I. INTRODUCTION
In this paper I argue that vagueness in law does not raise any special issues about interpretation, that is, I am defending the following thesis:
No-Interpretation Thesis:
Vagueness in law does not call for a special interpretation of the law.
This goes against a strong trend in legal theory where vagueness in the statute is seen to call for special interpretation. Ronald Dworkin is a well-known friend of this view. Here, for instance, is a quote from his 1978 paper, “Is There Really No Right Answer in Hard Cases?”:
It is open for a lawyer to argue, for example, as I have myself, that the impact of the statute on the law is determined by asking which interpretation, of different interpretations admitted by the abstract meaning of the term, best advanced the set of principles and policies that provides the best political justification for the statute at the time it was passed.Footnote 1
Scott Soames defends a quite different role for interpretation, and yet he, too, argues that vagueness in law calls for interpretation of the law:
the job of interpretation is simply to discern [the content of the law] from the linguistic meaning of the text, plus the context in which it was produced. If [the lawmakers didn't implicitly encoded a discretionary standard into the law], then the job of the interpretation is to formulate one—thereby creating new law by precisification.Footnote 2
The difference between the views of Dworkin and Soames, on the one hand, and my view, on the other, is variously relevant for legal theory and practice. It has consequences for our understanding of precedent in the case of a ruling in a borderline case; it is also relevant for the idea of judge-made law and for the relation between legal language and ordinary language. I do not, however, discuss these aspects in detail but focus on two things: first, I bring out clearly the difference between the two kinds of views, focusing on the formulation by Scott Soames, and second, I offer some arguments in favor of the No-Interpretation thesis. But before I get to the difference between the two kinds of views I describe, for purposes of exposition, a hypothetical scenario and two distinct cases of vagueness which it gives rise to.
Scenario
A law is passed that makes it illegal to paint houses pink. Call it the No Pink Houses law. A group of people, who call themselves the Pink Society, decides to undermine the law. They buy a huge amount of red paint and divide it into a thousand equal parts, enumerated P1, P2, P3, . . . P1000, each of which is large enough to paint a whole house. Then they add a splash of white paint into P1, they add two splashes of white paint into P2, three splashes of white into P3, and so on. The result is that P1 is still clearly red, whereas P1000 is virtually white. The colors in between form a spectrum from red through pink to white.Footnote 3 Then the members of the Pink Society paint their houses in the various colors ranging from red through pink to white. It should be clear enough that very many members of the Pink Society have broken the No Pink Houses law.
Case One
It is a general principle in the application of law that like cases should be treated alike.Footnote 4 The ideal of fairness rests on this principle as well as the relevance of precedent. But given this principle, vagueness can easily generate a problem. Consider the following course of events:
Suppose that Dick painted his house with P100, which is clearly pink. He is charged with breaking the law by painting his house pink and is taken to court and convicted. He does not object to this but demands that like cases should be treated alike and points to Harry, who painted his house with P99. Now Harry is taken to court and convicted of breaking the law. And just like Dick, he demands that like cases should be treated alike and points to Tom, who used P98 for his house. And so on. In the end Susan, who painted her house with P1, which is clearly red and not pink, faces the judge. On the principle that like cases should be treated alike, she, too, is convicted of having broken the No Pink Houses law.
Susan then appeals the case and is acquitted on the grounds that her house is clearly red and not pink. The same happens to Judy, who used P2 for her house, and Sally, who used P3. And so on, until we come to Tom, Dick, and Harry. On the grounds that like cases should be treated alike, they are also acquitted, even though their houses are clearly pink.
Case Two
Consider now the principle that criminal liability does not arise unless one can be shown to be guilty. Consider Smith, who is a member of the Pink Society and painted his house with P20, which is on the border of pink and red. Think only of this single instance (not the series from P1 to P1000) and suppose that the authorities charge Smith with violating the No Pink Houses law. While painting the house, Smith was unsure whether he was violating the law, but now, when facing the charges, he is convinced that he will win the case. Or, even better, he is convinced that the authorities do not even have a case against him. Because the color of his house is a borderline instance of pink, it cannot be shown to be pink rather than red, and hence Smith cannot be shown to have violated any law.Footnote 5 If it cannot be established that someone has violated a law, it is clear that he is not liable for violating the law. This is puzzling since it seems to transform all borderline cases of violation of the law into clear cases of nonviolation of the law.Footnote 6
II. VAGUENESS
In classical semantics, a predicate divides the set of entities into two subsets: those entities to which the predicate applies and those to which it does not apply. Once vagueness is taken into account, one cannot assume that a predicate like “tall” divides the set of entities into two complementary subsets, those entities to which it applies and those to which it does not apply. We get entities with a third status, the borderline cases. So we have the tall men, for instance Al Gore, we have the nontall men, for instance Woody Allen, and then there are those who are neither tall nor nontall (or they are at least neither clearly tall nor clearly nontall, or they are questionable members of either group, or they are borderline cases of being tall, or. . .).
The first systematic attempts to deal with vagueness focused on certain consequences of vagueness for logic. Instead of two-valued logic—that is, logic that operates on two truth values, T and F—a third value, I, standing for “indeterminate”, was introduced, and philosophers tried to recover as much of classical logic as possible.Footnote 7 However, it soon became evident that this would not do and truth-functionality would have to be given up. Consider, for instance, the sentence:
(1) Harry is tall.
Supposing that Harry is a borderline case of a tall man; this sentence should get the truth value I. Given truth-functionality, a sentence of the form “If A, then B” should receive the truth value “I” when both parts of the sentence have the truth value “I”. But the sentence:
(2) If Harry is tall, then Harry is tall
should, all the same, be true. Or replace one occurrence of “tall” with “short,” and sentence (2) becomes false even if the parts of the sentence have the truth value I. Once considerations of this kind come to mind, questions arise showing that in order to say something sensible about the logic of vagueness—or the logic for a vague language—one needs a semantic or metaphysical theory of vagueness. In 1958 Henrik Mehlberg advanced an outline of such a theory:
A vague term may be characterized tentatively as one the correct use of which is compatible with several distinct interpretations. The term “Toronto” is vague because there are several methods of tracing the geographic limits of the city designated by this name, all of them compatible with the way the name is used. It may be interpreted, for instance, either as including some particular tree on the outskirts of the city or as not including it.Footnote 8
Mehlberg then goes on to discuss the relevance of this idea for the truth of sentences containing vague words:
A statement including vague terms may nevertheless be either true or false if its truth-value is not affected by the multiplicity of their admissible interpretations. Such a statement is true (or false, as the case may be) if it remains true (or false) under every admissible interpretation of the vague terms it contains. Thus, although both “Toronto” and “Canada” are vague terms, it is nevertheless true that Toronto is in Canada, because this statement remains true under any admissible interpretation of the two geographical terms it contains. . . . The statement “The number of trees in Toronto is even” becomes true under some of the admissible interpretations of its subject and false under the remaining interpretations; it is therefore neither true nor false.Footnote 9
Mehlberg notes that although bivalence is affected by vagueness, the law of the excluded middle need not be affected. And this leads to the conclusion that “the ordinary connection between the truth value of a disjunction and the truth values of its members does not apply to statements with vague terms.”Footnote 10 Mehlberg offers the following disjunction as an example of a statement that is true even if both its disjuncts are indeterminate:
(3) The number of trees in Toronto is even or the number of trees in Toronto is odd.
In 1975 Kit Fine offered an account of vagueness and a logic to go with it, based on ideas that are entirely in line with those of Mehlberg.Footnote 11 But Fine goes much further in developing them, changing intuitive ideas into a worked-out theory of meaning and logic for vague natural language. The basic idea behind his account is a two-level structure called supervaluation.Footnote 12
First Level
Vague terms admit of precisifications. The term “red” is vague, but it could, in principle at least, be made more precise without affecting its definite extension or definite antiextension. To make a term more precise consists in reducing the set of objects to which the term neither applies nor fails to apply.
Second Level
Truth in a language—say truth in English—is defined with respect to acceptable precisifications of the vague terms in the language rather than with respect to the terms as they in fact stand. In an acceptable precisification, logic and semantics are classical. The fact of vagueness in natural language means that for very many predicates there are various different acceptable precisifications that are, as far as determinate extensions and antiextensions of the vague terms are concerned, equally good.Footnote 13
For illustration, consider a series of houses whose color ranges from red through pink to white. To simplify things, suppose we single out five houses from the series:
P1 is definitely red, P20 and P30 are on the border of red and pink, whereas P100 is definitely pink and P1000 is definitely white. A complete precisification of “red,” “pink,” and “white” consists in drawing a sharp line dividing the spectrum into sections to which the predicates “red,” “pink,” and “white” apply without a reminder. This might be done in different ways. One precisification would have “red” apply to P20 and “pink” apply to P30, another would have “red” apply to both P20 and P30, and the third would have “pink” apply to both. But no matter how the spectrum is divided, the sentence:
(4) P20 is red or P20 is pink
is going to be true. In one precisification it is true because P20 is red; in another it is true because P20 is pink. The same is the case for the sentence:
(5) P20 is red or P20 is not red.
In one precisification the sentence is true since P20 is red, in another precisification it is true because it is not red. However, the sentence:
(6) P20 is red
is true in some precisification but false in others. Truth in a language—say, truth in English—is then defined in the following way:
A sentence is true only if it is true in all admissible precisifications; a sentence is false only if it is false in all acceptable precisifications; a sentence is indeterminate only if it is true in some admissible precisifications and false in others.
This treatment of vagueness has many virtues. First, it respects the commonsense intuition that terms such as “heap,” “red,” and “pink” are vague while it can at the same time accommodate the intuition that sentences such as (4) and (5) are true whereas (6) is not true.Footnote 14
Fine's account of vagueness can illuminate many aspects of the problems of vagueness in law. It may actually have more appeal in the legal context than in the context of ordinary language. This is a point that Scott Soames, for instance, acknowledges and that I discuss further in Section V.
III. VAGUENESS AND LAW
In considering how vagueness affects law, one might have in mind any of three questions:
(Q1) Where does vagueness appear in the law?
(Q2) What is the relevance of vagueness as it appears in the law?Footnote 15
(Q3) What resources are there to deal with vagueness in the law?
In what follows I am concerned mainly with (Q3), which I call the question of applicability. I discuss questions (Q1) and (Q2) mainly insofar as they are relevant for addressing the question of applicability.
One possible answer to the question of applicability is that there is nothing special about law with respect to vagueness that sets it apart from other activities that rely on linguistic practices. If that is right, law is still variously affected by vagueness, since language is variously so affected. And just as there may be no answer to an ordinary question of the form “Is this pink?” where “this” refers to an item whose color is on the border between red and pink, so a comparable question in the context of law might lack an answer. This view has been contested, most notably by Ronald Dworkin, who claims to defend the “unpopular view” that even in cases affected by vagueness there may be a right answer.Footnote 16
The “unpopularity” of Dworkin's view derives, no doubt, from the fact that it seems to contradict the obvious, namely that in the face of vagueness it may be indeterminate whether a rule applies or not. As Timothy Endicott puts it, “One possible objection to the right answer thesis [Dworkin's thesis] is that if the language of a legal rule is vague, there is no right answer to a question of whether the rule applies in a borderline case.”Footnote 17
This objection to Dworkin's view rests on the presupposition that judgments in legal language are affected by the fact of vagueness in more or less the same way as judgments in ordinary language are affected. But that need not be so.
One thing to notice is that the relevance of vagueness in law is different from the relevance of vagueness in ordinary language and communication. Usually when issues of vagueness arise in ordinary language, one need not worry about them too much. It does not really matter whether a particular house is described using the predicate “pink” or the predicate “red.” Such vagueness may certainly cause theoretical problems for philosophers and linguists but rarely does it have any practical consequences. It is at most a cause for a minor quarrel in the family, and unless the members are excessively stubborn, this quarrel will run its course in one afternoon. So one might say that vagueness does not matter that much in real life. It is not even clear why strict uniformity among the users of a language should be considered to be of a particular value. In law, however, uniformity is not only of value but of fundamental importance, for we do not want matters of rights and duties and guilt or innocence to vary among the citizens without a material difference to account for the variation. (Case one above highlights exactly this.)
Further—and here we come to the kind of reasons Dworkin uses to defend his view—there might be resources to deal with vagueness in the legal framework that are not available in language generally. This is what Dworkin is getting at when he draws a distinction between the fact of vagueness and the consequences of vagueness:
It is a very popular idea among lawyers that the vagueness of the language they use guarantees that inevitably there will be no right answer to certain legal questions. But the popularity of this idea is based on a failure to discriminate between the fact and the consequences of vagueness in canonical legal language.Footnote 18
The reasoning behind Dworkin's idea derives from the idea that in order to apply a legal rule in a particular case, one must be able to show that the rule applies in the case in question. This should entail that criminal liability does not arise in a borderline case, which, in turn, would narrow the range of the law to exclude the borderline cases. (This is the point made by case two above.) The fact of vagueness would then not cause gaps in the law.
There is, I think, something quite right here in what Dworkin is saying. In legal discourse there are extra resources to deal with vagueness. In ordinary language, unlike legal discourse, there is no special demand that issues involving borderline cases be resolved one way or the other. Sometimes one might want to maximize truth, at other times one might want to maximize falsehood, and in some cases one might want to maximize indeterminacy. Moreover, in ordinary language former decisions need not impose restrictions on later decisions; that is, in certain circumstances a particular shade of color might be considered pink rather than red and in other circumstances it might be considered red rather than pink without any norm of language or communication being violated. In short, there is no general norm of language use that encourages speakers to maximize truth, falsity, or indeterminacy or to strive for strict uniformity either across time or across different cases at the same time.
However, I do not think that Dworkin's argument succeeds in supporting his conclusion. The argument falters on the fact of higher-order vagueness, as Timothy Endicott stresses: “The objection to Dworkin's argument is that it ignores the problem of higher-order vagueness.”Footnote 19 Endicott actually overstates his point, for Dworkin does not ignore higher-order vaguenessFootnote 20 though he does not deal convincingly with it.Footnote 21 I do not go into the arguments of Endicott and Dworkin here, for they are not essential for my purposes. After all, I am not defending Dworkin's thesis that there is virtually always a right answer in legal cases, only the thesis that vagueness does not call for a special interpretation of the law. But Dworkin's claim that there are extra resources to deal with vagueness is going to be relevant when we consider the plausibility of the view that vagueness in law calls for a special kind of interpretation: legal interpretation.
IV. VAGUENESS AND JUDICIAL DECISIONS
It belongs to natural law, according to Aquinas, “to avoid offending those among whom one has to live.”Footnote 22 Suppose now that in a certain society this means, among other things, that people should not paint their houses in colors that are offensive to their neighbors. The No Pink Houses law is then introduced to give specific content to the claim that people should not paint their houses in colors that are offensive to their neighbors. But as we have seen, the No Pink Houses law is itself vague.
One question now is whether a judge could go one step further toward a more determinate classification of conduct as admissible or nonadmissible. Could a judge, for instance, legitimately rule that a house painted in a particular shade of pink is a violation of the law whereas a house that is painted in a color that is ever so slightly more toward red is not a violation of the law? Rephrasing this as a general question about legitimacy of judicial decisions and borrowing the notion of precisification from the supervaluation framework, we might ask:
(Q1) Can a judge make legitimate distinctions that are a matter of indifference as far as the statute goes, as long as the distinctions fall within an admissible precisification?
Or, to put things differently:
(Q2) Can a judge make legitimate distinctions that are not grounded in the law that justifies the making of such distinctions, as long as the distinctions constitute admissible precisifications of the law?
(Q3) If a precisification of a law is admissible, is it not open to a judge to make such a precisification even if it involves making distinctions that are not prescribed by the law itself?
Let me call the view that we get by answering these questions in the affirmative the precisification view. I take the following thesis about law to be a consequence of the precisification view:
(C1) A law can warrant the making of a distinction even if the distinction as such is not grounded in the law.
We can also formulate the precisification view as a thesis about the role of judges:
(C2) In a court ruling, judges can legitimately make distinctions that are not grounded in the law.
The precisification view differs from that of Dworkin in an important respect. Dworkin's idea is that in legal disputes there is a right answer and what makes the answer right is that it is dictated by the law and the standard resources and practices within the legal system. In Dworkin's view the role of the judge is to find the right answer. Things turn around in the precisification view; there may be no single right answer to be found, though there is a right answer to be made. A judge who rules in a borderline case gives an answer, and if the answer is legitimate—that is, if it constitutes an admissible precisification—then the answer is also a right answer.
The rationale behind the precisification view, whether it is understood as a thesis about law or about judges, is quite simple: to borrow the idea of an admissible precisification from the supervaluation view and use it as a defining constituent for the notion of legitimate decision. The notion of admissible precisification comes from a linguistic view about vagueness in language. People say that the thoughts and practices of speakers of a language determine (together with some other facts) that certain entities fall under a predicate such as “pink” and that certain other entities do not fall under the predicate. Then there are those entities of which it is not settled by the thoughts and practices of speakers whether they fall under the predicate or not. These are the borderline cases. However, one can imagine that the thoughts and practices of the speakers could be made more definitive so as to reduce the range of borderline cases without altering the status of those entities that are determined to fall under the predicate or determined not to fall under it. The result would be a precisification of the predicate, and since it does not alter anything that has already been decided, it is called admissible. It only affects those cases that have not yet been decided.Footnote 23
Likewise, we could think of a ruling in a borderline case as admissible on the grounds that the case was, so to speak, waiting for a decision to be made. A law such as the No Pink Houses law determines the status of certain houses; the pink houses constitute a violation of the law. But the presence of the vague predicate “pink” in the law makes it an open question whether a house whose color is on the border of red and pink constitutes a violation of the law. This means that there are certain decisions that are left to the judges to make, namely, decisions regarding borderline cases.Footnote 24
In the next section, I discuss the relation between precisification and legal interpretation. The former is a notion defined in the formal supervaluation framework, whereas the latter belongs to a long-standing institutional tradition of law.
V. PRECISIFICATION AND INTERPRETATION
From the perspective of a supervaluation view of vagueness, there is an important difference between the legal context and ordinary language. In the former there is someone in particular who is in a position to precisify the vague legal language, whereas in the latter no one is in such a position. In ordinary language there is not as much of a division of labor as in the legal context, and as far as there is a division of labor in ordinary language, its function is not to reduce vagueness. As a result, even when vagueness in ordinary language has clear practical consequences, little can be done in the way of precisification, for no one is in a position to bring about a precisification. If the range of borderline cases gets narrower over time, that will just be a feature of general use of language and not a result of someone's decision. In the legal case, things are quite different. Facing a decision in a borderline case—say, having to decide whether the owner of the house painted with P30 has broken the No Pink Houses law—the judge might (after exhausting all the relevant resources) simply rule that a house painted with P30 constitutes a violation of the law. In this case the judge would have made a right decision, since (i) he made a decision and (ii) his decision was legitimate. This is the central thesis of the precisification view described above.
The precisification view as spelled out above does, however, leave open the question of how one should understand the notion of a precisification, that is, it does not specify whether or how a ruling in a borderline case affects the meaning of the law. One could take the idea of precisification quite literally, maintaining that a ruling in a borderline case precisifies the law by interpreting it in such a way as to make its meaning more definite. I refer to this view as the Interpretation thesis:
Interpretation Thesis:
By making a decision in a borderline case, a judge offers an interpretation of the statute with respect to the case in question and thereby changes the meaning of the statute.
According to the Interpretation thesis, a legal decision in a borderline case may be judged to be right not because it follows from the law but because it effects an admissible precisification of the law. This view of the role of the judge has certain implications both for our understanding of the lawmaking role of judges and for the truth conditions of legal language. If a ruling by a judge in a borderline case precisifies the law, then the sentence:
(7) P30 is pink
becomes true because it was decided by the judge that it should be true. The customary semantic features of the constituent parts do not determine that the sentence is true; they leave it indeterminate. It is only the special role of the judge and the fact that her decision has consequences for the status of sentence (7) that result in (7) being true.
The Interpretation thesis is defended by Scott Soames:
the first step in adjudicating the case is to examine the legislative history to discover whether the lawmakers implicitly encoded a [discretionary] standard into the law. If they did, then the content of the law determinates the outcome of the case, and the job of interpretation is simply to discern that content from the linguistic meaning of the text, plus the context in which it was produced. If they didn't adopt such a standard, then the job of interpretation is to formulate one—thereby creating new law by precisification.Footnote 25
Soames then continues:
This generation of new content by filling gaps in the original text is another respect in which legal interpretation is importantly different from some other kinds of linguistic interpretation. This special gap-filling role is justified by the fact that the court has no alternative, if it is to decide the case at all, and by the fact that legislators are well aware of this aspect of adjudication, and often deliberately employ vague language with future judicial precisification in mindFootnote 26
The idea here is that by ruling in a borderline case, a judge changes the law—she makes it more precise by eliminating some borderline cases. This means, for instance, that once a decision in a borderline case has been made, then, if a similar case comes up again, it will not be a borderline case but a clear case of either being subject to the law or not, depending on the previous ruling. This is also the understanding of George C. Christie in “Vagueness and Legal Language”:
vague language may also be used because a law giver may have a general idea of what he wants to accomplish but may be uncertain as to what specific conduct to prohibit. He decided on some vague general standard which can evolve through a series of individual applications, a general standard which can even change in content as the nature of society changes. Perhaps the most famous of such general standards are the “due process” provisions in the fifth and fourteenth amendments to the [American] Constitution.Footnote 27
If we accept this understanding of precisification, we should say that by ruling in a particular way, a judge changes the law by assigning different and more specific meaning to the statute. The judge changes the law by giving a specific content to a vague statement. Here the notion of an interpretation is indeed appropriate, even if different from some other kind of linguistic interpretation, as Soames points out.
VI. APPLICATION OR INTERPRETATION
In the previous section I describe the view that a decision in a borderline case involves a special kind of interpretation—legal interpretation. The Interpretation thesis may seem a natural way of applying the supervaluation framework to vagueness in law, but there is another—and more modest—understanding of the precisification view available. According to this view, a ruling in a borderline case involves only an application of the law and does not result in any change in the meaning of the law. I call this the Application thesis.
Application Thesis:
By making a decision in a borderline case, a judge applies the vague statute to the borderline case without changing the meaning of the statute.
Both the Interpretation thesis and the Application thesis rely on the notion of admissible precisification as defined in a supervaluation account of vagueness. But although the formalism of supervaluation is well developed, how one should understand it in the legal case is unclear (nor is it clear how one should understand it in the linguistic case). In particular, it does not determine whether one should take a precisification of a vague term to be an interpretation of the term (and thereby affect the meaning of the term) or whether a precisification of a vague term involves only an application of the term (and thereby affects only what we do with language and not its meaning).
The word “interpretation” might get in the way here. Sometimes “interpretation” is used so generously that it is impossible to apply a term or a rule sensibly without interpreting it.Footnote 28 When I draw a distinction between interpreting the law and applying the law, I presuppose a narrower concept of interpretation, one where a distinction between interpreting a law and simply applying it makes sense. My understanding of “interpretation” goes back to Wittgenstein's dictum that there is a way of understanding a rule that is not an interpretation and which shows itself in the rule being followed or broken on particular occasions.Footnote 29 I suggest those who insist that all application of a word (or a rule) involves an interpretation of the word should view the difference between the Application thesis and the Interpretation thesis as a difference between two kinds of interpretations, one that leaves the meaning of the word intact and another that changes the meaning of the word. The following example should bring out the difference between the two theses.
Suppose a judge rules that a particular house painted with P20—and thus borderline between red and pink—is a violation of the No Pink Houses law. A little later another house painted with the same color is discovered in a different neighborhood. Now, according to the Interpretation thesis, the ruling in the latter case should be straightforward, for although the newly discovered house used to be a borderline case of the No Pink Houses law, it is not so anymore. It is simply a violation of the law.
According to the Application thesis things are not so straightforward. The latter house will still be a borderline instance of constituting a violation of the law, and although the previous ruling is certainly relevant, it is relevant only as a precedent and not as an event that has changed the meaning of the statute. The owner of the latter house might argue in his defense that the cases were not comparable due to various nonlinguistic features—say, the neighborhood being more shady, or nearby houses having similar colors, making the relevant comparisons different, or the people in the neighborhood being generally more tolerant of offending colors, and so on—so that the house should not be considered to constitute a violation of the law. Moreover, even if the two cases were utterly comparable, the latter house would still constitute a borderline case, which should be relevant when deciding on a sentence for the violation, for a sentence in a borderline case should, other things being equal, be milder than a sentence in a clear case, even if the clear case is just a minor violation.Footnote 30
My point is that there is substantial difference between the Interpretation thesis and the Application thesis. However, one might still question this by alluding to context sensitivity of vague words. Contextualists' approaches to vagueness hold that it is characteristic of vague words that their extensions may change from one context to another. This could accommodate the view that a ruling in a borderline case makes the statute more precise (by giving an interpretation of it) and yet allow that the extension of the relevant words changes from one context to another. According to contextualism about meaning, a house in the series from red through pink to white may change status on different occasions of evaluation, that is, a house that is considered pink when compared to a house that is ever so slightly more pink may be considered borderline pink on an occasion when compared to a house that is ever so slightly redder.Footnote 31 But the legal case does not allow for such a variation, for a house that has been judged pink and has thereby given rise to a conviction of the owner cannot be considered a moment later borderline pink for similar legal purposes. This shows that contextualism has less appeal in the legal case than in the case of ordinary language.
However, one might suggest a more modest version of contextualism according to which the meaning of a word like “pink” could be sensitive to certain kinds of context changes, such as going from one neighborhood to another. This, however, would not undermine the difference between the Interpretation thesis and the Application thesis because one can think of features that would be relevant for the question of violation of the law in a borderline case without being relevant for the meanings of the relevant terms. The tolerant character of the people in one neighborhood and the intolerant character of the people in another could be such a feature.
I do not intend the above example to support one thesis rather than the other, only to underline their difference. The following example by R.M. Sainsbury does, on the other hand, speak in favor of application rather than interpretation:
No one can criticize an art materials shop for organizing its tubes of paint on various shelves, including one labeled “red” and another “yellow,” even though there is a barely detectable, or perhaps even in normal circumstances undetectable, difference between the reddest paint on the shelf marked “yellow” and the yellowest paint on the shelf marked “red.”Footnote 32
A worker in an art materials shop does not narrow or close the gap between the meaning of the words “yellow” and “red” by organizing the tubes as described in the example. One should not describe the practice of arranging paint one way rather than another as involving an interpretation of the relevant words. By arranging the clearly red tubes on the shelf marked “red” and the clearly yellow tubes on the shelves marked “yellow,” the worker shows that he understands the words and, moreover, one might say that he applies them so as to bring order to the arrangement of tubes of paint in the store. The same goes for the tubes whose color is on the border between red and yellow. If several tubes of paint were sold from the shelf marked “yellow,” the worker might rearrange the tubes and take a few of the yellowest tubes on the shelf marked “red” and move them to the shelf marked “yellow.” We should not say that by rearranging the tubes, the person has reinterpreted the words “red” and “yellow.” The meanings of the words guide his action, that is, he applies the words “red” and “yellow.” When the tubes have been rearranged after some yellow tubes have been sold, what changes is not the meaning of the meaning but simply the circumstances of use.
Moving to the legal case, I must of course acknowledge that by ruling in a borderline case a judge may set a precedent and thereby narrow the space of legitimate rulings in the future. But although the space of legitimate rulings has been reduced by the setting of a precedent, the statute (or the legal text, what the law says) may be just as vague as before; what was a borderline case before the ruling is a borderline case after the ruling. Notice that the worker in the art materials shop might be working in circumstances that are normative in a manner similar to the legal case. He might, for instance, have to be consistent in the way he arranges materials in the shop, be able to offer good reasons for why he arranges things one way rather than another, and in general, he might be following the general demand that customers should be able to recognize that there is a reasonable order to the things in the store. All this might be in place, and yet he might rearrange the tubes after having sold some of the yellow ones without changing the meaning of the words “red” and “yellow.”
This example speaks in favor of application rather than interpretation. Are there other examples that pull us in the opposite direction? The example by Christie about the due process provisions—granting individuals a right not to be punished without a due process of law (Fifth Amendment) and a right to equal protection (Fourteenth Amendment)—does actually suggest that what happens in a judicial ruling is interpretation resulting in a changed law and not just the application of a law that leaves the meaning of the law intact. In fact it seems reasonable that the due process provisions must be interpreted and reinterpreted as time evolves and society changes. In a footnote Christie notes: “It is obvious that ‘due process of law’ and ‘the equal protection of the laws’ evolve in meaning with the development of man.”
But we should be careful here. Are the due process provisions an instance of vague law or are they perhaps rather an instance of ambiguous law? Notice that the vagueness (if such there is) does not give rise to a typical paradox of vagueness. Of course there is vagueness associated with the due process provisions, for questions about how much aid one must be able to get, or how good an opportunity one must have, or how much respect one must be shown, and so on all give rise to vagueness. But that is not what Christie has in mind, for the question that he considers is whether due process means adopting standard A or standard B, where A and B may be quite different standards. Here the case of the due process provisions is more like the case of sacrilegious contracts discussed by DworkinFootnote 33 and hence is better described as a case of ambiguity than a case of vagueness.Footnote 34
Another example from Christie is instructive in this respect, for it has been taken to involve either vagueness or ambiguity:
the attempt to distinguish among slight negligence, negligence, gross negligence and recklessness . . . are familiar. Many of the attempts were prompted by the provisions of automobile guest statutes that exempted the owner and/or driver from liability for negligence but imposed liability for either gross negligence or recklessness. The manner in which these statutes were framed indicated that the legislature felt that not only were gross negligence and recklessness distinguishable but that gross negligence could be distinguished from negligence. Judge Magruder's description of Chief Justice Rugg's method of distinguishing among negligence, gross negligence and recklessness is instructive: It was simply the difference “among a fool, a damned fool, and a God-damned fool.”Footnote 35
The difference between negligence, gross negligence, and recklessness is described as a typical vague difference or a difference susceptible to vagueness. And the description of Chief Justice Rugg's method of drawing a distinction shows an attempt simply to apply another vague measure. But, Christie continues a little later:
Eventually many courts threw up their hands and declared that there were no degrees of negligence, that there was no practicable difference among slight negligence, negligence and gross negligence. The only workable distinction was between negligence and recklessness. The distinction between the two was said to lie in the fact that negligence was to be determined solely by objective standards, whereas recklessness required a type of mens rea. It was this mens rea which was said to distinguish recklessness from negligence and not any necessary difference in the objective manifestations of the behavior being examined.Footnote 36
Here we see how the courts turned an issue of vagueness into an issue of ambiguity—and resolved the ambiguity by selecting a specific meaning, that is, interpreting “recklessness” as involving mens rea. Where Chief Justice Rugg saw a spectrum involving vagueness extending from slight negligence to recklessness, many courts saw an ambiguity between two kinds of cases where one was decided by objective standards and the other by similar objective standards plus a certain frame of mind.
The understanding of Chief Justice Rugg does not sit well with the interpretation view. Suppose that in certain circumstances a person is judged to have shown negligence but not gross negligence. A little later a similar case comes up, and now a judge must decide whether the case in question is one involving slight negligence, negligence, or gross negligence. What can she do? She can and should look at the previous ruling, but that ruling will not provide her with a clearer meaning of the terms “slight negligence,” “negligence,” and “recklessness” as they appear in the statute, though it may provide her with various reasons for describing the case using one of those terms—reasons that might involve equally vague standards as to whether the person in question was a fool, a damned fool, or a God-damned fool. But the relevant reasons need not be the kind of reasons that would change the meaning of the statute but simply reasons that justify the application of a particular term to the case at hand. And all such reasons, accessible or inaccessible to the judge, may in the end fail to dictate a right answer. And yet the judge has to rule; she has to decide for the case at hand whether it was a violation of the law or not.Footnote 37
VII. LANGUAGE OF LAW AND ORDINARY LANGUAGE
I am offering an account of vagueness in law drawing on theories of supervaluation as an account of vagueness. This is what I call the precisification view. I also spell out two understandings of this view—the Interpretation thesis and the Application thesis—and offer some reasons for preferring the second to the first. According to the Interpretation thesis, the idea of a precisification, as developed by Kit Fine and others, should be taken quite literally—even more so than in the linguistic account of vagueness for ordinary language. According to the Application thesis, the idea of a precisification only helps to spell out a space for legitimate action; it is a kind of a background device that is helpful in specifying what can legitimately be done and what not. Applying the law in a borderline case does not, on this account, call for a specific interpretation of the law. Taken together with the precisification view, this leads to the thesis mentioned at the outset:
No-Interpretation Thesis:
Vagueness in law does not call for specific interpretation of the law.
I do not think my arguments in favor of the Application thesis—and thereby in favor of the No-Interpretation thesis—are conclusive; I admit that the issue is complicated and turns not only on an understanding of the notion of precisification but also on the distinction between ambiguity and vagueness. A proper judgment of the two alternative understandings of the precisification view would need to rely on an examination of more cases. But rather than doing that, let me change the focus and look more broadly at the relation between legal language, ordinary language, and justice.
One criticism of supervaluation as a model for vagueness in language is that the very notion of a precisification is out of place. That a word is vague is not really a defect of the word but an essential feature of it that derives from its circumstances of application, perceptual and cognitive skills of people, and the principles of communication in general. However that may be, the notion of precisification seems to fit much better in the legal context than in the context of ordinary language. One can easily think of actual precisifications in the context of law—this is one of the central ideas behind the Interpretation thesis. This suggests that one might use supervaluation, taken quite literally, as a model for dealing with vagueness in law. But it also suggests that the language of law and ordinary language could develop in different ways, and that might be problematic for reasons of justice.
Consider one of the fundamental ideas in John Rawls's theory of justice, namely, that of a well-ordered society. Three things are of central importance for this idea:
First. . .it is a society in which everyone accepts, and knows that everyone else accepts, the very same political conception of justice. . ..
Second, society's basic structure—that is, its main political and social institutions and the way they hang together as one system of cooperation—is publicly known, or with good reason believed, to satisfy those principles of justice.
Third . . . citizens have a normally effective sense of justice, that is, one that enables them to understand and apply the publicly recognized principles of justice, and for the most part to act accordingly as their position in society, with its duties and obligations, requires.Footnote 38
The worry now is that if the language of law departs from ordinary language, then that might undermine the society as well ordered. It is the third item here that is of special relevance. If the language of law departs from ordinary language in some fundamental sense, then it is unclear that the citizens can “understand and apply the publicly recognized principles of justice, and for the most part act accordingly as their position in society, with its duties and obligations, requires.”
The point can be illustrated by the above example of negligence. The publicly recognized duties and obligations demand that one should not show negligence and definitely not gross negligence and recklessness. But what does that mean? What kind of demand is this? The public content of this demand is given by the ordinary understanding of the terms “negligence,” “gross negligence,” and “recklessness.” These terms are vague in ordinary language and remain so despite various rulings of courts, for such rulings are not among the data that ordinary people rely on in either learning or applying those terms. The vagueness of the demand seems, therefore, to be an unavoidable feature of the publicly recognized duties and obligations. If the legal demand with respect to negligence and recklessness is made precise as a result of judicial interpretation of the law, this means that the principle of justice with respect to negligence and recklessness will not be publicly recognizable. And should something like that happen on a large scale, the principles of justice would not be publicly recognizable and the society in question would not be well ordered.
This observation should count in favor of the Application thesis rather than the Interpretation thesis, for the basic functioning of the legal system should not undermine it as a system of justice, that is, as a system that both constitutes and maintains a just society by regulating the behavior of the society. If this is right, then it seems that the Application thesis ought be preferred to the Interpretation thesis even if, as a matter of fact, legal practitioners take it to be the other way around.
