The Case Law System in America
Author(s): Karl N. Llewellyn, Paul Gewirtz, Michael Ansaldi Source: Columbia Law Review, Vol. 88, No. 5 (Jun.,  1988), pp. Published by: Columbia Law Review Association, Inc. Stable URL: http://www.jstor.org/stable/1122696
Accessed: 24/02/2010 07:52
989-1020
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=clra.
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.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Columbia Law Review Association, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Columbia Law Review.
STOR
http://www.jstor.org
THE CASE LAW SYSTEM IN AMERICA
By Karl N. Llewellyn*
Edited with an Introduction by Paul Gewirtz**
Translated from the German by Michael Ansaldi***
Introduction
Only a few American legal scholars have been unquestionably great. Karl Llewellyn, who died in 1962, is surely one of these. Enormously creative and influential in such diverse fields as contracts, commercial law, jurisprudence, and anthropology,1 Llewellyn was perhaps the most important of the "legal realists." He was a person of almost heroic intellectual ambition, yet was also actively involved with the practical affairs of the legal profession. His biographer has aptly called him "the most romantic of legal realists, the most down-to-earth of legal theorists."2
What follows are excerpts from a major "lost" book that Llewellyn wrote at the height of his powers. The book arose out of a course that Llewellyn gave in Germany in 1928-29, while on leave from the Columbia Law School faculty and visiting on the Leipzig Faculty of Law. The course was designed to introduce German lawyers and scholars to the American case law system. Supplemented with illustrative material that resembles a condensed American casebook, Llewellyn's lectures, which were given in the German language, were revised and published in Germany in 1933 under the title Präjudizienrecht und Rechtsprechung in Amerika. American scholars who knew German immediately recognized that it was a major work. The book was reviewed and analyzed in leading American law reviews.3  Professor Lon Fuller, for ex-
© 1988 Paul Gewirtz and Michael Ansaldi.
*   Late Professor of Law, University of Chicago Law School.
**   Professor of Law, Yale Law School.
***   Assistant Professor of Law, Florida State University College of Law.
1.  Some of his leading works include: The Bramble Bush (1930); Cases and Materials on the Law of Sales (1930); The Cheyenne Way (1941) (with E. Hoebel); The Common Law Tradition: Deciding Appeals (1960); Jurisprudence: Realism in Theory and Practice (1962) (collecting previously published essays on jurisprudential subjects); The First Struggle to Unhorse Sales, 52 Harv. L. Rev. 873 (1939); Across Sales on Horseback, 52 Harv. L. Rev. 725 (1939); On Our Case-Law of Contract: Offer and Acceptance (pts. 1 8c 2), 48 Yale L.J. 1, 779 (1938-39); The Rule of Law in Our Case-Law of Contract, 47 Yale L.J. 1243 (1938); On Warranty of Quality, and Society (Part 1), 36 Colum. L. Rev. 699 (1936); On Warranty of Quality, and Society (Part 2), 37 Colum. L. Rev. 341 (1937); What Price Contract?—An Essay in Perspective, 40 Yale L.J. 704 (1931). Llewellyn was also the Chief Reporter for and a principal drafter of the Uniform Commercial Code. He taught at Yale Law School from 1919-20 and 1922-24, at Columbia Law School from 1924-51, and at the University of Chicago Law School from 1951-62.
2.  W. Twining, Karl Llewellyn and the Realist Movement vii (1985).
3.  See, e.g., Radin, Case Law and Stare Decisis: Concerning Präjudizienrecht In Amerika, 33 Colum. L. Rev. 199 (1933) (article-length review essay).
989
990                            COLUMBIA LA W REVIEW               [Vol. 88:989
ample, viewed the book as so important that his well-known article on the American legal realist movement—an article he ''intended as a critical evaluation of legal realism generally"4—took Llewellyn's book as its main focus. Wrote Fuller:
[W]e lack as yet a comprehensive work which will both describe and apply the methods of legal realism, which can serve both as an exposition of the approach and as an exemplification of it ... . The nearest approach to such a work is to be found in a book recently published by Professor Llewellyn. [Präjudizienrecht]5
For all the attention the book received in the 1930s, however, it was never translated into English or published in the United States, and it has long been out of print in Germany.
My own study of the legal realists kept leading me to references to the book and eventually to the judgment that it should be translated and published for an English-language audience. The excerpts printed here give a flavor of the work. The entire book, which Professor Michael Ansaldi translated and I edited, is expected to be published in 1989.
The subject of the book is the American case law system—how cases are decided, how precedents are worked with, how various social actors perceive the case law system, and how such a system evolves. Case law was a lifelong concern of Llewellyn's, and in the evolution of Llewellyn's ideas The Case Law System in America stands between The Bramble Bush (1930) and The Common Law Tradition (1960). Perhaps because Llewellyn was addressing a non-American audience in this book, he presented his ideas in a more systematic and structured way than elsewhere, and he articulated his premises more clearly and explicitly. Inevitably, he introduced comparative law perspectives. Most importantly, he wrote a book that is full of new and arresting ideas, arrest-ingly presented.
In his correspondence, Llewellyn refers to his book as "a type of study which . . . has never been consistently carried out"—a "study of the sociology of case law."6 Central to this project, and to the excerpts here, is a demonstration that our case law system provides judges both leeways and constraints—a conclusion that depends upon a proper understanding of legal rules and precedent. Llewellyn argues that courts do not "apply" legal rules, but either expand or contract them; creation is unescapable (§ 52).7 The doctrine of precedent is two headed, providing one technique for narrowing an unwelcome precedent and an-
4.  Fuller, American Legal Realism, 82 U. Pa. L. Rev. 429, 430 (1934).
5.  Id.
6.  Letter from Karl Llewellyn to Walter Sharp (Mar. 7, 1930) (in the Karl Llewellyn papers, University of Chicago).
7.  This and other references to section numbers indicate sections of the complete book that are excerpted below.
1988]
LLEWELLYN
991
other to expand a welcome one. For Llewellyn, these contradictory techniques are each "correct," and in fact, their coexistence is necessary for the viability of the case law system. Furthermore, as Llewellyn argues in his brilliant discussion of concurring and dissenting opinions, there is leeway in deciding "what 'the' facts are" and how they are classified for purposes of legal analysis, with "each way of construing the facts [containing] a degree of violence to either the fact situation or the classifying category" (§ 42).
Llewellyn's focus on contradictions, dualities, and leeways in the law clearly anticipates some elements of modern critical theory. However, Llewellyn's point was not to trumpet law's indeterminacy, but to emphasize how an adequate measure of predictability and certainty in the case law is achieved nevertheless. In this regard, Llewellyn makes three points about "legal certainty" that stand out. First, the "operating technique" of judges and lawyers—a lawyer's craft and "feel"—is a powerful source of certainty in the law (§ 55). Second, it is the fact situations of cases rather than legal rules that guide decisions, producing a " 'sense of justice' in the individual case" (§ 56). This pressure from the facts is what "makes different judges, . . . despite their divergent analyses, generally . . . reach the same result" (§ 42), and thus act in predictable ways. Third, Llewellyn distinguishes between legal certainty for a lawyer (which means predictability of judicial decisions) and legal certainty for a layperson (which means congruence between legal rules and "real-life norms"). Almost paradoxically, Llewellyn argues that this latter sort of "certainty" requires "change" in the legal rules since social norms are continually changing (§ 58).
Pervading this work is a sense of Llewellyn's faith—faith in law, lawyering, and legal scholarship itself. By insisting that our case law retains predictability and certainty in spite of the legal realists' insights, Llewellyn affirms that something recognizable as "law" survives. In affirming "operating technique" as a source of predictability for the legal system, Llewellyn finds morality in the lawyer's craft. And in concluding with a statement of some modest pride in his own work, he offers "hope" that legal scholarship can bring us "closer to Life" and to "knowledge about the whole purpose of law, its utility to society in general" (§ 74).
RG.
* * * * *
992                            COLUMBIA LA W REVIEW               [Vol. 88:989
Section 8 How Case Law Rules are Applied
In American case law technique, two things should be distinguished: 1) deciding a case after all the relevant legal rules have been determined, and 2) determining what the relevant rules of law are. Let us turn initially to the first, and apparently the simpler of these.
In the main, the process in America is strikingly similar to the way cases are decided in Germany, whether the relevant rule comes from the constitution, statutes, or case law—particularly if the field of substantive law is well-tilled, and the wording of the relevant rules has already taken on a sharply defined cast. In both countries, we may pretty safely say, almost every case on appeal to a court of last resort could be decided just as easily, legally speaking, for the plaintiff as for the defendant (cf. § 42).* It is hard to get around this dilemma. If the proper outcome of the case were not really a matter of doubt, how is it that an honest, competent judge in the court of first instance could decide it 4'incorrectly"? Again, if the outcome of the case were not really in doubt, how is it that an honest, competent attorney could burden his client with the time and expense of an appeal if the trial court has rendered a "correct" judgment? The very fact that there is an appeal usually proves that doubts exist among professionals, unless the attorney is using the appeal merely as a dilatory tactic. While a party's bad faith or complications in the underlying facts may be behind a case's going to trial, it rarely is what lies behind an appeal.
Again in both countries, the opinion giving reasons for the judgment is almost like a lawyer's oral argument in form. The outcome, known to the opinion-writer before drafting the opinion, is presented as simply inevitable, whatever doubts the panel may have had in arriving at it. We know nothing for sure about opinions except that the actual decision came first; that opinions are justifications of those decisions; and, since judges are lawyers too, that the decisions were probably influenced to some extent by the legal considerations the opinions discuss (although they may not have been). So much for what we do not know. On the positive side, we also know another important thing: Since the judge goes on being a lawyer, and a learned one at that, the opinion he writes in this case will henceforth influence him in similar future cases,1 and the approach he takes now will give us some leverage
*   References in Llewellyn's text to section numbers indicate other sections of the work.   In these excerpts, citations to cases and other materials have generally been changed to conform with A Uniform System of Citation (14th ed. 1986).  [Editor's Note] 1. Cf. B. Cardozo, The Nature of the Judicial Process 21-23 (1921): In the life of the mind as in life elsewhere, there is a tendency toward the reproduction of kind.  Every judgment has a generative power.  It begets in its own image.   Every precedent, in the words of Redlich, has a "directive force for future cases of the same or similar nature."   Until the sentence was pronounced, it was as yet in equilibrium.   Its form and content were uncertain.
1988]
LLEWELLYN
993
for holding him to the same path in the future. We Common Lawyers say, of course, that the judge must follow precedent (except for borderline cases), while in Germany "following" precedent (like it or not) has settled in only as what courts do in practice. But in both countries a precedent is often, surprisingly, not followed—this happens, even if we are not consciously aware of it. Indeed, on very rare occasions, courts openly confess that they are departing from an earlier decision. Qui bene distinguit, bene decernit*
Section 8a The Status of Positive Jurisprudence
Thus, it looks as if we have reached a dead end: Applying deductive reasoning to legal rules apparently will not let us say anything certain about how a lawsuit will turn out when the law is in doubt. Furthermore, opinions apparently will not permit us to say anything certain about the process actually followed in reaching a decision, and least of all about how a later case will be decided. Thus, practical jurisprudence would appear to be at an end. But this is obviously not the case. Practical jurisprudence clearly exists, at least in embryo, as a practical skill, from which something scientific may one day emerge. Even in our current state of knowledge, it is clear that both of the above propositions are exaggerated. Yet each one is just as true as its opposite—and each is directly useful by making us aware of how wrong-headed its opposite is.  Much is accomplished just by that.
What more can one aim for? That is as yet unclear. A natural science of the law has not yet progressed much beyond the initial skepti-
Any one of many principles might lay hold of it and shape it. Once declared, it is a new stock of descent. It is charged with vital power. It is the source from which new principles or norms may spring to shape sentences thereafter. If we seek the psychological basis of this tendency, we shall find it, I suppose, in habit. Whatever its psychological basis, it is one of the living forces of our law. Not all the progeny of principles begotten of a judgment survive, however, to maturity. Those that cannot prove their worth and strength by the test of experience are sacrificed mercilessly and thrown into the void. The common law does not work from pre-established truths of universal and inflexible validity to conclusions derived from them deductively. Its method is inductive, and it draws its generalizations from particulars. The process has been admirably stated by Munroe Smith: "[T]he method of the lawfinding experts has always been experimental. The rules and principles of case law have never been treated as final truths, but as working hypotheses, continually retested in those great laboratories of the law, the courts of justice. Every new case is an experiment; and if the accepted rule which seems applicable yields a result which is felt to be unjust, the rule is reconsidered. It may not be modified at once, for the attempt to do absolute justice in every single case would make the development and maintenance of general rules impossible; but if a rule continues to work injustice, it will eventually be reformulated. The principles themselves are continually retested . . . ." *   "He who distinguishes well, decides well."  [Translator's Note]
994                            COLUMBIA LA W REVIEW               [Vol. 88:989
cism and the desire for knowledge. But I can propose the following useful course of action, one which can be nicely tested on case law: 1) As far as possible one should first study cases "from the front," and only then ex post facto—that is to say, first analyze the material in existence before the case was decided, to try to see what sort of decision was foreseeable; then one should compare and study the actual a) holding and b) opinion. 2) One should then study and categorize both "expected" and "unexpected" decisions a) by type and b) if possible, by outcome-determining factors. If one finds categories that hold up, covering some cases or even most of them, then at least to that extent one can structure what one knows, what one can predict, and ultimately the way to argue any pending case. After that, one can examine more closely what influence a prior decision had or will have on subsequent cases. Up until now, all this has remained only a practical skill, where a lawyer's store of experience and his knowledgeable "feel" can still provide more information than legal scholarship can. Above all, especially in America, the great Unknown still plays a role: i.e. the skill of the attorneys and the personality of the judge (but here the equation is a hard one to solve). Nevertheless, we are already at a point where science, even if it cannot eliminate experience and "feel," can at least lend them considerable support.
For example, one can set up the following case-types:
a)  a decision stemming from a judge's honest efforts to derive a conclusion from a rule (generally convinced that this is the only permissible way to behave; often linked to the view that law is a closed system of rules);
b)  a decision following not so much as a conclusion from a single rule as from a determinate "legal institution,"* a more or less malleable composite in the judge's mind of hundreds of rules (Usually in such cases, the judge roundly rejects all conceptual innovations, and deduces the result from whether the facts contain any of the hallmarks of the legal "institution.");
c)  a decision in which emotional factors of some kind—ethical, political, socio-political, economic, religious, etc.—are the deciding factor (But even so, judges can almost always come up with an acceptable account of their legal grounds for the decision.);
d)  a decision in which the judges more or less consciously behave in a policy-oriented fashion—possibly even citing the results of avail-
* In The Case Law System in America, Llewellyn uses the words "institutions" (Institute) or "legal institutions" (Rechtsinstitute) in a sense different from ordinary usage. By it he means, as he indicates in this section, something like an identifiable composite of any number of legal rules all bearing on the same general subject. For example, he cites "estoppel" as such an "institution." In this translation, the word "institution" is occasionally set off in quotation marks where it was thought necessary to draw attention to the special sense in which Llewellyn is using the word.   [Translator's Note]
1988]
LLEWELLYN
995
able social-science research—within the leeway bounded by rules and precedents.
We thus have four main types of cases (each of which has a tendency to blur into the next category). There is much they make clear, and, in predicting decisions, they take us from being astrologers to being astronomers.
Section 8b New Perspectives on Legal Uncertainty
Perhaps here is a way to make these ideas more concrete: A judge is a lawyer; he is also a human being. Both traits play a powerful part in all decisions. In legal circles, one tends to regard the judge's "human" side as the emotional, the legally incalculable factor. But now enter the sociologist, to show that man's so-called free will is really subject to constraints. By virtue of environment and upbringing, the ethical values affecting him, the thought patterns and mental images absorbed from his surrounding, a man is conditioned, limited and unconsciously constrained to such a degree that his so-called freedom of action seems little more than mechanical. Since the prevailing illusion is that man does possess free will, the sociologist rightly corrects that view by emphasizing the almost unbelievable regularity and determinacy of human behavior. But in seeing man as a product of his times and his rather narrow social circumstances, in seeing him indeed almost as a fungible commodity swept along by outside factors, the sociologist's eye scans whole regions, mass-scale phenomena lasting years and decades—the economic, religious and other forces that affect human behavior. I do not deny that the judge, as man, can be seen in this light.
In thinking about judges' decisionmaking, however, an exactly opposite illusion prevails: that of absolute certainty. Judges, lawyers and laymen all share the belief that it is possible to fashion a law knowa-ble in advance. The necessary corrective for this is to show the freedom of the judge. As man he is, indeed, constrained. As a man with legal training, who wants to base his decisions on existing rules of law, who wants in fact to be confined by them, and who, besides this, has so internalized the rules and institutions that he could no longer shake them off even if he wished—as lawyer, then, he is still further constrained. But here, the level of our analysis has shifted: what looks to the sociologist like constraint can look to the lawyer like completely free discretion. The sociologist sees the forest, the lawyer sees the trees. Decisionmaking is a matter oi individual cases, individual rules of law, individual decisions. To the lawyer—whose daily work they are— these individual matters loom large. To every specialist, the minute details of his speciality seem enormously significant; and for lawyers—as for all specialists—they are very important. But for just this reason they must liberate themselves from their peculiar professional illusion: that they alone have fashioned something truly permanent, something not
996
COLUMBIA LA W REVIEW               [Vol. 88:989
subject to the forces of eternal change. The legal socialization of the judge, who as man is already highly socialized, provides only a limited degree of predictability in dealing with these all-important particulars. In the law, however, it is precisely the resolution of particular cases that matters. This is not to underestimate the role that large-scale phenomena play. Compared to other social phenomena, the institution of judicial decisionmaking is indeed among the most conservative and inflexible. It is so stable that one can generally effect changes in case outcomes without even changing the verbal formulas employed. Such is its fixity that it almost always impedes other social change. Thus, one can readily pardon a layman for believing that law can be, ought to be, and indeed is, fixed and certain—seen as a whole, the law indeed supports this belief. But lawyers cannot be so readily pardoned for being thrown off by this deceptive image: For their purposes, indeed, it is completely untrue. Lawyers work, and should think, on the level of the virtually unpredictable individual case, at the constantly shifting boundaries of legal growth. In what a lawyer does, he is neither sociologist nor layman, but lawyer. Thus, the question facing him is always: When will the tendency towards stability prevail and when will it not? And if it doesn't, in which direction will the change be pointing? . . .
*****
Section 42
Separate Opinions: The Value of Dissents and Special
Concurrences for Scientific Observation of
the Law
The value of separate opinions1 to the scientific observer of a precedent system is exceptionally great. The judge who writes a separate opinion deals with the same fact situation, based on the same record, same precedents or statutes, and same attorneys' arguments, and has participated in the same panel deliberations, as the author of the court's opinion. His utterances constitute an equally legitimate historical source. No scholar looking at a case afterwards can make the same claim. Indeed, the separate opinion can give a scholar reviewing the case an otherwise unattainable glimpse into the essence of decision-
1. Separate opinions, long and short, are found in [the following cases appearing in the case supplement that is part of the complete book]: Adkins v. Children's Hosp., 261 U.S. 525 (1923); Etting v. Bank of U.S. 24 U.S. (11 Wheat.) 59 (1826); King v. Pauly, 159 Cal. 549, 115 P. 210 (1899); Friend v. Childs Dining Hall, Co. 231 Mass. 65, 120 N.E. 407 (1918); Petterson v. Pattberg, 248 N.Y. 86, 161 N.E. 428 (1928); Cammack v. J.B. Slattery 8c Bros., 241 N.Y. 39, 148 N.E. 781 (1925); People v. Davis, 231 N.Y. 60, 131 N.E. 569 (1921); MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (1916); Davis v. Rosenzweig Realty Co., 192 N.Y. 128, 84 N.E. 943 (1908); Elterman v. Hyman, 192 N.Y. 113, 84 N.E. 937 (1908); Canavan v. City of Mechanicville, 229 N.Y. 473, 128 N.E. 882, aff'g 190 A.D. 252, 180 N.Y.S. 62 (1920).
1988]
LLEWELLYN
997
making and the way it happens, into the nature of judicial thinking.2
An observer will first note that the statements of facts in the separate opinion and the main opinion do not precisely coincide; often they are at loggerheads. Yet when conclusions of law are put in deductive form, the statement of facts is always the central component of the deduction's minor premise.3 A case comes out one way or the other depending on how the fact situation is treated.4 Without needing a separate opinion to prove it, a scholar might already have suspected that the facts of the case undergo reshaping as the decision is being made, but especially as the opinion is being written. He knows that it is impossible for anyone to have objective knowledge of the facts as they really are. Out of one and the same mass of facts, each of us, based on individual experience, decides what "the" facts are.   What one has
2.  The frequency test [published as an Appendix to the complete work] which was performed on cases in volume 155 of the Northeastern Reporter shows dissenting votes being cast 42 times, but dissenting opinions only 14 times, and 5 separate concurrences out of 593 decisions (i.e. 3%, as against 18% in the cases I have reproduced in [the case supplement]). Thus, separate opinions are a relatively infrequent phenomenon. A number of courts show none at all for the period in question. But the significance of separate opinions is not to be measured by these percentages, certainly no more than, for example, the significance of the law in real life could be measured by the ratio between cases that come to litigation and the totality of transactions giving rise to legally enforceable consequences. As are courts to real life, so too is the separate opinion to decisionmaking: It appears only on the margins. But the possibility that it could appear has a strong influence on a high percentage of cases in which it does not appear—i.e. it has a sociological and political effect on how a decision is made.
The value of separate opinions to the scientific observer is still greater. The double or triple vantage point guarantees a new perspective, lends as it were a depth dimension to decisions. Once this is recognized, one attains new insight into even unanimous opinions, i.e. those which one would somehow be viewing with one eye shut were one unaware of separate opinions as a possibility.
3.  It is entirely appropriate, given their premises, for clear-thinking proponents of traditional deductive doctrine to emphasize that the touchstone for the resolution of a case is what the facts are found to be, and to fear any lessening of judges' workloads which removes this function from them. I for my part have a different opinion of the appropriate value to be assigned these factual determinations and the process by which judges make them. I would maintain that the issue here is not so much whether the facts as stated by the court mandate a particular result as whether the legal analysis presented as underlying the result accords or conflicts with the way the facts are stated. In the individual case, the chief merit of having judges decide the facts is that it makes the presence of facts more strongly felt. I attach only a second-order significance to the compulsion that forces judges to revise their initial analysis when necessary to achieve a particular result.
4.  Thus I am totally at a loss to understand the notion of a legal rule's being clear and stable, but its application doubtful and difficult. I can only evaluate rules by their effects. If the effect, i.e. the outcome in the individual case, is doubtful, then to me the rule appears fluctuating, uncertain, indeterminate. That the rule is always phrased the same way is of interest chiefly to explain why the rule is so often erroneously viewed as fixed and certain. Between these two points of view, incidentally, lies more than a mere difference in semantics. Cf. esp. Sauer, Die Frage der Kündigung des belgisch-chinesischen Handelsvertrages von 1865, in I Reichsgerichtspraxis im deutschen Rechtsleben 122, 146 (O. Schreiber ed. 1929).
998
COLUMBIA LA W REVIEW               [Vol. 88:989
learned to regard as important is what one sees, what one most readily notes about a situation. This is what one looks for. This is also what one has a tendency to emphasize, overlooking elements in themselves no less important. This is true of anyone's observation of any fact situation. It is most certainly true when it is a matter of classifying facts for purposes of description and further use. For lawyers in particular, such classifying is a tacit precondition for handling any legal dispute, for understanding fact situations. A lawyer has no wish to deal with isolated facts, with the Unique. He wants to deal with facts as instances of fact categories. Within the raw factual matter, he seeks out the few "essential" facts, those which are of legal significance because they fit into a legal fact category, thus providing a handle for "applying" a legal rule. We know this, we accept this, and call it good. But we are apt to slip into the belief that not only is there generally just one possible way to classify facts, but also that the particular classification made in a specific lawsuit has something necessary, something foreordained about it.
This is the reason the legal sociologist, like everyone, regards battles over "correct" legal doctrine with interest, but without being much engaged by them. To the sociologist it is clear—if the battles are worth paying attention to at all—that a fact situation admits of more than one of the constructions espoused; that each way of construing the facts will contain a degree of violence to either the fact situation or the classifying category. For the facts will typically require the making of some "adjustments" to a category which, before the court came to construe these facts, was not quite applicable to them. There is a slight shifting of either the facts or the category. And neither competing interpretation is "right" or "wrong." Rather, the interpretation either does or does not further a particular purpose, the interpreter tacitly choosing among various possible purposes (such as practicality of the solution in the real world, maintaining balance in the legal system, etc.)
We know too that a raw fact situation cannot be classified without shunting the bulk of the facts off to one side. Even this glossing over of certain "legally inessential" facts can easily come to seem utterly predictable and inevitable. And if, when studying court cases, one generally gets to see only a single, officially presented statement of the facts, if one takes this official statement as the basis for one's knowledge and criticism of the case—as for the most part we are bound to do, given how cumbersome it is to review court records—the "application" of the legal rule will seem deceptively simple. In the very determination of the facts, artful—even artificial—conclusions and deductions have been drawn. But when one sees a second judge at work, when at a significant point5 in his opinion facts emerge that were completely glossed over in
5. Cf. MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (1916); Adkins v. Children's Hosp., 261 U.S. 525 (1923); Petterson v. Pattberg, 248 N.Y. 86, 161 N.E. 428 (1928); Davis v. Rosenzweig Realty Co., 192 N.Y. 128, 84 N.E. 943 (1908).
1988]
LLEWELLYN
999
the majority's version, we cannot avoid realizing that everything is not quite as simple as it might first have seemed.
What is most obviously and patently true of the statement of facts is also true of the applicable rule of law.6 Even where statements of facts happen to agree, the legal analyses made of them diverge. Where a dissent is written, the differing analyses would indeed lead to differing outcomes. It is here that one realizes that, if decisionmaking is viewed as a matter of pure deduction from pre-existing, clearly defined rules of law, there is less to be said for the predictability of decisions than one might readily think.7 Yet even where the separate opinion is in the form of a concurrence, one inevitably reaches the same conclusion. For if the same fact situation can be analyzed by learned judges in a variety of ways, it is only a matter of chance that the analyses eventuate in the same result, if the legal analysis is really the decisive factor in the decision. It is hard to avoid concluding that, to a large extent, legal analysis is not a means for coming to a conclusion, but rather a means for justifying—justifying as lawyer, justifying as craftsman—a conclusion already reached on other grounds (cf. § 8).8 On this view of things, the separate concurrence assumes a level of significance beyond even a dissent. For here we can observe a majority of the panel, despite all their differences over the law, nonetheless reaching the same conclusion from the same fact situation. We are then forced to ask ourselves how this can be. In English law this is even more readily obvious than in American. In American law, separate concurrences are found only in cases where a judge is simply unable to approve of the majority's opinion.9 Owing to time pressures, or out of politeness, or to further the course of the panel's work over the long run, a judge frequently acquiesces in an analysis which in reality does not satisfy him.10 But where, as in Eng-
6.  Cf. Cammack v. J.B. Slattery & Bros., 241 N.Y. 39, 148 N.E. 781 (1925); People v. Davis, 231 N.Y. 60, 131 N.E. 569 (1921); Davis v. Rosenzweig Realty Co., 192 N.Y. 128, 84 N.E. 943 (1908); Friend v. Childs Dining Hall Co., 231 Mass. 65, 120 N.E. 407 (1918).
7.  What was said (§ 8b, end) about the "virtually unpredictable individual case" is also to be understood in this way, and is to be supplemented by what is said below about legal certainty.
8.  Naturally, this is neither to deny nor to underestimate the actual perceptible effects (frequently alluded to herein) of the traditional modes of thought which legal concepts and legal rules compel lawyers to adopt. Cf. the "perspectives," § 8b. It is worth recognizing that to speak of the process of decisionmaking can really only be misleading. Any ora-dimensional description (or prescription) will always be false (or fail). There is a variety of processes. All of them can be understood (and, in my opinion, also can be kept under control).
9.  Cf. Canavan v. City of Mechanicville, 229 N.Y. 473, 128 N.E. 882, aff'g 190 A.D. 252, 180 N.Y.S. 62 (1920); King v. Pauly, 159 Cal. 549, 115 P.210 (1911); Elterman v. Hyman, 192 N.Y. 113, 84 N.E. 937 (1908).
10.  A high-court judge once said to me: "Writing a separate opinion too often means losing the weight of one's vote on the panel." Another judge said: "One should not squander the ammunition of dissent."
1000                             COLUMBIA LA W REVIEW                [Vol. 88:989
land, all panel members typically express their own opinion individually, one finds that these opinions nearly always diverge from each other to some extent (often sharply), both as to which facts are emphasized and how these facts are construed. At the same time, one notes that opinions dissenting over the ultimate outcome are rather rare. In every appellate case, in every legal system, one sees the same phenomenon when a lower court's judgment is being affirmed. (Dissenting opinions occur chiefly when a judgment is being reversed.) It makes little difference that the judges are at different jurisdictional levels.
All of this leads one to ask: What does it mean to reach decisions "deductively"?11 What meaning should be ascribed to an opinion? What is the essence of decisionmaking? What is meant by "legal certainty"? This in turn raises the question: How can we know what legal certainty is before finding out how it is possible for four judges, putting four fundamentally different constructions on one and the same fact situation, still to come to the same conclusion? Can we make any progress in our understanding of legal certainty before we have even looked at how this happens? Or put another way: Why are differences over outcome so relatively infrequent in Anglo-American law? To paraphrase the earlier question, what is it that makes different judges, apparently despite their divergent analyses, generally still reach the same result?
Section 43
Separate Opinions: Their Legal, Political
and Social Value
Purely from the perspective of law's role in society, one great value of a separate opinion is precisely that it paves the way for new developments in the law (being in this respect like well-considered dictum, only more so).1 One virtue of a precedent system is that, where appropriate,
11. According to Georges Cornil, 1 Acta Acad. U. Jurisp. Comp. 110 (1928), the deductive explanation of cases "no longer misleads anyone." I do not share his optimism. I too believe that much in the process is in fact to be explained as deduction—at least where a fact situation is on all fours with a previously decided case or had explicitly been taken into consideration when a norm was established. And very similar to deduction is the process depicted in §§ 53 and 54.
1. The sort of separate opinions that are valuable in this way are those which seek to hasten on in advance of old law, but have yet to win over a majority of the panel. Cf. Cammack v. J.B. Slattery & Bros., 241 N.Y. 39, 148 N.E. 781 (1925) (Cardozo, J.). Such an opinion is very often prophetic. (But one may probably regard the Canavan case, 229 N.Y. 473, 128 N.E. 882, aff'g 190 A.D. 252, 180 N.Y.S. 62 (1920), as over and done with in New York—as naturally happens, in an individual state, with the majority of dissents.) On the other hand, a separate opinion which is only an echo of an outdated rule will generally not be given any attention by later judges, as these things go, and constitutes no more than an historical monument. MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (1916); Friend v. Childs Dining Hall Co., 231 Mass. 65, 120 N.E. 407 (1918); cf. B. Cardozo, supra § 8 n.l, at 79 ("Even as late as 1905, the decision in Loch-ner v. N.Y., 198 U.S. 45, still spoke in terms untouched by the light of the new spirit. It
1988]
LLEWELLYN
1001
judges are able to make small experiments with legal rules, which they can always revise to take account of new, unforeseen fact situations, until they arrive at an appropriate legal rule (or even a legal "institution").2 Far more grain is brought to the mill when a number of judges can experiment with the same difficult case independently of each other, for this stimulates legal inventiveness. When two asserted rules are in conflict, it will usually involve a situation in which both opinions have something right and something not right about them, and where both disputants have made the same methodological error, namely, framing a single rule so broadly that it encompasses unlike fact situations. The specific cases that actually occur to one judge are correctly decided under his rule. It is just that he has phrased his rule so broadly that the language encompasses the specific cases that occurred to the other judge too, which might have been decided better under the latter's rule. In every conflict, people think in terms of antinomies, of overbroad generalizations, of thesis and antithesis. Happily, though, it is easier to devise a needed synthesis, the third (and, until this point in the process, unrecognized) possible solution, if the antinomies are accentuated and fought over right at the very start, rather than if the judges wait until later to figure out what they are.3 A judge needing to decide a case which poses a novel issue for him gains more insight if he has two previously decided cases to peruse rather than only one.4 This gain in insight is especially evident when the earlier cases the judge
is the dissenting opinion of Justice Holmes, which men will turn to in the future as the beginning of an era. In the instance, it was the voice of a minority. In principle, it has become the voice of a new dispensation, which has written itself into law.") (footnote omitted). In Adkins v. Children's Hosp., 261 U.S. 525 (1923), the Lochner decision indeed came back to life 20 years later. But most probably this is an isolated occurrence, and Holmes' dissent is, most likely, what will endure. And even if the vast majority of his dissenting opinions remain dissenting positions, with only scholarly, historical and literary value, nonetheless Holmes is one of the few judges who owe their very high judicial reputations primarily to their dissenting opinions. These have now even been published. O.W. Holmes, The Dissenting Opinions of Mr. Justice Holmes (A. Lief ed. 1929).
2.  Cf. Legniti v. Mechanics & Metals Nat'l Bank, 230 N.Y. 415, 130 N.E. 597 (1921); Equitable Trust Co. v. Keene, 232 N.Y. 290, 133 N.E. 894 (1922). Cf. also the case of Wildenberger v. Ridgewood Nat'l Bank, 230 N.Y. 425, 130 N.E. 600 (1921).
3.  Nowhere does this come through more clearly than where a subordinate court presents a higher court with more than one opinion. In Clemens Horst Co. v. Biddell Bros., [1912] App.Cas. 18 (1911), the House of Lords practically adopted Kennedy's dissenting opinion in the Court of Appeal as its own. In Canavan v. City of Mechan-icville, 229 N.Y. 473, 128 N.E. 882, afF'g 190 A.D. 252, 180 N.Y.S. 62 (1920), the upper court's experiment goes further, and is better, than the lower court's.
4.  A separate opinion often has indirectly useful effects on the law. And as for its effects in a different state, cf. the problem in Canavan, 229 N.Y. 473, 128 N.E. 882, aff'g 190 A.D. 252, 180 N.Y.S. 62 (1920), in light of the in-depth discussions it contains, as contrasted with a situation where only one opinion exists. To a lesser extent this is also true of Cammack v. J.B. Slattery & Bros., 241 N.Y. 39, 148 N.E. 781 (1925); Friend v. Childs Dining Hall Co., 231 Mass. 65, 120 N.E. 407 (1918); and especially Petterson v. Pattberg, 248 N.Y. 86, 161 N.E. 428 (1928).
1002                          COLUMBIA LA W REVIEW               [Vol. 88:989
reads come from states other than his own, so that he need not regard the majority view taken in those cases as having the force of precedent.5
It certainly ought to be possible to derive similar benefit from comparing an upper-court opinion which reverses and the lower-court opinion reversed. Such is the nature of authority, however, that, as a general rule, a reversed opinion is virtually invisible. This cannot happen with a separate opinion, published together with the opinion of the court.
Of perhaps even greater social and political value, however, is the dissenting opinion.
In fact, separate opinions provide what those who use the phrase "public diplomacy" are seeking. Were the public ever to harbor a suspicion that courts were not acting aboveboard, the dissenting opinion guarantees the public that any bending of the law will see the light of day. And not just that. The dissenting opinion also guarantees the public that judges are on the job, that in the chambers where a panel's deliberations take place, judges join battle over the law, each judge feeling individually responsible for the panel's decisions.
It is not necessary for there to be such a degree of openness, either to make judges behave conscientiously or to evoke public confidence in the courts. Through its work, an upright judiciary will secure this confidence in any event. But the high regard enjoyed by American judges, particularly upper-court judges, and the central position which courts continue to occupy in our constitutional order, seem to show that judges have no need whatsoever to present a united front to the outside world in order for the law and the courts to be held in high esteem.  To the
5. Extremely interesting in this connection is how a dissent can occasionally work to strengthen the precedential force of a decision. In Boiko 8c Co. v. Atlantic Woolen Mills, Inc., 195 A.D. 207, 186 N.Y.S. 624 (1921), New York's intermediate appellate court (First Department) had let stand a complaint stating a cause of action seeking recovery of the purchase price (as opposed to damages from refusal to take delivery) based on a contract for sale and delivery, although after the goods had been appropriated to the contract, no delivery took place. In Berkshire Cotton Mfg. v. Cohen, 204 A.D. 397, 198 N.Y.S. 240 (1923), on a very similar set of facts, the same Department rejected an action for the purchase price, opining that if the Boiko court had considered the fact that the action had been brought in the form "goods sold and delivered," the court would certainly have dismissed the complaint in accord with its earlier decisions. In this way Boiko's innovation was to have been done away with. But the Court of Appeals, the highest appellate court, reversed, letting the complaint stand, citing Boiko. 236 N.Y. 364, 140 N.E. 726 (1923). It noted that one judge on the Boiko panel had written a dissent expressly focusing on the form of tjie complaint. One would have to assume, then, that the members of the Boiko panel could not have failed to consider this argument. Thus, the attempt by the Appellate Division in Berkshire Cotton to distinguish Boiko failed. (In passing, the Court of Appeals went on to concede that it had indeed affirmed the result in Bready v. Wechsler Co., 200 A.D. 78, 192 N.Y.S. 660 (1922) (a case arising in the interval between Boiko and Berkshire Cotton), wherein the intermediate court had enunciated the rule, now sought to be relied on in Berkshire Cotton, which departed from Boiko. But the ratio decidendi in Bready had had nothing to do with the form of the complaint.)
1988]
LLEWELLYN
1003
extent that separate concurrences or dissents actually play a discernible role in shaping the public's attitude towards the law in civil or criminal cases, they probably serve, as I said, to elevate the judiciary and its esteem considerably. One possible exception might have to be conceded: cases in which a sharply divided panel holds a statute unconstitutional. I call this a possible exception, not an absolutely certain one, because even the bitterest opponents of such a decision regularly submit to it. From this very fact alone there is a rise in esteem for the judiciary. I also call it a possible exception because a court must already be held in high regard if the legislature, administrative agencies and the public rapidly reconcile themselves to a 5-4 decision holding a law unconstitutional. But even were this one case really to constitute an exception, it nevertheless says nothing about the general run of cases. All of us agree that in the usual sort of civil or criminal case a majority of the panel, indeed a simple majority, should decide. When the constitutionality of a law is being challenged, the difficulty lies not with the very useful separate opinion, but rather with the fact that the private-law paradigm of a simple majority has simply been carried over unchanged into another area, to the politically significant area of restrictions on legislative power. Were, for example, a two-thirds majority or something like a 7-2 vote required to rebut the presumption in favor of a statute's constitutionality, dissenting opinions could pose no conceivable threat to the esteem enjoyed by the court.
But do not separate opinions evoke in the general public a sense of legal uncertainty? As far as I can tell, they do so only to the extent that the possibility of a decision's being reversed on appeal evokes a similar feeling. We know that the bar feels rather more comforted than not by the role of appellate courts as reviewers and reversers of trial courts. (This, incidentally, is something to think about in connection with legal certainty!) As for the public at large, it is only the rare case that ever becomes a topic of conversation, and in America constitutional cases far more often than any other type. But one must also bear in mind that separate opinions are found only in a small percentage of cases. The benefits they offer to legal scholarship manifest themselves precisely in cases where a separate opinion is most needed. But the fact that separate opinions are always a possibility, plus the fact that they appear infrequently, taking both together, must actually serve to reassure people about legal certainty.
There is another benefit to be borne in mind, which for the most part is still off in the future. If I am correct in my view of the true nature of legal certainty, separate opinions offer legal scholarship a great (and as yet largely unused) means for expanding and securing such certainty (cf. § 52 ff.).
What has already been said about separate opinions is also true, to a somewhat lesser degree, of the practice of naming the author of a panel opinion. Naturally, the provision of such information is not neces-
1004                          COLUMBIA LA W REVIEW               [Vol. 88:989
sary to fortify either the judicial temperament or the judicial sense of duty; one sees this in countries where judgments are issued anonymously. But providing such information surely works to further these ends. A named author must answer for his case not only to the whole panel, but to the Bar as a whole.6 Moreover, to see an author's name right at the beginning of an opinion is both meaningful and extremely helpful, particularly when the body of decided cases has grown large and unwieldy, or when case law has first begun to deal with a new area. A certain judge's opinions may be dubious, his dicta dismissed, his experiments regarded with the greatest skepticism, his utterances construed narrowly. Another judge's opinions may be acute and insightful, his dicta more valuable than many people's decisions, his intuition prophetic, his formulation of rules well-considered and confident. A leading judge thereby achieves greater esteem and, above all, far-reaching influence. The notation of who the author was has repeatedly been decisive for the law's formation in other states or in later times. This naturally is not without its dangers. "It seems to be the prerogative of a lofty mind not only to enlighten by its wisdom, but to enslave by its authority."7 But even the greatest cannot ultimately prevail with their errors, while their names give their happy insights the strength to secure the public good.8
*****
6.  Cf. Cardozo's attitude in his writings. When a judge is elected, especially for a short term of office, he may occasionally have to answer for his authorship to the lay public as well. It should be noted that in States with an elected judiciary, one detects this in court opinions only very rarely. The problem noted here only matters in connection with the election of judges.
7.  Brewster v. Hardeman, Ga. (Dudley) 138, 145 (1831) (Lamar, J.).
8.  The attitude taken towards this question is highly characteristic of how closely lawyers are tied to their national traditions. An American lawyer (even more so a judge) would fear a collapse of all judicial responsibility were signed opinions to be given up. He is unacquainted with the practice, as in Germany, of issuing judicial decisions anonymously; he is scarcely able to conceive of anything of the sort. In Massachusetts over the last few decades, very strong pressure has been felt against dissenting opinions, which is apparently to be ascribed to the Supreme Judicial Court's Chief Justice Rugg. In my calculations, there were no dissenting votes recorded from Massachusetts. (In Wilcox v. Massachusetts, 163 N.E. 251, a 4-3 vote was indicated, with Rugg in the minority. In 266 Mass. 230, 165 N.E. 429 (1929), the opinion was republished, without any indication of how the votes were divided. I am grateful to Professor Maguire of Harvard for pointing this out to me.) Professor Morgan has written to me: "There are some indications that the court's policy is to discourage dissents. A careful reading of the opinions, where no expressed dissent is recorded, often reveals that the court was not unanimous. An opinion rather frequently says that the majority of the court believe or are convinced." In the same context he writes: "If you have been reading Massachusetts decisions for the past few years, you will realize that the court has been woefully weak." He suggests a cause-and-efFect relationship. I have noted these same phenomena and harbor the same suspicion very strongly myself. It is a fact that we do not like having to put up with enforced unanimity towards the outside world.  It is something unknown to us.
But a good battle-lance can be turned the other way around. Many a German lawyer has displayed to me signs of moral indignation at the very thought of individually
1988]
LLEWELLYN
1005
Section 52
Certainty in Case Law;
Rules of Law Do Not Decide Doubtful Cases
The words "legal certainty" seem to evoke in most lawyers' minds an image of simply being able to apply an existing rule of law deductively. We are used to thinking like this, particularly since judicial opinions and legal discourse must always be dressed up in this way so as to be socially acceptable. My claim would be, though, that for the cases which occasion difficulties, this kind of legal certainty never has existed and never will exist (cf. § 8); that to strive for this kind of certainty is a waste of time; that legal certainty in reality consists of something quite different; and that, once we understand what legal certainty consists of, a goodly part of the debate about whether a judge is bound by legal rules or is free when deciding cases will disappear on its own.
The first question involves the nature of legal rules. To make the discussion more concrete, I shall first presuppose a rule always given an invariant wording. A rule of this sort, ultimately, does encompass many cases. This no one doubts. Insofar as these cases were known to the lawmaker before the rule was laid down, and insofar as circumstances have remained unchanged since that time, one can work with the rule deductively. But once the possibility of doubt arises, from that point on the situation is different. The issue is no longer what the existing content of the rule is, but whether the rule will or will not contain the doubtful case. If doubt is really present, the boundary of the rule in relation to the instant case will be unknown until a judicial decision has been rendered. Thus, the task of the judge is to reformulate the rule so that from then on the rule i/ndoubtedly includes the case or undoubtedly excludes it. "To apply the rule" is thus a misnomer. Rather one expands a rule or contracts it. One can only "apply" a rule after first freely
signed opinions, and even more so at the thought of published separate opinions. This happened in one case in a particularly lawyer-like fashion: Two gentlemen, who were active before the German Supreme Court, in fact amiably acknowledged that they knew who had written the various opinions and, in most cases, how the panel had divided up—and that this was very useful to them in their work. But they immediately announced their predictable "It can't be possible" attitude at the thought of an official publication of these attendant circumstances, first because such information was useless, and second because its publication was dangerous.
I am appending these trifling matters from both countries in such detail because a single sharp encounter with an individual's failure of imagination or with mutually inconsistent views held by one person can be more liberating than dozens of opera majora on comparative law. And without a partial liberation from "matters of course" (for me, the inevitability of case law, perhaps) a comparativist approach would be as sterile as a mule. To which another point must then be added: That only in the light shed by a given system's "matters of course" can an outsider go on to attain insight into the true significance of any individual feature of the system (one rule; one "institution"). Thus, wherever I can, I also need to highlight those "matters of course" in American law that German readers will find most astonishing.
1006                          COLUMBIA LA W REVIEW               [Vol. 88:989
choosing to include the instant case within it or to exclude the case from it.
Section 53 The Automatic Expansion of Word Content
I would like to start off with a hypothetical case that no one has any doubt about how to handle, but which no one had in mind when the rule was created. Nonetheless anyone who looked at the case and the rule at the same time would "recognize" that the case must be governed by the rule. Here I maintain that no real "interpretation" of the rule is taking place. Rather, there is an extension, a reformulation, a reading of something new into the rule. For the rule is made up of a series of words, a nexus of linguistic symbols. In itself the symbol is nothing. But the symbol is a means of referring to things or to mental constructs. To the extent that something has demonstrably been referred to by the symbol's original use, the symbol has content right from the start. But, to an indeterminate degree, every symbol also has what may be called latent content: it has an unambiguous and predictable (along with an ambiguous and not fully predictable) capacity for expanding. Once something new and different appears, something not thought of before, it can be felt to fit within existing categories. In this sense, every category in fact has an immanent expansive capacity.1 This expansive capacity is clearly predictable to some extent, but only to some extent. In our hypothetical case, every lawyer would expand the rule the same way. But when he does so, he does so instantaneously; he does not see the expansion. To him, it is as if the new case had always fallen within this category, just as if there had been a conscious intent that the new case should belong in this category. Nevertheless he has, even if unwittingly, expanded the category. Granted, the probability of this expansion was always present. Granted, a lawmaker, when he uses "word-symbols" at all, must approve of such expansion, since it is inevitable.
1. Sociologically immanent: i.e. if one presupposes determinate historical circumstances, existing legal practice, and lawyers trained in a particular tradition and particular doctrinal materials, one can predict with certainty the immediate inclusion of certain fact situations (which had not occurred until then) within particular categories. I simply fail to understand (and so cannot discuss) immanent meanings of words in any other sense. In other words: the foregoing discussion deals only with describing a linguistic symbol's semantic content at a given time, the content to which a lawmaker (or judge) has referred, or possibly can have referred by using the symbol: What a word-symbol signifies can extend beyond that which has already occurred or been conceived of only in a normative sense, not a descriptive sense. Now I have nothing against the idea of setting up norms, in advance, for what has not yet been conceived of. Such an idea is an exceedingly valuable normative tool—but has really very little to do with legal certainty. Legal certainty concerns the world of What Already Exists. Legal certainty means calcu-lability in advance of the actual rendering of a decision. And the data for such a calculation never extend beyond the symbol's or the legal rule's factual content at a given time, plus the factors allowing one to infer its expansion to cover new cases.
1988]
LLEWELLYN
1007
But no lawmaker can intend the inclusion of such a concrete case; the case, by hypothesis, has not yet been thought of. It is in this unconscious process of expansion, I believe, that the key to true legal certainty is to be found.
Section 54 Hidden Expansion in the Doubtful Case
Matters are no different, only more sharply highlighted, when a new case is such that one first must mull over whether to include it within an existing category, or in which existing category to include it. Doubt may arise from the fact that the case is not quite like other cases in the category, or because, while the case has all the characteristics of the category, it also has other features which cast doubt upon the criteria for inclusion in that category. Or on a technical legal analysis, the case may be no different from prior cases of the kind; but social conditions may have fundamentally changed. In both situations, decisional law will need to reformulate the rule. If the new case is brought under an old category, the category thereafter is broader than it had been. But if the new case is excluded, the category has acquired a more determinate boundary where earlier its boundary was uncertain and there was still the possibility of its extending to cover this or a similar case. Whichever way the decision goes, the symbol thus acquires a new content. It refers either to the New along with the Old, or, more narrowly limited, to the Old alone. Here too a judge can render a decision while honestly working to "interpret" the relevant word in the legal rule. For we all, lawyers not least, are mistaken about the nature of language. We regard language as if words were things with fixed content. Precisely because we apply to a new fact situation a well-known and familiar linguistic symbol, we lose the feeling of newness about the case; it seems long familiar to us. The word hides its changed meaning from the speaker. What political innovation cannot be introduced, provided that it is presented behind the mask of a familiar, acceptable linguistic symbol? Although in the law such innovation usually occurs on a smaller scale, this should not obscurď the fact that the process is the same. For the participant, the process is obscured by his using familiar symbols in his mind to represent the new fact situation, to visualize it, to manipulate it in his own thinking. And gradually, almost by necessity, the New is thereby identified with the Old. It is almost automatic for us to analyze and classify a new fact situation using categories with familiar names. We construe the facts. Only when the construction is attacked do we recognize the presence of doubt, of an innovation. And even then, for lawyers, recognition is significantly impeded because they operate with the black-letter notion that the answer must already exist, and that there can be just one answer. Accordingly, the language of legal analysis assumes that the applicability of certain categories is uncontestable:  "Here we are dealing with a power of attorney, a sale,
1008                          COLUMBIA LA W REVIEW               [Vol. 88:989
etc." Once again, awareness of the fact of innovation disappears. But one who works over a line of cases in retrospect recognizes that something new has been created.1 And can it have been old at the very moment it was created?
Thus, I conclude that the resolution of every doubtful or disputed case creates new law and refashions legal rules, even when the judge is aware only of "interpreting" a legal rule and has no desire to do anything but. Yet he has not interpreted', he has placed a case within the rule or outside it.
Therefore, legal certainty in the American sense, i.e. the prediction of the outcome of a lawsuit based on deduction from the existing content of a legal rule, is possible only in those cases where no real controversy should have existed at all. Deductive prediction can occur only where one party does not know the already-existing content of a legal rule. Dubious facts may bring about a lawsuit, i.e. a trial; but this will not be a genuine law suit, involving dispute over the law, even though it is settled by means of litigation. Such litigation is fundamentally different from a dispute over the rules to be applied or over their content.
Section 55
Judges' Operating Technique Guarantees Legal
Certainty Even Under a "Free Law" Theory
In spite of all this, the outcome of a dispute concerning the law is predictable to a truly amazing degree, and for that reason the law is (descriptively) certain. Pound correctly emphasizes that we must focus separately on two basic elements: 1) First, the rules of law, law's existing normative material (understanding the term broadly enough to include not only the rules' existing content, but also all the new content that every lawyer would undoubtedly add to them). 2) Second, the traditional way the law is handled, the operating technique of the trained lawyer as passed down to him (thus practice, not norm; way of acting, not verbal formula). In discussions of legal certainty, not enough emphasis is placed on this second element. Its great importance becomes clear when one thinks of the different results a layman and a trained lawyer derive from the identical legal rules, even if one hands the layman every rule that the lawyer intends to use in solving the problem.  Its impor-
1. Cf. J. Hedemann, Reichsgericht u. Wirtschaftsrecht passim (1929); H. Maine, Early History of Institutions 229 (1888) ("New combinations of circumstance are constantly arising, but in the first instance they are exclusively interpreted according to old legal ideas. A little later lawyers admit that the old ideas are not quite what they were before the new circumstances arose.").
It is a fundamental principle of geology that large-scale changes of the past are traceable to causes still at work today. Here lies, I think, a stimulating thought for legal scholars. Stimulation of a different sort is offered by analytic geometry: that the study of two neighboring points may throw more light on the whole than directly studying the whole.
1988]
LLEWELLYN
1009
tance is further apparent when one considers the phenomenon of the separate opinion in American and English law (§42, particularly towards the end). Naturally, the difference between the answers obtained by lawyers and laymen results in part from the fact that the lawyer's experience has already given these rules content, as indicated above, a content which will be very different from that which an untrained layman will read into the words. But no less important is the trained, tradition-determined manner of handling the material. This method is what enables a lawyer to solve the truly new case—to come up with a sure and skillful solution, one pretty much in harmony with those of other lawyers, based on legal rules which really do not cover the case. He has learnt to derive basic guidelines from legal rules, guidelines which do not enable him to derive the solution of the new case from old law, but which will bring the solution of the new case into harmony with the essence and spirit of existing law. It is precisely this art of solving the new case—pretty much the same, and deeply rooted, among lawyers everywhere—that ensures the continuity of the judicial decisions of a particular time with prior law.1 I might even venture to assert: Even if a trained lawyer believed himself totally free, only very rarely would he be able to loosen himself from these constraints, and then only to a limited extent. Note, for example, that when the natural-law philosopher proposes his ideal solutions, he again and again reverts to the positive law of his homeland. This assertion is true even more so of a judge, who feels himself not to be an omnipotent legislator, but at most an interstitial one, whose duty it is to adapt the legal system to current needs. What earlier was said in § 8b is again applicable here: The freest judge's space for movement continues to grow smaller, and must remain so.2 The constraints and the socialization resulting from his membership in society and from his legal training guarantee the continuity of decisions, the continuity of legal norms, and the predictability of the "freest" decisionmaking. To recognize this also means that so-called Free Law is free only within this small space, and in no way threatens a true legal certainty.3 But at the same time one also must recall the conclusion of the preceding paragraphs—that so-called de-
1.  In § 53, n.l, supra, I spoke of a sociologically immanent and predictable capacity of linguistic symbols or legal rules to expand. Here I wish to speak of a similar quality in a legal system. Only in this way, in my opinion, is the phenomenon discussed at the end of § 42 to be explained. Several legal rules, several legal "institutions" often cover a case; what is given is only the predictable outcome, for the plaintiff or the defendant—which clearly emerges from the trends prevailing at the moment. Cf. also B. Cardozo, supra §8n. 1.
2.  To this effect in our country, more than anyone else, the incomparable Holmes, in a wide variety of writings and opinions.
3.  Nonetheless, it has long since become clear that the exertions of the "Free Law" proponents even in the heat of battle were directed not against a true legal certainty or against the continuity of decisionmaking, but only against a "literalness" of interpretation, which corresponded neither to the nature of language nor to the best legal tradition.
1010                          COLUMBIA LA W REVIEW               [Vol. 88:989
ductively certain law is really nothing of the sort, but permits freedom of movement within a small space. The only noteworthy contribution of the Free Law movement is then to advocate that this freedom of movement be consciously understood and not exercised blindly. It will presently become clear (§ 56) why in my view the conscious freedom of a trained lawyer means a greater real certainty than a blindly literalist attitude makes possible (in all Europe I have heard of only one Bonjuge Mániou, and in my country of only one, on a high-court bench).4
Section 56 Judicial Intuition: Fact-Guided Decision
Pound adds ''intuition'' as a third factor in decisionmaking. He seems to mean that indescribable something which, above and beyond the available legal materials and the lawyering art, perhaps in spite of them, leads judges to sound results. In this context, he notes something quite true about our law: The result is generally correct, whether or not the judge could come up with good grounds for doing so. It is also true that, overall, judges have done their part better than academic lawyers have done theirs. This is not to say that judges have been even half as good systematizers as the scholars, but only that they have perceived and satisfied the needs of the law as a living element of society, even when their ability to fashion legal grounds for decisions has lagged behind. Meanwhile scholars have gone on constructing theories divorced from real life.
Certain aspects of this third element can be defined more precisely, particularly with regard to the new meanings words take on in changed circumstances. If one observes a new fact situation and is sensitive to its real-life meaning, then there is a sudden and (so to speak) ex post facto change in the meaning of one's prior life experience in that area, and thus a change of content in the words used to describe and regulate the area. I cannot explain this process; I only record its existence: The New illuminates and at the same time changes the Old. The ''intuition'' in this process lies in the judge's unconsciously using his prior experience and his sensitivity to the meaning of new fact situations. Where a judge fails to reinterpret the law soundly, it is almost always because he lacks this sensitivity. He sees the New well enough with his eyes, but fails to see what it means. This lack of sensitivity is almost entirely due to a lack of those experiences that might have permitted him to recognize what the new facts mean. A judge's intuition extends only as far as his experience and sensitivity.
4. Naturally I do not deny that the deeply-rooted belief that one is required to decide purely deductively has an influence on decisions. I am only asserting this: If a new case is before him, the judge must move, one way or another. To expand an old legal rule or legal concept is just as legally correct, just as much grounded in tradition, as its opposite. Thus, the judge cannot but decide freely, whether he "freely" decides or not.
1988]
LLEWELLYN
1011
The other side of intuition comes from the same source, but it is different in its effects. Here too it is a matter of judges' insight into new fact situations and their meaning, being generally referred to as the "sense of justice" in the individual case. I would like to call this "fact-guided decision": It presupposes that a legal formula to resolve the problem is not yet available when the decision is made. If a decision is rendered in this way, one sees the following sequence of events: a) understanding the facts; b) deciding on the basis of the facts ("the outcome must be this way"); c) searching for a legal justification; d) writing an opinion which contains a justification, a construing of rights. In such cases, the construction is purely a means to an already determined end. One collects the legal rules needed for a justification, twisting and turning them until they seem to yield the result already decided upon. It is impossible to know how often this process occurs in judicial decisionmaking. But this much is clear: If a judge is "tempted" even once to let the facts guide his decision, he will see how unexpectedly fertile legal concepts and ideas are. Furthermore, when judges engage in this "construing with a purpose," the borderline between what they do unconsciously and what they do with complete awareness is as fluid as it is wide. Whether conscious or not, it is all the same, because the process is not one that involves deductive legal certainty. The sort of legal certainty that does exist is to be found in the predictable reaction of a trained lawyer, or of a judge with a certain degree of life experience, to what the ultimate disposition of a fact situation should be. It is extremely important to emphasize that this does not involve any interpretation of legal rules. In other words, the legal rules do not lay down any limits within which a judge moves. Rather, they set down guidelines from which a judge proceeds towards a decision. They indicate the experiential basis and the approved direction for developing norms, and thus the foundations of existing law. From these one may diverge only a short distance in any particular case. They reveal the established complex with whose overall thrust the new decision will agree and (descriptively as well as normatively) must agree. This divergence-not-too-far, this agreement of direction is, in fact, guaranteed through legal training and the judicial conscience. Nonetheless, a legal rule functions not as a closed space within which one remains, but rather as a bough whose branches are growing; in short, as a guideline and not as a starting premise; not as inflexible iron armor which constrains or even forbids growth, but as a skeleton which supports and conditions growth, and even promotes and in some particulars liberates it.1
1. Descriptively speaking, I clearly do not regard a legal rule as something hard and fast. Rather, with every legal rule there are three elements at play: a core of completely fixed content, made up of what has already existed and what is already concretely intended; next, a fluctuating borderline area of possible expansions, which is determined in part by the way the rule happens to read; and, finally, a "trend" affecting these possibilities, which is determined in part by the fixed core, in part by neighboring legal
1012                          COLUMBIA LA W REVIEW               [Vol. 88:989
I freely concede that the type of legal training, as well as the type of legal system, can also affect the extent of growth. In a case law system, the room for creativity is somewhat expanded by very reason of the freer linguistic form of the rules. But, by contrast, where judges adhere fairly strictly to a doctrine that the rules of law laid down by the legislature are completely controlling (a doctrine that even we follow), a court will constantly hark back to the exact words of the rules. Thus, growth will come by expanding the meaning of linguistic symbols rather than by formulating new rules of law. On the other hand, under a case law system, the judge's sense of responsibility may have some restraining effect, since he is formulating a legal rule that must account for the past as much as provide a standard for the future. By and large I believe that the course of decisionmaking proves that the scope and pace of growth varies somewhat from time to time and place to place, both within a particular system and among different systems. But even where the scope for growth is smallest, this process of expansion and expansion is so readily apparent that it must be recognized in order to understand the nature of adjudication of any type.
*****
Section 58 Legal Certainty for Laymen
So much for lawyers. Far more important to me seems legal certainty for the man in the street. Here, knowing what decisionmaking really is about provides the only way of getting near the most feasible and satisfactory solution. For one should never forget that the average layman does not conform his actions to legal norms but to social norms.1 The latter often closely resemble legal norms, but seldom do they coincide. It is simply not possible to live by legal norms. The social norms governing laymen's behavior, however, do not stop undergoing a process of transformation and change simply because lawyers have laid down legal rules. Legal certainty in one's business dealings, legal certainty for laymen does not consist in a lawyer's ability to predict the outcome of a lawsuit or the legal consequences of some action.  Such
rules, in part by extra-legal conditions, and in part by the needs of those affected by the law at a given time. This description is clearly appropriate for case law rules, which for the most part do not even have a fixed wording, yet also for legal rules with fixed wordings, even if my treatment of linguistic symbols is only partly right. Only that many rules with fixed wordings view their borderline area as somewhat confined. The borderline area, however, is never "decreasingly small." And every expansion of the fixed core into what was previously borderline territory pushes the fluctuating border that much further out. 1. And if he adapts to "legal norms," it is mostly popular legal norms, which crudely and often falsely imitate true legal norms. For this reason, incidentally, in the old doctrine of customary law there is always a three-part ambiguity regarding the opinio necessitatis: Is it a necessity under lay law, lay custom or lawyer-made law? Cf. H. Isay, Rechtsnorm und Entscheidung 236 (1929).
1988]
LLEWELLYN
1013
predictability is much less significant to the layman: To him it only means legal certainty with respect to a particular lawsuit. For him, by far the more important type of legal certainty consists in knowing that some dealing of his would be treated—if matters ever came to litigation—in a manner that a reasonable man (i.e. a reasonable layman in the circumstances) might have anticipated, had he thought about the then-unforeseen dispute at the start. Insofar as this is not possible, because no layman could have so foreseen,2 what we have is not a normal case of real-life legal certainty, but a case that the social physician must treat as pathological. The cases raising important issues of real-life legal certainty, though, are those where the question is whether the judge will interpret the existing formulas with or without insight into the altered or new circumstances. And here one may safely claim: So-called legal certainty on the old lawyers' model means nothing but disillusionment for the man in the street with interests at stake in a lawsuit. Indeed, in cases involving a gradually expanding legal rule (discussed supra, §§ 53, 54), the very fact of its expansion clearly shows that a failure to expand the rule—i.e. chaining down the living present to the fiction of deductive legal certainty—would be unendurable. Only steady, unnoticed change in the content of legal rules—which comes down to jettisoning orthodox theory—gives that theory continued viability.
This line of argument can be summed up roughly as follows: All words (that is, linguistic symbols) and all rules composed of words continuously change meaning as new conditions emerge. A layman involved in a lawsuit—that is, a person interested in its outcome— fashions his norms, and consequently his expectations about the law, directly from the probabilities in real life, which change as a function of changing conditions. Legal rules provide certainty in the affairs of people whose interests are affected by the law if, in a lawsuit, they yield a result that accords with their real-life norms. Thus, the most important legal certainty for those whose interests the law affects depends on whether the judge can make the direction and degree of semantic change in a legal rule (or a verbal symbol used by the rule) keep up with the corresponding change in the real-life situation. But if the change sanctioned by the judge keeps up more or less, but not quite, then one speaks of the law's mild conservatism. If the change on the judge's part is noticeably not keeping up, then one speaks of a crisis in decisionmaking. And, finally, if the change on the judge's part is keeping up perfectly, neither judge nor layman realizes any change has occurred; that is left to later scholarly observation—and legal certainty prevails. It can prevail only through change.
We should obviously not overlook those situations where the law-
2. Perhaps because usage is still in flux, as so often, or because two interest-groups' fixed but different usages are at odds and as yet unsettled, or because the case is the sort that the judge must decide somewhat arbitrarily because of equally divided equities. This requires only norms for judicial decision, not legal norms for real life.
1014
COLUMBIA LA W REVIEW               [Vol. 88:989
yer's notion of legal certainty is also decisive for the layman, due to the practice of consulting a lawyer before concluding some transaction— where, for example, a piece of real property, setting up a corporation, etc. is involved. Thus, this chiefly involves matters of long-term rights, or the treatment of relationships at once continuing and impersonal (increasingly important nowadays). To the extent it is at all proper to speak here of laymen's legal certainty, it exists only as a reflection of lawyers' legal certainty. A person who consults no lawyer, or a bad one, must bear the consequences of his own stupidity. But it is precisely in such cases that certain segments of the Bar acquire insight into new social conditions, leading them to come up with new analyses and new interpretations of rules of law. Above all, it is the client-counseling lawyers who pave the way to understanding what adjudication is really all about, and whose work is in harmony with its true nature. It is time for legal theory to draw a lesson from their activity.
*****
Section 63
Black-Letter Law Versus my Approach: Towards a
Sociology of Doctrine
It will be easier to see where my views lead, and to direct more pointed criticism at them sooner, if one looks at their relation to the formal doctrines of black-letter law. Though I am no adherent of these doctrines, I might be permitted to venture a few steps in this direction, since, whatever we may say, a fundamentally sociological approach to the law is already threatening to make drastic inroads on standard legal doctrine.
I should probably begin by saying that the prevailing (though not unanimous) view is that legal doctrines have a directive, not merely a descriptive, function. One who criticizes the law's formal written doctrines for being untrue to life, however, is expressing a belief that doctrine is out of touch with how decisionmaking actually happens. I might express it like this: Seen from an empirical perspective, even legal doctrines are based, in the first instance, on a sociology of law. By this I mean not just that the operative legal rules can actually be found in law books, but that the content which the doctrines say these rules possess is drawn in the first instance from observation of actual decisionmaking, rather than from how legal scholars respond to the words of the rule.
Consequently, every legal rule proposed by those who uphold black-letter doctrine contains an implicit assertion—insofar as nothing to the contrary is explicitly emphasized—that court decisions in fact follow the stated rule, that the rule is thus sociologically valid. This implicit assertion becomes clearer when we note that most of the Ought-statements take the verbal form of Is-statements—even though the coincidence of judicial behavior with rule is so completely assumed that
1988]
LLEWELLYN
1015
judicial behavior is often not even separately discussed. It is no answer to this objection to say that Ought-statements do not claim to be Is-statements, that their only claim is normative. In fact, they almost always claim to be the prevailing norm; and 95% of the time this means a norm which is in fact being applied in decisionmaking, not only one which ought to be applied.
Here we encounter the first dramatic change that would have to be made in the typical black-letter discussion if the views expressed here were someday to be accepted: Never again could one tacitly assume the descriptive accuracy of a stated legal rule. Rather, one would always have to investigate how courts in fact proceed, one would always have to make explicit whether and to what extent an Ought-statement regarding a particular legal topic differed from an Is-statement about the corresponding court practice. Both should always be expressly recognized as complementary but distinct sorts of statements. Sometimes they are recognized as such, at times even frequently; but often they are not. A separate study of Is-statements (which may or may not coincide with the parallel Ought-statements!) is not an indispensable part of the orthodox black-letter approach of today. Rather, it only sometimes gets added as an afterthought.
Such an approach would have still further consequences for traditional legal doctrine. Were one to make a special study of the way black-letter scholars state legal rules—or, if a statute already supplies the language of the rule, the way black-letter scholars give content to this verbal formula—one would usually find that, even though this process is based on empirical observation, it only selectively reflects prior observation. However conscious or careful they are about doing it, black-letter lawyers cull only a few cases from among the relevant ones decided, plus a few of those discussed in the literature or hypothesized, perhaps first giving a few they thought up themselves. Rarely will they even have all the relevant cases in front of them; and should they have them all, they still regularly omit a number. Indeed, even when they do not have all the cases, they exclude many of the cases looked at. What remains becomes the core, the framework of their legal rule, or what they would maintain is its "correct'' content. To "test" their viewpoint, they will then typically hypothesize a few more cases, including or excluding them from the scope of their rule: What they are doing is making normative rules for specific cases before they occur. I believe I am saying no more than what empirical observation would reveal when I state: From that point on, black-letter scholars emphasize those decisions that harmonize with their version of the rule, revealing a marked tendency to look down the wrong end of the telescope in assessing the number and significance of contrary decisions. What they have done is somehow to forget both that their own procedure for framing legal rules has its basis in description, and that their rule has simplified what it meant to describe.   Taking a fundamentally sociological approach, though,
1016                          COLUMBIA LAW REVIEW               [Vol. 88:989
would make it much easier to avoid these pitfalls: A scholar would always have before him not only the observed cases that formed the original basis for the rule, but also the cases that had been observed since the rule was first formulated.
But the blade cuts deeper still. Were it acknowledged that rules can be framed only by accepting some cases and excluding others, we would have to defend our selection of cases to ourselves and to the legal world qua selection, and not simply assume that the selection is justified from the start. And if there is anything to be said for my views on the semantic content of word-symbols (§ 53) or on the limits to what "the" lawmaker can possibly have meant in a concrete instance, the defense of a selection of cases would chiefly have to revolve around the effects the selection might have on the outcome of actual cases. This would have to be done continually, as new concrete cases occurred, since norms, being abstract to some degree and framed before disputed cases ever occur, can never be assumed to have a desirable content unless they are subjected to constant testing.
Given the time constraints, this would to some degree mar the aesthetic side of any ideal "system," as it doubtless would mar elegantia juris too. But here a sociologist of law might wish to add a different kind of remark. Ideal systems are beautiful; elegance, an aesthetic criterion, affects the shape the law takes. Both bring the artist happiness; the work sings the master's praise. But when one constructs a system of ethics or political economy, no one suffers if they turn out to be untenable; the master reaps his joy even then. In law, however, elegantia juris comes only at the expense of litigants' interests. A sociologist can pay for his aesthetic sensibilities with his own sweat; a lawyer's are paid for with other people's lives, property, and fortunes.
My view of decisionmaking and legal rules ought to lead, then, to two important changes in the most common operating technique of the present day. As for its descriptive aspect, a rule would be deemed valid only to the extent it actually reflected the data, and its validity as description would be limited to how well it reproduced given data. But if we were instead to give a formal, doctrinal critique of some actual decision or the proposed decision of some hypothetical case, how well the decision logically accorded either with prior decisions or with the supposed meaning of some written rule would be only one criterion to judge the decision by.1  The other would be the social and political effects the
1. Simply because for all cases whose outcome is in doubt there will be more than one logical conclusion that can be squared with the norms in force. I am entirely unprepared to concede that, despite whatever changes the law is undergoing, a judge deciding a novel case is supposed to view the point to which the law had developed when this new case arose as though it were, for the moment, an inflexible Given, and dispositive in its wording. Rather, it is my opinion that, at any moment, one should recognize the presence in the law of competing trends relevant to the particular case at issue. It is an attorney's job to generalize one trend so that its formulation convincingly includes the
1988]
LLEWELLYN
1017
decision would have. This would necessarily place some restrictions on how far an individual scholar would be able to judge the merits of the decision as legal doctrine, because he could not arrive at a satisfactory judgment based solely on what his own life experiences happen to have been. To have any kind of basis forjudging the decision, he would first have to explore its effects on persons with a stake in the legal outcome. We lawyers are hardly accustomed to such trenchings on our domain. But I find it hard to see how else we can defend our value-judgments about particular legal disputes to interested parties or to society at large. Even where the question is whether a particular decision logically follows from earlier decisions or legal rules, my views would lead to a change in operating technique. For the legal rule used to measure the new decision could no longer be regarded as static, but rather as constantly undergoing a change of content. Thus it would hardly settle a problem even were it clear that the proposed way of resolving it was not contained in the prior content of the legal rule. Further study would be needed to determine whether it was possible to enlarge the rule's literal meaning to cover the new case too, or whether it was necessary or legitimate to reformulate the content of the words used in the rule. Above all else, what would matter under this approach would be whether the application of a rule to a particular fact pattern had the same or a different effect from the one it had in earlier cases.
I probably need not emphasize that the view presented here, as well as the operating technique which flows from it, is no more the prevailing view in our precedent system than it is in Germany. For even in America the viewpoint that we have repeatedly encountered, viz. that the court must apply only "valid existing law," is at the forefront of legal thinking. Holmes was probably the first to pave the way to new insight and a new point of view. Under the influence of modern logic and psychology, this view has recently been gaining ground not only among scholars2 but also increasingly among the most enlightened
new case too. Likewise, it is his adversary's job to frame the competing trend in words which lead to the opposite result. It is the judge's job to choose—and perhaps also to reformulate the victorious trend, more narrowly or broadly than espoused by the attorney. In short, it appears to me that legal change should be acknowledged at the very moment it is occurring, and in the normal doctrine of the law too.
2. Corbin, Underhill Moore, C. E. Clark, Douglas, Sturges, Walton Hamilton, Turner (Yale); R.R.B. Powell, Y. B. Smith, Patterson, Magill, Michael, Klaus, Handler (Columbia); Walter Wheeler Cook, Oliphant, Yntema, Hope (Johns Hopkins); T.R. Powell, Morgan, Nathan Isaacs (Harvard); McMurray, Kidd, Max Radin, Ballantine (California); Bigelow, Bogert (Chicago); Bohlen, Wright (Pennsylvania); Breckenridge (North Carolina)—and many more. Thus, such efforts have their roots primarily (though by no means exclusively) at Columbia, Yale and Johns Hopkins. Among scholars, it is more or less these same men who are simultaneously engaged in working out Is-statements about the actual course of decisionmaking, and in examining "legal facts" with the aim of critiquing of the Ought-sentences. (For the line of argument in an expanded form, see Llewellyn, A Realistic Jurisprudence—The Next Step, 30 Colum. L. Rev. 431 (1930)).
1018
COLUMBIA LA W REVIEW               [Vol. 88:989
judges.3 That it will continue to gain still more ground seems clear to me; but it is very far from orthodox.
On the other hand, I must hasten to underscore that the doctrinal "innovations" which would flow from acceptance of a fundamentally sociological attitude have nothing new about them at all. Each process discussed here is a run of the mill occurrence nowadays, but everywhere it wears the frock-coat and top-hat of black-letter law. In America, I know this is so; I assume it is no different in German law. Day after day, the courts and the supporters of black-letter law act along the lines here discussed. Not always, just frequently. Nowadays, nobody complains when this is what is actually taking place.
The only seemingly odd thing would be for that which now often occurs but seldom gets talked about to be talked about as often as it occurs. And something else: for people at that point to develop insight and expressly acknowledge that it is precisely this process which is normal, predictable and correct in all genuinely disputed cases, not orthodox legal theory's supposed "deduction." Creative innovations that are made every day but now go unnoticed could then be made consciously and as straightforwardly as possible, given the very limited means at a judge's disposal. When reaffirming the Old, one would not be acting on the basis of a supposed inability to change it, but on the basis of a conscious assessment that this is the best thing to do. In short, the lawyer's ideology would be changed to conform to his day-today actions.
Perhaps this would be a gross political error. Perhaps legal training could no longer rein in a judge who knows where he stands. Perhaps the continuity in case law decisionmaking, as well as the constraints of taking one's directions from a statute, would dissolve if a judge were to lose the belief that he was tied to one spot. Perhaps he must believe he is obeying in order for him to be a wise commander. Perhaps old Portalis was actually off the mark when he sought to assure us: on ne saurait comprendre combien cette habitude de science et de raison adoucit et regle ce pouvoir* We do not wish to make our judges into law-givers on the scale of the legislature. Perhaps we best achieve this by denying that judges actually have what is in fact the indispensable power to create law for specific cases—thereby inducing them, as far as possible, to exercise this power blindly, because unconsciously.
But we, the newer case law men, observe our judges and see some who still, in all candor, believe they cannot create new law; and we see others who know they must create it, since there is no other way. We
3. Holmes, Brandeis, Stone (Supreme Court of the United States); Cardozo, Pound, Lehmann, Crane (New York); Wheeler (Connecticut); Burch, Mason (Kansas); Hough (deceased), Learned Hand, Augustus Hand, Swan, Anderson (1st and 2d Circuits); etc.
* "It is hard to appreciate the extent to which this power is regulated and modified by a judge's habits of learning and thought."  [Translator's Note]
1988]
LLEWELLYN
1019
compare the way each judges, each creates. The first kind of judging can still get by without complaints from us. But with the other, one's heart is gladdened. From their training flows technique; from the judges' sense of duty flows continuity of law; from a conscious assumption of power flows responsibility and a constant distrust of one's own prejudices; from insight flows that series of small, fundamentally important steps forward which make friends of Law and Life.4
* * * * *
Section 74 Conclusion
This concludes my study of the American case law system. I have tried, for the most part, to avoid value and policy judgments. Where they could not be avoided, I have tried to label them in such a way that none of my own value judgments could be understood as statements of fact. Above all, I have attempted to keep the selection of cases and the way the facts and circumstances are presented from being influenced by my own socio-political views. The main goal of this book is what is stated in the preface: to present the actual facts about the culture studied in the way an anthropologist would; wherever possible, to present the matters depicted in such a way that those who adhere to a different theory will be able to refute me with my own raw materials. Only to the extent one approaches this goal can a study in the sociology of law, currently an undeveloped science, have any hope of lasting value.
I regret that the area studied here is such a narrow one. Only upper-court cases are treated; and then, only the relation between decisionmaking and legal rules. Thus, it only deals with the appellate level; with legal doctrine and the case law system at this appellate level; with the creation, development, handling and effects of legal rules at this level. There is nothing about the extremely important trial-court level, about how procedure affects the workings of legal rules; virtually nothing about the influence of legal rules on the practice of law, particularly the practice of those lawyers who are primarily client counselors. There is truly nothing about the influence of legal rules on the lives of people whose interests the law affects. While such a narrow field of study may seem rather paltry, I nonetheless would wish to say: It only
4. I have now let myself slip into polemics and must hasten to give the other side its due. The evidence supplied by our judiciary is not convincing, for two reasons. First, given the current state of legal doctrine, it is only the most insightful and intellectually capable judges who are able to reach this new level of understanding. And it is from them that one would have expected the better opinions anyway. Second, these judges are all lawyers who have grown to maturity under the old doctrine, and who have incorporated the old training's good points into their usual ways of doing things. After attaining the new insights, they could not be rid of them even if they wanted to. Nonetheless their evidence is stirring. Especially noteworthy in these enlightened judges is their liberation from unconscious prejudices of class, caste, etc.
1020                          COLUMBIA LA W REVIEW               [Vol. 88:989
looks that way. For a science is made up of only modest, small-scale works. In the area worked on here, at least we are able to have known facts before our eyes—which, in legal study, has not been all that common. Admittedly, the exclusive focus on appellate-court decisions and legal rules plays into the lawyer's peculiar prejudice that these decisions are precisely what matters, in and of themselves, regardless of the effects they may have on the society from which they spring. But perhaps it is precisely here that hope lies. Once we get to thinking about what these legal rules really are, what their meaning really is, what the nature even of supreme-court decisionmaking is, then we must already be drawing closer to Life and finding in ourselves the urge to obtain more first-hand knowledge about the whole purpose of law, its utility to society in general. But once our legal fraternity feels this urge within it, the smaller problems—like questions about the nature and growth of precedent—will be solved through a new wealth of illuminating facts.