Anthony Kennedy
Anthony Kennedy; drawing by David Levine


There is a particular way of reading and enforcing a political constitution, which I call the moral reading. Most contemporary constitutions declare individual rights against the government in very broad and abstract language, like the First Amendment of the United States Constitution, which provides that Congress shall make no law abridging “the freedom of speech.” The moral reading proposes that we all—judges, lawyers, citizens—interpret and apply these abstract clauses on the understanding that they invoke moral principles about political decency and justice. The First Amendment, for example, recognizes a moral principle—that it is wrong for government to censor or control what individual citizens say or publish—and incorporates it into American law. So when some novel or controversial constitutional issue arises—about whether, for instance, the First Amendment permits laws against pornography—people who form an opinion must decide how an abstract moral principle is best understood. They must decide whether the true ground of the moral principle that condemns censorship, in the form in which this principle has been incorporated into American law, extends to the case of pornography.

The moral reading therefore brings political morality into the heart of constitutional law.1 But political morality is inherently uncertain and controversial, so any system of government that makes such principles part of its law must decide whose interpretation and understanding will be authoritative. In the American system judges—ultimately the justices of the Supreme Court—now have that authority, and the moral reading of the Constitution is therefore said by its critics to give judges absolute power to impose their own moral convictions on the public. I shall shortly try to explain why that crude charge is mistaken. I should make plain first, however, that there is nothing revolutionary about the moral reading in practice. So far as American lawyers and judges follow any coherent strategy of interpreting the Constitution at all, they already use the moral reading.

That explains why both scholars and journalists find it reasonably easy to classify judges as “liberal” or “conservative”: the best explanation of the differing patterns of their decisions lies in their different understandings of central moral values embedded in the Constitution’s text. Judges whose political convictions are conservative will naturally interpret abstract constitutional principles in a conservative way, as they did in the early years of this century, when they wrongly supposed that certain rights over property and contract are fundamental to freedom. Judges whose convictions are more liberal will naturally interpret those principles in a liberal way, as they did in the halcyon days of the Warren Court. The moral reading is not, in itself, either a liberal or a conservative charter or strategy. It is true that in recent decades liberal judges have ruled more statutes or executive orders unconstitutional than conservative judges have. But that is because conservative political principles for the most part either favored or did not strongly condemn measures that could reasonably be challenged on constitutional grounds in those decades.

There have been exceptions to that generalization. Conservatives strongly disapprove, on moral grounds, the affirmative action programs that give certain advantages to minority applicants to universities or jobs, and conservative justices have not hesitated to follow their understanding of what the moral reading required in such cases.2 The moral reading helps us to identify and explain not only these large-scale patterns, moreover, but also more fine-grained differences in constitutional interpretation that cut across the conventional liberal-conservative divide. Conservative judges who particularly value freedom of speech, or think it particularly important to democracy, are more likely than other conservatives to extend the First Amendment’s protection to acts of political protest, even for causes that they despise, as the Supreme Court’s decision protecting flag-burners shows.3

So, to repeat, the moral reading is not revolutionary in practice. Lawyers and judges, in their day-to-day work, instinctively treat the Constitution as expressing abstract moral requirements that can only be applied to concrete cases through fresh moral judgments. As I shall argue later, they have no other real option except to do so. But it would indeed be revolutionary for a judge openly to recognize the moral reading, or to admit that it is his or her strategy of constitutional interpretation, and even scholars and judges who come close to recognizing it shrink back, and try to find other, usually metaphorical, descriptions of their own practice.

There is therefore a striking mismatch between the role the moral reading actually plays in American constitutional life and its reputation. It has inspired all the greatest constitutional decisions of the Supreme Court, and also some of the worst. But it is almost never acknowledged as influential even by constitutional experts, and it is almost never openly endorsed even by judges whose arguments are incomprehensible on any other understanding of their responsibilities. On the contrary, the moral reading is often dismissed as an “extreme” view that no really sensible constitutional scholar would entertain. It is patent that judges’ own views about political morality influence their constitutional decisions, and though they might easily explain that influence by insisting that the Constitution demands a moral reading, they never do. Instead, against all evidence, they deny the influence and try to explain their decisions in other—embarrassingly unsatisfactory—ways. They say they are just giving effect to obscure historical “intentions,” for example, or just expressing an overall but unexplained constitutional “structure” that is supposedly explicable in nonmoral terms.


This mismatch between role and reputation is easily explained. The moral reading is so thoroughly embedded in constitutional practice and is so much more attractive, on both legal and political grounds, than the only coherent alternatives, that it cannot readily be abandoned, particularly when important constitutional issues are in play. But the moral reading nevertheless seems intellectually and politically discreditable. It seems to erode the crucial distinction between law and morality by making law only a matter of which moral principles happen to appeal to the judges of a particular era. It seems grotesquely to constrict the moral sovereignty of the people themselves—to take out of their hands, and remit to a professional elite, exactly the great and defining issues of political morality that the people have the right and the responsibility to decide for themselves.

That is the source of the paradoxical contrast between mainstream constitutional practice in the United States, which relies heavily on the moral reading of the Constitution, and mainstream constitutional theory, which wholly rejects that reading. The confusion has had serious political costs. Conservative politicians try to convince the public that the great constitutional cases turn not on deep issues of political principle, which they do, but on the simpler question of whether judges should change the Constitution by fiat or leave it alone.4 For a time this view of the constitutional argument was apparently accepted even by some liberals. They called the Constitution a “living” document and said that it must be “brought up to date” to match new circumstances and sensibilities. They said they took an “active” approach to the Constitution, which seemed to suggest reform, and they accepted John Ely’s characterization of their position as a “noninterpretive” one, which seemed to suggest inventing a new document rather than interpreting the old one.5 In fact, this account of the argument was never accurate. The theoretical debate was never about whether judges should interpret the Constitution or change it—almost no one really thought the latter—rather it was about how it should be interpreted. But conservative politicians exploited the simpler description, and they were not effectively answered.

The confusion engulfs the politicians as well. They promise to appoint and confirm judges who will respect the proper limits of their authority and leave the Constitution alone, but since this misrepresents the choices judges actually face, the politicians are often disappointed. When Dwight Eisenhower, who denounced what he called judicial activism, retired from office in 1961, he told a reporter that he had made only two big mistakes as President—and that they were both on the Supreme Court. He meant Chief Justice Earl Warren, who had been a Republican politician when Eisenhower appointed him to head the Supreme Court, but who then presided over one of the most “activist” periods in the Court’s history, and Justice William Brennan, another politician who had been a state court judge when Eisenhower appointed him, and who became one of the most liberal and explicit practitioners of the moral reading of the Constitution in modern times.

Presidents Ronald Reagan and George Bush were both intense in their outrage at the Supreme Court’s “usurpation” of the people’s privileges. They said they were determined to appoint judges who would respect rather than defy the people’s will. In particular, they (and the platform on which they ran for the presidency) denounced the Court’s 1973 Roe v. Wade decision protecting abortion rights, and promised that their appointees would reverse it. But when the opportunity to do so came, three of the justice Reagan and Bush had appointed between them voted, surprisingly, not only to retain that decision in force, but to provide a legal basis for it that much more explicitly adopted and relied on a moral reading of the Constitution. The expectations of politicians who appoint judges are often defeated in that way, because the politicians fail to appreciate how thoroughly the moral reading, which they say they deplore, is actually embedded in constitutional practice. Its role remains hidden when a judge’s own convictions support the legislation whose constitutionality is in doubt—when a justice thinks it morally permissible for the majority to criminalize abortion, for example. But the ubiquity of the moral reading becomes evident when some judge’s convictions of principle—identified, tested, and perhaps altered by experience and argument—bend in an opposite direction, because then enforcing the Constitution must mean, for that judge, telling the majority that it cannot have what it wants.


Senate hearings considering Supreme Court nominations tend toward the same confusion. These events are now thoroughly researched and widely reported by the press, and they are often televised. They offer a superb opportunity for the public to participate in the constitutional process. But the mismatch between actual practice and conventional theory cheats the occasion of much of its potential value. (The hearings provoked by President Bush’s nomination of Judge Clarence Thomas to the Supreme Court, are a clear example.) Nominees and legislators all pretend that hard constitutional cases can be decided in a morally neutral way, by just keeping faith with the “text” of the document, so that it would be inappropriate to ask the nominee any questions about his or her own political morality. (It is ironic that Justice Thomas, in the years before his nomination, gave more explicit support to the moral reading than almost any other well-known constitutional lawyer has; he insisted that conservatives should embrace that interpretive strategy and harness it to a conservative morality.) Any endorsement of the moral reading—any sign of weakness for the view that constitutional clauses are moral principles that must be applied through the exercise of moral judgment—would be suicidal for the nominee and embarrassing for his questioners. In recent years, only the hearings that culminated in the defeat of Robert Bork seriously explored issues of constitutional principle, and they did so only because Judge Bork’s opinions about constitutional law were so obviously the product of a radical political morality that his convictions could not be ignored. In the confirmation proceedings of the present Justices Anthony Kennedy, David Souter, Thomas, Ruth Bader Ginsburg, and Stephen Breyer, however, the old fiction was once again given shameful pride of place.

The most serious result of this confusion, however, lies in the American public’s misunderstanding of the true character and importance of its constitutional system. As I have argued elsewhere, the American ideal of government not only under law but under principle as well is the most important contribution our history has given to political theory. Other nations and cultures realize this, and the American ideal has increasingly and self-consciously been adopted and imitated elsewhere. But we cannot acknowledge our own contribution, or take the pride in it, or care of it, that we should.

That judgment will appear extravagant, even perverse, to many lawyers and political scientists. They regard enthusiasm for the moral reading, within a political structure that gives final interpretive authority to judges, as elitist, antipopulist, antirepublican, and antidemocratic. That view rests on a popular but unexamined assumption about the connection between democracy and majority will, an assumption that American history has in fact consistently rejected. When we understand democracy better, we see that the moral reading of a political constitution is not antidemocratic but, on the contrary, is practically indispensable to democracy. I do not mean that there is no democracy unless judges have the power to set aside what a majority thinks is right and just. Many institutional arrangements are compatible with the moral reading, including some that do not give judges the power they have in the American structure. None of these varied arrangements, however, is in principle more democratic than others. Democracy does not insist on judges having the last word, but it does insist that they must not have it.

The Moral Reading

The clauses of the American Constitution that protect individuals and minorities from government are found mainly in the so-called Bill of Rights—the first ten amendments to the document—and the further amendments added after the Civil War. (I shall sometimes use the phrase “Bill of Rights,” inaccurately, to refer to all the provisions of the Constitution that establish individual rights, including the Fourteenth Amendment’s protection of citizens’ privileges and immunities and its guarantee of due process and equal protection of the laws.) Many of these clauses are drafted in exceedingly abstract moral language. The First Amendment refers to the “right” of free speech, for example, the Fifth Amendment to the process that is “due” to citizens, and the Fourteenth to protection that is “equal.” According to the moral reading, these clauses must be understood in the way their language most naturally suggests: they refer to abstract moral principles and incorporate these by reference, as limits on government’s power.

There is of course room for disagreement about the right way to restate these abstract moral principles, so as to make their force clearer to us, and to help us to apply them to more concrete political controversies. I favor a particular way of stating the constitutional principles at the most general possible level. I believe that the principles set out in the Bill of Rights, taken together, commit the United States to the following political and legal ideas: government must treat all those subject to its dominion as having equal moral and political status; it must attempt, in good faith, to treat them all with concern; and it must respect whatever individual freedoms are indispensable to those ends, including but not limited to the freedoms more specifically designated in the document, such as the freedoms of speech and religion. Other lawyers and scholars who also endorse the moral reading might well formulate the constitutional principles, even at a very general level, differently and less expansively than I just have however, and though here I want to explain and defend the moral reading, not my own interpretations under it, I should say something about how the choice among competing formulations should be made.

Of course the moral reading is not appropriate to everything a constitution contains. The American Constitution includes a great many clauses that are neither particularly abstract nor drafted in the language of moral principle. Article II specifies, for example, that the President must be at least thirty-five years old, and the Third Amendment insists that government may not quarter soldiers in citizens’ houses in peacetime. The latter may have been inspired by a moral principle: those who wrote and enacted it might have been anxious to give effect to some principle protecting citizens’ rights to privacy, for example. But the Third Amendment is not itself a moral principle: its content is not a general principle of privacy. So the first challenge to my own interpretation of the abstract clauses might be put this way. What argument or evidence do I have that the equal protection clause of the Fourteenth Amendment (for example), which declares that no state may deny any person equal protection of the laws, has a moral principle as its content though the Third Amendment does not?

This is a question of interpretation or, if you prefer, translation. We must try to find language of our own that best captures, in terms we find clear, the content of what the “framers” intended it to say. (Constitutional scholars use the word “framers” to describe, somewhat ambiguously, the various people who drafted and enacted a constitutional provision.) History is crucial to that project, because we must know something about the circumstances in which a person spoke to have any good idea of what he meant to say in speaking as he did. We find nothing in history, however, to cause us any doubt about what the framers of the Third Amendment meant to say. Given the words they used, we cannot sensibly interpret them as laying down any moral principle at all, even if we believe they were inspired by one. They said what the words they used would normally be used to say: not that privacy must be protected, but that soldiers must not be quartered in houses in peacetime.

The same process of reasoning—about what the framers presumably intended to say when they used the words they did—yields an opposite conclusion about the framers of the equal protection clause, however. Most of them no doubt had fairly clear expectations about what legal consequences the Fourteenth Amendment would have. They expected it to end certain of the most egregious Jim Crow practices of the Reconstruction period. They plainly did not expect it to outlaw official racial segregation in school—on the contrary, the Congress that adopted the equal protection clause itself maintained segregation in the District of Columbia school system. But they did not say anything about Jim Crow laws or school segregation or homosexuality or gender equality, one way or the other. They said that “equal protection of the laws” is required, which plainly describes a very general principle, not any concrete application of it.

The framers meant, then, to enact a general principle. But which general principle? That further question must be answered by constructing different elaborations of the phrase “equal protection of the laws,” each of which we can recognize as a principle of political morality that might have won their respect, and then by asking which of these it makes most sense to attribute to them, given everything else we know. The qualification that each of these possibilities must be recognizable as a political principle is absolutely crucial. We cannot capture a statesman’s efforts to lay down a general constitutional principle by attributing to him something neither he nor we could recognize as a candidate for that role. But the qualification will typically leave many possibilities open. It was once debated, for example, whether the framers intended to stipulate, in the equal protection clause, only the relatively weak political principle that laws must be enforced in accordance with their terms, so that legal benefits conferred on everyone, including blacks, must not be denied, in practice, to anyone.

History seems decisive that the framers of the Fourteenth Amendment did not mean to lay down only so weak a principle as that one, however, which would have left states free to discriminate against blacks in any way they wished so long as they did so openly. Congressmen of the victorious nation, trying to capture the achievements and lessons of a terrible war, would be very unlikely to settle for anything so limited and insipid, and we should not take them to have done so unless the language leaves no other interpretation plausible. In any case, constitutional interpretation must take into account past legal and political practice as well as what the framers themselves intended to say, and it has now been settled by unchallengeable precedent that the political principle incorporated in the Fourteenth Amendment is not that very weak one, but something more robust. Once that is conceded, however, then the principle must be something much more robust, because the only alternative, as a translation of what the framers actually said in the equal protection clause, is that they declared a principle of quite breathtaking scope and power: the principle that government must treat everyone as of equal status and with equal concern.

Two important restraints sharply limit the latitude the moral reading gives to individual judges. First, under that reading constitutional interpretation must begin in what the framers said, and, just as our judgment about what friends and strangers say relies on specific information about them and the context in which they speak, so does our understanding of what the framers said. History is therefore plainly relevant. But only in a particular way. We turn to history to answer the question of what they intended to say, not the different question of what other intentions they had. We have no need to decide what they expected to happen, or hoped would happen, in consequence of their having said what they did, for example; their purpose, in that sense, is not part of our study. That is a crucial distinction. We are governed by what our lawmakers said—by the principles they laid down—not by any information we might have about how they themselves would have interpreted those principles or applied them in concrete cases.

Second, and equally important, constitutional interpretation is disciplined, under the moral reading, by the requirement of constitutional integrity.6 Judges may not read their own convictions into the Constitution. They may not read the abstract moral clauses as expressing any particular moral judgment, no matter how much that judgment appeals to them, unless they find it consistent in principle with the structural design of the Constitution as a whole, and also with the dominant lines of past constitutional interpretation by other judges. They must regard themselves as partners with other officials, past and future, who together elaborate a coherent constitutional morality, and they must take care to see that what they contribute fits with the rest. (I have elsewhere said that judges are like authors jointly creating a chain novel in which each writes a chapter that makes sense as part of the story as a whole.)7 Even a judge who believes that abstract justice requires economic equality cannot interpret the equal protection clause as making equality of wealth, or collective ownership of productive resources, a constitutional requirement, because that interpretation simply does not fit American history or practice, or the rest of the Constitution.

Nor could he plausibly think that the constitutional structure commits any other than basic, structural political rights to his care. He might think that a society truly committed to equal concern would award people with handicaps special resources, or would secure convenient access to recreational parks for everyone, or would provide heroic and experimental medical treatment, no matter how expensive or speculative, for anyone whose life might possibly be saved. But it would violate constitutional integrity for him to treat these mandates as part of constitutional law. Judges must defer to general, settled understandings about the character of the power the Constitution assigns them. The moral reading asks them to find the best conception of constitutional moral principles—the best understanding of what equal moral status for men and women really requires, for example—that fits the broad story of America’s historical record. It does not ask them to follow the whisperings of their own consciences or the traditions of their own class or sect if these cannot be seen as embedded in that record. Of course judges can abuse their power—they can pretend to observe the important restraint of integrity while really ignoring it. But generals and presidents and priests can abuse their powers, too. The moral reading is a strategy for lawyers and judges acting in good faith, which is all any interpretive strategy can be.

I emphasize these constraints of history and integrity, because they show how exaggerated is the common complaint that the moral reading gives judges absolute power to impose their own moral convictions on the rest of us. Macaulay was wrong when he said that the American Constitution is all sail and no anchor,8 and so are the other critics who say that the moral reading turns judges into philosopher-kings. Our constitution is law, and like all law it is anchored in history, practice, and integrity. Still, we must not exaggerate the drag of that anchor. Very different, even contrary, conceptions of a constitutional principle—of what treating men and women as equals really means, for example—will often fit language, precedent, and practice well enough to pass these tests, and thoughtful judges must then decide on their own which conception does most credit to the nation. So though the familiar complaint that the moral reading gives judges unlimited power is hyperbolic, it contains enough truth to alarm those who believe that such judicial power is inconsistent with a republican form of government. The constitutional sail is a broad one, and many people do fear that it is too big for a democratic boat.

What is the Alternative?

Constitutional lawyers and scholars have therefore been anxious to find other strategies for constitutional interpretation, strategies that give judges less power. They have explored two different possibilities. The first, and most forthright, concedes that the moral reading is right—that the Bill of Rights can only be understood as a set of moral principles. But it denies that judges should have the final authority themselves to conduct the moral reading—that they should have the last word about, for example, whether women have a constitutional right to choose abortion or whether affirmative action treats all races with equal concern. It reserves that interpretive authority to the people. That is by no means a contradictory combination of views. The moral reading, as I said, is a theory about what the Constitution means, not a theory about whose view of what it means must be accepted by the rest of us.

This first alternative offers a way of understanding the arguments of a great American judge, Learned Hand. Hand thought that the courts should take final authority to interpret the Constitution only when this is absolutely necessary to the survival of government—only when the courts must be referees between the other departments of government because the alternative would be a chaos of competing claims to jurisdiction. No such necessity compels courts to test legislative acts against the Constitution’s moral principles, and Hand therefore thought it wrong for judges to claim that authority. Though his view was once an open possibility, history has long excluded it; practice has now settled that courts do have a responsibility to declare and act on their best understanding of what the Constitution forbids.9 If Hand’s view had been accepted, the Supreme Court could not have decided, as it did in its famous Brown decision in 1954, that the equal protection clause outlaws racial segregation in public schools. In 1958 Hand said, with evident regret, that he had to regard the Brown decision as wrong, and he would have had to take the same view about later Supreme Court decisions that expanded racial equality, religious independence, and personal freedoms such as the freedom to buy and use contraceptives. These decisions are now almost universally thought not only sound but shining examples of our constitutional structure working at its best.

The first alternative strategy, as I said, accepts the moral reading. The second alternative, which is called the “originalist” or “original intention” strategy, does not. The moral reading insists that the Constitution means what the framers intended to say. Originalism insists that it means what they expected their language to do, which as I said is a very different matter. (Though some originalists, including one of the most conservative justices now on the Supreme Court, Antonin Scalia, are unclear about the distinction.) 10 According to originalism, the great clauses of the Bill of Rights should be interpreted not as laying down the abstract moral principles they actually describe, but instead as referring, in a kind of code or disguise, to the framers’ own assumptions and expectations about the correct application of those principles. So the equal protection clause is to be understood as commanding not equal status but what the framers themselves thought was equal status, in spite of the fact that, as I said, the framers clearly meant to lay down the former standard not the latter one.

The Brown decision I just mentioned crisply illustrates the distinction. The Court’s decision was plainly required by the moral reading, because it is obvious now that official school segregation is not consistent with equal status and equal concern for all races. The originalist strategy, consistently applied, would have demanded the opposite conclusion, because, as I said, the authors of the equal protection clause did not believe that school segregation, which they practiced themselves, was a denial of equal status, and did not expect that it would one day be deemed to be so. The moral reading insists that they misunderstood the moral principle that they themselves enacted into law. The originalist strategy would translate that mistake into enduring constitutional law.

That strategy, like the first alternative, would condemn not only the Brown decision but many other Supreme Court decisions that are now widely regarded as paradigms of good constitutional interpretation. For that reason, almost no one now embraces the originalist strategy in anything like a pure form. Even Robert Bork, who remains one of its strongest defenders, qualified his support in the Senate hearings following his nomination to the Supreme Court—he conceded that the Brown decision was right, and said that even the Court’s 1965 decision guaranteeing a right to use contraceptives, which we have no reason to think the authors of any pertinent constitutional clause either expected or would have approved, was right in its result. The originalist strategy is as indefensible in principle as it is unpalatable in result, moreover. It is as illegitimate to substitute a concrete, detailed provision for the abstract language of the equal protection clause as it would be to substitute some abstract principle of privacy for the concrete terms of the Third Amendment, or to treat the clause imposing a minimum age for a President as enacting some general principle of disability for persons under that age.

So though many conservative politicians and judges have endorsed originalism, and some, like Hand, have been tempted to reconsider whether judges should have the last word about what the Constitution requires, there is in fact very little practical support for either of these strategies. Yet the moral reading is almost never explicitly endorsed, and is often explicitly condemned. If neither of the two alternatives I described is actually embraced by those who disparage the moral reading, what interpretive strategy do they have in mind? The surprising answer is: none. Constitutional scholars often say that we must avoid the mistakes of both the moral reading, which gives too much power to judges, and of originalism, which makes the contemporary Constitution too much the dead hand of the past. The right method, they say, is something in between which strikes the right balance between protecting essential individual rights and deferring to popular will. But they do not indicate what the right balance is, or even what kind of scale we should use to find it. They say that constitutional interpretation must take both history and the general structure of the Constitution into account as well as moral or political philosophy. But they do not say why history or structure, both of which, as I said, figure in the moral reading, should figure in some further or different way, or what that different way is, or what general goal or standard of constitutional interpretation should guide us in seeking a different interpretive strategy.11

So though the call for an intermediate constitutional strategy is often heard, it has not been answered, except in unhelpful metaphors about balance and structure. That is extraordinary, particularly given the enormous and growing literature in American constitutional theory. If it is so hard to produce an alternative to the moral reading, why struggle to do so? One distinguished constitutional lawyer who insists that there must be an interpretive strategy somewhere between originalism and the moral reading recently announced, at a conference, that although he had not discovered it, he would spend the rest of his life looking. Why?

I have already answered the question. Lawyers assume that the disabilities that a constitution imposes on majoritarian political processes are antidemocratic, at least if these disabilities are enforced by judges, and the moral reading seems to exacerbate the insult. If there is no genuine alternative to the moral reading in practice, however, and if efforts to find even a theoretical statement of an acceptable alternative have failed, we would do well to look again at that assumption.