The traditional rule for construing criminal statutes is the rule of lenity, a name given to a common law principle that "penal statutes should be strictly construed against the government or parties seeking to enforce statutory penalties and in favor of the persons on whom penalties are sought to be imposed."(1) The motivating purpose of the rule is to provide adequate notice to defendants (due process), and to reinforce the notion that only the legislature has the power to define what conduct is criminal and what conduct is not (separation of powers).(2)
Although widely accepted, the rule is by no means adhered to universally. The legislatures of many states, frustrated by what seemed to be unnaturally narrow judicial readings of criminal statutes, have eliminated the rule of lenity. New York and California did so more than a century ago.(3) Recent academic literature has been critical as well. Professors Jeffries and Kahan criticize the principle on the grounds that its application does not further its stated rationales--legislative primacy and fair notice.(4) Other commentators object to lenity as inappropriate in particular situations, such as RICO,(5) civil bankruptcy,(6) government corruption cases,(7) and environmental crimes.(8) Moreover, not even lenity's strongest supporters can believe in it without qualification. Courts have good reason not to want every instance in which statutory language underdetermines meaning to require acquittal.
This Article takes issue with lenity's critics, new and old. Critics either assume lenity in a strong version not now used by the courts, or have too much confidence in how well we can construe statutes without some overriding principle to resolve ambiguities. The rule of lenity has been narrowed dramatically over time in response to changes in the ways that courts generally interpret statutes. In seventeenth century England, a very broad rule of lenity was developed to thwart the will of a legislature bent on seeing statutory violators hanged.(9) American courts, notably through Chief Justice Marshall and Justice Story, adopted the rule early on, but with qualifications that gave additional deference to the intent of the legislature.(10) In the early part of the twentieth century, American courts began looking seriously at legislative history and at other extratextual materials as a routine aspect of the interpretation of statutes. Fueled largely by a series of opinions by Justice Frankfurter, the rule of lenity subsequently was narrowed further, relegating it to a tie breaker only after courts exhausted all other interpretive aids.(11)
Using advances from linguistics and cognitive psychology, this Article argues that the narrow, Frankfurter approach to lenity best approximates the way we use language and form concepts. Early attempts to impose a more mechanical rule of lenity gave inadequate recognition to the flexibility of our conceptual structure at the price of undermining legislative intent. The Frankfurter approach best serves the dual purposes of deference to legislative will and adequate notice to defendants.
Attempts to dispense with the rule altogether also come at great expense to widely held jurisprudential values. Language is sometimes vague and sometimes ambiguous. When we do not know how a legislature intended a statute to apply even after careful study, we have to look outside the statutory language for an answer. This Article argues that the narrow rule of lenity embodies important values with which we are unwilling to dispense. For this reason, courts in jurisdictions that have eliminated lenity legislatively continue to apply it anyway. For this same reason, we should not accept academic proposals to eliminate lenity, such as Professor Kahan's proposal to replace lenity with definitive interpretations by the Department of Justice.(12)
Section I of this Article explores the problem that the rule of lenity is intended to solve: What is it about the way we write criminal laws that leads to such uncertainty in our understanding of them? Coming to grips with this issue is a prerequisite for evaluating approaches to the interpretation of criminal statutes. The relevant literature on criminal statutes rarely addresses this problem, however, and the courts virtually never do, apart from general remarks about the indeterminacies of language. This Article begins by describing the interpretive problems in cognitive and linguistic terms, an analysis that serves as a framework for the remainder of the Article.
Section II describes in more detail the three different approaches to lenity noted above: the broadest version, used centuries ago by courts in England; the traditional American rule, developed by John Marshall; and the narrow, Frankfurter version. This Article shows that the narrow version grew from a shift in interpretive style at the beginning of this century, in which legislative history and other extratextual material began to play increasingly important roles in statutory construction. By applying lenity only after taking this extratextual material into account, courts are likely to find less uncertainty in statutory meaning, and therefore apply lenity less frequently. This Article proposes that Frankfurter, not wanting to concede that his narrow version of lenity was a departure from tradition, literally fabricated history to make it appear consistent with earlier cases, including Marshall's opinions.
Section III examines how these issues show themselves in the Rehnquist Court's approach to interpreting criminal statutes. For the most part, the Court uses the narrow version of lenity. Justice Scalia's efforts to unseat the Frankfurter approach and return lenity to its nineteenth century form have been largely unsuccessful. This section also discusses the manner in which the current Court gives broad scope to certain statutory terms in light of the cognitive framework described in Section I.
Section IV discusses attempts to eliminate lenity legislatively, and evaluates suggestions by scholars that lenity should be eliminated in favor of other values, including Professor Kahan's recent proposal that the Chevron doctrine should replace lenity. This section shows that these efforts fail both because lenity reflects such deeply entrenched values that courts continue to apply it in certain circumstances even when the legislature has enacted laws against it, and because they do not account adequately for cognitive difficulties in statutory interpretation. Section IV is followed by a brief conclusion.
A LINGUISTIC APPROACH TO PROBLEMS IN STATUTORY INTERPRETATION
Two issues predominate in disputes over the scope of criminal statutes: (1) ambiguity;(13) and (2) problems of conceptualization resulting from poor fit between the words of a statute and the events in the world. Courts historically have treated these interpretive problems quite differently from one another.
A criminal statute is ambiguous, in a narrow sense relevant here, when it refers to P, P can alternatively encompass either a or b, and it is beyond dispute that the defendant did a.(14) To illustrate with a classic example of ambiguity from the linguistic literature, the sentence "flying planes can be dangerous"(15) can mean either "it can be dangerous to fly planes" or "planes that are aloft can be dangerous." We recognize this ambiguity because we have tacit knowledge of the syntactic structures of the sentences we speak and hear. The sentence "flying planes can be dangerous" can be analyzed as having a syntactic structure associated with either reading. Our syntactic knowledge, however, takes us no further than this. We must infer from context which reading a speaker intended. If the context does not resolve the matter because it is insufficiently robust, we will simply suffer a partial failure in communication. When this happens in the interpretation of criminal statutes, it motivates application of the rule of lenity.
The same holds true for word meaning. When we hear the word "bank," we know that this word in English refers to the earth along a river, or to a financial institution. For communication to be successful, we must determine from context and everyday experience which of these two uses of the word the speaker had in mind. The speaker's choice of the word "bank" instead of the word "cheese" reduces the universe of possible interpretations enormously. To resolve any residual uncertainty, we rely on context and our knowledge of how the word ordinarily is used. The choice of word itself takes us only so far.
Despite the fact that lenity often is stated in terms of ambiguity, most lenity cases involve conceptual difficulties(16) as opposed to ambiguity. Some do, however, concern ambiguity. Consider, for example, Liparota v. United States,(17) which involved the interpretation of an ambiguous mens rea requirement in a criminal statute. In that case, a jury had found the defendant guilty of food stamp fraud.(18) The applicable statute reads in part: "[W]hoever knowingly uses, transfers, acquires, alters, or possesses coupons, authorization cards, or access devices in any manner contrary to [the statute] or the regulations" is subject to a fine and imprisonment.(19) Liparota, who owned a lunch counter in Chicago, had been purchasing food stamps for less than their face value from an undercover government agent.(20) The regulations clearly prohibited this.(21) Liparota admitted buying the food stamps, and admitted that he bought them knowingly, but he argued that the statute was ambiguous with respect to the scope of "knowingly."(22) The government took the position that "knowingly" should be construed to modify "uses, transfers, acquires, alters, [or] possesses coupons or authorization cards," but not to modify, "in any manner contrary to [the statute] or the regulations."(23) On the other possible reading, for which Liparota argued...