"The Relations Between the Courts and the Law."
Both The Legal Process and 'Relations' have two great virtues, on top of those that come simply from Hart's penetrating intellect and intimate knowledge of law. They see law, and American law in particular, as systems, as wholes composed of parts that work together. Second, they develop and use legal concepts that are both sufficiently abstract to describe phenomena that recur from one substantive area of law to another and sufficiently concrete to be readily applicable to those phenomena. They are applied legal theory, and models of successful legal scholarship.
This contribution seeks to apply Hart's method to three questions that fit on his agenda. All concern the ways in which courts and the law interact. The first concerns the ways in which courts and executive officials fit into the complex American legal system in which a single body of substantive law is administered by institutions of two different levels of government. The next two, closely related to one another, extend Hart's inquiry into the interactions between legal rules and government institutions by probing the standard short hand according to which courts make law in common-law systems and unmake law when they find unconstitutionality in systems that have American-style judicial review.
Although Hart's analytical concepts are general, his specific focus was on the American legal and constitutional system. For Henry Hart, that may have been more because of his readers' limitations than his own. In my case, that much parochialism results from the sound advice to all authors, write what you know about. Henry Hart knew about law in general, but most of us must be more narrowly focused.
I THE COURTS AND THE LAW AND THE EXECUTIVE AND THE LAW
On one important issue 'The Relations Between State and Federal Law' is boilerplate: it appears in a passage from a Supreme Court opinion that the Court itself quotes, leaving Hart's words embedded in multiple levels of quotation marks. 'The requirement that a state court of competent jurisdiction treat federal law as the law of the land does not necessarily
include within it a requirement that the State create a court competent to hear the case in which the federal claim is presented. The general rule, "bottomed deeply in belief in the importance of state control of state judicial procedure, is that federal law takes the state courts as it finds them". Hart, [The Relations Between State and Federal Law], 54 Colum. L. Rev. [489, 508 (1954)]; see also Southland Corp. v. Keating, 465 U.S. 1, 33 (O'Connor, J., dissenting); FERC v. Mississippi, 456 U. S. [742, 774 (1982)] (opinion of Powell, J.) The States thus have great latitude to establish the structure and jurisdiction of their own courts.' (4)
That passage includes two standard principles concerning the relations between state and federal law and state and federal courts. The first has to do with the law, and the complex, hierarchical legal system that Hart analyzed in such depth. The American legal system includes rules that originate with the States of the Union and with the United States as a separate law-making entity. The Supremacy Clause of Article VI makes that multiplicity into a hierarchy: if the two conflict, federal law prevails. As Hart stressed, federal law is not complete, but rather interstitial. (5) When federal law does not operate, state law continues to do so. Federal law may modify or displace or otherwise constrain state law, because of its hierarchical superiority. Together, state and federal law thus make 'one system of jurisprudence, which constitutes the law of the land for the state.' (6)
That system of jurisprudence is administered by two systems of courts, state and federal. (7) The same substantive rules apply in both. Matters are more complicated with respect to jurisdiction, procedure, and similar bodies of law. The question as to which the Court quoted itself quoting Hart in Johnson v Fankell was whether state courts, in applying federal substantive law, had to follow a federal rule regarding appeals. The answer to that question in Johnson was yes, but as the Court stressed by citing 'Relations', that was not a foregone conclusion.
A case central to Hart, Erie Railroad v Tompkins, (8) dealt with the converse situation, in which federal courts apply state law, as they often are called on to do. Erie held that in identifying the unwritten law of any State of the Union, the federal courts are absolutely bound by the state courts' account of that law. That is so, said Erie, even when the state cases apply principles of private law generally shared among common law jurisdictions. That case put an end to the so-called federal general common law, which arose when federal courts took their own view of the commonly shared unwritten principles of private law.
Shortly after Erie, the Court considered that case's application to cases in which a federal court is called on to provide equitable relief with respect to a substantive right created by state law. The Court in Guaranty Trust Co. v York (9) invoked the jurisprudential underpinnings of Erie and found that in those circumstances a federal court is 'in effect, only another court of the state' and 'cannot afford recovery if the right to recover is made unavailable by the State nor can it substantially affect the enforcement of the right as given by the State'. (10) Hart believed that the Court had seriously overemphasized the importance of uniformity of decision between state and federal court, as opposed to the need to recognize 'the state courts as organs of coordinate authority with other branches of the state government', which he believed to be the central ground of Erie. (11) That focus on uniformity and the avoidance of incentives to pick federal over state court had blinded the Court to the difference between the great issue of Erie and 'the collateral problem ... of the dividing line between state substantive law, both statutory and decisional, and federal remedial and procedural law'. (12) In Hart's view, it was no 'offense to the ideals of federalism for federal courts to administer, between citizens of different states, a juster justice than state courts, so long as they accept the same premises of underlying, primary obligation.' (13)
Hart was also seriously disappointed in Klaxon Co. v Stentor Electric Manufacturing Co., Inc., (14) in which the Court addressed the choice of law principles that federal courts should follow in deciding cases under state law after Erie. Relying on the need to maintain 'equal administration of justice in coordinate state and federal courts sitting side by side', the Court instructed federal trial courts to follow the choice of law principles of the state in which they sit. (15) Hart thought that questions of choice of law 'are essentially federal, in that they involve, by hypothesis, more than one state. To the solution of no other type of controversy is the diversity jurisdiction better suited.' (16)
Underlying Hart's criticisms of Guaranty Trust and Klaxon was his understanding of the interaction between the American federal system and different categories of legal rules. In The Legal Process, Hart and Sacks drew 'a major distinction between officially addressed and privately addressed directions'. (17) Rules of primary conduct--conduct not directly connected to litigation--are addressed to private people, while rules regarding official remedies are addressed to government officials. (18) In similar fashion, choice of law rules are addressed to courts. Hart did not deny that sophisticated parties might plan their affairs with knowledge of the law of remedies and conflicts, but for the analyst to focus on rules of that kind while neglecting norms about primary conduct would be to reason backwards not frontwards. (19)
The distinction between privately addressed and officially addressed rules in a federal system explains how Hart could endorse Erie without admiring Guaranty Trust or Klaxon. The one system of law administered by two systems of courts is the substantive law covered by Erie. The judges of the two systems of courts have rules addressed to them as officials, which retain their distinct character as state and federal law. A federal court administering federal remedial law to vindicate a state primary right thus might indeed be able to provide a juster justice than a state court, with different remedial capacities and principles, acting to protect the same right.
When Hart gave the Supreme Court the phrase that federal law takes the state courts as it finds them, he was applying this vision of federalism and the different relations of courts to primary and officially-addressed rules. He also provided some of the building blocks for a solution to a closely related problem. In Printz v United States, (20) the Court held that Congress could not 'commandeer' state executive officials by imposing on them affirmative obligations to carry out federal law. In support of the statute, the government had argued that such duties for state executive officials was parallel to the obligation of state court judges to apply federal law, which is well established. (21) Writing for the Court, Justice Scalia replied that state court judges are required to apply federal law by the Supremacy Clause, an obligation that says nothing about state executive officials. (22) By itself that answer is inadequate. The Supremacy Clause creates one system of jurisprudence that is the law of the land for the states and for all their branches of government.
A better account of Printz explains how the Constitution makes federal law supreme without turning the States into federal agencies subject to congressional control. As Hart explained, while state courts apply to parties the one system of primary law that reflects federal supremacy under Article VI, federal law takes those courts as they are created and empowered by state law. (23) If a case does not come within a state court's jurisdiction under state law, the court has no occasion to apply federal law, notwithstanding that law's supreme status. Under Erie, state law takes federal courts as it finds them, with their own jurisdictional limitations and. Hart believed, their own remedial capacities.
Applying Hart's insights to commandeering of executive officials requires some thought about the relations of the executive to the law. The work of executive officials is usually quite specific, and the affirmative duties to perform that work are similarly specific. Even large components of the federal executive, like the Department of Defense, are specialized in function. The Defense Department does not operate national parks any more than it operates state parks. As a result, executive officials administer some but by no means all legal rules. That is true for the states as well as the federal government. State executive officials have state-law duties that direct them to implement particular parts of the legal system, in particular those parts derived from state law.
State courts, by contrast, are charged by state law with deciding cases on the basis of the legal rules that govern the relations between the parties. That body of rules addressed to private parties and not to government actors, as Hart and Sacks would put it, reflects the principle of federal supremacy. As a result, when Congress makes or changes a federal statute that bears on a case, it changes the law that a state court adjudicating the case must look to. Congress thus can affect how state courts must decide without giving them any affirmative obligations with respect to their official function. Those affirmative obligations come from rules addressed to officials, not private people, and come from state law. Nothing similar happens with respect to state executive officials when Congress changes the body of law that is addressed to private people, because state executive officials do not have a state-law duty that requires that they look to that entire body of law. Rather, like federal executive officials they have much narrower affirmative duties to implement particular legal rules.
The difference between rules addressed to private people and rules addressed to officials, combined with the federal system that distinguishes between the two, further combined with the different functions of courts and executive officials, together defeat the analogy between state courts and state executive agencies discussed in Printz. This point is not conclusive as to the entire problem of commandeering state executive officials by Congress, but it does resolve one question. On that question, 'The Relations Between State and Federal Law' provides the structural template that can be applied to understand the two great principles of American constitutionalism, federalism and separation of powers.
II JUDICIAL LAW-MAKING
The Legal Process is concerned very much with the operation of courts that set and follow precedent, and 'Relations' includes state and federal common law in the legal universe that it explores. Although Hart of course was aware of the relationship between rules of judicial precedent and the ability of the courts to create what he called decisional law, he did not explore that relationship in much depth. He and Sacks did, however, stress the difference between statutory and decisional law, and the fact that the latter arises because of rules of stare decisis, (24) I will argue that an exploration of the origins of decisional law yields a better understanding of some of the issues with which Hart was concerned. My inquiry is inspired by his approach because it uses characteristically legal concepts in an attempt to explain the structure and operation of the legal system and the government institutions that work in it.
It is also inspired more specifically by Hart's rejection of the tendency to reason backwards from remedies to primary rules. (25) That direction of analysis may seem to rest on hard-headed realism: what really matters is what a court will actually do, so that is the place to start, and not with the principles the courts say they apply in reaching the conclusions that really matter. The statement that courts make law also may seem to be hard-headed realism: what matters is what the courts will do, so what they will do is the law. The same hard-headed realism, applied to constitutionalism, says that courts unmake law just as they make it. When a court holds a statutory rule unconstitutional, it makes that rule invalid.
The line between being hard-headed and being rock-headed is an important one. A careful examination of the reason one might say that courts make and unmake law is much more illuminating than the simple assertion that they do, and shows the limits of those assertions.
A Do Courts Make Law?
'I am not so naive (nor do I think our forebears were) as to be unaware that judges in a real sense "make" law. But they make it as judges make it, which is to say as though they were "finding" it--discerning what the law is, rather than decreeing what it is today changed to, or what it will tomorrow be.' (26) Justice Scalia's characteristically pointed formulation helps explain Hart and Sacks' distinction between statutes and judicial decisions. The latter are not free-standing law because they by hypothesis rest on and apply pre-existing legal principles. Because they are not law as statutes are, judicial decisions must have some other source of authority. That source is the principle that later courts follow earlier courts.
The hard-headed, realistic statement that courts make law thus must be qualified to reflect the reason it is true to say that they do so. Courts make law insofar as they set authoritative precedents, so the scope of their law-making power depends on the rules of precedent that later, sometimes other, courts will follow. (27) The American federal judicial system illustrates the point. One somewhat surprising feature of that system is that United States District Courts--the main trial courts in the federal system --do not regard district court opinions, even their own, as anything but persuasive authority. Elsewhere the federal judicial system exhibits quite strong but often limited rules of precedent, vertical and horizontal. District courts regard themselves as absolutely bound by the precedents of the court of appeals for their circuit, and all lower federal courts regard themselves as absolutely bound by precedents of the Supreme Court of the United States on federal questions. In general, the courts of appeals for the circuits sit in three-judge panels, although those courts consist of more than three judges, sometimes many more. Panels follow a rule of absolute stare decisis horizontally: their doctrine is that only the court sitting en banc may reject an existing precedent. So the federal courts are no strangers to binding precedent, but they are also no strangers to opinions that are not binding at all. (28)
For practical purposes, therefore, all American federal courts make law only for some future disputes, and the identity of the disputes is determined by the rules of precedent. That is true even for the Supreme Court of the United States. Under Erie, if the Court has occasion to pronounce on a question of state law, and the highest court of the relevant state later takes a different view, the Court itself and other federal courts will follow the more recent state decision. As a result, the answer to the question, does this court make law, can be, yes to some extent.
Not only do the rules of precedent determine whether a court makes law and how strongly, they also determine the role of judicial opinions. (29) It has long been standard practice in jurisdictions that build their law on the common law to find the precedential force of a decision in its ratio decidendi. The ratio need not be found exactly in the opinion of the earlier case, and may to some extent depart from it. (30) As Neil Devins and David Klein have recently pointed out, however, in recent decades American courts not of last resort, including in particular the federal circuit and district courts, have come much more to treat the text of appellate opinions as if they were statutes. (31) American courts are much less likely today to reconsider the reasoning of a precedent they regard as authoritative.
In addition to showing the importance of the rules of precedent, the new practice may seem paradoxical. A standard defining feature of the common law is that it is unwritten, and has and can have no canonical verbal formulation. (32) The common law is found in practice, and the same practice can have more than one description in words. That is the central difference between the common law and statutes, including statutes that codify it. One is captured in particular words and the other is not. But American courts are coming to regard particular words as the law, although those words are not found in statutes but in judicial opinions.
Some seeming paradoxes are actually sources of illumination, and I believe that such a seeming paradox is at work here. The illumination is to see that judicial opinions and the common law are indeed distinct, and that the relationship between them is indeed constituted by the rules of precedent. The practice of treating opinions like statutes is itself a practice, and opinions have legal effect similar to statutes just because they are so treated. (33) The practice of treating opinions like law has no justification other than judicial practice, so unless it is simply lawless it is lawful because the pattern of judicial decision in this respect constitutes the law. That pattern, like the common law, cannot be captured in any text; if courts were to deny that they were deciding that way they still would be doing so.
In the absence of any statutory or constitutional rules concerning the rules of precedent, those rules are found in the intrinsically unwritten practice of courts, especially lower courts. Judicial decisions that set precedent, whether their precedential effect is found mainly in their outcome or the words of the courts' opinions, can be said to be law insofar as those unwritten rules make them authoritative.
A careful (and I hope Haitian) analysis of the grounds of judicial law-making leads to two conclusions on issues of great interest to Hart, one of which is highly heterodox. The first is that the rule of Erie does not follow from the obligation of federal courts to apply state law. Erie may be thought to follow from that obligation, along with the further premise that the content of unwritten state law is necessarily found in the decisions of state courts about it. If the written law of a state produces that equation Erie is correct for the law of that state, but in the absence of written law, the argument must rest on the claim that state court decisions are law in a system of unwritten law. (34) That claim, however, is true only insofar as rules of precedent make it true. The question in Erie was whether state precedents were binding on federal courts. The principle that judicial decisions are law insofar as the rules of precedent make them so cannot be used in deciding what the rules of precedent are, because it depends on an answer to that question. Well-worn phrases like decisional law and judge-made law can lull even the sharpest analysts into passing over the complexities that such phrases capture and sometimes obscure through their brevity.
Scepticism about the reasoning of Erie (if not its result) comes and goes in American scholarship, and right now is fairly common. (35) On another issue important to Hart I will urge a position that is much more heterodox: federal law does not require that state courts follow the precedents of the Supreme Court of the United States, even with respect to federal law.
That state courts have such an obligation is such an ingrained assumption of American lawyers that many of them probably would be at a loss to explain why it is true. A natural first place to look is the Supremacy Clause of Article VI, which does indeed require that state courts decide according to The laws of the United States.' (36) Judicial decisions and opinions are law, however, only insofar as the rules of precedent make them so. Hence, as with Erie, the assumption that they are law cannot be used in determining the content of the rules of precedent.
Another natural place to look is the word 'supreme', which the Constitution uses in vesting the judicial power of the United States 'in one supreme Court and in such inferior Courts as the Congress may from time to time ordain and establish'. (37) The argument in favour of the standard position draws some force from the fact that 'supreme' is a description, not part of a name; statutes call that tribunal the Supreme Court of the United States, but the Constitution does not. The word, however, need not imply anything about the rules of precedent, because it has another implication that is important enough to justify its use: a supreme court is one not subject to further review in some important sense. (38)
A more plausible argument comes from the Supreme Court's appellate jurisdiction. The Constitution sets out the cases and controversies to which the judicial power of the United States extends, puts some of those in the Court's original jurisdiction, and then gives it appellate jurisdiction over all the others, subject to such exceptions as Congress may make. (39) The default arrangement, then, is one in which the Court has appellate jurisdiction over state and federal courts with respect to the great majority of the cases that any federal court can hear. Although the statutes establishing the federal judiciary have never given that Court all of that appellate jurisdiction, they have always given it authority to decide appeals from state courts in large categories of cases involving federal law. (40) While the details of that jurisdiction have varied somewhat, its focus on cases turning on federal questions, and its breadth within that category, have been a constant.
Today, appellate jurisdiction may seem naturally to imply an obligation on the part of the court lower in the hierarchy to follow the precedents of the court to which its decisions may be appealed. An important feature of the actual rules of precedent in the federal courts is that they strongly (but not perfectly) reflect that principle. District courts follow the precedents of the court of appeals to which their decisions may be appealed and regard the decisions of other federal circuit courts to be of only persuasive force. (41) Fairness to litigants strongly supports that rule of precedent. It is burdensome to require that a party who will prevail on appeal to go to the trouble of actually taking the appeal. If lower courts fail to follow the precedents of the courts that can reverse them, however, such burdensome appeals are encouraged rather than avoided.
That fact certainly can justify the familiar rule of precedent, including the rule that state courts follow by regarding themselves as bound by the Supreme Court's holdings as to federal law. The Constitution, however, does not mandate every reasonable rule. For example, it does not by itself prevent private parties from trying to wear one another out with pointless appellate proceedings. Courts and legislatures may adopt rules against frivolous appeals, but the Constitution itself does not contain any. To say that the standard rule of precedent about state courts is sound policy is by no means properly to attribute it to the Constitution.
Some additional thought about American legal practice when the Constitution was framed casts serious doubt on an inference to standard practice from the structure of the judiciary. While the Constitution's grant of appellate jurisdiction contemplates decisions in cases, it does not by itself contemplate judicial opinions. Early practice does not suggest that the production of opinions was regarded as part of the Supreme Court's constitutional task. Congress did not provide for an official Reporter of Decisions until 1817, more than a quarter century after it first organized the Court in the Judiciary Act of 1789. (42) In the Court's first decade, the Justices generally delivered seriatim opinions, rather than producing an opinion for the court as a whole. (43) Seriatim opinions require state courts to puzzle out the meaning of each case, making their task more difficult when performed in good faith while giving more room for manoeuvre to state judges seeking to minimize the constraint imposed by the Court's decisions. If the Justices thought that state courts had a constitutional obligation to follow their precedents, they probably would have formulated those precedents in a more user-friendly form.
This line of reasoning does not deny that judges, state and federal, expected that state courts generally would regard themselves as bound by the federal-law precedents of the Supreme Court. Reasonable state court judges would wish to spare litigants pointless appeals, and so when they could discern a clear trend in the Supreme Court's cases they would find precedential force in that trend, regardless of their own views. State courts' own practice of stare decisis, founded on considerations of policy, thus would lead them to give more-than-persuasive force to Supreme Court decisions when an appeal to the Court was available to litigants. That precedential force would come from state and not federal law.
The argument that derives rules of precedent from the Supreme Court's appellate jurisdiction encounters another difficulty, which in turn involves an issue of great importance to Hart. The judicial power of the United States extends to controversies between citizens of different states, without regard to the substantive law governing the dispute. (44) From the foundation of the Constitution to today, the so-called diversity jurisdiction has included many cases that turn only on the law of a particular state, with no element of federal law relevant. (45) In the jurisdictional statutes, Congress has never given the Supreme Court appellate jurisdiction over state courts in diversity cases. (46) Throughout the Constitution's history, federal diversity jurisdiction has always originated in the lower federal courts, with such appellate review in the Supreme Court as Congress provides. The Constitution, however, does not require that arrangement. Indeed, by default the Court would have appellate jurisdiction over diversity cases in state court. Congress has prevented that result by statute since the very first judiciary act in 1789, but could institute diversity-based appellate jurisdiction any time it chose.
Appellate jurisdiction based on the identity of the parties rather than the substance of the law fits very poorly with the principle that rules of precedent follow appellate jurisdiction. The same question of Virginia law, for example, can come before a Virginia court in a case involving two citizens of Virginia as well as in one involving a Virginian and a citizen of Maryland. (47) Were Congress to allow the Supreme Court to exercise its diversity appellate jurisdiction, the latter case but not the former would be subject to appeal to the Court. If the Court had a holding on the issue, rules of precedent based on appellate jurisdiction would bind the Virginia court to the federal holding in some cases and not in others.
Henry Hart himself had a solution to that problem. That solution is not consistent with a derivation of rules of precedent from appellate jurisdiction. Hart endorsed the long-standing practice that when the Supreme Court has appellate jurisdiction over a case from state court, that jurisdiction generally extends only to questions of federal law. (48) That practice is based primarily on the jurisdictional statutes, which have never given the Court its constitutionally-permissible appellate diversity jurisdiction, and always either explicitly or by judicial inference have limited its review to federal questions in cases in which the federal question supports jurisdiction but issues of state law were also decided below. (49) The central case, from the nineteenth century, is Murdock v Memphis. (50) The original judiciary act explicitly provided that when the Supreme Court of the United States exercised appellate jurisdiction over a state court, it was limited to review of issues of federal law. (51) In Murdock, the Court confronted an amendment to the jurisdictional statute that arguably eliminated the restriction and gave it appellate jurisdiction over the entire case, not just the federal questions. The Court found that the revised version of the statute implicitly retained the restriction, and thereby deliberately avoided the question whether the Constitution would allow a federal court to reverse a state court on a question of state law. (52)
Together, Erie and Murdock require federal courts to follow the views of state courts on state law. They thereby create a rule of precedent that cannot be derived from the rules of appellate jurisdiction. Federal courts are bound by the decisions of state courts concerning state law, even though no state court has appellate jurisdiction over a federal court. State courts are not bound by the decisions concerning state law of the one federal court that has jurisdiction over them, but rather that court is bound by the state-law interpretations of courts that it reviews. The principle that state law is for state courts may or may not be sound, but it cannot be derived from the system of appellate jurisdiction allowed by the Constitution. The legal and judicial system described so penetratingly in 'Relations' thus defies the principle that rules of precedent derive from appellate structures.
If state courts have an obligation to follow the precedents of the Supreme Court, it cannot be because that Court can reverse them on questions of federal law. Any such obligation must derive from structural and functional considerations. (53) At this point a fundamental question about the Constitution arises, one on which I with some diffidence depart from Hart. In my view, a strong practical argument that is not rooted in the Constitution's text is a consideration of policy, not a constitutional norm. American federalism very likely works better if the country has a single highest court with respect to federal law, and all other courts follow that tribunal's precedents. Practical considerations figure strongly in the formulation of rules of precedent, and state courts should look to them in developing their rules of precedent. In the absence of an actual rule of federal constitutional law, or a federal statute, those rules of precedent remain part of state and not federal law.
The Constitution, in my view, thus leaves the question of judicial precedent to sub-constitutional law, which for state courts on this issue is state law. When the courts involved are state courts, the Constitution takes their rules of precedent as it finds them.
B Do Courts Unmake Law?
In some constitutional systems a judicial body performing constitutional review can cause an enacted statute to cease to be legally operative after it has gone into force, in effect repealing it. (54) The Federal Convention considered creating a Council of Revision which would keep laws from going into force by exercising the veto function ultimately vested in the President and would include the judges of the supreme court. (53) According to Madison's notes of the Convention, a leading objection was that the judges, as expositors of the laws, should have no role in making them. (56)
As the distinction between the veto and a later case in court underlines, the form of judicial review that has developed under the Constitution takes place after a law has been adopted through a legislative process in which judges have no role. (57) It is common to say that when an American court finds a statutory or other subconstitutional rule to be unconstitutional, it invalidates the rule. (58) Taken literally and combined with the fact that courts act after legislation has been adopted, that formulation implies that courts unmake law the way legislatures do when they repeal a statute and the way the Constitutional Court of Austria does. Repeal or amendment causes a norm that previously had been binding to cease to be so, or to be binding only in an altered form. Repeals and amendments generally are not retroactive; they leave in place legal results already reached under the repealed or amended norm.
Despite common use of the terminology of invalidation, the canonical account of American judicial review does not exhibit it. According to Marbury v Madison, (59) judicial review results from the Constitution's hierarchical superiority to other law and the courts' obligation to decide cases according to law. In that hierarchy, when a lower-level norm contradicts a higher-level norm the lower-level norm is inoperative. Higher-level norms thus provide criteria of validity, just as the requirements of enactment do, and a federal statute that is contrary to the Constitution is no more an operative legal rule than is one that has been subject to a presidential veto that was not overridden. According to this account, courts do not bring invalidity about. Rather, they recognize it, just as they recognize that statutes have been adopted without themselves legislating.
But of course courts make law, as everyone knows, even in the form of recognizing existing law. So when they supposedly recognize that under the existing hierarchical law a sub-constitutional norm never was valid in the first place, they are really making it invalid. Like many seemingly sophisticated arguments, that one is too facile. Courts decide cases and often set precedents. Insofar as they may be said to make and change law, and to change a previously valid norm to one that is not valid, they do so through those processes. Insofar as deciding cases and setting precedents does not resemble legislation, courts do not unmake law any more than they make it. An examination of the actual practice of the American constitutional system shows that there is some force to references to judicial invalidation on constitutional grounds, but that force is quite limited.
In order to better analyze the question whether and to what extent courts engaging in constitutional review actually make norms invalid, I will identify the dimensions on which any divergence between the recognition and the creation of constitutional invalidity would manifest itself. The first is time. If constitutional invalidity is produced by the legal hierarchy itself, then it arises when an invalid norm is adopted and is recognized by the courts later if appropriate. According to this reasoning, findings of invalidity should be applied retroactively as measured from the point of decision. That is, a court that finds a norm to be invalid on constitutional grounds should decide the case as if it had never been operative. By contrast, if the court makes the norm invalid, then in any applications of it prior to the court's invalidation it should be treated as having been valid at the time.
In one important respect standard practice is consistent with the Marbury account and inconsistent with judicial invalidation. When private conduct is subject to a sub-constitutional norm, in an adjudication that takes place after the conduct, if the court finds the norm to be inconsistent with the Constitution it will treat the norm as void ab initio and will not penalize the private conduct in question. (60) A criminal defendant prosecuted under an unconstitutional rule may not be convicted.
In other contexts, however, a judicial decision may reasonably be said to change the applicable constitutional rule and so create a dividing line in time. Perhaps most familiar in American constitutional law today is the doctrine of qualified immunity for law enforcement officers. (61) When a police officer, for example, conducts a search that is not contrary to the Fourth Amendment under existing cases, the officer will enjoy immunity from damages even if the court at the time of trial decides that the search was unlawful. Once that principle is clearly established in judicial holdings, however, qualified immunity is no longer available in damages actions.
As the example of qualified immunity indicates, many doctrines are designed to deal with legal uncertainty. Often the rationale for those doctrines is to protect decisions that rely on reasonable judgments about the law, even if those judgments are inaccurate. In a system that features strong judicial precedent, reasonable judgments about the law will reflect existing precedents and will change as the stock of precedents changes. The doctrines concerning case law developments are not specifically concerned with constitutional issues, though they generally apply to them as to other sources of uncertainty. (62)
In a legal system with much uncertainty in which courts play an important role in reducing uncertainty, both by resolving concrete disputes and by setting precedents, it is not surprising that doctrines exist to manage uncertainty. In particular, it is not surprising that parties are not always strictly held to the law as identified by a court before which they appear. When a doctrine designed to protect reasonable expectations based on judicial pronouncements operates in a case involving constitutionality, it may produce an effect that resembles judicial repeal of a statutory rule. That result, which derives from principles unconnected to constitutional review as such, is quite different from a free-standing principle that in constitutional review, courts invalidate legal rules. Thus, although it is fair to say that judicial changes in the law through changes in applicable precedent often operate the way legislation operates with respect to time, it would be misleading to say more specifically that courts operate like legislatures when they find sub-constitutional rules to be invalid. The latter formulation would be misleading because it would suggest that the function of constitutional review is distinct from other aspects of the judicial function in which governing precedent may change.
Another dimension on which recognition and creation of invalidity diverge is space, or more accurately the scope of vertical stare decisis, which is roughly defined in geographic terms in the United States. If courts actually operate like legislatures, they should be able to nullify an existing rule throughout the country, just as Congress can. As I discussed above, the rules of precedent generally follow the lines of appellate authority. Many American courts have only limited appellate authority, and as a result only a limited ability in effect to make a law invalid prospectively. In particular, the precedent-setting authority of the federal courts of appeals is geographically limited. The Fourth Circuit cannot repeal a statute for the whole country, the way Congress can. For all federal courts but one, the analogy between constitutional review and legislative power breaks down.
A third and less obvious dimension on which recognition and creation of invalidity diverge involves the substantive rather than the geographic scope of precedent. The Supreme Court of the United States in Obergefell v Hodges (63) found that opposite-sex marriage requirements were unconstitutional. When Obergefell was decided, the Supreme Judicial Court of Massachusetts in Goodridge v Department of Public Health (64) had already recognized same-sex marriage in the law of that state. Had Massachusetts shortly thereafter changed its constitution to impose an opposite-sex marriage requirement, the new provision would have been unconstitutional under Obergefell.
A sequence of events like that is consistent with the hypothesis that the Constitution prevents statutory rules that are inconsistent with it from ever becoming operative law, and that courts set precedents concerning the meaning of the Constitution. It is not consistent with the hypothesis that statutory rules that are inconsistent with the Constitution are effective until a court acts on them the way a legislature does and thereby changes the applicable legal rules. In this respect, judicial invalidation on constitutional grounds is once again a figure of speech. What the courts actually do is more like adjudication than legislation.
Another dimension along which the difference between judicial recognition of and creation of constitutional invalidity would be manifest concerns separation of powers. If courts make statutory rules invalid, then those rules bind executive officials until a court engages in nullification. By contrast, if courts simply recognize preexisting invalidity, executive officials should be able to do so too. According to the simple logic of the legal hierarchy, executive officials should no more enforce or otherwise treat as binding statutory rules that are inconsistent with the Constitution. Rules meeting that description are invalid, full stop. (65) To the extent that American constitutional practice recognizes executive review, it aligns with the principle of preexisting nullity on constitutional grounds and not judicial nullification.
That practice recognizes some executive review, and not just executive acceptance of judicial nullification. For example, the standard position of the federal executive branch is that it may decline to implement acts of Congress that unconstitutionally trench on the prerogatives of the executive. (66) Executive review, at least in limited circumstances, is thus part of American constitutional practice. The only rationale I have encountered for it is that under the logic of Marbury executive officials recognize invalidity; to my knowledge, no one argues that they make rules invalid. It is conceivable that executive officials find but courts produce invalidity, but they are alike in that neither has the legislative power. That suggests that their positions are the same with respect to the law.
To be sure, executive review is subject to fierce criticism. When Presidents of the United States say that they will not carry out a statutory rule because it is unconstitutional, controversy almost always results. The executive must follow the law, goes the criticism. Only the courts may declare statutes unconstitutional or invalidate them, goes the natural follow-on that responds to invocation of invalidity. (67) From the critics, one might conclude that actual exercises of executive review are not normative practice, but lawless deviations from constitutional principle.
A unique judicial power to nullity statutory rules on constitutional grounds implies that executive review is improper, but the impropriety of executive review would not imply a unique judicial power to nullity statutory rules on constitutional grounds. The conclusion that in the absence of judicial invalidation the executive is bound by statutory rules without regard to their constitutionality could follow from a different premise. For example, judicial review requires the assumption that the judiciary must resolve constitutional questions for itself, and is not bound by the legislature's conclusion that its statutes comport with the Constitution. That assumption might be correct for the judiciary but not for the executive.
Independent judgment on legal questions, at least most legal questions, is certainly a plausible attribute for a judiciary. If the courts are not to decide legal questions for themselves, it is hard to see what they are for. (68) Similar reasoning does not apply to the executive. Executive officials might be bound by legislative judgments on legal questions while still performing their central function of implementation. An institution that builds roads and distributes benefits and institutes law suits does not need independent legal judgment as much as does an institution the point of which is conclusively to resolve legal disputes. Someone who endorses judicial review while rejecting executive review thus might reasonably believe that the asymmetry results from the differences between the two operational branches but not from a judicial power actually to annul previously valid rules. Condemnations of executive review, even if they are sound, thus do not imply that courts make rules invalid.
Along the various dimensions on which the difference between judicial recognition and judicial creation of invalidity would manifest itself, practice mainly but not exclusively aligns with the Marbury account, according to which the Constitution produces invalidity and the courts recognize that invalidity in deciding particular cases. Judicial invalidation on constitutional grounds is not an accurate description of American constitutional practice. It describes that practice roughly at best, and so is no more than a metaphor. The same is true of judicial law-making. Metaphors can be useful, but they can also be fatally misleading. Professor Hart insisted that we look beyond them, and showed us how.
(1) Grant Gilmore, The Death of Contract (1974) 57.
(2) Henry M Hart, Jr. and Albert M Sacks, The Legal Process: Basic Problems in the Making and Application of Law (1994). Although The Legal Process was first published in hard cover in 1994, that version is based on the Tentative Edition of 1958: at 1. The 1958 version existed only in mimeographed form, at vii-vii (Preface by former Dean of the Harvard Law School Erwin M Griswold).
(3) Henry M Hart, Jr., 'The Relations Between State and Federal Law' (1954) 54 Columbia Law Review 489. I will often refer to the article as 'Relations'. The other is Henry M Hart, Jr., 'The Power of Congress to Limit the Jurisdiction of Federal Courts: An Exercise in Dialectic' (1953) 66 Harvard Law Review 1362.
(4) Johnson v Fankell (1997) 520 U.S. 911, 919 (quoting Howlett v Rose (1990) 496 U.S. 356, 372 (quoting Hart)).
(5) 'Congress rarely enacts a complete and self-sufficient body of federal law.' Hart, above n 3, 498.
(6) Claflin v Houseman (1876) 93 U.S. 130, 137.
(7) 'Legal or equitable rights, acquired under either system of laws [state or federal], may be enforced in any court of either sovereignty competent to hear and determine such kind of rights and not restrained by its constitution in the exercise of such jurisdiction.' Ibid 136.
(8) (1938) 304 U.S. 64.
(9) (1946) 326 U.S. 99.
(10) Ibid 108-109.
(11) Hart, above n 3, 512.
(12) Ibid 512-513. Hart referred to citizens of different states because most of the cases in which federal courts apply state law arise in their jurisdiction based on diversity of state citizenship of the parties.
(13) Ibid 513.
(14) (1941) 313 U.S. 481.
(15) Ibid 496.
(16) Hart, above n 3, 514 (footnote omitted).
(17) Hart and Sacks, above n 2, 119.
(18) Ibid 118. Hart and Sacks go into considerable depth concerning the Hohfeldian analytical categories in their principal application, which is to primary conduct. Ibid 127-137.
(19) Hart and Sacks criticized Hohfeld and Corbin for suggesting 'that the right and the right of action necessarily went together so that if there was no right of action there was no primary right.... Lots of people have tried to think backwards in this way. It is the essence of clear analysis to see that it is backwards, and instead to think frontwards.' Ibid 136. They elaborated: 'The purposes of law are not expressed in the sanctions which it is feasible or expedient to impose.' (at 136).
(20) (1997) 521 U.S. 898.
(21) Ibid 928-929.
(22) Ibid. The Supremacy Clause says that the Constitution, federal treaties, and laws of the United States are the supreme law of the land 'and the judges in every state shall be bound thereby, any thing in the constitution or laws of any state to the contrary notwithstanding.' United States Constitution Article VI.
(23) 'Since 1789, rights derived from federal law could be enforced in state courts unless Congress confined their enforcement to the federal courts. This has been so precisely for the same reason that rights created by the British Parliament or by the Legislature of Vermont could be enforced in the New York courts. Neither Congress nor the British Parliament nor the Vermont Legislature has power to confer jurisdiction upon the New York courts. But the jurisdiction conferred upon them by the only authority that has power to create them and to confer jurisdiction upon them--namely the law-making power of the State of New York--enables them to enforce rights no matter what the legislative source of the right may be.' Brown v Gerdes (1944) 321 U.S. 178, 188 (Frankfurter J concurring).
(24) Hart and Sacks contrasted the authority of statutes and judicial decisions. 'The statute owes its authority to the settling power of the legislature, which gave that particular set of words the force of law. The decisional doctrine ... owes such authority as it may have in later cases to the doctrine of stare decisis.' Hart and Sacks, above n 2, 126.
(25) Ibid 119.
(26) James B Beam Distilling Company v Georgia (1991) 501 U.S. 529, 549 (Scalia J concurring in the judgment). Emphasis in the original.
(27) By calling a precedent authoritative or binding, I mean that a subsequent court will regard it as more than persuasive authority.
(28) See Evan H Caminker, 'Why Must Inferior Courts Obey Superior Court Precedent?' (1994) 46 Stanford Law Review 817, 823-825 (describing rules of precedent in American federal court system). My argument here builds on and complements Dean Caminker's foundational work on this topic. He concludes that although today's standard practice is justified, it does not flow as a matter of formal law from the Constitution or federal statute, but instead rests on practical considerations (at 821-822). That conclusion, which I share, supports my conclusion regarding the relations between state and federal law: as to the state courts, the rules of precedent, including vertical precedent, are matters of state law.
(29) With apologies to Commonwealth readers, I will use American terminology, according to which a court's decree is its judgment and the explanation of that decree is its opinion. The explanations of their own views given by judges who do not speak for the court as a whole are also called opinions, not judgments. It is thus familiar practice for American courts to issue a judgment but no opinion for the court, either because no explanation is prepared or because a majority of the judges are not able to agree on one.
(30) 'Moreover, even though the decisional doctrine is respected in later cases, its ratio decidenci is not imprisoned in any single set of words; and this gives it a flexibility that a statute does not have.' Hart and Sacks, above n 2, 126. The other Hart agreed. In the use of precedent in English law, 'there is no authoritative or uniquely correct formulation of any rule to be extracted from cases.' H L A Hart, The Concept of Law (2nd edition, 1994) 134.
(31) Neil Devins and David Klein, 'The Vanishing Common Law Judge?' (forthcoming, 2016) 165 University of Pennsylvania Law Review.
(32) 'The common law, like the system of real numbers, is a conceptual system--not a textual one. The concepts of negligence, of consideration, of reliance, are not tied to a particular verbal formulation, but can be restated in whatever words seem clearest in light of current linguistic conventions. Common law is thus unwritten in a profound sense.' Richard A Posner, 'Legal Formalism, Legal Realism, and the Interpretation of Statutes and the Constitution' (1986-1987) 37 Case Western Reserve Law Review 180, 186. If the common law is truly found in the practice of the courts, then even concepts like negligence and reliance do not necessarily capture it properly. The important point here is that the common law is unwritten.
(33) In saying that, I do not mean to slip into vulgar realism and assert that the law simply is what the courts do. Judicial behaviour can deviate from the law, and for that reason it is possible to criticize courts for being lawless. H L A Hart, above n 30, 141-147. 'We are not final because we are infallible, but we are infallible only because we are final.' Brown v Allen (1953) 344 U.S. 443, 540 (Jackson J concurring).
(34) No State's constitution provides that the decisions of the State's courts constitute the law of the State. Caleb H Nelson, 'A Critical Guide to Erie Railroad v Tompkins' (2013) 54 William and Mary Law Review 921, 980-981.
(35) See, eg, Suzanna Sherry, 'Wrong, Out of Step, and Pernicious: Erie as the Worst Decision of All Time' (2011) 39 Pepperdine Law Review 129.
(36) United States Constitution, Article VI
(37) United States Constitution, Article III, section I.
(38) The qualifier reflects the fact that state highest courts are properly described as supreme even though they are subject to the appellate jurisdiction of the Supreme Court of the United States with respect to some of their decisions. 'Supreme' thus can be short for 'highest within this system of judicial institutions'.
(39) United States Constitution, Article III, sections 1-2.
(40) Today, the Court has appellate jurisdiction over the state courts in a wide range of cases involving federal questions. 28 USC [section] 1257.
(41) See Evan H Caminker, above n 28, 824-825.
(42) Act of 3 March 1817, ch lxiii, 3 Stat 376.
(43) For example, Professor David Currie noted the difficulty posed in extracting a holding from the seriatim opinions in the early and still important case of Calder v Bull (1798) 3 U.S. 386, which raises the question whether the Ex Post Facto Clause of Article I, Section 10 applies only to criminal legislation. David P Currie, The Constitution in the Supreme Court: The First Hundred Years, 1789-1888 (1985) 44.
(44) 'The judicial power shall extend to ... controversies ... between citizens of different states.' United States Constitution, Article III. section 2.
(45) See, eg, Erie Railroad v Tompkins (1938) 304 U.S. 64 (non-federal tort claim and no applicable federal law).
(46) The first Judiciary Act confined the Supreme Court's appellate jurisdiction over state courts to federal questions. Act of 24 September 1789, ch xx, s 25, 1 Stat 73, 85-87. So does title 28 of the United States Code today. 28 USC [section] 1257.
(47) Federal diversity jurisdiction is not exclusive.
(48) Henry M Hart, Jr., above n 3, 503-504.
(50) (1874) 87 U.S. 590.
(51) Act of 24 September 1789, ch xx, s 25, 1 Stat 73, 85-87.
(52) (1874) 87 U.S. 625-627.
(53) Dean Caminker similarly concludes that the hierarchical structure of the judiciary does not require strong vertical stare decisis, but that the current practice must be and can be justified on normative grounds. Evan H Caminker, above n 28, 822.
(54) The Constitutional Court of Austria, designed by Hans Kelsen, operates that way. 'If a provision is found to be unconstitutional, the court has to repeal it [and the repeal] enters into force upon expiry of the day of publication.' Christoph Bezemek, 'A Kelsenian Model of Constitutional Adjudication' (2012) 67 Zeitschrift fur offentliches Recht 115, 127. By default the repeal is prospective only, except as to the case that gave rise to it. 'Unless the Constitutional Court states otherwise, a repeal has no ex tunc effect; the case because of which the proceedings were initiated being an exception.' At 127 (citations and footnote omitted).
(55) M Farrand (ed) 2 The Records of the Federal Convention of 1787 (revised edition, 1937) 71, 73 (proposal by delegate, later Supreme Court Justice, James Wilson).
(56) Ibid 74-75 (delegates Elbridge Gerry and Caleb Strong).
(57) United States Constitution, Article I, section 7 (federal statutes adopted by both houses of Congress and the President, who has a qualified veto).
(58) For example, in INS v Chadha (1983) 462 U.S. 919, the Supreme Court held that a legislative veto provision in the Immigration and Nationality Act, which authorized one House of Congress to overturn an immigration decision by the Attorney General, was unconstitutional. In dissent, Justice White said, 'Today the Court not only invalidates [section] 244(c)(2) of the Immigration and Nationality Act, but also sounds the death knell for nearly 200 other statutory provisions in which Congress has reserved a "legislative veto.'" At 967.
(59) (1803) 5 U.S. 137.
(60) This is especially clear when a court finds that a statutory rule imposing criminal punishment is inconsistent with a substantive provision of the Constitution like the First Amendment. Individuals who violate such a rule before any court has found the rule unconstitutional are not punished, so the court decides as if the statutory rule was void from the beginning. See, eg, United States v Eichman, (1990) 496 U.S. 310 (affirming dismissal of criminal indictment on grounds that the statute at issue, which had been adopted shortly before the defendants' violation of it and before any courts' decision on its constitutionality, was unconstitutional).
(61) See, eg, Pearson v Callahan (2009) 555 U.S. 223 (governing order in which issues are to be decided in a case involving a claim of official immunity).
(62) Hart and Sacks dealt at length with problems of change in judicial understandings, in particular the question whether courts should change their common-law doctrines retrospectively or prospectively only. Hart and Sacks, above n 2, 576-596.
(63) (2015) 133 S. Ct. 2584.
(64) (2003) 798 N.E. 2 941 (Mass).
(65) 'If we accept Alexander Hamilton's justification for judicial review--that unconstitutional statutes are simply not "law" and do not bind anyone--then there is no distinction between judge and President.' Frank H Easterbrook, 'Presidential Review' (1989-1990) 40 Case Western Reserve Law Review 905, 92) (footnote omitted).
(66) 'The President has enhanced responsibility to resist unconstitutional provisions that encroach upon the constitutional powers of the Presidency. Where the President believes that an enactment unconstitutionally limits his powers, he has the authority to defend his office and decline to abide by it, unless he is convinced that the Court would disagree with his assessment.' 'Presidential Authority to Decline to Execute Unconstitutional Statutes' (1994) 18 Opinions of the Office of Legal Counsel 199, 201.
(67) Frank H Easterbrook, above n 65, 906 (discussing hostile reactions to presidential refusals to enforce). More recent work by another proponent of executive review, Professor Saikrishna Prakash, brings the account of objections more up to date. Saikrishna Bangalore Prakash, 'The Executive's Duty to Disregard Unconstitutional Laws' (2008) 96 Georgetown Law Journal 1613, 1619-1620.
(68) In Marshall's famous phrase, 'It is emphatically the province and duty of the Judicial Department to say what the law is.' Marbury v Madison (1803) 5 U.S. 137, 177. Neither that statement nor the reasoning of Marbury more generally imply that only the judiciary has that duty, or that the judiciary's statements about what the law is bind other government actors outside the decision of any particular case.
John Harrison, James Madison Distinguished Professor of Law, University of Virginia.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Henry M. Hart Jr., Columbia Law Review, vol. 54, p. 489; Special Issue: My Favourite Law Review Article|
|Publication:||University of Queensland Law Journal|
|Date:||Jun 1, 2016|
|Previous Article:||Ronald Coase's "The Problem of Social Cost."|
|Next Article:||A.W.B. Simpson's "The Horwitz Thesis and the History of Contracts."|