The Ordinary Meaning of Rules
© Springer International Publishing Switzerland 2015Michał Araszkiewicz, Paweł Banaś, Tomasz Gizbert-Studnicki and Krzysztof Płeszka (eds.)Problems of Normativity, Rules and Rule-FollowingLaw and Philosophy Library11110.1007/978-3-319-09375-8_22
22. The Ordinary Meaning of Rules
University of the Pacific, McGeorge School of Law, Sacramento, USA
Brian G. Slocum
Judges typically claim that rules contained in legal texts are interpreted in accordance with their ordinary meaning. It follows that the constituent question of what makes some meaning the ordinary one and the evidential question of how the determinants of ordinary meaning are identified and conceptualized are of crucial importance to the interpretation of legal texts. While a comprehensive analysis of these questions is beyond the scope of this chapter, it is possible as well as important to outline how such questions must necessarily be approached. Certainly, there are a variety of ways in which courts habitually go beyond or reject the linguistic meaning of the relevant text. Normatively based desires to, for example, ensure fair notice or avoid constitutional questions may cause a court to give a text a legal meaning that does not correspond with its linguistic meaning. The ordinary meaning principle, though, is necessarily concerned with the linguistic meaning of the text and not normative matters. As such, certain views about meaning and interpretation can be rejected as being incorrect. In particular, certain claims made by actual intentionalists are fundamentally inconsistent with how the ordinary meaning doctrine must be conceptualized. In short, the intentionalist position that a text means what its author intended it to mean, as well as the associated claims about the nature of natural language that often accompany this assertion, must be rejected. Instead, the ordinary meaning doctrine must be explicated on the basis of systematicities and conventions of language.
KeywordsTextualismIntentionalismOrdinary meaningStatutory interpretationLegal interpretationLinguistics
22.1 The Ordinary Meaning Doctrine and the Problem of Actual Intentionalism
22.1.1 Judicial Reliance on Ordinary Meaning
By what standard should rules contained in legal texts be interpreted? The Supreme Court of the United States believes, as do numerous courts around the world, that rules should be identified on the basis of general principles of language usage that apply equally outside of the law. For instance, the Supreme Court recently reasoned, in rejecting what it viewed as an unordinary meaning, that just because a “dictionary definition is broad enough to encompass one sense of a word does not establish that the word is ordinarily understood in that sense” (emphasis in original).1 The judicial reliance on the ordinary meaning of language should not be surprising. A characteristic feature of legal texts is that they employ natural language in order to accomplish their purposes (Mattila 2002). Further, legal texts are widely viewed as a form of communication (McCubbins and Rodriguez 2011; Van Schooten 2007). Ideally, assuming that successful communication is the goal in most cases, these texts should be understood by different people in the same way. One aspect of this broad requirement is that legal texts should be understandable to the general public, as well as to judges and sophisticated practitioners. As Cappelen (2007, p. 19) explains, “[w]hen we articulate rules, directives, laws and other action-guiding instructions, we assume that people, variously situated, can grasp that content in the same way”. Such a goal would seem to require that absent some reason for deviation, such as words with technical or special legal meanings, the language used in legal texts should be viewed as corresponding with that used in non-legal communications (Mellinkoff 1963).
22.1.2 Framing Actual Intentionalism
Due to the centrality of the ordinary meaning doctrine to legal interpretation, the constituent question of what makes some meaning the ordinary one and the evidential question of how the determinants of the ordinary meaning of legal texts are identified are of crucial importance to the interpretation of legal texts. Both questions are greatly undertheorized in legal scholarship. That is, there are no systematic accounts of the constitutive and evidential questions. This chapter does not attempt to offer a systematic account of the ordinary meaning doctrine but rather addresses a few issues that are integral to it. Specifically, the chapter develops the claim that the interpretation of legal texts, and consequently the elaboration of legal rules, must necessarily be a hypothetical exercise that relies on conventions of language, which largely constitute the ordinary meaning doctrine. In establishing this position, particular attention is given to the claims of intentionalists regarding language and meaning. Intentionalist arguments deserve special attention because, if valid, they pose a challenge to any theory that gives the ordinary meaning concept a central role in determining the meaning of legal texts.
The ordinary meaning concept is necessarily based on an ‘objective’ view of meaning because its communicative content is not constituted by the content of the author’s communicative intentions. In contrast, intentionalism is based on a Gricean conception of communication where the communicative content of an utterance is the content that the speaker intends the hearer to understand by recognizing that very intention (Greenberg 2011b). The so-called ‘strong’ version of intentionalism posits both a constitutive claim that (1) the meaning of a conversational utterance or a text is identical to the speaker or author’s intended meaning (meaning thesis), and an epistemic claim that (2) interpretation consists of determining the speaker or author’s intended meaning (interpretation thesis). In the intentionalist model, the determination of authorial intention is thus indistinguishable from interpretation, which simplifies how intentionalists view the meaning of language (Fish 1999).
Consider a typical situation where a sentence may express two (or more) linguistic meanings represented by (A1) or (A2), and the author intends to express only one of the meanings. In considering the meaning of (A1) and (A2) distinctions can be made, at least according to non-intentionalists, among (i) the linguistic meaning or meanings that a sentence S has in a language L, (ii) the linguistic meaning or meanings that a sentence S conveys in a language L in its context of utterance, and (iii) the meaning that an utterance S in a language L expresses on a given occasion of use. The first category, (i), can be referred to as the word-sequence meaning or ‘sentence meaning’, which is the meaning of a sequence of words taken in the abstract following the relevant language’s syntactic and semantic rules. It excludes all information about the context in which the sentence was uttered or written and is treated as a token of a sentence type in a given language. The second category, (ii), can be referred to as the ‘utterance meaning’ (or ‘what is said’), which is the meaning a sequence of words conveys in its context of utterance. The third category, (iii), can refer to ‘utterer’s meaning’, which is the meaning the speaker or author had in mind to convey by use of the sequence of words. The ‘utterer’s meaning’ holds regardless of whether the meaning was recognized via lexical items and syntactic structures or other publicly available indicators of meaning, or, for some, regardless of whether the meaning was recognized at all by the relevant listeners or hearers.
For an intentionalist, ‘utterer’s meaning’ is all that matters (or, more strongly, all that can be said to exist), and in determining utterer’s meaning, one must ascertain the communicative intent of the author. A text can therefore never fail to mean what the author intends it to mean. Further, intentionalists argue that without resort to the author’s communicative intent, every text would mean everything its words could potentially mean in the language in which it was written. Thus, any sentence S would always mean both (A1) and (A2), and perhaps multiple other meanings. In light of these arguments, one would think that an intentionalist should be committed to the view that actual authorial intentions can be determined with “some epistemically respectable degree of warrant” (Iseminger 1996). For these reasons, the intentionalist meaning thesis rejects the traditional distinction between ‘sentence meaning’ (or ‘what is said’) and ‘utterer’s meaning’. Further, intentionalists argue that in legal contexts deviations from intentionalism by judges may sometimes be wise for reasons specific to law, such as ‘fair notice’, but such deviations nevertheless represent a rejection of a statute’s true meaning in favor of some artificial meaning.
Both the meaning thesis and the interpretation thesis pose a challenge to the ordinary meaning concept. If, as intentionalists argue, the meaning of an utterance or text is identical to the speaker or author’s intended meaning and the goal of interpretation is to uncover this meaning, it is not immediately clear why the ordinary meaning concept should have a prominent role in the determination of meaning. Any evidence, such as independent word meaning found in dictionaries or corpora, not connected to the author’s actual intent would be irrelevant (Azar 2007). Even worse, if words do not have established or conventional meanings in a language, the ordinary meaning concept would be incoherent since, by its very nature, it assumes that words have conventional meanings. Consider, for example, Fish’s (2005, p. 633) view that the definition of ‘interpretation’ is limited to the search for authorial intent because “words do not carry fixed or even relatively fixed meanings”. At the least, it is not clear why evidence of ordinary meaning should act as more than a pragmatic constraint on the ‘true meaning’ of the text, or perhaps serve as a concession that the true meaning of the text cannot be determined due to insufficient indications of authorial intent.
Due to the intentionalist view of language, the importance of the ordinary meaning concept to interpretation can only be realized through an examination of the determinants of meaning and the limits of textual interpretation, and a consequent rejection of the meaning and interpretation theses. As well, the constraints of textual interpretation, and the rejection of the meaning and interpretation theses, underscore the legitimacy and necessity of the ordinary meaning doctrine. As developed in this chapter, the ordinary meaning of a phrase or sentence must, constitutively, be a properly contextualized, public meaning, rather than an idiosyncratic meaning that can only be discerned, if at all, based on non-textual inferences from intent. Ordinary meaning must therefore be based on theories of linguistic phenomena (including conventions of meaning) external to the author and is thereby inherently theoretical. It relies on, to use Hogan’s (1996) term, a notion of “semantic objectivism” that posits that there are facts about what words mean and these facts are independent of what individual people mean by the words they use.
Intentionalists claim that it is “embarrassingly obvious” that textual meaning and authorial intention are the same thing (Campos 1996). Their position, however, cannot be sustained. The meaning and interpretation theses must be rejected because the interpretive process is not, as intentionalists would ideally conceive it, telementational where a thought is transferred from the author’s mind to the interpreter’s mind (Harris 2001). Levinson (2002) observes that our interests in literature are communicative ones but are not more narrowly conversational ones. The same is true with respect to legal texts. Texts are dissimilar from oral utterances in various important ways, and these differences are crucial to how texts can be interpreted. Most importantly, it changes how we conceptualize authorial intention. An author’s so-called semantic intentions—the intent to use a word or phrase and mean something by it rather something else it may mean—are relevant to what the author means, and indeed may be said to constitute it, but they are not determinative of what the text means. Contrary to intentionalist arguments, the interpretive process should not be viewed as involving merely an archaeological excavation-like search for authorial intent. Rather, it should be conceptualized as a theoretical and structured enterprise where, among other things, courts apply linguistic and cognitive science insights to determine the meaning of the text, taken from an interpreter’s standpoint.
This chapter seeks to explicate the principle that the ordinary meaning of legal rules found in texts is necessarily based on a hypothetical intentionalism that focuses on utterance meaning rather than utterer’s meaning. As a consequence, the arguments made by actual intentionalists about textual interpretation must be rejected. Specifically, the chapter highlights the important distinctions between oral conversations and texts, rejects various intentionalist arguments about the nature of natural language, and, finally, argues that a hypothetical view of interpretation is consistent with the current approaches to interpretation. In doing so, this chapter considers arguments about intentionalism and language made by literary theorists, as well as legal theorists (and linguists and philosophers). Of course, intentionalism is not monolithic. Different theories of intentionalism exist depending on the views and needs of a particular community and the kinds of communications that must be interpreted. For instance, legal theorists who advocate for an intentionalist view of legal interpretation undoubtedly have some justifications for intentionalism that differ from those offered by those who are concerned with literary interpretation. These concerns and motivations may have no direct correspondence to non-legal interpretation of texts. Nevertheless, the basic intentionalist position on the nature of language and interpretation is the same, regardless of the type of text at issue.
22.2 Actual Intentionalism and the Distinction Between Oral Conversations and Texts
Perhaps the biggest flaw in the intentionalist view of interpretation is that it does not adequately consider the fundamental differences between the interpretation of verbal utterances and texts. Yet, one of the most important, and constraining, realities of interpretation concerns the differences between oral and written communications. Grice’s intentionalism was built on a model of spoken language, but other intentionalists have dismissed the differences between spoken and written language. In fact, as Davies (2006) notes, intentionalists make the comparison central to their claims. Carroll (1992, pp. 117–118), for example, states that “[w]hen we read a literary text … we enter a relationship with its creator that is roughly analogous to a conversation.”
The different system of signs utilized in writing is driven by the separation, in typical texts, of the context of utterance from the context of interpretation. In typical written discourse, in contrast to typical spoken discourse, the context that concerns the physical setting in which an utterance is produced does not correspond to the physical setting in which the utterance is interpreted. This physical context (which Ivanic (1994) terms “Context A”) is not shared in typical written discourse because the interlocutors are separated in time and space.2 The text is thus a closed, unilateral speech act with no opportunity for linguistic replies that can help precisify indeterminate messages. Unlike the case with much oral communication, the reader of a text cannot immediately, or often ever, seek clarification from the writer of the text. This limitation on communicating via text caused Socrates to compare writing with paintings because like writing, “the creations of the painter have the attitude of life, and yet if you ask them a question they preserve a solemn silence” (Plato 360 B.C.E., 274–277).
In addition to the unilateral nature of a written communication, vital cues that allow for the ascertainment of a speaker’s intent in an ordinary conversation are not available to the interpreter of a text. With a text, the entire message must be expressed in words. Such a message cannot convey the paralinguistic cues that typically assist the listener in a conversation in discerning the speaker’s meaning. Thus, important cues such as winking, facial expressions, posture, stance, laughing and gestures are unavailable to help discern the meaning of the utterance. In addition, with a text the words must obviously be written, making unavailable cues from prosody (rhythm, rate, stress, pitch, pitch contour and intonation of speech), which serve an important role in determining meaning in ordinary conversation. Along with offering semantic clues, these prosodic cues contribute to the emotive or attitudinal quality of an utterance, and consequently its meaning.
Routine and important aspects of conversational interpretation are thus unavailable to the interpreter of a typical text. The lexico-grammatic choices may be the same in a given oral conversation and in a text, but the message sent by the speaker in an oral conversation can be modified by adjusting the paralinguistic and prosodic cues that are unavailable in textual utterances. Further, the lack of a shared physical context means that in typical written discourse there is no possibility for each interlocutor to monitor the other’s comprehension and degree of agreement and to adjust utterances accordingly. The lack of a shared physical context should not be underestimated because language comprehension hinges on voice-based and stereotype-dependent inferences about the speaker (Van Berkum et al. 2008). Language comprehension is immediately context-dependent and is processed by the same early interpretation mechanism in the brain that constructs sentence meaning based on just the words. Language comprehension thus takes very rapid account of the social context, particularly information regarding the speaker, which indicates the centrality of these contextual cues in the final interpretation given an utterance. Little is known about how the brain actually constructs an interpretation (Van Berkum 2008). Nevertheless, research indicates that the linguistic brain uses heuristics, particularly regarding the speaker, to arrive at the earliest possible interpretation. Instead of the standard two-step model of language processing where listeners first compute a local, context-independent meaning for the sentence and then determine what it really means given the wider communicative context and the particular speaker, Van Berkum et al. (2008), maintain that research supports a one-step model in which knowledge about the speaker is brought to bear immediately by the same fact-acting brain system that combines the meaning of individual words into a larger whole.
The differences described above between the interpretive cues available in a prototypical conversation and those available for a text are relevant to the interpretation of most texts and are particularly relevant to the interpretation of legal texts. For various reasons, legal texts are paradigmatic examples of autonomous texts which differ greatly from the prototypical personal communication context. Consider the interpretation of statutes. Unlike the typical oral conversation, the legislative context is impersonal. Legislators and interpreters do not, generally, know each other personally, making clarification impossible (absent further legislation, of course, which nevertheless cannot be seen as the continuation of a conversation). Contextual information regarding the legislators’ intent is often either not available or is of questionable value. For statutes, the typical source of contextual information termed ‘legislative history’ has been widely attacked as being unreliable and subject to manipulation. Part of the reason is that the legislative process is a far less cooperative one than is the typical communicative context. The common strategic behavior is such that some have questioned the applicability of Grice’s maxims of cooperation to legislation (Marmor 2012; Poggi 2011). Due to strategic behavior, and other legislative practices, the words of the legislative text may reflect, in a general sense, the intent of the legislature to modify the content of the law but may not have been chosen to implement the legislators’ communicative intentions. Further, the relevant interpreters, including judges, may have to interpret the text decades or centuries after it was created.
The unsponsored nature of statutes should have various consequences, including underscoring the necessity of the ordinary meaning doctrine. Interpreters must place sustained, if not exclusive, attention on textual rather than extra-textual authorial evidence. More precisely, the unsponsored nature of statutes would seem to require that textual meaning, as opposed to authorial meaning, be the primary determinant of the communicative meaning of a legal text. Designating legal texts as unsponsored also illustrates the unpersuasiveness of the common intentionalist arguments that use examples from ordinary conversations (see Fish 2005 for one such example). An intentionalist account of interpretation that fails to address the fundamental differences between ordinary conversations and legal texts undermines any argument that the meaning of a text is identical to the author’s intended meaning and that interpretation consists of discovering that meaning.
22.3 The Nature of Language and Intentionalist Arguments About Meaning
22.3.1 ‘Fixed’ Meanings
Viewing legal texts as unsponsored and not analogous to ordinary conversations supports a theory of interpretation that would give a prominent role to the ordinary meaning concept and its reliance on conventions of meaning, but intentionalists make other arguments about meaning not yet considered that, if persuasive, would undermine reliance on conventions. Stanley Fish (2005, p. 633), for example, argues that the definition of ‘interpretation’ is limited to the search for authorial intent because “words do not carry fixed or even relatively fixed meanings”. Thus, for Fish (635), “lexical items and grammatical structures by themselves will yield no meaning—will not even be seen as lexical items and grammatical structures—until they are seen as having been produced by some intentional agent”. Fish’s position that words do not have even relatively fixed meanings would, if true, seriously undermine any reliance on conventions of meaning. If fact, if taken seriously, it would counsel that authors are the only parties qualified to interpret their own works.
Notwithstanding the ideality of language, assertions about the absence of inherent lexical meaning or the reality of diachronic fluctuations in word meanings do not further an intentionalist agenda. In a narrow sense, any linguist or philosopher of language should agree with certain intentionalist assertions about language. With respect to natural language, it is the users of language that determine whether signs will convey meaning and what meaning the signs will convey (Cruse 1983). That is, the marks on a page that any given community of speakers agrees to interpret as words do not contain any inherent meaning. As Jackson (1995, pp. 18–19) observes, neither the sound d-o-g nor the written characters ‘dog’ inherently refer to a member of the subspecies canis lupus familiaris (which is itself a conventional term). Rather, the reference is natural, and seemly inherent, to English speakers due to the strength of the convention that ‘dog’ refers to [dog]. Saussure (2011, p. 67) referred to this phenomenon as the “arbitrary nature of the sign”. That is, there is an arbitrary connection between a ‘sound-image’, or signifier, and a ‘concept’, or signified. Saussure explained that “if words stood for pre-existing concepts, they would all have exact equivalents in meaning from one language to the next; but this is not true” (2011, p. 116)For the most part, the connection between signifier and signified is not rational or natural but is instead conventional. Thus, it would be just as rational for d-o-g to refer to a member of the group of organisms that consist of all gill-bearing aquatic craniate animals that lack limbs with digits (i.e., ‘fish’) as it would for it to refer to a member of the subspecies canis lupus familiaris.
While words have the meanings that are assigned to them by the users of language, these meanings are independent of any particular authorial intention. Indeed, as has been observed by various scholars, various combinations of words can create meaningful sentences that have never been uttered at all. The nature of language easily allows for such possibilities. Verbs select the events and nouns select the entities that allow us to individuate happenings and events in the world. Identifying the meanings of verbs and nouns requires distinguishing their meanings from the happenings and entities in the world that they describe in a particular use. The lexicalized, or context-invariant, meaning, determined on the basis of grammatical behavioral patterns in the data, consists of those properties that are shared across all uses or a verb or noun, regardless of context. When a noun is chosen to describe an entity in the world, a claim is being made that the entity has the attributes lexicalized by the noun, though it may also have other attributes. Two nouns may refer to the same entity, but in lexicalizing different attributes, they may construe it as an entity in different ways. Similarly, a single verb can be applied to quite different happenings, and two verbs may lexicalize distinct but partially overlapping sets of attributes yet be able to refer to the same happening out in the world. Nevertheless when a verb is chosen to describe an event, a comment is made regarding certain attributes present in the chain of happenings in the world being referred to by that verb. For example, Botne (2001) has produced a crosslinguistic study which shows that verbs of dying fall into four major types that can be differentiated according to which stages of the dying process they lexicalize. The Acute type lexicalize the point of death alone, while the Transitional type lexicalize the onset, the point of death, and the result.
Speakers are thus constrained by the lexicalized meaning of verbs (and nouns). Consider a situation where the speaker wishes to convey that a table has been dried. In choosing an appropriate verb, the speaker must consider that there are verb pairs where each member lexicalizes distinct sets of attributes in the same stream of happenings (Levin and Hovav 1995). The speaker could describe the scenario using the verb ‘wipe’ or the verb ‘dry’.
(1) a. Kelly wiped the table with a tea towel.
b. Kelly dried the table with a tea towel.
‘Wipe’ is a manner (as opposed to result) verb that involves nonscalar changes, which means that it is not characterizable in terms of a scale, that is, an ordered set of degrees along a dimension representing a single attribute. The lack of an ordering relation and complexity are the two properties that contribute to making a change nonscalar (Hovav and Levin 2010). The vast majority of nonscalar change verbs, such as ‘wipe’, involve a complex combination of many changes at once, so that there is no single, privileged scale of change. If the speaker uses ‘wiped’, as in (1a), the speaker will have said nothing about whether the table has actually been dried, though perhaps most hearers would assume that it has been.3