Did Making Over the Prisons Require Making Up the Law

Volume 84 Issue 5July 1999

Article 9

Did Making Over the Prisons Require Making Up the Law

Stephen P. Garvey

Part of theLaw Commons

BOOK REVIEW

DID MAKING OVER THE PRISONS REQUIRE MAKING UP THE LAW?

Stephen P. Garveyt

JUDICIAL POLICY MAKING AND THE MODERN STATE: How THE COURTS REFORMED AMERICA'S PRISONS. By Malcolm M. Fleekqt & Edward L. Rubin.ttt Cambridge: Cambridge University Press, 1998. Pp. xv, 490. $75.00.

Do judges make "policy"? If they do, does that jeopardize the rule of law? In Judicial Policy Making and the Modem State: How the Courts Reformed America's Prisons,' Malcolm Feeley and Edward Rubin say "yes" to the first question, and "no" to the second. Judges do make policy because they make legal doctrine, and they do so without threatening the rule of law.

Feeley and Rubin pursue these questions against the backdrop of the prison reform cases, which the authors describe as "the most striking example of judicial policy making in modem America."2 The book is divided into two parts. Part I, subtitled "The Case of Judicial Prison Reform,"3 provides several case studies. Two lengthy studies examine the reform of the Arkansas and Texas prison systems, and three shorter studies examine reform efforts involving the Colorado State Penitentiary, the Santa Clara CountyJails, and the United States Penitentiary at Marion, Illinois.4

Part II, subtitled "The Theory of Judicial Policy Making,"5 provides an original and insightful description of the judicial policymaking process from start to finish, beginning with how judges define the problem for which policy-new legal doctrine-ultimately will be the solution and ending with howjudges implement that solution. My goal here is to examine Feeley and Rubin's account of the process by which judges actually create new legal doctrine and their claim that judicial policy making is consistent with the rule of law.6

I HoLT v. SARVER

Judicial Policy Making describes five different case studies of prison reform, each with its own distinctive contour and character. I focus on the litigation that started it all: Holt v. Sarver.7 The judge who managed the litigation, Chief Judge Henley of the District Court of the Eastern District of Arkansas, is the hero-or villain-of that story.8

Cummins and Tucker Prison Farms, the objects of the Holt litigation, were regimes of private tyranny. Indeed, one easily could have mistaken them for antebellum plantations with guard overseers and inmate slaves.9 "For the ordinary convict," Judge Henley wrote in 1970, "a sentence to the Arkansas Penitentiary... amounts to a banishment from civilized society to a dark and evil world completely alien to the free world, a world that is administered by criminals

under unwritten rules and customs completely foreign to free world culture."'0

The superintendents of the "dark and evil world" inside Cummins and Tucker were inmate "trusty guards."" The trusties essentially ran the prison.'2 Only eight free-world guards were on hand to preserve order in a prison of almost 1000 inmates, and only two of them were on guard at night.13 Segregated by race, inmates slept in open barracks, where "creepers" roamed after sunset, assaulting other inmates and settling scores.'4

Inmates in today's prisons usually earn token wages for their labor.15 The inmates of Cummins and Tucker earned nothing. What little money they had they got by literally selling their blood.'6 Inmates who refused to work or who failed to work up to the trusties' standards were whipped with a leather strap.'7 When the federal courts enjoined the strap's use,'8 the prison used "isolation cells" instead.'9 These eight-by-ten-foot cells, eleven in number, held four men on average, but sometimes many more.20 Inmates in the isolation units lived on "grue," a mixture of "meat, potatoes, vegetables, eggs, oleo, syrup, and seasoning baked all together in a pan and served in four-inch squares."'2' Rehabilitation programs, at least at Cummins, were nonexistent.22

Judge Henley transformed this "dark and evil world" into a modem prison. Using the authority of the Eighth Amendment's ban on

cruel and unusual punishments,23 he was the first to hold that the conditions within a prison as a whole-that a prison itself-violated the Eighth Amendment. "After long and careful consideration," Judge Henley wrote in Holt, "the Court has come to the conclusion that the [Eighth and] Fourteenth Amendment[s] prohibit[ ] confinement under the conditions that have been described."24 He ordered the prison authorities "to make a prompt and reasonable start toward eliminating the conditions that have caused the Court to condemn the system."25

According to Feeley and Rubin, Judge Henley refused to tolerate the gap between his idea of what a prison should be and the prevailing legal doctrine that countenanced prisons like Cummins and Tucker. He bridged that gap using the authority of the Eighth Amendment to create new legal doctrine. He then deployed this doctrine to remake an institution, bringing the grim reality of life inside Cummins and Tucker into line with his own beliefs about what it should be.

According to one long-standing critique, "institutional litigation" of this sort possesses questionable legitimacy because it relies on innovative and extraordinary procedural and remedial mechanisms.26 But these supposedly extraordinary mechanisms actually are quite ordinary.27 Courts have long used similar devices in noninstitutional litigation. Institutional litigation like Holt was extraordinary, not because of the procedures or remedies the courts used, but because of the substantive rights they recognized.28

But perhaps the rights Holt recognized were themselves illegitimate. The Eighth Amendment, you might say, cannot fairly be interpreted to authorize a federal court to engage in the wholesale restructuring of state prisons. Yet this allegation has an obvious, if controversial, reply: properly interpreted, the Eighth Amendment did not simply permitJudge Henley to lead Cummins and Tucker into the civilized world; it positively required him to do so.

Feeley and Rubin offer a different reply. They concede that Judge Henley's remake of Cummins and Tucker was not based on an interpretation of the Eighth Amendment. On the contrary, it was the product of judicial policy making. But-and here's the rub-making policy is, they insist, a perfectly legitimate thing for judges to do, and that goes for Judge Henley.

Is it? Is judicial policy making legitimate? That's the crux of the matter. But before we get to that question, we should examine what, according to Feeley and Rubin, judicial policy making is all about. We should also take a closer look at what Judge Henley actually did in Holt. Was it policy making or interpretation?

II JUDICIAL POLICY MAKING: How JUDGES MAKE NEW LAw

According to Feeley and Rubin, when judges make policy they proceed along much the same path as do other policy makers, whether they are legislators, executives, or members of administrative agencies.29 Applied to judges, the "classic analysis" of the policy-making process proceeds in four stages: (1) defining the problem, (2) identifying the goal, (3) creating doctrine, and (4) implementing the solution.30 The tricky part for judicial policy makers is stage three because judges who "create doctrine" step outside their natural roles as fact finders and interpreters of law and enter a domain where the legitimacy of their action is arguably in doubt.

Feeley and Rubin break this process of doctrine creation into three basic steps: dissonance, integration, and coordination.3' From the perspective of a federal district court judge like Judge Henley, the process goes like this:

Step 1-Dissonance. At this stage judges experience a sense of dissonance between their own beliefs about how the world should be and how it can be under existing law. This dissonance is "probably more likely to occur when there are conflicts within doctrine, thereby weakening its force.'32 Faced with this dissonance, judges basically

have two options: either swallow hard and follow existing law, or try to find some way to bring the law and their beliefs into alignment.33 They are more apt to do the latter "when they have some assurance that the beliefs that motivate them are strongly felt and widely held."34 Step 2-Integration. At this stage the judge's creative juices start flowing. He or she has resolved to eliminate or overcome the dissonance of Step 1 by bringing the law into alignment with his or her own personal beliefs. The question is how to do it.

The key to integration is the notion of a "coordinating idea," which bridges the gap between the law as it is and the law as it ought to be. Coordinating ideas enable a judge to meld his or her own beliefs together with existing doctrine, allowing them to coexist without conflict35 But coordinating ideas do not actually create new doctrine unless and until they successfully pass through Step 3.36

Step 3-Coordination. At this stage the judge sends his or her coordinating idea out into the judicial market, where it competes with the coordinating ideas of otherjudges. Once a majority ofjudges endorse a particular coordinating idea, it becomes part of the institutional fabric of the law. It becomes, in short, new law.37

]

STEP1

Schematically, the process looks like this: Judge's Attitude

ISTE22

3 STEP 3

How did this process play out in Holt? It began of course with Judge Henley, who experienced dissonance between the kind of conditions under which he believed prisoners were entitled to live and the kind of conditions under which prevailing legal doctrine tolerated them living. Consequently, Judge Henley set out to integrate the two. A number of converging legal developments helped his efforts.

The courts once considered prisoners "slaves of the State,"38 and the so-called "'hands-off' doctrine" kept prisons and prison administrators free from the scrutiny of the federal courts.39 Prisoners at the time had no real way to obtain federal-court review of grievances related to the conditions of their confinement.40 Habeas corpus enabled them to challenge the validity of their convictions, but not the conditions of life inside the prison itself,41 and § 1983 was only available if a state official's alleged misconduct was actually unauthorized by state law.42 Indeed, before 1962, no one was certain the Eighth Amendment itself even applied to the states.

Over time, however, things began to change. The hands-off doctrine gradually gave way, and the Supreme Court eventually opened the door for prisoners to challenge state misconduct under § 1983, even if the state official's actions were in fact authorized by state law.43 The Court also removed doubts about the Eighth Amendment's application to the states.44 In short, by the time Judge Henley heard Holt, prisoners were no longer slaves of the state. They were entitled to some protection under the Eighth Amendment, though just how much remained to be seen.

Other doctrinal obstacles also receded. Article I judges had long thought that due respect for federalism meant they shouldn't interfere with how the states ran their prisons. They also thought that due respect for separation of powers principles meant that prison administration was a job for executives and legislatures, not for courts.

But as Feeley and Rubin persuasively show, by 1970 the doctrines of federalism and separation of powers had lost the strength to keep state prisons fully outside the purview of the federal courts.45 Thus, when Judge Henley looked inside the Arkansas penitentiary, the law no longer spoke in an unequivocal voice telling the federal courts to keep their hands off the state prisons. The stage was therefore set for an act of integration.

Integration is an act of judicial creativity.46 It produces and requires a coordinating idea, of which Feeley and Rubin identify four different kinds. At one end of the spectrum are labels, in the middle are analogies and metaphors, and at the opposite end are institutional reconceptualizations. A label is the "mere giving of a name to a vaguely discerned but previously unarticulated legal idea,"47 like the "right of privacy" first christened in Warren and Brandeis's famous article.48 An analogy "transfer[s] one body of law to a new subject area,"49 while a metaphor "create[s] a new image to characterize a situation."50 Finally and most dramatically, an institutional reconceptualization is a "fully realized concept of an entire institution's legal status."51

Each of these kinds of coordinating ideas shares a few defining characteristics. A coordinating idea is: (1) fully realized-it "must be presented as an end state, a definitive right, obligation, qualification, or exception";52 (2) delimited-it must "represent[ ] a delimited change in doctrine, generally involving a single right or obligation";53 (3) directional-it "will generally move in the direction that judges

perceive prior doctrinal innovations as establishing";54 and (4) implementable-the "subject institution" must be able "to understand the coordinating idea,"55 which must also "be capable of being implemented incrementally.'56

The courts achieved integration in the prison reform cases through an institutional reconceptualization-the "idea of a moral, legally justifiable prison."57 This idea emerged fiully formed in the "very first case-the Arkansas litigation-and was simply refined and clarified in subsequent decisions."58 Judge Henley's reconceptualization of the prison in fact relied on two distinct coordinating ideas: rehabilitation and bureaucratization.59 A constitutionally legitimate prison provides inmates with the chance to become rehabilitated and preserves order through bureaucratic rationality.

The final step in the judicial policy-making process, coordination, was one Judge Henley could not take by himself.60 Integration is a one-man act, but coordination is a collective one.61 Faced with the dissonance between their beliefs and the law,62 judges must decide whether to attempt integration or not. If they try to integrate, they release the resulting doctrinal innovation into the judicial public domain, where other judges will either embrace it or reject it. If rejected, the doctrine eventually will wither and die, usually meeting its

demise on appeal. If embraced, it will become part of the -doctrinal corpus,63 and the judiciary as an institution will have created a new doctrine. The courts then will develop and refine this doctrine through the normal processes ofjudicial interpretation.64

III

Holt plainly used the Eighth Amendment in a way no other case had used it before. Holt was also a controversial decision. Yet as Feeley and Rubin themselves suggest, the fact that a judicial decision is dramatic or controversial doesn't necessarily mean it is an example of policy making.65 Interpretations can produce outcomes just as dramatic or controversial as the results of policy making.

According to Feeley and Rubin, the distinction between policy making and interpretation rests basically on (1) what ajudge actually does with the legal material available to him or her66 and (2) what a judge experiences while doing it.67 Holt, claim Feeley and Rubin, is guilty of policy making on both counts.

First, when judges engage in interpretation, they use existing law to guide their decision. When they make policy, however, they use law as a grant ofjurisdiction.68 For example, Feeley and Rubin claim that in Holt the Eighth Amendment functioned not as a source of standards guiding Judge Henley's decision, but as a jurisdictional grant

that invited him to make new law governing the state's prison system.69 Along the same lines, they suggest that Holt was policy making because it transformed the Eighth Amendment from a previously nonjusticiable provision into one that supplied the basis for restructuring a complex social institution7" and because it deployed detailed remedial decrees.71 "The orders that the courts imposed were," they say, "far too specific to have been derived from four little eighteenth-century words."72

Second, judges have one kind of experience when they interpret the law, but they have a different kind of experience when they make policy.73 In fancier terms, the phenomenology of interpretation differs from the phenomenology of policy making. The idea is this: if we were to askJudge Henley what he saw himself doing when he decided Holt, he should-if he were being candid and honest-tell us he was making new doctrine and not simply interpreting the requirements of old doctrine.74

A. What Judge Henley Did

I'm sympathetic to the claim thatJudge Henley's remedial orders were not the product of interpretation. But at the risk of sounding naive,75 I'm not so sure the same is true of the underlying right he tried to secure with those orders-the right to minimally decent conditions of confinement. Applying Feeley and Rubin's criteria, we need to look first at what Judge Henley actually did, which requires

taking a look at where the law, especially the Eighth Amendment, stood circa 1970.

Background-First, the law had generally been moving in a direction broadly in keeping with the decision Judge Henley ultimately reached in Holt. For example, Monroe v. Pape,76 decided nearly a decade before Holt, "gave to individual citizens a viable remedy in the federal courts for deprivations of federally protected rights by persons acting under color of law."77 Similarly, Robinson v. Calfornia,78 decided in 1962, held the states accountable for their actions under the Eighth Amendment.79 Both decisions set the stage for Holt.

Justiciability-Second, none of the standard justiciability doctrines-"the prohibition against advisory opinions, standing, ripeness, mootness, and the political question doctrine"80-kept the Eighth Amendment out of the federal courts altogether. Moreover, the Supreme Court had been applying the Eighth Amendment well before Holt and the other prison-reform cases came along. True, the Eighth Amendment was hardly a major constitutional player in 1970, but neither was it completely moribund. A few constitutional provisions, most notably the Guarantee Clause,81 are famous for being empty vessels.82 But the Eighth Amendment is not empty today, nor was it empty in 1970.

Doctrine-Third, Judge Henley did not decide Holt on a clean doctrinal slate. He relied heavily, for example, on Jackson v. Bishop,83 which the Eighth Circuit decided just a little over a year before Holt. In Jackson, the Eighth Circuit held that prison officials could under no circumstances use corporal punishment to enforce prison discipline.84 Speaking for the court in Jackson, then-Judge Harry Blackmun arrived at this conclusion after a careful review of Eighth Amendment doctrine as it looked on the eve of Holt.

This is what he saw: In Carey v. Settle,85 the Eighth Circuit had pledged to intervene whenever prison discipline "shock[ed] general

CORNELL LAW REVEW

conscience or [became] intolerable in fundamental fairness, and so ... amount[ed] to [the] illegal administration of prison sentence."86 Likewise, in Lee v. Tahash,87 the Eighth Circuit added: "It may be observed... that penal admeasurements made by general conscience and sense of fundamental fairness doubtless will not be without some relationship to the humane concepts and reactions of present-day social climate."88

Judge Blackmun also sought guidance from extant Supreme Court doctrine. For example, the Court's 1910 decision in Weems v. United States89 condemned a sentence of fifteen years at "hard and painful labor" as unconstitutionally disproportionate to the crime of making a "false entry in a public record."90 Moreover, in Trop v. Dulles,91 ChiefJustice Warren, writing for a plurality of the Court's members some forty-eight years after Weems, famously observed that the Eighth Amendment "must draw its meaning from the evolving standards of decency that mark the progress of a maturing society."92 According to the Chief Justice, at the Amendment's core was "nothing less than the dignity of man."93

Judge Blackmun concluded Jackson's survey of the prevailing doc-

trinal landscape:

In summary, then, so far as the Supreme Court cases are concerned, we have a flat recognition that the limits of the Eighth Amendment's proscription are not easily or exactly defined, and we also have clear indications that the applicable standards are flexible, that disproportion, both among punishments and between punishment and crime, is a factor to be considered, and that broad and idealistic concepts of dignity, civilized standards, humanity, and decency are useful and usable.94 According to Feeley and Rubin, the "four little eighteenth-century words" of the Eighth Amendment itself were the only interpretive sourcesJudge Henley had available to consult when he decided Holt.95

]MAKING OVER THE PRISONS

If so, then I can understand how they might think that the Amendment functioned in Holt as little more than a grant ofjurisdiction. By themselves, those words provide little guidance. But when Judge Henley decided Holt, those words did not stand alone. A body of doctrine and precedent had begun to form around those words, including the "broad and idealistic concepts of dignity, civilized standards, humanity, and decency."96

Remedy--Fourth, the reliefJudge Henley ordered in Holt was -at

first anyway-quite unremarkable. Initially, he simply ordered the defendants "to make a prompt and reasonable start toward eliminating the conditions that have caused the Court to condemn the System and to prosecute their efforts with all reasonable diligence to completion as soon as possible."97 Resistance from prison authorities may have forced him to issue more detailed remedial orders later, but that wasn't the judge's fault. Moreover, the underlying rightJudge Henley intended his orders to secure remained the same, no matter how detailed the orders themselves eventually became.

So why can't we say Holt was an interpretation of the Eighth Amendment, albeit a bold one? Some might say Holt fails as an interpretation because it strays too far from the Amendment's text or from the historical intent of its authors. According to Justices Scalia and Thomas, for example, the Eighth Amendment's original meaning shows that the Amendment doesn't apply to prison conditions at all.98 Feeley and Rubin sometimes talk as though they too think the only valid constitutional interpretation is a textual one,99 or at least one that excludes any appeal to "moral norms" or "social principles."100 At other times, however, they seem to take a more generous view of

the resources available to the interpretive task. For example, they speak with apparent approval of-though never openly endorsePhilip Bobbitt's account of constitutional interpretation, which includes "prudential" and "ethical" arguments as well as textual and historical ones.101

But if Feeley and Rubin limit the available modes of constitutional interpretation to text and history, so that the appearance of nontextual and nonhistorical arguments serve as markers for policy making, they need to say why we should accept such a modest view of interpretation. On the other hand, if they don't limit the available modes of constitutional interpretation to text and history, and instead allow recourse to other approaches as well, it becomes correspondingly more difficult to see what's wrong with placing Holt on, or at least near to, the interpretation side of the interpretation/policy-making divide.

Finally, consider this: according to Feeley and Rubin, both the constitutional right of privacy and the prevailing doctrines of First Amendment law owe their existence to judicial policy making.102 We probably could add the constitutional law governing police efforts to secure confessions from criminal suspects, which has given us the famous and familiar Miranda warnings,103 as well as the ornate doctrinal web securing our right to be free from unreasonable searches and seizures.'04 In contrast, Feeley and Rubin describe Brown v. Board of

Education'05 as a case of interpretation, not policy making. "[O]ne must concede,". they say, "that integration of the public schools was at least a plausible reading of the constitutional text rather than an act of policy making by the courts."106 But if that's right, what fairly distinguishes Brown from Holt or from any of the other doctrinal innovations Feeley and Rubin treat as policy making?

B. What Judge Henley Thought

So much for whatJudge Henley did when he decided Holt. What about the second basis on which Feeley and Rubin rest the distinction between policy making and interpretation-judicial phenomenology? Whatever he did in Holt, what did Judge Henley think he was doing? According to Feeley and Rubin, he must have thought he was doing something other than interpretation. Again, I'm not so certain.

Of course, no one knows what Judge Henley actually thought he was doing because (so far as I know) he's never told anyone. Still, we might try to draw some inferences from his opinion itself. For example, if Judge Henley had shifted gears from interpretation to policy making, one might expect to see the shift reflected somewhere in his opinion. But if the gears are shifted, the passenger hardly notices. Indeed, Holt reads to me more like the account of a mind engaged in a complex and controversial interpretation of an assortment of legal materials, with the process eventually leading to the creation or identification of a new and controversial constitutional doctrine.

Admittedly, one could say that Judge Henley was just acting strategically, not telling us that he was making policy or that he managed to kid himself into thinking he really was just interpreting the Eighth Amendment. For example, Feeley and Rubin say that rather than admit they were making policy, the judges in the prison cases "[m] ore often.., resorted to the usual argument that they were simply interpreting the Constitution, most specifically the Eighth Amendment's prohibition against 'cruel and unusual punishment.107

At the same time, however, Feeley and Rubin reject accounts that portray judges as unreflective or dishonest. They refuse, for example, to ascribe judges' "general unwillingness to state that they are engaged in policy making... to cynicism or duplicity."'08 It would in fact be difficult for them to assert otherwise. Remember that their theory of judicial policy making relies in part on the phenomenology

ofjudging.1°9 This phenomenological approach allows them to reject public-choice theories of adjudication, which claim that judges really act to advance their own self-interest,'10 as well as critical theories of adjudication, which claim that judges act to advance the interests of their class, race, or gender."1' These theories, Feeley and Rubin argue, are untrue to what judges actually experience when they adjudicate.".2

Yet if phenomenology counts, and if judicial opinions give some indication of that phenomenology, then why not give more credit to Judge Henley's opinion in Holt? Why not take Holt at face value? Feeley and Rubin can't rely on judicial phenomenology to dismiss the claims of public-choice and radical theorists and then dismiss judicial phenomenology when it jeopardizes their own theory, unless they tell us what the difference is. They must take the bitter with the sweet.

Perhaps, as Feeley and Rubin suggest, judges like Judge Henley don't think they are making policy (even when they are) because they "lack [the] concepts or vocabulary with which to describe the policymaking process.""13 Presumably the idea here is that once judges have the right vocabulary, they will experience a conceptual epiphany. After reading Judicial Policy Making, Judge Henley might say: "Oh... Now I see. What I was really doing was making policy. How could I have been so blind?" But I suspect he might say something more like the following: 'Well, I admit I was breaking new ground, but I was trying all along to be faithful to my best understanding of what the Eighth Amendment required."

Much of what I've said to this point may appear pedantic. After all, Feeley and Rubin stress that the boundary between policy making and interpretation, however clear in theory, is fuzzy in practice. Sometimes, perhaps often, policy making and interpretation will exist

side by side in the same opinion.114 Consequently, it should be relatively easy to show that a specific opinion partakes of both modes of adjudication. But that leads to the question: what difference does it make what we call the process by which Judge Henley reached his decision in Holt?

The way we characterize that process does of course make a difference if, like Feeley and Rubin, we want to shed special light on a particular form of judicial decision making. Indeed, despite my misgivings, I certainly could be convinced, as Feeley and Rubin urge, that "policy making" is really a more useful term for describing what was going on in Holt than is "interpretation."

On the other hand, I think it probably makes little difference to judicial practice. Analytic accounts of a practice typically aren't supposed to reform that practice. Judges have done and probably will continue to do what Judge Henley did in Holt, whether we choose to call it policy making or interpretation. Of course, I suppose the label we use could cause judicial practice to change if it caused judges to become more self-aware of the rules they already implicitly follow when they make policy. Self-awareness may in fact lead judges to make better policy. Indeed, Feeley and Rubin suggest that 'Judicial policy making... can be improved only if it is acknowledged.""5

Still, if the ultimate question is the legitimacy of whatJudge Henley did, then I have a hard time seeing what turns on the label-policy making or interpretation-we use. I agree that labels can carry normative baggage that often goes undetected. For instance, we generally think interpretation is prima facie legitimate, while policy making is prima facie illegitimate. But the underlying question of legitimacy remains, no matter what the label is.

IV WAs JOLLT LEGIrIMATE?

Judicial policy making, say Feeley and Rubin, is perfectly legitimate because it is perfectly compatible with the rule of law. But, they claim, in order to appreciate why that's so, one must first reconceptualize the rule of law itself.

According to Feeley and Rubin, most of us think the rule of law demands that judges be subject to external constraints that are "gen-

eral, clear, well-accepted and congruent with the legal order."116 Consequently, when judges make policy they "are operating without constraint to produce a body of variable, unpredictable, and personally motivated orders."117 The political branches of government are free to make law because their legitimacy flows not from fidelity to pre-existing law but from electoral accountability. Judges don't enjoy that luxury. Thus, according to the conventional view, judges who make policy violate the rule of law because judicial policy making is not subject to the requisite "external constraints."

Feeley and Rubin insist that this line of argument is mistaken because it erroneously relies on an "interpretation-oriented"118 conception of the rule of law, according to which legitimacy is a matter of fidelity to a pre-existing text. In a modern administrative state, however, this "premodern"'19 conception of the rule of law is of "questionable relevance"120 because texts constrain very little of the law making that goes on in the modern state. But, suggest Feeley and Rubin, text is not the only source of constraint on judicial policy making and thus is not the only way to secure the rule of law.

Feeley and Rubin claim that judges are constrained when they make policy, but not by text. Instead, the process through which they make policy and the institutional nature of the judiciary as a whole constrain them. As Feeley and Rubin put it, the constraint on judicial policy making comes from the "internal dynamics of the coordination process."'21 Judges can make new doctrine only if they collectively succeed in coordination, and they can collectively succeed in coordination only if the coordinating ideas they individually develop "meet[ ] the rather demanding conditions of realization, delimitation, and directionality."'122 In its modern form, the rule of law is a matter of process and institutional structure, not of fidelity to a pre-existing text.123

Finally, once we realize that "oppression can be contained by supervision and culture, instead of formal mechanisms,"'24 our worries about judicial policy making should dissipate. Indeed, according to Feeley and Rubin, judicial policy making has been constrained like this from the start, only we didn't realize it because we were looking for the relevant source of constraint in the wrong place. We were looking for it in one text or another, but of course it was never there. Equipped with our new understanding of the rule of law, we can at last see why judicial policy making poses no real threat to the rule of law.

In essence, Feeley and Rubin urge us to look from afar at what Judge Henley did. We need, they say, to step back and putJudge Henley's decision in its larger institutional context. From this perspective, we can see that Judge Henley could get away only with so much. He could go only so far in translating his own beliefs into law. His reconceptualization of the prison could win the acceptance of his fellow judges and become new doctrine only if it satisfied the constraints of the policy-making process. If Judge Henley's reconceptualization failed to carry the day, his effort to make new doctrine would have been stillborn.

Fair enough. But now consider Holt not from afar, but up close and personal. Consider it from the perspective of Warden Sarver, who was on the receiving end ofJudge Henley's orders.'25 The rest of us might sleep easier knowing that the judicial policy-making process set limits on Judge Henley and others like him, but Warden Sarver has a different concern. For him, the real question is not so much about constraint as it is about authority. Warden Sarver might well acknowledge that judges generally need the endorsement or acquiescence of other judges in order to successfully create new doctrine, but what he really wants to know is by what right-by what authority-did Judge Henley create the doctrine that emerged from Holt?

To answer that question, we need to better specify what the rule of law means. First, it can mean thatjudges are limited or constrained in just how far they can realistically impose their will on the rest of us. Call this a "power-based" conception of the rule of law. Second, it can mean that for whatever reason judicial decision making is generally predictable enough to allow us to arrange our lives and affairs to avoid

the exercise ofjudicially directed state coercion. Call this a "predictability-based" conception of the rule of law.126 Finally, it can meanand here's Warden Sarver's worry-that any judicially authorized exercise of state coercion is justified by law. Call this a 'Justificationbased" conception of the rule of law.127

Feeley and Rubin explain why judicial policy making can coexist with the first two conceptions of the rule of law, but not with the third. Judicial policy making, they explain, doesn't violate the power-based conception because a single judge can in fact go only so far in translating his or her will into law. The demands of the coordination process and the institutional nature of the judiciary in particular impose practical limits on judicial willfulness. Nor does judicial policy making violate the predictability-based conception of the rule of law because the judicial policy-making process generally follows a relatively well-defined path, which Feeley and Rubin have done so much to illuminate. But Feeley and Rubin don't really explain how, if at all, judicial policy making can coexist with the justification-based conception of the rule of law.

The question of justification becomes most urgent when the judges making the policy are members of the jurisdiction's highest court, including the Justices of the United States Supreme Court. The coordination process and judicial hierarchy limit lower court judges, but not members of the top courts. Yet if the primary limitation on judicial policy making derives from the coordination process itself, and if the policy making of the highest courts bypasses that process, then policy making from the top down "revive us] the doubts about whether doctrine creation violates the rule of law."'28 Think, Feeley and Rubin remind us, of Roe v. Wade.129

To quiet these doubts, Feeley and Rubin turn to Alexander Bickel.130 Bickel famously urged the Supreme Court's members to make liberal use of the "passive virtues" as a way to shepherd the Court's delicate legitimacy. According to Bickel, the free exercise of the passive virtues enables the Court to avoid a decision on the merits of controversial cases, including cases that otherwise would call on it to make policy.'3' Bickel's praise for passivity attracted a variety of

criticisms. Feeley and Rubin emphasize two. First, passivity would turn the Court into a "craven, poll-watching, dispute-dodging political actor." Second, it would "sanction[ ] insincere subterfuges."132

Yet however persuasive these objections to Bickel's passive virtues might be when lodged against controversial cases that can only be decided through interpretation, they lose much of their punch, claim Feeley and Rubin, when lodged against cases calling for policy making. A strategy of taking such cases only after they have "percolated" in the lower courts is, they claim, statesman-like and commonsensical, not craven.'33 Likewise, refusing to decide a case for which the legal doctrine needed to reach a resolution doesn't yet exist hardly amounts to an insincere subterfuge.34

But Judge Henley was not on the Supreme Court, and I don't think Feeley and Rubin would encourage district court judges to be aggressive practitioners of the passive virtues. So we still don't have an answer for Warden Sarver. We can call Judge Henley's decision in Holt an example of policy making, or we can call it an example of interpretation. Either way, Warden Sarver will still want to know: was Judge Henley's order-an order backed by state coercion-justified by law?

Some scholars, like Ronald Dworkin, probably would say it was, though Dworkin probably would insist also thatJudge Henley was simply interpreting the Eighth Amendment and not making policy.'35 According to the conventional (some might say caricatured) understanding of Dworkin's well-known theory of adjudication,136 every legal question in principle has one right answer.137 Dworkin's model of adjudication is of course an ideal: real-life judges lack the wherewithal to pull off what it requires of them. Only 'Judge Hercules," Dworkin's imaginary ideal judge, has all that it demands.'38 But if Judge Henley managed to pull off this Herculean task in Holt, then his decision was in fact justified by law.

At the opposite end of the spectrum are critical scholars who think law is radically indeterminate. Mature legal systems, they say, have a variety of features that make it implausible to argue, as does Dworkin, that all legal questions have one right answer.39 On the contrary, they claim no legal question has one right answer. Even "easy" questions only look easy because judges and lawyers oftentimes just happen to agree on what the answer is. But any such agreement is based on the common ideology they share, not on the law.'40 The rule of law, they urge, should be exposed for the fiction it is. We may applaud whatJudge Henley did in Holt, but the law did not justify his doing it. Indeed, the law itself never justifies any outcome.

Feeley and Rubin are, I think, somewhere in the middle, and rightly so. They think neither that law always justifies an outcome, nor that it never does. In other words, adjudication, including judicial policy making, is neither always fully determinate, as Dworkin thinks, nor always radically indeterminate, as critical scholars think. It is instead moderately indeterminate, sometimes providing an answer and sometimes providing only more or less of an answer. Moreover, according to Feeley and Rubin, once we recognize the existence of the rules that govern the policy-making process, judicial outcomes that looked indeterminate are really less indeterminate than we might originally have they're still somewhat 41 indeterminate.'

thought,

though

So, returning to Warden Sarver's question: was Judge Henley's decision in Holt, however indeterminate, nonetheless justified by law? I think it was.

We can say that the answer to the legal question presented in Holt was legally indeterminate if the legal reasons Judge Henley offered to support the answer he gave did not uniquely determine his answer. Moreover, the available legal reasons include legal sources (e.g., legal rules and principles), interpretive operations that judges legitimately can perform on those sources in order to generate further legal sources, and rational operations that judges can perform on the legal sources and on the applicable facts to generate particular outcomes in particular cases.142

Furthermore, a particular legal source or interpretive operation is part of the legal wherewithal of a particular legal system if it falls within the scope of what H.L.A. Hart famously called the "rule of recognition."143 As Hart described it, the rule of recognition "is in effect a form of judicial customary rule" that owes its existence to the social fact that "it is accepted and practised in the law-identifying and lawapplying operations of the courts."144 In other words, the rule of recognition enables judges to identify or recognize the legal reasonsthe legal sources and interpretive operations-on which they legitimately can rely to r~solve a particular case or controversy.

Now, how does all this relate to Feeley and Rubin's account of judicial policy making? As I see it, the heretofore hidden rules that constitute and regulate the judicial policy-making process are among the interpretive operations judges in our legal system actually use to decide cases. Moreover, ajudge's reliance on these policy-making, interpretive operations is valid or legitimate because those operations derive their legal status and validity from the rule of recognition.145 If so, then Feeley and Rubin's account of judicial policy making really amounts to a partial account of the content of our rule of recognition. The judicial policy-making process is therefore legitimate because "it is accepted and practised in the law-identifying and law-applying operations of the courts.'46 Consequently, insofar as the outcome in Holt was the outcome of policy making, Judge Henley could legitimately embark on that process because the rule of recognition authorized him to do so.

Still, the result of that process no doubt remained indeterminate. The legal reasons available to Judge Henley, including the interpretive operations that constitute the policy-making process, probably did not uniquely determine the outcome in Holt. Moreover, you might think that this residual indeterminacy casts doubt on the legitimacy of Judge Henley's action. If the legitimate exercise of state coercion required that a judicial rule or outcome be uniquely justified, then you would probably be right.

But here I'm inclined to agree with Jules Coleman and Brian Leiter, who argue that the legitimate exercise of state coercion re-

quires only that a judicial rule or outcome be justified, not that it be uniquely justified. As they put it: "The problem with coercion is its use to enforce outcomes that are notjustified; it is not that coercion is being employed to enforce justified outcomes that happen not to be uniquely warranted."147 Consequently, if the outcome Judge Henley reached in Holt was a justifiable exercise of state coercion-and I think it was, even if it was not uniquely justified-then Warden Sarver cannot fairly complain that he was subject to the illegitimate exercise of state power.

So my reply to Warden Sarver would be something like this: Judge Henley's decision in Holt was a justifiable and therefore legitimate exercise of state coercion against him because (1) the process of judicial policy making (if that's what we want to call it) derives its validity from the rule of recognition and (2) the policy (if that's what you want to call it) that Judge Henley articulated in Holt was a legally justified policy, even if not uniquely justified.

Of course, Warden Sarver still might be unhappy. He might concede that the outcome in Holt was ajustified and legitimate exercise of state coercion, but he might nonetheless complain that, however legitimate, it was nonetheless an unwise or undesirable exercise of state coercion. Indeed, even ifjudicial policy making, as Feeley and Rubin suggest, is hard-wired into the modem administrative state,148 we can still ask if we'd be better off in the long run if judges lacked the authority to make policy or if their authority to make policy was more limited.149 In other words, we can always ask if we'd be better off with a different rule of recognition-one that placed judicial policy making within certain limits or out of bounds altogether.

Feeley and Rubin spend relatively little time addressing this question. Ending the state-sanctioned abuse of inmates and reducing the risk of inmate-on-inmate violence were of course all for the good, but any more global assessment of the desirability of the prison-reform cases, they fairly say, would require developing a full-blown normative account of the prison's role in modem penal policy.150 Having exhaustively described the process of judicial policy making and having tried to defend its legitimacy, Feeley and Rubin can certainly be ex-

cused for not having launched into a full-scale defense of its desirability. Still, some general thoughts on the circumstances under which judicial policy making is a desirable mode ofjudicial decision making would have been helpful.151

Feeley and Rubin have done an exceptional job of illuminating the process ofjudicial policy making. Their account of that process is complex and sophisticated, taking us well beyond prior accounts, which typically rest content to say relatively unhelpful things like "policy making requires judgment" or "it requires pragmatism."152 For that accomplishment, we owe them a considerable debt.

Likewise, they do a convincing job of describing the various constraints and limits under which judicial policy making operates. But their defense of judicial policy making's legitimacy would have been more complete if they also would have explained why judicial policy making and the doctrines that result from it are justified by law. I've tried to sketch one such explanation above, but if that explanation fails, then Feeley and Rubin need to supply us-and Warden Sarverwith another.

--- FOOTNOTES ---

INTRODUCTION

t Associate Professor of Law, Cornell Law School. I thank Gary Simson and Theo-

dore Eisenberg for helpful comments on earlier versions of this review.

I- Claire Sanders Clements Dean's Professor of Political Science and Law, University

of California at Berkeley, Boalt Hall School of Law.

j-ft Professor of Law, University of California at Berkeley, Boalt Hall School of Law. Professor Rubin has subsequently joined the faculty at the University of Pennsylvania Law School.

I MALcoLM M. FEELEY & EDWARD L. RUBINJUDICIAL POLICY MAKING AND THE MODERN

STATE: How THE COURTS REFORMED AMERICA'S PRISONS (1998).

Id at 13; see id. at 336 (describing "prison reform cases[] as the high-water mark of

judicial policy making").

3 Id. at 27. 4 See i. at 51-145.

Id. at 145.

6 Much of Chapter 6 builds on Edward Rubin & Malcolm Feeley, Creating Legal Doc-

trine, 69 S. CAL. L. REv. 1989 (1996).

7 309 F. Supp. 862 (E.D. Ark. 1970). This case is generally known as Holt II, but I will refer to it in the text simply as Holt. Holt I involved an earlier challenge to specific conditions inside Cummins and Tucker, see Holt v. Sarver, 300 F. Supp. 825 (E.D. Ark. 1969), whereas Holt H involved a challenge to the conditions of confinement with each of the prison farms as a whole. See Holt, 309 F. Supp. at 364. Of course the litigation surrounding the Arkansas Penitentiary did not begin with Holt I, nor did it end with Holt I. For earlier proceedings, see Courtney v. Bishop, 409 F.2d 1185 (8th Cir. 1969); Jackson v. Bishop, 404 F.2d 571 (8th Cir. 1968), vacating 268 F. Supp. 804 (E.D. Ark. 1967); and Talley v. Stephens, 247 F. Supp. 683 (E.D. Ark. 1965). For later proceedings, see Finney v. Mabr, 458 F. Supp. 720 (E.D. Ark. 1978); Finney v. Mabry, 455 F. Supp. 756 (E.D. Ark. 1978); and Finney v. Hutto, 410 F. Supp. 251 (E.D. Ark. 1976), affd, 548 F.2d 740 (8th Cir. 1977), affd, 437 U.S. 678 (1978).

My decision to focus on a single case and the decision-making process of a single district courtjudge doesn't square well with Feeley and Rubin's emphasis on judicial policy making as an institutional process, but I hope my narrower focus does not lead me to misconstrue their theory.

8 Cf. FEE=& RuBan, supra note 1, at 226 ("Integration of the developing doctrine of

civil rights with this attitude of discomfort and dismay about the conditions in state prisons was first achieved by Judge Henley in the Arkansas litigation."). Judge Henley elevated to the Eighth Circuit during the course of the Arkansas litigation.

9 Cummins Farm was originally a plantation. See id. at 52. For a thoughtful discussion of the "plantation model" upon which many postbellum southern prisons were built, see id. at 150-58.

10 Holt 309 F. Supp. at 381.

Id. at 373-76. See id. See id at 373.

"1 12 13 14 Id. at 376 (internal quotation marks omitted). 15 Cf GAIL S. FUNKE ET AL., ASSETS AND LL&BiUTnEs OF CORRECiIONAL INDUSrREs 15

tbl.2-2 (1982) (listing wage rates for prisoner labor in several states as of 1972).

16 17

See Holt, 309 F. Supp. at 371. Inmates were also tortured with the infamous "Tucker telephone." FEELEY & RUBIN, supra note 1, at 56 (internal quotation marks omitted). As Feeley and Rubin describe it:

The "Tucker telephone" was a device that generated electricity by means of a hand crank. Electrodes from this device were attached to the extremities of the prisoners, including their genitals, and then cranked by guards. The resulting shocks, apart from being excruciatingly painful, could bum their bodies and cause seizures and death. Prominently placed in a small building (the "telephone booth") at Tucker Farm, the "telephone" was a standard punishment for all sorts of infractions.

Id. at 56 n.*.

18 SeeJackson v. Bishop, 404 F.2d 571, 581 (8th Cir. 1968) (Blackmun, J.) (ordering the district court below to enter a decree "restraining the Superintendent of the Arkansas State Penitentiary and all personnel of the penitentiary system from inflicting corporal punishment, including the use of the strap, as a disciplinary measure").

See id. at 59 n.*.

19 FEELEY & RUBIN, supra note 1, at 59. 20 21 Holt v. Sarver, 300 F. Supp. 825, 832 (E.D. Ark. 1969). 22

See Holt v. Sarver, 309 F. Supp. 362, 378 (E.D. Ark. 1970).

See U.S. CONST. amend. VIII ("Excessive bail shall not be required, nor excessive

fines imposed, nor cruel and unusual punishments inflicted.").

24 Holt, 309 F. Supp. at 381.

Id. at 383. See, e.g., Abram Chayes, The Role of the Judge in Public Law Litigation, 89 HAxv. L. REv. 1281, 1289 (1976) ("[This] new model ofjudicial action and the judicial role... depart[s] sharply from received conceptions."); William A. Fletcher, The Discretionay Constitution: Institutional Remedies andJudicial Legitimacy, 91 YALE L.J. 635, 637 (1982) (arguing "that since trial court remedial discretion in institutional suits is inevitably political in nature, it must be regarded as presumptively illegitimate" unless "the political bodies that should ordinarily exercise such discretion are seriously and chronically in default"). But cf James A. Henderson,Jr., Comment: Settlement Class Actions and the Limits ofAdjudication, 80 Coauu_. L. Ray. 1014, 1017 (1995) ("Although powerful arguments can be made that public law litigation is beyond the legitimate limits of adjudication, several factors combine to justify, at least arguably, judicial intervention in such cases." (footnote omitted .

See Theodore Eisenberg & Stephen C. Yeazell, The Ordinaiy and the Extraordinaty in Institutional Litigation, 93 HAtv. L. REv. 465, 474 (1980) ("In fact, one finds 'extraordinary' elements in the old litigation and discovers that the new litigation presents relatively little that is genuinely novel.").

See id. at 516 ("We do not deny that courts [involved in institutional litigation] have been up to new things, but we conclude that the novelty flows from the new rights created rather than from the remedies employed.").

29 30 31 32

See FEELEY & RUBIN, supra note 1, at 146. Id. at 149, 204. See id. at 221-33. Id. at 223.

4- Gaproduces felt need --

E

to integrate or acquiesce

xisting Doctrine

Dissonance

Integration

Creation of Coordinating Idea

Coordinating Idea Enters Judicial Public Domain

See id. at 223-24. A third option would be to "ignore the doctrine and make a [but b]ehavior of this

decision that is governed entirely by [the judge's] attitudes ... nature is extremely rare, given the power of role expectations." Id. at 223.

35 36

Id. at 219. See id. at 223-24. See id. at 226 ("While the interaction of doctrine, attitude, and integration occurs within a single judge's mind, a new doctrine, however conceptual its character, is not an individual idea.").

See id. at 226-33 (discussing the coordination process).

38 Ruffin v. Commonwealth, 62 Va. (21 Gratt) 790, 796 (1871). 39 Note, Beyond the Ken of the Courts: A Critique ofJudicial Refusal to Review the Complaints

of Convicts, 72 YALE LJ. 506, 506 (1963).

40 41

See id. See id. at 510 (noting as a major limitation to habeas corpus "the restriction that the writ is only available to contest the legitimacy of one's confinement and is not available to test the legitimacy of the mode or manner of confinement").

See id. at 512 ("It is well established that unjustified violence by a state official amounting to denial of due process of law constitutes grounds for action under [this) provision( ]." (emphasis added .

See Monroe v. Pape, 365 U.S. 167, 172 (1961) ("[I]n enacting [§ 1983, Congress) meant to give a remedy to parties deprived of constitutional rights, privileges and immunities by an official's abuse of his position.").

See Robinson v. California, 370 U.S. 660, 667 (1962) (applying the Eighth Amendment to the states through the Fourteenth Amendment and holding that "a state law which imprisons a person [guilty only of narcotics addiction] ... inflicts a cruel and unusual punishment in violation of the Fourteenth Amendment").

45 Feeley and Rubin discuss how the rise of the administrative state transformed the doctrines of federalism and separation of powers in Chapter 5 (federalism), see FEELEY & RUBIN, supra note 1, at 149, and Chapter 7 (separation of powers), see id. at 297. Their arguments in those chapters are important contributions that deserve much more attention than the passing remarks I make here. For an earlier discussion of federalism and separation of powers that reaches much the same conclusion, see Eisenberg & Yeazell, supra note 27, at 495-510. As Eisenberg and Yeazell put it, "nothing in the twistings of [separation of powers or federalism] doctrines create[d] any fatal obstacles to [institutional] litigation." Id. at 495.

See FEELEY & RUBIN, supra note 1, at 222 ("Integration is the starting point for a

theory of creativity within an institution, and specifically for a theory of judicial creativity.").

Id. at 238. See Samuel D. Warren & Louis D. Brandeis, The Right to Privacy, 4 HAgv. L. REv. 193

47 48 (1890).

49 FEELEY & RUBIN, supra note 1, at 239; see also id. at 237 ("When using an analogy, one takes a complete and established approach to one subject and applies it, in its entirety, to another.").

50 M at 239; see also id. at 237 ("IWhen using a metaphor, one takes a complete image and applies it to a situation as means of conceptualizing that situation in its entirety.").

Id. at 238. Id. at 233. Id. at 234.

Id. at 236.

54 55 1& at 259. Feeley and Rubin call this requirement "comprehensibility." Id. 56

Id, at 262. Feeley and Rubin call this requirement "incrementalism." Id. Comprehensibility and incrementalism, both components of implementability, are especially important when the coordinating idea is an institutional reconceptualization. See id. at 355 ("When the coordinating idea consists of an institutional conceptualization, as it did in the prison cases, there is an additional set of constraints involving the implementation process.").

57 Id. at 239. 58 59 See id. at 252-58 (discussing rehabilitation as a coordinating idea); id. at 271-75

Id.

(discussing bureaucratization as a coordinating idea).

I confess that I'm not sure I fully understand why Feeley and Rubin identify rehabilitation and bureaucratization as the two reconceptualizations that animated the prison reform cases. For a variety of reasons, both of these ideas have limited normative and rhetorical appeal. Instead of bureaucratization, I find more appealing the idea that prison life should be subject to the minimal procedural requirements of the "rule of law" or of "due process." Likewise, I find the idea that inmates are entitled to decent conditions of confinement more appealing than the idea that they are entitled to rehabilitation.

See id. at 226 ("While the interaction of doctrine, attitude, and integration occurs within a single judge's mind, a new doctrine, however conceptual its character, is not an individual idea.").

61 The prison cases involved horizontal coordination. Indeed, according to Feeley and Rubin, the "prison reform cases are an extreme example of horizontal coordination, and acquire additional interest thereby." Id. at 232. Coordination can also work vertically, when the supreme court of ajurisdiction steps in and either affirms or preempts an ongoing process of coordination in the lower courts. See id. at 229-30.

62 Federal judges throughout the South were experiencing the same kind of disso-

nance Judge Henley experienced. See id. at 220-21.

WAS HoLT PoLIcy MAKING OR INTERPRETATION?

See id. at 227 ("The idea that will become the new legal .doctrine is one that ulti-

mately prevails as a means of integration for the majority ofjudges.").

See id. at 209 ("Once created, the doctrine can certainly be elaborated by the more

familiar process of interpretation ....").

65 66 Feeley and Rubin put it thus:

See id at 12-13 ("[N]ot all activist decisions involve policy making.").

The difference between policy making and interpretation is generally apparent from examining the rationale that appears in the opinion and the result the judge has reached. When the judge is interpreting a legal text, the opinion will be replete with textual references, and will attempt to link those references to the result by linguistic analyses, historical accounts of meaning, more general analysis of structure, purpose, or the drafters' intent, and citations of prior decisions that relied on these interpretive techniques. When the judge is making public policy, such references will be absent, and in their place will be discussions of moral norms, social principles, nonlegal sources, nonauthoritative legal texts, and citations of prior decisions that feature such discussions.

Id.

See infra notes 107-13 and accompanying text. See FEEL=" & RUBIN, supra note 1, at 5 ("[Piolicy making is distinguished from interpretation because it treats the text as a source ofjurisdiction, not a guide to decision."); i& at 215 ("]udges fashion legal doctrine when they decide, for one reason or another, that existing texts provide them only with a grant ofjurisdiction, and not with any particularized guidance for the case at hand."); id. at 337 (noting that after identifying the relevant subject area through interpretation, judges engaged in policy making. "pay no farther attention to the text, generally because the text conveys no useful information").

See id. at 14 ("The Eighth Amendment was relevant to the prison conditions cases, however-not as a source of standards, but as a basis for judicial jurisdiction."); id. at 206 ("The broad language of the cruel and unusual punishment clause must be seen as a grant of jurisdiction, a mandate that courts should somehow concern themselves with prisons because prisons are a form of punishment.").

See id. at 14 ("[T]he claim that one can find so elaborate a set of standards in a previously nonjusticiable provision makes hash of any coherent theory of interpretation."). See id. ("The orders that the courts imposed were far too specific to have been derived from four little eighteenth-century words, no matter how carefully those words were read or how long their meaning pondered.").

72 73

Id. at 14. See id. at 7-8 ("Interpretation and policy making are different experiences for the

judge and are perceived differently by others ....

74 Cf id. at 338 ("When [judges] sit down to write opinions like Roe v. Wade, Griswold v. Connecticut Hennigsen v. Bloomfield Motors, or the prison reform cases, they must have the subjective sense that they are doing something different from the interpretation of a text." (footnotes omitted .

75 Cf id. at 207 ("No thoughtful observer-indeed, no thoughtless observer eitherhas ever been persuaded that federal judges were simply interpreting the Eighth Amendment in the prison cases." ); id. at 209 ("The judges who decided the prison cases never persuaded anyone that they were deriving their standards from the Eighth Amendment ... ."); id. at 338 (noting that no "educated observer" would be "fooled" into thinking the prison reform cases (among others) were based only on an "interpretation of a text").

365 U.S. 167 (1961).

76 77 Holt v. Sarver, 309 F. Supp. 362, 367 (E.D. Ark. 1970) (citing Monroe v. Pape, 365

U.S. 167 (1961 .

78 370 U.S. 660 (1962). 79 See id. at 666-67. 80 81

ERWIN CHEMEmRNsy, FEDERALJURISDICr1ON § 2.1, at 42 (2d ed. 1994). See U.S. CONST. art. IV, § 4 ("The United States shall guarantee to every State in this

Union a Republican Form of Government. .. ").

82 Cf. Erwin Chemerinsky, Cases Under the Guarantee Clause Should BeJustidable, 65 U. CoLo. L. REv. 849, 852 (1994) ("[I] t is time for the Guarantee Clause to be resurrected and given a meaningful role in contemporary constitutional law.").

404 F.2d 571 (8th Cir. 1968). See id. at 579 ("[W]e have no difficulty in reaching the conclusion that the use of the strap in the penitentiaries of Arkansas is punishment which, in this last third of the 20th century, runs afoul of the Eighth Amendment.

83 84

..

351 F.2d 483 (8th Cir. 1965).

356 U.S. 86 (1958) (plurality opinion). Id. at 102, quoted in Jackson, 404 F.2d at 579. Id. at 100, quoted in Jackson, 404 F.2d at 579.

Id. at 485, quoted in Jackson, 404 F.2d at 578. 352 F.2d 970 (8th Cir. 1965). Id. at 972, quoted inJacksom, 404 F.2d at 578. 217 U.S. 349 (1910).

86 87 88 89 90 Jackson, 404 F.2d at 578 (citing Weems, 217 U.S. at 373). 91 92 93 94 Jackson, 404 F.2d at 579. The Supreme Court later quoted Jackson with approval in its first case dealing with the relationship between the Eighth Amendment and conditions of confinement. See Hutto v. Finney, 437 U.S. 678, 685 (1978) ("The Eighth Amendment's ban on inflicting cruel and unusual punishments.. .prohibits penalties... that transgress today's '"broad and idealistic concepts of dignity, civilized standards, humanity, and decency."'" (quoting Estelle v. Gamble, 429 U.S. 97, 102 (1976) (quotingJacksn, 404 F.2d at 579 . 95

FEELEY & RUBIN, supra note 1, at 14.

96 Jackson, 404 F.2d at 579. 97 Holt; 309 F. Supp. at 383. 98

See Helling v. McKinney, 509 U.S. 25, 40 (1993) (Thomas,J., dissenting) ("I believe that the original meaning of 'punishment,' the silence in the historical record, and the 185 years of uniform precedent shift the burden of persuasion to those who would apply the Eighth Amendment to prison conditions. In my view, that burden has not yet been discharged."); see also Harmelin v. Michigan, 501 U.S. 957, 962-65 (1991) (opinion of Scalia, J., joined by Rehnquist, C.J.) (relying on historical analysis to conclude that the "Eighth Amendment contains no proportionality guarantee"). But see FEELEY & RUBIN, supra note 1, at 207 (criticizing Justice Thomas's historical analysis).

99 See, e.g., FEELEY & RUBIN, supra note 1, at 242 ("The constraint the judges experienced in the prison cases did not derive from a text, but from the internal dynamics of the coordination process."); id. at 358 ("[Flidelity to texts is often regarded as the sine qua non ofjudicial legitimacy, because it is the means of binding judges to the people's will. .. "); cf, e.g., id. at 11 ("Sometimes,judges base their decision on their own best efforts to understand an authoritative text; at other times, however, they base their decision on their sense of the best public policy.").

Id. at 338; cf. id. (claiming that judicial opinions based on interpretation of legal texts "will be replete with textual references" whereas those based on policy making will include "discussions of moral norms, social principles.... and citations of prior decisions that feature such discussions").

Id. at 6-7. Feeley and Rubin say that Bobbitt's theory "certainly does describe the practice of constitutional interpretation." Id. at 7 (emphasis added). Moreover, they argue that Bobbitt's "prudential" mode of argument identifies a "separate policy-making approach." Id. at 12. But if the practice of constitutional interpretation includes policy making, then the distinction between interpretation and policy making collapses. Bobbitt presents his theory in Pmi-np BOBBIr, CONSTrrrTIONAL FATE (1982), and PHILIP BOBBrT, CONSTITauTONAL INTERPRETATION (1991). See also Symposium, Philip Bobbitt's Constitutional Interpretation, 72 Tax. L. REv. 1703 (1994).

See FEELEY & RUBIN, supra note 1, at 4. According to Feeley and Rubin:

IJiudicial policy making produced the constitutional right of privacy decisions such as Griswold v. Connecticut and Roe v. Wade, the common law right of privacy and publicity decisions, the free speech decisions, the mental hospital reform decisions, many federal antitrust decisions, and the decisions creating implied warranties for consumer products.

Id. (footnotes omitted).

See Miranda v. Arizona, 384 U.S. 436, 478-79 (1966). Compare Joseph D. Grano,

Prophylactic Rules in Criminal Procedure: A Question of Article II Legitimacy, 80 Nw. U. L. Ray. 100, 163-64 (1985) (concluding that "prophylactic rules [that] function[ ] as conclusive presumptions[, as does Miranda,] ... must be outside the scope of the federal courts' article III lawmaking authority"), with David A. Strauss, The Ubiquity of Prophylactic Rules, 55 U. CHI. L. REv. 190, 209 (1988) (arguing that prophylactic rules like those contained in Miranda are ubiquitous in constitutional law and concluding that Miranda "reflected, at bottom, a traditional approach to the interpretation of the Constitution").

104 For an in-depth discussion of Fourth Amendment doctrine, see WAYNE R. 1AFAVE,

SEARCH AND SFizuRE: A TRaraSE ON THE FOURTH AMENDMENT (1996).

347 U.S. 483 (1954). FEELEY & RUBIN, supra note 1, at 13. Id. at 14.

Id. at 338.

See id. at 233 ("The conceptual account given here is itself a theory, and it is derived

from the more general phenomenological idea that people construct systems of meaning to control both their perceptions and their actions.").

See id. at 232 ("Public choice theory would focus on the self-interest of the judges,

ascribing the prison reform cases, perhaps, to their desire to maximize their salaries or their leisure time.").

See id. at 218 (rejecting "the notion, sometimes implied in critical legal studies or critical race theory, that the federal judiciary constitutes a conscious conspiracy"); cf. id. at 232 ("Marxist scholars might attribute the judiciary's actions to the elite's desire for more efficient mechanisms to control the lower classes ....

See id. at 233 ("[Tlhese deeper explanations, whether based on public choice, Marxism, or a variety of other theories, do not describe the terms in which the actors themselves thought.").

Id. at 360; see also id. at 338-39 ("[J]udges will continue to make policy in certain

circumstances, and.., they will continue to be circumspect, and perhaps insincere about it, until a legal discourse for judicial policy making is developed.").

See id. at 11 ("Judicial policy making and interpretation are... separate functions,

despite their obvious overlaps and frequent simultaneity of operation."); id. at 337 (noting that "interpretation and policy making mix together").

115 Id at 361. But cf Scott Altman, Beyond Candor, 89 MicH. L. REv. 296, 299 (1990)

(arguing that "judges holding inaccurate beliefs about their decisions might decide better than they would with a clearer understanding of their actions").

116 117 118 119 120

FEELEY & RUBIN, supra note 1, at 346. Id. at 347. Id. at 21. Id. at 349. Id. at 21. Id. at 242; see also id. at 351 ("The rule of law remains a valid norm, and an important one, but it has been transformed by administrative reality and modem social theory from the requirement of fixed ind preestablished rules to one of socially embedded constraints on the actions of government officials.").

Id. at 355. See id. at 243 ("What is being constrained is not each individual judge but the judiciary as an institution."); id. at 244 ("It is not the individual decision maker who is constrained, but the institution in its general operation."); id. at 346 ("Judges are constrained by the internal dynamics of the policy-making process, specifically the need to implement policy by means of ideas that coordinate individual judges' efforts to integrate personal attitudes and preexisting doctrine."); id. at 352 (explaining that the constraining force "is real and results from social attitudes and hierarchical supervision, not from governmental

structure"); id. at 353 ("The substantive constraints that limit the judicial policy-making process do not flow directly from social norms, but rather are inherent in the techniques by which that process is performed.").

124 125

Id. at 351. In reality, Warden Sarver was a willing participant in Judge Henley's efforts to

restructure Cummins and Tucker. See, e.g., id. at 61 ("[I] twas widely believed at the timeand still is today-that Sarver helped the plaintiffs' attorneys draft their complaints against him.").

126 Cf Jules L. Coleman & Brian Leiter, Determinacy, Objectivity, and Authority, in LAW

AND INTERPRETATION 203, 232 (Andrei Marmor ed., 1995) ("[I]ndeterminacy will not pose the threat to liberalism one would otherwise expect, provided that indeterminate judicial decisions are nevertheless reliably predictable.").

127 Cf id. at 236 ("In order for the coercive power of the state to be legitimately em-

ployed, judges' decisions must be justified... by the available set of legal reasons.").

FEELEY & RUBIN, supra note 1, at 246. 410 U.S. 113 (1973). See FFELEY & RUBIN, supra note 1, at 246-47.

128 129 130 131 SeeAlexander M. Bickel, Foreword: The Passive Virtues, 75 HARv. L. REv. 40, 79 (1961) (emphasizing "the wide area of choice open to the Court in deciding whether, when, and

how much to adjudicate"); see also ALEXANDER M. BICKEL, THE LEAST DANGEROUS BRANCH: THE SuPREME COURT AT THE BAR OF POLITICS 111-98 (1962) (Chapter 4: "The Passive Virtues").

132 133

FEELEY & RUBIN, supra note 1, at 247. See id.

See id.

134 135 Cf RONALD DwORKIN, TAKING RIGHTS SERIOUSLY 136 (1977) ("[T]he Court can enforce what the Constitution says only by making up its own mind about what is cruel ...."). 136 Dworkin has modified his theory over time. He calls its most recent instantiation

"law as integrity." RONALD DWORIuN, LAW'S EMPIRE 225-75 (1986).

See generally Ronald Dworkin, No Right Answer?, 53 N.Y.U. L. REv. 1 (1978) (defending the "one right answer" thesis). Dworkin no longer makes this strong claim. See Coleman & Leiter, supra note 126, at 214 & n.23.

See DwoRmN, supra note 136, at 105 (introducing Hercules, "a lawyer of superhu-

man skill, learning, patience and acumen").

139 Cf.Joseph William Singer, The Player and the Cards: Nihilism and Legal Theory, 94 YAtE

LJ. 1, 14-19 (1984) (explaining why legal doctrine is most likely indeterminate).

140 Cf iti at 19-25 (explaining why legal doctrine is predictable despite being

indeterminate).

141 Cf FEm= & RUBIN, supra note 1, at 240 ("The creation of doctrine clearly repre-

sents a mode of legal decision making where doctrine is indeterminate ....

I draw these ideas from Coleman & Leiter, supra note 126, at 213, which I think is

one of the best and most complete discussions around of the relationship between legal determinacy and authority.

143 H.L.A. HART, THE CONCEPT OF LAW 100-10 (2d ed. 1994) (describing features of the rule of recognition). For an account of the rule of recognition that differs from Hart's, see Jules Coleman, Authority and Reason, in THE AUTONOMY OF LAW 287, 288, 296-97 (Robert P. George ed., 1996) (outlining "an alternative account of the rule of recognition that emphasizes the fact that the rule of recopnition is a convergent social practice among officials").

144 HART, supra note 143, at 256. 145 Cf. Kent Greenawalt, The Rule of Recognition and the Constitution, 85 MICH. L. REV. 621, 655-56 (1987) (arguing that interpretive "standards need to be accorded some place among ultimate or derivative criteria for determining law").

146 HART, supra note 143, at 256.

147 Coleman & Leiter, supra note 126, at 237. For a reply to Coleman and Leiter's defense of law's legitimate authority in the face of its moderate indeterminacy, see Mark Tushnet, Defending the Indeterminacy Thesis, 16 QuNmNPIAc L. Rxv. 339 (1996).

See FEELEY & RUBIN, supra note 1, at 336 (noting the "brute existence" of judicial

policy making).

149 See, e.g., Henderson, supra note 26, at 1017-18 (setting forth criteria under which

institutional litigation is "at least arguably" desirable).

See FEELEY & RUBIN, supra note 1, at 26 ("[W]e do not offer any conclusion about whether judicial intervention into state prisons was good or bad-that is, whether it achieved a net social benefit. The main reason is that the criteria by which such a determination could be made have not been established.").

CONCLUSION

a

,

151 Feeley and Rubin do suggest that judicial policy making is desirable insofar as it typically results only in a widely-held norm being imposed on a recalcitrant hold-out. See id. at 202-03. So, for example, the prison cases resulted in the courts' imposing a national norm relating to proper prison conditions on the recalcitrant South. See id.

See Rubin & Feeley, supra note 6, at 1989-90.