Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 17 September 2019

Constitutional and Statutory Interpretation

Abstract and Keywords

This article discusses relatively established theories with respect to statutory and constitutional interpretation. Written constitutions and statutes provide authoritative directions for officials and citizens within liberal democracies. The article mentions that descriptive and normative theories connect with each other in critical respects. Statutory interpretation involves the construction and application of provisions adopted by legislatures. The theoretical questions about interpreting statutes and constitutions suggest more general questions about the meaning of human communications; and scholars of philosophy of language, linguistics, literary theory, and religious hermeneutics discuss analogous issues. This article discusses an important issue in statutory interpretation that is the nature and status of legislative intent. A vital aspect of the issue concerns the sources on which judges should draw. This article deals with central features of American constitutionalism as the situation within which to consider problems of constitutional interpretation.

Keywords: statutory interpretation, constitutional interpretation, legislatures, American constitutionalism

1 Introduction

Statutes and constitutions declare legal duties and legal rights. Judges and executive officials must interpret what they provide. Controversy exists about how official interpreters do, and should, perform this task and about how to theorize their practice. This chapter suggests what is relatively settled about statutory and constitutional interpretation, and what is subject to genuine disagreement.

Written constitutions and statutes provide authoritative directions for officials and citizens within liberal democracies (as well as other forms of government). These directions are not merely advisory; others are legally bound to comply. When courts interpret statutes and constitutions, they ordinarily construe a provision that seems directly applicable in accordance with other relevant provisions, related legal doctrines, and legal principles governing interpretation.

In common law countries, one very important principle of law is that courts should follow precedents that they or superior courts have established in earlier cases. Though it could be otherwise, courts follow precedents that concern constitutional and statutory law, as well as judicially developed common law. The doctrine of precedent—with its distinction between holdings that bind and non-binding dictum, (p. 269) and its conditions for overruling—is a vital aspect of statutory and constitutional interpretation. Though I say a few words about precedents in statutory and constitutional cases, I do not explain the practice of precedent, which is treated elsewhere.

Anyone who discusses legal interpretation must face a preliminary question about what constitutes ‘interpretation’. I take an inclusive approach, one that fits traditional legal usage; but I should note some narrower possibilities. One approach is that ‘interpretation’ is not required when the text is clear and decision is simple; it takes place only when decision is difficult.

Another approach is that ‘interpretation’ involves discerning the original meaning of a statutory or constitutional provision. Other criteria may be relevant to how wise judges should decide cases, but they involve something other than interpretation. According to this account, the question of how a free speech ruling would fit with current practice might be relevant to what a court should do, but it would have nothing to do with what the First Amendment means.1

Yet a third approach distinguishes between discerning conventional practices and usages of language, for example, determining what lawyers understand by the privilege against self-incrimination, and ‘interpreting’ the broad significance of a text or practice, deciding how the privilege against self-incrimination relates to basic notions of fairness in the criminal process and to personal autonomy in a free society. Some authors argue that judges rarely interpret in this grander sense;2 others claim that such interpretation is at the root of what judges do.3 About this dispute, one can say that what lies on the surface of judicial action in many cases, including some difficult cases, is textual exegesis or the parameters of a settled practice, not interpretation in this deeper sense. In other cases, opinions reveal that judges are assessing the underlying rationale of texts and practices.

A final approach divides the interpretation of concepts from their application. On this view, a determination of whether a legal concept applies or does not apply to a borderline case is not itself a matter of interpretation.

In discussing the bases for judicial decisions, I do not restrict the label ‘interpretation’ to difficult cases, to inquiries about original meaning, or to evaluations of the deep significance of a text or practice. Also, because the practical import of a concept often depends on the situations to which it is applied, I treat questions of application as aspects of ‘interpretation’. Thus, deciding whether a search of an automobile's glove compartment is ‘unreasonable’ involves interpretation of the constitutional ban on unreasonable searches and seizures.

I do not include as interpretation, however, the discretionary application of concepts that plainly leave a wide range of choice to courts or administrative agencies.

(p. 270) Thus, if a statute requires that power companies be able to charge a ‘fair rate’, an agency or court may set one of a number of rates as ‘fair’ in the circumstances. The choice of the exact rate to charge is not mainly one of interpreting the statute (though some rates will be precluded by the statute's meaning).

Statutory and constitutional interpretation share commonalities. Both involve the construction of legally authoritative texts. For each, one must puzzle about the relationship between theory and practice. Judges, other officials, lawyers, and scholars interpret statutory and constitutional provisions. Although they must decide whether some materials, such as legislative history, matter for interpretation, they may proceed without any full-blown self-conscious theory that fits together everything they take into account. Numbers of scholars, and some judges, have developed theories, more or less comprehensive, about the point of statutory or constitutional interpretation. These theories explain why various criteria for decision should be given weight. How do such theories relate to desirable practice?

The most sceptical approach to theory of this sort is that it is crude and unhelpful—that judges and critics alike would do better to focus on nuances of the practices in which they are involved, forgoing the presumption of believing that comprehensive theories are insightful and can guide. The most ambitious claims for theory are that it is indispensable for sound decision, that all judges should think carefully about the theoretical justifications for their practice. Among a range of intermediate positions, perhaps the most straightforward is the following. Theory can be valuable and influential over time, but it belongs mainly to scholars and law journals. The performance of most judges, lacking much theoretical training, will not be improved if they try to work through deeper questions about their practice. In their actual decisions, judges will often implicitly side with one theory against another, but at least for many problems of theory, judges need not try to resolve them self-consciously.

People who think carefully about theory and practice will develop nuances in these positions. Judges are better equipped to think about some theoretical issues than others. Certain competing theories have sharply variant practical implications; other competing theories about how to conceptualize what judges do, or should do, have no opposing implications for practice. I avoid the question of just how self-conscious judges and lawyers should be about theoretical inquiries. I assume that efforts to build comprehensive theories can be valuable, that many theoretical disagreements have practical implications, and that over time the acceptance of theories influences practice. But I also assume that many able judges immersed in practice can decide wisely, without self-consciously working out many of the deeper theoretical implications of the choices they make.

One general point about theories of statutory and constitutional interpretation deserves mention. A theorist might distinguish a descriptive account of how judges do interpret statutes within any legal system from an appraisal of how they should interpret. He might, for example, say that in the United States, judges give weight to what actual legislators subjectively intended about the coverage of statutes, but they (p. 271) should not do so. However, descriptive and normative theories connect with each other in critical respects. Within legal systems, people rely to a degree on the continuation of existing practices; reasonable reliance constitutes a strong argument for maintaining prevailing interpretive practices. A second connection between descriptive and normative accounts is that when practices themselves are uncertain or various judges engage in different practices, a theorist may conceptualize present practices in light of what he thinks are sound practices, blurring normative and descriptive elements. Whether this aspect of interpreting practices is inevitable, it is common. Finally, when judges or legal scholars recommend shifts from particular legal practices, they claim support in broader features of the legal and political system. Almost always they say that if one understands the fundamental nature of the system, one will see that the practices they challenge are out of line. Thus, they rely on the system itself—a kind of ‘is’—to support an ‘ought’—the proposed changes in practice. Thus, threads of ‘is’ and ‘ought’ are interwoven in most theories of statutory and constitutional interpretation.

Although statutory and constitutional interpretation resemble each other in many respects, typical instances of the two forms of interpretation differ significantly. The differences concern the authority of constitutional and statutory provisions, the political legitimacy of the bodies enacting them, the generality of the textual language, the age of the provisions, and the ease with which political bodies can override what the courts decide. Any analysis of the two forms of interpretation must attend to these differences. In the United States, constitutional interpretation receives great public attention, and lies close to the heart of our political system. But statutory interpretation is also highly important. More significantly for our purposes, what statutory interpretation involves is less complex than what constitutional interpretation involves. Because the nuances of constitutional interpretation are most easily understood against a background of statutory interpretation, I begin with the latter. But most of the theoretical issues I pose for interpreting statutes have very close analogues in respect to constitutional interpretation.

2 Statutory Interpretation

2.1 The General Parameters

Statutory interpretation involves the construction and application of provisions adopted by legislatures. The language of statutes is clear in its import for many circumstances. Citizens, perhaps with the help of simplifying directives or numerical tables, are able to learn their legal duties. Landlords, for example, can find out how (p. 272) many days notice they must give tenants to leave if they wish to take over apartments for their own use. With tax forms, people who have calculated their income know how much income tax they owe. When legal duties are clear, courts usually need not declare them, although judges do instruct juries about legal duties even when the only dispute in a case concerns what factually occurred.

When statutory rights and duties are unclear, courts usually are not the initial interpreters. Private individuals and companies, and their lawyers, decide what they should do and what benefits they may claim from others. Administrative agencies issue detailed regulations implementing more general statutory provisions. Often these agencies also engage in quasi-adjudicative decisions, resolving contested issues before they reach court. But in common law countries, courts are the final interpreters of what statutes provide. (In many civil law systems, separate courts deal with issues of administrative law, but that is rare in common law systems.)

Among the most fundamental theoretical issues in statutory interpretation are these three: is meaning fixed at enactment or does it evolve? How far is meaning determined by reader understanding, how far by a legislative intent discerned partly by other elements? What is the comparative importance of specific narrow objectives and broader legislative purposes? Accompanying these theoretical questions are related questions about the sources courts should use to determine statutory meaning. Similar theoretical issues and questions about sources arise in constitutional interpretation.

The theoretical questions about interpreting statutes and constitutions suggest more general questions about the meaning of human communications; and scholars of philosophy of language, linguistics, literary theory, religious hermeneutics, and other fields discuss analogous issues. Many of them also ask whether the meaning of a communication can change over time, whether original meaning is determined by what a speaker intends or what a listener understands, whether specific objectives, such as allowing the death penalty, matter more than broad purposes, such as eliminating cruel punishments. Various answers offered for the most general philosophic questions or for particular fields outside law do not resolve the crucial issues about statutory interpretation. We need to see why.

Many ordinary communications have immediate practical purposes. The speaker wants to convey his thoughts or feelings or to request some action. Communication and response are fleeting in time. Other communications last. What some poets write is read centuries later. People in political authority also issue directives that last.

In ordinary discourse, speakers try to communicate in light of what listeners will understand; listeners apprehend what speakers say in light of what they think the speakers' purposes are. The intent of speakers and the understanding of listeners coalesce. But things can go wrong or a communication may be incomplete. Something goes wrong when a speaker's intent differs from how the listener understands his words. A communication is incomplete if it fails to indicate how a subject it covers (in some sense) should be resolved. A simple example of incompleteness would be parents, (p. 273) going out for the evening, who instruct their teenage daughter to go to bed at 11 o'clock when the 9 o'clock movie is over. The parents do not realize that on this night the 9 o'clock movie on television lasts until 12 o'clock. The daughter has to decide whether her parents were mainly concerned about an 11 o'clock bedtime or were willing to have her watch the end of that movie. Suppose the parents actually want her to go to bed at 11 o'clock but the daughter reasonably concludes that, because it is Saturday and her parents have allowed her to stay up until midnight on some other Saturdays, she may watch until 12 o'clock. What did the instructions themselves mean?

One approach to the problems of failure, incompleteness, and endurance over time is that the intention of the speaker, or writer, controls. He communicated; perhaps the communication means what he intended. An alternative approach is that the communication means what a listener would understand it to mean. For this purpose, one might ask about the meanings words typically express or focus on a reasonable listener in this context. On either approach, if the speaker says, ‘Shut the door’, when he wants a window shut, the speaker's intent is ‘Shut the window’, but his communication means ‘Shut the door’ (at least if a door is available to shut). Once we introduce the understanding of listeners or readers, the question arises whether the meaning of a communication can shift over time as readers of later years interpret words written earlier. And, if reader understanding is crucial, what if different readers have different understandings? Analysis becomes more complex still if the writer wants and expects various readers to understand his communication differently. What does such a communication mean?

A theorist might deal with these problems by developing a general theory about the meaning of communications. He would need an account that covers private letters—for which he might be tempted to find that meaning is fixed when the letter is written—and poetry written for a large audience—for which he would be attracted to an idea of meaning that differs with various readers and changes over time. Once our theorist arrived at a general theory about meaning, he would fit statutes into that theory.

We can quickly realize that what might constitute ‘meaning’ according to some general philosophic theory might not be exactly what courts should regard as controlling for the practical consequences of a legal system, a system in which the state uses its police power to coerce compliance. For instance, someone drawn to an author's intent approach to meaning might recognize that old statutory language fails to give modern citizens fair warning as to just what behaviour is prohibited. He might conclude that someone should not now be treated as a criminal for acting in a way that she would not realize is forbidden by the language, though the original understanding of the language would cover that action.

If the theorist wished to maintain a constant approach to meaning, he would need to distinguish between what a statute ‘means’ and everything else that judges should deem relevant in determining the statute's practical effect. He would distinguish (p. 274) statutory interpretation as an inquiry into meaning from assessment of everything that matters for a final decision in a statutory case.

Proceeding in this way raises two powerful objectives. The first lies in general philosophic considerations. Why should we assume that meaning is constant over a wide range of communications? Why is it not possible that meaning for letters is different from meaning for poetry? Perhaps the most persuasive general philosophic account is that the ‘meaning’ of communications depends on the nature of the communications involved—that many important inquiries about meaning must be field specific?

The second general objection lies in traditions of law I have already mentioned. Lawyers speak of the meaning of a statute as conforming with how a statute should be applied. They do not say, ‘The meaning of the statute supports plaintiff 's case, but defendant wins because of other considerations’. (Courts do use such language in the law of trusts, which contains special principles when a charitable trust cannot be implemented according to its terms.) With statutes, judges and lawyers speak of meaning and practical application as coinciding. Were legal theorists to differentiate ‘meaning’ from all else that matters for how a provision should be applied, they would employ a dichotomy unfamiliar to lawyers and judges. That, in itself, is not a crushing objection, but it is a serious disadvantage.

Given the tradition in law that desirable performance and application follow statutory meaning, it is convenient to treat interpretation of meaning as involving all that judges take into account when they decide how to understand a statutory provision. Thus, legal traditions, as well as doubts that any general theory can plausibly treat all communications similarly, support the view that statutory interpretation, and the ‘meaning’ of statutes, includes all the considerations that move judges to develop an understanding of statutory provisions as they apply in practice.

We can see then, that a theory about how judges should interpret statutes, a theory about how they should decide what statutes mean, cannot depend entirely on some general philosophy about the use of language. Such a philosophy may illuminate various subjects, such as how people expect authoritative rules of different types to be understood; but the main ingredients of a theory of statutory interpretation depend on analysis of legal systems, and more precisely, on analysis of how courts and legislatures should relate to each other within particular political orders. What is apt for liberal democracies with an independent judiciary may not be apt for a dictatorship in which interpretative officials are directly responsible to the ruler. Not only are interpretive approaches likely to differ in the two kinds of systems, different interpretive approaches will fit best with the normative premises of each system. More importantly, what is apt for Great Britain, with its cabinet system of government and careful legislative drafting, may not be apt for the United States, with its separate political branches and uneven drafting. Issues about statutory interpretation need to be resolved mainly in terms of political and legal theory, not some general philosophy of language.

(p. 275) 2.2 Evolutionary versus Fixed Meaning

Should the meaning of statutes be regarded as fixed or subject to change? When judges interpret statutes soon after their enactment, circumstances will not have changed enough to raise most questions about a possible evolution in meaning; these questions arise with older enactments. To understand opposing positions, we need to be clear what is not in dispute.

If someone asks whether the meaning that judges assign to some statutory provisions has changed over time, the answer is ‘yes’. The debated questions are whether judges do, and should, employ an interpretive methodology that embraces evolving meaning. Should judges implicitly say, ‘We now hold this provision to mean x, but perhaps it should have been given a different meaning shortly after adoption’? An evolutionist says, ‘Yes, statutes are pieces of a complex legal system, and their meaning appropriately changes, as does the meaning of common law principles’. An originalist says, ‘No; statutes should be taken to mean what they originally meant, and they should be applied accordingly. Legislatures should change statutes, not courts.’

The gulf between the typical ‘evolutionist’ and the typical ‘originalist’ is less wide than these fictional remarks suggest. Legislatures adopt some provisions with open-ended phrases that definitely envision that those who apply the law will make judgments consonant with changing circumstances; the originalist agrees that the coverage of such provisions appropriately changes over time.

Even for provisions without this intrinsic flexibility, authoritative judicial precedents may establish a meaning that differs from what judges now deciding a case believe was the probable original meaning. In that event, originalists acknowledge that modern judges should (generally) follow the precedents rather than insisting on their understanding of the original meaning. Similarly, when administrative agencies have assigned a reasonable meaning to unclear statutory provisions, courts may adhere to that meaning, rather than adopt the meaning they would, on balance, find to be the most likely original meaning. Such deference to administrative interpretation may be warranted partly because the agency has a good sense of original meaning or is implementing flexible language, but the deference is also based on a view that administrative agencies are partners of the courts in interpreting statutes and developing the law. Thus, courts may follow agency interpretations not only when they think agencies have discretion to interpret one way or another but also, sometimes, when they think the agency interpretations are probably mistaken. (Exactly when courts should defer to agencies is controversial; originalists regard the circumstances for deference as more restricted than do evolutionists.)

Finally, originalists acknowledge that fair warning matters and that unclear statutory provisions should be fitted into the existing corpus of law. Especially in criminal cases, judges should not employ an original meaning if that now fails to give adequate warning. And significant changes in the surrounding body of law can affect how judges should understand an unclear statutory provision.

(p. 276) In these various respects, most originalists accede to forms of evolution in meaning. They acknowledge not only that such changes in meaning do occur but that they should occur. What evolutionists must concede to originalists is that judges rarely talk about changed meaning.

What then divides ‘originalists’ from ‘evolutionists’? They may disagree over the circumstances when judges should accept meanings other than those they would discern as original meanings, over what factors judges may take into account, and over the weight to be given to reasons in favour of change. They may also disagree about how to conceptualize judicial acceptance of a non-original meaning.

On the conceptual point, an originalist might claim that the legal ‘meaning’ of a statutory provision does not really change, that what modern judges are doing is deferring to the mistaken interpretations of meaning by others (as when they follow precedents or administrative rulings) or giving effect to considerations other than meaning (as when they decline to apply language that now fails to give fair warning).

If we realize that author's intention does not necessarily determine meaning, that reader understanding can matter for what language means, we have no reason to rule out the possibility that modern reader understanding can make a difference, and thus no reason to rule out the possibility that legal meaning can change over time. And, once we see that judges appropriately treat provisions as having a meaning that varies from the original one, we will find no reason in general philosophy to doubt that the meaning can actually change. From the standpoint of legal and political theory, what is of prime significance is how judges actually treat statutory provisions, not whether they see themselves (or theorists see them) as accepting changed meaning or deciding on the basis of considerations other than actual meaning.

Originalists differ with evolutionists over substance in being more likely to accede to changes in meaning only when they think the coverage of provisions is genuinely uncertain in the first place. Except perhaps when deferring to precedents, an originalist judge will not decide against a meaning she thinks is evident, or plain. That is, an administrative decision or change in the surrounding body of law will not lead her to deviate from a plain meaning. An evolutionist will find more provisions to be uncertain in coverage, and thus susceptible to change in meaning, than will an originalist. An evolutionist will also sometimes be willing to accept change in meaning, although the original meaning was not uncertain.

Originalists are likely to deny that changed social facts, other than changes in the law, can generate changed statutory meanings. They will deny that the shifting normative appraisals of judges are a proper basis for new interpretations, unless a provision explicitly leaves room for such appraisals.

Finally, even when an originalist concedes that judges may take account of arguments that are not based on original meaning, he will give more weight to the indicators of original meaning than will an evolutionist, who may regard original meaning as having slight significance as a statute ages.

(p. 277) All originalists and most evolutionists share one important point. They reject an account of meaning that fits a common modern version of meaning for poetry and other literary works. If meaning depends on reader understanding, why not say that meaning actually varies with the understanding of each reader? If you and I find different meanings in a poem, perhaps all one can say is that the poem has different meanings. Not only does the poem have different meanings, individual readers may have no good reason to try to find a meaning that will satisfy other readers.

A sceptic about legal reasoning may believe that the meaning of legal texts similarly comes down to the subjective reactions of individual readers. On this view, the reasons judges assign to adopt one meaning over another are a cloak that conceals subjective reaction. In its most extreme form, this position is definitely mistaken. Given the conventions of natural language and of law, many statutory provisions, cast in fairly precise terms, do require one understanding in context and exclude others. But this reality leaves open the possibility that over some range in which reasonable competing arguments can be made about meaning, interpretation and meaning are determined by individual subjective responses. Since the language of literary works, even poems, also exerts constraint on the meaning people reasonably find in them, the sceptic can conclude that law is really not so different in this respect from poetry.

Few theorists, however, urge this analogy to poetry as apt from the inside—as a guide for those who make authoritative decisions about legal meaning. Perhaps a judge should try to get the personal feel of a statute, as a reader gets the feel of a poem; but the judge cannot rest there. Her decision, with that of colleagues, determines whom the state will coerce; and it establishes meaning for future cases. The judge must consider whether her response is idiosyncratic or whether the reasons that move her would or should move other judges. The judge aims for a meaning that other judges would also appropriately find. Whether the meaning is ‘original’ or based partly on non-original factors, the judge seeks an interpretation that is sound for judges generally.

2.3 Readers' Understanding and Legislators' Intent

One of the most widely discussed issues in statutory interpretation is the nature and status of legislative intent. On examination, this issue turns largely on the relative importance of legislators' ideas about what they have enacted and readers' understandings of enactments. A vital aspect of the issue concerns the sources on which judges should draw.

To unpeel the elements of this broad issue, it helps to begin with the reminder that, in ordinary communication, speaker's (writer's) intent and listener (reader) understanding will substantially coalesce. What the listener understands is what the speaker has meant to communicate. If an outsider is aware of a communication and (p. 278) its social context, his best evidence of what the speaker intends is what a typical listener would understand by the speaker's words. An outsider could have special knowledge about a particular speaker and listener that would lead him to conclude that their understandings diverge, but this would be unusual. The outsider will be likely to reach differing judgments about speaker intent and listener understanding only if he has information about the speaker's intent that is not available to the listener. Commonly that information consists of remarks by the speaker made before or after she has communicated, about what she was trying to say. To revert to our example of parents telling their teenage daughter to go to bed at 11 o'clock when the 9 o'clock movie is over, the parents might say to friends, ‘We've got really concerned about how tired she is. We're insisting that she stay up no later than 11.’ The friends would be confident that the parents did not intend their daughter to watch the last hour of the movie that ends at 12 o'clock, although the daughter might be left in doubt by the instructions themselves.

How does this analysis apply to statutory interpretation? If judges rely entirely on the statutory text understood in light of social circumstances that are known to readers, they will have little basis to conclude that legislators intended something different from what readers would understand. Any theoretical debate about the comparative importance of the intent of legislators and the understanding of readers would lack practical significance. The theoretical debate has practical bite if judges may use sources to discern what legislators intended that ordinary readers would not employ.

2.3.1 Evidences of Legislative Intent

Individual legislators may declare their aims before or after they adopt a law. Everyone is suspicious of post-enactment statements, which are not subject to review and disagreement by other legislators before an act is passed. As a consequence, the main source for evidence of legislative intent (apart from text and context) is what is called legislative history, materials from the process of enactment that indicate how a bill is understood. The primary sources are reports of committees that screen and revise bills, statements made by sponsors on the floor of the legislature about what bills mean, and actual changes in the texts of bills as they proceed towards passage. These materials are, in a sense, available to readers; but a search of them is time-consuming and it is usually hard to extract what is relevant for a particular legal issue.

If a judge consults these materials, but assumes they do not also underlie the understanding of ordinary readers of the statutory text, she may determine that legislators were trying to do something at variance with what a typical reader of the text would conclude. The debate over the role of legislators' intent and reader understanding in statutory interpretation, thus, links substantially to the debate about the use of legislative history.

Those who assume that reader understanding is what matters for statutory meaning are likely to accord legislative history little or no role. Those who believe (p. 279) legislative intent has independent significance are likely to support the use of legislative history.

We must be careful, however, not to equate the place of legislative history with the more theoretical question about legislative intent. Some theorists who believe the actual intent of legislators about the scope of the law they have enacted is irrelevant still find value in the judicial use of legislative history. That history may reveal independent facts that contribute to sound interpretation and may indicate something about broader public attitudes. Further, pieces of legislative history may appropriately have a conventional weight in interpretation that does not depend on the states of mind of legislators.

Some of the arguments against use of legislative history do not directly challenge the possible significance of legislators' intents. The most common argument against judicial use of legislative history is that its exploration is time consuming and usually unproductive and that most bits of modern legislative history represent, not the considered views of important legislators, but the interests of lobbyists who have persuaded legislative assistants to insert constructions of provisions that are favourable to their interests.

I consider the issue of legislative intent on the assumption that some materials may be available to discern that intent that would not form a part of reader understanding. On this assumption, should legislators' intent control, should reader understanding control, or is each independently relevant?

2.3.2 The Obvious Importance of Reader Understanding

We can begin with reader understanding. Without doubt it has importance. Laws restrict human behaviour, and the legislature has no power to restrict behaviour without adopting statutes. If legislators must adopt statutes to restrict behaviour, they cannot successfully do so if citizens lack a basis in the law to apprehend the restrictions. More to the point for difficult cases, if statutory language leaves uncertain which acts are forbidden and which allowed, a reader's judgment about the language should matter for a judge's best interpretation of a provision. At least when modern reader understanding is not readily distinguishable from how readers at the time of adoption would have understood language, a law's original meaning should not rest exclusively on legislators' intent, it should include reader understanding. (One might believe that for very old laws what should matter is modern reader understanding along with what legislators originally intended, that original reader understanding should have little or no independent significance.)

Depending on the kind of statute adopted, the relevant ‘reader’ might be an ordinary person, an expert in a field such as atomic energy, or a lawyer. A criminal law directed at ordinary people differs from procedural rules for lawyers, and both differ from legislation regulating highly technical subjects. Now, considering that ordinary people rarely read statutes, someone might doubt whether their understanding, as opposed to that of lawyers, should ever matter. But people should be able to (p. 280) understand on their own why they have committed a crime or violated someone's rights, even if they have not read a statute in advance. In any event, for most purposes, the understanding of lawyers will not differ greatly from the understanding of ordinary well-educated speakers of English.

How should a judge conceive the reader of statutory language? For provisions that are complex but yield their meaning to careful analysis, she should assume a reader who is intelligent and makes the effort to unlock the complicated puzzle of the language. But there are no experts in the meaning of ordinary words and phrases (except those who comprehend the way others understand—such as compilers of dictionaries). The judge's basic task is to try to understand the text as would readers who confront it. (One can conceive of readers constructed normatively who give the best reading possible to statutory language, but this is to move away from a genuine reader understanding approach towards one in which a judge constructs the best normative reading.)

A judge who asks how a reader would understand a text must pay some attention to social context and legislative purpose. A reader will view a text against social context. He will have some sense of the problems a law was meant to address, and some idea of legislative purposes, and will interpret the language accordingly. Thus, a judge using a reader approach will need to determine what sources to use to discern the social context of a statute, and the purposes a reader would attribute to the statute. The judge might rely, for example, on analyses of social problems in newspapers and magazines or a President's State of the Union address.

For modern legislation, and modern reactions to older legislation, a judge may be bold enough to assume that she adequately replicates what a reader would understand. Her reading can be a reader's reading. But this will not work for the original understanding of much older legislation. A judge cannot intuitively grasp all of what the reader of a time long past would have understood; she needs to make a conscious attempt to recapture older understandings. But just how should she do this? Should she posit a single reader of the period or recognize that different readers may have understood crucial words or phrases differently? Should she conceive of a well-educated reader or one with a rudimentary education? Should her reader represent normatively desirable attitudes or average attitudes? To take an extreme example, if readers in 1830 assumed that members of the ‘white race’ were superior to all others, is a provision of uncertain coverage, adopted in 1830, now to be read accordingly? And exactly what states of mind of a reader, present or past, should our judge take as the attitudes that count for understanding? I return to these problems after surveying the more familiar analogous problems about legislative intent.

2.3.3 Deviation from Textual Language

Another important interpretive problem about textual understanding warrants mention. When, if ever, should judges construe language differently from what the (p. 281) words of a provision seem on the face of it to say? There is general agreement that judges should fill in omitted words, when no one doubts that a word, such as ‘no’, has been omitted by mistake. More controversially, judges also try to interpret language in a way that corresponds with basic policies underlying a statute and with fundamental principles of justice. In one famous case, the Supreme Court construed a law that barred employers from making contracts with aliens ‘to perform labor or service of any kind in the United States’.4 Although the specific statutory language was broad enough to cover the church's contract with a British minister, the Court said the statute was aimed at manual labourers, not at brain toilers, and especially not at ministers of the gospel. Justice Antonia Scala, a prominent ‘textualist’ (i.e. an originalist who relies on textual meaning, not legislative intent), has expressed powerful disagreement with the approach of that case:5 when specific words are relatively clear, courts should follow them, leaving legislatures to correct their own mistakes.

2.3.4 The Possible Relevance of Legislators' Intents

Judges in various common law countries have adopted different postures towards legislators' intent and the resources of legislative history. For most of the twentieth century British courts declined to look at internal legislative materials; the American Supreme Court used them extensively to discern the objectives of Congress. At the beginning of a new century, a majority of Supreme Court Justices still rely on legislative history, but others, led by Justice Scalia, do not. Among the reasons for opposing that use are these: (1) The materials of legislative history are an unreliable indication of attitudes within the legislature. (2) Legislators can regulate only by voting on statutes; they should not control meaning by inserting their points of view in legislative history. (3) Judicial reliance on legislative history contravenes the necessity of adoption of legislation by two houses and the executive. (4) The concept of legislative intent is itself misconceived.

The first ground of opposition has great practical importance, but presents few troublesome theoretical questions. The materials of legislative history may be a good or bad reflector of what legislators believe. Judicial investigation of the materials may or may not often be illuminating about specific understandings of particular provisions and about broader purposes that underlie a statute. Judges and scholars have to assess the value of legislative history against the costs of judges considering it and of lawyers combing it for relevant information. Part of this evaluation involves a decision as to whether committee reports need reflect the views of legislators themselves to be significant or may appropriately be given weight because they reflect the views of active staff members.

(p. 282) The other three objections are ones that, in principle, legislators' intentions should not make a difference. These objections raise general questions about instructions and rules, and narrower questions about political life and legal systems.

In respect to the general questions about instructions and rules, suppose that one person has a relationship of authority over another. If the subject's job were to try to carry out all the desires of the person with authority, the latter's verbal instructions would mainly serve as a means to indicate her desires. When the subject recognizes that the authority's instructions were unclear, or somehow knows that she did not intend results clearly called for by the instructions, the subject will be guided by the authority's intent. The argument that the implications of the instructions themselves should control has force only if the subject's job is to follow instructions, not to carry out an authority's every perceived wish.

Legislatures can act only by adopting the language of statutes; citizens need not do what legislators wish if the legislators have failed to adopt statutory language. Thus, the argument that only the statutory language should count passes this initial hurdle that instructions may be merely evidence of desires.

However, even when people are at liberty to do what they wish, or think is best, unless limited by instructions, relationships vary. A subject may have no relevant interests of his own opposed to those of the person in authority over him, and the authority may possess greater competence or may enjoy a greater legitimacy because she has been assigned responsibility by others. A patient has entrusted his medical care to a doctor. The doctor gives instructions about treatment to a nurse and to a doctor who is filling in for her. The nurse is less expert; the second doctor, though as skilled and experienced as the first doctor, has less responsibility for this patient's treatment. Both the nurse and second doctor should follow the first doctor's instructions according to her intent, if they happen to discover that intent.

In other circumstances, one person has authority to limit what a subject can do by issuing instructions, but the latter has interests of his own and the liberty to pursue these interests, except as instructed to the contrary. Employees, for instance, may be free to dress as they choose, except as their bosses restrict them. A worker might reasonably say he is not successfully restricted unless the terms of his boss's instructions do so, that he need not act in accord with her intent, if that intent is more confining than the apparent force of the instructions taken by themselves. The worker may assert that he has a range of liberty that can only be confined by instructions that by themselves tell him what to do. The lesson of these informal analogies is that the importance of the intent of someone issuing instructions depends on the reasons why the person has authority and on that person's relation to those subject to her.

Laws vary among themselves in respect to relations between legislatures and those subject to the laws. Some statutes are instructions to the executive branch from the legislature, a body with greater political authority, to carry out broad public policies. Other statutes restrict the liberty of individuals in ways the individuals disapprove or dislike. If an inquiry about relationships were to determine whether the language of (p. 283) statutes should itself be controlling or legislative intent should also count, the right answer would depend on the statute involved. Instead of concluding that intent is irrelevant because legislatures can regulate only by adopting language, judges would consider all those directly affected by a statute, before deciding whether they should reach beyond the guiding textual language to rely on an intent discoverable from outside the text.

The worry that reliance on legislative history in the United States bypasses a formal process of approval rests partly on the assumption that the Constitution, requiring acceptance by two legislative bodies and the President, was designed to make legislation difficult. Judges who use legislative history give weight to indications of meaning that have not been approved in the required way. It is a partial answer to this worry that judges usually employ legislative history to reach conclusions that are unclear from statutory provisions themselves. Using such material to interpret what has been adopted is not quite the same as allowing members in one house to legislate by their legislative history. Against an ideal of adoption by two houses and the executive, judicial use of legislative history to infer legislative intent may have a flaw, but the flaw is not grave enough to reject legislative history on that ground.

2.3.5 How to Understand Legislative Intent

The claim that legislative intent is wholly misconceived, that the concept is actually incoherent, is yet another objection to courts relying upon legislative history to reveal intent. To evaluate this objection, we must examine just how legislative intent might be understood. It might reflect the actual mental states of legislators, embody an objective notion of a reasonable legislator, or function as a construct based on conventions about what materials count. Many opinions read as if legislative intent reflects the actual mental states of legislators, but that idea has been powerfully challenged.

I suggest how we might best develop a mental states version of legislative intent, and offer some qualifications, before inquiring whether adoption of an objective or conventional approach avoids the perceived difficulties with a mental states account.

One might take the concept of legislative intent as an illustration of the broad theoretical issue of what constitutes a group intent. One would try to sketch the parameters of ‘group intents’ and then decide whether the attitudes of legislators towards a piece of legislation qualify. This way of proceeding, however, would be misguided.

As to whether a group can ever have an ‘intent’, the answer is ‘yes’. It is not that the group as such has a mental state; it is that members of a group share the same mental state. If all (or nearly all) members of a group knowingly share an intent that relates to function of the group, the group (in this sense) has that intent. Our parents had a shared intent about the bedtime of their teenage daughter and all the members of a team might share an intent to play in a certain way.

Under these strict conditions, legislators would rarely have a shared intent about the coverage of specific statutory provisions. We might proceed to ask whether (p. 284) groups may have intents under less stringent conditions that legislators would more often satisfy. But the crucial question for interpretation is not whether a legislature has a group intent, it is whether judges should take into account the attitudes of legislators in interpreting statutes.

This exercise might be appropriate, perhaps because the intents of a minority of individual legislators should matter, even if the conditions for having a group intent are not satisfied. (And the exercise might be inappropriate even if the conditions are satisfied.) Once we focus on what courts may profitably take into account, we shall see not only that ‘group intent’ is a red herring, we can grasp that judges might be influenced by something more complex than a simple, single mental state held by some number of legislators.

Some clarifications are in order. (1) The relevant mental states might concern broad purposes of statutes as well as beliefs about the coverage of particular provisions. (2) If judges reach conclusions about actual mental states, these must be based on probabilities, not certainties. (3) Judges will base judgments about actual mental states mainly on expressions or manifestations of some kind. It would be a reasonable requirement, given the nature of legislation, that these manifestations must occur prior to adoption of legislation.

When one considers what mental state should count as an intent, two obvious candidates are ‘hopes’ and ‘expectations’. But each proves inapt. A legislator might agree to language that she knows is designed to prohibit particular behaviour. She does not favour this prohibition, but has gone along with it in order to obtain support for aspects of the statute she endorses. She hopes the courts will construe the provision contrary to the way in which she understands it. Her hopes do not reflect her relevant intent. Suppose another legislator expects courts, unsympathetic to the aims of legislation, to construe a provision contrary to his hopes and to his understanding of what it accomplishes. His expectations do not reflect his relevant intent.

The state of mind that is much more relevant than hopes or expectations is the legislator's understanding about what the statute does, his opinion about how courts should interpret its language. It might be objected that this approach sets up a vicious circle. The legislator's understanding is based on how he thinks courts should interpret, on everything courts should take into account. Courts would interpret on the basis of legislators' understandings of how they (the courts) should interpret.

If there is a circle at all here, it is not vicious, because the judicial function certainly encompasses much more than guesses about what legislators think judges should do. Legislators' intents are only one basis for judicial interpretation. Judges are influenced but not controlled by how they think legislators believe they should interpret. Further, judges may decide to give more weight to legislators' views about the straightforward significance of statutory language than to their views about the subtle nuances of judicial interpretation.

What if a legislator believes courts should interpret in one way, but he accepts as settled a contrary practice courts have established? Judges should be guided by the (p. 285) legislator's view of how specific language or broad purposes should be understood, given interpretive practices he takes as settled. (In cases that are otherwise very close, legislators' hopes might also figure to some degree.)

Whose views should count and what weight should judges accord them? Contrary to what is sometimes assumed, the views of those who vote against a bill may matter. A legislature is a co-operative body, not simply a collection of majority voters. A legislator may participate in drafting legislation and endorse what a particular provision does, but vote against the final bill because it has other objectionable features. The legislator's view of what the provision does should be relevant. Even if she opposed that very provision, her understanding carries importance if she has joined in the effort to formulate the language.

The views of legislators who have actually considered a provision matter more than the views of passive participants (even if one could guess accurately what the passive participants thought). If one takes the temper of a group, one ordinarily gives special weight to the views of those who are most active and best informed. Moreover, legislators themselves do approve (or would approve if they thought about it) a system in which their own views matter most for the statutes with which they are most actively involved.

Given the conditions of modern legislation, do any views count other than those of legislators? In so far as members choose to rely on the judgments of their staff and of members of the executive branch who propose legislation, courts should also take those views into account. Indeed, if a judge thinks of legislation as a complex process involving many actors other than the legislators themselves, she may deem significant the views of staff members and executive officials who have participated in drafting, even apart from some notion of implicit delegation by legislators.

Judges who consider the actual views of legislators are under strong pressure to consider their hypothetical views as well. By hypothetical views, I mean the views actual legislators would have had if they had asked themselves questions they did not address. One reason judges might address hypothetical attitudes concerns passive and silent legislators. Often judges may have little basis to know whether these legislators actually had a view. But judges may be able to say that if the committee members and other dominant legislators had a particular opinion, it either was shared (an actual understanding) or would have been shared (a hypothetical understanding) by most other legislators. Probably the views of an informed minority should be relevant by themselves. But if virtually all those who express themselves agree, judges who count hypothetical understandings can (usually) assume that most legislators would have had a similar view.

Public choice theorists have objected to judges relying on hypothetical votes. Because of the arbitrary order in which options may be considered and because of strategic voting, one cannot be confident how any legislator would have voted on an issue, unless one knows the context of the vote. Reliance on the kind of hypothetical understanding I have suggested seems largely (though perhaps not entirely) to avoid (p. 286) these problems. The question posed is what a legislator would have thought about a provision, given the provision (and surrounding law) as it was when adopted. Answering this question may be hard but it is not much subject to the difficulties with hypothetical votes.

A mental states version of legislative intent, as developed here, is feasible; it is not incoherent. If judges try to discern and combine the crucial understandings in a nuanced way, they will concededly be unable to formulate exactly what they are doing, but this inability characterizes much practical reasoning. Judges may, none the less, be able to make sensitive assessments both of legislators' understandings and of how these should weigh against other interpretive factors.

Giving conventional weight to pieces of legislative history may be an alternative to a mental states approach. The idea of conventional weight is that certain sources have conventionally, over the years, come to have a significance for interpretation. On this account, courts assign explanatory importance to committee reports and other statements, not because these connect in any particular way to mental states, but because earlier judicial interpreters have treated them in this way. A serious problem with this account as a radical alternative to a mental states approach is that the pieces of legislative history that are accorded the most weight are just those that have best reflected the attitudes of key legislators. Further, a strong argument for giving these pieces of history less weight than in the past has been that they no longer well reflect legislators' attitudes. An account that draws some connection between conventional weight and mental states is more appealing than one that conceives a total divorce. Here is one such account.

The materials that courts have given the most weight are those that have best indicated the attitudes of key legislators. In typical cases, judges have a rough idea how much weight to assign to committee reports, sponsors' statements, and other indications of intent, and they do not worry much about just how accurately the materials reflect mental states in individual cases. However, were judges to possess some powerful indication in a case that standard materials poorly reflect mental states, and that other materials are more revealing, judges would consider these other materials. A rough correlation, thus, exists between conventional weight and mental states. Were judges to perceive the correlations as changing drastically, conventional weight would change.

A different alternative to a mental states account is the construction of a reasonable legislator. Legislative intent would be the intent a reasonable legislator would have. Just how far this construction escapes inquiry into mental states depends on how the reasonable legislator is conceived. If the reasonable legislator is taken as the average legislator, judges could make judgments (mainly empirical judgments) about what he would think only by considering the likely attitudes of most actual legislators, or the attitudes of most people who might occupy the position of legislator. Questions about whose and what mental states matter would not go away, though they would be somewhat transformed. If the reasonable legislator is taken as superior to most (p. 287) ordinary legislators—better informed, more careful, more attuned to public values—then the judge's conclusion about what the reasonable legislator would think would depend less on assumptions about ordinary legislators. The judicial inquiry would then be mainly normative. Even in that event, the reasonable legislator would presumably be responsive to social problems actually perceived by people. In constructing the purposes that a reasonable legislator would ascribe to a statute, a judge would have to evaluate which problems were perceived by citizens and legislators. That aspect of the inquiry would be empirical, and would concern itself with mental states.

2.3.6 Reader Understanding and Relevant Mental States

It remains to return briefly to reader understanding approaches, to examine some problems about mental states that should be clearer now that we have explored various approaches to legislative intent.

Readers understand statutory provisions in context. They need to grasp the social context of legislation in order to understand its purposes. If the purposes are disputed, readers must assess them according to their estimate of the likely attitudes of legislators, based on whatever sources the readers use, or according to some construct of how a reasonable legislator would react. Even the latter construct, as I have indicated, does not wholly escape inquiry into likely mental states.

More serious difficulties face the judge who is constructing the relevant reader. Readers of statutory texts may have hopes, expectations, and views about proper interpretation that may split in the same manner as the views of legislators. A judge who employs a reader understanding approach must stand ready in some instances to say what states of mind matter. And if one asks how readers of a statute, living at the time of its adoption, would have understood a provision, the answer may be that different readers would have understood it differently.

Here is an example. Title VII of the 1964 Civil Rights Act provides that employers and unions may not ‘discriminate’ on grounds of race, gender, and so on. One sense of ‘discriminate’ involves any unfavourable classification; but some people think that one cannot ‘discriminate’ against members of the dominant group, that classifications that favour those who have previously been oppressed do not discriminate against those who have benefited from oppression. The difference over the meaning of ‘discriminate’ has an obvious bearing on whether the crucial statutory sections allow employers and unions to undertake affirmative action that favours members of previously victimized groups. Looking to the purposes of the law will not settle this issue, because some readers will discern a primary purpose to help the previous victims of discrimination, while other readers will believe the main aim was to eliminate all categorization along the lines of race and gender.

The judge who adopts a ‘reader understanding’ approach must decide implicitly how to weight the views of different readers or how to combine those readers into one (p. 288) representative reasonable reader. In either event, problems of mental states and their combinations are present in reader approaches, as well as in inquiries about legislative intent. The difficulties of resolving such problems cannot be a decisive reason to consider reader understanding to the exclusion of legislative intent.

2.4 The Relevance of Purpose

A crucial question about any approach to the meaning of a statute is how much weight to give to purposes, as contrasted with the language of particular provisions and indications of specific intent. Most legislation is the product of compromise; it is not surprising that some provisions may be out of line with a law's broad purposes. If the language of a provision clearly indicates a result, courts will give it effect even if it lies in tension with a statute's more general aims.

The more troublesome interpretive problems are these. (1) The language of the provision could be read in one of two ways; the first conforms with indications in the legislative history about what results were intended under the provision, the second better carries out the act's broad purposes revealed by the legislative history. (2) The statutory language could bear a number of meanings; one of these seems the most natural reading of the specific provision, but another fits better with underlying purposes.

The first problem can be formulated in this way. When judges look outside the text to discern the aims of the legislature, should they pay more attention to purposes or specific intent? The argument in favour of specific intent is that what committee reports and other legislative materials say about specific resolutions is more focused than what they say about purposes. One argument for emphasis on purposes is that statements about purpose are less subject to manipulation by interested legislators than are claims about specific resolutions. Another argument is that the development of a rational law is best served by judges concentrating on legislative purposes; this is especially true as statutes age and conditions change. In their influential work on statutory interpretation, Henry Hart and Albert Sacks urged that the primary use of legislative history was to ascertain legislative purposes.6 Critics have objected that this approach is not faithful to legislative compromise and encourages judges to follow their own policy views under the guise of finding amorphous purposes.

Issues about the place of purpose may also arise when judges do not rely on legislative history. Should judges prefer the reading of a narrow provision that is most natural even when that reading fits awkwardly with purposes that are stated in a pre-amble or are obviously implicit in the body of the statute as a whole? Judges and scholars who emphasize the priority of text tend to be concerned with restricting (p. 289) judicial discretion; they believe judges should generally stick to what the immediately relevant text seems to say rather than relying heavily on vague statements of purpose.

3 Constitutional Interpretation

3.1 Issues and Possibilities

Virtually every problem about statutory interpretation also comes up in constitutional interpretation, but crucial differences between most statutes and most constitutional provisions affect the dominant issues and persuasive positions. Three central features of the American Constitution are that it is much older than most statutes, it has many open-ended and vague provisions, and it is difficult to amend.

Partly as a consequence of these features, people worry about the political legitimacy of judges interpreting constitutional provisions and imposing their interpretations on the rest of the government. These features and concerns are related in the following (made up) attack on frequent judicial invalidation of statutes: ‘Why should non-elected courts prevent legislatures from enacting the will of the people, relying on an ancient document whose wording is unclear and whose adoption was by unrepresentative politicians?’

In this chapter, I treat these central features of American constitutionalism as the setting within which to consider problems of constitutional interpretation. But I should emphasize that matters are different in some other constitutional regimes. The American Constitution was adopted in 1789 and the Bill of Rights followed two years later. The balance of federal and state power was significantly altered by the Fourteenth Amendment, adopted in 1868 after the Civil War. As a consequence, the bulk of the national constitution was adopted more than 200 years ago, and its most important later amendment is more than 130 years old. Only male land holders had the vote in most states when the Constitution and Bill of Rights were adopted, and the great majority of blacks were living in slavery. Representation was broader by 1868, but women lacked the vote.

The American Constitution is very difficult to amend. The most common method requires a two-thirds vote of each house of Congress and approval in three-quarters of the state legislatures. The difficulty of formal amendment is one reason why the early provisions have shown such staying power.

The language of many of the crucial provisions of the Constitution is sparse. Congress has the power ‘To regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes …’7 ‘Congress shall make no law affecting (p. 290) the establishment of religion or prohibiting the free exercise thereof. …’8 This language does not provide a detailed code about how to treat problems.

The constitutions of many countries, and of many American states, are more modern, more detailed, and more amendable. These differences could well influence just how their provisions should be interpreted.

Before I turn to troublesome issues of constitutional interpretation, I note some kinds of systems in which ‘constitutions’ have a status different from that in the United States, and I touch on some of the constitutional responsibilities of non-judicial officials.

Some legal systems, such as Great Britain, lack a comprehensive written constitution. Judges may invoke settled constitutional practices when they interpret laws, but they cannot say that legislation enacted by Parliament is invalid. The legislature acting with the approval of the king or queen, is supreme. Some countries have written constitutions that are understood to constrain legislators and other officials. But courts do not enforce provisions of the written constitution against legislative acts; rather they accept those acts as valid.

In the United States, the prevailing assumption has been that an enacted part of the Constitution is valid until it is repealed by subsequent amendment. In India, by contrast, provisions of the Constitution have been declared invalid because they conflict with the principles of more important parts of the Constitution.9

Bruce Ackerman has proposed that the American Constitution may be amended by other than the formal means, through a combination of legislative action and public approval.10 The Supreme Court has never acknowledged such authority, and, given the difficulties of judicial delineation of when informal amendment has taken place, it seems unlikely to do so. But a theorist may none the less contend that written judicial opinions that give a modern cast to ancient provisions are disingenuous efforts to conceal the reality of informal amendment. In what follows, I concentrate on interpretation of the written constitution, neglecting both the possibility that enacted, unrepealed constitutional provisions may be invalid, and the possibility that what, on occasion, courts really interpret is an unacknowledged informal amendment.

My discussion focuses on constitutional interpretation in systems in which courts interpret a written constitution and have the authority to declare statutes and executive actions (such as unlawful searches) unconstitutional. But this focus should not (p. 291) obscure the reality that even in such systems other officials also interpret the constitution. Legislators decide whether a bill they might adopt passes constitutional muster and high police officers decide what investigative techniques are constitutionally permitted.

In the United States, whole areas of constitutional law are left to non-judicial bodies. The Constitution provides that Presidents may be impeached, convicted, and removed from office for ‘Treason, Bribery, or other high Crimes and Misdemeanors’.11 Deciding how to classify forms of official misbehaviour involves constitutional interpretation; but, according to traditional understanding, courts will not review determinations that the House of Representatives and the Senate make about impeachable behaviour. Thus, even in systems in which courts enforce most provisions of written constitutions, they may not enforce every aspect of those constitutions.

Within the areas in which courts are active, two interesting questions arise about how other officials should regard their constitutional responsibilities. The first is how far legislators and executive officials should regard themselves as constrained by the constitutional determinations of courts. Without doubt, they should comply with what judges decide in any individual case, but should they also take a court's announced rule of law as controlling their behaviour more generally?

In all common law countries, other officials are largely guided by judicial interpretations; but suppose members of Congress think the Supreme Court has made a mistake in holding a statute invalid? May they appropriately re-enact similar legislation? When the Dred Scott decision held the Missouri Compromise invalid, because it purported to free slaves who had been brought by masters into free territories, Abraham Lincoln responded that he did not take the principles the Supreme Court announced as ones Congress had to follow.12 It is one thing to say that courts should apply the Constitution in their decisions, another to say that their opinions bind other branches of government. In recent decades, the Supreme Court has spoken as if its decisions lay down authoritative principles for all branches of government.13

A second question about attitudes of non-judicial officials is how they should regard the Constitution as regulating their behaviour.14 In some areas of American constitutional law, a statute's validity can depend on the reasons why legislators adopted it. For example, a law enacted in order to promote a religious objective violates the Establishment Clause. But judges are hesitant to decide that religious objectives underlay a statute that can be justified on non-religious grounds, and, in any event, the religious objectives of a few legislators would not render an entire statute (p. 292) invalid. Suppose a single legislator would not vote for a law that provides aid to private religious schools, except that she believes the aid will promote religious truth. She knows that other legislators have secular reasons for supporting the bill—better education for children—and that a court will not declare the law invalid because it is based on an impermissible purpose. At least if she accepts the idea that the Constitution bars religious objectives for statutes, she probably has a constitutional obligation not to vote for the law, even if she knows that her vote in favour will not lead to judicial invalidation. I do not further pursue these questions about how non-judicial officials should conceive their fidelity to the Constitution.

3.2 Fundamental Questions of Political Legitimacy

Judicial review of the validity of legislative action under the federal Constitution generates two related problems of legitimacy. One problem, the more straightforward, concerns judicial interpretation. The Constitution is a kind of higher law, a law that all actors in the government should observe. If legislators and the executive are faithful to their roles and oaths of office, they will try to comply with the Constitution. In many respects the precise requirements of the Constitution are not clear. If Congress adopts a law and the President signs it, two branches of government have made a judgment that the law is constitutional. Why should the judiciary, a third non-elected branch, be able to say that the law is invalid, and, given the difficulty of amendment, effectively prevent such a law from being implemented? Why should judges not regard the decisions that the coordinate branches reach about the constitutionality of their own activities as conclusive? When legislators adopt statutes, is not the deliberate decision of the supreme elected body more important than the opinions of a majority of nine justices?

A closer analysis must distinguish among kinds of constitutional issues. Ours is a polity that is federal and whose national executive is independent from Congress. Some federal body needs to be able to say when states have overstepped their powers vis-à-vis the national government, and, within the national government, one may need an adjudicative body to settle disputes about authority between the legislature and executive. The Supreme Court and other federal courts are appropriate bodies to perform these functions.

The more doubtful instances of review occur when someone claims that the national Congress has exceeded its power vis-à-vis the states, that Congress has infringed individual rights, or that a state has infringed individual rights. The system of government would not break down if courts accepted these exercises of legislative power.

Someone might complain that any sketch of an argument against review based on political authority misses the point. Whatever we might recommend for a country (p. 293) adopting a new constitution, the United States has a constitution that gives courts the authority to review all kinds of constitutional issues. If the Constitution confers this authority, courts must exercise it.

Conceivably arguments against review were once relevant to analysis of whether the Constitution provided for judicial review of Congressional statutes, a power that is not explicit in the constitutional text. But that issue was settled long ago, in the famous case of Marbary v Madison,15 and probably according to the design of the framers. Given that review is now firmly entrenched, its wisdom and political legitimacy might seem irrelevant to how modern judges should perform their duties.

Here is how the concern about political legitimacy connects to theories of interpretation. If judicial review of statutes is itself suspect, then judges who exercise this power should do so very cautiously. In the case of reasonable doubt, they should accept what the legislature has done.

In one of the best-known and most influential law review articles, James Bradley Thayer argued that Congress has the primary (though not exclusive) authority to interpret the constitutionality of its actions, and that its determinations are entitled to respect.16 Courts can disregard a statute only ‘when those who have the right to make laws have not merely made a mistake, but have made a very clear one,—so clear that it is not open to rational question’.17 Courts should not make their own straightforward judgments about constitutionality, but should reach a ‘conclusion as to what judgment is permissible to another department which the constitution has charged with the duty of making it’.18 For long periods of the past century, various critics of the Court's decisions restricting congressional power vis-à-vis the states and upholding economic and individual (non-economic) rights have claimed that judges should restrain themselves and defer to the resolutions of legislatures.19

Related issues of political legitimacy, often raised as objections to judges imposing their wills too freely on legislators, go back to the basic idea of constitutional constraint. Why should a modern democratic government be limited in its decisions by the judgments of unrepresentative men who died long ago? We need to separate the three strands of this complaint: the inappropriateness of constitutional restraint; the unfairness of one generation restricting later generations; and the unrepresentativeness of a country's founders.

Should modern governments ever be restrained by written constitutions? In principle, there is no good reason why societies should not declare that some dispositions are off limits for transient majorities of legislative bodies; and for many societies this may be a wise policy. At a time of sober reflection, the members of a society, or the (p. 294) great majority of their representatives, may decide to take certain political issues off the table of legislative choice, to prevent the legislature from giving way to passions or parochial interests. Such a reservation is particularly important to preserve the fairness of the political process itself and the equitable consideration of the interests of all citizens.

Should one generation be able to bind another? This troubling question yields no simple answer. A minor point is that, in truth, a society does not have separate generations but interlocking generations. At no single moment in time does the generation of constitution-makers give way to the next generation, and so on. But the problem that we live under a constitution whose major parts were adopted by persons no longer alive remains.

The best answer to the question of why we should feel bound is that a constitution must prove itself over time as valuable in the life of each generation. One reason why an old constitution is now valuable, and thus is binding, is that it provides a shared set of answers to fundamental political questions. A society must have some ground rules, and it is very difficult to get agreement on new ones. Moreover, as American history shows, a sense of the historical continuity of a constitution can have considerable value for a feeling of community. The legitimacy of the Constitution derives partly from its being in place and partly from its ancient origin. But it legitimately constrains the people of our time because it contributes to our social life.

The problem of unrepresentativeness has a similar answer. What counts over time is what the Constitution provides and how well that fits the notions of justice of subsequent generations. If some very important interests were not represented, that is a reason to be sceptical both that those interests were then fairly treated and that similar interests now receive appropriate constitutional protection. But the test for us now is content, not fairness of representation in origin.

We can easily see how these concerns about the legitimacy of an overarching constitution could affect interpretation. If someone doubts the wisdom of much restraint of democratically elected bodies, he will want the restraints that do exist to be understood modestly. Thus, the fact that a written constitution restrains present majorities may combine with the countermajoritarian feature of judicial review to counsel restraint by courts, to counsel broad acceptance of what legislatures have chosen. Concerns about the legitimacy of an old Constitution can work similarly, to promote a relaxed view of its restrictions. Perhaps judges should exercise judicial restraint, not invalidating all legislation the Constitution was designed to prevent and certainly not reaching out to invalidate other legislation an originalist would deem acceptable.

Worries about unrepresentativeness might work differently, to suggest a variable approach to interpretation. Perhaps judges should ask whether any particular provision resulted from unfair representation and whether the provision continues to impinge unfairly on interests like those that were not represented. If the answer to both questions is ‘yes’, judges might minimize or expand the import of the provision (p. 295) accordingly. Here is a crude example. The original Constitution provides that ‘No state shall pass any … Law impairing the Obligation of Contracts’.20 This formulation may have been a consequence of unjust representation of property and business interests. That might be a reason to interpret the restriction narrowly, not to prevent modern legislative relief for poor debtors.

Beliefs about failures of representation could lead judges, instead, to broaden vague constitutional protections to protect interests that were not adequately represented when the provisions were adopted. Thus, the modern Supreme Court has interpreted the Equal Protection Clause of the Fourteenth Amendment to protect equality for women, although any such protection was far from the minds of the men who adopted the amendment.

3.3 The Countermajoritarian Difficulty

A substantial amount of discourse about constitutional law in the twentieth century has taken up what has been referred to as the ‘countermajoritarian difficulty’. The main focus has been on the undemocratic character of judicial review, but we have seen that the nature of an enduring written constitution itself presents a problem for majoritarian democracy. When, during the first third of the century, a politically conservative Supreme Court was striking down federal and state legislation designed to promote economic justice, many liberals and legal scholars supported Thayer's approach: judges should defer to the decisions of legislatures. In widely read work published in the 1950s and 1960s, Alexander Bickel urged, among other things, that the Supreme Court should be cautious about confronting the political branches, that it should employ ‘passive virtues’ to avoid deciding many issues for which the principled resolution would be rejection of popular legislative initiatives.21

During the late 1930s and 1940s, the Supreme Court began to read the Constitution to provide little restraint when Congress impinged on state power and when states regulated economic interests. At the same time, the Court provided increasing protection under the free speech and free press clauses, under the criminal process sections of the Bill of Rights, and, for blacks and other minorities, under the Equal Protection Clause of the Fourteenth Amendment. Within the Supreme Court and outside, people advanced justifications for the Court's reading some rights expansively and others narrowly. Formulations differed, but the basic idea was that courts need to protect the integrity of the political process and to assure that powerful majorities do not gang up on disadvantaged minorities. This approach was first (p. 296) suggested in a Supreme Court opinion in a famous footnote, footnote 4 in United States v Carolone Products;22 its fullest scholarly development was by John Hart Ely, who wrote that open-ended claims of the Constitution should be understood as ‘representation reinforcing’.23

During the last part of the twentieth century, some scholars, most notably Ronald Dworkin, have challenged the underlying premises of the ‘countermajoritarian difficulty’. Dworkin has argued that the fundamental nature of democracy is not having decisions made by majoritarian votes in legislatures, but affording equal concern and respect for citizens.24 On his view, if courts reach resolutions that promote equal concern and respect, there is no sacrifice of democratic values.

If one assumes that legislatures are more fairly representative of the population than are courts, this position is overstated. When disagreements exist over what is just and useful in a society, there is some benefit to having these disagreements debated and resolved in politically responsible bodies. Jeremy Waldron has argued that, since no one is morally infallible, it is more appropriate for the moral views of democratically elected legislators to implement broad constitutional language than for a few unelected judges to do so.25 Democracy has something to do with who makes decisions, and the most democratic forms of decision are by the people themselves or their elected representatives. If we believed that, over time, we could project constitutional outcomes from legislatures and courts that were equally good and just and that yielded equally clear constitutional principles, we should probably choose decision by legislatures. So, judicial invalidation of legislation carries some cost in terms of democratic values.

Offsetting gains, however, may make such a practice appropriate and desirable in a democracy. Most simply, on some important issues, one may reasonably trust courts more than legislatures. Many constitutional guarantees are designedly countermajoritarian. If temporary majorities, or legislators seeking political gain at the expense of hated minorities, are likely to give way to passion in suppressing unpopular speech, that is a solid basis to have courts determining the constitutional boundaries of free speech.

A more subtle point involves the interplay of legislative and judicial decision. Judicial invalidation may promote some values even if legislatures and courts are equally protective of those values. Free speech provides an apt illustration. If Congress chooses to allow speech to be free, a court will not declare that choice to be invalid. Those who prefer that speech be suppressed rarely have a constitutional argument that it must be suppressed. Thus, whenever Congress chooses in favour of (p. 297) freedom, its choice will stand. The issue of invalidation emerges when Congress chooses to suppress speech. If the courts defer to Congress, the choice to suppress will rarely be overturned. Non-deferential courts will invalidate more statutes that suppress speech. Thus, active judicial invalidation will promote greater freedom of speech than the alternative, even if Congress cares as much about free speech as the courts.

I should enter two caveats, however. This conclusion would not follow if the practice of judicial invalidation caused Congress to care less about free speech, to leave its protection to the courts. Further, my analysis does not apply when both parties to a political or legal dispute have competing claims of constitutional right, as when a free exercise claim to special treatment is met with an argument that special treatment would violate the Establishment Clause.

Any analysis of judicial review in an actual political system should be realistic about what legislatures are like in that system. No one believes that legislatures come close to perfectly representing constituents. If legislatures are too often insensitive to minorities, bend to vested interests, and are touched by corruption (and one does not think that would change much if judicial review ended), one may conclude that courts are not so much less representative than legislatures. (Of course, one must also look at courts as they are, not as they ideally should be.) An appraisal of what set of institutions and authorities best fulfils democratic values needs to take account of actualities as well as ideals.

Finally, a possible virtue of judicial resolution is its principled manner. Members of society may learn about fundamental values from the sustained, disciplined analysis that judicial opinions provide. On the other hand, the argumentative form of those opinions tends to put down those who hold opposing views, and may discourage accommodation and compromise.

In summary, although robust judicial review may involve some cost in terms of values of democracy, that cost may be substantially outweighed by the contributions review can make to democracy. It may be instructive in this regard that among liberal democracies since World War II, the movement among national governments (Germany, Canada, and the Republic of South Africa are notable examples) and in transnational organizations, such as the European Community, has been towards judicial review, not away from it.

Questions about legitimacy illuminate how issues we examined for statutory interpretation vary in their constitutional context. Should constitutional interpretation be evolutionary or originalist? We may contrast evolutionary interpretation with strict originalism and moderate originalism, terms I develop below. I suggest that strict originalism is not defensible, and that a plausible form of moderate originalism is not too different in practice from a moderate evolutionary approach.

(p. 298) 3.4 Undisputed Features of Constitutional Interpretation

It is helpful to begin with some undisputed features of constitutional interpretation. Much more often than is true with statutes, the crucial constitutional language in a case comes down to a general formulation, such as the two religion clauses I have quoted. With respect to most of the key phrases in the Constitution, extensive prior litigation has developed more specific doctrines than the constitutional language provides. Much constitutional adjudication turns out to be more about the outcomes and doctrines of closely related cases than any fresh examination of the original clauses.26 A student of free speech law meets concepts like viewpoint discrimination, public forum, and compelling interest that are crucial modern doctrines but that have no place in the Free Speech Clause itself, in the history that preceded it, or in the contemporaneous writings of the framers. I do not say these are inappropriate principles, only that they are not to be found in any easy way in the textual provision.

Another undisputed feature of constitutional interpretation is that protections must be read to deal with some circumstances that the founding generation did not foresee. In the most extreme examples, a new technology creates a reality no one then conceived. Thus, the First Amendment guarantees freedom of the press; people in 1789 had no idea of radio or television, much less the internet. The Fourth Amendment protects against unreasonable searches; no one then imagined electronic surveillance. Constitutional principles must apply to technical realities beyond the ken of the framers' generation.

Some provisions of the Constitution may have been designedly left open to changing values. A number of Supreme Court decisions have explicitly treated the cruel and unusual punishment clause in this way.

As many, many cases reflect, original understanding (of some kind) is relevant; the controversial issue is whether original understanding (of some kind) should control.

Existing bodies of constitutional law also matter. An originalist thinks that modern judges should aim to carry out the original understanding, that Supreme Court Justices should never have self-consciously departed from that understanding. To put the point this way, however, leaves open what modern Justices should do about solidly entrenched bodies of constitutional law that they believe have departed from the original understanding. Should they now declare that the Free Speech and Free Press Clauses cover only prior restraints, if they think that was the original understanding? Should they overrule Brown v Board of Education if they think people of the time assumed racially segregated schools were consistent with the Equal Protection Clause of the Fourteenth Amendment? A very few scholars have suggested that the modern Supreme Court should abandon bodies of law that are out of line with the original understanding, but most originalist scholars and all originalist Justices assume that judges work within the context of precedents and doctrines built up over (p. 299) time. Although justices should not build upon and expand mistaken directions, they should not throw away all settled standards that fit poorly with original understanding.27 Thus, virtually everyone agrees that some adherence to precedent and settled doctrine should qualify an originalist approach to constitutional interpretation.

Finally, at least among judges and most scholars, there is agreement that decisions in constitutional cases should be principled. In a widely read lecture that became controversial because of its challenge to the Supreme Court's reasoning in Brown v Board of Education,28 Herbert Wechsler asserted that a primary requisite of judicial reasoning is that it be genuinely principled.29 Although Wechsler's phrase ‘neutral principles’ led some to believe that he thought judges could decide cases without making controversial assessments of value, that was not his thesis. Rather, he indicated that a judicial opinion should honestly state the grounds of decision, that the decision should rest on a general principle that reaches other cases; that the judges should be willing to decide as the principle indicates.30 Thus, if in a case involving a civil rights activist whose speech led some members of his audience to hit police officers, the Court said one cannot be punished for speech unless one actually urges others to commit crimes, it should announce such a principle only if it thinks it should apply the principle to members of the Ku Klux Klan and Communist Party as well.

When I say that judges and most scholars agree that constitutional decisions, like other judicial decisions, should be principled in this respect, I mean that they agree that this is the ordinary standard at which judges should aim. People disagree over how demanding the counsel of candour should be,31 over the circumstances, if any, in which candour should give way to the interests of achieving majority opinions or concealing the degree of a court's innovation.32

A more radical critique of judicial principles is that they conceal just how political constitutional decisions are. As this critique is made by critical legal scholars, critical feminists, and critical race theorists, the claim is that apparently neutral principles serve the interests of the dominant class (gender) (race) at the expense of those who (p. 300) are oppressed.33 Such a critique may underlie proposals to shift responsibilities away from courts,34 it could be a basis to introduce more explicit political evaluations or ‘narrative’ approaches into judicial opinions.35 But, given the long tradition of judicial opinions that claim to state principles that rise above immediate political controversies, no one expects a radical shift away from opinions that rely on principles. And most observers continue to think that the aspiration to decide according to principles is a worthy one.

With these features of constitutional adjudication in hand, we can address what divides originalists from evolutionists who believe in a ‘Living Constitution’.

3.5 Original Understanding

If one talks about original understanding, just what kind of understanding matters? An originalist must address this question, but so also must anyone who thinks that original understanding counts for interpretation, even if it is not necessarily decisive. When we face this general question, a range of more discrete questions emerges: whose understanding, what subject of understanding, what kind of attitude, what level of understanding, what reason for following that understanding? Some of these questions are closely similar to ones we examined for statutory interpretation; others vary significantly in the constitutional context.

3.5.1 Why Follow Original Intent?

The question why judges should follow original intent bears on how the other questions might be answered. I shall here put aside the question of why original understanding should be given some weight, and focus on the question whether the original understanding should be afforded decisive significance.

One answer to the question of why judges should concentrate on original understanding is that this strategy of interpretation restrains judges, and unconstrained judicial discretion is an evil greatly to be feared. Another answer is that this approach allows desirable latitude for the political branches. A third answer is that officials must be faithful to the Constitution, and originalism represents fidelity. I shall take these answers in reverse order.

(p. 301) The possibility of fidelity ties in with the problems of legitimacy we have already examined. As I indicated earlier, our obligation to comply with the Constitution comes from its value in the lives of our and future generations. We have no duty of compliance to people who lived long ago; our duties lie to our contemporaries and our successors. But, it might be said, our generation has accepted the Constitution as the structure of government, therefore we owe other members of our present society fidelity to it. This conclusion alone is not enough to get us to originalism.

We have to ask first how our generation has accepted the Constitution. If judges have engaged in non-originalist decisions, say protecting a right to have abortions and forbidding classifications that disfavour women, and most citizens approve of these decisions, it can hardly be said that originalism is a central aspect of the constitutionalism that is now agreed upon.36

Originalism must be defended as achieving desirable values, not on the ground that modern acceptance of the Constitution creates some obvious duty of fidelity to the original understanding.

By itself, originalism does not necessarily lead courts to be deferential to the political branches. Nevertheless, under either or both of two plausible premises, originalism fits well with the idea that legislatures should have broad latitude. The simplest premise is one Thayer advanced. The constitutional design was to make Congress the primary judge of its own authority. A deferential attitude towards legislation conforms with that design.

The second premise is more complex. In most important areas, the original understanding restricted Congress and state legislatures less than the modern Supreme Court has done. Thus, for example, the Free Speech Clause and the Equal Protection Clause had much less scope than the modern Court has given them. If the Supreme Court followed an originalist philosophy, legislatures would have greater freedom than they would be likely to have under a competing approach.

This conclusion is not true across the board. Under the original understanding of the Commerce Clause, Congress would be more restricted than it is now, and an originalist approach to the ‘Contracts Clause’ would limit state legislatures more than has modern Supreme Court interpretation.

Even in areas where originalist interpretation would restrict legislatures less than have modern judicial decisions, it would be more restrictive than some alternatives. Most obviously, it would be more restrictive of legislatures than an interpretive approach under which judges gave extreme deference to all legislative determinations.

Whether or not originalism promotes deference, it does restrict judges, or so proponents like Justice Antonin Scalia claim.37 On their view, judges must seek an (p. 302) objective original understanding rather than writing their own moral and political proclivities into the Constitution.

Just what kind of originalism someone adopts will depend partly on the reasons that he embraces originalism. Similarly, for someone who is not an originalist, the reasons to give original understanding some weight will help determine what original understanding matters. With these insights, we can tackle narrower, more specific questions about original understanding.

3.5.2 Adopters and Readers

The question about ‘whose understanding’ resembles the same question about statutory interpretation, but has less practical significance. We may start with the divide between readers and ‘writers’. Most Supreme Court cases that refer to views at the time the Constitution was adopted have concentrated on the Framers (including, notably, in church-state cases, Thomas Jefferson, who did not participate in drafting the Bill of Rights, but who did draft the influential predecessor, the Virginia Statute for Religious Liberty). The Constitution was drafted and prepared at the Philadelphia Convention and ratified within state conventions. Since ratification was crucial to the Constitution's coming into force, there is no ground to disregard the views of ratifiers in preference to those at the Philadelphia Convention. Thus, if we seek the intent of those who adopted the Constitution, we must include all the adopters (proposers and ratifiers). Similarly, for constitutional amendments, we must include not only those in Congress who voted on the amendments but those in state legislatures that approved them.

Set against the adopters, we have the general population of ‘readers’. We might ask how readers of particular constitutional language would have understood it. In contrast to most ordinary statutory language, the reader's understanding of the most widely interpreted parts of the Constitution could not be formed primarily from a parsing of what terms mean in ordinary English. (This comment is also true about some statutory language.) To understand what the right not to ‘be compelled in any criminal case to be a witness against himself’38 means, for example, one would require a sense of the practices to which the language refers.39 Similarly, the relation of the Free Speech and Free Press Clauses to the law of seditious libel is a matter of understanding history more than of understanding ordinary English.

For some matters in the original Constitution, people's actual views may have been informed as much by writings such as the Federalist Papers as by the language of the document. This is not to say that common understanding of words may never play a part—a generous sense of the term ‘necessary’ plays a role in the construction of the (p. 303) Necessary and Proper Clause in McCulloch v Maryland40—but it figures much less than in textual interpretation of statutes.

We have little evidence that the proposers viewed provisions differently from the ratifiers and the citizenry, and less evidence that the ratifiers had a view different from citizens at large. Of course, we do not know that all these groups had similar views, but judges must rely on the views they find expressed. Thus, so long as judicial conclusions do not depend on the opinions of one or two prominent founders, judges need worry less about exactly whose understanding matters than with respect to statutory interpretation.

With constitutional amendments that were preceded by detailed Congressional debates that would not have been available to state legislatures, interpreters may have more basis to gauge a special understanding of the proposers that may not have been shared by ratifiers or citizens.

If the question whose understanding counts has limited practical bite, it nevertheless possesses theoretical interest. Issues of fair notice to readers of a time long past now seem irrelevant. One can, of course, maintain that law is public, and that original reader understanding continues to be of some importance; but if the Constitution, and most amendments, represent a coherent design of government, perhaps what those responsible for the design understood should now matter more than how provisions would have been seen by typical readers of the public 200 years ago.

This observation may suggest a more general point: that when a document ages, the understanding of those who adopted it should increase in significance vis-à-vis the understanding of original readers. If this is a sound view, the crucial ‘original understanding’ may itself shift over time, as the views of adopters increase in importance relative to those of original readers.

One significant issue is whether the views of framers expressed after adoption should have any significance. A leading example is use in church-state cases of a letter of Jefferson's written after the Bill of Rights was adopted, which uses the phrase ‘wall of separation’.41 One might say that judges should consider no statements that were unavailable to the adopters at the time they acted, since such statements could not indicate meaning to the adopters and could not then be rebutted by those with a different understanding. However, some of these later statements may shed light on understandings at the time amendments were adopted.

On occasion, courts refer to legislation enacted shortly after the passage of constitutional provisions as evidence of how the provisions were understood. This usage is less subject to the idiosyncracies of individual proposers, because it takes majorities in both houses to produce legislation.

(p. 304) 3.5.3 What Attitude?

By ‘what kind of attitude’, I refer to the distinction between hopes, expectations, and beliefs about how a provision should be understood. I assume that, as with statutory interpretation, the most important state of mind is how someone, adopter or citizen, believed that a provision should be understood.

3.5.4 Attitudes about Coverage and Interpretation

This brings us to the two most troublesome questions: what subject of understanding and what level of understanding? By the subject of understanding, I mean roughly to distinguish attitudes about interpretation from attitudes about coverage. Most of the adopters, and at least some lawyers in the general public, entertained ideas about how judges should interpret authoritative legal language, as well as ideas about what the Constitution would be taken to require. Of course, none had experience with written constitutions as the structure for federal governments, but they were familiar with statutes and colonial charters. Some scholars claim that up to the time the Constitution was adopted, judges had interpreted the language of statutes flexibly, paying little or no attention to what the legislators who adopted the statute intended and not reading the language in a strict or narrow way.42 If so, it is reasonable to suppose that people who adopted or read the Constitution conceived that it would be interpreted accordingly, conceived, that is, that the content of its provisions would be construed in an evolutionary way.

If this hypothesis is accurate, judges who stick rigidly to the views of the founding generation about the content of constitutional provisions are not being faithful to their overall understanding. Thus, if the framers wrote provisions designed to encourage later judges to make their own moral evaluations—what Ronald Dworkin calls ‘the moral reading of the Constitution’43—then judges who freely implement shifting moral judgments are following the original understanding. For someone whose reason for originalism is fidelity to the founders, this historical question about their interpretive ideas carries great significance.

But an originalist who thinks Justices should adhere to the original views about coverage might deflect this historical argument in the following way. Whatever members of the founding generation believed, our well-founded notions of desirable judicial performance and restraint lead us to conclude that judges interpreting authoritative texts should stick to what the adopters and original readers believed about the scope of the language. Thus, originalist judges should follow original views about content, not original views about the nature of judicial authority. The plausibility of this position lies in the claimed values of originalism which we have already examined.

(p. 305) 3.5.5 Levels of Understanding

The related question about levels of understanding is perhaps even more central. This question parallels the distinction between purpose and specific intent in statutory interpretation. Members of the founding generation had views about specific practices the Constitution forbade; they also had ideas about the purposes of the provisions and the reasons for their adoption. Suppose that modern judges believe the reasons and purposes behind a prohibition cover practices the founding generation thought were acceptable and that they, the judges, need not stretch the actual textual language greatly to cover the practices. Should judges declare these practices to be unconstitutional? According to what we may call a narrow or strict originalism, modern judges should take as crucial the practices the founders thought were forbidden.44 According to a moderate originalism, modern judges should be mainly guided by purposes and reasons. A moderate originalist might emphasize a very general level of judgment—protect whatever speech promotes liberal democracy and the spread of truth—or some intermediate middle level—protect speech whose prohibition raises the dangers the founders feared from the crime of seditious libel.

The more abstract the level of reference, the harder in practice it becomes to distinguish a judge who is a moderate originalist from one who accepts evolutionary interpretation of a ‘living constitution’. The ‘moral reading’ approach is an example. If a judge finds in the Constitution certain very general values, such as the equality of citizens, and applies these to modern circumstances according to her own moral evaluation, is she a moderate originalist or an evolutionist? Does this depend on her own self-understanding, whether she thinks she is engaged in originalist interpretation or something else?

This example may suggest doubt that this categorization between moderate originalists and evolutionists is itself very useful. Perhaps for constitutional interpretation, instead of employing these general labels, we would do better to focus on exactly what it is that judges should take from the founders and what evaluations they should draw from contemporary moral and political standards or make on the basis of their own moral and political judgment.

One effort in this direction is Laurence Lessig's idea of translation, that modern judges should try to understand constitutional judgments in the context of the founder's generation and then apply those judgments to the very different contexts of modern life.45 In Lessig's own hands, the notion of translation gives modern judges (p. 306) considerable latitude, but one can imagine a version that instructs judges to stick closely to the founders' values, leaving judges much less room to respond to changing assumptions about values.

Judges who make a point of calling themselves originalists tend to emphasize the specific practices the founders thought the Constitution forbade. As I have said, virtually all originalists do acknowledge that constitutional principles may apply to new technologies, but a narrow or strict originalist would insist that the principles applied to the new technologies should (so far as possible) be the principles accepted by the founding generation. Such an approach, faithfully followed, puts more constraint on modern judges than does a moderate originalism that allows for greater development of new principles as social conditions and values change. The major problem with narrow or specific practice originalism is that it may be too constraining when a written Constitution is hard to amend.

I can make these abstract points best in terms of some specific examples. The adopters of the Fourteenth Amendment, and people at the time, did not think the Equal Protection Clause forbade public schools segregated by race (they also did not think the clause forbade special measures to favour former slaves). The decision in Brown v Board of Education is hard to defend according to narrow originalism. However, a moderate originalist may take the Equal Protection Clause as representing judgments that blacks should not be treated by the state as inferior and should have access equal to whites for important state services; recognizing the modern place of public schools, the universal motivating assumptions about racial subordination that lie behind segregation, and the damage that a stigma of inferiority can cause for children, he may find it easy to conclude that school segregation is now unconstitutional.

At the time the First Amendment was adopted, no one thought it restricted damages for ordinary defamation (libel and slander). In the 1960s huge awards for libel of officials in southern states threatened free discussion of the civil rights movement and official reactions to marches and demonstrations. The Supreme Court decided, in New York Times v Sullivan,46 that the First Amendment was meant to bar the crime of seditious libel, a crime that penalized harsh criticism of the government. If public officials could successfully sue newspapers for innocent mistakes of fact regarding government activities, criticism of government could be sharply curtailed. Thus, the Court protected defamation of public officials in their public functions, except when statements were knowingly false or made in reckless disregard of the truth. Again, the result was one a narrow originalist could not easily defend, but one that fitted comfortably with moderate originalism.

The controversial decision creating a constitutional right to abortion was plainly mistaken according to narrow originalists and even difficult to justify for moderate (p. 307) originalists.47 In the crucial case of Roe v Wade,48 the Supreme Court relied mainly on the Fourteenth Amendment's Due Process Clause. Although a few courts prior to that amendment had said that ‘Due Process’ includes some substantive limits as well as procedural requirements, the clause's main thrust was procedural, and one could not draw out of it any general protection of autonomous decision in personal matters. The decision is consistent with moderate originalism only if one casts the relevant constitutional values at a highly abstract level.

The application of the Equal Protection Clause to women presents somewhat similar issues. People in the generation of the Fourteenth Amendment, who denied women the vote, barred them from vocations, and sharply restricted their rights of property, did not assume the Equal Protection Clause would preclude classifications treating women and men differently. One needs to cast the original constitutional value at a high level of abstraction to cover equality for women (and for some other disadvantaged groups to whom protection has been extended). At least in this instance, however, the constitutional language comfortably fits the extension, since the clause itself is not explictly restricted.

An interesting test of anyone's theory is how they regard the argument that capital punishment may be ‘cruel and unusual’. People in the founding generation and at the time of the Fourteenth Amendment clearly accepted capital punishment, and the Constitution itself definitely contemplates its imposition. For example, the Fifth and Fourteenth Amendments forbid the taking of ‘life, liberty, or property without due process of law’ (thus, implying that life may be taken with due process). Someone can argue that the founders themselves had a flexible view of ‘cruel punishment’ and might have envisioned that some day capital punishment, among other punishments, could be so regarded. But one may wonder whether a moderate originalism should allow the Constitution to be interpreted to reject practices whose permissibility the Constitution itself clearly presupposes.

3.5.6 Structural Inferences

I have written thus far as if judges are always focusing on single constitutional provisions (or two closely related provisions, such as the Free Exercise and Establishment Clauses), but an accepted form of constitutional argument relies on fundamental structures of the Constitution.49 Just as one statutory provision is interpreted in light of surrounding provision and the purpose of the statute as a whole, so also one consitutional provision is properly interpreted in light of other provisions and the basic structures of the Constitution. But occasionally the Supreme Court has gone further, (p. 308) and declared a constitutional right that does not rest on any particular provision. Prior to adoption of the Fourteenth Amendment, the Supreme Court declared that states could not interfere with the freedom of citizens to travel to the seat of the national government.50 In the majority opinion in Griswold v Connecticut,51 Justice Douglas found a right of married couples to use contraceptives in the ‘penumbras’ of a number of the Bill of Rights.

Any of the various theories about legitimate constitutional interpretation can find some place for arguments from structure, but a strict originalist will give them less scope than might a moderate originalist or evolutionary theorist. A strict originalist would find an argument from structure to be persuasive only if it is persuasive about the rights and duties the Constitution was originally understood to create.52 A moderate originalist could declare rights beyond those originally conceived, so long as the new rights flowed from the values and practices of the original Constitution. An evolutionary theorist could decide that new rights fitted well with the Constitution as a whole, as it is now understood.

Given the difficulty of amendment, the symbolic importance of the Constitution, and the desirability of continuity over time, during which sweeping changes in social relations and dominant values occur, a rigorously narrow originalist approach to constitutional interpretation is misguided. A plausible originalism must focus on broad constitutional principles, and how these may be applied to changing conditions. If a constitution is to survive over centuries with relatively few amendments, judges either must practise a moderate originalism that sometimes (at least) takes original constitutional values at a fairly high level of abstraction, or they must practise an evolutionary approach that places considerable emphasis on continuity with values underlying the Constitution. At the edges, these two forms of practice become nearly indistinguishable.

3.6 The Modern Reader

It remains to comment on two aspects of constitutional interpretation I have thus far touched upon only briefly. One aspect is the role of the modern reader; the other aspect is the common law quality of much constitutional adjudication.

When one thinks of most constitutional issues, modern reader understanding of the text matters less than it does for typical statutory interpretation. This is so for two reasons. Constitutional provisions mainly restrict government actions. Citizens do (p. 309) not usually rely on them in the way that they rely on what statutes forbid, allow, and give them a right to. I do not mean to say such reliance never occurs. A newspaper publisher might violate a law prohibiting publication of information on how to make bombs, in the belief that the law infringes freedom of the press. The wisdom of the publisher's constitutional judgment may depend on the accuracy of his confidence that he will not be punished. But citizens do not often rely in this way on the privilege against self-incrimination or the establishment clause. Their individual behaviour is not guided by their constitutional interpretation. Fair warning is less important for most constitutional claims than for most statutory claims.

The second reason why modern reader understanding of the text matters less in constitutional cases is that neither modern readers nor their lawyers come to the constitutional text standing by itself. If they did, the general words of most crucial clauses would fail to give much guidance. In any event, modern understanding is deeply influenced by interpretive decisions over the years. Justices may reasonably worry about withdrawing protections previous decisions have afforded, as with the constitutional right to have abortions, but the issue is not how modern readers would grasp the constitutional text standing alone, the issue concerns the reliance of citizens on what the modern Court has announced.

3.7 Common Law Aspects of Constitutional Interpretation

For judges, as well as citizens, constitutional adjudication is incremental. Constitutional rights in practice have more to do with what courts have decided over the years than with what the original document says and what members of the founding generation believed. As so stated, this point is perfectly consistent with originalist interpretation. If clauses of constitutions (and statutes) are highly general, courts attempting to carry out the original understanding may need to develop sets of ancillary doctrines and distinctions not found in the text. After many cases have been decided, new decisions may apply and refine these doctrines without any reference back to the original understanding.

But the common law quality of much constitutional adjudication also suggests a model of restraint that differs from originalism. Judges developing the common law are restrained by the holdings and principles of prior decisions. Perhaps, as David Strauss has urged,53 the most important restraint on judges in constitutional adjudication is of the same kind. This restraint can exist even if decisions have developed the law in a way that is hard to justify on originalist grounds.

The self-conscious evolutionist, thus, has some answer to the charge that he abandons judicial restraint. He can point out that originalist judges are less constrained (p. 310) than originalist theory supposes, both because it is so difficult to pin down original understandings and because the only plausible version of originalism, some version of moderate originalism, leaves so much latitude to modern judges in choosing the level of principles and in applying those principles to modern life. Originalist approaches afford even less constraint because judges must decide how much to accept of the existing body of decisional law that they do not think can be justified on originalist grounds.

The evolutionist, who thinks judges should recognize a ‘Living Constitution’, believes it is much better simply to admit that in many areas the law has drifted from what can be justified by any genuinely constraining originalism, and to recognize that constraint lies in judges preserving reasonable continuity with what their predecessors have done. On this vision, constitutional law develops out of an original constitution, but over time develops in some respects that are not easy to connect to the words of that constitution or the understandings of the founding generation. Such a pattern is not to be regretted as a series of errors that, for reasons of continuity, judges should not now correct, but as what amounts to healthy development of law for a constitution that endures for centuries.

Notes:

(1) See Gary Lawson, ‘On Reading Recipes … and the Constitution’, Georgetown Law Journal, 85 (1997), 1823–4.

(2) Dennis Patterson, Law and Truth (New York: Oxford University Press, 1996).

(3) Ronald Dworkin, Law's Empire (Cambridge, Mass.: Belknap Press, 1986).

(4) Church of the Holy Trinity v United States, 143 US 457 (1892).

(5) Antonin Scalia, A Matter of Interpretation: Federal Courts and the Law: An Essay (Princeton, NJ: Princeton University Press, 1996), 18–23.

(6) Henry M. Hart, Jr. and Albert M. Sacks, The Legal Process: Basic Problems in the Making and Application of Law (Cambridge, Mass.: tentative edition, 1958).

(7) US Constitution, Art. I, §8. cl. 3

(8) US Constitution, Amendment I.

(9) Gopalan, A.K. v State of Mad., AIR 1950 SC 27, 93; Moinuddin v Uttar Pradesh, AIR 1960 All. 484; Venkataramara v Mysore, AIR 1958 SC 255. See M. Jain, Indian Constitutional Cases, 4th edn. (Bombay: N. M. Tripathi Private Ltd., 1987), 853.

(10) We the People: Vol. I, Foundations (Cambridge, Mass.: Belknap Press, 1991); We the People: Volume II Transformations (Cambridge, Mass.: Belknap Press, 1998). See Akhil Reed Amar, The Bill of Rights: Creation and Reconstruction (New Haven: Yale University Press, 1998). On the unsuitability of Ackerman's theory for judicial recognition, see Greenawalt, ‘Dualism and its Status’, Ethics, 104(3) (1994), 480–99.

(11) US Constitution, Art. II, §4.

(12) See the Seventh Joint Debate at Alton, 15 Oct. 1858. Cited in Harold Holzer (ed.), The Lincoln-Douglas Debates (New York: Harper Collins, 1993), 360–2.

(13) See e.g. Cooper v Aaron 358.

(14) See generally Paul Brest, ‘The Conscientious Legislator's Guide to Constitutional Interpretation’, Stanford Law Review, 27 (1975), 585.

(16) Marbury v Madison, 5 US 137, 1 Cranch 137 (1803).

(16) Thayer, ‘The Origin and Scope of the American Doctrine of Constitutional Law’, Harvard Law Review, 7 (1893), 129, 136.

(17) ibid. 144.

(19) For a recent argument against judicial review, see Mark Tushnet, Taking the Constitution Away From the Courts (Princeton, NJ: Princeton University Press, 1999).

(20) US Constitution, Art. I, § 10.

(21) See, especially, Alexander Bickel, The Least Dangerous Branch: The Supreme Court at the Bar of Politics, 2nd edn. (New Haven: Yale University Press, 1986); Alexander Bickel, The Morality of Consent (New Haven: Yale University Press, 1975).

(22) 304 US 144, 152 n. 4 (1938).

(23) Ely, Democracy and Distrust: A Theory of Judicial Review (Cambridge, Mass.: Harvard University Press, 1980).

(24) Dworkin, Freedom's Law: The Moral Reading of the American Constitution (Cambridge, Mass.: Harvard University Press, 1996). A different formulation of this view might be that democracy is morally legitimate only to the extent that it embodies equal concern and respect.

(25) Jeremy Waldron, Law and Disagreement (Oxford: Oxford University Press, 1999).

(26) See Charles Fried, ‘Constitutional Doctrine’, Harvard Law Review, 107 (1994), 140.

(27) See Henry P. Monaghan, ‘Stare Decisis and Constitutional Adjudication’, Columbia Law Review, 88 (1988), 723. No one of whom I am aware is less willing to accept ‘mistaken’ precedents than Raoul Berger, ‘Original Intent and Boris Bittler’, Indiana Law Journal, 66 (1991), 723.

(28) 347 US 483 (1954).

(29) Herbert Wechsler, ‘Toward Neutral Principles of Constitutional Law’, Harvard Law Review, 73 (1959), 1.

(30) See Kent Greenawalt, ‘The Enduring Significance of Neutral Principles’, Columbia Law Review, 78 (1978), 982; Martin Golding, ‘Principled Decision Making and the Supreme Court’, Columbia Law Review, 63 (1963), 35.

(31) See David Shapiro, ‘In Defense of Judicial Candor’, Harvard Law Review, 100 (1987), 731; Henry Monaghan, ‘Taking Supreme Court Decisions Seriously’, Maryland Law Review, 39 (1979), 1, 22–6; Scott c. Idleman, ‘A Prudential Theory of Judicial Candor’, Texas Law Review, 73 (1995), 1307.

(32) See Eugene V. Rostow, The Sovereign Prerogative: The Supreme Court and the Quest for Law (New Haven: Yale University Press, 1962).

(33) See e.g. Charles R. Lawrence III, ‘Race, Multiculturalism, and the Jurisprudence of Transformation’, Stanford Law Review, 47 (1995), 819. Catherine MacKinnon, Toward a Feminist Theory of the State (Cambridge, Mass.: Harvard University Press, 1989); ead., ‘Feminism, Marxism, Method, and the State: Toward Feminist Jurisprudence’, Signs, 8 (1983), 635.

(34) See Tushnet, n. 19 above.

(35) On narrative generally, see e.g. Richard Delgado, ‘Storytelling for Oppositionists and Others: A Plea for Narrative’, Michigan Law Review, 87 (1989), 2411; Kim Lane Scheppele, ‘Foreword: Telling Stories’, Michigan Law Review, 87 (1989), 2073; Julius G. Getman, ‘Voices’, Texas Law Review, 66 (1988), 577; Lynne N. Henderson, ‘Legality and Empathy’, Michigan Law Review, 86 (1987), 1574.

(36) Matters become somewhat more complicated if one thinks that courts consistently employ an originalist rhetoric that persuades citizens, who do not quite acknowledge that a number of decisions they like fail under originalist standards.

(37) Antonin Scalia, A Matter of Interpretation.

(38) US Constitution, Amendment V.

(39) For example, the language seems clearly to provide that a criminal defendant need not testify, but may a witness in a civil case be compelled to admit to crimes? The language itself is not revealing.

(40) 17 US (4 Wheat.) 316 (1819).

(41) Everson v Board of Education, 330 US 1, 16 (1947). The expression was used earlier in Reynolds v United States, 98 US 145 (1878).

(42) See H. Jefferson Powell, ‘The Original Understanding of Original Intent’, Harvard Law Review, 98 (1985), 885.

(43) Dworkin, Freedom's Law, n. 24 above.

(44) An illuminating categorization of various approaches is Paul Brest, ‘The Misconceived Quest for the Original Understanding’, Boston University Law Review, 60 (1980), 204. See also Daniel A. Farber, ‘The Originalism Debate: A Guide for the Perplexed’, Ohio State Law Journal, 49 (1989), 1085.

(45) Lawrence Lessig, ‘Fidelity in Translation’, Texas Law Review, 71 (1993), 1165;Understanding Changed Readings: Fidelity and Theory’, Stanford Law Review, 47 (1995), 395.

(46) New York Times v Sullivan, 376 US 254 (1964).

(47) If the Ninth Amendment or the Privileges and Immunities Clause of the Fourteenth Amendment was designed to protect some pervasive zone of personal life, a moderate originalist might believe that abortion now fits within that zone.

(48) 410 US 113 (1973).

(49) See Charles Black, Structure and Relationship in Constitutional Law (Baton Rouge: Louisiana State University Press, 1969).

(50) Crandall v Nevada, 73 US 35 (1867).

(51) 381 US 479 (1985). It may be worth noting that only one other Justice joined the Douglas opinion without joining any other. For discussion of recent cases, see Brannon P. Denning and Glenn Harlan Reynolds, ‘Comfortably Penumbral’, Boston University Law Review, 77 (1997), 1089.

(52) More precisely, one would have to consider the understanding at the time the most recent relevant provisions were adopted.

(53) David A. Strauss, ‘Common Law Constitutional Interpretation’, University of Chicago Law Review, 63 (1996), 877.