Some puzzles about the rule model

We have suggested that, at least under some conditions, judges can best settle controversy within a community, implement community values, and provide coordination for private action by announcing determinate rules to govern future judicial decision-making and treat­ing prior rules as governing their decision-making.

5.4.1 Making rules

Judges often state rules in their opinions that are intended to be, or at least taken to be, authoritative for later judges.^[54] Typically, how­ever, judicial rules are narrower in scope than many legislative rules. Legislatures issue rules in response to social and economic problems that come to their attention in a variety of ways, including the demands of constituents, the need to respond to unanticipated problems, or the desire to plan for the future. Their rules normally are prospective and as comprehensive as necessary to solve the problems that motivate them. In contrast, to the extent that judges issue rules at all, they do so in the context of particular disputes brought before them by litigants. Most often, judicial rules appear in the course of a description and analysis of questions raised by the current litigation. The rule applies retroactively to the parties before the court, and the aim is to resolve that dispute and similar future disputes, without establishing a policy to govern a broad field of activity.

One question about the rule model of judicial decision-making is why the rule-making practices of judges differ from those of legislatures.

Conceptually, there is no barrier to a court's announcing a prospective rule, applicable to future cases only. Suppose, for example, that the judge has before her a nuisance case involving a pet bear in a residen­tial neighborhood. No previous nuisance cases have raised the problem of wildlife kept as pets in residential areas. The judge reflects on the situation and reaches three conclusions. First, in the great majority of cases, pet bears pose an unacceptable risk to those who live in their vicinity. Second, this is a context in which the need for settlement and coordination supports a determinate rule: pet bears in residential neighborhoods are nuisances per se. With a rule in place, neighbors will not need to litigate whenever an owner brings home a wild pet, and pet owners will not make financial and emotional investment in unacceptable pets. Third, the defendant's bear is a friendly, declawed animal that has spent its life in captivity, poses no threat to neighbors, and should not be removed from its accustomed home.

At this point, assuming plenary judicial power to make rules, the judge has at least three options. She can reject the current nuisance action, allow the bear to stay, and decline to issue a rule. She can announce a rule that pet bears in residential neighborhoods are nuisances per se and order removal of the bear. Or she can announce a per se rule but specify that it applies only prospectively, that is, to future bears.

There is a case to be made for the third choice. The judge has, in the course of hearing arguments and considering evidence, gathered infor­mation that suggests not only who should win but also what a good rule would be, in light of a pattern of past cases mentioned by parties. She may conclude that she can both decide fairly between the parties (bear wins), and also use the general information she has acquired about bears in residential neighborhoods, in combination with her own reasoning skills, to announce a no-bear rule that will serve sig­nificant settlement and coordination purposes in the future. So why not split her decision between rule and outcome? By doing so, she can achieve settlement and provide coordination for the future, while also saving a seemingly harmless bear. Maybe we should distrust the judge's decision to exempt the bear on the ground that it may reflect cognitive bias toward salient facts. But assume the court has already considered this possibility, discounted her conclusions accordingly, and continues to believe that she should make the third choice in this case, saving this bear but banning future bears. This looks like a reasoned decision, and there is no obvious reason why it is not within the judge's power to make it.

Nevertheless, most people familiar with common law practice would be surprised if the judge made the third choice. Courts simply do not decide a case one way and at the same time announce a rule for future cases that would require a different result in the current case. Occasionally, and controversially, they may overrule existing judicial rules prospectively, following the rule in the current case but prescrib­ing a different result for the future (Levy 1960; Eisenberg 1988).^[55] Prospective overruling, however, at least has the virtue of protecting litigants who have relied on a rule that was in effect at the time they acted; in the case of the bear there is no existing rule on which the bear owner might have relied.

Another common practice among judges is to avoid announcing rules that stray too far from what is needed to resolve a current dispute. Rules, by nature, are general and determinate; therefore, it is not pos­sible to limit the predicates of judicial rules to the facts of particular disputes. Nevertheless, judges tend to avoid rules that extend signifi­cantly beyond the exigencies of the case, and when they do announce broad rules, other courts may call the rules ‘dicta' and refuse to give them authoritative effect (see Burton 1995; Simpson 1961).

For example, suppose again that a judge is presiding over a case involv­ing a pet bear in a residential neighborhood. In the course of hear­ing arguments and examining past cases, she becomes interested in a broader range of nuisance problems in residential neighborhoods, and particularly the problem of lawn-mower noise. She concludes many disputes among homeowners could be solved by a rule limiting lawn-mowing to the hours between two and four o'clock on Saturday afternoons. Assume there are no constitutional or legislative limita­tions on judicial rule-making that prevent the judge from issuing a rule that in the future, lawn-mowing outside these hours is a nuisance per se. Nevertheless, the judge might hesitate to announce a lawn-mowing rule.

In this case, the explanation may lie in natural reasoning. Under the rule model of judicial decision-making, both the judge's resolution of the case before her and her decision whether to announce a rule for future cases are governed by natural reasoning, and natural reasoning itself may tend to limit the scope of judicial rule-making. On full reflec­tion, the judge may decide not to announce the rule, both because it is outside the scope of her primary role as adjudicator and because there is no reason to think that the parties presenting evidence in her cur­rent case have gathered a full set of relevant facts about the costs and benefits of lawn-mowing at various hours.

5.4.2 Identifying precedent rules

A second question about judicial rules is: what acts and statements by prior judges count as precedent rules? The primary functions of rules are to settle controversy and to enable actors to coordinate their conduct, based on the assumption that most or all will follow the rule. To perform these functions effectively, rules must be general enough to govern classes of cases and determinate enough to be understood without reference to the values they are designed to serve. A judicial statement that does not meet these criteria may operate as a decisional standard, but it is not a rule for our purposes. A rule, as we use the term, must be authoritative in the sense that it preempts further delib­eration and mandates certain results.

Legislative rules normally are easy to identify, simply by tracing the process of their enactment. Judicial rules are harder to recognize, both because courts traditionally have been reluctant to legislate overtly, and because judicial opinions tend to focus on the immediate task, which is to resolve a particular dispute. A typical opinion contains a narrative description of the dispute, references to past cases, and a more or less complete explanation of the court's reasoning. The court may never explicitly announce a rule for future cases. Nevertheless, the equivalent of a posited, authoritative rule is often detectable in explanatory remarks and citations. As long as the court has a rule in mind and the rule is capable of restatement in determinate form, posit­ing that rule can occur in an informal way.

As noted, judicial rules are often implicit, rather than explicit, in judi­cial opinions. Although judicial rules are not always laid down explic­itly in the manner of a legislative rule, it is important to distinguish between rules implicitly posited by prior judges and rules that later judges draw, or ‘abduce', from the data of past decisions.

Here is a simple example: a judge hears a case in which homeowners seek an injunction against the opening of a drag-racing track, on the ground of nuisance. The plaintiffs cite a prior case in which a judge enjoined the operation of an outdoor paintball arena on the ground that the paintball arena would constitute a nuisance in a residential neighborhood. In her opinion, the judge in the paintball case stated that neighbors had presented substantial evidence that paintball is a noisy sport. She also cited cases holding, respectively, that a weekly outdoor trumpet concert, an amusement park, and an ice cream truck were nuisances in residential neighborhoods but a daily outdoor chess tournament was not, and she mentioned that these enjoined activities were noisy. The judge in the case of the drag-racing track might infer that the judge in the paintball case intended a rule, ‘noisy activities occurring on a regular basis in a residential neighborhood are nuisances'.

By comparison, suppose that there is no paintball case, and no similar case in which a prior judge referred to noise. Instead, the judge in the case of the drag-racing track finds an array of past cases holding, respectively, that a weekly trumpet concert, an amusement park, and an ice cream truck were nuisances and a chess tournament was not. In this case, the judge has no basis for inferring a pre-existing rule against noisy activities because no prior case mentions noise as a relevant concern. She can posit a rule to this effect, but there is no precedent rule in effect because there is no evidence that any single prior judge (or judicial panel) intended to establish a rule against noisy activities.

A related question is how to interpret judicial rules. As we discussed at length in Chapter 2, we equate the meaning of any rule with the meaning the rule-maker intended. Creating an authoritative rule is an intentional act. At least ideally, it occurs after the rule-maker has brought her expertise and her powers of reason to bear on both the problem to be solved by the rule and the form an authoritative solution should take. It follows, for us, that in order to preserve the lawmaker's authority and capture the exercise of reason that went into making the rule, the meaning later courts give to the rule should be understood as the meaning the rule-maker intended to communicate through the language used in the rule.

These general principles apply to judicial rules as well as legisla­tive rules. Judicial rules come into existence when they are posited by a judge and they mean what that judge intended them to mean. Sometimes intended meaning will include consequences of the rule's words that the judge did not consider and that do not further the pur­pose the judge had in mind; these consequences are nonetheless a part of the intended meaning of the rule.

A further difficulty, which judicial rules share with legislative judicial rules, is that in many cases the rule-maker will be not one judge, but a panel of judges. As we explained in Chapter 3, our understanding of intended meaning leads to the conclusion that unless a majority of judges agreed on a particular implication of the rule's words, that implication is not within the meaning of the rule. Fortunately, the practice among judges of writing majority, concurring, and dissenting opinions helps to clarify what can count as an intended meaning.

5.4.3 Limiting judicial rule-making

Beyond what we have discussed here, the rule model of judicial deci­sion-making places no limits on judicial power to announce rules. Yet, not all judicial rules will be good rules. The judge who posited the rule may have been distracted by the compelling facts of the case before her, or the parties may have failed to provide sufficient evidence about the future consequences of the rule.^[56]

One possible response to the problem of unsatisfactory common law rules would be to place limits on judicial rule-making power. Judicial rules might be open to disqualification if parties to a later litigation established that the rule-maker did not deliberate adequately before issuing a rule. Alternatively, judicial rules might be treated as non­binding unless and until they passed a test of acceptance over time or were ‘taken up' by a sufficient number of future judges (see Postema 2002; Simpson 1973). In effect, the process of acceptance would elevate the rule from one judge's non-binding statement to the authoritative product of collective wisdom.

Both these strategies, however, are inconsistent with the idea of a rule as a means of settling controversy and supporting coordination through an exercise of authority. The first possibility, in which the prior judge's deliberation leading to a rule would be subject to later review, is inconsistent with the nature of rules: a rule is an authorita­tive directive to be followed without further reflection. The second, in which the rule gains authority through acceptance, is inconsistent with the character of a rule as an intentional act. In the next section we consider instead an alternative way to limit the effect of questionable precedent rules, by giving later judges a power to overrule. Of course, overruling also undermines the settlement and coordination value of judicial rules. But it does so in a more straightforward way and there­fore results in a more rational picture of judicial authority.

5.4.4 Overruling precedent rules

As just described, we view a judicial rule as an authoritative rule from the time it is announced, subject to the power of later judges to overrule. Judicial power to issue binding rules implies that, as long as the rule is intended to have the effect of a rule and is sufficiently determinate to operate as a rule, its status as a rule is not subject to gainsaying the quality of the judge's reasoning or to a requirement of acceptance by subsequent courts. However, judicial power to make rules also implies a power to make new rules. Thus, subject to whatever hierarchical ordering may be in place among courts, judges can override rules they find to be flawed. In their role as adjudicators, they must follow prior rules even if they believe the underlying reasons for the rule require a

parties); State v. Shack, 277 A.2d 369 (1971) (announcing a limitation on owner’s right against trespass, although no active party represented the owner). different result in the case now before them. But in their role as rule­makers, they can eliminate prior rules or replace them with new rules.

The rule model would not be successful if future judges lacked the power to overrule. Because judges are primarily adjudicators and therefore prone to give more weight to immediate facts than to long­term consequences, their rules may be poorly conceived. Alternatively, well-conceived rules may become obsolete as circumstances change. It follows that judicial power to make authoritative rules, without judicial power to overrule authoritative rules, would often lock in error.

Overruling of precedent rules is appropriate in just two circumstances. First, for the reasons just mentioned, the rule may not be justified as a rule. In other words, it may cause more error over time by prescribing wrong outcomes than it prevents by settling controversy or enabling actors to coordinate their conduct. Second, the rule may be justified as a rule and yet not be optimally effective: it may prevent more error than it causes, but prevent less error, or cause more error, than an alternative rule. If instead, the rule is a good rule and no better rule is available, but it generates what appears to be the wrong result in a particular case, overruling is not appropriate. Errors of this type are an inevitable feature of any rule and not a reason to eliminate or replace the rule.

Suppose, for example, a judge has before her a claim of nuisance against the owner of a pit bull. A prior case announced the rule that ‘pit bulls in residential neighborhoods are nuisances per se'. If this rule is superior to any alternatives, the judge should apply it even if she determines that the pit bull in her case is a gentle, well-trained animal unlikely to cause harm. If, on the other hand, evidence indicates that the prior rule was based on faulty reasoning by a judge presiding over a rare case of mauling, she should overrule the prior rule, either by announcing a modified rule, or by declaring that per se rules are not appropriate for the regulation of dogs.

Under a rule model, the power of judges to make and unmake judicial rules is coextensive with the power of a legislature to make and repeal legislation. Traditionally, however, judges have been reluctant to over­rule precedent rules without strong reason to do so (see Summers 1997; Lamond 2005). This may be a sensible attitude for judges, even if their power to overrule is plenary.

The dilemma judges face in dealing with precedent rules is that, unlike legislators, judges are both rule-followers and rule-makers. As rule­followers, judges must treat the rules as authoritative and follow them as written, without consideration of the reasons that lie behind them. As rule-makers, judges have power to overrule existing rules, and in deciding whether to exercise this power they must consider whether the rule in question, as applied to all cases that fall within its terms, will serve the reasons that support it. The difficulty is that if judges consult the reasons supporting the rule for the purpose of deciding whether to overrule it, they will be tempted to consider also whether the outcome of the rule in the particular case before them serves those same rea­sons. This is not the right question: if the rule is a good rule over the run of cases to which it applies, the judge should leave it in place.

To illustrate, suppose a prior court has announced a rule that pit bulls are nuisances per se in residential neighborhoods, based on a balance between the welfare that owners derive from their dogs and the risk of injury to others. Once this rule is in place, a later judge cannot decide for a particular owner on the ground that in this case, welfare outweighs risk. But the judge is free to decide that overall, the welfare pit bulls provide to their owners outweighs the risks they present to neighbors. In theory, these two conclusions are distinct: one concerns the outcome of the current case, and the other concerns the overall performance of the rule. In practice, however, it will often be difficult for the later judge to compartmentalize in this way.

One possible solution to this problem is a rule governing occasions for overruling. A court (or a legislature) might announce a rule such as ‘overrule precedent rules that have not been followed for thirty years', or ‘overrule precedent rules that have been questioned by later judges in ten or more cases'. One difficulty with rules of this kind is that they may function well in some areas of law but not in others: thirty years is a long time in tort law but not in property law. Another problem, at least if the overruling rule is itself a judicial rule, is a problem of regress. The overruling rule itself assumes that judges have a general power to overrule rules; if so, there is no reason why a later judge could not overrule the current overruling rule and announce a substitute. Any subsequent judge could do the same, leaving no source of stability in the handling of judicial rules.

It appears, then, that in the area of overruling, natural reasoning is complicated by the dual role of judges, and rules for overruling are unlikely to be effective. Probably the best response is an overruling practice by which judges overrule precedent rules if but only if they are significantly flawed. This is not a rule; nor is it a direction for unqualified exercise of moral and practical reasoning. Instead it sets a standard (‘significant') that relies on judges' own sense of significance. In other words, it delegates the process of reasoning about precedent rules to future judges, subject to the proviso that they should look only for flaws that they deem significant. A solution of this type is not ideal, but it appears to be the only practical alternative to never overruling precedent rules, or freely discarding precedent rules.

The problem of when courts should overrule obsolete or suboptimal precedent rules arises not only in the context of common law adjudica­tion but also in the course of applying canonical texts such as constitu­tions and statutes. Courts establish precedent rules under canonical texts in two ways. First, if the text in question is itself a comparatively vague standard that amounts to a delegation of power to engage in natural reason within boundaries set by the text, courts applying the standard may translate it over time into more determinate rules. A later court, however, might conclude that these interpretive rules should be overruled on the ground that they do not properly reflect the meaning of the statute. Second, if the text itself takes the form of a rule, but the words of the rule contain an ambiguity, a precedent court may adopt a new formulation of the rule, intended to clarify the original. In this case, a later court might conclude that the precedent court has misinterpreted the original canonical text and consequently overrule the prior judicial rule. Both cases raise the question of what attitude the later court should take toward overruling the earlier inter­pretive conclusion: should the later court overrule the prior judicial clarification whenever it believes the clarification is not faithful to the text, or overrule only when it also concludes that errors caused by the prior court's mistake are greater than the loss of settlement value that overruling will cause, or only when the earlier court's error was ‘significant'?