THE VIRTUES OF REDUNDANCY IN LEGAL THOUGHT

38 Cleveland State Law Review 153 (1990)

I. INTRODUCTION

             Redundancy has had a bad reputation among legal intellectuals. When someone says, for example, that  the ninth and tenth amendments are redundant, we can be pretty sure that this person attaches little importance to these constitutional provisions. Listen to one of the definitions of redundant provided by the Oxford English Dictionary: "superabundant, superfluous, excessive." 1 In this essay, I propose that legal theorists pay serious attention to the concept of redundancy used by engineers. I explain how redundancy- in this special sense- is essential to any intellectual enterprise in which we try to reach action-guiding conclusions, including the enterprise of law. I will describe the virtues of redundancy in legal thought.

            My interest in the virtues of redundancy grows out of my interest in the social function of the liberal conception of justice and the rule of law.2 By "justice," I mean the correctness of the end or results of a legal system; by the "rule of law," I mean the formal and procedural correctness  of the means used to reach these results. Put another way, justice is the substantive correctness of a legal decision; the rule of law is the formal and procedural correctness of a legal decision. Although the liberal conception of justice and the rule of law is far more specific, we need not settle on an exact conception of justice and the rule of law here. The virtues of redundancy would arise regardless of what specific content were given these two concepts provided that we believe that both justice find the rule of law are distinct values that a legal system needs to respect.

           For as soon as we acknowledge the legitimacy of both justice and the rule of law, we are confronted with potential conflicts between these concepts when attempting to decide the merits of a particular act.  Because these concepts are inherently different from each other, an analysis based on justice may lead us in one direction while an analysis of the rule of. law may lead in another.  The possibility-indeed the inevitability- of, such conflicts leads legal intellectuals to establish a hierarchy of values such that in cases of conflict one value governs.  So some may consider a substantive concept of justice to be primary while others emphasize the rule of law.  Thus some judicial conservatives accuse their leftist adversaries of being "result-oriented," while others on the left accuse their conservative opponents as being "formalist." Each camp is willing to jettison the less-favored value whenever it conflicts with the value on, which it places a premium. 3

            To appreciate the virtues of redundancy in legal thought, we must first consider why an exclusive reliance on either justice or  the rule of law is misplaced.  Put another way, I want to explain why it is useful to rely on two (or more) different modes of analysis even though in many cases these modes of analysis are "superfluous" or  redundant insofar as they generate the same result and even though such reliance will at times yield conflicting evaluations.  Indeed, I want to claim that such conflicts serve an important function.  Let me begin by briefly summarizing my thesis.

II. CONVERGENCE  AND CONFIDENCE

                 We all carry with us an extensive set of beliefs or opinions about most matters that relate to our lives.  We critically analyze our beliefs because we know that it is quite possible that any of the opinions we currently hold may be in error.  Yet because we also know that any of our methods of critical analysis are fallible, we rely on different methods of analysis that have proven their value over time and that we have reason to think capture something that is true about the world we live in.  The perceived soundness of our modes of analysis does not depend merely on their being widely accepted, but, rather, we generally accept as "sound" those methods of critical analysis that have been recommended by both our experience and our reason.  Conversely, we reject as unsound even popular methods of analysis-such as astrology-that we have no reason to think capture anything that is true about the world.

              My thesis is that the degree of confidence we have in any of our beliefs largely depends upon the degree to which the different methods we use to critically assess our beliefs converge on the same conclusion.  The greater the number of different sound methods of evaluation that converge on a single conclusion, the more confident we can be in that conclusion. Conversely, a conflict between competing sound modes of analysis over a particular conclusion should lessen our confidence in it and motivate us to search for a better approach.  The virtue of adopting multiple or redundant modes of analysis is, then, two-fold: (a) convergence (or agreement) among them supports greater confidence in our conclusions; and (b) divergence (or conflict) signals the need to reexamine critically the issue in a search for reconciliation.  In sum, convergence begets confidence, divergence stimulates discovery.

                We may think of this as a "checks and balances" theory of decision-making.  Traditional constitutional theory justifies a separation of powers on the grounds that each distinct branch of government is supposed to act as a "check" on the others.  Although any given decisionmaker can err or abuse its power, we cannot always detect when error or abuse has occurred.  Consequently, we can be more confident in actions taken by a system in which distinct institutions with different constituencies and procedures must concur on a single course of action than one that relied exclusively on a single decisionmaker.  This is not to say that the failure to take action in the absence of unanimity may not sometimes prove to be a problem.  Rather, the idea is simply that, whatever the costs of inaction, we can be more confident in the actions we take when all agree than when one or more branch of government is willing to oppose a particular course of action.

                 Similarly, in law an "easy" case is one in which all our methods of analysis reach the same result giving us a high degree of confidence in this result.  For example, we can be more confident that a particular statute creates a duty of obedience when we conclude that its substance is consistent with an analysis of justice, that it has the appropriate formal qualities, and that it was enacted by sound procedures. 4 None of the relevant methods of analysis is unnecessary or superfluous because it is the convergence of the results of multiple analytic methods that provides this confidence and, in the absence of multiple modes of analysis, we cannot observe convergence.

                  If being "redundant" is synonymous with being unnecessary or superfluous, then neither mode of analysis is redundant.  But if, however, the term "redundancy" is used as in engineering to refer to parallel systems that work in the event another system fails, 5 then redundancy has its virtues.  For example, multiple control and navigation systems are designed into an airplane in the event one or more systems fail.  We have far more confidence in the safety of, such  a plane than we would in a plane that relied exclusively on a single control system or on a single navigation system.  So too in analyzing the legitimacy of a statute, if we were to do away with any of our sound modes of analysis, we may still reach the correct result, but we will lose something valuable: the confidence that our judgment is correct.

               Analogously, in evidence law we heighten our confidence by permitting the introduction of so-called "cumulative" evidence.  For example, when the identity of an assailant is at issue, we permit several witnesses each to testify that the defendant was the culprit.  All other things being equal, when many witnesses converge independently on the same person, we can be more confident in their identification than when a single witness identifies a defendant, or when different witnesses disagree about whether the defendant was the assailant.  In sum, cumulative or "redundant" evidence is relevant and admissible because it can affect the confidence of the fact finder in reaching a conclusion.

                   The virtues of redundancy extend to moral reasoning as well.  For example, to determine the requirements of justice (as distinct from the rule of law), we may begin with a historical inquiry to identify the rights that have evolved over time in legal systems that adhere to sound procedures.  We may then evaluate these evolved rights against the results recommended by an abstract deontological moral rights analysis.  Finally, we may "check" the results of these two modes of analysis by evaluating the social consequences of adhering to the rights they recommend. 6 Today we would call a person who relied exclusively on the first of these modes of analysis a traditionalist or historicist; one who relied exclusively on the second, a Kantian; and one relying exclusively on the third method, a Utilitarian. 7 That there is no accepted term to describe one who accepts the importance of using all three modes of analysis as checks on one another indicates the extent to which the virtues of redundancy in moral reasoning are unappreciated,

                   Some may see a subtle difference between my use of the term redundancy and its use in engineering.  In my account we constantly evaluate our conclusions by a variety of different analytic methods to achieve some degree of confidence in them or to identify areas that require further discovery.  In an airplane, however, we normally rely exclusively on a "primary" system, and shift to the redundant or "back-up" system only when our primary system fails.  Moreover, when we shift, we shift completely to the back-up system and rely upon it exclusively until we land.  If there is a difference in usage, however, it is more a matter of degree than of kind. 8 On the one hand, in engineering, redundancy is a broad concept encompassing a variety of approaches; redundant systems are not always primary and secondary in nature but may operate simultaneously-that is, two or more systems may be said to be "on line" redundant.  Multiple systems are said to "mirror" each other's operation so that fault can be detected by divergence among the systems and, once detected, corrected.  This kind of redundant system is sometimes called "fault tolerant." 9

                 On the other hand, in both law and moral reasoning we generally adopt one mode of analysis as primary and then check it against other secondary modes of analysis.  For example, we may primarily rely on doctrinal analysis to determine legal results, except in comparatively rare and difficult cases when we must reconsider the results suggested by a doctrinal analysis in the more theoretical light of justice and the rule of law.  Elsewhere I have described the legal reasoning process of applying doctrine consisting of rules and principles to facts and of checking doctrine against theory as "three-dimensional." 10 In this way, the primary system of legal thought is doctrinal-what the general public thinks of as "the law"with legal theory and such deeper theoretical concepts as justice and the rule of law operating in a secondary or "back-up" manner.

                Another potential difference between engineering and legal redundancy is that while redundant subsystems in most engineering applications tend to be identical to each other, in law we avoid system failure by relying on modes of analysis that differ from each other.  This is because the typical engineering problem is to cope with subsystem failures caused by component failures when there is no doubt that a subsystem, if operational, would accurately perform its designated function.  In law (and in unusual engineering applications), however, the problem is that even a fully functioning mode of analysis can err or be abused so redundancy requires multiple subsystems of analysis each constructed according to different criteria to detect such errors or abuse.  The engineering concept of redundancy is broad enough to encompass both strategies of achieving fault tolerance.

III. WHY THE VIRTUES OF REDUNDANCY ARE OVERLOOKED

                  At least three reasons explain why the virtues of redundancy are so commonly overlooked.  First, moral philosophers and legal intellectuals do not spend much time worrying about easy cases where differing modes of analysis converge.  Their primary concern is with hard cases in which sound methods of analysis recommend conflicting results.  The curiosity of moral philosophers is piqued when the best analysis of moral rights seems to conflict with the best analysis of consequences.  Such cases then become their primary and often their exclusive concern.  Similarly, legal thinkers are provoked when, for example, adherence to the formalities of the rule of law seems to require the injustice of releasing a guilty criminal or of refusing to enforce a contractual commitment.  Nevertheless, a relentless focus on hard cases where competing modes of analysis diverge can be like staring into the sun.  It can cause one to see nothing but hard case-, and blind one to the importance of cases where convergence exists.

               Second, modern intellectuals are trained to accept the principle of parsimony-or "Ockham's razor" 11 -leading them to seek the minimally sufficient account of any conclusion.  Where two explanations reach the same result in a particular case, they go on to offer increasingly difficult cases to determine which of the explanations of the easy case is "doing the real work." In economic terms, they seek the single "marginal" factor that tips the balance in one direction or the other and consequently ignore the many "inframarginal" factors that are needed to get to the margin.  My thesis is that multiple analyses are necessary to engender confidence in our conclusions and to stimulate the growth of legal and moral knowledge and therefore, if this thesis is right, multiple analyses would survive Ockham's razor.  But because the concern of intellectuals is usually to identify the singular analysis that "best" accounts for a particular result-not an analysis that begets confidence or stimulates discovery-the virtues of redundancy are usually neglected.

                The source of this preoccupation with finding a single best account provides a third reason why the virtues of redundancy are overlooked.  Intellectuals in many disciplines, from law to philosophy to economics, are often oblivious to the serious practical problems of knowledge and interest that pervade actual decisionmaking. 12 By arguing that their favored method of analysis should be the sole touchstone of decisionmaking despite the likelihood that even the best of methods will sometimes be mistaken, many intellectuals underestimate the practical problem of knowledge.  By assuming the good faith of those employing evaluative methods of which they approve and ignoring the possibility that analysts will deliberately or unconsciously skew the results of their analysis, many intellectuals underestimate the practical problem of interest.  In sum, many, if not most, intellectuals tend to pay little attention to the need to make decisions in the face of either fallible or biased analysis or both.  But it is the pervasiveness of human fallibility and bias that renders redundancy a virtue.

IV. ACHIEVING CONVERGENCE BETWEEN COMPETING MODES OF ANALYSIS

                Suppose we accept the twin virtues of redundancy: convergence begets confidence and divergence stimulates discovery.  When a conflict between modes of analysis appears, how exactly are we to deal with the tension?  In the balance of this article, I discuss some well-known examples of a divergence between justice and the rule of law to illustrate various waYs we deal with such divergence.  In this section, I discuss the different ways to resolve divergence and achieve convergence.  In the next section, I will suggest how we cope with seemingly irreconcilable conflicts between sound modes of analysis.

                A conflict between normally sound modes of analysis provides important information about the existence of a possible problem either 'with the application of our analytic techniques or with the techniques themselves.  Without such conflicts we would have no reason to reconsider either our analyses or our analytic methods.  To consider how conflicts between different modes of analysis can be productive, let us consider cases in which justice seems to be in conflict with the rule of law.

                 In the introduction to this article I offered the following distinction between justice and the rule of law: Justice refers to the correctness of the end or results of a legal system; the rule of law refers to the formal and procedural correctness of the means used to reach these results.  To be sure, we sometimes use the term "justice" to refer both to the correctness of a result as well as to the propriety of the procedure by which it was reached.  According to this usage, there are two categories of justice: "substantive" justice on the one hand and "procedural" justice on the other.  However, since the liberal value of the rule of law is the traditionally accepted term that corresponds to procedural justice or "legality," we are left with the term "justice" to refer to the correctness of result as assessed by criteria other than the rule of law.

                This way of distinguishing between justice and the rule of law reveals that a conflict can occur when an end we think to be correct conflicts with a means we also think to be correct.  For example, suppose that police search a murder suspect's apartment and obtain vital evidence-bloody sheets and clothes.  Because such evidence is both reliable and relevant to the charge, it would ordinarily be admissible in a criminal trial of the suspect.  Because it is both relevant and reliable, the evidence provided by the sheets and clothes serves the just end of establishing the guilt of a murderer.  Now suppose that the search was "unlawful" because it was instigated by a voluntary statement made by the defendant after having been arrested without probable cause.  Although the admission of such evidence is just with respect to the end of convicting a murderer, it violates the rule of law because it was obtained by procedurally flawed means.  Considerations of justice argue in favor of admitting this evidence, while rule of law considerations argue for exclusion.  Although I shall continue to concentrate on this example, the same type of conflict arises elsewhere.  For example, the statute of frauds requires that to be enforceable certain promises be in writing.  The absence of a writing undermines the legitimacy of enforcing even a promise we are quite sure was made.  How are such conflicts to be resolved?  I shall consider three ways of resolving conflicts by achieving convergence.

A. Achieving Convergence by Retracing One's Steps

                 The most obvious and probably the most common way to relieve a tension that develops between competing methods of analysis is to retrace one's steps.  When addition of a column of figures yields a sum that is unexpectedly high or low we usually recalculate to see if we might have made a mistake.  When we reach two different sums in two tries, we usually try a third time to see which of the two conclusions was right.  Of course, the divergence between our initial calculation and our expectations presupposes some competing source of expectations other than formal calculation.

                 For example, when figuring my travel expenses for a business trip I may find the calculated sum to be high or low according to my experience on past trips or perhaps according to a rough running total I kept in my head during the trip.  To disregard knowledge based on my past experience when it conflicts with knowledge based on a formal computation would be irrational, if recalculating the figures is easy enough to do.  Suppose that when I recalculate my expenses I find that I had previously erred in my formal computation.  Convergence of conclusions between my formal calculations and my experience is now attained.  I need reexamine neither my formal method of calculation nor my experience.

                To illustrate how we can retrace our steps in law, consider a case I worked on as a criminal prosecutor, involving a motion by a defendant to suppress his confession to murder.  Chicago Police Department homicide detectives were at a loss to solve a vicious murder of a husband and wife.  The man and woman had been hog-tied in their apartment and repeatedly shot in their heads in the presence of their young children.  Having no leads whatsoever, the detectives began interviewing all arrested persons to see if any of them had heard anything "on the street." The defendant, Joseph Cooley, had been taken into custody on an entirely unrelated burglary charge.  Cooley told the homicide detectives that he had heard about the crime and that a person he knew only by a nickname had committed it.  The police asked for a description of this person and where they might find him, and Cooley provided the information.  When the police were unable to locate the person, Cooley agreed to help the police look for him.  Eventually, the police began to suspect that the person described by Cooley was fictitious and that Cooley himself might have committed the crime.  At this point, they advised Cooley of his constitutional rights and he confessed to having committed the murder with a friend of his named Sammy Bynum.  When questioned about the murder and confronted with Cooley's statement, B-,-num also confessed.

                 It was well-settled that a statement by a person who had not been advised of his constitutional rights made during a "custodial interrogation" could not be used at trial by the prosecution. 13 Although Cooley had confessed only after he had been advised of his constitutional rights, the defense attorneys' motion to suppress the confession relied on what is known as the "fruit of the poisonous tree" doctrine because evidence (the fruit) obtained by the police due to their illegal conduct (the poisonous tree) is subject to suppression despite the fact that, viewed in isolation from the earlier illegality, this evidence was obtained legally. 14  Here Cooley's confession was made only after the police had confronted him with evidence that his earlier statements were lies.  Thus his "legal" statement given after the warning (the fruit) was a direct product of the earlier statements he made before he had been advised of his rights (the poisonous tree).  The defense argument was simple and well-supported by prior cases: because Cooley was unquestionably in custody at the time he first spoke to the police, the lack of warnings before these statements rendered his later confession inadmissible.

                  I searched long and hard for cases to justify admission of the confession but to no avail.  I was haunted by the thought that, notwithstanding the rules of criminal procedure as I understood them, what the police had done somehow made sense.  After several frustrating days, I pulled back from my research and thought about the doctrine upon which the defense was relying.  What evil was this rule supposed to address?  The rule was intended primarily to prevent the inherent coercion of a custodial interrogation from leading to false confessions.  Also, the repeated act of reciting these rights serves to remind police officers that suspects in custody have constitutional rights that the police are legally bound to respect.  Moreover, there is a sense that, wholly apart from its effect on confessions, it is simply unjust for police to mistreat suspects in their custody.  For all these reasons, a rule of thumb had developed among both police and criminal lawyers that any statement made by a defendant while in custody is inadmissible unless, prior to making the statement, the defendant had been apprised of his or her rights.

                   In my case the defendant was unquestionably in custody. . . but wait!  The rule of thumb rested solely on whether or not the defendant was in custody.  The cases upon which this rule of thumb was based, however, spoke in terms of a "custodial interrogation." In this case, perhaps the police conduct seemed reasonable because the normal coercion incident to an interrogation (custodial or otherwise) was absent.  The police had no reason to suspect this defendant and no reason to mistreat him to gain the information they desired; they were simply eliciting information from every criminal suspect with whom they came in contact. (Although the defendant was undoubtedly motivated to fabricate his story to obtain some consideration on his burglary charge, this type of motivation was not considered by courts to be unduly coercive, unless the police induced the statement by making promises of a benefit.) Notwithstanding that the defendant was in custody, perhaps this was not an interrogation at all; perhaps it was a custodial conversation.  In these circumstances, the fact that the defendant was in custody did not create, as it usually does, the element of coercion and abuse that the rule was devised to address.

                  With this theory in mind, I went back to the books and reviewed several examples of statements that had been admitted into evidence despite the fact that the defendant was in official custody when the statements were made.  In each of these "exceptional" cases the elements of an interrogation appeared to be lacking.  Although I thought they established the plausibility of my theory, these cases were not substantially similar to mine, and none made the explicit distinction between interrogation and conversation that I was planning to urge upon the judge.  Then I happened upon the case of United States v. Wiggins. 15

                  In Wiggins, the police were baffled in their effort to solve a murder and the defendant was in custody for an unrelated burglary.  In the course of questioning every person arrested for a crime, they spoke with the defendant, Charles Wiggins, who told them he had heard who had done it.  When the police were unable to find the person Wiggins had described, they asked if he would help them locate him.  Eventually, the police came to suspect Wiggins and they advised him of his constitutional rights.  Wiggins then confessed.  The only salient differences between my case and Wiggins were the charge for which the defendant was initially arrested (auto theft, not burglary) and the fact that in Wiggins the defendant initiated one of the pre-warning statements.

                   In Wiggins, the Court of Appeals for the District of Columbia held that, although the initial statements were made while the defendant was in custody and before he had been advised of his rights, these statements were not the product of an interrogation, but were conversational in nature.  In its ruling, the court cited no precedent with any comparable facts; as near as I could determine at the time, no appellate court in the country had so held up to then.  Relying on my theory (except it wasn't really my theory anymore) and Wiggins, I asked the judge to deny the defendant's motion to suppress the statement and he did.  After the motion was denied both Cooley and Bynum pled guilty to the crime.

                 Although I gained much satisfaction from my work on this case, I do not believe that my effort was extraordinary.  To the contrary, I think it is a rather typical example of how lawyers deal with cases where they perceive a conflict between a just outcome and the rule of law.  This story suggests how a better understanding of the law is provoked by tensions between methods of analysis.  When there is an apparent conflict between justice and the rule of law, we may be moved to retrace our steps.  The fault may not lie in our settled abstract conceptions either of justice (murderers should be convicted) or of the rule of law (confessions obtained improperly should be suppressed), but in our specific understanding of these conceptions.  Here an apparent tension between the justice of convicting the murderers and ensuring that the police adhere to the rule of law motivated a reexamination of the rule as applied to the facts of the case.  Also informing this reexamination was a perceived conflict between a rule of thumb that had developed in practice and my intuition that the police had acted reasonably.

             Examining the theoretical basis of the rule (no statements made during custodial interrogations without warnings) revealed a conflict between the original rule and a rule of thumb (no statements made while in custody without warnings) that had evolved in practices conflict that I and many others had not previously perceived.  I concluded that, in light of the underlying theory of the original rule, the rule of thumb was seriously misleading in this case. (It remains, however, a serviceable rule of thumb for most cases. 16) I did not discover a new exception to the rule; indeed, there was no need to do so.  Nor was there any need to question the merits of the original rule or the theory that supported it.  All it took to convert a hard case into a case that even the defense attorneys implicitly conceded was comparatively easy, was to gain a better understanding of the rule itself. 17

B. Achieving Convergence by Rethinking the Rule of Law

               Sometimes resolving a conflict between our modes of analysis requires us to alter our methods.  For example, a hard case in which justice appears to conflict with the rule of law may require that we modify our conceptions of either justice or the rule of law or both. At us consider again the exclusionary rule for evidence gained as a result of an illegal search.  In such a circumstance, the apparent lack of convergence is between the means employed (the illegal search) and the end being sought (the conviction of the guilty).  The trauma caused by this conflict may, however, motivate us to reconsider the merits of the exclusionary rule itself.  Without a sense of the injustice of releasing a guilty offender we would have no reason to search for some alternative to the exclusionary rule.

           Let us assume that the principal reason for excluding from a trial evidence obtained from an illegal search is to deter police from conducting illegal searches in defiance of the Constitution. 18 Although the object of the rule is the deterrence of illegal searches, the conflict between this rule and justice only arises when an illegal search reveals incriminating evidence.  Indeed, the exclusionary rule provides no remedy whatever to the victim of an illegal search when that search fails to uncover incriminating evidence.  Innocent victims of such searches simply have no evidence to suppress.  Confining legal relief solely to those persons who are in possession of incriminating evidence while denying any relief to those persons against whom no illegally seized evidence was found may seem even a bit perverse.  Only when it thwarts justice does the exclusionary rule apply.

                  Now suppose there was a different remedy that could be afforded to all victims of police misconduct and that did not conflict with the justice of using reliable evidence of guilt.  The remedy that immediately presents itself is monetary damages-the normal way that tort law compensates for the most serious of physical and emotional injuries and even for death.  Suppose it was shown that requiring police departments to compensate the victims of their employees' illegal conduct-including those victims against whom no incriminating evidence was found-in an efficient administrative setting (while allowing the prosecution to use the illegallyseized evidence at trial) would deter misconduct as well as, if not better than, the remedy of exclusion.  We might then resolve this conflict between justice and the rule of law and achieve convergence by adjusting our conception of one of these modes of analysis-in this case our conception of the remedy we use to rectify violations of the rule of law.  On the other hand, had the illegal police conduct undermined the reliability of the evidence obtained, as in the case of a coerced confession, such evidence would still be inadmissible.  Even if some coerced confessions are in fact truthful and accurate, coercion is likely to decrease the reliability of confessions in ways that cannot be effectively discerned by a court and the end of justice will not be well-served by relying on such tainted evidence.

                 Of course, it may be that restitution would not deter police misconduct as effectively as the exclusionary rule. 19 My point is simply that, if restitution and exclusion are roughly comparable means of deterring police misconduct, restitution is superior because, unlike the exclusionary rule, it would resolve a tension between justice and the rule of law.

C. Achieving Convergence by Rethinking Justice

                  Just as convergence may be achieved by altering the means to better fit the ends, we may sometimes reconsider our ends in light of the means that such ends require.  With drug prohibition, for example, the police are required to enforce a ban against consensual adult behavior on the grounds that such conduct is unjust.  Suppose, however, that due to the lack of a complaining witness in most drug cases (and other consensual crimes) this end simply cannot be obtained without conducting wholesale illegal searches or by creating so many exceptions to the prohibition on unreasonable searches that the Fourth Amendment is seriously weakened as a barrier to government misconduct.  The seriousness of this conflict between means and end may call into question the justice of the end of drug prohibition.

                Of course, simply highlighting the apparent conflict between the justice of drug prohibition and the means that are needed to enforce it is not dispositive.  We would not be moved to reexamine the law prohibiting murder, notwithstanding that such a crime also lacks a complaining witness.  Still, the trauma created by a profusion of intrusive police tactics might well provoke a more careful scrutiny of the underlying end of such laws.  Such scrutiny may reveal pertinent distinctions between consensual conduct and murder that lead us to conclude, not only that the end does not justify these means, but that the end itself is not justified.  Initially provoked by a serious conflict between justice and the rule of law we may come to conclude that punishing drug consumption is an unjust interference with choices that lie within the moral jurisdiction of individuals.20 By revealing difficulties that need to be addressed, conflicts between such competing modes of analysis as justice and the rule of law make possible the growth of legal knowledge.  Perhaps a perceived conflict can be addressed by a recalculation; perhaps we need to reformulate one or more of our modes of analysis.  When such conflicts exist we can neither act in complete confidence nor abandon our search for some resolution.  This pursuit of convergence in the face of occasional conflicts between redundant modes of analysis accounts for the evolutionary improvement of both our conclusions and our analytic techniques.

V. COPING WITH INTRACTIBLE CONFLICTS BETWEEN COMPETING MODES OF ANALYSIS

                  Assuming one accepts my account of the two virtues of redundancy, some may still object that the thesis presented here is trivial.  Everyone knows that convergence begets confidence, and divergence stimulates discovery, they might say; what do we do when conflicting modes of analysis resist our most diligent efforts at reconciliation as they seem to do in so many areas of law?  Moreover, such intractable conflicts force us to choose between our modes of analysis and, by choosing, reveal which mode is really fundamental to the others.

It is not clear, however. how this objection undermines the thesis presented here unless it is assumed that convergence and the confidence it engenders are extremely rare or that conflicts between modes of analysis rarely lead to the discovery of new knowledge.  Such claims would need to be substantiated, not merely asserted.  In sum, the problem of intractable conflict may be worth serious consideration without rendering the thesis presented here at all trivial.  But dealt with how?  Because my views on this issue are still developing, my description of the method by which we handle seemingly intractable conflicts-what I will call presumptive reasoning-will be brief and merely suggestive. 21

                 We normally handle unyielding conflicts by adopting an operative presumption.  For example, we may presume that an action recommended by all relevant modes of analysis is legitimate and that an action about which modes of analysis intractably conflict is illegitimate.  Either presumption can be rebutted under special circumstances, but unless such circumstances are established, we adhere to the outcome recommended by the operative presumption.

               In constitutional law, when the legislative, executive, and judicial branches of government disagree on the constitutionality of a federal statute, we presume such a statute to be illegitimate and refuse to enforce it. 22 Of course, we also recognize that in special and rather unusual circumstances this presumption of invalidity can be overcome, such as when a super-majority of Congress overrides a presidential veto, or when a constitutional amendment overcomes a Supreme Court opinion.  Similarly, as was discussed above, we generally require that, to be admissible, evidence must be both relevant to proving guilt and legally obtained.  Evidence that fails either test is presumed to be illegitimate.  In both of these examples, requiring convergence entails that a deficiency in any factor will result in a prima facie judgment that a particular action either enforcing a statute or admitting evidence-is illegitimate.

               These are examples of a presumption of invalidity that applies when our multiple modes of analysis diverge, but this is not the only type of presumption that can handle such conflict.  A presumption of priority designates one of several modes of analysis as presumptively governing a situation in which differing modes of analysis conflict.  For example, in constitutional law, the Supreme Court has adopted what it calls a "presumption of constitutionality" 23 which largely defers to a judgment of the legislative and executive branches that a statute is constitutional. 24 Such a presumption is supposed to determine the outcome of a case unless, for example, a statute infringes rights that the Court considers to be fundamental. 25
 

                 We may also use a presumption of priority to determine what justice requires when faced with a conflict between a moral rights and a consequentialist analysis.  Suppose, for example, that a moral rights analysis recommends against interfering with an individual's freedom of choice, while a consequentialist analysis recommends such interference.  We might presume such interference to be unjust, while allowing that an overpowering consequentialist case for interference might rebut our normal presumption favoring a moral rights analysis.  Permitting a moral rights analysis to prevail in the absence of convergence does not eliminate any use for consequentialist analysis, and permitting consequentialist analysis in extraordinary circumstances to rebut the normal presumption favoring moral rights does not reduce this type of approach to pure consequentialism.  In such a scheme, the moral rights analysis is setting the agenda; only outcomes that survive a moral rights analysis are then subjected to consequentialist scrutiny.

                Presumptive reasoning, it should be emphasized, does not eliminate hard cases caused by conflicting modes of analysis.  However, by converting a case that remains hard in principle into one that is easy to decide in practice, presumptive reasoning can provide a practical way of coping with conflicts among competing modes of analysis in the absence of an ultimate resolution of the conflict.  Cases where the exclusionary rule requires the guilty to escape conviction are hard to accept, but not hard to decide.  Regardless of the injustice caused by its suppression, until we resolve this conflict, illegally obtained evidence is simply inadmissible; call the next case.

The use of presumptive reasoning is pervasive.  The process by which our set of operative presumptions evolves is a subject worthy of more extended discussion than I shall attempt here.  Suffice it to say that, as presumptions proliferate, it is likely that such proliferation will provide previously unavailable information that may permit reform of the inherited set of default decisions.  That is, as the set of exceptions to a previously serviceable presumption is articulated, it may eventually become clear to observers and participants that the presumption itself requires modification, perhaps even reversal, so as to capture the outcomes now expressed as exceptions to the rule.

              In contract law, for example, simple promises are presumptively unenforceable.  For a time, only those promises supported by bargained-for consideration were presumed to be enforceable.  Exceptional cases in which promises were enforced in the absence of consideration were clustered in an ad hoc doctrinal category known as promissory estoppel.  As this category has expanded, the original presumption limiting enforcement to promises supported by bargained-for consideration has been widely criticized (but not yet replaced).

If, however, the existence of bargaining is viewed not as the sine qua non of contractual obligation, but as evidence of the parties' intention to alienate rights, then there may be other ways of manifesting such intent besides bargaining.  In many, if not most, promissory estoppel cases, there exists some other manifestation of intent to be legally bound- perhaps a formality of some kind-even though a bargain is lacking.  By replacing the presumption based on bargaining with a presumption based oil a manifested intention to be legally bound, most of the formerly exceptional promissory estoppel cases could be "justly" decided by the prevailing  presumption.26

                In sum, the process of developing an increasingly elaborate set of presumptions and exceptions can sometimes provide the knowledge necessary to simplify the system of presumptions in a way that, without a complicated past, would have been nearly impossible to conceive from the armchair.  Of course, once discovered, the new system of presumptions may still be critically evaluated to see whether it serves the values of justice and the rule of law.

                An evolutionary account of presumptions must include not only how we come to change our presumptions in the light of new information, but also how it is that some presumptions-such as the presumption of innocence-have proved to be so remarkably stable over time.  It may be that certain features of social life are so ubiquitous that they can accurately be characterized as part of the unchanging nature of human social life.  To the extent these problems remain constant, evolved ways of handling them -such as the presumptions specified by the liberal conceptions of justice and the rule of law-are likely to remain relatively stable as well.

VI. CONCLUSION: REDUNDANCY AND IMPARTIALITY

                  We are accustomed to seemingly endless debates between schools of thought erected around single touchstones for truth.  I say the ultimate touchstone is justice; you say it is the rule of law.  I say it is rights; you say it is consequences.  I say it is the individual; you say it is the group.  I say it is liberty; you say it is responsibility.  The analysis presented in this article suggests that these debates endure because both sides fail to appreciate the important role played by conflicting modes of analysis.

                  All of these dichotomous categories of analysis are conceptual tools that have proved useful to understanding, but each mode of analysis captures a necessarily partial view of the world we inhabit.  Because each view is partial, however, it is not circular to use each side of a dichotomy (or trichotomy) to criticize the other.  Viewing a given phenomenon from a variety, of partial, though sound, perspectives provides a clearer picture of the underlying phenomenon itself.  The more different sound modes of analysis we consult, the more impartial our vision becomes, The values associated with justice and the rule of law are two such partial, but sound, perspectives of law.  Much as our brain's ability to fuse the slightly different two-dimensional perspectives of each of our eyes enables us to see in three dimensions., a superior vision of law results from the fusion of justice and the rule of law.

FOOTNOTES:

* Visiting Professor of Law, Northwestern University School of Law; Professor and Norman Edna Freehling Scholar, Illinois Institute of Technology, Chicago-Kent College of Law. An earlier version of this paper was presented at  the "Conference on the Practical Applications of Jurisprudential Thought" held October 19-21, 1989, at  Cleveland State University College of Law, at the Legal Theory Workshop at Columbia University School of Law and at faculty workshops of at Northwestern school of Law and Western New England School of Law. I benefitted greatly from these discussion as well as from comments by Ron Allen, Jim Haddad, Linda Hirshman,  Keith Hylton, Dale Nance, Michael Perry, and Marshall Shapo. Special thanks are due to my colleague, Anita Bernstein, for her extensive comments and suggestions which helped to improve this article immeasurably. Finally, I received financial support for the project  of which this paper is part from the Mars! hall Ewell Fund of IIT Chicago-Kent College of Law and from the Humane Studies Foundation.

1 8 OXFORD ENGLISH DICTIONARY 319 (1970)

2 The analysis presented here is part of a larger work in progress in which I discuss how the liberal conceptions of Justice- based on certain notions of individual rights- and the rule of law- based on certain formal features of legality- have evolved as solutions to the pervasive social problems of knowledge, interest, and power. I summarize the thesis in  Barnett, Foreword: Can Justice and the Rule of Law be Reconciled? 11 HARV. J. L. & PUB. POL'Y 597 (1988).

3 See id. at 597-99.  

4  See Barnett, Foreword: The Ninth Amendment and Constitutional Legitimacy, 64 CHI-KENT L. REV. 37 (1989).

5 See, e.g., McGRAW-HILL DICTIONARY OF SCIENTIFIC AND TECHNICAL TERMS 1582 (4th ed. 1989): "[R]edundancy ... Any deliberate duplication or partial duplication of circuitry or information to decrease the probability of a system or communication failure." The second edition of the Oxford English Dictionary now includes this meaning.  See 13 OXFORD ENGLISH DICTIONARY (2d ed.) 438 (1989) ("Engin.  The incorporation of extra parts in the design of a mechanical or electronic system in such a way that its function is not impaired in the event of a failure.").

6 See Barnett, Foreword: Of Chickens and Eggs- The Compatibility of Moral Rights and Consequentialist  Analysis, 12 HARV.  J. L. & PUB.  POL'Y  611 (1989).

7 I suggest elsewhere that traditional liberal natural rights thinkers used all three modes of analysis to achieve confidence in their conclusions.  See id. at 626-630.

8 I thank my brother, Howard Barnett, for tutoring me in the following rudiments of engineering terminology.

9  See, e.g., MCGRAW HILL DICTIONARY OF SCIENTIFIC  AND TECHNICAL TERMS, supra note 5, at 698: [Fault tolerance ... The capability of a system to perform in accordance with design specifications even when undesired changes in the internal structure or external environment occur.

10 See Barnett, Why We Need legal Philosophy, 8 HARV.  J. L. & PUB.  POL'Y 1, 9-10 (1985).

11 See, e.g., A DICTIONARY OF PHILOSOPHY 164 (2d ed. 1986): Ockham's razor.  Principle propounded by William of Ockham, that . . . one should choose the simplest hypothesis that will fit the facts.  A stronger form claims that only what cannot be dispensed with is real and that to postulate other things is not only arbitrary but mistaken.

12  I summarize the problems of knowledge and interest in Barnett, supra note 2, at 599-618.

13 The rule was formulated in Miranda v. Arizona, 384 U.S. 436 (1966).

14 See W. LaFave & J. Israel, 1 CRIMINAL PROCEDURE 760-68 (1984).

15 509 F.2d 454 (D.C. Cir. 1975).

16 Some may favor retaining for every case the rule of thumb which relies solely on the factor of custody over this interpretation of the original rule which requires both custody and interrogation.  Making the necessity of warnings turn on whether a conversation was or was not an interrogation will inject an increased element of uncertainty into the enforcement of the rule-uncertainty for both the police and the courts.  While a plausible objection, I shall not consider it here except to note that it illustrates why the rule of law is likely to clash with justice.  The increased ease and certainty of application that attaches to the less discriminating rule of thumb comes at the price of more conflicts with justice.

17 Since I litigated this case, the U.S. Supreme Court has resolved the tension between justice and the rule of law in this situation by holding that the "fruit of the poisonous tree" doctrine does not apply to Miranda violations.  Thus, the postwarning statement by Cooley would no longer be tainted by the earlier statements made with no warning, provided that all the statements were voluntary. See Oregon v, Elstad, 470 U.S. 298 (1985).

18 The following discussion of the exclusionary rule is taken from Barnett, Resolving the Dilemma of the Exclusionary Rule: An Application of Restitutive Principles of Justice, 32 EMORY L.J. 937 (1983).

19 And other objections may be made against this proposal as well.  I consider some of them elsewhere.  See id.

20 For a more extensive consideration of this issue, see Barnett, Caring the Drug-Law Addiction: The Harmful Side Effects of Legal Prohibition, in DEALING WITH DRUGS 73 (R.  Hamowy ed. 1987),

21 For an account of how presumptive reasoning operates in legal decision-making, see Epstein, Pleadings and Presumptions, 40 U, CHI.  L. REV. 556 (1973).

22 See Barnett, supra note 4.

23 See United States v. Carolene Products, 304 U.S. 144, 152 n.4.

24 As the Court in Carolene Products explained: [T]he existence of facts supporting the legislative judgment is to be presumed, for regulatory legislation affecting ordinary commercial transactions is not to be pronounced unconstitutional unless in light of the facts made known or generally assumed it is of such a character as to preclude the assumption that it rests upon some rational basis, within the knowledge and experience of the legislators. 304 U.S. at 152.

25 See 304 U.S. at 152 n.4 ("There may be narrower scope for operation of the presumption of constitutionality when legislation appears on its face to be within a specific prohibition of the Constitution, such as those of the first ten amendments. . . ."). Other circumstances listed by the Court that may rebut the normal presumption of constitutionality include legislation that affects the political processes or legislation directed at religious groups or motivated by prejudice against discrete and insular minorities.

26   This discussion greatly simplifies these issues of contract theory.  For instance, the presumption favoring enforcement of bargained-for promises displaced an earlier exclusive reliance on a presumption favoring formal contracts under sea).  For an elaboration of these and other related issues, see Barnett, A Consent Theory of Contract, 86 COLUM.  L. REV. 269, 287-89, 312-317 (1986); Barnett & Becker, Beyond Reliance: Promissory Estoppel, Contract Formalities and Misrepresentation, 15 HOFSTRA L. REV. 443 (1987).