Reasoning from legal principles

We have argued that analogical reasoning and related approaches to precedent decisions cannot be based purely on factual similarity or dis­similarity between particular past and present cases, without reference to supporting principles that give substance to similarity.

Legal principles differ from legal ‘reasons' of the type described in the previous section. At least ideally, legal reasons are reasons that explain particular past cases and are discoverable from the facts and outcomes of those cases. Current judges may be able to avoid legal reasons estab­lished in precedent cases, either by distinguishing them in the context of a current case or, possibly, by overriding them. In theory, however, they are reasons put into place by past judges. Legal principles, in contrast, are the work of current judges, who survey and interpret the body of past decisions that appear to bear on a current case. They are present judges' best exposition of what has happened in the past.

The idea that judges decide cases by reasoning from legal principles has a venerable history and a strong resonance for many lawyers and judges (see Hart and Sacks 1994; Pound 1922; 1940). Bankowski et al. describe legal principles as:

principles partly emergent from practice and custom, partly constructed out of moral or ideological elements that bring together practice and contempo­rary values in a coherent order...

Law students are taught to reason in this way, judicial opinions often follow this pattern, and traditional academic commentary uses related methods to explain existing law and propose reform. In this section, we will argue that the idea of legal principles has played a leading part in the mystification of legal reasoning. In our view, legal principles, like case-to-case analogies, are fallacious justifications for current results.

6.2.1 Dworkinian principles

Possibly the best-known account of the process of reasoning from legal principles comes from Ronald Dworkin (Dworkin 1986, 230-2, 254-8; Dworkin 1978, 22-31, 115-18). Dworkin described legal principles as the morally best principles capable of explaining a substantial percent­age of legal decisions. In Dworkin's account, legal principles must meet two criteria. The first is a threshold requirement of ‘fit' with existing legal materials, including legislation and prior judicial decisions. The second is that the resulting principle must conform to moral ideals as closely as is possible, consistently with the requirement of fit. In other words, a well-crafted legal principle is the product of adjustment between two potentially incompatible criteria: precedent and reason.

Legal principles are not rules and do not dictate outcomes in the manner of rules. Instead, they are ‘starting points' for legal decision­making. In the case of a rule, only one rule can provide the answer to a legal question: if multiple rules appear to govern a case, the judge must determine which one of them is authoritative.

Although no legal principle carries the force of a governing rule, legal principles are authoritative in the sense that the combination of prin­ciples pertinent to a current case, weighed as the judge deems appro­priate, determines the judge's decision. Another important feature of legal principles, at least in Dworkin's version of the common law, is that only legal principles determine outcomes. Other legal materials, such as rules or standards stated in past opinions, are simply data points from which the judge constructs the legal principles that will provide an answer in her case.^[60] In effect, principles immanent in past legal materials and revealed by current judges, constitute the common law.

The primary way in which judges make use of legal principles is to derive a solution to a new case from a collection of past decisions. Suppose, for example, that a judge is considering a nuisance case against a prop­erty owner who plans to turn a shuttered, indoor, retail space into a recreational facility featuring giant floor-level trampolines. The parties to the suit refer to a number of prior nuisance cases set in mainly resi­dential neighborhoods: in one line of cases, courts enjoined an outdoor archery range and an outdoor paintball arena but declined to enjoin an outdoor tennis club, an indoor bowling alley, a golf course, and an indoor rifle range. In another line of cases, courts ordered homeown­ers to remove a pet bear, a pet ocelot, and a pet crocodile from their homes. In another, courts permitted a day care center, a halfway house serving non-violent parolees, and a well-maintained sewage treatment plant to operate. The opinions in some of these cases referred to the possibility of increased noise and traffic, sometimes stating that these difficulties were unavoidable in modern life, and sometimes expressing concern about congestion and danger to pedestrians.

In addition to supporting current judgments, legal principles can pro­vide a basis for avoiding rules announced in past cases. Suppose that the court in the trampoline case studies the prior case in which a court case enjoined an archery range from operating in a residential neighborhood. The archery range was next door to a public school, and the case arose after a stray arrow came dangerously close to hitting a child walking to school. In its published opinion, the precedent court announced a rule that ‘recreational facilities within one mile of a public school are nuisances per se unless maintained by the school for the use of students'. The proposed trampoline facility appears to fall within the scope of the rule announced in the archery case; it would be located three quarters of a mile from the school (although not on the same street), and it would not be maintained as a school facility. Thus, it falls within the scope of the precedent rule; and if the rule is authoritative, the judge in the trampoline case must enter an injunction.

In a regime governed by legal principles, however, a rule of this type does not operate as an authoritative rule but only as evidence sup­porting a legal principle of indeterminate weight.

Accounts of what sort of legal data support a legal principle vary among commentators. Dworkin, for example, states that ‘fitting what judges did is more important than fitting what they said' (Dworkin 1986, 284). Thus, at least in the context of judge-made common law, only the facts and outcomes of past decisions are relevant, and current judges are free to disregard rules or particular statements that appear in prior opinions. On this understanding, legal principles are similar to forms of analogical reasoning in which judges extract ‘reasons' for decision from past facts and outcomes. Unlike legal reasons, however, legal principles are formed from an amalgam of multiple past determi­nations and constructed by the current judge.

Other accounts of legal principles assume that judges constructing principles will take account of a broader set of legal materials, includ­ing not only facts and outcomes but also rules and principles stated in prior opinions (see Hart and Sacks 1994; Perry 1997). This form of principle-construction, however, should not be confused with deduction from authoritative rules.

Although the form of statements appearing in past cases may affect the extent to which they can function as rules, the main difference between legal principles and rules is not a matter of form but of the role the past statement plays in the current judge's decision. If a cur­rent judge declines to enjoin an indoor trampoline park based on a prior judicial statement that activities that pose no significant risk to health or safety are not nuisances, the judge is following a rule. If the judge reaches the same decision on the ground that past judicial state­ments, together with facts and outcomes of past decisions, support a principle that activities that pose no significant risk to the health or safety of residents are not nuisances, the judge is constructing and applying a legal principle.

Another way to describe the difference between principles and rules is that a precedent rule is posited by a prior judge and remains in place as law until it is overruled or replaced. The prior judge is the lawmaker and exercises lawmaking authority by announcing the rule. A legal principle, in contrast, is not posited but organic. The princi­ples applicable to any given case are generated by the current judge, but the judge does not posit them. Instead, she abduces them from prior material. In the course of formulating legal principles, the judge exercises judgment, but only in the limited sense necessary to extract, assess, and weigh principles immanent in existing legal materials. She does not engage in unconstrained moral and empirical judgment in the manner of a lawmaker. The dimension of fit with prior legal material also means that legal principles may change over time as the body of legal decisions changes.

One consequence of the difference between legal rules and legal prin­ciples is that a rule dictates the results of all cases that fall within its terms. Because rules are authoritative in this way, and also determinate in form, only one rule can govern a particular set of facts. Conflicts among competing rules must be solved by overruling all but one of the competing rules. Legal principles do not dictate outcomes; instead they are reasons that have weight when they come in conflict with other principles. In any given case, the outcome depends on a bal­ance among applicable principles that is carried out by the current judge. Thus, if nuisance cases indicate both that landowners should have significant liberty to determine the use of their property and that landowners should not engage in activities that pose a significant risk to human health or safety, a judge deciding a case brought to enjoin a construction of an indoor rifle range must decide both the extent to which each principle is implicated by the plan and also which principle carries greater weight.

At the same time, legal principles are not moral principles. The dimen­sion of fit requires a legal principle to conform to the pattern of past decisions even if, as a consequence, the resulting principle is morally flawed. Legal principles need only be the morally best principles that pass the threshold of fit.

Thus, given that there will always be past outcomes or past statements that the current judge views as morally or empirically incorrect, legal principles are not ideal principles; they are the most morally attrac­tive incorrect principles that fit the background of prior decisions (Alexander and Sherwin 2001). The reasoning process involved in formulating legal principles is structurally similar to the process of reflective equilibrium: in each case, the judge begins with particular outcomes and formulates a principle to support them. Substantively, however, the two are very different. A judge reasoning to reflective equilibrium begins with her own judgment about a particular problem and then refines it by formulating explanatory principles and testing those principles against additional examples. Throughout the process, the reasoning is her own and she is free to reject any and all particular judgments that cannot be explained by what she confidently believes is a correct moral principle.

In contrast, when a judge formulates a legal principle, her task is to derive a principle from decisions made by past judges, and, on some accounts, from statements in past opinions. Until the threshold of fit has been met, it does not matter whether these decisions and state­ments were right or wrong. Even if not posited in an authoritative sense, they constitute the body of prior ‘law' to which the principle must conform. The principle the current judge formulates must con­form, to some degree that is never fully specified, to the pattern of past legal decisions and statements (Dworkin 1986, 230-1, 255; Dworkin 1978, 116-17). Accordingly, it cannot be, and does not aspire to be, a moral principle: it is the morally best principle that fits a morally imperfect background of law. Nor is it a form of posited law: neither past judges nor the current judge is its author, and there is no reason to believe that either past judges or the current judge would choose it as the correct standard for decision.

Returning to the nuisance example: Suppose that the current judge believes that the morally and empirically best rule for nuisance cases is that property owners must not engage in activities that place sig­nificant new burdens on the surrounding community unless their activities serve community interests in a way that at least justifies the burdens they impose. The activity at issue is an indoor paintball arena, and the current judge believes that paintball has no redeeming social value that would justify an increase in traffic and noise in the area and the negative psychological effects of war games on adolescents. In prior decisions, courts enjoined owners from keeping wild animals as pets and enjoined the operation of an archery range; they allowed a private tennis club, a bowling alley, a private golf course, a trampoline facility, a day care center, a halfway house for non-violent offenders, and a sewage treatment plant. The judge probably can disregard the rifle range as a mistake without violating the requirement of fit. The day care center and the sewage treatment plant may be justifiable on the ground that they serve important community interests. But four of the activities approved in prior cases - the private tennis court, the pri­vate golf course, and possibly the trampoline facility and the bowling alley - appear to impose burdens without significant social benefit. If so, the threshold of fit requires the judge to modify her ideal principle: owners must not engage in activities that impose new burdens without either compensating social benefits or compensating private economic benefits. The judge is left with a legal principle, one that is morally incorrect in her view, but one that is the most morally attractive but morally incorrect principle available. She may also have to tolerate a new paintball arena.

Legal principles may nevertheless be appealing because they seem to allow a compromise between unconstrained natural reasoning and compliance with authoritative rules, and to give meaning to the idea of a common law that is broader than the pronouncements of particu­lar courts or legislative bodies. Judges who formulate and apply legal principles are constrained not only by formal enactments but by limits placed on them by institutional history and legal process. At the same time, application of legal principles does not require judges to set aside moral values or abstain from exercising their power of reason as they decide cases. They must formulate and apply the most morally attrac­tive principles that fit with institutional history, but in doing so they can discard at least some of the past's mistakes.

Legal principles, then, promise a body of law shaped by internal coher­ence. Legal principles allow judges to maintain consistency among past, present, and future decisions and across doctrinal boundaries. A regime of coherent legal principles also has the advantage of providing an answer, although not necessarily a unique answer, to every dispute, while also ensuring that the answer has some basis in pre-existing law. The right outcome is also an outcome that fits what has come before.

Or so it may seem, for those who are willing to accept an almost-right outcome that almost fits what has come before. We do not accept the claims made for this form of decision-making. Even assuming that the compromise involved in reasoning from legal principles is acceptable, we do not believe that legal principles impose meaningful constraints on judicial reasoning. And to the extent that they do constrain judicial reasoning, we believe they reflect a vice, not a virtue, of coherentist approaches to legal decision-making.

6.2.2 Problems of weight and problems of fit

Working primarily from Ronald Dworkin's description of legal prin­ciples, we will demonstrate in two ways that legal principles are not capable of constraining judicial decisions in any meaningful way. In Dworkin's account, judges must construct principles that ‘fit' the pre­existing body of legal materials and then ‘weigh' these principles in the context of a current case to determine what outcome is required. We argue that both the concept of proper fit and the concept of weight lack meaningful content when applied to real problems of adjudication.

‘Fit' is a relationship between legal principles and prior legal practice and decisions, which judges must maintain as a matter of political morality: ‘anyone who accepts law as integrity must accept that the actual political history of his community will sometimes check his other political convictions' (Dworkin 1986, 255). At the same time, the appropriate degree of fit between principles and prior practice is a question to be determined by each judge, case by case. The constraint imposed by the fit requirement ‘is not the constraint of external hard fact or of interpersonal consensus' (Dworkin 1986, 257).

Fit, then, is a vague idea: nothing in the concept itself indicates where the threshold lies. A further problem is that, even assuming that judges generally converge on the degree of fit required, a judge unsatisfied with prior legal history and practice can always devise a legal principle that fits perfectly with past cases and also apply what the judge views as a correct moral principle to decide a current case. To do this, the judge needs only to state the moral principle she believes to be correct, then conform her decision to past practice by adding an exception for the precise facts of prior cases that went the other way (see Alexander and Kress 1995).

For example: a judge is deciding a nuisance case involving a proposed indoor paintball arena. She believes that the correct moral principle for land use in residential neighborhoods is that ‘landowners must not engage in activities that impose significant burdens on neighbors unless the activity in question provides compensating benefits to the commu­nity'. This principle, alone, would lead her to enjoin the paintball arena from opening. However, past decisions X, Y, and Z permitted a private tennis club, a bowling alley, and a private golf course to operate, and our judge concludes that none of these permitted activities provides a community benefit that justifies the traffic problems they create. One option open to the judge is to endorse a legal principle, ‘landowners should be at liberty to use their property in any manner that does not pose a significant risk to health or safety'. The judge, however, believes that this principle is morally flawed because it does not adequately take account of other sorts of burdens, such as traffic and noise, and does not consider compensating community benefits. So she chooses instead to escape the dilemma by adopting a principle, ‘landowners must not engage in activities that impose significant burdens on neigh­bors unless the activity in question provides compensating benefits to the community, except that they may operate the particular tennis club, bowling alley, and golf course authorized in cases X, Y, and Z’. This principle fits prior practice, not just roughly, but perfectly; and it produces what the judge believes to be a correct moral outcome in her case and future cases.

Dworkin indicates that this type of maneuver would not be consistent with the goal of ‘integrity' in legal decision-making: the resulting prin­ciple does not conform to the ideal of law as a coherent political prac­tice. Integrity, however, is an elusive requirement that Dworkin admits may be pursued by different judges in different ways (Dworkin 1986, 254-8). Apart from the vagueness of the term integrity in this context, formulating appropriate legal principles is also complicated by the fact that the only cases in which legal principles actually impose constraint on judicial reasoning are cases in which the judge believes the outcome the legal principle requires is morally or empirically wrong. In the circumstances, the judge may well feel justified in choosing a correct moral principle subject to exceptions that correspond to past mistakes over a principle that fits more seamlessly with past decisions but, in the judge's view, is wrong. If so, the work-around we have just described may be the morally best way to achieve the maximum combination of fit and moral justification.

Our second argument for the incoherence of legal principles pertains to the problem of weighing competing principles. When legal deci­sions are governed by rules, the judge's task is to determine which rule is authoritative in the circumstances of a case and then apply that rule to the facts. Legal principles operate differently: multiple principles, pointing in different directions, may apply to a single case. The judge's task is to weigh their moral strength in context. Of course, weigh­ing moral strength is greatly complicated by the fact that, due to the requirement of fit, legal principles are likely to be morally imperfect. Even when the mix of eligible legal principles includes some that are morally ideal, the judge normally will need to assign weight to non­ideal as well as ideal principles in order to achieve a sufficient fit with prior legal material. Accordingly, the question the judge must answer is: what weight should a morally incorrect principle have in competi­tion with other correct and incorrect principles?

Past decisions do not themselves determine the weight of the principle they support. The current judge might count the number of decisions that support a particular legal principle, but the number of support­ing decisions does not tell the judge what weight a morally incorrect principle embedded in past cases should have as a reason for decision in the current case. The judge might turn for guidance to correct moral principles (or what she deems to be correct moral principles), but correct moral principles will always dictate that incorrect principles should have no weight at all. It seems, then, that there is no plausible standard for determining the weight of incorrect legal principles: their weight can only be a function of ungoverned intuition. Ultimately, therefore, legal principles cannot exert any sensible form of control over the judge's decision about what to do.

So far, we have addressed our arguments to Dworkin's account of legal principles, in which the current judge constructs a legal principle from the data of past decisions and uses it to resolve a pending dispute. Suppose instead that we adopt a less ambitious, convention-based view of legal principles: principles generally agreed upon within the legal profession count as legal principles (see Fiss 1982). Might these con­vention-based principles act as useful guides to legal decision-making? We do not think so.

To count as a legal principle, a convention-based principle must oper­ate in the manner of a legal principle rather than an authoritative rule. It must be organic rather than posited, changing as the body of pro­fessional opinion and practice changes. The principle's content must be determined at least in part by coherence with past decisions. The principle must exert weight in the process of decision-making rather than prescribe a result. The difference between this type of principle and the legal principles Dworkin discussed is that the weight, as well as the content, of a convention-based principle is established by profes­sional agreement rather than constructed by a current judge.

There are two ways to conceive of the source of convention-based principles. First, they might arise directly from consensus among legal professionals about principles governing decision-making: both the content of each principle and its weight are constituted by profes­sional agreement. Alternatively, convention-based legal principles may be discernible in patterns of agreement among legal profession­als about how particular cases should be decided: convention-based principles are the principles that best explain professional consensus about outcomes.

The first proposal, that legal principles can arise through professional consensus about the principles themselves, is implausible. The legal profession has no authority to posit principles, and if it did the result would be a set of rules rather than a set of organic principles that change as professional consensus changes. If, on the other hand, con­vention-based principles are not posited but arise from a changeable consensus about legal principles, it is difficult to see how the governing principles can predate, and determine the outcome of, a particular dispute.

Suppose instead we take the second tack, that legal principles arise through professional consensus about past outcomes: they are the best principles consistent with outcomes previously accepted by the profes­sion. One difficulty with this approach is that it depends on the unlikely event of consensus among legal actors about a range of outcomes. To the extent that members of the profession fail to agree on outcomes, there are no governing legal principles. A further problem is that when professionals do agree about outcomes, their consensus may be attrib­utable to moral agreement rather than institutional history, and so fail to support a legal principle.

6.2.3 Pernicious effects of legal principles

To this point, we have argued that legal principles are not logically defensible and do not impose any genuine constraint on decision­making by current judges. At this point, we will assume both that it is possible to derive legal principles from past decisions and that legal principles can guide decision-making in a meaningful way. Our argument here is that, to the extent legal principles do guide decision­making, their effect on the quality of judicial decisions is more likely to be harmful than helpful.

No method of decision-making is perfect. The quality of unconstrained moral reasoning depends on the reasoning skills of judges. Judicial reasoning is likely to be imperfect, not only because judges are human, but because their reasoning may be influenced by the special salience of the particular dispute they must decide and the sympathies they feel toward parties before them. At the same time, judges understand the gravity of the matters that come before them. They are also familiar with the history of legal decision-making and the background of estab­lished legal rules, and can use this knowledge in the process of reason­ing to reflective equilibrium (see Penner forthcoming). Accordingly, judges are likely to reason fairly well.

Authoritative rules will also produce flawed results. To guide future conduct and decisions effectively, rules must be stated in determi­nate terms. As a result, they will sometimes require outcomes that do not serve the moral and empirical reasons that motivated the rule. Determinate, authoritative rules, however, have benefits that can at least compensate for the moral and empirical mistakes they cause in some cases. Because they are both determinate and authoritative, well- designed rules can streamline decision-making, settle moral contro­versy, and enable private actors to coordinate their conduct in salutary ways.

Legal principles - the morally best set of principles that fit the pat­tern, and possibly the reasoning, of prior decisions - provide no such compensating benefits. They cannot settle controversy or help to coordinate decision-making because the elements of fit and weight are unstable: they are determined case by case by individual judges and change as the background of legal decisions changes. The process by which judges select and weigh legal principles is at least as complex as the process of moral and empirical reasoning, and the risk of judicial error is at least as great. The resulting principles do not even aspire to be morally and empirically correct, and the likelihood of variation among judges is too high to support private decision-making. In short, the method of decision according to legal principles is morally defi­cient and functionally unreliable.

Not only do legal principles fail to provide the benefits associated with determinate, authoritative rules; at least in Dworkin's account, they also override rules. Rules announced in prior judicial opinions are simply one type of material that a newly hatched legal principle is supposed to fit. Moreover, because the requirement of fit is only one part of the formula for selection of legal principles, once the threshold of fit is met, the current judge can disregard existing rules that might otherwise stand in the way of a favored legal principle. Determinate, authoritative rules provide significant benefits within our legal system. In a regime of legal principles, however, there can be no authoritative rules.

6.2.4 Why proposed justifications for legal principles fail

Despite the negative effects of legal principles just described, various normative arguments have been made on their behalf. The notion of decision according to specifically legal principles fits the ideal of common law as an evenhanded system of governance that has worked itself pure over time. In our view, however, legal principles are not the way to realize this or any other ideal.

Dworkin's primary argument in favor of legal principles holds that principles derived from the pattern of past decision-making promote a form of equal treatment that Dworkin refers to as ‘integrity' (Dworkin 1986, 225-32). The key point in this argument is that requirement of fit between legal principles and past decisions results in like treatment of similarly situated past and present litigants. More comprehensively, decisional principles tied to past legal determinations unite the body of law and create ‘a coherent conception of justice and fairness' that applies to all parties in all cases.

In our view, like treatment of litigants over time is not necessarily a moral good. Equal treatment is a moral virtue in various settings, such as distribution of public benefits, imposition of punishment, or application of moral principles. Common law reasoning, however, does not fall into this category. One problem is that litigants' situations are rarely identical; another is that case descriptions in past judicial opinions may omit important facts. The main difficulty, however, is that, although equal treatment may require consistent application of moral principles to different parties, it does not require that past moral mistakes in legal decision-making be applied equally to future litigants. A past error is a cause for regret but not a reason in the name of ‘equal­ity' for additional moral wrongs.

Particularly in commercial or criminal contexts, those who act under law rightly expect consistency in legal decision-making. Legal princi­ples, however, are not a reliable source of consistency over time. At least in Dworkin's model, the requirement of fit with past decisions means that principles will change as the background of prior outcomes changes, and the processes of constructing and weighing principles means that principles will vary among judges. The alternative we have supported, which combines unconstrained moral reasoning with authoritative judicial rules, at least allows for determinate rules in areas where consistency is particularly important. Legal principles not only fail to provide reliable guidance but also defeat the effect of rules by treating them merely as data points for construction of principles.

Another argument made on behalf of legal principles, related to the argument from equal treatment, is that legal principles avoid retroac­tive decision-making. Natural and rule-based models of the common law treat judges as independent decision-makers, resolving cases on grounds that may not have appeared in prior opinions or on the basis of newly minted rules. Judges, in other words, act in the manner of legislators. In contrast, legal principles exist prior to their application to particular disputes, as the best moral principles that explain the body of prior decisions. Coherence with the past ensures that all new cases will be resolved on the basis of pre-existing rights.

There are several reasons why this argument fails. The most evident of these is that, although legal principles are derived from a pattern of past decisions, they do not predate current decisions in a way that matters morally. Legal principles are constructed in part from prior material, but they remain highly indeterminate until announced and applied by a current court. They may predate the decision they con­trol in some sense, but not in a way that prevents unfair surprise. Authoritative rules, on the other hand, address the problem of surprise directly by dictating a result whenever the rule's predicate facts are present. In the absence of a rule, natural reasoning does not predate disputes but it can and should take account directly of the possibility that, given the history of legal decisions, a particular outcome might constitute an unfair surprise.

Another reason why legal principles do not protect against retroac­tivity is that, even in the hands of like-minded courts, their content will change over time (Kress 1984; Alexander and Kress 1995). Legal principles are the morally best principles that pass a threshold of fit with prior material. Given this threshold, judges will discard material they view as morally mistaken just up to the point at which the result­ing principle no longer would qualify as a legal principle. The resulting legal principle, in other words, will always be the best principle that fits the minimum allowable number of past decisions. Meanwhile, new material will constantly be added to the body of law. As this occurs, judges will discard more past mistakes and account for new mistakes, and the underlying body of law will change accordingly. Because legal principles are organic in this way, actors before a court may face a legal principle that is significantly different from the principle in place at the time of action or the time they planned their activities. If so, the method of reasoning from legal principles may be less rather than more capable of avoiding pernicious retroactivity than unconstrained natural reasoning.

6.2.5 Summary: why legal principles do not and should not have a role in judicial decision-making

For many lawyers, the idea of legal principles that place prior legal materials in their best moral light seems to capture an important part of legal reasoning. Legal principles preserve the wisdom and insights of past judges while also distinguishing judicial decisions from legislation. As a manner of logic, however, legal principles cannot operate in the matter their proponents suggest. The threshold of fit is too indeterm­inate and malleable, and the notion that legal principles have weight apart from their intrinsic moral value is too elusive to sustain the claim that legal principles derived from prior legal material can and should guide judges in reaching decisions.

If judges took the requirement of fit very seriously, seeking principles or ‘reasons' that explain the body of prior legal material in a logi­cal manner and without awkward manipulation, past decisions might exert more power over current outcomes (see Lamond forthcoming; Horty 2011). The effect, however, would still be to incorporate past errors without obtaining the certainty associated with authoritative rules. In any event, a strict standard of fit, at least when applied in a way that takes account of innovations in multiple jurisdictions, will eventually result in incoherence.

For these reasons, we have eliminated from our account of legal rea­soning the entire apparatus of legal principles. To the extent that analogical reasoning purports to derive answer from similarities and differences among past decisions rather than refer to past decisions in a process of reflective equilibrium, we also exclude the method of analogy from our account of coherent legal decision-making. We will suggest in the next chapter that various traditional practices that include comparing facts and outcomes of prior cases and deriving ten­tative principles from judicial opinions can have practical value in the process of judicial decision-making. But legal principles and analogies should not, and cannot, determine the outcomes of cases. Judges who purport to resolve disputes on these grounds are either engaging in natural reasoning under the pretense of applying legal reasons or legal principles, or reasoning deductively from informally posited rules.