Karl Llewellyn, The Bramble Bush 45-49, 66-69 (Oceana ed. 1951)

Now for a while I am going to risk confusion for the sake of talking simply. I am going to treat as the rule of the case the ratio decidendi, the rule the court tells you is the rule of the case, the ground, as the phrase goes, upon which the court itself has rested its decision. For there is where you must begin, and such refinements as are needed may come after.

The court, I will assume, has talked for five pages, only one of which portrayed the facts assumed. The rest has been discussion. And judgment has been given for the party who won below: judgment affirmed. We seek the rule.

The first thing to note is this: no rule can be the ratio decidendi from which the actual judgment (here: affirmance) does not follow. Unless affirmance follows from a rule, it cannot be the rule which produced the actual holding of affirmance. But that holding is the decision, and the court speaks ex cathedra only as to the dispute decided, and only as to the decision it has made. At this point, too, I think you begin to see the bearing of the procedural issue. There can be a decision (and so an ex cathedra ration) only as to a point which is before the court. But points come before a court of review by way of specific complaint about specific action of the court below, and in no other way. Hence nothing can be held which is not thus brought up.

You will have noted that these two statements are not quite the same. For the losing party may have complained of five, or fourteen, different rulings by the court below, but the final judgment below is affirmed or reversed but once. [Llewellyn then asks how we should interpret an appellate court’s reversal of the judgment on each of the five issues raised in the appeal. Merely because reversal on only one point would have sufficed to reverse the judgment didn’t make the ruling on the other points mere dicta, because they were called for by the facts and issues in the case. Still, the ruling on any one of the points carries less weight than it would if it had been on the only issue in the case, perhaps because the court may not have ‘sweated’ over each point as much as it would had the point been the only issue in the case. Suppose, too, the judgment is reversed on one point and affirmed on the other four. The points affirmed are not dicta but their authority is weak since they do not meet the other test: the actual judgment does not logically follow from them.]

But our troubles with the ration decidendi are not over. We meet forthwith a further formal rule. Our judge states his facts, he argues his position, he announces his rule. And lo, he seems but to have begun. Once clean across the plate. But he begins again, winds up again, and again he delivers his ratio – this time, to our puzzlement, the words are not the same. At this point it is broader than it was before, there it is narrower. And like as not he will warm up another time, and do the same job over – differently again. I have never made out quite why this happens. A little, it may be due to a lawyer’s tendency to clinch an argument by summarizing its course, when he is through. A little, it may be due to mere sloppiness of composition, to the lack, typical of our law and all its work, of a developed sense of form, juristic or esthetic, for what the Romans knew as elegantia. Sometimes I get a wry suspicion that the judge repeats because he is uneasy on his ground, that he lifts up his voice, prays his conclusion over loud and louder, to gain and make conviction, much like an advertiser bare of argument except his slogan. At other times I feel as I read opinions the thrill of adventure in an undiscovered country; the first and second statements of the ratio, with all that has led up to them, are like the first and second reconnoiterings of strange hills, like first and second chartings of what has been found and what surmised – knowledge and insight growing as the opinion builds to its conclusion. But whatever the reason, recurrent almost-repetition faces us; also the worry that the repetition is seldom exact. Which phrasing are we then to tie to?

Perhaps in this, as in judging how far to trust a broadly-stated rule, we may find guidance in the facts the court assumes. Surely this much is certain. The actual dispute before the court is limited as straitly by the facts as by the form which the procedural issue has assumed. What is not in the facts cannot be present for the decision. Rules which proceed an inch beyond the facts must be suspect.

But how far does that help us out? What are the facts? The plaintiff’s name is Atkinson and the defendant’s Walpole. The defendant, despite his name, is an Italian by extraction, but the plaintiff’s ancestors came over with the Pilgrims. The defendant has a schnauzer-dog named Walter, red hair, and $30,000 worth of life insurance. It is about an auto accident. The defendant’s auto was a Buick painted pale magenta. He is married. His wife was in the back seat, an irritable somewhat faded blonde. She was attempting back seat driving when the accident occurred. He had turned around to make objection. In the process the car swerved and hit the plaintiff. The sun was shining; there was a rather lovely dappled sky low to the West. The time was late October on a Tuesday. The road was smooth, concrete. It had been put in by the McCarthy Road Work Company. How many of these facts are important to the decision? How many of these facts are, as we say, legally relevant? Is it relevant that the road was in the country or the city; that it was concrete or tarmac or of dirt; that it was a private or a public way? Is it relevant that the defendant was driving a Buick, or a motorcar, or a vehicle? Is it important that he looked around as the car swerved? Is it crucial? Would it have been the same if he had been drunk, or had swerved for fun, to see how close he could run by the plaintiff, but had missed his guess?

Is it not obvious that as soon as you pick up this statement of the facts to find your legal bearings you must discard some as of no interest whatsoever, discard others as dramatic but as legal nothings? And is it not clear, further, that when you pick up the facts which are left and which do seem relevant, you suddenly cease to deal with them in the concrete and deal with them instead in categories which you, for one reason or another, deem significant? It is not the road between Pottsville and Arlington; it is "a highway." It is not a particular pale magenta Buick eight, number 732507, but "a motorcar," and perhaps even "a vehicle." It is not a turning around to look at Adoree Walpole, but a lapse from the supposedly proper procedure of careful drivers, with which you are concerned. Each concrete fact of the case arranges itself, I say, as the representative of a much wider abstract category of facts, and it is not in itself but as a member of the category that you attribute significance to it. But what is to tell you whether to make your category "Buicks" or "motorcars" or "vehicles"? What is to tell you to make your category "road" or "public highway"? The court may tell you. But the precise point that you have up for study is how far it is safe to trust what the court says. The precise issue you are attempting to solve is whether the court’s language can be taken as it stands, or must be amplified, or must be whittled down.

This brings us at last to the case system. For the truth of the matter is a truth so obvious and trite that it is somewhat regularly overlooked by students. That no case can have a meaning by itself! Standing alone it gives you no guidance. It can give you no guidance as to how far it carries, as to how much of its language will hold water later. What counts, what gives you leads, what gives you sureness, that is the background of the other cases in relation to which you must read the one. They color the language, the technical terms, used in the opinion. But above all they give you the wherewithal to find which of the facts are significant, and in what aspect they are significant, and how far the rules laid down are to be trusted.

Here, I say, is the foundation of the case system. For what, in a case class, do we do? We have set before you, at either the editor’s selection or our own, a series of opinions which in some matter are related. They may or may not be exactly alike in their outcome. They are always supposedly somewhat similar on their legally relevant facts. Indeed, it is the aspects in which their facts are similar which give you first guidance as to what classes of fact will be found legally relevant, that is, will be found to operate alike, or to operate at all, upon the court. On the other hand, the states of facts are rarely, if ever, quite alike. And one of the most striking problems before you is: when you find two cases side by side which show a difference in result, then to determine, what difference in their facts, or what difference in the procedural set-up, has produced that difference in result. Those are the two problems which must be in your mind as you examine the language of the opinions. I repeat them. First, what are the significant categories of facts, and what is their significance to the court? Second, what differences in facts or in procedural set-up produce differences in the court’s action when the situations are otherwise alike?

* * *

We turn first to what I may call the orthodox doctrine of precedent, with which, in its essence, you are already familiar. Every case lays down a rule, the rule of the case. The express ratio decidendi is prima facie the rule of the case, since it is the ground on which the court chose to rest its decision. But a later court can reexamine the case and can invoke the canon that no judge has power to decide what is not before him, can, through examination of the facts or of the procedural issue, narrow the picture of what was actually before the court and can hold that the ruling made requires to be understood as thus restricted. In the extreme form this results in what is known as expressly "confining the case to its particular facts." This rule holds only of redheaded Walpoles in pale magenta Buick cars. And when you find this said of a past case you know that in effect it has been overruled. Only a convention, a somewhat absurd convention, prevents flat overruling in such instances. It seems to be felt as definitely improper to state that the court in a prior case was wrong, peculiarly so if that case was in the same court which is speaking now. It seems to be felt that this would undermine the dogma of the infallibility of courts. So lip service is done to that dogma, while the rule which the prior court laid down is disemboweled. The execution proceeds with respect, with mandarin courtesy.

Now this orthodox view of the authority of precedent – which I shall call the strict view – is but one of two views which seem to me wholly contradictory to each other. It is in practice the dogma which is applied to unwelcome precedents. It is the recognized, legitimate, honorable technique for whittling precedents away, for making the lawyer, in his argument, and the court, in its decision, free of them. It is a surgeon’s knife.…

…[W]hen you turn to the actual operations of the courts, or, indeed, to the arguments of lawyers, you will find a totally different view of precedent at work beside this first one. That I shall call, to give it a name, the loose view of precedent. That is the view that a court has decided, and has decided authoritatively, any point or all points on which it chose to rest a case, or on which it chose, after due argument, to pass. No matter how broad the statement, no matter how unnecessary on the facts or the procedural issues, if that was the rule the court laid down, then that the court has held. Indeed, this view carries over often into dicta, and even into dicta that are grandly obiter. In its extreme form this results in thinking and arguing exclusively from language that is found in past opinions, and in citing and working with that language wholly without reference to the facts of the case which called the language forth.

Now it is obvious that this is a device not for cutting past opinions away from judges’ feet, but for using them as a springboard when they are found convenient. This is a device for capitalizing welcome precedents. And both the lawyers and judges use it so. And judged by the practice of the most respected courts of ordinary statute, this doctrine of precedent is like the other, recognized, legitimate, honorable.

What I wish to sink deep into your minds about the doctrine of precedent, therefore, is that it is two-headed. It is Janus-faced. That it is not one doctrine, nor one line of doctrine, but two, and two which, applied at the same time to the same precedent, are contradictory of each other. That there is one doctrine for getting rid of precedents deemed troublesome and one doctrine for making use of precedents that seem helpful. That these two doctrines exist side by side. That the same lawyer in the same brief, the same judge in the same opinion, maybe using the one doctrine, the technically strict one, to cut down half the older cases that he deals with, and using the other doctrine, the loose one, for building with the other half. Until you realize this you do not see how it is possible for law to change and develop, and yet to stand on the past. You do not see how it is possible to avoid the past mistakes of courts, and yet to make use of every happy insight for which a judge in writing may have found expression …

Nor, until you see this double aspect of the doctrine-in-action, do you appreciate how little, in detail, you can predict out of the rules alone; how much you must turn, for purposes of prediction, to the reactions of the judges to the facts, and to the life around them …

Applying this two faced doctrine of precedent to your work in a case class you get, it seems to me, some such result as this. You read each case from the angle of its maximum value as precedent, at least from the angle of its maximum value as a precedent of the first water. You will recall that I recommended taking down the ratio decidendi in substantially the court’s own words. You see now what I had in mind. On the other hand, you will read each case for its minimum value as a precedent, to set against the maximum. In doing this you have your eyes out for the narrow issue in the case, the narrower the better. The first question is, how much can this case fairly be made to stand for by a later court to whom the precedent is welcome: You may well add – although this will be slightly flawed authority – the dicta which appear to have been well considered. The second question is, how much is there in this case that cannot be got around, even by a later court that wishes to avoid it?

You have now the tools for arguing from that case as counsel on either side of a new case. You turn then to the problem of prediction. Which view will the same court, on a later case on slightly different facts, take: will it choose the narrow or the loose: Wish use will be made of this case by one of the other courts whose opinions are before you? Here you will call to your aid the matter of attitude that I have been discussing. Here you will use all that you know of individual judges, or trends in specific courts, or, indeed, of the trend in the line of business, or in the situation, or in the times at large – in anything which you may expect to become apparent and important to the court in later cases. But always, and always, you will bear in mind that each precedent had not one value, but two, and that the two are wide apart, and that whichever value a later court assigns to it, such assignment will be respectable, traditionally sound, dogmatically correct. Above all, as you turn this information to your own training you will, I hope, come to see that in most doubtful cases precedents must speak ambiguously until the court has made up its mind whether each one of them is welcome or unwelcome. And the job of persuasion which falls upon you will call, therefore, not only for providing a technical ladder to reach on authority the result that you contend for, but even more, if you are to have your use of the precedents made as you propose it, the job calls for you, on the facts, to persuade the court your cases is sound.

People – and they are curiously many – who think that precedent produces or ever did produce a certainty that did not involve matters of judgment and of persuasion, or who think that what I describe involves improper equivocation by the courts or departure from the court-ways of some golden age – such people simply do not know our system of precedent in which they live.

Editor’s note: Why does Llewellyn keep talking about the "facts the court assumes," rather than "the facts the court finds"?

Herman Oliphant,^a A Return to Stare Decisis, 14 American Bar Association Journal 71-73, 159 (1928).

…[A] most profound change … has been slowly and imperceptibly creeping into our treatment of problems in Anglo-American law, a fundamental change which merits careful study in order that we may recognize its presence, measure its extent, and judge its consequences. Let me anticipate my conclusions by asserting that we are well on our way toward a shift from following decisions to following so-called principles, from stare decisis to what I shall call stare dictis; by saying that this shift has far-reaching and unfortunate consequences for both the art of judicial government and the science of law, and by proposing a return toward the ancient doctrine of stare decisis.

Stare decisis analyzed. Support for this position will be found by examining that doctrine. It asserts not one thing, but two. For one thing, it asserts that prior decisions are to be followed, not disregarded. But it also asserts that we are to follow the prior decisions and not something else.

The First Meaning of the Doctrine. Most discussions of the doctrine of stare decisis have emphasized the first of these two assertions. In those we are told of the advantages and disadvantages of the doctrine. It has been pointed out how, on the one hand, it makes the law applicable to future transactions certain and the future decisions of judges predictable; and again, how it gives us justice according to law and not according to the whims of men. On the other hand, it has been shown that to follow it gives us a measure of inflexibility in our law, resisting changes needed to meet changing conditions. We are all familiar with these and other broad implications of this branch of the doctrine and have considered the necessary choice between conflicting advantages which its acceptance or rejection involves. The vigor of this branch of the ancient doctrine has been weakened but little. Something in the cases is being followed. This whole aspect of the matter is mentioned here only to be set aside.

The Second Meaning of the Doctrine. … There seems to have been little study of this phase of the doctrine, – of just what it is in prior decisions which is to be followed. General statements that the decision is to be looked for, that dicta are of slight weight and offer no certain guide can be turned to at many places in the books and are familiar to all. Students beginning their law study are told these things in a general way and then are left to an apprenticeship among the cases to discover largely for themselves their fuller meaning. Yet this matter is the one most vital and difficult factor conditioning the soundness of their scholarship. It is because the word decision may mean any one of many things that it is perilous to leave the matter thus unarticulated.

What Does the Case Decide? In the first place, a court, in deciding a case, may throw out a statement as to how it would decide some other case. Now if that statement is a statement of another case which is as narrow and specific as the actual case before the court, it is easily recognized as dictum and given its proper weight as such. In the second place the court may throw out a broader statement, covering a whole group of cases. … But so long as that statement does not cover the case before the court, it is readily recognized as being not a decision, much less the decision of the case. It is dictum, so labeled and appraised. But in the third place, a court may make a statement broad enough to dispose of the case in hand as well as to cover a few or many other states of fact. Statements of this sort may cover a number of fact situations ranging from one other to legion. Such a statement is sometimes called the decision of the case. Thereby the whole ambiguity of that word is introduced and the whole difficulty presented.

If a more careful usage limits the word decision to the action taken by the court in the specific case before it, i.e., to the naked judgment or order entered, the difficulty is not met; it is merely shifted. Stare decisis thus becomes useless for no decision in that limited sense can ever be followed. No identical case can arise. All other cases will differ in some circumstances – in time, if in no other, and most of them will have differences that are not trivial. Decision in the sense meant in stare decisis must, therefore, refer to a proposition of law covering a group of fact situations … as a minimum, the fact situation of the instant case and at least one other.

To bring together into one class even this minimum of two fact situations however similar they may be, always has required and always will require^b an abstraction. If Paul and Peter are to be thought of together at all, they must both be apostles or be thought of as having some other attribute in common. Classification is abstraction. An element or elements common to the two fact situations put into one class must be drawn out from each to become the content of the category and the subject of the proposition of law which is thus applied to the two cases.

But such a grouping may include multitudes of fact situations so long as a single attribute common to them all can be found. Between these two extremes lies a gradation of groups of fact situations each with its corresponding proposition of law, ranging from a grouping subtending but two situations to those covering hosts of them. This series of groupings of fact situations gives us a parallel series of corresponding propositions of law, each more and more generalized as we recede farther and farther from the instant statement of facts and include more and more fact situations in the successive groupings. It is a mounting and widening structure, each proposition including all that has gone before and becoming more general by embracing new states of fact. For example, A’s father induces her not to marry B as she promised to do. On a holding that the father is not liable to B for so doing, a gradation of widening propositions can be built, a very few of which are:

There can be erected on the action taken by a court in any case such a gradation of generalizations and this is commonly done in the opinion. Sometimes it is built up to dizzy heights by the court itself and at times by law teachers and writers, it is reared to those lofty summits of the absolute and the infinite.

Where on that gradation of propositions are we to take our stand and say "This proposition is the decision of the case within the meaning of the doctrine of stare decisis?" Can a proposition of law of this third type ever become so broad that, as to any of the cases it would cover, it is mere dictum?

A Question of Double Difficulty. That would be difficult enough if it ended there. But just as one and the same apple can be thrown into any one of many groups of barrels according to its size, color, shape, etc., so also there stretches up and away from every single case in the books, not one possible gradation of widening generalizations, but many multitudes of radii shoot out from it, each pair enclosing one of an indefinite number of these gradations of broader and broader generalizations. For example, a contract for wages contains a stipulation that it shall be non-assignable by the employee. A court holds that the laborer can assign anyway and that his assignee can sue the employer for the wages regardless of the stipulation. This holding can serve as the apex of many triangles of generalizations. At the base of one will be a broad generalization treating the claim as property and asserting the alienability of property; at the base of another will be an equally broad generalization having to do with contractual stipulations opposed to public policy and the base of a third will be a similarly wide generalization concerning the liquidation of claims in the labor market. Others could be enumerated and other cases similarly analyzed. That is not needed, for we all know of at least one case appearing in the case books of more than one subject upon which securely rests more than one inverted pyramid of favorite theory.

A student is told to seek the "doctrine" or "principle" of a case, but which of its welter of stairs shall he ascend and how high up shall he go? Is there some one step on some one stair which is the decision of the case within the meaning of the mandate of stare decisis? That is the double difficulty. Each precedent considered by a judge and each case studied by a student rests at the center of a vast and empty stadium. The angle and distance from which that case is to be viewed involves the choice of a seat. Which shall be chosen? Neither judge nor student can escape the fact that he can and must choose. To realize how wide the possibilities and significant the consequences of that choice are is elementary to an understanding of stare decisis. To ask whether there exists a coercion of some logic to make that choice either inevitable or beneficent, searches the significance of stare decisis in judicial government and the soundness of scholarship in law. This question is real and insistent. It is one which should be asked explicitly and faced squarely. …

But there is a constant factor in the cases which is susceptible of sound and satisfying study. The predictable element in it all is what courts have done in response to the stimuli of the facts of the concrete cases before them. Not the judges’ opinions, but which way they decide cases will be the dominant subject matter of any truly scientific study of law. This is the field for scholarly work worthy of best talents because the work to be done is not the study of vague and shifting rationalizations but the study of such tough things as the accumulated wisdom of men taught by immediate experience in contemporary life, – the battered experience of judges among brutal facts. The response of their intuition of experience to the stimulus of human situations is the subject-matter having that constancy and objectivity necessary for truly scientific study. When we pin our attention to this, we may more freely criticize what courts have said but we shall more cautiously criticize what they have done realizing, as we shall, that they were exposed to the impact of more facts than we.

This surer thing for scholarly purpose is also the inner secret of what is soundest in the enfeebled stare decisis in judicial government of today. With eyes cleared of the old and broad abstractions which curtain our vision, we come to recognize more and more the eminent good sense in what courts are wont to do about disputes before them. Judges are men and men respond to human situations. When the facts stimulating them to the action taken are studied from a particular and current point of view, which our present classification prevents, we acquire a new faith in stare decisis. From this viewpoint we see that courts are dominantly coerced not by the essays of their predecessors but by a surer thing, – by an intuition of fitness of solution to the problem, – and a renewed confidence in judicial government is engendered. To state the matter more concretely, the decision of a particular case by a thoughtful scholar is to be preferred to that by a poorly trained judge but the decision of such a judge in a particular case is infinitely to be preferred to a decision of it preordained by some broad "principle" laid down by the scholar when this and a host of other cases had never even occurred to him.

One sampling of this proposed subject-matter of a real science of law must suffice. There are two lines of old cases involving the validity of promises not to compete. They are considered in square conflict. But when the opinions are ignored and the facts re-examined, all the cases holding the promise invalid are found to be cases of employees’ promises not to compete with their employers after a term of employment. Contemporary guild regulations not noticed in the opinions made their holdings eminently sound. All the cases holding the promises valid were cases of promises by those selling a business and promising not to compete with the purchasers. Contemporary economic reality made these holdings also eminently sound. The distinction between those two lines of cases is not even hinted at in any of the opinions but the courts’ intuition of experience led them to follow it with amazing sureness and the law resulting fitted life. That is a sample of the stuff capable of scientific study. …

1. To what if any extent is Oliphant’s essay in conflict with Llewellyn’s view?

I'm including the excerpt from Goodhart's article, fut I'd advise reading it only if you are particularly interested in the subject. I don't intend to discuss it at the enrichment session. As you read the Notes and Questions at the end, ignore the references to Goodheart.

Arthur L. Goodhart, Determining the Ratio Decidendi of a Case, 40 Yale L.J. 161-183 (1920)

…The initial difficulty with which we are faced is the phrase "ratio decidendi" itself. With the possible exception of the legal term "malice," it is the most misleading expression in English law, for the reason which the judge gives for his decision is never the binding part of the precedent. The logic of the argument, the analysis of prior cases, the statement of the historical background may all be demonstrably incorrect in a judgment, but the case remains as a precedent nevertheless. It would not be difficult to cita a large number of leading cases, both ancient and modern, in which one or more of the reasons given for the decision can be proved to be wrong; but in spite of this these cases contain valid and definite principles which are as binding as if the reasoning on which they are based was correct.

In Priestly v. Fowler^d the famous or infamous doctrine of common employment was first laid down. Of this case it has been well said, "Lord Abinger planted it, Baron Alderson watered it, and the Devil gave it increase." Yet the case is still law in England (although limited in effect by the Employers Liability Act of 1880) in spite of the fact that the two reasons on which Lord Abinger based his judgment are palpably incorrect. The first reason is that any other rule would be "absurd." This argument is always a dangerous one upon which to base a judgment and in this instance it is, unfortunately, the rule in Priestly v. Fowler which as proved to be not only absurd but also unjust. The second reason given by Lord Abinger is that by his contract of service a servant impliedly consents to run the risk of working with negligent fellow-servants. In fact, of course, a servant does not consent to run the risk; the implication was invented by the judge himself.…

For that matter, by what may seem a strange method to those who do not understand the theory of the Common Law, it is precisely some of those cases which have been decided on incorrect premises or reasoning which have become the most important in the law. New principles, of which their authors were unconscious or which they have misunderstood, have been established by these judgments. Paradoxical as it may sound, the law has frequently owed more to its weak judges than it has to its strong ones. A bad reason may often bake good law.… Our modern law of torts has been developed to a considerable extent by a series of bad arguments, and our property law is in many instances founded on incorrect history. To state this is not, however, to question the authority of that law. It is clear, therefore, that the first rule for discovering the ratio decidendi of a case is that it must not be sought in the reasons on which the judge has based his decision.…

Having stated its reasons for reaching a certain conclusion, the court frequently sums up the result in a general statement of the law on the point at issue. Can we find the principle of the case in this proposition of law, this comprehensive expression of the rule involved, which students underline with such enthusiasm in their casebooks? Thus in the chapter on Judgments in Halsbury’s The Laws of England (18:210), the rule is given as follows:

Professor Morgan of the Harvard Law School, in his valuable book The Study of Law, says:

If these statements are to be understood in their literal sense, it is respectfully submitted that the words are misleading, for it is not the rule of law "set forth" by the court, or the rule "enunciated" as Halsbury puts it, which necessarily constitutes the principle of the case. There may be no rule of law set forth in the opinion, or the rule when stated may be too wide or too narrow. In appellate courts, the rule of law set forth by the different judges^e may have no relation to each other. Nevertheless, each of these cases contains a principle which can be discovered on proper analysis.

So also a case may be a precedent, involving an important principle of law, although the court has given judgment without delivering an opinion. At the present time, although occasionally an appellate court will affirm without opinion a case which involves an interesting point, we rarely find a case of any importance in which an opinion has not been written. In the past, however, especially during the Year Book period, we find a great number of cases in which there were no opinions and in which the principle therefore must be sought elsewhere.

Of more frequent occurrence in recent cases is the practice of delivering an opinion, but at the same time being careful not to state any general principle of law ….

Again, a case may contain a definite principle, although the expression of it in the opinion may not be strictly accurate. In Rex v. Fenton [1 Lew. C.C. 179 (1830)] the prisoner caused the death of a man by wantonly throwing a large stone down a mine. In his charge to the jury Tindal, C.J., said:

The principle of the case was correct, although the statement of it was too wide, as was held in the later case of Regina v. Franklin [15 Cox C.C. 163 (1883)]. In that case the prisoner threw a box belonging to a refreshment stall keeper into the sea, thereby killing a swimmer. The point at issue was whether, apart from the question of negligence, the prisoner was guilty of manslaughter, his act having been a wrongful one. Field, J., said:

A striking example of an overstatement of the principle involved in a case may be found in Riggs v. Palmer [115 N.Y. 506 (1889)]. The court held that a legatee, who had murdered his testator, could not take under the will, because no one shall be permitted "to take advantage of his own wrong, or to found any claim upon his own iniquity, or to acquire property by his own crime." It would, of course, be possible to give a large number of situations in which this statement would be wrong or doubtful. Would it apply, for example, if the legatee had negligently killed the testator in a motor accident? …

On the other hand the rule of law may be stated in too narrow a form. In Barwick v. English Joint Stock Bank [L.R. 2 Ex. 259 (1867)] the defendant’s bank manager fraudulently induced the plaintiff to accept a valueless guarantee. In delivering the judgment of the court, Willes, J., said:

It was generally believed that this statement of the law was correct until, forty-five years later, the House of Lords in Lloyd v. Grace, Smith & Co. [1912 A.C. 716] held that it was too narrow. The words "and for the master’s benefit" were merely descriptive of the facts in the Barwick case, and not a necessary part of the principle involved. The House of Lords did not disapprove of the principle of the Barwick case, but held that "it is … a mistake to qualify it by saying that it only applies when the principal has profited by the fraud."

When we consider the appellate courts it becomes even more obvious that the principle of the case cannot necessarily be found in the rule of law enunciated, for it is not infrequent to find that, although the judges may concur in the result, they differ widely in their statements of the law.^e… Nevertheless these cases cannot be ignored as precedents on the ground that the rules of law set forth cannot be reconciled.

Since, therefore, the principle of the case is not necessarily found in either the reasoning of the court or in the proposition of law set forth, we must seek some other method of determining it. Does this mean that we can ignore the opinion entirely and work out the principle for ourselves from the facts of the case and the judgment reached on those facts? This seems to be the view of a certain American school of legal thought represented by Professor Oliphant. According to him it is what the judge does and not what he says that matters. He writes:

Undoubtedly this theory has the attractiveness of simplicity. No longer will we have to analyze the sometimes lengthy and difficult opinions of the judges; all that we are concerned with are the facts and the conclusions. The judge who writes an opinion will be wasting both his own time and ours, for it is not what he says but what he does that matters. We can ignore the vocal behaviour of the judge, which sometimes fills many pages, and concentrate upon his nonvocal behaviour which occupies but a few lines.

Unfortunately I believe there is a fallacy in Professor Oliphant’s argument which will prevent our following this convenient course. The fallacy lies in suggesting that the facts of a case are a constant factor, that the judge’s conlusion is based upon the fixed premise of a given set of facts. We do not have to be philosophers to realize that facts are not constant but relative. The crucial question is "What facts are we talking about?" The same set of facts may look entirely different to two different persons. The judge founds his conclusions upon a group of facts selected by him as material from among a larger mass of facts, some of which might seem significant to a layman, but which, to a lawyer, are irrelevant. The judge, therefore, reaches a conclusion upon the facts as he sees them. It is on these facts that he bases his judgment, and not on any others. It follows that our task in analyzing a case is not to state the facts and the conclusion, but to state the material facts as seen by the judge and his conclusions based on them. It is by his choice of material facts that the judge creates law. A congeries of facts are presented to him; he chooses those which he considers material and rejects those which are immaterial, and then bases his conclusion upon the material ones. To ignore his choice is to miss the whole point of the case. Our system of precedent becomes meaningless if we say that we will accept his conclusion but not his view of the facts. His conclusion is based on the material facts as he sees them, and we cannot add or subtract from them by proving that other facts existed in the case. It is, therefore, essential to know what the judge has said about his choice of the facts, for what he does has a meaning for us only when considered in relation to what he has said. A divorce of the conclusion from the material facts on which that conclusion is based is illogical, and must lead to arbitrary and unsound results.

The first and most essential step in the determination of the principle of a case is, therefore, to ascertain the material facts on which the judge has based his conclusion. Are there any rules that will help us in isolating those material facts? It is obvious that none can be found which will invariably give us the desired result, for if this were possible then the interpretation of cases, which is one of the most difficult of the arts, would be comparatively easy. The following tentative suggestions may, however, prove of some aid to the student faced with his first case-book.

If there is no opinion, or if the opinion does not contain a statement of the facts, then we must assume that all the facts given in the report are material except those which on their face are not.^f Thus the facts of person, time, place, kind, and amount are presumably immaterial unless stated to be material. As a rule the law is the same for all persons, at all times, and at all places within the jurisdiction of the court. For the purposes of the law a contract made between A and B in Liverpool on Monday involving the sale of a book worth £10 is identical with a similar contract made between C and D in London on Friday involving the sale of a painting worth £100,000.…

If there is an opinion which gives the facts, the first point to notice is that we cannot go behind the opinion to show that the facts appear to be different in the record. We are bound by the judge’s statement of the facts even though it is patent that he has misstated them, for it is on the facts as he, perhaps incorrectly, has seen them that he has based his judgment.…

Two … cases illustrate this point in an interesting manner. In Smith v. London and South Western Ry. [L.R. 6 C.P. 14 (1870)], Kelly, C.B., Channell, B., and Blackburn, J., each assumed a fact "that no reasonable man would have foreseen that the fire would get to the plaintiff’s cottage." We lose the whole point of their judgments if we attempt to explain them by showing that a reasonable man should have foreseen that the fire might reach the cottage.… Similarly in In Re Polemis and Furness, Withy & Co., [(1912) 3 K.B. 560] the Court of Appeal was bound by the arbitrators’ finding of fact that a reasonable man would not have anticipated that a plank falling into the hold of a steamer filled with petrol vapour might cause an explosion. This finding of fact is probably incorrect, but we cannot ignore it if we are to determine the true principle of the judgments based on it. As has already been said, if we are not bound by the facts stated by the judge it would be wholly illogical to be bound by his conclusion on those facts.

Moreover, such a course would be most inconvenient, for it would then become necessary when citing an important casee to go through the record so as to be certain that the facts as given by the court were correct. In view of the vast number of precedents existing on almost any disputed point of law the task of the common law lawyer is sufficiently difficult at the present time; if he must also consult the record in every case to determine the actual facts his work will be overwhelming. The emphasis which American law libraries are now placing on collecting the whole records in leading cases may prove to be a dangerous one, for such collections tend to encourage a practice which is inconvenient in operation and disastrous in theory.

Although it is comparatively rare to find any real conflict between the facts given and those in the record, it is of frequent occurrence to find that the facts in the opinion fail to include some of the facts in the record. Under these circumstances there are two possible explanations of the omission: (1) the fact was considered by the court but was found to be immaterial, or (2) the fact in the record was not considered by the court as it was not called to its attention by counsel or was for some other reason overlooked. Which of the two explanations is the correct one will depend on the circumstances of the particular case. If counsel have referred to the fact in the course of their arguments this is strong evidence that the fact has not been overlooked by has been purposely omitted. For this reason the practice in the Law Reports of giving a short summary of counsel’s speeches is of particular value. But if it is clear that a certain fact, however material it may have been, was not considered by the court, then the case is not a precedent in future cases in which a similar fact appears.… It must be noted, however, that the burden of showing that a fact has been overlooked is a heavy one, for as a rule a material fact does not escape the attention of counsel and of the court.

Having, as a first step, determined the facts of the case as seen by the judge, it is then necessary to discover which of these facts he has found material for his judgment. This is far more difficult than the first step, for the judge may fail to label his facts. It is only the strong judge, one who is clear in his own mind as to the grounds for his decision, who invariably says, "on facts A and B and on them alone I reach conclusion X." Too often the cautious judge will include in his opinion facts which are not essential to his judgment, leaving it for future generations to determine whether or not these facts constitute a part of the ratio decidendi. The following guides may, however, be followed in distinguishing between material and immaterial facts.

(1) As was stated above in discussing the principle of a case in which there is no opinion, the facts of person, time, place, kind, and amount are presumably immaterial. This is true to an even greater extent when there is an opinion, for if these facts are held to be material particular emphasis will naturally be placed upon them.

(2) All facts which the court specifically states are immaterial must be considered so. [Professor Goodheart here refers to a case in which the evidence showed that defendant maintained an illegal drinking place, at which acts disturbing the peace occurred. The New York Court of Appeals ruled that the charge of maintaining a public nuisance was sufficiently sustained by proof of an illegal drinking place, and that the evidence of disorderly conduct introduced at the trial was immaterial.-Editor]

(3) All facts which the court impliedly treats as immaterial must be considered immaterial. The difficulty in these cases is to determine whether a court has or has not considered the fact immaterial. Evidence of this implication is found when the court, after having stated the facts generally, then proceeds to choose a smaller number of facts on which it bases its conclusion. The omitted facts are presumably held to be immaterial. In Rylands v. Fletcher [L.R. 3 H.L. 330 (1868)] the defendant employed an independent contractor to make a reservoir on his land. Owing to the contractor's negligence in not filling up some disused mining shafts, the water escaped and flooded the plaintiff's mine. The defendant was held liable. Is the principle of the case that a man who builds a reservoir on his land is liable for the negligence of an independent contractor? Why then is the case invariably cited as laying down the broader doctrine of "absolute liability"? The answer is found in the opinions. After stating the facts as above, the judges thereafter ignored the fact of the contractor's negligence, and based their conclusions on the fact that an artificial reservoir had been constructed. The negligence of the contractor was, therefore, impliedly held to be an immaterial fact.…

It is obvious from the above cases that it is essential to determine what facts have been held to be immaterial, for the principle of a case depends as much on exclusion as it does on inclusion. It is under these circumstances that the reasons given by the judge in his opinion, or his statement of the rule of law which he is following, are of peculiar importance, for they may furnish us with a guide for determining which facts he considered material and which immaterial. His reason may be incorrect and his statement of the law too wide, but they will indicate to us on what facts he reached his conclusion. . . .

(4) All facts which are specifically stated to be material must be considered material. Such specific statements are usually found in cases in which the judges are afraid of laying down too broad a principle. Thus in Heaven v. Pender [11 Q.B.D. 503 (1883)] the plaintiff, a workman employed to paint a ship, was injured because of a defective staging supplied by the defendant dock owner to the shipowner. Brett, M. R., held that the defendant was liable on the ground that:

Cotton and Bowen, L.JJ., agreed with the Master of the Rolls that the defendant was liable, but the material facts. on which they based their judgment were: (1) that the plaintiff was on the staging for business in which the dock owner was interested, and (2) he "must be considered as invited by the dock owner to use the dock and all appliances provided by the dock owner as incident to the use of the dock." The principle of the case cannot, therefore, be extended beyond the limitation of these material facts.

(5) If the opinion does not distinguish between material and immaterial facts then all the facts set forth in the opinion must be considered material with the exception of those that on their face are immaterial. There is a presumption against wide principles of law, and the smaller the number of material facts in a case the wider will the principle be. Thus if a case like Hambrook v. Stokes [(1925) 1 K.B. 141], in which a mother died owing to shock at seeing a motor accident which threatened her child, is decided on the fact that a bystander may recover for injury due to shock, we have a broad principle of law. If the additional fact that the bystander was a mother is held to be material we then get a narrow principle of law. Therefore, unless a fact is expressly or impliedly held to be immaterial, it must be considered material.

(6) Thus far we have been discussing the method of determining the principle of a case in which there is only a single opinion, or in which all the opinions are in agreement. How do we determine the principle of a case in which there are several opinions which agree as to the result but differ in the material facts on which they are based? In such an event the principle of the case is limited to the sum of all the facts held to be material by the various judges. A case involves facts A, B and C, and the defendant is held liable. The first judge finds that fact A is the only material fact; the second that B is material, the third that C is material. The principle of the case is, therefore, that on the material facts A, B and C the defendant is liable. If, however, two of the three judges had been in agreement that fact A was the only material one, and that the others were immaterial, then the case would be a precedent on this point, even though the third judge had held that facts B and C were the material ones. The method of determining the principle of a case in which there are several opinions is thus the same as that used when there is only one. Care must be taken by the student, however, to see that the material facts of each opinion are stated and analyzed accurately, for sometimes judges think that they are in agreement on the facts when they concur only in the result.

Having established the material and the immaterial facts of the case as seen by the court, we can then proceed to state the principle of the case. It is to be found in the conclusion reached by the judge on the basis of the material facts and on the exclusion of the immaterial ones. In a certain case the court finds that facts A, B and C exist. It then excludes fact A as immaterial, and on facts B and C it reaches conclusion X. What is the ratio decidendi of this case? There are two principles: (1) In any future case in which the facts are A, B and C, the court must reach conclusion X, and (2) in any future case in which the facts are B and C the court must reach conclusion X. In the second case the absence of fact A does not affect the result, for fact A has been held to be immaterial. The court, therefore, creates a principle when it determines which are the material and which are the immaterial facts on which it bases its decision.

It follows that a conclusion based on a fact the existence of which has not been determined by the court, cannot establish a principle. We then have what is called a dictum. If, therefore, a judge in the course of his opinion suggests a hypothetical fact, and then states what conclusion he would reach if that fact existed, he is not creating a principle. The difficulty which is sometimes found in determining whether a statement is a dictum or not is due to uncertainty as to whether the judge is treating a fact as hypothetical or real. When a judge says, "In this case, as the facts are so and so, I reach conclusion X," this is not a dictum, even though the judge has been incorrect in his statement of the facts. But if the judge says, "If the facts in this case were so and so then I would reach conclusion X," this is a dictum, even though the facts are as given. The second point frequently arises when a case involves two different sets of facts. Having determined the first set of facts and reached a conclusion on them, the judge may not desire to take up the time necessarily involved in determining the second set. Any views he may express as to the undetermined second set are accordingly dicta. If, however, the judge does determine both sets, as he is at liberty to do, and reaches a conclusion on both, then the case creates two principles and neither is a dictum.… On the other hand, if in a case the judge holds that a certain fact prevents a cause of action from arising, then his further finding that there would have been a cause of action except for this fact is an obiter dictum. By excluding the preventive fact the situation becomes hypothetical, and the conclusion based on such hypothetical facts can only be a dictum.

Having established the principle of a case, and excluded all dicta, the final step is to determine whether or not it is a binding precedent for some succeeding case in which the facts are prima facie similar. This involves a double analysis. We must first state the material facts in the precedent case and then attempt to find those which are material in the second one. If these are identical, then the first case is a binding precedent for the second, and the court must reach the same conclusion as it did in the first one. If the first case lacks any material fact or contains any additional ones not found in the second, then it is not a direct precedent. Thus, in Nichols v. Marsland [(L.R. 10 Ex. 255 (1875)] the material facts were similar to those in Rylands v. Fletcher [supra] except for the additional fact that the water escaped owing to a violent storm. If the court had found that this additional fact was not a material one, then the rule in Rylands v. Fletcher would have applied. But as it found that it was a material one, it was able to reach a different conclusion.

[The author then rejects the objection that the "material facts" doctrine "leaves us with hardly any general legal principles, for facts are infinitely various . . ." This variety is reduced by categorization (as Llewellyn demonstrated). Nor do judges nowadays, the author asserts, circumvent precedents by finding, arbitrarily, that additional facts are material. And if judges determine, erroneously, that certain material facts exist, the decision should nevertheless be authoritative in relation to the facts assumed: "it is better to suffer this mistake, which may prove of benefit to the law as a whole, however painful its results may have been to the individual litigant, than to throw doubt on every precedent on which our law is based."–Editor]

Conclusion. The rules for finding the principle of a case can, therefore, be summarized as follows:

(1) The principle of a case is not found in the reasons given in the opinion.

(2) The principle is not found in the rule of law set forth in the opinion.

(3) The principle is not necessarily found by a consideration of all the ascertainable facts of the case and the judge's decision.

(4) The principle of the case is found by taking account (a) of the facts treated by the judge as material, and (b) his decision as based on them.

(5) In finding the principle it is also necessary to establish what facts were held to be immaterial by the judge, for the principle may depend as much on exclusion as it does on inclusion.

The rules for finding what facts are material and what facts are immaterial as seen by the judge are as follows:

(1) All facts of person, time, place, kind and amount are immaterial unless stated to be material.

(2) If there is no opinion, or the opinion gives no facts, then all other facts in the record must be treated as material.

(3) If there is an opinion, then the facts as stated in the opinion are conclusive and cannot be contradicted from the record.

(4) If the opinion omits a fact which appears in the record this may be due either to (a) oversight, or (b) an implied finding that the fact is immaterial. The second will be assumed to be the case in the absence of other evidence.

(5) All facts which the judge specifically states are immaterial must be considered immaterial.

(6) All facts which the judge impliedly treats as immaterial must be considered immaterial.

(7) All facts which the judge specifically states to be material must be considered material.

(8) If the opinion does not distinguish between material and immaterial facts then all the facts set forth must be considered material.

(9) If in a case there are several opinions which agree as to the result but differ as to the material facts, then the principle of the case is limited so as to fit the sum of all the facts held material by the various judges.

(10) A conclusion based on a hypothetical fact is a dictum. By hypothetical fact is meant any fact the existence of which has not been determined or accepted by the judge.

Julius Stone, The Ratio of the Ratio Decidendi, 22 Modern Law Review 597, 603-608 (1959)

. . . If the ratio of a case is deemed to turn on the facts in relation to the holding, and nine facts (a)-(j) are to be found in the report there may (so far as logical possibilities are concerned) be as many rival rationes decidendi as there are possible combinations of distinguishable facts in it. What is more, each of these "facts" is usually itself capable of being stated at various levels of generality all of which embrace "the fact" in question in the precedent decision, but eac~ of which may yield a different result in the different fact-situation of a later case. The range of "facts" of Donoghue v. Stevenson, [a 1932 House of Lords decision imposing liability upon the manufacturer of an opaque bottle of ginger beer found to contain a dead snail, for injury (shock and gastro-enteritis) to the plaintiff, a Scotch widow who drank from the bottle given her by one who purchased it from a retailer who in turn purchased it from the manufacturer–Editor] standing alone, might be over-simplified into a list somewhat as follows, each fact being itself stated at alternative levels.

(a) Fact as to the Agent of Harm. Dead snails, or any snails, or any noxious physical foreign body, or any noxious foreign element, physical or not, or any noxious element.

(b) Fact as to Vehicle of Harm. An opaque bottle of ginger beer, or an opaque bottle of beverage, or any bottle of beverage, or any container of commodities for human consumption, or any containers of any chattels for human use, or any chattel whatsoever, or any thing (including land or buildings).

(c) Fact as to Defendant's Identity. A manufacturer of goods nationally distributed through dispersed retailers, or any manufacturer, or any person working on the object for reward, or any person working on the object, or anyone dealing with the object.

(d) Fact as to Potential Danger from Vebicle of Harm. Object likely to become dangerous by negligence, or whether or not so.

(e) Fact as to Injury to Plaintiff. Physical personal injury, or nervous or physical personal injury, or any injury.

(f) Fact as to Plaintiff's Identity. A Scots widow, or a Scotswoman or a woman, or any adult, or any human being, or any legal person.

(g) Fact as to Plaintiff's Relation to Vehicle of Harm. Donee of purchaser, from retailer who bought directly from the defendant, or the purchaser from such retailer, or the purchaser from anyone, or any person related to such purchaser or other person, or any person into whose hands the object rightfully comes, or any person into whose hands it comes at all.

(h) Fact as to Discoverability of Agent of Harm. The noxious element being not discoverable by inspection of any intermediate party, or not so discoverable without destroying the saleability of the commodity, or not so discoverable by any such party who had a duty to inspect, or not so discoverable by any such party who could reasonably be expected by the defendant to inspect, or not discoverable by any such party who could reasonably be expected by the court or a jury to inspect.

(j) Fact as to Time of Litigation. The facts complained of were litigated in 1932, or any time before 1932, or after 1932, or at any time.

Let us first consider the question of "materiality" apart from any view on that matter "explicitly" or "implicitly" manifest in the precedent court's opinion. As to none of these facts (a)-(j), and as to none of the several alternative levels of statement of each of them, could it be said on the basis of the report of Donoghue v. Stevenson alone that it was on its face not "material" (in the logical sense) to the holding in that case. Even as to the time of litigation, as to which we are most tempted to say that this at least must be "immaterial" on the face of it, we must be careful to avoid a petitio principii.^h Are we really prepared to assert with dogmatism that Donoghue v. Stevenson should have been and would in fact have been, so decided in 1800?…

Does it then overcome this difficulty to define "materiality" as Professor Goodhart in effect does, in terms of the precedent court's explicit or implicit assertion as to which of facts (a)-(j) are material? Or to insist that the question, What are "material facts" by which we determine the prescriptive ratio of a case? is always to be determined according to the view of the precedent court, and not according to the view of the later court or observer. (Indeed, in defending his position in 1959 [(1959) 22 Modern Law Review] this distinction becomes almost its central bastion.) Yet there will often be the gravest doubt as to what facts the precedent court "explicitly or implicitly" "determined" to be material. There will often be inconsistent indications from what is expressed or implicit, even in a one-judge court. Such inconsistencies as between the concurring judgments in appellate courts are notoriously also a consistent and fruitful source of legal uncertainty and change .…

Yet these are not the most crucial difficulties with Professor Goodhart's system. The crucial ones arise rather from the several alternative levels of statement of each "material fact" of the precedent case, ranging from the full unique concreteness of that actual case, through a series of widening generalizations.…

. . . Is it reasonable to assume that courts using language appropriate to the case before them do, or could, address themselves in their choice of language to all the levels of generality at which each "material" fact (a)-(j) of the concrete case is capable of statement, not to speak of the possible combinations and variations of these facts, and the implications of all these for as yet unforeseen future cases? . . . And to admit also that level which might be "Implicit" in the former judgment would in most cases be merely to impute to the precedent court a choice of levels of generalised statement (and therefore of the reach of the ratio in the instant case) which must in reality be made by the instant later court . . .

Notes and Questions

1. For what purposes might a lawyer or law student want to know the "rule" of a case? Think of at least three. Would you expect the best possible statement of the "rule" of the case to be the same for each one?

2. Can you see a way or ways in which Professors Oliphant and Goodheart view "the law" differently? Note that Professor Oliphant was writing in the United States, which contained at the time more than 48 separate, sovereign jurisdictions, and where all courts have the power (however infrequently they may exercise it) of overruling their own prior decisions. Professor Goodheart was writing in England, where at the time the House of Lords (the highest appellate court in the country) took the position that it had no power to overrule its prior decisions.

3. Imagine a totally corrupt system of courts, in which the winning party in any lawsuit was the party who offered the judge the largest money bribe for decision. Imagine, in addition, that the judges nevertheless wrote opinions, just as judges do now, purporting to explain the results of the cases in terms or principles of law (making no reference to the bribes). What would Llewellyn view as the "holding" of a case in that system? Oliphant? Goodheart?

4. Professor Oliphant adverts, in his final paragraph, to cases involving the validity of promises not to compete. He asserts that the cases "are considered in square conflict," speaking of the principles stated in the opinions. He also asserts that in fact they stand, unanimously, for the proposition that an employer cannot enforce upon an employee a promise not to compete, but one who buys a business to another may enforce such a promise if the seller made it. How would Goodheart view those cases?

5. What psychological assumptions are the authors making about the process by which judges reach their decisions? Are they warranted?

6. Since Goodhart's critics do not reject the precedent system itself, they agree that an appellate court deciding a case should examine prior opinions. Beyond this point, many questions arise to which answers don't come easily. Do they think a court, in examining precedents, should not look for the things Goodhart would have them look for? If so, what do they think the court should look for? What guidance as "rules of law" do they think precedent opinions offer? Is any guidance to come from other sources?