A familiar jurisprudential view is that statutes have the content and apply the way the

Transcription

1 Soames, Legislative Intent, and the Meaning of a Statute Barbara Baum Levenbook North Carolina State University 1 A familiar jurisprudential view is that statutes have the content and apply the way the legislature intended. Scott Soames has challenged this view in one form, while giving credence to another. Although the burden of his most recently published paper is that legislative intention in the form of legislative purpose does not determine statutory content (antecedent to authoritative interpretation of them), he repeats his earlier claim that there are some legislative intentions that do. I focus on this latter claim. I maintain that Soames inflates the role of the legislative intentions he thinks count and appears to ignore a source of pragmatic information that does the bulk of the work in determining how a statute applies. The purpose of this paper is not to challenge what appears to be Soames s theory of legal content or his implicit normative theory of adjudication. Like other legal philosophers, I view a statute s legal content as its legal significance, the contribution it makes to the legal requirements, prohibitions, immunities, privileges, and powers created by a jurisdiction s law. I want to separate out the question of a statute s content (and how it applies) from the question of what its (contribution to) legal content is. However, I cannot resist adding that insofar as Soames thinks that legal content of a statutory text must always include what is asserted in that text and conveyed by issuing it, 2 he is wrong. 1 I am grateful to my colleague David Auerbach for his valuable suggestions on an earlier draft of this paper. 2 He seems to be committing to this claim in an earlier paper, Soames (2009), pp There he says that the content of the law enacted is what was said in enacting it, subject to authorized precisifications in certain limited circumstances (ibid, pp ). 1

2 One can show that he is wrong without adopting any controversial theory of the nature of law (or of legal content), such as Dworkin s. One reason everyone who theorizes about law must recognize is that in our constitutional system, the asserted content of a statute may fail to contribute to legal content, because it violates a provision of the constitution. (The Kansas statute central to Brown v. Board of Education (347 U.S. 483 [1954] is a good example. The parties to the lawsuit understood what was asserted in the statute, but one side denied that what was asserted had any implications for the law.) Moreover, it isn t controversial that there sometimes are many things in a jurisdiction s law other than constitutional provisions - - previous judicial decisions, other statutes, entrenched legal doctrines - - that modify the effect of a statute on the law even at the time of the statute s creation or adoption. For example, a jurisdiction may have the doctrine of desuetude, according to which any legal requirement established by a statute ceases to exist after a period of non- enforcement, even if the statute hasn t been repealed, interpreted authoritatively, or declared invalid in a court of law. The statute s legal content at that point is nil, although it still has linguistic content. Yet another example comes from the American doctrine of preemption of state statutes on a subject by the existence of a federal law, which may come first. Once more, a state statute may have no legal content despite its linguistic content. Anyone with a sufficient legal education could add to my examples. Shall we say that the background practice of desuetude or the existence of a federal statute contributes to what is asserted by the statute when it is created or adopted? These background things are not something the addressees of many statutes that is, ordinary citizens and law subjects can be expected to know about, and so, to presuppose, or to recognize as a presupposition by the statute 2

3 adopter(s). So how does the statute convey the information they add? Moreover it doesn t seem that the following claim- schema is necessarily or conceptually false: X understands what the statute asserts, but doesn t understand the law on its subject, because he doesn t have knowledge of other things in the law that predate the statute. 3 So I think the view that what is asserted in a statute is its legal content is a nonstarter (even though in some cases the law will be exactly what is asserted). But the view that what is asserted in a statute is its linguistic content is an interesting theory. One of the ways in which one would expect philosophers of language to be helpful to philosophers of law is by offering insights into the linguistic content of legal materials such as statutes. So, rather than prolonging a discussion of legal content, I propose to view Soames as presenting an account of the linguistic content of a statute. In doing so, I will challenge his idea that legislative intent of a certain sort plays a decisive pragmatic role in determining what the legislators conveyed by their statute and committed themselves to in issuing it. The structure of this paper is as follows. In the first section, I will lay out Soames s account of the intentions that determine the linguistic content of a statute. In the second section, I will briefly lay out my own account, which I have presented elsewhere (Levenbook 2006). In the final section, I will argue for the advantages of my own account. 3 Moreover, to borrow a charge from Dworkin, Soames s view does not fit American legal practice all that well. The materials that Soames thinks may provide important pragmatic information about the legal content of a statute are not generally seen by executives signing statutes into law in this country, nor do such executives have a practice of calling for them. (See discussion below for a description of these materials.) How, then, can an executive with a veto know the legal content of the statute with sufficient reliability to have good grounds for deciding whether to veto or sign it? 3

4 Soames s Account Let me initially agree with Soames that the linguistic content of the statute is what the lawmakers committed themselves to, in adopting the statutory language. (Soames 2009, p. 410) Soames makes an assumption that what legislators (or the legislature?) committed themselves to in a statute is what a statute conveys, and I will not challenge this. But Soames also assumes that what a statute conveys is identical with what the legislators asserted or said in the statute. He draws on general ideas of theoretical linguistics and philosophy of language to point out that what is asserted can differ from the semantic content of a remark. [W]hat language users intend to say, assert, or stipulate is a crucial factor, along with linguistic meanings of the words they use, in constituting what they do say, assert, or stipulate, he reminds us (Soames, forthcoming). If one applies this view to statutes, as Soames does, intentions must contribute to statutory content. Specifically, this is what Soames has to say about these intentions: ILLOCUTIONARY INTENTIONS Soames points out, against textualists, that illocutionary intentions of the legislators are crucial in determining that lawmakers said or asserted anything by issuing the statutory text. He also claims that illocutionary intentions determine what the lawmakers asserted (Soames 2011, p. 43). The illocutionary intentions that count are intentions to say, assert, or stipulate that P, by enabling one s audience to recognize one s intention to do so. (Soames, forthcoming). 4 I don t dispute that illocutionary intentions to say or assert, period, are necessary to establish that lawmakers said or asserted something by issuing the statutory text. In the case of statutes that are directives, I won t 4 Compare with Soames 2011, p. 43: the illocutionary intentions that [give] the law its content are to say, assert, or stipulate that P, in part by virtue of one s audience recognizing one s intention to do so.. 4

5 dispute that the illocutionary intention of the legislature, or shared illocutionary intentions of the legislators, to direct conduct by a certain text by enabling their addressees to recognize their intention to do so is necessary in determining that the text is a statutory directive. And I won t dispute another kind of general illocutionary intention to be discussed, below. But Soames wants more specific content in the illocutionary intentions that determine the statutory content. Consequently, in this broad category of illocutionary intentions, Soames includes the following: ASSERTIVE INTENTIONS FOR ELLIPTICAL TEXTS Assertive or stipulative intentions are necessary to recognize, Soames thinks, in order to understand what was (literally) said, asserted or stipulated, even for legislative acts (Soames 2011, pp ). That is because they (or, rather, their content) is implicitly incorporated into asserted, or stipulative content (Soames 2011, p. 47). Those who assert or utter presuppose that their assertive intentions will be recognized by their audience (Soames 2011, p. 47). So if the speech act is an assertion, assertive or stipulative intentions are necessary to recognize in order to understand the assertion s propositional content (but not, I take it, in order to understand that it was an assertion, as opposed to a warning, an ironic statement, an evaluation, and so on). Soames maintains that there are cases in which the assertive or stipulative intentions of law- makers fill a gap in the semantic content of a text that is nonspecific on a crucial point (Soames 2011, p. 52, note omitted). In more informal terms, Soames thinks that some statutory texts may be elliptical in 5

6 their semantic content, but they may have a nonelliptical content, nonetheless. 5 Assertive intentions of the legislators 6 ( the lawmakers ) may exist that would fill in the gap. Soames s main example is the federal statute involved in Smith v. United States. 7 The statute 8 stated that certain stiffer penalties must be visited upon a defendant who: during and in relation to... [a] drug trafficking crime[,] uses... a firearm. In this case, the defendant offered to barter an (unloaded) automatic weapon for drugs, and was convicted and given the specific punishment under the statute. The phrase uses a firearm is the part of the statutory text that Soames believes to be semantically incomplete. He explains: [T]he ordinary meaning of uses an N is silent about how the thing in question is used. Because of this, when the phrase occurs in a sentence, the resulting assertion must be completed, either by the content provided by an explicit qualifying phrase such as as a weapon, or as an item of barter or (when no such qualifying phrase is present) by content that is presupposed by those using the sentence to assert or stipulate something. (Soames, forthcoming) In other words, the semantic meaning of the text, in the context in which it was adopted, does not answer the question of for what purpose a weapon must be used; and failure to answer that question results in failure to answer the question of whether the statute applies to Smith s act, once the (nonlegal) empirical facts about Smith s act and its circumstances are established. For these answers, one turns, Soames holds, to the legislature s assertive intentions. These intentions, he thinks, are to 5 Soames calls the latter legal content, which one can do to distinguish it from semantic content. However, there is another phrase he uses the linguistically based content of the relevant legal texts which will do this job adequately. Soames (2009, p. 403). 6 Soames talks of intentions of legislators, as of an aggregated group, not of the legislature (Soames 2009, p. 421). For reasons that cannot be gone into here, this may be a mistake when assertive intentions are in question U.S. 223 (1993) 8 Title 18 U.S.C. 924(c)(1) 6

7 convey additional content that the legislature or legislators presupposed. That content is to be determined by inference. In the case in question, Soames presumes that Congress took itself to be prohibiting a particular but narrow use of the firearms as weapons, unless there is something in the congressional record to the contrary (Soames 2009, p. 414). OTHER ASSERTIVE INTENTIONS Filling a semantic gap is only one function assertive intentions may have. Soames remarks that there are cases in which assertive intentions include a restriction of a general class named in a text that the speaker presupposes to be recognized by his audience in the context. (Soames 2011, p. 47) The result is, Soames says, that what is asserted differs from what the text literally means (Soames, forthcoming). 9 There are also cases in which the statutory text contains a crucial term that is vague, but lawmakers include something that would otherwise be a borderline case in its extension (Soames 2009, ; Soames 2011, p. 52). 10 Presumably, they may do this including through assertive intentions, evidence of which comes from something other than the statutory text. Finally, Soames mentions that legislators may use terms in the text of the statute referentially in a way at odds with their semantic meaning (in a way that misdescribes, taken literally). (Soames 2009, pp ) Presumably, they have assertive intentions to do so. He cites a case, Nix v. Hedden (149 U.S. 304 [1893]), in which a statute imposed a tariff on imported vegetables in their natural state but did not mention taxing fruits. The tariff was imposed at the Port of New York on imported tomatoes. Soames then imagines that the lawmakers were using vegetables referentially to refer to, 9 It should be added that Soames doesn t think this happens with Constitutional amendments, but his discussion of the point leaves it possible it might happen in some cases (e.g., statutory texts?) 10 Can they have an assertive intention to include something clearly not in the borderline of the vague predicate? That would be a case of misdescribing what they have in mind. But how would it, on Soames s view, affect the content of the statute? 7

8 among other things, tomatoes, although tomatoes are, scientifically speaking, fruits of the plant (Soames 2009, p. 410). Other kinds of alleged legislative intentions do not, in Soames s view, determine statutory content antecedent to some official interpretation of it. For instance, legislative purpose is often what is meant in jurisprudential talk of legislative intent. But for Soames, searching for this kind of legislative intention is not a way of finding antecedent law, nor is it a way to use pragmatic information to determine what the legislature asserted by issuing the text. 11 It is a way of making sure that judicial sharpening, precisifying, or altering of antecedent law is consistent with deference to the legislature. In short, according to Soames, illocutionary intentions, including specific assertive intentions, contribute to statutory content; but legislative purpose does not (until it is used in an official alteration of statutory content by judicial lawmakers ). This view reverses Justice Rehnquist s claim on behalf of the Supreme Court over time that:.congress intended what it enacted. (United States Railroad Retirement Board v. Fritz, 449 U.S. 166 (198) at 179.) For Soames, Congress enacted in part what it intended. And therein lies the problem. For Soames s account distorts the truth about statutory content, however attractively it might mirror insights about nonlegal, private speech. As I have indicated, the view both inflates the role of legislative intention and deflates the role of other sources of pragmatic information in determining statutory content and how a statute applies. 11 What Soames means by purpose is, he says, the chief reasons publicly offered to justify and explain a statute s adoption. (Soames 2011, p. 54) 8

9 Problems with Soames s View Even if Soames s account weren t distorting, there would be problems with it. Let us consider what is supposed to be the epistemic basis for ascribing assertive intentions to a legislature or to legislators. Soames says that there is an inference involved and that it is an inference from information about the context of utterance (Soames 2009, pp. 414, 422). It is an inference to what the legislature [understood] itself to be asserting in the statute (Soames 2009, p. 415) or what lawmakers took themselves to be asserting (Soames, forthcoming) or to the content the lawmakers had.in mind (Soames 2011, p. 47). 12 But from what or with the aid of what in the context of utterance does one draw the inference? It is clearly not (as I think it should not be) from the psychological histories of the individual legislators. Interviews with legislators or examinations of their diaries or letters to establish what they were thinking at the time they voted for the legislation (or against it?) is never mentioned. The inference is primarily from something like principles of rationality (specialized for linguistic contexts) for the contexts in which statutes are issued, something like Grice s maxims for conversational implicatures (Soames 2009, p. 422). (However, Soames does not attempt to describe the content of principles, or maxims, for legislative asserting.) Soames writes as if the result of using these principles or maxims on the context in which a statute is produced is presumptive evidence of what the lawmakers did understand themselves to be asserting.. (Soames 2009, p. 416). 13 But like all evidence, it is defeasible. The presumption can be defeated, according to Soames, by legislative 12 Or what they, in fact, recognized themselves to be strictly and literally saying in adopting the statute (Soames 2011, p. 48) 13 But he also talks about illocutionary intentions reasonably attributed to users of the phrases (Soames, forthcoming.) 9

10 history in the case of the federal statute, by what is said (by whom? By how many? 14 ) in the debate in the Congressional record (Soames 2009, p. 414). 15 What else can defeat the presumption is left almost entirely unclear (conference committee reports? Prior drafts?), with one exception. 16 Now, it may be perfectly fine to think of assertive intentions in a conversational case in this way, as given by what the speaker understood him- or herself to be asserting. Indeed, I am willing to agree that ordinarily, a single speaker in a conversation whose remark has a semantic gap may be presumed to have intended it filled in a way given by, say, Gricean maxims, unless, of course, there is evidence from elsewhere in that conversation to the contrary. And I am willing to agree that knowing the linguistic content of an assertive remark requires drawing the appropriate inferences about what the speaker took him- or herself to assert. But it isn t perfectly fine to think of the content of a statute that way, nor to think of knowledge of its content along these lines. If we do, we make nonsense of the rules of legal systems such as the English legal system until 1992 in which judges attempting to apply a statute are forbidden to consult legislative history. 17 Such rules would seem self- defeating on Soames s view if the purpose were to discern what the lawmakers asserted by the statute. For the rules would forbid checking for defeaters of the relevant presumptions, for the implicit inclusion of borderline cases, and 14 Congress has 535 members. 15 One should note that there isn t much of an official record of debates or conference committees and the like in most state legislatures in the U.S. (Summers 1991, p. 424), so we won t have this evidence and cannot check for defeaters very often with regard to American state and municipal statutes. Even if we could, the legislative record of debate cannot reveal a bent rule, so that the semantic gap in the text is filled, e.g., in one way on even numbered months and in another way on odd numbered months. But why not, if the debate reveals that some of the debaters thought it was or wanted it filled in the first way and some in the second? 16 In discussing one of Fuller s statutory examples, Soames remarks that what the lawmakers would have said about filling in a semantic gap if they had adverted to it when in fact they hadn t doesn t count (Soames 2009, pp ). Hence, if there were evidence of these individual dispositions, that evidence presumably would not be a defeater of the presumption in question. 17 According to Bix (1993, p. 187): The English courts try to `discover Parliamentary intent by focusing almost exclusively on the text of the statute itself. And see ibid., fn. 11 and its associated text for the history of the English attitude toward looking at records of Parliamentary debate. 10

11 so on. 18 The point is that legal systems are not always geared to discovering what the legislature took itself to have asserted or what legislators did, assuming it makes sense to talk this way. But it does not follow that legal systems aren t always geared to discovering, when doubts arise, the contents of statutes. There are other suspect consequences of thinking of the content of a statute the way Soames does. If we let what some members of the legislature say in debates determine what the assertive intentions are for the group when the intentions cannot be discovered by the semantics of the text alone, we leave no conceptual room for the debaters to misunderstand that to which the text will commit them. Why think they are infallible? And then how shall we handle the assumption by the best professional legislative drafters in this country that one of their jobs is to check drafts to warn the prospective sponsors of to what the bill, if passed, would commit the legislature in advance of official debate? When a good drafter learns that a legislative sponsor has an assertive intention that would, e.g., defeat a presumption when the text has a semantic gap, the drafter s response is not to leave the text alone in the hope that the legislature will adopt that assertive intention. A good drafter would regard such behavior as irresponsible and unprofessional. The drafter does not take the position that what the draft bill, if passed, would provide for is yet to be determined once the debate is over, in case defeaters of a presumption arise. Instead, the drafter typically responds that the text of the bill will not provide for what the sponsor has in mind, and suggests an alteration to the text. In addition, the assumption that statutory content is identical with what the legislature asserted (antecedent to alterations by judicial lawmakers ) and the assumption that assertion depends upon the 18 And see remarks in fn. 3, supra. 11

12 absence or presence of specific assertive intentions make it less than rational to expect compliance with the dicey parts of the alleged content when the statute is merely promulgated to ordinary lay addressees, as statutes very often are, when the statute is glossed by someone who does not have full access to the debate record and the gloss is widely relied upon (as statutes very often are in driving manuals, guides to the ordinary taxpayer, government informational pamphlets on Your Rights as a Consumer, newspaper reports, and the like), and when the statute is so old that the record of debate on it, if it exists, isn t readily available to most of the addressees of the statute. Or, rather, it will be less than rational unless the addressees have ready access to a heuristic device that warns them of the deviations from the semantic meaning in the context of the adoption of the statute. And how often does that occur? Yet legal officials act as if they expect compliance, nonetheless or, more to the point, as if the demand for compliance were perfectly reasonable, under the circumstances. Are we to assume that this is a gross error or heartless disingenuity on their part? I suggest that Soames is misled by trying to force a statute to reflect a little too much into the shape of a conversational remark, despite his awareness that the two situations have relevant dissimilarities. The problems here suggest that statutes aren t utterances or, at least, they aren t like paradigm cases of utterances. Statutory language is canonical, so the textual language is meant to be resistant to something. (Alteration in the face of assertive intentions of the original legislators, perhaps?) Statutes are aimed at addressees unknown and unknowable, in social and economic situations not anticipated, who know and can imagine little of the contexts of the statutes adoption, and who may bring new presuppositions to their reading of the statutes. 19 Statutes are in modern legal 19 As I remarked elsewhere (Levenbook 2006, p. 73), statutes, unlike these conversational directives, are typically adopted and applied in at- remove contexts. Much of the nonlinguistic information in conversational contexts 12

13 systems products of a group acting as an institution on institutional rules about what counts as producing it and what counts as its textual content. Utterances, even those intended as directives, don t seem to have these characteristics; or at least, standard cases of utterances do not. 20 These differences should make one suspicious about assuming that the only context that counts in providing pragmatic information about what is communicated by the statute (or what the legislature committed itself to by its text) is the context of the statute s official issuance, and that this context operates in much the same way, pragmatically, for both statutes and conversational utterances. Determining a Statute s Content: The limitations of illocutionary intention and what else is needed I have no quarrel with the conclusion Soames comes to in Smith v. United States, namely, that the statute does not apply to the case in which a gun is traded for drugs. Nor do I have a quarrel with the idea that it is the pragmatic enrichment of semantic meaning that determines enough of the statutory content to determine how the statute applies. The quarrel comes at the next stage. Soames, as we have seen, treats a statute s content (antecedent to any official alteration) as what was said in enacting it, and finds that in the context of enactment (Soames 2009, p. 417). He takes some facts including the conversation leading up to the utterance, facial expressions, tones of voice, and so on is typically missing for statutory audiences or addressees. 20 Insofar as some nonlegal directives such as university or departmental regulations, rules adopted by condominium associations, the laws of the game of rugby - - have these characteristics, their content seems to be best thought of more along the lines of statutes than of conversational utterances. I find it interesting that I d say something similar about my students written remarks in their essays. I know perfectly well what they have asserted when I put a question mark in the margin or write What do you mean? they ve asserted something nonsensical, ambiguous, crucially incomplete, or opaque because ungrammatical. It has a regrettable linguistic content. I am politely reminding them that they failed to assert what they took themselves to be asserting. Some other record of what they took themselves to be asserting at the time would not make me change my grading, since the examination book is the canonical text for this purpose. 13

14 about what we might call, broadly, legislative history to both limit and help establish what a statute conveys, by establishing legislative illocutionary intentions that include specific assertive intentions. This emphasis on the context of the issuance of the statute is, I suggest, misplaced. There is no arguing with Soames that, if it makes sense to impute illocutionary intentions to a legislature as a whole, or to impute shared illocutionary intentions to legislators, as it well might, such intentions at the time of the issuing of the statute make that text authoritative. But I want to hold that the illocutionary intentions that count are quite general and unspecific. The intentions are shared intentions on the part of those voting for it to change the text of a bill (with or without amendments) into a statute, a legislative or group intention to issue the resulting statute, and (assuming the statute has the semantic form of a directive) a group intention that the statute be recognized as intended to direct conduct in accordance with its content, whatever that is. 21 But I deny that Soames s more detailed assertive intention, if it exists, is authoritative when it comes to what the statute conveys and to what the legislature committed itself in adopting it. To explain what does the work Soames imagines specific assertive intentions to do, we must turn to what is known by the large class of addressees, both actual and potential, of a statute. I will confine my discussion to statutes that purport to direct conduct, to be directives. In large part, the addressees of such statutes are not officials; that is, most of the statutes do not purport to directly tell officials what to do. Many statutes are addressed to legal laymen, and purport to tell them what to do. I take the prototypical directive statutes to be of this kind. At any rate, I have proposed that the pragmatic information that either enhances or modifies the semantic content of the text and determines what the statute conveys and to what the legislature committed itself when passing the 21 I am indebted to my colleague David Auerbach for this last idea. 14

15 statute is determined from this audience- relative point of view. Such an account will not confine us to the linguistic content of a statute in the context of its adoption by legislators, since statutes direct conduct across sometimes large spans of time. I am not entirely certain whether such an account generates anything comfortably called the linguistic content of a statute, since this phrase implies that there is only one linguistic content, until it is authoritatively altered by other legal officials. Perhaps such an account generates a range of linguistic contents over a long span of time. That might sound like a familiar idea to philosophers of language until I add that on my account, this range may vary even when the semantic and syntactic rules do not. My account yields more than a claim about a variable linguistic content. For my account explains something else that might vary over a long period of time, namely the set of actual and potential act- tokens to which a directive statute applies, even for a statute whose linguistic content is vague. So I prefer to talk about how a statute applies, or to what kinds of possible act- tokens a statute that is a directive applies. The set of eligible act- tokens, I hold, can vary through time and social context. Elsewhere, I have explained that members of the society targeted by the statute are participants in certain social practices I call practices of salient array- identification. I said there: Whole directives have clusters of act- tokens that they are about. When a word or a phrase is used in a directive, we can think of the directive as having a map of arrays that may or may not bear a tight relation to the map of a constituent word or phrase in its standard orientation and standard contexts. The map for the directive could, for example, be the result of reorienting or reconfiguring the map for a constituent word or phrase used in a standard context..there are broad social practices that make some clusters in a map of act- tokens for a directive salient under certain conditions and others obviously peripheral.. [Like any other source of pragmatic information, these array- setting practices 15

16 make use of cues and components that aren t semantic.] They may include knowledge of how people do things, what is expected, the usual, and the norm. (Levenbook 2006, p. 95) Something like this addressee ability, the ability to detect socially salient arrays of a directive text, is necessary in order to account for how statutes successfully epistemically guide ordinary law subjects (with no special expertise in relation to the statute) subjects at considerable remove from the lawmakers, even on Soames s view. Suppose he holds that statutes can accurately guide the conduct of these subjects because there can be accurate glosses of the statutes linguistic contents, written by people who have access to the special information Soames thinks they will need that current addressees do not. A gloss is accurate, I have observed, to the degree that it applies to all and only the same act- tokens as the statute does. So if the gloss has linguistic content that is highly accurate, we do not, presumably, have to check its history for unexpressed assertive intentions by the gloss writers. (That would be the start of an infinite regress, because then it would permit glosses on glosses, and so on.) So it must be that those who are addressees of the statute can apply the gloss by relying on its text using their linguistic competences and knowledge of their cultural and social setting alone. So addressees must have the ability to detect or assign arrays of act tokens to a text without requiring additional evidence of the authors assertive intentions. (See Levenbook 2006, p. 103) To illustrate the social salience theory of statutory applications, let us consider three examples. The first is a different provision of the federal statute in Smith v. United States. That statute further provides:..nor shall the term of imprisonment imposed [for using a firearm] under this subsection run concurrently with any other term of imprisonment including that imposed for the... drug trafficking crime in which the firearm was used or carried." (18 U.S.C. 924(c)(1)) The three defendants in United States v. Gonzales (520 U.S. 1, 117 S. Ct [1997]) were convicted in both state and federal courts of 16

17 several crimes, including the federal crime of using a firearm during drug trafficking offenses. The sentences for the other crimes were assigned concurrently, but the question was whether the sentence for using a firearm during drug trafficking might also run concurrently. Assuming for the moment that the only issue was the linguistic meaning of the statute (i.e., that here the linguistic content is the legal content), the answer is clear. It is clear because a sentencing for use of a firearm consecutively with other sentences, even state sentences, is within the socially salient array of act- tokens for the relevant portion of the statute, and a sentencing for use of a firearm concurrently is clearly nowhere near it, indeed, is in the socially salient array of act- tokens forbidden by it. So it will surprise no one that Justice O Connor, for the Court, held that the text of the statute is straightforward, and requires a consecutive sentence. (She also pointedly remarked that there is no reason to survey the legislative history, which a lower appellate court had done to justify the opposite result. 22 It follows on the majority view that even if such a survey could be used to produce some evidence, as Justice Steven claimed, of a likely Congressional intent to limit any other term of imprisonment to any other federal terms of imprisonment only, that is irrelevant.) The second is a Canadian bylaw that required all drug shops to be closed at 10 p.m. on each and every day of the week. In R. v. Liggetts- Finlay Drug Stores, Ltd., the defendants closed their drug shop at 10 p.m. and reopened it a few minutes later. 23 As I said elsewhere (Levenbook 2006, p. 96): Shutting down a drug store and then reopening minutes later is nowhere near the salient array of [what is permitted by the terms of] the Canadian bylaw, as even the defendants must have recognized U.S.1 at 6. We therefore follow the text, rather than the legislative history (520 U.S. 1 at 8) 23 (1919) 3 WLR 1025; discussed in R. Cross (1987) at

18 (Consequently, it will surprise no one to learn that the court in R. v. Liggetts- Finlay dismissed the defendants contention that they had complied with the bylaw.) Should we rest this result on the claim that the bylaw is elliptical? It is true that one can always ask, Close for how long? So there is a semantic gap in the statutory text. Perhaps there exist assertive intentions that might seem to plug this gap. However, we do not need them, and they are of limited value if of value at all. I concede that for the day or for the business day might fruitfully be added to the text; its addition would be something toward a completion. And perhaps there are some linguistic rules that would impute the appropriately amended intention to the legislature, absent something to the contrary in official records. But the addition in question won t explain how we know that, if the bylaw specified closing at 11:45 p.m., closing at 11:45 p.m. and reopening at midnight plus one minute wouldn t be in conformity with the bylaw, either. I suggest that the addition in this case has the cart before the horse. If the phrase closing for the (business) day is the obvious choice, or if it is used in the official record of the debate on the subject, that is because it is a loose way of describing the socially salient array of act- tokens for the bylaw. The third illustration is from Michael Moore, who imagines a federal statute that makes it a crime to obstruct or retard the passage of the mail, or any driver or carrier. Moore (1981, p. 277) imagines that a county sheriff arrests a murder suspect who happens to be a mail carrier while he is on his route. If it is clear, as I think it is, that a lawful arrest of a mail- carrying murder suspect is not in an array of act- tokens socially salient as forbidden by the statute, it is abundantly clear how the content of the statute applies to this case. Arresting a mail carrier is a clear case of obstructing the passage of the mail, so vagueness is not at issue here. Moreover, I think there s a strong case to be made that the result isn t reached by filling in a semantic gap in the statute. We aren t waiting for a further legislature 18

19 answer to the question, Obstructing by doing what? as if a phrase beginning with by and followed by the names of activities was missing after the word obstructing. As I said elsewhere (Levenbook 2006, p. 93): What seems to be doing the work of making the arrest permissible is an understood exception to the stated rule; yet specific exception clauses cannot, even in principle, be given exhaustively. This suggests that a mechanism different from ellipsis must be brought in to explain the obstructing mail example. The practice of social salient array identification may operate in complete ignorance of, and even in conflict with, some of the specific assertive intentions of the original adopters of the texts in question. One of Soames s examples is Hart s fictitious statute reading, No vehicles are permitted in the park, which might be seen to have a semantic gap or two (permitted when in the park?). Nonetheless, legislators cannot make that statute permit the driving through the park of family cars full of commuters at evening rush hour by having held some intention to do so, never expressed in the text of the statute that was adopted by both houses of a bicameral legislature and signed by the executive. Driving family cars containing commuters through a park at any hour is precisely the sort of act one would locate in the socially salient array of what is forbidden by the fictitious statute that is, until and unless much in our social conditions changes radically. Similarly, if the legislators actually shared an intention in the case of the federal statute involved in Smith v. United States to add a punishment to drug deals accomplished by bartering for firearms, so long as the text of the federal statute in question is held constant, that shared assertive intention may be wholly irrelevant to what the linguistic content of the statute in question is in a given broad social context. (It is another question entirely whether that shared intention is wholly irrelevant to the contribution to the law made by the statute.) If the salient 19

20 array contains only drug deals using firearms as weapons, then the statute applies its extra penalties only to them, at least in the present social context. Again, if a statutory text provides for a tax on vegetables, and the legislators were unanimously using vegetables referentially to refer to tomatoes, but no legislated piece of text indicates this fact and tomatoes aren t in the socially salient array, then the statutory content does not provide for a tax on tomatoes. In all these cases, the legislators failed to create a statute communicating what they took it to assert. They failed to commit themselves to what they took themselves to commit themselves in adopting the statutes in question. I should add that the account so far is meant to apply in the simple case, and that I recognize complications. There will be complications whenever a statute has been authoritatively interpreted at odds with both its semantic meaning and with its linguistic content, however the latter is understood. There will be complications if, when legislators share one of Soames s specific assertive intentions, they are able to communicate the intention to the addressees in some way other than the text (e.g., by interviews with reporters, in government agency helpful pamphlets, and so on), or by a semantic inconsistency in the text. 24 This communication may cause the socially salient array to shift accordingly. There will be complications when a statute makes explicit reference to an earlier provision of a criminal or civil code, without repeating this earlier provision, and declares that it applies to some new area of conduct. 25 Obviously, no array will be socially salient for the later statute unless both provisions are kept in mind, and that might not be something the addressees ordinarily do. (But it could be something that gloss writers will do.) 24 Brink (1989, p. 187) provides an example: There will be no.[fishing] within fifty miles of shore; this regulation includes whales and dolphins. 25 See, e.g., provisions of chapters 8 and 9 of the Swedish Criminal Code, discussed in Peczenik and Bergholz (1991), p

21 There are two additional sources of complications: Some statutes use terms with a technical meaning, and although some of the technical meanings may be from the industry or activity purportedly being regulated (and so, easily recognizable by addressees), some of them may be legal technical terms whose semantic meanings are recognizable only by legal experts (e.g., in taxation codes). some statutes might use a scientific expression, such as carcinogenic, or cocaine, where what is carcinogenic or is cocaine is fully determinable only by the methods of science. The latter two complications can, I think, be handled with the help of an idea familiar to philosophers of language, that of deferential conventions. I have elsewhere borrowed the view that there are deferential conventions among speakers of English for terms like carcinogenic and cocaine (Levenbook 2006, pp ). They are conventions to defer to experts on the referents of those terms. If anything is salient when a statute outlaws cocaine possession or the addition of carcinogenic substances to food products, it is that instances of cocaine or carcinogenic substances are to be determined by experts. Something similar applies when a tax statute uses an obviously legal term. (It should be noted that experts sometimes affect what is socially salient to addressees more directly. As I have mentioned elsewhere, consider the role expert input has had in the enlightened view of what spreading germs or practicing safe sex refers to among ordinary law subjects.) Advantages of My Account Suppose Soames s idea of assertive and illocutionary intentions is adapted so that the intentions are imputed and indefeasible. This kind of imputed intention exists regardless of whether the legislature gave the matter much thought, to use a Soames phrase, and even if the target audience has some evidence that it did not, or at least, did not come to a common understanding on the point. 21

22 The idea will be to be able to say that legislators to have the assertive intentions to assert the content that is established on the basis of the pragmatic considerations I have described. Perhaps one can talk about an assertive intention in this way. However, it isn t helpful. Since evidence about the legislators or the legislative history that hasn t already contributed to what is socially salient cannot defeat the imputation, the real work is being done by the social salience. We might as well eliminate the middleman. I have been arguing that the role of the legislative intentions in determining the linguistic content of statutes is more restricted than Soames believes, and that pragmatic information determining the correct application of statutes comes from sources other than the ones Soames identifies. Of course, if one takes the phrase sources other than the ones Soames identifies broadly enough, the foregoing sentence isn t an accurate conclusion of my argument. I haven t shown that socially salient arrays come from nothing Soames has discussed. For Soames alludes to Gricean- style maxims as setting presuppositions about illocutionary intentions. Suppose that it is proposed that the social salience of arrays is itself a product of maxims producing presuppositions (about what? legislative intentions?) for pragmatic information for statutory directives. Perhaps a set of maxims for statutory directives might be discovered that yields all and only the socially salient arrays of act tokens for the statute throughout its lifetime. I have my doubts, and I do not think these maxims been discovered yet. More to the point, I have a quarrel with the restriction to the context of utterance for use of these maxims, as my discussion to follow should make abundantly clear, if it isn t clear already. I conclude this paper with some reasons to prefer my account. 22

23 It should be clear that Soames s view is originalist about what the legislature conveys and committed itself to in adopting the statute. 26 Not so mine. On my account, what the legislature conveys by its statute and how a statute applies can change over time even though no one in legal authority has altered its pragmatically- enhanced meaning. The broadest difference in the accounts occurs when one considers the enlightened application of an elderly statute. Suppose, to use language from the British Rent Act of 1920, a statute extends protection from rent increases held by a protected tenant who dies intestate to a member of tenant's family residing with the tenant at death. In 1920, what is clearly not in the British socially salient array of someone who is a member of the family is a long- standing live- in lover of the same sex. 27 But by 2012, let us suppose, a large majority of the British public counts a lesbian couple with children as a family. 28 On my account, the statute applies in a new way in 2012, provided that member of the family hasn t become a technical legal term. In order to say that the linguistic content has altered even though no further lawmakers have altered it, Soames would have to hold: (1) the semantic rules for member of the family have changed in this period, and (2) the original lawmakers had the illocutionary intention that the statute be recognized as intended to direct conduct in accordance with its semantic content, whatever that turns out to be. I leave it to philosophers of language to determine if (1) is defensible. Claim (2) assumes something about assertive intentions that may be inconsistent with Soames s evidence- based approach to them. Failing either of these claims, the best Soames could say is that the linguistic content ought to be altered so that it does apply in the new way - - given the presupposition that Parliament in 1920 had 26 As he expressly admits in his forthcoming paper. 27 See the facts of Fitzpatrick v. Sterling Housing Association Ltd. (2001) 1 A.C This is reportedly true about Americans, according to Powell (2010). 23

24 illocutionary intentions based on something like Gricean maxims and the semantic meanings of the time, or given evidence of specific assertive intentions at the time incompatible with the present- day salient array. Something similar can be said about a variation on a statute that Brink (1988, pp ) imagines concerning toxic substances. Let us assume that the statute prohibits toxic substances in food and drink, and further assume that it is produced in the late 19 th century and never amended or authoritatively interpreted in a court of law. Because of deferential conventions, my account can say that, as the experts to which the conventions defer change their scientific and medical theories, the list of substances banned changes over time. But Soames has difficulties holding that the statute now bans, e.g., cocaine in soft drinks, or that it doesn t still ban something we now know not to be toxic to humans (e.g., saccharine). Moreover, either one would have to be an historian of science, medicine, or the legislature in question to tell what it bans, on Soames s view, or one must have ready access to glosses written by someone who was such an historian. As I hinted earlier, this creates an epistemic problem for those addressees who wish to comply. My account presents a happier epistemic picture than does Soames s; and that is an advantage, I think, in explaining how law subjects can apply the elderly statute correctly. My account does not make that application contingent on law subjects being advised by those well versed in history (legislative or otherwise), or on sheer luck and guesswork. This, I think, fits better what the law subjects themselves think about their ability to identify and apply the statute, although that consideration is hardly decisive. Indeed, insofar as Soames s account requires law subjects, in order to actually comply with, or know, how or whether the statute applies beyond its semantic meaning (or in spite of it) to have been advised by those well versed in the institutional history, my account has another advantage. It can 24