solutions to the legalcoordination game.

In Figure 3, Option 1 might be the U.S. Constitution witha Bill of Rights (most preferred by Column); Option 2,Parliamentary sovereignty with an independent judiciary(most preferred by Row), and Option 3 a Hobbesian sovereign.Again, we see that accepting a Hobbesian sovereign (<R3,C3>) is worse than living in a state of nature.

C1 C2 C3

R1

21

00

00

R2

00

12

00

R3

00

00

1-1-

Figure 3: Three Social Contracts

The problem here is that <R3, C3> is not a coordination point.It is not a coordination equilibrium—indeed it is not anequilibrium solution at all. The intuitive idea of coordinationas “doing the same thing” departs from the formal idea of acoordination equilibrium. In the ordinary language sense itlooks as if we coordinate, but <R3, C3> clearly is not inequilibrium, so in the formal sense it is not a coordinationequilibrium. Thus the analysis of coordination games doesnot justify the conclusion that all ways of “doing the samething” are better than no coordination because, formally,some types of “action-in-concert” (LD: 108) are instances ofnon-coordination. Thus, just because we have a commonpolicy X on matter M, it does not follow that X is a coordinationequilibrium. The less harshly we describe the “state of nature,”the more common policies actually fail to be coordinationequilibria. And even if some ways of doing the same thingare genuine coordination equilibria, that is not enough: onlyif we have the Pareto dominance of coordination (§II), canwe present a plausible case for the authority of law.

The upshot of all this is that other things equal, the moreattractive the no-agreement points, the more possiblecommon policies—legal systems—will fail to Pareto dominatethem. This has deep consequences for Law andDisagreement. Waldron advances a number of argumentsagainst entrenched constitutional rights, but consider whathe tells us is “the deepest reasons of liberal principle for beingunhappy about a Bill of Rights”:

When a provision is entrenched in a constitutionaldocument, the claim-right (to liberty or provision)

that it lays down is compounded with an immunityagainst legislative change. Those who possess theright now get the additional advantage of its beingmade difficult or impossible to alter their legalposition. That can sound attractive; but, as W. N.Hohfeld emphasized, we should always look at bothsides of any legal advantage [citation omitted]. Thecorrelative to the claim-right is of course the dutyincumbent upon officials and others to respect anduphold the right. And the term correlative to theconstitutional immunity is what Hohfeld would calla disability; in effect, a disabling of the legislaturefrom its normal functions of revision, reform, andinnovation in the law. To think that a constitutionalimmunity is called for is to think oneself justified indisabling legislators in this respect (and, thus,indirectly, in disabling the citizens whom theyrepresent) (LD: 221).

We can now see that, surprisingly, the disability to whichWaldron objects is soundly grounded in the legal coordinationgame. Assume that we do not embrace a Hobbesian state ofnature, but we make the no-agreement points (regarding civilsociety) less harsh—say people order the outcomes in a waycloser to a Lockean story. Given this Lockean sort of ordering,legal systems that are understood by some as making themworse off than they are in the state of nature—say becausepeople believe such systems violate fundamental rights to life,liberty and property—will not be genuine coordinationequilibria at all. Moreover, even if some arise as coordinationequilibria because of the choices of others (e.g., <R3, C3> inFigure 2), they may not Pareto-dominate all versions of thestate of nature. It follows, then, that insofar as the authority oflaw derives from its role in improving everyone’s lot by helpingto achieve coordination in the face of disagreement (LD: 108),it has no authority if it selects a way of acting together, X, thatis not Pareto superior to some state of nature, Y. X would notbe a solution to the legal coordination game. Its way of actingtogether does not improve the lot of some: we cannot expectthem to grant authority to states that fail to improve on thestate of nature. The very thing to which Waldron objects—the legislative disability—is precisely what follows from thelegal coordination game’s account of political authority.Entrenched rights identify “actions-in-concert” that are notPareto-dominant to all states of nature, and so cannot besolutions to the legal coordination game.

It appears, then, that unless Waldron wishes to take theHobbesian route of characterizing all no-agreement pointsas horrendous, and so claim that all ways of acting togetherare coordination equilibria that Pareto dominate all ways of“going it alone,” he cannot both assert that the key tolegislative authority is the way in which it assists in solvingthe legal coordination problem and that there is somethingdisturbing in principle about legislative disabilities. If it is thecase that legislation has coordinative authority only insofaras it is required to solve the legal coordination game, then inall those cases in which its directives would fail to do sobecause the directives are not Pareto superior to some no-agreement point, it does not have authority.

IV. Legislation as CoordinationWaldron’s main concern, however, is not a coordinationanalysis of political life in general, but of politics, andespecially of legislation. Waldron is, of course, right that there


— 125 —

LAW

are some things on which we need to coordinate, and forsome matters the coordination account is insightful. We all(or, very nearly all) agree that we need some laws of property,and that any regime of property rights is better than a free-for-all. Waldron, though, isn’t just out to show that some ofour current laws are coordination points (again, vis-à-vis thestate of nature), but that we can understand debates aboutnew laws, and our acceptance of their authority, as impurecoordination games. “A piece of legislation deserves respectbecause of the achievement it represents in thecircumstances of politics: action-in-concert in the face ofdisagreement” (LD: 108).

The story, to recap, is this. We are considering a range ofpossible laws {X, Y, Z} with regard to some matter M (LD:107). We each rank the alternatives differently (hence ourdisagreement). If the law enacted, Y, is the solution to a legalcoordination game it must be the case that Y Pareto-dominates every no-agreement point. Now it would certainlyseem that the obvious way to characterize the no-agreementpoints is that they all constitute no legislation at all on matterM, so each unilaterally acts in her preferred way. Thus, forWaldron to show that new legislation solves the legalcoordination problem, it must be the case that “no law at allon M” ranks below every proposed (possible?) law ineveryone’s preference ordering. And that appears to be whatWaldron does suppose: “Suppose too,” he adds, “that we allknow that M requires a common policy…” (LD: 107).

This supposition, though, is manifestly contrary to fact.Overwhelmingly, in debates about new legislation a significantnumber of people believe that no common policy on M isrequired, or at least that no common policy is certainly betterthan many of the proposed ways of acting together. GivenWaldron’s very permissive conception of the reasonable, onalmost any issue there are reasonable citizens who believethat no common policy should be pursued. Debates aboutabortion, drug laws, environmental policy, trade policy,pornography, affirmative action and stem cell research areall examples: many would insist that some of the proposalsoffered by their fellow citizens are distinctly worse than nolegislation at all. With pornography, for instance, classicalliberals would insist that no common policy at all is thepreferred option: people can read it, buy it, leer at it orwhatever, if doing so is consistent with their other rights (suchas property rights). Some people will prohibit it in theirbuildings, others won’t: each going her way is the preferredoption.

Over a very wide range of political issues, then, it wouldseem that for each and every proposal P in the set of options,a number of citizens will rank it as inferior to some no-agreement points: (1) those who prefer all no-agreementpoints to all agreement points, and so rank P and all otherslaws behind every way of going it alone, such as classicalliberals in our pornography case; (2) those who prefer some,but not all, ways of going it alone to all ways of acting togetherand (3) those who prefer some non-P option to some noagreements, but prefer some no agreements to P. It thusseems almost impossible for any new law to be a solution tothe legal coordination game.

Ah, one might say, but then no new law—liberty—is thecommon policy. Allowing porn (or abortion, or not regulatingpollution or stem cell research) is itself a policy, so there isno getting away from a common policy. This prima facie

tempting reply totally trivializes the idea of a coordinationgame and Waldron’s use of game theory. On this view everycell in Figure 1 is a coordinated outcome; it is just thatsometimes we coordinate by doing things differently. Thereno longer is a no-agreement point, for every cell constitutesa “coordinated” outcome. That clearly won’t do, for not everycell satisfies the requirements of being a coordinationequilibrium.

Let us consider the problem more carefully. Suppose thatat time t we are playing a coordination game, and have arrivedat an equilibrium. Suppose that the game involves simply thefirst two rows and first two columns of Figure 4 (the shadedcells); and we have an equilibrium at <R1, C1>, a certain setof laws regarding property rights and personal rights. Nowsuppose at time t+1 a proposal is made to add to the set oflaws a statute against selling pornography (the unshadedcells).

C1 C2 C3

R1

21

00

00

R2

00

12

00

R3

00

00

5.15.1

Figure 4: Old Laws and New Laws

Although <R3, C3> is a Pareto-dominant equilibrium point (visa vis all uncoordinated outcomes), it does not Pareto-dominate the current coordination equilibrium. It does notsolve any sort of coordination problem, because there alreadyis a coordination equilibrium. Even though Row prefers <R3,C3> to the status quo, she would not defect from <R1, C1>,and of course Column has no incentive at all to move to <R3,C3>.

A factor upsetting the current equilibrium would do thetrick. If we are no longer at an equilibrium point, then we donot have to be moved away from one. Thus, it has beensuggested to me that we might imagine that the current lawsregarding property rights are no longer in equilibrium because,say, environmentalists refuse to obey them any longer,believing that they support environmental policies thatenvironmentalists view as wrong. Hence theenvironmentalists might violate the property rights of loggingfirms or petroleum companies. Thus we require a new lawto regain coordination. Such a situation is depicted in Figure5 (again, the original game is the shaded cells, the game att+1 includes all the cells).

C1 C2 C3

R1

21

00

00

R2

00

12

00

R3

1-2.1

00

5.15.1

Figure 5: Destabilizing New Preferences


— 126 —

LAW

Now given Row’s new preferences, what was previously anequilibrium <R1, C1> no longer is, and the new law <R3, C3>is indeed in equilibrium. But this account takes as its startingpoint that Row, the environmentalist, undergoes a preferencechange such that he prefers each going their own way, i.e.,<R3, C1>, to acting together, <R1, C1>. Notice, though, thatthis provides a coordination account of new laws only byundermining Waldron’s main claim: that though we all havedifferent preferred common policies, we all agree that acommon policy is better than each going it alone. Row doesnot think this in Figure 5 at t+1: Row now would prefer goingit alone unless Row gets his preferred outcome—and that iswhy Row violates the property rights of owners under thecurrent regime.

This account of new laws is thus paradoxical. If there isno preference change, the current laws are in equilibrium,and there will be no movement to a new law. On the otherhand, there can movement to new laws if some change theirpreferences and begin to disobey the current laws such thatwe no longer have coordination. But this requires that someprefer going it alone to having some (i.e., status quo) commonpolicies, and that is precisely counter to Waldron’s idea of the“circumstances of politics.”

Of course exogenous factors may be such that newproblems arise. More sophisticated evolutionary models canexplain why small changes in circumstances and behaviorcan lead to new equilibria. I do not wish to insist on a staticmodel (though, it should be noted, Waldron gives no cluewhatsoever as to how he might wish to include an analysis ofdynamic equilibria into his theory of law). Even in the type ofsimple games we are considering, however, we can build in arationale for changes of equilibria, e.g., when a new law wouldPareto-dominate the current law, and so everyone agreesthat the addition of the new law is better, as in Figure 6.

C1 C2 C3

R1

21

00

00

R2

00

12

00

R3

00

00

5.25.2

Figure 6: Pareto-dominance of a New Law

Again, the current game is in the shaded area, but because ofexogenous factors the third row/column become available.Although <R1, C1> is still in equilibrium, it is not in strongequilibrium as <R3, C3> Pareto-dominates it. We can easilyimagine, then, that a political authority might move us fromthe former to the latter. This suggests a different model ofnew legislation: rather than seeing it as selecting coordinationpoints and so helping to solve the legal coordination game,we might see politics and legislation as moving us around thematrix, from one coordination equilibrium to another. Atone point we have <R1, C1>; then we take a vote and moveto <R2, C2>, which is not Pareto-superior, but simply preferredby a majority. New laws, then, do not help solve coordinationproblems, they move us from one coordination equilibriumto another.

This, I think, is the most attractive interpretation ofWaldron’s account. But its plausibility depends on theequivocation between “ways of doing the same thing” and“coordination equilbria” that I pointed out in section III.Consider Figure 7:

C1 C2 C3 C4

R1

21

00

00

00

R2

00

32

00

00

R3

00

00

5.25.2

00

R4

00

00

00

1-4

Figure 7: Legislation v. Coordination?

Suppose we start at <R1, C1>; as in the previous figure,legislation can move us to a new coordination point, <R2,C2>, that Pareto-dominates it. But legislation is majoritarian,so it can also move us to a coordination point in which themajority (Row) endorses egalitarian measures, taking somegains away from Column, hence it can move us <R3, C3>.But for the same reasons it can move us <R4, C4>, which isnot a coordination equilibrium at all. Once we allow thatlegislation can move us around the matrix, including to a pointthat is not Pareto-superior to the status quo, there is no reasonto suppose that the legislation will really identify acoordination equilibrium. Only by erroneously supposing that,by necessity, every way of doing the same thing is acoordination equilibrium, could that seem plausible.

V. A New Law or a Return to The State Of Nature?The claim that legislative politics can be modeled as a legalcoordination game might be saved if we can identifysufficiently nasty no-agreement points. If, for example, itplausibly could be maintained that all no-agreement pointsrepresent a return to the state of nature, then new legislationon M might be usefully modeled on the legal coordinationgame. For then M might plausibly be said to Pareto-dominateall no-agreement points. But, political rhetoric aside, it is arare issue indeed in which the cost of inaction is politicaldisintegration. Not even Hobbes’ theory endorses this claim.It is at least plausible to interpret Hobbes as maintaining thatany defection regarding any law threatens a return us to thestate of nature. Suppose this is Hobbes’ view. If so, then civilsociety itself is at stake with every decision whether to obeyevery law. However, even on this “Hobbesian” view, ex anteit is not the case in a Hobbesian state that all new legislationon M Pareto-dominates no legislation about M at all. As longas subjects have preference orderings, many are apt to preferno action at all to the selected option. To be sure, once thesovereign makes his choice we might have something thatlooks like a coordination game: unilateral defection wouldno longer make sense. And of course, the Hobbesian casethus understood turns on “extremely strong assumptions”—which is a nice way of saying they are false.

A variant of this interpretation of new legislation as alegal coordination game has been suggested. Consider what


— 127 —

LAW

it might mean to say there is no law at all on matter M—forexample, what might it mean to say there is no law at allregarding pornography. I have been supposing it means thatthe current legal system remains as it now is, but no additionallaw regarding pornography is introduced. But another possibleinterpretation employs a more modest version of the returnto the state of nature view. Imagine that we return to the stateof nature in all matters regulating pornography. Thus thebaseline is not all current laws that protect owners of adultbookshops, property rights, privacy, etc., but a return to thestate of nature regarding all these matters, so that each personis now at liberty to do anything she wants with pornography,from reading it, to burning it and, presumably, burning adultbookstores and pornographers. Call this the hole in the webof law interpretation: we imagine no agreement points asabsences of any legal regulations relating to this matter. Ifthat is the baseline, then any new law regarding pornographywill apt to be Pareto-superior to a hole in the web of law.Again, the trick is to depict the no-agreement points in asufficiently dastardly fashion so that anything would be better.

But this seems an awfully odd way to conceive of ourcoordination problem. Go back to Figure 1, the simple battleof the sexes problem. According to the hole in the web oflaw interpretation, although we now have a coordinativeequilibrium (say <R1, C1>), moving to <R2, C2> also can beexplained in terms of solving a coordination problem as it isbetter than either of the uncoordinated outcomes. Supposethat we are now at the fights (<R1, C1>) and my wife proposesthat going to the ballet solves our coordination problem,because it is, after all, better off then no coordinated outcomeat all! But if we are now in a coordinative equilibrium, wehave already solved the coordination problem; saying thatother equilibria are better than no coordination at all hardlygives us a coordination reason to move from where we are.Indeed, it misunderstands the concept of an equilibrium.Once we are at the fights, there is no coordination problemto be solved, even if my wife would have preferred the ballet.To be sure, we might think of legislation as simply moving usfrom one coordination equilibrium to another, but we havealready considered and rejected that interpretation.

One suggestion that has been put to me is that this modelof law might be insightful if all laws contained sunset clauses,so that at some point we were indeed thrown back into thestate of nature. That, at least, would solve the puzzle of whywe are constantly leaving equilibria. Numerous objectionsconfront such an interpretation, most devastating of which isthat our legal system seldom contains such clauses, so thecoordination account could not explain our system. “Themodel would work if our system was different than it is” isnot an especially compelling defense of a model of our legalsystem. More technically, this analysis would understand lawas an iterated impure coordination game; those sorts of gameshave their own range of solutions that are not discussed byWaldron or other legal philosophers employing game theory.

VI. Coordination Games, Collective Action and theRule of LawAlthough Waldron explicitly models the rule of law andlegislation on coordination games, often it seems that heunderstands them differently, as collective action problems(LD: 201). Indeed, that he believes we have an obligationbased on fairness to do our part in collective action suggests

the possibility of free riders and prisoner’s dilemmas (LD:239ff). So law might be understood as achieving a publicgood, which we all wish but would prefer to obtain withoutpaying the costs of obedience. This, of course, points to avery different way of understanding the rule of law. Whateverproblems there may be with understanding law as a type ofcollective action problem, it has one great advantage overcoordination accounts: it makes sense of the centrality ofsanctions to law. As Leslie Green effectively demonstratedwith respect to an earlier generation of coordination theoriesof law, it is exceedingly difficult for coordination theories toplausibly account for the role of sanctions in law. Finnisacknowledged this: if law was an impure coordination game,“deviance would not be so much wrong as irrational—unintelligible.” If we are in a coordination equilibrium, no onehas an incentive to unilaterally defect, even if some wouldprefer other equilibria. To be sure, it is possible to work insanctions in a round-about way—perhaps they are signalsletting people know what are the coordination points are.But it is hard to see how this accounts for the role of lawenforcement in our political society. It is certainly a stretch.Given the amount spent on policing and punishment, it wouldbe an extraordinarily expensive way of letting people knowwhat the law is.

VII. Conclusion: Law and Coordination GamesI have not argued here that it is unreasonable or erroneous tounderstand the law as coordinative. Perhaps Finnis is correctthat there is a distinct, non-technical, notion of coordinationthat can be appealed to—some way that makes sense of notwanting to be played for a “sucker” by non-cooperators (onecannot be played for a sucker in a coordination game). Andwithout doubt one of the functions the law plays is to allowus to gain benefits from coordinating our activities. But it doesnot follow that this long-standing insight is enhanced by tryingto employ game theory, and especially if one claims, as doesWaldron, that a specific sort of game is the primary model forunderstanding all of law. My point here is not the banal onethat any model of legal authority is apt to leave somethingout: as Waldron reminds us, models of authority have to bepurchased “wholesale, not retail” (LD: 121). Rather, I haveargued that the whole attempt to employ game theory in thisway is misguided. It fails to appreciate the requirements for asolution to the game, what would have to be true for law toactually be an equilibrium solution, and how our currentunderstanding of the law would actually be undermined bydepicting it as an equilibrium solution. Perhaps a much moresophisticated version, building in an analysis of dynamicequilibria, could do the job, but surely simple static gamesare not the place to look.

Endnotes* Philosophy Department and Murphy Institute, Tulane University.Earlier versions of this paper were delivered at the 2001 meeting ofthe American Political Science Association and to the Law Faculty atthe Universidad Torcuato Di Tella. I would like to especially thankJeremy Waldron, Larry Alexander and Guido Pincione for theircomments and suggestions.1. Jeremy Waldron, Law and Disagreement (Oxford: OxfordUniversity Press, 1999), p. 102. Hereafter referred to as “LD.”2. R. Duncan Luce and Howard Raiffa, Games and Decisions (NewYork: John Wiley and sons, 1957), p. 90.3. John M. Finnis, “Law as Coordination,” Ratio Juris, vol. 2 (March1989): 97-104.


— 128 —

LAW

4. David Lewis, Convention (Cambridge, MA: Harvard UniversityPress, 1969), p. 15. Emphasis in original. See also Jean Hampton,Hobbes and the Social Contract Tradition (Cambridge: CambridgeUniversity Press, 1986), p. 138.5. This is not the same as saying that each equilibrium point is astrong equilibrium. An equilibrium is in strong equilibrium when nooutcome Pareto-dominates it; I am only requiring that nouncoordinated outcome Pareto-dominates any coordinatedoutcome. See Peter C. Ordershook, Game Theory and Political Theory(Cambridge: Cambridge University Press, 1986), p. 305. Cf. LeslieGreen’s condition (2c): “(almost) everyone prefers that everyoneconform to some [norm] R rather than not conform to any.” “Law,Coordination and the Common Good,” Oxford Journal of LegalStudies, vol. 3 (Winter 1983): 299-324 at p. 302.6. See Noel B. Reynolds, “Law as Convention,” Ratio Juris, vol. 2(March 1989): 105 –120 at p. 107 and Green, “Law, Coordination andthe Common Good,” p. 301. In a footnote to LD, Waldron suggeststhat much of the force of Green’s criticism of modeling law as acoordination game is only directed against the particular versionGreen was attacking (Finnis). I believe that this under-appreciatesthe force of many of Green’s insightful points. Waldron does notpause to explicitly respond to Green’s detailed criticisms.7. See Lewis, Convention, pp. 10ff.8. It should be stressed that it only matters what they believe, and sohow they order the outcomes. Nothing depends here on the claimthat there actually are such rights. See Green, “Law, Coordinationand the Common Good,” pp. 309ff.9. Emphasis added.10. On this see David Estlund, “Jeremy Waldron on Law andDisagreement,” Philosophical Studies, vol. 99 (2000): 111–128.11. See above, note 6.12. At the roundtable on his Law and Disagreement at the 2001American Political Science Association annual meeting, SanFrancisco, CA.13. Again, an analysis of dynamic equilibria could explain why smallchanges in behavior might gradually move us to a new equilibrium.This holds the possibility for a more sophisticated coordinationanalysis of law, though it is not clear it would overcome the problemsof the baseline that I explored earlier in this paper. Although I cannotgo into it here, such a dynamic analysis seems more consistent witha stress on the common law (as opposed to explicit legislation), as adynamic equilibrium incrementally responding to exogenouschanges.14. See e.g., Nabil Al Najjar, “A Theory of Forward Induction in FinitelyRepeated Games,” Theory and Decision, vol. 38 (1995): 173-193.15. See Edna Ullmann-Margalit, “Is Law a Coordinative Authority?”Israel Law Review, vol. 16(1982): 350-55. But cf. Green, “Law,Coordination and the Common Good,” p. 317.16. Green, “Law, Coordination and the Common Good,” p. 317ff.17. Finnis, “Law as Coordination,” p. 99.18. Ibid., p. 102

RECENT ARTICLES OF INTEREST

—ABSTRACTS—

Richard A. Epstein. “Privacy, Publication,and the First Amendment: The Dangers ofFirst Amendment Exceptionalism,” 52Stanford Law Review 1003-1047 (2000).

It is paradoxical but true that the law rightly treats thedisclosure of information as both one of the greatest goodsand one of the greatest evils. On the one hand, the disclosureof information is of the greatest importance as a means toenabling human beings to make informed decisions and avoidall sorts of mishaps. On the other hand, it is essential for thesuccess of many endeavors that private information be keptprivate. Striking the proper balance between the norm ofdisclosure and the norm of privacy is a problem the law mustsolve. This paper assesses the relative merits of two mutuallyinconsistent views of First Amendment law with respect totheir capacity to aid in solving this problem. Unfortunately,both views can be found embodied in current case law. Thefirst view sees the role of First Amendment law as preservingcommon law rules which protect freedom of speech and thepress as part of a larger conception of freedom. The second,here called “First Amendment exceptionalism,” viewsspeaking and writing as peculiarly privileged activities, andpotentially affords those who speak and write (eg., membersof the press) rights that others do not have. The first view canaccommodate the privacy norm by prohibiting publication ofinformation that was obtained in violation of common lawduties (information, for instance, that was obtained bytrespass, fraud, or breach of confidence or contract). Thelatter view tends to encourage the creation of exceptions tosuch duties, and thus eliminates potential solutions to theconflict between the norms of disclosure and privacy.

Mahoney, Paul G. “The Common Law andEconomic Growth: Hayek Might Be Right,”30 Journal of Legal Studies 503-524 (2001).

There is evidence, the author maintains, that financial marketscontribute to economic growth and that legal institutionscontribute tot he growth of financial markets. There is alsoevidence that countries whose legal systems derive from thecommon law tradition provide superior investor protections,on average, than those whose legal systems derive from theFrench civil law tradition. The question is: Why?—for there isnot much substantive difference, in fact, in legal rules incountries with the two types of legal system.

The answer, the author believes, lies in the different viewsin the two sorts of system about the roles of the private sectorand government, and this, in turn, implicates differences inthe origins, in the 17th and 18th centuries, of the common lawand civil law systems. During the disturbances in England inthe 17th century the judges enlisted on the side of Parliamentand against the centralizing tendencies of the monarchy. More


— 129 —

LAW

specifically, they both insisted that the Crown was subject tothe common law, and created a law of property thattransformed landowners from tenants of the king to modernowners. In 18th century France, in contrast, the judges enlistedon the side of the monarchy and thus became targets of therevolution. In the resulting political system the judiciary wasplaced under the control of the government, in order to insurethat the “general will,” the policies of the elected government,could be implemented without judicial interference. Thesehistorical happenstances account, the author believes, for themore individualizing tendencies of the common law and themore centralizing tendencies of the civil law—as reflected inthe sharp distinction between public and private law that iscentral in the latter but not the former.

Using techniques of regression analysis, the authorpresents data, based on a sample of 102 countries, in supportof the claim that average annual growth in real per capitaGDP is greater in countries inspired by the common law thanin civil law countries. He claims that common law countries,between 1960 and 1992, grew, on average, 0.71% per yearfaster than civil law countries. The explanation, he argues, ismore dispersed governmental power and, as a result, superiorprotection for property and contract rights in common lawcountries. He argues also that the data show that the relationis causal.

Merrill, Thomas W. and Smith, Henry E.“What Happened to Property in Law andEconomics?” 111 Yale Law Journal 357-398(2001).

The authors argue that under the influence of the legal realistsand of Hohfeld, Coase, Calabresi and Melamed, and the law-and-economics movement, the idea of property hasmetamorphosed from a right in rem to a package of rights inpersonam. Whereas an in rem right was understood to involvea special type of relation of a person to a thing that confersrights that are good against the world, the modern “bundle ofrights” analysis takes property to be a variable list of rightsagainst identified persons with respect to particular resources.Some writers, for example, see ‘property rights’ (or ‘propertyrules’) merely as devices for establishing baselines forexchanges, contractual or otherwise, leading to highest-valueuses of resources.

Though there is much that is salutary in the bundle-of-legal-relations analysis, the authors maintain that there arefeatures of property it can’t adequately explain. For example,all modern property systems standardize property by limitingthe forms of ownership; and whereas nuisance lawsometimes seeks to accommodate the best uses of resources,in other cases it gives one party a blanket right to exclude.The explanation of these and other cases that are mysteriousfrom the bundle-of-rights point of view is that there aresubstantial informational costs associated with property.Where only a limited number of people are involved, the costsof ascertaining what uses of resources are permitted are low.But where large and indefinite numbers of third parties areinvolved, the in personam approach creates highinformational costs, whereas the in rem approach, involvesrelatively low informational costs—for the idea of rights that

are good against the world immediately lets third partiesknow what they may and may not do with respect to someresource. Property must be seen as settling relationships notonly among a small number of individuals, but between anowner and the world at large.

The authors speculate that the in personam analysis ofproperty is an epiphenomenon of a society that has solvedthe problem of order (security) and thinks of property onlyfrom the standpoint of uses that maximize welfare. But thehard nut of social order, and the role of property in achievingit, is still an essential element in an adequate understandingof property.

Lewis, Penney. “Rights Discourse andAssisted Suicide,” 27 American Journal of Law& Medicine 45-99 (2001).

Lewis provides a comprehensive, almost encyclopedic,survey of the philosophical and legal debate over assistedsuicide in terms of different types of rights. She firstenumerates rights supporting a right to assisted suicide,including liberty, autonomy or self-determination, privacy,dignity, equality, freedom of conscience and religion, andproperty. She then surveys rights invoked against assistedsuicide including a right to life, equality or equal protection,property, and autonomy. Next, she systematically considersthe various critiques of rights discourse, including argumentsthat it is indeterminate, that it suppresses and distorts thedebate, that it is uncompromising and impedesaccommodation of competing interests, that it is conclusoryand over-simplifying, and that it is absolutist. The broadercontext of rights discourse also is considered as it relates toand informs the assisted suicide debate.

Dworkin, Ronald. “Do Values Conflict? AHedgehog’s Approach,” 43 Arizona Law Review251-259 (2001).

Dworkin examines and rejects the claim that the politicalvalues of equality and freedom conflict. He proposes usingdynamic conceptions of “equality” and “freedom” to showthat they can be reconciled. These concepts (and others,such as “democracy” and “community”) are interpretiveconcepts which are the subject of continuing dialogue inpolitics and other communities of discourse, not static naturalkinds or “flat” definitions. After considering and dismissingvarious objections to his position, he shows how his approachrenders “bizarre” political claims that taxes which are usedto promote equality in society violate a right to freedom ofthe taxpayers.


— 130 —

LAW

Eugene Volokh. “Freedom of Speech andInformation Privacy: The TroublingImplications of a Right to Stop People fromSpeaking About You,” 52 Stanford Law Review1049-1124 (2000).

Various commentators on the law have proposed a serious“right of personal privacy,” conceived as a “right to controlinformation about ourselves.” Such a right would be a rightto restrict the speech of others. Such restrictions wouldrequire existing exceptions to the First Amendment to bestretched further than they have so far. In themselves, onemight think that the recommended restrictions on speechwould be worthwhile: Though they would also restrict one’sown speech, they promise to bring with them the assurancethat others may not circulate information about one thatmakes one uncomfortable. These considerations, however,have to do only with the intended consequences of theproposed right. This article argues that there are unintendedconsequences that give us reason for concern, consequencesthat are based on the fact that law develops by precedentand analogy. All of the major arguments for such a right—that it only restricts commercial speech, that it protectspersonal dignity, that it is needed to protect an already-recognized civil right, or that it passes muster under a “context-sensitive balancing” rule—would just as well justify curtailingpresently permitted speech in other areas.

Liam Murphy, “Beneficence, Law, andLiberty: The Case of Required Rescue,” 89Georgetown Law Journal 605-65 (2001).

The Anglo-American legal tradition has been peculiarlyunreceptive to the admitted moral duty to provide an easyrescue. It has long recognized that legal and moral duty areclosely related, but in the department of duties requiring notforbearance but positive action, it has often taken refuge inthe position that “with purely moral obligations the law doesnot deal” (Buch v Armory Mfg.). Liam Murphy finds thisposition to be at first glance curious, given that utilitarianthought (which routinely sanctions duties to aid) arose andhas flourished within this selfsame anglophone tradition.

But the very proximity of utilitarianism, and itsproblematic “serpent windings” (Kant’s term), may in factexplain the hostility, as Murphy notes before proceeding toshow that the tradition is mistaken in refusing to allow (mild)criminal sanctions for failure to perform an easy rescue, butprobably correct in refusing to allow a tort recovery. Themistaken reasons to refuse to recognize a legal duty of easyrescue have come in a variety of forms, most being generaland substantive, others being based on formal or purportedlyconceptual features of criminal law or of tort.

The principal substantive worries arise from theanglophone fixation on liberty, and the thought that a duty ofeasy rescue, like “positive” duties as a genus, is especiallyinvasive of liberty. This worry leads Millians like Joel Feinbergto insist, as a precondition of accepting the duty to rescue assuitable for legal enforcement, upon recasting failures to aidas a species of harmings. Murphy points out that Mill himself

(rightly) thought it unnecessary to deny the patent truth thata duty to aid is a species of duties of beneficence. Enforcingbeneficence is neither paternalistic nor merely moralistic, andso was not intended to come under the harm principle’sstricture, Murphy argues.

Lockeans have had a reservation different from thescruple of the Millians: unlike Millians, Lockeans typicallyassert that liberty is not merely a value but a natural right, andtherefore regard the idea of coercively enforceable positiveduties as contrary to a right to negative liberty (or “freedomfrom interference”). Closer inquiry shows, however, that theLockean (or libertarian) is not so much concerned with libertyas with a preferred schedule of rights. It is simply false,Murphy reminds us, to assert that positive duties are moreburdensome than negative ones–an error that both Millianand Lockean liberals are prone to. Michael Moore’ssuggestion that negative duties remove only one option, whilepositive duties remove all but one, fallaciously fixes on themoment of acting. “A positive duty to take out the garbageonce a week is very different from the positive duty to becomea garbage collector” [636]. Affluent “haves” are likely to assessthe relative burdens of positive and negative duties verydifferently from “have nots.” The negative legal duty not tosteal is “far more restrictive of the liberty of the destitute thanany positive legal duty [to aid that] anyone has seriouslyproposed,” [637] as Murphy says. The Lockean’s attachmentto private property also undercuts her claim to limit libertyonly for the sake of liberty: only a “strongly moralizedconception” [641] of liberty can serve to reconcile propertyrights with the “no limiting liberty except for liberty’s sake”principle.

The real problems with the legal duty of easy rescue werenoticed by Macaulay in his 19th century note to the IndianPenal Code: 1) there is no principled way to distinguish rescuecases from the far wider genus of failures to benefit, and 2)there is a “problem of demands,” as Murphy terms it. Theproblem of demands is that of drawing a line between thedegree of sacrifice required of the bearer of the legal duty,and the degree that the bearer will be excused from making–from (in other words) the legally supererogatory. Murphydismisses the suggestion that artful draftsmanship can handlethe problem of demands: no defense of the legal duty canstand apart from a plausible general account of the moralduty of beneficence–an ambitious undertaking indeed. As tothe first problem, Murphy endorses Peter Unger’s critique ofappeals to the moral significance of proximity, as relied uponby Frances Kamm and Jeremy Waldron, among others, toisolate rescuings within the broader category of aidings.Murphy is similarly unimpressed by efforts along Rawlsian“dualist” lines to erect a firewall between issues of institutionaldesign and matters of personal morality: “If the only way tojustify a legal duty to rescue is by appeal to a fundamentalprinciple that we reject as unacceptably demanding as aprinciple of personal conduct, the case for the legal duty mustfail” [647].

The problem of demands has drawn extensivephilosophical attention over the several decades, as one ofthe two grave weaknesses of consequentialist moral theories(the other being their failure to take seriously “theseparateness of persons,” as Bernard Williams has charged–a problem that Murphy apparently does not find to beseparately implicated in this case). Murphy discusses three


— 131 —

LAW

responses to the problem of demands: 1) stick with an“optimizing” formula and simply accept, a la Unger, thatstringent demands may come; 2) fashion a “nonoptimizing”or “limited” principle of beneficence that draws acceptablelines; or 3) pursue, with Murphy, the idea of “a ‘collectiveprinciple of beneficence,’ according to which people mustsustain only that amount of sacrifice they would be requiredto sustain if everyone were doing their part” [652].

Choosing from even this shortlist is, as Murphy concedes,“an extremely complicated matter” [651] but, happily, thereis an “overlapping consensus” among the three that shouldsettle upon a criminal duty of easy rescue.

Murphy traces how the principles should reach thisconclusion. He first argues that, against an assumedbackground of taxes, transfers, and fire and police protection,a straightforward utilitarian case can be made for confiningthe legal duty to emergency situations, and this is a resultavailable under each of the three formulations of the moralprinciple of beneficence, regarded as an aim of the criminallaw. As for the problem of demands, the challenge for eachof the three principles is to scale its demands so that they arereasonably congruent with the legal duty limited to easy (notstrenuous) rescues: what justification is there for a moral dutythat lets the non-rescuer off the hook as soon as she faces“more-than-minimal” risks? Surely (Murphy assumes) the lawmay not demand more of us than morality does, but moralitydoes tend to become frightfully demanding.

The optimizing principle would justify sacrificing a milliondollar yacht to save a life–a demand that many will see (rightlyor wrongly) as too demanding. Murphy appeals to (but doesnot fully detail) Law-and-Economics scholar Saul Levmore’sargument that an optimal legal rule would combine a smallpenalty with a promise of compensation for the rescuer.Rationalizing the small penalty is made easier by appeal tothe (folk?) distinction between doing and allowing: deterringkillings requires a stiffer penalty because killings (and theirfruits) are largely within the control of the would-be killer. Incontrast, lettings-die call for little deterrence because one wholets another die is little able to bring about the satisfaction ofrage or greed in this fashion (recall James Rachels’ outréexample of the man who lets his cousin drown in the bathtubin order to assure his inheritance). Thus, acceptingoptimization as an aim of the criminal law seems to requireprecisely the legal duty of easy rescue, carrying a modestpenalty, that is found in the criminal codes of manynonanglophone countries and in a few of the United States.True, an optimizing principle of personal morality woulddemand more, but that is just what makes the optimizingprinciple implausible as a general moral principle.

Murphy turns to the nonoptimizing and the collectiveprinciples of beneficence, and must ask (given his rejectionof Rawlsian dualism), what could justify as much as a duty ofeasy rescue as a matter of personal morality–given theproblem of demands? Murphy admits the difficulty of thematter here, especially given that he has (elsewhere) workedout the (perhaps startling) conclusion that the limited andthe collective principles, as applied to personal morality, willnot invariably require even the cheap and easy rescue. This,Murphy admits, “may seem to ruin the case for legal duties torescue” [660], but he quickly adds that a legal provision forcompensation “makes all the difference” and thus “theproblem of demands is, in the end, sidestepped at the level

of institutional design” [660]. The “sidestep” transforms theeasy/cheap/sure-to-succeed but sometimes morally optionalrescue into a morally required, costless rescue–thus achievingan overlapping consensus with the optimizing principle.

The velocity of Murphy’s discussion at this crux is perhapsdisconcerting. Footnotes [660-61, nn. 237-38] inform us thathis investigation has disclosed that, appearancesnotwithstanding, the relevant moral duty is not one to avoidwrongdoing, but to cultivate character of a preferred kind (theeasy-rescuing kind, presumably). It is puzzling how a moralduty to cultivate a disposition “always to perform easyrescues” comports with the absence of any “generalrequirement to perform easy rescues” [660-61 n.238]; and itis equally puzzling how Murphy’s solution is in the endconsistent with his declared rejection of Rawlsian dualism.Murphy concedes that his conclusion on the subject of thecriminal duty “is not irresistible” [661].

So much for the criminal duty. The counterpart tort dutyis not welcome to Murphy, largely because the tort principleof compensatory damages in cases of wrongful death willinevitably tend to impose a significantly greater burden onthe nonfeasant defendant than a mild and antecedently moredefinite criminal-law sanction, especially if that is monetaryonly. Not only is an added tort inducement unneeded if thecriminal-law duty is assumed already to exist, it would havethe dubious effect of encouraging riskier and costlier rescueefforts. Murphy also adverts to, but does not elaborate,structural principles of corrective justice that influentialtheorists such as Ernest Weinrib (and, one might add, ArthurRipstein) have found to be inhospitable to recognizing a tortliability.

In his conclusion, Murphy acknowledges the temptationto make two mistakes. The first mistake would be to infer,from the absurdity of enacting a general legal duty ofbeneficence, that there is no important moral duty ofbeneficence operating outside the special and rare contextof emergency rescue. This mistake is especially perniciousin our legalistic culture and very nonideal circumstances, inwhich states fall far short of guaranteeing the general welfare.The second mistake would be think that the overlappingconsensus among competing principles of beneficence,found in the case of the individual’s duty to rescue, justifiesthe hope that a similar consensus could let us avoid choosingamong the competitors when we turn (as we must) to thequestion of the state’s responsibility for welfare. The state’svastly greater power to do good sets it apart from the case ofthe individual as benefactor and, as Murphy forcefully argues,if there is no good liberty-based argument against positivelegal duties on the individual’s part, there is a fortiori none onthe state’s.


— 132 —

LAW

RECENT BOOKS OF INTEREST

Benson, Peter The Theory of Contract: New Essays.Cambridge University Press, 2001.Burgess, Bonnie B. Fate of the Wild: The Endangered SpeciesAct and the Future of Biodiversity. University of Georgia Press,2001.Cohen, Michael H. Beyond Complementary Medicine: Legaland Ethical Perspectives on Health Care and Human Evolution.University of Michigan Press, 2000.Dershowitz, Alan M. Supreme Injustice: How the High CourtHijacked Election 2000. Oxford University Press, 2000.Dickson, Del The Supreme Court in Conference (1940-1985):The Private Discussions Behind Nearly 300 Supreme CourtDecisions. Oxford University Press, 2001.Fallon, Jr., Richard H. Implementing the Constitution. HarvardUniversity Press, 2001.Fletcher, George P. Our Secret Constitution: How LincolnRedefined American Democracy. Oxford University Press,2001.Kahn, Paul W. Law and Love: The Trials of King Lear. YaleUniversity Press, 2000.Katz, Michael B. The Price of Citizenship: Redefining theAmerican Welfare State. Metropolitan Books, 2001.Marks, Susan The Riddle of All Constitutions: InternationalLaw, Democracy, and the Critique of Ideology. OxfordUniversity Press, 2000.Mayer, Kenneth R. With a Stroke of the Pen: Executive Ordersand Presidential Power. Princeton University Press, 2001.Posner, Richard A. Breaking the Deadlock: The 2000 Election,the Constitution, and the Courts. Princeton University Press,2001.Rhode, Deborah L. In the Interests of Justice: Reforming theLegal Profession. Oxford University Press, 2000.Rubenfeld, Jed Freedom and Time: A Theory of ConstitutionalSelf-Government. Yale University Press, 2001.Zimring, Franklin E., Hawkins, Gordon, & Kamin, SamPunishment and Democracy: Three Strikes and You’re Out inCalifornia. Oxford University Press, 2001.

APA Newsletters€¦ · Theodore Benditt, Editor Spring 2002 Volume 01, Number 2 APA NEWSLETTER ON Philosophy and Law FROM THE EDITOR Theodore Benditt University of Alabama at Birmingham

Documents