Morality and the Law

H. L. A. Hart became Professor of Jurisprudence in Oxford in 1952, and since that time he has dominated and transformed legal theory. He has persuaded a generation of academic lawyers to face issues of legal and moral principle that their predecessors had ignored, and to attack these issues with the precision and skills of contemporary moral philosophy. I shall try to explain what these issues of principle are, and why Hart’s work has made so great a difference, but in order to do so I must describe the subject of jurisprudence itself, and the theories of jurisprudence that held the field at the time Hart began.

When lawyers argue cases, or advise clients, or draft laws to meet specific social goals, they face problems that are technical, in the sense that there is general agreement within the profession as to what sort of argument or evidence is relevant. But sometimes lawyers must deal with problems that are not technical in this sense, and there is no general agreement on how to proceed. One example is the ethical problem that is presented when a lawyer asks, not whether a particular law is effective, but whether it is fair. Another example is the conceptual puzzles that arise when lawyers try to describe the law in concepts that are unclear. A lawyer may want to say, for instance, that the law of torts holds men liable only for damage caused by their faults. Another lawyer may challenge this statement, and the issue between them may be a disagreement not about fact or doctrine, but about what fault means. Or two lawyers may disagree whether the Supreme Court, in the 1954 segregation case, was following established principles or making new law; and the issue between them may turn on what principles are and what it means to apply them. It is unclear how conceptual issues like these are to be resolved; certainly they lie beyond the ordinary techniques of the practicing lawyers.

Lawyers call these recalcitrant questions “jurisprudential,” and they disagree, as one would expect, on whether it is important to resolve them. Law schools generally provide special courses, called “Jurisprudence” or “Legal Theory” or something of the sort, devoted to their study, but since the distinguishing mark of these issues is just that there is no agreement on what sort of issues they are, and what techniques of study they require, these courses vary widely in the methods they use. The method chosen, moreover, influences the choice of the particular issues selected for study, though this choice is also affected by intellectual fashion and public affairs. Just now, for example, the question of whether men have a moral obligation to obey the law figures prominently in jurisprudence courses throughout the country; but a decade ago almost no one mentioned that issue.

In the 1950s, the dominant approach to jurisprudence in England and America was what one might call a professional approach. The lawyers who taught jurisprudence recognized that jurisprudential questions, like those I have listed, were troublesome just because they were not amenable to ordinary legal techniques; but they proposed nevertheless to meet this difficulty by picking out those aspects of the questions that could be treated with these techniques while ignoring the rest. When lawyers deal with the technical questions I mentioned, they use a combination of three particular skills. Lawyers are trained to analyze statutes and judicial opinions to extract legal doctrine from these official sources. They are trained to analyze complex factual situations in order to summarize the essential facts accurately. And they are trained to think in tactical terms, to design statutes and legal institutions that will bring about particular social changes decided upon in advance. The professional approach to jurisprudence tried to reformulate jurisprudential issues so that one or more of these skills could be brought to bear. This approach produced only the illusion of progress, and left the genuinely important issues of principle in the law untouched.

To sustain this serious charge I must describe where jurisprudence stood at mid-century. In England the subject was taught out of standard textbooks like Salmond on Jurisprudence and Paton on Jurisprudence. Most of these texts were devoted to what they called analytical jurisprudence, which they carefully distinguished from “ethical jurisprudence,” or the study of what the law ought to be. By analytical jurisprudence they meant the careful elaboration of the meaning of certain terms (like “fault,” “possession,” “ownership,” “negligence,” and “law”) that are fundamental to law in the sense that they appear not just in one or another branch but throughout the range of legal doctrine. These concepts, like those I mentioned earlier, are troublesome because lawyers use them even though they do not understand exactly what they mean.

But the English texts attacked these concepts, not by elucidating their meaning in ordinary speech, but rather by using conventional doctrinal methods to demonstrate their specifically legal meaning as revealed in cases and statutes. They studied the opinions of judges and legal experts and extracted from them summaries of the various legal rules and doctrines in which these troublesome concepts appeared; but they did little to connect these rules with the various non-legal judgments about fault, possession, etc., that the layman makes.

If we ask why lawyers argue about these concepts, however, we can see why this emphasis on doctrine appears irrelevant. A lawyer worries about the concept of fault, not because he is unaware of how the courts have used the term, or what the rules for determining legal fault are, but because he uses the non-legal concept of fault to justify or criticize the law. He believes—as a matter of habit or conviction—that it is morally wrong to punish someone for something not his fault; he wants to know whether the law offends this moral principle in holding an employer liable for what his employee does, or in holding a negligent driver liable for the death of a man he ran down if the injury was slight but the victim was a haemophiliac. He knows these facts of legal doctrine very well, but he is unclear whether the facts clash with the principle. Is harm a man’s fault if it is committed by someone under his charge, or if it results from his act because of circumstances he could not possibly foresee? These questions call for an analysis of the moral concept of fault, not the legal concept that the lawyer already understands; but it is just the moral use of the concept that the doctrinal approach of English jurisprudence ignored.

The record of American jurisprudence is more complex. It devoted itself largely to one issue that English theory had, in comparison, neglected: How do courts decide difficult or controversial lawsuits? Our courts had played a larger role than the English courts in reshaping nineteenth-century law to the needs of industrialization, and our constitution made legal issues out of problems that in England were political only. In England, for example, the issue of whether minimum wage legislation is fair was a political issue, but in America it was a constitutional, that is, judicial, issue as well. American lawyers were therefore pressed harder to furnish an accurate description of what the courts were doing, and to justify this if they could; the call was most urgent when the courts appeared to be making new and politically controversial law instead of simply applying old law as orthodox legal theory required.

Early in this century, John Chipman Gray and, later, Oliver Wendell Holmes published skeptical accounts of the judicial process, debunking the orthodox doctrine that judges merely apply existing rules. This skeptical approach broadened, in the 1920s and ’30s, into the powerful intellectual movement called “legal realism.” Its leaders (Jerome Frank, Karl Llewelyn, Wesley Sturges, and Morris and Felix Cohen, among others) argues that orthodox theory had gone wrong because it had taken a doctrinal approach to jurisprudence, attempting to describe what judges do by concentrating on the rules they mention in their decision. This is an error, the realists argued, because judges actually decide cases according to their own political or moral tastes, and then choose an appropriate legal rule as a rationalization. The realists asked for a “scientific” approach that would fix on what judges do, rather than what they say, and the actual impact their decisions have on the larger community.

The main line of American jurisprudence followed this call for realism, and avoided the doctrinal approach of the English texts. It emphasized the two other professional skills—the lawyer’s skills at marshalling facts and at designing tactics for social change. We can trace the later impact of realism more clearly if we distinguish these two techniques. The emphasis on facts developed into what Roscoe Pound of Harvard called sociological jurisprudence; he meant the careful study of legal institutions as social processes, which treats a judge, for example, not as an oracle of doctrine, but as a man responding to various sorts of social and personal stimuli. Some lawyers, like Jerome Frank and Pound himself, attempted to carry out this sort of study, but they discovered that lawyers do not have the training or statistical equipment necessary to describe complex institutions in other than an introspective and limited way. Sociological jurisprudence therefore became the province of sociologists.

The emphasis on tactics had a more lasting effect within the law schools. Scholars like Myres McDougal and Harold Lasswell at Yale, and Lon L. Fuller, Henry Hart, and Albert Sachs at Harvard, though different from one another, all insisted on the importance of regarding the law as an instrument for moving society toward certain large goals, and they tried to settle questions about the legal process instrumentally, by asking which solutions best advanced these goals.

But this emphasis on fact and strategy ended by distorting jurisprudential issues in much the same way as the English doctrinal approach distorted them, that is, by eliminating just those issues of moral principle that form their core. This failure emerges if we consider in greater detail the central problem that the sociologists and instrumentalists discussed: Do judges always follow rules, even in difficult and controversial cases, or do they sometimes make up new rules and apply them retroactively?

Lawyers have argued this issue for decades, not because they are ignorant of the sorts of decisions judges make or the reasons they give, but because they are unclear what the concept of following rules really means. In easy cases (when a man is charged with violating a statute that forbids driving over sixty miles an hour, for example) it seems right to say that the judge is simply applying a prior rule to a new case. But can we say this when the Supreme Court overturns precedent and orders the schools desegregated, or outlaws procedures that for decades the police have been using and the courts condoning? In these dramatic cases the Court gives reasons—it does not cite statutes, but it does appeal to principles of justice and policy. Does that mean that the Court is following rules after all, although of a more general and abstract quality? If so, where do these abstract rules come from, and what makes them valid? Or does it mean that the Court is deciding the case in accordance with its own moral and political beliefs?

The lawyers and laymen who ask these questions are not worrying aimlessly or out of idle curiosity; they know that judges wield great political power, and they are concerned with whether that power is justified, in general or in particular cases. They are not necessarily persuaded that judges who make up new rules are acting improperly. But they want to know how far the justification for judicial power available in easy cases—that the judge is applying standards already established—extends to hard ones, and therefore how much, and what sort of, supplementary justification these hard cases require.

The question of justification has important ramifications, because it affects not only how far judicial authority extends, but the extent of an individual’s political and moral obligation to obey judge-made law. It also affects the grounds on which a controversial opinion may be challenged. If it makes sense to say that a judge should follow existing standards in hard cases, then it makes sense for a conscientious objector to argue that the judge has made a mistake in the law when he holds the draft constitutional. But if judges can only make new law in hard cases, that claim is nonsense. So though the question of whether judges follow rules may sound linguistic, it reveals concerns that are in the last degree practical.

I have spelled out these implications to show that here, as in the case of the concept of fault, there are issues of moral principle that lie beneath an apparently linguistic problem. Critics of law accept, again either by habit or through conviction, the principle that a judicial decision is fairer if it represents the application of established standards rather than the imposition of new ones. But they are unclear what counts as applying established standards, and they express this uncertainty by asking whether judges are really following rules, in some sense at least, even in novel cases. Jurisprudence should respond to this concern by exploring the nature of moral argument, trying to clarify the principle of fairness which the critics have in mind to see whether judicial practice does, in fact, satisfy that principle.

But American jurisprudence made no such attempt. The sociologists, for their part, refused to talk about following rules, on the ground that that concept was too vague to be studied in experimental or quantitative ways. The very fact that men are unable to agree on what following a rule means, they pointed out, disqualifies that concept for science; if each investigator were to use his own sense of the term, there would be no objective data and no joint progress. So sociologists like Glendon Schubert, C. Herman Pritchett, and Stuart Nagel substituted questions that seemed related and more precise: Do judges from particular economic or social backgrounds, or from particular sorts of legal practice, or with particular political affiliations, or particular value schemes, tend to decide in favor of corporate defendants? Do the judges in the Supreme Court form parties that stick together when cases involving race, or labor unions, or antitrust are decided? These empirical questions seemed relevant, because if social background or prior allegiance determines a judge’s decision, this suggests that he is not following rules.

But in fact this information, though interesting and useful for other purposes, throws little light on the issues of principle that inspired the original question. Lawyers need no evidence to show that judges disagree, and that their decisions often reflect their background and temperament. They are puzzled, however, as to whether this means only that the judges differ on the nature and point of fundamental legal principles, or whether it demonstrates that there are no such principles. If it means the former, then this argues that judges are trying to follow rules, as they see them, and that people who disagree with their decisions may still be right on the law; if it means the latter then this argument, as I said, is absurd. The lawyers are also uncertain whether the fact of divergence, on either account, is to be regretted, or accepted as inevitable, or applauded as dynamic, and how all this connects with the crucial issues of political obligation and law enforcement that lawyers face. The sociological approach, in reframing the question, eliminated just those aspects that bear on all these issues.

The instrumental branch of post-realism also reframed the question, though in a different way. Henry Hart and Sachs, in their brilliant materials on the legal process, suggested that conceptual questions about rules could be bypassed by putting the issue this way: How should judges reach their decisions in order best to advance the goals of the legal process? But their hope that this would avoid the puzzles about rules was vain, because it proved impossible to state the goals of the legal process without those problems appearing at a later stage. If we state the goal of the process in some vacuous way (by saying that the law should do justice, or advance the just state) then the question is inescapable whether, as many suppose, justice requires decisions according to prior rules, and this question, in turn, requires an analysis of what it is to follow a rule. If we attempt to state some more particular or precise goal (that the legal process should increase the gross national product, for example) then the exercise loses its point, because there is no warrant for assuming that any such particular goal is the proper exclusive concern of law.

We may argue (as some writers did) that the law will be economically more efficient if judges are allowed to take the economic impact of their decision into account; but this will not answer the question of whether it is fair for them to do so, or whether we can regard economic standards as part of existing law, or whether decisions based on economic impact carry more or less moral weight for that reason. Suppose that a judge is persuaded, for example, that the automobile industry will prosper if he repeals an old rule and invents a new one for its benefit, and that the general economy will benefit if the automobile industry does. Is this a good reason for changing the rule? We cannot decide this sort of question by analysis that simply relates means to ends.

So the various branches of the professional approach to jurisprudence failed for the same underlying reason. They ignored the crucial fact that jurisprudential issues are at their core issues of moral principle, not legal fact or strategy. They buried these issues by insisting on a conventional legal approach. But if jurisprudence is to succeed, it must expose these issues and attack them as issues of moral theory.

The single fact explains H. L. A. Hart’s importance and success. Hart is a gifted moral philosopher; he has an extraordinary instinct for issues of principle, and a marvelous lucidity in setting them out. In his first book, The Concept of Law, for example, he raised the issue of whether judges follow rules in a way that made plain the connection between this problem and the moral issue of when it is proper for one man to charge another with an obligation. He offered an analysis of the rules that our community follows, as a matter of convention, in making and criticizing arguments about moral obligation, and argued that judges follow much the same rules in reasoning about legal obligation. In another book, Causation in the Law, Hart and a co-author, A. M. Honore, discussed the conceptual puzzles about fault, which I mentioned earlier, but unlike Hart’s predecessors, they undertook to explain the ordinary as well as the strictly legal meanings of that concept. Like Hart’s colleagues in the Oxford school of philosophy, particularly J. L. Austin, they used the study of ordinary language to show the ways in which members of the community habitually ascribe fault and responsibility to one another; and they then used these conventional judgments to explain, for example, the legal rule that holds a man fully liable if he injures a haemophiliac.

They pointed out that ordinary language draws a distinction between unusual circumstances existing at the time a man acts and unlikely coincidences that arise afterward. It distinguishes, for example, the case in which the careless driver slightly injures a man who dies because he is haemophiliac from the case in which a careless driver slightly injures a man who dies of blood poisoning through a doctor’s negligence. Most people would say that in the first case the careless driver caused the death and that it was his fault; but they would not say this in the second case. This distinction, in turn, reflects a popular conception of causation: the ordinary man distinguishes a causally effective act as an act that operates upon a stage already set; contemporaneous circumstances, like the blood disease, are part of the stage setting and not competing causes. But later events, like a doctor’s negligence, are interventions that break the causal chain. So, the legal rule is comprehensible as an extension of popular theories of morality and cause.

But Hart has not been content simply to explain the law by showing how it incorporates the ordinary man’s moral judgments. He views this sort of analysis as a necessary preliminary to a critical evaluation both of the law and of the popular morality on which it rests. Until we are clear what moral practice or judgment the law reflects, we cannot criticize it intelligently; but once we are clear about this it remains to ask whether this practice or judgment is sensible, or sound, or consistent with the other principles the law claims to serve.

Hart’s most recent book, Punishment and Responsibility, is an excellent example of this process of criticism. The book reprints a series of his essays on jurisprudential issues in criminal law; most of these essays treat the problem of whether a man should be excused from liability for a criminal act because of his mental state. Should he be excused (or should his penalty be reduced) if his act was an accident, or if he acted negligently rather than deliberately, or if he was mentally ill? The law generally grants an excuse, or at least mitigates punishment, in such cases, but some contemporary critics argue that this policy is wrong.

If the criminal law aimed at vengeance and retribution, they say, the point of these mental defenses would be obvious, for there is no satisfaction to be gained from taking revenge on someone who acted by mistake or was insane. But if the law’s goals are only to prevent further harm by the criminal, and deter others by his example, the defenses seem counterproductive. We could prevent harm more by jailing the accident-prone driver than by jailing the man who murders his father for an inheritance; and we would increase the deterrent power of the law if we accepted no excuses whatever, and did not encourage potential criminals to hope that they could fake an insanity plea if caught. So the critics argue that the law should sharply limit these defenses, on the ground that the defenses increase the expense of trials and legal education, that they are abused, and that their purpose in an enlightened criminal system seems obscure.

Hart disagrees. He begins his response by reminding us that it is wrong to assume that the criminal law (or any other branch of the law) has a set of goals that are overriding, in the sense that every feature of the law must be tailored to these goals. The criminal law aims at preventing crime, to be sure, but it must pursue this aim subject to principles that may limit its efficiency in reaching them; it would be wrong to punish an innocent man as a hostage even if to do so would in fact reduce crime. We must understand the mental defenses in that light, so the fact that they may interfere with crime prevention is not a conclusive argument. But this point is only negative and leaves open the question whether the mental defenses are in fact justified, or ought to be changed. Hart attacks this issue in the manner I described; he begins by asking whether the mental defenses reflect any moral tradition, or any general aim or policy, of the community.

He considers first the suggestion of some criminal lawyers, like Professor Jerome Hall, that the point of the mental defense is to insure that the law punishes as criminals only men who are morally blameworthy on conventional standards. This attractive notion has initial plausibility. In ordinary life we do not blame someone who has done harm if he acted unintentionally or inadvertently (unless, perhaps, he was also careless); nor do we do so if we believe that he is suffering from a serious mental disorder. It is therefore plausible to suppose that judges and legislators would carry these attitudes into the criminal law, in the form of a doctrine that men should not be punished under these conditions, even though it would be more efficient to do so.

Hart rejects this theory, however, on the ground that there are many crimes—for example, the failure to abide by the English rail transport regulations—which are not in themselves morally blameworthy. The existence of such crimes, he argues, shows that the law has no general purpose to condemn only blameworthy acts, and so proves this could not be the point of the mental defenses.

But Hart is wrong $$$ the blameworthiness theory in this way; he is confusing, I think, two grounds on which a violation of law might be morally wrong. It might be wrong to break a law because the act the law condemns (killing, for example) is wrong in itself. Or it might be wrong, even though the act condemned is not wrong in itself, just because the law forbids it, and railroad legislation may be a case in point. Perhaps it is doubtful whether it is wise or fair for England to nationalize the railroads; it might still be true that once the law is passed everyone has a moral obligation to obey it.

Of course, it does not follow that a man is morally to blame every time he does what the law prohibits. He might not be blameworthy because the law is so unfair or unjust that the normal moral obligation to obey the law has lapsed, an argument that is made by some conscientious objectors to the draft. (It is arguable that the point of the due process and other clauses of the United States Constitution is to guard against a man’s being punished in that case.) Or he may not be blameworthy because his act was committed by accident, or inadvertently, or because he had a mental disease and so was not responsible for his conduct. Professor Hall’s argument, which Hart rejects too quickly, is that the point of the mental defenses is to guard against punishing a man in this case.

Even if Hall is right, however, the critical questions remain, because we must ask whether our conventional attitudes about blame and punishment are really relevant to the law. Those who doubt the value of the mental defenses argue that since the goal of the criminal law is to reform and deter, and not to punish, these conventions are irrelevant, and the mental defenses should be abandoned. They drive this point home by proposing to drop the word “punishment,” and speak of “treatment” instead. If a man has committed a crime, they argue, then how society should treat him—whether he should be confined, or hospitalized, or released—should depend on which course would best prevent a repetition. It confuses the issue, on this account, to ask whether he was morally blameworthy in doing what he did, because treatment might be unnecessary even if he was, and advisable even if he was not. We must ask whether the mental defenses serve any purpose that is relevant to this revised notion of the criminal law.

In one of the early essays, “Legal Responsibility and Excuses,” reprinted in the present book, Hart offers this suggestion: The mental defenses increase each man’s control over his own fate, by reducing the number of occasions on which the law will interfere with his freedom in a way he could not have predicted from his own deliberate acts. If the defenses were repealed, we would all have to live with the feat that some accident or piece of inadvertence would send us to jail, or involve us in a lengthy, expensive, and degrading trial. By virtue of the defenses, we can count on the fact that in general we will be prosecuted only for acts done with the awareness that prosecution might follow, which has the incidental benefit that those who are punished have had at least the satisfaction of taking and carrying out a decision to break the law.

But this is a weak argument, if it is addressed entirely to the increased personal security that the mental defenses afford, because this increased security is minimal. After all, the community has accepted a great many decisions that make life much more perilous, such as the decision to foster competition in commerce, to permit automobiles and to wage war. These decisions vastly increase the likelihood that particular men will suffer harm which they could not foresee and which does not flow from their deliberate acts; but society still accepts these decisions, and runs these risks, for the sake of some goal of efficiency or profit or national policy. If, as Hart is willing to suppose, eliminating the mental defenses would increase the efficiency of the criminal law in preventing crime, then this would add to the ordinary citizen’s personal safety and control over his own fate in a measure that would presumably outweigh the increased risk of his being subject to liability for an accident.

Hart is more successful, I think, when he provides a different and more general justification for the mental defenses in a later essay, “Punishment and the Elimination of Responsibility”: “Human society is a society of persons; and persons do not view themselves or each other merely as so many bodies moving in ways which are sometimes harmful and have to be prevented or altered. Instead persons interpret each other’s movements as manifestations of intentions….” Elsewhere, and in the same vein, he says that the law would treat people as means rather than ends if it abandoned these defenses.

These statements unite the legal doctrines with a wide range of moral traditions. The principle they urge is that the government must treat its citizens with the respect and dignity that adult members of the community claim from each other. The government may restrain a man for his own or the general good, but it may do so only on the basis of his behavior, and it must strive to judge his behavior from the same standpoint as he judges himself, that is, from the standpoint of his intentions, motives, and capacities. Men generally feel that they have chosen to act as they have, but they do not feel this to be so in particular circumstances of accident, compulsion, duress, or disease. And each of us makes this distinction not only for himself but in judging how to respond to others he regards with any respect. Even a dog, Holmes said, knows the difference between being kicked and being tripped over.

The criminal law might be more efficient if it disregarded this troublesome distinction, and jailed men or forced them to accept treatment whenever this seemed likely to decrease future crime. But that, as Hart’s principle suggests, would cross the line that separates treating someone else as a fellow human being from treating him as a resource for the benefit of others, and there can be no more profound insult, under the conventions and practices of our community, than that. The insult is as great whether the process is called one of punishment or treatment. It is true that we sometimes restrain and give treatment to a man just because we believe that he does not have control over his conduct. We do this under civil commitment statutes and typically, after a man is acquitted of a serious crime on grounds of insanity. But we ought to recognize the compromise with principle that this policy involves; and should treat a man against his will only when the danger he presents is vivid, not whenever we calculate that it would probably reduce crime if we did.

Of course this line of argument raises more questions than it settles. Some philosophers think, on the basis of contemporary physiology and psychology, that this phenomenological distinction between choice and compulsion makes no sense. They believe that all human behavior is determined by factors beyond individual control, so that the feelings of free choice we often have are just illusions. But the scientific evidence for this is far from conclusive, and even those who think it weighty must decide how the law should respond until the case is proved (if it ever is or can be). If we accepted the view that all behavior is determined, for example, would it follow that we should abandon entirely the idea that human beings have rights that their government is morally bound to respect? If we chose not to go so far, either because the scientific evidence is inconclusive, or because we are reluctant to abandon the notion of rights in any event, would it not be inconsistent to abandon the mental defenses in the name of science? If, on the other hand, we retain these defenses, and accept as their basis the phenomenological distinction between choosing or not choosing to do something, as this argument suggests, how should that guide our approach to troublesome cases, like that of the psychopath? Is the psychopath in control of himself, according to conventional standards of behavior, or is his case a half-way house, which explains our confusion? There is no space here to puruse these issues (some of which Hart discusses), and I mention them only to show that the approach to jurisprudence that emphasizes principle cannot stop at simply showing the links between legal and social practice, but must continue to examine and criticize social practice against independent standards of consistency and sense.

The mental defenses are not the only controversial aspects of the criminal law. Just now the public is even more concerned with those rules of criminal procedure—regarding interrogation, confessions, and preventive detention, for example—that protect the alleged criminal at some cost to police efficiency. It might be useful to point out the value of a more philosophic approach to these issues than academic lawyers have yet provided. So far the liberal position has been presented chiefly in instrumental terms. The liberal argues that the proper goals of the criminal law include the the protection of individual freedom as well as the prevention of crime, and that the procedural safeguards strike a balance between these two goals. But this way of putting the point suggests that a balance between the two goals is in order; it encourages others to ask why the majority of law-abiding citizens should not strike the balance further on the side of its own protection.

The liberal is placed in a difficult position by that question. He might reply that he, personally, values the liberty of others more than his own increased security, but he would have to admit that this is a minority position. He might, argue that the majority would itself be better off in the long run by promoting freedom at the cost of a little security, but this argument, though popular with liberals, is plainly wrong. The criminal law presents more of a threat to the black narcotics addict than to the middle-class white, and there is little reason to suppose that interrogating the former without counsel, or locking him up pending trial, will even in the long run affect the latter’s liberty.

Here again Hart’s general approach is helpful. The liberal position should be argued, it would suggest, by emphasizing moral principles that act as constraints on the law rather than citing the law’s conflicting goals. It should fix on legal doctrines that are embedded in our traditions (like the doctrines that no man may be forced to condemn himself, and that a man is presumed innocent until proved guilty) to support the claim that society has no right to interrogate a man without a lawyer, and that an accused suspect is entitled to be free before his trial, whether the majority benefits or not. Of course there may be conflicts between these principles and practical needs, but these are not occasions for fair compromise, but rather, if the principles must be dishonored, for shame and regret.

Those who take a different view, and want to increase police efficiency, accept the doctrines I mentioned, such as the privilege against self-incrimination, but deny that they guarantee the particular rights the liberals claim. They would argue that this privilege, for example, protects a man from being tortured to extract a confession, but does not entitle him to withdraw a voluntary confession just because it was given without reflection. So the controversy must be pressed in philosophical terms: does the use of unadvised confessions, or preventive detention, contradict the moral principles underlying the established doctrines? I think they do, but it remains for jurisprudence to construct the bridges between legal and moral theory that support that claim.

Perhaps the principle Hart cited, that the government must show the minimum of respect even to accused criminals and treat them as humans rather than as opportunities, will help establish that a contradiction exists. This principle, for example, informs the doctrine that a man is innocent until proved guilty, and helps to explain why it seems wrong to imprison a man awaiting trial on the basis of a prediction that he might commit further crimes if released on bail. For any such prediction, if it is sound, must be based on the view that an individual is a member of a class having particular features, which class is more likely than others to commit crime. The prediction, that is, must be actuarial, like the prediction an insurance company makes about the likelihood of teenagers to have automobile accidents. But it is unjust to put someone in jail on the basis of a judgment about a class, however accurate, because that denies his claim to respect and treatment as an individual.

Alan Dershowitz, in a recent article in this journal, said that the judgment that someone will commit a crime is like the judgment that he has committed a crime, in that both are judgments of probability that might be wrong. But this ignores the crucial difference that the latter is a judgment entirely about what a particular person did; and, with rare exceptions, in deciding this we do not allow evidence about what people like him commonly do. This sort of evidence, however, is necessary to the sort of prediction Dershowitz describes, and so that prediction necessarily involves a break with the requirement of individual respect that a judgment of guilt does not.

It does not follow that we should never try to decide whether a given man will probably commit a crime. It does follow that when we act on such a prediction, and jail him in advance, we should be aware that we are acting against a principle that is not only a legal tradition but one of the moral ground rules of the community.

It is the responsibility of lawyers to hold legal policy and legal action to principle. But this is a responsibility that lawyers have found it difficult to fulfill, in large part because important principles have been obscured by conceptual confusions and poor legal theory. H. L. A. Hart has taught lawyers to recognize issues of moral principle that have been ignored, and has shown, by his example, how steady and clear thought can clarify these principles and improve our law as an instrument of justice. His work now dominates the discipline that his genius has transformed.