Every theory of interpretation is based on semantic theory.1 The object of legal interpretation is a text, whether enacted into law or not, that is expressed through the medium of language. The text is communicative; it is designed to establish a legal norm to which people will conform their behavior. The goal of interpretation is to understand the language of the text. Because the limits of language set the limits of interpretation, we must understand the essence of language and the problems it raises. The difficulty is that there is no one theory for understanding language; each theory evaluates language from its own perspective.2 Subjective and objective theories diverge on the issue. Objective theories include extensional and intentional theories,3 each of which incorporates both ontological and epistemological approaches.
Language is a system of signs or symbols by which we think and through which we communicate. Natural language is the system of signs or symbols customarily used by members of a given society to communicate with each other. Life experience teaches—and every theory of law or semantics must take this into account—that people communicate. Semantic Towers of Babel are the exception. Language has meaning. While the words themselves have no intrinsic meaning, they have an accepted meaning among those who speak the same language.4 “Words, as mere symbols, are nothing but an empty shell with no meaning.”5
Language, however, is more than a collection of disembodied symbols; words have meaning in a given language. Saying “automobile” and saying “cigarette” are two different things, in the framework of a linguistic community. The word “automobile” and the word “cigarette” mean different things, even if that meaning is not intrinsic to the words themselves. We could call a four-legged creature that barks “cat.” But in a given language, at a given time, and in a given context, words have a given meaning. The meaning comes from the fact that, if we present the symbol (a four-legged creature that barks) and the word (dog), a speaker of English connects the two. When an English speaker hears or reads the word “dog,” he or she thinks of that same four-legged animal that barks.6 This is the power of a language—those who use it, share it, and it allows people to transmit meanings attached to the different symbols. Of course, anyone can express his or her thoughts through a private code. But in order to share those thoughts with someone else, that person will have to provide a key for deciphering the code.
The meaning that words have in a language is not static.7 Language changes. Relying on the works of various scholars, Zarfati lists five factors that change language: history, linguistics, social factors, psychological factors, and the influence of foreign languages.8 When we evaluate the communicative power of language, we take into account a given system of assumptions that are external to the language itself. The key question, of course, is which meaning of the language the interpreter should consult: the meaning it had on the day the text was written, or the meaning it has on the day it is interpreted? Language by itself cannot give an answer. Both meanings together establish the range of semantic possibilities, and the choice between them is guided by the criteria of the system of interpretation.9 Purposive interpretation gives language the legal meaning that best realizes the purpose of the norm. It may be the meaning of the language at the time the text was created, or the meaning at the time the text is interpreted, depending on the particular case at hand.10
Judge Hand’s remark that “Words are such temperamental beings”11 is as true today as it was when he first said it. Language is temperamental because the agreements people make with each other, based on language, are neither fixed nor precise. Natural language is not mathematics. Language has no single, uncontroversial meaning.12 Meaning changes with context. As Professor Thayer said at the end of the nineteenth century, we are not in
Lawyer’s Paradise where all words have a fixed, precisely ascertained meaning . . . and where, if the writer has been careful, a lawyer, having a document referred to him, may sit in his chair, inspect the text, and answer all questions without raising his eyes.13
Indeed, the best drafter cannot create a text that, when read in any circumstance, will impart the same single meaning to all readers. There is always “maneuvering room” in which understandings about the meaning of language are equivocal. This is especially true of general language referring to people and things not identified in the text. Language does facilitate communication among people, but its limitations prevent the communication from even approaching perfection. Language can mean more than one thing even in a single context; language can be ambiguous; language can be vague. A good drafter can try to limit the ambiguity if he or she chooses. Sometimes there is reason to create an ambiguous text.14 When a drafter does not want ambiguity, the best he or she may do is select words that, in particular contexts, give a single meaning for the primary and typical situations that the drafter wishes to address in the text.15
Language is ambiguous when a word or sentence has more than one meaning. The ambiguity may be a function of semantics or syntax. Semantic ambiguity occurs when a word in the sentence is ambiguous. A common example is the use of the words “and” and “or.” Usually, “and” connects, while “or” separates. But we sometimes use “and” to present alternatives, and “or” to denote combination.16
Syntactic ambiguity occurs when the meaning of a sentence is ambiguous because of the structure or order of its words. The phrase “institutions or corporations for public use” is ambiguous because it can mean either of two things: (1) “institutions for public use” or “corporations for public use”; or (2) “institutions” or “corporations for public use.”
Ambiguity is unavoidable and natural to natural language. Often, we can resolve the ambiguity by viewing the language in context. There are many potential contexts, however. A theory of interpretation must determine the relevant context. Purposive interpretive establishes the purpose of the norm grounded in the text as the relevant context, but that context does not always resolve the ambiguity. In those cases, judicial discretion is necessary.
Vagueness exists when there is uncertainty about the application of a word or sentence to details.17 The uncertainty may be of several kinds: First, the word or sentence may not include the necessary and sufficient conditions for its application. For example, a reference to a crime with moral turpitude is vague because “moral turpitude” is not self-defining.18 A second kind of vagueness occurs when it is unclear whether a particular condition is necessary for the provision to apply. For example, the word “intent” is vague, because it is not clear if it extends to a situation in which a person does not wish to achieve a particular result by his or her action, but he or she knows, with a reasonable degree of certainty, that his or her action will cause that result. A third kind of vagueness occurs when a word or clause applies to the kinds of things that do not have a clearly defined scope or natural end point.19 For example, words like “day,” “night,” “far,” “near,” “forest,” “thicket,” “mountain,” and “hill” are vague. They pose a boundary problem with no semantic solution.20 Sometimes, but not always, it is clear from the context whether a provision belongs to one or another end of the spectrum of possibilities. The context itself may be vague, however. In any case, a theory of interpretation is needed to decide which context is relevant. According to purposive interpretation, the relevant context is the purpose of the norm in question. Still, in exceptional circumstances, that criterion will not clarify the vagueness, and the interpreter must exercise discretion.
Context is likely to resolve a text’s ambiguity and vagueness.21 But what counts as “context” for this purpose? The literature distinguishes between intrinsic and extrinsic context.22 The intrinsic context is the textual context, that is, the other phrases surrounding the phrase being interpreted. It is the chapter that contains the clause being interpreted, or the entire document (constitution, statute, contract, will) containing the chapter. The extrinsic context is broader. It includes every context beyond the textual one. It includes the history of the text, the status of the law before and after the text was written, the general framework of the law in the given legal system, and the legal system’s social background and fundamental principles. As Justice Sussman said, referring to the text of a statute, “A word in a statute is a creature of its environment. Its character is a function of its context.”23
The key question regarding context (internal or extrinsic) is: What does the interpreter seek from the context? There are usually many different contexts. We should focus on the context that is relevant to interpretive activity. To find this context, we should evaluate the goal of the interpretation. The purpose of interpretation determines the relevance of the context. For example, if the goal of interpretation were to find the most aesthetically pleasing meaning,24 the relevant context would be different than if the goal of interpretation were to actualize the values at the core of the text being interpreted. Purposive interpretation is based on the view that the goal of legal interpretation is to realize the purpose of the text being interpreted. Purpose is the relevant context. The internal and extrinsic contexts are just means and devices through which we uncover the purpose at the core of the text. When we study context, we are not looking for beauty. We are looking for purpose. We need to know which of the different contexts the interpreter should use, and what that context means. Interpretive theory is necessary to provide an answer. According to purposive interpretation, the relevant context is that which provides information about the purpose at the core for the text. In choosing among the range of potential contexts, the interpreter in law should focus on the one that provides information about the purpose at the core of the legal text.
The nature of language is such that a legal text often has more than one semantic meaning. However, the number of semantic meanings is not infinite; people do successfully communicate. Language has a limited ability to bear meaning. It cannot sustain meaning beyond what is accepted in a given legal community or in the parties’ private code. A text is not a fortress to be conquered with a dictionary, but it does have walls, beyond whose confines it has no meaning. As Professor Radin noted,
Words are certainly not crystals, as Mr. Justice Holmes has wisely and properly warned us, but they are after all not portmanteaus. We can not quite put anything we like into them.25
The judge must give the language of the text a meaning that will not “tear” the “surface casing”26 of the words or sentences.27 Interpretation beyond the language of the text loses its interpretive character and must draw its legitimacy from non-interpretive doctrines.
A text in private law expresses the autonomy of the private will.28 The parties to a contract may formulate the contract as they wish. They may use a private code.29 They may decide that, in their contract, a horse is a donkey, and a donkey is a horse. Similarly, a testator may formulate his or her will in his or her own private code.30 He or she may call the wine cellar a “library,” and the bottles of wine, “books.” In that case, the range of semantic possibilities includes the private meaning as well as the accepted meaning in the given language. A judge applies the relevant interpretive criteria to choose between those meanings. According to purposive interpretation, the decisive factor is the testator’s intent (in a will) and the joint intent of the parties (in a contract), so that the meaning they intend will usually trump. Of course, the private code or language cannot remain entirely private. Those who use a private code must give the interpreter a key to understanding it. Once they do, the interpreter should use the key interpret the text according to the intent of the author (for a will) or authors (for a contract).
The process becomes more complicated in the context of a statute or constitution, which are public documents,31 for use by the public as a whole. A piece of legislation is not a linguistic or logical riddle. It is intended to send a message. It therefore must speak in a language that its target audience understands, that is, in the public language.32 A legislature that speaks in riddles or in language that is not clear misses the point and the essence of legislation. It tarnishes the rule of law by writing legal instructions that are not understandable to its target audience but nevertheless impose sanctions on members of that audience who violate those instructions. The legislature—and the legal system—have a basic obligation to formulate statutes in a language that is understandable to members of the given society.33 Having said that, and subject to constitutional restraints, interpreters must consider language in both its private and its public meaning, the combination of which determines the range of semantic possibilities. The choice between these meanings depends on the relevant interpretive criteria; purposive interpretation makes its selection based on the purpose at the core of the text.
The meaning of a text is explicit when it is transmitted to the reader through the dictionary meaning of the language, understood in its relevant (internal and external) textual context.34 The meaning of a text is implicit when it is transmitted to the reader but is not part of the dictionary meaning of the language.35 Implicit language is written into the text in invisible ink. It is written between the lines. Dickerson uses the following example36 to demonstrate implicit meaning: A son asks his mother, “May I go to the pool, and afterward, to the movies?” The mother replies, “You may go to the pool.” Implicit in her answer is that the son may not go to the movies (implied negation or “negative arrangement”). Her son arrives at this meaning not from the dictionary meaning of the language, but rather from the use of explicit language, in its extrinsic and internal context. Professor Cross offers another example:37 Richard tells Simon, “As a present, I’m giving you my apartment in London, my apartment in New York, and the furniture in my apartment in London.” Implicit in Richard’s words is that he does not give Simon the furniture in his New York apartment. These are obvious examples, but meaning can also be implied in less obvious cases. Consider a constitution whose various provisions establish the authority of the branches of government and protections of human rights. We can say that the constitution implicitly establishes the principles of separation of powers and judicial independence.38
The semantic meaning of a text (constitution, statute, contract, and will) is both the explicit and implicit meaning. The implicit meaning is also called the ulterior,39 ellipsis,40 or tacit41 meaning. It is part of the range of semantic meanings, just like the explicit meaning. An interpreter learns the implicit meaning through logic, intelligence, and reason. The question is not whether the implicit meaning is critical or necessary, clear or apparent; the question is whether the implicit meaning is reasonable, logical, and warranted.42
Readers are likely to disagree over the boundaries of “implication.” Do constitutional provisions for democratic and egalitarian elections imply recognition of a constitutional right to freedom of political expression? The Supreme Court of Australia said yes.43 Is that result implicit? Are we crossing the boundary between “implicit” meaning and a textual gap that the court fills? Whatever the definition, in principle, the text speaks to us explicitly and implicitly. In case of a conflict between the two meanings, the explicit meaning trumps. More precisely, when a text has an explicit meaning relating to a particular issue, there is no room to infer a contradictory implicit meaning relating to the same issue: expressum facit cessare tacitum.44 Indeed, like the explicit meaning, the implicit meaning determines the meaning that the text can bear. It does not determine the legal meaning of the text. An interpreter extracts the text’s legal meaning from the range of semantic meanings (explicit and implicit), using the purpose of the norm as the criterion for his or her selection.
“Implied meaning” can mean two things. On the one hand, it can be the meaning implicit in the text. On the other hand, implied meaning can fill in a gap in the text.45 There is a big difference between these two meanings. Implied meaning as implicit meaning does not add “what is” to “what is not.” Implied meaning (as implicit meaning) is part of the text: verba illata est inesse vindentur (implicit words are considered to be included). Referring to statutory language, Bennion said that
It is a fact of language, indeed a fact of life, that every statement consists not only of what is expressed but also of what is implied. Neither portion is more compelling than the other.46
Determining the implied meaning (as implicit meaning) of a text is part of interpreting it. On the other hand, an implied meaning (or term) exists when the text contains a gap. The court fills in the gap—adds “what is” to “what isn’t”—by adding implied language. According to my theory, this supplementation is part of interpretation in the broad sense, governed by non-interpretive doctrines.
The range of semantic possibilities includes the language’s totality of meanings. It includes the natural and ordinary meaning and the exceptional and special meaning. Of course, the presumption is that the purpose of a norm is expressed in its natural and ordinary language.47 This presumption stems from the fact that creating a legal text is a communicative activity. The presumption is, however, rebuttable. Sometimes, the interpreter should give the language of the text an exceptional and special meaning, if that is necessary to achieve the purpose at the core of the text in question. On the semantic level, exceptional and special meanings should be treated like natural and ordinary meanings; the difference is in the rebuttable presumption that the purpose of the norm will be achieved through the natural and ordinary language.
An interpreter may use a dictionary48 as a source of the meaning that a text can bear.49 A dictionary, either general or legal,50 can help the interpreter (as a linguist) determine the range of semantic possibilities. It does not determine the legal meaning of the text. As Hart and Sachs noted:
A dictionary, it is vital to observe, never says what meaning a word must bear in a particular context. Nor does it ever purport to say this. An unabridged dictionary is simply an historical record, not necessarily all-inclusive, of the meanings which words in fact have borne, in the judgment of the editors, in the writings of reputable authors.51
The dictionary determines what meaning may be given to the text; it does not determine what meaning must be given to the text. It is a linguistic tool, not a criterion for understanding the legal language of the text. The words of a statute are not fortresses to be conquered with dictionaries.52 The same is true of all types of legal texts.
Most legal systems have canons for understanding legal language.53 For example, it is an accepted canon that the meaning of a word or phrase depends on its environment: noscitur a sociis. Legal systems also commonly assume the “no” of the language from the totality of its “yes” (expressio unius est exclusio alterius). Similarly, if a legal text contains a string of details followed by a general phrase—like a statute about “trucks, buses, private cars, or other vehicles”—one can assume that the general phrase, “other vehicles,” includes additional items of the same type (vehicles that move on land, not ships and airplanes): ejusdem generis. What is the legal status of these canons? Some treat them as legal norms that interpreters should use to determine a text’s legal meaning. I disagree. The canons reflect semantic rules, not legal rules.54 They are not unique to law but rather are rules of general applicability for understanding language. They are based on experience, logic, and correct use of language. They are common to all legal systems and systems of interpretation. They belong to the field of language.55 They help determine the range of a legal text’s semantic possibilities. They determine the meaning that the text is capable of bearing.56 They do not determine the legal meaning that the text bears. As Hart and Sachs said:
Maxims should not be treated, any more than a dictionary, as saying what meaning a word or group of words must have in a given context. They simply answer the question whether a particular meaning is linguistically permissible.57
Canons are important because they determine the limits beyond which, in the absence of a private lexicon, interpretation cannot operate. They are tools in the larger project of determining the text’s legal meaning.
Some scholars object to linguistic canons on the grounds that they contradict each other, to the point where they become useless.58 Those scholars are correct to object to attempts to use the canons to determine the legal meaning of a text,59 but there is nothing wrong with using the canons to understand the semantic meaning of a text.60 The canons are a compilation of rules of grammar and language.61 Viewed as tools to help understand the meaning that a text can bear in its language, the canons are of some use.
The structure and logic62 of language allow us to infer the “no” of the text from its “yes”: expressio unius est exclusio alterius.63 This canon is an example of implicit semantic meaning, which may be negative or positive. When we can infer a negative meaning from the “yes” of the text, we can rule out the existence of a gap and recognize the existence of a negative arrangement. As Cross writes:
It is doubtful whether the maxim does any more than draw attention to a fairly obvious linguistic point, viz. that in many contexts the mention of some matters warrants an inference that other cognate matters were intentionally excluded.64
Of course, logic governs the scope of this canon. If I say that people are mortal, you should not infer that animals live forever. Even when it is possible to infer “no” from “yes,” that choice does not necessarily become the text’s legal meaning. Perhaps the “yes” itself is included merely as a precaution, to remove doubt (ex abundanti cautela);65 perhaps it is only as an example; perhaps it reflects sloppy writing.66
Another accepted canon says that if the text contains a string of specifics followed by a general, supplementary expression, one may understand the general expression to extend to additional specifics of the kind enumerated in the text. For example, if a contract provides that a renter is not liable for rent in case of fire, flood, storm, or other unexpected incident, the phrase “other unexpected incident” may not be understood to include a situation in which it was impossible to use the rental because of construction problems.67 This linguistic canon says that a list of specifics may restrict an otherwise general supplementary phrase. The canon employs logical deduction, based on common sense and life experience, to arrive at implicit meaning. It is not, however, the criterion for determining the text’s legal meaning, which depends on the purpose at the core of the norm.
1 See M. Landau, “Hamishpat B’mishpat: Balshanut Mishpatit—Habalshan B’sherut Hamishpatan [Linguistics and Law],” 22 Iyunei Mishpat 37 (1999); G.L. Williams, “Language and the Law,” 61 Law Q. Rev. 71, 179, 293, 384, (1945); 62 Law Q. Rev. 387 (1946);P. Goodrich, “The Role of Linguistics in Legal Analysis,” 47 Mod. L. Rev. 523 (1984); Goodrich, “Law and Language: An Historical and Critical Introduction,” 11 J. Law & Soc. 173 (1984). Williams discusses the connection between language and interpretation in the fourth article of this series. See also Law and Linguistics Conference, 73 Wash. U. L.Q. 785 (1995). The end of the issue contains an extensive bibliographic list (p. 1311).
2 F. Palmer, Semantics: A New Outline (1976); D. Taylor, Explanation and Meaning (1970).
3 The Theory of Meaning (G. Parkinson ed., 1968) makes this division.
4 A. Ross, On Law and Justice 112 (M. Dutton trans., 1959).
5 Sussman, “Miksat Mitaamei Parshanut [Interpretation],” Sefer Hayovel L’Pinchas Rosen 148 (1962).
6 Ogden and Richards, supra p. 3, note 3 at 11. For a critique of the plain exposition of this approach, see A. Dickerson, “Referentia Meaning: The Static Aspects,” 10 Jurimetrics J. 58 (1969–70).
7 See Justice Holmes’s comments in Towne v. Eisner, 245 U.S. 418, 425 (1918).
8 G. Zarfati, Hebrew Semantics 118 (2d. ed. 1968).
9 Bennion, supra p. 6, note 13 at 698.
10 Other interpretive methods, of course, reach different conclusions. Subjective interpretive methods focus on the meaning the language had at the time the text was created. Objective interpretive methods focus on the meaning of the language at the time of interpretation.
11 L. Hand, The Spirit of Liberty 157 (I. Dilliard ed., 3d ed. 1960). See also Z. Chafee, “The Disorderly Conduct of Words,” 41 Colum. L. Rev. 381 (1941).
12 B. Bix, Law, Language, and Legal Determinacy (1993). See also J. Steyn, “Written Contracts: To What Extent May Evidence Control Language?” 41 Current Legal Probs. 23 (1988); Slim v. Daily Telegraph Ltd. [1968] 2 Q.B. 157, 171.
13 J. Thayer, A Preliminary Treatise on Evidence at the Common Law 428 (1898).
14 On passing legislation that is deliberately ambiguous, see J. Grundfest and A.C. Pritchard, “With Multiple Personality Disorders: The Value of Ambiguity in Statutory Design and Interpretation,” 54 Stan. L. Rev. 627 (2002).
15 Seaford Court Estates v. Asher [1949] 2 All E.R. 155, 164. Friedmann attributes to me the view that there is no such thing as a plain text. D. Friedmann, “Liparshanut Hamunach ‘Parshanut’ Vihearot Lipsak Din Apropis [Notes on the Apropis Decision],” 7 Hamishpat 21 (2002). Indeed, I claim that there is no such thing as a text that is plain under every possible circumstance. However, I naturally think that a text can be plain in most circumstances that the author of the text sought to address. Of course, the determination that the text is plain can be done only after the text has been interpreted. See supra p. 12.
16 A. Dickerson, The Fundamentals of Legal Drafting 76 (1965).
17 J. Waldron, “Vagueness in Law and Language: Some Philosophical Issues,” 82 Cal. L. Rev. 509 (1994); G. Christie, “Vagueness and Legal Language,” 48 Minn. L. Rev. 885 (1964); Bix, supra p. 9, note 22.
18 W. Alston, Philosophy of Language 88 (1964).
19 Moore, supra p. 24, note 67 at 195. Moore calls this “degree-vagueness.”
20 Hobbs v. London & S.W. Ry. Co. (1875) L.R. 10 Q.B. 111, 121. See also Lavery v. Pursell (1888) 39 Ch. D. 508, 517; Mayor of Southport Corp. v. Morriss [1893] 1 Q.B. 359, 361; Attorney-General v. Brighton & Hove Co-operative Supply Ass’n [1900] 1 Ch. 276, 282.
21 For the difference between them, see 2 E.A. Farnsworth, Farnsworth on Contracts 239 (1990).
22 See Dickerson, supra p. 3, note 3 at 103.
23 H.C. 65/58, Shalit v. Interior Minister, 23(2) P.D. 477, 513. See also Attorney General v. Prince Ernest Augustus of Hanover [1957] A.C. 436, 461.
24 On aesthetics in law, see P. Schlag, “The Aesthetics of American Law,” 115 Harv. L. Rev. 1047 (2002).
25 Radin, supra p. 25, note 74 at 866.
26 C.A. 6339/97, supra p. 20, note 54 at 253.
27 D. Payne, “The Intention of the Legislature in the Interpretation of Statutes,” 9 Current Legal Probs. 96, 104 (1956); G. Williams, “The Meaning of Literal Interpretation,” 131 New L.J. 1128 (1981).
28 J.H. Wigmore, Evidence in Trials at Common Law 223 (J.H. Chadbourn ed., 4th rev. ed. 1981).
29 See Corbin, supra p. 9, note 20 at 155.
30 See Wigmore, supra p. 103, note 28 at 231. See also Re Rowland [1963] 1 ch. 1, 9; C.A. 1900/96, supra p. 61, note 1 at 828.
31 W.J. Brennan, “Construing the Constitution,” 19 U.C. Davis L. Rev. 2 (1985).
32 When the author of a text uses a private code, instead of public language, he must append an additional text that deciphers the code. Such a text is considered an interpretive clause in various statutes, or even an “Interpretive Act” accepted in common law countries.
33 See L. Fuller, The Morality of Law 63 (rev. ed. 1969).
34 Dickerson, supra p. 3, note 3 at 40.
35 Id. at 40.
36 Id. at 41.
37 Cross, supra p. 3, note 3 at 134.
38 Reference re Remuneration of Judges of the Provincial Court [1997] 3 S.C.R. 3.
39 G.L. Williams, “Language of the Law IV,” 61 Law Q. Rev. 384, 400 (1945).
40 See Bennion, supra p. 77, note 64 at 382. See also Bennion, supra p. 77, note 64 at 250.
41 Restatement (Second) of Contracts §97 (1981).
42 For a different approach, see Lord Westbury’s opinion in Parker v. Tootal (1865) 11 H.L. 143, 161.
43 See supra p. 70.
44 See Bennion, supra p. 77, note 64 at 383, 963.
45 Williams, supra p. 105, note 39 at 14.
46 Bennion, supra p. 77, note 64 at 382.
47 See infra p. 173.
48 See S. Thumma and J. Kirchmeier, “The Lexicon Has Become a Fortress: The United States Supreme Court’s Use of Dictionaries,” 47 Buff. L. Rev. 227 (1999).
49 See Lewison, supra p. 42, note 130 at 90.
50 See D. Mellinkoff, “The Myth of Precision and the Law Dictionary,” 31 UCLA L. Rev. 423 (1983).
51 Hart and Sachs, supra p. 3, note 3 at 1190.
52 Cabell v. Markham, 148 F. 2d 737, 739 (2d. Cir. 1945).
53 Infra p. 170. On the distinction between linguistic canons and textual canons, see W. Eskridge and P. Frickey, “Statutory Interpretation as Practical Reasoning,” 42 Stan. L. Rev. 321 (1990).
54 Sunstein, supra p. 13, note 31 at 454 (“[T]hey operate as rules of syntax or grammar”).
55 Bennion, supra p. 77, note 64 at 897. Justice M. Cheshin distinguishes between “weak” and “strong” rules of interpretation. The former derive from language, logic, and reason. The latter derive from the society’s fundamental principles. H.C. 5012/97 Matach Health, Support, and Welfare Services Ltd. v. Ministry of Health, 52(1) P.D. 49, 61.
56 Law Comm. (no. 21), The Interpretation of Statutes.
57 Hart and Sachs, supra p. 3, note 3 at 1191.
58 Llewellyn, supra p. 54, note 191 at 395.
59 See Frankfurter, supra p. 19, note 50 at 544.
60 W. Lattin, “Legal Maxims and Their Use in Statutory Interpretations,” 26 Geo. L.J. 1 (1937); G. Miller, “Pragmatics and the Maxims of Interpretation,” 1990 Wis. L. Rev. 1179, 1202.
61 The different canons can contradict each other. Resolving the conflict is done according to criteria external to the canons themselves. In purposive interpretation, it is done according to the purpose of the text.
62 Sutherland, supra p. 78, note 70 at 203.
63 E. Mureinik, “Expressio Unius: Exclusio Alterius,” 104 S.Afr. L.J. 264 (1987).
64 Cross, supra p. 3, note 3 at 140.
65 Bennion, supra p. 6, note 13 at 976.
66 See A. Samuels, “The Eiusdem Generis Rule in Statutory Interpretation,” [1984] Statute L. Rev. 180.
67 Saner v. Bilton (1878) 7 Ch. D. 815.