公 法 评 论

et revelabitur quasi aqua iudicium et iustitia quasi torrens fortis



Ronald Dworkin, "Natural Law Revisited"

34 U. FLA. L. REV. 165 (1982)

[Editor's Note: Ronald Dworkin is Professor of Jurisprudence at Oxford as well as a member of the faculty of New York University School of Law. Dworkin's writings have pursued a controversial solution to the problems that have been exposed in the foregoing materials.]

1. What is Naturalism?

Everyone likes categories, and legal philosophers like them very much. So we spend a good deal of time, not all of it profitably, labeling ourselves and the theories of law we defend. One label, however, is particularly dreaded: no one wants to be called a natural lawyer. Natural law insists that what the law is depends in some way on what the law should be. This seems metaphysical or at least vaguely religious. In any case it seems plainly wrong. If some theory of law is shown to be a natural law theory, therefore, people can be excused if they do not attend to it much further.
In the past several years, I have tried to defend a theory about how judges should decide cases that some critics (though not all) say is a natural law theory and should be rejected for that reason. I have of course made the pious and familiar objection to this charge, that it is better to look at theories than labels. But since labels are so much a part of our common intellectual life it is almost as silly to flee as to hurl them. If the crude description of natural law I just gave is correct, that any theory which makes the content of law sometimes depend on the correct answer to some moral question is a natural law theory, then I am guilty of natural law. I am not now interested, I should add, in whether this crude characterization is historically correct, or whether it succeeds in distinguishing natural law from positivist theories of law. My present [164]concern is rather this. Suppose this is natural law. What in the world is wrong with it?

A. Naturalism

I shall start by giving the picture of adjudication I want to defend a name, and it is a name which accepts the crude characterization. I shall call this picture naturalism. According to naturalism, judges should decide hard cases by interpreting the political structure of their community in the following, perhaps special way: by trying to find the best justification they can find, in principles of political morality, for the structure as a whole, from the most profound constitutional rules and arrangements to the details of, for example, the private law of tort or contract. Suppose the question arises for the first time, for example, whether and in what circumstances careless drivers are liable, not only for physical injuries to those whom they run down, but also for any emotional damage suffered by relatives of the victim who are watching. According to naturalism, judges should then ask the following questions of the history (including the contemporary history) of their political structure. Does the best possible justification of that history suppose a principle according to which people who are injured emotionally in this way have a right to recover damages in court? If so, what, more precisely, is that principle? Does it entail, for example, that only immediate relatives of the person physically injured have that right? Or only relatives on the scene of the accident, who might themselves have suffered physical damage?

Of course a judge who is faced with these questions in an actual case cannot undertake anything like a full justification of all parts of the constitutional arrangement, statutory system and judicial precedents that make up his "law." I had to invent a mythical judge, called Hercules, with superhuman powers in order even to contemplate what a full justification of the entire system would be like.(1) Real judges can attempt only what we might call a partial justification of the law. They can try to justify, under some set of principles, those parts of the legal background which seem to them immediately relevant, like, for example, the prior judicial decisions about recovery for various sorts of damage in automobile accidents. Nevertheless it is useful to describe this as a partial justification璦s a part of what Hercules himself would do璱n order to emphasize that, according to this picture, a judge should regard the law he mines and studies as embedded in a much larger system, so that it is always relevant for him to expand his investigation by asking whether the conclusions he reaches are consistent with what he would have discovered had his study been wider.

It is obvious why this theory of adjudication invites the charge of natural law. It makes each judge's decision about the burden of past law depend on his judgment about the best political justification of that law, and this is of course a matter of political morality. Before I consider whether this provides a fatal defect in the theory, however, I must try to show how the theory might work in practice. It may help to look beyond law to other enterprises in which participants extend a discipline into the future by re-examining its past. This process is in fact charac-[165]teristic of the general activity we call interpretation, which has a large place in literary criticism, history, philosophy and many other activities. Indeed, the picture of adjudication I have just sketched draws on a sense of what interpretation is like in these various activities, and I shall try to explicate the picture through an analogy to literary interpretation.(2) I shall, however, pursue that analogy in a special context designed to minimize some of the evident differences between law and literature, and so make the comparison more illuminating.

B. The Chain Novel

Imagine, then, that a group of novelists is engaged for a particular project. They draw lots to determine the order of play. The lowest number writes the opening chapter of a novel, which he then sends to the next number who is given the following assignment. He must add a chapter to that novel, which he must write so as to make the novel being constructed the best novel in can be. When he completes his chapter, he then sends the two chapters to the next novelist, who has the same assignment, and so forth. Now every novelist but the first has the responsibility of interpreting what has gone before in the sense of interpretation I described for a naturalist judge. Each novelist must decide what the characters are "really" like; what motives in fact guide them; what the point or theme of the developing novel is; how far some literary device or figure consciously or unconsciously used can be said to contribute to these, and therefore should be extended, refined, trimmed or dropped. He must decide all this in order to send the novel further in one direction rather than another. But all these decisions must be made, in accordance with the directions given, by asking which decisions make the continuing novel better as a novel.

Some novels have in fact been written in this way (including the soft-core pornographic novel NAKED CAME THE STRANGER) though for a debunking purpose, and certain parlor games, for rainy weekends in English country houses, have something of the same structure. But in this case the novelists are expected to take their responsibilities seriously, and to recognize the duty to create, as far as they can, a single unified novel rather than, for example, a series of independent short stories with characters bearing the same names. Perhaps this is an impossible assignment; perhaps the project is doomed to produce, not simply an impossibly bad novel, but no novel at all, because the best theory of art requires a single creator, or if more than one, that each have some control over the whole. (But what about legends and jokes? What about the Old Testament, or, on some theories, the ILLIAD?) I need not push that question further, because I am interested only in the fact that the assignment makes sense, that each of the novelists in the chain can have some sense of what he or she is asked to do, whatever misgivings each might have about the value or character of what will then be produced.

The crucial question each must face is this. What is the difference between continuing the novel in the best possible way, by writing plot and development that can be seen to flow from what has gone before, and starting a fresh novel with characters having the same names? Suppose you are a novelist well down [166]the chain, and are handed several chapters which are, in fact, the first sections of the Dickens short novel, A CHRISTMAS CAROL. You consider these two interpretations of the central character: that Scrooge is irredeemably, inherently evil, and so an example of the degradation of which human nature is intrinsically capable, or that Scrooge is inherently good, but progressively corrupted by the false values and perverse demands of high capitalist society. The interpretation you adopt will obviously make an enormous difference in the way you continue the story. You aim, in accordance with your instructions, to make the continuing novel the best novel it can be; but you must nevertheless choose an interpretation that makes the novel a single work of art. So you will have to respect the text you have been given, and not choose an interpretation that you believe the text rules out. The picture that text gives of Scrooge's early life, for example, might be incompatible with the claim that he is inherently wicked. In that case you have no choice. If, on the other hand, the text is equally consistent with both interpretations, then you do have a choice. You will choose the interpretation that you believe makes the work more significant or otherwise better, and this will probably (though not inevitably) depend on whether you think people like Scrooge are in fact, in the real world, born bad or corrupted by capitalism.

Now consider a more complex case. Suppose the text does not absolutely rule out either interpretation, but is marginally less consistent with one, which is, however, the interpretation you would pick if they both fit equally well. Suppose you believe that the original sin interpretation (as we might call it) is much the more accurate depiction of human nature. But if you choose that interpretation you will have to regard certain incidents and attributions established in the text you were given as "mistakes." You must then ask yourself which interpretation makes the work of art better on the whole, recognizing, as you will, that a novel whose plot is inconsistent or otherwise lacks integrity is thereby flawed. You must ask whether the novel is still better as a novel, read as a study of original sin, even though it must now be regarded as containing some "mistakes" in plot, than it would be with fewer "mistakes" but a less revealing picture of human nature. You may never have reflected on that question before, but that is no reason why you may not do so now, and once you make up your mind you will believe that the correct interpretation of Scrooge's character is the interpretation that makes the novel better on the whole.

C. The Chain of Law

Naturalism is a theory of adjudication not of the interpretation of novels. But naturalism supposes that common law adjudication is a chain enterprise sharing many of the features of the story we invented. According to naturalism, a judge should decide fresh cases in the spirit of a novelist in the chain writing a fresh chapter. The judge must make creative decisions, but must try to make these decisions "going on as before" rather than by starting in a new direction as if writing on a clean slate. He must read through (or have some good idea through his legal training and experience) what other judges in the past have written, not simply to discover what these other judges have said, or their state of mind when they said it, but to reach an opinion about what they have collectively done, in the way that each of our novelists formed an opinion about the collective novel so far written. Of course, the best interpretation of past judicial decisions is the interpretation that shows these in the best light, not aesthetically but politically, as coming as close to the correct ideals of a just legal system as possible. Judges [167]in the chain of law share with the chain novelists the imperative of interpretation, but they bring different standards of success璸olitical rather than aesthetic璽o bear on that enterprise.
The analogy shows, I hope, how far naturalism allows a judge's beliefs about the personal and political rights people have "naturally"璽hat is, apart from the law璽o enter his judgments about what the law requires. It does not instruct him to regard these beliefs as the only test of law. A judge's background and moral convictions will influence his decisions about what legal rights people have under the law. But the brute facts of legal history will nevertheless limit the role these convictions can play in those decisions. The same distinction we found in literary interpretation, between interpretation and ideal, holds here as well. An Agatha Christie mystery thriller cannot be interpreted as a philosophical novel about the meaning of death even by someone who believes that a successful philosophical novel would be a greater literary achievement than a successful mystery. It cannot be interpreted that way because, if it is, too much of the book must be seen as accidental, and too little as integrated, in plot, style and trope, with its alleged genre or point. Interpreted that way it becomes a shambles and so a failure rather than a success at anything at all. In the same way, a judge cannot plausibly discover, in a long and unbroken string of prior judicial decisions in favor of the manufacturers of defective products, any principle establishing strong consumers' rights. For that discovery would not show the history of judicial practice in a better light; on the contrary it would show it as the history of cynicism and inconsistency, perhaps of incoherence. A naturalist judge must show the facts of history in the best light he can, and this means that he must not show that history as unprincipled chaos.

Of course this responsibility, for judges as well as novelists, may best be fulfilled by a dramatic reinterpretation that both unifies what has gone before and gives it new meaning or point. This explains why a naturalist decision, though it is in this way tied to the past, may yet seem radical. A naturalist judge might find, in some principle that has not yet been recognized in judicial argument, a brilliantly unifying account of past decisions that shows them in a better light than ever before. American legal education celebrates dozens of such events in our own history. In the most famous single common law decision in American jurisprudence, for example, Cardozo reinterpreted a variety of cases to find, in these cases, the principle on which the modern law of negligence was built.(3)

Nevertheless the constraint, that a judge must continue the past and not invent a better past, will often have the consequence that a naturalist judge cannot reach decisions that he would otherwise, given his own political theory, want to reach. A judge who, as a matter of political conviction, believes in consumers' rights may nevertheless have to concede that the law of his jurisdiction has rejected this idea. It is in one way misleading to say, however, that he will be then forced to make decisions at variance with his political convictions. The principle [168]that judges should decide consistently with principle, and that law should be coherent, is part of his convictions, and it is this principle that makes the decision he otherwise opposes necessary.

D. Interpretation in Practice

In this section I shall try to show how a self-conscious naturalist judge might construct a working approach to adjudication, and the role his background moral and political convictions would play in that working approach. When we imagined you to be a novelist in the chain novel, several pages ago, we considered how you would continue the first few chapters of A Christmas Carol. We distinguished two dimensions of a successful interpretation. An interpretation must "fit" the data it interprets, in order not to show the novel as sloppy or incoherent, and it must also show that data in its best light, as serving as well as can be some proper ambition of novels. Just now, in noticing how a naturalist judge who believed in consumers' rights might nevertheless have to abandon the claim that consumers' rights are embedded in legal history, we relied on the same distinction....A naturalist judge would be forced to reject a politically attractive interpretation, we supposed, simply because he did not believe it fit the record well enough. If fit is indeed an independent dimension of success in interpretation, then any judge's working approach would include some tacit conception of what "fit" is, and of how well a particular interpretation must fit the record of judicial and other legal decisions in order to count as acceptable.

This helps us to explain why two naturalist judges might reach different interpretations of past judicial decisions about accidents, for example. They might hold different conceptions of "fit" or "best fit," so that, for instance, one thinks that an interpretation provides an acceptable fit only if it is supported by the opinions of judges in prior cases, while the other thinks it is sufficient, to satisfy the dimension of fit, that an interpretation fit the actual decisions these judges reached even if it finds no echo in their opinions. This difference might be enough to explain, for example, why one judge could accept an "economic" interpretation of the accident cases璽hat the point of negligence law is to reduce the overall social costs of accidents瓀hile another judge, who also found that interpretation politically congenial, would feel bound by his beliefs about the requirement of fit to reject it.

At some point, however, this explanation of differences between two judges' theories of the same body of law would become strained and artificial. Suppose Judge X believes, for example, that pedestrians ought to look out for themselves, and have no business walking in areas in which drivers are known normally to exceed the legal speed limit. He might rely on this opinion in deciding that "our law recognizes no general right to recover whenever someone is injured by a speeding driver while walking on a highway where most drivers speed." If Judge Y reaches a different judgment about what the law is, because he believes that pedestrians should be entitled to assume that people will obey the law even when there is good evidence that they will not, then it would strain language to explain this difference by saying that these judges disagree about the way or the degree in which an interpretation of the law must fit past decisions. We would do better to say that these judges interpret the law differently, in this instance, because they bring different background theories of political morality to their interpretations, just as two art critics might disagree about the correct interpretation of impressionism because they bring different theories about the value of art to that exercise.

Any naturalist judge's working approach to interpretation will recognize this distinction between two "dimensions" of interpretations of the prior law, and so we might think of such a theory as falling into two parts. One part refines and develops the idea that an interpretation must fit the data it interprets. This part takes up positions on questions like the following. How many decisions (roughly) can an interpretation set aside as mistakes, and still count as an interpretation of the string of decisions that includes those "mistakes"? How far is an interpretation better if it is more consistent with later rather than earlier past decisions? How far and in what way must a good interpretation fit the opinions judges write as well as the decisions they make? How far must it take account of popular morality contemporary with the decisions it offers to interpret? A second part of any judge's tacit theory of interpretation, however, will be quite independent of these "formal" issues. It will contain the substantive ideals of political morality on which he relies in deciding whether any putative interpretation is to be preferred because it shows legal practice to be better as a matter of substantive justice. Of course, if any working approach to interpretation has these two parts, then it must also have principles that combine or adjudicate between them.

This account of the main structure of a working theory of interpretation has heuristic appeal. It provides judges, and others who interpret the law, with a model they might use in identifying the approach they have been using, and self-consciously to inspect and improve that model. A thoughtful judge might establish for himself, for example, a rough "threshold" of fit which any interpretation of data must meet in order to be "acceptable" on the dimension of fit, and then suppose that if more than one interpretation of some part of the law meets this threshold, the choice among these should be made, not through further and more precise comparisons between the two along that dimension, but by choosing the interpretation which is "substantively" better, that is, which better promotes the political ideals he thinks correct. Such a judge might say, for example, that since both the foreseeability and the area-of-physical-risk interpretations rise above the threshold of fit with the emotional damage cases I mentioned earlier, foreseeability is better as an interpretation because it better accords with the "natural" rights of people injured in accidents.

The practical advantages of adopting such a threshold of fit are plain enough. A working theory need specify that threshold in only a rough and impressionistic way. If two interpretations both satisfy the threshold, then, as I said, a judge who uses such a theory need make no further comparisons along that dimension in order to establish which of them in fact supplies the "better" fit, and he may therefore avoid many of the difficult and perhaps arbitrary decisions about better fit that a theory without this feature might require him to make. But there are nevertheless evident dangers in taking the device too seriously, as other than a rule-of-thumb practical approach. A judge might be tricked into thinking that these two dimensions of interpretations are in some way deeply competitive with one another, that they represent the influence of two different and sometimes contradictory ambitions of adjudication.
He will then worry about those inevitable cases in which it is unclear whether some substantively attractive interpretation does indeed meet the threshold of fit. He will think that in such cases he must define that threshold, not impressionistically, as calling for a "decent" fit, but precisely, perhaps everything will then turn on whether that interpretation in fact just meets or just fails the crucial test. This [170]rigid attitude toward the heuristic distinction would miss the point that any plausible theory of interpretation, in law as in literature, will call for some cross influence between the level of fit at which the threshold is fixed and the substantive issues involved. If an interpretation of some string of cases is far superior "substantively" it may be given the benefit of a less stringent test of fit for that reason.

For once again the underlying issue is simply one of comparing two pictures of the judicial past to see which offers a more attractive picture, from the standpoint of political morality, overall. The distinction between the dimensions of fit and substance is a rough distinction in service of that issue. The idea of a threshold of fit, and therefore a lexical ordering between the two dimensions, is simply a working hypothesis, valuable so far as the impressionistic characterization of fit on which it depends is adequate, but which must be abandoned in favor of a more sophisticated and piecemeal analysis when the occasion demands.

Of course the moment when more sophisticated analysis becomes necessary, because the impressionistic distinction of the working theory no longer serves, is a moment of difficulty calling for fresh political judgments that may be hard to make. Suppose a judge faces, for the first time, the possibility of overruling a narrow rule followed for some time in his jurisdiction. Suppose, for example, that the courts have consistently held, since the issue was first raised, that lawyers may not be sued in negligence. Our judge believes that this rule is wrong and unjust, and that it is inconsistent in principle with the general rule allowing actions in negligence against other professional people like doctors and accountants. Suppose he can nevertheless find some putative principle, in which others find [satisfaction] though he does not, which would justify the distinction the law has drawn. Like the principle, for example, that lawyers owe obligations to the courts or to abstract justice, [so that] it would be unfair to impose on them any legal obligation of due care to their clients. He must ask whether the best interpretation of the past includes that principle in spite of the fact that he himself would reject it.

Neither answer to this question will seem wholly attractive to him. If he holds that the law does include this putative principle, then this argument would present the law, including the past decisions about suits against lawyers, as coherent; but he would then expose what he would believe to be a flaw in the substantive law. He would be supposing that the law includes a principle he believes is wrong, and therefore has no place in a just and wise system. If he decides that the law does not include the putative principle, on the other hand, then he can properly regard this entire line of cases about actions against lawyers as mistakes, and ignore or overrule them; but he then exposes a flaw in the record of a different sort, namely that past judges have acted in an unprincipled way, and a demerit in his own decision, that it treats the lawyer who loses the present case differently from how judges have treated other lawyers in the past. He must ask which is, in the end, the greater of these flaws; which way of reading the record shows it, in the last analysis, in the better and which in the worse light.
It would be absurd to suppose that all the lawyers and judges of any common law community share some set of convictions from which a single answer to that question could be deduced. Or even that many lawyers or judges would have ready at hand some convictions of their own which could supply an answer without further ado. But it is nevertheless possible for any judge to confront issues like these in a principled way, and this is what naturalism demands of him. He [171]must accept that in deciding one way rather than another about the force of a line of precedents, for example, he is developing a working theory of legal interpretation in one rather than another direction, and this must seem to him the right direction as a matter of political principle, not simply an appealing direction for the moment because he likes the answer it recommends in the immediate case before him. Of course, there is, in this counsel, much room for deception, including self-deception. But in most cases it will be possible for judges to recognize when they have submitted some issue to the discipline this description requires and also to recognize when some other judge has not. * * *

2. Is It Delusion?

A. Internal and External Scepticism

I have been describing naturalism as a theory about how judges should decide cases. It is of course a further question whether American (or any other) judges actually do decide cases that way. I shall not pursue that further question now. Instead, I want to consider certain arguments that I expect will be made against naturalism simply as a recommendation. In fact, many of the classical objections to "natural law" theories are objections to such theories as models for, rather than descriptions of, judicial practice. I shall begin with what might be called the sceptical attack.

I put my description of naturalism in what might be called a subjective mode. I described the question which, according to naturalism, judges should put to themselves, and answer from their own convictions. Someone is bound to object that, although each judge can answer these questions for himself, different judges will give different answers, and no single answer can be said to be objectively right. "There are as many different 'best' interpretations as there are interpreters, he will say, because no one can offer any argument in favor of one interpretation over another, except that it strikes him as the best, and it will strike some other interpreter as the worst. No doubt judges (as well as many other people) would deny this. They think their opinions can have some objective standing, that they can be either true or false. But this is delusion merely."

What response can naturalism, as I have described it, make to this sceptical challenge? We must begin by asking what kind of scepticism is in play. I have in mind a distinction which, once again, might be easier to state if we return to a literary analogy. Suppose we are studying Hamlet and the question is put by some critic whether, before the play begins, Hamlet and Ophelia have been lovers. This is a question of interpretation, and two critics who disagree might present arguments trying to show why the play is, all things considered, more valuable as a work of art on one or the other understanding about Hamlet and Ophelia. But plainly a third position is possible. Someone might argue that it makes no difference to the importance or value of the play which of these assumptions is made about the lovers, because the play's importance lies in a humanistic vision of life and fate, not in any detail of plot or character whose reading would be affected by either assumption. This third position argues that the right answer to this particular question of interpretation is only that there is no right answer; that there is no "best" interpretation of the sexual relationship between Hamlet and Ophelia, only "different" interpretations, because neither interpretation would make the play more or less valuable as a work of art. This might strike you (it does me) as exactly the right position to take on this particular issue. It is, in a sense, a sceptical position, because it denies "truth" both to [172]the proposition that Hamlet slept with Ophelia, and to the apparently contrary proposition that he did not. But if this is scepticism, it is what we might call internal scepticism. It does not challenge the idea that good arguments can in principle be found for one interpretation of Hamlet rather than another. On the contrary it relies on an interpretive argument璽hat the value of the play lies in a dimension that does not intersect the sexual question璱n order to reach its "sceptical" position on that question.

Contrast the position of someone who says that no one interpretation of any work of art could ever succeed in showing it to be either really better or really worse, because there is not and cannot be any such thing as "value" in art at all. He means that there is something very wrong with the enterprise of interpretation (at least as I have described it) as a whole, not simply with particular issues or arguments within it. Of course he may have arguments for his position, or think he has; but these will not be arguments that, like the arguments of the internal sceptic, explicitly assume a positive theory of the value of art in general or of a particular work of art. They will be a priori, philosophical arguments attempting to show that the very idea of value in art is a deep mistake, that people who say they find a work of art "good" or "valuable" are not describing any objective property, but only expressing their own subjective reaction. This is external scepticism about art, and about interpretation in art.

B. The Threat of Scepticism

If a lawyer says that no one interpretation of the legal record can be "objectively" the correct interpretation, he might have external scepticism in mind. He might mean that if two judges disagree about the "correct" interpretation of the emotional damages cases, because they hold different theories of what a just law of negligence would be like, their disagreement is for that reason alone merely "subjective," and neither side can be "objectively" right. I cannot consider, in this essay, the various arguments that philosophers have offered for external scepticism about political morality. The best of these arguments rely on a general thesis of philosophy that might be called the "demonstrability hypothesis." This holds that no proposition can be true unless the means exist, at least in principle, to demonstrate its truth through arguments [convincing] to everyone who understands the language and is rational. If the demonstrability hypothesis is correct, then external scepticism is right about a great many human enterprises and activities; perhaps about all of them, including the activities we call scientific. I know of no good reason to accept the demonstrability hypothesis (it is at least an embarrassment that this hypothesis cannot itself be demonstrated in the sense it requires) and I am not myself an external sceptic. But rather than pursue the question of the demonstrability hypothesis, I shall change the subject.
Suppose you are an external sceptic about justice and other aspects of political morality. What follows about the question of how judges should decide cases? About whether naturalism is better than other (more conservative or more radical) theories of adjudication? You might think it follows that you should take no further interest in these questions at all. If so, I have some sympathy with your view. After all, you believe, on what you take to be impressive philosophical grounds, that no way of deciding cases at law can really be thought to be any better than any other, and that no way of interpreting legal practice can be preferred to any other on rational grounds. The "correct" theory of what judges [173]should do is only a matter of what judges feel like doing, or of what they believe will advance political causes to which they happen to be drawn. The "correct" interpretation of legal practice is only a matter of reading legal history so that it appeals to you, or so that you can use it in your own political interests. If you are convinced of these externally sceptical propositions, you might well do better to take up the interesting questions raised by certain sociologists of law璹uestions about the connection between judges' economic class and the decisions they are likely to reach, for example. Or to take up the study of strategies for working your will on judges if you ever come to argue before them, or on other judges if you ever join the bench yourself. Your external scepticism might well persuade you to take up these "practical" questions and set aside the "theoretical" questions you have come to see as meaningless.

But it is worth noticing that philosophers who say they are external sceptics rarely draw that sort of practical conclusion for themselves. Most of them seem to take a rather different line, which I do not myself fully understand, but which can, I think, fairly be represented as follows. External scepticism is not a position within an enterprise, but about an enterprise. It does not tell us to stop making the kinds of arguments we are disposed to make and accept and act on within morality or politics, but only to change our beliefs about what we are doing when we act this way. Imagine that some chessplayers thought that chess was an "objective" battle between forces of light and darkness, so that when black won good had triumphed in some metaphysical sense. External sceptics about chess would reject this view, and think that chess was entertainment merely; but they would not thereupon cease playing chess or play it any differently from their deluded fellow players. So external sceptics about political morality will still have opinions and make arguments about justice; they will simply understand, in their philosophical moments, that when they do this they are not discovering timeless and objective truths.

If you are an external sceptic who takes this attitude, you will have driven a wedge between your external scepticism and any judgments you might make about how judges should decide cases, in general, or about what the best justification is of some part of the law, in particular. You will have your own opinions about these matters, which you will express in arguments or, if you are an academic lawyer, in law review articles or, if you are a judge, in your decisions. You may well come to believe that the best interpretation of the emotional damages cases shows them to be grounded in the principle of foreseeability, for example. When you retreat to your philosophical study, you will have a particular view about the opinions you expressed or exhibited while you were "playing the game." You will believe that your opinions about the best justification of the emotional damage cases were "merely" subjective opinions (whatever that means) with no basis in any "objective" reality. But this does not itself provide any argument in favor of other opinions about the best interpretation. In particular, it does not provide any argument in favor of the internally sceptical opinion that no interpretation of the accident cases is best.

Of course, your external scepticism leaves you free to take up that internally sceptical position if you believe you have good internal arguments for it. Suppose you are trying to decide whether the best interpretation of the emotional damage cases lies in the principle that people in the area of physical risk may recover for emotional damage, or the broader principle that anyone whose emotional dam-[174]age was foreseeable may recover. After the most diligent search and reflection asking yourself exactly the questions naturalism poses, you may find that the case for neither of these interpretations seems to you any stronger than the case for the other. I think this is very unlikely, but that is beside the present point, which is only that it is possible. You would be internally sceptical, in this way, about any uniquely "correct" interpretation of this group of cases; but you would have supplied an affirmative argument, beginning in your naturalistic theory, for that internally sceptical conclusion. It would not have mattered whether you were an external sceptic, who nevertheless "played the game" as a naturalist, or an external "believer" who thought that naturalism was stitched into the fabric of the universe. You would have reached the same internally sceptical conclusion, on these assumed beliefs and facts, in either case.

What is, then, the threat that external scepticism poses to naturalism? It is potentially very threatening indeed, not only to naturalism, but to all its rival theories of adjudication as well. It may persuade you to try to have nothing to do with morality or legal theory at all, though I do not think you will succeed in giving up these immensely important human activities. If this very great threat fails (as it seems to have failed for almost all external sceptics) then no influence remains. For in whatever spirit you do enter any of these enterprises環owever firmly your fingers may be crossed璽he full range of positions within the enterprise is open to you on equal terms. If you end in some internally sceptical position of some sort, this will be because of the internal power of the arguments that drove you there, not because of your external sceptical credentials.

We must now consider another possibility. The sceptical attack upon naturalism may in fact consist, not in the external scepticism I have been discussing, but in some global form of internal scepticism. I just conceded the possibility that we might find reason for internal scepticism about the best interpretation of some particular body of law. Suppose we had reasons to be internally sceptical about the best interpretation of any and all parts of the law? It is hard to imagine the plausible arguments that would bring us to that conclusion, but not hard to imagine how someone with bizarre views might be brought to it. Suppose one holds that all morality rests on God's will, and had just decided that there is no God. Or he believes that only spontaneous and unreflective decisions can have moral value, and that no judicial decision can either be spontaneous or encourage spontaneity. These would be arguments not rejecting the idea or sense of morality, as in the case of external scepticism, but employing what the author takes to be the best conception of morality in service of a wholesale internally sceptical position. If this position were in fact the right view to take up about political morality, then it would always be wrong to suppose that one interpretation of past judicial decisions was better than another, at least in cases when both passed the threshold test of fit. Naturalism would therefore be a silly theory to recommend to judges. So the threat of external scepticism [sic: presumably "internal skepticism" was intended here], it materializes, is in fact much greater than the threat of external scepticism. But (as the examples I chose may have suggested) I cannot think of any plausible arguments for global internal scepticism about political morality.
Of course, nothing in this short discussion disputes the claim, which is plainly true, that different judges hold different political moralities, and will therefore disagree about the best justification of the past. Or the claim, equally true, that [175]there will be no way for any side in such disagreements to prove that it is right and its opponents wrong. The demonstrability thesis (as I said) argues from these undeniable facts to general external scepticism. But even if we reject that thesis, as I do, the bare fact of disagreement may be thought to support an independent challenge [to] naturalism, which does not depend on either external or internal scepticism. For it may be said that whether or not there is an objectively right answer to the question of justification, it is unfair that the answer of one judge be accepted as final when he has no way to prove, as against those who disagree, that his position is better. This is part of the argument from democracy to which we must now turn.

3. Is It Undemocratic?

So if we are to reject naturalism, in favor of some other positive theory of adjudication, this cannot be by virtue of any general appeal to external scepticism as a philosophical doctrine. We need arguments of substantive political morality showing why naturalism is unwise or unjust. In the remaining sections of the essay I shall consider certain arguments, that I have either heard or invented, to that effect. Of course arguments against naturalism must compare it, unfavorably, with some other theory, and arguments that might be effective in the context of one such comparison would be self-defeating in another.

I shall consider, first, the arguments that might be made against naturalism from the standpoint of what I believe is a more positivist theory of adjudication, though nothing turns on whether this theory is properly called positivism. Someone might propose, as an alternative to naturalism, that judges should decide cases in the following way. First, they should identify the persons or institutions which are authorized to make law by the social conventions of their community. Next, they should check the record of history to see whether any such persons or institutions have laid down a rule of law whose language unambiguously covers the case at hand. If so, they should decide that case by applying that rule. If not璱f history shows that no rule has been laid down deciding the case either way璽hen they should create the best rule for the future, and apply it retrospectively. The rule they thus create would then become, for later judges, part of the record endorsed by convention, so that later judges facing the same issue could then find, in that decision, language settling the matter for them. We might call this theory of adjudication "conventionalism."

Some people are drawn to conventionalism, over naturalism, because they think the former is more democratic. It argues that people only have the rights, in court, that legislators and judges, whom convention recognizes to have legislative power, have already decided to give them. Naturalism, on the other hand, assigns judges the power to draw from judicial history rights that no official institution has ever sanctioned before, and to do so on no stronger argument than that the past is seen in a better light, according to the convictions of the judges, if these rights are presupposed. This seems the antithesis of what democracy requires.
But this argument mistakes the cases in which a conventionalist and a naturalist are likely to disagree. Conventionalist judges can dispose of cases at the first stage, by copying the decisions already made by elected officials, only in those cases in which some statute exactly in point unambiguously dictates a particular result. Any conscientious naturalist is very likely to make exactly the same deci-[176]sions in those "simple" cases, so conventionalism cannot be more democratic because it decides these differently. The two styles of adjudication will normally recommend different decisions only when some fresh judicial judgment is required which goes beyond what the legislature has unarguably said, either because the statute in play is open to different interpretations, or because no particular statute is in play at all. But in these "hard" cases the difference between the two theories of adjudication cannot be that one defers to the legislature's judgment while the other challenges that judgment. Because, by hypothesis, there is no legislative judgment that can be treated in either of these ways. Conventionalism argues that the judge must, in these "hard" cases, choose the rule of decision which best promotes the good society as he conceives it. It is hardly more democratic for judges to rely on their own convictions about the best design of the future than to rely instead on their convictions about the best interpretation of the past.

So the argument from democracy in favor of conventionalism over naturalism seems to come to nothing. But we should consider one possible counter-argument. I have been assuming that conventionalism and naturalism will designate the same cases as "easy," that is, as cases in which no fresh judgment is required by the judge. But perhaps a naturalist has more room than a conventionalist to deny that an apparently "easy" case really is. Consider the following example. Since naturalism encourages a judge to rely on his own convictions about which interpretation shows the past in the best light, it permits outrageous political convictions to generate outrageous judicial decisions. Suppose a naturalist judge believes that majority will is tyranny. He believes that our political institutions should be arranged so that statutes are enforced only when they have been enacted by a two-thirds vote. He acknowledges that he cannot apply this principle unless it provides an acceptable fit with past practice, but he sets the threshold of fit low enough, in perfect good faith, so as to be able conscientiously to claim that all counter-examples (all cases in which statutes passed by a bare majority have been enforced by the courts) are "mistakes."

This is no doubt possible. Nothing in the design of naturalism insures that a judge with silly or mad opinions will not be appointed; but nothing in the design of conventionalism insures that either; and conventionalism will not prevent him from reaching preposterous decisions once appointed. A conventionalist judge needs a concept of convention. He must decide, for example, whether it is a convention of our society that the Constitution should be followed, and nothing in the structure of conventionalism can insure that a judge will in fact reach the correct answer to that question. No theory of adjudication can guarantee that only sensible decisions will be reached by judges who embrace that theory. We can protect ourselves from madness or gross stupidity only by independent procedures governing how judges are to be appointed, how their decisions may be appealed and reversed, and how they may be removed from office if this should appear necessary.

But it may now be said that naturalism would encourage anti-democratic decisions from judges who hold, not mad, but plausible and even attractive political convictions, and who deploy perfectly sensible theories about how much of the past an interpretation must fit. For naturalism leaves no doctrine or practice immune from re-examination. We may use an earlier example as an illustration. Suppose a firm line of cases has rejected the idea that clients may sue lawyers who are negligent. Conventionalism is then committed (so it might be said) to [177]continuing that doctrine until it is reversed by legislation, which seems the democratic solution. But naturalism encourages judges to put this line of cases in a wider context, and ask whether the rule refusing recovery against negligent lawyers would not itself be rejected by the best justification of the rest of the law, which allows recovery for negligent injury of almost every other kind. So a naturalist might be led to overrule these cases, which a conventionalist would leave for the legislature to review.

Indeed there is nothing in the theory of naturalism, as I described it, which would prevent an intelligent and sensible naturalist from taking the same line with certain statutes. Suppose an old statute makes blasphemy a crime and, though it has not been enforced in centuries, it is suddenly revived by a public prosecutor anxious to make a splash. A naturalist judge might well develop a theory of obsolescence, even though this had never been recognized in the jurisdiction before. He might say that the best interpretation of judicial practice as a whole yields the following qualification to the rule that statutes are always to be enforced. "Old statutes quite at variance with the spirit of the present time, which would not be enacted by a present legislature, and which have not been employed since ancient times, are unavailable as grounds of criminal prosecution." If prosecutors have not tried to revive old statutes in the fairly recent past, this qualification would be consistent with judicial practice, and it might plausibly be thought to show that practice in a better light, as both more rational and more closely tying what counts as valid legislation to the will of the people.

So both in the case of precedent and legislation a competent naturalist judge might find certain cases hard, and amenable to the command of imaginative reinterpretation, which a conventionalist must concede to be easy even when the obvious answer is unattractive. So perhaps naturalism would sometimes produce "novel" decisions by sensible judges that conventionalism would discourage. But is it right to say that naturalism is for this reason less "democratic." A minimally competent naturalist judge would begin his argument by recognizing, indeed, insisting, that our political system is a democracy; he would continue by arguing that democracy, properly understood, is best served by a coherent rather than an unprincipled private law of negligence, and by an institution of legislation that is sensitive rather than obdurate to changes in popular morality. So the disagreement between naturalism and conventionalism about which cases are really "easy" is not a disagreement between those who oppose and those who respect democracy; it is rather the more familiar disagreement about what democracy really is. When the disagreement is seen in this light, it is far from apparent that the naturalist has the worst of the argument. In the next section, I shall argue that naturalism respects, better than its rivals, a right that has seemed to many people crucial to the idea of democracy, which is the right each person has to be treated, by his government, as an equal.

4. Is It Crazy?

A. Instrumentalism

We must turn now to the arguments that might be made against naturalism, not from the standpoint of conventionalism, but from the different direction of a more radical theory I shall call instrumentalism. This theory encourages judges always to look to the future: to try to make the community as good and wise and [178]just a community as it can be, with no essential regard to what it has been until now. Of course instrumentalist judges will differ, among themselves, about the correct model of the good community. Some will define this in almost exclusively economic terms. They will think that a rich community is for that reason a good community. Others will take a more utilitarian line, and emphasize the importance of general happiness over total wealth. But still others will insist on the importance of personal and political rights, and will therefore provide, in their account of the good society, that certain fundamental interests of individuals, like liberty of conscience or a decent standard of living, be respected at the cost of general wealth or average happiness.

An instrumentalist judge will see himself or herself as an officer of government charged with contributing to the good society according to his or her conception of what that is. Of course a sensible instrumentalist judge will acknowledge the importance of institutional factors as either an obstacle or opportunity in this enterprise. He will understand, in particular, that the rules he fashions must work together with the rules provided by other institutions and other officials, so that he is constrained by what we might call consistency in strategy. If the legislature and other judges have laid down rules in the past that he is powerless to overrule, for example, he must not create rules of his own which, operating alongside those established rules, would produce chaos. For that would make the community worse not better off through his efforts. But instrumentalism denies that judges should be constrained by the past in any less pragmatic way than that. It denies, in particular, that they should also seek consistency in principle, by making their decisions conform to the best interpretation, as the naturalist conceives this, of the past. Naturalism insists that the past should be allowed to cast a shadow over the future beyond the pragmatic requirements of strategy. Instrumentalism condemns this as irrational.
In order to bring out the difference between the two theories, consider this situation. You think that it would be best, all things considered, if no one were ever allowed to recover damages for emotional injury. You think this because you believe that actions for emotional damage involve the risk of fraud, and force insurance premiums higher than the optimum for economic efficiency. Of course you think, as part of this view, that no one has what we might call a moral right that the law provide damages for emotional injuries. If you thought anyone did have such a right, then you would think that the good society should recognize that right and enforce it by producing the appropriate legislation even at the cost of efficiency. But since you think people have no such moral right you think that society would be better off, on the whole, if it provided no legal right to such damages.

Now suppose you are an instrumentalist judge faced with a suit by a mother who suffered emotional injury when she heard, on the telephone, that her son had been run down by a careless driver. You find, when you search the books, that the other judges of your jurisdiction have consistently awarded recovery for emotional damages to relatives who actually saw physical damage to someone they loved. Of course you think that all these decisions were wrong. You would be tempted to overrule the whole line of decisions if you could, but suppose this is beyond your power. The line might include decisions of the highest court of the state, for example, and you might be sitting in a lower court. You will nevertheless grasp the opportunity to limit the damage these cases do to the community's welfare, according to your convictions, by declaring that only relatives who actu-[179]ally saw the injury may recover for emotional damage. This will create no practical contradictions, or inconsistency in strategy.
What objection could there be to this instrumentalist solution to the problem, assuming as you do, that it conduces to a better state of affairs, on the whole, than the opposite decision? A naturalist might be led by his naturalism to the opposite decision, even if he shared your assumptions about the best state of affairs. He would be unable to find any principled distinction between seeing and hearing about an accident, and he would be forced to concede that the best justification of the past recognizes a judicial right to recover for emotional damage if that damage was reasonably foreseeable. Of course he might try to show (as the naturalist judge in the example I considered earlier was able to show about actions in negligence against lawyers) that allowing recovery for emotional damage was inconsistent with some broader line of cases. But suppose he could not show this, as indeed he is unlikely to be able to do. He would then be forced to decide the present case for the plaintiff mother, therefore compounding the damage to the future. What could be the possible sense or other merit in that? This is the basis of the instrumentalist charge: that insofar as naturalism requires different decisions from those an instrumentalist would reach, naturalism is crazy.

Naturalism seems to assume that in these circumstances it would be for some reason unfair to decide against her. But why? She has (by hypothesis) no moral right to a rule allowing her damages. On the contrary, the situation would be better if no one were ever required to pay damage for her sort of injury. The fact that our judicial process has made one mistake is no good argument for making that mistake more general. Of course a naturalist cannot say that it would be unfair to decide against the mother because most judges in the past have behaved as naturalists. It would beg the present question to say that this provides a reason why a judicial decision that offends naturalism is unfair. For the question at issue is whether it is unfair to reach a decision which offends the best interpretation of the past. If we want to sustain naturalism as against instrumentalism, we must argue that the fact that a given principle figures in the best justification of legal practice as a whole provides a reason for extending that principle into the future, and we must not rely on that very claim in making our case for it. But how can we then argue the case? What can we say to the instrumentalist who claims, reasonably enough, that two mistakes are worse than one?

B. The Political Order

The naturalist might begin his reply by noticing that the dispute now in play is wider than simply a dispute about how judges should decide cases. Naturalism assumes and instrumentalism denies that the members of a community can have rights and duties against one another, and against the community as such, just by virtue of the political history of the community. That they can have rights and duties they would not have if that history had been different. But this is an idea familiar not only to lawyers but to our general political rhetoric. Politicians say that America is a democracy, and therefore that certain things ought and ought not to be done. Or that America respects the rule of law, and therefore that Congress should not enact certain laws.

We should give a name to the idea behind this rhetoric. Let us say that the set of political rights people have just by virtue of the political history of their community constitutes the "political order" of the community. Naturalism recognizes [180]that communities have political orders, and offers an account of what a political order is. A community's political order is provided by the principles assumed in the best interpretation (in the sense we have been using) of its concrete political structures, practices and decisions. Naturalism supposes that people have a right to have this order enforced, in court, on demand. It is not true that every rule of law a legislature or court adopts is part of the political order, properly understood. The best interpretation of the order as a whole may show this particular rule inconsistent with the rest, and so a "mistake" that should be ignored in stating what the order really is. But if it is indeed part of the genuine political order, properly understood, that people suffering emotional injury are entitled to damages against the tortfeasor, then someone who has suffered such damage is for that reason entitled to a judicial order to that effect.

Of course naturalism is a theory about judicial rights, that is, about the rights people have to win law suits. It takes no position about how far the political order furnishes or constrains the rights people have to particular legislation in their favor, or their rights to revolt or otherwise to establish a very different political order. If the political order includes a constitution which, properly interpreted, disables the legislature from changing the present order in certain ways, then people do have judicial rights, under this order, that the courts not enforce legislation which contradicts these commands. But naturalism, as such, leaves the legislature otherwise free to improve the present order, both in detail and, if appropriate, radically. The idea, that people have an abstract judicial right to the enforcement of the present order, imposes a kind of conservatism on politics; but this is a conservatism imposed on adjudication alone.
Instrumentalism challenges not simply naturalism's conception of a political order, but the concept of a political order itself. It denies the fact that political history that has taken a certain form can ever be the ground of a genuine right or duty at least against a court. This is the upshot of the instrumentalist's thesis that there are no judicial rights by virtue of the judicial past. He believes that a judge is never obliged, by the nature of the past, to work against the best solution for the future. The instrumentalist argues that the idea that judges are constrained in this way is irrational. Of course he recognizes each society has a distinctive political past, and concedes that most people believe their rights and duties are, at least in some ways, a function of the past. But the instrumentalist holds that this opinion is silly.

Now what arguments does a naturalist have available in reply? We might begin by considering one familiar argument a naturalist might be tempted to make, though only to reject it. Someone might argue that judges should never attempt to change the political order because this would require them to make judgments of political morality which ought to be left to the peoples' elected representatives. So judges should accept the popular idea of a political order, and enforce that order as history presents it to them, for that reason. This is like the (bad) argument we supposed a conventionalist might make against naturalism; in any case it is not an argument a naturalist can make against instrumentalism because, according to naturalism, a judge must make decisions of political morality in order to decide what the political order, properly construed, really is. We labored that point in our description of how a naturalist judge would go about deciding which interpretation of the past was the best interpretation. There is, for the naturalist, a crucial distinction between interpreting and improving the political order of the community, but these are both activities which engage the judge's moral sense.

For much the same reason the naturalist cannot use another familiar argument often made in favor of judicial conservatism. It is sometimes said that judges do great damage to social efficiency when they surprise litigants by changing established rules of law. Once again this is an argument that a conventionalist might be tempted to employ against naturalism. But it is unavailable to naturalism because nothing insures that a naturalist judge's interpretation of the past will not prove surprising. A naturalist is charged with discovering and enforcing the best interpretation of his community's political structure and past decisions, but the interpretation he believes best may be (as we saw in the example of Cardozo's decision in McPherson v. Buick) interpretation that has occurred to no one else. In any case, the argument is a bad argument against instrumentalism for a different reason. This argument supposes that a novel decision, such as an instrumentalist might make, will in fact be unwise, pragmatically, for the future. But if this is really so, then an instrumentalist is ready to take that into account in deciding which decision will be best for the future. We noticed that an instrumentalist will look to the past, not as a source of rights, but strategically, to discover whether his judgment will in fact have the beneficial effects on the future he supposes. If disregarding some established line of precedent will actually diminish efficiency, because it will discourage people from counting on established rules of law in planning their affairs, then this is exactly the kind of strategic consideration instrumentalism stands ready to acknowledge.
A naturalist must find his defense of naturalism璷f his idea that the standing political order is a source of judicial rights璭lsewhere. He must meet the instrumentalist's challenge directly, by showing why people can have genuine political rights just by virtue of the actual political history of their community, and why these rights hold with special force in litigation. Can we find such an argument for naturalism? We must begin by stipulating a general requirement of justice in government. Any government must treat its citizens as equals, as equally entitled to concern and respect. Of course this general requirement is very abstract. Different people璦nd different societies瓀ill have different views of what it is to treat people as equals. But we can nevertheless speak of a general duty of government to treat its citizens this way, and derive from this two distinct and more concrete responsibilities. The first is the responsibility, in creating a political order, to respect whatever underlying moral and political rights citizens may have in the name of genuine equality. The second is the obligation to extend whatever political order it does create equally and consistently to everyone.

These obligations are distinct because they can be fulfilled or violated independently. A society may develop a conception of justice that we, as critics of that society, reject. In its pursuit of efficiency or other collective goals, it may violate rights we think people have as individuals, but it may nevertheless enforce that conception consistently and, in that sense, fairly, allowing to everyone the resources, opportunities, and protections each is entitled to have under the theory it has adopted. It may, on the other hand, put in place an admirable political order; it may adopt a general scheme of principles and institutions, which we, as critics, approve as exactly what justice requires; but it may nevertheless fail to enforce that scheme consistently, so that some people do not have the resources and opportunities the public order requires them to have.

Once we recognize both the fact and the independence of these two rights, we see how it is possible that a government might commit the following special form of injustice. It might deny to some people a right it has, but need not have, extended to others. But that is exactly what the instrumentalist judge I just imagined does in denying the mother her suit for emotional damages. In one sense the situation that follows his decision is an improvement over the situation that would have resulted had he decided for the mother. If he is right in thinking that allowing recovery for emotional damage is not required by morality, and that it is damaging to the economy, then there will be less "unnecessary" damage to economic efficiency. But the plaintiff mother in this case nevertheless has a complaint. Though she has no right to a legal regime under which people in her position recover damages, she does have a right that the legal regime in force be consistently applied to her. Otherwise society fails to give her justice according to its conception of what justice requires, and that is a failure to treat her with equal concern and respect. One of her political rights has been violated.

So the naturalist's approach to this case is correct, and the instrumentalist's wrong, because the former respects and the latter violates the plaintiff's right to be treated as an equal. This is enough to make out what I am presently most anxious to show: that instrumentalism is wrong in assuming that the political order cannot be an independent source of rights. Of course the case I supposed as an example made it easier to demonstrate that point; it is implausible to think that a negligent driver has a moral right not to have the law recognize emotional damages. So the defendant driver seemed to have no proper objection to a naturalist judge's decision in favor of the mother. We can easily imagine cases, however, in which even the best interpretation of the community's law would show that it failed to recognize a substantive right someone ought to have.

Suppose, for example, that someone sues for damages for invasion of privacy, but even the most sophisticated interpretation of the law of the community fails to reveal a principle sustaining any such right. Now the situation reveals a conflict (as we might put it) between the two rights that follow from the abstract right to justice. A naturalist judge, who denies the action, will have upheld the defendant's right to a consistent application of the public order, but failed to uphold the plaintiff's right to a better public order. Naturalism insists that the function of courts, at least in a political society meeting minimum standards of justice, is to address the former rather than the latter right. No doubt more argument is necessary (which I cannot supply here) to sustain that choice. Once the two rights are distinguished, however, and both recognized, that choice is not crazy. Naturalism is not, as the present objection supposed, irrational.
But what if the condition I just mentioned is not met? What if the best interpretation of the legal system (or some important part of it) shows it to be wicked? Suppose that the most sophisticated interpretation of our Constitution, at the time of the Fugitive Slave Acts, contained no principle in virtue of which slaves had a right to be free, so that even a naturalist judge would have had to recognize those unfortunate statutes as perfectly constitutional. An instrumentalist might well want to say that here, at least, instrumentalism would provide a better guide to decision, because it would advise the judge to ignore the constitutional structure, if he could get away with this, and find some way to thwart the Acts. But naturalism has the virtue, even in cases like this one, of bringing to the surface an issue of political morality that cannot be ignored.

Of course a constitutional structure that permits slavery is deeply defective. It violates people's first political right: the right to a public order that treats them as equals. The more difficult issue is this: is there any room, in this sorry picture, for the slaveholders' second right? Does the slaveholder whose slaves have escaped have any right, however weak, that the constitutional system be enforced on his behalf, as it is on behalf of the slaveowner who has managed to keep his slaves imprisoned at home? If you were a naturalist judge, you might think that he does. In this case you would have to decide the Fugitive Slave Cases for the slaveowners even though you despise them and deplore that constitution, and even though you privately work for a constitutional amendment or even for the civil war. But you might also come to the opposite conclusion. You might think that no one can have any right, even a weak right, to the equal benefit of wicked laws. In that case you would decide against the slaveowners if you could, because the underlying reason for your concern with the past, which is people's abstract rights to institutional consistency, would have exhausted its power. It would not matter if you put your conclusion in the terminology of older natural law theories, and said that the Fugitive Slave Acts were not really law. Or if you used the language of modern positivism, and said that though they were law they were too evil to be enforced. For the important issue is not what you say but what you do, and though naturalism does not in itself answer the difficult moral question I posed, it does tell you what consequences for our decision follow from the answer you give to that question.

C. The Two Ideals

Perhaps you will allow me a summary of this last part of my argument. Our political system admits of two ideals; it is imperfect in two ways. It stands in the shadow of an external ideal, which is the ideal of a perfectly just and effective system. This is the challenge it offers to legislation, and, beyond that, to the political will and sense of justice of the community which has the standing power to make it better, closer to the external ideal of what a political system should be. But unless it is a very bad political system it stands also in the shadow of a different, internal ideal, which is the ideal of itself made pure. This is the challenge it offers to adjudication: the challenge of making the standards that govern our collective lives articulate, coherent and effective. Naturalism insists on the difference between the two ideals, and makes that difference the nerve of the rule of law.
People will disagree about what the internal ideal of our order is like, perhaps just as much as they disagree about what external ideal our order should pursue. Indeed they will disagree about the former precisely because they disagree about the latter. So no one will have any guarantee that, if he should come to court, those who judge him according to naturalism will reach the result that he himself thought was the best interpretation of our order when he acted. That is inevitable in any community which recognizes what is plainly true: that people have rights beyond the rights conventionalism recognizes, that is, that they have rights beyond the strict and narrow limits within which everyone agrees what these rights are. But naturalism at least takes the actual political order, properly interpreted, as the common standard, so that citizens are encouraged to put to themselves the same questions that officials who adjudicate their disputes will ask in judging them. No doubt this practice will cause surprise and disappointment, even despair. No doubt it will produce injustice. Its virtue is that it seems less vulnerable, in all these respects, than available alternatives for bringing the rule of principle to an imperfect world.

We can, as a community, strive towards these two ideals at the same time, though through different institutions and practice. We embrace the two ideals as an agenda for sustained and continuing debate. We have no hope璦nd indeed no wish璽hat the debate will end. We understand that the decision of political officials must be accepted, from time to time. But we insist that this is only because someone's decisions must be accepted and not because these decisions come guaranteed for accuracy. We know that the quality of the debate is itself, quite apart from any agreement it might produce, something that makes ourselves and our community better. This is the image we should have of politics and of our lives in politics. Our courts play a distinct sovereign and indispensable role in this image. They are the forum of the second ideal.


1. At an earlier time, Professor Fuller had suggested a less literary analogy than the chain novel of Professor Dworkin. Fuller supposed:
An inventor of useful household devices dies leaving the pencil sketch of an invention on which he was working at the time of his death. On his deathbed he requests his son to continue work on the invention, though he dies without having had a chance to tell the son what purpose the invention was to serve or anything about his own plans for completing it. In carrying out his father's wish the son's first step would be to decide what the purpose of the projected invention was, what defect or insufficiency of existing devices it was intended to remedy. He would then try to grasp the underlying principle of the projected invention, the "true reason of the remedy"....With these problems solved he would then proceed to work out what was essential to complete the design for the projected device.
L. FULLER, THE MORALITY OF LAW 84-85 (rev. ed. 1969). Also, Fuller argued that his desideratum about avoiding logically contradictory rules can be extended to the need to avoid rules that are inconsistent in underlying principle. Is there any significant difference between Fuller's and Dworkin's approach? Which do you think is better?

2. Notice that Dworkin's theory goes beyond an account of pure common-law decision making; he would apply the same methodology in cases involving the interpretation of statutes. What do you suppose Dworkin would think about the decision in Riggs v. Palmer? See RONALD DWORKIN, LAW'S EMPIRE 15-20 & passim (1986) (discussing Riggs in depth). The methodology can also be applied in the context of constitutional issues: How, for example, might it work in the context of the fugitive slave controversy sampled in Prigg? See Ronald Dworkin, The Law of the Slave-Catchers (reviewing R. COVER, JUSTICE ACCUSED (1975)), TIMES LITERARY SUPPLEMENT, Dec. 5, 1975, at 1437 (arguing that the fugitive slave laws could have been considered unconstitutional "mistakes" when viewed in terms of the interpretation of the constitution as a whole).

This page was last modified on February 18, 1999.

The Justice and the Legal System website is administered by L. Daniel Liutikas.
Entire contents ?1998-99. All rights reserved