PHIL 168: Philosophy of Law UCSD; Fall 2015 Prof. David O. Brink Handout #2: Hart's Model of Rules and Legal Realism

Similar documents
LEGAL POSITIVISM AND NATURAL LAW RECONSIDERED

Dworkin, selections from Taking Rights Seriously. Dworkin identifies these three propositions as forming the core of the legal positivist position:

Originalism and Constructive Interpretation

Philosophy 34 Spring Philosophy of Law. What is law?

The Identity of Legal Systems

Proceduralism and Epistemic Value of Democracy

In 1978, Congress established the Foreign Intelligence Surveillance Court, which reviews warrants related to national security investigations.

The Determinacy of Republican Policy: A Reply to McMahon

Waltz s book belongs to an important style of theorizing, in which far-reaching. conclusions about a domain in this case, the domain of international

Positivism and the Separation of Realists from Their Skepticism: Normative Guidance, the Rule of Law, and Legal Reasoning

T1 INTRODUCTION... 7 WHAT IS IT?... 7 TYPES... 7 THE RULE OF LAW...

TOPIC: - THE PLACE OF KELSONS PURE THEORY OF LAW IN

INTRODUCTION TO READING & BRIEFING CASES AND OUTLINING

Ducking Dred Scott: A Response to Alexander and Schauer.

Meeting Plato s challenge?

The Rights and Wrongs of Taking Rights Seriously

Law and Politics Book Review Sponsored by the Law and Courts Section of the American Political Science Association.

NASH EQUILIBRIUM AS A MEAN FOR DETERMINATION OF RULES OF LAW (FOR SOVEREIGN ACTORS) Taron Simonyan 1

Rawls versus the Anarchist: Justice and Legitimacy

LAWS2249 Legal Theory 2 nd Semester 2009

REALIST LAWYERS AND REALISTIC LEGALISTS: A BRIEF REBUTTAL TO JUDGE POSNER

Phil 115, June 20, 2007 Justice as fairness as a political conception: the fact of reasonable pluralism and recasting the ideas of Theory

Judicial Legislation, by Fred V. Cahill

Ekaterina Bogdanov January 18, 2012

Common law reasoning and institutions

Phil 116, April 5, 7, and 9 Nozick, Anarchy, State, and Utopia

ANARCHISM: What it is, and what it ain t...

Notes toward a Theory of Customary International Law The Challenge of Non-State Actors: Standards and Norms in International Law

PART 3. Being CHAPTER 11 BEING REALISTIC 375 CHAPTER 12 BEING COMMITTED TO JUSTICE 405 CHAPTER 13 BEING ETHICAL 435

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

POLITICAL CULTURE CITIZENS ORIENTATIONS TOWARD THE POLITICAL SYSTEM, THE POLITICAL AND POLICYMAKING PROCESS AND THE POLICY OUTPUTS AND OUTCOMES

CHAPTER 1 PROLOGUE: VALUES AND PERSPECTIVES

IN BRIEF LEGAL PHILOSOPHY. Ontario Justice Education Network

Though several factors contributed to the eventual conclusion of the

Elliston and Martin: Whistleblowing

THE IDEA OF A CONSTITUTION: A PLEA FOR STAATSRECHTSLEHRE. David Dyzenhaus

POSITIVISM WITH REFERENCE AMERICAN REALISM

John Rawls THEORY OF JUSTICE

Political Society By John Locke From From Second Treatise Of Government, Chapter

Socio-Legal Course Descriptions

Justification and Judicial Repsonsibility

1100 Ethics July 2016

Book Review: The Hart-Fuller Debate in the Twenty-First Century, by Peter Cane (ed)

SAMPLE HIGHER ORDER QUESTIONS STUDENT SCALE QUESTIONS TEST ITEM SPECIFICATION NOTES. How did the benchmark help me. better understand?

RESPONSE TO JAMES GORDLEY'S "GOOD FAITH IN CONTRACT LAW: The Problem of Profit Maximization"

Disagreement, Error and Two Senses of Incompatibility The Relational Function of Discursive Updating

Activity Three: The Enlightenment ACTIVITY CARD

Phil 115, May 24, 2007 The threat of utilitarianism

POLITICAL AUTHORITY AND PERFECTIONISM: A RESPONSE TO QUONG

Phil 115, June 13, 2007 The argument from the original position: set-up and intuitive presentation and the two principles over average utility

Successfully Defending Patents In Inter Partes Reexamination And Inter Partes Review Proceedings Before the USPTO. Matthew A. Smith 1 Sept.

Holmes and Hand. By Patrick Ward. Member of the Class of 2014 at Elon University School of Law

ANARCHY AND POWER What Causes War? Ch. 10. The International System notes by Denis Bašić

LL&V plot summary: weeks one and two

Definition: Institution public system of rules which defines offices and positions with their rights and duties, powers and immunities p.

Jus in Bello through the Lens of Individual Moral Responsibility: McMahan on Killing in War

Transitional Justice, Retributive Justice and Accountability for Wrongdoing Abstract

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

Introduction[1] The obstacle

Chapter Two: Normative Theories of Ethics

AN EGALITARIAN THEORY OF JUSTICE 1

IS LAW DETERMINED BY MORALITY? Dworkin and Inclusive Legal Positivism

E-LOGOS. Rawls two principles of justice: their adoption by rational self-interested individuals. University of Economics Prague

Hart s View Criminal law should only act on bare minimum and it should not extend into the private realm

Politics between Philosophy and Democracy

Foreword to Reviews (Books on the Law of Contracts)

The legal philosophy of Ronald Dworkin.

Advisory Committee on Enforcement

The Possible Incommensurability of Utilities and the Learning of Goals

Is Rawls s Difference Principle Preferable to Luck Egalitarianism?

S.L. Hurley, Justice, Luck and Knowledge, (Cambridge, MA: Harvard University Press, 2003), 341 pages. ISBN: (hbk.).

The Politics of reconciliation in multicultural societies 1, Will Kymlicka and Bashir Bashir

Last time we discussed a stylized version of the realist view of global society.

Alberta (Attorney General) v. Krushell, 2003 ABQB 252 Date: Action No

Good Morning Finance 270. Finance 270 Summer The Legal & Regulatory Environment of Business

Review of Michael E. Bratman s Shared Agency: A Planning Theory of Acting Together (Oxford University Press 2014) 1

Legal Reasoning, the Rule of Law, and Legal Theory: Comments on Gerald Postema, Positivism and the Separation of the Realists from their Skepticism

Comments: Individual Versus Collective Responsibility

SUPREME COURT OF THE UNITED STATES

Chapter 1: Theoretical Approaches to Global Politics

CHAPTER 1 PROLOGUE: VALUES AND PERSPECTIVES

Choosing Among Signalling Equilibria in Lobbying Games

Self-Judging Self-Defense

Political Obligation 2

Volume 60, Issue 1 Page 241. Stanford. Cass R. Sunstein

Economic philosophy of Amartya Sen Social choice as public reasoning and the capability approach. Reiko Gotoh

JURISPRUDENCE: PHILOSOPHY ABOUT STUDY OF LAW

DISSENTING OPINIONS. Yale Law Journal. Volume 14 Issue 4 Yale Law Journal. Article 1

Fill in the matrix below, giving information for each of the four Enlightenment philosophers profiled in this activity.

The limits of background justice. Thomas Porter. Rawls says that the primary subject of justice is what he calls the basic structure of

The Morality of Conflict

21 Recommendations. For Uniformed Police In 21 st Century

Phil 290, February 8, 2011 Christiano, The Constitution of Equality, Ch. 2 3

7) For a case to be heard in the Supreme Court, a minimum of how many judges must vote to hear the case? A) none B) one C) nine D) five E) four

Reading Finnis Natural Law Theory in the Shadow of Hart

PROBLEMS OF CREDIBLE STRATEGIC CONDITIONALITY IN DETERRENCE by Roger B. Myerson July 26, 2018

IN THE COURT OF APPEAL IN THE MATTER OF THE LEGAL PROFESSION ACT CHAP 90:03 AND

Spain Espagne Spanien. Report Q192. in the name of the Spanish Group. Acquiescence (tolerance) to infringement of Intellectual Property Rights

ACCESS TO JUSTICE FOR CHILDREN: CROATIA

-- The search text of this PDF is generated from uncorrected OCR text.

Transcription:

Draft of 9-23- 15 PHIL 168: Philosophy of Law UCSD; Fall 2015 Prof. David O. Brink Handout #2: Hart's Model of Rules and Legal Realism Hart develops his own conception of the nature of law in the wake of his critique of Austin s command theory. Hart sees law as a body of rules and treats the union of primary and secondary rules as the key to the science of jurisprudence. LAW AS THE UNION OF PRIMARY AND SECONDARY RULES A system of primary rules, such as a system of positive morality, suffers three sorts of problems. There can be uncertainty about which primary rules are part of the system. Disputes can arise about the interpretation of the primary rules. The primary rules are comparatively static. Hart claims that these problems can be solved only by introducing secondary rules (or meta- rules) into the system - - that is, rules about the rules. He recognizes three kinds of secondary rules. A rule of recognition resolves uncertainty about which rules are primary rules by specifying criteria for membership among the primary rules. Rules of adjudication address concerns about the interpretation and application of the primary rules by establishing authoritative tribunals to adjudicate disputes and interpret and apply the primary rules. Rules of change ensure that the system of primary rules is dynamic by specifying procedures for modifying the primary rules in light of changing circumstances, beliefs, and values. Hart thinks that the addition of secondary rules to a system of primary rules marks the step from the pre- legal world into the legal world (91). Primary rules are valid law iff they have the right pedigree, that is, they satisfy the criteria in the rule of recognition. Secondary rules are valid law by regulating the conduct of officials in the system in the right way (112-13). It s possible and might be useful to write down the system s secondary rules in a document, such as a written constitution. But this is not necessary. What s important is that there be operative in the system a rule of recognition that officials adopt. Most officials must adopt an internal, and not merely external, attitude to the secondary rules of the system, viewing the rules as justified and being disposed to criticize departures from them. By contrast, it is not necessary for ordinary citizens to adopt the internal point of view, though if too many adopt a purely external point of view that might render the system unstable. Notice that there is an apparent circularity for Hart if the rule of recognition depends for its adoption by the officials in the system and if we need to specify who the officials in the system are by appeal to the rule of recognition. Hart can avoid circularity if he at least initially characterizes the officials not in terms of legal authority but rather by some sociological criteria, say, involving patterns of deference or dispositions to obey. LEGAL POSITIVISM Like Austin, Hart is a legal positivist, because he thinks that law is a social artifact and that the existence of law is a matter of social pedigree that does not guarantee adequate moral content. Primary rules are valid law iff they have the pedigree laid down by the rule of recognition. In most systems that we are familiar with, that is a pedigree involving institutional enactment. But

institutions, such as legislatures and constitutional conventions, can adopt immoral primary rules. This ensures, as Austin says, that the existence of law is one thing, its merit or demerit another. Does Hart qualify this defense of legal positivism? After all, he also defends what he calls the minimum content of natural law. This is the idea that given pervasive, but contingent, facts of human nature involving the circumstances of justice - - including scarcity of goods, limited altruism, and approximate equality of physical strength and intelligence - - there is a certain necessity that legal systems contain and enforce some modest norms of mutual non- aggression and cooperation (192-200). The general observance of these norms of cooperation and non- aggression is mutually beneficial, and law is necessary to insure this general observance. The alternative is anarchy, which is mutually disadvantageous, or as Thomas Hobbes (1588-1679) says nasty, brutish, and short. However, this concession to natural law is fully compatible with legal positivism, in particular, with immoral laws and legal systems (200). Systems that meet the minimum content of natural law can nonetheless engage in selective but significant injustice, denying these basic rights to political minorities, as we witnessed historically in Nazi Germany, South African Apartheid, and Jim Crow. HART S THEORY OF ADJUDICATION Hart s model of rules seems to support a familiar picture of adjudication (Ch. 7). At least in morally decent legal systems, it is presumably the obligation of the courts to interpret and apply the system s primary rules. Hart claims that legal rules are formulated in general terms. A municipal ordinance might forbid vehicles in the park. A constitutional provision might guarantee due process of law or prohibit unreasonable searches and seizures or cruel and unusual punishment. Hart thinks that whereas some cases - - easy cases - - clearly fall within the meaning of such general terms, in other cases - - hard cases - - it s deeply contested whether they fall within the meaning of such terms. Hart claims that the law is determinate in easy cases and that (in morally decent legal systems) courts have a duty to apply pre- existing law. But he thinks that the law is not determinate in hard cases and that judges can only decide such cases by exercising a quasi- legislative capacity that he calls judicial discretion. LEGAL REALISM AND THE MODEL OF RULES It helps to understand and appreciate Hart s theory of adjudication to see how it grows out of his critical engagement with American Legal Realism. The American Legal Realists were a group of American legal scholars and judges in the first half of the twentieth century, including Oliver Wendell Holmes (1841-1935), John Chipman Gray (1839-1915), Karl Llewellyn (1893-1962), and Jerome Frank (1889-1957), who sought to revolutionize legal scholarship and pedagogy. The legal realists were reacting to the study of legal doctrine via the case method, made influential by Christopher Langdell (1826-1906) at Harvard in the late nineteenth century. The case method tended to treat the study of law in an insular way as a codified set of principles that could be derived from the study of cases. By contrast, the realists were interested in viewing the law in larger social context - - paying attention to the social antecedents and consequences of legislation and litigation. When applied to adjudication, the realists made both descriptive claims about the social, economic, and political influences on litigation and judicial decision- making and normative claims that it was appropriate and desirable for courts to consider the social impact of their decisions. This aspect of the jurisprudence of legal realism has had an enduring legacy for legal research and pedagogy. Much of the interdisciplinary research and pedagogy in the legal academy today, including the very different and influential sub- fields of Law and Economics and Critical Legal Studies, is the product of legal realist innovations. With this interdisciplinary approach to the law Hart has no quarrel. But many of the legal realists also made general and surprising claims about the nature of law intended to challenge 2

traditional claims made by analytical jurists, such as Austin. It is to these philosophical claims about the law that we now turn. The realists criticized the model of rules, which they associated with what they called "mechanical jurisprudence" and associated with the case method. The mechanical jurist apparently thinks that all cases are easy cases and that legal reasoning takes a simple syllogistic form. 1. All Fs are forbidden (legal rule). 2. A's action x is F (statement of fact). 3. Hence, x is forbidden (judicial decision). Some legal realists are skeptical about whether there are exceptionless rules that can serve as the major premises in such syllogisms. They are rule- skeptics. Instead of identifying the law with legal rules, they think we should identify it in some way with the decisions of courts. Holmes appeals to the "bad man's point of view and identifies the law with predictions of what courts will do. Gray appeals to the final authority of judges in interpreting the law and identifies law with what courts decide. LEGAL REALISM AS VERIFICATIONISM ABOUT THE LAW Ironically, rule- skepticism represents a kind of philosophical anti- realism about the law. To see this, we need to understand realism and anti- realism as philosophical doctrines. Realism about a domain typically claims (a) that there are facts or truths of a certain kind and (b) that they are independent of our beliefs about them and our verification procedures. For instance, realism about subatomic particles says that there really are subatomic particles, independently of the role they play in making macroscopic predictions. Psychological realism implies that there are real internal mental states of agents, such as beliefs and desires, and that these aren t just useful ways of predicting people s behavior. Moral realism says that there are moral facts and truths, independently of the moral beliefs and emotions of appraisers. Anti- realists deny one of these two claims, either asserting that there are no such facts (nihilism) or insisting that the facts or truths in question can and should be reduced to our evidence (verificationism). So, for instance, operationalists claimed that temperature is not a real magnitude but just whatever would result from our measurements with thermometers. Behaviorism claimed that talk of mental states was just a way of talking about actual or probable behavior. And moral anti- realists either deny that there are moral facts or relativize them to emotional responses or beliefs of appraisers. As such, realism implies fallibility, because it implies that there are facts that are in principle independent of our evidence. By contrast, anti- realist views deny or at least restrict fallibility. For example, verificationist views eliminate the possible gap between truth and our evidence by reducing the facts or truths in question to our evidence or verification procedures. The model of rules is realistic (in this sense), because it treats the decisions of courts (and predictions about the decisions of courts) as reliable but fallible evidence about what the law is (what the rules require). By contrast, rule skepticism is anti- realistic (verificationist), because it collapses the distinction between the law and our evidence about the law and implies judicial infallibility. LEGAL VERIFICATIONISM'S EXPLANATORY PROBLEMS A realist recognizes standards that we might conform to (or fail to conform to) and that might guide our responses; the verificationist does not. For this reason, legal realism seems a poor fit from the perspective of both observers and participants, because participants seem to use these standards to guide their behavior. It's hard for observers of the legal process to square legal realism with the way legal principals act as if there are pre- existing legal rules. Judges write opinions as if they were applying constitutional, statutory, or common law rules and not as if they were legislators designing 3

desirable policy; they write as if there are legal standards that they are bound to apply. Litigants demand a decision as a matter of pre- existing legal right; they do not lobby for legislative favors. Lawyers present arguments for these legal rights that rely on the language of statutes and constitutional provisions and previous cases. If legal realism were true, we should expect legal principals to behave in a much different way. Legal realism also fails to make sense of the legal process from the point of view of the judge or other interpreters and participants. The commonsense view and the view contained in our separation of powers doctrine is that it is the function of the judiciary to interpret and apply the law. But if there are no pre- existing legal rules to interpret or apply, what is it that judges are supposed to be doing? Should a judge see herself as trying to decide what she is going to decide? If so, it's unclear how she should deliberate. It's also unclear how parties to a dispute before her should frame their arguments; there seems to be no reason to frame their arguments in terms of the best construction of existing legal materials. These explanatory difficulties reveal ways in which legal realism is committed to counter- intuitive claims. In fact, legal realism seems to be an error theory, committed to the proposition that legal observers and principals are all making a systematic mistake about the law. Why should we embrace the legal realist s error theory? THE BAD MAN'S POINT OF VIEW Holmes defends the predictive theory as a consequence of the bad man's point of view. It is true that the bad man's reason for being interested in legal rules is derivative in a way that his interest in the decisions of courts is not. However, it doesn't follow that the law includes only the decisions of courts and not the legal rules unless we assume that the law is only what the bad man cares about non- derivatively. Even if we assumed the bad man s point of view, it s not clear that we should identify the law with what he cares about non- derivatively. The bad man is also interested in bridges not collapsing on him and collapsing on his enemies. But that doesn t mean that engineering is to be identified with predictions about bridge collapse. With both law and engineering, it seems more plausible to identify law and engineering with bodies of rules that the bad man might value instrumentally. Also, the bad man s point of view is only one perspective. Why privilege it? Why not take the perspective of the law- abiding citizen, who wants to know what the legal rules are so that she may conform her behavior to the law? This perspective would seem to support the model of rules, not rule- skepticism. THE ARGUMENT FROM LEGAL POSITIVISM Holmes and Gray also defend legal realism as a consequence of avoiding what they see as confusing law and morality. Their argument has something like this structure. 1. The alternative to legal realism is to recognize unenforced rights. 2. Unenforced rights imply natural law. 3. Hence, either legal realism or natural law is true. 4. Natural law theory is false (legal positivism is true). 5. Hence, legal realism is true. Is this argument sound? It certainly seems valid. If so, its soundness depends on the truth of its premises. Of course, one might dispute (4) and defend natural law. But if we treat legal positivism and natural law as mutually exclusive and jointly exhaustive, this would concede that the legal positivist must be a legal realist. That result would be troubling to Hart and others. In fact, we should reject (2). The unenforced rights in question need not be dictates of natural law or Holmes s brooding omnipresence in the sky (whatever that might be). They can be understood as those legal claims supported by the relevant legal rules. When judges fail to apply these rules, there are unenforced legal 4

rights. Austin or Hart can explain the existence of such rights. Hence, recognizing them does not entail natural law. THE ARGUMENT FROM JUDICIAL FINALITY AND PRECEDENT Gray defends verificationist claims about the law by appeal to the fact that in legal systems such as our own, it is the courts that have the authoritative say on what the laws mean. There are two things worth distinguishing here. One issue concerns finality. Courts (ultimately the highest court) have (has) the final say in a particular case about what legal rights the parties to that case have. Another issue concerns precedent. Other things being equal, a court's ruling in a relevantly similar case has the status of law; subsequent interpretations of the law must take that decision into account with other legal materials. But finality does not imply infallibility. Hart's analogy with the finality of umpires in baseball is helpful (142-46). Because umpires have the final say on the interpretation and application of the rules of baseball, we might be tempted to say that the rules of baseball are whatever the umpire says they are. But this confuses the umpire's finality with his infallibility. Indeed, it's hard to explain how the players regulate their behavior or the umpire is to make decisions if it is not normally as a good faith effort to follow the rules of the game. Baseball is not scorer s discretion, and the law is not what courts say it is. Nor does precedential value imply infallibility. We can agree that a bad decision can add to the law while insisting that that decision was not supported by the law that existed at the time of the decision. The forward- looking claim on behalf of bad interpretation and the backward- looking claim against it are fully compatible. Indeed, if the significance of precedent is that subsequent interpretations of the law must take that decision into account with other legal materials, then precedents are among the pre- existing legal rules that any future decision must take into account. But then precedent counts against legal verificationism, not in favor of it. HARD CASES AND INDETERMINACY Legal realism is implausible as a general theory of law if only because there are legal rules that judges can apply in easy cases. But legal realism's indeterminacy thesis is more plausible when restricted to hard cases in which, for a variety of reasons, it is controversial what the law requires or whether the law applies to the case at hand. If we remember that the legal realists were legal scholars studying classic cases that posed deep issues of principle and appellate court judges focusing on cases involving difficult issues of law (rather than fact), we can see that their legal diet consisted almost exclusively of hard cases. So, even though the realists formulate their claims about the law in quite general terms, perhaps these claims were really only intended to apply to hard cases. So construed, realist claims about legal indeterminacy and judicial legislation are not so outlandish and are much more plausible. Indeed, Hart himself clearly accepts this more restricted realist claim. At least in morally decent legal systems, Hart believed that courts should interpret and apply the law by applying these primary rules. Hart thought that there were often good reasons for law- makers to enact laws that employed general terms - - such as anti- competitive practices, due process, and unreasonable search and seizure - - rather than trying to give an exhaustive specification of all the actions and activities that the law should regulate. But, Hart claimed, general terms are essentially open textured, with the result that cases could be divided into easy cases, to which the legal rules clearly apply, and hard cases, in which it was controversial whether the rule applies (119-20). Hart believed that hard cases are legally indeterminate (124, 252). Judges cannot decide such cases by applying the law but only by exercising a quasi- legislative capacity that he called judicial discretion. He makes clear that this sort of judicial legislation need not and should not be arbitrary; it should reflect characteristic judicial virtues of impartiality, neutrality, and principled decision- making (124, 200, 273). But such resolution must ex hypothesi be based on extra- legal considerations. Hart s argument for judicial discretion in hard cases has something like this form. 5

1. The law consists of legal rules formulated in general terms. 2. All general terms are open- textured: though they contain a core of settled meaning, they also have a periphery where their meaning is not determinate. 3. Controversial or hard cases, about which reasonable people with legal training disagree, fall within the open texture of legal terms within existing legal rules. 4. Hence, hard cases are legally indeterminate. 5. Hence, courts could decide hard cases only on extra- legal (e.g. moral and political) grounds. 6. Hence, in hard cases courts must exercise judicial discretion and make, rather than apply, law. Consider Hart s example of a municipal ordinance prohibiting vehicles in the park. The core meaning of the term vehicle applies to my SUV and my motorcycle. So if I am caught in the park doing doughnuts in my SUV or wheelies on my motorcycle, my case is an easy case, determinately prohibited by the legal rules. But vehicle is an open- textured concept. It is unclear whether it applies to bicycles, skateboards, Segways, and roller blades. Cases involving the use of these devices in the park would be hard cases and, according to Hart, legally indeterminate. Courts could decide such cases, he thinks, only by exercising the quasi- legislative capacity of judicial discretion. The law is gappy, but these gaps are gradually filled in over time by the exercise of judicial discretion. CONCERNS ABOUT JUDICIAL DISCRETION We might worry about the move from indeterminacy to judicial discretion in the inference from (5) to (6). Of course, if courts are to decide legally indeterminate cases, they must exercise discretion. But should they decide such cases? A democratic worry is that judicial discretion involves judge- made law and, as such, violates the separation of powers doctrine that requires law- makers to be politically accountable as, in principle, legislators are and many judges are not. Alternatively, we might worry that judicial discretion is unfair insofar as it would hold people legally liable for conduct whose legal status at the time was, by hypothesis, indeterminate. Are these worries decisive? Hart thinks not (see his postscript, 273). In response to the democratic worry, we might distinguish between a generalized discretion that encourages judges to make law wherever there is none and a more restricted discretion that encourages them to legislate but only interstitially, at the margins of existing legal rules. Whereas a generalized discretion would clearly offend the separation of governmental powers or functions, a restricted judicial discretion, of the sort Hart contemplates, would do so much less. In response to the fairness worry, it can be argued that retrospective legislation is unfair only where people's expectations about what they are permitted to do were reasonable. But in hard cases, where, ex hypothesi, the law is unclear, it is not clear that parties could have reasonable expectations that retrospective legislation might upset. 6