Some challenges to the determinacy of intended meanings

2.3.1 The multiplicity of the rule-maker's intentions

The first analysis of the determinacy of intended meanings that we consider, which comes from Gregory Bassham, suggests that when the rule-maker issues his rule ‘No wild animals are allowed within one thousand feet of private residences', he has several intentions regarding what ought to be done, not one, and these various intentions can con- flict.^[19] Bassham distinguishes between, on the one hand, the rule-making authority's beliefs about what he has bound others to do or to omit and, on the other hand, what other changes in the world that the rule-maker expects or hopes will be accomplished by so binding them (Bassham 1992).

Within the category of the rule-maker's beliefs about the legally bind­ing effects of his enactment, Bassham distinguishes among scope beliefs, counterfactual scope beliefs, and semantic intentions. Scope beliefs are the rule-maker's actual beliefs about what changes he has adverted to and determined should be implemented (‘No ferocious mountain lions...').

Counterfactual scope beliefs are the beliefs about such binding effects that the rule-maker would have held had he considered the particular application (of his text) in question (Bassham 1992) (‘No Malaysian sun bears...'). Bassham gives as an example of a counterfactual scope belief the belief the framers of the United States Constitution would have held on whether skyjacking is an ‘infamous' crime within the meaning of the grand jury clause of the Fifth Amendment. Obviously, the framers would not have had actual beliefs about skyjacking in 1791.

The rule-makers' semantic intentions are their intentions regarding the meaning of their legal texts (Bassham 1992). Semantic intentions may depart from scope and counterfactual scope beliefs. For example, a legislature that bans toxic substances may have in mind a particular concept of toxicity, or a particular definition or description of toxicity - semantic intentions - as well as some particular examples of toxic substances that they believed they were banning (scope beliefs). It may turn out, however, that the particular examples of toxic substances may not in fact be toxic according to the contemplated definition of toxicity. Or it may turn out that both the particular examples of toxic substances and the contemplated definition of toxicity are inconsist­ent with the best theory of the nature of toxicity as a concept. (If the rule-maker enacts a rule referring to bears, it may be that some of the exemplars the rule-maker has in mind do not fall within the definition of ‘bear' he is employing; and his definition of ‘bear' may be at odds with the best scientific account of what defines bears as a particular family - Ursidae - of mammals.)

Bassham labels as ‘moderate intentionalists' those who would follow the authorities' semantic intentions when these conflict with the scope and counterfactual scope intentions (Bassham 1992). And he distin­guishes ‘spare' semantic intentions, which refer to the true nature of a term's referent, from ‘rich' semantic intentions, which refer to the user's definition of the term.

Although Bassham would deal with these various types of authori­ties' intentions - scope; counterfactual scope; and semantic, rich and spare - by choosing which one should be dispositive (Bassham 1992), we suggest that the question is not one of the interpreter's choice.

Choosing which one of Bassham's categories of intentions should be the basis of interpretation appears artificial for two reasons. First, whenever the various intentions conflict, there is no reason not to ask which intentions the rule-makers would wish to dominate in such a case. If semantic intentions conflict with scope intentions, there is no reason to choose semantic intentions as what the authorities ‘intended' if the authorities would have chosen their scope intentions to domi­nate their semantic ones.

Second, we believe that Bassham's distinctions - and other analogous ones - ultimately collapse. Consider, for example, the distinction he makes between sense and reference within semantic intentions. Ultimately, what a term refers to - its reference - cannot be deter­mined without a definition or description (its sense). But definitions and descriptions purport to be of something (Shain 1993). (We would not know the thing that ‘gold' refers to without some description of it, though likewise we think that gold can be misdescribed.) Moreover, the distinction between semantic intentions and more particu­lar scope beliefs and counterfactual scope intentions breaks down because our definitions and descriptions can be fallible generalizations from particulars, and the particulars can be fallible inferences from generalizations.

In the end, what we want to know - given some fact situation, the rule-makers' semantic understandings, the true nature of that to which their terms refer, and so on - is what rule-makers determined the binding effect of their action should be. Because the rule-makers' exemplars may be inconsistent with the definitions of the terms they employ, and the definitions may be inconsistent with the true nature of the terms' referents, the question is: Which did they intend to domi­nate in cases of such inconsistency?^[20] In some cases, perhaps referents will dominate definitions and exemplars, while in others, definitions or exemplars will dominate.^[21]

Our proposal is akin to Bassham's notion of counterfactual scope beliefs and intentions but broader and, as we explained earlier, not strictly ‘counterfactual'.^[22] What the rule-makers mean by an enact­ment is what they intend its binding effects to be over a range of cir­cumstances, not all of which - indeed, hardly any of which - they have adverted to.

2.3.2 Levels of generality of intentions

Our conclusions that the meaning of the rule-maker's rules should be equated with the rule-maker's intended legally binding effects raises the question: at what level of generality should intent be defined for purposes of interpretation? The rule-maker's intent about what should be done can be understood as referring to particular cases, categories of cases, immediate ends, or ultimate ends. Thus, his intent in issuing the rule, ‘No wild animals...', can be described as preventing an imagined scene of a ferocious wolverine being near a home with several small children, preventing wild animal injuries more generally, promoting safe and secure residences, promoting well-being, governing well, or doing what is right. Each of these is, in a sense, a true description of what the rule-maker intends. Moreover, they are only some of the points along a continuum of generality, with an infinite number of gradations in between.

We have already concluded that the relevant intention must be some­thing more than an intention to govern the particular cases to which the rule-maker adverted. At the other end of the continuum, our account of the functions performed by the rule-maker and his rules places important limits on the extent to which the interpreter can refer to the full range of the rule-maker's intentions in issuing a rule. Rules cannot be equated with the more general ends they serve with­out losing the benefits of their form.

In this regard, there is a difference between the interpretation of rules and the interpretation of vague standards (such as, perhaps, ‘due pro­cess of law'). In the case of a standard, the role of the rule-maker is to identify ends and values to be pursued while saying very little about the means of pursuing them. In other words, the rule-maker is not attempting a complete settlement of controversy. Accordingly, until more determinate rules have been issued to give content to the stand­ards, there is no reason to limit the range of interpretive inquiry into the rule-maker's intent. Indeed, insofar as a legal norm is a standard, it has delegated authority to those who apply it. It is they who must translate values into determinate actions. The rule-maker's intended meaning is to delegate to others the task of giving content to the stand­ard. Of course, the rule-maker may intend that the standard be cabined or qualified in certain ways, in which case the rule-maker's intended meanings of such limitations or directives are material. However, any other intention of his beyond these is quite immaterial. There is really nothing else to ‘interpret' once one has determined that a legal norm is a standard.

Our principal concern in this chapter, however, is with interpreta­tion of determinate rules - rules that are designed to settle disputes and thus that curtail the controversial consideration of the best means for promoting certain values or ends. Rules work by attempting to translate moral principles into concrete terms and thus reducing the uncertainty, error, and controversy that result when individuals follow their own unconstrained moral judgment. If the meaning of the rules is equated with the moral principles that the rules are designed to serve, there are, in effect, no rules and no means for curbing moral error. Put differently, an important part of what the rule-maker intends in issuing a rule is that it be a serious rule and not just an expression or reminder of the principles that motivated him to issue it.

If, for example, the rule-maker has issued a rule, ‘No wild animals...', in order to promote safety, it is not open to ‘interpret' the rule to forbid cows near residences, no matter how great a threat they pose to safety. Nor can one conclude that a particular jaguar is permitted because it causes no offense to the purpose (safety) that led the rule­maker to issue the rule. These may be instances in which the rule is under-inclusive or over-inclusive when measured against the under­lying moral principles, but under-inclusiveness and over-inclusiveness do not affect the rule's meaning: they are simply consequences of the rule-maker's deliberate choice to use a rule.

We would be overstating our case, however, if we insisted that the interpreter must never refer to the purposes underlying the rule­maker's rules. We have already noted that the basic process by which people grasp the meaning of others' words involves some reference to the speaker's purpose(s) in using the words. That is how the inter­preter might know that the rule-maker did not intend his ‘no-wild­animal' rule to prohibit escaping a fire with a pet bear by a route less than one thousand feet from a residence (Alexander and Sherwin 2001; Greenawalt 1992; Greenawalt 1997a). But none of this means that the rule-maker's intended meaning for his rule is synonymous with his more general purposes, the most general of which is always the Spike Lee purpose - ‘Do the right thing' (Alexander 1995). A rule's purpose informs its intended meaning but is not its intended meaning.

2.3.3 ‘Translating' the rule-maker's rule in light of his mistakes

We are at this juncture in the argument. The interpretation of the rule-maker's rules must be consistent with the rule-maker's role as a governing authority. That role is to determine what ought to be done, and his determination is what he intends to communicate to citizens and officials through his rules. The meaning of the rules, therefore, just is what the rule-maker wishes to communicate through them - his intended meaning. And although various inconsistent beliefs and intentions may be embodied in the rule-maker's rules, when that is the case, the meanings of the rules are still what binding effects the rule­maker intends the rule to have, at the time he promulgates it.

Searching for the rule-maker's intended meanings - which intended meanings, we have argued, are the key to interpreting legal texts - leads to the further problem of distinguishing between translations and corrections of the rule-maker's intent. In imagining what the rule­maker would say about the binding effects of his rules, we can, as we have said, envision his conceding that ‘I meant X, but I now see that X was a mistake'. In our view, this should not alter the conclusion that the rule means X. But it raises the following issue: Is it possible to distinguish between (1) what the rule-maker did determine with respect to a factual situation that he was not adverting to at the time he authored his communication about what ought to be done, and (2) what the rule-maker should have determined with respect to such a situation? In other words, can we distinguish between the rule-maker's intended meanings and what appear from the subjects' perspective to be the meanings that should have been intended? We have insisted on this possibility, but are we correct?

Consider one of the most carefully argued and persuasive attempts to describe proper interpretation in circumstances not envisioned by the authoring authorities. In ‘Fidelity in Translation', Lawrence Lessig equates interpretation in such circumstances to translation (Lessig 1993). Translation, says Lessig, requires two steps. The first step is to read the text for the meaning it carries in its original context - how the authorities envisioned their determination would operate. The second step is to translate that meaning into the current context of applica­tion. Interpretation as translation requires that the meaning of a text be preserved as the contexts of application change.

Lessig then goes on to describe translation of legal texts - how mean­ing in application is preserved through contextual changes (Lessig 1993). The original context from which meaning is derived consists of the presuppositions of the rule-makers about matters of fact, about matters of law, and about values. For example, the framers of the Fifth Amendment posited a privilege against self-incrimination in a factual setting in which there were no police forces and police interrogators such as those we find today. Therefore, even if the framers envisioned the clause's application to be restricted to inter­rogations at or before trial by magistrates or prosecutors, their mean­ing for the clause is preserved by applying it to the modern context of interrogation - namely, custodial interrogation of the defend­ant by the police. A difference in factual presuppositions between 1791 and today gets us from the Fifth Amendment to Miranda v.

Arizona¹¹ (requiring a recital of rights before a custodial interrogation), such that we can say that the Miranda decision is what the fram­ers of the Fifth Amendment determined. In Lessig's terms, Miranda is a faithful translation of the Fifth Amendment's privilege against self-incrimination.

Lessig gives as an example of a change in legal presuppositions the case where a legal text is implicitly premised on the existence of other legal doctrines, doctrines that in the interim between the promulgation of the text and the application in question are judicially overruled or legislatively repealed (Lessig 1993). As an example of a change in legal presuppositions that arguably requires a change in application, Lessig points to the expansion of federal power under modern judicial inter­pretations of the commerce clause, a change in legal presuppositions that arguably affects whether the Tenth Amendment - reserving to the states all powers not delegated to the federal government (US Const. amend. X) - which might originally have been nothing but a truism, now has some affirmative legal bite in order to preserve a meaningful domain of state sovereignty.

The problem with Lessig's approach, as Lessig himself recognizes, is that it threatens to efface the distinction between interpretation - fidelity in translating the rule-makers' determinations - and emenda­tion, correcting those determinations that are, from the interpreter's point of view, mistakes (Lessig 1993). To see why this is so, imagine any case in which the interpreter has a view on what ought to be done, a view that superficially conflicts with what the authorities have deter­mined. To take one of Lessig's examples (Lessig 1993), consider that in 1864 Congress wrote a provision into the United States Code limiting the fee that a veteran could pay an attorney for representation in a veterans' benefit suit to ten dollars. In 1864 ten dollars could purchase adequate legal services. Congress's intention was merely to limit what attorneys could charge veterans, not to exclude attorneys from veter­ans' benefits proceedings altogether. Today, however, because of over a century of inflation, the ten-dollar limitation would operate to exclude legal representation. The question then is whether fidelity in transla­tion requires us to read ‘ten dollars' to mean the cost of the amount of legal service that ten dollars would have bought in 1864. ^[23]

Lessig is aware of the importance of this question, and he addresses it in the context of setting limits to his model of translation (Lessig 1993)· The important limit for our purposes is what Lessig calls ‘struc­tural humility'. Humility in translation requires that the translator not improve the original text, that is, not correct mistakes that the text contains. The translator's job is to find equivalence in meaning between contexts, not too improve the meaning.

Not all improvements are inappropriate, however, but only improve­ments that affect the task for which we hold the text's authors respon­sible (Lessig 1993). Thus, it is okay for a translator of poetry to improve the poet's handwriting but not his poetry. We do not judge the poet by his handwriting. On the other hand, it is not okay to improve the handwriting in a child's paper if penmanship is part of the child's work to be evaluated. Lessig summarizes the point: ‘What humility requires, then, is a claim about the background understanding of what it is the author is being held responsible for. Against this background, humility counsels the translator to stay clear of presuppositions that touch the author's responsibility.'

Of what presuppositions of legal rule-makers must legal interpreters stay clear? Lessig answers that they must stay clear of ‘political' or value presuppositions but not factual or legal ones (Lessig 1993, 1253-4). In other words, the interpreter, in faithfully translating a legal text, may correct for factual and legal presuppositions that turn out to be false, but not for value presuppositions with which the interpreter disagrees.

How does Lessig's analysis operate in the context of Congress's ten- dollar cap on attorneys' fees? Remember, the important questions for Lessig are whether there has been change in presuppositions (a change that would have resulted in a different text), and whether the changed presuppositions are factual or evaluative. The presupposi­tion that Lessig attributes to the 1864 Congress - that ten dollars will purchase adequate legal representation - surely appears to be a fac­tual one in Lessig's schema. Therefore, the faithful interpreter, who holds Congress responsible for its value choices but not its factual beliefs, would ‘interpret' ten dollars to mean whatever dollar amount is now necessary to purchase what ten dollars would have purchased in attorneys' services in 1864. Moreover, one would suppose that similar results might be in order for the constitutional requirement that the president be 35 and serve four-year terms, and that senators serve six-year terms, or even that posted speed limits are to be taken literally.

Lessig does not justify his conclusion that we should hold authoring legal rule-makers responsible for their value presuppositions but not for their factual and legal presuppositions. Lessig realizes that cor­recting for all now-perceived-as-erroneous presuppositions of the authoring legal rule-makers in effect makes the interpreters, not the authoring legal rule-makers, responsible for governance (Lessig 1993). Thus, he attempts to cabin such corrections by distinguishing between correcting erroneous value beliefs and correcting erroneous factual beliefs. Lessig's distinction, however, is unsatisfactory, both because the relation of fact and value is too messy, and because - as the attor­neys' fees example suggests - we wish to make the authoring rule­makers responsible for factual determinations.

To take the latter point first, many disagreements that interpreters have with the rule-makers whose laws they must interpret are disa­greements about matters of fact. Does the 55-mile-per-hour speed limit save lives? Will congressional term limits increase political accountability? Will ceilings on insurance rates actually work as they are forecast to do? And although Lessig's examples of fidelity in trans­lation involved correcting factual and legal presuppositions that were correct when made but became incorrect over (a relatively long) time, there is no reason in principle that we can see why Lessig would not want interpreters to correct factual and legal presuppositions that were erroneous when made. Or conversely, if we hold rule-makers responsi­ble for failure accurately to assay the present, why should we not hold them responsible for failing accurately to predict the future?

In Lessig's scheme, it turns out to be impossible for the legal rule­makers to say that the law they enacted was premised on a factual or legal error that should cause it to be repealed. For, faithfully inter­preted, the law's factual and legal presuppositions are all corrected; repeals because of factual or legal error are never necessary. Because, however, we believe that such repeals are often necessary, there must be something amiss in Lessig's view of interpretation. Put differently, we think there is frequently a gap between what the authoring rule­makers did determine ought to be done and what, in light of the facts, they should have determined. Lessig's model of fidelity in translation, however, renders such a gap a conceptual impossibility.^[24]

Lessig's distinction between factual and legal presuppositions, which can be corrected by the interpreter, and value presuppositions, which cannot be corrected, is also problematic. It seriously underestimates the number of value disagreements that ultimately can be reduced to disagreements about facts or to mistakes in reasoning. For example, Lessig argues that something has gone amiss if we correct as erroneous the constitutional framers' presupposition that bicameralism is a ‘better' form of govern­ment than unicameralism (Lessig 1993). That presupposition was a value judgment for which the framers should be deemed responsible. It is hard to imagine, however, how that ‘value' judgment is anything other than a judgment about facts - that is, about how efficient, liberty-protective, and responsive the two types of government will be. And the value pre­suppositions behind the preference for the superior form of government in terms of efficiency, liberty-protection, and responsiveness have argu­ably not changed a bit since the constitutional framing.

Lessig himself admits that the line between value and factual presuppo­sitions will not be clear or even a stable one (Lessig 1993). Ultimately, the important question is: for what do we hold the authoring legal rule­makers responsible? Unfortunately, Lessig makes no case for hold­ing legal rule-makers responsible for their values but not their factual beliefs. Indeed, if the legal rule-makers' role is to make shared abstract moral notions much more determinate, the role of rule-makers in determining factual matters is in some sense more important than their role in determining values.

We began this discussion of Lessig by asking whether it is possible to distinguish between what the rule-makers did determine and what the rule-makers should have determined with respect to factual situations to which they did not advert. Lessig fails to distinguish between two different situations: when what the authors determined changes with the context of application, and when what they determined is shown to be mistaken by the context of application. Lessig's fidelity in transla­tion effaces that distinction with respect to factual distinctions and assumes it with respect to value determinations.

Yet, if we are correct about the existence of intended meanings, the distinction between rule-makers' actual intended meanings and what they should have intended is possible in principle. The obstacle that remains is whether the inquiry into intended meanings is sufficiently determinate to support a set of meaningful rules.

2.3.4 The determinacy of intended meaning: the ‘Kripkenstein' critique

We come now to an argument that expresses skepticism about infer­ring intended meanings from the quite limited content of a mental state, namely, the rule-maker's mental state at the time he enacts the rule. Presumably, the content of that mental state is supposed to be normative for the interpreters of the rule, including the rule-maker himself. For example, the content of the rule-maker's mind when he promulgates the ‘no-wild-animal' rule is supposed to make it true that he intended to prohibit red pandas even if these were not in his mind at the time. Yet how can such a momentary and limited mental state be normative for such ascriptions of intention?

What we are raising is what is labeled the ‘Kripkenstein' critique of determinate intentions because it is based on Saul Kripke's interpreta­tion of Wittgenstein (Kripke 1982). That critique rests on the recogni­tion that no mental state content, present or past, can by itself ever make it true that by uttering certain words, one has intended some future act. For example, when we issue the rule ‘Add 68 and 57', what makes it the case that by ‘add' we intend that arithmetic operation that will produce the answer 125? According to the critique, nothing in our past uses of ‘add' precludes the possibility that we might now mean an arithmetic operation that produces the answer 5. For what we did in the past is, for example, consistent with a meaning of ‘add' that produces the result ‘125 for every day before today, 5 for every day thereafter'.

The conclusion to draw from the Kripkenstein critique is not the skeptical one that determinate intentions and rules are an impossibil­ity because there are no mental facts that can anchor determinacy (Humphrey 1999). Rather, the conclusion to draw - and that is almost universally drawn, though by differing routes - is that determinate intentions and rules are matters of knowing how rather than knowing that (Yablon 1987; Zapf and Moglen 1996; Marmor 1999). We learn through interaction with others how to follow rules, including those we set for ourselves (Yablon 1987). When we ‘interpret' what we have intended with respect to situations to which we have not fully adverted - which situations exist for all intentions - we do not look for mental facts in addition to those we call the intention; rather, we just grasp, as we have learned to do, the full range of what we intended in light of our actual mental state and its context (Schauer 1991a; Ripstein 1996). Assertions about intentions are not like assertions about the speed of light, true or false independently of our social reality. Rather, assertions about intentions and their products, such as rules, are anchored in part by actual mental states and in part by social practices.

Indeed, despite the Kripkenstein critique of determinacy, we experi­ence determinacy of intentions and communication daily. We follow rules correctly, be they mathematical or linguistic; and with many rules, we rarely disagree about what they require (Schauer 1991a; Yablon 1987). Whether we are Marxists or monarchists, we stop at stop signs, put commas after introductory dependent clauses, and get 125 when we add 67 and 58. Determinate rules are an everyday fact of life. What the Kripkenstein critique accomplishes is not the undermin­ing of determinacy; rather, at most it forces us to seek the ground of determinacy not in mental states alone but in mental states coupled with skills learned in the course of forms of life.

The issue, we believe, that lies at the core of the controversies about interpretation is not the shopworn question of how transparent inten­tions are toward their objects (If I intend x, and x entails y, do I intend y?), an issue that seems to be one of how to use the term ‘intention' and not one of what intentions are like in the world. The basic issue rather lies within the realm of transparency itself. If we say that in intending to ban wild animals, with a lion as our exemplar, we intended to outlaw Malaysian sun bears, which we did not have in mind and may never have seen, heard of, or imagined, are we reporting a fact about the world, or are we applying certain norms of interpretation to our own past act?^[25]

We believe it is in part doubts about the facticity of intentions regard­ing non-contemplated applications that lead theorists to abandon authorial intentions as the key to interpretation. Intention skeptics ultimately reject any distinction between ‘What did you intend regard­ing unadverted-to situation S?' and ‘What would you have intended regarding S had you adverted to it?' The latter question has no single answer, however, because what you would have determined regarding S had you adverted to it depends upon what you would have believed about S. And because you could have believed any number of things about S, anything from what we believe about S to very different beliefs, there is no fact of the matter about what you would have determined. And because what you did determine regarding S is what you would have determined, there is no fact of the matter about what you did determine.

To be a skeptic about the facticity of intentions regarding situations not adverted to, however, makes one a skeptic about the facticity of all intentions (see Fish 1989). Intentions are future directed, yet that future, no matter how accurately imagined at the time of intending, will frequently, if not almost always, at the time of consummation be somewhat different in relevant ways. More importantly, the set of intentions regarding situations not adverted to contains all intentions about which we think the authorities were mistaken (because mistakes always indicate a failure to advert at some level of analysis). Therefore, skepticism about the facticity of intentions regarding situations not adverted to leads to a denial that there is a fact of the matter about the authorities' intentions in any case in which those intentions could restrict us from doing what is right by our own lights. Such skepticism completely undermines the role of a rule-making authority.

Yet even the intention skeptics should be loath to see all intentions fall into the category of norm-governed interpretations about which there is no fact of the matter. Not only must we say that, although the role of legal authorities is to determine what ought to be done in the future, there is never any fact of the matter about what they determine; we must also deny the legal authorities even the more limited roles that most theorists would grant them, such as determining the words to be interpreted, or the language of those words.^{[26] [27]}

We are left then in the following predicament. The facticity of inten­tions is problematic. There are surely cases where even the one whose intention is in question cannot answer confidently that he did or did not intend that¹ And even when he is confident that he did or did not intend that, it is not at all clear on what such confidence is based. On the other hand, we are confident in most cases about what we intended, including cases where we believe we should have intended something different. And the price of denying the facticity of inten­tions is quite high. If the author's intention cannot extend beyond his world as he sees it at the moment of authoring, there is nothing but the norms and beliefs of the interpreter to determine what ought to be done at any later time, norms and beliefs that themselves cannot be projected authoritatively into the future.

We ourselves are not skeptics about the facticity of intentions. We believe that there are real, not hypothetical, intentions about mat­ters not specifically adverted to, and that whether or not a rule that bans wild animals in residential neighborhoods applies to a species of animal of which the rule-makers were unaware is a question of fact.

2.4 Conclusion

We have argued that when interpreters, whether they be ordinary citi­zens or officials, are faced with a canonical legal text, their task is to determine the lawmaker's intended meaning of that text - what the lawmaker intended to communicate through that text regarding what the citizens and officials are and are not bound to do or refrain from doing. We have based our argument on the essential similarity between understanding canonical legal rules and understanding requests from Mom or from our children that we treat as normative for what we should do. Mom and the kids may misspeak or speak in ways that, in the absence of our knowledge of what they intend to signify, would be ambiguous or incomprehensible. So, too, may lawmakers. But if we take the intended meaning as normative, legal texts containing such infelicities are no more problematic than infelicitous instructions left us by Mom or the kids.

In fleshing out our case for treating reasoning from canonical legal texts as essentially no different from reasoning from other requests or demands that we take to be normative, such as those issued by Mom or the kids, we have pointed out that intended meanings are

what must be done in such cases. But in cases in which the authorities did have intentions, they did determine what must be done even if they come to believe they were mistaken in so determin­ing.

not undermined by the finitude of all mental states. Nor are they undermined by the possibility that they will be infelicitous vis-à-vis the rule-maker's purposes for his rule - either generally, in the case of an ill-conceived rule, or in specific applications. In some cases, the infelicity will be such that we are sure the rule-maker did not intend that meaning, and in other cases the infelicity will raise doubts in the mind of the rule-maker himself whether his intended meaning covers the example. In these cases, the rule does not apply, but the reason is a conflict with or failure of the rule-maker's intended meaning, not a conflict with the rule's purposes. Nor are intended meanings under­mined by the multiplicity of the rule-maker's purposes, or the varying levels of generality at which they might be characterized, or the pos­sibility of conflict among these purposes and between more general and less general characterizations of purposes. Nor are intended mean­ings undermined by changes in the factual, legal, or moral landscape. Finally, intended meanings are not undermined by Kripkenstein skep­ticism about rule-following.

However, what if the intended meaning of a canonical legal text is quite unjust? Or what if it is quite opaque to the average citizen? Or what if the rule-maker is a multi-member institution whose members intend different and perhaps inconsistent meanings at the time they enact the rule? We deal with these problems of equating the meaning of canonical legal texts with their authors' intended meanings in the following chapter.