How Judges Reason: The Logic of Adjudication

by Joel Levin

This is an explanation of how individuals think and reason when they are behaving as judges. The end product of judges’ reasoning is often called, somewhat imprecisely, the common law, and its shape and influences are varied and divergent. In some ways, the flow of judicial reasoning is predictable in the way any reasoning system can be predicted. One looks to implication and consistency, and adds rules of inference, elegance, and simplicity. However, there are two significant reasons why judicial reasoning is not only unusual, but complex and difficult to predict.

First, getting legal concepts and lines of precedent straight and in good order is not enough. Justice and fairness always count. Even the basic terminology of the legal system reflects this. There are Justice Departments, ministers of justice, and (architecturally) justice centers, halls of justice, and justice buildings; but there are not rule departments, ministers of legal standards, precedent centers, halls of strict construction of past decisions, or norm buildings. There was a Justice Holmes and Justice Cardozo, but never a Chief of Law-and-Order Holmes or Associate Stare-Decisis-Enforcer Cardozo. If a rule offends fairness, neither consistency nor years of hallowed history bars its reconsideration. For example, under old English law, a debtor buying a house, making 29 years and 11 months of payments on a 30-year mortgage forfeited all monies if the final payment was one day late. So offensive to justice was this rule that an “equity of redemption” was imposed to allow the debtor to keep, at least, his past equity paid.

Second, legal reasoning is irreducibly democratic. The influences of individual judges, and the lawyers and litigants who influence them, count without regard to a decision’s improbability, unreasonability, inconsistency, or even stupidity. This is unlike other areas where, for example, my view of quantum mechanics and the view of a Cal Tech physics professor would never be placed on the same footing. If both of us were appointed to a court, however, then at least in one important sense, our decisions would each count, and count significantly (although not necessarily equally).

* * * *

Law is a puzzle. To the ordinary individual, uninitiated in the mysteries and ritual of the legal process, the variety and complexity of law in modern societies can be bewildering. His reaction to any particular legal situation is likely to vary widely, depending on which aspect of law he encounters. A policeman accusing him of theft; a store manager refusing to allow the return of an unsatisfactory purchase; an executor informing him of a piece of land now his; a zoning clerk demanding the removal of a newly and dearly acquired nude statue; a tax collector scrutinizing with disapproval the deduction recommended by his broker; an attorney advising him of a claim brought against him by the milkman crippled by his daughter’s rollerskate: each elicits a different reaction. Moreover, it is unlikely that the ordinary observer could readily formulate a set of beliefs or sketch a general and unified conceptual picture of how these different legal bits might fit together, and what he thinks about them as a whole.

Yet, it is even more unlikely that this observer would be without any cohesive set of ideas and opinions at all. His stock of beliefs is likely to reflect his experience and his reflectiveness. If forced to distill his ideas and to formulate a general statement concerning law, and further pushed to focus his attention on the judicial aspect of law – say, by making him a party to a suit – he would likely discuss the question of the freedom of judges to decide cases. He might settle on this issue simply because it is such a popularly controversial one, or he might see it as a necessary first step in constructing a coherent justification of the various fragmentary, eclectic, and possibly opposing ideas he holds. He might reason that to answer divergent legal questions, it is first necessary to understand under what constraints those professionally called upon to answer these questions labor.

In addressing the issue of judicial constraint, this ordinary individual could embrace either of two popularly held positions (or he might embrace both, although they are contradictory, with the knowledge that he is in a dilemma but unable to see his way out). He might think that judges are highly bound by precedent, statutes, regulations, or just generally, rules, and that in any situation, some specific, ascertainable rule applies. He, as a citizen, and a judge as a judge, each simply follow the rules. Law is complex, and rule ascertainment is not always easy. However, if one had the time, energy, and expertise – if one were, in Ronald Dworkin’s term, Herculean – one could discover the rule with some certainty.1 The role of the judge is to acquire the expertise and expend the necessary time and energy.

The alternate position would also start with an acknowledgment of the existence of a myriad of statements, precedents, regulations, and rules. It would look to their vagueness, their ambiguity, and their incompleteness and would measure those against the need to answer the questions that arise in adjudication. Given these difficulties, the alternate position would say that judges may do pretty much as they please. It might also point out that this is what judges do anyway. Judges must justify their positions, but justifications are easily and sophistically manufactured.

These two positions are in fact widely held, often by the same individual when he initially (pre-theoretically) considers different situations. That such a split is so wide could be passed off as a demonstration of the general ignorance and confusion of the legal laity (or at least half of it, if one position is sound). The difficulty, however, is that such a split is reflected in lawyers’ briefs and judges’ decisions, as well as in learned treatises and jurisprudential discussions. The split has been incorporated in a sophisticated way into a whole judicial literature, and has given rise to schools of advocacy, as well as schools of reconciliation.

Trying to analyze and ultimately to resolve this issue leads to a focus sufficiently abstract to encompass the breadth of the law. Abstraction may be the luxury of the fanciful theorist in a society stable, simple and homogeneous. Perhaps mountain herders do not require a very elaborate theory to understand the proper decisions and criteria for those decisions pronounced by the legal remedy-agent or judge. Modern society – however complex, heterogeneous, changing and law-laden – makes theory necessary rather than luxurious.

The purpose of abstraction or theory, then, must be to allow either the judge hearing a case or appeal, or the interested layman or lawyer addressing the issue facing the judge, to reach a decision. Clearly, however, not just any decision will do. There may be a certain relativism concerning the power or authority of a judge: regardless of integrity, ability or understanding, the fact of the decision remains. But this fact is not enough. Judges employ criteria; they may be themselves judge for their ability to make good or correct or coherent decisions; and, most importantly, society at large behaves in accordance with some expected set of decisional outcomes and criteria which should be used in reaching those decisions.

Understanding which outcomes are correct or what generally is to count as law requires criteria for reaching those outcomes. A three level system is needed to explain judicial reasoning. Three levels are employed because two different kinds of criteria are used to yield or produce answers to questions that can come before the courts. The answers themselves occupy the first level. Criteria that furnish the answers populate the second level. The larger beliefs and attitudes which tell what criteria can be used to reach a legal decision (as opposed to those respecting taste, politics, morality, religion, etc.) populate the third level. Here, the members of this third set will generally be called “justifications”.

The claim here is not that this tertiary schema is descriptive of how judges or those answering questions addressable to judges actually reason, at least not always. Moreover, this claim is not a prescriptive one telling how such reasoning should work if it is done right. Rather, the three level schema and its attendant theory of judicial pluralism provide a framework for analyzing judicial reasoning. Such reasoning is perceived to be done well or badly, to have reached correct or incorrect conclusions, to have been based on defensible or indefensible reasons, and to be expected or surprising. Society and its members look at judicial decisions and are able to comment on and debate them. An explanation of the ability to do so is in order.

However, before setting forth the propositional schema, several caveats should be given. First, the fact that some proposition can be found on one level does not exclude it from inclusion on another level. The principle of sovereign immunity – “the king can do no wrong” – might be stated as an answer to a tort question to a judge facing the issue of liability to a vassal gored by a royal bull (first order). It might be the criterion used to decide issues of civil liability in its form of a general legal principle or a limitation on a court’s jurisdiction (second order). Finally, it might be believed, either because of divine right or as part of a general political theory concerning the separation of powers (third order).2

The same proposition is used in all those levels (or orders), but it is being used in a logically different way. There is no reason to be concerned that such varying placement occurs or that a single proposition might do different kinds of work in the same schema if called upon to do so. Setting out all possible propositions is not the point, any more than it would be in physics. Rather, the propositions – as answers, criteria and justifications – are there (and ascertainable) when needed.

This brings up the second, and related, point. There is no magic in the number three. There could be more or fewer levels. It is a basic truth in mathematical logic that any higher order set can be incorporated into a lower order set. More obviously, the number of levels could multiply as each possible reason becomes a ground for a new level. The breadth of the reasons would then determine the number of levels.

The reason for three is that it reflects the major types of reasons one intuitively employs. Whether, for example, the language of a statute is embodied in a first or second order proposition is largely irrelevant. In general, statutes provide criteria for answering justiciable questions. If one draws the second order more abstractly to include only statutory law rather than individual statutes (i.e., all statutes but not a listing including, e.g. The Sherman Antitrust Act)3, that is unimportant. In general, it would be interesting only if statutes are sought to answer questions, and are sought because of certain third order political or social beliefs.

The final caveat concerns the existence of a third level. It might be thought that, as the function of the third or justificatory level is solely to yield the criteria of the second order, its content is trivial. Who could doubt that what have at least since Gray been called “sources of the law” 4 – statutes, past decisions, constitutions, etc. – make up the criteria, pure and simple. However, even if that were so (and many controversies about the role of disputed sources such as custom, desuetude,5 and the role of equity throw such a contention into doubt), the ordering among the criteria would still need to be given.

1The term “Herculean” comes from Ronald Dworkin’s article “Hard Cases,” 88 Harv. L. Rev. 1057 (1975), reprinted as Chapter 4 of Taking Rights Seriously (1977), discussed generally in this work in Chapter VI.

2The principle of sovereign immunity is that no state or government may be sued without its own consent. This doctrine originated with the monarchical, theological, and tyrannical divine right tenet of “the King can do no wrong.” Throughout its existence, it has been pressed into service to justify a number of otherwise unjustifiable (and usually unjust) decisions. It has even reached constitutional status in a nation (the U.S.) otherwise not known for favoring royal prerogatives. See, for example, The Schooner Exchange v. McFaddon, 7 Cranch 116 (U.S.) (1812); Keifer & Keifer v. Reconstruction Finance Corp., 306 U.S. 381, 59 S. Ct. 516 (1939); Monaco v. Mississippi, 292 U.S. 313, 54 S. Ct. 745 (1934). The principle also lives in reciprocal agreements between nations that citizens of one nation cannot sue a second nation without its permission. See, for example, 28 U.S.C. Section 1604.

315 U.S.C. Section 1-7.

4John Chipman Gray, The Nature and Source of the Law (2nd ed. 1921).

5It should be noted that while the formal application of desuetude – which is the disuse or discontinuing application of a statute – is relatively rare, the finding of exceptions and implicit partial or complete overruling of statutes is, if not routine, far from uncommon.

Back to publications.