The possibility of determinate rules

Another important background feature of our analysis of legal rea­soning is the assumption that rules can provide determinate answers to legal questions in a significant number of cases.

Critics of rule-oriented legal theory have, in various ways, challenged the assumption that rules can communicate determinate instructions to their subjects. Some critics are broadly skeptical about the capacity of law to constrain decision-making (Altman 1986). Others believe in the possibility of legal constraint but argue that constraint comes not from rules but from professional norms or specialized modes of reasoning, such as reasoning by analogy (Weinreb 2005).

Particularly among proponents of analogical reasoning, the claim of indeterminacy often takes the form of an assertion that legal rules, being general, cannot determine their own application to particular cases (Burton 1995; Weinreb 2005). This argument obviously runs contrary to our own conception of rule-oriented decision-making, in which the critical feature of serious rules is precisely their capacity to dictate their own application to particular cases. The indeterminacy claim might also seem puzzling to an ordinary rule subject, for whom many rules appear to provide comprehensible instructions about what to do.

What, then, does it mean to say that rules cannot determine their own application? One way to understand this claim is that the full exten­sion of a rule - all cases to which it applies - is never clear from the rule's terms.

Another interpretation of the claim that rules are indeterminate is that it rests on a general claim about language. It may be that, in a certain technical sense, the words of a rule have no ‘meaning' apart from their use in particular cases because there are no facts in the world that correspond to the meaning of abstract language (Kripke 1982). This argument is linguistically interesting but unimportant for purposes of legal reasoning. Whatever the true nature of linguistic meaning, basic social understandings allow courts and rule subjects to make sense of the language of rules in many of their applications.

Assume, for example, that the governing rule prohibits the keeping of wild animals ‘within one thousand feet of a private residence without the owner's consent'. This rule contains some tricky words: owner­ship is a complicated legal construct, and a full definition of consent involves contestable moral conclusions (Westen 2004). Yet, the more typical forms of ownership are widely known, and most people under­stand that, in a case of disputed land use, consent normally means express permission. Thus, in at least some instances, and probably in many, the words of the no-wild-animal rule, coupled with minimal linguistic and social expertise on the part of rule subjects, dictate the rule's application. As Frederick Schauer puts it, among members of a community who share a language and a sense of its ‘universal con­text', words and their intended meanings have ‘semantic autonomy' (Schauer 1991a).

A more significant version of the claim that legal rules cannot deter­mine their own application is the claim that the meaning of any rule depends on its purpose. On this view, rules are promulgated as means for realizing certain underlying values and ends, and the only way to ascertain their application to particular cases is to ask what those values and ends require in the circumstances (Fuller 1958). Assuming the no­wild-animal rule is designed to protect the safety of surrounding resi­dents, an animal owner, or a court, might conclude that it should not apply to a very docile, well-caged bear 999 feet from a single residence occupied by a retired lion tamer. Thus, even in a linguistically simple case, the words of the rule do not determine whether an act is legally permissible.

In our view, this argument overlooks the settlement function of serious rules. Given the possibility that those who apply rules will err in assess­ing the implications of a rule's purposes for individual cases, the best way to promote those purposes may be to identify a course of action that, if universally followed, will result in fewer errors overall. In other words, the benefits of the rule as a means of advancing purposes and realizing certain values come precisely from its semantic autonomy - the independence of what it prescribes from the purposes it serves. At best, the argument that rules are indeterminate because their meaning in particular cases depends on their purposes expresses a contestable view about the best way to pursue social ends rather than a logical property of rules (Schauer 1991a).

Another variant of the indeterminacy argument takes a different form but is ultimately similar in effect. Rule skeptics sometimes assert that rules cannot determine the outcomes of particular cases because the application of any rule depends on a prior classification of facts (Burton 1995; Weinreb 2005). For example, Steven Burton states that, at the point of application of a rule, ‘[t]he connection between the abstract class and the case remains to be drawn...

Burton has something more in mind than the obvious truth that the outcome of any decision depends on the decision-maker's skill and integrity in finding facts.^[11] Rather, his claim appears to be that the deci­sion-maker must judge which facts count as important features of the case in order to determine whether the case fits within the words of the rule. However, if we are correct that the words of a rule, read in light of common social understandings about usage and context, have semantic autonomy, it should follow that rules themselves pick out the important features of individual cases. In contrast, Burton may be using the term ‘classification' to refer to an assessment of how specific facts relate to the underlying purposes of the rule: if, and only if, certain facts are important to the purposes of the rule, or to the overall question of what outcome is best, should they be classified as falling within the terms of the rule (Burton 1995)? If this is the argument, however, it suffers from the same weakness as the argument from the purposive interpretation of rules: it depends on an inadequate view of the operation of rules.

The last indeterminacy argument we address is an argument about the body of legal rules as a whole. Centuries of legislative and judicial rule-making have produced a tangled accumulation of rules. Even if we assume that individual rules have a degree of semantic autonomy, the number and complexity of existing rules, combined with a certain amount of interpretive play, make it likely that, in a case of any dif­ficulty, two or more different rules will point to different outcomes. As a consequence, legal rules do not determine the outcome of particular cases: decision-makers face a choice among rules, a choice for which the rules themselves provide no guidance (Llewellyn 1960a).

The claim of indeterminacy is significant, but we do not think it seri­ously threatens the possibility of governance by rules. As Frederick Schauer has pointed out, the extent of overlap among rules is an empir­ical question (Schauer 1991a). Moreover, rather than simply choosing among rules that appear to conflict, judges can and do avoid conflict by ranking and refining the rules. The very fact that legal actors try to reconcile conflicting rules belies the suggestion that the multiplicity of rules undermines legal constraint.

A related argument against rule-oriented views of law holds that, even if rules are capable of conveying determinate meaning, simple rules that permit deductive reasoning are rare in our legal system. More typically, legal norms are phrased as broad standards that call for ‘reasonable' conduct or ‘fair' dealing between parties. Standards of this type are not susceptible to deductive reasoning and, the argument goes, require courts to engage in alternative methods such as analogi­cal or ‘a fortiori’ reasoning.

The prevalence of standards is in part an empirical question, to which we have no answer. We do not believe, however, that legal standards pose a problem for our view of law and legal reasoning. Our descrip­tion of legal reasoning admits of two and only two forms of reasoning about law: deductive reasoning from determinate rules and natural reasoning, consisting of unconstrained moral and empirical reasoning. A legal standard that tells actors or judges to ‘do what is right' is simply an official decision to invite all-things-considered natural reasoning (Alexander and Sherwin, forthcoming).

In practice, legal standards are rarely if ever this open ended - they are likely to tell actors: ‘drive at a reasonable speed in light of risk to other drivers', or ‘do not engage in activities that interfere unreasonably with others' use and enjoyment of their land', or ‘do not take advantage of a contractual partner's reasonable mistake' (see Zipursky 2015).

Within these fairly determinate limits on future decision-making, the effect of the standard is to call on future actors or decision-makers to engage in natural reasoning, that is, ordinary normative and empirical reasoning. Natural reasoning is channeled and limited, but otherwise it is normative and empirical reasoning of the usual kind. No analogies or abductions or ‘legal principles' are required.

1.4